Re: [gentoo-dev] Re: Re: [gentoo-core] crap use flags in the profiles
On Wed, 2005-08-31 at 09:18 -0400, Stephen P. Becker wrote: Notice that for almost everything, amd64 is barely behind x86...just a minor version number/revision or two at most. That's the ATs hard at work keeping us current ;) -- Homer Parker Gentoo/AMD64 Arch Tester Strategic Lead [EMAIL PROTECTED] -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] Re: Re: [gentoo-core] crap use flags in the profiles
On Wed, 31 Aug 2005 05:36:52 -0700 Duncan [EMAIL PROTECTED] wrote: | No offense intended, but as a user, I /like/ to actually know that a | package keyworded for my arch (segment) is known to work on it in full | (IMHO) uncrippled amd64 form, not in some (IMHO) crippled 32-bit | special case. If we went the other way and removed x86 keywording | from everything that failed in 64-bit mode, including all 32-bit only | codecs and the like, x86(32) arch(segment) folks would rightly be | wailing in protest. | | Again, no offense intended, but unless you have some magic way to fix | that situation, perhaps the MIPS devs and users are willing to live | with that problem on MIPS, but neither x86(32) users nor amd64 users | (and by this I'm including devs, which are obviously users as well) | are interested in being saddled with an unnecessary problem, when the | current situation avoids it, or I expect the amd64 keyword would have | never been added. It's not magic. We've been handling packages that work on sparc64 but not sparc32 for years with a single keyword. Just because you (and, from the looks of things, most of the x86 and amd64 developers) don't know about some of portage's features doesn't mean they don't exist :) -- Ciaran McCreesh : Gentoo Developer (Vim, Shell tools, Fluxbox, Cron) Mail: ciaranm at gentoo.org Web : http://dev.gentoo.org/~ciaranm pgp0fUPxcW4oR.pgp Description: PGP signature
Re: [gentoo-dev] Re: Re: [gentoo-core] crap use flags in the profiles
Stephen P. Becker wrote: [Wed Aug 31 2005, 08:18:53AM CDT] We don't live with that problem on MIPS because it doesn't exist. If something doesn't work in one spot, we dont' stable keyword it...simple as that. Also keep in mind that for some stuff, we don't have to test on both. For example, we have no supported little endian machines that are capable of running X, therefore, we don't care about testing X there. See how it works? So, the basic suggestion is that x86 and amd64 would both use the same keyword, but that for cases such as valgrind pre-3.0, which didn't work at all on amd64, then those package are profile-masked, and there's separate amd64 and x86 profiles (as there are now) to handle those distinctions? -g2boojum- -- Grant Goodyear Gentoo Developer [EMAIL PROTECTED] http://www.gentoo.org/~g2boojum GPG Fingerprint: D706 9802 1663 DEF5 81B0 9573 A6DC 7152 E0F6 5B76 pgp848fu89eTh.pgp Description: PGP signature
Re: [gentoo-dev] Re: Re: [gentoo-core] crap use flags in the profiles
On Wed, 2005-08-31 at 16:32 +0100, Ciaran McCreesh wrote: It's not magic. We've been handling packages that work on sparc64 but not sparc32 for years with a single keyword. Just because you (and, from the looks of things, most of the x86 and amd64 developers) don't know about some of portage's features doesn't mean they don't exist :) This was kinda my point. Developers on sparc and mips are aware of these issues. Developers on amd64 and x86 are not. Forcing the merging of these KEYWORDS would destroy the QA that already is done on these architectures, as most of our developer pool would need some serious training to be able to do this. I know that *I* don't know how to do this myself (yet), but I'm going to look into it. Anyway, can we *please* quit hijacking this thread, as this isn't a x86 vs. other arches thread and rather a thread about profiles and their USE flags. If you want to discuss x86's defficiencies, start your own thread, please. -- Chris Gianelloni Release Engineering - Strategic Lead/QA Manager Games - Developer Gentoo Linux signature.asc Description: This is a digitally signed message part