Public SourceMage Discussion List

[Seth Alan Woolley <seth AT positivism.org>]

On Wed, Feb 16, 2005 at 06:38:38PM +0100, Benoit PAPILLAULT wrote:
> Seth Alan Woolley a écrit :
> >There's a bug for it, and most of the problems with gcc-3.4 are 
> >resolved.  It's my anticipation reading the bug that it will be in test 
> >soon, and in stable after that in time for an official ISO release.
> 
> OK :-).
> 
> >Wait, there's a difference between what you do in test isos and what you 
> >do in stable isos.  For a stable iso, there shouldn't be a custom 
> >grimoire at all.  For a test iso, it should use stable griomire with a 
> >fixes overlay.  When all the fixes are in stable, you drop the overlay 
> >grimoire, freeze the stable grimoire except for ISO-critical bugfixes, 
> >and then release.
> 
> I've always used stable grimoire so far (plus manual tweaking for some 
> spells). I like your method where I could work in parallel with the 
> stable grimoire being updated.
> 
> >>I'm going to ask that for glibc. I'm pretty surprised that apparently 
> >>nobody will take care of the bugs I reported on glibc devel version 
> >>since they don't seem mandatory to solve the initial issue (installed 
> >>system not able to rebuild).
> >
> >
> >If they are mandatory, can you explain why they are mandatory rather 
> >than declaring that they (your specific fixes rather than proposed 
> >alternatives) are mandatory?
> 
> No. They are not mandatory for the specific problem being solved. My 
> general feeling is that I'm loosing my time reporting bugs with bugfix 
> since those bugfix are not used.

True, there's a tradeoff, which is often why I discuss the problem in 
irc with somebody of Higher Authority before I start dumping time into 
it.  I know how you feel after submitting a patch and somebody else 
coming along and showing a "different" way.  As long as they are prompt 
about it, it's the same to me (but that's probably your issue -- the 
promptness).

> 
> >>I'm pretty worried about this point because, as part of my ISO job, we 
> >>have discussed about using INSTALL_ROOT feature to solve some problem of 
> >>the current ISO generation script and i'm currently investigating it. 
> >>So, if the glibc spell in devel do not implement this feature, either I 
> >>need to have my own grimoire (with the possibility of re-creating the 
> >>initial problem), either I leave.
> >
> >
> >There are two bugs about this:
> >
> >http://bugs.sourcemage.org/show_bug.cgi?id=7209
> >http://bugs.sourcemage.org/show_bug.cgi?id=8242
> >
> >Do you have a patch that fixes the 8242 issue?
> 
> Yes. It's in bug #8242 itself :-)

Ah, right.  I should have read closer; I was looking for a unified diff.

> 
> >Are you using INSTALL_ROOT right now?  Is this holding back your push to 
> >a release?
> 
> I'm investigating it in fact as this is the method that was agreed on 
> for future ISO. Currently, all the spell I've tested works OK or 
> requires minor fix to work with INSTALL_ROOT. So, this is quite a 
> promising work ... it would be a pity if all the work done on glibc 
> would produce a spell not respecting INSTALL_ROOT feature.

That's great.  I think He who shall not be named had fixed most of the 
install_root issues with basesystem spells long ago, and any that don't 
work are a result of bit rot.

> 
> >Reading that bug indicates that you proposed fixes and the sorcery and 
> >grimoire team leaders worked on an alternative method that appears to 
> >have worked.  If there's some reason that's not the case, can you note 
> >it in the bug?  The last three comments are from people other than you, 
> >so it seems there is some feedback happening.
> 
> Yes, there are feedbacks, which is good. But more than 3 months later, 
> nothing have been pushed to stable and what is going to be pushed to 
> stable is missing some fixes (not mandatory, but could help the ISO team 
> A LOT!). Since people are actively testing what's currently in devel, I 
> think we can discuss and have those fixes integrated before everything 
> is moved to stable :-)
> 
> By the way, since I've access to devel, I could integrate those fixes 
> myself. Would that be OK? (I think I already ask for that? :-)).

I'm not the proper person to authorize that (Arwed is, I believe).

Maybe he'll pipe in a bit later?  He might be at work.

-- 
Seth Alan Woolley [seth at positivism.org], SPAM/UCE is unauthorized
Key id EF10E21A = 36AD 8A92 8499 8439 E6A8  3724 D437 AF5D EF10 E21A
http://smgl.positivism.org:11371/pks/lookup?op=get&search=0xEF10E21A
Security Team Leader Source Mage GNU/Linux http://www.sourcemage.org

Attachment: pgp2ExpdMR2df.pgp
Description: PGP signature