in reply to OT: Mozilla .9.9 (Linux)
there's really not much anyone here can do to help besides pointing you at the mozilla release page and bugzilla unless you post some code.
having it inaccessible (private or offline) or too large to post is a poor excuse. if you can't debug something, you should always reduce it to a minimal test case. make a copy of the codebase and start removing parts of the code, testing each time. each time you remove a piece, if the error is still reproduced, leave that piece out. if it makes the error go away, put it back in and start taking out other pieces. the goal is to eliminate every possible variable until you wind up with a tiny piece of code by changing one or two lines you can either reproduce the error or not. then post that if it hasn't become obvious where the problem is (99% of the time it will).
i don't mean to be too rough but i'm constantly amazed at how few people seem to be able to grasp this basic strategy. there's no magic secret to debugging code; you just have to go through it systematically and painstakingly. debuggers and logging (print statements or otherwise) can help but nothing i've found works as well as reducing it to a minimal test case.
|
|---|