[00:19] yofel: [ Philip Muškovac ] [00:19] * Bump build-dep on python-sip-dev to >= 4.14 [00:19] in pykde4 [00:19] yofel: do you remember why? [00:20] It would have been to prevent archive skew. [00:20] so, can be dropped? [00:20] Yes. We should be the same as Debian. [00:20] because pykde4 for quantal is FTBFS'ing [00:20] ack [00:21] Hmmm. [00:22] I'm wrong. [00:22] oh? [00:22] if(SIP_VERSION STRLESS "040e00") # These version numbers also appear in ../CMakeLists.txt [00:22] message(FATAL_ERROR "The version of SIP found is too old. 4.14 or later is needed.") [00:22] endif() [00:22] drat :/ [00:22] So that's something you need to keep. [00:22] too late, I uploaded for quantal [00:22] but it'll still fail obviously [00:22] Sorry [00:23] that's fine, it's in staging [00:23] backport python-sip-dev? [00:23] I wonder how much trouble that'll cause [00:25] * shadeslayer backports from raring [00:26] oh hurray [00:26] pykde4_4.10.90-0ubuntu1~ubuntu12.10~ppa1_source.changes rejected [00:26] no harm done :) [00:26] You'll need python-qt4 too then. [00:26] okay [00:26] if(PYQT4_VERSION STRLESS "040905") # These version numbers also appear in ../CMakeLists.txt [00:26] message(FATAL_ERROR "The version of PyQt found is too old. 4.9 or later is required.") [00:26] endif() [00:27] Don't recall what's there, but you generally want sip4 and python-qt4 paired with what upstream released together. [00:28] I see [00:28] I've backported the official raring packages [00:28] should be enough [00:31] http://qa.kubuntu.co.uk/ninjas-status/build_status_4.10.90_quantal.html < mostly sorted out :) [00:38] hm, everything should auto resolve [00:38] will test later today then [03:44] I am on vacation. [03:44] Sweet. [04:16] Anybody home? [05:23] manchicken: No. [05:26] ScottK: lies! [05:26] I need something to do. [05:26] Do people use kdevelop? [05:26] Some do. [05:29] What about Qt Develop? It looks like that's shooting at Xcode or Android Studio, seems like it's mainly pointed at Ubuntu Touch and qml. [05:33] The Ubuntu Touch stuff is (as you might guess) an Ubuntu only plugin that's added. [05:33] It is mostly aimed at Qt/QML. [05:34] no JT :( [05:34] someone do something to bug 1183615 plz :( [05:34] bug 1183615 in kde4libs (Ubuntu) "package kdelibs5-plugins (not installed) failed to install/upgrade: não pode copiar dados extráidos para './usr/lib/kde4/libexec/ksendbugmail' para '/usr/lib/kde4/libexec/ksendbugmail.dpkg-new': fim de ficheiro ou stream inesperado" [Undecided,New] https://launchpad.net/bugs/1183615 [05:34] apachelogger: But you have a willing coder awaiting enlightement. [05:35] :O [05:35] and that at 7:35 [05:35] manchicken is looking for work to do. [05:37] manchicken: https://trello.com/c/8T4OXfLX could be something [05:37] essentially bzr clone lp:kubuntu-debug-installer [05:39] right now it tries to find debug packages using the already active repositories, what we need it to do is allow us to add repositories on-the-fly [05:39] e.g. see https://wiki.ubuntu.com/DebuggingProgramCrash#Debug_Symbol_Packages [05:39] Bonus points for detecting if XMir and declining to file kwin bugs if it is. [05:40] TLDR: most debug packages are autogenerated and put in a separate package repository, so to get access that other repoistory needs to be added to apt [05:40] ^^ [05:40] ScottK: I think that is another card TBH [05:41] But we can't add it by default because if everyone enabled that repository, it would get crushed. [05:41] ^ that [05:42] so I was thinking ... debug-installer gets started -> tries to find debug packages for files -> fails -> adds the ddebs repository -> updates apt cache -> tries to find debug packages -> failure/success, regardless it removes the repository again -> updates cache [05:43] but I guess you could just play around with it a bit see where the code takes you [05:45] ScottK: trello says we should discuss the MRE list on kubuntu-devel btw [05:45] https://trello.com/c/JT78Qcz3 [05:46] OK. [05:55] * manchicken looks... [06:00] !info user-manager saucy [06:00] user-manager (source: user-manager): user management tool for the Plasma workspace. In component universe, is extra. Version 0.0~git20130223-0ubuntu2 (saucy), package size 41 kB, installed size 203 kB [06:01] :/ [06:01] afiestas_: didn't you do a release? [06:02] apachelogger: So we're trying to modify the crash handler so that it will install dbgsym packages in the absence of dbg packages? [06:02] yes [06:02] well [06:02] TBH it probably should just always use dbgsym [06:03] dbgsym is just the debug symbols for the existing binaries? [06:03] Does it replace the binaries with ones using -g flags? [06:04] nope, essentially it's the debug symbols of the binaries [06:05] Is that a supplemental file, similar to a dSYM file on OS X? [06:05] apachelogger: A new change though is that if a -dbg package exists, the -dbgsym is and empty dependency package that depends on the -dbg. [06:05] I've never done debug symbols on GNU/Linux without it being compiled in. [06:05] No point in going to the -dbgsym if the -dbg is available. [06:06] so the binaries are built with -g by default and then some packaging tool strips all debug symbols into a shadow binary (i.e. instead of one binary you have two, one with the functional code and one with the debug code). gdb can then put those files back together and allow debugging of a binary that technically has no debug symbols [06:06] manchicken: dsym sounds similar enough [06:06] apachelogger: Okay, that makes sense. [06:06] ScottK: actually even more so [06:06] But that actualy doesn't matter for this project, right? We're just dealing with package installs. [06:06] if dbgsym is just a dep on dbg it makes even more sense to only go for dbgsym [06:07] Maybe. [06:07] Yes, just package installs. [06:07] that way you have consistent handling and the package relationship will take care of -dbg packages [06:07] OK [06:07] so we install only -dbgsym but indeed -dbg gets in as well [06:07] manchicken: pretty much [06:07] Let me look at this existing package. I think step 1 is figuring out what we already have. [06:08] yeah, bzr clone lp:kubuntu-debug-installer [06:09] it's a really simple program ... it gets a list of files that could not be traced (i.e. no debug symbols present) and it uses qapt to find the packages containing those files, then it tries to find the -dbg/-dbgsym package associated [06:09] manchicken: general code refactoring is also very welcome, that was just quickly hacked together and since part refactored once or twice ^^ [06:11] apachelogger: While it's loading (slow internet), is this C++, Ruby, Python, other? [06:11] manchicken: c++ [06:12] bug 103378 [06:12] bug 103378 in kubuntu-settings (Ubuntu) "impossible mount floppy device in /media" [Undecided,Confirmed] https://launchpad.net/bugs/103378 [06:12] I'm getting a permission error. [06:12] Mount Wars 2: The Floppy Strikes Back [06:13] Evidently it's attempting to send my SSH key and then rejecting it. [06:14] manchicken: launchpad.net [06:14] check that the right public key is available there [06:14] and that the right key is used on your end I guess [06:14] I'm gonna need to generate a new one I think, I doubt that I still have the old one lying around... [06:14] I'll check my old FSF thumb drive, that's probably where I would have put it :) [06:16] Also make sure the username is right. [06:17] Host bazaar.launchpad.net [06:17] User $USERNAME [06:17] in ~/.ssh/config [06:18] If your LP username and your local username aren't the same. [06:29] Making a new key it is! [06:30] My FSF thumb drive is corrupt :) [06:30] Glad I only stored keys on it. [06:34] I'm in business! [06:34] I think [06:35] That said, this DL is going to interrupt The Doctor. [06:36] NO, NOT THE DOCTOR :O [06:36] Yeah, I just started the new one. S1E1. [06:37] I know I'm behind the times... but there are only so many geek-culture references I can miss before I can stop ignoring things. [06:37] shadeslayer, Riddell: https://trello.com/c/DplmVapI that does nto seem complete at all........................ [06:38] manchicken: ^^ [06:38] apachelogger: I don't understand what you mean by "^^" [06:38] manchicken: I'm probably a close to two decades older than you and I watched the Doctor in college. [06:38] How can you have missed that one? [06:39] ScottK: I am 31... and three children old. [06:39] Right, so 1 year short of two decades older and three also. [06:40] Heh [06:40] http://www.urbandictionary.com/define.php?term=%5E%5E [06:41] * valorie started on season one as well, but I'm halfway through [06:41] the Beeb has done a nice job on cleaning up those old tapes [06:42] You're a brave man, trusting anything to urbandictionary.com. [06:42] I wasn't sure if you were doing the ^_^ thing, or if you were trying to point up. [06:43] Not brave. apachelogger's the right kind of insane for urbandictionary. [06:44] Nice [06:44] manchicken: pointing up is ↑ [06:44] compose key magic [06:44] Oh, cmake... I forgot about cmake. [06:45] seamake [06:45] I like the compose key. [06:45] today in apachelogger's package wonderland: flooding launchpad with projects to import git repos [06:45] I still pretend like I speak a little french, it's useful for that. [06:46] oh la la [06:46] Oais :) [06:46] Ouais [06:47] Please note: I only pretend to speak a little French, never said I could type. [06:48] uhm [06:48] yofel_, Riddell: it would appear we lost the kwallet kde-runtime patches [06:49] I've got a buddy who is the language chair of Illinois Wesleyan University. He's helping me improve my language skills. The goal is to get work to transfer me to Montreal or Paris for a year. [06:51] paris ftw? [06:52] well, unless one is into maple syrup I guess [06:52] omnomnom [06:52] I'm actually not picky. [06:52] Montreal is beautiful though [06:52] I want to go somewhere to actually use my French, not terribly picky where. [06:53] my friend from Paris tells me that the french is not much like what is spoken in France [06:53] Montreal would be preferred since it's within driving distance, however. [06:53] and Paris is not? :O [06:53] ah well, I guess Paris is beautiful in a different way ^^ [06:53] more like what was spoken a few hundred years ago [06:53] yes, Paris is amazingly beautiful [06:53] Yeah, my buddy says that Québecois french is much more antiquated than French french. [06:53] good morning [06:54] Morning. [06:58] LMFAO, garbage can eats a dude... obligatory burp... [07:01] It's been a while since I've seen C++ [07:01] Outside of that stuff we have at work, where we use C++ as nothing more than OO around our C. [07:03] manchicken: their french hasn't received updates for quite a while I guess [07:03] imagine the security problems [07:03] quassel is only lagging a bit [07:12] apachelogger: Indeed! They're a commonwealth afterall. [07:14] They need a hotkey to flip between workspaces like you can Alt-Tab [07:16] manchicken: meta+tab? [07:16] depending on what you mean by workspace [07:17] apachelogger: Only rotates through them. Doesn't start with the most recently active one like alt+tab does. [07:20] ah yeah, that doesn't exist === jacky is now known as jalcine [07:26] It should. I use workspaces a lot. [07:31] manchicken: file a wish bug on bugs.kde.org I guess [07:33] apachelogger: Well, since it's a feature I would use, it's a clear omission which warrants the highest priority. [07:38] I agree. [07:38] doesn't change the process though :P [07:41] Also a problem. [07:41] I am loving Dr Who. [07:41] How did I live this long without this cheesy and amazing sci-fi? [07:43] manchicken: it was all leading up to this day [07:52] grumble... why won't this thing build? [07:52] CMake Error at CMakeLists.txt:10 (kde4_add_executable): [07:52] I know I'm missing something obvious... I just don't know what. [07:55] sudo apt-get install install kdelibs5-dev? [07:55] meh, GL issues in virtualbox -.