summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* CuFail appends a newline; reserve blank lines to seperate test suites.wrowe2002-12-311-2/+2
| | | | | | | | This bit of formatting looks lousy either way, but this seems more consistent to me. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64250 13f79535-47bb-0310-9956-ffa450edef68
* I like Ryan's formatting of the notimpl stuff, do the same for failure.wrowe2002-12-311-1/+1
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64249 13f79535-47bb-0310-9956-ffa450edef68
* Clarify what is going on here... Win32 can only recover the 'primary'wrowe2002-12-311-1/+15
| | | | | | | | | | | | group of an existing token. An arbitrary user sid can't be an authenticated token, so apr_uid_get cannot return the group. There may be a way to do so, but I hadn't found it. For now, register this case as not implemented; it would probably be better if both the apr_uid_get and apr_uid_current quit dealing in gids altogether. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64248 13f79535-47bb-0310-9956-ffa450edef68
* APR_STATUS_IS_FOO() tests must provide at _least_ the canonical test forwrowe2002-12-311-3/+7
| | | | | | | APR_FOO amoungst other platform-specific flavors. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64247 13f79535-47bb-0310-9956-ffa450edef68
* We can't optimize away this test! Just because a path didn't containwrowe2002-12-311-1/+1
| | | | | | | | | | backslashes doesn't mean that both paths were not absolute. There probably is a 'safe' optimization for that block, but I'm not seeing it this morning. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64246 13f79535-47bb-0310-9956-ffa450edef68
* Look at the status before the pointer resultwrowe2002-12-311-1/+1
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64245 13f79535-47bb-0310-9956-ffa450edef68
* ABS_ROOT is simpler to follow if it is what its name implieswrowe2002-12-311-16/+16
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64244 13f79535-47bb-0310-9956-ffa450edef68
* A second stab at getting apr_thread_cond_timedwait to work. Someonewrowe2002-12-311-2/+2
| | | | | | | | | else should look at this code, please, and let me know if it follows the correct behavior. It passes the tests, but that presumes the tests themselves are sufficient. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64243 13f79535-47bb-0310-9956-ffa450edef68
* Agreed with Ryan... it's entirely redundant to run the tests twice wherewrowe2002-12-311-11/+6
| | | | | | | | | | libraries and modules are identical (although you may by defining LIB_NAME the same as MOD_NAME)... and it makes no sense to emit a not-implemented message in this case. Drop the exception, and exclude the _library() fns if LIB_NAME is undefined. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64242 13f79535-47bb-0310-9956-ffa450edef68
* Fix apr_poll behavior on Darwin/Win32 (now passing testpoll.)wrowe2002-12-311-0/+4
| | | | | | | | | | | We were getting spurious returned events because the select-based poll implementation wasn't zeroing out previous results before setting the current ones and returning. Submitted by: Garrett Rooney <rooneg@electricjellyfish.net> git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64241 13f79535-47bb-0310-9956-ffa450edef68
* Use apr_uint32_t rather than apr_size_t for the maximum hash table size.jerenkrantz2002-12-302-11/+11
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64240 13f79535-47bb-0310-9956-ffa450edef68
* Fix uninitialized variable dereference from a recent commitbrianp2002-12-301-1/+1
| | | | | | | | | that was causing a segfault at httpd startup Submitted by: Erik Abele <erik@codefaktor.de> Reviewed by: Brian Pane git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64239 13f79535-47bb-0310-9956-ffa450edef68
* Fix up code in thread_cond. First always escape the case where we outrightwrowe2002-12-291-23/+72
| | | | | | | | | | | | | fail a single WaitForSingleObject(..., INFINITE) (yes, it can happen) and properly unwind the *correct* error when we fail the nested Wait. Also clean up the usage where rv is an apr result, and res in an NT result to conform a bit better with other code. Finally implements thread_cond_wait_with_timeout. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64238 13f79535-47bb-0310-9956-ffa450edef68
* Get the new DSO error values working on Windows. Windows now passes allrbb2002-12-291-0/+8
| | | | | | | DSO tests. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64237 13f79535-47bb-0310-9956-ffa450edef68
* No pthread_setconcurrency here on Darwin.wrowe2002-12-291-1/+1
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64236 13f79535-47bb-0310-9956-ffa450edef68
* Close the last testdir emit by always calling FindFirstFile on open/rewindwrowe2002-12-292-3/+45
| | | | | | | | | invocations. We don't fill-in the finfo until we invoke apr_dir_read the second time and onwards (following the initial invocation of apr_dir_read by the apr_dir_open or apr_dir_rewind functions.) git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64235 13f79535-47bb-0310-9956-ffa450edef68
