[00:33] [A [01:23] argh how many times does Wendy's need to inspect one restaurant in 8 week? [01:24] I guess the answer is 5, I just got finished with one tonight, so the release notes I planned will be a little delayed [01:24] Riddell: ScottK et al ^^ [02:07] apachelogger: is installing the gnomey u1 client before trying your kde one still a requirement? [02:07] apachelogger: and do UDFs work yet? if so, how do i set one up? [03:45] claydoh: Nice. Sounds great. === shadesalyer_ is now known as shadeslayer_ [06:45] kronos_: dude.. on the IRC so early? [06:45] nothing to do @ IIT? :P [06:45] shadeslayer_: sort of .. [06:46] hehe :P [06:46] shadeslayer_: m bunking a psychology class :D [06:46] haha :) [06:46] ive gtg now [06:46] ciao [06:46] shadeslayer_: ciao ... [06:49] maco: no, yes, systemsettings -> ubuntuone -> Folders [07:04] apachelogger: does that require maverick? cuz using your ppa on lucid i dont have a folders setting [07:04] yes, maverick [07:05] * apachelogger tries to get desktopcouch and the syncdaemon patched up in maverick this very moment since he will prolly not be able to do a whole lot on the weekend === yofel_ is now known as yofel [10:35] * Riddell wonders why the release team meeting has been cancelled [10:38] heh cause no release :D [10:38] lol [10:38] sup peoples? [10:40] wow do we have an new kubuntu.org page? [10:40] looks good [11:10] debfx: yes lots of ? icons in kate [11:11] shadeslayer: fancy doing the new kdepim beta? [11:40] <[4-tea-2]> Howdy. I got sent here because I found a data-destroying bug in digiKam 1.3 (KDE 4.4.92, all from the kubuntu PPA). [11:41] <[4-tea-2]> I reported it for digiKam on the KDE bug thingie, but I guess I was kinda late, with KDE 4.5 coming out in a few days. I kinda hope to be in time to have it fixed for 10.10, though. Anyone interested? [11:43] [4-tea-2]: 4.5 is unrelated to the digikam release, upstream are likely to be best placed to fix it [11:43] [4-tea-2]: what's the bug number in bugs.kde.org ? [11:43] <[4-tea-2]> Gimme a sec. [11:44] <[4-tea-2]> https://bugs.kde.org/show_bug.cgi?id=246743 [11:44] KDE bug 246743 in Tags "digiKam ate my image (file truncated to 0 byte after tagging)" [Major,Unconfirmed] [11:47] <[4-tea-2]> I was under the impression the digiKam was a "part of KDE" and would be released in-sync. God to know it isn't so. [11:47] <[4-tea-2]> s/God/Good/ [11:47] [4-tea-2]: KDE is a large community, some parts of KDE software are released together, many others do their own releases [11:48] digikam does its own release [11:48] <[4-tea-2]> gotcha [11:48] [4-tea-2]: that looks like a significant problem indeed but upstream will be best placed to deal with it, if you notice a fix or other comment from him do let us know, we might be able to patch our packages [11:50] <[4-tea-2]> Will do, thanks for your advice! [11:58] o/ [11:58] now here is an interesting thing I noticed... [11:58] well, actually two [11:58] for one thing [11:58] knotify eats almost 20 MiB of RAM (unshared that is) [11:59] of those 20 around 10 are because of phonon (i.e. sound notifications) [11:59] and of thsoe 10 are around 4 from libavcodec itself [11:59] I find that a bit outrages really [11:59] secondly [12:00] krunner uses about 5 MiB of unshared memory for libqalculate [12:00] and that thing is only used for the =1+1 stuff [12:00] now as much as I like this feature [12:00] 5 MiB makes me roll with my eyes really wildly [12:01] so just to have that said ... by turning off sound support in knotify and the caluclator in krunner one can save 15 MiB of RAM [12:01] ...that is quite a bit... [12:01] * yofel would rather have akregator not taking 300MiB after reading a few feeds... [12:01] now to continue on those odd observations [12:02] ubuntuone-syncd from a cold start eats more than flipping mysql [12:03] and to blame myself too [12:04] either ubuntuone-statusnotifier leaks memory in the lib OR I am very very sloppy [12:04] once connected and active that beasty has a heap of > 19 MiB [12:04] in fact, I think it is growing ... which supports the theory of leaking [12:50] Riddell, shall we announce a Kubuntu Translations Day for next week at some point today [12:51] dpm: what date did we agree on? [12:51] Riddell, it's on