diff options
Diffstat (limited to 'ext')
186 files changed, 6498 insertions, 1778 deletions
diff --git a/ext/ftp/ftp.stub.php b/ext/ftp/ftp.stub.php index f4812655a4..698cfe215c 100644 --- a/ext/ftp/ftp.stub.php +++ b/ext/ftp/ftp.stub.php @@ -2,133 +2,57 @@ /** @generate-function-entries */ -/** @return resource|false */ -function ftp_connect(string $hostname, int $port = 21, int $timeout = 90) {} +function ftp_connect(string $hostname, int $port = 21, int $timeout = 90): FTPConnection|false {} #ifdef HAVE_FTP_SSL -/** @return resource|false */ -function ftp_ssl_connect(string $hostname, int $port = 21, int $timeout = 90) {} +function ftp_ssl_connect(string $hostname, int $port = 21, int $timeout = 90): FTPConnection|false {} #endif -/** @param resource $ftp */ -function ftp_login($ftp, string $username, string $password): bool {} - -/** @param resource $ftp */ -function ftp_pwd($ftp): string|false {} - -/** @param resource $ftp */ -function ftp_cdup($ftp): bool {} - -/** @param resource $ftp */ -function ftp_chdir($ftp, string $directory): bool {} - -/** @param resource $ftp */ -function ftp_exec($ftp, string $command): bool {} - -/** @param resource $ftp */ -function ftp_raw($ftp, string $command): array {} - -/** @param resource $ftp */ -function ftp_mkdir($ftp, string $directory): string|false {} - -/** @param resource $ftp */ -function ftp_rmdir($ftp, string $directory): bool {} - -/** @param resource $ftp */ -function ftp_chmod($ftp, int $permissions, string $filename): int|false {} - -/** - * @param resource $ftp - * @param string $response - */ -function ftp_alloc($ftp, int $size, &$response = null): bool {} - -/** @param resource $ftp */ -function ftp_nlist($ftp, string $directory): array|false {} - -/** @param resource $ftp */ -function ftp_rawlist($ftp, string $directory, bool $recursive = false): array|false {} - -/** @param resource $ftp */ -function ftp_mlsd($ftp, string $directory): array|false {} - -/** @param resource $ftp */ -function ftp_systype($ftp): string|false {} - -/** - * @param resource $ftp - * @param resource $stream - */ -function ftp_fget($ftp, $stream, string $remote_filename, int $mode = FTP_BINARY, int $offset = 0): bool {} - -/** - * @param resource $ftp - * @param resource $stream - */ -function ftp_nb_fget($ftp, $stream, string $remote_filename, int $mode = FTP_BINARY, int $offset = 0): int {} - -/** @param resource $ftp */ -function ftp_pasv($ftp, bool $enable): bool {} - -/** @param resource $ftp */ -function ftp_get($ftp, string $local_filename, string $remote_filename, int $mode = FTP_BINARY, int $offset = 0): bool {} - -/** @param resource $ftp */ -function ftp_nb_get($ftp, string $local_filename, string $remote_filename, int $mode = FTP_BINARY, int $offset = 0): int {} - -/** @param resource $ftp */ -function ftp_nb_continue($ftp): int {} - -/** - * @param resource $ftp - * @param resource $stream - */ -function ftp_fput($ftp, string $remote_filename, $stream, int $mode = FTP_BINARY, int $offset = 0): bool {} - -/** - * @param resource $ftp - * @param resource $stream - */ -function ftp_nb_fput($ftp, string $remote_filename, $stream, int $mode = FTP_BINARY, int $offset = 0): int {} - -/** @param resource $ftp */ -function ftp_put($ftp, string $remote_filename, string $local_filename, int $mode = FTP_BINARY, int $offset = 0): bool {} - -/** @param resource $ftp */ -function ftp_append($ftp, string $remote_filename, string $local_filename, int $mode = FTP_BINARY): bool {} - -/** @param resource $ftp */ -function ftp_nb_put($ftp, string $remote_filename, string $local_filename, int $mode = FTP_BINARY, int $offset = 0): int|false {} - -/** @param resource $ftp */ -function ftp_size($ftp, string $filename): int {} - -/** @param resource $ftp */ -function ftp_mdtm($ftp, string $filename): int {} - -/** @param resource $ftp */ -function ftp_rename($ftp, string $from, string $to): bool {} - -/** @param resource $ftp */ -function ftp_delete($ftp, string $filename): bool {} - -/** @param resource $ftp */ -function ftp_site($ftp, string $command): bool {} - -/** @param resource $ftp */ -function ftp_close($ftp): bool {} - -/** - * @param resource $ftp - * @alias ftp_close - */ -function ftp_quit($ftp): bool {} - -/** - * @param resource $ftp - * @param int|bool $value - */ -function ftp_set_option($ftp, int $option, $value): bool {} - -/** @param resource $ftp */ -function ftp_get_option($ftp, int $option): int|bool {} +function ftp_login(FTPConnection $ftp, string $username, string $password): bool {} +function ftp_pwd(FTPConnection $ftp): string|false {} +function ftp_cdup(FTPConnection $ftp): bool {} +function ftp_chdir(FTPConnection $ftp, string $directory): bool {} +function ftp_exec(FTPConnection $ftp, string $command): bool {} +function ftp_raw(FTPConnection $ftp, string $command): array {} +function ftp_mkdir(FTPConnection $ftp, string $directory): string|false {} +function ftp_rmdir(FTPConnection $ftp, string $directory): bool {} +function ftp_chmod(FTPConnection $ftp, int $permissions, string $filename): int|false {} + +/** @param string $response */ +function ftp_alloc(FTPConnection $ftp, int $size, &$response = null): bool {} +function ftp_nlist(FTPConnection $ftp, string $directory): array|false {} +function ftp_rawlist(FTPConnection $ftp, string $directory, bool $recursive = false): array|false {} +function ftp_mlsd(FTPConnection $ftp, string $directory): array|false {} +function ftp_systype(FTPConnection $ftp): string|false {} + +/** @param resource $stream */ +function ftp_fget(FTPConnection $ftp, $stream, string $remote_filename, int $mode = FTP_BINARY, int $offset = 0): bool {} + +/** @param resource $stream */ +function ftp_nb_fget(FTPConnection $ftp, $stream, string $remote_filename, int $mode = FTP_BINARY, int $offset = 0): int {} +function ftp_pasv(FTPConnection $ftp, bool $enable): bool {} +function ftp_get(FTPConnection $ftp, string $local_filename, string $remote_filename, int $mode = FTP_BINARY, int $offset = 0): bool {} +function ftp_nb_get(FTPConnection $ftp, string $local_filename, string $remote_filename, int $mode = FTP_BINARY, int $offset = 0): int {} +function ftp_nb_continue(FTPConnection $ftp): int {} + +/** @param resource $stream */ +function ftp_fput(FTPConnection $ftp, string $remote_filename, $stream, int $mode = FTP_BINARY, int $offset = 0): bool {} + +/** @param resource $stream */ +function ftp_nb_fput(FTPConnection $ftp, string $remote_filename, $stream, int $mode = FTP_BINARY, int $offset = 0): int {} +function ftp_put(FTPConnection $ftp, string $remote_filename, string $local_filename, int $mode = FTP_BINARY, int $offset = 0): bool {} +function ftp_append(FTPConnection $ftp, string $remote_filename, string $local_filename, int $mode = FTP_BINARY): bool {} +function ftp_nb_put(FTPConnection $ftp, string $remote_filename, string $local_filename, int $mode = FTP_BINARY, int $offset = 0): int|false {} +function ftp_size(FTPConnection $ftp, string $filename): int {} +function ftp_mdtm(FTPConnection $ftp, string $filename): int {} +function ftp_rename(FTPConnection $ftp, string $from, string $to): bool {} +function ftp_delete(FTPConnection $ftp, string $filename): bool {} +function ftp_site(FTPConnection $ftp, string $command): bool {} +function ftp_close(FTPConnection $ftp): bool {} + +/** @alias ftp_close */ +function ftp_quit(FTPConnection $ftp): bool {} + +/** @param int|bool $value */ +function ftp_set_option(FTPConnection $ftp, int $option, $value): bool {} +function ftp_get_option(FTPConnection $ftp, int $option): int|bool {} diff --git a/ext/ftp/ftp_arginfo.h b/ext/ftp/ftp_arginfo.h index e1a9d770bc..30251b87ee 100644 --- a/ext/ftp/ftp_arginfo.h +++ b/ext/ftp/ftp_arginfo.h @@ -1,14 +1,14 @@ /* This is a generated file, edit the .stub.php file instead. - * Stub hash: 4957601533b387d70ebb15811821fc06507cdbc2 */ + * Stub hash: 151310a13eeeb2bbf4df1ab38a3a29506b26a570 */ -ZEND_BEGIN_ARG_INFO_EX(arginfo_ftp_connect, 0, 0, 1) +ZEND_BEGIN_ARG_WITH_RETURN_OBJ_TYPE_MASK_EX(arginfo_ftp_connect, 0, 1, FTPConnection, MAY_BE_FALSE) ZEND_ARG_TYPE_INFO(0, hostname, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, port, IS_LONG, 0, "21") ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, timeout, IS_LONG, 0, "90") ZEND_END_ARG_INFO() #if defined(HAVE_FTP_SSL) -ZEND_BEGIN_ARG_INFO_EX(arginfo_ftp_ssl_connect, 0, 0, 1) +ZEND_BEGIN_ARG_WITH_RETURN_OBJ_TYPE_MASK_EX(arginfo_ftp_ssl_connect, 0, 1, FTPConnection, MAY_BE_FALSE) ZEND_ARG_TYPE_INFO(0, hostname, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, port, IS_LONG, 0, "21") ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, timeout, IS_LONG, 0, "90") @@ -16,60 +16,60 @@ ZEND_END_ARG_INFO() #endif ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_login, 0, 3, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, username, IS_STRING, 0) ZEND_ARG_TYPE_INFO(0, password, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_ftp_pwd, 0, 1, MAY_BE_STRING|MAY_BE_FALSE) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_cdup, 0, 1, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_chdir, 0, 2, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, directory, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_exec, 0, 2, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, command, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_raw, 0, 2, IS_ARRAY, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, command, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_ftp_mkdir, 0, 2, MAY_BE_STRING|MAY_BE_FALSE) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, directory, IS_STRING, 0) ZEND_END_ARG_INFO() #define arginfo_ftp_rmdir arginfo_ftp_chdir ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_ftp_chmod, 0, 3, MAY_BE_LONG|MAY_BE_FALSE) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, permissions, IS_LONG, 0) ZEND_ARG_TYPE_INFO(0, filename, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_alloc, 0, 2, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, size, IS_LONG, 0) ZEND_ARG_INFO_WITH_DEFAULT_VALUE(1, response, "null") ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_ftp_nlist, 0, 2, MAY_BE_ARRAY|MAY_BE_FALSE) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, directory, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_ftp_rawlist, 0, 2, MAY_BE_ARRAY|MAY_BE_FALSE) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, directory, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, recursive, _IS_BOOL, 0, "false") ZEND_END_ARG_INFO() @@ -79,7 +79,7 @@ ZEND_END_ARG_INFO() #define arginfo_ftp_systype arginfo_ftp_pwd ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_fget, 0, 3, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_INFO(0, stream) ZEND_ARG_TYPE_INFO(0, remote_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "FTP_BINARY") @@ -87,7 +87,7 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_fget, 0, 3, _IS_BOOL, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_nb_fget, 0, 3, IS_LONG, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_INFO(0, stream) ZEND_ARG_TYPE_INFO(0, remote_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "FTP_BINARY") @@ -95,12 +95,12 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_nb_fget, 0, 3, IS_LONG, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_pasv, 0, 2, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, enable, _IS_BOOL, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_get, 0, 3, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, local_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO(0, remote_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "FTP_BINARY") @@ -108,7 +108,7 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_get, 0, 3, _IS_BOOL, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_nb_get, 0, 3, IS_LONG, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, local_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO(0, remote_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "FTP_BINARY") @@ -116,11 +116,11 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_nb_get, 0, 3, IS_LONG, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_nb_continue, 0, 1, IS_LONG, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_fput, 0, 3, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, remote_filename, IS_STRING, 0) ZEND_ARG_INFO(0, stream) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "FTP_BINARY") @@ -128,7 +128,7 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_fput, 0, 3, _IS_BOOL, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_nb_fput, 0, 3, IS_LONG, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, remote_filename, IS_STRING, 0) ZEND_ARG_INFO(0, stream) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "FTP_BINARY") @@ -136,7 +136,7 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_nb_fput, 0, 3, IS_LONG, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_put, 0, 3, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, remote_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO(0, local_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "FTP_BINARY") @@ -144,14 +144,14 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_put, 0, 3, _IS_BOOL, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_append, 0, 3, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, remote_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO(0, local_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "FTP_BINARY") ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_ftp_nb_put, 0, 3, MAY_BE_LONG|MAY_BE_FALSE) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, remote_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO(0, local_filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "FTP_BINARY") @@ -159,20 +159,20 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_ftp_nb_put, 0, 3, MAY_BE_LONG|MA ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_size, 0, 2, IS_LONG, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, filename, IS_STRING, 0) ZEND_END_ARG_INFO() #define arginfo_ftp_mdtm arginfo_ftp_size ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_rename, 0, 3, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, from, IS_STRING, 0) ZEND_ARG_TYPE_INFO(0, to, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_delete, 0, 2, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, filename, IS_STRING, 0) ZEND_END_ARG_INFO() @@ -183,13 +183,13 @@ ZEND_END_ARG_INFO() #define arginfo_ftp_quit arginfo_ftp_cdup ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_ftp_set_option, 0, 3, _IS_BOOL, 0) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, option, IS_LONG, 0) ZEND_ARG_INFO(0, value) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_ftp_get_option, 0, 2, MAY_BE_LONG|MAY_BE_BOOL) - ZEND_ARG_INFO(0, ftp) + ZEND_ARG_OBJ_INFO(0, ftp, FTPConnection, 0) ZEND_ARG_TYPE_INFO(0, option, IS_LONG, 0) ZEND_END_ARG_INFO() diff --git a/ext/ftp/php_ftp.c b/ext/ftp/php_ftp.c index a6dca8c874..30bffbdd8c 100644 --- a/ext/ftp/php_ftp.c +++ b/ext/ftp/php_ftp.c @@ -29,13 +29,15 @@ #include "ext/standard/info.h" #include "ext/standard/file.h" +#include "Zend/zend_exceptions.h" +#include "Zend/zend_interfaces.h" #include "php_ftp.h" #include "ftp.h" #include "ftp_arginfo.h" -static int le_ftpbuf; -#define le_ftpbuf_name "FTP Buffer" +static zend_class_entry *php_ftp_ce = NULL; +static zend_object_handlers ftp_object_handlers; zend_module_entry php_ftp_module_entry = { STANDARD_MODULE_HEADER_EX, @@ -56,15 +58,42 @@ zend_module_entry php_ftp_module_entry = { ZEND_GET_MODULE(php_ftp) #endif -static void ftp_destructor_ftpbuf(zend_resource *rsrc) -{ - ftpbuf_t *ftp = (ftpbuf_t *)rsrc->ptr; +typedef struct _php_ftp_object { + ftpbuf_t *ftp; + zend_object std; +} php_ftp_object; + +static inline zend_object *ftp_object_to_zend_object(php_ftp_object *obj) { + return ((zend_object*)(obj + 1)) - 1; +} + +static inline php_ftp_object *ftp_object_from_zend_object(zend_object *zobj) { + return ((php_ftp_object*)(zobj + 1)) - 1; +} + +static zend_object* ftp_object_create(zend_class_entry* ce) { + php_ftp_object *obj = zend_object_alloc(sizeof(php_ftp_object), ce); + zend_object *zobj = ftp_object_to_zend_object(obj); + obj->ftp = NULL; + zend_object_std_init(zobj, ce); + object_properties_init(zobj, ce); + zobj->handlers = &ftp_object_handlers; - ftp_close(ftp); + return zobj; +} + +static void ftp_object_destroy(zend_object *zobj) { + php_ftp_object *obj = ftp_object_from_zend_object(zobj); + + if (obj->ftp) { + ftp_close(obj->ftp); + } } PHP_MINIT_FUNCTION(ftp) { + zend_class_entry ce; + #ifdef HAVE_FTP_SSL #if OPENSSL_VERSION_NUMBER < 0x10101000 && !defined(LIBRESSL_VERSION_NUMBER) SSL_library_init(); @@ -76,7 +105,18 @@ PHP_MINIT_FUNCTION(ftp) #endif #endif - le_ftpbuf = zend_register_list_destructors_ex(ftp_destructor_ftpbuf, NULL, le_ftpbuf_name, module_number); + INIT_CLASS_ENTRY(ce, "FTPConnection", NULL); + php_ftp_ce = zend_register_internal_class(&ce); + php_ftp_ce->ce_flags |= ZEND_ACC_FINAL | ZEND_ACC_NO_DYNAMIC_PROPERTIES; + php_ftp_ce->create_object = ftp_object_create; + php_ftp_ce->serialize = zend_class_serialize_deny; + php_ftp_ce->unserialize = zend_class_unserialize_deny; + + memcpy(&ftp_object_handlers, &std_object_handlers, sizeof(zend_object_handlers)); + ftp_object_handlers.offset = XtOffsetOf(php_ftp_object, std); + ftp_object_handlers.dtor_obj = ftp_object_destroy; + ftp_object_handlers.clone_obj = NULL; + REGISTER_LONG_CONSTANT("FTP_ASCII", FTPTYPE_ASCII, CONST_PERSISTENT | CONST_CS); REGISTER_LONG_CONSTANT("FTP_TEXT", FTPTYPE_ASCII, CONST_PERSISTENT | CONST_CS); REGISTER_LONG_CONSTANT("FTP_BINARY", FTPTYPE_IMAGE, CONST_PERSISTENT | CONST_CS); @@ -143,7 +183,8 @@ PHP_FUNCTION(ftp_connect) ftp->use_ssl = 0; #endif - RETURN_RES(zend_register_resource(ftp, le_ftpbuf)); + object_init_ex(return_value, php_ftp_ce); + ftp_object_from_zend_object(Z_OBJ_P(return_value))->ftp = ftp; } /* }}} */ @@ -177,11 +218,19 @@ PHP_FUNCTION(ftp_ssl_connect) /* enable ssl */ ftp->use_ssl = 1; - RETURN_RES(zend_register_resource(ftp, le_ftpbuf)); + object_init_ex(return_value, php_ftp_ce); + ftp_object_from_zend_object(Z_OBJ_P(return_value))->ftp = ftp; } /* }}} */ #endif +#define GET_FTPBUF(ftpbuf, zftp) \ + ftpbuf = ftp_object_from_zend_object(Z_OBJ_P(zftp))->ftp; \ + if (!ftpbuf) { \ + zend_throw_exception(zend_ce_value_error, "FTPConnection is already closed", 0); \ + RETURN_THROWS(); \ + } + /* {{{ Logs into the FTP server */ PHP_FUNCTION(ftp_login) { @@ -190,13 +239,10 @@ PHP_FUNCTION(ftp_login) char *user, *pass; size_t user_len, pass_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rss", &z_ftp, &user, &user_len, &pass, &pass_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Oss", &z_ftp, php_ftp_ce, &user, &user_len, &pass, &pass_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* log in */ if (!ftp_login(ftp, user, user_len, pass, pass_len)) { @@ -215,13 +261,10 @@ PHP_FUNCTION(ftp_pwd) ftpbuf_t *ftp; const char *pwd; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "r", &z_ftp) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "O", &z_ftp, php_ftp_ce) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); if (!(pwd = ftp_pwd(ftp))) { php_error_docref(NULL, E_WARNING, "%s", ftp->inbuf); @@ -238,13 +281,10 @@ PHP_FUNCTION(ftp_cdup) zval *z_ftp; ftpbuf_t *ftp; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "r", &z_ftp) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "O", &z_ftp, php_ftp_ce) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); if (!ftp_cdup(ftp)) { php_error_docref(NULL, E_WARNING, "%s", ftp->inbuf); @@ -263,13 +303,10 @@ PHP_FUNCTION(ftp_chdir) char *dir; size_t dir_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rs", &z_ftp, &dir, &dir_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Os", &z_ftp, php_ftp_ce, &dir, &dir_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* change directories */ if (!ftp_chdir(ftp, dir, dir_len)) { @@ -289,13 +326,10 @@ PHP_FUNCTION(ftp_exec) char *cmd; size_t cmd_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rs", &z_ftp, &cmd, &cmd_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Os", &z_ftp, php_ftp_ce, &cmd, &cmd_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* execute serverside command */ if (!ftp_exec(ftp, cmd, cmd_len)) { @@ -315,13 +349,10 @@ PHP_FUNCTION(ftp_raw) char *cmd; size_t cmd_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rs", &z_ftp, &cmd, &cmd_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Os", &z_ftp, php_ftp_ce, &cmd, &cmd_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* execute arbitrary ftp command */ ftp_raw(ftp, cmd, cmd_len, return_value); @@ -337,13 +368,10 @@ PHP_FUNCTION(ftp_mkdir) zend_string *tmp; size_t dir_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rs", &z_ftp, &dir, &dir_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Os", &z_ftp, php_ftp_ce, &dir, &dir_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* create directory */ if (NULL == (tmp = ftp_mkdir(ftp, dir, dir_len))) { @@ -363,13 +391,10 @@ PHP_FUNCTION(ftp_rmdir) char *dir; size_t dir_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rs", &z_ftp, &dir, &dir_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Os", &z_ftp, php_ftp_ce, &dir, &dir_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* remove directorie */ if (!ftp_rmdir(ftp, dir, dir_len)) { @@ -390,13 +415,10 @@ PHP_FUNCTION(ftp_chmod) size_t filename_len; zend_long mode; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rlp", &z_ftp, &mode, &filename, &filename_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Olp", &z_ftp, php_ftp_ce, &mode, &filename, &filename_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); if (!ftp_chmod(ftp, mode, filename, filename_len)) { php_error_docref(NULL, E_WARNING, "%s", ftp->inbuf); @@ -415,13 +437,10 @@ PHP_FUNCTION(ftp_alloc) zend_long size, ret; zend_string *response = NULL; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rl|z", &z_ftp, &size, &zresponse) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Ol|z", &z_ftp, php_ftp_ce, &size, &zresponse) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); ret = ftp_alloc(ftp, size, zresponse ? &response : NULL); @@ -445,13 +464,10 @@ PHP_FUNCTION(ftp_nlist) char **nlist, **ptr, *dir; size_t dir_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rp", &z_ftp, &dir, &dir_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Op", &z_ftp, php_ftp_ce, &dir, &dir_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* get list of files */ if (NULL == (nlist = ftp_nlist(ftp, dir, dir_len))) { @@ -475,13 +491,10 @@ PHP_FUNCTION(ftp_rawlist) size_t dir_len; zend_bool recursive = 0; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rs|b", &z_ftp, &dir, &dir_len, &recursive) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Os|b", &z_ftp, php_ftp_ce, &dir, &dir_len, &recursive) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* get raw directory listing */ if (NULL == (llist = ftp_list(ftp, dir, dir_len, recursive))) { @@ -505,13 +518,10 @@ PHP_FUNCTION(ftp_mlsd) size_t dir_len; zval entry; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rs", &z_ftp, &dir, &dir_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Os", &z_ftp, php_ftp_ce, &dir, &dir_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* get raw directory listing */ if (NULL == (llist = ftp_mlsd(ftp, dir, dir_len))) { @@ -539,13 +549,10 @@ PHP_FUNCTION(ftp_systype) ftpbuf_t *ftp; const char *syst; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "r", &z_ftp) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "O", &z_ftp, php_ftp_ce) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); if (NULL == (syst = ftp_syst(ftp))) { php_error_docref(NULL, E_WARNING, "%s", ftp->inbuf); @@ -567,13 +574,10 @@ PHP_FUNCTION(ftp_fget) size_t file_len; zend_long mode=FTPTYPE_IMAGE, resumepos=0; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rrs|ll", &z_ftp, &z_file, &file, &file_len, &mode, &resumepos) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Ors|ll", &z_ftp, php_ftp_ce, &z_file, &file, &file_len, &mode, &resumepos) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); php_stream_from_res(stream, Z_RES_P(z_file)); XTYPE(xtype, mode); @@ -612,13 +616,10 @@ PHP_FUNCTION(ftp_nb_fget) size_t file_len; zend_long mode=FTPTYPE_IMAGE, resumepos=0, ret; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rrs|ll", &z_ftp, &z_file, &file, &file_len, &mode, &resumepos) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Ors|ll", &z_ftp, php_ftp_ce, &z_file, &file, &file_len, &mode, &resumepos) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); php_stream_from_res(stream, Z_RES_P(z_file)); XTYPE(xtype, mode); @@ -657,13 +658,10 @@ PHP_FUNCTION(ftp_pasv) ftpbuf_t *ftp; zend_bool pasv; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rb", &z_ftp, &pasv) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Ob", &z_ftp, php_ftp_ce, &pasv) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); if (!ftp_pasv(ftp, pasv ? 1 : 0)) { RETURN_FALSE; @@ -684,13 +682,10 @@ PHP_FUNCTION(ftp_get) size_t local_len, remote_len; zend_long mode=FTPTYPE_IMAGE, resumepos=0; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rpp|ll", &z_ftp, &local, &local_len, &remote, &remote_len, &mode, &resumepos) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Opp|ll", &z_ftp, php_ftp_ce, &local, &local_len, &remote, &remote_len, &mode, &resumepos) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); XTYPE(xtype, mode); /* ignore autoresume if autoseek is switched off */ @@ -749,13 +744,10 @@ PHP_FUNCTION(ftp_nb_get) int ret; zend_long mode=FTPTYPE_IMAGE, resumepos=0; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rss|ll", &z_ftp, &local, &local_len, &remote, &remote_len, &mode, &resumepos) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Oss|ll", &z_ftp, php_ftp_ce, &local, &local_len, &remote, &remote_len, &mode, &resumepos) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); XTYPE(xtype, mode); /* ignore autoresume if autoseek is switched off */ @@ -816,13 +808,10 @@ PHP_FUNCTION(ftp_nb_continue) ftpbuf_t *ftp; zend_long ret; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "r", &z_ftp) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "O", &z_ftp, php_ftp_ce) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); if (!ftp->nb) { php_error_docref(NULL, E_WARNING, "No nbronous transfer to continue"); @@ -859,13 +848,10 @@ PHP_FUNCTION(ftp_fput) php_stream *stream; char *remote; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rsr|ll", &z_ftp, &remote, &remote_len, &z_file, &mode, &startpos) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Osr|ll", &z_ftp, php_ftp_ce, &remote, &remote_len, &z_file, &mode, &startpos) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); php_stream_from_zval(stream, z_file); XTYPE(xtype, mode); @@ -908,13 +894,10 @@ PHP_FUNCTION(ftp_nb_fput) php_stream *stream; char *remote; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rsr|ll", &z_ftp, &remote, &remote_len, &z_file, &mode, &startpos) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Osr|ll", &z_ftp, php_ftp_ce, &remote, &remote_len, &z_file, &mode, &startpos) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); php_stream_from_res(stream, Z_RES_P(z_file)); XTYPE(xtype, mode); @@ -961,13 +944,10 @@ PHP_FUNCTION(ftp_put) zend_long mode=FTPTYPE_IMAGE, startpos=0; php_stream *instream; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rpp|ll", &z_ftp, &remote, &remote_len, &local, &local_len, &mode, &startpos) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Opp|ll", &z_ftp, php_ftp_ce, &remote, &remote_len, &local, &local_len, &mode, &startpos) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); XTYPE(xtype, mode); if (!(instream = php_stream_open_wrapper(local, mode == FTPTYPE_ASCII ? "rt" : "rb", REPORT_ERRORS, NULL))) { @@ -1014,13 +994,10 @@ PHP_FUNCTION(ftp_append) zend_long mode=FTPTYPE_IMAGE; php_stream *instream; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rpp|l", &z_ftp, &remote, &remote_len, &local, &local_len, &mode) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Opp|l", &z_ftp, php_ftp_ce, &remote, &remote_len, &local, &local_len, &mode) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); XTYPE(xtype, mode); if (!(instream = php_stream_open_wrapper(local, mode == FTPTYPE_ASCII ? "rt" : "rb", REPORT_ERRORS, NULL))) { @@ -1049,13 +1026,10 @@ PHP_FUNCTION(ftp_nb_put) zend_long mode=FTPTYPE_IMAGE, startpos=0, ret; php_stream *instream; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rpp|ll", &z_ftp, &remote, &remote_len, &local, &local_len, &mode, &startpos) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Opp|ll", &z_ftp, php_ftp_ce, &remote, &remote_len, &local, &local_len, &mode, &startpos) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); XTYPE(xtype, mode); if (!(instream = php_stream_open_wrapper(local, mode == FTPTYPE_ASCII ? "rt" : "rb", REPORT_ERRORS, NULL))) { @@ -1107,13 +1081,10 @@ PHP_FUNCTION(ftp_size) char *file; size_t file_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rp", &z_ftp, &file, &file_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Op", &z_ftp, php_ftp_ce, &file, &file_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* get file size */ RETURN_LONG(ftp_size(ftp, file, file_len)); @@ -1128,13 +1099,10 @@ PHP_FUNCTION(ftp_mdtm) char *file; size_t file_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rp", &z_ftp, &file, &file_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Op", &z_ftp, php_ftp_ce, &file, &file_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* get file mod time */ RETURN_LONG(ftp_mdtm(ftp, file, file_len)); @@ -1149,13 +1117,10 @@ PHP_FUNCTION(ftp_rename) char *src, *dest; size_t src_len, dest_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rss", &z_ftp, &src, &src_len, &dest, &dest_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Oss", &z_ftp, php_ftp_ce, &src, &src_len, &dest, &dest_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* rename the file */ if (!ftp_rename(ftp, src, src_len, dest, dest_len)) { @@ -1175,13 +1140,10 @@ PHP_FUNCTION(ftp_delete) char *file; size_t file_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rs", &z_ftp, &file, &file_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Os", &z_ftp, php_ftp_ce, &file, &file_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* delete the file */ if (!ftp_delete(ftp, file, file_len)) { @@ -1201,13 +1163,10 @@ PHP_FUNCTION(ftp_site) char *cmd; size_t cmd_len; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rs", &z_ftp, &cmd, &cmd_len) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Os", &z_ftp, php_ftp_ce, &cmd, &cmd_len) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); /* send the site command */ if (!ftp_site(ftp, cmd, cmd_len)) { @@ -1223,18 +1182,19 @@ PHP_FUNCTION(ftp_site) PHP_FUNCTION(ftp_close) { zval *z_ftp; - ftpbuf_t *ftp; + php_ftp_object *obj; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "r", &z_ftp) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "O", &z_ftp, php_ftp_ce) == FAILURE) { RETURN_THROWS(); } - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { - RETURN_THROWS(); + obj = ftp_object_from_zend_object(Z_OBJ_P(z_ftp)); + if (obj->ftp) { + ftp_quit(obj->ftp); + ftp_close(obj->ftp); + obj->ftp = NULL; } - ftp_quit(ftp); - RETURN_TRUE; } /* }}} */ @@ -1246,13 +1206,10 @@ PHP_FUNCTION(ftp_set_option) zend_long option; ftpbuf_t *ftp; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rlz", &z_ftp, &option, &z_value) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Olz", &z_ftp, php_ftp_ce, &option, &z_value) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); switch (option) { case PHP_FTP_OPT_TIMEOUT_SEC: @@ -1298,13 +1255,10 @@ PHP_FUNCTION(ftp_get_option) zend_long option; ftpbuf_t *ftp; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "rl", &z_ftp, &option) == FAILURE) { - RETURN_THROWS(); - } - - if ((ftp = (ftpbuf_t *)zend_fetch_resource(Z_RES_P(z_ftp), le_ftpbuf_name, le_ftpbuf)) == NULL) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Ol", &z_ftp, php_ftp_ce, &option) == FAILURE) { RETURN_THROWS(); } + GET_FTPBUF(ftp, z_ftp); switch (option) { case PHP_FTP_OPT_TIMEOUT_SEC: diff --git a/ext/ftp/tests/007.phpt b/ext/ftp/tests/007.phpt index 0520b37293..1ae879f901 100644 --- a/ext/ftp/tests/007.phpt +++ b/ext/ftp/tests/007.phpt @@ -180,36 +180,36 @@ try { fclose($ftp); ?> --EXPECT-- -ftp_login(): supplied resource is not a valid FTP Buffer resource -ftp_pwd(): supplied resource is not a valid FTP Buffer resource -ftp_cdup(): supplied resource is not a valid FTP Buffer resource -ftp_chdir(): supplied resource is not a valid FTP Buffer resource -ftp_exec(): supplied resource is not a valid FTP Buffer resource -ftp_raw(): supplied resource is not a valid FTP Buffer resource -ftp_mkdir(): supplied resource is not a valid FTP Buffer resource -ftp_rmdir(): supplied resource is not a valid FTP Buffer resource -ftp_chmod(): supplied resource is not a valid FTP Buffer resource -ftp_alloc(): supplied resource is not a valid FTP Buffer resource -ftp_nlist(): supplied resource is not a valid FTP Buffer resource -ftp_rawlist(): supplied resource is not a valid FTP Buffer resource -ftp_mlsd(): supplied resource is not a valid FTP Buffer resource -ftp_systype(): supplied resource is not a valid FTP Buffer resource -ftp_fget(): supplied resource is not a valid FTP Buffer resource -ftp_nb_fget(): supplied resource is not a valid FTP Buffer resource -ftp_pasv(): supplied resource is not a valid FTP Buffer resource -ftp_get(): supplied resource is not a valid FTP Buffer resource -ftp_nb_get(): supplied resource is not a valid FTP Buffer resource -ftp_nb_continue(): supplied resource is not a valid FTP Buffer resource -ftp_fput(): supplied resource is not a valid FTP Buffer resource -ftp_nb_fput(): supplied resource is not a valid FTP Buffer resource -ftp_put(): supplied resource is not a valid FTP Buffer resource -ftp_append(): supplied resource is not a valid FTP Buffer resource -ftp_nb_put(): supplied resource is not a valid FTP Buffer resource -ftp_size(): supplied resource is not a valid FTP Buffer resource -ftp_mdtm(): supplied resource is not a valid FTP Buffer resource -ftp_rename(): supplied resource is not a valid FTP Buffer resource -ftp_delete(): supplied resource is not a valid FTP Buffer resource -ftp_site(): supplied resource is not a valid FTP Buffer resource -ftp_close(): supplied resource is not a valid FTP Buffer resource -ftp_set_option(): supplied resource is not a valid FTP Buffer resource -ftp_get_option(): supplied resource is not a valid FTP Buffer resource +ftp_login(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_pwd(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_cdup(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_chdir(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_exec(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_raw(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_mkdir(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_rmdir(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_chmod(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_alloc(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_nlist(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_rawlist(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_mlsd(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_systype(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_fget(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_nb_fget(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_pasv(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_get(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_nb_get(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_nb_continue(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_fput(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_nb_fput(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_put(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_append(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_nb_put(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_size(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_mdtm(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_rename(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_delete(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_site(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_close(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_set_option(): Argument #1 ($ftp) must be of type FTPConnection, resource given +ftp_get_option(): Argument #1 ($ftp) must be of type FTPConnection, resource given diff --git a/ext/ftp/tests/CONFLICTS b/ext/ftp/tests/CONFLICTS deleted file mode 100644 index 254defddb5..0000000000 --- a/ext/ftp/tests/CONFLICTS +++ /dev/null @@ -1 +0,0 @@ -server diff --git a/ext/ftp/tests/dead-resource.phpt b/ext/ftp/tests/dead-resource.phpt new file mode 100644 index 0000000000..75a9931d54 --- /dev/null +++ b/ext/ftp/tests/dead-resource.phpt @@ -0,0 +1,24 @@ +--TEST-- +Attempt to use a closed FTPConnection +--SKIPIF-- +<?php +require 'skipif.inc'; +--FILE-- +<?php +require 'server.inc'; + +$ftp = ftp_connect('127.0.0.1', $port); +if (!$ftp) die("Couldn't connect to the server"); +var_dump(ftp_login($ftp, 'user', 'pass')); +var_dump(ftp_close($ftp)); + +try { + var_dump(ftp_login($ftp, 'user', 'pass')); + echo "Login did not throw\n"; +} catch (ValueError $ex) { + echo "Exception: ", $ex->getMessage(), "\n"; +} +--EXPECT-- +bool(true) +bool(true) +Exception: FTPConnection is already closed diff --git a/ext/ftp/tests/ftp_set_option_errors.phpt b/ext/ftp/tests/ftp_set_option_errors.phpt index 3f5476f6c3..5e5ef328f3 100644 --- a/ext/ftp/tests/ftp_set_option_errors.phpt +++ b/ext/ftp/tests/ftp_set_option_errors.phpt @@ -13,40 +13,22 @@ $ftp = ftp_connect('127.0.0.1', $port); ftp_login($ftp, 'user', 'pass'); $ftp or die("Couldn't connect to the server"); -try { - ftp_set_option($ftp, FTP_TIMEOUT_SEC, 0); -} catch (ValueError $exception) { - echo $exception->getMessage() . "\n"; +$options = [ + [ 'option' => FTP_TIMEOUT_SEC, 'value' => 0 ], + [ 'option' => FTP_TIMEOUT_SEC, 'value' => '0' ], + [ 'option' => FTP_USEPASVADDRESS, 'value' => ['1'] ], + [ 'option' => FTP_AUTOSEEK, 'value' => 'true' ], + [ 'option' => FOO_BAR, 'value' => 1 ], +]; +foreach ($options as $option) try { + var_dump(ftp_set_option($ftp, $option['option'], $option['value'])); +} catch (\Throwable $ex) { + echo "Exception: ", $ex->getMessage(), "\n"; } -try { - ftp_set_option($ftp, FTP_TIMEOUT_SEC, '0'); -} catch (TypeError $exception) { - echo $exception->getMessage() . "\n"; -} - -try { - ftp_set_option($ftp, FTP_USEPASVADDRESS, ['1']); -} catch (TypeError $exception) { - echo $exception->getMessage() . "\n"; -} - -try { - ftp_set_option($ftp, FTP_AUTOSEEK, 'true'); -} catch (TypeError $exception) { - echo $exception->getMessage() . "\n"; -} - -try { - ftp_set_option($ftp, FOO_BAR, 1); -} catch (ValueError $exception) { - echo $exception->getMessage() . "\n"; -} - -?> --EXPECT-- -ftp_set_option(): Argument #3 ($value) must be greater than 0 for the FTP_TIMEOUT_SEC option -ftp_set_option(): Argument #3 ($value) must be of type int for the FTP_TIMEOUT_SEC option, string given -ftp_set_option(): Argument #3 ($value) must be of type bool for the FTP_USEPASVADDRESS option, array given -ftp_set_option(): Argument #3 ($value) must be of type bool for the FTP_AUTOSEEK option, string given -ftp_set_option(): Argument #2 ($option) must be one of FTP_TIMEOUT_SEC, FTP_AUTOSEEK, or FTP_USEPASVADDRESS +Exception: ftp_set_option(): Argument #3 ($value) must be greater than 0 for the FTP_TIMEOUT_SEC option +Exception: ftp_set_option(): Argument #3 ($value) must be of type int for the FTP_TIMEOUT_SEC option, string given +Exception: ftp_set_option(): Argument #3 ($value) must be of type bool for the FTP_USEPASVADDRESS option, array given +Exception: ftp_set_option(): Argument #3 ($value) must be of type bool for the FTP_AUTOSEEK option, string given +Exception: ftp_set_option(): Argument #2 ($option) must be one of FTP_TIMEOUT_SEC, FTP_AUTOSEEK, or FTP_USEPASVADDRESS diff --git a/ext/ftp/tests/server.inc b/ext/ftp/tests/server.inc index 38599391d6..8bc836fc40 100644 --- a/ext/ftp/tests/server.inc +++ b/ext/ftp/tests/server.inc @@ -4,59 +4,21 @@ $socket = null; $errno = 0; $context = stream_context_create(array('ssl' => array('local_cert' => dirname(__FILE__).'/cert.pem'))); -for ($i=0; $i<10 && !$socket; ++$i) { - $port = rand(50000, 65535); - - $socket = @stream_socket_server("tcp://127.0.0.1:$port", $errno, $errstr, STREAM_SERVER_BIND|STREAM_SERVER_LISTEN, $context); -} -//set anther random port that is not the same as $port -do{ - $pasv_port = rand(50000, 65535); -}while($pasv_port == $port); - +$socket = stream_socket_server("tcp://127.0.0.1:0", $errno, $errstr, STREAM_SERVER_BIND|STREAM_SERVER_LISTEN, $context); if (!$socket) { echo "$errstr ($errno)\n"; die("could not start/bind the ftp server\n"); } -$pid = pcntl_fork(); +$socket_name = stream_socket_get_name($socket, false); +$port = (int) substr($socket_name, strrpos($socket_name, ':') + 1); -function pasv_listen($action){ - global $pasv_port, $tmp_file; - $tmp_file = 'nm2.php'; - $pid = pcntl_fork(); - if($pid === 0){ - $soc = stream_socket_server("tcp://127.0.0.1:$pasv_port"); - $fs = stream_socket_accept($soc, 3); - switch ($action) { - case 'fget': - case 'get': - //listen for 3 seconds 3 seconds - fputs($fs, "I am passive.\r\n"); - break; - case 'put': - file_put_contents($tmp_file, stream_get_contents($fs)); - break; - case 'list': - fputs($fs, "drwxr-x--- 3 owner group 4096 Jul 12 12:16 .\r\n"); - fputs($fs, "drwxr-x--- 3 owner group 4096 Jul 12 12:16 ..\r\n"); - fputs($fs, "drwxr-x--- 3 owner group 4096 Jul 12 12:16 public_ftp\r\n"); - break; - case 'list_null': - fputs($fs, "\r\n"); - break; - } - fclose($fs); - exit; - } -} +$pid = pcntl_fork(); if ($pid) { - function dump_and_exit($buf) { var_dump($buf); - fclose($GLOBALS['s']); exit; } @@ -90,10 +52,14 @@ if ($pid) { if (!$s) die("Error accepting a new connection\n"); + register_shutdown_function(function() use($pid, $s) { + fclose($s); + pcntl_waitpid($pid, $status); + }); + fputs($s, "220----- PHP FTP server 0.3 -----\r\n220 Service ready\r\n"); $buf = fread($s, 2048); - function user_auth($buf) { global $user, $s, $ssl, $bug37799; @@ -143,7 +109,6 @@ if ($pid) { $user = $m[1]; if ($user !== 'user' && $user !== 'anonymous') { fputs($s, "530 Not logged in.\r\n"); - fclose($s); exit; } @@ -163,7 +128,6 @@ if ($pid) { fputs($s, "230 User logged in\r\n"); } else { fputs($s, "530 Not logged in.\r\n"); - fclose($s); exit; } } @@ -432,25 +396,20 @@ if ($pid) { $i=0; if (empty($bug73457)) { - do { - if (!empty($ssl)) { - $soc = @stream_socket_server("tcp://127.0.0.1:$pasv_port", $errno, $errstr, STREAM_SERVER_BIND|STREAM_SERVER_LISTEN, $context); - } else { - $soc = @stream_socket_server("tcp://127.0.0.1:$pasv_port"); - } - /* Could bind port, Try another port */ - if (!$soc) { - $pasv_port = rand(50000, 65535); - } - $i++; - } while ($i<10 && !