=== bjf is now known as bjf[afk] === devildante is now known as devildante[afk] [00:42] kenvandine, around? === devildante[afk] is now known as devildante [01:16] RAOF - i'm not sure if you saw my earlier message in the scrollback [01:17] chrisccoulson: I did not - I only saw asac wanting me to be awake at 2:30am. [01:17] RAOF - ah, ok. it was a comment about bug 623700 [01:17] Launchpad bug 623700 in gnome-session (Ubuntu) (and 1 other project) "Edge condition in idle monitor (affects: 1) (heat: 8)" [Undecided,New] https://launchpad.net/bugs/623700 [01:18] i've not looked at it in that much detail, but i fixed what looks like a similar issue in gnome-power-manager earlier in the year [01:18] and it was actually a slightly different issue [01:18] what i found is that g-p-m was missing the negative alarm altogether [01:18] Ah, yes. I noticed that gnome-power-manager was setting appropriate transition thresholds. [01:19] so, what was happening is: [01:19] 1) Positive alarm triggers [01:19] 2) g-p-m updated it's idle state, then [01:19] Note that I've also fixed an xserver bug in the SYNC extension, which might be what you're about to describe :) [01:19] 3) presumably it got pre-empted by something else, and then [01:20] 4) by the time it ran again to set up the negative alarm, it has already fired [01:20] and so it thought the session was idle indefinately [01:20] Sweet! [01:20] That wasn't what I'd fixed in the xserver. [01:20] RAOF, http://git.gnome.org/browse/gnome-power-manager/commit/?id=0c171b962989d20b55f65ae016c584d3543f4801 [01:20] that's the issue i fixed last cycle [01:22] Why didn't you just set the negative transition trigger at the same time that you set the positive transition trigger? [01:24] RAOF - the original patch i submitted did that, but we had to back it out for some reason, and i'm trying to remember why [01:24] Rather than wait until the +ve transition had fired before setting up the -ve transition. [01:24] You might have hit the xserver bug where if you had a +ve and -ve transition with the same threshold there was a good chance you'd never get the -ve transition? [01:25] i'll try and remember what it broke ;) [01:27] i'm sure it was something weird like, if we set the -ve transition whilst the count was already lower than the threshold, then it ignored it, and never triggered at all [01:27] something weird like that [01:27] but my memory is sketchy now ;) [01:29] chrisccoulson: So, what I fixed in xserver was the following: if you had a -ve transition with the same threshold as the +ve transition, then if you caught the signal at *exactly* the threshold then the -ve transition wouldn't fire because X wouldn't notice IDLETIME going *above* the threshold. [02:40] If Michael Vogt or Sebastian Heinlein, who I think are the maintainers of gdebi are here, or if someone could forward it, I have a suggestion. Or perhaps there is a better place to ask? [02:41] My suggestion is this. Add the text "\n Do you trust the source of this package?" to the dialog which appears for confirmation when installing an unsigned package. [02:41] This would make it clear to new users in particular the nature of the confirmation. [03:08] RAOF, do you have multiple languages installed? [03:08] robert_ancell: Only multiple EN variants. [03:08] Why? [03:09] RAOF, I'm trying to text my gettext/polkit changes. But I didn't test if the translations worked before at all. Could you install language-pack-de-base and run software center and try and install a new package. Check if the dialog that asks for your password is in german. [03:10] actually, it might not honour the language because the polkit agent is started on login... [03:10] I need to restart to boot the new kernel, anyway. [03:10] RAOF, cancel that, I'm going to log in as full german and see if it works [03:10] brb [04:49] How do i persist changes to group.conf ? The changes I make for the ltsp users seem to be lost every now and then. ... may be whenever i run apt-get or on reboot [07:54] pitti: hey, are we frozen yet? === almaisan-away is now known as al-maisan [08:04] Can we not be frozen yet? :) [08:05] Or, more to the point, can I get a sponsor for another -ati upload to fix another DRI2 xserver crash? [08:07] * hyperair wonders how many hoops there are to jump through in order to get a new gpod into ubuntu at this point. [08:16] morning