Jani Taskinen wrote:
Dmitry Stogov kirjoitti:
Hi Jani,
Unfortunately this patch doesn't fix the bug.
I'm not sure if it fixes module dependency issues (let me know how to
check it), but the original bug has the same two problems.
1) Possible infinity recursion in php_log_err()
I did not
janiThu May 14 21:00:07 2009 UTC
Modified files:
/php-src/build order_by_dep.awk
Log:
- Fix harmless extra for() loop iteration
http://cvs.php.net/viewvc.cgi/php-src/build/order_by_dep.awk?r1=1.2r2=1.3diff_format=u
Index: php-src/build/order_by_dep.awk
janiThu May 14 21:23:44 2009 UTC
Modified files:
/php-src/build order_by_dep.awk
Log:
- Fix the harmless patch that caused harm :)
http://cvs.php.net/viewvc.cgi/php-src/build/order_by_dep.awk?r1=1.3r2=1.4diff_format=u
Index:
Hi Jani,
Unfortunately this patch doesn't fix the bug.
I'm not sure if it fixes module dependency issues (let me know how to
check it), but the original bug has the same two problems.
1) Possible infinity recursion in php_log_err()
2) Uninitialized DATEG(tzcache. It's initialized only during
Hi!
2) Uninitialized DATEG(tzcache. It's initialized only during RINIT() so
any error messages befor this may crash PHP.
This isn't the first time we see extension being used before
initialization. Maybe we need some infrastructure to prevent that.
--
Stanislav Malyshev, Zend Software
On Wed, 13 May 2009, Stanislav Malyshev wrote:
2) Uninitialized DATEG(tzcache. It's initialized only during RINIT() so any
error messages befor this may crash PHP.
This isn't the first time we see extension being used before initialization.
Maybe we need some infrastructure to prevent
Derick Rethans wrote:
On Wed, 13 May 2009, Stanislav Malyshev wrote:
2) Uninitialized DATEG(tzcache. It's initialized only during RINIT() so any
error messages befor this may crash PHP.
This isn't the first time we see extension being used before initialization.
Maybe we need some
Dmitry Stogov kirjoitti:
Hi Jani,
Unfortunately this patch doesn't fix the bug.
I'm not sure if it fixes module dependency issues (let me know how to
check it), but the original bug has the same two problems.
1) Possible infinity recursion in php_log_err()
I did not even try to fix that
Stanislav Malyshev kirjoitti:
Hi!
2) Uninitialized DATEG(tzcache. It's initialized only during RINIT()
so any error messages befor this may crash PHP.
This isn't the first time we see extension being used before
initialization. Maybe we need some infrastructure to prevent that.
Yes, we
Dmitry Stogov kirjoitti:
Derick Rethans wrote:
On Wed, 13 May 2009, Stanislav Malyshev wrote:
2) Uninitialized DATEG(tzcache. It's initialized only during RINIT()
so any
error messages befor this may crash PHP.
This isn't the first time we see extension being used before
initialization.
On Wed, 2009-05-13 at 00:45 +, Jani Taskinen wrote:
Log:
- Fixed bug #48247 (PHP crashes on errors during startup)
Looks good for 5.3 to me.
One comment though:
cat $ext_builddir/lib/timelib_config.h EOF
#ifdef PHP_WIN32
# include config.w32.h
#else
# include php_config.h
Johannes Schlüter kirjoitti:
On Wed, 2009-05-13 at 00:45 +, Jani Taskinen wrote:
Log:
- Fixed bug #48247 (PHP crashes on errors during startup)
Looks good for 5.3 to me.
Committed.
One comment though:
cat $ext_builddir/lib/timelib_config.h EOF
#ifdef PHP_WIN32
# include
janiWed May 13 00:45:57 2009 UTC
Added files:
/php-src/ext/date config0.m4
Removed files:
/php-src/ext/date config.m4
Modified files:
/ZendEngine2zend.c
/php-src/build order_by_dep.awk
13 matches
Mail list logo
