Friday, 2012-10-26

*** sonach has joined #mer00:00
*** trbs has quit IRC00:02
*** reels__ has joined #mer00:11
*** reels_ has quit IRC00:15
*** himamura has quit IRC00:18
*** Aurium has quit IRC00:20
*** clopez has joined #mer00:20
*** popey has joined #mer00:24
*** X-Fade has joined #mer00:25
*** tilgovi has quit IRC00:26
*** ronan22 has quit IRC00:32
*** popey has quit IRC00:33
*** blam_ has joined #mer00:51
*** kulve has quit IRC00:54
*** kulve has joined #mer00:54
*** andre__ has quit IRC01:09
*** himamura has joined #mer01:58
*** FSCV has joined #mer02:01
*** otep_ has quit IRC02:02
*** lardman|home has joined #mer02:11
*** lardman has quit IRC02:12
*** otep has joined #mer02:21
*** clopez has quit IRC02:47
*** DocScrutinizer05 has quit IRC03:03
*** DocScrutinizer05 has joined #mer03:03
*** himamura has quit IRC03:11
*** KaiRo_Mozilla has joined #mer03:12
*** KaIRC has quit IRC03:16
*** morphis|away is now known as morphis03:25
*** himamura has joined #mer03:30
*** pa has joined #mer03:32
*** phinaliumz has joined #mer03:36
*** dcthang has joined #mer03:43
*** phinaliumz has quit IRC03:52
*** phinaliumz has joined #mer03:53
*** furikku has joined #mer03:58
*** rdqfdx has joined #mer04:01
*** KaiRo_Mozilla has quit IRC04:01
*** morphis is now known as morphis|away04:02
*** Behold has quit IRC04:03
*** BeholdMyGlory has joined #mer04:03
*** ronan22 has joined #mer04:04
Bostiklbt: you're right, and I think I know why - gold linker use has to be separately enabled for gcc; I'll try one more hack and then see what happens with a privately built gcc04:13
*** kostaja has joined #mer04:13
*** himamura has quit IRC04:31
*** pa has quit IRC04:51
StskeepsBostik: ugh, really no way to individually say "use this ld" ?05:09
BostikStskeeps: not in gcc 4.6, it seems but I'm looking at how to hack it05:10
Bostikespecially as "nice" linker selection seems to depend on which build options have been available in the first place05:11
Stskeepsnice linker selection can be done in gcc, as long as you can switch on ld/gold on runtime.. i hope05:11
*** dakovaci has joined #mer05:17
*** Jucato has quit IRC05:18
Bostikhmm... I think I might be able to trick the build...05:18
*** kjokinie has quit IRC05:20
*** Jucato has joined #mer05:22
*** PMG has quit IRC05:22
*** FSCV has quit IRC05:26
*** VDVsx has joined #mer05:27
*** PMG has joined #mer05:38
*** phinaliumz has quit IRC05:43
*** kontio has joined #mer05:46
*** phinaliumz has joined #mer05:50
situMorning everyone05:51
*** pirut has quit IRC05:52
*** rcg has joined #mer06:02
*** ScriptRipper has quit IRC06:10
*** furikku has quit IRC06:13
*** furikku has joined #mer06:13
*** ScriptRipper has joined #mer06:15
*** pirut has joined #mer06:15
*** pirut has quit IRC06:16
*** pirut has joined #mer06:17
*** SfietKonstantin has joined #mer06:17
*** pirut has quit IRC06:19
*** pirut has joined #mer06:19
*** niqt has joined #mer06:27
*** pirut has quit IRC06:28
*** pirut has joined #mer06:30
*** popey has joined #mer06:32
rcghi ccssnet06:35
niqtmoring all06:42
*** pohly has joined #mer06:46
*** kjokinie has joined #mer06:46
rcgmorning niqt :)06:48
Bostikjumpin' jesus on a pogostick, qtwebkit is too large even for gold linker(!)06:51
Bostikcopied /usr/bin/ to /tmp/gold/ld, prepended PATH, and get... "/tmp/gold/ld: out of memory"06:52
Stskeepswtf :P06:52
*** veskuh has joined #mer06:52
Stskeeps--no-keep-memory in use?06:53
Bostikalthough for gold that's a no-op and only accepted to keep compatibility with regular ld06:55
Stskeepswhat happens if you do --hash-style=sysv ?06:55
Stskeepswe use gnu for sanity purposes but i'm on occasion not sure if anybody.. actually.. tests the innovations they make06:55
Stskeepslately i've run into DWARF-4 causing really bizarre linking issues06:55
BostikStskeeps: I'm not sure anyone really tests with targets where int32 becomes a problem :)06:56
Bostikbut I'll try with that option too06:57
Stskeepsi bet it's a conspiracy06:57
*** SfietKonstantin has quit IRC06:58
Bostikwhat other projects really hit that? webkit, chromium, gecko, ... oracle perhaps in-house?06:58
*** fcorrea has joined #mer06:59
*** gabriel9|work has joined #mer06:59
*** fcorrea has quit IRC07:00
*** fcorrea has joined #mer07:00
*** mikhas has joined #mer07:02
henrik__lbt: you did some magic with packaging. Cool07:03
*** ka6sox is now known as ka6sox-away07:07
*** lamikr has joined #mer07:09
*** stefan_schmidt_w has joined #mer07:10
*** jpetersen has joined #mer07:10
*** M4rtinK has joined #mer07:11
*** dakovaci has quit IRC07:14
*** dakovaci has joined #mer07:14
*** ali1234 has joined #mer07:18
*** M4rtinK has quit IRC07:33
*** calvaris has joined #mer07:33
*** leinir has quit IRC07:42
*** leinir_ has joined #mer07:42
*** leinir_ has quit IRC07:46
*** slaine has joined #mer07:47
*** leinir_ has joined #mer07:48
*** leinir_ has quit IRC07:48
*** leinir_ has joined #mer07:48
*** cxl000 has joined #mer07:49
*** leinir_ has quit IRC07:52
*** dakovaci has quit IRC07:55
*** leinir_ has joined #mer07:59
*** dakovaci has joined #mer08:00
*** leinir_ has quit IRC08:05
*** dakovaci has quit IRC08:06
*** Jucato has quit IRC08:10
*** leinir_ has joined #mer08:11
*** leinir_ has quit IRC08:11
*** leinir_ has joined #mer08:11
*** arc_mat_ has joined #mer08:12
*** leinir_ has quit IRC08:14
*** Jucato has joined #mer08:14
*** leinir_ has joined #mer08:14
*** arcean_ has joined #mer08:16
*** yunta has joined #mer08:19
*** notmart has joined #mer08:27
*** notmart has joined #mer08:27
*** BeholdMyGlory has quit IRC08:29
*** BeholdMyGlory has joined #mer08:30
*** Attie has joined #mer08:31
*** leinir_ is now known as leinir08:37
*** alh has quit IRC08:37
*** alh has joined #mer08:37
*** rcg has quit IRC08:55
*** pohly has quit IRC08:55
*** kimitake has quit IRC08:56
*** kimitake has joined #mer09:00
*** fcorrea has quit IRC09:08
*** popey has quit IRC09:08
*** phinaliumz has quit IRC09:11
*** vgrade has quit IRC09:13
*** rcg has joined #mer09:15
*** rc_ has joined #mer09:18
*** aportale has joined #mer09:18
*** mdfe has joined #mer09:19
*** sonach has left #mer09:20
*** mihero has quit IRC09:20
*** mihero has joined #mer09:21
*** rcg has quit IRC09:21
lbtaportale: seen auri?09:30
