So I found out today when I went on IRC that someone had remasked dbus, a package that I'm the maintainer for. No one told me, no bugs were filed. Nothing. They just up and did it themselves.
Jan 04 11:29:44 Cardoe when are we gonna get OOo binaries? Jan 04 11:29:47 Astinus JoseJX: by the way, did you guys ~ppc the new dbus 0.60 and then downgrade it last nigh$Jan 04 11:30:24 axxo 60 was package.masked Jan 04 11:30:25 JoseJX Cardoe: Ask the OO.o folks, they haven't updated the ppc page since 1.1.2 Jan 04 11:30:47 * georges has quit (Connection timed out) Jan 04 11:30:48 JoseJX Perhaps suka can look into what we'd need to do to build one for them. Jan 04 11:30:55 Astinus Somehow my PB managed to upgrade itself to 0.60 last night, then downgrade this morning$Jan 04 11:31:03 * wrobel is now known as wrobel_away Jan 04 11:31:03 * Astinus wondered what the hell was going on Jan 04 11:31:08 JoseJX Astinus: I'm not sure, check the cvs logs I guess. Jan 04 11:31:12 Cardoe JoseJX: nice.... Jan 04 11:31:13 axxo it was unmasked & then remasked by someone else Jan 04 11:31:22 Astinus axxo: Mmkay Jan 04 11:31:38 Astinus Well, for anyone who cares, it works fine on my PPC *shrug* Jan 04 11:31:49 Cardoe I wanna know who masked it Jan 04 11:32:01 Astinus Cardoe: you a PPC dev? Jan 04 11:32:13 Cardoe Astinus: no Jan 04 11:32:16 Cardoe I'm the dbus maintainer Jan 04 11:32:21 Astinus Ah, lol :P Jan 04 11:32:27 axxo 24 # John N. Laliberte <allanonjl@gentoo.org> (03 Jan 2006) Jan 04 11:32:27 axxo 25 # undo Cardoe's unmask due to possible unresolved issues Jan 04 11:32:27 axxo 26 >=sys-apps/dbus-0.60 Jan 04 11:32:40 Astinus 'Works for me' Jan 04 11:32:55 Astinus Cardoe: well, GCC 3.4.5 with gcj and dbus with gcj != working Jan 04 11:33:02 dmwaters hi all Jan 04 11:33:13 Cardoe Astinus: what's the issue? Jan 04 11:33:13 Astinus hey deedra Jan 04 11:33:44 Cardoe Astinus: I need bug reports. Jan 04 11:33:46 Astinus Cardoe: compile error, near the end, didn't have time to investigate to I just -gcj'd d$Jan 04 11:33:52 AllanonJL|W Cardoe: i remasked it Jan 04 11:34:00 Cardoe AllanonJL: excuse me? Jan 04 11:34:02 Cardoe AND WHY? Jan 04 11:34:08 Cardoe You do not have the right just to do that Jan 04 11:34:10 Cardoe There are no issues Jan 04 11:34:16 Cardoe except the one Astinus just mentioned to me Jan 04 11:34:19 Cardoe NO BUG REPORTS Jan 04 11:34:25 * Astinus hides Jan 04 11:34:32 axxo run boy run! Jan 04 11:34:40 * [ViRgiLiO] has quit (Read error: 104 (Connection reset by peer)) Jan 04 11:34:45 * [ViRgiLiO] (n=virgilio@84-122-101-222.onocable.ono.com) has joined #gentoo-dev Jan 04 11:35:01 Cardoe I know about the QT bindings not building properly. Jan 04 11:35:05 Cardoe It turns out it's a bug in the eclass Jan 04 11:35:10 Cardoe And it's the same issue with 0.50 Jan 04 11:35:14 Cardoe AND AND AND AND AND!!! Jan 04 11:35:18 Cardoe WITH THE STABLE VERSION Jan 04 11:35:20 Astinus Cardoe: Well you'll have a bug report off me in however long it takes OOo to recompile $Jan 04 11:35:38 Cardoe the QT people have had the bug report for SEVERAL weeks Jan 04 11:35:45 Cardoe so AllanonJL why did you mask it? Jan 04 11:36:07 ciaranm and why is Cardoe only finding out via irc that someone masked it? Jan 04 11:36:10 