**** BEGIN LOGGING AT Sat Mar 05 02:59:56 2011 Mar 05 04:44:18 Could one vac into a meego box and use it like normal? Mar 05 04:44:21 like from an ipad? Mar 05 05:16:17 "vac"? Mar 05 05:16:41 xiambax: you mean "vnc"? Mar 05 05:16:51 Whoops Mar 05 05:16:52 Yes Mar 05 05:16:52 :D Mar 05 05:18:14 meego cloud. Mar 05 05:18:19 would be a rad idea Mar 05 05:18:43 i just wanna know Mar 05 05:18:50 Im torn between exopc and new ipad Mar 05 05:20:18 thing is, vnc is suffering a lot on high-latency and low-bandwidth links Mar 05 05:20:45 mmmmm Mar 05 05:20:46 i see Mar 05 05:20:53 id get exopc if it was dualcore Mar 05 05:20:54 but it is not Mar 05 05:21:46 for example: it's possible to watch a movie over vnc, but i did it only once, over a gigabit ethernet network and it still wasn't exactly smooth Mar 05 05:25:58 Do you have any recommends for a tablet I can run meego 1.2 alpha on? Mar 05 05:26:11 2gb ram and dualcore preferably Mar 05 05:27:31 I don't want an ipad. but i do Mar 05 05:27:43 I want to phase out all my apple products I think though Mar 05 05:31:33 not really Mar 05 05:32:59 i'd love to see something based on the 9W dual-core version of the bobcat. or the tegra3 Mar 05 05:51:38 I just want something fast Mar 05 05:51:42 around 600 to 700 bucks Mar 05 06:23:03 hi, :) Mar 05 06:23:21 morning wassupnari Mar 05 06:23:45 actually afternoon in here :D Mar 05 06:24:24 anyway, i built 1.1.90 version image and running on my device.. Mar 05 06:24:33 but i cannot run apps.. Mar 05 06:25:04 there are some.. libEGL and GLib-CRITICAL warnings Mar 05 06:25:15 do you know the reason? Mar 05 06:25:34 right - quick checklist: 1) what processor 2) are you running as user 'meego'? Mar 05 06:27:35 1) armv7 architecture 2) yes Mar 05 06:27:58 you might need to rm /usr/lib/qt4/plugins/styles/libmeegointernalqtstyleplugin.so Mar 05 06:28:08 if you don't have a GLESv2 implementation Mar 05 06:28:16 okay, i will try it Mar 05 06:28:31 that plugin acts a little weird so Mar 05 06:29:11 hm.. there's no libmeegointernal*.so file in /usr/.../styles/ Mar 05 06:29:31 hmm Mar 05 06:29:54 what app are you trying to run? Mar 05 06:30:04 i just have libmeegotouchqtstyleplugin.so Mar 05 06:30:06 ah Mar 05 06:30:09 rm that one instead :P Mar 05 06:30:11 is that same? Mar 05 06:30:13 yes Mar 05 06:30:16 okay D) Mar 05 06:33:11 still.. apps are not working.. and warnings come up. Mar 05 06:33:30 ok, which app are you trying to run? Mar 05 06:33:58 sms, chat, people.. etc. :) Mar 05 06:34:10 hang on Mar 05 06:34:23 okay Mar 05 06:35:04 takes a little longer for me to find things at the moment, 7:30am here and i seem to not have any coffee in the house ;) Mar 05 06:35:35 export M_USE_SOFTWARE_RENDERING=1 Mar 05 06:35:47 lol, okay~ i have to go home, too :) Mar 05 06:35:48 and then try to run one of the apps, if you're doing this from command line Mar 05 06:35:57 okay Mar 05 06:36:14 i should really get a tegra2 device to test on myself.. Mar 05 06:36:50 hm.. still not working Mar 05 06:37:16 what exact warnings do you have? still libEGL? Mar 05 06:37:21 or now just glib Mar 05 06:37:31 still both Mar 05 06:37:58 hmmm. Mar 05 06:38:24 can i see the glib ones? Mar 05 06:38:27 somehow Mar 05 06:39:00 (process:1249): GLib-CRITICAL **: g_variant_get_type: assertion `value != NULL' failed (process:1249): GLib-CRITICAL **: g_variant_type_is_subtype_of: assertion `g_variant_type_check (type)' failed (process:1249): GLib-CRITICAL **: g_variant_builder_add_value: assertion `!GVSB(builder)->expected_type || g_variant_is_of_type (value, GVSB(builder)->expected_type)' failed (process:1249): GLib-CRITICAL **: g_variant_builder_end: asse Mar 05 06:39:29 libEGL warning: DRI2: failed to authenticate libEGL warning: DRI2: failed