[00:00] lol [00:00] dantti: where was const missing? [00:01] apachelogger: all around but the code still not work... [00:01] Oo [00:01] the array has the values but as QBAlist vanishes I think the copy vanishes too then.. [00:01] request = ippNewDefaultRequest(values.take("printer-name").toString(), isClass, operation); [00:01] that line is still missing a toUtf8 :P [00:03] apachelogger: it's not, ippNew... is not from cups [00:03] I do not understand :P [00:03] can you please put a toUtf8 there :P [00:04] also please commit and push and tell me what to do to test [00:04] apachelogger: ok now it works [00:04] lol [00:04] :P [00:04] don't forget my copyright :P [00:05] the copy flag must be for something else, the qba list is more "global" now, so it several qstringlist goes there oh [00:05] no it will be fine the list is just for not deleting the items.... [00:05] what list? [00:06] the qbalist? [00:06] please commit and push :P [00:06] and need to look at this [00:06] http://privatepaste.com/e955dd497a [00:06] ohhhhhhhh [00:07] dantti: maybe the copy only happens at cupsDoRequest? [00:07] apachelogger: maybe... [00:07] apachelogger: what's your copyright line? [00:07] Harald Sitter [00:08] the QBA should be fine eitherway, it will go out of scope after the request (while scoped) [00:08] dantti: mind that you will still need to delete the char ** [00:08] apachelogger: oh [00:08] or commit and push and let me do it ;) [00:08] * apachelogger fancies doing that :) [00:09] apachelogger: please pull [00:11] dantti: mind if I remove the commented code? [00:11] apachelogger: no, [00:11] we just need valgrind to tell us it does not leak :P [00:17] dantti: how can I test it still works? === apachelogger_ is now known as apachelogger [00:18] apachelogger: make install then plasmaengineexplorer --engine printmanager [00:18] ask for AllJobs [00:18] if you have any [00:19] why? qbaList->reserve(qbaList->size() + list.size()); [00:20] dantti: http://wstaw.org/m/2012/03/08/plasma-desktopeu2263.png [00:20] apachelogger: because we are on a loop, so the next qstringlist might need more space [00:21] yeah [00:21] I moved the QBA elsewhere :P [00:21] into the loop [00:21] turns out dorequest seems to do the copy [00:21] so we are fine [00:21] apachelogger: right, do you agree with me about the reserve part? [00:22] apachelogger: you have all data on that screenshot :) [00:22] oh [00:22] actually you are right [00:22] you have too many loops there :O [00:24] this is flawd then [00:24] dantti: we'd also need to grow the char** accordingly [00:24] apachelogger: no, you only pass one char ** [00:25] but inside while (i != values.constEnd()) { [00:25] I have no idea what values can contain [00:25] but if case QStringList can be true more than once we have a problem [00:25] like I have a QStringList of the attributes I was and a QStringList of something else, I'll ippAddStrings() two times with different char** list [00:26] yes, but they do not get copied until doRequest [00:26] so if you want to do it that you always pass different char** we'd still have to put them in a list [00:27] and then delete all char ** inside the list [00:27] otherwise we leak [00:27] apachelogger: hmm [00:27] right we need a list of char ** too :/ [00:27] http://paste.kde.org/435104/ [00:27] dantti: we should just make that a vector though [00:29] apachelogger: what if instead of new char * we just do char[list.size() +1]? [00:30] apachelogger: that way it get's deleted when out of scope no? [00:30] that could work still [00:30] where would you put that? [00:30] the only way this would work is if it were the same scope as qbalist [00:31] at which point we still have one char** and thus cannot process multiple qstringlists [00:32] apachelogger: I'd put inside qstringtoChar**() [00:32] or even pass it to your function... [00:33] yes but then it has too little scope [00:33] considering cups only copies the strings at doRequest it is likely that it only iterates the pointer at this point [00:33] by that time we'd have no pointer anymore though [00:33] so I fear storing them is the only solution [00:35] apachelogger: it copies