On Mon, 2011-11-28 at 18:41 +0200, Alan McKinnon wrote:
> My experience is different to both of yours. I too have been using
> Gentoo for many years and had good results with unstable. Hardly ever,
> if even at all, have I run into packages that would not compile at
> Build failures for me have always been some unusual configs on my end,
> usually strange USE flags. But I don't use any of the more exotic
> packages like those in sci- and games- so YMMV I guess.
I'm not saying that unstable is somehow bad, I'm just saying it's
sometimes... unstable.
I dont' have any "exotic" packages or configs either, but I do from time
to time encounter such problems as
1. Patches not included
2. Patches not applying
3. build failures because a patch in a previous revision is no
longer applicable in the new revision
4. build failures caused by upstream issues
5. build failures due bad ebuilds
6. incomplete DEPENDS or RDEPENDS(this actually happens quite more
frequently than i'd like)
7. Broken functionality (upstream bugs)
8. A dependency of a package was bumped, and that package doesn't
build against the bump.
Granted, when I test, I test hard. I depclean with build time
dependencies removed, to make sure packages have the correct DEPENDS. I
do an "emerge -e world" about once per month. I have a build system
that builds virtual appliances from scratch that help me find bugs
(granted, most of these VMs are in the stable tree so they actually find
bugs in stable and the stage3 tarballs). I set USE flags manually
instead of using the defaults. So, while that may be considered an
"unusual config" it should work and it helps me find bugs before they
get into stable.
But my feeling is, if you use the testing branch and you *don't* find
bugs, then you aren't testing hard enough :P
-a