to open any driver (search paths /usr/lib/dri) Mar 05 06:39:35 ok Mar 05 06:39:36 like above.. Mar 05 06:40:11 try to add -software to the command line Mar 05 06:40:40 i'v already try that Mar 05 06:41:10 ok Mar 05 06:41:14 * Stskeeps wonders what is wrong Mar 05 06:41:28 Running non-meego graphics system enabled MeeGo touch, forcing native graphicssystem Not loading meegotouch-qt-style for meegotouch app. unable to connect to com.meego.dialer Could not start messageserver Segmentation fault (core dumped) Mar 05 06:42:41 always thanx for your kindly help :D Mar 05 06:42:44 was it meego-tegra2-1.1.90.20110218.0311-mmcblk3p.raw.bz2 that i made? Mar 05 06:42:48 i have to go in a 20 min. Mar 05 06:42:56 no Mar 05 06:43:01 ok Mar 05 06:43:01 it's new one Mar 05 06:43:50 still hardfp image? Mar 05 06:45:19 ok, maybe it'd be easier if i could see the kickstart file (.ks), then i can answer some questions myself :) Mar 05 06:45:34 Heya Stskeeps Mar 05 06:45:43 hm.. okay Mar 05 06:45:54 i'll sent it by my email, is that okay? Mar 05 06:45:57 send Mar 05 06:46:02 yes that's ok Mar 05 06:46:26 have your morning coffee, and see you on Monday! :D Mar 05 06:46:30 see you Mar 05 06:48:26 well crap, if I manually setuid root X, uxlaunch boots Mar 05 06:48:30 why would that be? Mar 05 06:48:39 bad driver? Mar 05 06:48:44 emgd Mar 05 06:48:50 i use Xorg-server-setuid on n900 too Mar 05 06:49:02 well see it gets installed Mar 05 06:49:05 yet i still have to do that Mar 05 06:49:23 used mic2? Mar 05 06:49:26 yup Mar 05 06:49:32 -f raw or -f fs? Mar 05 06:49:45 -f loop, then rsync'ed -aHx Mar 05 06:49:50 to my target device Mar 05 06:49:56 and yes, i checked before, it's not setuid before either Mar 05 06:50:07 -a preserves permissions (archive) Mar 05 06:50:13 -H preserves hardlinks Mar 05 06:50:15 ah, that's why then Mar 05 06:50:23 ??? Mar 05 06:51:14 -f loop has a bug that ruins the setuid of binaries Mar 05 06:51:21 Stskeeps: what should I be doing? I need to move the filesystem onto a specially formatted stick, with a vfat partition Mar 05 06:51:26 awww fuck me Mar 05 06:51:45 just make a -f raw file then Mar 05 06:51:45 lovely, does -f fs work? Mar 05 06:52:01 in the end i just need to rsync the filesystem over Mar 05 06:52:15 ok Mar 05 06:52:25 man, this totally stuck me for the entire night Mar 05 06:52:40 Stskeeps: I owe you a keg. Mar 05 06:52:43 ;) Mar 05 06:53:11 let me just pastie you something Mar 05 06:54:12 if i can just get this sucker booting and stable, i can start building what i need to build :) Mar 05 06:55:09 http://pastie.org/1635299 Mar 05 09:46:28 moo raster Mar 05 09:48:58 RST38h: burrrp Mar 05 10:56:00 hey Mar 05 10:56:16 did you notice, ASUS anounced to create meego netbooks Mar 05 10:57:26 http://www.engadget.com/2011/03/04/asus-jerry-shen-pledges-3d-tablet-meego-and-android-netbooks/ Mar 05 11:24:17 WhiteBit, frals posted about that last night on twitter Mar 05 11:51:05 * WhiteBit expect to see asus phones with meego :) Mar 05 12:07:14 ~seen sivang Mar 05 12:07:15 Shiny! http://edition.cnn.com/2011/TECH/web/02/23/bring.back.firefly/index.html Mar 05 12:07:25 sivang <~sivan@unaffiliated/sivang> was last seen on IRC in channel #meego, 20h 2m 19s ago, saying: 'or whatever they were called back in 96'. Mar 05 12:49:12 http://ftp.twaren.net/Linux/Linpus/update/releases.n/meego/Packages/openal-soft-1.12.854-2.meego.i586.rpm -> I try this package on such a link, but not in the latest image, why ? :o http://wiki.meego.com/Release_Engineering/Plans/1.2 Mar 05 13:02:24 djszapi: what do you mean? Mar 05 13:03:25 WhiteBit: both Asus and Acer stated they would create Meego netbooks when Meego was released... Mar 05 13:03:27 https://bugs.meego.com/show_bug.cgi?id=14188 Mar 05 