sm-discuss AT lists.ibiblio.org
Subject: Public SourceMage Discussion List
List archive
- From: seth AT swoolley.homeip.net
- To: sm-discuss AT lists.ibiblio.org
- Subject: [SM-Discuss] prometheus triaging done
- Date: Sat, 22 Apr 2006 23:48:14 -0700
For those who have been waiting for me to finish going through the
grimoire -- I'm done!
Prometheus found 357 bugs that were not invalid and not dupes
96 installed.lst
2609 winner.lst
1784 skipped.lst (501 loser.lst)
4489 total
About 70% of the bugs filed by prometheus were valid, including the
pretty bad invalid start where I had a corrupted packages file. I
didn't get any of the kde section done because qt-x11 got messed up by a
misconfiguration, but I think I covered most of the grimoire. A second
run through should yield some interesting results.
A lot of the bugs were as a result of x86_64 issues. A lot were md5
issues. A lot were it trying to pull in z-rejected for me. A lot were
missing dependencies.
Actual broken compiles were rare. This means it's been a pretty good
showing for the grimoire team, GOOD JOB GUYS!
Now if we just get to work fixing these bugs I'll hold off on re-running
for a while until you can get them fixed and into stable-rc.
Be sure not to mark these fixed until it's fixed in stable, by the way!
Not in the comments that you fixed it in devel/test, request an
integration if you want. If we do it this way we can track what's
actually fixed in stable and I can spot regressions easier.
Thanks for all the hard work improving sourcemage!
In other news, a couple of sorcery bugs were hit, but they were mostly
already known and targetted for later releases. The major ones were:
cvs spells can sometimes not succeed even if cvs is a dependency due to
when the parallel summon can take place (already known).
summon has locking issues with corrupted spells (new).
Dependency resolution breaks in odd ways with confounding errors upon
catastrophic problems with spells in resolving DEPENDS (already known).
Good job sorcery team for making sorcery so reliable for this mass test!
As said before, there are over a dozen ways to improve prometheus to
make it so I don't have to spam everybody with edits to prometheus
reports every time it runs and to avoid duplicates, but for now, I think
I should go to sleep. :)
Seth
-
[SM-Discuss] prometheus triaging done,
seth, 04/23/2006
- Re: [SM-Discuss] prometheus triaging done, Andrew, 04/23/2006
Archive powered by MHonArc 2.6.24.