- [07:56] Naw, that didn't work [07:58] soee: please do NOT backport pykde [07:58] manchicken: paste complete output please [08:01] http://pastie.org/8105989 [08:03] tab fail [08:03] sry soee [08:03] yofel_: hello [08:04] shadeslayer: please don't backport pykde4 to << raring unless you want to backport the whole SIP stack with rebuilds and want to do the python << 3.3 porting [08:04] yofel_, np [08:04] yofel_: btw, fwiw, considering we have native packages I'd not put the packaging revision as the actual package revision as that will block uploads if the packaging changed but not the source [08:04] e.g. https://code.launchpad.net/~neon/+recipe/project-neon5-attica [08:05] yofel_, I still can't get dh_strip working, it just wasn't even executed during build. [08:05] apachelogger: er, what? [08:06] we intentionally have revno:packaging in there so you can upload a new package with new packaging [08:06] that won't work if it is -revno [08:06] needs to be +revno-0 [08:06] no [08:06] the packaging is never part of the program source [08:06] ever [08:07] Riddell: is there no debug package for /usr/lib/x86_64-linux-gnu/libQtGui.so.4.8.4 ? I installed libqt4-dbg but still don't get line numbers [08:07] it's le native [08:07] native packages have no revision [08:07] yofel_, what do you think will make it work? [08:07] otherwise they are not really native are they [08:07] and also packaging is always part of the prgram source for native [08:07] because, well, they are native [08:07] project-neon5-attica_0.0+git20130703+r417~8f3b796-16~saucy1.tar.xz already exists in Project Neon KDE Frameworks 5 [08:07] that's the damn same packaging revno [08:07] yes [08:08] doesn't change the fact that the revision is wrong [08:08] with native packaging the -XXX part is relevant too [08:08] yes [08:08] relevantly wrong [08:08] NO [08:08] apachelogger: Any idea? [08:08] you uploaded project-neon5-attica - 0.0+git20130703+r417~8f3b796-16~saucy1 [08:08] yes [08:08] I am not arguing attica [08:08] now you tried uploading the damn same thing twice [08:08] I am arguing the version being shit [08:08] it's not [08:08] you can't just rebuild unchanged things [08:09] !language | apachelogger [08:09] apachelogger: Please watch your language and topic to help keep this channel family-friendly, polite, and professional. [08:09] if you change something please bump the version [08:09] in all uploads I see revno:packaging is 16, so what changed? [08:09] ubottu: SFU will you [08:09] apachelogger: I am only a bot, please don't think I'm intelligent :) [08:10] obvious [08:10] what you just did is the completely wrong thing to do [08:11] yofel_: http://www.debian.org/doc/debian-policy/ch-controlfields.html#s-f-Version [08:11] see debian_revision [08:11] apachelogger: just to note: in native packaging, the "-foo" part is not ignored but part of the source version itself [08:12] look at the tar filename [08:12] yes [08:12] you just never committed something [08:12] it is still wrong [08:12] I forgot to push the branchy [08:12] so nothing changed [08:12] it is still wrong [08:12] no it's not [08:12] it's packaging, so it should be - to represent that [08:12] It is optional; if it isn't present then the upstream_version may not contain a hyphen. This format represents the case where a piece of software was written specifically to be a Debian package, where the Debian package source must always be identical to the pristine source and therefore no revision indication is required. [08:13] "where the Debian package source must always be identical to the pristine source" [08:14] what's a pristine source? [08:14] the upstream source? [08:14] the debian/ folder is not part of the upstream source [08:14] the unaltered source [08:15] if it were upsteram source it would say upsteram not pristine :P [08:15] well, for native packages that doesn't exist. We only do - to represent that we add the debian/ package later [08:15] sure it does [08:15] native package == pristine source [08:15] non-native package .orig.tar.