Friday, 2012-03-02

*** himamura has joined #mer00:11
*** raignarok has quit IRC00:29
*** tilgovi has quit IRC00:33
*** lynxis has joined #mer00:49
*** beyondcreed has joined #mer00:57
*** odin_ has joined #mer01:08
*** odin_ has quit IRC01:10
*** odin_ has joined #mer01:11
*** ZiQiangHuan has joined #mer01:17
*** KaiRo_Mozilla has quit IRC01:24
*** Alison_Chaiken has joined #mer01:28
*** lilstevie has quit IRC01:30
*** beyondcreed has quit IRC01:33
*** beford has joined #mer01:40
*** kavurt has quit IRC01:46
*** himamura has quit IRC02:09
*** bigbluehat has joined #mer02:29
*** jonnor_work has quit IRC02:43
*** eman` has joined #mer02:45
*** bigbluehat has quit IRC02:49
ZiQiangHuanStskeeps: When I play mp3 in my qtmediahub but no success, then  I checked the xsession-errors carefully and found a message "open: Permission denied". Maybe that's the reason. Any idea about that ?02:50
*** sonach has joined #mer02:54
*** rolandx1 has quit IRC03:22
*** ZiQiangHuan has quit IRC03:31
*** rolandx1 has joined #mer03:34
*** himamura has joined #mer03:54
*** jluisn has quit IRC04:02
*** wmarone has joined #mer04:14
*** wmarone_ has quit IRC04:15
*** lynxis has quit IRC04:52
Stskeepsmorn05:05
Stskeepssonach: could you tell ziqianghuan to look if he's using a /dev/ file for mp3 play and check the permissions of that?05:06
kulveInteresting question in the mailing list: "[mer-general] Building Firefox with Mer SDK" (cross-comping for ARM)05:07
Stskeepsi'm seeing it as a good sign that people are -ahead- of what we're going to do next05:08
Stskeepsie, sb2 enabled sdk05:08
Stskeepsit shows we have a pretty good feeling of what's actually needed to be done :P05:08
kulvetrue but it would be nice to say that "you can do it like this" instead of "you will be able to do it like this in the future"..05:10
Stskeepsyeah05:11
Stskeepssoon enough..05:11
kulveI'll be doing all my compilations for RPi in SDK locally so hopefully it's working soon ;)05:11
Stskeepsit'll be quite simple really - you'll be able to make a image with 'mic', init it as a sb2 target, run sb2 on top of it with fast zypper/rpm/gcc/bash etc and just build ahead05:12
sonachStskeeps: good morning! I will tell ziqianghuan about this:)05:14
kulveis the purpose of the platform SDK to provide the right versions of the right tools? Could I in theory just install the right tools into my debian stable and skip the platform SDK stuff (just run mic directly in my debian)?05:14
kulveI think that's what I did with meego (although I had problems with "build" package etc..)05:15
Stskeepskulve: the purpose of the platform sdk is to provide the right version of right tools as the problem was that a -lot- of people had all sorts of bizarre issues on various distributions05:15
kulveok, then I understood it correctly :)05:16
Stskeepswhich broke down the ability to ramp up people to use meego as there were more pages on how to fix those bizarre things than instructions how to use them ;)05:16
Stskeepsso yeah, sure05:17
kulvea bit OT for me but was there a plan to provide that as a bootable image for vmware etc.? Is it missing something else than a bootloader and the kernel?05:19
Stskeepsi'm hoping that we'll put out the reference images coupled with nemo projects' x86 adaptations, yeah05:20
Stskeepswith LLVMpipe05:20
kulveLLVMpipe?05:21
Stskeepsyeah, fast GLES software rendering05:21
Stskeepswe had 60fps on x86 virtualbox of libmeegotouch apps05:21
kulveok05:21
kulvethat's meant for running the target software?05:22
Stskeepssure05:23
kulveand does it work also for arm?05:23
Stskeepsllvmpipe doesn't work on arm at the moment, due to bugs in LLVM's jit :/05:23
kulveI really dislike the meego way when the doc let you assume you'll be able to do what ever and later it turns out and they were talking about x86 only..05:24
