PHP

UPDATE, 2:42a.m. - found the issue, seems to be some referencing stuff. Updated information in the extended entry. Last night I blogged about my frustration with a bug in my ActiveRecord implementation that caused Apache to segfault after calling the save() method (which, strangely enough, never occured on the first request after a fresh Apache restart, but on all subsequent requests). I had already given up on being able to find the bug and was thinking about a workaround - and that workaround seems to have identified the bug for me. But there's a catch. After internally building the correct SQL for the desired operation, a method named _executeQuery() is called, which looks up if there already is a previously used prepared statement object for this particular query. If there is, the existing object would be grabbed and simply executed again with the new bindings. If not, a new statement is prepared and written to the statement cache array. In my simple case where only one INSERT query had to be performed, that cache functionality was never used - yet, it caused Apache to segfault. I have rewritten the method so that it does not use prepared statements at all anymore. Now the question is: Is this a PDO bug or is it something else? Read on for the details.

The good news is: I'm on vacation! Yay Well, actually vacation is almost over... I have until Wednesday to become totally relaxed and yet there's so much left to do... need to toss out lots of old stuff, scratch the rust off an old Singer sawing machine rack to refit it as a table and things like that. At least in some respect my vacation has been a little bit productive - since last Wednesday I am now officially a Zend Certified Engineer (most probably the last one to ever do the PHP4 certification... well, I still had this voucher that would have gone void on August 31st - guess when the PHP5 certification is due to hit the road...). Of course, I also wanted to do lots of coding during my vacation - what better method to relax than a good few thousand lines of code, huh? No, I have no life... or do I? Anyway, the coding part was kind of put back when I hit a very nasty bug that I couldn't find yet. Worst of all, I have no idea if it's the stuff I wrote (more likely), parts of the Zend Framework (ruled out already), PHP itself (less likely) or Apache (likeliness hits rock bottom). In the unlikely case you should be interested in my pain, read on (this is a lenghty and rather uninteresting one - you have been warned!!).