everyone [08:16] morning [08:16] hey kiwinote [08:16] hi glatzor [08:17] hi mvo, mpt, vish [08:19] morning and471 [08:24] mail [08:24] hey [08:24] * mvo waves to and471 and glatzor and kiwinote [08:24] hi mvo [08:24] mvo, did you have any ideas about the software-properties thing? [08:24] and471: the source.list finding ? [08:25] mvo, yup, going from a sourcelistentry to the display_name [08:26] and471: hi! :) [08:27] mvo: could we rename update manager this cycle? :) [08:27] :) [08:28] last day! ;p [08:37] Good morning [08:38] Howdie pitti! [08:38] hey RAOF, how are you? [08:38] hi pitti [08:38] A little cold, actually :) [08:38] RAOF: I wanted to talk to you about this apport/intel work item [08:38] I should go and turn on the heater. [08:38] hey ajmitch [08:38] RAOF: wow, in Australia! [08:39] In my defense, it's the southern bit of Australia in winter :P [08:39] asac: seems it built by now [08:39] it is tasmania after all, right? [08:39] We even have snow! [08:39] (Although not right now) [08:39] mvo: frozen> no idea, but I don't think so [08:39] RAOF: still need -ati sponsoring? [08:39] pitti: Yes please. [08:40] http://cooperteam.net/Packages/xserver-xorg-video-ati_6.13.1-1ubuntu4.dsc [08:40] With bonus changes in the same dir. [08:40] mvo, I am going and coming back in about 4 hours, do you think you could have the software-properties thing working (or glatzor ?) [08:40] mvo, glatzor, so then I can implement it in software-properties? [08:40] mpt: hi! devildante's latest changes were at your request / approved by you? [08:41] pitti: Yeah, I saw your query re: the apport/intel thing. It's fine to defer that to Natty; our existing GPU hang infrastructure is perfectly acceptable. [08:41] mvo: did you change anything in the addons branch before merging or not? (just so I know if I need to update things in my branch) [08:42] RAOF: I'm interested in how much work that would actually be [08:42] kiwinote: please update, I don't think I changed much other than resolving conflicts [08:43] RAOF: perhaps you have some example bug reports where the apport info is good and bad? i. e. how to tell them apart? [08:44] RAOF: so we'd somehow need to track that we added drm.debug to grub, and delete it again once we get the next hang? [08:44] pitti: yeah [08:44] thanks [08:44] pitti: Yeah, that's about it. [08:45] It's pretty easy to tell you've added drm.debug to grub; just parse the kernel command-line. [08:45] sure [08:45] Although I guess there are all sorts of edge cases possible where users have set it themselves. [08:46] RAOF: well, in that case we wouldn't touch it, I guess [08:47] hey [08:47] RAOF: so, perhaps you can add links to a good and a bad bug, and I'll have a look? [08:47] bonjour seb128 [08:47] pitti: I'll see if I can find a pair. [08:47] hello pitti [08:58] mvo, I am going and coming back in about 4 hours, do you think you could have the sources.list thing working so then I can implement it (or glatzor ?) [09:13] mvo, bah, gotta go now, see ya [09:20] vish: hello, sorry for my late reply. renaming the whole thing is a lot of churn. it will also break translaitons (not a big deal though). I would prefere to wait for N because we may get a complettely new one then [09:22] hi! Just managed to track down the bug that was bothering me last night, so I'll just finish tidying up all the print statements, make a few last adjustments, do some more testing and then merge devildante's branch with his latest changes. Sorry for the delay ;) [09:22] mvo: ^ [09:22] kiwinote: sweet! [09:23] mvo: the only remaining biggish thing is that we get three transaction-finished signals, so the pane refreshes three times. (but that behaviour seemed to be in trunk before the addons merge as well) [09:24] mvo: np.. cool :) [09:24] kiwinote: yeah, that is quite possible (and anoying) [09:25] mpt: good morning. what is wrong with "Find it in the menu at"? I don't mind the exact words, but mentioning something about menu or anything like this seems like a win, no? [09:41] hm, I guess "menu" is not great as it might be confused with the menubar [09:45] Good morning mvo [09:46] mvo, does that code make a distinction between gnome-panel and Unity? I didn't see one, but then I didn't really know what to look for [09:47] Good morning kiwinote, yes, all requested by me [09:47] mpt: no, I don't think the unity feature we need is