Stskeepsyeah..05:24
*** kimitake_idle is now known as kimitake05:25
Stskeepsbut the goal is anyway to let people develop mer + mer products in a VM without a device, without it being slideshow05:25
kulveI propose that the documentation talks generally only about the stuff that works both for x86 and for arm. If something extra (like running the target binaries with GUI) works only for either one, that's something documented separately as experimental05:26
Stskeeps:nod:05:26
Stskeepsmakes sense05:27
*** lynxis has joined #mer05:27
*** lynxis has joined #mer05:27
Stskeepsthere'll be some footnotes for that in mer for sure, ie, availability of libGL API is x86 only05:27
Stskeepsand libGLESv2 is what you should use for portability05:27
kulveI guess the platform sdk doesn't support mips yet?05:28
Stskeepsi have a patch for 'mipsel' for the mic tools i wanted to upstream05:29
*** lynxis has quit IRC05:29
kulvetime for me to get to the office and continue with the ogles stuff05:30
Stskeepsalright05:30
Stskeepssee you in a bit then05:30
timophmorning05:31
Stskeepsmorn timoph05:31
* Stskeeps ponders to brew some coffee05:35
*** jukkaeklund has quit IRC05:35
*** ZiQiangHuan has joined #mer05:36
*** cxl000 has quit IRC05:39
ZiQiangHuanStskeeps: I'll check it as you just said.05:41
Stskeepsalright05:41
Stskeepsit sounds a bit like a /dev node with bad permissions you first see the problem when running as user05:41
ZiQiangHuanStskeeps: Maybe I run the command "gst-launch -0.10 -v playbin2 uri=file:///mnt/usb/movies/test1.mp3 " as the root user, qtmediahub is run by user mer05:42
StskeepsZiQiangHuan: yes05:43
Stskeepsyou should make sure your applications can run as user, running as root is bad :)05:43
ZiQiangHuanStskeeps: But how I can run the command "gst-launch-0.10" as user05:43
StskeepsZiQiangHuan: su mer05:44
Stskeepsand test that way05:44
*** machine2 has joined #mer05:55
Stskeepshello machine205:56
sonachStskeeps: where is the source code of 'qmlwebkitplugin' in MerDS's source code? I cannot find it...06:13
Stskeepssonach: qtwebkit i think06:13
Stskeepswarning: big big source code06:14
sonachI want to have a look at how qtwebkit is pluginned into qml.06:14
sonachok,06:14
Stskeepshttp://doc.qt.nokia.com/4.8-snapshot/qmlwebkit.html is documentation how to use it, http://doc.qt.nokia.com/4.7-snapshot/qml-extending.html is how to extend QML with C++ code06:15
Stskeepser, http://doc.qt.nokia.com/4.8-snapshot/qml-extending.html might be better06:15
Stskeepshttp://doc.qt.nokia.com/4.7-snapshot/qtwebkit-bridge.html is extending webkit/html5 with C++ backend things06:19
*** tomeff has quit IRC06:21
*** cxl000 has joined #mer06:31
*** jukkaeklund has joined #mer06:31
sonachStskeeps: thank you for the links!06:35
Stskeepsnp06:35
*** beford has quit IRC06:41
*** beford has joined #mer06:41
kulveStskeeps: so, I have pvrsrvkm.ko loaded and initialized succesfully and I don't know if I should get omaplfb.ko from somewhere. What to try next? :)06:48
Stskeepskulve: well, pvrsrvinit (?) works fine i presume06:49
Stskeepskulve: start up X, make sure libGLESv2.so.2 and libEGL.so.1 points to your SGX's libEGL.so libGLESv2.so06:49
Stskeepsand get mer-gfx-tests packae06:50
Stskeepskage06:50
kulveI run that via "/etc/init.d/rc.pvr start" and based on the debug prints the pvrsrvinit is run fine06:50
Stskeepsok06:50
Stskeepsand run 'glestest'06:51
kulvethe SGX's installation script created the proper libGLESv2.so.2 and libEGL.so.1 links, nice06:51
Stskeepsoh finally some sanity06:52
Stskeepsthe story is that the SGX drivers build system is so utterly screwed up that there's no way to get a proper soname in06:52
Stskeepseven intel ended up (when making modern sgx for their devices) just symlining06:52
Stskeepslinking06:52
