]> granicus.if.org Git - php/log
php
8 years agoCompatibility with Firebird & Oracle select syntax.
Dorin Marcoci [Sat, 17 Dec 2016 19:34:32 +0000 (21:34 +0200)]
Compatibility with Firebird & Oracle select syntax.

8 years agofix proto
Anatol Belski [Sun, 18 Dec 2016 23:50:29 +0000 (00:50 +0100)]
fix proto

8 years agoCursor is not opened on singleton selects.
Dorin Marcoci [Thu, 15 Dec 2016 08:57:45 +0000 (10:57 +0200)]
Cursor is not opened on singleton selects.

Test case for unregistered bug on FB3 singleton selects

Set error mode to warning instead of exception.

8 years agofix arg spec and datatype, follow up on 73594
Anatol Belski [Sun, 18 Dec 2016 20:16:35 +0000 (21:16 +0100)]
fix arg spec and datatype, follow up on 73594

8 years agoFixed bug #73594
Bruce Weirdan [Fri, 25 Nov 2016 02:14:14 +0000 (04:14 +0200)]
Fixed bug #73594

Skip the tests when local resolver does not behave

* Added SKIPIF sections to check local resolver
* Added test to check $authns parameter

8 years agoFix bug #46103
Nikita Popov [Sun, 18 Dec 2016 16:10:19 +0000 (17:10 +0100)]
Fix bug #46103

8 years agoAdd test for get_browser() in object mode
Nikita Popov [Sun, 18 Dec 2016 12:16:25 +0000 (13:16 +0100)]
Add test for get_browser() in object mode

All other tests only check array return values...

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Anatol Belski [Sat, 17 Dec 2016 19:45:22 +0000 (20:45 +0100)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  fix C89 compat

8 years agofix C89 compat
Anatol Belski [Sat, 17 Dec 2016 19:43:32 +0000 (20:43 +0100)]
fix C89 compat

8 years agoreplace default appveyor build version
Anatol Belski [Sat, 17 Dec 2016 12:12:34 +0000 (13:12 +0100)]
replace default appveyor build version

8 years agoadd Appveyor fast_fail
Anatol Belski [Sat, 17 Dec 2016 10:26:13 +0000 (11:26 +0100)]
add Appveyor fast_fail

8 years agoimprove error check
Anatol Belski [Sat, 17 Dec 2016 10:25:06 +0000 (11:25 +0100)]
improve error check

8 years agoturn mcrypt back on for Appveyor runs
Anatol Belski [Fri, 16 Dec 2016 21:59:47 +0000 (22:59 +0100)]
turn mcrypt back on for Appveyor runs

(cherry picked from commit fca2312fbecbc8eb44639d5141bfeae48976206b)

8 years agoadd build status badge
Anatol Belski [Fri, 16 Dec 2016 10:53:41 +0000 (11:53 +0100)]
add build status badge

(cherry picked from commit 3a886c9d20f450545a0d93bb4fc974e28a50df85)

8 years agoadd scripts for AppVeyor integration
Anatol Belski [Fri, 16 Dec 2016 10:48:08 +0000 (11:48 +0100)]
add scripts for AppVeyor integration

(cherry picked from commit c737d6fb495ee986f7819472db861aa9e1bf42b1)

8 years agoSilence warning from unhandled enum
Sara Golemon [Tue, 13 Dec 2016 05:46:49 +0000 (21:46 -0800)]
Silence warning from unhandled enum

(cherry picked from commit 57bbe2c140752f491b1fa24336b817bd48f65a93)

8 years agofix leaking streams and memory mapped files
Anatol Belski [Sun, 11 Dec 2016 01:52:02 +0000 (02:52 +0100)]
fix leaking streams and memory mapped files

(cherry picked from commit f1ff23095b1a4fe6d6a65331dda7832ae02eb1a1)

8 years agoFix stack management in ini scanner
Nikita Popov [Thu, 15 Dec 2016 21:16:44 +0000 (22:16 +0100)]
Fix stack management in ini scanner

8 years agomake timing check more forgiving in these tests
Anatol Belski [Thu, 15 Dec 2016 11:20:37 +0000 (12:20 +0100)]
make timing check more forgiving in these tests

Particularly on slower VMs, the sporadic fails can still happen.
The timing is kept in an uncritical range, but allows the tests
pass there. Mayby, it'd make sense to introduce a new group for
this kind of tests, so tests requiring exact time measurement
can be avoided on unsuitable environments.