your calendar, next Friday [12:54] next Friday is good, I'll e-mail kubuntu-devel [12:54] might do a wee blog too [12:56] uh! [12:56] * apachelogger has the leak [12:57] dpm: when do language packs get generated? [12:58] Riddell, twice a week except for freeze periods, IIRC, but I'd have to ask Arne when the next round is [12:59] cos the current KDE translations uploaded earlier this week are incorrect [12:59] Riddell: weird thing is that even when null.png is installed, kate still displays ? icons [12:59] and there was that bug with the desktop_.pot files which won't get fixed until we upload 4.5.0 on Tuesday [13:00] right [13:00] [ubuntuone-kde] Harald Sitter * apachelogger@ubuntu.com-20100806120045-zhz2kal1y8p7nelv * src/libs/ (SyncDaemonStatus.cpp SyncDaemonStatus.h) implement proper dtor to fix memleak [13:01] Riddell, if we could get new language packs on Thursday or so, that'd be good, if not, we can just delay the translations day. Let me check with Arne. [13:02] Actually, now that I'm thinking about it, there is a bug with the translations approver in LP right now [13:02] <\sh> apachelogger: ping [13:03] \sh: pogo [13:03] <\sh> apachelogger: you are the one who gace a talk about plasmoids in javascript...how do I get the attached layout from a QGraphicsViewItem like a Frame()? [13:03] dpm: one that would make doing translations QA not very useful? [13:05] hm [13:05] hmmmmmmm [13:06] \sh: how do you get hold of a QGraphicsViewItem in the javascript envrionment? [13:06] IIRC plasma does not even expose that [13:06] Riddell, yep, one of those nasty ones. I'm checking it out with danilo and Arne right now [13:06] <\sh> apachelogger: wait I'll give you an example code ;) [13:07] \sh: also, QGraphicsView inherits indirectly from QFrame ... so technically should be able to set it up there too [13:07] <\sh> apachelogger: well, what I mean is something like this: [13:08] dpm: wiki page https://wiki.kubuntu.org/Kubuntu/TranslationsDay [13:09] <\sh> function foo() { frame=new Frame(); framelayout=new LinearLayout(frame); return frame;} newframe=foo(); newframe.getLayout().additem(new Label().text="hello"); [13:09] <\sh> so my function is returning the container object...and now I need the attached layout of the container [13:09] newframe.layout() [13:09] <\sh> thx [13:10] Riddell, looks good! Let me wait for danilo's answer before announcing it though. If we cannot get a language pack with the 4.5.0 translations on time, we might want to delay the translations day a bit [13:10] <\sh> apachelogger: not defined layout() ;) [13:10] Oo [13:10] scary [13:11] \sh: apparently you cannot access the layout :O [13:12] <\sh> lol [13:13] full ack [13:13] anyhow [13:14] hmmmm [13:15] \sh: if I amnot mistaken the layout is not the layout of the frame anyway [13:15] but just parented by the frame, i.e. when the frame gets cleaned up the layout will go to [13:17] \sh: I recommend asking in #plasma-devel [13:18] <\sh> apachelogger: well, reading http://api.kde.org/4.x-api/kdelibs-apidocs/plasma/html/classPlasma_1_1Frame.html and http://doc.trolltech.com/4.6/qgraphicswidget.html#layout it should be exposed somehow [13:18] \sh: but I think the Frame is really meant to contain but one thing and not multiples [13:18] the javascript api deliberately limits access to functions [13:18] <\sh> apachelogger: tell that the scrollerwidget example in c++ ;) [13:18] what is not documented at http://techbase.kde.org/Development/Tutorials/Plasma/JavaScript/API is not supported [13:19] \sh: the C++ api targets a different kind of developer IMHO ^^ [13:19] \sh: also, even if the layout function was exposed it would not be of any use because it returns a QLayout which is not supported in the plasma::javascript api [13:20] what you could do is use the qtscriptbindings and use a real QFrame [13:20] <\sh> QGraphicsLayout * layout () const <- which is something like linearlayout [13:20] a bit hackish though [13:20] <\sh> apachelogger: the other way of solving my problem is going back to python ;) [13:20] \sh: something like will not suffice with the javascript api I am afraid [13:21] well [13:21] there is one argument that makes javascript or c++ the favorable approach which is that those are the only two engines