13:03:30 <_MeeeGoBot_> Bug 14188 nor, Undecided, ---, gavin.hindman, NEW, [FEA] Add openal(-soft) support for ARM (N900) Mar 05 13:03:36 jonnor: I meant that ^^ Mar 05 13:04:59 Yes, and what about it? Mar 05 13:05:19 the mentioned package is for x86, it cannot be used on ARM Mar 05 13:05:23 well, I did not see any arm package about it Mar 05 13:05:24 if that is what you tried Mar 05 13:05:59 that is why I opened the feature request. If I was blind do not hesitate to close the request and send me the availability. Mar 05 13:34:31 morn andre__ Mar 05 13:38:55 heja Mar 05 13:39:59 'lo Mar 05 13:40:46 what does it mean this QGestureManager::deliverEvent: could not find the target for gesture Mar 05 14:15:33 wathek, perhaps you did not link up the signal to a slot? Mar 05 14:15:44 it's an event Mar 05 14:15:48 lcuk, what signal? Mar 05 14:16:02 yep it's an event Mar 05 14:16:04 lol ok thiago Mar 05 14:16:13 wathek, perhaps you did not link up the event to a handler? Mar 05 14:16:26 lcuk, let me pastebin my code Mar 05 14:16:43 I remember seeing this error, but I have no clue what it is Mar 05 14:16:53 try asking ddenis or zalbisser when they're around Mar 05 14:17:34 lucian1, thiago, http://pastebin.com/f0a7YAS1 Mar 05 14:17:38 there's my code Mar 05 14:18:01 and I don't understand why I got the message "TAP GESTURE" twice Mar 05 14:18:08 http://qt.gitorious.org/qt/qt-aleksi/blobs/75599a71e957cae29ddd4d3df9e89d9d4edc0b3d/src/gui/kernel/qgesturemanager.cpp#line459 Mar 05 14:18:32 i know it is not mainline, but google shows where the warning originates from, just walk backwards a bit? Mar 05 14:19:25 zalbisser_, hi Mar 05 14:50:24 wathek: In bool FrmAccueil::viewportEvent(QEvent *event)... Mar 05 14:50:39 gabrbedd, yep ? Mar 05 14:50:40 wathek: Shouldn't the 2nd line be... Mar 05 14:50:47 { Mar 05 14:50:55 wathek: return gestureEvent(static_cast(event)); Mar 05 14:51:01 thiago: :-p Mar 05 14:51:33 gabrbedd, not really Mar 05 14:52:02 gabrbedd, gestureEvent is a method that I wrote Mar 05 14:53:25 wathek: Just a guess... but on that line you handled the event. And then again at the end of the function, you re-handle the event by invoking the parent class's handler (return QGraphicsView::viewportEvent(event)) Mar 05 14:54:28 gabrbedd, touchEvent are sent to children first and then to parents if they haven't been treated ? Mar 05 14:55:04 yes Mar 05 14:55:18 localised events are sent to the topmost widget at the coordinates Mar 05 14:55:25 if they aren't handled, they are sent to the parent Mar 05 14:55:53 but wathek /did/ handle it, and then still sent to the parent. Mar 05 14:56:11 thiago, ok Mar 05 14:56:30 gabrbedd, yep and since QViewport is the big parent Mar 05 14:56:38 wathek: For example: http://doc.qt.nokia.com/4.7-snapshot/gestures-overview.html#example-event-handling Mar 05 14:56:42 I think it's better if I use return true; Mar 05 15:01:38 gabrbedd, same problem I changed my function to Mar 05 15:02:09 http://pastebin.com/BuWHMQgs Mar 05 15:02:26 gabrbedd, I added the return true; but I still getting the same message : Mar 05 15:04:28 No, I meant to do this: http://pastebin.com/MMDATJ2X Mar 05 15:04:42 (that's a diff, by the way) Mar 05 15:10:41 gabrbedd, but that way I won't be able to add FingerItem objects Mar 05 15:12:19 wathek: ?? So when the event is received it is at the same time a TouchEvent and GestureEvent ? Mar 05 15:12:21 gabrbedd, maybe I should call the gestureEvent function after adding FingerItems Mar 05 15:12:27 yes Mar 05 15:12:33 euh ! Mar 05 15:12:42 yes Mar 05 15:13:58 gabrbedd, when a touch begin I'll add FingerItem to the scene if theres a touch update I'll update the position of the finger items and if the touch end I'll remove the finger items Mar 05 