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Matteo Beccati [Thu, 15 Dec 2016 08:31:54 +0000 (09:31 +0100)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  Skip tests when secure_file_priv dir not writable

8 years agoSkip tests when secure_file_priv dir not writable
Matteo Beccati [Thu, 15 Dec 2016 08:31:00 +0000 (09:31 +0100)]
Skip tests when secure_file_priv dir not writable

8 years agoAdd missing UPGRADING entry. Manual is updated before 7.0 release.
Yasuo Ohgaki [Thu, 15 Dec 2016 07:16:45 +0000 (16:16 +0900)]
Add missing UPGRADING entry. Manual is updated before 7.0 release.

8 years agoextend skip section
Anatol Belski [Wed, 14 Dec 2016 18:53:55 +0000 (19:53 +0100)]
extend skip section

The day light offset is calculated based on the system TZ, so with
an unsuitable one there's no way to workaround the false positives.

8 years agoPartially fix bug #70492
Nikita Popov [Wed, 14 Dec 2016 17:10:36 +0000 (18:10 +0100)]
Partially fix bug #70492

The libmagic portion is missing.

8 years agocleanup temporary data
Anatol Belski [Wed, 14 Dec 2016 11:33:13 +0000 (12:33 +0100)]
cleanup temporary data

8 years agoMore fixes for bug #73089
Nikita Popov [Tue, 13 Dec 2016 16:54:30 +0000 (17:54 +0100)]
More fixes for bug #73089

Insert a lot more DEREFs. Some may not be necessary, but better
safe than sorry. I'm thinking the soap_hash_str_find_deref function
should become part of the zend_hash API -- we need this in many
places.

8 years agoAdd a bit more test cases (edge cases) to the BIT test case
Andrey Hristov [Mon, 12 Dec 2016 19:52:36 +0000 (21:52 +0200)]
Add a bit more test cases (edge cases) to the BIT test case

8 years agoAdd a test case to test fetching of multiple rows with bit values.
Andrey Hristov [Mon, 12 Dec 2016 19:28:33 +0000 (21:28 +0200)]
Add a test case to test fetching of multiple rows with bit values.

8 years agoFix off by 1 problem.
Andrey Hristov [Mon, 12 Dec 2016 19:11:02 +0000 (21:11 +0200)]
Fix off by 1 problem.

The problem was manifestated only with BIT columns and only when more than
one row was fetched. The problem was coming from the fact that in pre-7.0
times mysqlnd was using a no-copy optimization. This optimization kept the
strings (and also the BIT mask equivalents as strings) in the packet and the
zval referred to them. 7.0+ zvals cannot use no-copy and always copy. Because
of this the allocated memory for the packet was reduced by 1 by the person who
ported the driver, but the starting address of the bit area wasn't reduced.
Because of this the bit_area started at wrong address and the length decoded
wrong.

8 years agobetter way to get ACP
Anatol Belski [Mon, 12 Dec 2016 00:54:22 +0000 (01:54 +0100)]
better way to get ACP

8 years agofix datatype for zpp, yet a followup on bug #73679
Anatol Belski [Mon, 12 Dec 2016 00:26:06 +0000 (01:26 +0100)]
fix datatype for zpp, yet a followup on bug #73679

8 years agoupdate NEWS
Anatol Belski [Sun, 11 Dec 2016 16:39:08 +0000 (17:39 +0100)]
update NEWS

8 years agoFixed bug #73679 DOTNET read access violation using invalid codepage
Anatol Belski [Sun, 11 Dec 2016 16:06:55 +0000 (17:06 +0100)]
Fixed bug #73679 DOTNET read access violation using invalid codepage

