This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: [gfortran] SIGH! How to alienate a contributor
On Sat, Nov 20, 2004 at 03:17:24AM +0100, Steven Bosscher wrote:
> On Saturday 20 November 2004 02:53, Steve Kargl wrote:
> > On Sat, Nov 20, 2004 at 02:37:54AM +0100, Steven Bosscher wrote:
> > > On Saturday 20 November 2004 02:24, Steve Kargl wrote:
> > > > The first small patch was available on Sat, 13 Nov 2004.
> > >
> > > Oh no!! A *whole* week! Say it ain't so!!
> >
> > I only have 2 weeks of annual leave that I can dedicate to
> > gfortran. One week is gone.
>
> I have more time over the whole year, but less in the normal
> week. In fact, other than in the weekends I don't do any GCC
> hacking at all (try not to, anyway). We just don't have any
> person working full time on gfortran.
>
> One problem is that it's hard to review patches that lack a
> proper explanation. For example, just now on IRC Paul pointed
> out to me that the argument to the UMASK function really should
> be forced to be a kind 4 integer (32 bits) to match mode_t from
> "man umask(2)" - so what I've just commited to CVS is in fact
> incorrect.
No. It's not incorrect. I cast the value in umask.c to
mode_t prior to calling umask(2). There is no guarantee
that mode_t is 32 bit.
> > Writer-after-approval isn't very valuable if no one is reviewing
> > patches. It's only after a scathing email that anyone responds
> > (with apologies to Janne).
>
> You cannot expect everyone to jump in line when you happen to
> find some time to do some hacking. We're all still volunteers.
> And as I told you off-list, I never received your reply to the
> comments Janne had on your patch.
Janne's comment was inconsequential and it was fixed. Whether
mask (or exit or unlink) was attributed to the F95 standard or
the GNU Fortran language has zero impact on whether the code works.
--
Steve