available on every plasma setup [13:21] python or ruby or webkit or foo might either not be built at all or not installed [13:23] \sh: does it work if you do the same thing with the layout outside the function? [13:24] <\sh> apachelogger: dunno...but I know how to workaround that now ;) [13:24] \sh: how? [13:27] <\sh> apachelogger: http://paste.ubuntu.com/474012/ <- like that ;) [13:28] ^^ [13:30] http://javascript.about.com/library/blmultir.htm [13:30] <\sh> apachelogger: that's ugly [13:31] well, it is javascript [13:31] <\sh> apachelogger: the js array version is much prettier ; [13:31] <\sh> apachelogger: I know ;) I did some coding with Qooxdoo ;) [13:31] <\sh> which is Qt for the web ;) [13:31] the ugliest about all of this is that there is no layout() function though [13:32] * apachelogger uses that function considerably often [13:35] Riddell, ok, I've talked to Danilo and Arne. Let's delay the Kubuntu Translations day a week (on the 20th of Aug). This way we'll ensure that the approver bug is fixed, that we've got a language pack containing all translations, and that we're past the next Launchpad release, which seems to be next Thursday [13:37] Riddell: ill do it tonight :) [13:37] i saw the mail @ college [13:39] and all the PC's had windows .. one had red hat... couldnt do much there :p [13:39] <\sh> apachelogger: btw..is it possible to use the HTTP extension for posting data to a web api? [13:39] <\sh> (inside this javascript foo?) [13:39] ill need testers :D [13:41] can someone off the top of their head think of translatable apps that are Kubuntu-specific, or that are KDE-ish and where the upstream is LP? I'd like to promote their translation, since they tend to remain hidden amongst the big list of translatable templates, and Ubuntu translators tend to concentrate on more GNOME-related translations [13:42] <\sh> apachelogger: forget all about javascript...I'm coding the stuff in c++... that's much easier even for me [13:42] dpm: ok i think it works this : KDE does translations -> LP strips during build -> we work on translations [13:43] i get a ton of mail when that happens from rosseta :D [13:44] now what is a foo.lsm file [13:44] shadeslayer, I know the workflow, notice that I haven't mentioned KDE upstream apps. I'm asking about applications such as kubuntu-firefox-installer and so on [13:44] ohhh those... [13:44] :-) [13:44] no idea on those.. thats all the info i have :P [13:44] no worries, hopefully someone can come up with a list [13:46] dpm: https://launchpad.net/kcm-gtk https://launchpad.net/kubuntu-notification-helper [13:46] * shadeslayer gets back to Behind Enemy Lines [13:46] great, another one on the list, thanks debfx [13:46] well, two, sorry [13:48] \sh: you can include any javscript library/module foo in your plasmoid [13:49] dpm: kubuntu-debug-installer [13:53] apachelogger: need some advise on http://gitorious.org/~megabigbug/rekonq/megabigbugs-clone/commit/c6ec469aba250f0fc658f5968c2ae7517abfd6ff [13:53] cool, thanks apachelogger [13:53] apachelogger: we can pass url.url() instead of all that manually right? [13:54] shadeslayer: what is url? [13:54] url is a KUrl object [13:54] dpm: I think that is it, the other tools are old and supposedly getting replaced by JontheEchidna's work on C++ implementations [13:55] shadeslayer: I do not think so [13:56] do not think what? :D [13:56] right [13:56] if I am not mistaken KUrl::url() will return the whole url [13:56] but in the code at hand a manual composition of the host only is used [13:56] like www.google.com [13:56] just that ^ [13:56] shadeslayer: but I think you can let kurl or qurl build that [13:56] http://www.google.com [13:56] BUT [13:56] right [13:57] theres always one :P [13:57] if the url were http://www.google.com/?2lsl=111 then that would be the return value of .url() [13:57] which is not what you need at this point [13:58] ok... i got it.. we just need the http://www.google.com [13:58] IMHO what that change does is wrong too though [13:58] because it does not take into account the port (if available) [13:58] nor authentication (if available) [13:58] which of course would increase the hit rate [13:59] KURL::url().host() ? [13:59] or rahter KURL::host() [14:00] dpm: ok 20th August