Mail Archives: djgpp-workers/2001/09/20/13:11:33
> May I suggest, once again, to step with a debugger into the affected
> code and see what exactly fails there, and how? It can't be terribly
> hard to find the problem, or at least come up with some information
> that will allow to guess what might be the reason.
I just tried that again, and remebered why the 'debug it' approach won't
work. The problem relies on a subprogram getting FD 3 or 4 assigned,
but
if you run the debugger, IT will get FDs 3 and 4 (the subprogram usually
starts at FD 8 or so). So the issue might be manifesting in the
debugger,
but you won't be able to debug it there.
I'll try to get a minimal program put together over the weekend
(annoyingly,
a program that fopen()s a couple files and fprintf()s to them does not
show
any problems), and then add printfs to it (and libc) to (hopefully)
catch
the place where things go wrong.
- Raw text -