*** pohly has joined #mer07:04
*** jrayhawk has quit IRC07:05
*** machine2 has quit IRC07:08
*** machine2 has joined #mer07:09
StskeepsREMINDER: advisory board meeting at 12:00 UTC07:10
*** lamikr has joined #mer07:11
ZiQiangHuanStskeeps: memmap():open /dev/mem error!07:13
ZiQiangHuanStskeeps: But I changed /dev/mem to "crw-r--r-- 1 root kmem 1, 1 Jan  1 00:00 /dev/mem"07:14
StskeepsZiQiangHuan: eep, your hardware interfacing depends on accessing /dev/mem ?07:15
ZiQiangHuanStskeeps: Maybe07:15
StskeepsZiQiangHuan: that's not good hardware interfacing :)07:15
StskeepsZiQiangHuan: /dev/mem cannot be accessed by user07:15
Stskeepsas it would be a big security hole07:16
*** norayr has joined #mer07:16
ZiQiangHuanStskeeps: yes. But how can I solve this problem?07:16
Stskeepsnot sure07:16
Stskeepsyou can run things as root but that's not good either07:16
kulveit's so nice that the mer shell over the serial console has proper .. shell. E.g. ctrl-r works even over a boot :)07:17
Stskeepskulve: yeah.. no promises it's going to stay like that, or well, you'd probably have a choice between busybox and the usual gnu toolset07:17
ZiQiangHuanStskeeps: you mean "chmod" can't solve this ?07:17
StskeepsZiQiangHuan: correct07:17
ZiQiangHuanStskeeps: How I can make qtmediahub run by root07:18
ZiQiangHuanStskeeps: I see QMH is run by user mer now07:18
StskeepsZiQiangHuan: moment07:18
StskeepsZiQiangHuan: /etc/sysconfig/uxlaunch   ,, add user=root07:20
Stskeepsinstead of what is user= now07:20
ZiQiangHuanStskeeps: I will have a try07:20
Stskeepsthis is a security problem, but since you're prototyping.07:20
Stskeeps.07:20
kulveStskeeps: http://pastie.org/3502277 (I have no idea about the DRI2 and I guess GLX is something not used with OpenGL ES?)07:21
Stskeepskulve: looks like your libEGL/libGLES is mesa?07:22
ZiQiangHuanStskeeps: yes. Maybe we will talk about this with the device's vendors07:22
*** machine2 has quit IRC07:22
kulveStskeeps: yeah. I'll take back that proper linking comment ;)07:23
*** Eren has joined #mer07:23
Erenhi folks07:23
*** niqt has joined #mer07:23
Stskeepsmorn eren07:23
ErenStskeeps: what's your timezone?07:24
StskeepsEren: UTC+107:25
ErenStskeeps: oh, you wake up early :)07:26
niqtmorn all07:26
StskeepsEren: i've been up for 2h30m07:26
Erenniqt: morning07:28
ZiQiangHuanStskeeps: The content in /etc/sysconfig/uxlaunch is all started with "#", It seems this file isn't be used now07:34
StskeepsZiQiangHuan: it's used07:35
ZiQiangHuanStskeeps: you mean it's be checked by systemd ?07:35
StskeepsZiQiangHuan: it's used by uxlaunch07:37
*** veskuh has joined #mer07:45
*** jrayhawk has joined #mer07:46
*** kthomas_vh_ has quit IRC07:47
*** kthomas_vh_ has joined #mer07:47
*** mdavey is now known as mdavey_away07:48
ZiQiangHuanStskeeps: I changed it to http://pastie.org/3502377, I found the qtmeiahub is still started by mer07:48
StskeepsZiQiangHuan: without the spaces07:50
Stskeepsjust user=root07:50
*** slaine has quit IRC07:52
ZiQiangHuanStskeeps: I can't log in my system now07:52
ZiQiangHuanudevd[62]: unable to receive ctrl connection: Function not implemented07:53
*** mdavey_away has quit IRC07:53
Stskeepswhat kernel version?07:53
ZiQiangHuanStskeeps: 2.6.3507:54
Stskeepsyou might need http://git.angstrom-distribution.org/cgi-bin/cgit.cgi/meta-texasinstruments/plain/recipes-kernel/linux/linux-omap4-2.6.35.7/0001-ARM-6329-1-wire-up-sys_accept4-on-ARM.patch?id=05d8fa46706b147e91655651f3c479ae36107eb207:55
Stskeepsin your kernel07:55
*** leinir has quit IRC07:59
*** kimitake has quit IRC08:05
*** Alison_Chaiken has quit IRC08:05
sonachStskeeps: Is there any method to change the permission of '/dev/mem'?08:07
Stskeepssonach: disabling http://cateee.net/lkddb/web-lkddb/STRICT_DEVMEM.html -might- work08:11
