#21776 [Opn->Fbk]: Heisenbug after database code changes.

ID: 21776 Updated by: iliaa@php.net Reported By: marcus@wordtracker.com -Status: Open +Status: Feedback Bug Type: Unknown/Other Function Operating System: Windows and Linux PHP Version: 4.2.3 New Comment: Please try using this CVS snapshot: http://snaps.php.net/php4-STABLE-latest.tar.gz For Windows: http://snaps.php.net/win32/php4-win32-STABLE-latest.zip Previous Comments: ------------------------------------------------------------------------ [2003-01-23 14:07:28] marcus@wordtracker.com Cannot get a backtrace from the instructions on bug submission. The PHP session does not actually crash, generating a core dump, but exits with a fatal error. The text of the error is gibberish, but nevertheless it does not die. How can we stop or investigate the process at the point of the error. It looks like PHP's symbol table gets corrupted during an earler part of the code. Is there anyway we can investigate this further. yours, Marcus. ------------------------------------------------------------------------ [2003-01-20 10:46:44] georg@php.net Not enough information was provided for us to be able to handle this bug. Please re-read the instructions at http://bugs.php.net/how-to-report.php If you can provide more information, feel free to add it to this bug and change the status back to "Open". Thank you for your interest in PHP. ------------------------------------------------------------------------ [2003-01-20 10:31:29] marcus@wordtracker.com This is the bug from hell. It is random crashing in a section of our unit tests. On Windows the crash manifests itself as... The instruction "(1)" referenced memory at "(2)" which is not writable where (1) is :0x77fcb9b1 (once), 0x77fcb892 (twice), 0x77fcb9fb (once) on a random attempt of two dozen tries. There is too much code to post here and we cannot yet isolate it - our individual unit tests all pass. We are working on a backtrace from a Linux machine and will add to this bug soon. The code that was changed that started generating these problems was mysql related. There are a lot of references in that code (it is a persistent object library). The problem manifests itself in Linux with random fatal errors such as unknown function where the function name is simply "<()" !? It also happens both with version 3 and 4 of MySQL and versions 4.2.3 and 4.3.0 of PHP. It is a complete show stopper. yours, Marcus. ------------------------------------------------------------------------ -- Edit this bug report at http://bugs.php.net/?id=21776&edit=1