[gentoo-dev] [RFC] Maintainer Timeout
Hi, I want to propose a Maintainer Timeout such as FreeBSD. If a maintainer or herd does not fix (or assign/comment) a bug in a reasonable amount of time (2 weeks? 3 month?) any developer can fix it (or a pre established group of developers such as QA) Thanks for your time :) -- Timothy `Drizzt` Redaelli - http://dev.gentoo.org/~drizzt/ FreeSBIE Developer, Gentoo Developer, GUFI Staff There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence. -- Jeremy S. Anderson signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] [RFC] Maintainer Timeout
I want to propose a Maintainer Timeout such as FreeBSD. If a maintainer or herd does not fix (or assign/comment) a bug in a reasonable amount of time (2 weeks? 3 month?) any developer can fix it (or a pre established group of developers such as QA) There's a little difference between does not fix and assign/comment, I'd appreciate if you could be a bit more verbose on that ;) wkr, Tobias signature.asc Description: Dies ist ein digital signierter Nachrichtenteil
Re: [gentoo-dev] [RFC] Maintainer Timeout
Tobias Scherbaum wrote: I want to propose a Maintainer Timeout such as FreeBSD. If a maintainer or herd does not fix (or assign/comment) a bug in a reasonable amount of time (2 weeks? 3 month?) any developer can fix it (or a pre established group of developers such as QA) There's a little difference between does not fix and assign/comment, I'd appreciate if you could be a bit more verbose on that ;) In theory if a maintainer assign a commit it will works on it, maybe we should use 2 different timeouts. one for not assigned bugs and one (longer) for assigned. -- Timothy `Drizzt` Redaelli - http://dev.gentoo.org/~drizzt/ FreeSBIE Developer, Gentoo Developer, GUFI Staff There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence. -- Jeremy S. Anderson signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] [RFC] Maintainer Timeout
Timothy Redaelli wrote: [Fri Feb 02 2007, 04:17:32AM CST] I want to propose a Maintainer Timeout such as FreeBSD. If a maintainer or herd does not fix (or assign/comment) a bug in a reasonable amount of time (2 weeks? 3 month?) any developer can fix it (or a pre established group of developers such as QA) It seems to me that this is an area where common sense and trust really should win out. If you've made a reasonable effort to contact the maintainer, to no avail, and you can competently fix the bug, then go fix it and e-mail the maintainer what you've done. What's a reasonable time? Well, that's where the common sense comes in. If somebody just broke libpng and all of its dependencies, then you don't really want to spend much time waiting. (Of course, in such an extreme case it might be best to drop the problem on QA.) On the other hand, if the bug is a version bump, then the waiting time had best be a good long time, since there may be a very good reason why the new version isn't in the tree and the maintainer just left on his or her honeymoon. (Although in that case it wouldn't hurt for the dev to package.mask the offending package without putting it in the tree, since that way there would be a record that the package was deliberately left out, and there'd be a warning if a hapless user tried to add it via an overlay.) I have mixed feelings on the notion of ownership of ebuilds. When Gentoo had only a handful of devs, the tree was almost entirely collectively owned by all devs, with baselayout and portage being the only packages that were labelled don't touch unless you really know what you're doing. Pretty much everybody who had time fixed bugs in ebuilds in the tree, whether the ebuilds belonged to them or not. Today we have many more packages that require specialist knowledge to maintain (either because the individual packages are extremely complicated, or because they are part of an integrated system of packages), and the devs that maintain those packages are understandably touchy about having other devs break those ebuilds. For the majority of packages in the tree, though, I'd like to encourage their maintainers to be less possessive. If somebody wants to fix bugs in ebuilds that I maintain, go right ahead. Just don't break it in the process, please. -g2boojum- -- Grant Goodyear Gentoo Developer [EMAIL PROTECTED] http://www.gentoo.org/~g2boojum GPG Fingerprint: D706 9802 1663 DEF5 81B0 9573 A6DC 7152 E0F6 5B76 pgpYNlMKFSf0i.pgp Description: PGP signature
Re: [gentoo-dev] [RFC] Maintainer Timeout