Stskeepsbut again, it's not a good way to use /dev/mem in the first place :)08:11
Stskeepsyou should be able to get uxlaunch to start as root08:12
sonachStskeeps: why now Mer starts Xorg with user "mer"?08:14
StskeepsSage__: was i586 still blocked?08:14
sonachStskeeps: I find "user --name mer  --groups audio,video --password rootme" in ks file. Does this matters?08:16
*** smoku has joined #mer08:17
Stskeepssonach: yes, but uxlaunch usually autodetects08:18
Stskeepssonach: set user=root in /etc/sysconfig/uxlaunch08:18
*** ericlr has joined #mer08:18
sonachStskeeps: set user=root and patch the 2.6.35 kernel would solve this. is that right?08:19
Stskeepswell, i'm a little surprised it won't boot up earlier08:19
Stskeepsbut yes, one of the ways08:19
sonachStskeeps: OK:) Now we just start Xorg and qtmediahub after login to test other stuff...08:20
*** sigkill_ has joined #mer08:20
sonachStskeeps: after login, it is root:)08:20
Stskeepsok08:20
kulveStskeeps: not making good progress: http://pastie.org/350250008:20
Stskeepsok, can you try with glestest instead?08:21
kulveI think the mer-gfx-tests complain about libGL.so? I don't have that anywhere. And the sgx_init_test says "FAIL - PVRSRV_ERROR_NO_DC_DEVICES_FOUND"08:21
Stskeepsi'll check eglinfo08:21
*** kimitake has joined #mer08:21
kulveI tested both, just sec08:21
kulve# glestest08:22
kulveNo protocol specified08:22
kulvecannot connect to X server08:22
Stskeepsand this was with export DISPLAY=:0 ?08:22
kulveyes08:22
Stskeepstry to do it as 'mer' user08:22
kulvehttp://pastie.org/3502521 "Unable to initialize EGL"08:23
Stskeepsmuch better08:23
*** s1gk1ll has quit IRC08:23
kulvesomething flashes on the screen08:23
kulveand I am still missing the omaplfb of which I know nothing about08:23
Stskeepscan i see your powervr.ini and ls -al /usr/lib/libpvrPVR*08:23
kulveWindowSystem=libpvrPVR2D_FRONTWSEGL.so08:24
*** leinir has joined #mer08:24
Stskeepsright, that's -probably- wrong :)08:24
kulvehttp://pastie.org/350252608:24
Stskeepsset it to X11WSEGL instead, ie,  libpvrPVR2D_X11WSEGL.so08:24
kulve-bash: vi: command not found08:25
kulve-bash: vim: command not found08:25
kulve:(08:25
kulve-bash: emacs: command not found08:25
kulve!08:25
Stskeepsit's a minimal image08:25
Stskeepsnano or sed ;)08:25
kulveno nano either. Sed do exists :)08:26
kulveI used echo08:26
Stskeepsor cat ;)08:27
kulvesame thing: http://pastie.org/350255008:29
kulveI assume I don't need to initialize pvr before starting X?08:29
kulvecurrently I do the initialisation manually afterwards08:29
kulve"PVRSRVMetricsTimeNow: not implemented"08:30
Stskeepsokay, can you sanity check the libGLESv2.so.2 links and libEGL.so.1 and then look into how possibly you could build omaplfb.ko ?08:31
Stskeepstest with sgx_init_test08:32
kulvesgx_init_test still complains about the same (PVRSRV_ERROR_NO_DC_DEVICES_FOUND)08:33
kulveand the links poin to libGLESv2.so.1.1.16.3977 and libEGL.so.1.1.16.397708:33
Stskeepsok08:33
Stskeepscan i see the Makefile for sgx kernel driver?08:34
*** M4rtinK has joined #mer08:34
kulvefind -iname makefile\* | wc -l08:35
kulve1908:35
Stskeepssigh..08:35
Stskeeps:P08:35
StskeepsGraphics_SDK_4_03_00_02/08:35
Stskeeps ?08:35
kulvethe kernel stuff is in GFX_Linux_KM08:36
Stskeepsyes, just wondering what version it was08:36
kulvethat one, yes08:36
Stskeepsis it 4.03.00.02 ?08:36
Stskeepso08:36
Stskeepsk08:36
kulveDISPLAY_CONTROLLER =omaplfb08:36
kulvethat's in https://netst.org/pub/linux/ARM/OMAP/TI/Graphics_SDK_4_03_00_02/GFX_Linux_KM/eurasiacon/build/linux/omap3630_linux/makefile.shared_conf08:36
Stskeepswhere did you pick up the pvrsrvkm.ko from?08:37
*** ZiQiangHuan_ has joined #mer08:38
*** ZiQiangHuan has quit IRC08:38
*** ZiQiangHuan_ is now known as ZiQiangHuan08:38
kulveit's build with make and installed to my specfied directory. I guess it's copied from ./eurasiacon/binary_omap3630_linux_debug/pvrsrvkm.ko during the install phase08:38