8 years agofix uninitialized member
Anatol Belski [Thu, 8 Dec 2016 16:15:06 +0000 (17:15 +0100)]
fix uninitialized member

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Ferenc Kovacs [Thu, 8 Dec 2016 06:09:41 +0000 (07:09 +0100)]
Merge branch 'PHP-5.6' into PHP-7.0

8 years agoupdate NEWS
Ferenc Kovacs [Thu, 8 Dec 2016 06:05:32 +0000 (07:05 +0100)]
update NEWS

8 years agogit clean section
Anatol Belski [Thu, 8 Dec 2016 02:24:18 +0000 (03:24 +0100)]
git clean section

8 years agoupdate NEWS
Anatol Belski [Wed, 7 Dec 2016 20:22:34 +0000 (21:22 +0100)]
update NEWS

8 years agoFixed bug #72931 PDO_FIREBIRD with Firebird 3.0 not work on returning statement
Dorin Marcoci [Wed, 7 Dec 2016 20:06:11 +0000 (21:06 +0100)]
Fixed bug #72931 PDO_FIREBIRD with Firebird 3.0 not work on returning statement

8 years agoFixed typo in skipif error message
Markus Staab [Mon, 5 Dec 2016 19:57:03 +0000 (20:57 +0100)]
Fixed typo in skipif error message

8 years agoCheck if PHP is in execution state.
Dmitry Stogov [Wed, 7 Dec 2016 12:10:59 +0000 (15:10 +0300)]
Check if PHP is in execution state.

8 years agoupdate NEWS
Anatol Belski [Tue, 6 Dec 2016 19:25:27 +0000 (20:25 +0100)]
update NEWS

8 years agoFixed issue getting executable lines from custom wrappers
Bob Weinand [Tue, 6 Dec 2016 17:27:31 +0000 (18:27 +0100)]
Fixed issue getting executable lines from custom wrappers

See also krakjoe/phpdbg#161

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Anatol Belski [Tue, 6 Dec 2016 15:17:28 +0000 (16:17 +0100)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  fix leak, take 2

8 years agofix leak, take 2
Anatol Belski [Tue, 6 Dec 2016 15:12:39 +0000 (16:12 +0100)]
fix leak, take 2

8 years agofix leak, take on 7.x
Anatol Belski [Tue, 6 Dec 2016 13:42:59 +0000 (14:42 +0100)]
fix leak, take on 7.x

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Anatol Belski [Tue, 6 Dec 2016 13:36:52 +0000 (14:36 +0100)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  fix leak, take on 5.6

8 years agofix leak, take on 5.6
Anatol Belski [Tue, 6 Dec 2016 13:34:27 +0000 (14:34 +0100)]
fix leak, take on 5.6

8 years agoFixed bug #73663 ("Invalid opcode 65/16/8" occurs with a variable created with list())
Xinchen Hui [Tue, 6 Dec 2016 09:17:59 +0000 (17:17 +0800)]
Fixed bug #73663 ("Invalid opcode 65/16/8" occurs with a variable created with list())

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Stanislav Malyshev [Tue, 6 Dec 2016 06:33:33 +0000 (22:33 -0800)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  This still leaks memory, I don't have enough knowledge in WDDX code to fix them :(

8 years agoThis still leaks memory, I don't have enough knowledge in WDDX code to fix them :(
Stanislav Malyshev [Tue, 6 Dec 2016 06:32:59 +0000 (22:32 -0800)]
This still leaks memory, I don't have enough knowledge in WDDX code to fix them :(

8 years agoFix bug #73645 - int/size_t confusion
Stanislav Malyshev [Tue, 6 Dec 2016 06:16:00 +0000 (22:16 -0800)]
Fix bug #73645 - int/size_t confusion

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Stanislav Malyshev [Tue, 6 Dec 2016 05:58:55 +0000 (21:58 -0800)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  Fix bug #73631 - Invalid read when wddx decodes empty boolean element

8 years agoFix bug #73631 - Invalid read when wddx decodes empty boolean element
Stanislav Malyshev [Tue, 6 Dec 2016 05:40:55 +0000 (21:40 -0800)]
Fix bug #73631 - Invalid read when wddx decodes empty boolean element

8 years agoFixed behavior of failing compound assignments (they shouldn't change the source...
Dmitry Stogov [Mon, 5 Dec 2016 18:45:08 +0000 (21:45 +0300)]
Fixed behavior of failing compound assignments (they shouldn't change the source value when exception thrown during type converion).