* Add the new error codes for DSO failures.rbb2002-12-291-0/+5
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64234 13f79535-47bb-0310-9956-ffa450edef68
* standardize on a couple of easy to test for error codes for DSO errors.rbb2002-12-296-17/+19
| | | | | | | This doesn't work fully on Windows yet, that is the next commit. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64233 13f79535-47bb-0310-9956-ffa450edef68
* apr_dir_make_recursive is now implemented. The unix code is a littlewrowe2002-12-291-1/+45
| | | | | | | | wonky, lots of redundant strdups. See this code if you care to streamline that port a bit. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64232 13f79535-47bb-0310-9956-ffa450edef68
* Only APR_FINFO_DIRENT is guarenteed to be supported... other platformswrowe2002-12-291-5/+5
| | | | | | | | | may be missing things such as APR_FINFO_GROUP, APR_FINFO_INODE etc. It would be healthy to add a {seperate} test that APR_FINFO_MIN at least succeeds, but that should not be part of the mainline tests. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64231 13f79535-47bb-0310-9956-ffa450edef68
* Remove some code that we can never reach. If we fail therbb2002-12-291-12/+1
| | | | | | | | CuAssertPtrEquals, the test suite will longjmp. If we pass it, then the pointer is NULL, and we won't go to the internals of the if statement. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64230 13f79535-47bb-0310-9956-ffa450edef68
* Get us building on unix again.rbb2002-12-291-1/+1
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64229 13f79535-47bb-0310-9956-ffa450edef68
* Defer allocation of the apr_file_t until the file is successfully openedwrowe2002-12-291-33/+38
| | | | | | | (and the thread_mutex, when required, was successfully created.) git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64228 13f79535-47bb-0310-9956-ffa450edef68
* Add an internal Win32 apr_file_open flag APR_OPENINFO to allow APR itselfwrowe2002-12-293-16/+32
| | | | | | | | | to access a Win32 file or directory without READ/WRITE access, for various GetFileInformationEx() and security descriptor access. Now the testfile suite passes on Win32. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64227 13f79535-47bb-0310-9956-ffa450edef68
* apr_file_open(f, "directory", ..., p) is not portable, test for it.wrowe2002-12-291-18/+22
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64226 13f79535-47bb-0310-9956-ffa450edef68
* Readd the test removed since rev 1.60. Let others untangle the implications,wrowe2002-12-291-0/+22
| | | | | | | | since only two coders care {again} and don't share the same opinion this discussion is entirely moot. Note this test fails on Win32. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64225 13f79535-47bb-0310-9956-ffa450edef68
* Reclassify LIB and LIB2 as MOD and LIB (corresponding to loadable moduleswrowe2002-12-291-34/+41
| | | | | | | | and libraries, respectively.) Re-introduce a single CuNotImpl for those cases where LIB is undefined. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64224 13f79535-47bb-0310-9956-ffa450edef68
* Revert the bit from rev 1.31 that Rbb objects to... continue to testwrowe2002-12-291-0/+3
| | | | | | | undefined/undocumented behavior. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64223 13f79535-47bb-0310-9956-ffa450edef68
* Apply a patch to fix make check in test from Matt Kraai, extended to runthommay2002-12-291-4/+4
| | | | | | | | | testall during make check. Submitted by: Matt Kraai <kraai@alumni.cmu.edu> Reviewed by: Thom May git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64222 13f79535-47bb-0310-9956-ffa450edef68
* New ignore goodnesswrowe2002-12-291-1/+6
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64221 13f79535-47bb-0310-9956-ffa450edef68
* Get testall building within the studio environment for debugging on win32.wrowe2002-12-293-17/+389
| | | | | | | | This is BS to maintain two build files, but as long as everyone shares the responsibility for keeping them in sync, I can ignore this from now on. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64220 13f79535-47bb-0310-9956-ffa450edef68
* After recovering the 'manditory' bits, confirm that fields are identicalwrowe2002-12-291-27/+61
| | | | | | | | | | | | | | | | one-by-one for diagnostics. You can't simply memcmp() all of the fields, such as ->valid, because some platforms don't obtain the same information from both the stat() of a filename and the getfileinfo() of an open handle. Even the filename is "unknowable" against a file handle, since one of many filenames could have been used to access that file. Finally, there is one test that (correctly) fails, the APR_FINFO_PROT values on Win32 don't agree. I suspect we are really getting the perms of the 'open file handle' rather than the 'filesystem object'. Researching. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64219 13f79535-47bb-0310-9956-ffa450edef68
* Remove a bogus test. All of the original architects have argued againstwrowe2002-12-291-22/+0
