[00:00] bla [00:00] yeah, most likely :( [00:01] meh [00:01] just when I had an easy bug on my assigned list :( [00:02] JontheEchidna++ [00:02] beautiful code :) [00:02] :) [00:02] awesome to read this after days of ugly codes [00:02] which class are you admiring, if I may ask? [00:03] installevent [00:03] not the largest of classes though ;) [00:03] just beautiful to look at [00:03] * apachelogger doenst dare looking at the hook ^^ [00:04] heh [00:04] hook got better, thanks to amachai [00:04] inocnsistency! [00:04] JontheEchidna: hook sometimes puts oneline if's into brackets and sometimes not [00:04] whiles are all without breakets [00:05] * apachelogger notes that in his opinion it is much easier to read with brackets [00:05] hook is also the only class to have 3 authors :P [00:06] JontheEchidna: some emtpy lines in the large if monsters would probably help wonders [00:06] JontheEchidna: take a look at astyle [00:06] wonderful tool [00:06] they use it on kdelibs [00:06] you can find the kdelibs options somewhere in techbase I think [00:06] * apachelogger will use that on the date&time kcm soon [00:07] I think what we're using is pretty close to what amarok's doing, but amarok does that spaces-inside-parenthesis thing, and we don't [00:07] because that also got like 3 authors, but like 5 different code styles ;) [00:07] JontheEchidna: how about specing a kubuntu code style guide? [00:08] for UDS? [00:08] yep [00:08] or for maverick in general? [00:08] kk [00:08] well [00:08] UDS == UDS + general [00:08] I'm not quite sure how to get a blueprint to be a "UDS-official" spec [00:08] and we really should brainstorm on getting rid of the installevent stuff and move it to kde(libs) [00:09] JontheEchidna: mark it for that sprint [00:09] and get it discussed by someone ;) [00:09] how do I get it on the schedule, I suppose is the question [00:10] JontheEchidna: I think there is a wiki page on blueprinting [00:10] IMHO blueprinting is way too much limited on usage in UDS context though [00:10] then again the lp UI is so crappy you wouldnt want to use it every day ;) [00:15] omg [00:15] the screensaver configuration stuff dates back to 1996 [00:17] that means it was written in KDE2 and forgotten [00:18] no, KDE 1!!! [00:18] and then got ported along :) [00:18] old != bad [00:18] probably ported with qt1to2 [00:18] then promptly forgotten ;) [00:18] no [00:19] knetworkconf is an example of that happening with qt3to4 [00:19] well, a bad example [00:19] TBH the screensaver kcm is a good example of how code can be maintained for that long [00:19] some compability libs are used [00:20] but other than that it doesnt look THAT bad [00:20] oh [00:20] our mm patches are still bugged [00:21] * apachelogger totally forgot about that [00:22] * apachelogger loves it how the systray goes bonkers when you restart kded ^^ [00:35] * apachelogger thinks the installevent hates him [00:50] apachelogger: https://wiki.kubuntu.org/Kubuntu/Specs/MaverickCodestylePolicy [01:01] JontheEchidna: I think we should only use // style comments [01:02] that /**/ stuff is odd IMHO [01:02] that's apidocs stuff [01:02] also limiting to // helps with consistent code appearance a bit [01:02] you'll only see /**/ in headers [01:02] JontheEchidna: what is apidocs stuff? [01:02] /**/ [01:03] no, those are C multiline comments :P [01:03] and in the presented code examples they are used outside headers [01:03] which are used as doxygen syntax in headers [01:03] ooh, the AmarokProcIO one [01:03] JontheEchidna: that sure can be configured [01:03] doxygen can use // comments too :) [01:04] see :P [01:04] * apachelogger would opt for // only, and only that [01:04] the only time I think /* */ is justified is for large comment blocks, like the license block in a header [01:04] simply because it's easier when writing it [01:05] no one writes license