$soc); - + if (!empty($ssl)) { + $soc = stream_socket_server("tcp://127.0.0.1:0", $errno, $errstr, STREAM_SERVER_BIND|STREAM_SERVER_LISTEN, $context); + } else { + $soc = stream_socket_server("tcp://127.0.0.1:0"); + } if (!$soc) { echo "$errstr ($errno)\n"; die("could not bind passive port\n"); } + + $soc_name = stream_socket_get_name($soc, false); + $pasv_port = (int) substr($soc_name, strrpos($soc_name, ':') + 1); } else { - $pasv_port=1234; + $pasv_port = 1234; } $p2 = $pasv_port % ((int) 1 << 8); @@ -548,7 +507,6 @@ if ($pid) { dump_and_exit($buf); } } - fclose($s); exit; } diff --git a/ext/gd/gd.c b/ext/gd/gd.c index 88d04824ea..d811b66866 100644 --- a/ext/gd/gd.c +++ b/ext/gd/gd.c @@ -57,8 +57,6 @@ #include "gd_compat.h" -static int le_gd_font; - #include <gd.h> #include <gd_errors.h> #include <gdfontt.h> /* 1 Tiny font */ @@ -235,6 +233,75 @@ static void php_gd_object_minit_helper() php_gd_image_object_handlers.offset = XtOffsetOf(php_gd_image_object, std); } +static zend_class_entry *gd_font_ce = NULL; +static zend_object_handlers php_gd_font_object_handlers; + +typedef struct _php_gd_font_object { + gdFontPtr font; + zend_object std; +} php_gd_font_object; + +static php_gd_font_object *php_gd_font_object_from_zend_object(zend_object *zobj) +{ + return ((php_gd_font_object*)(zobj + 1)) - 1; +} + +static zend_object *php_gd_font_object_to_zend_object(php_gd_font_object *obj) +{ + return ((zend_object*)(obj + 1)) - 1; +} + +static zend_object *php_gd_font_object_create(zend_class_entry *ce) +{ + php_gd_font_object *obj = zend_object_alloc(sizeof(php_gd_font_object), ce); + zend_object *zobj = php_gd_font_object_to_zend_object(obj); + + obj->font = NULL; + zend_object_std_init(zobj, ce); + object_properties_init(zobj, ce); + zobj->handlers = &php_gd_font_object_handlers; + + return zobj; +} + +static void php_gd_font_object_free(zend_object *zobj) +{ + php_gd_font_object *obj = php_gd_font_object_from_zend_object(zobj); + + if (obj->font) { + if (obj->font->data) { + efree(obj->font->data); + } + efree(obj->font); + obj->font = NULL; + } + + zend_object_std_dtor(zobj); +} + +static zend_function *php_gd_font_object_get_constructor(zend_object *object) +{ + zend_throw_error(NULL, "You cannot initialize a GdFont object except through helper functions"); + return NULL; +} + +static void php_gd_font_minit_helper() +{ + zend_class_entry ce; + INIT_CLASS_ENTRY(ce, "GdFont", class_GdFont_methods); + gd_font_ce = zend_register_internal_class(&ce); + gd_font_ce->ce_flags |= ZEND_ACC_FINAL | ZEND_ACC_NO_DYNAMIC_PROPERTIES; + gd_font_ce->create_object = php_gd_font_object_create; + gd_font_ce->serialize = zend_class_serialize_deny; + gd_font_ce->unserialize = zend_class_unserialize_deny; + + /* setting up the object handlers for the GdFont class */ + memcpy(&php_gd_font_object_handlers, &std_object_handlers, sizeof(zend_object_handlers)); + php_gd_font_object_handlers.clone_obj = NULL; + php_gd_font_object_handlers.free_obj = php_gd_font_object_free; + php_gd_font_object_handlers.get_constructor = php_gd_font_object_get_constructor; + php_gd_font_object_handlers.offset = XtOffsetOf(php_gd_font_object, std); +} /********************************************************* * @@ -265,19 +332,6 @@ PHP_INI_BEGIN() PHP_INI_END() /* }}} */ -/* {{{ php_free_gd_font */ -static void php_free_gd_font(zend_resource *rsrc) -{ - gdFontPtr fp = (gdFontPtr) rsrc->ptr; - - if (fp->data) { - efree(fp->data); - } - - efree(fp); -} -/* }}} */ - /* {{{ php_gd_error_method */ void php_gd_error_method(int type, const char *format, va_list args) { @@ -302,8 +356,8 @@ void php_gd_error_method(int type, const char *format, va_list args) /* {{{ PHP_MINIT_FUNCTION */ PHP_MINIT_FUNCTION(gd) { - le_gd_font = zend_register_list_destructors_ex(php_free_gd_font, NULL, "gd font", module_number); php_gd_object_minit_helper(); + php_gd_font_minit_helper(); #if defined(HAVE_GD_FREETYPE) && defined(HAVE_GD_BUNDLED) gdFontCacheMutexSetup(); @@ -614,7 +668,6 @@ PHP_FUNCTION(gd_info) /* {{{ Load a new font */ PHP_FUNCTION(imageloadfont) { - zval *ind; zend_string *file; int hdr_size = sizeof(gdFont) - sizeof(char *); int body_size, n = 0, b, i, body_size_check; @@ -625,7 +678,7 @@ PHP_FUNCTION(imageloadfont) RETURN_THROWS(); } - stream = php_stream_open_wrapper(ZSTR_VAL(file), "rb", IGNORE_PATH | IGNORE_URL_WIN | REPORT_ERRORS, NULL); + stream = php_stream_open_wrapper(ZSTR_VAL(file), "rb", IGNORE_PATH | REPORT_ERRORS, NULL); if (stream == NULL) { RETURN_FALSE; } @@ -704,13 +757,8 @@ PHP_FUNCTION(imageloadfont) } php_stream_close(stream); - ind = zend_list_insert(font, le_gd_font); - - /* Adding 5 to the font index so we will never have font indices - * that overlap with the old fonts (with indices 1-5). The first - * list index given out is always 1. - */ - RETURN_LONG(Z_RES_HANDLE_P(ind) + 5); + object_init_ex(return_value, gd_font_ce); + php_gd_font_object_from_zend_object(Z_OBJ_P(return_value))->font = font; } /* }}} */ @@ -1584,7 +1632,7 @@ static void _php_image_create_from(INTERNAL_FUNCTION_PARAMETERS, int image_type, } - stream = php_stream_open_wrapper(file, "rb", REPORT_ERRORS|IGNORE_PATH|IGNORE_URL_WIN, NULL); + stream = php_stream_open_wrapper(file, "rb", REPORT_ERRORS|IGNORE_PATH, NULL); if (stream == NULL) { RETURN_FALSE; } @@ -1883,7 +1931,7 @@ PHP_FUNCTION(imagexbm) im = php_gd_libgdimageptr_from_zval_p(imgind); if (file != NULL) { - stream = php_stream_open_wrapper(file, "wb", REPORT_ERRORS|IGNORE_PATH|IGNORE_URL_WIN, NULL); + stream = php_stream_open_wrapper(file, "wb", REPORT_ERRORS|IGNORE_PATH, NULL); if (stream == NULL) { RETURN_FALSE; } @@ -2695,42 +2743,21 @@ PHP_FUNCTION(imagefilledpolygon) /* }}} */ /* {{{ php_find_gd_font */ -static gdFontPtr php_find_gd_font(int size) +static gdFontPtr php_find_gd_font(zend_object *font_obj, zend_long font_int) { - gdFontPtr font; + if (font_obj) { + return php_gd_font_object_from_zend_object(font_obj)->font; + } - switch (size) { - case 1: - font = gdFontTiny; - break; - case 2: - font = gdFontSmall; - break; - case 3: - font = gdFontMediumBold; - break; - case 4: - font = gdFontLarge; - break; - case 5: - font = gdFontGiant; - break; - default: { - zval *zv = zend_hash_index_find(&EG(regular_list), size - 5); - if (!zv || (Z_RES_P(zv))->type != le_gd_font) { - if (size < 1) { - font = gdFontTiny; - } else { - font = gdFontGiant; - } - } else { - font = (gdFontPtr)Z_RES_P(zv)->ptr; - } - } - break; - } - - return font; + switch (font_int) { + case 1: return gdFontTiny; + case 2: return gdFontSmall; + case 3: return gdFontMediumBold; + case 4: return gdFontLarge; + case 5: return gdFontGiant; + } + + return font_int < 1 ? gdFontTiny : gdFontGiant; } /* }}} */ @@ -2740,14 +2767,15 @@ static gdFontPtr php_find_gd_font(int size) */ static void php_imagefontsize(INTERNAL_FUNCTION_PARAMETERS, int arg) { - zend_long SIZE; + zend_object *font_obj; + zend_long font_int; gdFontPtr font; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "l", &SIZE) == FAILURE) { - RETURN_THROWS(); - } + ZEND_PARSE_PARAMETERS_START(1, 1) + Z_PARAM_OBJ_OF_CLASS_OR_LONG(font_obj, gd_font_ce, font_int) + ZEND_PARSE_PARAMETERS_END(); - font = php_find_gd_font(SIZE); + font = php_find_gd_font(font_obj, font_int); RETURN_LONG(arg ? font->h : font->w); } /* }}} */ @@ -2801,17 +2829,24 @@ static void php_gdimagecharup(gdImagePtr im, gdFontPtr f, int x, int y, int c, i static void php_imagechar(INTERNAL_FUNCTION_PARAMETERS, int mode) { zval *IM; - zend_long SIZE, X, Y, COL; + zend_long X, Y, COL; char *C; size_t C_len; gdImagePtr im; - int ch = 0, col, x, y, size, i, l = 0; + int ch = 0, col, x, y, i, l = 0; unsigned char *str = NULL; + zend_object *font_obj; + zend_long font_int; gdFontPtr font; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "Olllsl", &IM, gd_image_ce, &SIZE, &X, &Y, &C, &C_len, &COL) == FAILURE) { - RETURN_THROWS(); - } + ZEND_PARSE_PARAMETERS_START(6, 6) + Z_PARAM_OBJECT_OF_CLASS(IM, gd_image_ce) + Z_PARAM_OBJ_OF_CLASS_OR_LONG(font_obj, gd_font_ce, font_int) + Z_PARAM_LONG(X) + Z_PARAM_LONG(Y) + Z_PARAM_STRING(C, C_len) + Z_PARAM_LONG(COL) + ZEND_PARSE_PARAMETERS_END(); im = php_gd_libgdimageptr_from_zval_p(IM); @@ -2826,9 +2861,8 @@ static void php_imagechar(INTERNAL_FUNCTION_PARAMETERS, int mode) y = Y; x = X; - size = SIZE; - font = php_find_gd_font(size); + font = php_find_gd_font(font_obj, font_int); switch (mode) { case 0: @@ -4092,7 +4126,7 @@ static gdIOCtx *create_stream_context_from_zval(zval *to_zval) { return NULL; } - stream = php_stream_open_wrapper(Z_STRVAL_P(to_zval), "wb", REPORT_ERRORS|IGNORE_PATH|IGNORE_URL_WIN, NULL); + stream = php_stream_open_wrapper(Z_STRVAL_P(to_zval), "wb", REPORT_ERRORS|IGNORE_PATH, NULL); if (stream == NULL) { return NULL; } diff --git a/ext/gd/gd.stub.php b/ext/gd/gd.stub.php index fe0d5658ee..0ef410fc1e 100644 --- a/ext/gd/gd.stub.php +++ b/ext/gd/gd.stub.php @@ -2,13 +2,12 @@ /** @generate-function-entries */ -final class GdImage -{ -} +final class GdImage {} +final class GdFont {} function gd_info(): array {} -function imageloadfont(string $filename): int|false {} +function imageloadfont(string $filename): GdFont|false {} function imagesetstyle(GdImage $image, array $style): bool {} @@ -187,17 +186,17 @@ function imageopenpolygon(GdImage $image, array $points, int $num_points_or_colo function imagefilledpolygon(GdImage $image, array $points, int $num_points_or_color, ?int $color = null): bool {} -function imagefontwidth(int $font): int {} +function imagefontwidth(GdFont|int $font): int {} -function imagefontheight(int $font): int {} +function imagefontheight(GdFont|int $font): int {} -function imagechar(GdImage $image, int $font, int $x, int $y, string $char, int $color): bool {} +function imagechar(GdImage $image, GdFont|int $font, int $x, int $y, string $char, int $color): bool {} -function imagecharup(GdImage $image, int $font, int $x, int $y, string $char, int $color): bool {} +function imagecharup(GdImage $image, GdFont|int $font, int $x, int $y, string $char, int $color): bool {} -function imagestring(GdImage $image, int $font, int $x, int $y, string $string, int $color): bool {} +function imagestring(GdImage $image, GdFont|int $font, int $x, int $y, string $string, int $color): bool {} -function imagestringup(GdImage $image, int $font, int $x, int $y, string $string, int $color): bool {} +function imagestringup(GdImage $image, GdFont|int $font, int $x, int $y, string $string, int $color): bool {} function imagecopy(GdImage $dst_image, GdImage $src_image, int $dst_x, int $dst_y, int $src_x, int $src_y, int $src_width, int $src_height): bool {} diff --git a/ext/gd/gd_arginfo.h b/ext/gd/gd_arginfo.h index 4d59767b4e..ea61b08374 100644 --- a/ext/gd/gd_arginfo.h +++ b/ext/gd/gd_arginfo.h @@ -1,10 +1,10 @@ /* This is a generated file, edit the .stub.php file instead. - * Stub hash: 59256d0de105f1a2f5d5fc1e20f8090031b42a76 */ + * Stub hash: 23b7c8395c628239fb10d746ca6bba72871ee9f3 */ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_gd_info, 0, 0, IS_ARRAY, 0) ZEND_END_ARG_INFO() -ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_imageloadfont, 0, 1, MAY_BE_LONG|MAY_BE_FALSE) +ZEND_BEGIN_ARG_WITH_RETURN_OBJ_TYPE_MASK_EX(arginfo_imageloadfont, 0, 1, GdFont, MAY_BE_FALSE) ZEND_ARG_TYPE_INFO(0, filename, IS_STRING, 0) ZEND_END_ARG_INFO() @@ -380,14 +380,14 @@ ZEND_END_ARG_INFO() #define arginfo_imagefilledpolygon arginfo_imagepolygon ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_imagefontwidth, 0, 1, IS_LONG, 0) - ZEND_ARG_TYPE_INFO(0, font, IS_LONG, 0) + ZEND_ARG_OBJ_TYPE_MASK(0, font, GdFont, MAY_BE_LONG, NULL) ZEND_END_ARG_INFO() #define arginfo_imagefontheight arginfo_imagefontwidth ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_imagechar, 0, 6, _IS_BOOL, 0) ZEND_ARG_OBJ_INFO(0, image, GdImage, 0) - ZEND_ARG_TYPE_INFO(0, font, IS_LONG, 0) + ZEND_ARG_OBJ_TYPE_MASK(0, font, GdFont, MAY_BE_LONG, NULL) ZEND_ARG_TYPE_INFO(0, x, IS_LONG, 0) ZEND_ARG_TYPE_INFO(0, y, IS_LONG, 0) ZEND_ARG_TYPE_INFO(0, char, IS_STRING, 0) @@ -398,7 +398,7 @@ ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_imagestring, 0, 6, _IS_BOOL, 0) ZEND_ARG_OBJ_INFO(0, image, GdImage, 0) - ZEND_ARG_TYPE_INFO(0, font, IS_LONG, 0) + ZEND_ARG_OBJ_TYPE_MASK(0, font, GdFont, MAY_BE_LONG, NULL) ZEND_ARG_TYPE_INFO(0, x, IS_LONG, 0) ZEND_ARG_TYPE_INFO(0, y, IS_LONG, 0) ZEND_ARG_TYPE_INFO(0, string, IS_STRING, 0) @@ -831,3 +831,8 @@ static const zend_function_entry ext_functions[] = { static const zend_function_entry class_GdImage_methods[] = { ZEND_FE_END }; + + +static const zend_function_entry class_GdFont_methods[] = { + ZEND_FE_END +}; diff --git a/ext/hash/bench.php b/ext/hash/bench.php index 436390f7d5..846163315b 100755 --- a/ext/hash/bench.php +++ b/ext/hash/bench.php @@ -1,81 +1,12 @@ #!/usr/bin/env php <?php -/* - This gives rather interesting results :) +error_reporting(E_ALL); - Measures on a Notebook P4M-1.7 256MB Windows 2000: - sha1 0.556691 - tiger160,3 0.774469 - tiger192,3 0.776314 - tiger128,3 0.777004 - ripemd128 0.896674 - sha256 1.011164 - md5 1.016032 - tiger160,4 1.056617 - tiger128,4 1.063101 - tiger192,4 1.069258 - haval160,3 1.125099 - haval128,3 1.125679 - haval224,3 1.128017 - haval192,3 1.130026 - haval256,3 1.134846 - ripemd160 1.150693 - haval128,4 1.686261 - haval192,4 1.687274 - haval160,4 1.693091 - haval256,4 1.699323 - haval224,4 1.743094 - haval160,5 2.003452 - haval192,5 2.008341 - haval256,5 2.009048 - haval128,5 2.009555 - haval224,5 2.015539 - sha384 3.370734 - sha512 3.381121 - whirlpool 6.912327 - snefru 9.268168 - - Measures on a Desktop P4-2.4 512MB Debian (Linux-2.4): - md5 0.147739 - haval128,3 0.317006 - haval192,3 0.317524 - haval256,3 0.317526 - haval160,3 0.323035 - haval224,3 0.333318 - ripemd128 0.353447 - sha1 0.376200 - ripemd160 0.413758 - sha256 0.435957 - haval160,4 0.452357 - haval224,4 0.454531 - haval128,4 0.458026 - haval256,4 0.459051 - haval192,4 0.468094 - haval128,5 0.524262 - haval160,5 0.529573 - haval224,5 0.533655 - haval256,5 0.534446 - haval192,5 0.543726 - tiger128,3 0.577975 - tiger160,3 0.579951 - tiger192,3 0.597111 - tiger192,4 0.781408 - tiger160,4 0.801243 - tiger128,4 0.812239 - sha512 1.298627 - sha384 1.313607 - whirlpool 1.556159 - snefru 5.703742 - -*/ - -error_reporting(E_ALL&~E_NOTICE); - -$data = file_get_contents(__FILE__); +$data = str_repeat("\x00", (int) ($argv[1] ?? (2 * 1024))); $time = array(); foreach (hash_algos() as $algo) { - $time[$algo] = 0; + $time[$algo] = 0; } for ($j = 0; $j < 10; $j++) { @@ -92,4 +23,3 @@ asort($time, SORT_NUMERIC); foreach ($time as $a => $t) { printf("%-12s %02.6f\n", $a, $t); } -?> diff --git a/ext/imap/php_imap.c b/ext/imap/php_imap.c index c561a5e72c..4c72c90ff6 100644 --- a/ext/imap/php_imap.c +++ b/ext/imap/php_imap.c @@ -2758,7 +2758,7 @@ PHP_FUNCTION(imap_sort) } array_init(return_value); - if (slst != NIL && slst != 0) { + if (slst != NULL) { for (sl = slst; *sl; sl++) { add_next_index_long(return_value, *sl); } diff --git a/ext/mbstring/libmbfl/filters/mbfilter_7bit.c b/ext/mbstring/libmbfl/filters/mbfilter_7bit.c index 42ed23326d..331dc8e743 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_7bit.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_7bit.c @@ -31,6 +31,8 @@ #include "mbfilter.h" #include "mbfilter_7bit.h" +static int mbfl_filt_ident_7bit(int c, mbfl_identify_filter *filter); + const mbfl_encoding mbfl_encoding_7bit = { mbfl_no_encoding_7bit, "7bit", @@ -42,6 +44,12 @@ const mbfl_encoding mbfl_encoding_7bit = { NULL }; +const struct mbfl_identify_vtbl vtbl_identify_7bit = { + mbfl_no_encoding_7bit, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_7bit +}; + const struct mbfl_convert_vtbl vtbl_8bit_7bit = { mbfl_no_encoding_8bit, mbfl_no_encoding_7bit, @@ -75,6 +83,16 @@ int mbfl_filt_conv_any_7bit(int c, mbfl_convert_filter *filter) { if (c >= 0 && c < 0x80) { CK((*filter->output_function)(c, filter->data)); + } else { + mbfl_filt_conv_illegal_output(c, filter); + } + return c; +} + +static int mbfl_filt_ident_7bit(int c, mbfl_identify_filter *filter) +{ + if (c >= 0x80) { + filter->flag = 1; } return c; } diff --git a/ext/mbstring/libmbfl/filters/mbfilter_7bit.h b/ext/mbstring/libmbfl/filters/mbfilter_7bit.h index d0dfe2c518..f842a2ad3b 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_7bit.h +++ b/ext/mbstring/libmbfl/filters/mbfilter_7bit.h @@ -34,6 +34,7 @@ #include "mbfilter.h" extern const mbfl_encoding mbfl_encoding_7bit; +extern const struct mbfl_identify_vtbl vtbl_identify_7bit; extern const struct mbfl_convert_vtbl vtbl_8bit_7bit; extern const struct mbfl_convert_vtbl vtbl_7bit_8bit; diff --git a/ext/mbstring/libmbfl/filters/mbfilter_armscii8.c b/ext/mbstring/libmbfl/filters/mbfilter_armscii8.c index 38c8f98bd1..e2fa8d86ea 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_armscii8.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_armscii8.c @@ -38,7 +38,7 @@ const mbfl_encoding mbfl_encoding_armscii8 = { mbfl_no_encoding_armscii8, "ArmSCII-8", "ArmSCII-8", - (const char *(*)[])&mbfl_encoding_armscii8_aliases, + mbfl_encoding_armscii8_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_armscii8_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_ascii.c b/ext/mbstring/libmbfl/filters/mbfilter_ascii.c index 43d659a46c..4e02c26aac 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_ascii.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_ascii.c @@ -39,7 +39,7 @@ const mbfl_encoding mbfl_encoding_ascii = { mbfl_no_encoding_ascii, "ASCII", "US-ASCII", /* preferred MIME name */ - (const char *(*)[])&mbfl_encoding_ascii_aliases, + mbfl_encoding_ascii_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_ascii_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_big5.c b/ext/mbstring/libmbfl/filters/mbfilter_big5.c index 1e59924092..95c75a37e3 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_big5.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_big5.c @@ -59,7 +59,7 @@ const mbfl_encoding mbfl_encoding_big5 = { mbfl_no_encoding_big5, "BIG-5", "BIG5", - (const char *(*)[])&mbfl_encoding_big5_aliases, + mbfl_encoding_big5_aliases, mblen_table_big5, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_big5_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp1251.c b/ext/mbstring/libmbfl/filters/mbfilter_cp1251.c index 470adf1a08..af05b5ec6d 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_cp1251.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_cp1251.c @@ -39,7 +39,7 @@ const mbfl_encoding mbfl_encoding_cp1251 = { mbfl_no_encoding_cp1251, "Windows-1251", "Windows-1251", - (const char *(*)[])&mbfl_encoding_cp1251_aliases, + mbfl_encoding_cp1251_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_cp1251_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp1252.c b/ext/mbstring/libmbfl/filters/mbfilter_cp1252.c index a274e4b796..1f6717d2ac 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_cp1252.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_cp1252.c @@ -39,7 +39,7 @@ const mbfl_encoding mbfl_encoding_cp1252 = { mbfl_no_encoding_cp1252, "Windows-1252", "Windows-1252", - (const char *(*)[])&mbfl_encoding_cp1252_aliases, + mbfl_encoding_cp1252_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_cp1252_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp1254.c b/ext/mbstring/libmbfl/filters/mbfilter_cp1254.c index 7e4bb4b148..eebd5b322a 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_cp1254.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_cp1254.c @@ -39,7 +39,7 @@ const mbfl_encoding mbfl_encoding_cp1254 = { mbfl_no_encoding_cp1254, "Windows-1254", "Windows-1254", - (const char *(*)[])&mbfl_encoding_cp1254_aliases, + mbfl_encoding_cp1254_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_cp1254_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp5022x.c b/ext/mbstring/libmbfl/filters/mbfilter_cp5022x.c index 3879f9eb5b..798487d342 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_cp5022x.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_cp5022x.c @@ -59,7 +59,7 @@ const mbfl_encoding mbfl_encoding_cp50220 = { mbfl_no_encoding_cp50220, "CP50220", "ISO-2022-JP", - (const char *(*)[])NULL, + NULL, NULL, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_cp50220_wchar, @@ -70,7 +70,7 @@ const mbfl_encoding mbfl_encoding_cp50220raw = { mbfl_no_encoding_cp50220raw, "CP50220raw", "ISO-2022-JP", - (const char *(*)[])NULL, + NULL, NULL, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_cp50220raw_wchar, @@ -585,7 +585,7 @@ mbfl_filt_conv_wchar_cp50220_ctor(mbfl_convert_filter *filt) ctx->last.opaque = ctx; ctx->last.data = filt->data; filt->filter_function = vtbl_tl_jisx0201_jisx0208.filter_function; - filt->filter_flush = vtbl_tl_jisx0201_jisx0208.filter_flush; + filt->filter_flush = (filter_flush_t)vtbl_tl_jisx0201_jisx0208.filter_flush; filt->output_function = (output_function_t)ctx->last.filter_function; filt->flush_function = (flush_function_t)ctx->last.filter_flush; filt->data = &ctx->last; diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp51932.c b/ext/mbstring/libmbfl/filters/mbfilter_cp51932.c index da046c86b7..9b893bc638 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_cp51932.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_cp51932.c @@ -68,7 +68,7 @@ const mbfl_encoding mbfl_encoding_cp51932 = { mbfl_no_encoding_cp51932, "CP51932", "CP51932", - (const char *(*)[])&mbfl_encoding_cp51932_aliases, + mbfl_encoding_cp51932_aliases, mblen_table_eucjp, MBFL_ENCTYPE_MBCS, &vtbl_cp51932_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp850.c b/ext/mbstring/libmbfl/filters/mbfilter_cp850.c index 5d6b265fc1..d9d705f70f 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_cp850.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_cp850.c @@ -35,7 +35,7 @@ const mbfl_encoding mbfl_encoding_cp850 = { mbfl_no_encoding_cp850, "CP850", "CP850", - (const char *(*)[])&mbfl_encoding_cp850_aliases, + mbfl_encoding_cp850_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_cp850_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp866.c b/ext/mbstring/libmbfl/filters/mbfilter_cp866.c index 504903bef0..9c548627b0 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_cp866.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_cp866.c @@ -39,7 +39,7 @@ const mbfl_encoding mbfl_encoding_cp866 = { mbfl_no_encoding_cp866, "CP866", "CP866", - (const char *(*)[])&mbfl_encoding_cp866_aliases, + mbfl_encoding_cp866_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_cp866_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp932.c b/ext/mbstring/libmbfl/filters/mbfilter_cp932.c index 3cbca1cb64..511d864465 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_cp932.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_cp932.c @@ -60,7 +60,7 @@ const mbfl_encoding mbfl_encoding_cp932 = { mbfl_no_encoding_cp932, "CP932", "Shift_JIS", - (const char *(*)[])&mbfl_encoding_cp932_aliases, + mbfl_encoding_cp932_aliases, mblen_table_sjis, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_cp932_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp936.c b/ext/mbstring/libmbfl/filters/mbfilter_cp936.c index df6d77907f..d73954267f 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_cp936.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_cp936.c @@ -59,7 +59,7 @@ const mbfl_encoding mbfl_encoding_cp936 = { mbfl_no_encoding_cp936, "CP936", "CP936", - (const char *(*)[])&mbfl_encoding_cp936_aliases, + mbfl_encoding_cp936_aliases, mblen_table_cp936, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_cp936_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_euc_cn.c b/ext/mbstring/libmbfl/filters/mbfilter_euc_cn.c index 53a8aac2a6..0a7edaa927 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_euc_cn.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_euc_cn.c @@ -59,7 +59,7 @@ const mbfl_encoding mbfl_encoding_euc_cn = { mbfl_no_encoding_euc_cn, "EUC-CN", "CN-GB", - (const char *(*)[])&mbfl_encoding_euc_cn_aliases, + mbfl_encoding_euc_cn_aliases, mblen_table_euccn, MBFL_ENCTYPE_MBCS, &vtbl_euccn_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_euc_jp.c b/ext/mbstring/libmbfl/filters/mbfilter_euc_jp.c index 252a10a58d..61da27f043 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_euc_jp.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_euc_jp.c @@ -60,7 +60,7 @@ const mbfl_encoding mbfl_encoding_euc_jp = { mbfl_no_encoding_euc_jp, "EUC-JP", "EUC-JP", - (const char *(*)[])&mbfl_encoding_euc_jp_aliases, + mbfl_encoding_euc_jp_aliases, mblen_table_eucjp, MBFL_ENCTYPE_MBCS, &vtbl_eucjp_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_euc_jp_2004.c b/ext/mbstring/libmbfl/filters/mbfilter_euc_jp_2004.c index 30d869cea5..c6c708b337 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_euc_jp_2004.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_euc_jp_2004.c @@ -40,7 +40,7 @@ const mbfl_encoding mbfl_encoding_eucjp2004 = { mbfl_no_encoding_eucjp2004, "EUC-JP-2004", "EUC-JP", - (const char *(*)[])&mbfl_encoding_eucjp2004_aliases, + mbfl_encoding_eucjp2004_aliases, mblen_table_eucjp, MBFL_ENCTYPE_MBCS, &vtbl_eucjp2004_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_euc_jp_win.c b/ext/mbstring/libmbfl/filters/mbfilter_euc_jp_win.c index 5798bff0fc..7c1aebdea5 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_euc_jp_win.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_euc_jp_win.c @@ -69,7 +69,7 @@ const mbfl_encoding mbfl_encoding_eucjp_win = { mbfl_no_encoding_eucjp_win, "eucJP-win", "EUC-JP", - (const char *(*)[])&mbfl_encoding_eucjp_win_aliases, + mbfl_encoding_eucjp_win_aliases, mblen_table_eucjp, MBFL_ENCTYPE_MBCS, &vtbl_eucjpwin_wchar, @@ -319,8 +319,6 @@ mbfl_filt_conv_wchar_eucjpwin(int c, mbfl_convert_filter *filter) s1 = 0x2172; } else if (c == 0xffe2) { /* FULLWIDTH NOT SIGN */ s1 = 0x224c; - } else if (c == 0xff5e) { /* FULLWIDTH TILDE */ - s1 = 0x2141; } else { s1 = -1; c1 = 0; diff --git a/ext/mbstring/libmbfl/filters/mbfilter_euc_kr.c b/ext/mbstring/libmbfl/filters/mbfilter_euc_kr.c index e9b08e08c1..0c44a99860 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_euc_kr.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_euc_kr.c @@ -58,7 +58,7 @@ const mbfl_encoding mbfl_encoding_euc_kr = { mbfl_no_encoding_euc_kr, "EUC-KR", "EUC-KR", - (const char *(*)[])&mbfl_encoding_euc_kr_aliases, + mbfl_encoding_euc_kr_aliases, mblen_table_euckr, MBFL_ENCTYPE_MBCS, &vtbl_euckr_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_euc_tw.c b/ext/mbstring/libmbfl/filters/mbfilter_euc_tw.c index c76eeb36bf..74bca29d47 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_euc_tw.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_euc_tw.c @@ -60,7 +60,7 @@ const mbfl_encoding mbfl_encoding_euc_tw = { mbfl_no_encoding_euc_tw, "EUC-TW", "EUC-TW", - (const char *(*)[])&mbfl_encoding_euc_tw_aliases, + mbfl_encoding_euc_tw_aliases, mblen_table_euctw, MBFL_ENCTYPE_MBCS, &vtbl_euctw_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_gb18030.c b/ext/mbstring/libmbfl/filters/mbfilter_gb18030.c index b94d664f40..ec6e3973a5 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_gb18030.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_gb18030.c @@ -41,7 +41,7 @@ const mbfl_encoding mbfl_encoding_gb18030 = { mbfl_no_encoding_gb18030, "GB18030", "GB18030", - (const char *(*)[])&mbfl_encoding_gb18030_aliases, + mbfl_encoding_gb18030_aliases, NULL, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_gb18030_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_htmlent.c b/ext/mbstring/libmbfl/filters/mbfilter_htmlent.c index 7176fd4717..95a0817103 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_htmlent.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_htmlent.c @@ -57,7 +57,7 @@ const mbfl_encoding mbfl_encoding_html_ent = { mbfl_no_encoding_html_ent, "HTML-ENTITIES", "HTML-ENTITIES", - (const char *(*)[])&mbfl_encoding_html_ent_aliases, + mbfl_encoding_html_ent_aliases, NULL, MBFL_ENCTYPE_GL_UNSAFE, &vtbl_html_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso2022_jp_ms.c b/ext/mbstring/libmbfl/filters/mbfilter_iso2022_jp_ms.c index 1e9757eae6..32f4072ce2 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso2022_jp_ms.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso2022_jp_ms.c @@ -42,7 +42,7 @@ const mbfl_encoding mbfl_encoding_2022jpms = { mbfl_no_encoding_2022jpms, "ISO-2022-JP-MS", "ISO-2022-JP", - (const char *(*)[])&mbfl_encoding_2022jpms_aliases, + mbfl_encoding_2022jpms_aliases, NULL, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_2022jpms_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso2022jp_mobile.c b/ext/mbstring/libmbfl/filters/mbfilter_iso2022jp_mobile.c index 32ea47712e..00f0ddeda8 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso2022jp_mobile.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso2022jp_mobile.c @@ -44,7 +44,7 @@ const mbfl_encoding mbfl_encoding_2022jp_kddi = { mbfl_no_encoding_2022jp_kddi, "ISO-2022-JP-MOBILE#KDDI", "ISO-2022-JP", - (const char *(*)[])&mbfl_encoding_2022jp_kddi_aliases, + mbfl_encoding_2022jp_kddi_aliases, NULL, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_2022jp_kddi_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_1.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_1.c index 1549ba81bf..899234f3b0 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_1.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_1.c @@ -36,7 +36,7 @@ const mbfl_encoding mbfl_encoding_8859_1 = { mbfl_no_encoding_8859_1, "ISO-8859-1", "ISO-8859-1", - (const char *(*)[])&mbfl_encoding_8859_1_aliases, + mbfl_encoding_8859_1_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_1_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_10.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_10.c index 12cd960a88..e82c52f4b6 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_10.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_10.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_8859_10 = { mbfl_no_encoding_8859_10, "ISO-8859-10", "ISO-8859-10", - (const char *(*)[])&mbfl_encoding_8859_10_aliases, + mbfl_encoding_8859_10_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_10_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_13.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_13.c index 25526263b8..ec5d5348a3 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_13.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_13.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_8859_13 = { mbfl_no_encoding_8859_13, "ISO-8859-13", "ISO-8859-13", - (const char *(*)[])&mbfl_encoding_8859_13_aliases, + mbfl_encoding_8859_13_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_13_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_14.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_14.c index 06a2e4cd98..ac5b9a40d9 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_14.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_14.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_8859_14 = { mbfl_no_encoding_8859_14, "ISO-8859-14", "ISO-8859-14", - (const char *(*)[])&mbfl_encoding_8859_14_aliases, + mbfl_encoding_8859_14_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_14_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_15.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_15.c index 82b40796b8..7d9e21e5c3 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_15.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_15.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_8859_15 = { mbfl_no_encoding_8859_15, "ISO-8859-15", "ISO-8859-15", - (const char *(*)[])&mbfl_encoding_8859_15_aliases, + mbfl_encoding_8859_15_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_15_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_16.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_16.c index 80f1461364..5c8ae964be 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_16.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_16.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_8859_16 = { mbfl_no_encoding_8859_16, "ISO-8859-16", "ISO-8859-16", - (const char *(*)[])&mbfl_encoding_8859_16_aliases, + mbfl_encoding_8859_16_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_16_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_2.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_2.c index 86044e15d3..deacf0c0f4 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_2.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_2.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_8859_2 = { mbfl_no_encoding_8859_2, "ISO-8859-2", "ISO-8859-2", - (const char *(*)[])&mbfl_encoding_8859_2_aliases, + mbfl_encoding_8859_2_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_2_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_3.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_3.c index 64eb9b5cf9..5cd3bfce1d 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_3.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_3.c @@ -31,13 +31,15 @@ #include "mbfilter_iso8859_3.h" #include "unicode_table_iso8859_3.h" +static int mbfl_filt_ident_iso8859_3(int c, mbfl_identify_filter *filter); + static const char *mbfl_encoding_8859_3_aliases[] = {"ISO8859-3", "latin3", NULL}; const mbfl_encoding mbfl_encoding_8859_3 = { mbfl_no_encoding_8859_3, "ISO-8859-3", "ISO-8859-3", - (const char *(*)[])&mbfl_encoding_8859_3_aliases, + mbfl_encoding_8859_3_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_3_wchar, @@ -47,7 +49,7 @@ const mbfl_encoding mbfl_encoding_8859_3 = { const struct mbfl_identify_vtbl vtbl_identify_8859_3 = { mbfl_no_encoding_8859_3, mbfl_filt_ident_common_ctor, - mbfl_filt_ident_true + mbfl_filt_ident_iso8859_3 }; const struct mbfl_convert_vtbl vtbl_8859_3_wchar = { @@ -119,9 +121,6 @@ int mbfl_filt_conv_wchar_8859_3(int c, mbfl_convert_filter *filter) } n--; } - if (s <= 0 && (c & ~MBFL_WCSPLANE_MASK) == MBFL_WCSPLANE_8859_3) { - s = c & MBFL_WCSPLANE_MASK; - } } if (s >= 0) { @@ -132,3 +131,11 @@ int mbfl_filt_conv_wchar_8859_3(int c, mbfl_convert_filter *filter) return c; } + +static int mbfl_filt_ident_iso8859_3(int c, mbfl_identify_filter *filter) +{ + if (c >= 0xA0 && !iso8859_3_ucs_table[c - 0xA0]) { + filter->status = 1; + } + return c; +} diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_4.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_4.c index 12a9f7cb10..263036f27b 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_4.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_4.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_8859_4 = { mbfl_no_encoding_8859_4, "ISO-8859-4", "ISO-8859-4", - (const char *(*)[])&mbfl_encoding_8859_4_aliases, + mbfl_encoding_8859_4_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_4_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_5.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_5.c index 70496d07e7..b9a6132029 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_5.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_5.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_8859_5 = { mbfl_no_encoding_8859_5, "ISO-8859-5", "ISO-8859-5", - (const char *(*)[])&mbfl_encoding_8859_5_aliases, + mbfl_encoding_8859_5_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_5_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_6.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_6.c index 59d981e9c2..8b1fc4bbad 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_6.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_6.c @@ -31,13 +31,15 @@ #include "mbfilter_iso8859_6.h" #include "unicode_table_iso8859_6.h" +static int mbfl_filt_ident_iso8859_6(int c, mbfl_identify_filter *filter); + static const char *mbfl_encoding_8859_6_aliases[] = {"ISO8859-6", "arabic", NULL}; const mbfl_encoding mbfl_encoding_8859_6 = { mbfl_no_encoding_8859_6, "ISO-8859-6", "ISO-8859-6", - (const char *(*)[])&mbfl_encoding_8859_6_aliases, + mbfl_encoding_8859_6_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_6_wchar, @@ -47,7 +49,7 @@ const mbfl_encoding mbfl_encoding_8859_6 = { const struct mbfl_identify_vtbl vtbl_identify_8859_6 = { mbfl_no_encoding_8859_6, mbfl_filt_ident_common_ctor, - mbfl_filt_ident_true + mbfl_filt_ident_iso8859_6 }; const struct mbfl_convert_vtbl vtbl_8859_6_wchar = { @@ -119,9 +121,6 @@ int mbfl_filt_conv_wchar_8859_6(int c, mbfl_convert_filter *filter) } n--; } - if (s <= 0 && (c & ~MBFL_WCSPLANE_MASK) == MBFL_WCSPLANE_8859_6) { - s = c & MBFL_WCSPLANE_MASK; - } } if (s >= 0) { @@ -132,3 +131,11 @@ int mbfl_filt_conv_wchar_8859_6(int c, mbfl_convert_filter *filter) return c; } + +static int mbfl_filt_ident_iso8859_6(int c, mbfl_identify_filter *filter) +{ + if (c >= 0xA0 && !iso8859_6_ucs_table[c - 0xA0]) { + filter->status = 1; + } + return c; +} diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_7.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_7.c index 73ead533ea..e87ae0ec36 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_7.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_7.c @@ -31,13 +31,15 @@ #include "mbfilter_iso8859_7.h" #include "unicode_table_iso8859_7.h" +static int mbfl_filt_ident_iso8859_7(int c, mbfl_identify_filter *filter); + static const char *mbfl_encoding_8859_7_aliases[] = {"ISO8859-7", "greek", NULL}; const mbfl_encoding mbfl_encoding_8859_7 = { mbfl_no_encoding_8859_7, "ISO-8859-7", "ISO-8859-7", - (const char *(*)[])&mbfl_encoding_8859_7_aliases, + mbfl_encoding_8859_7_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_7_wchar, @@ -47,7 +49,7 @@ const mbfl_encoding mbfl_encoding_8859_7 = { const struct mbfl_identify_vtbl vtbl_identify_8859_7 = { mbfl_no_encoding_8859_7, mbfl_filt_ident_common_ctor, - mbfl_filt_ident_true + mbfl_filt_ident_iso8859_7 }; const struct mbfl_convert_vtbl vtbl_8859_7_wchar = { @@ -119,9 +121,6 @@ int mbfl_filt_conv_wchar_8859_7(int c, mbfl_convert_filter *filter) } n--; } - if (s <= 0 && (c & ~MBFL_WCSPLANE_MASK) == MBFL_WCSPLANE_8859_7) { - s = c & MBFL_WCSPLANE_MASK; - } } if (s >= 0) { @@ -132,3 +131,11 @@ int mbfl_filt_conv_wchar_8859_7(int c, mbfl_convert_filter *filter) return c; } + +static int mbfl_filt_ident_iso8859_7(int c, mbfl_identify_filter *filter) +{ + /* These bytes are not mapped to any character in ISO-8859-7 */ + if (c == 0xAE || c == 0xD2 || c == 0xFF) + filter->status = 1; + return c; +} diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_8.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_8.c index 05ca13f330..3436d8d3ea 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_8.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_8.c @@ -31,13 +31,15 @@ #include "mbfilter_iso8859_8.h" #include "unicode_table_iso8859_8.h" +static int mbfl_filt_ident_iso8859_8(int c, mbfl_identify_filter *filter); + static const char *mbfl_encoding_8859_8_aliases[] = {"ISO8859-8", "hebrew", NULL}; const mbfl_encoding mbfl_encoding_8859_8 = { mbfl_no_encoding_8859_8, "ISO-8859-8", "ISO-8859-8", - (const char *(*)[])&mbfl_encoding_8859_8_aliases, + mbfl_encoding_8859_8_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_8_wchar, @@ -47,7 +49,7 @@ const mbfl_encoding mbfl_encoding_8859_8 = { const struct mbfl_identify_vtbl vtbl_identify_8859_8 = { mbfl_no_encoding_8859_8, mbfl_filt_ident_common_ctor, - mbfl_filt_ident_true + mbfl_filt_ident_iso8859_8 }; const struct mbfl_convert_vtbl vtbl_8859_8_wchar = { @@ -118,9 +120,6 @@ int mbfl_filt_conv_wchar_8859_8(int c, mbfl_convert_filter *filter) } n--; } - if (s <= 0 && (c & ~MBFL_WCSPLANE_MASK) == MBFL_WCSPLANE_8859_8) { - s = c & MBFL_WCSPLANE_MASK; - } } if (s >= 0) { @@ -131,3 +130,11 @@ int mbfl_filt_conv_wchar_8859_8(int c, mbfl_convert_filter *filter) return c; } + +static int mbfl_filt_ident_iso8859_8(int c, mbfl_identify_filter *filter) +{ + if (c >= 0xA0 && !iso8859_8_ucs_table[c - 0xA0]) { + filter->status = 1; + } + return c; +} diff --git a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_9.c b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_9.c index 074f90bef2..a2b87d1174 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_iso8859_9.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_iso8859_9.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_8859_9 = { mbfl_no_encoding_8859_9, "ISO-8859-9", "ISO-8859-9", - (const char *(*)[])&mbfl_encoding_8859_9_aliases, + mbfl_encoding_8859_9_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8859_9_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_koi8r.c b/ext/mbstring/libmbfl/filters/mbfilter_koi8r.c index 7790be65bb..db00a73c4f 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_koi8r.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_koi8r.c @@ -39,7 +39,7 @@ const mbfl_encoding mbfl_encoding_koi8r = { mbfl_no_encoding_koi8r, "KOI8-R", "KOI8-R", - (const char *(*)[])&mbfl_encoding_koi8r_aliases, + mbfl_encoding_koi8r_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_koi8r_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_koi8u.c b/ext/mbstring/libmbfl/filters/mbfilter_koi8u.c index a62a1c596b..54c06d2663 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_koi8u.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_koi8u.c @@ -36,7 +36,7 @@ const mbfl_encoding mbfl_encoding_koi8u = { mbfl_no_encoding_koi8u, "KOI8-U", "KOI8-U", - (const char *(*)[])&mbfl_encoding_koi8u_aliases, + mbfl_encoding_koi8u_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_koi8u_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_qprint.c b/ext/mbstring/libmbfl/filters/mbfilter_qprint.c index d42cdca883..8f806def25 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_qprint.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_qprint.c @@ -37,7 +37,7 @@ const mbfl_encoding mbfl_encoding_qprint = { mbfl_no_encoding_qprint, "Quoted-Printable", "Quoted-Printable", - (const char *(*)[])&mbfl_encoding_qprint_aliases, + mbfl_encoding_qprint_aliases, NULL, MBFL_ENCTYPE_GL_UNSAFE, NULL, @@ -110,8 +110,7 @@ int mbfl_filt_conv_qprintenc(int c, mbfl_convert_filter *filter) } if (s <= 0 || s >= 0x80 || s == 0x3d /* not ASCII or '=' */ - || ((filter->status & MBFL_QPRINT_STS_MIME_HEADER) != 0 && - (mbfl_charprop_table[s] & MBFL_CHP_MMHQENC) != 0)) { + || ((filter->status & MBFL_QPRINT_STS_MIME_HEADER) && mime_char_needs_qencode[s])) { /* hex-octet */ CK((*filter->output_function)(0x3d, filter->data)); /* '=' */ n = (s >> 4) & 0xf; diff --git a/ext/mbstring/libmbfl/filters/mbfilter_sjis.c b/ext/mbstring/libmbfl/filters/mbfilter_sjis.c index dc5c2e4882..d993c89d7f 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_sjis.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_sjis.c @@ -63,7 +63,7 @@ const mbfl_encoding mbfl_encoding_sjis = { mbfl_no_encoding_sjis, "SJIS", "Shift_JIS", - (const char *(*)[])&mbfl_encoding_sjis_aliases, + mbfl_encoding_sjis_aliases, mblen_table_sjis, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_sjis_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_sjis_2004.c b/ext/mbstring/libmbfl/filters/mbfilter_sjis_2004.c index 04e1d3af17..e611f3f846 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_sjis_2004.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_sjis_2004.c @@ -45,7 +45,7 @@ const mbfl_encoding mbfl_encoding_sjis2004 = { mbfl_no_encoding_sjis2004, "SJIS-2004", "Shift_JIS", - (const char *(*)[])&mbfl_encoding_sjis2004_aliases, + mbfl_encoding_sjis2004_aliases, mblen_table_sjis, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_sjis2004_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_sjis_mac.c b/ext/mbstring/libmbfl/filters/mbfilter_sjis_mac.c index dc3d3692f4..ff8d01937f 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_sjis_mac.