Astinus 'possible unresolved issues' Jan 04 11:36:20 * Astinus goes Bugzie trawling Jan 04 11:36:22 Cardoe ciaranm: that's my other question Jan 04 11:36:28 AllanonJL|W I had message dyou last night Cardoe Jan 04 11:36:32 AllanonJL|W you went to watch a movie Jan 04 11:36:36 AllanonJL|W remember? Jan 04 11:36:57 Cardoe You asked me last night if I was around Jan 04 11:37:02 Cardoe I said no I was going to the movies Jan 04 11:37:05 Cardoe I said private message me Jan 04 11:37:08 Cardoe or write me an e-mail Jan 04 11:37:14 Cardoe and I'd check it when I got back in 2+ hrs Jan 04 11:37:17 Cardoe and you did NEITHER Jan 04 11:37:19 spyderous fox2mike: eh? Jan 04 11:37:28 Cardoe AllanonJL: You complain about me wanting stuff to happen NOW Jan 04 11:37:34 Astinus Cardoe: bugzie, just search for dbus, there's a lot of 'NEW' bugs Jan 04 11:37:36 Cardoe But I don't even demand less then 2 hrs. Jan 04 11:37:57 Cardoe Astinus: a lot of them aren't dbus bugs Jan 04 11:38:16 spyderous ferringb: cool, thanks. Jan 04 11:38:18 Cardoe Astinus: there's 1 new one Jan 04 11:38:33 Cardoe Astinus: there's the 2 QT bugs.. which are just newer versions of the old bugs. Jan 04 11:38:40 joem Cardoe: shortly after you unmasked it there were several devs in here complaining about$Jan 04 11:38:43 Cardoe Astinus: that are waiting on the QT herd to fix their eclass Jan 04 11:38:45 * wrobel_away is now known as wrobel Jan 04 11:38:52 Cardoe joem: No there weren't Jan 04 11:38:56 Cardoe joem: because I was here. Jan 04 11:39:01 Cardoe and there is no up/down Jan 04 11:39:03 Anarchy called laziness Jan 04 11:39:13 Anarchy most packages have patches to use new api in 0.60 Jan 04 11:39:19 joem Cardoe: ok shortly might not be the right word...but when I woke up yesterday Jan 04 11:39:20 Cardoe because EVERY package that depends on it I've masked properly. Jan 04 11:39:32 Cardoe And I've provided PATCHES TO EVERY APP that needs it. Jan 04 11:39:41 Cardoe They're in bugzilla. Jan 04 11:39:57 joem assigned to the right maintainers? Jan 04 11:40:00 Cardoe yes Jan 04 11:40:10 Cardoe Spoken to some of them on IRC even Jan 04 11:40:11 joem I didn't see one for totem or rhythmbox Jan 04 11:40:26 Anarchy would you like me to post patch for totem Jan 04 11:40:27 Cardoe Because it's available via upstream Jan 04 11:40:38 * grahl04 (n=grahl@p5489611C.dip.t-dialin.net) has joined #gentoo-dev Jan 04 11:40:38 * ChanServ gives channel operator status to grahl04 Jan 04 11:40:44 joem Cardoe: ok so when you say you provided patches you mean you didn't really Jan 04 11:40:48 Anarchy Cardoe, is right totem does need a bump Jan 04 11:40:50 Cardoe joem: rythembox is there Jan 04 11:40:54 Cardoe http://bugs.gentoo.org/show_bug.cgi?id=114457 Jan 04 11:40:55 Cardoe There Jan 04 11:40:58 Cardoe there's the patch Jan 04 11:41:00 Cardoe right there in the bug Jan 04 11:41:02 latexer and bmpx bug open about dbus? Jan 04 11:41:05 GenBot Cardoe: Bug 114457; "rhythmbox-0.9.2 fails to compile"; [Gentoo Linux :: Ebuilds]; {NEW$Jan 04 11:41:17 Astinus bug 114458 Jan 04 11:41:23 GenBot Astinus: Bug 114458; "totem-1.2.0-r2 fails to compile w/ dbus-0.60"; [Gentoo Linux :: G$Jan 04 11:41:30 Cardoe latexer: there's a patch included. Jan 04 11:41:35 