=== ximion2 is now known as ximion [00:21] hmm persia, any idea how to FORCE make to look for includes somewhere else? CPPFLAGS=-I/usr/src/linux-headers-2.6.31-608-imx51 doesnt work === ximion is now known as ximion2 [00:23] oh [00:23] It found it [00:23] but FTBFS [00:23] http://paste.kde.org/86929 [00:23] :S [00:29] * Quintasan goes to bed [00:29] Good night. [00:29] sweet dreams [00:29] no faceplants! [02:03] Quintasan, I don't think that it isn't finding includes. For example, if I create foo.c as "#include \n\nint main() {\n return 0;\n}\n" and call `gcc -o foo foo.c`., I'll get "foo.c:1:22: fatal error: myheader.h: No such file or directory". [06:20] morning [08:12] Quintasan: thanks for the fix! [08:12] didrocks: Sup, no problem :) [08:12] Quintasan: ok, I'll add the sni patch as well and then build :) [08:36] apachelogger: persia did some licensing magic with guys on #linaro and we might get the driver into archive [08:38] Not done yet: there is only commitment to discuss what is required to have a sane license. [09:21] morning [09:28] GRRRRR [09:28] bambee: \o [09:28] yo! === hunger_ is now known as hunger [09:31] persia, apachelogger: driver built! [09:33] and installed [09:33] * Quintasan tries [09:43] * Quintasan does the victory dance [09:43] persia, apachelogger: X starts with the imx driver [09:45] getting sleep was good, wasn't it? [09:45] :-) [09:46] Most likely it was :) [09:46] Division by zero in kernel. [09:46] [<8003e3fc>] (unwind_backtrace+0x0/0xe0) from [<8020599c>] (Ldiv0+0x8/0x10) [09:46] Division by zero in kernel. [09:46] [<8003e3fc>] (unwind_backtrace+0x0/0xe0) from [<8020599c>] (Ldiv0+0x8/0x10) [09:46] lol [09:48] ... [09:48] It works, but nothing changed [09:50] * Quintasan awaits apachelogger arrival [09:52] linaro@linaro:~$ DISPLAY=:0 glxinfo | grep direct [09:52] direct rendering: Yes [09:52] yeah [09:52] * Quintasan goes to play HoN [09:56] Nice! [10:01] Wait. How did you get HoN for armel? [10:05] * apachelogger wonders that too [10:06] Quintasan: did you put the intstructions on the quickstart board wiki page? [10:06] I don't think the instructions are sane enough yet. Supposedly, installing the hwpack installs the right bits (not that this helps with images or anything) [10:13] persia: well, just so that one can replicate it onto different dev setups ;) [10:14] If you like. paulliu was talking about getting something into multiverse, which I think would be better. [10:15] certainly [10:16] Of course, if you want hawtn355 *NOW*, then you'll have to bug Quintasan :) [10:17] (or repeat the discussion with paulliu) [10:21] apachelogger: There is no improvement after installing the driver [10:21] :/ [10:37] Quintasan: course not [10:37] you need to start with --opengl or somesuch [10:40] apachelogger: how much do you have to pay if you overdraw? [10:42] too much === apachelogger is now known as rohansgoogle [10:42] [1] .prn [10:42] [2] more .prn [10:42] [3] ~20 USD [10:42] [4] fluffy === rohansgoogle is now known as apachelogger [10:43] apachelogger: i didn't find any info on this on google btw, and the prepaid site says that you can't overdraw, so conflicting info everywhere [10:43] right, except that there is currency exchange involved [10:43] which makes just about everything possible [10:44] "can't" implies that the system security is perfect. This is the real world. Doing so usually requires ingenuity (but coincidence can happen) [10:44] like flying chickens [10:44] heh [10:48] apachelogger: start what with --opengl? [10:49] plasma-mobile --opengl [10:52] apachelogger: http://paste.kde.org/86995 [10:54] QEgl::display(): Cannot initialize EGL display: "Success (0x3000)" [10:54] QEglContext::chooseConfig(): Could