works better than portage, except for an ebuild bug introduced by #488776 is there anything you would expect to be tested in particular before stabilizing?
The main reason it's never gone stable is because of the "30 day" rule... We've sometimes pushed out important releases where we know *exactly* what's changed and what the impact is, and we'd rather not have to wait 30 days to get this to users because of a rule that was designed for typical upstreams rather than us...
I don't know of any such rule and I don't follow it. It must be some kind of "good practice".
I'll call for maintainer timeout in 2 weeks.
I'm closing this RESO WONTFIX. I feel that stabilising an alternate PM in the Gentoo tree is endorsing that PM which we should not do with any aside from Portage. If Paludis is proposed (and accepted) as replacing Portage as the default or someone gives me an extremely convincing argument as to why i'm wrong, then i will happily revisit my decision, but until then...Not so much.
Well rpm is stable...
(In reply to Ciaran McCreesh from comment #5) > Well rpm is stable... OK, a package manager that virtual/package-manager resolves to. Does that help, Mr. Pedantic?
(In reply to Jeff (JD) Horelick from comment #4) > I'm closing this RESO WONTFIX. > > I feel that stabilising an alternate PM in the Gentoo tree is endorsing that > PM which we should not do with any aside from Portage. > > If Paludis is proposed (and accepted) as replacing Portage as the default or > someone gives me an extremely convincing argument as to why i'm wrong, then > i will happily revisit my decision, but until then...Not so much. There has never been any special policy about stabilizing packages except that the package has to be stable. The same argument could be applied to systemd. Just because openrc is the so called "default" doesn't mean systemd must not be stable. That does not make sense and just makes it unnecessarily harder for stable arch users.
so, care to elaborate why keeping a working piece of software without major bugs indefinitely in ~arch improves user experience?
TBH, I have seen the light and this weekend, switched all my machines to paludis (I'm getting to the age where good core design and correctness seems to trump simplicity and bad code design). That isn't strictly relevant, but i've decided that we will go stable with this since most of us on this ticket are now in agreement. However, I went through standard archtesting procedure for paludis and I found a bug that will block stabilisation. The tests do not pass on Gentoo. Granted, we could simply RESTRICT="test" but that's a cheap, shitty solution. The 2 failing tests are: FAIL: process_TEST FAIL: pty_TEST Setting FEATURES="-userpriv -usersandbox" (probably userpriv is the culprit here) caused the tests to succeed. However, userpriv and usersandbox are in the list of FEATURES to be enabled for archtesting. Given that the tests succeed as root, perhaps it is worth doing RESTRICT="test".
Correction to last comment. Tests pass just fine on a new VPS, but not on my VPS built off an image from December 2009 with almost 5 years of cruft from regular use. I'm calling the test failure NOTABUG unless archtesters can also reproduce. I will file a STABLEREQ bug for sys-apps/paludis-2.0.0 ASAP
This was obsoleted by bug #520874 which is the paludis-2.0.0 STABLEREQ
