Re: [RFC] ready for cygport to default to gcc4?

["Charles Wilson" <cygwin at cwilson dot fastmail dot fm>]

Yaakov (Cygwin/X) wrote:
>Charles Wilson wrote:
>> gcc4 is still an *experimental* release:
>> 
>> We shouldn't default to using it until the gcc maintainer is confident
>> enough in it to promote it officially. I know it's a bit of a
>> chicken/egg problem, but there you go.
>
> 1) A few maintainers have already started using gcc4 for their packages.
>  While I thought that doing so with 4.3.2-1 on 1.5 was a bit premature,
> not so now.

Well, we disagree. Big surprise, I know. I believe it is premature to
use any test or experimental release when building an official package
[*], until the maintainer of that package removes the experimental
designation (or, when using setup's test:/curr: facilities, promotes it
to curr:).

[*] short of some overriding reason, such as: (a) two dependent packages
with different maintainers must coordinate their release -- so SOMEBODY
has to go to curr: first, or (b) a new package just plain won't work
unless the test:/experimental dependency is used. Like octave vs. gcc4,
IIRC.

> 2) How should Dave be able to stabilize gcc4 if nobody is testing it?
> Same goes for cygwin-1.7 for that matter, and I think I've been doing my
> part testing both.

Yes, you have. This is what I meant when I said "chicken/egg problem". 
I believe you have been doing The Right Thing: (semi-)private builds of
your packages using the experimental gcc4 -- not cygwin-mirror-system
public releases of your packages using that compiler.  At this point, if
you're happy with the way your packages are behaving with respect to the
new compiler, it's perfectly legitimate for you to lobby Dave to remove
the experimental designation from gcc4.

But, IMO, it is not legitimate to try to de-facto override Dave's
decision as the gcc maintainer, and MAKE gcc4 the default compiler via
your privileged position as the maintainer of cygport, if gcc's own
maintainer still believes gcc4 is experimental.

> Bottom line, we need to look at what the distro should like when
> cygwin-1.7 goes stable, and set a direction.

Yes, and we're (slowly) getting there.  BTW, I do not believe the
following thread
"[RFC] ABI bump for building with gcc4 ?"
http://cygwin.com/ml/cygwin-apps/2009-03/msg00033.html
ever reached a resolution.  Only you, Dave, and I participated...any
other maintainers have an opinion?

FWIW, I'm starting to lean in your direction (towards a flag day
release) on that regard, but I'd really like to hear from Corinna and/or
cgf on that issue. Also, a flag day release requires the participation
of ALL maintainers, so it's not a decision that just a few of us can
make on our own.

--
Chuck