On Fri, 2007-02-02 at 10:19 -0600, Grant Goodyear wrote: I have mixed feelings on the notion of ownership of ebuilds. When Gentoo had only a handful of devs, the tree was almost entirely collectively owned by all devs, with baselayout and portage being the only packages that were labelled don't touch unless you really know what you're doing. Pretty much everybody who had time fixed bugs in ebuilds in the tree, whether the ebuilds belonged to them or not. Today we have many more packages that require specialist knowledge to maintain (either because the individual packages are extremely complicated, or because they are part of an integrated system of packages), and the devs that maintain those packages are understandably touchy about having other devs break those ebuilds. For the majority of packages in the tree, though, I'd like to encourage their maintainers to be less possessive. If somebody wants to fix bugs in ebuilds that I maintain, go right ahead. Just don't break it in the process, please. Thanks Grant: this is--hands down--the best, clearest and most apropos response in this entire thread. I can only echo your feelings on it (and I would hope that those who've interacted with me know that). -- Seemant Kulleen Developer, Gentoo Linux -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [RFC] Maintainer Timeout
On Fri, 2 Feb 2007 10:19:21 -0600 Grant Goodyear [EMAIL PROTECTED] wrote: [lots of good stuff] I was going to respond to Timothy's proposal in much the same way - but Grant has said everything much better than I would have done! +lots Grant :) -- Kevin F. Quinn signature.asc Description: PGP signature
Re: [gentoo-dev] [RFC] Maintainer Timeout
Timothy Redaelli wrote: Hi, I want to propose a Maintainer Timeout such as FreeBSD. If a maintainer or herd does not fix (or assign/comment) a bug in a reasonable amount of time (2 weeks? 3 month?) any developer can fix it (or a pre established group of developers such as QA) Thanks for your time :) I think if everyone abides by two rules (which I thought were pretty evident, but maybe not). Don't be an asshole Don't screw up* Don't be an asshole means you should make a good faith effort to contact a maintainer or herd before touching stuff. I remember a time when portage was generating duff sha256 checksums and marienz and I were working on fixing them; this was right around the time that Xorg 7 (or 7.1) was coming out and we had to touch all the modular builds. Of course little did we know by touching the ebuilds we messed up all of the checksums for the source tarballs (because we had older tarballs locally; such that the local copy was older than the one on the mirrors). So even simple things like fixing a duff digest can be harmful if you are not careful. How long you wait depends on the severity of the problem. Don't just ping people on irc, send mail. If they get pissed, you should have some CYA material (thats cover your ass, for those who don't know). If they come complaining about it you have the MAIL-ID of the e-mail you sent them. Don't Screw Up means you shouldn't try to touch crap you have no friggin idea how to fix. ASK QUESTIONS FOR THE LOVE OF GOD. Most of my learning here has just been from incessantly annoying the hell out of the developers with more experience than me (especially marienz, vapier, and flameeyes). If you think your idea *might* work, test it out. Get others in -dev to test it out. Get someone in the herd to look at it, get ANYONE competent to look at it. And the Corollary: be honest when someone asks you a question to which you do not know the answer. The worst case is asking someone a question to validate your idea and them agreeing with you just for the hell of it, as opposed to actually examining the issue at hand. This is one of those cases where the more you know about the tree and it's workings the better equipped you are to deal with issues like this. We have a lot of developers who do not know as much as what I'll term 'the old skool devs' did (myself included). That makes keeping the 'everyone owns all the packages' idea around more difficult; in the end we have a team of developers that are not as well equipped to deal with issues. * Corollary: if you do screw up, take responsibility and fix it/find someone to help you. DON'T MAKE IT WORSE. -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [RFC] Maintainer Timeout
Alec Warner wrote: Don't be an asshole Don't screw up* * Corollary: if you do screw up, take responsibility and fix it/find someone to help you. DON'T MAKE IT WORSE. Indeed, agreed. An excellent response. signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] [RFC] Maintainer Timeout
Grant Goodyear wrote: [something I'd say in a less effective way] +1 -- Luca Barbato Gentoo/linux Gentoo/PPC http://dev.gentoo.org/~lu_zero -- gentoo-dev@gentoo.org mailing list