Stskeepsok08:39
kulveI can find omaplfb files only under omap3430 and dc_ti81xx, not from omap363008:39
Stskeepseurasiacon/build/linux/omap3630_linux/kbuild/Makefile:EXTRA_SUBDIRS += $(EURASIAROOT)/services4/3rdparty/dc_omap3430_linux08:44
Stskeepswth is going on here..08:44
kulvehmm that 3party 3430 directory include only .. yes..08:44
Stskeepsit should really build the omaplfb08:48
kulveso it seems08:48
*** raignarok has joined #mer08:48
*** dcthang has joined #mer08:48
kulvenot sure if those MAkefiles expect different path in EURASIAROOT..08:48
kulveI'll have longer lunch break now but I'll continue debugging that then08:49
Stskeepsok08:49
*** raignarok has quit IRC08:52
*** Anssi138 has joined #mer08:59
*** raignarok has joined #mer09:01
*** jukkaeklund has quit IRC09:02
*** Eren has quit IRC09:08
*** slaine has joined #mer09:09
*** lamikr has quit IRC09:12
Stskeepsgood morning slaine09:13
*** eocanha has joined #mer09:14
slaineMorning09:16
*** notmart has joined #mer09:18
*** notmart has joined #mer09:18
*** lamikr has joined #mer09:22
lbtmorning all09:23
Stskeepsmorn lbt09:23
*** jukkaeklund has joined #mer09:23
lbtmade some good progress on git packaging yesterday :)09:24
lbtfeeling happy about that09:24
lbtNeed to just check where the SDK is - I need to update the sdk in releases/09:25
lbtit's broken09:25
lbtAFAICT the fix I pushed yesterday to sdk-chroot fixes it though09:26
ZiQiangHuanStskeeps: now I can play mp3 in my qtmediahub09:27
slainemorning lbt09:28
StskeepsZiQiangHuan: congratulations09:28
lbthey slaine09:29
slaineI need coffee and food STAT09:30
* lbt sips coffee from almost-not-leaking coffee maker09:30
*** mdavey has joined #mer09:31
*** sonach has left #mer09:33
lbtso Stskeeps - how is 486 tools schedule looking?09:35
lbttools=cross09:35
Stskeepslbt: i found a bug and when changing this morning09:35
Stskeepsso waiting for updatesstorm to go through09:36
lbtOK - that soon eh?09:36
lbt:)09:36
Stskeepsi was considering a nap09:36
Stskeeps:P09:36
lbtI usually do that at ~3pm09:36
*** harbaum has joined #mer09:37
*** M4rtinK has quit IRC09:38
*** DocScrutinizer has quit IRC09:53
*** DocScrutinizer has joined #mer09:53
*** himamura has quit IRC09:56
*** bergie has joined #mer10:03
*** KaIRC has joined #mer10:04
*** sledges has joined #mer10:11
*** sledges has joined #mer10:11
sledgesgood mornign!10:11
Stskeepsgood morning sledges10:11
*** kimitake is now known as kimitake_idle10:11
* sledges just got SATA disk attached to imx53 running KDE Plasma PA210:11
sledgesinstead of µSD card10:11
*** jstaniek_QFridge has joined #mer10:11
sledgeswhat a speed boost!10:11
Stskeepshow fast bootup?10:12
sledgesbootup still slowish (but faster)10:12
sledgesbecause some calculations are performed, not only plain disk read10:12
Stskeeps:nod10:12
sledges<30>systemd[1]: Startup finished in 4s 731ms 204us (kernel) + 1min 30s 543ms 254us (userspace) = 1min 35s 274ms 458us.10:13
Stskeepsstill, not bad10:13
*** ericlr has quit IRC10:13
sledgesbut that is before KDE PA starts booting10:13
sledgeswhich is another 2-3 minutes10:13
sledgesstill an awesome down from 6-710:14
*** raignarok has quit IRC10:14
Stskeeps:nod10:14
Sage__Stskeeps: https://build.pub.meego.com/project/show?project=CE%3AMW%3AShared yes10:15
sledgesnow a big problem: the tilt sensor is acting as input device10:15
StskeepsSage__: huh10:15
sledges:) so if i rotate the board, i see the mouse rotating (and everytime being reset to the centre of the screen)10:15
StskeepsSage__: for the -old- repo?10:15
Sage__Stskeeps: yes the static release10:16
StskeepsSage__: ah.. the old version had a problem10:16
Sage__oh10:18
