Re: [gentoo-dev] Review for initial systemd.eclass
On Sat, 30 Apr 2011 23:17:42 +0200 Michał Górny mgo...@gentoo.org wrote: Here's the second version. I committed this version to gx86. All packages providing systemd units are free to use it now, yet I'm still open to comments. -- Best regards, Michał Górny signature.asc Description: PGP signature
Re: [gentoo-dev] ACE gcc and libc dependency
On Tue, May 3, 2011 at 7:15 PM, Mike Gilbert floppymas...@gmail.com wrote: On Tue, May 3, 2011 at 11:53 AM, Paweł Hajdan, Jr. phajdan...@gentoo.org wrote: On 5/3/11 5:27 PM, Kfir Lavi wrote: In the ebuild there is no mention of runtime dependency like gcc or glibc. See http://devmanual.gentoo.org/general-concepts/dependencies/index.html#implicit-system-dependency Why sys-devel/gcc don't have a library version without the actual compiler? This question may be a bit hard to understand, at least for me. Other distros package libstdc++ separately from gcc. On Gentoo, the one package (sys-devel/gcc) provides both. Maybe we should add a new USE flag that will install just the libraries. (USE=justlibs) Kfir
Re: [gentoo-dev] ACE gcc and libc dependency
On Wed, 4 May 2011 14:09:03 +0300 Kfir Lavi lavi.k...@gmail.com wrote: Maybe we should add a new USE flag that will install just the libraries. (USE=justlibs) Then you'd have to update nearly every package that deps upon GCC to have a [-justlibs] dependency. Arguably not so bad for GCC, but in general it's a huge problem, and it's the reason [client] and [server] flags aren't popular. (And if features like those are desired, please don't start updating zillions of ebuilds with [-blah] flags. Ask for an EAPI feature that lets flags in IUSE be marked as require me this way unless something explicitly says otherwise...) -- Ciaran McCreesh signature.asc Description: PGP signature
Re: [gentoo-dev] Review for initial systemd.eclass
On Wed, May 04, 2011 at 12:54:45PM +0200, Michał Górny wrote: On Sat, 30 Apr 2011 23:17:42 +0200 Michał Górny mgo...@gentoo.org wrote: Here's the second version. I think the inherit multilib line is no longer needed. I committed this version to gx86. All packages providing systemd units are free to use it now, yet I'm still open to comments. Does this mean we can or should file bugs like [1] and [2] against packages? I am in particular thinking of avahi, consolekit, and soon networkmanager. Thanks, Henry [1] https://bugs.gentoo.org/show_bug.cgi?id=365941 [2] https://bugs.gentoo.org/show_bug.cgi?id=365943
Re: [gentoo-dev] Git migration?
Am Mittwoch 04 Mai 2011, 00:18:08 schrieb Alec Warner: ask on the gentoo-scm list? -A I'd say this is of enough general interest, and the relevant people are for sure also subscribed here... -A -- Andreas K. Huettel Gentoo Linux developer - kde, sci, arm, tex dilfri...@gentoo.org http://www.akhuettel.de/ signature.asc Description: This is a digitally signed message part.
Re: [gentoo-dev] Git migration?
On Wed, May 4, 2011 at 9:23 AM, Andreas K. Huettel dilfri...@gentoo.org wrote: Am Mittwoch 04 Mai 2011, 00:18:08 schrieb Alec Warner: ask on the gentoo-scm list? -A I'd say this is of enough general interest, and the relevant people are for sure also subscribed here... -A -- Andreas K. Huettel Gentoo Linux developer - kde, sci, arm, tex dilfri...@gentoo.org http://www.akhuettel.de/ Andreas, While what you say is likely true, what occurs with your request is fragmentation of the information. That said, I think the majority of the information you seek is available in the gentoo-scm archives. From what I know about this, I think the issue is time and the ancillary utilities, like repoman, that will need to be updated. Beyond that documentation is required. See this thread for more info. http://archives.gentoo.org/gentoo-scm/msg_98932c55ec10fcc5445ab950e62b12dc.xml I know we are all pretty excited to migrate to git, but there is little reason to rush something this monumental. It seems far better to take the time to get it absolutely correct, technically (the best kind), the first time. I am certain, in any case, that your interest and involvement in this effort would be appreciated. Thanks, Matthew -- Matthew W. Summers Gentoo Foundation Inc.
[gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in sys-kernel/git-sources: ChangeLog git-sources-2.6.39_rc6.ebuild
On Wed, 4 May 2011 15:13:18 + (UTC), Mike Pagano (mpagano) wrote: mpagano 11/05/04 15:13:18 Modified: ChangeLog Added:git-sources-2.6.39_rc6.ebuild Log: Version bump snip K_EXTRAEINFO=This kernel is not supported by Gentoo due to its unstable and experimental nature. If you have any issues, try a matching vanilla-sources ebuild -- if the problem is not there, please contact the upstream kernel developers at http://bugme.osdl.org and on the linux-kernel mailing list to report the problem so it can be fixed in time for the next kernel release. pkg_postinst() { postinst_sources } Hello Mike, http://bugme.osdl.org is 404, do you mean bugzilla.kernel.org? I've tried to ask you on IRC once about this in the recent past but we didn't connect. Thanks, Jeremy
Re: [gentoo-dev] Review for initial systemd.eclass
On Wed, 4 May 2011 15:56:32 +0200 Henry Gebhardt hsggebha...@googlemail.com wrote: On Wed, May 04, 2011 at 12:54:45PM +0200, Michał Górny wrote: On Sat, 30 Apr 2011 23:17:42 +0200 Michał Górny mgo...@gentoo.org wrote: Here's the second version. I think the inherit multilib line is no longer needed. Thanks, fixed. I committed this version to gx86. All packages providing systemd units are free to use it now, yet I'm still open to comments. Does this mean we can or should file bugs like [1] and [2] against packages? I am in particular thinking of avahi, consolekit, and soon networkmanager. I'd wait with that until these two packages get updated. Then feel free to proceed, though I think I'll file some of them myself. -- Best regards, Michał Górny signature.asc Description: PGP signature
Re: [gentoo-dev] Re: [rfc] Rendering the official Gentoo logo / Blender 2.04, Python 2.2
On 05/01/2011 06:10 AM, Ulrich Mueller wrote: Could you bisect Blender to find out why it doesn't work with the new version? If 2.26 still produced good results, 2.37a already does not. Bisecting involves fixing compilation for each version. I stopped getting 2.30 to compile because it seemed to take forever (longer than fixing 2.26, 2.37a and 2.40 together) and two people had their hands on a port of the logo to Blender 2.57 by then, one of them still has. It's too early to give details. What I can say is that personally I would want a very close match in case of a Blender-based replacement, closer than what I have seen so far. It still seems possible though. Best, Sebastian
Re: [gentoo-dev] Re: [rfc] Rendering the official Gentoo logo / Blender 2.04, Python 2.2
On 05/01/2011 08:06 AM, Michał Górny wrote: Isn't it possible to create a better SVG then? It may be. Of the three variants trying to match the Blender version that I have seen so far, none is a replacement of equal quality on the bling scale to my impression. They feel like tradeoffs, not like the real thing. Maybe they try to come too close to the ray-traced rendering, but I'm not sure if I really want to propose a different direction either. I think such a variant would be much more portable and reproducible than blender files. What I dislike about the idea of moving to a new logo is that we would give up part of our culture just because we were unable to move it from past to present to future. Imagine this dialog: A: Hey guys, I noticed you have a new logo? B: Yeah, blender rendering changed - so we dropped it. I don't really want to be B in that dialog. I see the pragmatic aspect of moving to SVG but it also has the taste of giving up to me. To vercome that taste, a very strong replacement would be needed. If we replace the Blender g we may also need a substitute for the red-white Blender gentoo as seen at http://sources.gentoo.org/cgi-bin/viewvc.cgi/*docroot*/images/gentoo-new.gif if just for the sake of consistency. I am wondering what effect the Blender nature of a logo does have on the capability and will of people to create fan art based on it compared to an SVG version. It seems like there is only a handful of 3D Gentoo wallpapers but does that mean it would have been more with an SVG version, instead? On what levels could SVG work as a catalyst? If we ported the logo to Blender 2.57 now: what can we do to not be running after Blender rendering changes for all time or to reduce their impact on us? Is this a natural cost or an evil one? Just my 2 cents. Best, Sebastian
Re: [gentoo-dev] Re: [rfc] Rendering the official Gentoo logo / Blender 2.04, Python 2.2
On 05/01/2011 06:10 AM, Ulrich Mueller wrote: Could you bisect Blender to find out why it doesn't work with the new version? I tried a few more versions now. While Blender 2.31a still applies the reflection texture, Blender 2.32 does not anymore. At least on http://download.blender.org/source/ these two appear as consecutive releases. Both of these seem to run fine with Python 2.4.6, which is still in Gentoo. Without good image diffs, I cannot tell for sure if the rendering has changed since Blender 2.04. Best, Sebastian
Re: [gentoo-dev] Re: [rfc] Rendering the official Gentoo logo / Blender 2.04, Python 2.2
On Thu, 05 May 2011 04:31:10 +0200 Sebastian Pipping sp...@gentoo.org wrote: I think such a variant would be much more portable and reproducible than blender files. What I dislike about the idea of moving to a new logo is that we would give up part of our culture just because we were unable to move it from past to present to future. Imagine this dialog: A: Hey guys, I noticed you have a new logo? B: Yeah, blender rendering changed - so we dropped it. Well, the concept is to have the same logo but in new format. Of course, that would require some of professional work to ensure best-match rendering if even possible. And I don't think it's really giving up. It's just moving to the future, putting back good, ol' things in the museum. However, it'd be still best to fix it. If we replace the Blender g we may also need a substitute for the red-white Blender gentoo as seen at http://sources.gentoo.org/cgi-bin/viewvc.cgi/*docroot*/images/gentoo-new.gif if just for the sake of consistency. True. I am wondering what effect the Blender nature of a logo does have on the capability and will of people to create fan art based on it compared to an SVG version. It seems like there is only a handful of 3D Gentoo wallpapers but does that mean it would have been more with an SVG version, instead? On what levels could SVG work as a catalyst? As I see the concept of a logo, it is bound to a pretty strict rendering. It's fun to have it in 3D but I'm not sure if, say, rotating it or changing lighting would still make it the same logo. -- Best regards, Michał Górny signature.asc Description: PGP signature