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_sjis_mac.c @@ -46,7 +46,7 @@ const mbfl_encoding mbfl_encoding_sjis_mac = { mbfl_no_encoding_sjis_mac, "SJIS-mac", "Shift_JIS", - (const char *(*)[])&mbfl_encoding_sjis_mac_aliases, + mbfl_encoding_sjis_mac_aliases, mblen_table_sjis, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_sjis_mac_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_sjis_mobile.c b/ext/mbstring/libmbfl/filters/mbfilter_sjis_mobile.c index 68084a0e0a..d5015a92a8 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_sjis_mobile.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_sjis_mobile.c @@ -47,7 +47,7 @@ const mbfl_encoding mbfl_encoding_sjis_docomo = { mbfl_no_encoding_sjis_docomo, "SJIS-Mobile#DOCOMO", "Shift_JIS", - (const char *(*)[])&mbfl_encoding_sjis_docomo_aliases, + mbfl_encoding_sjis_docomo_aliases, mblen_table_sjis, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_sjis_docomo_wchar, @@ -58,7 +58,7 @@ const mbfl_encoding mbfl_encoding_sjis_kddi = { mbfl_no_encoding_sjis_kddi, "SJIS-Mobile#KDDI", "Shift_JIS", - (const char *(*)[])&mbfl_encoding_sjis_kddi_aliases, + mbfl_encoding_sjis_kddi_aliases, mblen_table_sjis, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_sjis_kddi_wchar, @@ -69,7 +69,7 @@ const mbfl_encoding mbfl_encoding_sjis_sb = { mbfl_no_encoding_sjis_sb, "SJIS-Mobile#SOFTBANK", "Shift_JIS", - (const char *(*)[])&mbfl_encoding_sjis_sb_aliases, + mbfl_encoding_sjis_sb_aliases, mblen_table_sjis, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_sjis_sb_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_sjis_open.c b/ext/mbstring/libmbfl/filters/mbfilter_sjis_open.c index 2535036e23..b3ec28740b 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_sjis_open.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_sjis_open.c @@ -60,7 +60,7 @@ const mbfl_encoding mbfl_encoding_sjis_open = { mbfl_no_encoding_sjis_open, "SJIS-win", "Shift_JIS", - (const char *(*)[])&mbfl_encoding_sjis_open_aliases, + mbfl_encoding_sjis_open_aliases, mblen_table_sjis, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_sjis_open_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_ucs2.c b/ext/mbstring/libmbfl/filters/mbfilter_ucs2.c index 68172efbe6..258c564aae 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_ucs2.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_ucs2.c @@ -30,13 +30,15 @@ #include "mbfilter.h" #include "mbfilter_ucs2.h" +static int mbfl_filt_ident_ucs2(int c, mbfl_identify_filter *filter); + static const char *mbfl_encoding_ucs2_aliases[] = {"ISO-10646-UCS-2", "UCS2" , "UNICODE", NULL}; const mbfl_encoding mbfl_encoding_ucs2 = { mbfl_no_encoding_ucs2, "UCS-2", "UCS-2", - (const char *(*)[])&mbfl_encoding_ucs2_aliases, + mbfl_encoding_ucs2_aliases, NULL, MBFL_ENCTYPE_WCS2BE, &vtbl_ucs2_wchar, @@ -65,6 +67,24 @@ const mbfl_encoding mbfl_encoding_ucs2le = { &vtbl_wchar_ucs2le }; +const struct mbfl_identify_vtbl vtbl_identify_ucs2 = { + mbfl_no_encoding_ucs2, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_ucs2 +}; + +const struct mbfl_identify_vtbl vtbl_identify_ucs2be = { + mbfl_no_encoding_ucs2be, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_ucs2 +}; + +const struct mbfl_identify_vtbl vtbl_identify_ucs2le = { + mbfl_no_encoding_ucs2le, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_ucs2 +}; + const struct mbfl_convert_vtbl vtbl_ucs2_wchar = { mbfl_no_encoding_ucs2, mbfl_no_encoding_wchar, @@ -237,3 +257,10 @@ int mbfl_filt_conv_wchar_ucs2le(int c, mbfl_convert_filter *filter) return c; } + +static int mbfl_filt_ident_ucs2(int c, mbfl_identify_filter *filter) +{ + /* Input string must be a multiple of 2 bytes */ + filter->status = (filter->status + 1) % 2; + return c; +} diff --git a/ext/mbstring/libmbfl/filters/mbfilter_ucs2.h b/ext/mbstring/libmbfl/filters/mbfilter_ucs2.h index bbf567a493..f6d9b8b772 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_ucs2.h +++ b/ext/mbstring/libmbfl/filters/mbfilter_ucs2.h @@ -35,6 +35,9 @@ extern const mbfl_encoding mbfl_encoding_ucs2; extern const mbfl_encoding mbfl_encoding_ucs2be; extern const mbfl_encoding mbfl_encoding_ucs2le; +extern const struct mbfl_identify_vtbl vtbl_identify_ucs2; +extern const struct mbfl_identify_vtbl vtbl_identify_ucs2be; +extern const struct mbfl_identify_vtbl vtbl_identify_ucs2le; extern const struct mbfl_convert_vtbl vtbl_ucs2_wchar; extern const struct mbfl_convert_vtbl vtbl_wchar_ucs2; extern const struct mbfl_convert_vtbl vtbl_ucs2be_wchar; diff --git a/ext/mbstring/libmbfl/filters/mbfilter_ucs4.c b/ext/mbstring/libmbfl/filters/mbfilter_ucs4.c index fcdd20d615..396bae91c5 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_ucs4.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_ucs4.c @@ -36,7 +36,7 @@ const mbfl_encoding mbfl_encoding_ucs4 = { mbfl_no_encoding_ucs4, "UCS-4", "UCS-4", - (const char *(*)[])&mbfl_encoding_ucs4_aliases, + mbfl_encoding_ucs4_aliases, NULL, MBFL_ENCTYPE_WCS4BE, &vtbl_ucs4_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_uhc.c b/ext/mbstring/libmbfl/filters/mbfilter_uhc.c index 5dc4aa81a9..c86612bac2 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_uhc.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_uhc.c @@ -59,7 +59,7 @@ const mbfl_encoding mbfl_encoding_uhc = { mbfl_no_encoding_uhc, "UHC", "UHC", - (const char *(*)[])&mbfl_encoding_uhc_aliases, + mbfl_encoding_uhc_aliases, mblen_table_uhc, MBFL_ENCTYPE_MBCS, &vtbl_uhc_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf16.c b/ext/mbstring/libmbfl/filters/mbfilter_utf16.c index 620e8a76f7..c23a432695 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_utf16.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_utf16.c @@ -30,13 +30,18 @@ #include "mbfilter.h" #include "mbfilter_utf16.h" +static int mbfl_filt_ident_utf16(int c, mbfl_identify_filter *filter); +static int mbfl_filt_ident_utf16le(int c, mbfl_identify_filter *filter); +static int mbfl_filt_ident_utf16be(int c, mbfl_identify_filter *filter); +static int mbfl_filt_conv_utf16_wchar_flush(mbfl_convert_filter *filter); + static const char *mbfl_encoding_utf16_aliases[] = {"utf16", NULL}; const mbfl_encoding mbfl_encoding_utf16 = { mbfl_no_encoding_utf16, "UTF-16", "UTF-16", - (const char *(*)[])&mbfl_encoding_utf16_aliases, + mbfl_encoding_utf16_aliases, NULL, MBFL_ENCTYPE_MWC2BE, &vtbl_utf16_wchar, @@ -65,13 +70,31 @@ const mbfl_encoding mbfl_encoding_utf16le = { &vtbl_wchar_utf16le }; +const struct mbfl_identify_vtbl vtbl_identify_utf16 = { + mbfl_no_encoding_utf16, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_utf16 +}; + +const struct mbfl_identify_vtbl vtbl_identify_utf16le = { + mbfl_no_encoding_utf16le, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_utf16le +}; + +const struct mbfl_identify_vtbl vtbl_identify_utf16be = { + mbfl_no_encoding_utf16be, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_utf16be +}; + const struct mbfl_convert_vtbl vtbl_utf16_wchar = { mbfl_no_encoding_utf16, mbfl_no_encoding_wchar, mbfl_filt_conv_common_ctor, NULL, mbfl_filt_conv_utf16_wchar, - mbfl_filt_conv_common_flush, + mbfl_filt_conv_utf16_wchar_flush, NULL, }; @@ -91,7 +114,7 @@ const struct mbfl_convert_vtbl vtbl_utf16be_wchar = { mbfl_filt_conv_common_ctor, NULL, mbfl_filt_conv_utf16be_wchar, - mbfl_filt_conv_common_flush, + mbfl_filt_conv_utf16_wchar_flush, NULL, }; @@ -111,7 +134,7 @@ const struct mbfl_convert_vtbl vtbl_utf16le_wchar = { mbfl_filt_conv_common_ctor, NULL, mbfl_filt_conv_utf16le_wchar, - mbfl_filt_conv_common_flush, + mbfl_filt_conv_utf16_wchar_flush, NULL, }; @@ -127,111 +150,89 @@ const struct mbfl_convert_vtbl vtbl_wchar_utf16le = { #define CK(statement) do { if ((statement) < 0) return (-1); } while (0) -/* - * UTF-16 => wchar - */ int mbfl_filt_conv_utf16_wchar(int c, mbfl_convert_filter *filter) { - int n, endian; - - endian = filter->status & 0xff00; - switch (filter->status & 0x0f) { - case 0: - if (endian) { - n = c & 0xff; - } else { - n = (c & 0xff) << 8; - } - filter->cache |= n; - filter->status++; - break; - default: - if (endian) { - n = (c & 0xff) << 8; + /* Start with the assumption that the string is big-endian; + * If we find a little-endian BOM, then we will change that assumption */ + if (filter->status == 0) { + filter->cache = c & 0xFF; + filter->status = 1; + } else { + int n = (filter->cache << 8) | (c & 0xFF); + if (n == 0xFFFE) { + /* Switch to little-endian mode */ + filter->filter_function = mbfl_filt_conv_utf16le_wchar; + filter->cache = filter->status = 0; } else { - n = c & 0xff; - } - n |= filter->cache & 0xffff; - filter->status &= ~0x0f; - if (n >= 0xd800 && n < 0xdc00) { - filter->cache = ((n & 0x3ff) << 16) + 0x400000; - } else if (n >= 0xdc00 && n < 0xe000) { - n &= 0x3ff; - n |= (filter->cache & 0xfff0000) >> 6; - filter->cache = 0; - if (n >= MBFL_WCSPLANE_SUPMIN && n < MBFL_WCSPLANE_SUPMAX) { - CK((*filter->output_function)(n, filter->data)); - } else { /* illegal character */ - n &= MBFL_WCSGROUP_MASK; - n |= MBFL_WCSGROUP_THROUGH; + filter->filter_function = mbfl_filt_conv_utf16be_wchar; + if (n >= 0xD800 && n <= 0xDBFF) { + filter->cache = n & 0x3FF; /* Pick out 10 data bits */ + filter->status = 2; + return c; + } else if (n >= 0xDC00 && n <= 0xDFFF) { + /* This is wrong; second part of surrogate pair has come first */ + CK((*filter->output_function)(n | MBFL_WCSGROUP_THROUGH, filter->data)); + } else if (n != 0xFEFF) { CK((*filter->output_function)(n, filter->data)); } - } else { - int is_first = filter->status & 0x10; - filter->cache = 0; - filter->status |= 0x10; - if (!is_first) { - if (n == 0xfffe) { - if (endian) { - filter->status &= ~0x100; /* big-endian */ - } else { - filter->status |= 0x100; /* little-endian */ - } - break; - } else if (n == 0xfeff) { - break; - } - } - CK((*filter->output_function)(n, filter->data)); + filter->cache = filter->status = 0; } - break; } return c; } -/* - * UTF-16BE => wchar - */ int mbfl_filt_conv_utf16be_wchar(int c, mbfl_convert_filter *filter) { int n; switch (filter->status) { - case 0: + case 0: /* First byte */ + filter->cache = c & 0xFF; filter->status = 1; - n = (c & 0xff) << 8; - filter->cache |= n; break; - default: - filter->status = 0; - n = (filter->cache & 0xff00) | (c & 0xff); - if (n >= 0xd800 && n < 0xdc00) { - filter->cache = ((n & 0x3ff) << 16) + 0x400000; - } else if (n >= 0xdc00 && n < 0xe000) { - n &= 0x3ff; - n |= (filter->cache & 0xfff0000) >> 6; - filter->cache = 0; - if (n >= MBFL_WCSPLANE_SUPMIN && n < MBFL_WCSPLANE_SUPMAX) { - CK((*filter->output_function)(n, filter->data)); - } else { /* illegal character */ - n &= MBFL_WCSGROUP_MASK; - n |= MBFL_WCSGROUP_THROUGH; - CK((*filter->output_function)(n, filter->data)); - } + + case 1: /* Second byte */ + n = (filter->cache << 8) | (c & 0xFF); + if (n >= 0xD800 && n <= 0xDBFF) { + filter->cache = n & 0x3FF; /* Pick out 10 data bits */ + filter->status = 2; + } else if (n >= 0xDC00 && n <= 0xDFFF) { + /* This is wrong; second part of surrogate pair has come first */ + CK((*filter->output_function)(n | MBFL_WCSGROUP_THROUGH, filter->data)); + filter->status = 0; } else { - filter->cache = 0; CK((*filter->output_function)(n, filter->data)); + filter->status = 0; } break; + + case 2: /* Second part of surrogate, first byte */ + filter->cache = (filter->cache << 8) | (c & 0xFF); + filter->status = 3; + break; + + case 3: /* Second part of surrogate, second byte */ + n = ((filter->cache & 0xFF) << 8) | (c & 0xFF); + if (n >= 0xD800 && n <= 0xDBFF) { + /* Wrong; that's the first half of a surrogate pair, not the second */ + CK((*filter->output_function)((0xD8 << 10) | (filter->cache >> 8) | MBFL_WCSGROUP_THROUGH, filter->data)); + filter->cache = n & 0x3FF; + filter->status = 2; + } else if (n >= 0xDC00 && n <= 0xDFFF) { + n = ((filter->cache & 0x3FF00) << 2) + (n & 0x3FF) + 0x10000; + CK((*filter->output_function)(n, filter->data)); + filter->status = 0; + } else { + CK((*filter->output_function)((0xD8 << 10) | (filter->cache >> 8) | MBFL_WCSGROUP_THROUGH, filter->data)); + CK((*filter->output_function)(n, filter->data)); + filter->status = 0; + } } return c; } -/* - * wchar => UTF-16BE - */ int mbfl_filt_conv_wchar_utf16be(int c, mbfl_convert_filter *filter) { int n; @@ -253,38 +254,53 @@ int mbfl_filt_conv_wchar_utf16be(int c, mbfl_convert_filter *filter) return c; } -/* - * UTF-16LE => wchar - */ int mbfl_filt_conv_utf16le_wchar(int c, mbfl_convert_filter *filter) { int n; switch (filter->status) { case 0: + filter->cache = c & 0xff; filter->status = 1; - n = c & 0xff; - filter->cache |= n; break; - default: - filter->status = 0; - n = (filter->cache & 0xff) | ((c & 0xff) << 8); - if (n >= 0xd800 && n < 0xdc00) { - filter->cache = ((n & 0x3ff) << 16) + 0x400000; - } else if (n >= 0xdc00 && n < 0xe000) { - n &= 0x3ff; - n |= (filter->cache & 0xfff0000) >> 6; - filter->cache = 0; - if (n >= MBFL_WCSPLANE_SUPMIN && n < MBFL_WCSPLANE_SUPMAX) { - CK((*filter->output_function)(n, filter->data)); - } else { /* illegal character */ - n &= MBFL_WCSGROUP_MASK; - n |= MBFL_WCSGROUP_THROUGH; - CK((*filter->output_function)(n, filter->data)); - } + + case 1: + if ((c & 0xfc) == 0xd8) { + /* Looks like we have a surrogate pair here */ + filter->cache += ((c & 0x3) << 8); + filter->status = 2; + } else if ((c & 0xfc) == 0xdc) { + /* This is wrong; the second part of the surrogate pair has come first + * Flag it with `MBFL_WCSGROUP_THROUGH`; the following filter will handle + * the error */ + n = (filter->cache + ((c & 0xff) << 8)) | MBFL_WCSGROUP_THROUGH; + CK((*filter->output_function)(n, filter->data)); + filter->status = 0; } else { - filter->cache = 0; + CK((*filter->output_function)(filter->cache + ((c & 0xff) << 8), filter->data)); + filter->status = 0; + } + break; + + case 2: + filter->cache = (filter->cache << 10) + (c & 0xff); + filter->status = 3; + break; + + case 3: + n = (filter->cache & 0xFF) | ((c & 0xFF) << 8); + if (n >= 0xD800 && n <= 0xDBFF) { + CK((*filter->output_function)((0xD8 << 10) | (filter->cache >> 10) | MBFL_WCSGROUP_THROUGH, filter->data)); + filter->cache = n & 0x3FF; + filter->status = 2; + } else if (n >= 0xDC00 && n <= 0xDFFF) { + n = filter->cache + ((c & 0x3) << 8) + 0x10000; CK((*filter->output_function)(n, filter->data)); + filter->status = 0; + } else { + CK((*filter->output_function)((0xD8 << 10) | (filter->cache >> 10) | MBFL_WCSGROUP_THROUGH, filter->data)); + CK((*filter->output_function)(n, filter->data)); + filter->status = 0; } break; } @@ -292,9 +308,6 @@ int mbfl_filt_conv_utf16le_wchar(int c, mbfl_convert_filter *filter) return c; } -/* - * wchar => UTF-16LE - */ int mbfl_filt_conv_wchar_utf16le(int c, mbfl_convert_filter *filter) { int n; @@ -315,3 +328,126 @@ int mbfl_filt_conv_wchar_utf16le(int c, mbfl_convert_filter *filter) return c; } + +static int mbfl_filt_conv_utf16_wchar_flush(mbfl_convert_filter *filter) +{ + int status = filter->status; + int cache = filter->cache; + filter->status = filter->cache = 0; + + if (status) { + /* Input string was truncated */ + CK((*filter->output_function)(cache | MBFL_WCSGROUP_THROUGH, filter->data)); + } + + if (filter->flush_function) { + (*filter->flush_function)(filter->data); + } + + return 0; +} + +static int mbfl_filt_ident_utf16(int c, mbfl_identify_filter *filter) +{ + if (filter->status == 0) { + if (c >= 0xfe) { /* could be a byte-order mark */ + filter->status = c; + } else { + /* no byte-order mark at beginning of input; assume UTF-16BE */ + filter->filter_function = mbfl_filt_ident_utf16be; + return (filter->filter_function)(c, filter); + } + } else { + unsigned short n = (filter->status << 8) | c; + filter->status = 0; + + if (n == 0xfeff) { + /* it was a big-endian byte-order mark */ + filter->filter_function = mbfl_filt_ident_utf16be; + } else if (n == 0xfffe) { + /* it was a little-endian byte-order mark */ + filter->filter_function = mbfl_filt_ident_utf16le; + } else { + /* it wasn't a byte-order mark */ + filter->filter_function = mbfl_filt_ident_utf16be; + (filter->filter_function)(n >> 8, filter); + return (filter->filter_function)(c, filter); + } + } + return c; +} + +static int mbfl_filt_ident_utf16le(int c, mbfl_identify_filter *filter) +{ + switch (filter->status) { + case 0: /* 1st byte */ + filter->status = 1; + break; + + case 1: /* 2nd byte */ + if ((c & 0xfc) == 0xd8) { + /* Looks like a surrogate pair */ + filter->status = 2; + } else if ((c & 0xfc) == 0xdc) { + /* This is wrong; the second part of the surrogate pair has come first */ + filter->flag = 1; + } else { + filter->status = 0; /* Just an ordinary 2-byte character */ + } + break; + + case 2: /* 3rd byte */ + filter->status = 3; + break; + + case 3: /* 4th byte */ + if ((c & 0xfc) == 0xdc) { + filter->status = 0; + } else { + filter->flag = 1; /* Surrogate pair wrongly encoded */ + } + break; + } + + return c; +} + +static int mbfl_filt_ident_utf16be(int c, mbfl_identify_filter *filter) +{ + switch (filter->status) { + case 0: /* 1st byte */ + if ((c & 0xfc) == 0xd8) { + /* Looks like a surrogate pair */ + filter->status = 2; + } else if ((c & 0xfc) == 0xdc) { + /* This is wrong; the second part of the surrogate pair has come first */ + filter->flag = 1; + } else { + /* Just an ordinary 2-byte character */ + filter->status = 1; + } + break; + + case 1: /* 2nd byte, not surrogate pair */ + filter->status = 0; + break; + + case 2: /* 2nd byte, surrogate pair */ + filter->status = 3; + break; + + case 3: /* 3rd byte, surrogate pair */ + if ((c & 0xfc) == 0xdc) { + filter->status = 4; + } else { + filter->flag = 1; /* Surrogate pair wrongly encoded */ + } + break; + + case 4: /* 4th byte, surrogate pair */ + filter->status = 0; + break; + } + + return c; +} diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf16.h b/ext/mbstring/libmbfl/filters/mbfilter_utf16.h index 4aa88619c7..601779add4 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_utf16.h +++ b/ext/mbstring/libmbfl/filters/mbfilter_utf16.h @@ -33,6 +33,9 @@ extern const mbfl_encoding mbfl_encoding_utf16; extern const mbfl_encoding mbfl_encoding_utf16be; extern const mbfl_encoding mbfl_encoding_utf16le; +extern const struct mbfl_identify_vtbl vtbl_identify_utf16; +extern const struct mbfl_identify_vtbl vtbl_identify_utf16le; +extern const struct mbfl_identify_vtbl vtbl_identify_utf16be; extern const struct mbfl_convert_vtbl vtbl_utf16_wchar; extern const struct mbfl_convert_vtbl vtbl_wchar_utf16; extern const struct mbfl_convert_vtbl vtbl_utf16be_wchar; diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf32.c b/ext/mbstring/libmbfl/filters/mbfilter_utf32.c index b936ab4102..1d9f479595 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_utf32.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_utf32.c @@ -30,13 +30,18 @@ #include "mbfilter.h" #include "mbfilter_utf32.h" +static int mbfl_filt_ident_utf32(int c, mbfl_identify_filter *filter); +static int mbfl_filt_ident_utf32le(int c, mbfl_identify_filter *filter); +static int mbfl_filt_ident_utf32be(int c, mbfl_identify_filter *filter); +static int mbfl_filt_conv_utf32_wchar_flush(mbfl_convert_filter *filter); + static const char *mbfl_encoding_utf32_aliases[] = {"utf32", NULL}; const mbfl_encoding mbfl_encoding_utf32 = { mbfl_no_encoding_utf32, "UTF-32", "UTF-32", - (const char *(*)[])&mbfl_encoding_utf32_aliases, + mbfl_encoding_utf32_aliases, NULL, MBFL_ENCTYPE_WCS4BE, &vtbl_utf32_wchar, @@ -65,13 +70,31 @@ const mbfl_encoding mbfl_encoding_utf32le = { &vtbl_wchar_utf32le }; +const struct mbfl_identify_vtbl vtbl_identify_utf32 = { + mbfl_no_encoding_utf32, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_utf32 +}; + +const struct mbfl_identify_vtbl vtbl_identify_utf32be = { + mbfl_no_encoding_utf32be, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_utf32be +}; + +const struct mbfl_identify_vtbl vtbl_identify_utf32le = { + mbfl_no_encoding_utf32le, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_utf32le +}; + const struct mbfl_convert_vtbl vtbl_utf32_wchar = { mbfl_no_encoding_utf32, mbfl_no_encoding_wchar, mbfl_filt_conv_common_ctor, NULL, mbfl_filt_conv_utf32_wchar, - mbfl_filt_conv_common_flush, + mbfl_filt_conv_utf32_wchar_flush, NULL, }; @@ -91,7 +114,7 @@ const struct mbfl_convert_vtbl vtbl_utf32be_wchar = { mbfl_filt_conv_common_ctor, NULL, mbfl_filt_conv_utf32be_wchar, - mbfl_filt_conv_common_flush, + mbfl_filt_conv_utf32_wchar_flush, NULL, }; @@ -111,7 +134,7 @@ const struct mbfl_convert_vtbl vtbl_utf32le_wchar = { mbfl_filt_conv_common_ctor, NULL, mbfl_filt_conv_utf32le_wchar, - mbfl_filt_conv_common_flush, + mbfl_filt_conv_utf32_wchar_flush, NULL, }; @@ -289,3 +312,143 @@ int mbfl_filt_conv_wchar_utf32le(int c, mbfl_convert_filter *filter) return c; } + +static int mbfl_filt_conv_utf32_wchar_flush(mbfl_convert_filter *filter) +{ + if (filter->status & 0xF) { + /* Input string was truncated */ + CK((*filter->output_function)(filter->cache | MBFL_WCSGROUP_THROUGH, filter->data)); + } + + if (filter->flush_function) { + (*filter->flush_function)(filter->data); + } + + filter->status = filter->cache = 0; + return 0; +} + +static int mbfl_filt_ident_utf32(int c, mbfl_identify_filter *filter) +{ + /* The largest valid codepoint is 0x10FFFF; we don't want values above that + * Neither do we want to see surrogates + * For UTF-32 (not LE or BE), we do also need to look for a byte-order mark */ + switch (filter->status) { + case 0: /* 1st byte */ + if (c == 0xff) { + filter->status = 1; + return c; + } + filter->filter_function = mbfl_filt_ident_utf32be; + break; + + case 1: /* 2nd byte */ + if (c == 0xfe) { + filter->status = 2; + return c; + } + filter->filter_function = mbfl_filt_ident_utf32be; + (filter->filter_function)(0xff, filter); + break; + + case 2: /* 3rd byte */ + if (c == 0) { + filter->status = 3; + return c; + } + filter->filter_function = mbfl_filt_ident_utf32be; + (filter->filter_function)(0xff, filter); + (filter->filter_function)(0xfe, filter); + break; + + case 3: /* 4th byte */ + if (c == 0) { + /* We found a little-endian byte-order mark! */ + filter->status = 0; + filter->filter_function = mbfl_filt_ident_utf32le; + return c; + } + filter->filter_function = mbfl_filt_ident_utf32be; + (filter->filter_function)(0xff, filter); + (filter->filter_function)(0xfe, filter); + (filter->filter_function)(0, filter); + break; + } + + return (filter->filter_function)(c, filter); +} + +static int mbfl_filt_ident_utf32le(int c, mbfl_identify_filter *filter) +{ + switch (filter->status) { + case 0: /* 1st byte */ + filter->status = 1; + break; + + case 1: /* 2nd byte */ + if (c >= 0xD8 && c <= 0xDF) { + filter->status = 4; /* might be surrogate if we are in BMP */ + } else { + filter->status = 2; + } + break; + + case 2: /* 3rd byte */ + if (c > 0x10) { + filter->flag = 1; /* too big */ + } + filter->status = 3; + break; + + case 3: /* 4th byte */ + if (c) { + filter->flag = 1; /* too big */ + } + filter->status = 0; + break; + + case 4: /* 3rd byte, previous byte looked like surrogate */ + if (!c) { + filter->flag = 1; /* yep, it's a surrogate */ + } + filter->status = 3; + } + return c; +} + +static int mbfl_filt_ident_utf32be(int c, mbfl_identify_filter *filter) +{ + switch (filter->status) { + case 0: /* 1st byte */ + if (c) { + filter->flag = 1; /* too big */ + } + filter->status = 1; + break; + + case 1: /* 2nd byte */ + if (c > 0x10) { + filter->flag = 1; /* too big */ + } if (c) { + filter->status = 4; /* not in the BMP */ + } else { + filter->status = 2; + } + break; + + case 2: /* 3rd byte */ + if (c >= 0xD8 && c <= 0xDF) { + filter->flag = 1; /* reserved range for surrogates */ + } + filter->status = 3; + break; + + case 3: /* 4th byte */ + filter->status = 0; + break; + + case 4: /* 3rd byte, not in BMP */ + filter->status = 3; + } + return c; +} diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf32.h b/ext/mbstring/libmbfl/filters/mbfilter_utf32.h index 14b9fa4823..de1235e451 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_utf32.h +++ b/ext/mbstring/libmbfl/filters/mbfilter_utf32.h @@ -33,6 +33,9 @@ extern const mbfl_encoding mbfl_encoding_utf32; extern const mbfl_encoding mbfl_encoding_utf32be; extern const mbfl_encoding mbfl_encoding_utf32le; +extern const struct mbfl_identify_vtbl vtbl_identify_utf32; +extern const struct mbfl_identify_vtbl vtbl_identify_utf32be; +extern const struct mbfl_identify_vtbl vtbl_identify_utf32le; extern const struct mbfl_convert_vtbl vtbl_utf32_wchar; extern const struct mbfl_convert_vtbl vtbl_wchar_utf32; extern const struct mbfl_convert_vtbl vtbl_utf32be_wchar; diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf7.c b/ext/mbstring/libmbfl/filters/mbfilter_utf7.c index b54bcf2b9c..ed931a8b1f 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_utf7.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_utf7.c @@ -51,7 +51,7 @@ const mbfl_encoding mbfl_encoding_utf7 = { mbfl_no_encoding_utf7, "UTF-7", "UTF-7", - (const char *(*)[])&mbfl_encoding_utf7_aliases, + mbfl_encoding_utf7_aliases, NULL, MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE, &vtbl_utf7_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf7imap.c b/ext/mbstring/libmbfl/filters/mbfilter_utf7imap.c index c8fe70fc7f..4c1cd8712f 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_utf7imap.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_utf7imap.c @@ -27,14 +27,64 @@ * */ +/* Modified UTF-7 used for 'international mailbox names' in the IMAP protocol + * Also known as mUTF-7 + * Defined in RFC 3501 5.1.3 (https://tools.ietf.org/html/rfc3501) + * + * Quoting from the RFC: + * + *********************************************************************** + * In modified UTF-7, printable US-ASCII characters, except for "&", + * represent themselves; that is, characters with octet values 0x20-0x25 + * and 0x27-0x7e. The character "&" (0x26) is represented by the + * two-octet sequence "&-". + * + * All other characters (octet values 0x00-0x1f and 0x7f-0xff) are + * represented in modified BASE64, with a further modification from + * UTF-7 that "," is used instead of "/". Modified BASE64 MUST NOT be + * used to represent any printing US-ASCII character which can represent + * itself. + * + * "&" is used to shift to modified BASE64 and "-" to shift back to + * US-ASCII. There is no implicit shift from BASE64 to US-ASCII, and + * null shifts ("-&" while in BASE64; note that "&-" while in US-ASCII + * means "&") are not permitted. However, all names start in US-ASCII, + * and MUST end in US-ASCII; that is, a name that ends with a non-ASCII + * ISO-10646 character MUST end with a "-"). + *********************************************************************** + * + * The purpose of all this is: 1) to keep all parts of IMAP messages 7-bit clean, + * 2) to avoid giving special treatment to +, /, \, and ~, since these are + * commonly used in mailbox names, and 3) to ensure there is only one + * representation of any mailbox name (vanilla UTF-7 does allow multiple + * representations of the same string, by Base64-encoding characters which + * could have been included as ASCII literals.) + * + * RFC 2152 also applies, since it defines vanilla UTF-7 (minus IMAP modifications) + * The following paragraph is notable: + * + *********************************************************************** + * Unicode is encoded using Modified Base64 by first converting Unicode + * 16-bit quantities to an octet stream (with the most significant octet first). + * Surrogate pairs (UTF-16) are converted by treating each half of the pair as + * a separate 16 bit quantity (i.e., no special treatment). Text with an odd + * number of octets is ill-formed. ISO 10646 characters outside the range + * addressable via surrogate pairs cannot be encoded. + *********************************************************************** + * + * So after reversing the modified Base64 encoding on an encoded section, + * the contents are interpreted as UTF-16BE. */ + #include "mbfilter.h" #include "mbfilter_utf7imap.h" +static const char *mbfl_encoding_utf7imap_aliases[] = {"mUTF-7", NULL}; + const mbfl_encoding mbfl_encoding_utf7imap = { mbfl_no_encoding_utf7imap, "UTF7-IMAP", NULL, - NULL, + mbfl_encoding_utf7imap_aliases, NULL, MBFL_ENCTYPE_MBCS, &vtbl_utf7imap_wchar, @@ -106,7 +156,7 @@ int mbfl_filt_conv_utf7imap_wchar(int c, mbfl_convert_filter *filter) case 0: if (c == 0x26) { /* '&' shift character */ filter->status++; - } else if (c >= 0 && c < 0x80) { /* ASCII */ + } else if (c >= 0x20 && c <= 0x7E) { /* ASCII */ CK((*filter->output_function)(c, filter->data)); } else { /* illegal character */ s = c & MBFL_WCSGROUP_MASK; @@ -145,7 +195,15 @@ int mbfl_filt_conv_utf7imap_wchar(int c, mbfl_convert_filter *filter) } } else { filter->cache = n; - CK((*filter->output_function)(s, filter->data)); + /* Characters which can be expressed as literal, ASCII characters + * should not be Base64-encoded */ + if (s < 0x20 || s > 0x7E || s == '&') { + CK((*filter->output_function)(s, filter->data)); + } else { + s &= MBFL_WCSGROUP_MASK; + s |= MBFL_WCSGROUP_THROUGH; + CK((*filter->output_function)(s, filter->data)); + } } break; @@ -177,7 +235,15 @@ int mbfl_filt_conv_utf7imap_wchar(int c, mbfl_convert_filter *filter) } } else { filter->cache = n; - CK((*filter->output_function)(s, filter->data)); + /* Characters which can be expressed as literal, ASCII characters + * should not be Base64-encoded */ + if (s < 0x20 || s > 0x7E || s == '&') { + CK((*filter->output_function)(s, filter->data)); + } else { + s &= MBFL_WCSGROUP_MASK; + s |= MBFL_WCSGROUP_THROUGH; + CK((*filter->output_function)(s, filter->data)); + } } break; @@ -204,7 +270,15 @@ int mbfl_filt_conv_utf7imap_wchar(int c, mbfl_convert_filter *filter) } } else { filter->cache = 0; - CK((*filter->output_function)(s, filter->data)); + /* Characters which can be expressed as literal, ASCII characters + * should not be Base64-encoded */ + if (s < 0x20 || s > 0x7E || s == '&') { + CK((*filter->output_function)(s, filter->data)); + } else { + s &= MBFL_WCSGROUP_MASK; + s |= MBFL_WCSGROUP_THROUGH; + CK((*filter->output_function)(s, filter->data)); + } } break; diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf8.c b/ext/mbstring/libmbfl/filters/mbfilter_utf8.c index 9e97d5b069..3b7a76de47 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_utf8.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_utf8.c @@ -57,7 +57,7 @@ const mbfl_encoding mbfl_encoding_utf8 = { mbfl_no_encoding_utf8, "UTF-8", "UTF-8", - (const char *(*)[])&mbfl_encoding_utf8_aliases, + mbfl_encoding_utf8_aliases, mblen_table_utf8, MBFL_ENCTYPE_MBCS, &vtbl_utf8_wchar, diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf8_mobile.c b/ext/mbstring/libmbfl/filters/mbfilter_utf8_mobile.c index ef8d4cc075..8f50d22781 100644 --- a/ext/mbstring/libmbfl/filters/mbfilter_utf8_mobile.c +++ b/ext/mbstring/libmbfl/filters/mbfilter_utf8_mobile.c @@ -46,7 +46,7 @@ const mbfl_encoding mbfl_encoding_utf8_docomo = { mbfl_no_encoding_utf8_docomo, "UTF-8-Mobile#DOCOMO", "UTF-8", - (const char *(*)[])&mbfl_encoding_utf8_docomo_aliases, + mbfl_encoding_utf8_docomo_aliases, mblen_table_utf8, MBFL_ENCTYPE_MBCS, &vtbl_utf8_docomo_wchar, @@ -57,7 +57,7 @@ const mbfl_encoding mbfl_encoding_utf8_kddi_a = { mbfl_no_encoding_utf8_kddi_a, "UTF-8-Mobile#KDDI-A", "UTF-8", - (const char *(*)[])&mbfl_encoding_utf8_kddi_a_aliases, + mbfl_encoding_utf8_kddi_a_aliases, mblen_table_utf8, MBFL_ENCTYPE_MBCS, &vtbl_utf8_kddi_a_wchar, @@ -68,7 +68,7 @@ const mbfl_encoding mbfl_encoding_utf8_kddi_b = { mbfl_no_encoding_utf8_kddi_b, "UTF-8-Mobile#KDDI-B", "UTF-8", - (const char *(*)[])&mbfl_encoding_utf8_kddi_b_aliases, + mbfl_encoding_utf8_kddi_b_aliases, mblen_table_utf8, MBFL_ENCTYPE_MBCS, &vtbl_utf8_kddi_b_wchar, @@ -79,7 +79,7 @@ const mbfl_encoding mbfl_encoding_utf8_sb = { mbfl_no_encoding_utf8_sb, "UTF-8-Mobile#SOFTBANK", "UTF-8", - (const char *(*)[])&mbfl_encoding_utf8_sb_aliases, + mbfl_encoding_utf8_sb_aliases, mblen_table_utf8, MBFL_ENCTYPE_MBCS, &vtbl_utf8_sb_wchar, diff --git a/ext/mbstring/libmbfl/filters/unicode_prop.h b/ext/mbstring/libmbfl/filters/unicode_prop.h index 6eaefbc2e3..1ecf42c556 100644 --- a/ext/mbstring/libmbfl/filters/unicode_prop.h +++ b/ext/mbstring/libmbfl/filters/unicode_prop.h @@ -19,148 +19,136 @@ * Suite 330, Boston, MA 02111-1307 USA * * The authors of this file: PHP3 Internationalization team - * - */ - -/* character property table */ -#define MBFL_CHP_CTL 0x01 -#define MBFL_CHP_DIGIT 0x02 -#define MBFL_CHP_UALPHA 0x04 -#define MBFL_CHP_LALPHA 0x08 -#define MBFL_CHP_MMHQENC 0x10 /* must Q-encoding in MIME Header encoded-word */ -#define MBFL_CHP_MSPECIAL 0x20 /* RFC822 Special characters */ - -/* - * Unicode table */ -static const unsigned char mbfl_charprop_table[] = { -/* NUL 0 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC, -/* SCH 1 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* SIX 2 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* EIX 3 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* EOT 4 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* ENQ 5 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* ACK 6 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* BEL 7 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* BS 8 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* HI 9 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* LF 10 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* VI 11 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* FF 12 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* CR 13 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* SO 14 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* SI 15 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* SLE 16 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* CSI 17 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* DC2 18 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* DC3 19 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* DC4 20 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* NAK 21 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* SYN 22 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* EIB 23 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* CAN 24 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* EM 25 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* SLB 26 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* ESC 27 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* FS 28 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* GS 29 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* RS 30 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* US 31 */ MBFL_CHP_CTL | MBFL_CHP_MMHQENC , -/* SP 32 */ MBFL_CHP_MMHQENC , -/* ! 33 */ 0 , -/* " 34 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* # 35 */ MBFL_CHP_MMHQENC , -/* $ 36 */ MBFL_CHP_MMHQENC , -/* % 37 */ MBFL_CHP_MMHQENC , -/* & 38 */ MBFL_CHP_MMHQENC , -/* ' 39 */ MBFL_CHP_MMHQENC , -/* ( 40 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* ) 41 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* * 42 */ 0 , -/* + 43 */ 0 , -/* , 44 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* - 45 */ 0 , -/* . 46 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* / 47 */ 0 , -/* 0 48 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* 1 49 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* 2 50 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* 3 51 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* 4 52 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* 5 53 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* 6 54 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* 7 55 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* 8 56 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* 9 57 */ MBFL_CHP_DIGIT | MBFL_CHP_MMHQENC , -/* : 58 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* ; 59 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* < 60 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* = 61 */ 0 , -/* > 62 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* ? 63 */ MBFL_CHP_MMHQENC , -/* @ 64 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* A 65 */ MBFL_CHP_UALPHA , -/* B 66 */ MBFL_CHP_UALPHA , -/* C 67 */ MBFL_CHP_UALPHA , -/* D 68 */ MBFL_CHP_UALPHA , -/* E 69 */ MBFL_CHP_UALPHA , -/* F 70 */ MBFL_CHP_UALPHA , -/* G 71 */ MBFL_CHP_UALPHA , -/* H 72 */ MBFL_CHP_UALPHA , -/* I 73 */ MBFL_CHP_UALPHA , -/* J 74 */ MBFL_CHP_UALPHA , -/* K 75 */ MBFL_CHP_UALPHA , -/* L 76 */ MBFL_CHP_UALPHA , -/* M 77 */ MBFL_CHP_UALPHA , -/* N 78 */ MBFL_CHP_UALPHA , -/* O 79 */ MBFL_CHP_UALPHA , -/* P 80 */ MBFL_CHP_UALPHA , -/* Q 81 */ MBFL_CHP_UALPHA , -/* R 82 */ MBFL_CHP_UALPHA , -/* S 83 */ MBFL_CHP_UALPHA , -/* T 84 */ MBFL_CHP_UALPHA , -/* U 85 */ MBFL_CHP_UALPHA , -/* V 86 */ MBFL_CHP_UALPHA , -/* W 87 */ MBFL_CHP_UALPHA , -/* X 88 */ MBFL_CHP_UALPHA , -/* Y 89 */ MBFL_CHP_UALPHA , -/* Z 90 */ MBFL_CHP_UALPHA , -/* [ 91 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* \ 92 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* ] 93 */ MBFL_CHP_MMHQENC | MBFL_CHP_MSPECIAL , -/* ^ 94 */ MBFL_CHP_MMHQENC , -/* _ 95 */ MBFL_CHP_MMHQENC , -/* ` 96 */ MBFL_CHP_MMHQENC , -/* a 97 */ MBFL_CHP_LALPHA , -/* b 98 */ MBFL_CHP_LALPHA , -/* c 99 */ MBFL_CHP_LALPHA , -/* d 100 */ MBFL_CHP_LALPHA , -/* e 101 */ MBFL_CHP_LALPHA , -/* f 102 */ MBFL_CHP_LALPHA , -/* g 103 */ MBFL_CHP_LALPHA , -/* h 104 */ MBFL_CHP_LALPHA , -/* i 105 */ MBFL_CHP_LALPHA , -/* j 106 */ MBFL_CHP_LALPHA , -/* k 107 */ MBFL_CHP_LALPHA , -/* l 108 */ MBFL_CHP_LALPHA , -/* m 109 */ MBFL_CHP_LALPHA , -/* n 110 */ MBFL_CHP_LALPHA , -/* o 111 */ MBFL_CHP_LALPHA , -/* p 112 */ MBFL_CHP_LALPHA , -/* q 113 */ MBFL_CHP_LALPHA , -/* r 114 */ MBFL_CHP_LALPHA , -/* s 115 */ MBFL_CHP_LALPHA , -/* t 116 */ MBFL_CHP_LALPHA , -/* u 117 */ MBFL_CHP_LALPHA , -/* v 118 */ MBFL_CHP_LALPHA , -/* w 119 */ MBFL_CHP_LALPHA , -/* x 120 */ MBFL_CHP_LALPHA , -/* y 121 */ MBFL_CHP_LALPHA , -/* z 122 */ MBFL_CHP_LALPHA , -/* { 123 */ MBFL_CHP_MMHQENC , -/* | 124 */ MBFL_CHP_MMHQENC , -/* } 125 */ MBFL_CHP_MMHQENC , -/* ~ 126 */ MBFL_CHP_MMHQENC , -/* DEL 127 */ MBFL_CHP_MMHQENC +/* Should we use QPrint-encoding in MIME Header encoded-word? */ +static const unsigned char mime_char_needs_qencode[] = { +/* NUL 0 */ 1, +/* SCH 1 */ 1, +/* SIX 2 */ 1, +/* EIX 3 */ 1, +/* EOT 4 */ 1, +/* ENQ 5 */ 1, +/* ACK 6 */ 1, +/* BEL 7 */ 1, +/* BS 8 */ 1, +/* HI 9 */ 1, +/* LF 10 */ 1, +/* VI 11 */ 1, +/* FF 12 */ 1, +/* CR 13 */ 1, +/* SO 14 */ 1, +/* SI 15 */ 1, +/* SLE 16 */ 1, +/* CSI 17 */ 1, +/* DC2 18 */ 1, +/* DC3 19 */ 1, +/* DC4 20 */ 1, +/* NAK 21 */ 1, +/* SYN 22 */ 1, +/* EIB 23 */ 1, +/* CAN 24 */ 1, +/* EM 25 */ 1, +/* SLB 26 */ 1, +/* ESC 27 */ 1, +/* FS 28 */ 1, +/* GS 29 */ 1, +/* RS 30 */ 1, +/* US 31 */ 1, +/* SP 32 */ 1, +/* ! 33 */ 0, +/* " 34 */ 1, +/* # 35 */ 1, +/* $ 36 */ 1, +/* % 37 */ 1, +/* & 38 */ 1, +/* ' 39 */ 1, +/* ( 40 */ 1, +/* ) 41 */ 1, +/* * 42 */ 0, +/* + 43 */ 0, +/* , 44 */ 1, +/* - 45 */ 0, +/* . 46 */ 1, +/* / 47 */ 0, +/* 0 48 */ 1, +/* 1 49 */ 1, +/* 2 50 */ 1, +/* 3 51 */ 1, +/* 4 52 */ 1, +/* 5 53 */ 1, +/* 6 54 */ 1, +/* 7 55 */ 1, +/* 8 56 */ 1, +/* 9 57 */ 1, +/* : 58 */ 1, +/* ; 59 */ 1, +/* < 60 */ 1, +/* = 61 */ 0, +/* > 62 */ 1, +/* ? 