joem Cardoe: ok that bug is assigned to somebody not in the metadata for rb and doesn't have$Jan 04 11:41:40 latexer ok, but not commited to portage. Jan 04 11:41:54 Cardoe latexer: I'm not the maintainer Jan 04 11:41:57 latexer Cardoe: 'patches available by searching bugzilla' doesn't seem to cut it, in my book. Jan 04 11:42:00 Cardoe joem: and totem has the patch Jan 04 11:42:05 Cardoe latexer: what do you want me to do? Jan 04 11:42:13 Cardoe WHAT does everyone want me to do? Jan 04 11:42:14 latexer Cardoe: so you coordinate those being available for unmasking at the same time you unma$Jan 04 11:42:17 Astinus Cardoe: mailbomb their inboxes :P Jan 04 11:42:20 Cardoe I can't FORCE developers to act Jan 04 11:42:28 Cardoe latexer: I asked for packages to be provided with the packages Jan 04 11:42:30 Cardoe but masked Jan 04 11:42:31 Anarchy Cardoe, I am on your side Jan 04 11:42:33 Cardoe and I would mask everything out Jan 04 11:42:40 * [ViRgiLiO] has quit (Read error: 104 (Connection reset by peer)) Jan 04 11:42:43 Cardoe latexer: ALL current versions of all packages BLOCK on the new dbus Jan 04 11:42:46 * [ViRgiLiO] (n=virgilio@84-122-101-222.onocable.ono.com) has joined #gentoo-dev Jan 04 11:42:49 joem Cardoe: totem doesn't have the patch as far as I can see Jan 04 11:43:05 Cardoe latexer: so it's not like your system will pull down the new dbus unless you have versi$Jan 04 11:43:06 Anarchy joem, it is right there on bug report Jan 04 11:43:07 Cardoe http://bugs.gentoo.org/show_bug.cgi?id=114458 Jan 04 11:43:11 Cardoe it's on the bug report Jan 04 11:43:15 latexer Cardoe: so you keep pressuring them to commit it. Jan 04 11:43:17 Astinus joem: That totem bug has a lot of patches all over it. Jan 04 11:43:23 joem Cardoe: ok I'm talking about in the tree Jan 04 11:43:32 Cardoe joem: Commit the patch Jan 04 11:43:35 Cardoe what can I tell you. Jan 04 11:43:37 Astinus !metadata totem Jan 04 11:43:37 jeeves Astinus: Package: media-video/totem Herd: gnome, video, gstreamer Maintainer: gnome, v$Jan 04 11:43:45 Cardoe I asked in #gentoo-desktop Jan 04 11:43:46 joem !herd video Jan 04 11:43:46 jeeves joem: (video) arj, flameeyes, hanno, kzimmerm, lordvan, lu_zero, luckyduck, phosphan, p$ Jan 04 11:43:46 latexer Cardoe: just unmasking it and letting everyone who upgrades dbus have major issues doesn't seem like a solution. Jan 04 11:43:49 Anarchy joem, then I guess I can bitch about devs closing bugs as fixed even tho they are not in the tree Jan 04 11:43:50 Cardoe latexer: NO Jan 04 11:43:54 Cardoe latexer: I did NOT DO THAT Jan 04 11:44:02 Cardoe latexer: I BLOCKED on dbus-0.60 Jan 04 11:44:16 Cardoe ALL progs that had dbus in their depends Jan 04 11:44:21 Cardoe we added a blocker on 0.60 Jan 04 11:44:22 latexer Cardoe: so you're telling me that 'emerge -uD dbus' won't upgrade dbus on folks? Jan 04 11:44:24 joem Cardoe: you did...the blocking causes shitty up/downgrades for people Jan 04 11:44:27 latexer regardless of the block? Jan 04 11:44:36 Cardoe latexer: I added the blocker Jan 04 11:44:47 Mr_Bones_ Halcy0n: ping Jan 04 11:44:54 Cardoe a blocker won't upgrade Jan 04 11:45:19 Astinus I had an up/downgrade, but that's only because someone remasked the fscker :/ It worked too :P