aportalelbt: He's on spontaneous vacation, today :)09:31
*** felipec has quit IRC09:31
lbtaportale: OK - do you know if he's pushed anything anywhere?09:31
lbtI was planning on syncing git and running through where we're up to09:31
*** hazchemix has joined #mer09:32
lbthenrik__: ping09:32
aportalelbt: That's the repo/branch . He committed an initial settings Ui for mer target handling and ssh key selection.09:33
lbtthat was it - ty09:34
aportalelbt: Can we use the mer logo in Qt Creator? The option page still needs an icon.09:35
lbtI have svg's too09:35
lbterm is looking empty09:36
aportalelbt: And tis one looks a bit.. white: :)09:38
lbtit's so pure09:38
lbtbut it really is the logo09:38
aportalelbt: Ah. against a non-white background it becomes visible. Looks nice indeed.09:40
aportalelbt: That's just what we need for an icon. We can still change it later, of course.09:41
henrik__lbt: pong09:41
*** rc_ has quit IRC09:43
*** rcg-work has joined #mer09:43
lbthenrik__: so lets see about pushing the web admin side forward today and maybe doing an image with it?09:44
yuntalbt: do we need hostname changing in sb2_manage?09:45
henrik__lbt: i'll send pull req for sdk-app09:45
*** TuTizz has joined #mer09:45
lbthenrik__: do you want to maintain it?09:45
henrik__lbt: why not09:46
lbtyunta: I wondered that - not yet09:46
*** TuTizz has quit IRC09:46
*** TuTizz has joined #mer09:46
lbthenrik__: I'll explain how to do gitpkg when you need to09:46
henrik__lbt: thanks. how about update targets?09:47
henrik__there was some discussion with yunta about that09:47
lbtyep, comment out that function for now09:47
lbtwe'll do remove/re-download09:47
lbtaportale: there's a big background on that page too now - it has an odd artefact - I'll upload a clean one sometime09:50
lbtaportale: so far Mer has always used that blue colour background so it may be useful09:51
*** fcorrea has joined #mer09:52
*** fcorrea has quit IRC09:56
*** fcorrea has joined #mer09:56
*** clopez has joined #mer09:59
*** popey has joined #mer09:59
*** TuTizz has quit IRC10:00
*** popey has quit IRC10:06
*** Milhouse has quit IRC10:06
*** Milhouse has joined #mer10:06
*** vgrade_ has quit IRC10:08
*** jpetrell has quit IRC10:13
*** popey has joined #mer10:15
*** vgrade has joined #mer10:18
lbtaportale: so looking at the target selection code...10:21
lbtit's checking for bin/qmake which is wrong I think10:21
aportalelbt: Are you in the mer plugin?10:23
* aportale looks at patch10:24
lbtQFileInfo qmake = QFileInfo(targetDir, QLatin1String("bin/qmake"));10:24
aportalelbt: I assume that is the "fake qmake".10:25
lbtwrong window :)10:25
lbtbut, no, it's not supposed to be there - that could be just a sysroot with just headers/qml10:26
aportalelbt: I thought we moved away from that idea and are now presenting Qt Creator the complete target.10:27
aportalelbt: The idea is that a "fake qmake" (shell script) needs to be copied into the mer target, so that Qt Creator uses that one in order to detect the Qt Version10:28
lbtthat wasn't my understanding10:28
lbtso that fake qmake is managed by sb210:28
lbtthe shell is shipped with qt creator10:28
lbtwrapper, sorry10:29
* Bostik officially hates gcc/ld option interactions10:29
lbtso it's not in the target10:29
yuntalbt: I've sent pull request your way.10:29
lbtty yunta10:30
aportalelbt: The magic: Qt creator let's the user select a Mer target and copies the fake qmake into it. In the same step, the Qt version is created.10:30
lbtwhy copy the 1liner?10:30
lbtseems odd10:30
lbtand you do not want to think there are binaries in the sysroot10:30
lbtthere aren't10:30
lbtthey are hidden in the vm10:31
lbtunder sb210:31
aportalelbt: It is more than that 1-liner. It echos out a few lines wit hpaths.10:31
lbtit issues an ssh command to VM to do sb2 qmake --query ?10:31
aportalelbt: are the libs (e.g. not in the target?10:31
lbtaportale: yes10:33
aportalelbt: Ideally, it would not need to use ssh. We don't want to requrire the user to have the VM running before starting Qt Creator10:33
aportalelbt: So Yes means they are not threre.10:33
lbtthe libs happen to be in the sysroot - but I thought we didn't need them. Just proper access to all headers and QML10:34
aportalelbt: we need in order to detect the architecture10:35
aportalelbt: We actually just need the first few bytes of it10:35
lbtaportale: OK, that's new :)10:35
aportalelbt: Now, I ned to understand the following: Do you intend to habe the targets10:36
aportalelbt: ...have the targets on the host?10:36
lbtI'm not sure it's a problem tbh10:36
lbtseeing the .so10:36
lbtI was looking further ahead to windows/mac10:36
lbtwhere we may extract the minimal data from sysroot10:37
aportalelbt: Aurindam understood that therefore, You would not want to have the stripped down sysroot,10:37
lbtI wanted that to be headers/qml only10:37
lbtright - so for current shared folder/linux work - we share a dir10:37
aportalelbt: Ok. You want both. Aurindam thought you just want the full target10:37
lbtbut we discussed places this may fail and we'd fall back to extract10:38
lbtso phase 1 is full target10:38
aportalelbt: Do you see reasons for it to fail on Windows?10:38
lbtit's windows10:38
*** stezz has joined #mer10:38
lbtwe've discussed using non-vbox solutions too10:39
aportalelbt: What makes it fail on windows? The file system?10:39
