Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 9 January 2013 20:23, Diego Elio Pettenò flamee...@flameeyes.eu wrote: On 09/01/2013 13:20, Duncan wrote: Are the git migration blockers at such a point that we can get an ETA yet? PLEASE ALL STOP DETOURING EVERY DAMN TOPIC OUT THERE WITH THE GIT MIGRATION, THANK YOU VERY MUCH. And yes I know it's not polite to scream. At this point I DON'T CARE. -- Diego Elio Pettenò — Flameeyes flamee...@flameeyes.eu — http://blog.flameeyes.eu/ Please stop behaving like a troll. If you don't want to discuss the git migration, then simply ignore that part of the conversation. -- Cheers, Ben | yngwin Gentoo developer Gentoo Qt project lead, Gentoo Wiki admin
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 9 January 2013 23:16, Duncan 1i5t5.dun...@cox.net wrote: Diego Elio Pettenò posted on Wed, 09 Jan 2013 13:23:13 +0100 as excerpted: On 09/01/2013 13:20, Duncan wrote: Are the git migration blockers at such a point that we can get an ETA yet? PLEASE ALL STOP DETOURING EVERY DAMN TOPIC OUT THERE WITH THE GIT MIGRATION, THANK YOU VERY MUCH. And yes I know it's not polite to scream. At this point I DON'T CARE. A single reply, possibly even a single sentence, is what I asked for, specifically stating I didn't intend it to become a subthread. Instead, I got... a content-free troll, in ALL CAPS, even.. =:( And here I am adding to it. I was successfully trolled too and know it but still can't help it. =:^( But I should have known there's nothing more likely to bring out a troll than a post saying I don't want a subthread... Sorry everyone. 100% honest, I /just/ wanted a single one one sentence answer and had no intention of... this! ARGH! No need to be sorry. Diego was way out of line. -- Cheers, Ben | yngwin Gentoo developer Gentoo Qt project lead, Gentoo Wiki admin
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 01/08/2013 11:49 PM, Duncan wrote: Zac Medico posted on Tue, 08 Jan 2013 23:36:59 -0800 as excerpted: Thought: Do the CVS keyword expansion in repoman, and then feed the expanded file to CVS for commit. This gives you a fixed file, which you can then generate your manifest against. Yeah, I guess that will work, if we disable the keyword expansion on the CVS server (is it possible?). If I'm reading him correctly, Douglas is suggesting to do the expansion in repoman, so the keyword that are normally expanded aren't even there to expand any longer. I guess we could use the cvs -ko option [1] on all files. That's apparently what prevents $Header expansion for $PORTDIR/skel.ebuild. Then expansion doesn't need to be disabled on the server, since the ebuild has no expandable keywords to trigger it. That leaves the ability to still use the keywords intact, if there's a special-case need for them or something. The expanded form still contains the $Header string that triggers the expansion though, so I we need to use the cvs -ko option or something like it. Of course that assumes that the keywords are suitably distinct such that they won't ordinarily be found in the pre-expanded lines. Whether that's actually the case or not I've no idea... Well, I'd suggest to simply drop the keywords. Weren't we planning to drop the CVS keywords for the git migration, anyway? [1] http://www.cs.utah.edu/dept/old/texinfo/cvs/cvs_16.html#SEC55 -- Thanks, Zac
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 09-01-2013 00:31:04 -0800, Zac Medico wrote: Of course that assumes that the keywords are suitably distinct such that they won't ordinarily be found in the pre-expanded lines. Whether that's actually the case or not I've no idea... Well, I'd suggest to simply drop the keywords. Weren't we planning to drop the CVS keywords for the git migration, anyway? Yeah, but I'd really appreciate it if they could stay for as long as we're on CVS, so my scripts that use the version number to retrieve diffs and apply them to the Prefix' tree versions keep on working. Fabian -- Fabian Groffen Gentoo on a different level signature.asc Description: Digital signature
