[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Nmh-workers] too many uninitialized variables in nmh-1.1 RC4
From: |
Josh Bressers |
Subject: |
Re: [Nmh-workers] too many uninitialized variables in nmh-1.1 RC4 |
Date: |
Fri, 25 Nov 2005 22:20:06 -0500 |
>
> > I often have a manfiest constant WHO_CARES,
> >
> > #define WHO_CARES (void*)012345
>
> #define WHO_CARES 0xdeadbeef
>
> int foo = WHO_CARES;
>
> Has the advantage that if you *do* make it through the code without actually
> setting a known good value, you'll probably get a SIGSEGV or other blatantly
> obvious indication that things have gone very very awry. '12345' doesn't
> always
> do that for you....
This thread has reached a point of silliness. The warning in question is
just a warning, it doesn't mean there is a problem, it means gcc thinks
there may be something that the developer should look at. Chances are good
that the logic does have a flaw. gcc will analyze the source and warn when
it thinks it's possible for a variable to remain unassigned before it's
used. Setting a variable to an arbitrary value to get a warning to go away
is a recipe for disaster. If this really bothers you, turn off -Wall. If
you want to do clever things for debugging purposes, you should investigate
the malloc perturbing code in newer glibc.
--
JB