*** imunsie has quit IRC10:39
lbtlet me put it this way - I'm covering our options for the time being10:39
lbtso not making assumptions now that we regret later10:40
lbtI expect something to go wrong there and want to make an alternative as easy as possible10:40
aportalelbt: Most likely we will have to keep the targets in the VM for Windows.10:41
lbtsome things I have a high confidence in - ie an ssh command to the VM is very unlikely to ever fail10:41
lbtaportale: that's a possibility (though I don't think it's a certainty)10:41
lbtso - you did say that the VM may not be running at QtC startup10:42
lbtI think we should assume it is for phase 110:42
lbtbut I think it's an interesting point that I'd not considered10:42
lbtgiven you can't build or setup projects without the VM I think we just need to review what happens there during phase 2 development10:44
aportalelbt: Let's discuss this with Aurindam on Monday and choose a feasable path. Aurindam took the "full target instead of sysroot subset" path as a given and started dealing wit hthat.10:44
lbtyep - but I think we may have misunderstood 'full target'10:45
lbtthe cross-compile stuff means it's probably not10:45
lbtit has libs but no executables10:45
lbtand if it did they could be arm10:45
lbtso running bin/* doesn't really make sense10:46
lbtit should certainly be considered read-only10:46
aportalelbt: The fake qmake does not need to be in bin/ It could be in the target root10:46
aportalelbt: Just needs to be in a defined location where it does not clash with the target10:47
lbtmmm could it not be in a creator private scripts/ dir ?10:47
lbtI'd prefer that10:47
aportalelbt: we need one per tearget10:47
lbtit's generic10:47
lbtthough you mentioned magic before10:48
lbt"Qt creator let's the user select a Mer target and copies the fake qmake into it. In the same step"10:48
aportalelbt: Exaclty. That is solving some chicken-egg problems10:48
aportalelbt: Having the script in the target also implicitly gives us the information about the kit's "sysroot" path10:49
lbtI don't see the qmake in the git tree10:49
aportalelbt: Frankly, this is a hack on top of hacks :) But arent't we all proud hackers?10:49
*** Saila has joined #mer10:50
aportalelbt: the fake qmake is not there, yet. Aurindams commit is just an initial one10:50
lbtOK - but I think I'd like to see something slightly different10:51
lbtI need to read the code a bit more10:51
lbtthis was the first thing to block me :)10:51
lbtaportale: thanks a lot for the discussion btw - we need to do this to ensure we understand what's going on10:53
lbtbad timing for auri - shame he didn't have monday off when the VM wasn't working but that's life eh?10:53
aportalelbt: Aurindam should have taken vacation the whole week before wednesday :)10:53
*** Saila has quit IRC10:54
lbtyeah - wasn't great10:55
aportalelbt: We need a very clear description about your planned changes to Mer setups. The targets-on-host information did apparently not arrive here 100% correctly.10:57
*** gabriel9 has joined #mer10:57
aportalelbt: We thought that you uchose that in order to avoid the "partial sysroot extraction" hack.10:57
*** gabriel9|work has quit IRC10:58
*** stezz has quit IRC10:59
lbtaportale: not really, that hasn't changed at all. The shared folders approach just avoids the extract/copy phase. There's no change to what data was planned to be exposed.11:00
lbtI think the problem is that you can now see private data and are using it11:01
*** fcorrea has quit IRC11:03
aportalelbt: Private headers and stuff, right? Do they need to be in the target? I would have assumed that they were pimpl'ed out.11:03
lbtno, sorry, not that at all11:04
lbtI meant just because you see the full target root filesystem (including /bin /var/log etc) doens't change the design of using just the info from a minimal sysroot (ie /usr/include/* )11:05
aportalelbt: Ah. OK. I think we should definitely be able to switch to the minimal sysroot when we have some time ond oxygen.11:06
lbtbecause later on we may need to go back to minimal sysroot - I don't recall saying we wouldn't11:06
lbtyes - so things like this qmake set off warning bells - that's all11:06
lbtthen "oh, we need to see" :)11:07
lbtso I'm just spending time now so we don't spend more time later11:07
*** stezz has joined #mer11:07
Jake9xxseems my history log got purged, what was the command line for mic nowadays to make a vm iamge sudo mic livecd.. ?11:08
lbtnb - I'd also be happy just to document these areas11:08
lbtJake9xx: I use       sudo mic create fs /mer/mer/devel/mer-sdk/nemo-handset-armv7hl-n950-0.20120920.1.NEMO.2012-10-07.1-sysroot.ks --pkgmgr=yum --arch armv7hl -o /srv/mer/targets/nemo_n911:08
lbtso mic create help also works11:08
Jake9xxlivecd, please - this is a vm image11:08
Jake9xxmic create livecd.. I remember Stskeeps had some params there as well..11:09
lbtI don't do livecd - just real images11:09
lbtI use raw for SDK VM11:09
Jake9xxlbt: I'm working on the mesa11:09
Jake9xxmesa llvm to be exact11:09
lbtsudo mic create raw /mer/mer/devel/mer-sdk/sdk-kickstarter-configs/kickstarts/mer-sdk-nemo-i486-vm.ks --pkgmgr=yum --arch i486  should work11:09
lbtthat boots to vbox11:09
Jake9xxarch i586 ?11:09
lbtno messing about with liveboot/install11:10
Bostikcollect2: ld terminated with signal 6 [Aborted]11:10
Bostikterminate called after throwing an instance of 'std::bad_alloc'11:10
*** stezz has quit IRC11:10
* lbt puts an arm round Bostik : there there, it's not that bad really11:10
lbtJake9xx: 586 yes if you use atom11:11
Jake9xxlbt: coolio, if you don't specify arch it probes it and is doing i686 now11:11
lbtJake9xx: yeah, that wont work iirc11:11
Jake9xxlbt: I beg to differ, we'll see11:12
lbtaportale: so listing all the hacks is also OK11:12
Jake9xxtrying to get rid of our swrast...11:12