| | | | | | | | | | | | | | | | testing for bogus (invalid) combinations of bits. This test checked a POSIX behavior of requiring the caller to request read and/or write access. On Win32 and some other platforms that support metadata access, opening a file internally for no read or write access is entirely valid. E.g. to modify the meta information, especially about directories on platforms that don't support opening 'directories' for read/write access, APR already uses apr_file_open() to handle the file metadata access. I see no compelling reason to emulate POSIX's fail-on-no-read-or-write-request behavior. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64218 13f79535-47bb-0310-9956-ffa450edef68
* Clean up these dup2 tests by retaining the original stderr file whilewrowe2002-12-291-21/+39
| | | | | | | | testing the substitution of an alternate stderr (with dup2) and finally replacing stderr with it's original, rather than closing it. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64217 13f79535-47bb-0310-9956-ffa450edef68
* First; once any apr object is closed, the results are undefined.wrowe2002-12-291-32/+12
| | | | | | | | | Second; if a platform does not use distinct 'libraries' and 'modules', the four _non_module tests (should have been named _library tests) don't apply; this includes Netware, Win32, BeOS and OS2. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64216 13f79535-47bb-0310-9956-ffa450edef68
* You cannot close files that failed to open.wrowe2002-12-281-2/+10
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64215 13f79535-47bb-0310-9956-ffa450edef68
* Fix several bugs where handle creation failed. Win32 doesn't (and otherwrowe2002-12-281-7/+11
| | | | | | | platforms shouldn't) accept NULL as apr_file_close. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64214 13f79535-47bb-0310-9956-ffa450edef68
* Grudgingly accept that folks may want to merge paths containing wildcardwrowe2002-12-282-1/+13
| | | | | | | | | characters... fail similarly to ENOENT/ENOTDIR, where we return the path but emit the error as a precaution. Folks interested in the results of an invalid path merge need to look for those APR_STATUS_IS_xxx() results. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64213 13f79535-47bb-0310-9956-ffa450edef68
* Introduce APR_EPATHWILD to designate that a given apr_file_[l]stat()wrowe2002-12-282-1/+8
| | | | | | | | was against a filename that included wildcards (on platforms where wildcards are invalid filename characters, e.g. OS2, Win32 etc.) git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64212 13f79535-47bb-0310-9956-ffa450edef68
* Several StdHandles may be defined as the same handle, handle that case.wrowe2002-12-282-10/+15
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64211 13f79535-47bb-0310-9956-ffa450edef68
* Remove the segfault from testfile.c and introduce new breakage.wrowe2002-12-281-0/+15
| | | | | | | | | | | | | If the filehandle we are closing corresponds to a standard handle, we must protect against these dup2'ed handles in file_cleanup and invalidate the corresponding StdHandle entries, or they may point (in the future) to a newly reused handle identifier. Why testfile.c thinks it can close a std handle and then simply obtain it again is beyond me. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64210 13f79535-47bb-0310-9956-ffa450edef68
* OS/2: Fix file open flag manipulation so that APR_EXCL actually works. Thisbjh2002-12-261-1/+1
| | | | | | | also fixes a number of other testfile failures. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64209 13f79535-47bb-0310-9956-ffa450edef68
* OS/2: Fill out the possible OS error codes for APR_STATUS_IS_EEXIST.bjh2002-12-261-1/+4
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64208 13f79535-47bb-0310-9956-ffa450edef68
* I forgot the ;.rbb2002-12-241-1/+1
| | | | | | | Submitted by: Ben Laurie <ben@algroup.co.uk> git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64207 13f79535-47bb-0310-9956-ffa450edef68
* Fix an infinite loop in apr_poll_socket_clear.rbb2002-12-241-0/+1
| | | | | | | Submitted by: INOUE Seiichiro <inoue@ariel-networks.com> git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64206 13f79535-47bb-0310-9956-ffa450edef68
* Allow apr_hash to have greater than int number of elements.jerenkrantz2002-12-233-11/+15
| | | | | | | (serf_spider needs a ridiculously large hash table.) git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64205 13f79535-47bb-0310-9956-ffa450edef68
* As silly as this stub appears, it at least provides that apr may bewrowe2002-12-231-0/+41
| | | | | | | build (trivially) outside of any other build schema. git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64204 13f79535-47bb-0310-9956-ffa450edef68
* Fixed prototype mismatchbnicholes2002-12-201-1/+1
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64203 13f79535-47bb-0310-9956-ffa450edef68
* NetWare makefilebnicholes2002-12-201-0/+255
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64202 13f79535-47bb-0310-9956-ffa450edef68
* Making the APR tests run on NetWarebnicholes2002-12-204-1/+28
| | | | git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@64201 13f79535-47bb-0310-9956-ffa450edef68
View Lorry Controller Status