Stskeepsi'll look at it later today if it's ok10:18
*** raignarok has joined #mer10:19
*** jonnor_work has joined #mer10:24
*** jukkaeklund has quit IRC10:29
sledgeshm, overall KDE PA2 on imx53 (from SATA HDD) boottime - 2min13sec !10:32
Stskeepsdid you have GLES up yet?10:32
*** beford has quit IRC10:33
*** jluisn has joined #mer10:34
sledgesnope :)10:34
Stskeepsk10:34
sledgesi had to make it run faster (not from sd-card) before moving ahead10:34
kulveGraphics_SDK_4_03_00_02 is a mess, imo10:34
sledgesbut can already show it to my boss :)10:34
Stskeepskulve: yeah..10:35
kulvehttps://netst.org/pub/linux/ARM/OMAP/TI/Graphics_SDK_4_03_00_02/GFX_Linux_KM/eurasiacon/build/linux/omap3630_linux/makefile.shared_conf enables SUPPORT_OMAP3430_OMAPFB3 if kernel > 2.6.2910:37
Stskeepsand that DC doesn't exist, right10:37
kulvein that case https://netst.org/pub/linux/ARM/OMAP/TI/Graphics_SDK_4_03_00_02/GFX_Linux_KM/eurasiacon/build/linux/omap3630_linux/kbuild/Makefile includes dc_omapfb3_linux instead of dc_omap3430_linux10:37
kulveand that doesn't exist, yes10:37
kulveI explicitly added it but it won't compile.10:37
Stskeepsis DSS2 enabled in kernel?10:38
kulveSYS_CFLAGS += -DPVR_HAS_BROKEN_OMAPFB_H sounds a bit odd..10:38
kulveCONFIG_OMAP2_DSS=y10:38
StskeepsDSS or DSS2 ?10:38
kulve$ grep -c DSS2 .config10:38
kulve010:38
Stskeepsnm10:38
StskeepsDSS, yeah10:38
sledgesthe files /usr/lib/*GLES* are there, but how to make them do stuff?10:40
kulveomaplfb_linux.c:145: error: implicit declaration of function 'omap_dispc_request_irq' and that one seems to be exported in drivers/video/omap/dispc.c10:40
Stskeepssledges: get mer-gfx-tests and run glestest as user with export DISPLAY=:010:41
Stskeepskulve: perhaps a header to be included10:41
kulveyes10:41
kulvefirst thing one should always do (but I always forgot) after untarring a vendor tar.gz is: git init && git add . && git commit -m "..." -a10:43
Bostikoh yes10:44
Bostikvery convenient ;)10:44
* lbt notes that this may soon be part of standard packaging10:45
lbtalthough I may have to rebrand "packaging" because so many developers are scared of it...10:46
*** beford has joined #mer10:46
*** lizardo has joined #mer10:47
Bostikmust have something to do with associating "packaging" with "productisation and deployment"10:51
lbteither that or "being organised"10:51
lbtI suspect the latter :)10:51
BostikI personally find it extremely inconvenient to have anything lying outside $HOME that doesn't come from packages10:52
lbt*g* by definition you're a sysadmin10:52
Bostikthe only exceptions I make are the kernel and my manual mplayer builds10:52
Bostikhell, even all my own software I've written for personal use has proper packaging :)10:52
kulvefor me $HOME is often on NFS and thus I often use /usr/local as well10:52
lbthehe ... in that case I welcome you with open arms :)10:53
lbtI used to manage with /everything10:53
kulveStskeeps: omaplfb: Maximum number of swap chain buffers: 210:53
lbtbut I have more than that now :D10:53
kulvesome progress again10:53
Stskeepskulve: progress :)10:54
*** drussell has quit IRC10:54
kulveand a kernel crash..10:54
Stskeepsboom10:55
kulveI no nothing about systemd. If I want a quick and dirty way to run "/etc/init.d/rc.pvr start" during the boot, where to write it?10:55
StskeepsREMINDER: Mer advisory board meeting in #mer-meeting in about an hour, http://www.mail-archive.com/mer-general@lists.merproject.org/msg00276.html10:57
kulveStskeeps: http://pastie.org/350329410:59
Stskeepskulve: does it show anything on screen?10:59
kulveyes10:59
*** ericlr has joined #mer10:59
Stskeepskulve: congratulations :)10:59
Stskeepskulve: now try qmlviewer -opengl ?11:00
kulvealthough it does it like qmlview: seems to be using 800x480 although my display is 480x80011:00
kulvethanks11:00
