JY's Weblog

 jeudi 18 avril 2002

Fighting a bug is like playing chess with Murphy. When you correct the bug, it's a checkmate to Murphy.
  

Of course if you didn't understand anything regarding my bug, don't forget that I'm a nerd.
  

BUG of the day : I was fighting with a bug in a Perl CGI script. Here is something tricky I discovered that I want to share with you (of course, you are my only friend ). The script sent me the right html page when I was testing from IE5/MacOSX but when testing from IE5/Windows I was getting a 500 error code. How is that?! Oh my god! How can the cgi script behave differently on the server side when using 2 different browsers?! Here is what I learnt (I hope I'm right in my analysis!): If the browser receives garbage text just before the http header (yes, before the Content-type thing), it will behave differently. In fact my script would write warning messages before the http header (warnings coming from perl -w with use strict), but it would in fact send back the right html content. IE5/OS X didn't show all the warning text that was before the http header, while IE5/Windows was printing a 500 http error code. This was a tricky bug!
  

Apple releases Mac OS X 10.1.4
  

perl.com: XSP, Taglibs and Pipelines
  

Knowledge management and professional development
  

Huumm. Here is what my Google box looks like at the moment. I guess Googlers are playing with the api right now??
  

I took a screenshot of the api timeout result on a Radio weblog.