15:18:39 gabrbedd, I still have the same message even after doing this : http://pastebin.com/suQhMUkJ Mar 05 15:29:12 wathek: I suggest you do a backtrace to find out what's triggering the events. Mar 05 15:29:31 wathek: My guess is that your custom gesture recognizer is actually sending two events. Mar 05 15:29:56 two Events ?! Mar 05 15:30:08 gabrbedd, let me pastebin my code Mar 05 15:30:22 wathek: Nope... I'm done. Mar 05 15:30:25 :-) Mar 05 15:30:26 lol Mar 05 15:30:27 ok Mar 05 15:30:38 wathek: I need to do some other stuff. :-) Mar 05 15:30:47 gabrbedd, ok thank you so mcuh Mar 05 15:30:49 much Mar 05 15:32:04 wathek: best of luck to you! Mar 05 15:32:09 :) Mar 05 15:59:15 thiago so how to solve it for meego then with cmake ? Mar 05 15:59:32 the flags should be automatic from the OBS Mar 05 15:59:39 in fact, they should be the compiler default Mar 05 15:59:45 are you using the default ARM compiler for MeeGo? Mar 05 15:59:52 are you using madde? Mar 05 16:00:13 I am using this image actually: http://repo.meego.com/MeeGo/builds/trunk/1.1.90.5.20110301.7/images/meego-handset-armv7l-n900-devel/meego-handset-armv7l-n900-devel-1.1.90.5.20110301.7-mmcblk0p.raw.bz2 Mar 05 16:00:21 mic-chroot and build. Mar 05 16:02:16 thiago: I can add -Wno-psabi though, would that be a dirty hackery ? Mar 05 16:02:50 thiago: current softfp doesn't have a proper baseline, but at least hardfp one will have --with-arch=armv7-a and such in default compiler so this shouldn't appear in future :P Mar 05 16:03:19 -Wno-psabi is just to shut up the compiler about the encoding of va_list changing Mar 05 16:03:41 Stskeeps: is djszapi using the right image for building ARM? Mar 05 16:03:46 because -march=armv7-a seems trickier to me... Mar 05 16:04:06 I guess that should be covered by cmake. Mar 05 16:04:30 thiago: .. sortof, he's doing same as you would compile on device. CFLAGS="-march=armv7-a -mfpu=vfpv3" should be exported and cmake picking this up Mar 05 16:05:31 most of us just use OBS or madde :P Mar 05 16:06:59 thiaog: I thought that would be one point about cmake, to build for the correct target with the correct target flag. Mar 05 16:07:04 * flag = option Mar 05 16:07:12 * thiago, even :) Mar 05 16:11:09 how can I change the state of a QGesture ? Mar 05 16:11:16 how does it get changed ? Mar 05 16:12:09 djszapi: I'm not sure cmake picks up CXXFLAGS Mar 05 16:12:24 it does Mar 05 16:12:36 is it set then with the proper values? Mar 05 16:12:44 fex. SET(CMAKE_CXX_FLAGS_DEBUG "-Wall -g3 -ggdb -O0" CACHE STRING "Debug options." FORCE) Mar 05 16:13:04 well, I think it is a cmake bug :) Mar 05 16:13:11 and I should not set anything. Mar 05 16:13:37 DocScrutinizer, see bug 7790 Mar 05 16:13:41 <_MeeeGoBot_> Bug https://bugs.meego.com/show_bug.cgi?id=7790 nor, Medium, 1.0.90.4, ilkka.koskinen, ASSI, N900 turns warm quickly after boot Mar 05 16:13:46 especially the comments Mar 05 16:14:40 is what I'm doing to create a gesture right : gesturerecognizer : http://pastebin.com/LMMZSVFE and tapgesture : http://pastebin.com/Bh2ytmsk Mar 05 16:14:54 cause I don't understand I cannot get the Gesture working Mar 05 16:14:55 :( Mar 05 16:18:35 thiago: Actually the same with -Wno-psabi Mar 05 16:19:21 the note about va_list is still there? Mar 05 16:20:35 thiago: seems that went away which means they were different issues: http://paste.kde.org/6552/ Mar 05 16:20:56 like I told you Mar 05 16:21:18 16:56 < thiago> djszapi: the error is caused by you having forgotten -march=armv7-a Mar 05 16:21:25 you don't pay attention when I tell you stuff, right? Mar 05 16:21:51 ohh, I thought it can be solved with the other option. Mar 05 16:22:07 but it seems both are needed. Mar 05 16:22:31 16:55 < thiago> djszapi: that note from