63 */ 1, +/* @ 64 */ 1, +/* A 65 */ 0, +/* B 66 */ 0, +/* C 67 */ 0, +/* D 68 */ 0, +/* E 69 */ 0, +/* F 70 */ 0, +/* G 71 */ 0, +/* H 72 */ 0, +/* I 73 */ 0, +/* J 74 */ 0, +/* K 75 */ 0, +/* L 76 */ 0, +/* M 77 */ 0, +/* N 78 */ 0, +/* O 79 */ 0, +/* P 80 */ 0, +/* Q 81 */ 0, +/* R 82 */ 0, +/* S 83 */ 0, +/* T 84 */ 0, +/* U 85 */ 0, +/* V 86 */ 0, +/* W 87 */ 0, +/* X 88 */ 0, +/* Y 89 */ 0, +/* Z 90 */ 0, +/* [ 91 */ 1, +/* \ 92 */ 1, +/* ] 93 */ 1, +/* ^ 94 */ 1, +/* _ 95 */ 1, +/* ` 96 */ 1, +/* a 97 */ 0, +/* b 98 */ 0, +/* c 99 */ 0, +/* d 100 */ 0, +/* e 101 */ 0, +/* f 102 */ 0, +/* g 103 */ 0, +/* h 104 */ 0, +/* i 105 */ 0, +/* j 106 */ 0, +/* k 107 */ 0, +/* l 108 */ 0, +/* m 109 */ 0, +/* n 110 */ 0, +/* o 111 */ 0, +/* p 112 */ 0, +/* q 113 */ 0, +/* r 114 */ 0, +/* s 115 */ 0, +/* t 116 */ 0, +/* u 117 */ 0, +/* v 118 */ 0, +/* w 119 */ 0, +/* x 120 */ 0, +/* y 121 */ 0, +/* z 122 */ 0, +/* { 123 */ 1, +/* | 124 */ 1, +/* } 125 */ 1, +/* ~ 126 */ 1, +/* DEL 127 */ 1 }; diff --git a/ext/mbstring/libmbfl/mbfl/mbfilter.c b/ext/mbstring/libmbfl/mbfl/mbfilter.c index 793dd8e078..68cb39fa90 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfilter.c +++ b/ext/mbstring/libmbfl/mbfl/mbfilter.c @@ -200,7 +200,6 @@ size_t mbfl_buffer_converter_feed(mbfl_buffer_converter *convd, mbfl_string *str size_t n; unsigned char *p; mbfl_convert_filter *filter; - int (*filter_function)(int c, mbfl_convert_filter *filter); ZEND_ASSERT(convd); ZEND_ASSERT(string); @@ -212,9 +211,8 @@ size_t mbfl_buffer_converter_feed(mbfl_buffer_converter *convd, mbfl_string *str filter = convd->filter1; if (filter != NULL) { - filter_function = filter->filter_function; while (n > 0) { - if ((*filter_function)(*p++, filter) < 0) { + if ((*filter->filter_function)(*p++, filter) < 0) { return p - string->val; } n--; @@ -396,18 +394,6 @@ const mbfl_encoding *mbfl_encoding_detector_judge(mbfl_encoding_detector *identd } n--; } - - /* fallback judge */ - if (!encoding) { - n = identd->filter_list_size - 1; - while (n >= 0) { - filter = identd->filter_list[n]; - if (!filter->flag) { - encoding = filter->encoding; - } - n--; - } - } } return encoding; @@ -486,7 +472,7 @@ mbfl_convert_encoding( const mbfl_encoding * mbfl_identify_encoding(mbfl_string *string, const mbfl_encoding **elist, int elistsz, int strict) { - int i, num, bad; + int i, bad; size_t n; unsigned char *p; mbfl_identify_filter *flist, *filter; @@ -495,12 +481,9 @@ mbfl_identify_encoding(mbfl_string *string, const mbfl_encoding **elist, int eli /* flist is an array of mbfl_identify_filter instances */ flist = ecalloc(elistsz, sizeof(mbfl_identify_filter)); - num = 0; if (elist != NULL) { for (i = 0; i < elistsz; i++) { - if (!mbfl_identify_filter_init2(&flist[num], elist[i])) { - num++; - } + mbfl_identify_filter_init2(&flist[i], elist[i]); } } @@ -511,7 +494,7 @@ mbfl_identify_encoding(mbfl_string *string, const mbfl_encoding **elist, int eli if (p != NULL) { bad = 0; while (n > 0) { - for (i = 0; i < num; i++) { + for (i = 0; i < elistsz; i++) { filter = &flist[i]; if (!filter->flag) { (*filter->filter_function)(*p, filter); @@ -520,7 +503,7 @@ mbfl_identify_encoding(mbfl_string *string, const mbfl_encoding **elist, int eli } } } - if ((num - 1) <= bad && !strict) { + if ((elistsz - 1) <= bad && !strict) { break; } p++; @@ -531,7 +514,7 @@ mbfl_identify_encoding(mbfl_string *string, const mbfl_encoding **elist, int eli /* judge */ encoding = NULL; - for (i = 0; i < num; i++) { + for (i = 0; i < elistsz; i++) { filter = &flist[i]; if (!filter->flag) { if (strict && filter->status) { @@ -544,7 +527,7 @@ mbfl_identify_encoding(mbfl_string *string, const mbfl_encoding **elist, int eli /* fall-back judge */ if (!encoding) { - for (i = 0; i < num; i++) { + for (i = 0; i < elistsz; i++) { filter = &flist[i]; if (!filter->flag && (!strict || !filter->status)) { encoding = filter->encoding; diff --git a/ext/mbstring/libmbfl/mbfl/mbfilter_8bit.c b/ext/mbstring/libmbfl/mbfl/mbfilter_8bit.c index f4c74946c1..41081aa883 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfilter_8bit.c +++ b/ext/mbstring/libmbfl/mbfl/mbfilter_8bit.c @@ -43,13 +43,19 @@ const mbfl_encoding mbfl_encoding_8bit = { mbfl_no_encoding_8bit, "8bit", "8bit", - (const char *(*)[])&mbfl_encoding_8bit_aliases, + mbfl_encoding_8bit_aliases, NULL, MBFL_ENCTYPE_SBCS, &vtbl_8bit_wchar, &vtbl_wchar_8bit }; +const struct mbfl_identify_vtbl vtbl_identify_8bit = { + mbfl_no_encoding_8bit, + mbfl_filt_ident_common_ctor, + mbfl_filt_ident_true +}; + const struct mbfl_convert_vtbl vtbl_8bit_wchar = { mbfl_no_encoding_8bit, mbfl_no_encoding_wchar, diff --git a/ext/mbstring/libmbfl/mbfl/mbfilter_8bit.h b/ext/mbstring/libmbfl/mbfl/mbfilter_8bit.h index acdf640fd7..64369b999d 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfilter_8bit.h +++ b/ext/mbstring/libmbfl/mbfl/mbfilter_8bit.h @@ -35,5 +35,6 @@ #include "mbfilter.h" MBFLAPI extern const mbfl_encoding mbfl_encoding_8bit; +extern const struct mbfl_identify_vtbl vtbl_identify_8bit; #endif /* MBFL_MBFILTER_8BIT_H */ diff --git a/ext/mbstring/libmbfl/mbfl/mbfilter_pass.c b/ext/mbstring/libmbfl/mbfl/mbfilter_pass.c index 3d7cffe9c1..e43f746ecc 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfilter_pass.c +++ b/ext/mbstring/libmbfl/mbfl/mbfilter_pass.c @@ -38,7 +38,7 @@ const mbfl_encoding mbfl_encoding_pass = { mbfl_no_encoding_pass, "pass", NULL, - (const char *(*)[])&mbfl_encoding_pass_aliases, + mbfl_encoding_pass_aliases, NULL, 0, NULL, diff --git a/ext/mbstring/libmbfl/mbfl/mbfl_convert.c b/ext/mbstring/libmbfl/mbfl/mbfl_convert.c index 6a46c2b541..af6a3bba93 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfl_convert.c +++ b/ext/mbstring/libmbfl/mbfl/mbfl_convert.c @@ -136,7 +136,7 @@ static void mbfl_convert_filter_common_init(mbfl_convert_filter *filter, const m filter->filter_ctor = vtbl->filter_ctor; filter->filter_dtor = vtbl->filter_dtor; filter->filter_function = vtbl->filter_function; - filter->filter_flush = vtbl->filter_flush; + filter->filter_flush = (filter_flush_t)vtbl->filter_flush; filter->filter_copy = vtbl->filter_copy; (*filter->filter_ctor)(filter); diff --git a/ext/mbstring/libmbfl/mbfl/mbfl_encoding.c b/ext/mbstring/libmbfl/mbfl/mbfl_encoding.c index 3ecb06553d..f740f7aec1 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfl_encoding.c +++ b/ext/mbstring/libmbfl/mbfl/mbfl_encoding.c @@ -199,44 +199,32 @@ static const mbfl_encoding *mbfl_encoding_ptr_list[] = { NULL }; -/* encoding resolver */ -const mbfl_encoding * -mbfl_name2encoding(const char *name) +const mbfl_encoding *mbfl_name2encoding(const char *name) { - const mbfl_encoding *encoding; - int i, j; + const mbfl_encoding **encoding; - if (name == NULL) { - return NULL; - } - - i = 0; - while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL){ - if (strcasecmp(encoding->name, name) == 0) { - return encoding; + for (encoding = mbfl_encoding_ptr_list; *encoding; encoding++) { + if (strcasecmp((*encoding)->name, name) == 0) { + return *encoding; } } - /* serch MIME charset name */ - i = 0; - while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL) { - if (encoding->mime_name != NULL) { - if (strcasecmp(encoding->mime_name, name) == 0) { - return encoding; + /* search MIME charset name */ + for (encoding = mbfl_encoding_ptr_list; *encoding; encoding++) { + if ((*encoding)->mime_name) { + if (strcasecmp((*encoding)->mime_name, name) == 0) { + return *encoding; } } } - /* serch aliases */ - i = 0; - while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL) { - if (encoding->aliases != NULL) { - j = 0; - while ((*encoding->aliases)[j] != NULL) { - if (strcasecmp((*encoding->aliases)[j], name) == 0) { - return encoding; + /* search aliases */ + for (encoding = mbfl_encoding_ptr_list; *encoding; encoding++) { + if ((*encoding)->aliases) { + for (const char **alias = (*encoding)->aliases; *alias; alias++) { + if (strcasecmp(*alias, name) == 0) { + return *encoding; } - j++; } } } @@ -244,63 +232,45 @@ mbfl_name2encoding(const char *name) return NULL; } -const mbfl_encoding * -mbfl_no2encoding(enum mbfl_no_encoding no_encoding) +const mbfl_encoding *mbfl_no2encoding(enum mbfl_no_encoding no_encoding) { - const mbfl_encoding *encoding; - int i; + const mbfl_encoding **encoding; - i = 0; - while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL){ - if (encoding->no_encoding == no_encoding) { - return encoding; + for (encoding = mbfl_encoding_ptr_list; *encoding; encoding++) { + if ((*encoding)->no_encoding == no_encoding) { + return *encoding; } } return NULL; } -enum mbfl_no_encoding -mbfl_name2no_encoding(const char *name) +enum mbfl_no_encoding mbfl_name2no_encoding(const char *name) { - const mbfl_encoding *encoding; - - encoding = mbfl_name2encoding(name); - if (encoding == NULL) { - return mbfl_no_encoding_invalid; - } else { - return encoding->no_encoding; - } + const mbfl_encoding *encoding = mbfl_name2encoding(name); + return encoding ? encoding->no_encoding : mbfl_no_encoding_invalid; } -const char * -mbfl_no_encoding2name(enum mbfl_no_encoding no_encoding) +const char *mbfl_no_encoding2name(enum mbfl_no_encoding no_encoding) { - const mbfl_encoding *encoding; - - encoding = mbfl_no2encoding(no_encoding); - if (encoding == NULL) { - return ""; - } else { - return encoding->name; - } + const mbfl_encoding *encoding = mbfl_no2encoding(no_encoding); + return encoding ? encoding->name : ""; } -const mbfl_encoding ** -mbfl_get_supported_encodings(void) +const mbfl_encoding **mbfl_get_supported_encodings(void) { return mbfl_encoding_ptr_list; } -const char * -mbfl_no2preferred_mime_name(enum mbfl_no_encoding no_encoding) +const char *mbfl_no2preferred_mime_name(enum mbfl_no_encoding no_encoding) { - const mbfl_encoding *encoding; + return mbfl_encoding_preferred_mime_name(mbfl_no2encoding(no_encoding)); +} - encoding = mbfl_no2encoding(no_encoding); - if (encoding != NULL && encoding->mime_name != NULL && encoding->mime_name[0] != '\0') { +const char *mbfl_encoding_preferred_mime_name(const mbfl_encoding *encoding) +{ + if (encoding->mime_name && encoding->mime_name[0] != '\0') { return encoding->mime_name; - } else { - return NULL; } + return NULL; } diff --git a/ext/mbstring/libmbfl/mbfl/mbfl_encoding.h b/ext/mbstring/libmbfl/mbfl/mbfl_encoding.h index 0b39153c9d..f945b5dd03 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfl_encoding.h +++ b/ext/mbstring/libmbfl/mbfl/mbfl_encoding.h @@ -145,19 +145,19 @@ typedef struct _mbfl_encoding { enum mbfl_no_encoding no_encoding; const char *name; const char *mime_name; - const char *(*aliases)[]; + const char **aliases; const unsigned char *mblen_table; unsigned int flag; const struct mbfl_convert_vtbl *input_filter; const struct mbfl_convert_vtbl *output_filter; } mbfl_encoding; -MBFLAPI extern const mbfl_encoding * mbfl_name2encoding(const char *name); -MBFLAPI extern const mbfl_encoding * mbfl_no2encoding(enum mbfl_no_encoding no_encoding); +MBFLAPI extern const mbfl_encoding *mbfl_name2encoding(const char *name); +MBFLAPI extern const mbfl_encoding *mbfl_no2encoding(enum mbfl_no_encoding no_encoding); MBFLAPI extern enum mbfl_no_encoding mbfl_name2no_encoding(const char *name); -MBFLAPI extern const mbfl_encoding ** mbfl_get_supported_encodings(); -MBFLAPI extern const char * mbfl_no_encoding2name(enum mbfl_no_encoding no_encoding); -MBFLAPI extern const char * mbfl_no2preferred_mime_name(enum mbfl_no_encoding no_encoding); - +MBFLAPI extern const mbfl_encoding **mbfl_get_supported_encodings(); +MBFLAPI extern const char *mbfl_no_encoding2name(enum mbfl_no_encoding no_encoding); +MBFLAPI extern const char *mbfl_no2preferred_mime_name(enum mbfl_no_encoding no_encoding); +MBFLAPI extern const char *mbfl_encoding_preferred_mime_name(const mbfl_encoding *encoding); #endif /* MBFL_ENCODING_H */ diff --git a/ext/mbstring/libmbfl/mbfl/mbfl_ident.c b/ext/mbstring/libmbfl/mbfl/mbfl_ident.c index 18144eb030..a48762ee3d 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfl_ident.c +++ b/ext/mbstring/libmbfl/mbfl/mbfl_ident.c @@ -78,6 +78,7 @@ #include "filters/mbfilter_iso8859_13.h" #include "filters/mbfilter_iso8859_14.h" #include "filters/mbfilter_iso8859_15.h" +#include "filters/mbfilter_iso8859_16.h" #include "filters/mbfilter_base64.h" #include "filters/mbfilter_qprint.h" #include "filters/mbfilter_uuencode.h" @@ -152,6 +153,7 @@ static const struct mbfl_identify_vtbl *mbfl_identify_filter_list[] = { &vtbl_identify_8859_13, &vtbl_identify_8859_14, &vtbl_identify_8859_15, + &vtbl_identify_8859_16, &vtbl_identify_armscii8, &vtbl_identify_cp850, &vtbl_identify_jis_ms, @@ -159,97 +161,82 @@ static const struct mbfl_identify_vtbl *mbfl_identify_filter_list[] = { &vtbl_identify_cp50221, &vtbl_identify_cp50222, &vtbl_identify_gb18030, + &vtbl_identify_7bit, + &vtbl_identify_utf16, + &vtbl_identify_utf16le, + &vtbl_identify_utf16be, + &vtbl_identify_8bit, + &vtbl_identify_ucs2, + &vtbl_identify_ucs2be, + &vtbl_identify_ucs2le, + &vtbl_identify_utf32, + &vtbl_identify_utf32be, + &vtbl_identify_utf32le, &vtbl_identify_false, NULL }; -/* - * identify filter - */ -const struct mbfl_identify_vtbl * mbfl_identify_filter_get_vtbl(enum mbfl_no_encoding encoding) +const struct mbfl_identify_vtbl* mbfl_identify_filter_get_vtbl(enum mbfl_no_encoding encoding) { - const struct mbfl_identify_vtbl * vtbl; - int i; + const struct mbfl_identify_vtbl** vtbl; - i = 0; - while ((vtbl = mbfl_identify_filter_list[i++]) != NULL) { - if (vtbl->encoding == encoding) { - break; + for (vtbl = mbfl_identify_filter_list; *vtbl; vtbl++) { + if ((*vtbl)->encoding == encoding) { + return *vtbl; } } - return vtbl; + return NULL; } mbfl_identify_filter *mbfl_identify_filter_new(enum mbfl_no_encoding encoding) { mbfl_identify_filter *filter = emalloc(sizeof(mbfl_identify_filter)); - if (mbfl_identify_filter_init(filter, encoding)) { - efree(filter); - return NULL; - } - + mbfl_identify_filter_init(filter, encoding); return filter; } mbfl_identify_filter *mbfl_identify_filter_new2(const mbfl_encoding *encoding) { mbfl_identify_filter *filter = emalloc(sizeof(mbfl_identify_filter)); - if (mbfl_identify_filter_init2(filter, encoding)) { - efree(filter); - return NULL; - } - + mbfl_identify_filter_init2(filter, encoding); return filter; } - -int mbfl_identify_filter_init(mbfl_identify_filter *filter, enum mbfl_no_encoding encoding) +void mbfl_identify_filter_init(mbfl_identify_filter *filter, enum mbfl_no_encoding encoding) { const mbfl_encoding *enc = mbfl_no2encoding(encoding); - return mbfl_identify_filter_init2(filter, enc ? enc: &mbfl_encoding_pass); + mbfl_identify_filter_init2(filter, enc ? enc : &mbfl_encoding_pass); } -int mbfl_identify_filter_init2(mbfl_identify_filter *filter, const mbfl_encoding *encoding) +void mbfl_identify_filter_init2(mbfl_identify_filter *filter, const mbfl_encoding *encoding) { - const struct mbfl_identify_vtbl *vtbl; - - /* encoding structure */ filter->encoding = encoding; - - filter->status = 0; - filter->flag = 0; - filter->score = 0; + filter->status = filter->flag = filter->score = 0; /* setup the function table */ - vtbl = mbfl_identify_filter_get_vtbl(filter->encoding->no_encoding); + const struct mbfl_identify_vtbl *vtbl = mbfl_identify_filter_get_vtbl(filter->encoding->no_encoding); if (vtbl == NULL) { vtbl = &vtbl_identify_false; } filter->filter_ctor = vtbl->filter_ctor; filter->filter_function = vtbl->filter_function; - /* constructor */ (*filter->filter_ctor)(filter); - - return 0; } void mbfl_identify_filter_delete(mbfl_identify_filter *filter) { - if (filter == NULL) { - return; - } - - efree((void*)filter); + efree(filter); } void mbfl_filt_ident_common_ctor(mbfl_identify_filter *filter) { - filter->status = 0; - filter->flag = 0; + filter->status = filter->flag = 0; } +/* A (useless) filter which says that _every_ string is invalid in a certain encoding. + * Obviously, that cannot be true. Remove after all encodings have proper identify filters */ int mbfl_filt_ident_false(int c, mbfl_identify_filter *filter) { filter->flag = 1; /* bad */ @@ -262,6 +249,7 @@ void mbfl_filt_ident_false_ctor(mbfl_identify_filter *filter) filter->flag = 1; } +/* For encodings in which _every_ possible input string is valid */ int mbfl_filt_ident_true(int c, mbfl_identify_filter *filter) { return c; diff --git a/ext/mbstring/libmbfl/mbfl/mbfl_ident.h b/ext/mbstring/libmbfl/mbfl/mbfl_ident.h index c9afd03fb4..39be183ef3 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfl_ident.h +++ b/ext/mbstring/libmbfl/mbfl/mbfl_ident.h @@ -58,8 +58,8 @@ MBFLAPI extern const struct mbfl_identify_vtbl * mbfl_identify_filter_get_vtbl(e MBFLAPI extern mbfl_identify_filter * mbfl_identify_filter_new(enum mbfl_no_encoding encoding); MBFLAPI extern mbfl_identify_filter * mbfl_identify_filter_new2(const mbfl_encoding *encoding); MBFLAPI extern void mbfl_identify_filter_delete(mbfl_identify_filter *filter); -MBFLAPI extern int mbfl_identify_filter_init(mbfl_identify_filter *filter, enum mbfl_no_encoding encoding); -MBFLAPI extern int mbfl_identify_filter_init2(mbfl_identify_filter *filter, const mbfl_encoding *encoding); +MBFLAPI extern void mbfl_identify_filter_init(mbfl_identify_filter *filter, enum mbfl_no_encoding encoding); +MBFLAPI extern void mbfl_identify_filter_init2(mbfl_identify_filter *filter, const mbfl_encoding *encoding); MBFLAPI extern void mbfl_filt_ident_common_ctor(mbfl_identify_filter *filter); MBFLAPI extern void mbfl_filt_ident_false_ctor(mbfl_identify_filter *filter); diff --git a/ext/mbstring/libmbfl/mbfl/mbfl_memory_device.c b/ext/mbstring/libmbfl/mbfl/mbfl_memory_device.c index 09746efcd0..cd064e5ee8 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfl_memory_device.c +++ b/ext/mbstring/libmbfl/mbfl/mbfl_memory_device.c @@ -37,106 +37,67 @@ /* * memory device output functions */ -void -mbfl_memory_device_init(mbfl_memory_device *device, size_t initsz, size_t allocsz) +void mbfl_memory_device_init(mbfl_memory_device *device, size_t initsz, size_t allocsz) { - if (device) { - device->length = 0; - device->buffer = NULL; - if (initsz > 0) { - device->buffer = emalloc(initsz); - device->length = initsz; - } - device->pos = 0; - if (allocsz > MBFL_MEMORY_DEVICE_ALLOC_SIZE) { - device->allocsz = allocsz; - } else { - device->allocsz = MBFL_MEMORY_DEVICE_ALLOC_SIZE; - } - } + device->buffer = (initsz > 0) ? emalloc(initsz) : NULL; + device->length = initsz; + device->pos = 0; + device->allocsz = MAX(allocsz, MBFL_MEMORY_DEVICE_ALLOC_SIZE); } -void -mbfl_memory_device_realloc(mbfl_memory_device *device, size_t initsz, size_t allocsz) +void mbfl_memory_device_realloc(mbfl_memory_device *device, size_t initsz, size_t allocsz) { - if (device) { - if (initsz > device->length) { - device->buffer = erealloc(device->buffer, initsz); - device->length = initsz; - } - if (allocsz > MBFL_MEMORY_DEVICE_ALLOC_SIZE) { - device->allocsz = allocsz; - } else { - device->allocsz = MBFL_MEMORY_DEVICE_ALLOC_SIZE; - } + if (initsz > device->length) { + device->buffer = erealloc(device->buffer, initsz); + device->length = initsz; } + device->allocsz = MAX(allocsz, MBFL_MEMORY_DEVICE_ALLOC_SIZE); } -void -mbfl_memory_device_clear(mbfl_memory_device *device) +void mbfl_memory_device_clear(mbfl_memory_device *device) { - if (device) { - if (device->buffer) { - efree(device->buffer); - } - device->buffer = NULL; - device->length = 0; - device->pos = 0; + if (device->buffer) { + efree(device->buffer); } + device->buffer = NULL; + device->length = device->pos = 0; } -void -mbfl_memory_device_reset(mbfl_memory_device *device) +void mbfl_memory_device_reset(mbfl_memory_device *device) { - if (device) { - device->pos = 0; - } + device->pos = 0; } -void -mbfl_memory_device_unput(mbfl_memory_device *device) +void mbfl_memory_device_unput(mbfl_memory_device *device) { if (device->pos > 0) { device->pos--; } } -mbfl_string * -mbfl_memory_device_result(mbfl_memory_device *device, mbfl_string *result) +mbfl_string* mbfl_memory_device_result(mbfl_memory_device *device, mbfl_string *result) { - if (device && result) { - result->len = device->pos; - mbfl_memory_device_output('\0', device); - result->val = device->buffer; - device->buffer = NULL; - device->length = 0; - device->pos= 0; - if (result->val == NULL) { - result->len = 0; - result = NULL; - } - } else { - result = NULL; - } - + result->len = device->pos; + mbfl_memory_device_output('\0', device); + result->val = device->buffer; + device->buffer = NULL; + device->length = device->pos = 0; return result; } -int -mbfl_memory_device_output(int c, void *data) +int mbfl_memory_device_output(int c, void *data) { mbfl_memory_device *device = (mbfl_memory_device *)data; if (device->pos >= device->length) { /* reallocate buffer */ - size_t newlen; if (device->length > SIZE_MAX - device->allocsz) { /* overflow */ return -1; } - newlen = device->length + device->allocsz; + size_t newlen = device->length + device->allocsz; device->buffer = erealloc(device->buffer, newlen); device->length = newlen; } @@ -145,20 +106,15 @@ mbfl_memory_device_output(int c, void *data) return c; } -int -mbfl_memory_device_strcat(mbfl_memory_device *device, const char *psrc) +int mbfl_memory_device_strcat(mbfl_memory_device *device, const char *psrc) { return mbfl_memory_device_strncat(device, psrc, strlen(psrc)); } -int -mbfl_memory_device_strncat(mbfl_memory_device *device, const char *psrc, size_t len) +int mbfl_memory_device_strncat(mbfl_memory_device *device, const char *psrc, size_t len) { - unsigned char *w; - if (len > device->length - device->pos) { /* reallocate buffer */ - size_t newlen; if (len > SIZE_MAX - MBFL_MEMORY_DEVICE_ALLOC_SIZE || device->length > SIZE_MAX - (len + MBFL_MEMORY_DEVICE_ALLOC_SIZE)) { @@ -166,50 +122,41 @@ mbfl_memory_device_strncat(mbfl_memory_device *device, const char *psrc, size_t return -1; } - newlen = device->length + len + MBFL_MEMORY_DEVICE_ALLOC_SIZE; + size_t newlen = device->length + len + MBFL_MEMORY_DEVICE_ALLOC_SIZE; device->buffer = erealloc(device->buffer, newlen); device->length = newlen; } - w = &device->buffer[device->pos]; + unsigned char *w = &device->buffer[device->pos]; memcpy(w, psrc, len); device->pos += len; return 0; } -int -mbfl_memory_device_devcat(mbfl_memory_device *dest, mbfl_memory_device *src) +int mbfl_memory_device_devcat(mbfl_memory_device *dest, mbfl_memory_device *src) { - return mbfl_memory_device_strncat(dest, (const char *) src->buffer, src->pos); + return mbfl_memory_device_strncat(dest, (const char*)src->buffer, src->pos); } -void -mbfl_wchar_device_init(mbfl_wchar_device *device) +void mbfl_wchar_device_init(mbfl_wchar_device *device) { - if (device) { - device->buffer = NULL; - device->length = 0; - device->pos= 0; - device->allocsz = MBFL_MEMORY_DEVICE_ALLOC_SIZE; - } + device->buffer = NULL; + device->length = 0; + device->pos = 0; + device->allocsz = MBFL_MEMORY_DEVICE_ALLOC_SIZE; } -void -mbfl_wchar_device_clear(mbfl_wchar_device *device) +void mbfl_wchar_device_clear(mbfl_wchar_device *device) { - if (device) { - if (device->buffer) { - efree(device->buffer); - } - device->buffer = NULL; - device->length = 0; - device->pos = 0; + if (device->buffer) { + efree(device->buffer); } + device->buffer = NULL; + device->length = device->pos = 0; } -int -mbfl_wchar_device_output(int c, void *data) +int mbfl_wchar_device_output(int c, void *data) { mbfl_wchar_device *device = (mbfl_wchar_device *)data; @@ -228,11 +175,10 @@ mbfl_wchar_device_output(int c, void *data) return -1; } - device->buffer = erealloc(device->buffer, newlen*sizeof(int)); + device->buffer = erealloc(device->buffer, newlen * sizeof(int)); device->length = newlen; } device->buffer[device->pos++] = c; - return c; } diff --git a/ext/mbstring/libmbfl/mbfl/mbfl_memory_device.h b/ext/mbstring/libmbfl/mbfl/mbfl_memory_device.h index 7e86db3e06..b8a188ec84 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfl_memory_device.h +++ b/ext/mbstring/libmbfl/mbfl/mbfl_memory_device.h @@ -69,6 +69,4 @@ MBFLAPI extern void mbfl_wchar_device_init(mbfl_wchar_device *device); MBFLAPI extern int mbfl_wchar_device_output(int c, void *data); MBFLAPI extern void mbfl_wchar_device_clear(mbfl_wchar_device *device); - - #endif /* MBFL_MEMORY_DEVICE_H */ diff --git a/ext/mbstring/libmbfl/mbfl/mbfl_string.c b/ext/mbstring/libmbfl/mbfl/mbfl_string.c index a036356dd2..760634e65d 100644 --- a/ext/mbstring/libmbfl/mbfl/mbfl_string.c +++ b/ext/mbstring/libmbfl/mbfl/mbfl_string.c @@ -31,31 +31,22 @@ #include "mbfl_string.h" #include "mbfilter_pass.h" -/* - * string object - */ -void -mbfl_string_init(mbfl_string *string) +void mbfl_string_init_set(mbfl_string *string, const mbfl_encoding *encoding) { - string->encoding = &mbfl_encoding_pass; - string->val = (unsigned char*)NULL; + string->encoding = encoding; + string->val = NULL; string->len = 0; } -void -mbfl_string_init_set(mbfl_string *string, const mbfl_encoding *encoding) +void mbfl_string_init(mbfl_string *string) { - string->encoding = encoding; - string->val = (unsigned char*)NULL; - string->len = 0; + mbfl_string_init_set(string, &mbfl_encoding_pass); } -void -mbfl_string_clear(mbfl_string *string) +void mbfl_string_clear(mbfl_string *string) { - if (string->val != (unsigned char*)NULL) { + if (string->val) { efree(string->val); } - string->val = (unsigned char*)NULL; - string->len = 0; + mbfl_string_init_set(string, NULL); /* Poison it so any attempt to reuse will fail hard */ } diff --git a/ext/mbstring/mbstring.c b/ext/mbstring/mbstring.c index b6395aecdd..e02b237bf7 100644 --- a/ext/mbstring/mbstring.c +++ b/ext/mbstring/mbstring.c @@ -1497,9 +1497,6 @@ PHP_FUNCTION(mb_preferred_mime_name) } /* }}} */ -#define IS_SJIS1(c) ((((c)>=0x81 && (c)<=0x9f) || ((c)>=0xe0 && (c)<=0xf5)) ? 1 : 0) -#define IS_SJIS2(c) ((((c)>=0x40 && (c)<=0x7e) || ((c)>=0x80 && (c)<=0xfc)) ? 1 : 0) - /* {{{ Parses GET/POST/COOKIE data and sets global variables */ PHP_FUNCTION(mb_parse_str) { @@ -2764,8 +2761,7 @@ PHP_FUNCTION(mb_encoding_aliases) array_init(return_value); if (encoding->aliases != NULL) { - const char **alias; - for (alias = *encoding->aliases; *alias; ++alias) { + for (const char **alias = encoding->aliases; *alias; ++alias) { add_next_index_string(return_value, (char *)*alias); } } @@ -3649,7 +3645,7 @@ PHP_FUNCTION(mb_send_mail) if (!suppressed_hdrs.cnt_type) { mbfl_memory_device_strncat(&device, PHP_MBSTR_MAIL_MIME_HEADER2, sizeof(PHP_MBSTR_MAIL_MIME_HEADER2) - 1); - p = (char *)mbfl_no2preferred_mime_name(tran_cs->no_encoding); + p = (char *)mbfl_encoding_preferred_mime_name(tran_cs); if (p != NULL) { mbfl_memory_device_strncat(&device, PHP_MBSTR_MAIL_MIME_HEADER3, sizeof(PHP_MBSTR_MAIL_MIME_HEADER3) - 1); mbfl_memory_device_strcat(&device, p); @@ -3658,7 +3654,7 @@ PHP_FUNCTION(mb_send_mail) } if (!suppressed_hdrs.cnt_trans_enc) { mbfl_memory_device_strncat(&device, PHP_MBSTR_MAIL_MIME_HEADER4, sizeof(PHP_MBSTR_MAIL_MIME_HEADER4) - 1); - p = (char *)mbfl_no2preferred_mime_name(body_enc->no_encoding); + p = (char *)mbfl_encoding_preferred_mime_name(body_enc); if (p == NULL) { p = "7bit"; } diff --git a/ext/mbstring/tests/data/8859-1.txt b/ext/mbstring/tests/data/8859-1.txt new file mode 100644 index 0000000000..3a55afefc4 --- /dev/null +++ b/ext/mbstring/tests/data/8859-1.txt @@ -0,0 +1,292 @@ +# 8859-1.TXT +# Date: 2015-12-02 20:19:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-1:1998 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-1:1998 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-1 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-1 order. +# +# Version history +# 1.0 version: updates 0.1 version by adding mappings for all +# control characters. +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x00A1 # INVERTED EXCLAMATION MARK +0xA2 0x00A2 # CENT SIGN +0xA3 0x00A3 # POUND SIGN +0xA4 0x00A4 # CURRENCY SIGN +0xA5 0x00A5 # YEN SIGN +0xA6 0x00A6 # BROKEN BAR +0xA7 0x00A7 # SECTION SIGN +0xA8 0x00A8 # DIAERESIS +0xA9 0x00A9 # COPYRIGHT SIGN +0xAA 0x00AA # FEMININE ORDINAL INDICATOR +0xAB 0x00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC 0x00AC # NOT SIGN +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x00AE # REGISTERED SIGN +0xAF 0x00AF # MACRON +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x00B1 # PLUS-MINUS SIGN +0xB2 0x00B2 # SUPERSCRIPT TWO +0xB3 0x00B3 # SUPERSCRIPT THREE +0xB4 0x00B4 # ACUTE ACCENT +0xB5 0x00B5 # MICRO SIGN +0xB6 0x00B6 # PILCROW SIGN +0xB7 0x00B7 # MIDDLE DOT +0xB8 0x00B8 # CEDILLA +0xB9 0x00B9 # SUPERSCRIPT ONE +0xBA 0x00BA # MASCULINE ORDINAL INDICATOR +0xBB 0x00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC 0x00BC # VULGAR FRACTION ONE QUARTER +0xBD 0x00BD # VULGAR FRACTION ONE HALF +0xBE 0x00BE # VULGAR FRACTION THREE QUARTERS +0xBF 0x00BF # INVERTED QUESTION MARK +0xC0 0x00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 0x00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 0x00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 0x00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 0x00C6 # LATIN CAPITAL LETTER AE +0xC7 0x00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 0x00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB 0x00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC 0x00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD 0x00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE 0x00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF 0x00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 0x00D0 # LATIN CAPITAL LETTER ETH (Icelandic) +0xD1 0x00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xD2 0x00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 0x00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 0x00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 0x00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x00D7 # MULTIPLICATION SIGN +0xD8 0x00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 0x00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA 0x00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB 0x00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x00DD # LATIN CAPITAL LETTER Y WITH ACUTE +0xDE 0x00DE # LATIN CAPITAL LETTER THORN (Icelandic) +0xDF 0x00DF # LATIN SMALL LETTER SHARP S (German) +0xE0 0x00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 0x00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 0x00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 0x00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 0x00E6 # LATIN SMALL LETTER AE +0xE7 0x00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 0x00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB 0x00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC 0x00EC # LATIN SMALL LETTER I WITH GRAVE +0xED 0x00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE 0x00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF 0x00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF0 0x00F0 # LATIN SMALL LETTER ETH (Icelandic) +0xF1 0x00F1 # LATIN SMALL LETTER N WITH TILDE +0xF2 0x00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 0x00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 0x00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 0x00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x00F7 # DIVISION SIGN +0xF8 0x00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 0x00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA 0x00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB 0x00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x00FD # LATIN SMALL LETTER Y WITH ACUTE +0xFE 0x00FE # LATIN SMALL LETTER THORN (Icelandic) +0xFF 0x00FF # LATIN SMALL LETTER Y WITH DIAERESIS diff --git a/ext/mbstring/tests/data/8859-10.txt b/ext/mbstring/tests/data/8859-10.txt new file mode 100644 index 0000000000..0ec67d72ee --- /dev/null +++ b/ext/mbstring/tests/data/8859-10.txt @@ -0,0 +1,292 @@ +# 8859-10.TXT +# Date: 2015-12-02 21:53:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-10:1998 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 October 11 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-10:1998 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-10 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-10 order. +# +# Version history +# 1.0 version new. +# 1.1 corrected mistake in mapping of 0xA4 +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x0104 # LATIN CAPITAL LETTER A WITH OGONEK +0xA2 0x0112 # LATIN CAPITAL LETTER E WITH MACRON +0xA3 0x0122 # LATIN CAPITAL LETTER G WITH CEDILLA +0xA4 0x012A # LATIN CAPITAL LETTER I WITH MACRON +0xA5 0x0128 # LATIN CAPITAL LETTER I WITH TILDE +0xA6 0x0136 # LATIN CAPITAL LETTER K WITH CEDILLA +0xA7 0x00A7 # SECTION SIGN +0xA8 0x013B # LATIN CAPITAL LETTER L WITH CEDILLA +0xA9 0x0110 # LATIN CAPITAL LETTER D WITH STROKE +0xAA 0x0160 # LATIN CAPITAL LETTER S WITH CARON +0xAB 0x0166 # LATIN CAPITAL LETTER T WITH STROKE +0xAC 0x017D # LATIN CAPITAL LETTER Z WITH CARON +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x016A # LATIN CAPITAL LETTER U WITH MACRON +0xAF 0x014A # LATIN CAPITAL LETTER ENG +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x0105 # LATIN SMALL LETTER A WITH OGONEK +0xB2 0x0113 # LATIN SMALL LETTER E WITH MACRON +0xB3 0x0123 # LATIN SMALL LETTER G WITH CEDILLA +0xB4 0x012B # LATIN SMALL LETTER I WITH MACRON +0xB5 0x0129 # LATIN SMALL LETTER I WITH TILDE +0xB6 0x0137 # LATIN SMALL LETTER K WITH CEDILLA +0xB7 0x00B7 # MIDDLE DOT +0xB8 0x013C # LATIN SMALL LETTER L WITH CEDILLA +0xB9 0x0111 # LATIN SMALL LETTER D WITH STROKE +0xBA 0x0161 # LATIN SMALL LETTER S WITH CARON +0xBB 0x0167 # LATIN SMALL LETTER T WITH STROKE +0xBC 0x017E # LATIN SMALL LETTER Z WITH CARON +0xBD 0x2015 # HORIZONTAL BAR +0xBE 0x016B # LATIN SMALL LETTER U WITH MACRON +0xBF 0x014B # LATIN SMALL LETTER ENG +0xC0 0x0100 # LATIN CAPITAL LETTER A WITH MACRON +0xC1 0x00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 0x00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 0x00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 0x00C6 # LATIN CAPITAL LETTER AE +0xC7 0x012E # LATIN CAPITAL LETTER I WITH OGONEK +0xC8 0x010C # LATIN CAPITAL LETTER C WITH CARON +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x0118 # LATIN CAPITAL LETTER E WITH OGONEK +0xCB 0x00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC 0x0116 # LATIN CAPITAL LETTER E WITH DOT ABOVE +0xCD 0x00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE 0x00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF 0x00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 0x00D0 # LATIN CAPITAL LETTER ETH (Icelandic) +0xD1 0x0145 # LATIN CAPITAL LETTER N WITH CEDILLA +0xD2 0x014C # LATIN CAPITAL LETTER O WITH MACRON +0xD3 0x00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 0x00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 0x00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x0168 # LATIN CAPITAL LETTER U WITH TILDE +0xD8 0x00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 0x0172 # LATIN CAPITAL LETTER U WITH OGONEK +0xDA 0x00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB 0x00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x00DD # LATIN CAPITAL LETTER Y WITH ACUTE +0xDE 0x00DE # LATIN CAPITAL LETTER THORN (Icelandic) +0xDF 0x00DF # LATIN SMALL LETTER SHARP S (German) +0xE0 0x0101 # LATIN SMALL LETTER A WITH MACRON +0xE1 0x00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 0x00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 0x00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 0x00E6 # LATIN SMALL LETTER AE +0xE7 0x012F # LATIN SMALL LETTER I WITH OGONEK +0xE8 0x010D # LATIN SMALL LETTER C WITH CARON +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x0119 # LATIN SMALL LETTER E WITH OGONEK +0xEB 0x00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC 0x0117 # LATIN SMALL LETTER E WITH DOT ABOVE +0xED 0x00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE 0x00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF 0x00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF0 0x00F0 # LATIN SMALL LETTER ETH (Icelandic) +0xF1 0x0146 # LATIN SMALL LETTER N WITH CEDILLA +0xF2 0x014D # LATIN SMALL LETTER O WITH MACRON +0xF3 0x00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 0x00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 0x00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x0169 # LATIN SMALL LETTER U WITH TILDE +0xF8 0x00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 0x0173 # LATIN SMALL LETTER U WITH OGONEK +0xFA 0x00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB 0x00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x00FD # LATIN SMALL LETTER Y WITH ACUTE +0xFE 0x00FE # LATIN SMALL LETTER THORN (Icelandic) +0xFF 0x0138 # LATIN SMALL LETTER KRA diff --git a/ext/mbstring/tests/data/8859-11.txt b/ext/mbstring/tests/data/8859-11.txt new file mode 100644 index 0000000000..5334e3c276 --- /dev/null +++ b/ext/mbstring/tests/data/8859-11.txt @@ -0,0 +1,286 @@ +# 8859-11.TXT +# Date: 2015-12-02 21:55:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-11:2001 to Unicode +# Unicode version: 3.2 +# Table version: 2.0 +# Table format: Format A +# Date: 2002 October 7 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-11:2001 characters map into Unicode. +# +# ISO/IEC 8859-11:2001 is equivalent to TIS 620-2533 (1990) with +# the addition of 0xA0 NO-BREAK SPACE. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-11 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-11 order. +# +# Version history: +# 2002 October 7 Created +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x0E01 # THAI CHARACTER KO KAI +0xA2 0x0E02 # THAI CHARACTER KHO KHAI +0xA3 0x0E03 # THAI CHARACTER KHO KHUAT +0xA4 0x0E04 # THAI CHARACTER KHO KHWAI +0xA5 0x0E05 # THAI CHARACTER KHO KHON +0xA6 0x0E06 # THAI CHARACTER KHO RAKHANG +0xA7 0x0E07 # THAI CHARACTER NGO NGU +0xA8 0x0E08 # THAI CHARACTER CHO CHAN +0xA9 0x0E09 # THAI CHARACTER CHO CHING +0xAA 0x0E0A # THAI CHARACTER CHO CHANG +0xAB 0x0E0B # THAI CHARACTER SO SO +0xAC 0x0E0C # THAI CHARACTER CHO CHOE +0xAD 0x0E0D # THAI CHARACTER YO YING +0xAE 0x0E0E # THAI CHARACTER DO CHADA +0xAF 0x0E0F # THAI CHARACTER TO PATAK +0xB0 0x0E10 # THAI CHARACTER THO THAN +0xB1 0x0E11 # THAI CHARACTER THO NANGMONTHO +0xB2 0x0E12 # THAI CHARACTER THO PHUTHAO +0xB3 0x0E13 # THAI CHARACTER NO NEN +0xB4 0x0E14 # THAI CHARACTER DO DEK +0xB5 0x0E15 # THAI CHARACTER TO TAO +0xB6 0x0E16 # THAI CHARACTER THO THUNG +0xB7 0x0E17 # THAI CHARACTER THO THAHAN +0xB8 0x0E18 # THAI CHARACTER THO THONG +0xB9 0x0E19 # THAI CHARACTER NO NU +0xBA 0x0E1A # THAI CHARACTER BO BAIMAI +0xBB 0x0E1B # THAI CHARACTER PO PLA +0xBC 0x0E1C # THAI CHARACTER PHO PHUNG +0xBD 0x0E1D # THAI CHARACTER FO FA +0xBE 0x0E1E # THAI CHARACTER PHO PHAN +0xBF 0x0E1F # THAI CHARACTER FO FAN +0xC0 0x0E20 # THAI CHARACTER PHO SAMPHAO +0xC1 0x0E21 # THAI CHARACTER MO MA +0xC2 0x0E22 # THAI CHARACTER YO YAK +0xC3 0x0E23 # THAI CHARACTER RO RUA +0xC4 0x0E24 # THAI CHARACTER RU +0xC5 0x0E25 # THAI CHARACTER LO LING +0xC6 0x0E26 # THAI CHARACTER LU +0xC7 0x0E27 # THAI CHARACTER WO WAEN +0xC8 0x0E28 # THAI CHARACTER SO SALA +0xC9 0x0E29 # THAI CHARACTER SO RUSI +0xCA 0x0E2A # THAI CHARACTER SO SUA +0xCB 0x0E2B # THAI CHARACTER HO HIP +0xCC 0x0E2C # THAI CHARACTER LO CHULA +0xCD 0x0E2D # THAI CHARACTER O ANG +0xCE 0x0E2E # THAI CHARACTER HO NOKHUK +0xCF 0x0E2F # THAI CHARACTER PAIYANNOI +0xD0 0x0E30 # THAI CHARACTER SARA A +0xD1 0x0E31 # THAI CHARACTER MAI HAN-AKAT +0xD2 0x0E32 # THAI CHARACTER SARA AA +0xD3 0x0E33 # THAI CHARACTER SARA AM +0xD4 0x0E34 # THAI CHARACTER SARA I +0xD5 0x0E35 # THAI CHARACTER SARA II +0xD6 0x0E36 # THAI CHARACTER SARA UE +0xD7 0x0E37 # THAI CHARACTER SARA UEE +0xD8 0x0E38 # THAI CHARACTER SARA U +0xD9 0x0E39 # THAI CHARACTER SARA UU +0xDA 0x0E3A # THAI CHARACTER PHINTHU +0xDF 0x0E3F # THAI CURRENCY SYMBOL BAHT +0xE0 0x0E40 # THAI CHARACTER SARA E +0xE1 0x0E41 # THAI CHARACTER SARA AE +0xE2 0x0E42 # THAI CHARACTER SARA O +0xE3 0x0E43 # THAI CHARACTER SARA AI MAIMUAN +0xE4 0x0E44 # THAI CHARACTER SARA AI MAIMALAI +0xE5 0x0E45 # THAI CHARACTER LAKKHANGYAO +0xE6 0x0E46 # THAI CHARACTER MAIYAMOK +0xE7 0x0E47 # THAI CHARACTER MAITAIKHU +0xE8 0x0E48 # THAI CHARACTER MAI EK +0xE9 0x0E49 # THAI CHARACTER MAI THO +0xEA 0x0E4A # THAI CHARACTER MAI TRI +0xEB 0x0E4B # THAI CHARACTER MAI CHATTAWA +0xEC 0x0E4C # THAI CHARACTER THANTHAKHAT +0xED 0x0E4D # THAI CHARACTER NIKHAHIT +0xEE 0x0E4E # THAI CHARACTER YAMAKKAN +0xEF 0x0E4F # THAI CHARACTER FONGMAN +0xF0 0x0E50 # THAI DIGIT ZERO +0xF1 0x0E51 # THAI DIGIT ONE +0xF2 0x0E52 # THAI DIGIT TWO +0xF3 0x0E53 # THAI DIGIT THREE +0xF4 0x0E54 # THAI DIGIT FOUR +0xF5 0x0E55 # THAI DIGIT FIVE +0xF6 0x0E56 # THAI DIGIT SIX +0xF7 0x0E57 # THAI DIGIT SEVEN +0xF8 0x0E58 # THAI DIGIT EIGHT +0xF9 0x0E59 # THAI DIGIT NINE +0xFA 0x0E5A # THAI CHARACTER ANGKHANKHU +0xFB 0x0E5B # THAI CHARACTER KHOMUT diff --git a/ext/mbstring/tests/data/8859-13.txt b/ext/mbstring/tests/data/8859-13.txt new file mode 100644 index 0000000000..b613dc00e1 --- /dev/null +++ b/ext/mbstring/tests/data/8859-13.txt @@ -0,0 +1,291 @@ +# 8859-13.TXT +# Date: 2015-12-02 22:03:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-13:1998 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-13:1998 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-13 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-13 order. +# +# Version history +# 1.0 version: created +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x201D # RIGHT DOUBLE QUOTATION MARK +0xA2 0x00A2 # CENT SIGN +0xA3 0x00A3 # POUND SIGN +0xA4 0x00A4 # CURRENCY SIGN +0xA5 0x201E # DOUBLE LOW-9 QUOTATION MARK +0xA6 0x00A6 # BROKEN BAR +0xA7 0x00A7 # SECTION SIGN +0xA8 0x00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xA9 0x00A9 # COPYRIGHT SIGN +0xAA 0x0156 # LATIN CAPITAL LETTER R WITH CEDILLA +0xAB 0x00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC 0x00AC # NOT SIGN +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x00AE # REGISTERED SIGN +0xAF 0x00C6 # LATIN CAPITAL LETTER AE +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x00B1 # PLUS-MINUS SIGN +0xB2 0x00B2 # SUPERSCRIPT TWO +0xB3 0x00B3 # SUPERSCRIPT THREE +0xB4 0x201C # LEFT DOUBLE QUOTATION MARK +0xB5 0x00B5 # MICRO SIGN +0xB6 0x00B6 # PILCROW SIGN +0xB7 0x00B7 # MIDDLE DOT +0xB8 0x00F8 # LATIN SMALL LETTER O WITH STROKE +0xB9 0x00B9 # SUPERSCRIPT ONE +0xBA 0x0157 # LATIN SMALL LETTER R WITH CEDILLA +0xBB 0x00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC 0x00BC # VULGAR FRACTION ONE QUARTER +0xBD 0x00BD # VULGAR FRACTION ONE HALF +0xBE 0x00BE # VULGAR FRACTION THREE QUARTERS +0xBF 0x00E6 # LATIN SMALL LETTER AE +0xC0 0x0104 # LATIN CAPITAL LETTER A WITH OGONEK +0xC1 0x012E # LATIN CAPITAL LETTER I WITH OGONEK +0xC2 0x0100 # LATIN CAPITAL LETTER A WITH MACRON +0xC3 0x0106 # LATIN CAPITAL LETTER C WITH ACUTE +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 0x0118 # LATIN CAPITAL LETTER E WITH OGONEK +0xC7 0x0112 # LATIN CAPITAL LETTER E WITH MACRON +0xC8 0x010C # LATIN CAPITAL LETTER C WITH CARON +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x0179 # LATIN CAPITAL LETTER Z WITH ACUTE +0xCB 0x0116 # LATIN CAPITAL LETTER E WITH DOT ABOVE +0xCC 0x0122 # LATIN CAPITAL LETTER G WITH CEDILLA +0xCD 0x0136 # LATIN CAPITAL LETTER K WITH CEDILLA +0xCE 0x012A # LATIN CAPITAL LETTER I WITH MACRON +0xCF 0x013B # LATIN CAPITAL LETTER L WITH CEDILLA +0xD0 0x0160 # LATIN CAPITAL LETTER S WITH CARON +0xD1 0x0143 # LATIN CAPITAL LETTER N WITH ACUTE +0xD2 0x0145 # LATIN CAPITAL LETTER N WITH CEDILLA +0xD3 0x00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 0x014C # LATIN CAPITAL LETTER O WITH MACRON +0xD5 0x00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x00D7 # MULTIPLICATION SIGN +0xD8 0x0172 # LATIN CAPITAL LETTER U WITH OGONEK +0xD9 0x0141 # LATIN CAPITAL LETTER L WITH STROKE +0xDA 0x015A # LATIN CAPITAL LETTER S WITH ACUTE +0xDB 0x016A # LATIN CAPITAL LETTER U WITH MACRON +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x017B # LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xDE 0x017D # LATIN CAPITAL LETTER Z WITH CARON +0xDF 0x00DF # LATIN SMALL LETTER SHARP S (German) +0xE0 0x0105 # LATIN SMALL LETTER A WITH OGONEK +0xE1 0x012F # LATIN SMALL LETTER I WITH OGONEK +0xE2 0x0101 # LATIN SMALL LETTER A WITH MACRON +0xE3 0x0107 # LATIN SMALL LETTER C WITH ACUTE +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 0x0119 # LATIN SMALL LETTER E WITH OGONEK +0xE7 0x0113 # LATIN SMALL LETTER E WITH MACRON +0xE8 0x010D # LATIN SMALL LETTER C WITH CARON +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x017A # LATIN SMALL LETTER Z WITH ACUTE +0xEB 0x0117 # LATIN SMALL LETTER E WITH DOT ABOVE +0xEC 0x0123 # LATIN SMALL LETTER G WITH CEDILLA +0xED 0x0137 # LATIN SMALL LETTER K WITH CEDILLA +0xEE 0x012B # LATIN SMALL LETTER I WITH MACRON +0xEF 0x013C # LATIN SMALL LETTER L WITH CEDILLA +0xF0 0x0161 # LATIN SMALL LETTER S WITH CARON +0xF1 0x0144 # LATIN SMALL LETTER N WITH ACUTE +0xF2 0x0146 # LATIN SMALL LETTER N WITH CEDILLA +0xF3 0x00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 0x014D # LATIN SMALL LETTER O WITH MACRON +0xF5 0x00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x00F7 # DIVISION SIGN +0xF8 0x0173 # LATIN SMALL LETTER U WITH OGONEK +0xF9 0x0142 # LATIN SMALL LETTER L WITH STROKE +0xFA 0x015B # LATIN SMALL LETTER S WITH ACUTE +0xFB 0x016B # LATIN SMALL LETTER U WITH MACRON +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x017C # LATIN SMALL LETTER Z WITH DOT ABOVE +0xFE 0x017E # LATIN SMALL LETTER Z WITH CARON +0xFF 0x2019 # RIGHT SINGLE QUOTATION MARK diff --git a/ext/mbstring/tests/data/8859-14.txt b/ext/mbstring/tests/data/8859-14.txt new file mode 100644 index 0000000000..3c3a02466b --- /dev/null +++ b/ext/mbstring/tests/data/8859-14.txt @@ -0,0 +1,292 @@ +# 8859-14.TXT +# Date: 2015-12-02 22:05:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-14:1998 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Markus Kuhn <http://www.cl.cam.ac.uk/~mgk25/> +# Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-14:1998 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-14 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-14 order. +# +# Version history +# 1.0 version: created +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x1E02 # LATIN CAPITAL LETTER B WITH DOT ABOVE +0xA2 0x1E03 # LATIN SMALL LETTER B WITH DOT ABOVE +0xA3 0x00A3 # POUND SIGN +0xA4 0x010A # LATIN CAPITAL LETTER C WITH DOT ABOVE +0xA5 0x010B # LATIN SMALL LETTER C WITH DOT ABOVE +0xA6 0x1E0A # LATIN CAPITAL LETTER D WITH DOT ABOVE +0xA7 0x00A7 # SECTION SIGN +0xA8 0x1E80 # LATIN CAPITAL LETTER W WITH GRAVE +0xA9 0x00A9 # COPYRIGHT SIGN +0xAA 0x1E82 # LATIN CAPITAL LETTER W WITH ACUTE +0xAB 0x1E0B # LATIN SMALL LETTER D WITH DOT ABOVE +0xAC 0x1EF2 # LATIN CAPITAL LETTER Y WITH GRAVE +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x00AE # REGISTERED SIGN +0xAF 0x0178 # LATIN CAPITAL LETTER Y WITH DIAERESIS +0xB0 0x1E1E # LATIN CAPITAL LETTER F WITH DOT ABOVE +0xB1 0x1E1F # LATIN SMALL LETTER