Jan 04 11:45:32 latexer Cardoe: ok, that part makes sense to me. Jan 04 11:45:51 Cardoe latexer: The only complaints of up/down that I had were after the mask. Jan 04 11:46:02 Cardoe latexer: and there was 1 package that DIDN'T even list dbus in the depends. Jan 04 11:46:02 joem Cardoe: no it was masked because of complaints Jan 04 11:46:09 joem Cardoe: lemme check log 1 sec Jan 04 11:46:10 Cardoe latexer: and the developer fixed it afterwards Jan 04 11:46:21 Cardoe latexer: but he didn't add the blocker Jan 04 11:46:27 AllanonJL|W can we move forward with this instead of arguing? Jan 04 11:46:40 Cardoe latexer: so... after the fact.. someone broke it. Jan 04 11:46:52 Astinus (14:00) <@*slarti> joem: okay, Cardoe unmasked dbus-0.60 which is great and everything, except the current versions of tot$Jan 04 11:46:56 Astinus (14:01) <@*slarti> joem: which means everyone who synced in the last few hours with gnome-base/gnome merged and dbus in US$Jan 04 11:47:00 Astinus (14:01) <@*slarti> joem: which means everyone who synced in the last few hours with gnome-base/gnome merged and dbus in US$Jan 04 11:47:00 * FieldySnuts (n=nikki@tor/session/x-bef6c474ff646570) has joined #gentoo-dev Jan 04 11:47:06 Astinus oops, odd paste issue, but *shrug* that was a complaint Jan 04 11:47:27 Cardoe how do they have an "odd blocker" they can't get around Jan 04 11:47:30 Cardoe it blocks on 0.60 Jan 04 11:47:34 Cardoe so it would install 0.50 Jan 04 11:47:39 Cardoe and keep 0.50 installed Jan 04 11:47:47 solar !meta bitchx Jan 04 11:47:48 jeeves solar: Package: net-irc/bitchx Herd: net-irc Maintainer: net-irc Jan 04 11:47:53 solar !herd net-irc Jan 04 11:47:54 jeeves solar: (net-irc) gregf, swegener Jan 04 11:48:00 solar !seen swegener Jan 04 11:48:02 GenBot swegener was last seen 25 hours, 52 minutes and 36 seconds ago, saying 'brix: the system occasionally hangs itself up. i d$Jan 04 11:48:02 glbt solar, swegener is right here! Jan 04 11:48:09 solar swegener: ping Jan 04 11:48:12 Cardoe slarti: is WRONG Jan 04 11:48:25 Cardoe AllanonJL: joem: so you guys basically took slarti's word... and didn't test it yourselves? Jan 04 11:48:31 Cardoe And that's 1 person complaining Jan 04 11:48:39 Cardoe after at least a dozen developers said it works for them Jan 04 11:48:43 joem Cardoe: no I know for a fact that totem was broke Jan 04 11:48:48 joem because there is no patch applied Jan 04 11:48:49 Cardoe I even asked several arches to test. Jan 04 11:48:50 Astinus (14:02) <@*joem> Cardoe: compnerd`: re ^^ can you not unmask dbus without patching apps that unmasking it will break Jan 04 11:49:05 Cardoe joem: there was a blocker. Jan 04 11:49:07 Cardoe compnerd added it. Jan 04 11:49:13 Cardoe So totem was not broken. Jan 04 11:50:19 Cardoe latexer: so what did I do wrong? Jan 04 11:50:23 Cardoe I provided the proper blockers. Jan 04 11:50:23 joem I consider the blocker to be broken since it can cause up/down issues and there was a patch that could have been easily ap$Jan 04 11:50:26 joem same with other apps Jan 04 11:50:39 Cardoe joem: Well I asked several times in #gentoo-desktop Jan 04 11:50:42 * [ViRgiLiO] has quit (Read error: 104 (Connection reset by peer)) Jan 04 11:50:47 * [ViRgiLiO] (n=virgilio@84-122-101-222.onocable.ono.com) has joined #gentoo-dev Jan 04 11:50:52 Astinus joem: There were more up/down issues caused by unmasking it, then remasking it. Jan 04 11:51:02 Cardoe joem: the blocker DOES NOT cause up/down Jan 04 11:51:07 joem antarus: I know that they both caused issues Jan 04 11:51:09 Cardoe a blocker doesn't let you upgrade to begin with Jan 04 11:51:18 Cardoe joem: well guess what? Jan 04 11:51:23 Cardoe You broke EVERYONE'S systems Jan 04 11:51:26 Cardoe by blocking it Jan 04 11:51:33 Cardoe because now they have to revdep-rebuild AGAIN Jan 04 11:51:35 Cardoe by downgrading Jan 04 11:51:40 Cardoe and then when I unmask it Jan 04 11:51:41 joem Cardoe: I didn't break anything Jan 04 11:51:44 Cardoe they have to revdep-rebuild Jan 04 11:51:52 Cardoe joem: the blocker is the proper way Jan 04 11:52:14 Cardoe >=sys-apps/dbus-0.35 Jan 04 11:52:14 Cardoe !>=sys-apps/dbus-0.60 )" Jan 04 11:52:21 Cardoe The blocker prevents ANY upgrade Jan 04 11:52:26 Cardoe so there's NO up/down issue Jan 04 11:52:37 * Ticho_ (i=ticho@gentoo/developer/ticho) has joined #gentoo-dev Jan 04 11:52:37 * ChanServ gives channel operator status to Ticho_ Jan 04 11:52:41 Cardoe The only up/down issue was for people that unmasked the package locally. Jan 04 11:52:48 Cardoe and then emerged it. Jan 04 11:52:48 * Ticho has quit (Read error: 104 (Connection reset by peer)) Jan 04 11:52:51 Cardoe and then took the mask off. Jan 04 11:52:56 Cardoe because it came out of mask. Jan 04 11:53:05 Cardoe and then they had to downgrade because of the blocker Jan 04 11:53:18 Cardoe And that's not my fault if people unmask packages locally. Jan 04 11:53:36 Cardoe joem: fix your packages. Jan 04 11:53:39 Cardoe I'm unmasking it again. Jan 04 11:54:03 Anarchy *snickers at Cardoe * Jan 04 11:54:31 Cardoe joem: and you're right... Jan 04 11:54:39 joem Cardoe: so your telling me that if I emerge dbus, it won't upgrade to 0.60 and then downgrade when i emerge totem? Jan 04 11:54:43 Cardoe it's my fault that bug wranglers assigned the bug wrong Jan 04 11:54:57 Cardoe joem: You're 100% it will. Jan 04 11:55:04 Cardoe AND THAT'S WHAT ALL PACKAGES in portage do with blockers Jan 04 11:55:10 Cardoe THEY WILL PULL THE PROPER PACKAGE Jan 04 11:55:16 Cardoe if there is a blocker Jan 04 11:55:21 Cardoe look at k3b Jan 04 11:55:27 Cardoe It downgrades you on a couple of packages Jan 04 11:55:35 Cardoe Because it doesn't work with the latest revs Jan 04 11:55:37 Cardoe Now... Jan 04 11:55:42 joem Cardoe: but see thats bad for users...when there are knowingly patches that could fix that which should be added prior Jan 04 11:55:51 joem can we add those before unmasking again so it won't be an issue at all? Jan 04 11:56:13 * zzam has quit ("KVIrc 3.2.0 'Realia'") Jan 04 11:56:14 Cardoe I just told you Jan 04 11:56:16 Cardoe I'm gonna unmask Jan 04 11:56:18 Cardoe so go add them Jan 04 11:56:51 joem see there is that attitude again..you could say "yea I'll wait 20 min so there isn't a sync issue for some users" Jan 04 11:57:33 Cardoe latexer: Did I just unmask them? Jan 04 11:57:33 Cardoe no Jan 04 11:57:37 Cardoe er joem Jan 04 11:57:48 Cardoe am I going to unmask them in the next 3 minutes Jan 04 11:57:49 Cardoe no Jan 04 