aportalelbt: Yes. I will document them asap.11:13
lbtaportale: can you pastie the qmake script that you're copying into the target? or is it the simple one we've been using?11:13
aportalelbt: It is a more elaborated one, and it is on Aurindam's machine :/11:14
lbtOK - we've been working on it too and it now works with old creator to build using sb211:14
lbtwe just added SB2_TARGET env variable to creator11:15
aportalelbt: Nice.11:15
lbtand ensured that passed into a wrapper in Mer SDK and that works fine11:15
aportalelbt: Is that variable set manually in the project settings?11:18
henrik__lbt: any pointers for the packaging stuff. I did osc copypack (i assume that was right thing to do). i have updated git, how i work with _src. I think i read some instructions somewhere but cannot find it now11:19
lbthenrik__: OK - so you need a tag (or use a branch if it's a WIP) called 0.2 on master11:20
lbtmake sure it's clean11:21
lbtthen checkout pkg-mer branch and update _src so the last field is 0.211:21
henrik__and the git tree is in place in my working dir (home:...:)11:21
lbtthen run gp_mkpkg11:22
lbtyou should also edit the yaml to change the version11:22
lbtand add an entry to  .changes11:22
lbtrun specify to update the spe11:23
lbt*then* run gp_mkpkg :)11:23
henrik__lbt: thanks.11:23
lbtlet me know what I got wrong :/11:24
henrik__i will :)11:24
lbt for more info11:24
henrik__that was it :)11:24
*** popey has quit IRC11:25
*** dcthang has quit IRC11:25
Jake9xxlbt, was there a way to look into the @ package collectiond in ks file?11:26
Jake9xxlbt I need to see where rpm x comes from11:27
aportalelbt: When would the scripts for minimal sysroot extraction be ready? Are you working on them?11:27
lbtaportale: no, phase211:28
lbtwhy do we need them?11:28
*** KaIRC has joined #mer11:28
lbtJake9xx: I look in the repo gzipped xml files or the source11:28
lbtJake9xx: I'm not aware of any tool to expand them - it would be nice though11:31
lbtI think mic could usefully do that in the log11:32
*** morphis|away is now known as morphis11:38
*** FlameReaper has joined #mer11:39
*** ScriptRipper has quit IRC11:44
*** FlameReaper has quit IRC11:45
*** ScriptRipper has joined #mer11:49
*** lamikr has quit IRC11:51
*** fcorrea has joined #mer11:52
*** faenil has joined #mer11:54
*** fcorrea has quit IRC11:55
*** fcorrea has joined #mer11:55
Jake9xxlbt: it would. parsing gzipped xml's is nasty :)11:57
*** arcean_ has quit IRC11:59
*** popey has joined #mer11:59
*** pa has joined #mer12:03
*** popey has quit IRC12:05
*** fcorrea has quit IRC12:06
*** popey has joined #mer12:08
*** lamikr has joined #mer12:10
*** Venemo has joined #mer12:11
*** arcean has joined #mer12:13
*** FlameReaper has joined #mer12:14
*** Sfiet_Konstantin has joined #mer12:16
*** jluisn has joined #mer12:17
*** fcorrea has joined #mer12:19
vgradelbt, looks like it may be of use this weekend12:21
Jake9xxlbt: actually if you make ks with kickstarter, in the beginning of ks you'll find the params12:22
*** FlameReaper has quit IRC12:25
*** trbs has joined #mer12:27
*** ronan22 has quit IRC12:30
lbtvgrade: :) I'm glad - let me know which bits need more explaining or need simplifying12:31
*** morphis is now known as morphis|away12:31
vgradelbt: I want to package this , which uses git modules12:32
vgradesub-modules that is12:32
lbtit depends how git archive works with them12:32
Jake9xxvgrade: welcome to yaml hell :)12:33
*** lizardo has joined #mer12:35
vgradeJake9xx: Ive done yaml before but git sub modules are new to me.  Seems they allow you to refer to a particular version of an existing external git repo12:37
lbtvgrade: just looking we'd need to do some work in gitpkg to make it 'just work' but it would be do-able12:42
*** pa has quit IRC12:44
vgradelbt, might be some good git training for me (I'm only on git clone page)12:45
lbtwell, git-pkg does some interesting things12:45
*** FlameReaper has joined #mer12:48
*** himamura has joined #mer12:49
*** niqt has quit IRC12:50
*** arcean has quit IRC12:51
*** arcean has joined #mer12:52
*** arcean has joined #mer12:52
*** rcg-work has quit IRC12:56
*** gabrbedd has joined #mer12:57
*** fcorrea has quit IRC12:59
*** fcorrea has joined #mer13:01
*** crevetor has quit IRC13:06
*** jayrulez has joined #mer13:08
*** Aurium has joined #mer13:09
lbtaportale: ok so the fake qmakes13:11
lbtyour stub one is providing some -query info?13:11
aportalelbt: Yes. It provides the absolute paths on the host, so tha Qt Creator can find headers, libs, etc.13:12
aportalelbt: And if it is not called with "-query", it will call the real qmake (or the wrapper in that case)13:12
lbtso I think I wanted see some kind of generic path manipulation13:13
lbteg how do you handle errors?13:14
lbtif $HOME/src/myfile/code.cpp has a syntax error in build13:14
lbtdoesn't creator need to highlight the 'right' source on the host13:14
*** kostaja has quit IRC13:15
lbtso I'd think that mechanism should handle the query results too13:15
aportalelbt: There are already file path mapping heuristics implemented somewhere in Qt Creator. Could be that they need to be tweaked for the Mer VM case, but this kind of issue is no news.13:15
lbtOK - so should that handle --query ?13:16
aportalelbt: The call of 'qmake -query' happens at the first when Qt Creator enumerates the registered Qt Versions.13:17
aportalelbt: ...or in the moment where the Qt version gets manually added.13:18
lbtand is that prior to the mapping  ?13:18
aportalelbt: The mapping happens on the fly when the user clicks on an error message/warning.13:18
*** FSCV has joined #mer13:19
aportalelbt: And the mapping uses the information that had initially be "-queri"ed that first time13:19
lbtOK - so I'd think the value entered for "target sysroot location" is all you need13:20