gcc is annoying as hell. Just add -Wno-psabi Mar 05 16:22:39 I did tell you to add both Mar 05 16:22:47 yeah, sorry I misunderstood it. Mar 05 16:22:57 then file a bug against cmake developers, definitely. Mar 05 16:23:02 or at least asking on their mailing list. Mar 05 16:23:16 ok Mar 05 16:23:28 it should be really hided internally. Mar 05 16:23:45 yes, like Stskeeps said, the compiler should have it by default Mar 05 16:24:00 but I think MeeGo sets it in the environment Mar 05 16:24:06 neither cmake nor qmake honour the environment Mar 05 16:24:59 thiago: should I add you in the 'CC' ? Mar 05 16:25:06 no Mar 05 16:25:12 k Mar 05 16:25:27 softfp gcc causes cmake to need CXXFLAGS with -march=armv7-a set Mar 05 16:26:14 he should be using the hardfp one Mar 05 16:26:22 (1.1.90 is softfp right now), no "armv7hl" in the image name Mar 05 16:26:30 oh Mar 05 16:27:54 will hopefully be armv7hl very soon now. QA reported no noticable difference in bugs/new bugs between softfp and hardfp Mar 05 16:35:08 thiago: I do not know how other kde projects were compiled on meego arm then. When does it happen ? Mar 05 16:35:31 what's "it" that you want to know the time of it happenning? Mar 05 16:36:19 Error: selected processor does not support `ldrex r4,[r3]' etc Mar 05 16:36:41 that's an error Mar 05 16:36:43 btw: http://www.cmake.org/Wiki/CMake_Cross_Compiling -> * CMake cannot autodetect the target system Mar 05 16:37:01 it happens when you compile code with the wrong processor architecture Mar 05 16:37:14 your Qt was compiled for ARMv6 or later Mar 05 16:38:59 right Mar 05 16:41:15 interesting, this project builds just fine in fremantle scratchbox tho., is it managed by scratxhbox there then ? Mar 05 16:41:32 yes Mar 05 16:41:51 that is armv6 iirc Mar 05 16:42:26 so is it a meego image bug then ? Mar 05 16:42:43 no Mar 05 16:42:51 couldn't it be handled from the meego image by default ? Mar 05 16:42:57 like sb does. Mar 05 16:43:04 I thought madde did it Mar 05 16:43:17 your buildsystem didn't pick up the settings though, or it bypassed madde Mar 05 16:45:02 well what I did is just 1) Download 2) mic-chroot 3) git clone 4) build Mar 05 16:45:22 his environment doesn't have proper CXXFLAGS exported like OBS does it (he's ignoring me for some reason, so please state he needs to export proper CXXFLAGS= for that to work) Mar 05 16:45:22 I do not use any madde. Mar 05 16:45:57 djszapi: you should use madde and pay attention to what Stskeeps is saying Mar 05 16:46:55 djszapi: if you don't want to use madde, then set the proper flags to the proper compiler yourself Mar 05 16:46:57 thiago: there is no madde inside the image. Mar 05 16:47:07 no, there isn't Mar 05 16:47:09 madde is in the SDK Mar 05 16:47:12 yes. Mar 05 16:47:31 I am not sure that is a more mature way to do it..how stable/working is that ? Mar 05 16:47:39 pretty good Mar 05 16:48:36 what is the drawback compared to this image way ? Mar 05 16:49:30 it's meant for compiling Qt-based applications Mar 05 16:49:45 it won't have all the devel packages, just those that Qt-based apps most often need Mar 05 16:50:10 so cannot I cmake there ? Mar 05 16:50:12 * use Mar 05 16:50:56 I don't know, I've never tried Mar 05 16:51:04 oh well Mar 05 16:52:21 djszapi: export CXXFLAGS="-O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -Wformat -Wformat-security -fmessage-length=0 -march=armv7-a -mtune=cortex-a8 -mlittle-endian -mfpu=vfpv3 -mfloat-abi=softfp -D__SOFTFP__" (maybe CFLAGS too) before running cmake, like is being done in https://build.meego.com/package/rawlog?arch=armv7el&package=meegotouch-applauncherd&project=Trunk&repository=standard . You Mar 05 16:52:21 my problem is more like that I should warn each developer to use the toolchain