go to bug id or search bugs for
I have a cron job set to regularly run the following:
/usr/local/bin/pecl upgrade --ignore-errors timezonedb
Earlier today it picked up the new 2015.6 timezonedb update from PECL and PHP 5.6.10 (at the time installed) immediately started spitting segmentation faults under all cases.
I upgraded to 5.6.12 in an attempt to solve the problem and the issue still persists.
Add a Patch
Add a Pull Request
we have many centos 6 cpanel server with php 5.5.28 ( compiled with easyapache )and timezonedb 2015.6 ( from pecl ) we have segfault of php request with many cms , downgrading timezonedb to 2015.5 we have no problem .
so i can confirm bug with timezonedb 2015.6
PHP just cut out at midnight server-time (EDT). Loads of segmentation faults. Tried rebuilding a bunch of stuff, no dice. Tracked it down to the timezonedb.so extension, commented it out, server came back up.
PHP 5.6.11 (cli) (built: Aug 15 2015 01:08:46)
Copyright (c) 1997-2015 The PHP Group
Zend Engine v2.6.0, Copyright (c) 1998-2015 Zend Technologies
with Zend OPcache v7.0.6-dev, Copyright (c) 1999-2015, by Zend Technologies
Running on FreeBSD 10.1-RELEASE-p16 in a jail.
Ports built with clang36 -O3 with gold linker in case it matters.
It looks like the issue caused by condition
#if PHP_VERSION_ID >= 70000
within timezonedb.h (of course when php version is less than 7)
Adding "#define PHP_VERSION_ID 70000" before "#if PHP_VERSION_ID >= 70000" produces correct (no SEGSERV) timezonedb.so.
Checked on PHP 5.5.27, CentOS 6.6
This should be fixed in both GIT master (and PHP-5.6), and the PECL release 2015.6.1
I can confirm this is fixed, thank you.