*** savaaZ has quit IRC13:20
lbtand it's easy to do the parse inside the generic wrapper13:20
lbtnot parse, substitution13:21
*** rcg-work has joined #mer13:21
lbtnb - I was also expecting that to happen in C++ because it needs to run on windows too13:22
lbtso in a bash script seems wrong13:23
*** FlameReaper has quit IRC13:23
lbtbut anyhow... yep, understand that. How does the script get the 2 strings? it needs the sysroot location and it needs the internal path to subst13:23
*** Hoolxi has joined #mer13:29
aportalelbt: Sorry, I did not your messages till now. WIll answer13:31
aportalelbt: Right, the "fake qmake" might eventually need to be c++ in order to work on Windows.13:33
lbtwell handling the reply from the ssh call13:35
aportalelbt: How the script gets the sysroot location? It looks at it's own location assuming that it is in a defined location in the sysroot. If that's what you meant13:35
lbtaportale: so we need to pass that in as well as SB2_TARGET13:36
lbthenrik__: nice13:36
*** pohly has joined #mer13:36
aportalelbt: Because the fake qmake script/executable will not be in the target, correct?13:36
henrik__lbt: good instructions. except i managed to make many mistakes while doing it. Next round will be less painfull13:37
lbtideally (and especially if that's the only reason to put it there :) )13:37
lbthenrik__: seen ?13:37
henrik__lbt: yes. that part was ok for me. the pkg-mer part was where i did all the mistakes13:39
aportalelbt: Hmm. I can understand that you find the script in the target an ungly idea. However, with the SB2_TARGET variable approach, I am not sure if we can set it just before the "-query" call. We will have to modify the Qt Version handling UI and add a environment variable editor for the qmake call.13:40
lbtaportale: maybe it would be good to write down what order things are called in?13:41
aportalelbt: Unfortunately the sysroot is a fiels of a "kit". And the Qt Version (and the -query) is there, before the kit exists.13:41
lbtsequence diagram or similar13:41
lbtI see13:42
aportalelbt: Can do. But in a nutshell: this is the chicken-egg problem.13:42
lbtnb.. if you don't know sysroot, how do you call $sysroot/bin/qmake -query ?13:42
*** blaroche has quit IRC13:43
aportalelbt: With an absolute path to the sysrott specific fake qmake13:43
lbtI guess you have 'sysroot' hardcoded into the qmake path in the Qt version13:43
lbtI have a suggestion13:43
aportalelbt: The qmake which Aurindam planned to plant into Mer Targets13:43
lbtso Mer Target preceds Qt Version ?13:44
lbtbecause that's a good place to set SB2_TARGET too13:44
lbtit has 'name' 'Location' .. add 'sb2 target name' which defaults to 'name'13:45
aportalelbt: Are you able to keep that in memory till Monday? :) I'd like Aurindam thinking about this, aswell.13:47
lbtI think a sequence diagram is a high prio13:48
lbtif you're around before me (probable) then menton it to auri too13:48
lbtnot infinite detail13:48
lbtjust the main config and QtC init stuff13:48
*** jstaniek_QFridge has joined #mer13:52
aportalelbt: Sure. And there is already some documentation of how and in which order kits are set up. Done by the fine people at KDAB:
aportalelbt: Page 213:53
lbtneat - and this is just to stop me making dumb suggestions :)13:53
lbtso where does Mer Target fit in ?13:54
lbtIdeally prior to Compiler13:54
lbtyep - that'd make sense13:55
lbtWe could even use that to provide auto-detected compiler and Qt version13:56
*** ssirkia1 has left #mer13:57
*** Sfiet_Konstantin has quit IRC14:00
*** Venemo_ has joined #mer14:02
*** Venemo has quit IRC14:02
*** Venemo_ is now known as Venemo14:02
*** faenil has quit IRC14:05
aportalelbt: Yes, the Mer target would resolve the order problems and as-well auto-create all the other settings.14:09
lbtOk, I'll look at the src around there (for fun)14:10
aportalelbt: Aaaand. Actually, I have wage ideas about how to achieve that without injecting scripts into the Mer Targets. But Aurindam has to sanity-check it :)14:10
lbtneat - I think if we follow the data flows we should understand better14:10
*** blaroche has joined #mer14:17
*** VDVsx has quit IRC14:18
*** pa has joined #mer14:25
*** niweber has quit IRC14:26
*** rcg-work has quit IRC14:27
*** savaaZ has joined #mer14:32
*** kontio has quit IRC14:39
*** Hoolxi has quit IRC14:42
*** Venemo has quit IRC14:43
*** jkt has joined #mer14:44
*** ortylp has joined #mer14:45
jkthi there; I'm using some code from the Qt's messaging framework in my application and have found a bug in it14:45
jktit looks like the upstream at is more or less dead (no commits since this summer)14:45
Stskeepsw00t: ^ , vdvsx14:46
jktwhere shall I post the patch? It affects the RFC2047 decoding and is pretty minimal.14:46
jktThe guidelines for contributing refers to OSC setup, which is not really something I'd like to do here :)14:47
Stskeepswould be best to go upstream naturally14:47
jktyeah, well, "where is the upstream" is the question14:47
jktI got no definitive answer on #qt-labs14:47
Stskeepsif you come by daytime, try to catch vdvsx14:48
jktStskeeps: which TZ is he in?14:48
jktI realize that it's Friday, but it's only 16:48 here (and there) :)14:49
Stskeepsyeah.. :P14:49
Stskeeps17:48 to be exact, so maybe he's on tonight14:50
jktanyway, is the patch (should apply to qmf's src/libraries/qmfclient/qmailmessage.cpp) and has some unit tests14:51
Stskeepsjkt, i think there was also work to move qmf to gerrit, but unsure how that's going14:51
*** yunta has quit IRC14:51
*** pa has quit IRC14:51
jktStskeeps: okay, I'll try to stay around for a while14:51
jktStskeeps: would that be qt-proj's gerrit, or something mer-specific?14:52
Stskeepsqt proj14:53
Stskeepswe try to work upstream as we can but sometimes backports are needed14:53
jktthanks for the information14:54
