Subject: Re: Warnings and builds...
To: None <email@example.com>
From: Mike Long <firstname.lastname@example.org>
Date: 05/06/1997 10:43:55
>Date: Mon, 5 May 1997 21:38:29 -0500 (CDT)
>From: Peter Seebach <email@example.com>
>Would anyone object to the patches it would take to make -current
>build cleanly with -Wformat? I was having the inevitable "proc
>size mismatch" (it appears my Sun3 kernel has fallen out of date
>again), and I noticed that the code uses %d to print sizeof()...
No objection to the concept, although the actual patches would have to
be judged on their own merit.
>Of course, there's no *good* answer in that case (You can't tell
>whether sizeof()'s return should be %u or %lu), but I am fairly
>confident that %d is "wrong". (Arguably, we don't care unless
>struct proc is expected to outgrow INT_MAX.)
Whenever an integer of unknown size is fed to *printf(), you should
cast it first. In this case, you should use:
printf("%lu", (unsigned long)sizeof(FOO));
>So... Is this pointless nit-picking? A good idea?
I think it's a good idea.
>Feedback appreciated, especially from people on the platforms (alpha
>comes to mind) where %d might actually be *wrong* for size_t, if
>size_t were unsigned long...
%d is wrong on the i386, which uses `unsigned int' for size_t.
Mike Long <firstname.lastname@example.org> http://www.shore.net/~mikel
"Every normal man must be tempted at times to spit on his hands,
hoist the black flag, and begin slitting throats." -- H.L. Mencken