-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Duncan wrote: > >> As usual you seem to have the gist of it. I actually am an AT, so if it >> works for me I'll see what the state seems to be and keyword STABLE in >> bugzilla if it makes sense (noting the caveats you listed). Will be a >> day or two though so if anybody else beats me to it I'm sure the powers >> that be will be happy to see one less package falling behind x86... :)
Ok, I ended up dumping my --resume list to a file so that I can work on the side. timidity++ works fine with the new version and I'm stable for the most part across the board. However, I'm not sure it is necessarily wise to stablize this ebuild - for whatever reason it seems to be in a state of flux - only a few days ago it was modified. Stable ebuilds should not be modified in general - they should be bumping revisions on them. Otherwise what is the point in testing it? The maintainer would probably need to branch off a new dev version of the ebuild (considering ppc is already using it they should probably do this anyway). > > In terms of falling behind x86, how are we actually doing compared to them > since they actually got an arch team and arch testers, and made it policy > that maintainers didn't keyword for x86 unless they got permission from x86 > to do their own packages? With more QA now, I'm guessing things will be a > bit slower for them, and we shouldn't do to bad against them now. Of > course, quality will be higher on x86 than it was before as well. Anyway, > I'm wondering how the stats are going since they actually became a proper > arch, arch-team and all, as I've not seen any numbers or comments on the > topic since then. > Right now it actually isn't too bad, but things were falling way behind for a while. Actually, it seemed to me that a number of these packages were libraries and perl packages. I think that this was in part because they can't be directly tested - you need to find something else that makes heavy use of the library and see if it breaks. Even then you're not really testing it fully. So, I think the tendency is for people to not want to stick their necks out and potentially break something else. The current numbers are that only 129 ebuilds are behind x86 in stable. It used to be in the 1000-2000 range, and clearly a lot of devs have pitched in to reduce this. I'd be curious how x86 compares to amd64 - they could easily be this far behind us. Unfortunately, my copy of imlate stopped working a number of portage upgrades ago - need to hunt a fresher one down (just a script that compares two archs and lists all packages which are stable on one and not the other). -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFE/AmcG4/rWKZmVWkRAmuVAKCBvmNVGDpHb3340U2zM+rOHdLVqwCgqbcy g/SNdhtLC8mmtn5scnbxaoU= =x5J7 -----END PGP SIGNATURE-----
smime.p7s
Description: S/MIME Cryptographic Signature