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 09/01/2013 10:09, Fabian Groffen wrote: Yeah, but I'd really appreciate it if they could stay for as long as we're on CVS, so my scripts that use the version number to retrieve diffs and apply them to the Prefix' tree versions keep on working. Since we're discussing adding this on Portage side — why not having it generate a per-package serial number instead? That could probably make more sense for everything... -- Diego Elio Pettenò — Flameeyes flamee...@flameeyes.eu — http://blog.flameeyes.eu/ signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 09-01-2013 10:14:21 +0100, Diego Elio Pettenò wrote: On 09/01/2013 10:09, Fabian Groffen wrote: Yeah, but I'd really appreciate it if they could stay for as long as we're on CVS, so my scripts that use the version number to retrieve diffs and apply them to the Prefix' tree versions keep on working. Since we're discussing adding this on Portage side — why not having it generate a per-package serial number instead? That could probably make more sense for everything... Yes, except that there would be no way to ask the VCS to retrieve the version for a given serial number, other than doing some guessing and checking the guess was right. The VCS no longer is in control of the number, so there is no guarantee numbers go up/down or with step 1. I guess, when implementing a serial number for a package we best do that server-side, when we generate the rsync image. -- Fabian Groffen Gentoo on a different level signature.asc Description: Digital signature
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 09/01/2013 13:20, Duncan wrote: Are the git migration blockers at such a point that we can get an ETA yet? PLEASE ALL STOP DETOURING EVERY DAMN TOPIC OUT THERE WITH THE GIT MIGRATION, THANK YOU VERY MUCH. And yes I know it's not polite to scream. At this point I DON'T CARE. -- Diego Elio Pettenò — Flameeyes flamee...@flameeyes.eu — http://blog.flameeyes.eu/
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On Wed, Jan 9, 2013 at 7:20 AM, Duncan 1i5t5.dun...@cox.net wrote: Zac Medico posted on Tue, 08 Jan 2013 23:42:39 -0800 as excerpted: Weren't we planning to drop the CVS keywords for the git migration, anyway? Are the git migration blockers at such a point that we can get an ETA yet? Something rough, like are we talking near midyear, late this year or early next, maybe two years out, or not close enough to even approach a reasonable guess, yet. Because if it's actually looking like this year, then anything like switching on cvs to serial numbers as mentioned elsewhere, probably isn't worth the trouble. Let's focus on the git switch and get 'er done! ++ There really isn't that much left - certainly not in comparison to adding serial numbers to ebuilds and building all kinds of logic that relies on them, in which case all of that logic becomes ADDITIONAL blockers on the git migration. It really makes no sense to add more and more logic to the VCS when we're going to be ripping it all out and replacing it. Rich
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On Wed, 09 Jan 2013 13:23:13 +0100 Diego Elio Pettenò flamee...@flameeyes.eu wrote: PLEASE ALL STOP DETOURING EVERY DAMN TOPIC OUT THERE WITH THE GIT MIGRATION, THANK YOU VERY MUCH. Translation: We all know that there are lots of things that would be a hell of a lot easier if we weren't the only project in the world still using CVS, but the Git migration is never going to happen, so mentioning it just makes everyone angry. -- Ciaran McCreesh signature.asc Description: PGP signature
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 09/01/2013 13:37, Ciaran McCreesh wrote: Translation: We all know that there are lots of things that would be a hell of a lot easier if we weren't the only project in the world still using CVS, but the Git migration is never going to happen, so mentioning it just makes everyone angry. No, since you don't seem to speak proper English I'll translate it properly: I don't really give a fuck about the Git migration at this point because I have other things to care about, somebody else is taking care of it. On the other hand, I'm tired of half the threads on this bloody mailing list to end up discussing back-and-forth about the Git migration, especially when half the people posting have no idea of what's going on, are not involved in the migration, or expect said migration to solve all the issues of the world, poverty included. -- Diego Elio Pettenò — Flameeyes flamee...@flameeyes.eu — http://blog.flameeyes.eu/ signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On Wed, Jan 9, 2013 at 7:42 AM, Diego Elio Pettenò flamee...@flameeyes.eu wrote: On 09/01/2013 13:37, Ciaran McCreesh wrote: Translation: We all know that there are lots of things that would be a hell of a lot easier if we weren't the only project in the world still using CVS, but the Git migration is never going to happen, so mentioning it just makes everyone angry. No, since you don't seem to speak proper English I'll translate it properly: I don't really give a fuck about the Git migration at this point because I have other things to care about, somebody else is taking care of it. On the other hand, I'm tired of half the threads on this bloody mailing list to end up discussing back-and-forth about the Git migration... The problem with this logic is that the topic is still quite relevant to Gentoo, and building more and more