not find a suitable EGL configuration [10:54] something still is not quite right [11:08] apachelogger: http://i.imgur.com/yIMti.png [11:20] oh you touch my tralala [11:27] didrocks: there is no need to merge qtcreator, all changes are in the Debian package [11:27] but it needs a new qtwebkit version [11:27] debfx: nice! so basically, once the new Qt is published, we just need to sync? [11:28] (qtwebkit coming with the new Qt merge, right?) [11:29] either we need to modify the build-depends in qtcreator or we wait for qtwebkit 2.2 [11:29] *2.2 beta 1 [11:30] debfx: maybe I'm misleading, is it something different from libqt4-webkit ? [11:30] libqtwebkit4 I mean [11:31] (I don't find qtwebkit-source in debian) [11:32] didrocks: qtcreator build-depends on libqtwebkit-dev (>= 2.1.0~2011week13-2) [11:32] the source package is called qtwebkit in debian [11:32] apachelogger: Any ideas about that EGL [11:32] debfx: ok, so we just need to merge/sync from there? [11:32] Quintasan: well, on PVR you'd need to fiddle witht eh libEGL so files no idea how it works for freescale [11:33] a glimps at the ubuntu example image surely would help [11:34] didrocks: 2.2 beta 1 will probably be released on monday, I'd like to wait until it's packaged in debian and then merge the package [11:35] debfx: oh sure, no hurry then :) [11:35] do you use qtwebkit in unity-2d? [11:40] yofel: ping [11:44] Quintasan: if you could write down what you did to get it building I might be able to beat it into shape over the weekend [12:11] debfx: sorry, didn't see, no we don't [12:23] apachelogger: install the .38 kernel and headers [12:23] Quintasan: wiki! [12:23] apachelogger: I'll tell you and you write it down [12:24] I am gone in a bit moving my entire life north [12:30] apachelogger: https://wiki.ubuntu.com/ARM/iMX53QuickStartBoard [12:30] There you go [13:07] DarkwingDuck: Would you please have a look at Princess Butterfly's email to kubuntu-devel re docs and get things straightened out? [13:26] apachelogger: pong [13:28] * yofel takes a look at kdepim [13:38] yofel: How's the splitting going? [13:40] as you see on the wiki, not much progress. I'll do some more later, first I need to fix kdepim [13:43] !ninjas [13:43] Ninja Time! apachelogger, bulldog98, debfx, JontheEchidna, Lex79, maco, neversfelde, nhandler, Quintasan, rgreening, Riddell, ScottK, stalcup, txwikinger, yofel [13:43] You all will notice from the wiki that yofel isn't getting a lot of help. [13:44] This is a big job, so please try to find a little time to knock part of it off: https://wiki.kubuntu.org/Kubuntu/Ninjas/Packaging [13:48] Getting to it [13:51] yofel: I'll be taking okular [13:51] sure, thanks [13:53] Tonio_: ^^^ you too. [13:55] btw. can we recompress oxygen-icons? Using xz would reduce the size from 356MB -> 201MB [13:57] ScottK: ^ [13:58] yofel: The source package? [13:58] yep [13:58] It's better not to. We should keep the source the same as it's shipped unless there's a strong reason to change. [13:58] k [13:58] I'll do oxygen-icons if it's a problem. [13:59] That's not too hard for me to upload. [14:26] Don't I basically have to rewrite rules for okular? [14:27] you have to rewrite the whole package except for the install and symbol files and the binary part of the control file [14:28] At least I already created a project for okular to hold the packaging branch, so you don't have to do that [14:30] * Quintasan goes for a quick walk [14:39] !@#$%$$#@@!