file or some cmake variable explicitely, that is why this toolchain file way is a bit tricky. If I were the only developer it could work. Mar 05 16:52:43 carstenek: if he wants to ignore you, it's his problem Mar 05 16:53:15 thiago: it gets a bit annoying after seeing him ask over and over again for several days when you give the answer right away :) Mar 05 16:53:17 djszapi: then fix the toolchain Mar 05 16:53:35 carstenek: kick him out for not harassing the channel Mar 05 16:53:39 s/not // Mar 05 16:53:40 thiago meant: carstenek: kick him out for harassing the channel Mar 05 16:53:42 thiago: what do you mean ? Mar 05 16:53:58 djszapi: Stskeeps was giving you the answer. Did you try it? Mar 05 16:54:12 not yet. Mar 05 16:54:17 why not? Mar 05 16:54:59 I am checking that link... Mar 05 16:57:41 btw, that long post is broken here. Mar 05 16:57:57 how about the other times he told you stuff to do? Mar 05 16:57:58 nothing after this: "...https://build.meego.com/package/rawlog?arch=armv7el&package=meegotouch-applauncherd&project=Trunk&repository=standard . You" Mar 05 16:58:08 never mind the rest of the comment. the first part is what matters. Mar 05 16:58:09 well if he does not piss me off, sure. Mar 05 16:58:29 last time I asked something about meego building and I was sent away. Mar 05 16:58:52 djszapi: you ignore people under your own peril. Mar 05 16:59:10 djszapi: if you don't see an answer because you ignored a person, that's your problem. Mar 05 16:59:18 you don't get to ask again and again, if an answer was given. Mar 05 16:59:19 thiago: it is not a problem for me. Mar 05 16:59:25 someone else will help, I am fine with that. Mar 05 16:59:28 no Mar 05 16:59:35 if someone gave you the answer, you act on it Mar 05 16:59:47 how if I do not see that ? Mar 05 16:59:47 this includes people you ignored. Mar 05 16:59:50 not my problem Mar 05 16:59:54 you will act on it Mar 05 17:00:25 if you ignore answers, you'll be kicked for harassing Mar 05 17:00:56 that is also harrassing if I ask something meego related building and I am sent away. Mar 05 17:01:52 I don't know what happened before Mar 05 17:01:58 all I know is that you were ignoring Stskeeps's answer Mar 05 17:02:06 true Mar 05 17:03:22 djszapi: long story short: last time you contacted me in private message you acted very rudely to being told exactly the right solution for a problem and being asked to ask the question publically. Mar 05 17:03:41 djszapi: you're working with MeeGo ARM stuff, which is my area of work and i'm helping people that ask in public channels about this. Mar 05 17:03:59 thiago: the export did not solve the issue though... Mar 05 17:04:10 djszapi: then it looks like cmake doesn't honour that setting Mar 05 17:04:18 djszapi: so how about we get past this unproductive stuff and listen to this as it will save us both very much time? Mar 05 17:04:41 carstenek: yeah, I was like that and I tried to do it the right way then Mar 05 17:04:52 but I got offensive answers all the time then. Mar 05 17:05:10 I did not see the change on your side and the last when I asked about meego building I was just sent away by you immediately. Mar 05 17:05:17 offensive answers should be avoided regardless Mar 05 17:05:18 and I think it was not fair by any mean. Mar 05 17:05:32 but you do seem to ignore questions and not pay attention to what you're being told Mar 05 17:05:32 djszapi: ok, what did you consider offensive? Mar 05 17:05:36 this is not of today Mar 05 17:05:48 djszapi: we have public logs and i think we should get over this. Mar 05 17:05:56 thiago: I did not get any solution answer :) Mar 05 17:06:00 thiago: still the same :p Mar 05 17:06:28 djszapi: tell cmake to pass the -march=armv7-a flag to the compiler Mar 05 