*** arcean_ has joined #mer14:55
*** ortylp has quit IRC14:55
*** VDVsx has joined #mer14:56
*** veskuh has quit IRC14:56
StskeepsVDVsx: jkt has a qmf patch, could you help him find proper upstream/review etc?14:58
lbtjkt: it's worth posting to the mer-general mailing list too14:58
lbtthat way it's around if we need to pull it to a carried patch14:58
jktlbt: I understand the point, but I'd prefer not to get too much involved at this point (subscribing to a mailbox and such); I've simply found a bug, patched it locally and want to report it to $upstream15:00
jktI don't have time to get involved with Mer at this point, sadly15:00
jktVDVsx:, test case at
lbtok, np - we have the gitorious patch15:00
*** arcean_ has quit IRC15:00
jktVDVsx: the patch has to be hand-applied to src/libraries/qmfclient/qmailmessage.cpp15:01
jktthe best way would be to extract these low-level bits out from that monster file, IMHO15:01
jktso that we can share patches and bugfixes more easily15:01
*** odin_ has quit IRC15:01
*** ced117 has joined #mer15:02
VDVsxjkt, why hand-applied ? looks quite small to me15:02
jktVDVsx: it's a different file with a different white spacing, sadly15:03
VDVsxjkt, qmf upstream is currently moving to gerrit and is not possible to make merge request in gitorious15:03
VDVsxso we have to wait, I can take care of it for mer only15:03
jktVDVsx: shall I talk to Don Sanders, or is he completely out now?15:04
VDVsxjkt, well, I had the same rights as Don, so I guess he can't do much anymore as well15:04
jktwell, I remember his name because he blogged about QMF and we've exchanged a few mails some time ago, that's why I'm asking15:05
VDVsxjkt, we also left Nokia, QMF code is now maintained by digia15:06
VDVsxjkt, where is your bug tracker, to have a look to the bug referred there (#553)15:07
*** gabriel9 has quit IRC15:09
jktah, and to be a bit more on topic here, the application's got a QML interface using Harmattan's components15:10
jktCONFIG+=harmattan for qmake15:11
jkthmm, redirects to his Nokia e-mail address15:11
*** mikhas has quit IRC15:11
jktI don't have plans to install Mer on my N950, but if someone is interested in packaging or giving it a try, please feel free to play with it a bit15:13
*** kostaja has joined #mer15:13
*** Behold has joined #mer15:13
phaeronI think someone already mentioned doing so in #nemomobile15:13
VDVsxjkt, humm, I think I already saw this bug before, probably some patch got lost somewhere, have to take a look15:14
VDVsxjkt, this is qmf upstream right ?15:14
VDVsxthe one from gitourious15:14
*** odin_ has joined #mer15:15
phaeronbut it was rcg who talked about it  , but not sure if they are the same15:15
*** Sfiet_Konstantin has joined #mer15:16
*** BeholdMyGlory has quit IRC15:16
*** vgrade has quit IRC15:17
*** stefan_schmidt_w has quit IRC15:18
*** fcorrea has quit IRC15:18
*** fcorrea_ has joined #mer15:18
*** fcorrea has joined #mer15:18
jktphaeron: well, someone sent me a patch for building via MeeGo's OBS, but it was rather crude and unconditionally disabled the desktop builds15:18
jktVDVsx: the code in Trojita comes from QMF on Gitorious, yep15:19
VDVsxjkt, btw how you handle accounts in your app, via qmf as well ?15:20
jktVDVsx: I have a standalone IMAP implementation; the only 3rd party code I ship is low-level stuff like decoding of IMAP's modified utf7, or RFC2047 processing15:21
*** hazchemix has quit IRC15:21
jktVDVsx: it supports just a single account for now15:22
*** jstaniek_QFridge has quit IRC15:22
jktVDVsx: it's using Qt's tree models rather heavily, along with lazy loading and stuff like that, so it feel pretty fast on reasonable IMAP servers15:22
VDVsxsoon qmf in mer will have support for accounts-qt, can also be an option for you in the future15:22
jktit has support for ESEARCH, QRESYNC and what not15:22
jktVDVsx: I'll see if it's usable on desktop15:23
VDVsxjkt, yes, it is15:24
VDVsxalready on ubuntu15:24
*** mikhas has joined #mer15:24
*** sroedal has quit IRC15:24
*** sroedal has joined #mer15:24
VDVsxmaybe in others didn't check15:24
jktVDVsx: but I suspect it would be of a limited functionality because Trojita can use weird things like connect to the IMAP server via QProcess (so that you can set up the SSH keys along ssg-agent and let your MUA launch a QProcess doing `ssh foo dovecot --exec-mail imap` etc)15:25
* VDVsx kicks mikhas :P15:25
VDVsxjkt, in that case, I think will limit you a bit15:25
mikhasouch, that hurt!15:26
* mikhas gets beer15:26
*** NIN101 has joined #mer15:26
*** pa has joined #mer15:26
*** jayrulez has quit IRC15:29
*** pa has quit IRC15:31
mikhasVDVsx, next time you go to Pt (in 2013 or so), you should tell me beforehand, I havent been there in 13 years now15:34
mikhastime to visit your home country again!15:34
VDVsxmikhas, im going now in Christmas time, then not sure when again :)15:35
VDVsxthere's a big conference there next month, I think cybette has there last year, but not sure what are the topics this year15:36
kallecarllbt: ping15:41
*** jayrulez has joined #mer15:54
*** jkt has left #mer16:00
*** fcorrea has quit IRC16:02
*** reels__ has quit IRC16:09
*** morphis|away is now known as morphis16:11
*** spiiroin has quit IRC16:12
*** pirut has quit IRC16:15
*** stezz has joined #mer16:17
*** GeneralAntilles has quit IRC16:22
*** jayrulez_ has joined #mer16:23
*** jayrulez has quit IRC16:23
*** khetzal has joined #mer16:25
*** stezz has quit IRC16:29
*** GeneralAntilles has joined #mer16:30
sledgesdo you know any good Intel Atom dev board (sized like pandaboard?)16:32
Stskeepswell, the old dcfl2something weren't bad16:33
Stskeepsbut maybe a bit bigger16:33
Stskeepsthan pb16:33
Stskeepsdidn't they have superMUC or something?16:33
kallecarlw00t: any progress on the presentation for Qt Dev Days?16:34
