<rlb>Is there a guile 1.8 -> 2.0 conversion guide anywhere? <civodul>rlb: not really, but almost everything should be in NEWS under "Changes in 2.0.0" <rlb>So I'm addressing the guile-1.8 removal question now, and wanted to make sure I include any relevant information from people here -- what was the opinion with respect to g-wrap/guile-cairo/guile-gnome-platform again? i.e. was the opinion that they should be removed from debian, removed from jessie, or ...? <mark_weaver>daviid's position was that g-wrap/guile-cairo/guile-gnome-platform should be removed from jessie if not updated to use guile 2.0. <mark_weaver>all of those packages support guile 2 now, which makes it a shame though :-( <mark_weaver>guile-cairo in particular is very easy to build on guile 2 <rlb>and how should I represent that to the release managers -- i.e. is daviid the upstream now? <mark_weaver>rlb: did you see wingo ask you whether he should cut a new guile-cairo release? <mark_weaver>that one is so easy, and quite useful for producing graphics, that it would be a shame not to include it. <mark_weaver>but I guess we should have moved on all of this much earlier, so I'm not blaming anyone. <mark_weaver>the real problem here is that rotty_ seems to have lost interest in guile lately, and is not maintaining his debian guile packages. <tadni`>I mean, besides maybe guile-cario, are any of these bindings really being used at all? :^P <tadni`>They're still on the 2.x branch of GNOME, that alone, I think most limit it's use. <rlb>wrt wingo, right -- I suppose the question is "who's going to maintain it"? i.e. it's a reasonable argument that we should be inclined against having packages in debian that don't have an active maintainer for X years. <mark_weaver>tadni`: if they're not being used much, it's probably because they are not widely packaged. they are very useful. <rlb>last non-NMU for guile-cairo was 2011-06 <rlb>and before that 2007-03 *rlb is not throwing stones (glass houses and all), but it's a question that's likely to be asked. <mark_weaver>obviously, these packages need to be maintained by someone in debian. <tadni`>mark_weaver: I mean, I expect a resurgence of development, eventually, for guile-GNOME -- but yeah, the fact that we might miss a major distro release is a bit of a bummer. <mark_weaver>I guess we'll just have to rely on packages outside of debian for now. <rlb>we may not -- I'm not at all sure where this will end up -- though obviously I have some inclination toward avoiding having to deal with guile-1.8 for X more years. <rlb>mark_weaver: well, another option is to just leave them in unstable (maybe guile-1.8 included), but not in jessie. <tadni`>What programs are still heavily dependent on 1.8? Lilypond? <mark_weaver>rlb: I think daviid is probably right that any packages that support guile 2.0 upstream but are built against 1.8 in debian should be removed. <rlb>so would I say that "some of the guile upstream people have recommended..." when I'm describing the situation? *rlb is just trying to gather a status update <mark_weaver>rlb: sure, you can say that's my recommendation, as one of the guile maintainers. <mark_weaver>rlb: but I'm only making that recommendation for packages that already support guile 2.0 upstream. <mark_weaver>my rationale is that in such cases, I'd rather people build the upstream package from source against guile 2.0 than to continue using guile 1.8. <rlb>mark_weaver: and did you mean removed entirely, or just from jessie, or no pref? <tadni`>rlb: I meant outside of Debian, like actually still require -- require, 1.8. <tadni`>Not just being built on it for some reason or another. *rlb only knows what dak tells him ;> <mark_weaver>rlb: that's a good question. I guess it depends on whether removing them would significantly increase the difficulty of adding them back to debian after they have been updated. <rlb>mark_weaver: certainly wouldn't be *as* easy iirc. <rlb>though it could be faster I suppose -- depends on which takes longer, going through the proper maintainer MIA process, or getting a "new" package approved. <rlb>of course someone could start the mia groundwork now... <rlb>mark_weaver: wrt