8 years agofix test
Anatol Belski [Sun, 4 Dec 2016 21:42:39 +0000 (22:42 +0100)]
fix test

This test depends on connection flags modifable from teh env,
that can affect the exact function called.

8 years agofix test
Anatol Belski [Sun, 4 Dec 2016 21:27:17 +0000 (22:27 +0100)]
fix test

On Windows, either doubled backslashes or forward slashes are supported
on the server side.

8 years agofix ident
Anatol Belski [Sun, 4 Dec 2016 19:59:12 +0000 (20:59 +0100)]
fix ident

8 years agofix test
Anatol Belski [Sun, 4 Dec 2016 16:28:43 +0000 (17:28 +0100)]
fix test

8 years agoenforce the test ini usage for the child process
Anatol Belski [Sun, 4 Dec 2016 13:01:40 +0000 (14:01 +0100)]
enforce the test ini usage for the child process

8 years agoFix PHP-7.0 build (wrong fix for the branch)
Bob Weinand [Sat, 3 Dec 2016 22:37:23 +0000 (23:37 +0100)]
Fix PHP-7.0 build (wrong fix for the branch)

8 years agoFix leak when using $this outside of object context
Bob Weinand [Sat, 3 Dec 2016 17:23:58 +0000 (18:23 +0100)]
Fix leak when using $this outside of object context

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Christoph M. Becker [Sat, 3 Dec 2016 13:18:00 +0000 (14:18 +0100)]
Merge branch 'PHP-5.6' into PHP-7.0

8 years agoFix minor typo
Adrien Crivelli [Wed, 30 Nov 2016 09:19:35 +0000 (18:19 +0900)]
Fix minor typo

8 years agoFix leak in shift_right_function
Nikita Popov [Fri, 2 Dec 2016 16:03:05 +0000 (17:03 +0100)]
Fix leak in shift_right_function

8 years agoFix test of wrong operator in self_xor.phpt
Bob Weinand [Fri, 2 Dec 2016 14:41:47 +0000 (15:41 +0100)]
Fix test of wrong operator in self_xor.phpt

8 years agoEliminated useless exception check
Dmitry Stogov [Fri, 2 Dec 2016 12:39:33 +0000 (15:39 +0300)]
Eliminated useless exception check

8 years agoFixed behavior of failing compound assignments (they shouldn't change the source...
Dmitry Stogov [Fri, 2 Dec 2016 12:13:55 +0000 (15:13 +0300)]
Fixed behavior of failing compound assignments (they shouldn't change the source value when exception thrown during type converion).

8 years agoFixed behavior of failing compound assignments (they shouldn't change the source...
Dmitry Stogov [Fri, 2 Dec 2016 10:58:44 +0000 (13:58 +0300)]
Fixed behavior of failing compound assignments (they shouldn't change the source value).

8 years agoFixed possible memory leak in &=, |=, ^=.
Dmitry Stogov [Fri, 2 Dec 2016 09:34:46 +0000 (12:34 +0300)]
Fixed possible memory leak in &=, |=, ^=.