*** calvaris has quit IRC16:37
*** faenil has joined #mer16:37
sledgesdcfl does not yield in google :{16:39
Stskeepsaren't there some good IVI boards?16:40
sledgesCongatex is not anymore maintained :(16:40
sledgesalso "cheap" would be nice :D16:40
Stskeepswelcome to the world of device making..16:40
Stskeepssituation i've been in, hello, here's a board, costs 200.000 eur16:41
Stskeepsor some similarly ridicolous number16:41
Stskeepsmy omapzoom2 was like 2k usd16:42
*** plfiorini has joined #mer16:42
*** vgrade has joined #mer16:42
Stskeepsevening plfiorini, vgrade16:42
vgradehi Stskeeps16:42
vgradeplfiorini: hey16:42
plfiorinihi vgrade16:43
vgradeplanning to package hawaii this weekend16:43
sledgesouch, they start from $999, same with tegra. I mean, still, mostly ARMs look cheap: iMX, Bb, Pb, RPi16:45
vgradeand may be look at some older hardware16:45
*** popey has quit IRC16:45
*** rcg has joined #mer16:45
vgradehi sledges , still on the MeeGo UX16:46
sledges:D hey vgrade, no, fought enough for Mer, it looks like boss is giving it a chance, hey hey hey! with ArCon-IVI upfront ;)16:46
* Stskeeps thinks anybody with the existing middleware for ivi could do a sane UI for one with lipstick16:46
* sledges has just actually won another fight - vs Tizen this time :)16:47
vgradesledges: nice work16:47
sledgesthanks, and not even one LOC :)16:47
Stskeepsyocto's growing up nicely, happy to see that16:48
sledgescan it embrace Mer(& Co.) soon?16:48
Stskeepsyocto's is it's own breed16:48
Stskeepsand covers some areas mer really doesn't want to touch16:49
kallecarlanyone know the whereabouts of w00t?16:50
sledgesis what it embraces i assume16:51
sledgesa beast that yocto's tamed16:51
vgradesledges: Intel have some nice IVI boards, don't know pricing though16:51
sledgesstill probably won't beat pandaboard with price, vgrade16:52
*** Attie has quit IRC16:52
sledgesand if we find an Intel based board, then my boss will go ahead with Tizen :D so ssssh :)16:52
vgradethis for production or just devel16:53
sledgesso might enter production if succeeds16:53
sledges(well, a low volume, or just the idea will be sold, no HW involved further)16:53
*** BeholdMyGlory has joined #mer16:53
*** lamikr has quit IRC16:54
*** Behold has quit IRC16:56
*** CosmoHill has joined #mer17:10
*** jayrulez_ has quit IRC17:11
*** jayrulez has joined #mer17:11
*** M4rtinK has joined #mer17:13
*** rubdos has joined #mer17:16
*** slaine has quit IRC17:18
*** aportale has quit IRC17:20
*** FSCV has quit IRC17:24
*** tilgovi has joined #mer17:27
Bostikhahaaa! qtwebkit, submit before your master!17:30
Bostikgot the library finally linked, now it's just a matter of finetuning the rest with a 25kg sledgehammer17:31
lbtseriously good stuff :D17:32
Bostikwhen we meet face to face, someone owes me a beer ... or ten17:32
lbtso.... gory details?17:32
lbtoh yes17:32
*** FSCV has joined #mer17:32
Bostikgory details? I took a *very* good look at the failing linker invocation and noticed that there was ONE item I could still drop - WebKit2, the split-process plugin architecture section17:33
BostikI realised that library alone must contain a good number of similar symbols as the rest of library and took it out on a desperate whim17:34
lbtOK, yes17:34
lbtso, how do we deal with this moving fwd?17:34
Bostiknon-debug object files, no WebKit2, gold linker and heavily stripped components... it fits into 32-bit linker17:34
Bostikeventually we will need a 64-bit build environment, there's no way around that17:35
lbtx32 arch for mer maybe?17:35
*** Sfiet_Konstantin has quit IRC17:35
lbtno, that's the inverse isn't it?17:36
lbtin which case how will Webkit work on that?17:36
Bostikgood question, but one I'm not ready to think about right at this moment17:36
lbtno, I suspect "beer" is topmost?17:36
BostikI still have the whole armv* stuff to deal with, but the first one is now *DOWN*17:37
Bostikyes, have one right here17:37
*** Behold has joined #mer18:04
*** KaIRC has quit IRC18:06
*** BeholdMyGlory has quit IRC18:07
*** ka6sox-away is now known as ka6sox18:09
*** Sfiet_Konstantin has joined #mer18:09
vgradeBostik: well done18:11
situBostik: you used gold for the final build ?18:12
*** notmart has quit IRC18:13
*** jstaniek has joined #mer18:14
*** Sfiet_Konstantin has quit IRC18:21
*** lbt is now known as lbt_away18:22
*** NIN101 has quit IRC18:22
*** andre__ has joined #mer18:25
*** andre__ has joined #mer18:25
*** khetzal has quit IRC18:29
*** BeholdMyGlory has joined #mer18:31
*** mikhas has quit IRC18:32
*** FSCV has quit IRC18:33
*** Behold has quit IRC18:34
Bostiksitu: yep, and I trick it into use because gcc 4.6 doesn't do it (at least the version in mer didn't)18:37
*** jayrulez has quit IRC18:39
Bostikbasically "copy to path under /tmp as 'ld'; prepend said path in $PATH to ensure it's the first discovered, pray"18:40
*** tilgovi has quit IRC18:41
*** Behold has joined #mer18:46
*** pvilja has joined #mer18:48
vgradeBostik: can you use alternatives, bottom of this page
*** BeholdMyGlory has quit IRC18:49
*** clopez has quit IRC18:54
*** khetzal has joined #mer18:54
*** morphis is now known as morphis|away18:55
*** alien_ has joined #mer18:56
*** nitrate_ has joined #mer19:03
*** lizardo has quit IRC19:03
Bostikvgrade: alternatives require root, which the build process is not supposed to have :)19:05
*** nitrate__ has quit IRC19:07
*** lamikr has joined #mer19:09
*** ortylp has joined #mer19:12
*** wmarone has quit IRC19:16
*** wmarone has joined #mer19:16
vgradeBostik: ah ok19:17
*** cat_x301 has quit IRC19:17
*** cat_x301 has joined #mer19:18
*** rubdos has quit IRC19:21
*** alien_ has quit IRC19:24
*** pohly has quit IRC19:24
*** arc_mat_ has quit IRC19:25
CosmoHillhey vgrade and Bostik19:31
vgradeHi CosmoHill19:32
vgradegood week?19:32
CosmoHillyeah pretty much19:33