11:57:53 Cardoe I'm waiting for you to patch your stuff Jan 04 11:57:56 Cardoe so go do it Jan 04 11:58:02 Cardoe before you continue to accuse me of more crap. Jan 04 11:58:06 fox2mike spyderous: ?? Jan 04 11:58:09 joem .. Jan 04 11:58:10 Cardoe because right now.. my system is broken Jan 04 11:58:13 Cardoe thanks to your mask Jan 04 11:58:30 Cardoe my gnome-applet's just all crashed. Jan 04 11:58:48 Cardoe joem: So go patch Jan 04 11:58:50 spyderous 08:08 #gentoo-dev: <@fox2mike> spyderous: where did they go? keithp and friend? Jan 04 11:58:51 Cardoe let me know when you're done. Jan 04 11:58:52 joem Cardoe: yea I really feel for you..do you know off hand of any other pkgs that have the mask and a patch? Jan 04 11:59:05 Cardoe joem: All of them. Jan 04 11:59:11 Cardoe joem: And it's the proper way to do it Jan 04 11:59:14 Cardoe NOT the wrong way Jan 04 11:59:16 joem all of what? Jan 04 11:59:24 Cardoe joem: all of the progs that have issues with dbus Jan 04 11:59:33 agaffney does thunderbird require openldap installed locally for ldap addressbooks to function? Jan 04 11:59:56 joem I don't know what programs have issues Jan 04 12:00:02 joem besides those two Jan 04 12:03:16 Cardoe joem: there are no issues Jan 04 12:03:20 Cardoe there is a proper blocker Jan 04 12:03:38 brainiac-ghost just unmask the damn thing Jan 04 12:04:04 brainiac-ghost it works on all programs i've used
Sounds like a QA issue to me.
(In reply to comment #2) > Sounds like a QA issue to me. I don't think so. As long as a package is actively maintained, everyone else has to ask the maintainer doing non-trivial/non-arch changes. It's a problem that some devs think they can touch the whole tree as they like, instead contacting the maintainer. Doug, you failed communication-wise, too, by not cc'ing the individuals you're angry with, giving them the chance to point out there position.
(In reply to comment #3) > (In reply to comment #2) > > Sounds like a QA issue to me. > > I don't think so. As long as a package is actively maintained, everyone else > has to ask the maintainer doing non-trivial/non-arch changes. It's a problem > that some devs think they can touch the whole tree as they like, instead > contacting the maintainer. Right, and the QA team should be dealing with that.
(In reply to comment #4) > Right, and the QA team should be dealing with that. We have no active QA (at least a formally active one) as far as I'm aware, and no: This is a genuine developer relations (handling problems between people) issue, not a QA (caring about the "product") one.
(In reply to comment #5) > (In reply to comment #4) > > Right, and the QA team should be dealing with that. > > We have no active QA (at least a formally active one) as far as I'm aware, You are not aware. > no: This is a genuine developer relations (handling problems between people) > issue, not a QA (caring about the "product") one. He filed the bug about others masking without contacting him. That clearly is QA.
QA isnt about devs touching packages of other devs
(In reply to comment #7) > QA isnt about devs touching packages of other devs I think it is. I'll let the other devrel members take over, since I don't want to deal with what I consider a QA issue.