8 years agoFix valgrind issue when throwing due to yield from aborted generator
Bob Weinand [Thu, 1 Dec 2016 14:21:28 +0000 (15:21 +0100)]
Fix valgrind issue when throwing due to yield from aborted generator

8 years agofork tests
Anatol Belski [Thu, 1 Dec 2016 13:52:18 +0000 (14:52 +0100)]
fork tests

8 years agoFixed bad merge
Dmitry Stogov [Thu, 1 Dec 2016 13:21:57 +0000 (16:21 +0300)]
Fixed bad merge

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Dmitry Stogov [Thu, 1 Dec 2016 12:50:05 +0000 (15:50 +0300)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  Workaround for GCC-4.9.2 bug
  oops, changed in wrong place

8 years agoWorkaround for GCC-4.9.2 bug
Dmitry Stogov [Thu, 1 Dec 2016 12:46:52 +0000 (15:46 +0300)]
Workaround for GCC-4.9.2 bug

8 years agoimprove skipif
Anatol Belski [Thu, 1 Dec 2016 12:32:10 +0000 (13:32 +0100)]
improve skipif

8 years agoSimpler overflow check
Dmitry Stogov [Thu, 1 Dec 2016 07:30:02 +0000 (10:30 +0300)]
Simpler overflow check

8 years agoMerge branch 'PHP-7.0' of https://git.php.net/push/php-src into PHP-7.0
Nuno Lopes [Wed, 30 Nov 2016 19:41:26 +0000 (14:41 -0500)]
Merge branch 'PHP-7.0' of https://git.php.net/push/php-src into PHP-7.0

8 years agoFix run-tests.php to always include TEST_PHPDBG_EXECUTABLE if phpdbg available
Bob Weinand [Wed, 30 Nov 2016 10:44:10 +0000 (11:44 +0100)]
Fix run-tests.php to always include TEST_PHPDBG_EXECUTABLE if phpdbg available

8 years agoFixed bug #73615 (phpdbg without option never load .phpdbginit at startup)
Bob Weinand [Wed, 30 Nov 2016 01:56:47 +0000 (02:56 +0100)]
Fixed bug #73615 (phpdbg without option never load .phpdbginit at startup)

8 years agoFixed bug #73586 (php_user_filter::$stream is not set to the stream the filter is...
Dmitry Stogov [Mon, 28 Nov 2016 09:54:47 +0000 (12:54 +0300)]
Fixed bug #73586 (php_user_filter::$stream is not set to the stream the filter is working on).

8 years agoFixed calling generators through magic __call()
Dmitry Stogov [Mon, 28 Nov 2016 09:20:19 +0000 (12:20 +0300)]
Fixed calling generators through magic __call()

8 years agooops, changed in wrong place
Stanislav Malyshev [Mon, 28 Nov 2016 00:11:41 +0000 (16:11 -0800)]
oops, changed in wrong place

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Stanislav Malyshev [Sun, 27 Nov 2016 23:38:22 +0000 (15:38 -0800)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  add NEWS

8 years agoadd NEWS
Stanislav Malyshev [Sun, 27 Nov 2016 23:38:09 +0000 (15:38 -0800)]
add NEWS

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Stanislav Malyshev [Sun, 27 Nov 2016 23:36:29 +0000 (15:36 -0800)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  Fix #68447: grapheme_extract take an extra trailing character

8 years agoMerge branch 'pull-request/1974' into PHP-5.6
Stanislav Malyshev [Sun, 27 Nov 2016 23:11:17 +0000 (15:11 -0800)]
Merge branch 'pull-request/1974' into PHP-5.6

* pull-request/1974:
  Fix #68447: grapheme_extract take an extra trailing character

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Stanislav Malyshev [Sun, 27 Nov 2016 22:53:27 +0000 (14:53 -0800)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  Fix #73549: Use after free when stream is passed to imagepng

8 years agoFix #73549: Use after free when stream is passed to imagepng
Christoph M. Becker [Thu, 17 Nov 2016 12:44:30 +0000 (13:44 +0100)]
Fix #73549: Use after free when stream is passed to imagepng

If a stream is passed to imagepng() or other image output functions,
opposed to a filename, we must not close this stream.