Stskeepsmight be a framebuffer definition thing11:00
kulveno fbset..11:00
kulve[    14.156] (--) FBDEV(0): Virtual size is 480x800 (pitch 480)11:01
Stskeepsframebuffer rotation wrong, or something, maybe11:01
kulverotation = 0, var->rotate = 011:02
kulveI'll debug that later11:02
kulveqmlviewer is quite smooth with -opengl :)11:03
Stskeepsgood11:03
sledgesglxinfo and egltest kill Xserver :)11:03
Stskeepssledges: don't use those, use glestest in mer-gfx-tests11:03
sledgessorry that's what i used11:03
sledgesjust don't remember the names right11:04
Stskeepslooks like you don't have correct libEGL/libGLESv2 then11:04
sledgesglestest and eglinfo11:04
sledgesbut they are not referencing them11:04
Stskeepsdo you have your vendor's libEGL.so and libGLESv2.so and them named as libGLESv2.so.1 and libEGL.so.1 ?11:04
sledgesprobably it is not an immediate linkage11:04
sledgesyes11:04
sledgesi look at strace11:04
Stskeepssure it's your vendors and not mes?11:05
sledgesand they are opening other files than that11:05
Stskeepsmesa11:05
sledgesits from smoku:imx:ha11:05
sledgesi will try the latest11:05
*** Free-MG has joined #mer11:06
*** phaeron has joined #mer11:09
Stskeepsgood morning phaeron11:09
phaerongood morning11:09
Stskeepsyour kid feeling better yet?11:10
phaeronStskeeps: he's ok already for a few days. there was another emergency that kept me away these days11:10
Stskeepsalright11:11
Stskeepshope you're alright and the emergency over with11:11
phaeronwell at least handled. wife's father operation on Sunday most probably11:12
Stskeeps:nod:11:12
*** drussell has joined #mer11:16
kulveStskeeps: so, next steps? :)11:16
Stskeepskulve: well, go take a look at nemo .ks and take it from there i guess11:16
*** JoseLuis has joined #mer11:16
kulveooh, nemo. Finally :)11:17
*** Free-MG has quit IRC11:17
*** csslayer has joined #mer11:17
Stskeepsi have some doubts if texture from pixmap will work correctly, but i guess you'll see11:21
kulvedon't ruin my expectations! ;)11:22
*** dcthang has quit IRC11:27
*** koo4 has joined #mer11:28
kulveStskeeps: (or Sage__:) should I try with or without @Nokia N950 Support and @Nokia N950 Proprietary Support? I.e. should I include those and then just copy my SGX stuff over those? Or leave those out and just copy my SGX stuff to the rootfs?11:32
Stskeepswithout11:32
Stskeepsleave them out totally, it's a different hw adaptation :)11:32
Stskeepsdon't even add the repos for them11:33
Sage__kulve: as Stskeeps said. Should not install any of those packages to your image11:34
*** nicola_ has joined #mer11:34
*** niqt has quit IRC11:34
*** nicola_ is now known as niqt11:35
kulveack11:36
kulveand I guess all armv7hl -> armv7l?11:37
Stskeepsyeah11:38
kulveshould I add these from the qmlviewer.ks: xorg-x11-drv-evdev xorg-x11-drv-fbdev xorg-x11-server-Xorg-setuid -xorg-x11-server-Xorg and maybe @Mer Minimal Xorg?11:41
Stskeepsyeah, probably11:42
Stskeepsyou can probably switch to you and zuh's omapfb eventually11:42
kulveand related to my rpm arch problems, the nemo does these hacks (that I did with my meego chroot as well): "echo -n 'armv7l-meego-linux' > /etc/rpm/platform" and "echo 'arch = armv7l' >> /etc/zypp/zypp.conf"11:42
Stskeepsok11:43
Stskeepsshould autodetect for rpm at least11:43
*** jukkaeklund has joined #mer11:43
Stskeepsbut you mentioned you had issues11:43
kulveyes11:43
*** ssirkia1 has joined #mer11:44
*** ssirkia has quit IRC11:44
*** ali1234 has quit IRC11:44
*** KaiRo_Mozilla has joined #mer11:45
*** ssirkia1 has quit IRC11:45
*** ssirkia has joined #mer11:46
kulvethose .ks files have something like this in them: "part / --size 500 --ondisk sda --fstype=ext4". Could I just skip that, when doing a tar.gz that I just untar to a sdcard?11:46
lbt*g*11:47
lbtkulve: I had the same concern11:47