the strings not the char** [00:35] yes but to access the strings it needs the char ** [00:35] apachelogger: it goes one by one (actually I don't even think we need the +1 and \0... [00:35] (the strings are at arbitrary locations in the memory [00:35] ) [00:35] apachelogger: that's what the QBAlist is for [00:36] yes but cups cannot access it [00:36] cups only has the char ** address [00:36] so it tries values[0] values[1] values[2] .... [00:36] apachelogger: cups do att[1].text = values[i]; [00:36] in a loop [00:36] that is what I am saying [00:36] so we can delete values** [00:36] values is not valid memory anymore! [00:36] not the values [00:37] values ** is values for cups! [00:37] just the array [00:37] no it has copied the values on the list [00:37] but only at doRequest [00:37] not a deep copy [00:37] just a reference [00:37] man this cups is crap [00:37] yup [00:37] printd++ [00:38] apachelogger: try that or push so I can give it a try [00:39] so [00:39] doing [00:39] for (int i = 0; i < list.size(); ++i) { [00:39] values[i] = 0; [00:39] } [00:39] delete values; [00:39] right after the addStrings should work? [00:40] apachelogger: http://privatepaste.com/bd04fa9ced [00:40] apachelogger: yes [00:40] apachelogger: shouldn't it be delete [] values; [00:41] * dantti is a bit lost on the pointers :P [00:41] too late to answer that question [00:41] wouldn't delete [] delete the content? [00:41] i.e. our QBA data [00:41] * apachelogger would totally use malloc :P [00:42] dantti: perhaps you are right [00:42] apachelogger: dunno :P [00:42] valgrind will know ;) [00:43] apachelogger: it's funny that _cupsStrAlloc() is called when we do ippAddStrings() [00:45] apachelogger: hmm it seems that the IPP_TAG_COPY actually does not copy [00:45] not having that would copy.... [00:46] the COPY would then mean we are passing a copy.... [00:46] Oo [00:46] wth? [00:46] is there documentation for cups? [00:47] apachelogger: look at line 543 from my last paste [00:47] apachelogger: there is.. [00:48] funny documentation ^^ [00:48] http://www.cups.org/documentation.php/api-httpipp.html#ippAddStrings and http://www.cups.org/documentation.php/api-httpipp.html#ipp_tag_e [00:48] if ((int)type & IPP_TAG_COPY) [00:48] value->string.text = (char *)values[i]; [00:48] apachelogger: that's the source code :P [00:49] that looks like an immediate deep copy to me? [00:49] eh [00:49] 1st level deep copy ^^ [00:49] so we can throw the values away [00:49] that is a good thing [00:49] wouldn't that be faster? [00:50] tho using the qbaList might do the opposite... [00:51] hm? [00:51] opposite? [00:51] it's just fine [00:51] inside the do-while we create a QBAList [00:51] inside we stuff all our QBAs [00:51] apachelogger: don't you think it's better to have cups allocating that? so we don't have to worry about QBAList growing... & friends [00:51] temporarily we have a char ** to carry data from the QBAs to ippaddstrings [00:52] that function 1st level copies the string addresses [00:52] dantti: oh [00:52] ohhhh [00:52] I didn't read beyond that line ^^ [00:52] lol [00:52] lol [00:53] dantti: yeah, so dropping the COPY flag should make deep copies [00:53] apachelogger: yup :) [00:53] lemme fiddle the code accordingly [00:53] apachelogger: dunno why gnomies did that :P [00:54] apachelogger: I guess we will still need the qdbaList or we qstrdup then delete them... [00:54] qstrdup is no option [00:54] we need the qbalist [00:54] lol [00:54] but with smaller scope [00:55] qstrdup would mandate 2 string copies [00:55] right [00:55] once from qstring -> QBA and a second time from QBA -> qstrdup [00:55] which is why we introduced that list to begin with ;) [00:56] it still works for me [00:56] cool [00:56] dantti: it is delete [] btw [00:57] since we new an array we need to delete one [00:57] obviously ;) [00:57] :) [00:58] I have a bunch of QLatin1Strings() change to push too :P [00:59] apachelogger: it would be awesome if cups emited a DBus signal when new jobs arive... [00:59] oh yes, you should use way more QLatin1String :) [01:00] at the time of the writting I didn't know about them.... [01:00] dantti: you