implemented, we could (relatively) easy hide the thing if unity is running [09:47] mpt: thanks, and good morning ;) [09:48] njpatel, hi, did the DX team get time to implement an API that either (a) reveals a particular launcher in Unity or (b) tells you where to find a particular launcher in Unity? [09:48] I think we should aim for navigation in the addons, i.e. clicking on the pkgname (as it was before) [09:49] mpt, heh, no :) [09:49] it breaks history currently, but that is a bug that needs attention, I think its massively more useful with that [09:50] mpt: nzmm did some tweaks to the where-is-it-installed bits, you may want to check trunk [09:50] seb128, hey, I keep missing you guys. Should we update evolution? (was going to ask didrocks in case it affected UNE) [09:51] mvo, I'm happy with re-adding that navigation, but as you say, it would increase the importance and scope of bug 426999 [09:51] Launchpad bug 426999 in software-center (Ubuntu) "Navigating to item via search doesn't show its department/subsection (affects: 1) (dups: 1) (heat: 15)" [Medium,Triaged] https://launchpad.net/bugs/426999 [09:52] robert_ancell, hey [09:52] robert_ancell, I'm not sure, to 2.30.3 at least yes [09:52] robert_ancell, btw your polkit change, can we update cdbs to add the domain? [09:52] the way it's doing for .desktop and .server [09:53] ie some sed rules to update the .policy at build time? [09:53] mpt: right, my feeling is that usefulness(navigation+426999) > usefulness(no-navigation) (but of course fixing that bug as well is the ideal state) [09:53] hey ... so the ALT key was fixed in gnome-terminal ... now paste doesnt work anymore :( ... ctrl+shift+v [09:53] robert_ancell, good morning, are you still maintaining the "Featured" list for USC? [09:53] seb128, I replied to that email. I assumed we can't do that easily? But if it works for .desktop then sure [09:53] mpt, no, I think mvo is doing that [09:53] robert_ancell, what email? [09:53] seb128, from David [09:53] robert_ancell, I just read yours about adding a .in [09:53] lol [09:53] robert_ancell, oh ok, I didn't see those yet [09:53] lets say we joint maintain it ;) [09:54] robert_ancell, see /usr/share/cdbs/1/rules/langpack.mk [09:54] seb128, I haven't looked at how people are using .policy in their build systems, but it's going to need intltool integration right? [09:54] robert_ancell, evo to 2.30.3 yes [09:54] robert_ancell, otherwise updates to do: gnome-bt, anjuta [09:54] robert_ancell, brasero as well [09:55] seb128, ok, I'll look at it tomorrow (evo and the polkit stuff) [09:55] robert_ancell, I though your patch would use gettext if the the domain is there? [09:55] mvo, mpt, ok, joint maintaining it is. Is this about Unity? [09:55] seb128, yeah it does [09:55] robert_ancell, ok, so we can just sed it [09:56] robert_ancell, see /usr/share/cdbs/1/rules/langpack.mk [09:56] robert_ancell, yeah, the DX team were asking me yesterday about including some sort of Unity intro as a featured item [09:56] robert_ancell, there is a regexp grepping the makefile to get the domain [09:56] robert_ancell, that should work for most of the GNOME softwares [09:57] I have no strong opinion either way, but the spec at https://wiki.ubuntu.com/DesktopTeam/Lucid/FeaturedApps looks like its not really qualified [09:57] mpt, I'm not oposed in particular. It feels to me like featured apps is moving towards an iOS/Android experience where there is not a fixed list so much as advertisements in the software center [09:58] "intro" doesn't seem qualified [09:58] or at least is not an application [09:58] some days I hate bzr [09:58] I'm not convinced this is the right vehicle to encourage people to try a new desktop experience - the method of switching to it is non-trivial [09:59] it just needs to download for half an hour to tell you that there is no commong ancestor with what you try to merge on [09:59] mpt, how do you feel about it? [10:00] robert_ancell, I was skittish about including the vanilla Unity package, because unlike anything else in Featured, it dramatically changes your environment [10:00] robert_ancell, I suggested having some sort of interface for switching in and out. Though of course it's a bit late for that. [10:01] mpt, yeah, I think that would be a requirement [10:01] mpt, does OMG not drive enough people to try it? [10:02] robert_ancell, mpt: is that