Re: Preliminary patches based on dereference of null pointer reports

[Samuel Thibault]

Marin Ramesa, le Sun 17 Nov 2013 07:01:24 +0100, a écrit :
> These two are preliminary patches generated from dereference of null
> pointer reports. Let's see how this goes. If I got it right, I'll
> continue to work on this. If I got it wrong, I'll let someone more
> experienced deal with these kind of warnings.

Well, it's not so much a question of experience.  With this kind of
issue, you have to look on a broader scale than just the code at stake.
Sometimes just an additional test makes sense, but quite often the issue
actually reveals a more profound matter.  Again, here experience can
of course tell what it is, but not having experience is not a problem:
simply read the source code.  For your second patch, for instance,
look for ifps and fp_thread assignations (use coccigrep for that),
there aren't many.  Just to be more convincing: I had no clue what was
supposed to happen to ifps, I've never dealt with that part of the code.
So I've just grepped etc. as I mentioned above.  Just do it, that's how
things work.

Samuel