hacks into how we handle packages to make up for the fact that we're not using git means that there are going to be more and more hacks to undo when the time for the switch actually comes. It is really not productive to have 3 people working on switching to git, and 10 people working on adding more blockers to the migration. Logic that depends on cvs keywords in ebuilds should be considered deprecated now. Logic that creates serial numbers that are going to change in six months should be considered deprecated before it is implemented (unless nobody cares that they're going to change). especially when half the people posting have no idea of what's going on, are not involved in the migration, or expect said migration to solve all the issues of the world, poverty included. Well, they'll solve this particular set of issues, and therefore the topic is quite relevant. Those who are knowledgeable are welcome to reply - such replies will be more helpful than blanket don't talk about git posts. Rich
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 01/09/2013 01:09 AM, Fabian Groffen wrote: On 09-01-2013 00:31:04 -0800, Zac Medico wrote: Of course that assumes that the keywords are suitably distinct such that they won't ordinarily be found in the pre-expanded lines. Whether that's actually the case or not I've no idea... Well, I'd suggest to simply drop the keywords. Weren't we planning to drop the CVS keywords for the git migration, anyway? Yeah, but I'd really appreciate it if they could stay for as long as we're on CVS, so my scripts that use the version number to retrieve diffs and apply them to the Prefix' tree versions keep on working. If we had a live cvs - git sync, then I'd suggest that you migrate your scripts to use that instead. However, it looks like this one is not synced regularly (last sync was about 2 months ago): http://git-exp.overlays.gentoo.org/gitweb/?p=exp/gentoo-x86.git;a=summary -- Thanks, Zac
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 09-01-2013 05:06:15 -0800, Zac Medico wrote: If we had a live cvs - git sync, then I'd suggest that you migrate your scripts to use that instead. However, it looks like this one is not synced regularly (last sync was about 2 months ago): http://git-exp.overlays.gentoo.org/gitweb/?p=exp/gentoo-x86.git;a=summary This is if stuff. If it wasn't if, we'd already done it, of course. -- Fabian Groffen Gentoo on a different level signature.asc Description: Digital signature
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On Wed, 09 Jan 2013 05:06:15 -0800 Zac Medico zmed...@gentoo.org wrote: On 01/09/2013 01:09 AM, Fabian Groffen wrote: On 09-01-2013 00:31:04 -0800, Zac Medico wrote: Of course that assumes that the keywords are suitably distinct such that they won't ordinarily be found in the pre-expanded lines. Whether that's actually the case or not I've no idea... Well, I'd suggest to simply drop the keywords. Weren't we planning to drop the CVS keywords for the git migration, anyway? Yeah, but I'd really appreciate it if they could stay for as long as we're on CVS, so my scripts that use the version number to retrieve diffs and apply them to the Prefix' tree versions keep on working. If we had a live cvs - git sync, then I'd suggest that you migrate your scripts to use that instead. However, it looks like this one is not synced regularly (last sync was about 2 months ago): http://git-exp.overlays.gentoo.org/gitweb/?p=exp/gentoo-x86.git;a=summary One quite good but not official mirror is: https://bitbucket.org/lmnd/gentoo-x86/overview If anyone wanted to use gx86 via git, I'd suggest them using this. However, this is not a Gentoo project and we can't guarantee that someday it won't disappear or stop being updated. -- Best regards, Michał Górny signature.asc Description: PGP signature
Re: [gentoo-dev] Re: Should portage tree CVS impose a commit moratorium during snapshot creation?
On 01/09/2013 12:31 AM, Zac Medico wrote: I guess we could use the cvs -ko option [1] on all files. That's apparently what prevents $Header expansion for $PORTDIR/skel.ebuild. Actually, we should use -kb rather than -ko, since -kb disables transformations entirely [1]. The -ko mode is identical except that it allows transformation of line endings, and we don't need or want any such transformations. It turns out that repoman has recognized the -kb mode for a very long time [2], and when it recognizes that there are no keywords to expand, it commits the Manifest in the same commit as the rest of the files (behaving exactly like it does in git repositories). [1] http://cvsgui.sourceforge.net/howto/cvsdoc/cvs_12.html#SEC102 [2] http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=d59466926e57f1226fe654deb109812a1cd19389 -- Thanks, Zac