QA could (had it not been done already) be involved with tidying up the mess caused by the mask being removed. Dealing with the people who screwed up, however, isn't a QA thing. Unless devrel would like confirmation that there was indeed a screwup...
(In reply to comment #6) > > We have no active QA (at least a formally active one) as far as I'm aware, > > You are not aware. We have an alias, but an active group with strategic goals, procedures, measures, planned, regular quality checks on the tree - all that is missing. So no, there's nothing I'd name QA.
The way this should work is: * Somebody screws up * QA is contacted * QA helps resolve the issue and usually it's left as that However, the touching somebody else's packages is another issue and the ombudsman will get involved.
Copying relevant parties. Any ombudsman feel free to comment or mediate, but I'm inclined to agree with Donnie, at least at first reading. (By the way, ombudsman@gentoo.org seems to be broken, which is why the long list.)
(In reply to comment #8) > (In reply to comment #7) > > QA isnt about devs touching packages of other devs > > I think it is. I'll let the other devrel members take over, since I don't want > to deal with what I consider a QA issue. > My understanding was that QA's job is to be authoritative on the right way to do things with the tree, and to assist in fixing honest mistakes when feasible. If the issue had been that the offending devs had mis-masked dbus and didn't understand that it had caused breakage, then that would be a QA issue, since the QA folks can be quite effective in explaining such errors. People touching other folks packages without permission, on the other hand, is more a devrel issue, it would seem to me, although only if it's a repeated problem and not a one-off error.
Back to the bug at hand, rather than a meta-debate on the nature of this kind of thing... Here's roughly what happened, from my point of view: - Cardoe unmasked dbus-0.60 (which was masked on account of packages not yet supporting the API changes). To be fair to Cardoe, there *were* patches in bugzilla with fixes to support the upgrade, but some (e.g. Totem) weren't applied. It's also true that a blocker on dbus-0.60 was added to the totem ebuilds. - I synced pretty much straight after this happened, by chance. When upgrading, I then had a problem -- the version of totem depended on by the gnome-base/gnome metapackage blocks dbus-0.60. So, when emerge -uDN'ing world, there is a blocker that can't be resolved without a) masking dbus-0.60 or b) unmerging gnome-base/gnome and totem, which is not really a possibility. I pinged the gnome herd about this and told them about (Astinus' paste of my complaint is truncated, I'll try to attach logs). So no, I don't think i was "WRONG", Cardoe.
I second Grant's opinion. The problem here was just a lack of communication between devs. Someone should have made better efforts to contact Cardoe before dbus was masked to work out whether it was best to mask dbus, block the dbus upgrade in totem, or what ever. At the very least someone should have told Cardoe that they had to mask dbus and exactly what the issue was. Nothing but a vague changelog entry "undo Cardoe's unmask due to possible unresolved issues" doesn't help very much.
slarti, you are still wrong. Because the tested provider was that the blocker resulted in people pulling in dbus-0.50. Which was tested. The only way we could duplicate any issue you had was to use -u and not -uD, which without deep depend checking resulted in the scope of the blocker not being considered in the full upgrade which caused people to upgrade to 0.60 and then subsequent emerges to get a blocker and be forced to downgrade. The solution provided was the best one available, and is still the best one available. Which was to block and only depend on dbus-0.50 which resulted in people only getting the new dbus if their emerge -auDv world did not have any packages which were incompatible installed. The resulting remask caused up/down issues and unresolved dependancy issues for the community at large because now people hard emerged updates to packages that were only working with dbus-0.60 so when a user now ran "emerge -auDv world", they got an error that there were no packages to satisfy a dependacy and now it was package.mask'd. Or if they were lucky and didn't have any of those packages, anything that called any of the DBUS api would crash, (just like gnome-applet's was in a repetitive crashing cycle for me) because of the changes to the API. So the re-masking was the worst part of the storm. In addition, since I'm the dbus maintainer, it would have been nice to have been notified. Especially a bug filed. That's my main issue, the fact that joem made some comments into #-dev while I was gone for the