blocks though :D [01:05] they get copied [01:05] apidocs can get a bit long winded, depending on what's being documented [01:06] but I agree that in .cpp files there should never be any non-license /**/ [01:06] well, by "writing" I mean /*\nCtrl-V\n*/ [01:07] JontheEchidna: ok, so lets go for // except for api documentation which must be /**/ [01:07] I think doxygen defaults to /** for C, and /** or /// for C++ [01:08] so both could work, although writing api docs with /** is a little easier [01:08] JontheEchidna: ideally /*\n [3 space indent] blah\n [3 space indent ]blahblah\hn */ [01:09] JontheEchidna: that said, maybe api documentation should be enforced too? [01:09] I'm all for that [01:10] groovy [01:11] http://people.ubuntu.com/~apachelogger/screencasts/kscreensaverinstall.ogv [01:11] also note how the name is nicely localized in the notifications (unlike the other implemetations of the patch ;) [01:12] ;) [01:12] question is if we want to break string freeze for that [01:13] kdesudo is huge liek xbox localized [01:13] its not the most prominent string though ... only the name of the thing to be installed [01:13] JontheEchidna: there is a bug in kdesudo too [01:13] it doesnt render the command for some reason [01:13] that always happens in combination with kdesudo --- foo -- I think [01:13] yeah, it gets cut off after a few args [01:14] MaverickCodestylePolicy updated [01:14] maybe it tries to parse te first -- or something [01:14] * apachelogger notes that kdesudo needs some love to get rocking with the codestyle ;) [01:15] ;) [01:15] did you see what I did with the use case names ;D [01:16] // Default constructor [01:16] Queuelist(QWidget *parent, const char *name = 0); [01:16] // Default destructor [01:16] ~QueueList(); [01:16] policy violation :P [01:17] /* [01:17] Default constructor [01:17] */ [01:17] and the like for destructor [01:17] probably needs @param in there too for the constructor [01:17] yeah [01:17] JontheEchidna: the slots format is messed up [01:18] carriage return in the wrong place, my fault [01:19] JontheEchidna: I should be contributor ;) [01:19] apachelogger: done [01:19] JontheEchidna: the copyright part implies that we create gpl-only software [01:20] apachelogger: do we want a specific licensing policy? [01:20] + it should be ensured that for each used license there is a complete copy stored along the source [01:20] too easy to forget [01:21] JontheEchidna: well, we could enforce KDE licensing [01:21] 'specially since Riddell's a big part of the KDE licensing policy team [01:21] yep [01:21] "Use KDE licensing policy, beotches [01:21] " [01:21] ^^ [01:22] you should probably specify if(...) vs. if (...), that's always a favourite :) [01:22] http://techbase.kde.org/Policies/Licensing_Policy [01:22] true [01:22] if () [01:22] works way better for long conditions [01:23] JontheEchidna: what do you think? [01:23] Personally I like if () [01:23] and the same for "for", "while" and "case" [01:23] yeah, applies to all that funky stuff [01:23] er "switch", not "case" [01:23] on a related note [01:24] zomg, the examples there use if() [01:24] brackets for oneliners, or not? [01:24] * apachelogger finds it much better to use brackets all and everywhere [01:24] my rule is, no brackets unless there is an else [01:24] brackets for consistency, imo [01:24] well, you need the brackets > 1 line anyway :S [01:25] *brackets everywhere for consistency [01:25] JontheEchidna: it also is quicker to add stuff [01:25] and one cant forget about making them [01:25] plus it makes it clear what belongs to that statement [01:25] I man one-line if's, but you could but it all one one line anyway, if (foo) bar(); [01:25] instead of relying on whitespace [01:25] s/man/mean/ [01:25] tsimpson: ah [01:26] apachelogger and JontheEchidna: Doesn't KDE have a coding style guide? Can we just use