* == pristine source [08:16] apachelogger: libapt is what I was missing. [08:16] look, I'm trying to represent something that doesn't match the debian package workflow in a sane way [08:16] manchicken: oh that makes sense, I guess we are not checking properly for it [08:16] yofel_: sure it does [08:16] it's a native package [08:17] apachelogger: I'm calculating not from a source package POV but from a recipe POW [08:17] *POV [08:17] dude [08:17] SIGH [08:17] whatever [08:17] !calm | apachelogger yofel_ [08:17] Sorry, I don't know anything about calm [08:17] !?!!??!??!!??!?!??!?!?!!?!??!?!!? [08:17] smartboyhw: dude [08:17] for all I care make it +neon{revno:packaging} [08:17] that would be fine a swell [08:18] I also learned about .kdev_include_paths. [08:18] no one knows what that does [08:18] manchicken: as that debug installer thing is kde, maybe ask JT when you see him how you can use QApt [08:19] apachelogger: version string is now "0.0+git{date}+r{revno}~{git-commit}+neon{revno:packaging}", good? [08:20] yah [08:20] <3 [08:23] I'm not entirely unfamiliar with QApt... I still have fond memories of it in Adept. [08:23] s/fond memories/horrible nightmares/ [08:23] manchicken meant: "I'm not entirely unfamiliar with QApt... I still have horrible nightmares of it in Adept." [08:23] I suspect/hope it has changed since 2008 though. [08:23] manchicken: not the same QApt [08:23] in fact not related at all [08:24] Yay [08:24] Not worse, right? [08:24] no? :P [08:24] There's the confidence I needed, right there :) [08:25] What's the usage for kubuntu-debug-installer? kubuntu-debug-installer `which kdesudo` [08:25] ? [08:26] it's usually called from inside Dr. Konqi from the 'Install debugging symbols' button in the backtrace view [08:26] not sure how that's done [08:27] It looks like it takes a program, or maybe a debug package. [08:28] list of file paths [08:28] kdesudo may be a bad example though ^^ [08:28] anyway [08:29] Yeah, it finds a package for a file, then it tries to guess the debug package. [08:29] manchicken: for real examples you could just start konqueror and kill -s ill `pidof konqueror` [08:29] that should bring up drkonqi [08:29] I think I found it. [08:29] kdesudo kubuntu-debug-installer `which kate` [08:30] BAM! [08:30] apachelogger: This is your work, yeah? [08:31] in cooperation with JontheEchidna, yeah [08:32] manchicken: no kdesudo needed btw [08:32] apachelogger: It's rather nice for what you described as a quick and dirty. Nice and simple. Not too noisy. [08:32] :) [08:33] What do we need to do with this now then? [08:33] It seems like dbgsym packages are already included. [08:33] manchicken: see the wiki page [08:33] yofel_, still here? [08:33] manchicken: you need to add an entire repository [08:33] yes, but gone now. sorry [08:33] Oooh, so we need the ability to have this thing add a repo before doing its deed? [08:34] yofel_, :( [08:34] manchicken: what it does right now is simply take dbgsym into account, but to make use of that the user has to manually add the repo [08:34] * smartboyhw really needs someone on advice of why the hell dh_strip isn't run. [08:34] Wait a minute... Did you just RTFM me? [08:35] manchicken: yeah... simply put echo "repourl" >> /etc/apt/sources.list.d/foo && apt-get update && finddbgsympackages && if (found) install && rm /etc/apt/sources.list.d/foo && apt-get update [08:36] i.e. add the repo (unless it is present already <- fun part), update cache, do guessing, possibly install stuff, revert changes (drop repo and update cache again) [08:40] apachelogger: That doesn't sound all that tricky, actually. [08:40] So we don't want to keep the repo in the list even if we install a package? [08:40] What about CVEs? [08:41] how do CVEs play into it? [08:42] and I'd imagine that once that is working we may want a checkbox somewhere to