introduction a maverick thing? [10:02] robert_ancell, depends what the goal is. If the goal is reporting bugs, then I'd guess they do. :-) (But I don't subscribe to those bug reports so I don't really know) [10:02] robert_ancell, mpt: that's a feature and feature freeze was 2 weeks ago [10:02] seb128, exactly [10:02] there is no way we do that this cycle [10:02] don't start wasting time on it [10:03] robert_ancell only has 1.5 week of work left this cycle [10:03] or 2 weeks [10:03] then holidays and we will be hard frozen by the time he's back [10:04] robert_ancell, btw quick question about shotwell, the remaining bugs on the specs, should they be dropped? [10:04] robert_ancell, I didn't try but what happens on first run? does it auto import your xdg dir? [10:04] seb128, don't worry, I'm not planning on working on it [10:04] ok ;-) [10:05] robert_ancell, that was one of the bugs on the spec [10:05] seb128, yeah, those bugs aren't going to be implemented in time, but they're not blockers [10:05] seb128, it offers to import from F-Spot or ~/Photos if I remember correctly. I'm happy with the current experience [10:06] ok [10:06] robert_ancell, I will just drop the bugs from the blueprint [10:06] so they stop showing up on the workitems list [10:06] import f-spot or photos seems fine [10:07] seb128, ok, can you add the list into the blueprint text if it is not? [10:07] seb128, do you know about bug 529696 - there's a lot of noise there [10:07] robert_ancell, I will [10:07] Launchpad bug 529696 in brasero (Ubuntu) (and 1 other project) "brasero can't copy audio cd (useless error message) (affects: 140) (dups: 11) (heat: 703)" [Medium,Confirmed] https://launchpad.net/bugs/529696 [10:08] mpt: any thoughts on the "Find it in " string ? to include something like "menu", "main menu", "panel" or anything like this? I'm fine using the version you put forward if you think its too difficult to communicate [10:08] robert_ancell, hum [10:09] robert_ancell, no, I know about bug #514782 [10:09] Launchpad bug 514782 in brasero (Ubuntu) (and 2 other projects) "Brasero fails to recognize erased DVD-RW media (dup-of: 581925)" [Medium,Incomplete] https://launchpad.net/bugs/514782 [10:09] Launchpad bug 581925 in udev (Ubuntu Maverick) (and 2 other projects) "cdrom_id: Does not recognize reblanked DVD+RW, DVD-RW and DVD-RAM (affects: 22) (dups: 4) (heat: 128)" [Medium,Fix released] https://launchpad.net/bugs/581925 [10:09] but that doesn't seem the same issue [10:09] robert_ancell, we should have a try to backport if it that has been fixed upstream [10:09] robert_ancell, btw how busy are you? you seem to merge again things we merged not so long ago [10:10] doesn't seem really useful work right now [10:10] robert_ancell, I'm not sure how you read your bug emails, just some random tasks for you if you are out of those [10:10] bug #623816 [10:10] Launchpad bug 623816 in rhythmbox (Ubuntu) "rhythmbox assert failure: rhythmbox: client.c:613: avahi_client_free: Assertion `client' failed. (affects: 1) (heat: 6)" [Low,Incomplete] https://launchpad.net/bugs/623816 [10:10] seems a libdmapsharing issue [10:11] robert_ancell, kees had a comment on the bug asking to turn on the testsuite for the lib as well at build time [10:11] seb128, not particularly busy. Is there anything in particular to look at. I'm not spotting a lot of bugs, e.g. https://launchpad.net/~desktop-bugs/+bugs?field.milestone%3Alist=21439 [10:11] robert_ancell, I will start adding things to the list [10:11] robert_ancell, otherwise there is the evo 2.20.3 updates [10:11] brasero to 2.31 [10:11] anjuta to 2.31 [10:11] yup, I'll do the updates [10:11] gnome-bt to 2.31 [10:11] so feel free to pick in those tomorrow [10:12] mvo, I didn't think it was difficult to communicate, I was just trying to make it shorter. :-) But I'll revert it since it looks wrong to you [10:12] I will start building a list of bugs as well today or tomorrow [10:13] vuntz - your gnome 2.31 script is tracking the GNOME 3 version of gnome-bluetooth [10:15] mpt: well, not wrong. I was just trying to make it easier for the user. I think we have some extra space there. like I said, no strong opinion, just trying to help a newbie that may need a little "push" to understand what "Applications" is meant [10:16] mpt: but either way is fine with me ("menu" may not be ideal because of the menubar, but I don't