that? [01:26] ScottK: no they dont [01:26] well [01:26] Oh. [01:26] only for kdelibs [01:26] for kdelibs they have [01:26] But this is very similar [01:26] OK. Why not use that? [01:26] not suited for apps IMHO [01:26] Instead of merely being "similar" [01:26] Why? [01:26] ScottK: because amarok also derived from it ;) [01:27] for example IIRC it features a 2 space indent, which IMHO is just horrible [01:27] no indeed it is 4 ^^ [01:27] Much easier when upstream just provides standard advice everyone can use (e.g. Python and PEP-8) [01:28] Our style is fully compliant with kdelibs, just with some more things specified [01:28] http://techbase.kde.org/Policies/Kdelibs_Coding_Style [01:28] ScottK: ours is more complete [01:28] indeed it implements the kdelibs style by a 100% [01:28] OK. [01:28] actually, kdelibs is using 4 space indent from the astyle command they give [01:29] JontheEchidna: [02:27:49] no indeed it is 4 ^^ [01:29] :P [01:29] oh [01:29] ^_^ [01:29] though I think it was 2 at some point or maybe I am confusing projects here [01:29] anyhow [01:29] kdelibs also gives an opinion on one-line if statements [01:29] JontheEchidna: it has some explicit statements that are worth adding to our style [01:29] apachelogger: yeah, like Qt includes [01:29] Use blank lines to group statements [01:29] Use only one empty line [01:30] No space after a cast [01:30] all very sensible things :) [01:30] Each variable declaration on a new line [01:30] JontheEchidna: ^^ [01:30] uberimportant to mention [01:31] eek, using endl in a kDebug() statement is not good practice. must fix the example [01:31] why must amarok use it's custom debug class? :( [01:31] * apachelogger notes that a team member from his most recent cpp exercise had a line that was like ... foo* bar, foobar, blah; [01:31] * apachelogger almost fell over from that [01:32] JontheEchidna: it doesnt anymore I think [01:32] in KDE 3 they had one because KDebug was essentially crap and mxcl came up with a much better thingy [01:32] I think that debug class is now also used in the lastfm client [01:33] they were using it in recent kde4 times, I know [01:33] * apachelogger notes to create an astyle script once the spec is accepted [01:33] JontheEchidna: then I really dont know, there must be some fancy advantage to it though [01:33] well, it's not beyond 'em to... overengineer things a tad from time to time ;) [01:34] but that's why it's Amarok, ya love it or hate it [01:34] ^^ [01:35] nothing compared to lancelot though [01:36] apachelogger: lancelot has a python meta-ui compiler. :S [01:36] combining the crappiness of .ui-like systems and python [01:37] lol, awesome :D [01:37] JontheEchidna: http://paste.ubuntu.com/410789/ [01:37] I made you addPackages function useful :P [01:37] \o/ [01:37] + m_screensaverPackages["kscreensaver"] = i18n("Set of default screensavers"); [01:37] a teeny tiny string... [01:37] hrm [01:37] decisions [01:38] IMHO getting one more item of my assigned list is worth that string :P [01:38] + chances are that most users will not see it anyway, systemsettings is a way too scary name ^^ [01:38] as long as you can get the red tape out of the way, I'm fine with it [01:38] it has the potential for awesome [01:39] well, once the freeze is over [01:39] * apachelogger also needs to fix the other patches anyway [01:39] they need to obtain the translated programName somehow [01:40] hardcoding that stuff scared me quite a bit last time I stumbled upon them [01:40] in fact [01:40] I think last time == when I implemented that stuff for konqueror [01:40] * apachelogger checks the konq patch [01:41] + dbusargs.append(KonqFactory::aboutData()->programName()); [01:42] see, I am a magician ;) [01:42] ;) [01:42] well, if implemented that properly, maybe I fixed dragon on the way [01:42] dbusargs.append("Dragon Player"); [01:42] omg [01:43] kaffeine also