Re: [PATCH v3] Y2038: make __mktime_internal compatibles with __time64_t

[Albert ARIBAUD <albert dot aribaud at 3adev dot fr>]

Hi Florian,

On Wed, 24 Oct 2018 13:25:59 +0200, Florian Weimer <fweimer@redhat.com>
wrote :

> * Albert ARIBAUD:
> 
> > Hi Paul,
> >
> > On Wed, 17 Oct 2018 18:44:05 -0700, Paul Eggert <eggert@cs.ucla.edu>
> > wrote :
> >  
> >> This patch looks OK to me, modulo the spelling fix.
> >> 
> >> In reviewing this patch I noticed a longstanding bug in mktime that 
> >> hasn't been reported. I just now reported it here:
> >> 
> >> https://sourceware.org/bugzilla/show_bug.cgi?id=23789
> >> 
> >> Fixing that bug may affect the patch.  
> >
> > Should I go and provide a fix to that bug, then, and resubmit my patch
> > over the fix?  
> 
> I do not have a strong preference here, as long as the bug fix goes into
> a separate patch.
> 
> (Despite what I just told to H.J., I think in general we should not
> force contributors to fix peripherally related bugs as part of the work
> they set out to do.)

I agree with this approach in the general case, however I would like to
point out that in this instance I had not read or felt Paul's comment
as forcing me to handle the bug in any way; on the contrary, I have
seized it as an opportunity to get acquainted with the bugzilla part of
the patch committing process as well as with adding tests to glibc,
since Y2038 will quite probably require adding such tests.

> Thanks,
> Florian

Cordialement,
Albert ARIBAUD
3ADEV