know if there is a good word for a "launcher menu" that is not geek-speek [10:16] ) [10:16] seb128, gtg, anything else you want to raise? [10:16] robert_ancell, one sec [10:17] robert_ancell, to reply to your polkit email [10:17] robert_ancell, the .policy are already in the POTFILES [10:17] seb128, oh also, I don't know if I'll be successful with the cheese/gst problem. I'm still spending some time blundering around, not solution yet. [10:17] they need to translate them to build the .policy at build time [10:17] seb128, oh, nice [10:17] ie they already have the translations [10:17] they just put them at build time in the xml [10:18] so we just need to add the sed magic to cdbs [10:18] as we do for desktop and server [10:18] .desktop .server I mean [10:18] seb128, so the translations are stored both in the xml and in .mo files? [10:18] yes [10:18] works for us :) [10:18] indeed ;-) [10:18] that's it [10:18] robert_ancell, oh [10:18] ok, should be easy [10:18] robert_ancell, in case you didn't see it be have a p.u.c ~platform [10:19] seb128, yes, thanks! [10:19] robert_ancell, http://people.canonical.com/~platform/desktop/versions.html [10:19] np [10:19] robert_ancell, so you should be able to do changes now [10:19] or add things [10:19] details are on the platform list [10:19] robert_ancell, that's it I think [10:19] need to update the irc channel title [10:19] yeah, I will do that today [10:19] I wanted to make sure the cronjob works before [10:20] seems it didn't update so I will check that today [10:20] seb128, oh, do you know if we will/should update dbus? [10:20] to 1.3? [10:20] yes [10:20] is that an unstable serie? [10:20] seems a bit late to me [10:20] pitti, ^ [10:21] robert_ancell, I would default to no [10:21] but I can check with Keybuk and pitti [10:21] I will let you know [10:21] ah, checked. it's debian experimental so versions is recommending it [10:21] how come some lines have 2 upstream versions? [10:21] but I suspect we shouldn't. I should probably update versions to handle that [10:22] right [10:22] I was going to say [10:22] let's focus on the few I listed before [10:22] and then on bug fixing [10:22] I will build a buglist for eow [10:22] seb128, they're the packages that are tracking GNOME 2.30, but it shows that there is a 2.31 package available [10:22] so next week we can spend time on those [10:22] robert_ancell, ok [10:22] the list has been getting shorter since GNOME 3 was delayed [10:23] I need to figure why nautilus-sendto tracks 2.90 [10:23] I think it's just evo, gnome-panel, gdm, that should be help back [10:23] http://people.gnome.org/~vuntz/tmp/versions/versions-2.32 doesn't [10:23] oh, that is weird [10:23] gnome-control-center gnome-settings-daemon gnome-applets [10:23] as well [10:24] though not sure about the 2 first ones [10:24] me neither [10:24] applets because they are tied to the panel right? [10:24] g-s-d has the polkit wrapper for the timezone [10:24] yes [10:24] our gnome-panel would not use the g-s-d one [10:24] I guess we could do the update but just not ship that wrapper for this cycle [10:24] seb128: I'm not sure; I don't usually follow dbus development [10:25] pitti, ok [10:25] robert_ancell, I think that's it from me, I will drop you an email if I think of anything else [10:25] later [10:25] bye [10:25] robert_ancell, I think you should have enough to do tomorrow with the few updates listed before and the polkit gettext domain [10:25] Au revoir [10:53] kiwinote: I work on your branch now (just fyi) [10:54] mvo: just pushed another revision [10:54] mvo: there are two remaining issues (also in trunk): - [10:55] mvo: -cancelling a policykit alert restores the wrong installation state [10:56] mvo: - after installing/removing the main app, any subsequent addon checkbox toggling doesn't display the addonsbar [10:56] kiwinote: did you figure out why recommends that are not len() == 1 are discarded? that does not seem to make a lot of sense to me [10:56] kiwinote: thanks for the info on the open bugs, are you looking into them or do you think we should put them into LP to keep track? [10:57] mvo: no, haven't investigated, but it may have been in the spec? [10:58] kiwinote: right, but why is it there? [10:58] mvo: the current trunk also doesn't allow install/remove via menu or listview, but have fixed that in my branch [10:58] mpt: do you happen to know