can't always get what you want :) [01:00] dantti: well, while you are at introducing QL1S, also do QStringBuilder :) [01:00] unfourtunatly [01:00] way too few people use that bugger [01:00] dantti: push'd btw [01:00] bugger? [01:00] cool [01:01] the QStringBuilder [01:01] * apachelogger makes breakfast now [01:01] eggs in the basket muhahaha [01:01] I don't see that class in the assistent [01:01] breakfast? [01:02] I though you were on the US? [01:03] more like EU [01:03] 2:03 here [01:03] perfect breakfast time ^^ [01:04] dantti: http://qt-project.org/doc/qt-4.8/qstring.html#more-efficient-string-construction [01:04] also http://labs.qt.nokia.com/2011/06/13/string-concatenation-with-qstringbuilder/ [01:04] lol.... 10pm here... [01:04] dinner time [01:05] apachelogger: the % huh? I tried but that doesn't compile... [01:11] included qstringbuilder? [01:14] apachelogger: :P no [01:26] apachelogger: thanks, valgrind is happy now :D [01:26] wohoo [01:27] apachelogger: btw pushed a few qstringBuilder changes [01:28] * apachelogger asks the kind launchpad to import the repo so he can kickoff a new snapshot [01:29] * dantti kquits-plasma desktop to have it not leaking [01:29] now I can start writting the colord kcm [01:31] first release that stuff :P [01:32] lol /me and my never released stuff :P [01:32] https://code.launchpad.net/~kubuntu-packagers/+recipe/print-manager-snapshot [01:32] oh, so you need forcing, great :P [01:33] debconf-kde needs a new release, Apper needs to leave playground, p-m too, colord needs to .... well start working :P [01:35] oh [01:35] dantti: qbaList->reserve(qbaList->size() + list.size()); [01:35] you might want to drop that first part [01:35] sure, we don't need that now :D [01:35] * apachelogger heads to bed [01:35] nini [01:36] o weird [01:36] dantti: if colord has bugs, tell RAOF [01:36] micahg: what is RAOF? [01:36] apachelogger: I can't cancel a jobs :P [01:36] dantti: RAOF is an ubuntu-desktopper [01:36] dantti: u broke it [01:37] micahg: it's not colord that has problems it is that the kded module I'm writting doesn't do all the needed stuff yet... [01:37] dantti: also if you did not have 3 loops the code might be lest prone to fall apart :P [01:37] so colord does not work was more like colord-kde doesn't work yet [01:38] apachelogger: lol... well I can take suggestions... [01:38] split it up into functions [01:38] for example [01:39] apachelogger: right I should do that (at least the do while loop could be in a function loop) [01:40] apachelogger: hmm I probably broke something else, the core is still working,, maybe my qstringbuilder stuff broke it... [01:40] just when I rolled a new snapshot :P [01:40] but really I am off now :P [01:41] hehe [01:42] apachelogger: fixed [03:25] yofel: Y U No Upload? [03:33] kde-runtime binaries accepted. [03:36] JontheEchidna: How goes? [03:36] Feel free to upload some KDE 4.8.1 from bzr. [03:37] pretty well [03:38] AIUI https://wiki.kubuntu.org/Kubuntu/Ninjas/Packagin is up to date if you want to grab something. Just make sure to use the .xz tarballs. [03:39] * ScottK does libksane [03:47] * JontheEchidna grabs analtiza [03:47] JontheEchidna: I've already stared that one. [03:47] k [03:47] Can you pick another. Sorry. [03:47] anybody touched smoke stuff? [03:47] Nope. [03:48] debfx: kactivities is now in the Kubuntu packageset. It took a manual override from cjwatson. [03:49] OK. That's done too. [03:50] That's probably about it for me for tonight. [03:53] [smokegen] Jonathan Thomas * echidnaman@kubuntu.org-20120308035304-c1shxomi8tlv3272 * debian/changelog releasing version 4:4.8.1-0ubuntu1 [03:56] [smokeqt] Jonathan Thomas * echidnaman@kubuntu.org-20120308035648-idcge3ep9kmc8q57 * debian/changelog releasing version 4:4.8.1-0ubuntu1 [04:09] apachelogger, the initial code is here https://gitorious.org/copyright-tools/copyright-tools [04:09] :-) [04:10] apachelogger, lots of stuff needs to be done [04:10] i will compile a todo list [07:06] hi give me address repository for test kde 4.8.1 [07:26] brestows1: for oneiric? [07:57] yes [08:08] brestows: it's in kubuntu-backports ppa [08:09] ok [08:09] update [08:25] guys why do you upload stuff with version 4:4.8.1a-something ??? [08:25] why the a there wasn’t any public version of it, so why not just dropping the a? [08:28] bulldog98: that means upstream made a tar which we put in kubuntu-ninjas [08:28] and then make a new tar so we had to update it in kubuntu-ninjas with a new version number [08:28] we could drop the a for the upload to the ubuntu archive [08:28] but that would mess up anyone who tested on kubuntu-ninjas [08:29] Riddell: that’s right, but isn’t there a better solution? [08:29] 4:4.8.1a-something is inelegant but not problematic [08:29] 4:4.8.1-something is more elegant but problematic for our testers [08:30] and it is a good idea to not annoy the testers, we do depend on them :) [08:30] sound’s like a plan :) [08:31] is there an alternative to "ubuntu-bug" that goes to b.k.o?`ie, something that collects all the files, reports and uploads them? then directs to the website and filing? [08:31] jussi: drkonqi [08:31] but otherwise no [08:32] Riddell: so I can run "drkonqi " ? [08:32] no you can copy and paste from drkonqi [08:32] dario might be receptive to ideas but actually integrating it with bugzilla needs a lot of coordination between sysadmins, bugzilla hackers and drkonqi (dario) [08:32] oh well thats annoying. perhaps we should create somethign like that? [08:33] there's a reason why Mark hired a team to make his own bug tracker, integration with a badly maintained project is hard [08:33] hrrr [08:33] see.... we need to get KDE to use LP :P [08:33] bugzilla is hardly a priority for mozilla,I don't know where their millions do get spent but it's not on a bug tracker [08:34] and KDE doesn't seem keen to switch to open source launchpad [08:34] we should take up a bribe fund ;) [08:35] switch to LP and we will give you lots of money :D [08:36] I fear KDE isn't likely to expect lots of money from canonical any time soon [08:36] but that's fine, alternatives for kubuntu are becoming apparant [08:37] hrm? [08:37] * Riddell flies off to Schipol [08:38] Riddell: have fun, thats one of the best airports in the world [08:39] (only city I actually enjoyed was amsterdam - I generally hate cities) [08:54] Riddell: hi [08:55] I add rep, run update, afte update version kde remained 4.7.1 [10:47] FYI: there's a datalos bug in 11.10 4.8 after 4.8.1. https://bugs.kde.org/show_bug.cgi?id=295484 I'll try if I can reproduce in precise [10:47] KDE bug 295484 in Mail Filter Agent "filtering removes email's content" [Critical,New: ] [11:30] ScottK: why some KDE packages need an override but others like libreoffice are in the packageset will always be a mystery to me. [11:41] apachelogger: Ariya is the right pic to have for ksplash [11:55] Riddell, i got sponsorship :-D \o/ [12:00] debfx: There is an algorithm (which I don't understand) based on rdepends and being seeded. Stuff like kdelibs is broadly used by so many things (e.g. ubiquity) that the algorithm lands it in core. The same happened with kactivities. LO, OTOH, doesn't have so many rdepends so it doesn't make it to core and ends up in Kubuntu and Desktop package sets. [12:02] Howdy all [12:11] https://plus.google.com/hangouts/extras/talk.google.com/ubuntu-dev# hanging out with holbach [12:18] tazz: Chatting on G+ with people that aren't involved in Kubuntu development is at least slightly random relative to the channel topic. [12:18] ScottK, ah sorry, i thought the dev stuff and internals were similar. [12:18] my apologies. [12:19] tazz: Not a big deal. [12:19] * ScottK looks at smokekde === greyback is now known as greyback|lunch === yofel is now known as yofel_ === yofel_ is now known as yofel === greyback|lunch is now known as greyback === bulldog98_ is now known as bulldog98 [14:32] when used class KUniqueApplication in konsole show warning message - "QDBusConnection: session D-Bus connection created before QCoreApplication. Application may misbehave" How this correct [14:32] Hi [14:40] brestows, I've seen that warning before, haven't had any misbehaviour so far. [14:41] jussi: if you hate cities you've clearly never been to edinburgh :) [14:41] hi brestows [14:41] BluesKaj: if used Qdbus, it does not always work [14:41] rbelem: great [14:41] hehe [14:42] which app brestows ? [14:43] Riddell: after update on ubuntu 11.10 always crash Muon [14:44] nevermind brestows , I've only used dbus when ssh'd into another pc [14:44] qdbus isn't in my knowledgebase [14:45] BluesKaj: I write app :) whitch use dbus connection and it don't work :( on qt 4.8 [14:46] brestows, right , I think :) [14:46] right , I think [14:46] BBL [14:53] bulldog98: seriously? [14:53] bulldog98: says who? [14:54] apachelogger: at least debian has the same stuff in it [14:54] eshat [14:54] Oo [14:54] the tarball has ariya [14:54] so *every* distro has it for ksplash [14:54] that does not mean it is what was intended by the artists [14:56] dantti: new snapshot building [14:57] apachelogger: did you pull? [14:57] why? [14:57] apachelogger: I've fixed a bug in configure dialog [14:57] last import was 1hr ago [14:57] https://code.launchpad.net/~neon/print-manager/master [14:57] ok, it's probably in :D [14:58] http://bazaar.launchpad.net/~neon/print-manager/master/revision/237 [14:58] yah [14:58] since I fixed a few hours after you went to bed [14:58] brestows: after updating to KDE SC 4.8.1? [15:00] Riddell: no :( I add rep that you given, did update system, version kde remained 4.7.1 [15:02] bulldog98: what is eshat working on? [15:02] brestows: then it is unclear what has been upgraded [15:02] apachelogger: oxygen [15:02] curious [15:02] so why is there this other thing/ [15:02] ? [15:02] the simple gradient [15:03] Riddell: could just put a standard update? [15:03] maybe yes [15:05] apachelogger: eshat said something about that that repro would be outdated [15:05] nice [15:06] bulldog98: you might want to propose that upstream starts using the outdated version then, because an ad-hoc vote amongst kubuntu-devel subscribers suggest ariya splashes give people vertigo [15:07] nixternal: feel free to use the fluffy-unicorn package to pick up where you left :P [15:07] rbelem: gitorious? :S [15:07] apachelogger: isn’t your voice more worth, than mine? [15:07] apachelogger, :-D [15:08] bulldog98: you speak on behalf of the kubuntu-devel subscribers, I'd say their combined voice is worth more than any single one ;) [15:08] rbelem: can't that move to github or git.kde or fdo if at all sensible [15:09] apachelogger, i think github is ok [15:09] rbelem: depends on whether you expect a lot of contributions from people who have a kde dev account [15:09] apachelogger, what is fdo? [15:09] in which case git.kde makes more sense TBH [15:09] freedesktop.org [15:10] ah! [15:10] Riddell: download update from kubuntu-ninja ppa, after install write about result [15:12] apachelogger, hum... but git.kde is not for kde only stuff? [15:14] apachelogger, fdo seems to be nice [15:14] rbelem: nice is a bold statement for fbo... [15:15] :-D [15:16] Mamarok, i never used fdo services [15:16] apparently KDE 4.8.1 is dripping in very slowly, I still only see part of the packages since yesterday [15:17] rbelem: git.kde.org depends on the project you want to host [15:17] rbelem: there is kdesupport [15:17] if it is somehow related to KDE it definitely is the place [15:18] rbelem: mind that fewer people have an fdo account [15:18] so from where I am standing you'd either go to github or git.kde [15:18] fdo is a right hassle [15:20] Mamarok, the project is a copyright scanner for source code. It will generate spdx and dep5 outputs [15:20] sounds like quite appropriate for git.kde.org [15:20] oh oh oh [15:20] or linuxfoundation [15:21] they also host crap like that [15:21] apachelogger, github seems to be more agnostic [15:21] :-D [15:21] ABI compat checker comes to mind [15:21] apachelogger, they already host another tool that does a similar job, fossology [15:21] rbelem: IMHO agnosticness is not a concern, from where you expect/want contributions is what matters ;) [15:22] apachelogger, hum... that's true [15:22] kde hosts like 30000 audio players ;) [15:22] fossology is written in c and php [15:23] Oo [15:23] eh [15:23] Oo [15:23] and needs rdbms [15:23] :O [15:23] :@ [15:23] -.- [15:23] /o\ [15:23] WTH [15:23]