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Matteo Beccati [Sun, 27 Nov 2016 09:20:38 +0000 (10:20 +0100)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  Fix occasionaly failing test

8 years agoFix occasionaly failing test
Matteo Beccati [Sun, 27 Nov 2016 09:20:13 +0000 (10:20 +0100)]
Fix occasionaly failing test

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Matteo Beccati [Sun, 27 Nov 2016 09:08:04 +0000 (10:08 +0100)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  Added missing array key to $JUNIT

8 years agoAdded missing array key to $JUNIT
Matteo Beccati [Sun, 27 Nov 2016 09:07:05 +0000 (10:07 +0100)]
Added missing array key to $JUNIT

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Stanislav Malyshev [Sat, 26 Nov 2016 23:36:57 +0000 (15:36 -0800)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  acinclude.m4: fix krb5-config detection and usage in PHP_SETUP_KERBEROS.
  Add more mbfl string size checks (bug #73505)

8 years agoacinclude.m4: fix krb5-config detection and usage in PHP_SETUP_KERBEROS.
Michael Orlitzky [Fri, 30 Sep 2016 23:47:20 +0000 (19:47 -0400)]
acinclude.m4: fix krb5-config detection and usage in PHP_SETUP_KERBEROS.

When building with kerberos support (--with-kerberos), a few libraries
and flags need to be added to various parts of the build system. The
most reliable way to get those flags is through the krb5-config
program that ships with both major implementations of kerberos. The
PHP_SETUP_KERBEROS macro in acinclude.m4 attempts to detect
krb5-config, and use it.

However, there's a bug in that macro. The --with-kerberos parameter
accepts a directory where the kerberos libraries can be found. When a
directory is given, it is stored in the PHP_KERBEROS variable. The
following test,

  if test "$PHP_KERBEROS" = "yes" && test -x "$KRB5_CONFIG"; then

thus fails whenever a directory is passed to --with-kerberos, since it
compares a directory name against the string "yes". This causes
krb5-config to go unused, and some unreliable fallback logic is
attempted instead. One consequence of this is that the Heimdal
kerberos implementation cannot be substituted for the MIT one, at
least when a directory is passed to --with-kerberos.

This commit reverses the logic and checks for "$PHP_KERBEROS" != "no".
To confirm that this fixes the issue, one can inspect the "-l" library
flags that get appended to the command-line. On a machine with Heimdal
and the unmodified acinclude.m4, running

  ./configure --with-openssl --with-kerberos=/usr

will log (for example) to config.log,

  configure:18082: checking for krb5-config
  configure:18101: found /usr/bin/krb5-config
  configure:18114: result: /usr/bin/krb5-config
  configure:18450: checking for RAND_egd
  configure:18450: cc ... conftest.c ... -lgssapi_krb5 -lkrb5 ...

which are the library names for the MIT implementation. After patching
acinclude.m4 to negate the logic, the same command on the same machine
outputs (to config.log):

  configure:18450: cc ... conftest.c -lgssapi -lheimntlm ...

These are the correct library names for the Heimdal implementation.

PHP-Bug: 73214

8 years agoAdd more mbfl string size checks (bug #73505)
Stanislav Malyshev [Sat, 26 Nov 2016 22:44:58 +0000 (14:44 -0800)]
Add more mbfl string size checks (bug #73505)

8 years agoMerge branch 'PHP-5.6' into PHP-7.0
Stanislav Malyshev [Sat, 26 Nov 2016 22:48:40 +0000 (14:48 -0800)]
Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6:
  Add more mbfl string size checks (bug #73505)

8 years agoAdd more mbfl string size checks (bug #73505)
Stanislav Malyshev [Sat, 26 Nov 2016 22:44:58 +0000 (14:44 -0800)]
Add more mbfl string size checks (bug #73505)

8 years agoFix #73612: preg_*() may leak memory
Christoph M. Becker [Sat, 26 Nov 2016 14:34:27 +0000 (15:34 +0100)]
Fix #73612: preg_*() may leak memory

We have to make sure that collectible zvals end up in the GC root
buffer, to avoid memory leaks.

8 years agoFix more size_t/int implicit conversions
Stanislav Malyshev [Fri, 25 Nov 2016 23:30:20 +0000 (15:30 -0800)]
Fix more size_t/int implicit conversions

Now the conversions are explicit and do checks. Not sure it's
the best way but at least we can see them now in the open.