why reocommends are discarded if there is more than one recommends? [10:58] mvo: dunno, mpt ^^ ? [10:58] dpm, hi [10:58] dpm, you might want to read the channel backlog [10:58] mvo, no [10:58] kiwinote: yeah, I noticed that fix [10:58] dpm, but what we said was basically what you wrote in your email [10:59] mpt: thanks, I will see if I can work with devildante on figuring it out [10:59] dpm, ie the potfiles.in and mo already have the translations and langpack.mk the magic to get the domain [10:59] kiwinote: it seems that its rather odd, either we should have them all or just ignore all recommends (as they are installed anyway by default) [10:59] mvo: indeed [10:59] kiwinote: that will simplify the implementation as well (ignoring recommends) [11:01] mvo: that sounds good, perhaps it may be good to check for any regressions though [11:01] kiwinote: yeah [11:02] kiwinote: there are quite a lot of packages that qualify for this criteria [11:02] mvo: can we tweak addon choosing after ui freeze? [11:02] ~2000 one if my count if correct [11:02] yeah, its bugfixing [11:03] I mean, if there are obvious bad ones in the list, that is a bug [11:03] mvo: ok, then if you leave that recommended check there for the meantime, then tomorrow/next week I'll investigate in to getting better addon lists [11:04] kiwinote: ok [11:05] mvo: although it doesn't really matter, so if you want to take it out now, feel free ;) [11:05] ok again ;) [11:05] I play with it a bit [11:14] hi seb128, ok, gotcha, I've just read the backlog [11:15] (re: policy files) [11:15] dpm, thanks for following on that [11:15] * dpm hugs seb128, and would hug robert_ancell if he were around [11:15] * seb128 hugs dpm [11:16] chrisccoulson: hey [11:16] chrisccoulson: don't worry about firefox updates [11:16] we got the extra builders [11:16] hi seb128 [11:16] that's good :) [11:22] kiwinote: hm, hm, I thin kthe whole alogrithm will need some serious re-thinking, I would not be supprised if we just go with enhances and suggests, its making rather odd suggests (synatic got a kdebase-bin recommends in some obscure way) [11:23] mvo: yeah, I had been thinking about enhances and suggests and then determine the other addons based on pkgname, ie so that randompkg has randompkg-doc as an addon etc [11:24] kiwinote: yeah, I play some more and add some test code/formal tests so that we have a baseline [11:25] mvo: great, thanks! [11:26] hmmmm, i seriously need to invest in some more RAM :/ [11:28] ram++ [11:29] to build firefox? [11:34] hello. can anyone tell me how to customize the login window in ubuntu lucic ? [11:35] mvo, kiwinote: Recommends should be showing up only if the package your're looking it is the *only* package in your sources that Recommends it. [11:36] mvo: just managed to fix those two bugs mentioned above [11:37] mvo: so that means that I'm happy with it going into trunk and being released whenever you are [11:38] kiwinote: nice, thanks [11:38] mvo, I have a bunch of small UI fixes here: https://code.launchpad.net/~mpt/software-center/fit-and-finish-2/+merge/33751 [11:39] mpt: I do you have examples so that we can add tests? the current code does not seem to do that (probably a bug) [11:40] kiwinote: thanks, merged into my addons branch [11:40] any help on customizing the login windows ? [11:41] kiwinote: you may want to merge back, some tweaks and tests, nothing much yet [11:41] mvo: yep, they looked fine [11:41] mvo, I don't know how to query the apt-cache for things like that ("Which packages recommend this package") [11:41] clear [11:42] mpt: thats fine, just give a hypothetical example [11:42] mpt: I'm writing tests for this stuff now and I need to be sure I understand the rational behind the rules and also figure out if/what packages apply for this [11:44] * mvo is off for lunch but reads backlog [11:52] mvo, come to think of it, I don't have a good example. I was treating Recommends like Depends there, assuming that they won't be interesting unless they're unique. But maybe that's not a good rule. [11:53] mvo, for example, synfigstudio Recommends synfig-examples, so it shows up as an add-on. synfig itself Suggests synfig-examples, but if it Recommended synfig-examples instead, under the current rule synfig-examples would show up as an add-on for neither of them. Which probably doesn't make sense. [11:54] kmess currently Recommends