evening does not constitute as a notification. The proper avenue for this is would have been a bug report, even if you did feel the need to mask it. Some way to notify the maintainer should have been attempted. In addition to all this, dbus-0.60 contains several important fixes required for emerging on a current Gentoo ~ARCH tagged system. Which resulted in a flurry of re-opened bugs for issues that were fixed by the 0.60 unmask. Now while joem & AllanonJL complained in IRC that they were not provided with info, the Gentopia herd or users that unmasked the ebuilds themselves did file bugs in Bugzilla about any applications that had issues with. Patches were either rewritten by people and attached or retreived straight from upstream and attached. In fact, some patches were written and later updated from upstream. I had even notified both joem & AllanonJL of any incompatiblities and gave them the link to the Tracker bug that was created for any dbus-0.60 issues. If they don't remember it's the conversation I had with them when I discussed that Gnome upstream has supported the new dbus API in the 2.12.2 which is why I'm pushing dbus 0.60 to be part of the stablization for Gnome, since well... Gnome itself needs it. It was also when I informed them of the issue that gnome-applets doesn't compile because missing modular X (xrdb) and libnotify depends (just to pinpoint 1 specific conversation). The fact that the bugs were assigned to the wrong individual by bug wranglers is not my fault nor is it really anyones. Bug wranglers assigns by the best possible information they had. The Gnome herd was also party to the Tracker and to the bugs that concerned Totem. Regardless of all of this, my issue is that had someone not complained about the remasking to me in IRC, I wouldn't have had a clue. File a bug report if there's a bug, no matter what. That's my issue.
Ok here is my issue since I'm cc'd and in the title: 1. I didn't remask the package but I endorsed the decision to do so(as did others) and don't think it should have been unmasked in the first place. 2.The patches in bugzilla for dbus-0.60 related issues weren't assigned to me or correct maintainers. Gentopia herd had a tracker bug with plenty of blockers that weren't resolved before unmasking dbus. Cardoe is wrong and Gnome herd was not assigned to the tracker bug (bug 114463). Nor were we on the totem bug. They could have been more responsible about this. It took me 10 minutes to apply patches after all this hit the fan today and the bugs were assigned to the right place. This would also have fixed the complaint mentioned in comment #15 Either way sorry for whatever problems were caused by bad communication.
OK, there's been a feud of sorts for a while between gentopia and gnome. I've been hoping for some sort of self-organising phenomenon between the two, but this seems to have brought everything to a head. That this bug is filed is a good thing, because hopefully we can draw up some way for both teams to co-exist. The gnome team has historically (at least since foser came on board and built his team) been the highest QA team in gentoo. Gentopia has had the noble goal of tracking project utopia in gentoo (and the stuff is fairly slick). These two goals are not mutually exclusive. But, as joem & others have noted, there's a definite lack of communication. I'd like a chance to talk with both Cardoe & the Gentopia team and AllanonJL, Joem & the gnome team individually and together over the next day or two. I've pinged both Cardoe and AllanonJL in cvs to that effect. Will report findings. I request devrel to not take any action against either party at this time.
(In reply to comment #16) > slarti, you are still wrong. Because the tested provider was that the blocker > resulted in people pulling in dbus-0.50. Which was tested. The only way we > could duplicate any issue you had was to use -u and not -uD, which without deep > depend checking resulted in the scope of the blocker not being considered in > the full upgrade which caused people to upgrade to 0.60 and then subsequent > emerges to get a blocker and be forced to downgrade. I did use -D -- I only ever call back the same upgrade command from shell history, and I've got no mention of -u without -D in my history either. When you tested the blocker, did you test it with the gnome metapackage or gnome-light with totem installed as part of world? In the latter case, yes, it would have worked.
I said I pinged them in cvs, which is of course impossible to do. I meant irc.
(In reply to comment #18) > > I request devrel to not take any action against either party at this time. > You have control.
I believe the parties involved in this have now come to a settlement... closing bug. Reopen if new issues arise or further clarification or mediation is required.