it is [14:00] Riddell, sounds good to me [14:00] txwikinger: where is that used? [14:00] shadeslayer: I think url.host() should be authority() [14:01] apachelogger: no idea.. QUrl has host if you only want the host of the url [14:01] you can also use authority I believe.. works both [14:01] no [14:01] host is only a subpart of authority [14:01] what would authority return? [14:01] authority will include authentication and port [14:02] apachelogger: right [14:02] on second thought maybe authentication is too much [14:02] on second thought of the second thought one should make a lookup with and without authentication and prefer former if that yields anything ;) [14:02] ah [14:02] there we are [14:02] shadeslayer: url.toString(FormattingOptions) [14:02] apachelogger: what are you guys doing anyway? [14:03] hmm [14:03] url.toString(QUrl::RemovePath|QUrl::RemoveQuery|QUrl::RemoveFragemnt|QUrl::StripTrailingSlash); [14:03] txwikinger: we need to get proper favicon support :p [14:03] in rekonq [14:03] ah [14:04] wasn't there somewhere an icon() function for that? [14:04] BUT.... qt webkit cache is b0rked [14:04] it downloads favicon for lp and google.. but not for wikipedia and other sites [14:04] well.. konq webcache b0rks with favicons too [14:05] txwikinger: in rekonq when you type stuff in urlbar... there are alot of search protocols.. their favicons dont get cached [14:07] shadeslayer: apachelogger: how can you save and reload all the session information with qtwebkit? [14:07] apachelogger: we just need to pass the host url to get favicon or the exact favicon url? [14:08] txwikinger: I dunno? foreach over all open pages and save their urls? [14:08] txwikinger: saveconfig or something? [14:08] shadeslayer: the favicon can be set in each webpage differently [14:08] WebCore::Image* image = WebCore::iconDatabase()->iconForPageURL(WebCore::KURL(url).string(), [14:08] WebCore::IntSize(16, 16)); [14:09] shadeslayer: I recon WebCore will try to get the favicon via default name ;) [14:09] or (if set) via the page itself [14:09] apachelogger: I mean the session information that I don't have to re-authenticate on a website after restarting the browser [14:09] apachelogger: that needs looking into then :D [14:09] * apachelogger is wondering why one must issue two calls to iconForUrl anyway since WebCore could easily check both cases [14:09] txwikinger: that is what a cookie is for? [14:10] apachelogger: where can I get the cookies in qtwebkit? [14:10] IIRC you cannot and should not store a HTTP authentication [14:10] since they are HTTP session bound [14:10] well konq does it [14:10] no [14:11] konq will throw a auth dialog, just with data filled out [14:11] unless it is a cookie [14:11] which is a completely different approach of authentication to begin with [14:11] well. i think it is a cookie what I am looking for [14:11] txwikinger: i think qtwebkit should store cookies automagically [14:11] kdewebkit supposedly uses kcookiejar [14:12] hmm.. so if I use kdewebkit instead, I can just use kcookiejar for it? [14:13] you should not have to worry about it [14:13] if you have to then there is a bug [14:13] the cookie storage units are tightly integrated with the core stuff [14:13] mostly one can just define an acceptability policy [14:13] like no cookies at all [14:14] or cookies are only valid for as long as the browser session lasts [14:14] well.. I want them be valid longer than the session [14:14] * txwikinger writing his own specialized browser [14:14] that ought to be default... [14:14] in any case [14:15] if you use kdewebkit everything should be just fine [14:15] and basically it is just qtwebkit with kioslaves and kcookiejar [14:15] ok. will try that [14:15] so I would have all the konq cookies available too? [14:16] I think so, yes [14:16] coo [14:16] cool [14:16] apachelogger: webcore is khtml? [14:16] no [14:17] lionel_: ^ [14:17] webcore is the heart of webkit [14:17] apachelogger: the name says it all :P [14:20] apachelogger: WebCore::IntSize(16, 16)); << thats in pixels? [14:20] supposedly [14:22] i do really need mentorz [14:28] hmm.. i still get Error org.freedesktop.DBus.Error.Spawn.ChildExited: Launch helper exited with unknown return code 1 [14:28] Riddell: can you