Front page | perl.perl5.porters |
Postings from October 2009
Re: future of maint
Thread Previous
|
Thread Next
From:
Nicholas Clark
Date:
October 11, 2009 13:02
Subject:
Re: future of maint
Message ID:
20091011200238.GW60303@plum.flirble.org
On Sun, Oct 11, 2009 at 08:50:12PM +0100, Dave Mitchell wrote:
> On Sun, Oct 11, 2009 at 08:09:55PM +0200, Rafael Garcia-Suarez wrote:
> > There are other kinds of portability fixes, which usually have less
> > impact on the code: that is fixes pertaining to new versions of
> > compilers. (notably gcc). I think that those should go in anyway.
>
> The trouble with those sorts of fixes is they carry the risk of breaking
> older platforms (the sort of platform nobody tests until after we
> release). I'm thinking especially of the sort of fix that adds some new
> compiler or linker option.
But don't we also run that risk with a .0 release?
If they're not tested, and we're not in a position to lean on the people who
have those platforms to test more when we need it, then we're also likely to
make changes in blead that inadvertently break them, which won't be detected
before we ship a .0, and if so, we'll never get a chance to fix them.
(Also, I was never too worried about breaking minor platforms that weren't
tested. The *real* point of a release candidate was to engage the S.E.P.
field. If it was broken in the release candidate, and you reported it, I would
fix it. If you didn't report it, well, that's *your* fault. Not mine.)
I'm wondering how long *after* 5.12.0 ships before someone complains that it
doesn't have suidperl, because that was the first time that they tried it.
Nicholas Clark
Thread Previous
|
Thread Next