F WITH DOT ABOVE +0xB2 0x0120 # LATIN CAPITAL LETTER G WITH DOT ABOVE +0xB3 0x0121 # LATIN SMALL LETTER G WITH DOT ABOVE +0xB4 0x1E40 # LATIN CAPITAL LETTER M WITH DOT ABOVE +0xB5 0x1E41 # LATIN SMALL LETTER M WITH DOT ABOVE +0xB6 0x00B6 # PILCROW SIGN +0xB7 0x1E56 # LATIN CAPITAL LETTER P WITH DOT ABOVE +0xB8 0x1E81 # LATIN SMALL LETTER W WITH GRAVE +0xB9 0x1E57 # LATIN SMALL LETTER P WITH DOT ABOVE +0xBA 0x1E83 # LATIN SMALL LETTER W WITH ACUTE +0xBB 0x1E60 # LATIN CAPITAL LETTER S WITH DOT ABOVE +0xBC 0x1EF3 # LATIN SMALL LETTER Y WITH GRAVE +0xBD 0x1E84 # LATIN CAPITAL LETTER W WITH DIAERESIS +0xBE 0x1E85 # LATIN SMALL LETTER W WITH DIAERESIS +0xBF 0x1E61 # LATIN SMALL LETTER S WITH DOT ABOVE +0xC0 0x00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 0x00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 0x00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 0x00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 0x00C6 # LATIN CAPITAL LETTER AE +0xC7 0x00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 0x00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB 0x00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC 0x00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD 0x00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE 0x00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF 0x00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 0x0174 # LATIN CAPITAL LETTER W WITH CIRCUMFLEX +0xD1 0x00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xD2 0x00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 0x00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 0x00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 0x00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x1E6A # LATIN CAPITAL LETTER T WITH DOT ABOVE +0xD8 0x00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 0x00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA 0x00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB 0x00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x00DD # LATIN CAPITAL LETTER Y WITH ACUTE +0xDE 0x0176 # LATIN CAPITAL LETTER Y WITH CIRCUMFLEX +0xDF 0x00DF # LATIN SMALL LETTER SHARP S +0xE0 0x00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 0x00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 0x00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 0x00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 0x00E6 # LATIN SMALL LETTER AE +0xE7 0x00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 0x00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB 0x00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC 0x00EC # LATIN SMALL LETTER I WITH GRAVE +0xED 0x00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE 0x00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF 0x00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF0 0x0175 # LATIN SMALL LETTER W WITH CIRCUMFLEX +0xF1 0x00F1 # LATIN SMALL LETTER N WITH TILDE +0xF2 0x00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 0x00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 0x00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 0x00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x1E6B # LATIN SMALL LETTER T WITH DOT ABOVE +0xF8 0x00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 0x00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA 0x00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB 0x00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x00FD # LATIN SMALL LETTER Y WITH ACUTE +0xFE 0x0177 # LATIN SMALL LETTER Y WITH CIRCUMFLEX +0xFF 0x00FF # LATIN SMALL LETTER Y WITH DIAERESIS diff --git a/ext/mbstring/tests/data/8859-15.txt b/ext/mbstring/tests/data/8859-15.txt new file mode 100644 index 0000000000..f21a763600 --- /dev/null +++ b/ext/mbstring/tests/data/8859-15.txt @@ -0,0 +1,294 @@ +# 8859-15.TXT +# Date: 2015-12-02 22:06:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-15:1999 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Markus Kuhn <http://www.cl.cam.ac.uk/~mgk25/> +# Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-15:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-15 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-15 order. +# +# Version history +# +# Version history +# 1.0 version: created +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x00A1 # INVERTED EXCLAMATION MARK +0xA2 0x00A2 # CENT SIGN +0xA3 0x00A3 # POUND SIGN +0xA4 0x20AC # EURO SIGN +0xA5 0x00A5 # YEN SIGN +0xA6 0x0160 # LATIN CAPITAL LETTER S WITH CARON +0xA7 0x00A7 # SECTION SIGN +0xA8 0x0161 # LATIN SMALL LETTER S WITH CARON +0xA9 0x00A9 # COPYRIGHT SIGN +0xAA 0x00AA # FEMININE ORDINAL INDICATOR +0xAB 0x00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC 0x00AC # NOT SIGN +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x00AE # REGISTERED SIGN +0xAF 0x00AF # MACRON +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x00B1 # PLUS-MINUS SIGN +0xB2 0x00B2 # SUPERSCRIPT TWO +0xB3 0x00B3 # SUPERSCRIPT THREE +0xB4 0x017D # LATIN CAPITAL LETTER Z WITH CARON +0xB5 0x00B5 # MICRO SIGN +0xB6 0x00B6 # PILCROW SIGN +0xB7 0x00B7 # MIDDLE DOT +0xB8 0x017E # LATIN SMALL LETTER Z WITH CARON +0xB9 0x00B9 # SUPERSCRIPT ONE +0xBA 0x00BA # MASCULINE ORDINAL INDICATOR +0xBB 0x00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC 0x0152 # LATIN CAPITAL LIGATURE OE +0xBD 0x0153 # LATIN SMALL LIGATURE OE +0xBE 0x0178 # LATIN CAPITAL LETTER Y WITH DIAERESIS +0xBF 0x00BF # INVERTED QUESTION MARK +0xC0 0x00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 0x00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 0x00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 0x00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 0x00C6 # LATIN CAPITAL LETTER AE +0xC7 0x00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 0x00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB 0x00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC 0x00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD 0x00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE 0x00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF 0x00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 0x00D0 # LATIN CAPITAL LETTER ETH +0xD1 0x00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xD2 0x00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 0x00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 0x00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 0x00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x00D7 # MULTIPLICATION SIGN +0xD8 0x00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 0x00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA 0x00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB 0x00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x00DD # LATIN CAPITAL LETTER Y WITH ACUTE +0xDE 0x00DE # LATIN CAPITAL LETTER THORN +0xDF 0x00DF # LATIN SMALL LETTER SHARP S +0xE0 0x00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 0x00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 0x00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 0x00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 0x00E6 # LATIN SMALL LETTER AE +0xE7 0x00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 0x00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB 0x00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC 0x00EC # LATIN SMALL LETTER I WITH GRAVE +0xED 0x00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE 0x00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF 0x00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF0 0x00F0 # LATIN SMALL LETTER ETH +0xF1 0x00F1 # LATIN SMALL LETTER N WITH TILDE +0xF2 0x00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 0x00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 0x00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 0x00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x00F7 # DIVISION SIGN +0xF8 0x00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 0x00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA 0x00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB 0x00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x00FD # LATIN SMALL LETTER Y WITH ACUTE +0xFE 0x00FE # LATIN SMALL LETTER THORN +0xFF 0x00FF # LATIN SMALL LETTER Y WITH DIAERESIS diff --git a/ext/mbstring/tests/data/8859-16.txt b/ext/mbstring/tests/data/8859-16.txt new file mode 100644 index 0000000000..4a05225a78 --- /dev/null +++ b/ext/mbstring/tests/data/8859-16.txt @@ -0,0 +1,293 @@ +# 8859-16.TXT +# Date: 2015-12-02 22:08:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-16:2001 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 2001 July 26 (header updated: 2015 December 02) +# Authors: Markus Kuhn <http://www.cl.cam.ac.uk/~mgk25/> +# +# Copyright (c) 1999-2001 Unicode, Inc. All Rights reserved. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-16:2001 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-16 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-16 order. +# +# Version history +# 1.0 version: created +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x0104 # LATIN CAPITAL LETTER A WITH OGONEK +0xA2 0x0105 # LATIN SMALL LETTER A WITH OGONEK +0xA3 0x0141 # LATIN CAPITAL LETTER L WITH STROKE +0xA4 0x20AC # EURO SIGN +0xA5 0x201E # DOUBLE LOW-9 QUOTATION MARK +0xA6 0x0160 # LATIN CAPITAL LETTER S WITH CARON +0xA7 0x00A7 # SECTION SIGN +0xA8 0x0161 # LATIN SMALL LETTER S WITH CARON +0xA9 0x00A9 # COPYRIGHT SIGN +0xAA 0x0218 # LATIN CAPITAL LETTER S WITH COMMA BELOW +0xAB 0x00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC 0x0179 # LATIN CAPITAL LETTER Z WITH ACUTE +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x017A # LATIN SMALL LETTER Z WITH ACUTE +0xAF 0x017B # LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x00B1 # PLUS-MINUS SIGN +0xB2 0x010C # LATIN CAPITAL LETTER C WITH CARON +0xB3 0x0142 # LATIN SMALL LETTER L WITH STROKE +0xB4 0x017D # LATIN CAPITAL LETTER Z WITH CARON +0xB5 0x201D # RIGHT DOUBLE QUOTATION MARK +0xB6 0x00B6 # PILCROW SIGN +0xB7 0x00B7 # MIDDLE DOT +0xB8 0x017E # LATIN SMALL LETTER Z WITH CARON +0xB9 0x010D # LATIN SMALL LETTER C WITH CARON +0xBA 0x0219 # LATIN SMALL LETTER S WITH COMMA BELOW +0xBB 0x00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC 0x0152 # LATIN CAPITAL LIGATURE OE +0xBD 0x0153 # LATIN SMALL LIGATURE OE +0xBE 0x0178 # LATIN CAPITAL LETTER Y WITH DIAERESIS +0xBF 0x017C # LATIN SMALL LETTER Z WITH DOT ABOVE +0xC0 0x00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 0x00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 0x00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 0x0102 # LATIN CAPITAL LETTER A WITH BREVE +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x0106 # LATIN CAPITAL LETTER C WITH ACUTE +0xC6 0x00C6 # LATIN CAPITAL LETTER AE +0xC7 0x00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 0x00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB 0x00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC 0x00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD 0x00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE 0x00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF 0x00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 0x0110 # LATIN CAPITAL LETTER D WITH STROKE +0xD1 0x0143 # LATIN CAPITAL LETTER N WITH ACUTE +0xD2 0x00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 0x00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 0x00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 0x0150 # LATIN CAPITAL LETTER O WITH DOUBLE ACUTE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x015A # LATIN CAPITAL LETTER S WITH ACUTE +0xD8 0x0170 # LATIN CAPITAL LETTER U WITH DOUBLE ACUTE +0xD9 0x00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA 0x00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB 0x00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x0118 # LATIN CAPITAL LETTER E WITH OGONEK +0xDE 0x021A # LATIN CAPITAL LETTER T WITH COMMA BELOW +0xDF 0x00DF # LATIN SMALL LETTER SHARP S +0xE0 0x00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 0x00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 0x00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 0x0103 # LATIN SMALL LETTER A WITH BREVE +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x0107 # LATIN SMALL LETTER C WITH ACUTE +0xE6 0x00E6 # LATIN SMALL LETTER AE +0xE7 0x00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 0x00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB 0x00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC 0x00EC # LATIN SMALL LETTER I WITH GRAVE +0xED 0x00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE 0x00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF 0x00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF0 0x0111 # LATIN SMALL LETTER D WITH STROKE +0xF1 0x0144 # LATIN SMALL LETTER N WITH ACUTE +0xF2 0x00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 0x00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 0x00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 0x0151 # LATIN SMALL LETTER O WITH DOUBLE ACUTE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x015B # LATIN SMALL LETTER S WITH ACUTE +0xF8 0x0171 # LATIN SMALL LETTER U WITH DOUBLE ACUTE +0xF9 0x00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA 0x00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB 0x00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x0119 # LATIN SMALL LETTER E WITH OGONEK +0xFE 0x021B # LATIN SMALL LETTER T WITH COMMA BELOW +0xFF 0x00FF # LATIN SMALL LETTER Y WITH DIAERESIS diff --git a/ext/mbstring/tests/data/8859-2.txt b/ext/mbstring/tests/data/8859-2.txt new file mode 100644 index 0000000000..20da9ba993 --- /dev/null +++ b/ext/mbstring/tests/data/8859-2.txt @@ -0,0 +1,292 @@ +# 8859-2.TXT +# Date: 2015-12-02 21:34:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO 8859-2:1999 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-2:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-2 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-2 order. +# +# Version history +# 1.0 version: updates 0.1 version by adding mappings for all +# control characters. +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x0104 # LATIN CAPITAL LETTER A WITH OGONEK +0xA2 0x02D8 # BREVE +0xA3 0x0141 # LATIN CAPITAL LETTER L WITH STROKE +0xA4 0x00A4 # CURRENCY SIGN +0xA5 0x013D # LATIN CAPITAL LETTER L WITH CARON +0xA6 0x015A # LATIN CAPITAL LETTER S WITH ACUTE +0xA7 0x00A7 # SECTION SIGN +0xA8 0x00A8 # DIAERESIS +0xA9 0x0160 # LATIN CAPITAL LETTER S WITH CARON +0xAA 0x015E # LATIN CAPITAL LETTER S WITH CEDILLA +0xAB 0x0164 # LATIN CAPITAL LETTER T WITH CARON +0xAC 0x0179 # LATIN CAPITAL LETTER Z WITH ACUTE +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x017D # LATIN CAPITAL LETTER Z WITH CARON +0xAF 0x017B # LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x0105 # LATIN SMALL LETTER A WITH OGONEK +0xB2 0x02DB # OGONEK +0xB3 0x0142 # LATIN SMALL LETTER L WITH STROKE +0xB4 0x00B4 # ACUTE ACCENT +0xB5 0x013E # LATIN SMALL LETTER L WITH CARON +0xB6 0x015B # LATIN SMALL LETTER S WITH ACUTE +0xB7 0x02C7 # CARON +0xB8 0x00B8 # CEDILLA +0xB9 0x0161 # LATIN SMALL LETTER S WITH CARON +0xBA 0x015F # LATIN SMALL LETTER S WITH CEDILLA +0xBB 0x0165 # LATIN SMALL LETTER T WITH CARON +0xBC 0x017A # LATIN SMALL LETTER Z WITH ACUTE +0xBD 0x02DD # DOUBLE ACUTE ACCENT +0xBE 0x017E # LATIN SMALL LETTER Z WITH CARON +0xBF 0x017C # LATIN SMALL LETTER Z WITH DOT ABOVE +0xC0 0x0154 # LATIN CAPITAL LETTER R WITH ACUTE +0xC1 0x00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 0x00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 0x0102 # LATIN CAPITAL LETTER A WITH BREVE +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x0139 # LATIN CAPITAL LETTER L WITH ACUTE +0xC6 0x0106 # LATIN CAPITAL LETTER C WITH ACUTE +0xC7 0x00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 0x010C # LATIN CAPITAL LETTER C WITH CARON +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x0118 # LATIN CAPITAL LETTER E WITH OGONEK +0xCB 0x00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC 0x011A # LATIN CAPITAL LETTER E WITH CARON +0xCD 0x00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE 0x00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF 0x010E # LATIN CAPITAL LETTER D WITH CARON +0xD0 0x0110 # LATIN CAPITAL LETTER D WITH STROKE +0xD1 0x0143 # LATIN CAPITAL LETTER N WITH ACUTE +0xD2 0x0147 # LATIN CAPITAL LETTER N WITH CARON +0xD3 0x00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 0x00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 0x0150 # LATIN CAPITAL LETTER O WITH DOUBLE ACUTE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x00D7 # MULTIPLICATION SIGN +0xD8 0x0158 # LATIN CAPITAL LETTER R WITH CARON +0xD9 0x016E # LATIN CAPITAL LETTER U WITH RING ABOVE +0xDA 0x00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB 0x0170 # LATIN CAPITAL LETTER U WITH DOUBLE ACUTE +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x00DD # LATIN CAPITAL LETTER Y WITH ACUTE +0xDE 0x0162 # LATIN CAPITAL LETTER T WITH CEDILLA +0xDF 0x00DF # LATIN SMALL LETTER SHARP S +0xE0 0x0155 # LATIN SMALL LETTER R WITH ACUTE +0xE1 0x00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 0x00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 0x0103 # LATIN SMALL LETTER A WITH BREVE +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x013A # LATIN SMALL LETTER L WITH ACUTE +0xE6 0x0107 # LATIN SMALL LETTER C WITH ACUTE +0xE7 0x00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 0x010D # LATIN SMALL LETTER C WITH CARON +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x0119 # LATIN SMALL LETTER E WITH OGONEK +0xEB 0x00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC 0x011B # LATIN SMALL LETTER E WITH CARON +0xED 0x00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE 0x00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF 0x010F # LATIN SMALL LETTER D WITH CARON +0xF0 0x0111 # LATIN SMALL LETTER D WITH STROKE +0xF1 0x0144 # LATIN SMALL LETTER N WITH ACUTE +0xF2 0x0148 # LATIN SMALL LETTER N WITH CARON +0xF3 0x00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 0x00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 0x0151 # LATIN SMALL LETTER O WITH DOUBLE ACUTE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x00F7 # DIVISION SIGN +0xF8 0x0159 # LATIN SMALL LETTER R WITH CARON +0xF9 0x016F # LATIN SMALL LETTER U WITH RING ABOVE +0xFA 0x00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB 0x0171 # LATIN SMALL LETTER U WITH DOUBLE ACUTE +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x00FD # LATIN SMALL LETTER Y WITH ACUTE +0xFE 0x0163 # LATIN SMALL LETTER T WITH CEDILLA +0xFF 0x02D9 # DOT ABOVE diff --git a/ext/mbstring/tests/data/8859-3.txt b/ext/mbstring/tests/data/8859-3.txt new file mode 100644 index 0000000000..5a6c5de547 --- /dev/null +++ b/ext/mbstring/tests/data/8859-3.txt @@ -0,0 +1,285 @@ +# 8859-3.TXT +# Date: 2015-12-02 21:39:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-3:1999 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-3:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-3 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-3 order. +# +# Version history +# 1.0 version: updates 0.1 version by adding mappings for all +# control characters. +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x0126 # LATIN CAPITAL LETTER H WITH STROKE +0xA2 0x02D8 # BREVE +0xA3 0x00A3 # POUND SIGN +0xA4 0x00A4 # CURRENCY SIGN +0xA6 0x0124 # LATIN CAPITAL LETTER H WITH CIRCUMFLEX +0xA7 0x00A7 # SECTION SIGN +0xA8 0x00A8 # DIAERESIS +0xA9 0x0130 # LATIN CAPITAL LETTER I WITH DOT ABOVE +0xAA 0x015E # LATIN CAPITAL LETTER S WITH CEDILLA +0xAB 0x011E # LATIN CAPITAL LETTER G WITH BREVE +0xAC 0x0134 # LATIN CAPITAL LETTER J WITH CIRCUMFLEX +0xAD 0x00AD # SOFT HYPHEN +0xAF 0x017B # LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x0127 # LATIN SMALL LETTER H WITH STROKE +0xB2 0x00B2 # SUPERSCRIPT TWO +0xB3 0x00B3 # SUPERSCRIPT THREE +0xB4 0x00B4 # ACUTE ACCENT +0xB5 0x00B5 # MICRO SIGN +0xB6 0x0125 # LATIN SMALL LETTER H WITH CIRCUMFLEX +0xB7 0x00B7 # MIDDLE DOT +0xB8 0x00B8 # CEDILLA +0xB9 0x0131 # LATIN SMALL LETTER DOTLESS I +0xBA 0x015F # LATIN SMALL LETTER S WITH CEDILLA +0xBB 0x011F # LATIN SMALL LETTER G WITH BREVE +0xBC 0x0135 # LATIN SMALL LETTER J WITH CIRCUMFLEX +0xBD 0x00BD # VULGAR FRACTION ONE HALF +0xBF 0x017C # LATIN SMALL LETTER Z WITH DOT ABOVE +0xC0 0x00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 0x00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 0x00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x010A # LATIN CAPITAL LETTER C WITH DOT ABOVE +0xC6 0x0108 # LATIN CAPITAL LETTER C WITH CIRCUMFLEX +0xC7 0x00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 0x00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB 0x00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC 0x00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD 0x00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE 0x00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF 0x00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD1 0x00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xD2 0x00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 0x00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 0x00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 0x0120 # LATIN CAPITAL LETTER G WITH DOT ABOVE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x00D7 # MULTIPLICATION SIGN +0xD8 0x011C # LATIN CAPITAL LETTER G WITH CIRCUMFLEX +0xD9 0x00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA 0x00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB 0x00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x016C # LATIN CAPITAL LETTER U WITH BREVE +0xDE 0x015C # LATIN CAPITAL LETTER S WITH CIRCUMFLEX +0xDF 0x00DF # LATIN SMALL LETTER SHARP S +0xE0 0x00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 0x00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 0x00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x010B # LATIN SMALL LETTER C WITH DOT ABOVE +0xE6 0x0109 # LATIN SMALL LETTER C WITH CIRCUMFLEX +0xE7 0x00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 0x00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB 0x00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC 0x00EC # LATIN SMALL LETTER I WITH GRAVE +0xED 0x00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE 0x00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF 0x00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF1 0x00F1 # LATIN SMALL LETTER N WITH TILDE +0xF2 0x00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 0x00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 0x00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 0x0121 # LATIN SMALL LETTER G WITH DOT ABOVE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x00F7 # DIVISION SIGN +0xF8 0x011D # LATIN SMALL LETTER G WITH CIRCUMFLEX +0xF9 0x00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA 0x00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB 0x00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x016D # LATIN SMALL LETTER U WITH BREVE +0xFE 0x015D # LATIN SMALL LETTER S WITH CIRCUMFLEX +0xFF 0x02D9 # DOT ABOVE diff --git a/ext/mbstring/tests/data/8859-4.txt b/ext/mbstring/tests/data/8859-4.txt new file mode 100644 index 0000000000..e2f5e725fd --- /dev/null +++ b/ext/mbstring/tests/data/8859-4.txt @@ -0,0 +1,292 @@ +# 8859-4.TXT +# Date: 2015-12-02 21:41:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-4:1998 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-4:1998 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-4 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-4 order. +# +# Version history +# 1.0 version: updates 0.1 version by adding mappings for all +# control characters. +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x0104 # LATIN CAPITAL LETTER A WITH OGONEK +0xA2 0x0138 # LATIN SMALL LETTER KRA +0xA3 0x0156 # LATIN CAPITAL LETTER R WITH CEDILLA +0xA4 0x00A4 # CURRENCY SIGN +0xA5 0x0128 # LATIN CAPITAL LETTER I WITH TILDE +0xA6 0x013B # LATIN CAPITAL LETTER L WITH CEDILLA +0xA7 0x00A7 # SECTION SIGN +0xA8 0x00A8 # DIAERESIS +0xA9 0x0160 # LATIN CAPITAL LETTER S WITH CARON +0xAA 0x0112 # LATIN CAPITAL LETTER E WITH MACRON +0xAB 0x0122 # LATIN CAPITAL LETTER G WITH CEDILLA +0xAC 0x0166 # LATIN CAPITAL LETTER T WITH STROKE +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x017D # LATIN CAPITAL LETTER Z WITH CARON +0xAF 0x00AF # MACRON +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x0105 # LATIN SMALL LETTER A WITH OGONEK +0xB2 0x02DB # OGONEK +0xB3 0x0157 # LATIN SMALL LETTER R WITH CEDILLA +0xB4 0x00B4 # ACUTE ACCENT +0xB5 0x0129 # LATIN SMALL LETTER I WITH TILDE +0xB6 0x013C # LATIN SMALL LETTER L WITH CEDILLA +0xB7 0x02C7 # CARON +0xB8 0x00B8 # CEDILLA +0xB9 0x0161 # LATIN SMALL LETTER S WITH CARON +0xBA 0x0113 # LATIN SMALL LETTER E WITH MACRON +0xBB 0x0123 # LATIN SMALL LETTER G WITH CEDILLA +0xBC 0x0167 # LATIN SMALL LETTER T WITH STROKE +0xBD 0x014A # LATIN CAPITAL LETTER ENG +0xBE 0x017E # LATIN SMALL LETTER Z WITH CARON +0xBF 0x014B # LATIN SMALL LETTER ENG +0xC0 0x0100 # LATIN CAPITAL LETTER A WITH MACRON +0xC1 0x00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 0x00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 0x00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 0x00C6 # LATIN CAPITAL LETTER AE +0xC7 0x012E # LATIN CAPITAL LETTER I WITH OGONEK +0xC8 0x010C # LATIN CAPITAL LETTER C WITH CARON +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x0118 # LATIN CAPITAL LETTER E WITH OGONEK +0xCB 0x00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC 0x0116 # LATIN CAPITAL LETTER E WITH DOT ABOVE +0xCD 0x00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE 0x00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF 0x012A # LATIN CAPITAL LETTER I WITH MACRON +0xD0 0x0110 # LATIN CAPITAL LETTER D WITH STROKE +0xD1 0x0145 # LATIN CAPITAL LETTER N WITH CEDILLA +0xD2 0x014C # LATIN CAPITAL LETTER O WITH MACRON +0xD3 0x0136 # LATIN CAPITAL LETTER K WITH CEDILLA +0xD4 0x00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 0x00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x00D7 # MULTIPLICATION SIGN +0xD8 0x00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 0x0172 # LATIN CAPITAL LETTER U WITH OGONEK +0xDA 0x00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB 0x00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x0168 # LATIN CAPITAL LETTER U WITH TILDE +0xDE 0x016A # LATIN CAPITAL LETTER U WITH MACRON +0xDF 0x00DF # LATIN SMALL LETTER SHARP S +0xE0 0x0101 # LATIN SMALL LETTER A WITH MACRON +0xE1 0x00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 0x00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 0x00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 0x00E6 # LATIN SMALL LETTER AE +0xE7 0x012F # LATIN SMALL LETTER I WITH OGONEK +0xE8 0x010D # LATIN SMALL LETTER C WITH CARON +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x0119 # LATIN SMALL LETTER E WITH OGONEK +0xEB 0x00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC 0x0117 # LATIN SMALL LETTER E WITH DOT ABOVE +0xED 0x00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE 0x00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF 0x012B # LATIN SMALL LETTER I WITH MACRON +0xF0 0x0111 # LATIN SMALL LETTER D WITH STROKE +0xF1 0x0146 # LATIN SMALL LETTER N WITH CEDILLA +0xF2 0x014D # LATIN SMALL LETTER O WITH MACRON +0xF3 0x0137 # LATIN SMALL LETTER K WITH CEDILLA +0xF4 0x00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 0x00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x00F7 # DIVISION SIGN +0xF8 0x00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 0x0173 # LATIN SMALL LETTER U WITH OGONEK +0xFA 0x00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB 0x00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x0169 # LATIN SMALL LETTER U WITH TILDE +0xFE 0x016B # LATIN SMALL LETTER U WITH MACRON +0xFF 0x02D9 # DOT ABOVE diff --git a/ext/mbstring/tests/data/8859-5.txt b/ext/mbstring/tests/data/8859-5.txt new file mode 100644 index 0000000000..a574c3feaf --- /dev/null +++ b/ext/mbstring/tests/data/8859-5.txt @@ -0,0 +1,292 @@ +# 8859-5.TXT +# Date: 2015-12-02 21:43:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO 8859-5:1999 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-5:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-5 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-5 order. +# +# Version history +# 1.0 version: updates 0.1 version by adding mappings for all +# control characters. +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x0401 # CYRILLIC CAPITAL LETTER IO +0xA2 0x0402 # CYRILLIC CAPITAL LETTER DJE +0xA3 0x0403 # CYRILLIC CAPITAL LETTER GJE +0xA4 0x0404 # CYRILLIC CAPITAL LETTER UKRAINIAN IE +0xA5 0x0405 # CYRILLIC CAPITAL LETTER DZE +0xA6 0x0406 # CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I +0xA7 0x0407 # CYRILLIC CAPITAL LETTER YI +0xA8 0x0408 # CYRILLIC CAPITAL LETTER JE +0xA9 0x0409 # CYRILLIC CAPITAL LETTER LJE +0xAA 0x040A # CYRILLIC CAPITAL LETTER NJE +0xAB 0x040B # CYRILLIC CAPITAL LETTER TSHE +0xAC 0x040C # CYRILLIC CAPITAL LETTER KJE +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x040E # CYRILLIC CAPITAL LETTER SHORT U +0xAF 0x040F # CYRILLIC CAPITAL LETTER DZHE +0xB0 0x0410 # CYRILLIC CAPITAL LETTER A +0xB1 0x0411 # CYRILLIC CAPITAL LETTER BE +0xB2 0x0412 # CYRILLIC CAPITAL LETTER VE +0xB3 0x0413 # CYRILLIC CAPITAL LETTER GHE +0xB4 0x0414 # CYRILLIC CAPITAL LETTER DE +0xB5 0x0415 # CYRILLIC CAPITAL LETTER IE +0xB6 0x0416 # CYRILLIC CAPITAL LETTER ZHE +0xB7 0x0417 # CYRILLIC CAPITAL LETTER ZE +0xB8 0x0418 # CYRILLIC CAPITAL LETTER I +0xB9 0x0419 # CYRILLIC CAPITAL LETTER SHORT I +0xBA 0x041A # CYRILLIC CAPITAL LETTER KA +0xBB 0x041B # CYRILLIC CAPITAL LETTER EL +0xBC 0x041C # CYRILLIC CAPITAL LETTER EM +0xBD 0x041D # CYRILLIC CAPITAL LETTER EN +0xBE 0x041E # CYRILLIC CAPITAL LETTER O +0xBF 0x041F # CYRILLIC CAPITAL LETTER PE +0xC0 0x0420 # CYRILLIC CAPITAL LETTER ER +0xC1 0x0421 # CYRILLIC CAPITAL LETTER ES +0xC2 0x0422 # CYRILLIC CAPITAL LETTER TE +0xC3 0x0423 # CYRILLIC CAPITAL LETTER U +0xC4 0x0424 # CYRILLIC CAPITAL LETTER EF +0xC5 0x0425 # CYRILLIC CAPITAL LETTER HA +0xC6 0x0426 # CYRILLIC CAPITAL LETTER TSE +0xC7 0x0427 # CYRILLIC CAPITAL LETTER CHE +0xC8 0x0428 # CYRILLIC CAPITAL LETTER SHA +0xC9 0x0429 # CYRILLIC CAPITAL LETTER SHCHA +0xCA 0x042A # CYRILLIC CAPITAL LETTER HARD SIGN +0xCB 0x042B # CYRILLIC CAPITAL LETTER YERU +0xCC 0x042C # CYRILLIC CAPITAL LETTER SOFT SIGN +0xCD 0x042D # CYRILLIC CAPITAL LETTER E +0xCE 0x042E # CYRILLIC CAPITAL LETTER YU +0xCF 0x042F # CYRILLIC CAPITAL LETTER YA +0xD0 0x0430 # CYRILLIC SMALL LETTER A +0xD1 0x0431 # CYRILLIC SMALL LETTER BE +0xD2 0x0432 # CYRILLIC SMALL LETTER VE +0xD3 0x0433 # CYRILLIC SMALL LETTER GHE +0xD4 0x0434 # CYRILLIC SMALL LETTER DE +0xD5 0x0435 # CYRILLIC SMALL LETTER IE +0xD6 0x0436 # CYRILLIC SMALL LETTER ZHE +0xD7 0x0437 # CYRILLIC SMALL LETTER ZE +0xD8 0x0438 # CYRILLIC SMALL LETTER I +0xD9 0x0439 # CYRILLIC SMALL LETTER SHORT I +0xDA 0x043A # CYRILLIC SMALL LETTER KA +0xDB 0x043B # CYRILLIC SMALL LETTER EL +0xDC 0x043C # CYRILLIC SMALL LETTER EM +0xDD 0x043D # CYRILLIC SMALL LETTER EN +0xDE 0x043E # CYRILLIC SMALL LETTER O +0xDF 0x043F # CYRILLIC SMALL LETTER PE +0xE0 0x0440 # CYRILLIC SMALL LETTER ER +0xE1 0x0441 # CYRILLIC SMALL LETTER ES +0xE2 0x0442 # CYRILLIC SMALL LETTER TE +0xE3 0x0443 # CYRILLIC SMALL LETTER U +0xE4 0x0444 # CYRILLIC SMALL LETTER EF +0xE5 0x0445 # CYRILLIC SMALL LETTER HA +0xE6 0x0446 # CYRILLIC SMALL LETTER TSE +0xE7 0x0447 # CYRILLIC SMALL LETTER CHE +0xE8 0x0448 # CYRILLIC SMALL LETTER SHA +0xE9 0x0449 # CYRILLIC SMALL LETTER SHCHA +0xEA 0x044A # CYRILLIC SMALL LETTER HARD SIGN +0xEB 0x044B # CYRILLIC SMALL LETTER YERU +0xEC 0x044C # CYRILLIC SMALL LETTER SOFT SIGN +0xED 0x044D # CYRILLIC SMALL LETTER E +0xEE 0x044E # CYRILLIC SMALL LETTER YU +0xEF 0x044F # CYRILLIC SMALL LETTER YA +0xF0 0x2116 # NUMERO SIGN +0xF1 0x0451 # CYRILLIC SMALL LETTER IO +0xF2 0x0452 # CYRILLIC SMALL LETTER DJE +0xF3 0x0453 # CYRILLIC SMALL LETTER GJE +0xF4 0x0454 # CYRILLIC SMALL LETTER UKRAINIAN IE +0xF5 0x0455 # CYRILLIC SMALL LETTER DZE +0xF6 0x0456 # CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I +0xF7 0x0457 # CYRILLIC SMALL LETTER YI +0xF8 0x0458 # CYRILLIC SMALL LETTER JE +0xF9 0x0459 # CYRILLIC SMALL LETTER LJE +0xFA 0x045A # CYRILLIC SMALL LETTER NJE +0xFB 0x045B # CYRILLIC SMALL LETTER TSHE +0xFC 0x045C # CYRILLIC SMALL LETTER KJE +0xFD 0x00A7 # SECTION SIGN +0xFE 0x045E # CYRILLIC SMALL LETTER SHORT U +0xFF 0x045F # CYRILLIC SMALL LETTER DZHE diff --git a/ext/mbstring/tests/data/8859-6.txt b/ext/mbstring/tests/data/8859-6.txt new file mode 100644 index 0000000000..fed0af3631 --- /dev/null +++ b/ext/mbstring/tests/data/8859-6.txt @@ -0,0 +1,249 @@ +# 8859-6.TXT +# Date: 2015-12-02 21:44:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO 8859-6:1999 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-6:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-6 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-6 order. +# +# Version history +# 1.0 version: updates 0.1 version by adding mappings for all +# control characters. +# 0x30..0x39 remapped to the ASCII digits (U+0030..U+0039) instead +# of the Arabic digits (U+0660..U+0669). +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA4 0x00A4 # CURRENCY SIGN +0xAC 0x060C # ARABIC COMMA +0xAD 0x00AD # SOFT HYPHEN +0xBB 0x061B # ARABIC SEMICOLON +0xBF 0x061F # ARABIC QUESTION MARK +0xC1 0x0621 # ARABIC LETTER HAMZA +0xC2 0x0622 # ARABIC LETTER ALEF WITH MADDA ABOVE +0xC3 0x0623 # ARABIC LETTER ALEF WITH HAMZA ABOVE +0xC4 0x0624 # ARABIC LETTER WAW WITH HAMZA ABOVE +0xC5 0x0625 # ARABIC LETTER ALEF WITH HAMZA BELOW +0xC6 0x0626 # ARABIC LETTER YEH WITH HAMZA ABOVE +0xC7 0x0627 # ARABIC LETTER ALEF +0xC8 0x0628 # ARABIC LETTER BEH +0xC9 0x0629 # ARABIC LETTER TEH MARBUTA +0xCA 0x062A # ARABIC LETTER TEH +0xCB 0x062B # ARABIC LETTER THEH +0xCC 0x062C # ARABIC LETTER JEEM +0xCD 0x062D # ARABIC LETTER HAH +0xCE 0x062E # ARABIC LETTER KHAH +0xCF 0x062F # ARABIC LETTER DAL +0xD0 0x0630 # ARABIC LETTER THAL +0xD1 0x0631 # ARABIC LETTER REH +0xD2 0x0632 # ARABIC LETTER ZAIN +0xD3 0x0633 # ARABIC LETTER SEEN +0xD4 0x0634 # ARABIC LETTER SHEEN +0xD5 0x0635 # ARABIC LETTER SAD +0xD6 0x0636 # ARABIC LETTER DAD +0xD7 0x0637 # ARABIC LETTER TAH +0xD8 0x0638 # ARABIC LETTER ZAH +0xD9 0x0639 # ARABIC LETTER AIN +0xDA 0x063A # ARABIC LETTER GHAIN +0xE0 0x0640 # ARABIC TATWEEL +0xE1 0x0641 # ARABIC LETTER FEH +0xE2 0x0642 # ARABIC LETTER QAF +0xE3 0x0643 # ARABIC LETTER KAF +0xE4 0x0644 # ARABIC LETTER LAM +0xE5 0x0645 # ARABIC LETTER MEEM +0xE6 0x0646 # ARABIC LETTER NOON +0xE7 0x0647 # ARABIC LETTER HEH +0xE8 0x0648 # ARABIC LETTER WAW +0xE9 0x0649 # ARABIC LETTER ALEF MAKSURA +0xEA 0x064A # ARABIC LETTER YEH +0xEB 0x064B # ARABIC FATHATAN +0xEC 0x064C # ARABIC DAMMATAN +0xED 0x064D # ARABIC KASRATAN +0xEE 0x064E # ARABIC FATHA +0xEF 0x064F # ARABIC DAMMA +0xF0 0x0650 # ARABIC KASRA +0xF1 0x0651 # ARABIC SHADDA +0xF2 0x0652 # ARABIC SUKUN diff --git a/ext/mbstring/tests/data/8859-7.txt b/ext/mbstring/tests/data/8859-7.txt new file mode 100644 index 0000000000..49131a4d94 --- /dev/null +++ b/ext/mbstring/tests/data/8859-7.txt @@ -0,0 +1,299 @@ +# 8859-7.TXT +# Date: 2015-12-02 21:47:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO 8859-7:2003 to Unicode +# Unicode version: 4.0 +# Table version: 3.0 +# Table format: Format A +# Date: 2003-Nov-12 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO 8859-7:2003 characters map into Unicode. +# +# ISO 8859-7:1987 is equivalent to ISO-IR-126, ELOT 928, +# and ECMA 118. ISO 8859-7:2003 adds two currency signs +# and one other character not in the earlier standard. +# +# Format: Three tab-separated columns +# Column #1 is the ISO 8859-7 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO 8859-7 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# Remap 0xA1 to U+2018 (instead of 0x02BD) to match text of 8859-7 +# Remap 0xA2 to U+2019 (instead of 0x02BC) to match text of 8859-7 +# +# 2.0 version updates 1.0 version by adding mappings for the +# three newly added characters 0xA4, 0xA5, 0xAA. +# +# 3.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x2018 # LEFT SINGLE QUOTATION MARK +0xA2 0x2019 # RIGHT SINGLE QUOTATION MARK +0xA3 0x00A3 # POUND SIGN +0xA4 0x20AC # EURO SIGN +0xA5 0x20AF # DRACHMA SIGN +0xA6 0x00A6 # BROKEN BAR +0xA7 0x00A7 # SECTION SIGN +0xA8 0x00A8 # DIAERESIS +0xA9 0x00A9 # COPYRIGHT SIGN +0xAA 0x037A # GREEK YPOGEGRAMMENI +0xAB 0x00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC 0x00AC # NOT SIGN +0xAD 0x00AD # SOFT HYPHEN +0xAF 0x2015 # HORIZONTAL BAR +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x00B1 # PLUS-MINUS SIGN +0xB2 0x00B2 # SUPERSCRIPT TWO +0xB3 0x00B3 # SUPERSCRIPT THREE +0xB4 0x0384 # GREEK TONOS +0xB5 0x0385 # GREEK DIALYTIKA TONOS +0xB6 0x0386 # GREEK CAPITAL LETTER ALPHA WITH TONOS +0xB7 0x00B7 # MIDDLE DOT +0xB8 0x0388 # GREEK CAPITAL LETTER EPSILON WITH TONOS +0xB9 0x0389 # GREEK CAPITAL LETTER ETA WITH TONOS +0xBA 0x038A # GREEK CAPITAL LETTER IOTA WITH TONOS +0xBB 0x00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC 0x038C # GREEK CAPITAL LETTER OMICRON WITH TONOS +0xBD 0x00BD # VULGAR FRACTION ONE HALF +0xBE 0x038E # GREEK CAPITAL LETTER UPSILON WITH TONOS +0xBF 0x038F # GREEK CAPITAL LETTER OMEGA WITH TONOS +0xC0 0x0390 # GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS +0xC1 0x0391 # GREEK CAPITAL LETTER ALPHA +0xC2 0x0392 # GREEK CAPITAL LETTER BETA +0xC3 0x0393 # GREEK CAPITAL LETTER GAMMA +0xC4 0x0394 # GREEK CAPITAL LETTER DELTA +0xC5 0x0395 # GREEK CAPITAL LETTER EPSILON +0xC6 0x0396 # GREEK CAPITAL LETTER ZETA +0xC7 0x0397 # GREEK CAPITAL LETTER ETA +0xC8 0x0398 # GREEK CAPITAL LETTER THETA +0xC9 0x0399 # GREEK CAPITAL LETTER IOTA +0xCA 0x039A # GREEK CAPITAL LETTER KAPPA +0xCB 0x039B # GREEK CAPITAL LETTER LAMDA +0xCC 0x039C # GREEK CAPITAL LETTER MU +0xCD 0x039D # GREEK CAPITAL LETTER NU +0xCE 0x039E # GREEK CAPITAL LETTER XI +0xCF 0x039F # GREEK CAPITAL LETTER OMICRON +0xD0 0x03A0 # GREEK CAPITAL LETTER PI +0xD1 0x03A1 # GREEK CAPITAL LETTER RHO +0xD3 0x03A3 # GREEK CAPITAL LETTER SIGMA +0xD4 0x03A4 # GREEK CAPITAL LETTER TAU +0xD5 0x03A5 # GREEK CAPITAL LETTER UPSILON +0xD6 0x03A6 # GREEK CAPITAL LETTER PHI +0xD7 0x03A7 # GREEK CAPITAL LETTER CHI +0xD8 0x03A8 # GREEK CAPITAL LETTER PSI +0xD9 0x03A9 # GREEK CAPITAL LETTER OMEGA +0xDA 0x03AA # GREEK CAPITAL LETTER IOTA WITH DIALYTIKA +0xDB 0x03AB # GREEK CAPITAL LETTER UPSILON WITH DIALYTIKA +0xDC 0x03AC # GREEK SMALL LETTER ALPHA WITH TONOS +0xDD 0x03AD # GREEK SMALL LETTER EPSILON WITH TONOS +0xDE 0x03AE # GREEK SMALL LETTER ETA WITH TONOS +0xDF 0x03AF # GREEK SMALL LETTER IOTA WITH TONOS +0xE0 0x03B0 # GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS +0xE1 0x03B1 # GREEK SMALL LETTER ALPHA +0xE2 0x03B2 # GREEK SMALL LETTER BETA +0xE3 0x03B3 # GREEK SMALL LETTER GAMMA +0xE4 0x03B4 # GREEK SMALL LETTER DELTA +0xE5 0x03B5 # GREEK SMALL LETTER EPSILON +0xE6 0x03B6 # GREEK SMALL LETTER ZETA +0xE7 0x03B7 # GREEK SMALL LETTER ETA +0xE8 0x03B8 # GREEK SMALL LETTER THETA +0xE9 0x03B9 # GREEK SMALL LETTER IOTA +0xEA 0x03BA # GREEK SMALL LETTER KAPPA +0xEB 0x03BB # GREEK SMALL LETTER LAMDA +0xEC 0x03BC # GREEK SMALL LETTER MU +0xED 0x03BD # GREEK SMALL LETTER NU +0xEE 0x03BE # GREEK SMALL LETTER XI +0xEF 0x03BF # GREEK SMALL LETTER OMICRON +0xF0 0x03C0 # GREEK SMALL LETTER PI +0xF1 0x03C1 # GREEK SMALL LETTER RHO +0xF2 0x03C2 # GREEK SMALL LETTER FINAL SIGMA +0xF3 0x03C3 # GREEK SMALL LETTER SIGMA +0xF4 0x03C4 # GREEK SMALL LETTER TAU +0xF5 0x03C5 # GREEK SMALL LETTER UPSILON +0xF6 0x03C6 # GREEK SMALL LETTER PHI +0xF7 0x03C7 # GREEK SMALL LETTER CHI +0xF8 0x03C8 # GREEK SMALL LETTER PSI +0xF9 0x03C9 # GREEK SMALL LETTER OMEGA +0xFA 0x03CA # GREEK SMALL LETTER IOTA WITH DIALYTIKA +0xFB 0x03CB # GREEK SMALL LETTER UPSILON WITH DIALYTIKA +0xFC 0x03CC # GREEK SMALL LETTER OMICRON WITH TONOS +0xFD 0x03CD # GREEK SMALL LETTER UPSILON WITH TONOS +0xFE 0x03CE # GREEK SMALL LETTER OMEGA WITH TONOS diff --git a/ext/mbstring/tests/data/8859-8.txt b/ext/mbstring/tests/data/8859-8.txt new file mode 100644 index 0000000000..ff068f7083 --- /dev/null +++ b/ext/mbstring/tests/data/8859-8.txt @@ -0,0 +1,258 @@ +# 8859-8.TXT +# Date: 2015-12-02 21:50:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-8:1999 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 2000-Jan-03 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-8:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-8 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-8 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# 1.1 version updates to the published 8859-8:1999, correcting +# the mapping of 0xAF and adding mappings for LRM and RLM. +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA2 0x00A2 # CENT SIGN +0xA3 0x00A3 # POUND SIGN +0xA4 0x00A4 # CURRENCY SIGN +0xA5 0x00A5 # YEN SIGN +0xA6 0x00A6 # BROKEN BAR +0xA7 0x00A7 # SECTION SIGN +0xA8 0x00A8 # DIAERESIS +0xA9 0x00A9 # COPYRIGHT SIGN +0xAA 0x00D7 # MULTIPLICATION SIGN +0xAB 0x00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC 0x00AC # NOT SIGN +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x00AE # REGISTERED SIGN +0xAF 0x00AF # MACRON +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x00B1 # PLUS-MINUS SIGN +0xB2 0x00B2 # SUPERSCRIPT TWO +0xB3 0x00B3 # SUPERSCRIPT THREE +0xB4 0x00B4 # ACUTE ACCENT +0xB5 0x00B5 # MICRO SIGN +0xB6 0x00B6 # PILCROW SIGN +0xB7 0x00B7 # MIDDLE DOT +0xB8 0x00B8 # CEDILLA +0xB9 0x00B9 # SUPERSCRIPT ONE +0xBA 0x00F7 # DIVISION SIGN +0xBB 0x00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC 0x00BC # VULGAR FRACTION ONE QUARTER +0xBD 0x00BD # VULGAR FRACTION ONE HALF +0xBE 0x00BE # VULGAR FRACTION THREE QUARTERS +0xDF 0x2017 # DOUBLE LOW LINE +0xE0 0x05D0 # HEBREW LETTER ALEF +0xE1 0x05D1 # HEBREW LETTER BET +0xE2 0x05D2 # HEBREW LETTER GIMEL +0xE3 0x05D3 # HEBREW LETTER DALET +0xE4 0x05D4 # HEBREW LETTER HE +0xE5 0x05D5 # HEBREW LETTER VAV +0xE6 0x05D6 # HEBREW LETTER ZAYIN +0xE7 0x05D7 # HEBREW LETTER HET +0xE8 0x05D8 # HEBREW LETTER TET +0xE9 0x05D9 # HEBREW LETTER YOD +0xEA 0x05DA # HEBREW LETTER FINAL KAF +0xEB 0x05DB # HEBREW LETTER KAF +0xEC 0x05DC # HEBREW LETTER LAMED +0xED 0x05DD # HEBREW LETTER FINAL MEM +0xEE 0x05DE # HEBREW LETTER MEM +0xEF 0x05DF # HEBREW LETTER FINAL NUN +0xF0 0x05E0 # HEBREW LETTER NUN +0xF1 0x05E1 # HEBREW LETTER SAMEKH +0xF2 0x05E2 # HEBREW LETTER AYIN +0xF3 0x05E3 # HEBREW LETTER FINAL PE +0xF4 0x05E4 # HEBREW LETTER PE +0xF5 0x05E5 # HEBREW LETTER FINAL TSADI +0xF6 0x05E6 # HEBREW LETTER TSADI +0xF7 0x05E7 # HEBREW LETTER QOF +0xF8 0x05E8 # HEBREW LETTER RESH +0xF9 0x05E9 # HEBREW LETTER SHIN +0xFA 0x05EA # HEBREW LETTER TAV +0xFD 0x200E # LEFT-TO-RIGHT MARK +0xFE 0x200F # RIGHT-TO-LEFT MARK diff --git a/ext/mbstring/tests/data/8859-9.txt b/ext/mbstring/tests/data/8859-9.txt new file mode 100644 index 0000000000..d0c9ff94e4 --- /dev/null +++ b/ext/mbstring/tests/data/8859-9.txt @@ -0,0 +1,294 @@ +# 8859-9.TXT +# Date: 2015-12-02 21:51:00 GMT [KW] +# © 2015 Unicode®, Inc. +# For terms of use, see http://www.unicode.org/terms_of_use.html +# +# Name: ISO/IEC 8859-9:1999 to Unicode +# Unicode version: 3.0 +# Table version: 2.0 +# Table format: Format A +# Date: 1999 July 27 (header updated: 2015 December 02) +# Authors: Ken Whistler <ken@unicode.org> +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-9:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-9 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-9 order. +# +# ISO/IEC 8859-9 is also equivalent to ISO-IR-148. +# +# Version history +# 1.0 version: updates 0.1 version by adding mappings for all +# control characters. +# 2.0 version: updates to copyright notice and terms of use; no +# changes to character mappings +# +# Updated versions of this file may be found in: +# http://www.unicode.org/Public/MAPPINGS/ +# +# Any comments or problems, contact us at: +# http://www.unicode.org/reporting.html +# +0x00 0x0000 # NULL +0x01 0x0001 # START OF HEADING +0x02 0x0002 # START OF TEXT +0x03 0x0003 # END OF TEXT +0x04 0x0004 # END OF TRANSMISSION +0x05 0x0005 # ENQUIRY +0x06 0x0006 # ACKNOWLEDGE +0x07 0x0007 # BELL +0x08 0x0008 # BACKSPACE +0x09 0x0009 # HORIZONTAL TABULATION +0x0A 0x000A # LINE FEED +0x0B 0x000B # VERTICAL TABULATION +0x0C 0x000C # FORM FEED +0x0D 0x000D # CARRIAGE RETURN +0x0E 0x000E # SHIFT OUT +0x0F 0x000F # SHIFT IN +0x10 0x0010 # DATA LINK ESCAPE +0x11 0x0011 # DEVICE CONTROL ONE +0x12 0x0012 # DEVICE CONTROL TWO +0x13 0x0013 # DEVICE CONTROL THREE +0x14 0x0014 # DEVICE CONTROL FOUR +0x15 0x0015 # NEGATIVE ACKNOWLEDGE +0x16 0x0016 # SYNCHRONOUS IDLE +0x17 0x0017 # END OF TRANSMISSION BLOCK +0x18 0x0018 # CANCEL +0x19 0x0019 # END OF MEDIUM +0x1A 0x001A # SUBSTITUTE +0x1B 0x001B # ESCAPE +0x1C 0x001C # FILE SEPARATOR +0x1D 0x001D # GROUP SEPARATOR +0x1E 0x001E # RECORD SEPARATOR +0x1F 0x001F # UNIT SEPARATOR +0x20 0x0020 # SPACE +0x21 0x0021 # EXCLAMATION MARK +0x22 0x0022 # QUOTATION MARK +0x23 0x0023 # NUMBER SIGN +0x24 0x0024 # DOLLAR SIGN +0x25 0x0025 # PERCENT SIGN +0x26 0x0026 # AMPERSAND +0x27 0x0027 # APOSTROPHE +0x28 0x0028 # LEFT PARENTHESIS +0x29 0x0029 # RIGHT PARENTHESIS +0x2A 0x002A # ASTERISK +0x2B 0x002B # PLUS SIGN +0x2C 0x002C # COMMA +0x2D 0x002D # HYPHEN-MINUS +0x2E 0x002E # FULL STOP +0x2F 0x002F # SOLIDUS +0x30 0x0030 # DIGIT ZERO +0x31 0x0031 # DIGIT ONE +0x32 0x0032 # DIGIT TWO +0x33 0x0033 # DIGIT THREE +0x34 0x0034 # DIGIT FOUR +0x35 0x0035 # DIGIT FIVE +0x36 0x0036 # DIGIT SIX +0x37 0x0037 # DIGIT SEVEN +0x38 0x0038 # DIGIT EIGHT +0x39 0x0039 # DIGIT NINE +0x3A 0x003A # COLON +0x3B 0x003B # SEMICOLON +0x3C 0x003C # LESS-THAN SIGN +0x3D 0x003D # EQUALS SIGN +0x3E 0x003E # GREATER-THAN SIGN +0x3F 0x003F # QUESTION MARK +0x40 0x0040 # COMMERCIAL AT +0x41 0x0041 # LATIN CAPITAL LETTER A +0x42 0x0042 # LATIN CAPITAL LETTER B +0x43 0x0043 # LATIN CAPITAL LETTER C +0x44 0x0044 # LATIN CAPITAL LETTER D +0x45 0x0045 # LATIN CAPITAL LETTER E +0x46 0x0046 # LATIN CAPITAL LETTER F +0x47 0x0047 # LATIN CAPITAL LETTER G +0x48 0x0048 # LATIN CAPITAL LETTER H +0x49 0x0049 # LATIN CAPITAL LETTER I +0x4A 0x004A # LATIN CAPITAL LETTER J +0x4B 0x004B # LATIN CAPITAL LETTER K +0x4C 0x004C # LATIN CAPITAL LETTER L +0x4D 0x004D # LATIN CAPITAL LETTER M +0x4E 0x004E # LATIN CAPITAL LETTER N +0x4F 0x004F # LATIN CAPITAL LETTER O +0x50 0x0050 # LATIN CAPITAL LETTER P +0x51 0x0051 # LATIN CAPITAL LETTER Q +0x52 0x0052 # LATIN CAPITAL LETTER R +0x53 0x0053 # LATIN CAPITAL LETTER S +0x54 0x0054 # LATIN CAPITAL LETTER T +0x55 0x0055 # LATIN CAPITAL LETTER U +0x56 0x0056 # LATIN CAPITAL LETTER V +0x57 0x0057 # LATIN CAPITAL LETTER W +0x58 0x0058 # LATIN CAPITAL LETTER X +0x59 0x0059 # LATIN CAPITAL LETTER Y +0x5A 0x005A # LATIN CAPITAL LETTER Z +0x5B 0x005B # LEFT SQUARE BRACKET +0x5C 0x005C # REVERSE SOLIDUS +0x5D 0x005D # RIGHT SQUARE BRACKET +0x5E 0x005E # CIRCUMFLEX ACCENT +0x5F 0x005F # LOW LINE +0x60 0x0060 # GRAVE ACCENT +0x61 0x0061 # LATIN SMALL LETTER A +0x62 0x0062 # LATIN SMALL LETTER B +0x63 0x0063 # LATIN SMALL LETTER C +0x64 0x0064 # LATIN SMALL LETTER D +0x65 0x0065 # LATIN SMALL LETTER E +0x66 0x0066 # LATIN SMALL LETTER F +0x67 0x0067 # LATIN SMALL LETTER G +0x68 0x0068 # LATIN SMALL LETTER H +0x69 0x0069 # LATIN SMALL LETTER I +0x6A 0x006A # LATIN SMALL LETTER J +0x6B 0x006B # LATIN SMALL LETTER K +0x6C 0x006C # LATIN SMALL LETTER L +0x6D 0x006D # LATIN SMALL LETTER M +0x6E 0x006E # LATIN SMALL LETTER N +0x6F 0x006F # LATIN SMALL LETTER O +0x70 0x0070 # LATIN SMALL LETTER P +0x71 0x0071 # LATIN SMALL LETTER Q +0x72 0x0072 # LATIN SMALL LETTER R +0x73 0x0073 # LATIN SMALL LETTER S +0x74 0x0074 # LATIN SMALL LETTER T +0x75 0x0075 # LATIN SMALL LETTER U +0x76 0x0076 # LATIN SMALL LETTER V +0x77 0x0077 # LATIN SMALL LETTER W +0x78 0x0078 # LATIN SMALL LETTER X +0x79 0x0079 # LATIN SMALL LETTER Y +0x7A 0x007A # LATIN SMALL LETTER Z +0x7B 0x007B # LEFT CURLY BRACKET +0x7C 0x007C # VERTICAL LINE +0x7D 0x007D # RIGHT CURLY BRACKET +0x7E 0x007E # TILDE +0x7F 0x007F # DELETE +0x80 0x0080 # <control> +0x81 0x0081 # <control> +0x82 0x0082 # <control> +0x83 0x0083 # <control> +0x84 0x0084 # <control> +0x85 0x0085 # <control> +0x86 0x0086 # <control> +0x87 0x0087 # <control> +0x88 0x0088 # <control> +0x89 0x0089 # <control> +0x8A 0x008A # <control> +0x8B 0x008B # <control> +0x8C 0x008C # <control> +0x8D 0x008D # <control> +0x8E 0x008E # <control> +0x8F 0x008F # <control> +0x90 0x0090 # <control> +0x91 0x0091 # <control> +0x92 0x0092 # <control> +0x93 0x0093 # <control> +0x94 0x0094 # <control> +0x95 0x0095 # <control> +0x96 0x0096 # <control> +0x97 0x0097 # <control> +0x98 0x0098 # <control> +0x99 0x0099 # <control> +0x9A 0x009A # <control> +0x9B 0x009B # <control> +0x9C 0x009C # <control> +0x9D 0x009D # <control> +0x9E 0x009E # <control> +0x9F 0x009F # <control> +0xA0 0x00A0 # NO-BREAK SPACE +0xA1 0x00A1 # INVERTED EXCLAMATION MARK +0xA2 0x00A2 # CENT SIGN +0xA3 0x00A3 # POUND SIGN +0xA4 0x00A4 # CURRENCY SIGN +0xA5 0x00A5 # YEN SIGN +0xA6 0x00A6 # BROKEN BAR +0xA7 0x00A7 # SECTION SIGN +0xA8 0x00A8 # DIAERESIS +0xA9 0x00A9 # COPYRIGHT SIGN +0xAA 0x00AA # FEMININE ORDINAL INDICATOR +0xAB 0x00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC 0x00AC # NOT SIGN +0xAD 0x00AD # SOFT HYPHEN +0xAE 0x00AE # REGISTERED SIGN +0xAF 0x00AF # MACRON +0xB0 0x00B0 # DEGREE SIGN +0xB1 0x00B1 # PLUS-MINUS SIGN +0xB2 0x00B2 # SUPERSCRIPT TWO +0xB3 0x00B3 # SUPERSCRIPT THREE +0xB4 0x00B4 # ACUTE ACCENT +0xB5 0x00B5 # MICRO SIGN +0xB6 0x00B6 # PILCROW SIGN +0xB7 0x00B7 # MIDDLE DOT +0xB8 0x00B8 # CEDILLA +0xB9 0x00B9 # SUPERSCRIPT ONE +0xBA 0x00BA # MASCULINE ORDINAL INDICATOR +0xBB 0x00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC 0x00BC # VULGAR FRACTION ONE QUARTER +0xBD 0x00BD # VULGAR FRACTION ONE HALF +0xBE 0x00BE # VULGAR FRACTION THREE QUARTERS +0xBF 0x00BF # INVERTED QUESTION MARK +0xC0 0x00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 0x00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 0x00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 0x00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 0x00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 0x00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 0x00C6 # LATIN CAPITAL LETTER AE +0xC7 0x00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 0x00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 0x00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA 0x00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB 0x00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC 0x00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD 0x00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE 0x00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF 0x00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 0x011E # LATIN CAPITAL LETTER G WITH BREVE +0xD1 0x00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xD2 0x00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 0x00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 0x00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 0x00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 0x00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 0x00D7 # MULTIPLICATION SIGN +0xD8 0x00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 0x00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA 0x00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB 0x00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC 0x00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD 0x0130 # LATIN CAPITAL LETTER I WITH DOT ABOVE +0xDE 0x015E # LATIN CAPITAL LETTER S WITH CEDILLA +0xDF 0x00DF # LATIN SMALL LETTER SHARP S +0xE0 0x00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 0x00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 0x00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 0x00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 0x00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 0x00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 0x00E6 # LATIN SMALL LETTER AE +0xE7 0x00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 0x00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 0x00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA 0x00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB 0x00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC 0x00EC # LATIN SMALL LETTER I WITH GRAVE +0xED 0x00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE 0x00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF 0x00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF0 0x011F # LATIN SMALL LETTER G WITH BREVE +0xF1 0x00F1 # LATIN SMALL LETTER N WITH TILDE +0xF2 0x00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 0x00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 0x00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 0x00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 0x00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 0x00F7 # DIVISION SIGN +0xF8 0x00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 0x00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA 0x00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB 0x00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC 0x00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD 0x0131 # LATIN SMALL LETTER DOTLESS I +0xFE 0x015F # LATIN SMALL LETTER S WITH CEDILLA +0xFF 0x00FF # LATIN SMALL LETTER Y WITH DIAERESIS diff --git a/ext/mbstring/tests/encoding_tests.inc b/ext/mbstring/tests/encoding_tests.inc new file mode 100644 index 0000000000..43c6addee6 --- /dev/null +++ b/ext/mbstring/tests/encoding_tests.inc @@ -0,0 +1,98 @@ +<?php + +// Common code for tests which focus on conversion and verification of text +// in some specific encoding + +function dbgPrint($str) { + $result = ''; + if (mb_check_encoding($str, 'ASCII')) + $result .= '"' . $str . '" '; + return $result . "(" . bin2hex($str) . ")"; +} + +function identifyValidString($goodString, $encoding) { + $result = mb_check_encoding($goodString, $encoding); + if (!$result) + die("mb_check_encoding failed on good $encoding string: " . dbgPrint($goodString)); +} + +function identifyInvalidString($badString, $encoding) { + $result = mb_check_encoding($badString, $encoding); + if ($result) + die("mb_check_encoding passed on bad $encoding string: " . dbgPrint($badString)); +} + +function testConversion($fromString, $toString, $fromEncoding, $toEncoding) { + $result = mb_convert_encoding($fromString, $toEncoding, $fromEncoding); + if ($result !== $toString) + die("mb_convert_encoding not working on $fromEncoding input: " . dbgPrint($fromString) . "\nExpected $toEncoding: " . dbgPrint($toString) . "\nActually got: " . dbgPrint($result)); +} + +function testValidConversion($fromString, $toString, $fromEncoding, $toEncoding) { + $illegalChars = mb_get_info('illegal_chars'); + testConversion($fromString, $toString, $fromEncoding, $toEncoding); + if (mb_get_info('illegal_chars') !== $illegalChars) + die("mb_convert_encoding incremented illegal_chars on valid $fromEncoding string: " . dbgPrint($fromString) . " when converting to $toEncoding"); +} + +function convertValidString($fromString, $toString, $fromEncoding, $toEncoding, $bothWays = true) { + testValidConversion($fromString, $toString, $fromEncoding, $toEncoding); + if ($bothWays) + testValidConversion($toString, $fromString, $toEncoding, $fromEncoding); +} + +function convertInvalidString($fromString, $toString, $fromEncoding, $toEncoding) { + $illegalChars = mb_get_info('illegal_chars'); + testConversion($fromString, $toString, $fromEncoding, $toEncoding); + if (mb_get_info('illegal_chars') <= $illegalChars) + die("mb_convert_encoding did not increment illegal_chars on invalid $fromEncoding string: " . dbgPrint($fromString) . " when converting to $toEncoding"); +} + +function testValidString($fromString, $toString, $fromEncoding, $toEncoding, $bothWays = true) { + identifyValidString($fromString, $fromEncoding); + convertValidString($fromString, $toString, $fromEncoding, $toEncoding, $bothWays); +} + +function testInvalidString($fromString, $toString, $fromEncoding, $toEncoding) { + identifyInvalidString($fromString, $fromEncoding); + convertInvalidString($fromString, $toString, $fromEncoding, $toEncoding); +} + +// Only for encodings where valid characters can be concatenated together in any +// way, without any escape sequences +// Also only for encodings which can be converted losslessly to and from $toEncoding +function testAllValidChars($charMap, $fromEncoding, $toEncoding) { + $goodChars = array_keys($charMap); + shuffle($goodChars); + while (!empty($goodChars)) { + $length = min(rand(5,10), count($goodChars)); + $fromString = $toString = ''; + while ($length--) { + $goodChar = array_pop($goodChars); + $fromString .= $goodChar; + $toString .= $charMap[$goodChar]; + } + + testValidString($fromString, $toString, $fromEncoding, $toEncoding); + } +} + +function testAllInvalidChars($badChars, $charMap, $fromEncoding, $toEncoding, $replacement) { + $badChars = array_keys($badChars); + shuffle($badChars); + $goodChars = array_keys($charMap); + shuffle($goodChars); + while (!empty($badChars)) { + if (empty($goodChars)) { + $goodChars = array_keys($charMap); + shuffle($goodChars); + } + $goodChar = array_pop($goodChars); + $fromString = array_pop($badChars) . $goodChar; + $toString = $replacement . $charMap[$goodChar]; + + testInvalidString($fromString, $toString, $fromEncoding, $toEncoding); + } +} + +?> diff --git a/ext/mbstring/tests/iso8859_encodings.phpt b/ext/mbstring/tests/iso8859_encodings.phpt new file mode 100644 index 0000000000..2ad647a589 --- /dev/null +++ b/ext/mbstring/tests/iso8859_encodings.phpt @@ -0,0 +1,75 @@ +--TEST-- +Exhaustive test of verification and conversion of ISO-8859-X text +--SKIPIF-- +<?php +extension_loaded('mbstring') or die('skip mbstring not available'); +if (getenv("SKIP_SLOW_TESTS")) die("skip slow test"); +?> +--FILE-- +<?php +srand(678); // Make results consistent +include('encoding_tests.inc'); +mb_substitute_character(0x25); // '%' + +function testValid($from, $to, $n) { + testValidString($from, $to, "ISO-8859-$n", 'UTF-16BE'); +} +function testInvalid($from, $to, $n) { + testInvalidString($from, $to, "ISO-8859-$n", 'UTF-16BE'); +} + +for ($n = 1; $n <= 16; $n++) { + if ($n == 11 || $n == 12) + continue; + + $toUnicode = array(); + $fromUnicode = array(); + + $fp = fopen(realpath(__DIR__ . "/data/8859-$n.txt"), 'r+'); + while ($line = fgets($fp, 256)) { + if ($line[0] == '#') + continue; + + if (sscanf($line, "0x%x\t0x%x", $byte, $codepoint) == 2) { + $toUnicode[$byte] = pack('n', $codepoint); + $fromUnicode[$codepoint] = chr($byte); + } + } + + // Try all ISO-8859-{$n} bytes (these are single-byte encodings) + for ($i = 0; $i < 256; $i++) { + if (isset($toUnicode[$i])) { + testValid(chr($i), $toUnicode[$i], $n); + } else { + testInvalid(chr($i), "\x00%", $n); + } + } + + // Try all Unicode codepoints up to 0xFFFF + for ($i = 0; $i <= 0xFFFF; $i++) { + if (isset($fromUnicode[$i])) { + testValid($fromUnicode[$i], pack('n', $i), $n); + } else if (($i & 0xfc00) != 0xd800) { + convertInvalidString(pack('n', $i), "%", 'UTF-16BE', "ISO-8859-{$n}"); + } + } + + echo "All is fine and dandy for ISO-8859-$n text\n"; +} + +?> +--EXPECT-- +All is fine and dandy for ISO-8859-1 text +All is fine and dandy for ISO-8859-2 text +All is fine and dandy for ISO-8859-3 text +All is fine and dandy for ISO-8859-4 text +All is fine and dandy for ISO-8859-5 text +All is fine and dandy for ISO-8859-6 text +All is fine and dandy for ISO-8859-7 text +All is fine and dandy for ISO-8859-8 text +All is fine and dandy for ISO-8859-9 text +All is fine and dandy for ISO-8859-10 text +All is fine and dandy for ISO-8859-13 text +All is fine and dandy for ISO-8859-14 text +All is fine and dandy for ISO-8859-15 text +All is fine and dandy for ISO-8859-16 text diff --git a/ext/mbstring/tests/mb_convert_encoding_failed_detection.phpt b/ext/mbstring/tests/mb_convert_encoding_failed_detection.phpt index b3c8ba10bb..45efe28766 100644 --- a/ext/mbstring/tests/mb_convert_encoding_failed_detection.phpt +++ b/ext/mbstring/tests/mb_convert_encoding_failed_detection.phpt @@ -1,5 +1,7 @@ --TEST-- mb_convert_encoding() when encoding detection fails +--INI-- +mbstring.strict_detection=1 --FILE-- <?php diff --git a/ext/mysqlnd/mysqlnd_ps.c b/ext/mysqlnd/mysqlnd_ps.c index 10f20eede8..55da87fdf0 100644 --- a/ext/mysqlnd/mysqlnd_ps.c +++ b/ext/mysqlnd/mysqlnd_ps.c @@ -109,7 +109,7 @@ MYSQLND_METHOD(mysqlnd_stmt, store_result)(MYSQLND_STMT * const s) } /* Position at the first row */ set->data_cursor = set->data; - } else if (result->stored_data->type == MYSQLND_BUFFERED_TYPE_ZVAL) { + } else if (result->stored_data->type == MYSQLND_BUFFERED_TYPE_C) { /*TODO*/ } diff --git a/ext/opcache/ZendAccelerator.c b/ext/opcache/ZendAccelerator.c index 22b484b789..56757895bb 100644 --- a/ext/opcache/ZendAccelerator.c +++ b/ext/opcache/ZendAccelerator.c @@ -802,7 +802,7 @@ static inline void kill_all_lockers(struct flock *mem_usage_check) /* errno is not ESRCH or we ran out of tries to kill the locker */ ZCSG(force_restart_time) = time(NULL); /* restore forced restart request */ /* cannot kill the locker, bail out with error */ - zend_accel_error(ACCEL_LOG_ERROR, "Cannot kill process %d!", mem_usage_check->l_pid); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "Cannot kill process %d!", mem_usage_check->l_pid); } mem_usage_check->l_type = F_WRLCK; @@ -2639,7 +2639,7 @@ static int zend_accel_init_shm(void) accel_shared_globals = zend_shared_alloc(sizeof(zend_accel_shared_globals) + sizeof(uint32_t)); } if (!accel_shared_globals) { - zend_accel_error(ACCEL_LOG_FATAL, "Insufficient shared memory!"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Insufficient shared memory!"); zend_shared_alloc_unlock(); return FAILURE; } @@ -2981,7 +2981,7 @@ static zend_result accel_post_startup(void) page_size = getpagesize(); # endif if (!page_size || (page_size & (page_size - 1))) { - zend_accel_error(ACCEL_LOG_FATAL, "Failure to initialize shared memory structures - can't get page size."); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Failure to initialize shared memory structures - can't get page size."); abort(); } jit_size = JIT_G(buffer_size); @@ -3001,7 +3001,7 @@ static zend_result accel_post_startup(void) break; case ALLOC_FAILURE: accel_startup_ok = 0; - zend_accel_error(ACCEL_LOG_FATAL, "Failure to initialize shared memory structures - probably not enough shared memory."); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Failure to initialize shared memory structures - probably not enough shared memory."); return SUCCESS; case SUCCESSFULLY_REATTACHED: #if defined(HAVE_JIT) && !defined(ZEND_WIN32) @@ -3014,7 +3014,7 @@ static zend_result accel_post_startup(void) break; case FAILED_REATTACHED: accel_startup_ok = 0; - zend_accel_error(ACCEL_LOG_FATAL, "Failure to initialize shared memory structures - can not reattach to exiting shared memory."); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Failure to initialize shared memory structures - can not reattach to exiting shared memory."); return SUCCESS; break; #if ENABLE_FILE_CACHE_FALLBACK @@ -3054,7 +3054,7 @@ static zend_result accel_post_startup(void) SHM_PROTECT(); } else if (!ZCG(accel_directives).file_cache) { accel_startup_ok = 0; - zend_accel_error(ACCEL_LOG_FATAL, "opcache.file_cache_only is set without a proper setting of opcache.file_cache"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "opcache.file_cache_only is set without a proper setting of opcache.file_cache"); return SUCCESS; } else { accel_shared_globals = calloc(1, sizeof(zend_accel_shared_globals)); @@ -4217,7 +4217,7 @@ static zend_persistent_script* preload_script_in_shared_memory(zend_persistent_s uint32_t checkpoint; if (zend_accel_hash_is_full(&ZCSG(hash))) { - zend_accel_error(ACCEL_LOG_FATAL, "Not enough entries in hash table for preloading. Consider increasing the value for the opcache.max_accelerated_files directive in php.ini."); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Not enough entries in hash table for preloading. Consider increasing the value for the opcache.max_accelerated_files directive in php.ini."); return NULL; } @@ -4269,7 +4269,7 @@ static zend_persistent_script* preload_script_in_shared_memory(zend_persistent_s } #endif if (!ZCG(mem)) { - zend_accel_error(ACCEL_LOG_FATAL, "Not enough shared memory for preloading. Consider increasing the value for the opcache.memory_consumption directive in php.ini."); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Not enough shared memory for preloading. Consider increasing the value for the opcache.memory_consumption directive in php.ini."); return NULL; } @@ -4587,7 +4587,7 @@ static int accel_preload(const char *config, zend_bool in_child) CG(map_ptr_last) = orig_map_ptr_last; if (EG(full_tables_cleanup)) { - zend_accel_error(ACCEL_LOG_FATAL, "Preloading is not compatible with dl() function."); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Preloading is not compatible with dl() function."); ret = FAILURE; goto finish; } @@ -4670,7 +4670,7 @@ static int accel_preload(const char *config, zend_bool in_child) zend_hash_sort_ex(&script->script.class_table, preload_sort_classes, NULL, 0); if (preload_optimize(script) != SUCCESS) { - zend_accel_error(ACCEL_LOG_FATAL, "Optimization error during preloading!"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Optimization error during preloading!"); return FAILURE; } @@ -4769,7 +4769,7 @@ static int accel_finish_startup(void) if (ZCG(accel_directives).preload && *ZCG(accel_directives).preload) { #ifdef ZEND_WIN32 - zend_accel_error(ACCEL_LOG_ERROR, "Preloading is not supported on Windows"); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "Preloading is not supported on Windows"); return FAILURE; #else int in_child = 0; @@ -4812,21 +4812,21 @@ static int accel_finish_startup(void) if (!ZCG(accel_directives).preload_user || !*ZCG(accel_directives).preload_user) { zend_shared_alloc_unlock(); - zend_accel_error(ACCEL_LOG_FATAL, "\"opcache.preload_user\" has not been defined"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "\"opcache.preload_user\" has not been defined"); return FAILURE; } pw = getpwnam(ZCG(accel_directives).preload_user); if (pw == NULL) { zend_shared_alloc_unlock(); - zend_accel_error(ACCEL_LOG_FATAL, "Preloading failed to getpwnam(\"%s\")", ZCG(accel_directives).preload_user); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Preloading failed to getpwnam(\"%s\")", ZCG(accel_directives).preload_user); return FAILURE; } pid = fork(); if (pid == -1) { zend_shared_alloc_unlock(); - zend_accel_error(ACCEL_LOG_FATAL, "Preloading failed to fork()"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Preloading failed to fork()"); return FAILURE; } else if (pid == 0) { /* children */ if (setgid(pw->pw_gid) < 0) { @@ -4847,7 +4847,7 @@ static int accel_finish_startup(void) if (waitpid(pid, &status, 0) < 0) { zend_shared_alloc_unlock(); - zend_accel_error(ACCEL_LOG_FATAL, "Preloading failed to waitpid(%d)", pid); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Preloading failed to waitpid(%d)", pid); return FAILURE; } diff --git a/ext/opcache/zend_accelerator_blacklist.c b/ext/opcache/zend_accelerator_blacklist.c index 9ae897b521..45ad0b2685 100644 --- a/ext/opcache/zend_accelerator_blacklist.c +++ b/ext/opcache/zend_accelerator_blacklist.c @@ -60,7 +60,7 @@ void zend_accel_blacklist_init(zend_blacklist *blacklist) blacklist->entries = (zend_blacklist_entry *) calloc(sizeof(zend_blacklist_entry), blacklist->size); if (!blacklist->entries) { - zend_accel_error(ACCEL_LOG_FATAL, "Blacklist initialization: no memory\n"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Blacklist initialization: no memory\n"); return; } blacklist->regexp_list = NULL; @@ -68,7 +68,7 @@ void zend_accel_blacklist_init(zend_blacklist *blacklist) static void blacklist_report_regexp_error(const char *pcre_error, int pcre_error_offset) { - zend_accel_error(ACCEL_LOG_ERROR, "Blacklist compilation failed (offset: %d), %s\n", pcre_error_offset, pcre_error); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "Blacklist compilation failed (offset: %d), %s\n", pcre_error_offset, pcre_error); } static void zend_accel_blacklist_update_regexp(zend_blacklist *blacklist) @@ -163,7 +163,7 @@ static void zend_accel_blacklist_update_regexp(zend_blacklist *blacklist) if (*c || i == blacklist->pos - 1) { if (*c) { if (!backtrack) { - zend_accel_error(ACCEL_LOG_ERROR, "Too long blacklist entry\n"); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "Too long blacklist entry\n"); } p = backtrack; } else { @@ -173,7 +173,7 @@ static void zend_accel_blacklist_update_regexp(zend_blacklist *blacklist) it = (zend_regexp_list*)malloc(sizeof(zend_regexp_list)); if (!it) { - zend_accel_error(ACCEL_LOG_ERROR, "malloc() failed\n"); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "malloc() failed\n"); return; } it->next = NULL; @@ -302,7 +302,7 @@ static void zend_accel_blacklist_loadone(zend_blacklist *blacklist, char *filena blacklist->entries[blacklist->pos].path_length = path_length; blacklist->entries[blacklist->pos].path = (char *)malloc(path_length + 1); if (!blacklist->entries[blacklist->pos].path) { - zend_accel_error(ACCEL_LOG_ERROR, "malloc() failed\n"); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "malloc() failed\n"); fclose(fp); return; } diff --git a/ext/opcache/zend_accelerator_debug.c b/ext/opcache/zend_accelerator_debug.c index f91c269975..a3db2a7215 100644 --- a/ext/opcache/zend_accelerator_debug.c +++ b/ext/opcache/zend_accelerator_debug.c @@ -28,9 +28,8 @@ #endif #include "ZendAccelerator.h" -void zend_accel_error(int type, const char *format, ...) +static void zend_accel_error_va_args(int type, const char *format, va_list args) { - va_list args; time_t timestamp; char *time_string; FILE * fLog = NULL; @@ -77,9 +76,7 @@ void zend_accel_error(int type, const char *format, ...) break; } - va_start(args, format); vfprintf(fLog, format, args); - va_end(args); fprintf(fLog, "\n"); fflush(fLog); @@ -98,3 +95,22 @@ void zend_accel_error(int type, const char *format, ...) } } + +void zend_accel_error(int type, const char *format, ...) +{ + va_list args; + va_start(args, format); + zend_accel_error_va_args(type, format, args); + va_end(args); +} + +ZEND_NORETURN void zend_accel_error_noreturn(int type, const char *format, ...) +{ + va_list args; + va_start(args, format); + ZEND_ASSERT(type == ACCEL_LOG_FATAL || type == ACCEL_LOG_ERROR); + zend_accel_error_va_args(type, format, args); + va_end(args); + /* Should never reach this. */ + abort(); +} diff --git a/ext/opcache/zend_accelerator_debug.h b/ext/opcache/zend_accelerator_debug.h index 234d0367de..61456096d1 100644 --- a/ext/opcache/zend_accelerator_debug.h +++ b/ext/opcache/zend_accelerator_debug.h @@ -29,5 +29,6 @@ #define ACCEL_LOG_DEBUG 4 void zend_accel_error(int type, const char *format, ...) ZEND_ATTRIBUTE_FORMAT(printf, 2, 3); +ZEND_NORETURN void zend_accel_error_noreturn(int type, const char *format, ...) ZEND_ATTRIBUTE_FORMAT(printf, 2, 3); #endif /* _ZEND_ACCELERATOR_DEBUG_H */ diff --git a/ext/opcache/zend_accelerator_hash.c b/ext/opcache/zend_accelerator_hash.c index 7cd9ef1ef4..b0293ae8f2 100644 --- a/ext/opcache/zend_accelerator_hash.c +++ b/ext/opcache/zend_accelerator_hash.c @@ -54,14 +54,14 @@ void zend_accel_hash_init(zend_accel_hash *accel_hash, uint32_t hash_size) /* set up hash pointers table */ accel_hash->hash_table = zend_shared_alloc(sizeof(zend_accel_hash_entry *)*accel_hash->max_num_entries); if (!accel_hash->hash_table) { - zend_accel_error(ACCEL_LOG_FATAL, "Insufficient shared memory!"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Insufficient shared memory!"); return; } /* set up hash values table */ accel_hash->hash_entries = zend_shared_alloc(sizeof(zend_accel_hash_entry)*accel_hash->max_num_entries); if (!accel_hash->hash_entries) { - zend_accel_error(ACCEL_LOG_FATAL, "Insufficient shared memory!"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Insufficient shared memory!"); return; } memset(accel_hash->hash_table, 0, sizeof(zend_accel_hash_entry *)*accel_hash->max_num_entries); diff --git a/ext/opcache/zend_shared_alloc.c b/ext/opcache/zend_shared_alloc.c index d32a70b7e8..f4cb4f55af 100644 --- a/ext/opcache/zend_shared_alloc.c +++ b/ext/opcache/zend_shared_alloc.c @@ -86,7 +86,7 @@ void zend_shared_alloc_create_lock(char *lockfile_path) fchmod(lock_file, 0666); if (lock_file == -1) { - zend_accel_error(ACCEL_LOG_FATAL, "Unable to create lock file: %s (%d)", strerror(errno), errno); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Unable to create lock file: %s (%d)", strerror(errno), errno); } val = fcntl(lock_file, F_GETFD, 0); val |= FD_CLOEXEC; @@ -98,7 +98,7 @@ void zend_shared_alloc_create_lock(char *lockfile_path) static void no_memory_bailout(size_t allocate_size, char *error) { - zend_accel_error(ACCEL_LOG_FATAL, "Unable to allocate shared memory segment of %zu bytes: %s: %s (%d)", allocate_size, error?error:"unknown", strerror(errno), errno ); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Unable to allocate shared memory segment of %zu bytes: %s: %s (%d)", allocate_size, error?error:"unknown", strerror(errno), errno ); } static void copy_shared_segments(void *to, void *from, int count, int size) @@ -231,14 +231,14 @@ int zend_shared_alloc_startup(size_t requested_size, size_t reserved_size) p_tmp_shared_globals = (zend_smm_shared_globals *) zend_shared_alloc(sizeof(zend_smm_shared_globals)); if (!p_tmp_shared_globals) { - zend_accel_error(ACCEL_LOG_FATAL, "Insufficient shared memory!"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Insufficient shared memory!"); return ALLOC_FAILURE; } memset(p_tmp_shared_globals, 0, sizeof(zend_smm_shared_globals)); tmp_shared_segments = zend_shared_alloc(shared_segments_array_size + ZSMMG(shared_segments_count) * sizeof(void *)); if (!tmp_shared_segments) { - zend_accel_error(ACCEL_LOG_FATAL, "Insufficient shared memory!"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Insufficient shared memory!"); return ALLOC_FAILURE; } @@ -252,7 +252,7 @@ int zend_shared_alloc_startup(size_t requested_size, size_t reserved_size) ZSMMG(shared_memory_state).positions = (int *)zend_shared_alloc(sizeof(int) * ZSMMG(shared_segments_count)); if (!ZSMMG(shared_memory_state).positions) { - zend_accel_error(ACCEL_LOG_FATAL, "Insufficient shared memory!"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Insufficient shared memory!"); return ALLOC_FAILURE; } @@ -263,7 +263,7 @@ int zend_shared_alloc_startup(size_t requested_size, size_t reserved_size) ZSMMG(reserved) = (char*)ZSMMG(shared_segments)[i]->p + ZSMMG(shared_segments)[i]->end; ZSMMG(reserved_size) = reserved_size; } else { - zend_accel_error(ACCEL_LOG_FATAL, "Insufficient shared memory!"); + zend_accel_error_noreturn(ACCEL_LOG_FATAL, "Insufficient shared memory!"); return ALLOC_FAILURE; } } @@ -340,7 +340,7 @@ void *zend_shared_alloc(size_t size) #if 1 if (!ZCG(locked)) { - zend_accel_error(ACCEL_LOG_ERROR, "Shared memory lock not obtained"); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "Shared memory lock not obtained"); } #endif if (block_size > ZSMMG(shared_free)) { /* No hope to find a big-enough block */ @@ -482,7 +482,7 @@ void zend_shared_alloc_lock(void) if (errno == EINTR) { continue; } - zend_accel_error(ACCEL_LOG_ERROR, "Cannot create lock - %s (%d)", strerror(errno), errno); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "Cannot create lock - %s (%d)", strerror(errno), errno); } break; } @@ -508,7 +508,7 @@ void zend_shared_alloc_unlock(void) #ifndef ZEND_WIN32 if (fcntl(lock_file, F_SETLK, &mem_write_unlock) == -1) { - zend_accel_error(ACCEL_LOG_ERROR, "Cannot remove lock - %s (%d)", strerror(errno), errno); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "Cannot remove lock - %s (%d)", strerror(errno), errno); } #ifdef ZTS tsrm_mutex_unlock(zts_lock); @@ -631,7 +631,7 @@ void zend_accel_shared_protect(int mode) for (i = 0; i < ZSMMG(shared_segments_count); i++) { DWORD oldProtect; if (!VirtualProtect(ZSMMG(shared_segments)[i]->p, ZSMMG(shared_segments)[i]->end, mode, &oldProtect)) { - zend_accel_error(ACCEL_LOG_ERROR, "Failed to protect memory"); + zend_accel_error_noreturn(ACCEL_LOG_ERROR, "Failed to protect memory"); } } #endif diff --git a/ext/pdo/pdo_dbh.c b/ext/pdo/pdo_dbh.c index 8917b9b0dc..00d20778e9 100644 --- a/ext/pdo/pdo_dbh.c +++ b/ext/pdo/pdo_dbh.c @@ -1293,8 +1293,12 @@ static int dbh_compare(zval *object1, zval *object2) static HashTable *dbh_get_gc(zend_object *object, zval **gc_data, int *gc_count) { pdo_dbh_t *dbh = php_pdo_dbh_fetch_inner(object); - *gc_data = &dbh->def_stmt_ctor_args; - *gc_count = 1; + zend_get_gc_buffer *gc_buffer = zend_get_gc_buffer_create(); + zend_get_gc_buffer_add_zval(gc_buffer, &dbh->def_stmt_ctor_args); + if (dbh->methods && dbh->methods->get_gc) { + dbh->methods->get_gc(dbh, gc_buffer); + } + zend_get_gc_buffer_use(gc_buffer, gc_data, gc_count); return zend_std_get_properties(object); } diff --git a/ext/pdo/php_pdo_driver.h b/ext/pdo/php_pdo_driver.h index 960ddec4ef..974eb0bad0 100644 --- a/ext/pdo/php_pdo_driver.h +++ b/ext/pdo/php_pdo_driver.h @@ -286,6 +286,10 @@ typedef int (*pdo_dbh_check_liveness_func)(pdo_dbh_t *dbh); * scope */ typedef void (*pdo_dbh_request_shutdown)(pdo_dbh_t *dbh); +/* Called when the PDO handle is scanned for GC. Should populate the get_gc buffer + * with any zvals in the driver_data that would be freed if the handle is destroyed. */ +typedef void (*pdo_dbh_get_gc_func)(pdo_dbh_t *dbh, zend_get_gc_buffer *buffer); + /* for adding methods to the dbh or stmt objects pointer to a list of driver specific functions. The convention is to prefix the function names using the PDO driver name; this will @@ -316,6 +320,7 @@ struct pdo_dbh_methods { pdo_dbh_get_driver_methods_func get_driver_methods; pdo_dbh_request_shutdown persistent_shutdown; pdo_dbh_txn_func in_transaction; + pdo_dbh_get_gc_func get_gc; }; /* }}} */ diff --git a/ext/pdo_dblib/dblib_driver.c b/ext/pdo_dblib/dblib_driver.c index 7f160a402f..e581642037 100644 --- a/ext/pdo_dblib/dblib_driver.c +++ b/ext/pdo_dblib/dblib_driver.c @@ -417,7 +417,8 @@ static const struct pdo_dbh_methods dblib_methods = { NULL, /* check liveness */ NULL, /* get driver methods */ NULL, /* request shutdown */ - NULL /* in transaction */ + NULL, /* in transaction */ + NULL /* get gc */ }; static int pdo_dblib_handle_factory(pdo_dbh_t *dbh, zval *driver_options) diff --git a/ext/pdo_firebird/firebird_driver.c b/ext/pdo_firebird/firebird_driver.c index c27a9e2ed5..c299907f0f 100644 --- a/ext/pdo_firebird/firebird_driver.c +++ b/ext/pdo_firebird/firebird_driver.c @@ -1005,7 +1005,11 @@ static const struct pdo_dbh_methods firebird_methods = { /* {{{ */ NULL, /* last_id not supported */ pdo_firebird_fetch_error_func, firebird_handle_get_attribute, - NULL /* check_liveness */ + NULL, /* check_liveness */ + NULL, /* get driver methods */ + NULL, /* request shutdown */ + NULL, /* in transaction */ + NULL /* get gc */ }; /* }}} */ diff --git a/ext/pdo_mysql/mysql_driver.c b/ext/pdo_mysql/mysql_driver.c index 641a9add1a..a83b12b872 100644 --- a/ext/pdo_mysql/mysql_driver.c +++ b/ext/pdo_mysql/mysql_driver.c @@ -577,7 +577,8 @@ static const struct pdo_dbh_methods mysql_methods = { pdo_mysql_check_liveness, NULL, pdo_mysql_request_shutdown, - pdo_mysql_in_transaction + pdo_mysql_in_transaction, + NULL /* get_gc */ }; /* }}} */ diff --git a/ext/pdo_oci/oci_driver.c b/ext/pdo_oci/oci_driver.c index 096a26575e..2c65fc973b 100644 --- a/ext/pdo_oci/oci_driver.c +++ b/ext/pdo_oci/oci_driver.c @@ -705,9 +705,10 @@ static const struct pdo_dbh_methods oci_methods = { pdo_oci_fetch_error_func, oci_handle_get_attribute, pdo_oci_check_liveness, /* check_liveness */ - NULL, /* get_driver_methods */ - NULL, - NULL + NULL, /* get_driver_methods */ + NULL, /* request_shutdown */ + NULL, /* in_transaction */ + NULL /* get_gc */ }; static int pdo_oci_handle_factory(pdo_dbh_t *dbh, zval *driver_options) /* {{{ */ diff --git a/ext/pdo_odbc/odbc_driver.c b/ext/pdo_odbc/odbc_driver.c index 9a254553d5..81e4915da5 100644 --- a/ext/pdo_odbc/odbc_driver.c +++ b/ext/pdo_odbc/odbc_driver.c @@ -384,7 +384,11 @@ static const struct pdo_dbh_methods odbc_methods = { NULL, /* last id */ pdo_odbc_fetch_error_func, odbc_handle_get_attr, /* get attr */ - NULL, /* check_liveness */ + NULL, /* check_liveness */ + NULL, /* get_driver_methods */ + NULL, /* request_shutdown */ + NULL, /* in_transaction */ + NULL /* get_gc */ }; static int pdo_odbc_handle_factory(pdo_dbh_t *dbh, zval *driver_options) /* {{{ */ diff --git a/ext/pdo_pgsql/pgsql_driver.c b/ext/pdo_pgsql/pgsql_driver.c index d800d65a2d..37680b8efa 100644 --- a/ext/pdo_pgsql/pgsql_driver.c +++ b/ext/pdo_pgsql/pgsql_driver.c @@ -1179,6 +1179,7 @@ static const struct pdo_dbh_methods pgsql_methods = { pdo_pgsql_get_driver_methods, /* get_driver_methods */ NULL, pgsql_handle_in_transaction, + NULL /* get_gc */ }; static int pdo_pgsql_handle_factory(pdo_dbh_t *dbh, zval *driver_options) /* {{{ */ diff --git a/ext/pdo_sqlite/sqlite_driver.c b/ext/pdo_sqlite/sqlite_driver.c index 26fa1c00eb..a8a96c7a13 100644 --- a/ext/pdo_sqlite/sqlite_driver.c +++ b/ext/pdo_sqlite/sqlite_driver.c @@ -695,6 +695,25 @@ static void pdo_sqlite_request_shutdown(pdo_dbh_t *dbh) } } +static void pdo_sqlite_get_gc(pdo_dbh_t *dbh, zend_get_gc_buffer *gc_buffer) +{ + pdo_sqlite_db_handle *H = dbh->driver_data; + + struct pdo_sqlite_func *func = H->funcs; + while (func) { + zend_get_gc_buffer_add_zval(gc_buffer, &func->func); + zend_get_gc_buffer_add_zval(gc_buffer, &func->step); + zend_get_gc_buffer_add_zval(gc_buffer, &func->fini); + func = func->next; + } + + struct pdo_sqlite_collation *collation = H->collations; + while (collation) { + zend_get_gc_buffer_add_zval(gc_buffer, &collation->callback); + collation = collation->next; + } +} + static const struct pdo_dbh_methods sqlite_methods = { sqlite_handle_closer, sqlite_handle_preparer, @@ -710,7 +729,8 @@ static const struct pdo_dbh_methods sqlite_methods = { NULL, /* check_liveness: not needed */ get_driver_methods, pdo_sqlite_request_shutdown, - NULL + NULL, /* in_transaction */ + pdo_sqlite_get_gc }; static char *make_filename_safe(const char *filename) diff --git a/ext/pdo_sqlite/tests/gc.phpt b/ext/pdo_sqlite/tests/gc.phpt new file mode 100644 index 0000000000..25407697e6 --- /dev/null +++ b/ext/pdo_sqlite/tests/gc.phpt @@ -0,0 +1,24 @@ +--TEST-- +GC support for PDO Sqlite driver data +--SKIPIF-- +<?php +if (!extension_loaded('pdo_sqlite')) print 'skip not loaded'; +?> +--FILE-- +<?php + +class Obj { + public $a; + public function callback() { } +} + +$obj = new Obj; +$obj->a = new PDO('sqlite::memory:'); +$obj->a->sqliteCreateFunction('func1', function() use ($obj) {}, 1); +$obj->a->sqliteCreateAggregate('func2', function() use ($obj) {}, function() use($obj) {}); +$obj->a->sqliteCreateCollation('col', function() use ($obj) {}); + +?