#$ [14:39] damn tests [14:39] yofel: Unable to find executable: /tmp/buildd/okular-4.6.90/obj-x86_64-linux-gnu/tests/shelltest.shell [14:39] Complaints to Dirk? [14:40] lemme try [14:40] want my packaging? [14:41] * Quintasan waits [14:41] ScottK: Is it okay to use dh8 there? [14:41] short rules etc [14:42] Yes, but try to stick with the style we got from Debian in the last merge. [14:42] Okay, awesome [14:42] We want this to be close to what they will eventually do. [14:42] Not enough ninjas online [14:43] ScottK: Also, any idea when Lex is coming back? [14:43] Nope. Last I heard was 'soon'. [14:43] Quintasan: It might be nice if you mailed him about it. [14:43] That way it's not just me asking. [14:43] * Quintasan includes hugs in his mail [14:44] Has someone an idea why I fail compiling kdevplatform in oneiric with no rule for Target »/usr/lib/librt.so« [14:44] yofel: !@#$% [14:44] It's really not there [14:44] git == magic [14:44] yeah, but where was it... [14:44] It's not in the tarball even [14:45] wait [14:45] tests/ [14:45] shelltest.cpp [14:45] :/ [14:46] ScottK: debfx: local build of the merged Qt is fine, and I tried the transition without any breakage. So uploading now [14:47] \o/ [14:47] Since okular is landing in main we can't ignore tests, can we? [14:47] nope [14:47] ... [14:48] .PHONY: override_dh_auto_test [14:48] in kdegraphics [14:48] :D [14:48] Or wait [14:48] I'm most likely using wrong stuff [14:48] include /usr/share/pkg-kde-tools/qt-kde-team/2/debian-qt-kde.mk [14:48] right? [14:48] didrocks: okay [14:48] right [14:48] * Quintasan rewrites [14:49] as well as the 2 vars that are defined for libokularcore0 [14:49] kde-runtime though, not kdebase-runtime [14:52] * Quintasan wonder if okular-dbg should depend on libokularcore0 [14:55] $(overridden_command) $(foreach p,$(shell dh_listpackages | grep ^lib),-p$p) -- -xkdebase-runtime <-- can someone explain this to me? [14:56] where did you get that from? [14:57] kdegraphics rules [14:58] ah, it tells dh_shlibdeps not to create any dependcies on kdebase-runtime [14:58] (I think) [14:58] change that to kde-runtime [14:59] has someone problems with multiarch in oneiric? [14:59] how does one even use multiarch? [14:59] yofel: http://wiki.debian.org/Multiarch/Implementation [14:59] follow the guide it will work on oneiric [15:00] but my problem is that cmake doesn’t like that [15:00] ah, to read for later [15:00] didrocks: Great. [15:01] bulldog98: are you sure that you cmake file is searching in /usr/lib/${CMAKE_ARCH_TRIPLET} ? [15:01] *your cmake file [15:01] yofel: I use the standard ones (trying to build calligra, kdevelop) [15:02] no idea [15:02] I'll look at it later once I've finshed reading through the kdepim merge and am done fixing it [15:02] yofel: ok [15:03] bulldog98: ah, now that you're there: is there a reason why you don't install usr/bin/akonadi_nepomuk_email_feeder ? [15:03] it's in the install file bug you commented it out [15:03] *but [15:03] yofel: no there is none must have been an accident [15:03] k [15:07] Which kit is it that controls privilege escalation for our GUI apps? Consolekit? [15:07] PolicyKit I think [15:07] That's the one. [15:08] Thanks. [15:08] It's pissing me off at the moment. [15:08] Why is it if you have multiple users that have admin rights it can't assume that if one of them is the one you are currently logged in as that's by default the user it should use? [15:08] So each time I deal with it I have to pick. [15:09] Let's make a sensible default and then let the user change if they want. [15:09] Where's the taco man when you need him? [15:10] ScottK: well especially when you tell it to rmember and it keeps asking.. even if had to select the first time. :( feel your pain [15:10] can't we just wrap things in kdesudo as we did before? policykit feels broken since a few releases ago [15:10] NO [15:10] yep, that's esp. annoying [15:10] I think getting poilcykit right