the maintenance of some of these packages, I'm very hesitant to take on more debian work right now -- though I suppose for packages that are well maintained upstream (with responsive upstreams), it might not be too much work. <rlb>though I'm also hesitant to maintain things I don't use myself. <mark_weaver>I think we should act on the assumption that those packages (g-wrap, guile-cairo, guile-gnome) are still active upstream and should eventually be part of debian again, when they have been updated to use guile 2.0. <rlb>suppose could also have collab-maint'ish setup for them... <mark_weaver>so, with that in mind, I don't have a strong opinion on whether they should remain in sid. whatever you think is best. <rlb>i.e. some set of people who can commit to a git repo, and then some subset of those that are maintainers and can upload to debian... <rlb>(maintainers or developers) <rlb>anyone heard anything more about lilypond's transition ( <rlb>I'm about to ping the debian maintainer) <rlb>mark_weaver: ok, thanks <mark_weaver>i wonder if I should apply to be a debian maintainer <rlb>mark_weaver: be happy to have the help ;> <rlb>mark_weaver: though there are also a lot of ways you might help without being a maintainer/developer... <mark_weaver>I've always been intimidated by tales of how long and difficult the process is. <mark_weaver>well, let's talk about this another time. currently, I have a 3.5 year old climbing all over me as I try to type this. probably time for a break :) <rlb>mark_weaver: I can't shed much light there (wrt process) -- when I joined it was quite a bit different. <rlb>mark_weaver: no problem -- thanks for the help <rlb>mark_weaver: but fwiw, I'm very glad I joined. ***micro__ is now known as micro
***micro is now known as Guest1705
***Guest1705 is now known as micro^
*rlb can't figure out why dak still thinks swig2.0 depends on guile-1.8... <rlb>(hmm, unless it uses g-wrap?) <daviid>mark_weaver, rlb, just red the conversation, I confirm that I personally recommend the removal of gwrap, guile-gwrap and guile-cairo from bebian [testing unstable] <rlb>looks like guile-cairo is just waiting on a package migration? *rlb checks to see why it hasn't migrated... <daviid>rlb: the latest guile-cairo tarball does _nor_ include the patches required to proper integrate it with guile-gnome and guile-clutter <daviid>it 'just' needs to be released, but it hasn't been done [in years] <rlb>daviid: so you're saying that that upload's not going to work? <rlb>i.e. the guile-cairo upload? <daviid>it will work for people using cairo on its own, but the guile-gnome [devel] and guile-gnome [clutter-devel] needs a guile-cairo built from a git clone <rlb>so that package will break the guile-gnome-platform package? *rlb may not understand the debian package mapping <daviid>unless the nmu was done from a guile-cairo git clone yes <daviid>rlb: imh, it is _much_ better, for the time beeing, to recommend building these packages from their respective latest git branch: g-wrap [devel, for the latest doc], guile-cairo [master], guile-gnome [devel, guile-clutter [guile-gnome clutter-devel branch] <rlb>mark_weaver: though I don't always enjoy *this* part of the job ;> <rlb>daviid: here's what I have in my summary on the topic (not sent yet): <rlb>"As a general comment, people on #guile, including one of the upstream <rlb>Guile maintainers, have said that at this point, they would prefer that <rlb>we remove packages from jessie that support Guile 2.0 upstream, but <rlb>still depend on 1.8 (though we might or might not keep them in <rlb>unstable). This was mentioned, in particular, with respect to g-wrap, <rlb>guile-cairo, and guile-gnome-platform." <daviid>don't forget to remove guile-gwrap as well <rlb>that's not part of g-wrap? <rlb>i.e. is one not the source package for the other? <rlb>hmm, well dak only cares about one? I'll check. <mark_weaver>is that one new enough to include in debian without causing problems with guile-gnome? <daviid>mark_weaver: i think it does not include latest pacthes wingo wrote 'for me' while binding clutter-1.10 <mark_weaver>I'm not sure those missing patches are sufficient justification for removing