Public SourceMage Discussion List

[Daniel Goller <dgoller AT satx.rr.com>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sun, 10 Dec 2006 08:42:16 +0100
Pieter Lenaerts <e-type AT sourcemage.org> wrote:

> Op zaterdag 09-12-2006 om 15:34 uur [tijdzone -0600], schreef Daniel
> Goller:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> > 
> > On Sat, 09 Dec 2006 20:02:52 +0100
> > Flavien Bridault <vlaaad AT sourcemage.org> wrote:
> > 
> > > Le samedi 09 décembre 2006 à 19:56 +0100, Pieter Lenaerts a écrit :
> > > > Op zaterdag 09-12-2006 om 08:25 uur [tijdzone -0600], schreef Daniel
> > > > Goller:
> > > > > Well, considering how often we force gcc3.4 we should not be too 
> > > > > proud of being able to use gcc4.1.
> > > > > For what it's worth, what hasn't been forced does play fine with 
> > > > > gcc4.1, and has done so for months.
> > > > > Have nothing to say to glibc2.5, since i haven't touched it.
> > > > > 
> > > > > We could either remove all the forcing of gcc3.4 first, and then 
> > > > > determine which needs 4.1 patches, or test things against 4.0 and 
> > > > > 4.1 at the time we do remove the forcing.
> > > > > There is nothing wrong with 4.1, some code does trigger a few extra 
> > > > > sloppyness indicators that even 4.0 still let's fly, but it's not a 
> > > > > lot.
> > > > > 
> > > > > To make things smoother in the future the forcing of gcc version 
> > > > > below what is in test must be discouraged, things do not get fixed 
> > > > > that way.
> > > > > This way a rebuild working actually has meaning.
> > > > > 
> > > > > We should move forward with gcc/glibc, and encourage to have the 
> > > > > gcc3.4 forcing removed next time people touch a spell that does it. 
> > > > > (After testing it of course.)
> > > 
> > > This anyway what spell maintainers are supposed to do since the
> > > beginning, right ?? 
> > > 
> > 
> > I guess not, if i see how Pieter replied. :/
> 
> then we're just missing each other's point.
> 
> our maintainers should indeed check when they update a spell, but you
> talk of removing all gcc3.4 forcing. it is not our job to clean up
> upstream code.
> 
> it is very normal to have this ${GCC_VERSION} forcing where needed iyam,
> and it is practically impossible to remove it.
> 
> of course you're right that maintainers should check on updates if the
> forcing is still needed.
> 
> my reaction was mainly to your stating that we should remove all gcc3.4
> forcing before moving over to gcc4.0. sorry if I wasn't clear.
> 
> 

I understood you quite well then, what's wrong with patching code so one does 
not need to carry around multiple gcc versions, unless one wishes to have 
more than one for some reason?
"Not our job" != "Let's not even apply already available patches" (as i had 
mentioned)
that is more what i was after, if you don't feel like fixing sloppy code, ok, 
but googling for patches is something we could all do more. Right? ;)

Daniel
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)

iD8DBQFFe8u++HaycIPdpbkRAi0yAKDGQHGORmlYYXF+YBsrInwlSC3pwACggxbB
a30ADWkStnd9Xl+0Z5rwRaM=
=v7kz
-----END PGP SIGNATURE-----