That was not the PITA. My method of finding it was to find the code that was being called right before the crash and try to fix that code, due to the fact that code must have caused the error since it was just called. Even now I am not really sure what the exact problem was since I just commented out the offending code and replicated the functionality with some other minor changes and the use of the standard library.
The crash on next day bug on the other hand is a total PITA, and still might exist, since it is totaly random and thus near impossible form me to track down.