carries that patch? [01:43] epic phail [01:43] then I think amarok, dragon and kaffeine need fixing [01:43] dbusargs.append("Dragon Player"); [01:43] er [01:43] * apachelogger is not going to care about arora which probably also got that patch [01:43] dbusargs.append("Kaffeine"); [01:43] yeah [01:43] stupid klipper [01:43] will fix that after freeze [01:44] does amarok have localized names? [01:44] no [01:44] not too big of a deal then, but it wouldn't hurt to fix it [01:45] So what does the kdelibs style lack? [01:45] I probably wont be able to sleep becaue of this :P [01:45] kubotu: time [01:45] apachelogger: Europe - Vienna - Thu Apr 08 02:45 CEST [01:45] * apachelogger is sure going t be glad when he can upload all the stuff fixed in https://bugs.edge.launchpad.net/~apachelogger/+assignedbugs [01:47] ScottK: how to name member variables, general function-naming guidelines, header include order, const correctness guidelines [01:47] OK. [01:47] Instead of replicating much of the kdelibs style, how about making the spec kdelibs style, plus this stuff. [01:48] Any of that stuff a candidate for adding to the kdelibs style? [01:49] I think there are advantages to having it all listed out in one location [01:51] Riddell: I posted my brainstorming ideas under your name :) ... didn't notice that before [01:51] also.. did ya get my ui file? [01:52] shtylman: actually, a lot of that stuff probably isn't Riddell. I saw a lot of other peeps in the last-edited field that dont' have a name listed [01:53] well, maybe not a lot, but [01:53] a few anyways [01:53] JontheEchidna: good...im not alone :) [01:57] JontheEchidna: That's true, but I can also imagine it being more work to maintain [01:58] Good style is good style. Can't change too much, I don't think [01:58] http://www.madatoms.com/site/blog/lost-spin-offs <-this [01:59] I'd totally watch "Rulers" and "Pros and Cons" [01:59] Clucks 'n Bucks too, since Hurley is awesome === dendro-afk is now known as dendrobates [04:38] apachelogger: The krandrtray thing doesn't revert either here. [09:37] Riddell: ping [09:39] hi agateau [09:39] Riddell: two things I'd like to see with you [09:39] 1. I noticed krandrtray uses non-standard entries in its ksni menu [09:39] it uses bold to indicate selected resolutions instead of radio buttons [09:40] but bold is not supported with dbusmenu, so there is no way to tell which menu item is selected [09:40] is it too late to include a fix for that? [09:40] where fix can be either: [09:41] port to radio buttons [09:41] disable dbusmenu for krandrtray (like I did for knm) [09:42] Riddell: what do you think? [09:46] may as well do it properly and port to radio buttons I'd think [09:46] ok [09:47] if it's possible to get it into lucid, I am going to work on it [09:47] yes it would be possible [09:47] great [09:47] unless nixternal has UI freeze issues but I doubt it's something we have explicitly documented [09:48] ok [09:48] I guess I can work on it anyway, the worse thing which could happen is having it only in kde 4.5 [09:50] Riddell: 2. have you seen the gsoc messages from Daniel Holbach? [09:50] Riddell: I am not sure I am the right person to answer them [09:50] (meaning, you are probably more qualified ;) ) [09:51] Riddell, agateau: mind the patch policy [09:51] no that knradrtray had much of an active upstream from what I saw :( [09:51] s/no/not [09:51] apachelogger: patch policy? [09:51] * agateau feels like he is about to learn about something he should know [09:52] agateau: https://wiki.kubuntu.org/Kubuntu/Specs/LucidPackaging [09:52] see Design part [09:52] you just confirmed my statement that it is not visible nough though ^^ [09:53] :) [09:53] ok, no problem with this [09:53] I would expect this patch to be easily upstreamed [09:54] agateau: just apply it there as well ;) [09:54] the code there is so ugly no one would