please take a look at http://people.ubuntu.com/~apachelogger/tmp/u1.debdiff [14:30] apachelogger: looks fine, let me check with ken if he's poking the ubuntu one people first [14:30] k [14:31] apachelogger: also i guess kenvandine is the packager of ubuntu-one so he should check you change too, he'd be likely to overwrite it with his next upload if he's not aware of it [14:42] 14:41 < kenvandine> Riddell, just talked to josh, he will get someone to merge them asap [14:42] 14:41 < kenvandine> Riddell, but i am ok with distro patch if that unblocks anything [14:42] apachelogger: ^^ [14:43] perfect [14:43] Riddell: thanks for poking :) [14:46] 14:42 < kenvandine> merging the branches doesn't mean they will release [14:46] 14:42 < kenvandine> although we might not want to upload a change like that on a friday :-D [14:46] Are there any python bindings for kdewebkit? [14:48] don't think so no [14:49] bah.. too bad [14:49] Riddell: did you see the backlog of want to be members? :D [14:50] https://edge.launchpad.net/~kubuntu-members/+members#proposed [14:50] shadeslayer: we always get those [14:51] shadeslayer: if you want you can e-mail them to ask if they're serious and where their wiki page is [14:51] hmm... well i know bulldog98 is away for a while :P [14:55] Riddell: err... the kdepim mail says beta 2 and yet the tarball is named beta 1 [14:55] Riddell: clearly, we are outnumbered [14:56] shadeslayer: unstable/kdepim/4.4.92/src/ [14:56] bah.. wrong folder :P [14:56] maco: who's we? [14:56] Riddell: me and you on the "yeah sure, normal users like html" thing [14:56] maco: uh.. bad idea to enable html on kmail by default [14:57] i think a relevant question is whether setting it to /display/ html also sets it to /send/ html [14:57] +1 for normal default behaviour [14:57] except that's not normal behaviour outside of geeky linux systems [14:58] Riddell: btw this is going again in experimental? [14:58] shadeslayer: yes [14:58] ive replied to html email and gotten a reply back asking what all that code is... normal people, upon seeing the code stuff in the preview pane, might think something's broken / the email corrupted, without even bothering to read the box [14:58] ok... [14:59] maco: Normal users also click on unsafe web sites and agree to install malware. I don't think that means we should make it easy for them to do it. [14:59] ScottK: do we disable javascript by default in browsers? [15:00] because i think "its not showing the colours in kmail" is likely to look just as broken as "the internet keeps telling me i need coffee" [15:00] maco: We don't, but the default browser experience and how to get to the more 'rich' experience is a lot less obvious in a browser. [15:01] default browser experience ... without javascript is a lot worse ... [15:01] Nightrose: while sitting in the train today I had the most horrible idea regarding the release script ... each part (source, l10n, docs...) could be based on an AbstractReleasePart class and each of those parts can have any number of subparts [15:02] apachelogger: that many abstractions make my head spin [15:02] so you have a part KDERelease which contains a part for Source and L10n and Doc and each of those can also contain multiple parts (e.g. if you want to do more complex stuff int he source fetching [15:02] Nightrose: well, it is quite dynamic while allowing high level of automation [15:03] like KDERelease can fill itself up with standard Source and L10n and Doc (if those are to be built...) and those will do appropriate things 90% of the time [15:03] * Nightrose nods innocently [15:04] but if someone desires to write their own messing with the source and stack that on top of the default one, then the particular change would be like 3 sloc [15:06] apachelogger: i cant frickin remember how to compare package versions with dpkg :S [15:06] pkg --compare-versions 4.5.0 > 4.5 isnt working [15:06] err... ^ dpkg [15:06] Nightrose: also it allows to wrap multiple releases into each other ... say you want to roll the amarok collection scanner independently ... then you would have one script that builds a Release but that Release contains another Release for the collectionbuild [15:07] Nightrose: anyhow, incredibly powerful without