Re: Another "Premature end of script" headers problem

["Dan Yamins" <dyamins@xxxxxxxxx>]

In my opinion, that would be unlikely, since it sometimes works, and sometimes not, and I guess the output file is always being written in the same place, no ?

Yes, of course you're right.  (Although something about the way the file is written by the .draw method could be dependent on the state of the way the file was written the last time, e.g. something about the set bit... For while, the problem wasn't happening "randomly," it looked like it happened every _other_ time I loaded the script.  Though now it appears to be happening "randomly.")

"premature end of script headers" happens when the cgi-bin script crashes before it has even started to send back a response to the browser.
So I guess your best option would be to add some logging into your script, and try to figure out when it crashes and for what reason.
Whatever your cgi-bin script prints to STDERR will usually find its way into the Apache error log for that server.

The only thing printed to my Apache error log was:

> [Mon Nov 17 17:26:41 2008] [error] [client 127.0.0.1] Premature end of script headers: IntegratedFileBrowser.cgi, referer: http://dataenvironment/System        > /CGI-Executables/main_gui.cgi?Path=../Data/Dan_Data/RandomData/ListOfAuthors.data/&Summary=../Data/Dan_Data/RandomData/ListOfAuthors.data       > /&Fragment=../Data/Dan_Data/RandomData/ListOfAuthors.data/&Representation=../Data/Dan_Data/RandomData/ListOfAuthors.data/

I did a bunch of error logging in my own script, and no failure occurred before the call to the ".draw" method using the third-party module I was using.  I was hoping that there might be some way I didn't have to go in the guts of that third-party module to do error logging, and that something about the pattern of the failure and the way it was happening (e.g. associated a method that writes files out) might give a clue as to the problem.


Thanks
Dan