care if you rm the whole thing ^^ [09:54] code in krandrtray? [09:55] krandr* as a whole [09:55] I've seen much worse [09:55] no doubt [09:55] but I have an easly upset stomach [09:55] :) [09:55] * agateau has too many scars from ugly code to be easily upset [09:55] agateau: if you get a minute or two please have a look at https://wiki.kubuntu.org/Kubuntu/Specs/MaverickCodestylePolicy [09:56] Riddell: ^ too please [09:56] it's mostly a combination of the kdelibs and amarok policies with some additional explicit requirements [09:57] it really looks like a copy of kdelibs coding style [09:58] Only I am not sure about kdelibs policy wrt indenting Q_OBJECT and public: blocks in classes [09:58] our spec says 90chars/line, kdelibs says 100 ;) [09:59] sounds like bike shedding (90 vs 100) :/ [09:59] dont be silly :P [09:59] * apachelogger wants 80 [09:59] ! [09:59] ^^ [09:59] yeah, and max 5 lines per function! [10:00] 50 [10:00] I would be bribable to go up to 75 for C-only code ;) [10:00] ugh, I can't read amarok source, there's way too much whitespace in it [10:00] vertical whitespace, mostly [10:00] that's why you need the 5 line limit! [10:02] agateau: yes I'll handle the gsoc e-mails [10:02] Riddell: ok great [10:04] apachelogger: I'm not sure why we need a code style policy, if we're modifying existing code then go with the existing codestyle, else use Qt policy same as KDE [10:05] Riddell: for our stuff only ... Qt and KDE styles feel to unprecise [10:05] I agree with Riddell [10:05] If it's good enough for kdelibs, it should be good enough for Kubuntu, no ? [10:08] Riddell, agateau: ScottK suggested that we coudl reference the kdelibs style and add addtional stuff (instead of currently reproducing it mostly) [10:08] JontheEchidna thought it might be good to have the list all in one place [10:09] Sput: you could run sed before reading ;) [10:09] * agateau would make the page a link to kdelibs page [10:09] apachelogger: or just not read such badly formatted code :P [10:09] lol [10:09] if you have the choice [10:10] * apachelogger never has and constantly ends up crying [10:10] * agateau hugs apachelogger [10:13] ok, krandrtray is more standard now [10:13] testing all options hurts the eyes =( [10:14] * apachelogger rehugs agateau [10:16] is it just me or doesn't krandrtray save the settings anymore? it doesn't even use the auto resolution by default :( [10:27] debfx: needs rewrite [10:28] debfx: or rather, needs to be changed to use kephal [10:28] though I am not exactly sure what kephal is meant to do , since it doesnt do anything right now [10:28] well, nothing other than watching for xrandr changes and emitting signals about that ^^ [10:30] Riddell: what have you decided to do with the brightness patch? [10:30] I'd like to upstream it but I don't know where to put the osdwidget class so it can be used by kmix and the battery applet [10:35] debfx: yes it should be upstreamed [10:36] debfx: since there's no obvious upstream to attend to it I suggest just posting to kde-core-devel [10:48] Riddell: ok, what about lucid? either the old patch should be dropped or updated to the new one [11:01] \o/ [11:01] * apachelogger just spent 94 EUR on train tickets === michaelk is now known as Guest58382 [11:02] emonkey: on may 20 I am from 7:20 until sometime in the afternoon in zurich! [11:02] after traveling for 12 hours or so :D [11:07] debfx: I'd like to update it but only after there's been agreement on kde-core-devel that it's a good thing [11:13] apachelogger: even better: bring emonkey to Randa :) [11:13] emonkey: be prepared to get kidnapped [11:14] Riddell: bug 558333 please [11:14] Launchpad bug 558333 in kubuntu-notification-helper "freeze exception for kscreensaver installation recommendation" [Undecided,New] https://launchpad.net/bugs/558333 [11:17] apachelogger: that only recommends screensavers