> +===DONE=== +--EXPECT-- +===DONE=== diff --git a/ext/pspell/pspell.c b/ext/pspell/pspell.c index 63732b77b0..4253f5f159 100644 --- a/ext/pspell/pspell.c +++ b/ext/pspell/pspell.c @@ -19,6 +19,7 @@ #endif #include "php.h" +#include "zend_interfaces.h" #include <stdlib.h> #include <ctype.h> @@ -48,58 +49,139 @@ static PHP_MINIT_FUNCTION(pspell); static PHP_MINFO_FUNCTION(pspell); -static int le_pspell, le_pspell_config; +static zend_class_entry *php_pspell_ce = NULL; +static zend_object_handlers php_pspell_handlers; +static zend_class_entry *php_pspell_config_ce = NULL; +static zend_object_handlers php_pspell_config_handlers; zend_module_entry pspell_module_entry = { STANDARD_MODULE_HEADER, - "pspell", ext_functions, PHP_MINIT(pspell), NULL, NULL, NULL, PHP_MINFO(pspell), PHP_PSPELL_VERSION, STANDARD_MODULE_PROPERTIES + "pspell", + ext_functions, + PHP_MINIT(pspell), + NULL, + NULL, + NULL, + PHP_MINFO(pspell), + PHP_PSPELL_VERSION, + STANDARD_MODULE_PROPERTIES, }; #ifdef COMPILE_DL_PSPELL ZEND_GET_MODULE(pspell) #endif -static void php_pspell_close(zend_resource *rsrc) +/* class PSpell */ + +typedef struct _php_pspell_object { + PspellManager *mgr; + zend_object std; +} php_pspell_object; + +static php_pspell_object *php_pspell_object_from_zend_object(zend_object *zobj) { + return ((php_pspell_object*)(zobj + 1)) - 1; +} + +static zend_object *php_pspell_object_to_zend_object(php_pspell_object *obj) { + return ((zend_object*)(obj + 1)) - 1; +} + +static zend_function *php_pspell_object_get_constructor(zend_object *object) { - PspellManager *manager = (PspellManager *)rsrc->ptr; + zend_throw_error(NULL, "You cannot initialize a PSpell object except through helper functions"); + return NULL; +} + +static zend_object *php_pspell_object_create(zend_class_entry *ce) +{ + php_pspell_object *obj = zend_object_alloc(sizeof(php_pspell_object), ce); + zend_object *zobj = php_pspell_object_to_zend_object(obj); + + obj->mgr = NULL; + zend_object_std_init(zobj, ce); + object_properties_init(zobj, ce); + zobj->handlers = &php_pspell_handlers; + + return zobj; +} + +static void php_pspell_object_free(zend_object *zobj) { + delete_pspell_manager(php_pspell_object_from_zend_object(zobj)->mgr); +} - delete_pspell_manager(manager); +/* class PSpellConfig */ + +typedef struct _php_pspell_config_object { + PspellConfig *cfg; + zend_object std; +} php_pspell_config_object; + +static php_pspell_config_object *php_pspell_config_object_from_zend_object(zend_object *zobj) { + return ((php_pspell_config_object*)(zobj + 1)) - 1; } -static void php_pspell_close_config(zend_resource *rsrc) +static zend_object *php_pspell_config_object_to_zend_object(php_pspell_config_object *obj) { + return ((zend_object*)(obj + 1)) - 1; +} + +static zend_function *php_pspell_config_object_get_constructor(zend_object *object) { - PspellConfig *config = (PspellConfig *)rsrc->ptr; + zend_throw_error(NULL, "You cannot initialize a PSpellConfig object except through helper functions"); + return NULL; +} - delete_pspell_config(config); +static zend_object *php_pspell_config_object_create(zend_class_entry *ce) +{ + php_pspell_config_object *obj = zend_object_alloc(sizeof(php_pspell_config_object), ce); + zend_object *zobj = php_pspell_config_object_to_zend_object(obj); + + obj->cfg = NULL; + zend_object_std_init(zobj, ce); + object_properties_init(zobj, ce); + zobj->handlers = &php_pspell_config_handlers; + + return zobj; } -#define PSPELL_FETCH_CONFIG do { \ - zval *res = zend_hash_index_find(&EG(regular_list), conf); \ - if (res == NULL || Z_RES_P(res)->type != le_pspell_config) { \ - zend_throw_error(NULL, "%s(): " ZEND_LONG_FMT " is not a PSPELL config index", get_active_function_name(), conf); \ - RETURN_THROWS(); \ - } \ - config = (PspellConfig *)Z_RES_P(res)->ptr; \ -} while (0) - -#define PSPELL_FETCH_MANAGER do { \ - zval *res = zend_hash_index_find(&EG(regular_list), scin); \ - if (res == NULL || Z_RES_P(res)->type != le_pspell) { \ - zend_throw_error(NULL, "%s(): " ZEND_LONG_FMT " is not a PSPELL result index", get_active_function_name(), scin); \ - RETURN_THROWS(); \ - } \ - manager = (PspellManager *)Z_RES_P(res)->ptr; \ -} while (0); +static void php_pspell_config_object_free(zend_object *zobj) { + delete_pspell_config(php_pspell_config_object_from_zend_object(zobj)->cfg); +} /* {{{ PHP_MINIT_FUNCTION */ static PHP_MINIT_FUNCTION(pspell) { + zend_class_entry ce; + INIT_CLASS_ENTRY(ce, "PSpell", class_PSpell_methods); + php_pspell_ce = zend_register_internal_class(&ce); + php_pspell_ce->ce_flags |= ZEND_ACC_FINAL | ZEND_ACC_NO_DYNAMIC_PROPERTIES; + php_pspell_ce->create_object = php_pspell_object_create; + php_pspell_ce->serialize = zend_class_serialize_deny; + php_pspell_ce->unserialize = zend_class_unserialize_deny; + + memcpy(&php_pspell_handlers, &std_object_handlers, sizeof(zend_object_handlers)); + php_pspell_handlers.clone_obj = NULL; + php_pspell_handlers.free_obj = php_pspell_object_free; + php_pspell_handlers.get_constructor = php_pspell_object_get_constructor; + php_pspell_handlers.offset = XtOffsetOf(php_pspell_object, std); + + INIT_CLASS_ENTRY(ce, "PSpellConfig", class_PSpellConfig_methods); + php_pspell_config_ce = zend_register_internal_class(&ce); + php_pspell_config_ce->ce_flags |= ZEND_ACC_FINAL | ZEND_ACC_NO_DYNAMIC_PROPERTIES; + php_pspell_config_ce->create_object = php_pspell_config_object_create; + php_pspell_config_ce->serialize = zend_class_serialize_deny; + php_pspell_config_ce->unserialize = zend_class_unserialize_deny; + + memcpy(&php_pspell_config_handlers, &std_object_handlers, sizeof(zend_object_handlers)); + php_pspell_config_handlers.clone_obj = NULL; + php_pspell_config_handlers.free_obj = php_pspell_config_object_free; + php_pspell_config_handlers.get_constructor = php_pspell_config_object_get_constructor; + php_pspell_config_handlers.offset = XtOffsetOf(php_pspell_config_object, std); + REGISTER_LONG_CONSTANT("PSPELL_FAST", PSPELL_FAST, CONST_PERSISTENT | CONST_CS); REGISTER_LONG_CONSTANT("PSPELL_NORMAL", PSPELL_NORMAL, CONST_PERSISTENT | CONST_CS); REGISTER_LONG_CONSTANT("PSPELL_BAD_SPELLERS", PSPELL_BAD_SPELLERS, CONST_PERSISTENT | CONST_CS); REGISTER_LONG_CONSTANT("PSPELL_RUN_TOGETHER", PSPELL_RUN_TOGETHER, CONST_PERSISTENT | CONST_CS); - le_pspell = zend_register_list_destructors_ex(php_pspell_close, NULL, "pspell", module_number); - le_pspell_config = zend_register_list_destructors_ex(php_pspell_close_config, NULL, "pspell config", module_number); + return SUCCESS; } /* }}} */ @@ -111,7 +193,6 @@ PHP_FUNCTION(pspell_new) size_t language_len, spelling_len = 0, jargon_len = 0, encoding_len = 0; zend_long mode = Z_L(0), speed = Z_L(0); int argc = ZEND_NUM_ARGS(); - zval *ind; #ifdef PHP_WIN32 TCHAR aspell_dir[200]; @@ -122,7 +203,6 @@ PHP_FUNCTION(pspell_new) #endif PspellCanHaveError *ret; - PspellManager *manager; PspellConfig *config; if (zend_parse_parameters(argc, "s|sssl", &language, &language_len, &spelling, &spelling_len, @@ -194,9 +274,8 @@ PHP_FUNCTION(pspell_new) RETURN_FALSE; } - manager = to_pspell_manager(ret); - ind = zend_list_insert(manager, le_pspell); - RETURN_LONG(Z_RES_HANDLE_P(ind)); + object_init_ex(return_value, php_pspell_ce); + php_pspell_object_from_zend_object(Z_OBJ_P(return_value))->mgr = to_pspell_manager(ret); } /* }}} */ @@ -207,7 +286,6 @@ PHP_FUNCTION(pspell_new_personal) size_t personal_len, language_len, spelling_len = 0, jargon_len = 0, encoding_len = 0; zend_long mode = Z_L(0), speed = Z_L(0); int argc = ZEND_NUM_ARGS(); - zval *ind; #ifdef PHP_WIN32 TCHAR aspell_dir[200]; @@ -218,7 +296,6 @@ PHP_FUNCTION(pspell_new_personal) #endif PspellCanHaveError *ret; - PspellManager *manager; PspellConfig *config; if (zend_parse_parameters(argc, "ps|sssl", &personal, &personal_len, &language, &language_len, @@ -298,26 +375,22 @@ PHP_FUNCTION(pspell_new_personal) RETURN_FALSE; } - manager = to_pspell_manager(ret); - ind = zend_list_insert(manager, le_pspell); - RETURN_LONG(Z_RES_HANDLE_P(ind)); + object_init_ex(return_value, php_pspell_ce); + php_pspell_object_from_zend_object(Z_OBJ_P(return_value))->mgr = to_pspell_manager(ret); } /* }}} */ /* {{{ Load a dictionary based on the given config */ PHP_FUNCTION(pspell_new_config) { - zend_long conf; - zval *ind; + zval *zcfg; PspellCanHaveError *ret; - PspellManager *manager; PspellConfig *config; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "l", &conf) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "O", &zcfg, php_pspell_config_ce) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_CONFIG; + config = php_pspell_config_object_from_zend_object(Z_OBJ_P(zcfg))->cfg; ret = new_pspell_manager(config); @@ -327,27 +400,24 @@ PHP_FUNCTION(pspell_new_config) RETURN_FALSE; } - manager = to_pspell_manager(ret); - ind = zend_list_insert(manager, le_pspell); - RETURN_LONG(Z_RES_HANDLE_P(ind)); + object_init_ex(return_value, php_pspell_ce); + php_pspell_object_from_zend_object(Z_OBJ_P(return_value))->mgr = to_pspell_manager(ret); } /* }}} */ /* {{{ Returns true if word is valid */ PHP_FUNCTION(pspell_check) { - size_t word_len; - zend_long scin; - char *word; + zval *zmgr; + zend_string *word; PspellManager *manager; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "ls", &scin, &word, &word_len) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "OS", &zmgr, php_pspell_ce, &word) == FAILURE) { RETURN_THROWS(); } + manager = php_pspell_object_from_zend_object(Z_OBJ_P(zmgr))->mgr; - PSPELL_FETCH_MANAGER; - - if (pspell_manager_check(manager, word)) { + if (pspell_manager_check(manager, ZSTR_VAL(word))) { RETURN_TRUE; } else { RETURN_FALSE; @@ -358,22 +428,20 @@ PHP_FUNCTION(pspell_check) /* {{{ Returns array of suggestions */ PHP_FUNCTION(pspell_suggest) { - zend_long scin; - char *word; - size_t word_len; + zval *zmgr; + zend_string *word; PspellManager *manager; const PspellWordList *wl; const char *sug; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "ls", &scin, &word, &word_len) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "OS", &zmgr, php_pspell_ce, &word) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_MANAGER; + manager = php_pspell_object_from_zend_object(Z_OBJ_P(zmgr))->mgr; array_init(return_value); - wl = pspell_manager_suggest(manager, word); + wl = pspell_manager_suggest(manager, ZSTR_VAL(word)); if (wl) { PspellStringEmulation *els = pspell_word_list_elements(wl); while ((sug = pspell_string_emulation_next(els)) != 0) { @@ -390,18 +458,16 @@ PHP_FUNCTION(pspell_suggest) /* {{{ Notify the dictionary of a user-selected replacement */ PHP_FUNCTION(pspell_store_replacement) { - size_t miss_len, corr_len; - zend_long scin; - char *miss, *corr; + zval *zmgr; + zend_string *miss, *corr; PspellManager *manager; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "lss", &scin, &miss, &miss_len, &corr, &corr_len) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "OSS", &zmgr, php_pspell_ce, &miss, &corr) == FAILURE) { RETURN_THROWS(); } + manager = php_pspell_object_from_zend_object(Z_OBJ_P(zmgr))->mgr; - PSPELL_FETCH_MANAGER; - - pspell_manager_store_replacement(manager, miss, corr); + pspell_manager_store_replacement(manager, ZSTR_VAL(miss), ZSTR_VAL(corr)); if (pspell_manager_error_number(manager) == 0) { RETURN_TRUE; } else { @@ -414,23 +480,21 @@ PHP_FUNCTION(pspell_store_replacement) /* {{{ Adds a word to a personal list */ PHP_FUNCTION(pspell_add_to_personal) { - size_t word_len; - zend_long scin; - char *word; + zval *zmgr; + zend_string *word; PspellManager *manager; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "ls", &scin, &word, &word_len) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "OS", &zmgr, php_pspell_ce, &word) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_MANAGER; + manager = php_pspell_object_from_zend_object(Z_OBJ_P(zmgr))->mgr; /*If the word is empty, we have to return; otherwise we'll segfault! ouch!*/ - if (word_len == 0) { + if (ZSTR_LEN(word) == 0) { RETURN_FALSE; } - pspell_manager_add_to_personal(manager, word); + pspell_manager_add_to_personal(manager, ZSTR_VAL(word)); if (pspell_manager_error_number(manager) == 0) { RETURN_TRUE; } else { @@ -443,23 +507,21 @@ PHP_FUNCTION(pspell_add_to_personal) /* {{{ Adds a word to the current session */ PHP_FUNCTION(pspell_add_to_session) { - size_t word_len; - zend_long scin; - char *word; + zval *zmgr; + zend_string *word; PspellManager *manager; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "ls", &scin, &word, &word_len) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "OS", &zmgr, php_pspell_ce, &word) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_MANAGER; + manager = php_pspell_object_from_zend_object(Z_OBJ_P(zmgr))->mgr; /*If the word is empty, we have to return; otherwise we'll segfault! ouch!*/ - if (word_len == 0) { + if (ZSTR_LEN(word) == 0) { RETURN_FALSE; } - pspell_manager_add_to_session(manager, word); + pspell_manager_add_to_session(manager, ZSTR_VAL(word)); if (pspell_manager_error_number(manager) == 0) { RETURN_TRUE; } else { @@ -472,14 +534,13 @@ PHP_FUNCTION(pspell_add_to_session) /* {{{ Clears the current session */ PHP_FUNCTION(pspell_clear_session) { - zend_long scin; + zval *zmgr; PspellManager *manager; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "l", &scin) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "O", &zmgr, php_pspell_ce) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_MANAGER; + manager = php_pspell_object_from_zend_object(Z_OBJ_P(zmgr))->mgr; pspell_manager_clear_session(manager); if (pspell_manager_error_number(manager) == 0) { @@ -494,14 +555,13 @@ PHP_FUNCTION(pspell_clear_session) /* {{{ Saves the current (personal) wordlist */ PHP_FUNCTION(pspell_save_wordlist) { - zend_long scin; + zval *zmgr; PspellManager *manager; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "l", &scin) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "O", &zmgr, php_pspell_ce) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_MANAGER; + manager = php_pspell_object_from_zend_object(Z_OBJ_P(zmgr))->mgr; pspell_manager_save_all_word_lists(manager); @@ -520,7 +580,6 @@ PHP_FUNCTION(pspell_config_create) { char *language, *spelling = NULL, *jargon = NULL, *encoding = NULL; size_t language_len, spelling_len = 0, jargon_len = 0, encoding_len = 0; - zval *ind; PspellConfig *config; #ifdef PHP_WIN32 @@ -577,23 +636,22 @@ PHP_FUNCTION(pspell_config_create) which is not what we want */ pspell_config_replace(config, "save-repl", "false"); - ind = zend_list_insert(config, le_pspell_config); - RETURN_LONG(Z_RES_HANDLE_P(ind)); + object_init_ex(return_value, php_pspell_config_ce); + php_pspell_config_object_from_zend_object(Z_OBJ_P(return_value))->cfg = config; } /* }}} */ /* {{{ Consider run-together words as valid components */ PHP_FUNCTION(pspell_config_runtogether) { - zend_long conf; + zval *zcfg; zend_bool runtogether; PspellConfig *config; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "lb", &conf, &runtogether) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Ob", &zcfg, php_pspell_config_ce, &runtogether) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_CONFIG; + config = php_pspell_config_object_from_zend_object(Z_OBJ_P(zcfg))->cfg; pspell_config_replace(config, "run-together", runtogether ? "true" : "false"); @@ -604,14 +662,14 @@ PHP_FUNCTION(pspell_config_runtogether) /* {{{ Select mode for config (PSPELL_FAST, PSPELL_NORMAL or PSPELL_BAD_SPELLERS) */ PHP_FUNCTION(pspell_config_mode) { - zend_long conf, mode; + zval *zcfg; + zend_long mode; PspellConfig *config; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "ll", &conf, &mode) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Ol", &zcfg, php_pspell_config_ce, &mode) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_CONFIG; + config = php_pspell_config_object_from_zend_object(Z_OBJ_P(zcfg))->cfg; /* First check what mode we want (how many suggestions) */ if (mode == PSPELL_FAST) { @@ -630,14 +688,14 @@ PHP_FUNCTION(pspell_config_mode) PHP_FUNCTION(pspell_config_ignore) { char ignore_str[MAX_LENGTH_OF_LONG + 1]; - zend_long conf, ignore = 0L; + zval *zcfg; + zend_long ignore = 0L; PspellConfig *config; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "ll", &conf, &ignore) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Ol", &zcfg, php_pspell_config_ce, &ignore) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_CONFIG; + config = php_pspell_config_object_from_zend_object(Z_OBJ_P(zcfg))->cfg; snprintf(ignore_str, sizeof(ignore_str), ZEND_LONG_FMT, ignore); @@ -648,22 +706,20 @@ PHP_FUNCTION(pspell_config_ignore) static void pspell_config_path(INTERNAL_FUNCTION_PARAMETERS, char *option) { - zend_long conf; - char *value; - size_t value_len; + zval *zcfg; + zend_string *value; PspellConfig *config; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "lp", &conf, &value, &value_len) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "OP", &zcfg, php_pspell_config_ce, &value) == FAILURE) { RETURN_THROWS(); } + config = php_pspell_config_object_from_zend_object(Z_OBJ_P(zcfg))->cfg; - PSPELL_FETCH_CONFIG; - - if (php_check_open_basedir(value)) { + if (php_check_open_basedir(ZSTR_VAL(value))) { RETURN_FALSE; } - pspell_config_replace(config, option, value); + pspell_config_replace(config, option, ZSTR_VAL(value)); RETURN_TRUE; } @@ -692,24 +748,22 @@ PHP_FUNCTION(pspell_config_data_dir) /* {{{ Use a personal dictionary with replacement pairs for this config */ PHP_FUNCTION(pspell_config_repl) { - zend_long conf; - char *repl; - size_t repl_len; + zval *zcfg; + zend_string *repl; PspellConfig *config; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "lp", &conf, &repl, &repl_len) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "OP", &zcfg, php_pspell_config_ce, &repl) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_CONFIG; + config = php_pspell_config_object_from_zend_object(Z_OBJ_P(zcfg))->cfg; pspell_config_replace(config, "save-repl", "true"); - if (php_check_open_basedir(repl)) { + if (php_check_open_basedir(ZSTR_VAL(repl))) { RETURN_FALSE; } - pspell_config_replace(config, "repl", repl); + pspell_config_replace(config, "repl", ZSTR_VAL(repl)); RETURN_TRUE; } @@ -718,15 +772,14 @@ PHP_FUNCTION(pspell_config_repl) /* {{{ Save replacement pairs when personal list is saved for this config */ PHP_FUNCTION(pspell_config_save_repl) { - zend_long conf; + zval *zcfg; zend_bool save; PspellConfig *config; - if (zend_parse_parameters(ZEND_NUM_ARGS(), "lb", &conf, &save) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS(), "Ob", &zcfg, php_pspell_config_ce, &save) == FAILURE) { RETURN_THROWS(); } - - PSPELL_FETCH_CONFIG; + config = php_pspell_config_object_from_zend_object(Z_OBJ_P(zcfg))->cfg; pspell_config_replace(config, "save-repl", save ? "true" : "false"); diff --git a/ext/pspell/pspell.stub.php b/ext/pspell/pspell.stub.php index cf15fb35eb..a10de4b9f9 100644 --- a/ext/pspell/pspell.stub.php +++ b/ext/pspell/pspell.stub.php @@ -2,7 +2,10 @@ /** @generate-function-entries */ -function pspell_new(string $language, string $spelling = "", string $jargon = "", string $encoding = "", int $mode = 0): int|false {} +final class PSpell {} +final class PSpellConfig {} + +function pspell_new(string $language, string $spelling = "", string $jargon = "", string $encoding = "", int $mode = 0): PSpell|false {} function pspell_new_personal( string $filename, @@ -11,38 +14,24 @@ function pspell_new_personal( string $jargon = "", string $encoding = "", int $mode = 0 -): int|false {} - -function pspell_new_config(int $config): int|false {} - -function pspell_check(int $dictionary, string $word): bool {} - -function pspell_suggest(int $dictionary, string $word): array|false {} - -function pspell_store_replacement(int $dictionary, string $misspelled, string $correct): bool {} - -function pspell_add_to_personal(int $dictionary, string $word): bool {} - -function pspell_add_to_session(int $dictionary, string $word): bool {} - -function pspell_clear_session(int $dictionary): bool {} - -function pspell_save_wordlist(int $dictionary): bool {} - -function pspell_config_create(string $language, string $spelling = "", string $jargon = "", string $encoding = ""): int {} - -function pspell_config_runtogether(int $config, bool $allow): bool {} - -function pspell_config_mode(int $config, int $mode): bool {} - -function pspell_config_ignore(int $config, int $min_length): bool {} - -function pspell_config_personal(int $config, string $filename): bool {} - -function pspell_config_dict_dir(int $config, string $directory): bool {} - -function pspell_config_data_dir(int $config, string $directory): bool {} - -function pspell_config_repl(int $config, string $filename): bool {} - -function pspell_config_save_repl(int $config, bool $save): bool {} +): PSpell|false {} + +function pspell_new_config(PSpellConfig $config): PSpell|false {} + +function pspell_check(PSpell $dictionary, string $word): bool {} +function pspell_suggest(PSpell $dictionary, string $word): array|false {} +function pspell_store_replacement(PSpell $dictionary, string $misspelled, string $correct): bool {} +function pspell_add_to_personal(PSpell $dictionary, string $word): bool {} +function pspell_add_to_session(PSpell $dictionary, string $word): bool {} +function pspell_clear_session(PSpell $dictionary): bool {} +function pspell_save_wordlist(PSpell $dictionary): bool {} + +function pspell_config_create(string $language, string $spelling = "", string $jargon = "", string $encoding = ""): PSpellConfig {} +function pspell_config_runtogether(PSpellConfig $config, bool $allow): bool {} +function pspell_config_mode(PSpellConfig $config, int $mode): bool {} +function pspell_config_ignore(PSpellConfig $config, int $min_length): bool {} +function pspell_config_personal(PSpellConfig $config, string $filename): bool {} +function pspell_config_dict_dir(PSpellConfig $config, string $directory): bool {} +function pspell_config_data_dir(PSpellConfig $config, string $directory): bool {} +function pspell_config_repl(PSpellConfig $config, string $filename): bool {} +function pspell_config_save_repl(PSpellConfig $config, bool $save): bool {} diff --git a/ext/pspell/pspell_arginfo.h b/ext/pspell/pspell_arginfo.h index 102907461c..f3b7b55310 100644 --- a/ext/pspell/pspell_arginfo.h +++ b/ext/pspell/pspell_arginfo.h @@ -1,7 +1,7 @@ /* This is a generated file, edit the .stub.php file instead. - * Stub hash: 1d201126072be285aa2dd3c0136cb7cc261f5841 */ + * Stub hash: efc4c20a1507cfba58ec265a2fe6f61b1e0f8a61 */ -ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_pspell_new, 0, 1, MAY_BE_LONG|MAY_BE_FALSE) +ZEND_BEGIN_ARG_WITH_RETURN_OBJ_TYPE_MASK_EX(arginfo_pspell_new, 0, 1, PSpell, MAY_BE_FALSE) ZEND_ARG_TYPE_INFO(0, language, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, spelling, IS_STRING, 0, "\"\"") ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, jargon, IS_STRING, 0, "\"\"") @@ -9,7 +9,7 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_pspell_new, 0, 1, MAY_BE_LONG|MA ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "0") ZEND_END_ARG_INFO() -ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_pspell_new_personal, 0, 2, MAY_BE_LONG|MAY_BE_FALSE) +ZEND_BEGIN_ARG_WITH_RETURN_OBJ_TYPE_MASK_EX(arginfo_pspell_new_personal, 0, 2, PSpell, MAY_BE_FALSE) ZEND_ARG_TYPE_INFO(0, filename, IS_STRING, 0) ZEND_ARG_TYPE_INFO(0, language, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, spelling, IS_STRING, 0, "\"\"") @@ -18,22 +18,22 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_pspell_new_personal, 0, 2, MAY_B ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, mode, IS_LONG, 0, "0") ZEND_END_ARG_INFO() -ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_pspell_new_config, 0, 1, MAY_BE_LONG|MAY_BE_FALSE) - ZEND_ARG_TYPE_INFO(0, config, IS_LONG, 0) +ZEND_BEGIN_ARG_WITH_RETURN_OBJ_TYPE_MASK_EX(arginfo_pspell_new_config, 0, 1, PSpell, MAY_BE_FALSE) + ZEND_ARG_OBJ_INFO(0, config, PSpellConfig, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_check, 0, 2, _IS_BOOL, 0) - ZEND_ARG_TYPE_INFO(0, dictionary, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, dictionary, PSpell, 0) ZEND_ARG_TYPE_INFO(0, word, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_MASK_EX(arginfo_pspell_suggest, 0, 2, MAY_BE_ARRAY|MAY_BE_FALSE) - ZEND_ARG_TYPE_INFO(0, dictionary, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, dictionary, PSpell, 0) ZEND_ARG_TYPE_INFO(0, word, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_store_replacement, 0, 3, _IS_BOOL, 0) - ZEND_ARG_TYPE_INFO(0, dictionary, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, dictionary, PSpell, 0) ZEND_ARG_TYPE_INFO(0, misspelled, IS_STRING, 0) ZEND_ARG_TYPE_INFO(0, correct, IS_STRING, 0) ZEND_END_ARG_INFO() @@ -43,12 +43,12 @@ ZEND_END_ARG_INFO() #define arginfo_pspell_add_to_session arginfo_pspell_check ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_clear_session, 0, 1, _IS_BOOL, 0) - ZEND_ARG_TYPE_INFO(0, dictionary, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, dictionary, PSpell, 0) ZEND_END_ARG_INFO() #define arginfo_pspell_save_wordlist arginfo_pspell_clear_session -ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_config_create, 0, 1, IS_LONG, 0) +ZEND_BEGIN_ARG_WITH_RETURN_OBJ_INFO_EX(arginfo_pspell_config_create, 0, 1, PSpellConfig, 0) ZEND_ARG_TYPE_INFO(0, language, IS_STRING, 0) ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, spelling, IS_STRING, 0, "\"\"") ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, jargon, IS_STRING, 0, "\"\"") @@ -56,27 +56,27 @@ ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_config_create, 0, 1, IS_L ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_config_runtogether, 0, 2, _IS_BOOL, 0) - ZEND_ARG_TYPE_INFO(0, config, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, config, PSpellConfig, 0) ZEND_ARG_TYPE_INFO(0, allow, _IS_BOOL, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_config_mode, 0, 2, _IS_BOOL, 0) - ZEND_ARG_TYPE_INFO(0, config, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, config, PSpellConfig, 0) ZEND_ARG_TYPE_INFO(0, mode, IS_LONG, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_config_ignore, 0, 2, _IS_BOOL, 0) - ZEND_ARG_TYPE_INFO(0, config, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, config, PSpellConfig, 0) ZEND_ARG_TYPE_INFO(0, min_length, IS_LONG, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_config_personal, 0, 2, _IS_BOOL, 0) - ZEND_ARG_TYPE_INFO(0, config, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, config, PSpellConfig, 0) ZEND_ARG_TYPE_INFO(0, filename, IS_STRING, 0) ZEND_END_ARG_INFO() ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_config_dict_dir, 0, 2, _IS_BOOL, 0) - ZEND_ARG_TYPE_INFO(0, config, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, config, PSpellConfig, 0) ZEND_ARG_TYPE_INFO(0, directory, IS_STRING, 0) ZEND_END_ARG_INFO() @@ -85,7 +85,7 @@ ZEND_END_ARG_INFO() #define arginfo_pspell_config_repl arginfo_pspell_config_personal ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_pspell_config_save_repl, 0, 2, _IS_BOOL, 0) - ZEND_ARG_TYPE_INFO(0, config, IS_LONG, 0) + ZEND_ARG_OBJ_INFO(0, config, PSpellConfig, 0) ZEND_ARG_TYPE_INFO(0, save, _IS_BOOL, 0) ZEND_END_ARG_INFO() @@ -133,3 +133,13 @@ static const zend_function_entry ext_functions[] = { ZEND_FE(pspell_config_save_repl, arginfo_pspell_config_save_repl) ZEND_FE_END }; + + +static const zend_function_entry class_PSpell_methods[] = { + ZEND_FE_END +}; + + +static const zend_function_entry class_PSpellConfig_methods[] = { + ZEND_FE_END +}; diff --git a/ext/pspell/tests/003.phpt b/ext/pspell/tests/003.phpt index e8677ae826..63291a8d42 100644 --- a/ext/pspell/tests/003.phpt +++ b/ext/pspell/tests/003.phpt @@ -34,7 +34,7 @@ var_dump(pspell_config_ignore($cfg, PHP_INT_MAX)); bool(false) Warning: pspell_new_config(): PSPELL couldn't open the dictionary. reason: The encoding "b0rked" is not known. This could also mean that the file "%sb0rked.%s" could not be opened for reading or does not exist. in %s003.php on line 9 -pspell_check(): 0 is not a PSPELL result index +pspell_check(): Argument #1 ($dictionary) must be of type PSpell, bool given --- bool(true) bool(true) diff --git a/ext/session/tests/003.phpt b/ext/session/tests/003.phpt index e7e4294375..4f9b37b75d 100644 --- a/ext/session/tests/003.phpt +++ b/ext/session/tests/003.phpt @@ -17,7 +17,7 @@ class foo { function method() { $this->yes++; } } -session_id("abtest"); +session_id("test003"); session_start(); session_decode('baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:1;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:1;}}'); diff --git a/ext/session/tests/004.phpt b/ext/session/tests/004.phpt index 4b81a088b2..c3ed0c5616 100644 --- a/ext/session/tests/004.phpt +++ b/ext/session/tests/004.phpt @@ -55,7 +55,7 @@ class foo { session_set_save_handler(array($hnd, "open"), array($hnd, "close"), array($hnd, "read"), array($hnd, "write"), array($hnd, "destroy"), array($hnd, "gc")); -session_id("abtest"); +session_id("test004"); session_start(); $_SESSION["baz"]->method(); $_SESSION["arr"][3]->method(); @@ -75,7 +75,7 @@ session_destroy(); ?> --EXPECT-- OPEN: PHPSESSID -READ: abtest +READ: test004 object(foo)#2 (2) { ["bar"]=> string(2) "ok" @@ -91,9 +91,9 @@ array(1) { int(2) } } -WRITE: abtest, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}} +WRITE: test004, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}} OPEN: PHPSESSID -READ: abtest +READ: test004 object(foo)#3 (2) { ["bar"]=> string(2) "ok" @@ -109,4 +109,4 @@ array(1) { int(2) } } -DESTROY: abtest +DESTROY: test004 diff --git a/ext/session/tests/005.phpt b/ext/session/tests/005.phpt index a7b42de7fa..2b51ceed94 100644 --- a/ext/session/tests/005.phpt +++ b/ext/session/tests/005.phpt @@ -56,7 +56,7 @@ class foo { session_set_save_handler(array($hnd, "open"), array($hnd, "close"), array($hnd, "read"), array($hnd, "write"), array($hnd, "destroy"), array($hnd, "gc")); -session_id("abtest"); +session_id("test005"); session_start(); session_decode($hnd->data); @@ -91,7 +91,7 @@ session_destroy(); ?> --EXPECT-- OPEN: PHPSESSID -READ: abtest +READ: test005 object(foo)#4 (2) { ["bar"]=> string(2) "ok" @@ -107,10 +107,10 @@ array(1) { int(2) } } -WRITE: abtest, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}} +WRITE: test005, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}} CLOSE OPEN: PHPSESSID -READ: abtest +READ: test005 object(foo)#2 (2) { ["bar"]=> string(2) "ok" @@ -127,10 +127,10 @@ array(1) { } } int(123) -WRITE: abtest, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:3;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:3;}}c|i:123; +WRITE: test005, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:3;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:3;}}c|i:123; CLOSE OPEN: PHPSESSID -READ: abtest +READ: test005 object(foo)#4 (2) { ["bar"]=> string(2) "ok" @@ -147,5 +147,5 @@ array(1) { } } int(123) -DESTROY: abtest +DESTROY: test005 CLOSE diff --git a/ext/session/tests/006.phpt b/ext/session/tests/006.phpt index 68cea7314a..e3c048748f 100644 --- a/ext/session/tests/006.phpt +++ b/ext/session/tests/006.phpt @@ -13,7 +13,7 @@ session.save_handler=files error_reporting(E_ALL); ob_start(); -session_id("abtest"); +session_id("test006"); session_start(); class a { diff --git a/ext/session/tests/009.phpt b/ext/session/tests/009.phpt index 10069c21d0..898cf76a5a 100644 --- a/ext/session/tests/009.phpt +++ b/ext/session/tests/009.phpt @@ -13,14 +13,14 @@ session.save_handler=files error_reporting(E_ALL); ob_start(); -session_id("abtest"); +session_id("test009"); ### Phase 1 cleanup session_start(); session_destroy(); ### Phase 2 $_SESSION["c"] does not contain any value -session_id("abtest"); +session_id("test009"); session_start(); var_dump($_SESSION); $_SESSION["name"] = "foo"; diff --git a/ext/session/tests/012.phpt b/ext/session/tests/012.phpt index c555d2ca1e..ad3f54962e 100644 --- a/ext/session/tests/012.phpt +++ b/ext/session/tests/012.phpt @@ -14,7 +14,7 @@ error_reporting(E_ALL); ### Absurd example, value of $_SESSION does not matter -session_id("abtest"); +session_id("test012"); session_start(); $_SESSION["_SESSION"] = Array(); $_SESSION = "kk"; diff --git a/ext/session/tests/013.phpt b/ext/session/tests/013.phpt index 32909eb58c..07fcea407b 100644 --- a/ext/session/tests/013.phpt +++ b/ext/session/tests/013.phpt @@ -12,10 +12,10 @@ session.save_handler=files <?php error_reporting(E_ALL); -session_id("abtest"); +session_id("test013"); session_start(); session_destroy(); -session_id("abtest2"); +session_id("test013-2"); session_start(); session_destroy(); diff --git a/ext/session/tests/014.phpt b/ext/session/tests/014.phpt index d8369cfec5..a0c6515243 100644 --- a/ext/session/tests/014.phpt +++ b/ext/session/tests/014.phpt @@ -14,7 +14,7 @@ session.save_handler=files <?php error_reporting(E_ALL); -session_id("abtest"); +session_id("test014"); session_start(); ?> diff --git a/ext/session/tests/015.phpt b/ext/session/tests/015.phpt index 527b86bc1d..788f1eac06 100644 --- a/ext/session/tests/015.phpt +++ b/ext/session/tests/015.phpt @@ -16,7 +16,7 @@ session.save_handler=files <?php error_reporting(E_ALL); -session_id("abtest"); +session_id("test015"); session_start(); ?> <a href="/link?<?php echo SID; ?>"> @@ -24,4 +24,4 @@ session_start(); session_destroy(); ?> --EXPECT-- -<a href="/link?PHPSESSID=abtest&PHPSESSID=abtest"> +<a href="/link?PHPSESSID=test015&PHPSESSID=test015"> diff --git a/ext/session/tests/018.phpt b/ext/session/tests/018.phpt index 5ec132b34f..d132abf717 100644 --- a/ext/session/tests/018.phpt +++ b/ext/session/tests/018.phpt @@ -16,7 +16,7 @@ session.save_handler=files error_reporting(E_ALL); -session_id("abtest"); +session_id("test018"); session_start(); ?> <form accept-charset="ISO-8859-15, ISO-8859-1" action=url.php> @@ -24,4 +24,4 @@ session_start(); session_destroy(); ?> --EXPECT-- -<form accept-charset="ISO-8859-15, ISO-8859-1" action=url.php><input type="hidden" name="PHPSESSID" value="abtest" /> +<form accept-charset="ISO-8859-15, ISO-8859-1" action=url.php><input type="hidden" name="PHPSESSID" value="test018" /> diff --git a/ext/session/tests/019.phpt b/ext/session/tests/019.phpt index d9a478245b..e3793a5522 100644 --- a/ext/session/tests/019.phpt +++ b/ext/session/tests/019.phpt @@ -23,7 +23,7 @@ class TFoo { } } -session_id("abtest"); +session_id("test019"); session_start(); $_SESSION["o1"] = new TFoo(42); diff --git a/ext/session/tests/020.phpt b/ext/session/tests/020.phpt index 267e52191c..f4ec38c8fb 100644 --- a/ext/session/tests/020.phpt +++ b/ext/session/tests/020.phpt @@ -17,7 +17,7 @@ session.save_handler=files error_reporting(E_ALL); -session_id("abtest"); +session_id("test020"); session_start(); ?> <a href="link.php?a=b"> @@ -25,4 +25,4 @@ session_start(); session_destroy(); ?> --EXPECT-- -<a href="link.php?a=b&PHPSESSID=abtest"> +<a href="link.php?a=b&PHPSESSID=test020"> diff --git a/ext/session/tests/021.phpt b/ext/session/tests/021.phpt index 4a97d7d32a..2532324fee 100644 --- a/ext/session/tests/021.phpt +++ b/ext/session/tests/021.phpt @@ -19,7 +19,7 @@ error_reporting(E_ALL); ini_set('session.trans_sid_hosts', 'php.net'); $_SERVER['HTTP_HOST'] = 'php.net'; -session_id("abtest"); +session_id("test021"); session_start(); ?> <form action="//bad.net/do.php"> @@ -60,11 +60,11 @@ session_destroy(); --EXPECT-- <form action="//bad.net/do.php"> <fieldset> -<form action="//php.net/do.php"><input type="hidden" name="PHPSESSID" value="abtest" /> +<form action="//php.net/do.php"><input type="hidden" name="PHPSESSID" value="test021" /> <fieldset> -<form action="../do.php"><input type="hidden" name="PHPSESSID" value="abtest" /> +<form action="../do.php"><input type="hidden" name="PHPSESSID" value="test021" /> <fieldset> -<form action="/do.php"><input type="hidden" name="PHPSESSID" value="abtest" /> +<form action="/do.php"><input type="hidden" name="PHPSESSID" value="test021" /> <fieldset> -<form action="/foo/do.php"><input type="hidden" name="PHPSESSID" value="abtest" /> +<form action="/foo/do.php"><input type="hidden" name="PHPSESSID" value="test021" /> <fieldset> diff --git a/ext/session/tests/023.phpt b/ext/session/tests/023.phpt index 23c7096f23..e9ce084256 100644 --- a/ext/session/tests/023.phpt +++ b/ext/session/tests/023.phpt @@ -17,7 +17,7 @@ class foo { function method() { $this->yes++; } } -session_id("abtest"); +session_id("test023"); session_start(); session_decode('baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:1;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:1;}}'); $baz = $_SESSION['baz']; diff --git a/ext/session/tests/024.phpt b/ext/session/tests/024.phpt index 174f2e2639..913956f5b3 100644 --- a/ext/session/tests/024.phpt +++ b/ext/session/tests/024.phpt @@ -56,7 +56,7 @@ class foo { session_set_save_handler(array($hnd, "open"), array($hnd, "close"), array($hnd, "read"), array($hnd, "write"), array($hnd, "destroy"), array($hnd, "gc")); -session_id("abtest"); +session_id("test024"); session_start(); $baz = $_SESSION['baz']; @@ -79,7 +79,7 @@ session_destroy(); ?> --EXPECTF-- OPEN: PHPSESSID -READ: abtest +READ: test024 object(foo)#%d (2) { ["bar"]=> string(2) "ok" @@ -95,9 +95,9 @@ array(1) { int(2) } } -WRITE: abtest, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}} +WRITE: test024, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}} OPEN: PHPSESSID -READ: abtest +READ: test024 object(foo)#%d (2) { ["bar"]=> string(2) "ok" @@ -113,4 +113,4 @@ array(1) { int(2) } } -DESTROY: abtest +DESTROY: test024 diff --git a/ext/session/tests/025.phpt b/ext/session/tests/025.phpt index 172ea579c6..4e91da9796 100644 --- a/ext/session/tests/025.phpt +++ b/ext/session/tests/025.phpt @@ -57,7 +57,7 @@ class foo { session_set_save_handler(array($hnd, "open"), array($hnd, "close"), array($hnd, "read"), array($hnd, "write"), array($hnd, "destroy"), array($hnd, "gc")); -session_id("abtest"); +session_id("test025"); session_start(); $baz = $_SESSION['baz']; $arr = $_SESSION['arr']; @@ -93,7 +93,7 @@ session_destroy(); ?> --EXPECTF-- OPEN: PHPSESSID -READ: abtest +READ: test025 object(foo)#%d (2) { ["bar"]=> string(2) "ok" @@ -109,10 +109,10 @@ array(1) { int(2) } } -WRITE: abtest, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}} +WRITE: test025, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:2;}} CLOSE OPEN: PHPSESSID -READ: abtest +READ: test025 object(foo)#%d (2) { ["bar"]=> string(2) "ok" @@ -129,10 +129,10 @@ array(1) { } } int(123) -WRITE: abtest, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:3;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:3;}}c|i:123; +WRITE: test025, baz|O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:3;}arr|a:1:{i:3;O:3:"foo":2:{s:3:"bar";s:2:"ok";s:3:"yes";i:3;}}c|i:123; CLOSE OPEN: PHPSESSID -READ: abtest +READ: test025 object(foo)#%d (2) { ["bar"]=> string(2) "ok" @@ -149,5 +149,5 @@ array(1) { } } int(123) -DESTROY: abtest +DESTROY: test025 CLOSE diff --git a/ext/session/tests/026.phpt b/ext/session/tests/026.phpt index b61039ee0c..f5abdac0a5 100644 --- a/ext/session/tests/026.phpt +++ b/ext/session/tests/026.phpt @@ -13,7 +13,7 @@ session.save_handler=files error_reporting(E_ALL); ob_start(); -session_id("abtest"); +session_id("test026"); session_start(); class a { diff --git a/ext/session/tests/027.phpt b/ext/session/tests/027.phpt index ebaf25467c..969bc5b872 100644 --- a/ext/session/tests/027.phpt +++ b/ext/session/tests/027.phpt @@ -13,14 +13,14 @@ session.save_handler=files error_reporting(E_ALL); ob_start(); -session_id("abtest"); +session_id("test027"); ### Phase 1 cleanup session_start(); session_destroy(); ### Phase 2 $_SESSION["c"] does not contain any value -session_id("abtest"); +session_id("test027"); session_start(); var_dump($_SESSION); $_SESSION["name"] = "foo"; diff --git a/ext/session/tests/CONFLICTS b/ext/session/tests/CONFLICTS deleted file mode 100644 index 30888ae480..0000000000 --- a/ext/session/tests/CONFLICTS +++ /dev/null @@ -1 +0,0 @@ -session diff --git a/ext/session/tests/bug41600.phpt b/ext/session/tests/bug41600.phpt index 79d5e12841..d4dfce9d88 100644 --- a/ext/session/tests/bug41600.phpt +++ b/ext/session/tests/bug41600.phpt @@ -17,7 +17,7 @@ session.save_handler=files error_reporting(E_ALL); -session_id("abtest"); +session_id("bug41600"); session_start(); ?> <a href="link.php?a=b"> @@ -25,4 +25,4 @@ session_start(); session_destroy(); ?> --EXPECT-- -<a href="link.php?a=b&PHPSESSID=abtest"> +<a href="link.php?a=b&PHPSESSID=bug41600"> diff --git a/ext/session/tests/bug42596.phpt b/ext/session/tests/bug42596.phpt index f27e86387e..83567bde5f 100644 --- a/ext/session/tests/bug42596.phpt +++ b/ext/session/tests/bug42596.phpt @@ -15,7 +15,7 @@ session.serialize_handler=php session.save_handler=files --FILE-- <?php -$sessdir = __DIR__.'/sessions/'; +$sessdir = __DIR__.'/bug42596/'; @rmdir($sessdir); mkdir($sessdir); $save_path = '0;0777;'.$sessdir; diff --git a/ext/session/tests/rfc1867.phpt b/ext/session/tests/rfc1867.phpt index 0e189b7aa6..a5ae8b2218 100644 --- a/ext/session/tests/rfc1867.phpt +++ b/ext/session/tests/rfc1867.phpt @@ -18,15 +18,15 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --COOKIE-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867 --GET-- -PHPSESSID=rfc1867-tests-get +PHPSESSID=rfc1867-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHP_SESSION_UPLOAD_PROGRESS" @@ -50,7 +50,7 @@ var_dump($_SESSION["upload_progress_" . basename(__FILE__)]); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867" bool(true) array(2) { ["file1"]=> diff --git a/ext/session/tests/rfc1867_cleanup.phpt b/ext/session/tests/rfc1867_cleanup.phpt index ae5dcbe8dc..9baa6144ba 100644 --- a/ext/session/tests/rfc1867_cleanup.phpt +++ b/ext/session/tests/rfc1867_cleanup.phpt @@ -18,15 +18,15 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --COOKIE-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867-cleanup --GET-- -PHPSESSID=rfc1867-tests-get +PHPSESSID=rfc1867-cleanup-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-cleanup-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHP_SESSION_UPLOAD_PROGRESS" @@ -50,7 +50,7 @@ var_dump(isset($_SESSION["upload_progress_" . basename(__FILE__)])); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-cleanup" bool(true) array(2) { ["file1"]=> diff --git a/ext/session/tests/rfc1867_disabled.phpt b/ext/session/tests/rfc1867_disabled.phpt index 5bedb1de88..8d8effd1f4 100644 --- a/ext/session/tests/rfc1867_disabled.phpt +++ b/ext/session/tests/rfc1867_disabled.phpt @@ -17,15 +17,15 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --COOKIE-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867-disabled --GET-- -PHPSESSID=rfc1867-tests-get +PHPSESSID=rfc1867-disabled-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-disabled-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="file1"; filename="file1.txt" @@ -44,7 +44,7 @@ var_dump(isset($_SESSION["upload_progress_" . basename(__FILE__)])); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-disabled" array(2) { ["file1"]=> array(5) { diff --git a/ext/session/tests/rfc1867_disabled_2.phpt b/ext/session/tests/rfc1867_disabled_2.phpt index d63748369d..c539c6eaea 100644 --- a/ext/session/tests/rfc1867_disabled_2.phpt +++ b/ext/session/tests/rfc1867_disabled_2.phpt @@ -17,15 +17,15 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --COOKIE-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867-disabled-2 --GET-- -PHPSESSID=rfc1867-tests-get +PHPSESSID=rfc1867-disabled-2-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-disabled-2-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="file1"; filename="file1.txt" @@ -44,7 +44,7 @@ var_dump(isset($_SESSION["upload_progress_" . basename(__FILE__)])); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-disabled-2" array(2) { ["file1"]=> array(5) { diff --git a/ext/session/tests/rfc1867_inter.phpt b/ext/session/tests/rfc1867_inter.phpt index b961bbcac3..fd28dfe07a 100644 --- a/ext/session/tests/rfc1867_inter.phpt +++ b/ext/session/tests/rfc1867_inter.phpt @@ -18,15 +18,15 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --COOKIE-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867-inter --GET-- -PHPSESSID=rfc1867-tests-get +PHPSESSID=rfc1867-inter-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-inter-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHP_SESSION_UPLOAD_PROGRESS" @@ -54,7 +54,7 @@ var_dump(isset($_SESSION["upload_progress_" . basename(__FILE__) . "_2"])); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-inter" array(2) { ["file1"]=> array(5) { diff --git a/ext/session/tests/rfc1867_no_name.phpt b/ext/session/tests/rfc1867_no_name.phpt index 98e8e7cf09..15877a664e 100644 --- a/ext/session/tests/rfc1867_no_name.phpt +++ b/ext/session/tests/rfc1867_no_name.phpt @@ -17,15 +17,15 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --COOKIE-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867-no-name --GET-- -PHPSESSID=rfc1867-tests-get +PHPSESSID=rfc1867-no-name-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-no-name-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="file1"; filename="file1.txt" @@ -44,7 +44,7 @@ var_dump(isset($_SESSION["upload_progress_" . basename(__FILE__)])); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-no-name" array(2) { ["file1"]=> array(5) { diff --git a/ext/session/tests/rfc1867_sid_cookie.phpt b/ext/session/tests/rfc1867_sid_cookie.phpt index 804a9ba308..85c28934f4 100644 --- a/ext/session/tests/rfc1867_sid_cookie.phpt +++ b/ext/session/tests/rfc1867_sid_cookie.phpt @@ -17,15 +17,15 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --COOKIE-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867-sid-cookie --GET-- -PHPSESSID=rfc1867-tests-get +PHPSESSID=rfc1867-sid-cookie-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-sid-cookie-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHP_SESSION_UPLOAD_PROGRESS" @@ -49,7 +49,7 @@ var_dump($_SESSION["upload_progress_" . basename(__FILE__)]); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-sid-cookie" bool(true) array(2) { ["file1"]=> diff --git a/ext/session/tests/rfc1867_sid_get.phpt b/ext/session/tests/rfc1867_sid_get.phpt index 17d1ebce0f..dfb192cb47 100644 --- a/ext/session/tests/rfc1867_sid_get.phpt +++ b/ext/session/tests/rfc1867_sid_get.phpt @@ -17,13 +17,13 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --GET-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867-sid-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-sid-get-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHP_SESSION_UPLOAD_PROGRESS" @@ -47,7 +47,7 @@ var_dump($_SESSION["upload_progress_" . basename(__FILE__)]); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-sid-get" bool(true) array(2) { ["file1"]=> diff --git a/ext/session/tests/rfc1867_sid_get_2.phpt b/ext/session/tests/rfc1867_sid_get_2.phpt index 4b32c6e352..33e4489cc8 100644 --- a/ext/session/tests/rfc1867_sid_get_2.phpt +++ b/ext/session/tests/rfc1867_sid_get_2.phpt @@ -17,15 +17,15 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --COOKIE-- -PHPSESSID=rfc1867-tests-cookie +PHPSESSID=rfc1867-sid-get-2-cookie --GET-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867-sid-get-2 --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-sid-get-2-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHP_SESSION_UPLOAD_PROGRESS" @@ -49,7 +49,7 @@ var_dump($_SESSION["upload_progress_" . basename(__FILE__)]); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-sid-get-2" bool(true) array(2) { ["file1"]=> diff --git a/ext/session/tests/rfc1867_sid_only_cookie.phpt b/ext/session/tests/rfc1867_sid_only_cookie.phpt index 0c376bb026..54897b91c8 100644 --- a/ext/session/tests/rfc1867_sid_only_cookie.phpt +++ b/ext/session/tests/rfc1867_sid_only_cookie.phpt @@ -17,15 +17,15 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --COOKIE-- -PHPSESSID=rfc1867-tests +PHPSESSID=rfc1867-sid-only-cookie --GET-- -PHPSESSID=rfc1867-tests-get +PHPSESSID=rfc1867-sid-only-cookie-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-sid-only-cookie-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHP_SESSION_UPLOAD_PROGRESS" @@ -49,7 +49,7 @@ var_dump($_SESSION["upload_progress_" . basename(__FILE__)]); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-sid-only-cookie" bool(true) array(2) { ["file1"]=> diff --git a/ext/session/tests/rfc1867_sid_only_cookie_2.phpt b/ext/session/tests/rfc1867_sid_only_cookie_2.phpt index 9b0aae9429..3fd46148d7 100644 --- a/ext/session/tests/rfc1867_sid_only_cookie_2.phpt +++ b/ext/session/tests/rfc1867_sid_only_cookie_2.phpt @@ -16,13 +16,13 @@ session.save_handler=files --SKIPIF-- <?php include('skipif.inc'); ?