summaryrefslogtreecommitdiff
path: root/TSRM
Commit message (Collapse)AuthorAgeFilesLines
* MFH: allocate_new_resource() handles the malloc() for us. Don't leak by ↵Sara Golemon2005-08-291-1/+0
| | | | mallocing twice.
* Fix compile failure with GCC 4.0.xfoobar2005-08-051-1/+1
|
* Bump up yearfoobar2005-08-036-6/+6
|
* Netware pthreads support is good enough. So removing all these fixes of old ↵Anantha Kesari H Y2005-07-161-18/+0
| | | | | | | times. -- Kamesh
* Reverting the DEFAULT_SLASH to / as it breaks the FreeBSD derived function ↵Anantha Kesari H Y2005-07-071-1/+1
| | | | | | | of NetWare LibC like glob. -- Kamesh
* gwtcwd of NetWare LibC gives a cwd with a volume information.Anantha Kesari H Y2005-07-071-0/+17
| | | | | | | | So using getcwdpath which gives with volume information. getcwdpath gives with directory seperator as \ which is against our DEFAULT_SLASH of /. So finding and replacing \ with / -- Kamesh
* COPY_WHEN_ABSOLUTE is not needed for NetWare.Anantha Kesari H Y2005-07-071-2/+0
| | | | | -- Kamesh
* - Unify the "configure --help" textsfoobar2005-05-291-1/+1
|
* Unterminated paren group in BETHREADS definition of tsrm_tls_set() macroSara Golemon2005-05-161-1/+1
|
* consolidate the various tls get/set calls for the different platforms intoWez Furlong2005-04-281-58/+23
| | | | | a single macro, at Andi's request.
* Add three new TSRM api functions:Wez Furlong2005-04-272-0/+96
| | | | | | | | | | TSRM_API void *tsrm_new_interpreter_context(void); TSRM_API void *tsrm_set_interpreter_context(void *new_ctx); TSRM_API void tsrm_free_interpreter_context(void *context); These can be used, with a suitable SAPI, to host multiple interpreters on the same thread.
* check should be on host_alias rather than on target_alias, but somehow works ↵Anantha Kesari H Y2005-04-271-1/+1
| | | | on configure generated using autoconf 2.57.
* MergeZeev Suraski2005-03-201-2/+2
|
* Revert bad fixfoobar2005-03-172-2/+2
|
* Fix Jani breakageRasmus Lerdorf2005-03-162-2/+2
|
* Add a comment to remind me to have a look at this at some point.Rasmus Lerdorf2005-02-271-1/+1
|
* - cross-compile fixfoobar2005-02-241-1/+9
|
* NetWare can use Autoconf based build system.Anantha Kesari H Y2005-02-221-11/+0
|
* NetWare LibC has lstat implementationAnantha Kesari H Y2005-02-211-2/+2
|
* NetWare can make use of autogenerated config header fileAnantha Kesari H Y2005-02-211-2/+0
|
* NetWare has pthreads implementationAnantha Kesari H Y2005-02-211-22/+0
|
* NetWare can use the tsrm_config.hAnantha Kesari H Y2005-02-211-1/+1
|
* NetWare LibC has a lstat implementationAnantha Kesari H Y2005-02-211-1/+1
|
* NetWare has a compliant pthread implementationAnantha Kesari H Y2005-02-011-14/+0
|
* finally use what we calculatedStefan Esser2005-01-151-1/+1
|
* - Fix outside-source-tree builds. Always include generated header filesfoobar2005-01-092-2/+2
| | | | | with #include <some_header.h> to make sure the correct file is used.
* Unnecesary check was removedDmitry Stogov2005-01-081-1/+1
|
* MFB_4_3: Quote macro names in AC_DEFUN()foobar2004-12-303-11/+11
|
* Fixed: virtual_popen in ZTS modeStefan Esser2004-12-021-4/+28
|
* Partly reverted because of problems with some realpath() implementationsStefan Esser2004-12-021-2/+4
|
* Fixed behaviour of TSRM in case of realpath() failureStefan Esser2004-12-011-7/+22
|
* removed redundant CLIB_STAT_PATCH checks. defined the IS_ABSOLUTE_PATH ↵Anantha Kesari H Y2004-10-081-6/+3
| | | | macros to suit the NetWare filesystem convention
* removed redundant NEW_LIBC checks and added the code to recognize the file ↵Anantha Kesari H Y2004-10-081-35/+13
| | | | naming convention in NetWare(supports Windows based drive names like sys: data: and unix like / (sys: being root)
* Fixed compiler warning.Ilia Alshanetsky2004-10-061-0/+1
|
* - Apply realpath() cache patch. We don't use it if we're in safe_mode andAndi Gutmans2004-10-052-17/+163
| | | | | | - friends (which are quite slow anyway). - If it proves to be stable I'll remove the #ifdef's in a few weeks.
* Apply the fixMarcus Boerger2004-07-151-10/+0
|
* - Fix typo although it isn't being builtAndi Gutmans2004-05-231-1/+1
|
* - Once ts_free_id() will be used, we cannot call ctors for deleted idsMarcus Boerger2004-05-131-3/+10
| | | | | when creating new threads.
* - More work on ts_free_id() but still keep it outMarcus Boerger2004-05-121-4/+11
|
* ifdef out the ts_free_id until php 5.1 and until it is completeMarcus Boerger2004-05-121-0/+4
|
* - Implement ts_free_id()Marcus Boerger2004-05-121-11/+32
|
* Disabled calling dtors on tsrm_shutdown - calling dtors in tsrm_shutdown makesStanislav Malyshev2004-03-301-0/+8
| | | | | | | modules registering TSRM ids to crash, if they have dtors, since the module is unloaded before tsrm_shutdown is called. Can be re-enabled after tsrm_free_id is implemented.
* Fixed define.Ilia Alshanetsky2004-03-091-0/+1
|
* Fixed bug #27338 (memory leak inside tsrm_virtual_cwd.c on win32).Ilia Alshanetsky2004-03-091-2/+7
|
* - Happy new year and PHP 5 for rest of the files too..foobar2004-01-083-6/+6
| | | | | # Should the LICENSE and Zend/LICENSE dates be updated too?
* - A belated happy holidays and PHP 5Andi Gutmans2004-01-083-6/+6
|
* It seems that making the NewAPIs.h header globally visible causesWez Furlong2003-12-223-9/+5
| | | | | | | problems when compiling the com_dotnet extension, so I'm moving it to be visible only to the VCWD code. Thanks Frank :-)
* Fix bug #26692; Could not find GetLongPathNameA under NT4.Wez Furlong2003-12-223-0/+20
| | | | | | | | | This fix relies on you having the Platform SDK headers. Using the new build system, the NewAPIs.h header will be detected automatically, however, for people building using .dsp files, you will need to add /DHAVE_NEWAPIS_H=1 to the TSRM project if you have those newer headers.
* Removed doubled loop from wrongly applied patchMarcus Boerger2003-12-141-5/+0
|
* Fix from Michael Sisolak (destructors weren't always properly called)Zeev Suraski2003-12-141-1/+6
|