17:06:30 that's your answer Mar 05 17:06:36 carstenek: also it was not fair sometimes on #meego.fi either. Mar 05 17:06:44 thiago: yeah, toolchain file... Mar 05 17:06:55 but the problem is that all the developers should aware of that. Mar 05 17:06:58 yes Mar 05 17:07:04 which means it's a toolchain problem Mar 05 17:07:20 still not handy like in case sb :p Mar 05 17:07:23 or, more to the point: Qt has hardcoded instructions to use ARMv6 instructions Mar 05 17:07:31 qmake is configured to know to pass the flag Mar 05 17:07:38 everything else that uses Qt needs to know that too Mar 05 17:08:12 yes, but I would be more glad about hiding it from the developer :o Mar 05 17:09:05 so teach cmake to extract the info from qmake and use it Mar 05 17:09:23 so propose something to the cmake developers. Mar 05 17:09:35 yes, you should propose something to them Mar 05 17:10:31 or wait for Qt 4.8. This particular problem will not happen there. Mar 05 17:11:15 cool, looks like problem is untangling itself. Mar 05 17:11:44 Qt 4.8 detects at runtime the compiler option and uses the appropriate instructions Mar 05 17:11:52 so the error won't happen Mar 05 17:12:04 however, if you don't pass -march=armv7-a, then you won't get thread-safe code Mar 05 17:12:05 thiago: mmh... Mar 05 17:12:10 so toolchain workaround till Qt 4.8 Mar 05 17:12:17 you'll get the unsafe ARMv5 or v6 code Mar 05 17:12:22 no Mar 05 17:12:29 oh? Mar 05 17:12:30 the toolchain still needs to be fixed. See what I've just said. Mar 05 17:12:40 I'm saying that in 4.8 it wouldn't error out. But it would still be wrong. Mar 05 17:14:07 yes. Mar 05 17:14:17 and also, it is Qt solution only, cmake is not just for Qt products. Mar 05 17:15:30 djszapi: and everybody knows that, so talk to the cmake folks Mar 05 17:16:07 yes. Mar 05 17:23:46 djszapi: I'm coming to the conversation late... but... I've heard that (historically) the Qt devs have told the CMake devs *not* to use qmake to detect Qt's build parameters. IIRC, they were told to use pkg-config. Mar 05 17:24:08 ooh... Mar 05 17:24:18 I have already posted it :p Mar 05 17:24:25 I can ask it from them, ty. Mar 05 17:24:27 gabrbedd: right, we did Mar 05 17:24:44 gabrbedd: but cmake and KDE devs were stubborn and refused to use pkg-config Mar 05 17:24:53 for good reason, though: it's crap outside Unix Mar 05 17:25:08 djszapi: Anyway... I think it helps to explain why thiago is "over it" when it comes to CMake's behavior. Mar 05 17:25:13 thiago: FULL ACK Mar 05 17:25:26 pkg-config's cross-compilation support is also lacking Mar 05 17:25:44 And qmake sucks in general Mar 05 17:25:48 last I checked, all distributions carry a patch to make it work. There has been no release containing the fix. Mar 05 17:26:01 RST38h: yeah, granted too. Mar 05 17:26:01 Quietly skipping source files that conditionally include themselves,for example Mar 05 17:26:35 why would anything include itself? Mar 05 17:27:45 for example, a multi-format display driver including itself with various #define BPP* values Mar 05 17:28:20 that's really weird Mar 05 17:28:23 * gabrbedd prefers CMake to qmake. But qmake is nice if you have a small, Qt-only project. Mar 05 17:28:28 both on the source and the qmake bug Mar 05 17:28:41 * RST38h prefers plain make,gmake at worst Mar 05 17:28:44 anyway, qmake is known to be limited. A replacement would be nice. Mar 05 17:28:50 RST38h: not the same kind of software. Mar 05 17:28:53 gabrbedd: cmake works just fine for small qt project tho. Mar 05 17:28:59 it's like comparing ls to midnight commander Mar 05 17:29:03 Builds me stuff, all I care about Mar 05 17:29:45 djszapi: Yes, but `qmake -project` is just too freaking easy. :-) Mar 05 17:30:01 cmake ../ is shorter :p Mar 05 17:30:23