> --GET-- -PHPSESSID=rfc1867-tests-get +PHPSESSID=rfc1867-sid-only-cookie-2-get --POST_RAW-- Content-Type: multipart/form-data; boundary=---------------------------20896060251896012921717172737 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests-post +rfc1867-sid-only-cookie-2-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHP_SESSION_UPLOAD_PROGRESS" diff --git a/ext/session/tests/rfc1867_sid_post.phpt b/ext/session/tests/rfc1867_sid_post.phpt index a126f9ba38..f22f153403 100644 --- a/ext/session/tests/rfc1867_sid_post.phpt +++ b/ext/session/tests/rfc1867_sid_post.phpt @@ -21,7 +21,7 @@ Content-Type: multipart/form-data; boundary=---------------------------208960602 -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHPSESSID" -rfc1867-tests +rfc1867-sid-post -----------------------------20896060251896012921717172737 Content-Disposition: form-data; name="PHP_SESSION_UPLOAD_PROGRESS" @@ -45,7 +45,7 @@ var_dump($_SESSION["upload_progress_" . basename(__FILE__)]); session_destroy(); ?> --EXPECTF-- -string(%d) "rfc1867-tests" +string(%d) "rfc1867-sid-post" bool(true) array(2) { ["file1"]=> diff --git a/ext/session/tests/session_basic1.phpt b/ext/session/tests/session_basic1.phpt index ffdcf6c5ec..13c21d3e85 100644 --- a/ext/session/tests/session_basic1.phpt +++ b/ext/session/tests/session_basic1.phpt @@ -17,7 +17,7 @@ ob_start(); echo "*** Testing basic session functionality : variation1 ***\n"; -$session_id = 'testid'; +$session_id = 'session-basic1'; session_id($session_id); $path = __DIR__; var_dump(session_save_path($path)); @@ -47,20 +47,20 @@ ob_end_flush(); *** Testing basic session functionality : variation1 *** string(0) "" *** Without lazy_write *** -string(6) "testid" +string(14) "session-basic1" bool(true) array(1) { ["lazy_write"]=> bool(false) } bool(true) -string(6) "testid" +string(14) "session-basic1" *** With lazy_write *** -string(6) "testid" +string(14) "session-basic1" bool(true) bool(true) -string(6) "testid" +string(14) "session-basic1" *** Cleanup *** -string(6) "testid" +string(14) "session-basic1" bool(true) bool(true) diff --git a/ext/session/tests/session_basic2.phpt b/ext/session/tests/session_basic2.phpt index cb9585303c..68230f89b7 100644 --- a/ext/session/tests/session_basic2.phpt +++ b/ext/session/tests/session_basic2.phpt @@ -19,7 +19,7 @@ ob_start(); echo "*** Testing basic session functionality : variation2 ***\n"; -$session_id = 'testid'; +$session_id = 'session-basic2'; session_id($session_id); $path = __DIR__; var_dump(session_save_path($path)); @@ -55,7 +55,7 @@ ob_end_flush(); *** Testing basic session functionality : variation2 *** string(0) "" *** Without lazy_write *** -string(6) "testid" +string(14) "session-basic2" bool(true) bool(true) bool(true) diff --git a/ext/session/tests/session_basic3.phpt b/ext/session/tests/session_basic3.phpt index f6ed10e296..5f6bd72b82 100644 --- a/ext/session/tests/session_basic3.phpt +++ b/ext/session/tests/session_basic3.phpt @@ -93,7 +93,7 @@ output_reset_rewrite_vars(); echo "*** Test trans sid ***\n"; ob_start(); -$session_id = 'testid'; +$session_id = 'session-basic3'; session_id($session_id); session_start(); // Should add session ID to relative URL only for SECURITY @@ -225,39 +225,39 @@ ob_end_flush(); *** Testing basic session functionality : variation3 use_trans_sid *** *** Test trans sid *** -<a href="/?PHPSESSID=testid">test</a> -<a href="/path?PHPSESSID=testid">test</a> -<a href="/path/?PHPSESSID=testid">test</a> -<a href="/path/?foo=var&PHPSESSID=testid">test</a> -<a href="../?PHPSESSID=testid">test</a> -<a href="../path?PHPSESSID=testid">test</a> -<a href="../path/?PHPSESSID=testid">test</a> -<a href="../path/?foo=var&PHPSESSID=testid">test</a> +<a href="/?PHPSESSID=session-basic3">test</a> +<a href="/path?PHPSESSID=session-basic3">test</a> +<a href="/path/?PHPSESSID=session-basic3">test</a> +<a href="/path/?foo=var&PHPSESSID=session-basic3">test</a> +<a href="../?PHPSESSID=session-basic3">test</a> +<a href="../path?PHPSESSID=session-basic3">test</a> +<a href="../path/?PHPSESSID=session-basic3">test</a> +<a href="../path/?foo=var&PHPSESSID=session-basic3">test</a> -<a href="/?PHPSESSID=testid#bar">test</a> -<a href="/path/?PHPSESSID=testid#bar">test</a> -<a href="/path/?foo=var&PHPSESSID=testid#bar">test</a> -<a href="../?PHPSESSID=testid#bar">test</a> -<a href="../path/?PHPSESSID=testid#bar">test</a> -<a href="../path/?foo=var&PHPSESSID=testid#bar">test</a> +<a href="/?PHPSESSID=session-basic3#bar">test</a> +<a href="/path/?PHPSESSID=session-basic3#bar">test</a> +<a href="/path/?foo=var&PHPSESSID=session-basic3#bar">test</a> +<a href="../?PHPSESSID=session-basic3#bar">test</a> +<a href="../path/?PHPSESSID=session-basic3#bar">test</a> +<a href="../path/?foo=var&PHPSESSID=session-basic3#bar">test</a> -<a href="/?foo&PHPSESSID=testid">test</a> -<a href="/?foo&PHPSESSID=testid#bar">test</a> -<a href="/?foo=var&PHPSESSID=testid">test</a> -<a href="/?foo=var&PHPSESSID=testid#bar">test</a> -<a href="../?foo&PHPSESSID=testid">test</a> -<a href="../?foo&PHPSESSID=testid#bar">test</a> -<a href="../?foo=var&PHPSESSID=testid">test</a> -<a href="../?foo=var&PHPSESSID=testid#bar">test</a> +<a href="/?foo&PHPSESSID=session-basic3">test</a> +<a href="/?foo&PHPSESSID=session-basic3#bar">test</a> +<a href="/?foo=var&PHPSESSID=session-basic3">test</a> +<a href="/?foo=var&PHPSESSID=session-basic3#bar">test</a> +<a href="../?foo&PHPSESSID=session-basic3">test</a> +<a href="../?foo&PHPSESSID=session-basic3#bar">test</a> +<a href="../?foo=var&PHPSESSID=session-basic3">test</a> +<a href="../?foo=var&PHPSESSID=session-basic3#bar">test</a> -<a href="file.php?PHPSESSID=testid">test</a> -<a href="file.php?foo&PHPSESSID=testid">test</a> -<a href="file.php?foo=var&PHPSESSID=testid">test</a> -<a href="file.php?foo=var&PHPSESSID=testid#bar">test</a> -<a href="../file.php?PHPSESSID=testid">test</a> -<a href="../file.php?foo&PHPSESSID=testid">test</a> -<a href="../file.php?foo=var&PHPSESSID=testid">test</a> -<a href="../file.php?foo=var&PHPSESSID=testid#bar">test</a> +<a href="file.php?PHPSESSID=session-basic3">test</a> +<a href="file.php?foo&PHPSESSID=session-basic3">test</a> +<a href="file.php?foo=var&PHPSESSID=session-basic3">test</a> +<a href="file.php?foo=var&PHPSESSID=session-basic3#bar">test</a> +<a href="../file.php?PHPSESSID=session-basic3">test</a> +<a href="../file.php?foo&PHPSESSID=session-basic3">test</a> +<a href="../file.php?foo=var&PHPSESSID=session-basic3">test</a> +<a href="../file.php?foo=var&PHPSESSID=session-basic3#bar">test</a> <a href="http://php.net">test</a> <a href="http://php.net/">test</a> @@ -310,19 +310,19 @@ ob_end_flush(); <a href="//php.net/some/path/file.php?foo=var">test</a> <a href="//php.net/some/path/file.php?foo=var#bar">test</a> -<form action="script.php" method="post"><input type="hidden" name="PHPSESSID" value="testid" /> +<form action="script.php" method="post"><input type="hidden" name="PHPSESSID" value="session-basic3" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> -<form action="../script.php" method="post"><input type="hidden" name="PHPSESSID" value="testid" /> +<form action="../script.php" method="post"><input type="hidden" name="PHPSESSID" value="session-basic3" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> -<form action="/path/script.php" method="post"><input type="hidden" name="PHPSESSID" value="testid" /> +<form action="/path/script.php" method="post"><input type="hidden" name="PHPSESSID" value="session-basic3" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> -<form action="../path/script.php" method="post"><input type="hidden" name="PHPSESSID" value="testid" /> +<form action="../path/script.php" method="post"><input type="hidden" name="PHPSESSID" value="session-basic3" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> @@ -341,5 +341,5 @@ ob_end_flush(); bool(true) *** Cleanup *** bool(true) -string(6) "testid" +string(14) "session-basic3" bool(true) diff --git a/ext/session/tests/session_basic5.phpt b/ext/session/tests/session_basic5.phpt index 50e6cde920..6a7511c317 100644 --- a/ext/session/tests/session_basic5.phpt +++ b/ext/session/tests/session_basic5.phpt @@ -24,7 +24,7 @@ ini_set('session.trans_sid_hosts','php.net,example.com'); echo "*** Testing basic session functionality : variation5 use_trans_sid ***\n"; echo "*** Test trans sid ***\n"; -$session_id = 'testid'; +$session_id = 'session-basic5'; session_id($session_id); session_start(); // Should add session ID to allowed hosts only for SECURITY @@ -238,116 +238,116 @@ ob_end_flush(); *** Testing basic session functionality : variation5 use_trans_sid *** *** Test trans sid *** -<a href="/?PHPSESSID=testid">test</a> -<a href="/path?PHPSESSID=testid">test</a> -<a href="/path/?PHPSESSID=testid">test</a> -<a href="/path/?foo=var&PHPSESSID=testid">test</a> -<a href="../?PHPSESSID=testid">test</a> -<a href="../path?PHPSESSID=testid">test</a> -<a href="../path/?PHPSESSID=testid">test</a> -<a href="../path/?foo=var&PHPSESSID=testid">test</a> - -<a href="/?PHPSESSID=testid#bar">test</a> -<a href="/path/?PHPSESSID=testid#bar">test</a> -<a href="/path/?foo=var&PHPSESSID=testid#bar">test</a> -<a href="../?PHPSESSID=testid#bar">test</a> -<a href="../path/?PHPSESSID=testid#bar">test</a> -<a href="../path/?foo=var&PHPSESSID=testid#bar">test</a> - -<a href="/?foo&PHPSESSID=testid">test</a> -<a href="/?foo&PHPSESSID=testid#bar">test</a> -<a href="/?foo=var&PHPSESSID=testid">test</a> -<a href="/?foo=var&PHPSESSID=testid#bar">test</a> -<a href="../?foo&PHPSESSID=testid">test</a> -<a href="../?foo&PHPSESSID=testid#bar">test</a> -<a href="../?foo=var&PHPSESSID=testid">test</a> -<a href="../?foo=var&PHPSESSID=testid#bar">test</a> - -<a href="file.php?PHPSESSID=testid">test</a> -<a href="file.php?foo&PHPSESSID=testid">test</a> -<a href="file.php?foo=var&PHPSESSID=testid">test</a> -<a href="file.php?foo=var&PHPSESSID=testid#bar">test</a> -<a href="../file.php?PHPSESSID=testid">test</a> -<a href="../file.php?foo&PHPSESSID=testid">test</a> -<a href="../file.php?foo=var&PHPSESSID=testid">test</a> -<a href="../file.php?foo=var&PHPSESSID=testid#bar">test</a> - -<a href="http://php.net/?PHPSESSID=testid">test</a> -<a href="http://php.net/?PHPSESSID=testid">test</a> -<a href="http://php.net/?PHPSESSID=testid#bar">test</a> -<a href="http://php.net/?foo&PHPSESSID=testid">test</a> -<a href="http://php.net/?foo&PHPSESSID=testid#bar">test</a> -<a href="http://php.net/?foo=var&PHPSESSID=testid">test</a> -<a href="http://php.net/?foo=var&PHPSESSID=testid#bar">test</a> -<a href="http://php.net/file.php?PHPSESSID=testid">test</a> -<a href="http://php.net/file.php?PHPSESSID=testid#bar">test</a> -<a href="http://php.net/file.php?foo&PHPSESSID=testid">test</a> -<a href="http://php.net/file.php?foo&PHPSESSID=testid#bar">test</a> -<a href="http://php.net/file.php?foo=var&PHPSESSID=testid">test</a> -<a href="http://php.net/file.php?foo=var&PHPSESSID=testid#bar">test</a> -<a href="http://php.net/some/path/file.php?PHPSESSID=testid">test</a> -<a href="http://php.net/some/path/file.php?foo&PHPSESSID=testid">test</a> -<a href="http://php.net/some/path/file.php?foo=var&PHPSESSID=testid">test</a> -<a href="http://php.net/some/path/file.php?foo=var&PHPSESSID=testid#bar">test</a> - -<a href="https://php.net/?PHPSESSID=testid">test</a> -<a href="https://php.net/?PHPSESSID=testid">test</a> -<a href="https://php.net/?foo=var&PHPSESSID=testid#bar">test</a> -<a href="https://php.net/file.php?PHPSESSID=testid">test</a> -<a href="https://php.net/file.php?foo=var&PHPSESSID=testid#bar">test</a> -<a href="https://php.net/some/path/file.php?PHPSESSID=testid">test</a> -<a href="https://php.net/some/path/file.php?foo=var&PHPSESSID=testid#bar">test</a> -<a href="https://php.net:8443/?PHPSESSID=testid">test</a> -<a href="https://php.net:8443/?PHPSESSID=testid">test</a> -<a href="https://php.net:8443/?foo=var&PHPSESSID=testid#bar">test</a> -<a href="https://php.net:8443/file.php?PHPSESSID=testid">test</a> -<a href="https://php.net:8443/file.php?foo=var&PHPSESSID=testid#bar">test</a> -<a href="https://php.net:8443/some/path/file.php?PHPSESSID=testid">test</a> -<a href="https://php.net:8443/some/path/file.php?foo=var&PHPSESSID=testid#bar">test</a> - -<a href="//php.net/?PHPSESSID=testid">test</a> -<a href="//php.net/?PHPSESSID=testid">test</a> -<a href="//php.net/?PHPSESSID=testid#bar">test</a> -<a href="//php.net/?foo&PHPSESSID=testid">test</a> -<a href="//php.net/?foo&PHPSESSID=testid#bar">test</a> -<a href="//php.net/?foo=var&PHPSESSID=testid">test</a> -<a href="//php.net/?foo=var&PHPSESSID=testid#bar">test</a> -<a href="//php.net/file.php?PHPSESSID=testid">test</a> -<a href="//php.net/file.php?PHPSESSID=testid#bar">test</a> -<a href="//php.net/file.php?foo&PHPSESSID=testid">test</a> -<a href="//php.net/file.php?foo&PHPSESSID=testid#bar">test</a> -<a href="//php.net/file.php?foo=var&PHPSESSID=testid">test</a> -<a href="//php.net/file.php?foo=var&PHPSESSID=testid#bar">test</a> -<a href="//php.net/some/path/file.php?PHPSESSID=testid">test</a> -<a href="//php.net/some/path/file.php?foo&PHPSESSID=testid">test</a> -<a href="//php.net/some/path/file.php?foo=var&PHPSESSID=testid">test</a> -<a href="//php.net/some/path/file.php?foo=var&PHPSESSID=testid#bar">test</a> - -<form action="script.php" method="post"><input type="hidden" name="PHPSESSID" value="testid" /> +<a href="/?PHPSESSID=session-basic5">test</a> +<a href="/path?PHPSESSID=session-basic5">test</a> +<a href="/path/?PHPSESSID=session-basic5">test</a> +<a href="/path/?foo=var&PHPSESSID=session-basic5">test</a> +<a href="../?PHPSESSID=session-basic5">test</a> +<a href="../path?PHPSESSID=session-basic5">test</a> +<a href="../path/?PHPSESSID=session-basic5">test</a> +<a href="../path/?foo=var&PHPSESSID=session-basic5">test</a> + +<a href="/?PHPSESSID=session-basic5#bar">test</a> +<a href="/path/?PHPSESSID=session-basic5#bar">test</a> +<a href="/path/?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="../?PHPSESSID=session-basic5#bar">test</a> +<a href="../path/?PHPSESSID=session-basic5#bar">test</a> +<a href="../path/?foo=var&PHPSESSID=session-basic5#bar">test</a> + +<a href="/?foo&PHPSESSID=session-basic5">test</a> +<a href="/?foo&PHPSESSID=session-basic5#bar">test</a> +<a href="/?foo=var&PHPSESSID=session-basic5">test</a> +<a href="/?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="../?foo&PHPSESSID=session-basic5">test</a> +<a href="../?foo&PHPSESSID=session-basic5#bar">test</a> +<a href="../?foo=var&PHPSESSID=session-basic5">test</a> +<a href="../?foo=var&PHPSESSID=session-basic5#bar">test</a> + +<a href="file.php?PHPSESSID=session-basic5">test</a> +<a href="file.php?foo&PHPSESSID=session-basic5">test</a> +<a href="file.php?foo=var&PHPSESSID=session-basic5">test</a> +<a href="file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="../file.php?PHPSESSID=session-basic5">test</a> +<a href="../file.php?foo&PHPSESSID=session-basic5">test</a> +<a href="../file.php?foo=var&PHPSESSID=session-basic5">test</a> +<a href="../file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> + +<a href="http://php.net/?PHPSESSID=session-basic5">test</a> +<a href="http://php.net/?PHPSESSID=session-basic5">test</a> +<a href="http://php.net/?PHPSESSID=session-basic5#bar">test</a> +<a href="http://php.net/?foo&PHPSESSID=session-basic5">test</a> +<a href="http://php.net/?foo&PHPSESSID=session-basic5#bar">test</a> +<a href="http://php.net/?foo=var&PHPSESSID=session-basic5">test</a> +<a href="http://php.net/?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="http://php.net/file.php?PHPSESSID=session-basic5">test</a> +<a href="http://php.net/file.php?PHPSESSID=session-basic5#bar">test</a> +<a href="http://php.net/file.php?foo&PHPSESSID=session-basic5">test</a> +<a href="http://php.net/file.php?foo&PHPSESSID=session-basic5#bar">test</a> +<a href="http://php.net/file.php?foo=var&PHPSESSID=session-basic5">test</a> +<a href="http://php.net/file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="http://php.net/some/path/file.php?PHPSESSID=session-basic5">test</a> +<a href="http://php.net/some/path/file.php?foo&PHPSESSID=session-basic5">test</a> +<a href="http://php.net/some/path/file.php?foo=var&PHPSESSID=session-basic5">test</a> +<a href="http://php.net/some/path/file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> + +<a href="https://php.net/?PHPSESSID=session-basic5">test</a> +<a href="https://php.net/?PHPSESSID=session-basic5">test</a> +<a href="https://php.net/?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="https://php.net/file.php?PHPSESSID=session-basic5">test</a> +<a href="https://php.net/file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="https://php.net/some/path/file.php?PHPSESSID=session-basic5">test</a> +<a href="https://php.net/some/path/file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="https://php.net:8443/?PHPSESSID=session-basic5">test</a> +<a href="https://php.net:8443/?PHPSESSID=session-basic5">test</a> +<a href="https://php.net:8443/?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="https://php.net:8443/file.php?PHPSESSID=session-basic5">test</a> +<a href="https://php.net:8443/file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="https://php.net:8443/some/path/file.php?PHPSESSID=session-basic5">test</a> +<a href="https://php.net:8443/some/path/file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> + +<a href="//php.net/?PHPSESSID=session-basic5">test</a> +<a href="//php.net/?PHPSESSID=session-basic5">test</a> +<a href="//php.net/?PHPSESSID=session-basic5#bar">test</a> +<a href="//php.net/?foo&PHPSESSID=session-basic5">test</a> +<a href="//php.net/?foo&PHPSESSID=session-basic5#bar">test</a> +<a href="//php.net/?foo=var&PHPSESSID=session-basic5">test</a> +<a href="//php.net/?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="//php.net/file.php?PHPSESSID=session-basic5">test</a> +<a href="//php.net/file.php?PHPSESSID=session-basic5#bar">test</a> +<a href="//php.net/file.php?foo&PHPSESSID=session-basic5">test</a> +<a href="//php.net/file.php?foo&PHPSESSID=session-basic5#bar">test</a> +<a href="//php.net/file.php?foo=var&PHPSESSID=session-basic5">test</a> +<a href="//php.net/file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> +<a href="//php.net/some/path/file.php?PHPSESSID=session-basic5">test</a> +<a href="//php.net/some/path/file.php?foo&PHPSESSID=session-basic5">test</a> +<a href="//php.net/some/path/file.php?foo=var&PHPSESSID=session-basic5">test</a> +<a href="//php.net/some/path/file.php?foo=var&PHPSESSID=session-basic5#bar">test</a> + +<form action="script.php" method="post"><input type="hidden" name="PHPSESSID" value="session-basic5" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> -<form action="../script.php" method="post"><input type="hidden" name="PHPSESSID" value="testid" />r +<form action="../script.php" method="post"><input type="hidden" name="PHPSESSID" value="session-basic5" />r <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> -<form action="/path/script.php" method="post"><input type="hidden" name="PHPSESSID" value="testid" /> +<form action="/path/script.php" method="post"><input type="hidden" name="PHPSESSID" value="session-basic5" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> -<form action="../path/script.php" method="post"><input type="hidden" name="PHPSESSID" value="testid" /> +<form action="../path/script.php" method="post"><input type="hidden" name="PHPSESSID" value="session-basic5" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> -<form method="post" action="http://php.net/script.php"><input type="hidden" name="PHPSESSID" value="testid" /> +<form method="post" action="http://php.net/script.php"><input type="hidden" name="PHPSESSID" value="session-basic5" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> -<form method="post" action="https://php.net/script.php"><input type="hidden" name="PHPSESSID" value="testid" /> +<form method="post" action="https://php.net/script.php"><input type="hidden" name="PHPSESSID" value="session-basic5" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> -<form method="post" action="//php.net/script.php"><input type="hidden" name="PHPSESSID" value="testid" /> +<form method="post" action="//php.net/script.php"><input type="hidden" name="PHPSESSID" value="session-basic5" /> <input type="text" name="test1"></input> <input type="text" name="test2" /> </form> @@ -436,5 +436,5 @@ ob_end_flush(); bool(true) *** Cleanup *** bool(true) -string(6) "testid" +string(14) "session-basic5" bool(true) diff --git a/ext/session/tests/session_commit_variation4.phpt b/ext/session/tests/session_commit_variation4.phpt index 2633f543af..b0c3ad4a94 100644 --- a/ext/session/tests/session_commit_variation4.phpt +++ b/ext/session/tests/session_commit_variation4.phpt @@ -12,7 +12,7 @@ ob_start(); echo "*** Testing session_commit() : variation ***\n"; var_dump(ini_get('session.use_strict_mode')); -var_dump(session_id("test")); +var_dump(session_id("session-commit-variation4")); var_dump(session_start()); var_dump(session_id()); var_dump(session_commit()); @@ -38,18 +38,18 @@ ob_end_flush(); string(1) "0" string(0) "" bool(true) -string(4) "test" +string(25) "session-commit-variation4" bool(true) -string(4) "test" +string(25) "session-commit-variation4" bool(true) string(1) "0" -string(4) "test" +string(25) "session-commit-variation4" bool(true) -string(4) "test" +string(25) "session-commit-variation4" bool(true) -string(4) "test" +string(25) "session-commit-variation4" bool(true) -string(4) "test" +string(25) "session-commit-variation4" bool(true) bool(true) string(1) "0" diff --git a/ext/session/tests/session_module_name_variation4.phpt b/ext/session/tests/session_module_name_variation4.phpt index 9b845151a8..7e58ddc6c4 100644 --- a/ext/session/tests/session_module_name_variation4.phpt +++ b/ext/session/tests/session_module_name_variation4.phpt @@ -15,6 +15,8 @@ echo "*** Testing session_module_name() : variation ***\n"; require_once "save_handler.inc"; $path = __DIR__; +$path = __DIR__ . '/session_module_name_variation4'; +@mkdir($path); session_save_path($path); session_module_name("files"); @@ -35,6 +37,7 @@ var_dump($_SESSION); var_dump(session_destroy()); ob_end_flush(); +rmdir($path); ?> --EXPECT-- *** Testing session_module_name() : variation *** diff --git a/ext/session/tests/session_save_path_variation4.phpt b/ext/session/tests/session_save_path_variation4.phpt index 8b3df52262..ec22e92051 100644 --- a/ext/session/tests/session_save_path_variation4.phpt +++ b/ext/session/tests/session_save_path_variation4.phpt @@ -14,7 +14,7 @@ ob_start(); echo "*** Testing session_save_path() : variation ***\n"; $initdir = __DIR__; -$sessions = ($initdir."/sessions"); +$sessions = ($initdir."/session_save_path_variation4"); chdir($initdir); @@ -38,7 +38,7 @@ ob_end_flush(); --CLEAN-- <?php $initdir = __DIR__; -$sessions = ($initdir."/sessions"); +$sessions = ($initdir."/session_save_path_variation4"); var_dump(rmdir($sessions)); ?> --EXPECTF-- diff --git a/ext/session/tests/session_save_path_variation5.phpt b/ext/session/tests/session_save_path_variation5.phpt index 446a472ab4..3977d0deab 100644 --- a/ext/session/tests/session_save_path_variation5.phpt +++ b/ext/session/tests/session_save_path_variation5.phpt @@ -15,7 +15,7 @@ session.name=PHPSESSID ob_start(); echo "*** Testing session_save_path() : variation ***\n"; $directory = __DIR__; -$sessions = ($directory."/sessions"); +$sessions = ($directory."/session_save_path_variation5"); chdir($directory); ini_set('open_basedir', '.'); @@ -36,7 +36,7 @@ ob_end_flush(); --CLEAN-- <?php $directory = __DIR__; -$sessions = ($directory."/sessions"); +$sessions = ($directory."/session_save_path_variation5"); var_dump(rmdir($sessions)); ?> --EXPECTF-- diff --git a/ext/session/tests/session_set_save_handler_basic.phpt b/ext/session/tests/session_set_save_handler_basic.phpt index 459f03c825..23865aa512 100644 --- a/ext/session/tests/session_set_save_handler_basic.phpt +++ b/ext/session/tests/session_set_save_handler_basic.phpt @@ -20,7 +20,8 @@ var_dump(session_module_name(FALSE)); var_dump(session_module_name("blah")); var_dump(session_module_name("foo")); -$path = __DIR__; +$path = __DIR__ . '/session_set_save_handler_basic'; +@mkdir($path); session_save_path($path); session_set_save_handler("open", "close", "read", "write", "destroy", "gc"); @@ -55,6 +56,7 @@ session_start(); session_destroy(); ob_end_flush(); +rmdir($path); ?> --EXPECTF-- *** Testing session_set_save_handler() : basic functionality *** diff --git a/ext/session/tests/session_set_save_handler_variation2.phpt b/ext/session/tests/session_set_save_handler_variation2.phpt index 224846466e..a4b7d4df6e 100644 --- a/ext/session/tests/session_set_save_handler_variation2.phpt +++ b/ext/session/tests/session_set_save_handler_variation2.phpt @@ -10,13 +10,15 @@ ob_start(); echo "*** Testing session_set_save_handler() : variation ***\n"; require_once "save_handler.inc"; -$path = __DIR__; +$path = __DIR__ . '/session_set_save_handler_variation2'; +@mkdir($path); session_save_path($path); var_dump(session_start()); var_dump(session_set_save_handler("open", "close", "read", "write", "destroy", "gc")); var_dump(session_destroy()); ob_end_flush(); +rmdir($path); ?> --EXPECTF-- *** Testing session_set_save_handler() : variation *** diff --git a/ext/session/tests/session_set_save_handler_variation3.phpt b/ext/session/tests/session_set_save_handler_variation3.phpt index be825cbe72..cc0ae1fb23 100644 --- a/ext/session/tests/session_set_save_handler_variation3.phpt +++ b/ext/session/tests/session_set_save_handler_variation3.phpt @@ -12,13 +12,15 @@ ob_start(); echo "*** Testing session_set_save_handler() : variation ***\n"; require_once "save_handler.inc"; -$path = __DIR__; +$path = __DIR__ . '/session_set_save_handler_variation3'; +@mkdir($path); var_dump(session_status()); session_save_path($path); var_dump(session_set_save_handler("open", "close", "read", "write", "destroy", "gc")); var_dump(session_destroy()); - ob_end_flush(); + +rmdir($path); ?> --EXPECTF-- *** Testing session_set_save_handler() : variation *** diff --git a/ext/session/tests/session_set_save_handler_variation4.phpt b/ext/session/tests/session_set_save_handler_variation4.phpt index e1b61f29a7..9930f7aaf2 100644 --- a/ext/session/tests/session_set_save_handler_variation4.phpt +++ b/ext/session/tests/session_set_save_handler_variation4.phpt @@ -23,7 +23,8 @@ function noisy_gc($maxlifetime) { } require_once "save_handler.inc"; -$path = __DIR__; +$path = __DIR__ . '/session_set_save_handler_variation4'; +@mkdir($path); session_save_path($path); session_set_save_handler("open", "close", "read", "write", "destroy", "noisy_gc"); @@ -42,13 +43,14 @@ var_dump($_SESSION); var_dump(session_destroy()); ob_end_flush(); +rmdir($path); ?> --EXPECTF-- *** Testing session_set_save_handler() : variation *** Open [%s,PHPSESSID] Read [%s,%s] GC [0] -2 deleted +1 deleted array(3) { ["Blah"]=> string(12) "Hello World!" diff --git a/ext/session/tests/session_set_save_handler_variation5.phpt b/ext/session/tests/session_set_save_handler_variation5.phpt index 9860dcc04a..7cf7814259 100644 --- a/ext/session/tests/session_set_save_handler_variation5.phpt +++ b/ext/session/tests/session_set_save_handler_variation5.phpt @@ -24,7 +24,8 @@ function noisy_gc($maxlifetime) { echo "*** Testing session_set_save_handler() : variation ***\n"; require_once "save_handler.inc"; -$path = __DIR__; +$path = __DIR__ . '/session_set_save_handler_variation5'; +@mkdir($path); var_dump(session_save_path($path)); echo "*** Without lazy_write ***\n"; @@ -48,6 +49,7 @@ var_dump(session_start()); var_dump(session_destroy()); ob_end_flush(); +rmdir($path); ?> --EXPECTF-- *** Testing session_set_save_handler() : variation *** diff --git a/ext/session/tests/session_set_save_handler_variation6.phpt b/ext/session/tests/session_set_save_handler_variation6.phpt index d65c0dc1c8..11af07f603 100644 --- a/ext/session/tests/session_set_save_handler_variation6.phpt +++ b/ext/session/tests/session_set_save_handler_variation6.phpt @@ -17,7 +17,8 @@ ob_start(); echo "*** Testing session_set_save_handler() : test write short circuit ***\n"; require_once "save_handler.inc"; -$path = __DIR__; +$path = __DIR__ . '/session_set_save_handler_variation6'; +@mkdir($path); session_save_path($path); session_set_save_handler("open", "close", "read", "write", "destroy", "gc", "create_sid", "validate_sid", "update"); @@ -53,6 +54,7 @@ session_start(); session_destroy(); ob_end_flush(); +rmdir($path); ?> --EXPECTF-- *** Testing session_set_save_handler() : test write short circuit *** diff --git a/ext/soap/tests/bug73037.phpt b/ext/soap/tests/bug73037.phpt index 4e72932862..8060575bd8 100644 --- a/ext/soap/tests/bug73037.phpt +++ b/ext/soap/tests/bug73037.phpt @@ -1,7 +1,5 @@ --TEST-- Bug #73037 SoapServer reports Bad Request when gzipped, var 0 ---CONFLICTS-- -server --SKIPIF-- <?php require_once('skipif.inc'); diff --git a/ext/soap/tests/custom_content_type.phpt b/ext/soap/tests/custom_content_type.phpt index 17a074f102..9172d3d6ca 100644 --- a/ext/soap/tests/custom_content_type.phpt +++ b/ext/soap/tests/custom_content_type.phpt @@ -8,8 +8,6 @@ SOAP customized Content-Type, eg. SwA use case echo "skip sapi/cli/tests/php_cli_server.inc required but not found"; } ?> ---CONFLICTS-- -server --FILE-- <?php diff --git a/ext/sockets/sockets.c b/ext/sockets/sockets.c index 0b21c4f684..c50d8d5d71 100644 --- a/ext/sockets/sockets.c +++ b/ext/sockets/sockets.c @@ -1045,11 +1045,7 @@ PHP_FUNCTION(socket_read) if (retval == -1) { /* if the socket is in non-blocking mode and there's no data to read, don't output any error, as this is a normal situation, and not an error */ - if (errno == EAGAIN -#ifdef EWOULDBLOCK - || errno == EWOULDBLOCK -#endif - ) { + if (PHP_IS_TRANSIENT_ERROR(errno)) { php_sock->error = errno; SOCKETS_G(last_error) = errno; } else { diff --git a/ext/sodium/config.m4 b/ext/sodium/config.m4 index 0676b66910..05d9ebd5a1 100644 --- a/ext/sodium/config.m4 +++ b/ext/sodium/config.m4 @@ -11,8 +11,9 @@ if test "$PHP_SODIUM" != "no"; then AC_DEFINE(HAVE_LIBSODIUMLIB, 1, [ ]) - dnl Add -Wno-type-limits as this may arise on 32bits platforms + dnl Add -Wno-type-limits and -Wno-logical-op as this may arise on 32bits platforms SODIUM_COMPILER_FLAGS="$LIBSODIUM_CFLAGS -Wno-type-limits" + AX_CHECK_COMPILE_FLAG([-Wno-logical-op], SODIUM_COMPILER_FLAGS="$SODIUM_COMPILER_FLAGS -Wno-logical-op", , [-Werror]) PHP_NEW_EXTENSION(sodium, libsodium.c sodium_pwhash.c, $ext_shared, , $SODIUM_COMPILER_FLAGS) PHP_SUBST(SODIUM_SHARED_LIBADD) fi diff --git a/ext/standard/array.c b/ext/standard/array.c index f99af84611..3a50e77100 100644 --- a/ext/standard/array.c +++ b/ext/standard/array.c @@ -1321,7 +1321,13 @@ PHP_FUNCTION(max) } /* }}} */ -static int php_array_walk(zval *array, zval *userdata, int recursive) /* {{{ */ +typedef struct { + zend_fcall_info fci; + zend_fcall_info_cache fci_cache; +} php_array_walk_context; + +static int php_array_walk( + php_array_walk_context *context, zval *array, zval *userdata, int recursive) { zval args[3], /* Arguments to userland function */ retval, /* Return value - unused */ @@ -1331,15 +1337,19 @@ static int php_array_walk(zval *array, zval *userdata, int recursive) /* {{{ */ uint32_t ht_iter; int result = SUCCESS; + /* Create a local copy of fci, as we want to use different arguments at different + * levels of recursion. */ + zend_fcall_info fci = context->fci; + /* Set up known arguments */ ZVAL_UNDEF(&args[1]); if (userdata) { ZVAL_COPY(&args[2], userdata); } - BG(array_walk_fci).retval = &retval; - BG(array_walk_fci).param_count = userdata ? 3 : 2; - BG(array_walk_fci).params = args; + fci.retval = &retval; + fci.param_count = userdata ? 3 : 2; + fci.params = args; zend_hash_internal_pointer_reset_ex(target_hash, &pos); ht_iter = zend_hash_iterator_add(target_hash, pos); @@ -1386,8 +1396,6 @@ static int php_array_walk(zval *array, zval *userdata, int recursive) /* {{{ */ if (recursive && Z_TYPE_P(Z_REFVAL_P(zv)) == IS_ARRAY) { HashTable *thash; - zend_fcall_info orig_array_walk_fci; - zend_fcall_info_cache orig_array_walk_fci_cache; zval ref; ZVAL_COPY_VALUE(&ref, zv); @@ -1400,28 +1408,20 @@ static int php_array_walk(zval *array, zval *userdata, int recursive) /* {{{ */ break; } - /* backup the fcall info and cache */ - orig_array_walk_fci = BG(array_walk_fci); - orig_array_walk_fci_cache = BG(array_walk_fci_cache); - Z_ADDREF(ref); GC_PROTECT_RECURSION(thash); - result = php_array_walk(zv, userdata, recursive); + result = php_array_walk(context, zv, userdata, recursive); if (Z_TYPE_P(Z_REFVAL(ref)) == IS_ARRAY && thash == Z_ARRVAL_P(Z_REFVAL(ref))) { /* If the hashtable changed in the meantime, we'll "leak" this apply count * increment -- our reference to thash is no longer valid. */ GC_UNPROTECT_RECURSION(thash); } zval_ptr_dtor(&ref); - - /* restore the fcall info and cache */ - BG(array_walk_fci) = orig_array_walk_fci; - BG(array_walk_fci_cache) = orig_array_walk_fci_cache; } else { ZVAL_COPY(&args[0], zv); /* Call the userland function */ - result = zend_call_function(&BG(array_walk_fci), &BG(array_walk_fci_cache)); + result = zend_call_function(&fci, &context->fci_cache); if (result == SUCCESS) { zval_ptr_dtor(&retval); } @@ -1458,33 +1458,22 @@ static int php_array_walk(zval *array, zval *userdata, int recursive) /* {{{ */ zend_hash_iterator_del(ht_iter); return result; } -/* }}} */ /* {{{ Apply a user function to every member of an array */ PHP_FUNCTION(array_walk) { zval *array; zval *userdata = NULL; - zend_fcall_info orig_array_walk_fci; - zend_fcall_info_cache orig_array_walk_fci_cache; - - orig_array_walk_fci = BG(array_walk_fci); - orig_array_walk_fci_cache = BG(array_walk_fci_cache); + php_array_walk_context context; ZEND_PARSE_PARAMETERS_START(2, 3) Z_PARAM_ARRAY_OR_OBJECT_EX(array, 0, 1) - Z_PARAM_FUNC(BG(array_walk_fci), BG(array_walk_fci_cache)) + Z_PARAM_FUNC(context.fci, context.fci_cache) Z_PARAM_OPTIONAL Z_PARAM_ZVAL(userdata) - ZEND_PARSE_PARAMETERS_END_EX( - BG(array_walk_fci) = orig_array_walk_fci; - BG(array_walk_fci_cache) = orig_array_walk_fci_cache; - return - ); - - php_array_walk(array, userdata, 0); - BG(array_walk_fci) = orig_array_walk_fci; - BG(array_walk_fci_cache) = orig_array_walk_fci_cache; + ZEND_PARSE_PARAMETERS_END(); + + php_array_walk(&context, array, userdata, 0); RETURN_TRUE; } /* }}} */ @@ -1494,26 +1483,16 @@ PHP_FUNCTION(array_walk_recursive) { zval *array; zval *userdata = NULL; - zend_fcall_info orig_array_walk_fci; - zend_fcall_info_cache orig_array_walk_fci_cache; - - orig_array_walk_fci = BG(array_walk_fci); - orig_array_walk_fci_cache = BG(array_walk_fci_cache); + php_array_walk_context context; ZEND_PARSE_PARAMETERS_START(2, 3) Z_PARAM_ARRAY_OR_OBJECT_EX(array, 0, 1) - Z_PARAM_FUNC(BG(array_walk_fci), BG(array_walk_fci_cache)) + Z_PARAM_FUNC(context.fci, context.fci_cache) Z_PARAM_OPTIONAL Z_PARAM_ZVAL(userdata) - ZEND_PARSE_PARAMETERS_END_EX( - BG(array_walk_fci) = orig_array_walk_fci; - BG(array_walk_fci_cache) = orig_array_walk_fci_cache; - return - ); - - php_array_walk(array, userdata, 1); - BG(array_walk_fci) = orig_array_walk_fci; - BG(array_walk_fci_cache) = orig_array_walk_fci_cache; + ZEND_PARSE_PARAMETERS_END(); + + php_array_walk(&context, array, userdata, 1); RETURN_TRUE; } /* }}} */ diff --git a/ext/standard/basic_functions.c b/ext/standard/basic_functions.c index 876ef347eb..33247c4cb3 100755 --- a/ext/standard/basic_functions.c +++ b/ext/standard/basic_functions.c @@ -463,8 +463,6 @@ PHP_RINIT_FUNCTION(basic) /* {{{ */ BG(strtok_last) = NULL; BG(ctype_string) = NULL; BG(locale_changed) = 0; - BG(array_walk_fci) = empty_fcall_info; - BG(array_walk_fci_cache) = empty_fcall_info_cache; BG(user_compare_fci) = empty_fcall_info; BG(user_compare_fci_cache) = empty_fcall_info_cache; BG(page_uid) = -1; @@ -1472,7 +1470,7 @@ PHPAPI int _php_error_log_ex(int opt_err, const char *message, size_t message_le return FAILURE; case 3: /*save to a file */ - stream = php_stream_open_wrapper(opt, "a", IGNORE_URL_WIN | REPORT_ERRORS, NULL); + stream = php_stream_open_wrapper(opt, "a", REPORT_ERRORS, NULL); if (!stream) { return FAILURE; } @@ -1953,12 +1951,8 @@ PHP_FUNCTION(highlight_string) hicompiled_string_description = zend_make_compiled_string_description("highlighted code"); - // TODO: Accept zend_string in highlight_string API. - zval str_zv; - ZVAL_STR_COPY(&str_zv, str); - highlight_string(&str_zv, &syntax_highlighter_ini, hicompiled_string_description); + highlight_string(str, &syntax_highlighter_ini, hicompiled_string_description); efree(hicompiled_string_description); - zval_ptr_dtor(&str_zv); EG(error_reporting) = old_error_reporting; diff --git a/ext/standard/basic_functions.h b/ext/standard/basic_functions.h index 97faf50fd9..34b2570175 100644 --- a/ext/standard/basic_functions.h +++ b/ext/standard/basic_functions.h @@ -66,8 +66,6 @@ typedef struct _php_basic_globals { char *strtok_last; char strtok_table[256]; size_t strtok_len; - zend_fcall_info array_walk_fci; - zend_fcall_info_cache array_walk_fci_cache; zend_fcall_info user_compare_fci; zend_fcall_info_cache user_compare_fci_cache; zend_llist *user_tick_functions; diff --git a/ext/standard/mail.c b/ext/standard/mail.c index 3080a3957b..4dfbfb03d0 100644 --- a/ext/standard/mail.c +++ b/ext/standard/mail.c @@ -371,7 +371,7 @@ void php_mail_log_to_syslog(char *message) { void php_mail_log_to_file(char *filename, char *message, size_t message_size) { /* Write 'message' to the given file. */ - uint32_t flags = IGNORE_URL_WIN | REPORT_ERRORS | STREAM_DISABLE_OPEN_BASEDIR; + uint32_t flags = REPORT_ERRORS | STREAM_DISABLE_OPEN_BASEDIR; php_stream *stream = php_stream_open_wrapper(filename, "a", flags, NULL); if (stream) { php_stream_write(stream, message, message_size); diff --git a/ext/standard/php_http.h b/ext/standard/php_http.h index 1df941ee00..3f764c18a6 100644 --- a/ext/standard/php_http.h +++ b/ext/standard/php_http.h @@ -25,6 +25,5 @@ PHPAPI void php_url_encode_hash_ex(HashTable *ht, smart_str *formstr, const char *key_prefix, size_t key_prefix_len, const char *key_suffix, size_t key_suffix_len, zval *type, const char *arg_sep, int enc_type); -#define php_url_encode_hash(ht, formstr) php_url_encode_hash_ex((ht), (formstr), NULL, 0, NULL, 0, NULL, 0, NULL) #endif diff --git a/ext/standard/tests/streams/bug64433.phpt b/ext/standard/tests/streams/bug64433.phpt index 56af241837..53fecc57f9 100644 --- a/ext/standard/tests/streams/bug64433.phpt +++ b/ext/standard/tests/streams/bug64433.phpt @@ -9,8 +9,6 @@ if(!$res) { die("skip could not open cli server script"); } ?> ---CONFLICTS-- -server --FILE-- <?php include __DIR__."/../../../../sapi/cli/tests/php_cli_server.inc"; diff --git a/ext/standard/tests/url/get_headers_error_003.phpt b/ext/standard/tests/url/get_headers_error_003.phpt index a8edc2cb21..7ee0b9e210 100644 --- a/ext/standard/tests/url/get_headers_error_003.phpt +++ b/ext/standard/tests/url/get_headers_error_003.phpt @@ -1,7 +1,5 @@ --TEST-- Test get_headers() function : test with context ---CONFLICTS-- -server --FILE-- <?php diff --git a/ext/zip/php_zip.c b/ext/zip/php_zip.c index 96baabacd4..58707d6f95 100644 --- a/ext/zip/php_zip.c +++ b/ext/zip/php_zip.c @@ -288,11 +288,13 @@ static int php_zip_add_file(ze_zip_object *obj, const char *filename, size_t fil } if (!expand_filepath(filename, resolved_path)) { + php_error_docref(NULL, E_WARNING, "No such file or directory"); return -1; } php_stat(resolved_path, strlen(resolved_path), FS_EXISTS, &exists_flag); if (Z_TYPE(exists_flag) == IS_FALSE) { + php_error_docref(NULL, E_WARNING, "No such file or directory"); return -1; } @@ -1163,6 +1165,7 @@ PHP_FUNCTION(zip_open) } if(!expand_filepath(ZSTR_VAL(filename), resolved_path)) { + php_error_docref(NULL, E_WARNING, "No such file or directory"); RETURN_FALSE; } @@ -1444,6 +1447,7 @@ PHP_METHOD(ZipArchive, open) } if (!(resolved_path = expand_filepath(ZSTR_VAL(filename), NULL))) { + php_error_docref(NULL, E_WARNING, "No such file or directory"); RETURN_FALSE; } diff --git a/ext/zip/tests/bug64342_0.phpt b/ext/zip/tests/bug64342_0.phpt index 43f60f5d7c..751fb9c4ba 100644 --- a/ext/zip/tests/bug64342_0.phpt +++ b/ext/zip/tests/bug64342_0.phpt @@ -38,6 +38,8 @@ DONE ?> --EXPECTF-- %s.txt + +Warning: ZipArchive::addFile(): No such file or directory in %s on line %d add failed close ok DONE diff --git a/ext/zip/tests/bug64342_1-mb.phpt b/ext/zip/tests/bug64342_1-mb.phpt index 4884c9fe5f..9efd046315 100644 --- a/ext/zip/tests/bug64342_1-mb.phpt +++ b/ext/zip/tests/bug64342_1-mb.phpt @@ -37,6 +37,7 @@ if ($zip->status == ZIPARCHIVE::ER_OK) { } @unlink($file); ?> ---EXPECT-- +--EXPECTF-- +Warning: ZipArchive::addFile(): No such file or directory in %s on line %d failed OK diff --git a/ext/zip/tests/bug64342_1.phpt b/ext/zip/tests/bug64342_1.phpt index 6b923d3ef1..1fff3da26d 100644 --- a/ext/zip/tests/bug64342_1.phpt +++ b/ext/zip/tests/bug64342_1.phpt @@ -37,6 +37,7 @@ if ($zip->status == ZIPARCHIVE::ER_OK) { } @unlink($file); ?> ---EXPECT-- +--EXPECTF-- +Warning: ZipArchive::addFile(): No such file or directory in %s on line %d failed OK |