much overhead on the coding side and yet a no brainer for those who trust in my supreme release script coding skillz ^^ [15:07] :D [15:07] *hug* [15:07] shadeslayer: gt, lt [15:07] dpkg --compare-version 1 gt 2; echo $? [15:07] yields 1 [15:07] shadeslayer: are you planing to split out the googletalk-call part of kopete into a separate package? [15:07] because 1 is not greater than 2 [15:08] Riddell: i cant wrap my head around it :S [15:08] shadeslayer: I think it just needs a kopete-plugin-googletalk package with the usr/bin/googletalk-call file in it no? [15:08] Riddell: IIRC it also links stuff into libkopete [15:09] hmm, that's a pain [15:09] Riddell: i can do that part but im usure if that will stop bringing in libmediaavcodec [15:09] * Riddell compiles it to check [15:09] and what apachelogger said [15:09] so either the build system gets changed around to not do that or a packag rebuild needs to happen (liek for quassel) [15:10] Riddell: also I think I just scared the ubuntuone people with my implementation of ubuntu-sso [15:10] of which the two around did apparently not hear until today [15:10] which kind of proofs my point that some stuff needs to be communicated via mailing lists [15:11] apachelogger: I suspect the idea of external contributions to ubuntu one is all new to them [15:11] looks very much so [15:11] apachelogger: i think i screwed up kdepim versioning [15:11] see right now we have 4.5~beta1 [15:12] but if 4.5.1 comes out... 4.5.1 < 4.5~beta1 [15:12] also I was thinking a bit about how to make owncloud sync stuff ... upon which I realized that desktopcouch is really just like akonadi but without the PIMish focus ;) [15:12] so I think owncloud ought to hook up with akonadi really << rbelem [15:13] supposedly that solves the problem of syncing mobile to desktop devices and vice versa [15:13] shadeslayer: no [15:13] shadeslayer: did you check that? [15:13] yes [15:13] shadeslayer@saphira:/media/Work/KDE/pim/pim/kdepim-4.5.0$ dpkg --compare-versions 4.5.1 gt 4.5~beta1; echo $? [15:13] 0 [15:13] that strikes me as odd TBH [15:14] 0 == true [15:14] whut! [15:14] means 4.5.1 is grater [15:14] s/grater/greater [15:14] :O [15:14] i thought 1 was true :P [15:14] because 4.5.0 gets seen by dpkg really as 4.5 [15:14] shadeslayer: not regarding application exit values [15:14] in shell 0 is true, just because shell likes to be confusing [15:14] heh ^ [15:16] * apachelogger thinks that really has to do with the fact that applications would have to report some sort of success or failure back to the OS and since success is hopefully more often the case 0 saves you one pointless bit per process exit :P [15:16] now I really wonder why apps exit with 0 on no error ^^ [15:18] Because they exit with the error code. [15:18] 0 is the non-error code. [15:20] Riddell: kdepim will take some time because ill refresh one of the patch.. and its huge [15:20] specifically kubuntu_01_install_headers.diff [15:20] ScottK: but why! [15:21] did the C standard authors just feel like it? [15:21] I'm not sure I can explain it. It just seems obviously sensible to me. 1 - ... are the error codes to tell you what went wrong. How could the non-error code be anything other than 0? [15:21] or is there a particular reason why 0 is preferrable?> [15:22] I'm not sure. I've never questioned it as it just made sense to me. [15:22] ^^ [15:23] * apachelogger diggs up the original C standard [15:25] could bug 609247 be acted upon while we wait for a review? [15:25] Launchpad bug 609247 in qapt (Ubuntu) "[MIR] qapt" [Undecided,New] https://launchpad.net/bugs/609247 [15:25] wow.. the patch applies cleanly :O [15:25] it's currently blocking an FTBFS upload [15:27] JontheEchidna: yes I'm going to look at that after kopete and pre-promote it if it's sane [15:27] apachelogger: so i guess you're saying that if i want to be able to have a netbook in sync with a laptop, its time for both to be upgraded? [15:28] Riddell: neat, thanks [15:28] maco: if you wish to use ubuntuone that is [15:28] apachelogger: right right... [15:28] Grrr [15:28] apachelogger: forward port u1 already :P [15:28] Quintasan: hey :D [15:29] I hate it when upstream ships debian/ inside tarball [15:29] shadeslayer: \o [15:29] shadeslayer: dude, you just need to build [15:29] Quintasan: yeah me too [15:29] Quintasan: but with 3.0 on our side... its no problem [15:29] it automatically moves the old debian folder out of the way and uses our packagin [15:30] huh? [15:30] How is that done? [15:30] Quintasan: source format 3.0 [15:30] * shadeslayer points to debian wiki [15:30] yeah, but where do I place our packaging? [15:31] Quintasan: just boot the upstream debian out of the source tree and replace it with ours [15:31] [muon] jmthomas * 1159935 * trunk/playground/sysadmin/muon/src/DetailsTabs/ (ChangelogTab.cpp ChangelogTab.h) Robustify changelog grabbing code. Thanks to Felix Geyer for the patch [15:32] AHA [15:32] cool story apachelogger [15:32] Finally, control is returned to the host environment. If the value of status is zero or [15:32] EXIT_SUCCESS, an implementation-deﬁned form of the status successful termination is [15:32] returned. If the value of status is EXIT_FAILURE, an implementation-deﬁned form [15:32] of the status unsuccessful termination is returned. Otherwise the status returned is [15:32] implementation-deﬁned. [15:34] ScottK: my take here is that the general speration is just 0 == success and 1 == error as per C which is probably random since it could also be 0 == fail and 1 == success [15:34] indeed EXIT_SUCCESS AND EXIT_FAILURE translate to 0 and 1... [15:35] apachelogger: If that were it, I'd agree, but I often see specific error codes used for different failure reasons. [15:35] ScottK: yeah, taht is where the implementation-defined form kicks in [15:36] 0 = error, 1 = works, 2+ = some other error would just be odd. [15:36] e.g. you could write a C compiler where exit(100) would really yield 1 [15:36] ScottK: I am not sure that was intentional [15:36] i think id go for 0 = success, and then any other number = a specific error [15:36] that way later if i get a bug report "it said error code 42" i can grep for 42 [15:36] also knowing C I would raelly think that they in particular did not think of additional error codes ;) [15:37] because from a host environment POV only success or error are interesting [15:37] id rather like 1 for success and anyother code for error :p [15:37] the additional error codes probably only come from people in the host environment being silly and implementing scripts around apps ;) [15:37] apachelogger: or wanting useful bug reports! [15:38] that is a bogus argument [15:38] as said [15:38] per spec exit(100) can yield anything in the host environment [15:38] since it is not specifically defined [15:38] btw anyone notice that spell check does not work in OOo anymore? [15:39] the MSVC could easilly make exit(100) cause a return value of 50 just because it feels like it ^^ [15:40] oh [15:41] ScottK, maco: on that note one could also go with exit != success && exit != error -> unknown_error [15:41] * Riddell spots libbluedevil in Debian New queue [15:42] Riddell: I already asked someone to push it through so we can sync it. [15:43] joy, we have two cwp's under two different source package names \o/ [15:43] I'll fix + do new cwp upstream release [15:43] JontheEchidna: CWP ? [15:43] !info plasma-widget-cwp [15:43] Package plasma-widget-cwp does not exist in lucid [15:43] !info plasma-widget-cwp maverick [15:43] ah widget [15:43] plasma-widget-cwp (source: plasma-widget-cwp): Customizable Weather Plasmoid (CWP). In component universe, is extra. Version 1.0.5-1 (maverick), package size 331 kB, installed size 1512 kB [15:44] !info plasma-widget-customizable-weather [15:44] plasma-widget-customizable-weather (source: plasma-widget-customizable-weather): a weather plasma widget that is highly customizable. In component universe, is optional. Version 1.0.0-0ubuntu1 (lucid), package size 221 kB, installed size 1504 kB [15:44] heh... [15:44] cwp is the one from Debian, so I think the other should be removed. [15:44] oh oh oh [15:44] Debian has a newer one though, we should sync. [15:45] Someone should request that. [15:45] ScottK, maco: I am not sure how this would look on hardware level but supposedly they kept the specified values there to a minimum to allow minimal overhead ... as it is you could drop the return values in a one bit register ^^ [15:45] Makes sense. [15:45]