path: root/ext
AgeCommit message (Collapse)AuthorFilesLines
2010-05-30- token_get_all now returns false on error, not "", returns array on successJohannes Schlüter1-2/+3
2010-05-30- Added recent Windows versions to php_uname and fix undefined windows versionPierre Joye1-4/+15
2010-05-30- wsPierre Joye1-1/+1
2010-05-30- More trait reflection workJohannes Schlüter4-1/+195
#- I'm not happy about ReflectionClass::getTraitAliases, yet
2010-05-29- Fixed ReflectionClass::isTrait() checking (to not identify abstract class ↵Felipe Pena2-1/+31
as Trait)
2010-05-29- Missing Trait check fixFelipe Pena2-2/+2
2010-05-29- Fixed wrong abstract class identification (it was identified as a Trait)Felipe Pena2-1/+55
2010-05-29- Add some folding marksJohannes Schlüter1-11/+26
2010-05-29- Make reflection aware of traitsJohannes Schlüter3-2/+96
2010-05-29- Fix testsJohannes Schlüter2-2/+2
2010-05-28Fixed bug #50976 (Soap headers Authorization not allowed)Dmitry Stogov1-3/+12
2010-05-28Fixed bug #50762 (in WSDL mode Soap Header handler function only being ↵Dmitry Stogov3-1/+83
called if defined in WSDL). (mephius at gmail dot com)
2010-05-27Fixed compiler warning (pointer mismatch)Kalle Sommer Nielsen1-1/+1
2010-05-27Update creditsJohannes Schlüter1-0/+1
2010-05-27Open server bug: mysqlnd works around it, libmysql fails. Let's skip because ↵Ulf Wendel1-9/+1
MySQL Server version tests are tricky and error prone.
2010-05-27Check before use, because mysqlnd can return NULLAndrey Hristov1-3/+6
and this will crash. However, this can happen only in case of OOM.
2010-05-27Don't use NULL pointer, always check before useAndrey Hristov1-1/+3
2010-05-27Test fixes and updates to work with a wider range of MySQL server configsUlf Wendel3-20/+28
2010-05-27Fix possible crashes, in case of OOM, due to half-bakenAndrey Hristov2-10/+27
2010-05-27Fix crashes in the allocator in case of OOM.Andrey Hristov1-4/+4
2010-05-27Fix deallocation which can crash if the object is half-baken.Andrey Hristov2-6/+11
Happens during OOM.
2010-05-26Remove unused variablePierrick Charron1-1/+0
2010-05-26Don't use conn->net without checking if it is NULL, becauseAndrey Hristov1-1/+3
it can be NULL in OOM conditions. This would cause a crash.
2010-05-26Fix testAndrey Hristov1-7/+11
2010-05-26Remove a function that has no usage. The zval cache was disabled/removedAndrey Hristov5-80/+0
in 5.3.0-RC3.
2010-05-26Add checks to check the underlying pointer in the resource.Andrey Hristov5-77/+98
It can happen that the underlying pointer is freed (like in the case of connection reconnect), and then it cannot be allocated anymore, which means that NULL will be used consequently by the functions that fetch the resource and the crash is immediate.
2010-05-26Added support for object references in recursive serialize() calls. FR #36424Michael Wallner9-245/+289
2010-05-26- Fixed bug #51899 (Parse error in parse_ini_file() function when empy value ↵Felipe Pena1-0/+37
followed by no newline)
2010-05-26- Fixed bug #51905 (ReflectionParameter fails if default value is an array ↵Felipe Pena1-0/+28
with an access to self::)
2010-05-25Handle the situation when MYSQLND_PROTOCOL's methods return NULL.Andrey Hristov4-27/+77
mysqlnd should not crash but gracefully return with an error.
2010-05-25Handle the situation when mysql_init()/mysqlnd_init()Andrey Hristov1-2/+10
return NULL. It is a rare case of OOM but PHP shouldn't crash, if possible.
2010-05-25- Fixed bug #51911 (ReflectionParameter::getDefaultValue() memory leaks with ↵Felipe Pena2-1/+23
constant array)
2010-05-25Gracefully handle OOM in mysqlnd_stmt_init. Release the handleAndrey Hristov1-15/+35
and return NULL.
2010-05-25Init should return status value because initialization can fail.Andrey Hristov2-8/+11
In this case mysqlnd_init() should return NULL to the end-user.
2010-05-25Use the specialized macro for this caseAndrey Hristov1-6/+6
2010-05-25Use only memory that has been allocatedAndrey Hristov1-6/+7
2010-05-25Access memory which was really allocatedAndrey Hristov1-25/+48
2010-05-25Fix shared buildAndrey Hristov2-2/+2
2010-05-25Handle OOM in the code that generates PS Execute requestsAndrey Hristov3-25/+64
to the server. Fixes crashes in case of OOM.
2010-05-25Fix metadata duplicatin which wasn't crash-safe in case of OOM.Andrey Hristov1-4/+32
2010-05-25Fixed compatibility with "Run Time Cache" patchDmitry Stogov1-0/+12
2010-05-25Changed test case to use our new dummy MX records.Martin Jansen1-2/+2
2010-05-24Check for -1 as well hereRasmus Lerdorf1-2/+2
2010-05-24Let's have a sane error message here instead of some cryptic zlib errorRasmus Lerdorf1-0/+4
2010-05-24Fixed testsIlia Alshanetsky19-1093/+1093
2010-05-24Fixed ZTS buildDmitry Stogov1-1/+1
2010-05-24Added caches to eliminate repeatable run-time bindings of functions, ↵Dmitry Stogov28-174/+738
classes, constants, methods and properties
2010-05-24Allow ereg to be built as shared on WindowsKalle Sommer Nielsen3-28/+39
## I know that we may remove ereg in trunk, but since the core doesn't depend on ## ereg, then I don't see any reason not to allow this
2010-05-21Ooops, return values...Sara Golemon2-2/+2
2010-05-21Add JSON_BIGINT_AS_STRING for json_decode() to parse large numbersSara Golemon6-16/+83
as strings rather than casting to double and loosing precision.