Re: I think I found the ODBC/Solid failures

> Based on the response this morning to my ODBC/Solid backtrace, I > carefully went over the script in question. The error last night was > from an odbc_free_result call, so I focused on those. I found several > places where I call odbc_free_result() even if there had not been any > rows returned from the odbc_exec(). I was calling odbc_free_result() > in any case, and some would have been empty. > > After meticulously cleaning these up, I cannot reproduce the failure. > It looks like this may have been causing the problem. What threw me > off is that this script has been running for 6 to 9 months, and it is > probably the busiest PHP script on the server. It was originally > written in PHP2, converted to PHP3, and to date has never faulted. I > just did not suspect the PHP3 script itself. > > To those that know the Solid code, does this sound like a probable > cause of the sporadic errors? Aha. Even if it is a problem in the Solid libs, it would be nice if we could somehow prevent the core dump by adding a bit of error checking to PHP. I am not familiar with the Solid API, but might there not be some way to prevent trying to free the result when the returned result is empty? Sounds like it might be a quick and easy check in odbc_free_result(). -Rasmus -- PHP Development Mailing List http://www.php.net/ To unsubscribe send an empty message to php-dev-unsubscribe@lists.php.net For help: php-dev-help@lists.php.net