*** arcean_ has joined #mer19:38
*** arcean has quit IRC19:39
*** arcean_ is now known as arcean19:39
*** furikku has quit IRC19:42
*** phdeswer has quit IRC19:43
*** tilgovi has joined #mer19:43
*** slaine has joined #mer19:54
Bostikhi there19:55
Stskeepsstill no good?19:55
*** ortylp has quit IRC19:55
BostikI won19:55
Stskeepsyou got it to build?19:55
Bostikread the backlog *g*19:55
Stskeepsoh dear19:56
Stskeepsi wonder how that'll play with evolution of any OS that uses webkit really19:57
Bostikor just wait for me to commit the whole spec+patch monstrocity and rant in the message19:57
Bostikthe fact that qtwebkit with webkit2 is impossible to link at all in 32bit userland really doesn't look good for most users19:58
Stskeepscan't those two somehow link independently?19:59
Stskeepsi mean, when you do a qtwebkit application you either go for one, or two19:59
Bostikthe "api" target apparently can't, I had to drop WebKit2 entirely and then it went through19:59
Bostikand yes, it was linking everything as shared20:00
Stskeepsapi target is a webkit thing, or qtwebkit?20:01
Stskeepssounds fun20:05
Stskeepsmer in .. 3.. .2.. 1?20:05
*** andre__ has quit IRC20:06
CosmoHillthat should be the boot loader20:07
BostikStskeeps: I think it might be qtwebkit only20:11
*** ortylp has joined #mer20:12
Stskeepswell just as a thought, it may be interesting to split it up in two .so's20:13
Bostikwe can start to figure out neat solutions and pester qt-releasing when we have caught up with them, I think20:14
Stskeepswho's the qtwebkit crowd anyhow?20:16
Stskeepslike, people you'd be able to talk to20:16
Bostikdamn good question20:16
Stskeepsi mean, when you cease to be able to build on 32-bit reference platforms..20:19
vgradeBostik: Stskeeps
Stskeepsthat looks interesting20:19
Bostikvgrade: thank you, that is a valuable link indeed20:19
* Stskeeps passes vgrade a cold beer20:20
vgradecheers, one for Bostik also20:20
*** andre__ has joined #mer20:20
*** ka6sox is now known as ka6sox-away20:33
* ljp updates his qtmobility20:37
*** khetzal has quit IRC20:41
* Stskeeps reads
Stskeepsthe title 'software engineer' always struck a bit weirdly to me from a danish perspective, ie, a computer scientist isn't an engineer :P20:46
Stskeepsthough i guess it's about the role you work in20:46
leinirStskeeps: weeell... that's an interesting one, though... my brother's uni course is Software Engineer... which /is/ an engineering approach to software development20:46
*** rdqfdx has quit IRC20:46
Stskeepsyeah.. i think i could take a similar one but never did20:46
leinirrather than the theoretical mathematics approach of computer science20:47
w00tTIL that Stskeeps never suffered software engineering20:47
Stskeepsi took evil stuff like science history and HCI20:47
Stskeepswhich proved to be very useful20:47
* w00t inserts nasty remark here20:47
w00t(I kid, I kid)20:47
Stskeepsthe systems couldn't really address a computer science student wanting to learn about science&technology :P20:47
Stskeepsand how it comes about in past20:47
leinirYou always get HCI now - even pure compsci has that as a requirement now :)20:48
leinirone semester's worth of a course of it, but still ;)20:48
leinirEvil, but as you've noted, useful :)20:48
Stskeepssocial computing was probably what i learnt the most from20:48
leinirMy course, of course, was informatics, not compsci - so i did three semesters of HCI stuffs ;)20:49
Stskeepsin the end: i'm just damn happy all that's over with now ;)20:49
*** Ian--- has joined #mer20:50
*** felipec has joined #mer20:51
*** ortylp has quit IRC20:52
*** Ian-- has quit IRC20:52
leinir*giggles* i really rather enjoyed my time at uni... could've done more, but... now i'm with KO and doing incredible things there as well, so i'm fine ;)20:54
w00ti've mixed feelings on not having studied20:55
w00tsometimes i wish i had20:55
w00tbut otoh in a lot of ways it's good i didn't, else i'd not be where i am20:56
*** kostaja has quit IRC20:56
Stskeepsi enjoyed when i was part of the start-uppy part of starting a research centre at university, where i was part of decisions, my views and ideas were listened to.. until there started coming in managers and me not being part of it anymore20:56
leinirOh yeah, of course important that i went to AAU, it's a rather unique way of studying :)20:57
Stskeepsbetter AAU than RUC..20:57
*** jpetersen has quit IRC20:57
leinir*laughs!* Yeaahh... RUC is sort of like AAU... done not quite as right ;)20:58
* Stskeeps 's proudly from, even though they lost my @@% diploma in the mail 3 times20:58
leinirpeople who i've met who attended RUC have tended to agree ;)20:58
leinir*facepalms* D'oh :P20:58
*** arc_mat_ has joined #mer21:04
*** lamikr has quit IRC21:06
*** arcean has quit IRC21:12
*** arcean has joined #mer21:14
*** jluisn has quit IRC21:15
leinirninis :)21:20
*** arcean has quit IRC21:25
*** arcean has joined #mer21:32
*** M13 has joined #mer21:39
*** arcean has quit IRC21:45
*** arcean has joined #mer21:46
*** fcorrea has joined #mer21:50
*** fcorrea has joined #mer21:51
*** gabrbedd has quit IRC22:05
*** VDVsx has quit IRC22:06
*** trbs has quit IRC22:10
*** ALo has joined #mer22:25
*** arcean has quit IRC22:26
*** dakovaci has joined #mer22:26
*** andre__ has quit IRC22:29
*** pa has joined #mer22:31
*** M4rtinK has quit IRC22:32
*** Jucato has quit IRC22:43
*** slaine has quit IRC22:44
*** pvilja has quit IRC22:52
*** pvilja has joined #mer22:53
*** rcg has quit IRC23:07
*** CosmoHill is now known as Cosmo|zzz23:09
Cosmo|zzznight night23:09
*** Aurium has quit IRC23:12
*** andre__ has joined #mer23:17
*** andre__ has joined #mer23:17
*** pa has quit IRC23:32
*** fcorrea has quit IRC23:37
*** tilgovi has quit IRC23:40
*** popey has joined #mer23:56

Generated by 2.11.0 by Marius Gedminas - find it at!