is the way to go. [15:10] KDESudo is broke even more codewise [15:11] great, choose between crap and garbage [15:12] ScottK: hum? The gnome frontend pick the current user running AFAIK. So it's technically possible [15:12] didrocks: Thanks. [15:12] ScottK: Any idea where usr/lib/kde4/kio_msits.so from okular should go? okular.install? [15:12] I agree if the KDE frontend doesn't do that, I can feel the pain :/ [15:13] Sounds like a good job for someone who knows a bit of C++. [15:13] Quintasan: it's in okular-extra-backends, leave it there [15:13] It's in listmissing :P [15:13] CHM backend goes there as well? [15:14] yofel: ^ [15:14] hm, great, the debian folks didn't install it [15:15] Quintasan: okular-extra-backends [15:15] moving [15:17] wth, can't find it in the install files now, but apt-file says okular-extra-backends [15:17] Fair enough [15:18] I did discover in my quick google about it that Fedora gives it it's own package while opensuse and mandriva include it in okular. [15:18] Debian has it in okular-extra-backends, so that's where we should put it. [15:20] * Quintasan builds [15:21] Should be ready in a minute [15:21] tests build? :P [15:21] ofc [15:21] :O [15:21] dpkg-buildpacakge -j32 FTW [15:21] . . . [15:21] * Quintasan needs more RAMZ [15:22] tmpfs will speed this up even more [15:22] k [15:22] new symbols! [15:24] One more and uploading [15:25] Too much test builds? What's that? [15:25] fooey, how did bulldog98 manage to create a kdepim packaging branch that's incompatible with the main one o.O [15:25] * yofel goes merging by hand [15:25] Don't ask me [15:26] yofel: have I really managed that? [15:26] http://paste.kde.org/87079 [15:30] yofel, ScottK: review please https://code.edge.launchpad.net/~kubuntu-packagers/okular/ubuntu [15:31] * Quintasan is sure that he didnt do anything stupid as usually [15:32] not fair, you actually had a usable copyright file to start with [15:33] :3 [15:33] I just removed files which were not there after split [15:33] sure that all the folks mentioned above actually have copyrights for okular? [15:33] * Quintasan kills it flies with fire [15:35] yofel: what does that error message mean? [15:35] yofel: crap [15:35] * Quintasan fixes [15:36] bulldog98: usually happens when you have a repos on launchpad in an old format and try to stack a branch with a recent format on it, no idea why it happens here though [15:37] * Quintasan notes licensecheck should have an option to list authors [15:37] Quintasan: --copyright [15:38] the output format isn't really usable though :/ [15:46] yofel: copyright updated [15:56] maybe change the packaging licensing to us, other than that, looks fine IMO [16:02] * yofel tries to continue with smoke while kdepim builds [16:04] yofel: Isn't 4.6.4 in ~kubuntu-ppa? [16:04] ~kubuntu-ppa/ppa [16:04] yofel: Can you get rid of it in ninjas? [16:04] sure [16:05] does clutter the page indeed [16:05] \o/ [16:05] i386 build of okular finished [16:05] * Quintasan takes a short break [16:06] already? [16:06] the oneiric builders somehow seem faster than the natty ones === mgraesslin_ is now known as mgraesslin [16:36] jussi: pokey pokey pokey [16:45] fregl: you rock! [16:47] * fregl blushes [16:47] and I thought it was didrocks :) [16:48] fregl: heh, thanks for the fixes! :-) [16:48] seems it was almost there from the first release, it's really nice :) [16:49] \o/ [16:49] fregl: btw, did you receive my email about the examples licensing of qt-at-spi? [16:49] * fregl pokes apachelogger [16:50] didrocks: uhm, let me check - right now I sometimes ignore mail for a while to not get too distracted... [16:50] fregl: no hurry anyway, especially not on a Friday evening. You should rather