from the screensaver module? it doesn't recommend to install screensavers from amarok? [11:17] Riddell: right [11:17] groovy [11:18] the screensavers from a whole different set of install types, so one needs to call the dbus interface with a package from that set to trigger recommendatin of the set :) [11:18] * Riddell approves [11:18] thanks [11:23] * Mamarok wonders what screensavers from Amarok would be... [11:24] Mamarok: nothing, it is just that one can easily use that code in a wrong way and end up with amarok recommending installation of screensavers ;) [11:24] sounds funny :) [11:25] well, currently it recommends installation of libdvdread which is not too sensible either IMHO [11:26] hm [11:26] JontheEchidna: I have but one question ... why does the current knh from lucid contain a build dir? :P [11:26] anyhow [11:27] not in sync with bzr :( [11:37] kubuntu: Harald Sitter * rapachelogger@ubuntu.com-20100408103733-3c7fu52g92svzf5n kdebase-workspace/debian/ (3 files in 2 dirs): [11:37] kubuntu: Add kubuntu_02_screensaver_restricted_install.diff to recommend installation [11:37] kubuntu: of kscreensaver package via k-n-h (LP: #273667) [11:42] kubuntu: Harald Sitter * rapachelogger@ubuntu.com-20100408104238-2odrc2rx5abo9em3 kdebase-workspace/debian/ (7 files in 2 dirs): merge with parent [12:03] mhhh [12:04] ahhhh [12:04] meeh [12:04] :( [12:05] oh well, kickoff doesnt like me either [12:05] curel world [12:35] Riddell: anyone tried lucid on arm yet? [12:37] freeflying: not as far as I know [12:38] Riddell: the experience was too bad [12:58] apachelogger, may 20? 7.20 am? [12:59] damn early ... what are you doing here? [12:59] freeflying: I saw a demo in February with Kubuntu Netbook that seemed to run, but I didn't see any interaction. [13:01] emonkey: Multimedia Sprint in Randa :) [13:01] Randa? Don't know that ... [13:01] emonkey: check here: http://community.kde.org/Multimedia/Sprint2010 [13:01] mom [13:01] Randa is close to Zermatt [13:01] persia: it could be run, but you know the experice is too bad [13:02] uh Zermatt? that's not that close ... [13:02] freeflying: That's annoying. Does something not work, or is it just slow? [13:03] persia: at the first boot up time, nepomuk used too much resources, make other application keep on crash [13:03] Hey guys I get a weird flashy effect the first time I click on Kicker (is it) the K menu. after that it's fine. [13:04] emonkey: much closer for you than all the people coming to the sprint [13:04] persia: I tried karmic on imx51, its not that bad [13:04] but yeah, as usual, the Swiss think far away when it's not the next town :) [13:04] Mamarok, true ... but I don't think that I've got enough money to get there ... [13:05] emonkey: well, check the prices, you might be astonished [13:06] freeflying: Sounds like a bug in nepomuk to me, but worth profiling. I suspect there's a way to make it work better. [13:07] ryanakca: how about updating "User of the moment" with that e-mail from Jay Franklin? [13:07] persia: :) actually, Qt4.6 shall perform better than gtk on arm [13:08] freeflying: I'm generally against competitive library comparison, but why? Do we know that it's compiled optimally in Ubuntu? [13:09] Mamarok, 113.- do you know how much that is for me? ;-) [13:10] emonkey: well, I sent a mail to various LUG lists talking about the SPrint in December, you could have been sponsored if you reigstered earlier [13:10] persia: well, forget what I've said :P [13:10] registered* [13:11] freeflying: But seriously, I expect that it needs a bug with some profiling information. Could be poor optimisation, non-optimal compiler flags, memory issues, etc. [13:11] Mamarok, But I think I'm not really that helpful there, that a sponsoring makes sense [13:11] why do you think that? [13:12] persia: if I can find another reference board, I will try more [13:13]