Re: mdocml version 1.12.0 available

[Joerg Sonnenberger <joerg@britannica.bec.de>]

On Sat, Oct 15, 2011 at 05:02:49PM +0200, Ingo Schwarze wrote:
> Hi Ulrich,
> 
> Ulrich Spörlein wrote on Thu, Oct 13, 2011 at 11:55:25AM +0200:
> > On Thu, 2011-10-13 at 02:50:45 +0200, Ingo Schwarze wrote:
> >> Joerg Sonnenberger wrote on Tue, Oct 11, 2011 at 04:56:41PM +0200:
> 
> >>> /home/joerg/work/NetBSD/cvs/src/external/bsd/mdocml/bin/mandoc/../../dist/mdoc_term.c:617:7: warning: Access to field 'prev' results in a dereference of a null pointer (loaded from variable 'n')
> >>>                 if (n->prev && MDOC_It == n->prev->tok) {
> >>>                     ^
> >>> /home/joerg/work/NetBSD/cvs/src/external/bsd/mdocml/bin/mandoc/../../dist/mdoc_term.c:611:7: warning: Access to field 'prev' results in a dereference of a null pointer (loaded from variable 'n')
> >>>                 if (n->prev && MDOC_It == n->prev->tok)
> >>>                     ^
> 
> >> I have no idea why clang thinks n might be NULL here.
> >> Do you understand that?
> >> Or is this just a false positive?
> 
> > print_bvspace could have been called with n being NULL.
> 
> No, i just checked all callers.
> All callers access n before calling print_bvspace,
> so it's a false positive.

That's not clear from the code flow and given that there are explicit
checks for NULL around, it punts. You could make this assumptions clear
with an assert.

Joerg
--
 To unsubscribe send an email to tech+unsubscribe@mdocml.bsd.lv