guile-cairo from debian, if we could have a slightly older guile-cairo in debian that works with guile 2.0 <daviid>mark_weaver: ok, but that guile-cairo won't work for guile-gnome <mark_weaver>civodul: we are discussing whether to recommend the removal of g-wrap, guile-cairo, and guile-gnome from debian, since the versions in debian are for guile 1.8 but the upstream versions already support 2.0. <mark_weaver>I agree that we should remove packages that support 2.0 upstream but use 1.8 in debian. <mark_weaver>but there's still a question of guile-cairo, which could be easily fixed to support guile 2.0 in debian, but it's missing some recent patches needed to work with the latest guile-gnome. *mark_weaver feels bad that all of these packages have supported guile 2.0 for a very long time, but the debian maintainer for those packages hasn't updated them :-( <daviid>mark_weaver: it's also because we did not release <mark_weaver>daviid: well, there are *plenty* of packages in debian that are taken from git. <mark_weaver>in fact, a lot of upstream devs are losing interest in cutting official releases, because debian will just use a git checkout instead, so why should they care? (grr) <daviid>anyway, my quizz is how will we explain that the debian guile-cairo pkg does work for guile-gnome? <mark_weaver>if someone is going to go through the trouble of compiling guile-gnome from source code, it's not much more work to say they have to rebuild guile-cairo also. <mark_weaver>on the other hand, guile-cairo is quite useful on its own *rlb still has some concern having a newly update package in jessie with no maintainer to handle bugs... <mark_weaver>yeah, the bottom line is that we need more help from debian devs/maintainers. <daviid>right, and from that point of view... said my opinion already... <civodul>mark_weaver, rlb: i think the guile-* packages in Debian should be updated to use Guile 2.0 when possible <mark_weaver>and maybe that means one of us needs to become one of those things, to help from the other side. <civodul>guile-cairo definitely works with 2.0 <daviid>g-wrap, guile-cairo and guile-gnome-platform all work with guile-2 <civodul>so either switch them to 2.0, or add a "-2" version of each <mark_weaver>civodul: our guile-cairo package in guix has a few substitutions to compile with guile-2, but it's quite trivial <civodul>mark_weaver: yes, and it's mostly about setting installation dirs <rlb>I could probably help fix them wrt nmu uploads, though it'd have to be fast... <rlb>I'll add that as a note to the status update. <rlb>mark_weaver: and with help I could probably handle more packages -- i.e. your help with 2.0 was quite effective (thanks). <rlb>but I'm not likely to have time to become a guile-gnome/platform/gwrap expert myself anytime soon. <rlb>mark_weaver: well once I hear back from don wrt lilypond, I'll send the status update, and we'll see -- it could turn out to be irrelevant if debian decides that we're just going to keep 1.8. <mark_weaver>if you have a little bit of time to spare, I think that updating guile-cairo to use guile-2.0 for jessie would be quite easy. <rlb>(well not completely irrelevant) <rlb>i.e. might still be good to have the other packages updated <rlb>mark_weaver: ok, we'll see -- so far it's taken me hours today just to get the status info collected and double-checked... <rlb>(and I still may have it wrong) <rlb>(have to check the package, then the unstable->testing propagation, then the buildds, etc.) <rlb>mark_weaver: if I decide to proceed, I'll probably just (badger y'all and then) add the relevant packages to my public git repos for now. <rlb>though eventually they might should be collab-maint (which I've never figured out :/) <mark_weaver>please don't hesitate to badger us if we can help get more guile-2.0 updates into jessie. <rlb>that might allow me to add others (even non-debian people?) as committers -- though with git that's not nearly as important... <rlb>and if I continue to use git-dpm -- other people might rather just let me handle it ;> <rlb>mark_weaver: ok -- thanks <daviid>rlb: mark_weaver, the guile-2 will or will not contain the patch that breaks guile-gnopme?