summaryrefslogtreecommitdiff
path: root/file_io
Commit message (Expand)AuthorAgeFilesLines
* This commit was manufactured by cvs2svn to create tag 'APR_0_9_4'.APR_0_9_4(no author)2003-09-262-12/+19
* Include apr_temp_dir_get() in OS/2 build.bjh2003-08-292-1/+3
* Add the path to the NetWare TMP directory to the list of test directories.bnicholes2003-08-251-0/+8
* Add apr_temp_dir_get(), a function to get the most suitable directory to placethommay2003-08-232-1/+168
* Implemented apr_file_mtime_set() for NetWarebnicholes2003-07-141-2/+41
* Finish the new timestamp-setting API, thanks to helpful tips fromsussman2003-07-083-7/+43
* New apr_file_mtime_set() API, implemented in unix and win32.sussman2003-07-073-1/+64
* Change a declaration that wouldn't compile under Borland C++brianp2003-06-251-1/+4
* When writing to pipes with a timeout set, handle the situationtrawick2003-06-051-2/+13
* Added flag APR_FILE_ATTR_HIDDEN for manipulating the "hidden" filebrane2003-05-244-1/+39
* Unless we are actively looking for a symlink, don't inform the apr_finfo_twrowe2003-05-191-1/+2
* Make sure that the current directory that is represented by the context is st...bnicholes2003-05-081-2/+18
* apr_file_gets(): Return APR_SUCCESS if any characters aretrawick2003-05-013-1/+20
* Call apr_file_flush() on a buffered file before getting the file info.bnicholes2003-04-031-0/+6
* OS/2: apr_stat() fixesbjh2003-03-311-5/+9
* Revert my recent change that concerns both Jeff Trawick and myself,wrowe2003-03-221-12/+6
* Given two solutions to the current mpm inheritence bugs,wrowe2003-03-221-8/+10
* dup2() is supported on NetWarebnicholes2003-03-201-8/+0
* Need to resolve the #define APR_INHERIT that is being used as a flagbnicholes2003-03-201-0/+1
* Fix "unused variable" compiler warning in apr_file_dup2 and eliminatejorton2003-03-191-6/+1
* Same fixes for netware as unix in pipe.cwrowe2003-03-191-1/+3
* apr_os_pipe_put() (where is the _file in that symbol name?) registerswrowe2003-03-191-1/+1
* Clean up a runover line, and reflect that apr_os_file_put() does notwrowe2003-03-191-2/+3
* A common sense logic change. If we are dup()ing or dup2()ingwrowe2003-03-191-4/+14
* Fix bug #5 of the evening; always register the proper child_cleanupwrowe2003-03-191-1/+3
* Fix bug #4 of the evening; it is altogether possible for the user to havewrowe2003-03-191-9/+13
* apr_p[c]alloc always succeeds, exit()s or calls the user defined callback.wrowe2003-03-191-3/+0
* Break the initial *new_file NULL checkwrowe2003-03-191-12/+12
* Bug Fix #1 .. Move the special-case of Netware into _dup_file() wherewrowe2003-03-191-4/+8
* Document an obvious problem lacking a simple solution - if we simplywrowe2003-03-181-0/+7
* Fix the 'test_buffered_write_size' test by ensuring that bufferedbrane2003-03-071-0/+6
* Eliminate a local variable in apr_unix_file_cleanup.jorton2003-03-061-3/+1
* Fix the 'test_buffered_write_size' test by ensuring that bufferedjorton2003-03-061-0/+6
* Make apr_file_inherit_set/unset work for pipes created by Unixjorton2003-03-061-0/+4
* Correct apr_file_gets() on OS2 and Win32 so that '\r's are no longerwrowe2003-03-042-6/+2
* Implemented the functionality for the APR_FILEPATH_NATIVE flag inbnicholes2003-03-041-0/+6
* OS/2: Don't return the apr_file_t if the open failed.bjh2003-03-021-1/+2
* Include filepath_util.c in OS/2 build.bjh2003-02-262-0/+2
* Move the apr_filepath_list_split_impl and apr_filepath_list_merge_implbrane2003-02-163-10/+5
* get the prototypes for these functions in the right placetrawick2003-02-132-12/+2
* CodeWarrior compiler can't handle initializing an array element with a variablebnicholes2003-02-121-1/+2
* Remove unused #include added in revision 1.39.brane2003-02-121-1/+0
* Added two new functions, apr_filepath_list_split and apr_filepath_list_merge,brane2003-02-124-0/+195
* Performance enhancement by trying to create and cache path contexts sobnicholes2003-02-071-2/+111
* Added a check to make sure that the path is not too long. Otherwise, just ba...bnicholes2003-01-281-6/+9
* Simplify and be more thorough. We only want to know how many bytes arewrowe2003-01-241-13/+10
* Expand the set/get semantic slightly... always allow the user to recoverwrowe2003-01-241-7/+15
* Eliminate open_nt_process_pipe and give apr_create_nt_pipe a very similarwrowe2003-01-241-5/+8
* Remove the stat info caching for NetWarebnicholes2003-01-142-205/+4
* Since deconstruct() does not actually validate the path, the check is meaning...bnicholes2003-01-091-7/+2