enjoy a beer for doing such a good work :-) [16:50] ah, right, I did [16:51] Kaleo: nice work as well on the unity-2d side! ;) [16:51] almost - today is Lønningspils - lots of trolls having a beer in norway ;) [16:51] :) [16:52] heh [16:52] didrocks: about the examples - actually they are so trivial, I'll simply put the lgpl header in there as well and be done with it. [16:52] fregl: excellent! that sounds like the easiest thing to do, right :) [16:53] back in the days Riddell already poked me to fix licenses... I thought I had done that everywhere... [16:55] I think no one never finishes to mess with licenses :-) [16:56] didrocks: ok, I hope they're all fine now, let me know if you find more stupid things I did there :) [16:57] fregl: I won't hesitate to bother you with that, don't worry :-) [16:57] damn :p [16:58] fregl: heh, thanks for the commit, it looks good ;) [16:58] yw [17:01] Quintasan: seems we have a FTBFS on arm for Qt. IIRC, there were some patch that were removed to prefer the debian's one [17:01] Quintasan: https://launchpad.net/ubuntu/+source/qt4-x11/4:4.7.3-3ubuntu1/+build/2589479/+files/buildlog_ubuntu-oneiric-armel.qt4-x11_4%3A4.7.3-3ubuntu1_FAILEDTOBUILD.txt.gz FYI [17:02] That was kubuntu_23_arm_memory_barriers.diff [17:03] Quintasan: can you have a look if we need that back? I'll sponsor you if needed [17:03] I think we'd better ask NCommander since I have next to no idea about ARM magic [17:05] sure, let's wait for him then, I have no idea on those arm patches as well [17:06] Besides, /tmp/ccCOH0VH.s:3278: Error: selected processor does not support Thumb mode `swp r6,r4,[r3]' [17:07] It's some Thumb2 IIRC [17:07] NCommander: http://bazaar.launchpad.net/~kubuntu-packagers/qt/ubuntu/revision/177 <- here are my changes [17:14] What happened to x-0003-Use-GCC-intrinsics-for-armv6-atomic-operations.patch [17:15] fabo told me we don't want this and want armv6_Include_explicitly_IT_instructions.patch instead [17:15] Quintasan: Look at the diff: http://launchpadlibrarian.net/74027553/qt4-x11_4%3A4.7.3-1ubuntu4_4%3A4.7.3-3ubuntu1.diff.gz - search for ldrex and see how many times it was included before and how many now. === emma is now known as em [17:16] Now ~2 times [17:16] Before ~6 times [17:16] I think armv6_Include ... replaces kubuntu_22_thumb2_support [17:16] ... [17:17] So I'm not sure, but I think that's likely the source of the problem. [17:17] ldrex 8 times in x-0003-Use-GCC-intrinsics-for-armv6-atomic-operations.patch [17:19] I also don't understand Fix_builds_with_compilers_without_--with-fpu_neon_as_default.patch [17:20] Actually that one may be OK. [17:23] Yeah. I think that's fine. [17:32] ScottK: pong [17:32] it's a day off in finland [17:32] Did I ping you? [17:32] just back home [17:32] Welcome. [17:33] ScottK: arm build failure, isn't it a ping ;) [17:33] Right. [17:33] x-0003-Use-GCC-intrinsics-for-armv6-atomic-operations.patch should have been deprecated [17:33] That's not included in our current build (that failed) [17:34] ScottK: Qt/armhf in debian is more or less Qt/ubuntu arm [17:34] Yep. [17:34] it builds as expected on armhf [17:34] So I'm confused. [17:35] and I built Qt on ubuntu arm builder successfully [17:35] fabo: It's not quite the same. The semantics for floating point differ quite a bit, which seems to affect alignment in odd ways. [17:36] FOund it [17:37] we're waiting :) [17:37] didrocks: +ifeq ($(DEB_HOST_ARCH),armhf) needs to include armel too [17:37] hum indeed :) [17:37] +ifeq ($(DEB_HOST_ARCH),armhf) [17:37] + extra_configure_opts += -arch armv6 [17:37] +endif [17:37] rules is a bit different for ubuntu [17:37] I'll fix. [17:38] ScottK: you're right :) [17:38] thanks ScottK [17:38] I had it in mind that you need to change this for ubuntu merge [17:38] Yep. [17:38] It's an easy one to miss. [17:39] * ScottK points ^^^ out to Quintasan. [17:39] ok, I can go back to my day off :) [17:39] Yes. Thanks for giving me the important clue. [17:40] fabo: enjoy ;) [17:40] ScottK: nice catch! [17:40] ScottK: Isn't that what I merged from Debian? [17:40] I didn't notice that one, sorry [17:40] Quintasan: Look at the diff. [17:40] It happens [17:41] iirc NCommander said -arch armv6 isn't necessary anymore [17:41] Yes, that's why I merged that change [17:41] I don't know why he said that. [17:41] For Qt you have to tell it to use the v6 stuff. [17:41] I guess we'll know for sure after I upload this. [17:42] Argh. armel, armhf it was before [17:42] * Quintasan notes that down [17:44] Quintasan: It was an easy thing to miss. [17:45] ScottK: Can you note that down in KUBUNTU-DEBIAN-DIFFERENCES? [17:45] BTW, I think referring to that in debian/changelog is cheating. [17:46] It's hard to tell later exactly what was there then. [17:48] If Ubuntu gets armhf, it will merge more cleanly. Ubuntu and Debian use "armel" to mean completely different things. [17:48] Hmm. [17:49] armhf == ARM hard float? [17:49] persia: We're including armhf compatibility in our packages for when that happens. [17:49] Yes [17:49] Updated the difference file too. [17:50] Basically anywhere Debian says armhf we should have both. [17:50] Okay. [17:51] ScottK, That's the best plan. No idea if/when, but at UDS lots of people wanted to do it. I think it depends on more buildds, more than anything else. [17:51] Not entirely coincidentally I'm helping in Debian make sure KDE builds on armhf. [17:53] Uploaded. [17:56] ScottK: Thanks. [17:57] once upon a time [17:57] Quintasan: did you get EGL to init? [17:57] apachelogger: Did you? [17:58] Take that as a "No." [17:58] No idea what to do with it [17:58] well, I am building packages right now [17:58] I'd rather not moot my setup as otherwise I had to rebuild me KDE [17:58] I'm not even sure what software is at fault [17:58] Is it the driver or KDE is doing some magic? [17:59] apachelogger: you pinged me before? [18:00] yofel: see kdepim related mail on kubuntu-devel list [18:00] sawn, working on it [18:00] I can reproduce that, no idea if the thing should be there though [18:00] ok perfect [18:00] it was before, so it's a regression [18:00] Quintasan: your driver setup is mooted most likely [18:00] Huh? [18:00] Quintasan: if you get the graphics sdk (which contains demo apps for gles) they will likely also not work [18:01] apachelogger: It works on your board? [18:01] no [18:01] but I had a similar issue with the n900 [18:02] what I find funny: earlier today freescale had the 11.01, 11.03 and all new 11.05 images for the quickstart [18:02] now I only can see the 11.01s [18:03] BTW, don't panic over what happens to the 4:4.7.3-3ubuntu1 builds for Qt. I had lamont kill them with fire. [18:03] trying to overwrite '/usr/lib/libEGL.so.1', which is also in package libegl1-mesa 7.10.2-0ubuntu2 [18:03] there we have it [18:03] boooh [18:04] * apachelogger introduces appropriate relationships [18:04] apachelogger: What are you exactly doing? [18:04] building packages [18:04] ... [18:05] didrocks: ^^^ (see my last comment) [18:06] ScottK: thanks :) [18:06] apachelogger: Quintasan: The OpenGL ES is provided by seperated library. It have to replace the mesa lib [18:07] I know [18:07] Maybe that's why it doesnt work here [18:07] perhaps [18:07] I still have the mesa-egl [18:07] Quintasan: did you install the amd-gpu-bin tar? [18:07] I did [18:07] the wiki only sez -x11 [18:07] They both contain the same stuff [18:08] they do? [18:08]