allow permanent use of the repo [08:42] the problem is ... the repo pretty much doubles the time apt-get update takes, so it's not something we'd want to introduce if someone files a crahs report once an installation lifetime [08:43] What if there's some update to the package, we have debug symbols installed, but now they're out of date. [08:43] OTOH if someone is a tester of pre-release software it's something they'd probably want to keep [08:43] manchicken: the package gets thrown out [08:43] Good point. [08:43] (in theory - that requires verification actually) [08:44] Well, either it'd get thrown out or the user would hold it in conflict preventing an update, right? [08:44] usually though since the packages are not pulled in by relationships with another package they have a weak dep chain and would get removed to satisfy the upgrade request [08:44] manchicken: yes [08:44] Then what you said sounds pretty reasonable. [08:45] And then when it's done, have some dialog that says "we had to add a repository in order to get the debug symbols, would you like to keep it or trash it?" and then do what they say. [08:45] if it doesn't work we can probably do some apt magic to lower the package score or something [08:45] such that they would get removed regardless [08:45] manchicken: yeah, something like that [08:45] Okay, cool. [08:46] I'm gonna go to bed now, I've got almost 0400 on the clock. The boys aren't going to let me sleep all day :) [08:46] manchicken: hehe, good night === yofel_ is now known as yofel [08:47] smartboyhw: re [08:47] I'll start playing with this tomorrow after I finish my SQL Server trigger for a client. [08:47] yofel, hello. I still can't get dh_strip working and as you did said, it doesn't even seem to be run. [08:48] smartboyhw: can you plese run debian/rules clean, tar everything up and upload it to my server? I need to try this myself [08:48] yofel, I forgotten how to get to it... [08:49] ubuntu@yofel.dyndns.org, ssh port 2224 [09:04] yofel, just pick https://launchpad.net/~smartboyhw/+archive/ppa/+files/libkgapi_2.0.1-0ubuntu1.dsc instead [09:05] * smartboyhw watches yofel through ssh:P [09:05] wrong syntax... [09:05] :O [09:06] makefiles required tabs, not spaces [09:06] *require [09:07] smartboyhw: http://www.gnu.org/software/make/manual/make.html#Rule-Syntax [09:08] smartboyhw: now it worked ;) [09:08] yofel, sorry:) [09:08] * smartboyhw hates tabbing BTW [09:08] :P [09:08] everyone does, just gnu make likes it -.- [09:08] yofel, -.- [09:08] I think I haven't done the copyright info though.... [09:08] when writing makefiles is the only time my tab key actually creates a tab character [09:09] yeah, I esp. have a behaviour override for makefiles in vim [09:11] * apachelogger blinks [09:11] :/ [09:11] yofel, copyright done. [09:12] Well, all files belong to him it seems:p [09:13] yofel: https://code.launchpad.net/~neon/+recipe/project-neon5-dbusmenu-qt suggestions? [09:13] yofel, review + upload? [09:13] Well, that's all GPL-2+ [09:14] Alright, there are some differences it seems [09:14] yep [09:15] yofel, the tests it seems [09:15] not only tests [09:15] Jan is missing in general [09:15] please add then I'll continue [09:15] * smartboyhw cheats and adds Jan to * only:P [09:17] yofel, done? [09:17] smartboyhw: add yourself to the packaging copyright ;) [09:18] yofel, ... [09:18] For all the packages I've worked, this is the first time I'm asked to add my own name into debian/* :P [09:19] yofel, yeah, so many files without copyright:P [09:20] yeah, but that's fine from what I see [09:20] so add yourself and we're done [09:20] I think [09:23] smartboyhw: ok, let me do one last thing [09:23] yofel, alright. [09:24] * smartboyhw waits for the grand finale:P [09:25] What's eatmydata!?!?! [09:26] makes fsync() and sync() a noop when used. Speeds up disk writes that don't need to be safe [09:27] this is an SSD though so the effect isn't as great as it is with a HDD [09:27]