konqueror, but I assume that's just a bug in kmess. [11:55] * mpt finds a bug in the size calculation [12:10] mpt: ok, thanks. so it sounds like its not too bad if the current implementation does not quite do what you initially speced, right? if you come up with a good example I'm happy to have a closer look, but the current code does not implement it, instead it discards everything that has more than one recommends or suggests (which does not make that much sense and is probably a oversight by devildante) [12:11] mvo, it's a live spec. If we find something that doesn't make sense, we should change it. :-) [12:11] mvo, so, we should show all Recommends? [12:13] mpt: I have not fully made up my mind yet, but it feels to me like we should not show recommends at all because they are installed anyway by default and if the user removes a recommend the app may not work that well as before [12:13] mpt: its a strong relationship than "addon". but I'm still researching it currently, this is so far just a feeling [12:19] mvo, Recommends are checked by default. If we don't show Recommends at all, and installing installs Recommends by default, the end result is that Recommends is just another name for Depends, no? [12:23] mpt: from software-centers POV yes. the wording in the policy is that they should be installed in all but the most unusal sittuations. its a policy decision from us if we want to allow removal too easily or not [12:25] mpt: I think we need to play a bit with it, from the code POV its not much difference [12:26] mvo, we should err on the side of false negatives (not showing things that we should show), rather than false positives (showing things that we shouldn't). So if you think we're better without Recommends for now, I'm quite happy with that [12:26] hi devildante :-) [12:26] hi mpt :) [12:26] mpt: thanks, I agree with that. I work on it in my branch and we can test both ways [12:27] hey devildante! [12:27] hi mvo ;) [12:27] oh, and I should hold up a sign saying "bugfix mode" from now on :) [12:27] "BUGFIX MODE" even :P [12:30] devildante: hi! [12:31] hi kiwinote :) [12:31] devildante: bug 624578 ;) [12:31] Launchpad bug 624578 in software-center (Ubuntu) ""Total size" is smaller with add-on selected than without (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/624578 [12:31] * devildante is happy he's working with good guys [12:33] mmnz_ has left the building [12:34] kiwinote, mpt: problem is I don't even see kipi-plugins in gwenview page :p [12:34] only see kipi-plugins-dbgsym because I enabled ddebs [12:38] devildante: true, by the looks of it I have fixed that specific addon in my branch actually [12:39] devildante: it may be worth trying to find another app though to see if you can reproduce the bug there though [12:40] kiwinote: okay, but later, I'm trying to solve a problem mpt gave me :p [12:40] devildante: ah, that's fine then ;) [12:40] devildante: by what path is kipi-plugins selected for gwenview? [12:41] mvo, kipi-plugins Enhances: gwenvie [12:41] w [12:42] I think we filter it out correctly now [12:42] kiwinote, filter it out? Why? === al-maisan is now known as almaisan-away === MacSlow is now known as MacSlow|lunch [12:42] it shouldn't be filtered [12:43] problem is only kipi-plugins-dbgsym is shown for me [12:43] devildante, mpt: ah, ok, then our filtering rules are too strict atm [12:43] and the only common point between it and kipi-plugins is: Enhances: digikam, gwenview, kphotoalbum, showimg [12:44] In my branch I didn't adjust any filtering rules, I only fixed some cases where the filtering rules weren't applied properly [12:44] kiwinote, first, if we can't trust Enhances:, the whole add-ons feature is probably a lost cause. Second, the spec should keep up to date with the implementation so that there's somewhere packagers (or people writing instructions for packagers) can learn the rules. [12:45] So if you change the filtering please let me know, and I'll update the spec. :-) [12:45] I think currently its a bit of a work-in-progress (the implementation just like the spec) [12:45] the current implementaion does not quite implement the spec either [12:46] ok, what do I need to change? [12:48] seb128, could perhaps someone from the desktop team look at bug 616856 and bug 613081? I think they might be easy ones to fix, but quite visible in localized systems. [12:48]