Sage__kulve: you need some kind of line in .ks file when doing tar.gz but it doesn't matter11:47
kulveok11:47
Sage__IIRC mic needs to know what kind of loop device it will create for the image build time.11:47
lbtthe argument is that the size is/can be used by mic to validate that you won't burst your device11:48
Stskeepskulve: your goal eventually is a replicate-able image at least11:49
*** KaIRC has quit IRC11:49
*** ssirkia has quit IRC11:54
*** ssirkia has joined #mer11:54
*** kurtul has joined #mer11:57
*** smoku has quit IRC11:57
StskeepsAdvisory board meeting in #mer-meeting now11:59
slaineHmmm, may as well setup fakeOBS on this server too11:59
lbtslaine: I usually run that on a discrete VM12:00
slaineThat's what I meant12:00
Stskeepsjukkaeklund: #mer-meeting?12:00
slaineI wasn't happy with using openSuSe on the phost, so reset that up last night12:01
jukkaeklundyes, thanks for reminding12:01
*** drussell has quit IRC12:01
slaineallocating the lvm's atm, so decided I should add a small one for fakeOBS12:01
*** drussell has joined #mer12:12
*** jarkko^ has quit IRC12:19
*** tomeff has joined #mer12:21
kulvewell, at least I see the nemo desktop :)12:22
kulvealthough I haven't started the SGX stuff..12:22
Stskeepsthat's more worrying ;)12:22
*** ZiQiangHuan has quit IRC12:24
kulveI made a new rootfs, so I don't have those GLES links etc. in place12:24
kulvewhere should I start the rc.pvr?12:24
Stskeepswe're in a meeting, let me answer afterwards12:24
kulveok12:25
*** aperezdc has quit IRC12:29
*** ali1234 has joined #mer12:35
*** himamura has joined #mer12:40
*** kurtul has quit IRC12:54
sledgesStskeeps, i'm one step closer: glestest shows blackscreen (same with qmlviewer -opengl), and eglinfo reports results12:58
sledgesso i believe opengl canvas is just rendering somewhere offscreen (KDE PA2 standing in a way?)12:58
Stskeepswhat does eglinfo say12:59
kulveI guess I should add something to /lib/systemd/system to run the rc.pvr before uxlaunch..13:02
*** djszapi has joined #mer13:03
djszapiHi! Is there a console environment apart from "osc" to test package builds locally to go through the trivial failures ?13:03
*** lamikr has quit IRC13:05
slainesledges: are you targeting an iMX53 based board ?13:06
*** ScriptRipper has joined #mer13:14
*** ScriptRipper has joined #mer13:14
kulveI'm getting SGX crashes after starting the PVR stuff manually: http://pastie.org/350401013:17
kulvealthough the uxlaunch is looping there anyway, so there are many things broken13:17
kulvebut I need to continue that on monday13:17
*** aperezdc has joined #mer13:21
*** phaeron has quit IRC13:25
*** zutesmog1 has left #mer13:35
*** bigbluehat has joined #mer13:45
*** Tm_T has quit IRC13:47
*** Tm_T has joined #mer13:47
*** jukkaeklund has quit IRC13:55
sledgesslaine: correct13:55
djszapilbt: Is there a console environment apart from "osc" to test package builds locally to go through the trivial failures ? Something like chroot environment or scratchbox(2) ?13:56
zumbidjszapi: osc chroot might do13:58
*** harbaum has quit IRC13:58
sledgesStskeeps: eglinfo: http://pastebin.com/b6DYNE2413:59
djszapizumbi: right13:59
djszapianybody having "osc" working on Archlinux in here ?14:00
eocanhaStskeeps: I'm following the instructions in http://wiki.merproject.org/wiki/Local_Mer but when I reach the step of14:02
eocanha"Copy the prjconf": osc -A merci meta prjconf fakeobs:Core:${BS_ARCH} | ...14:03
slainelbt, what was the breakdown of the vm's you wanted me to create again ?14:03
*** beford has quit IRC14:03
slaineI've got my phost setup, I've got a fakeobs one setup14:03
djszapianybody using Archlinux in here for development ?14:04
slaineI've got details for setting up 2 other vm's, one for fe/api and not sure what the other was for14:04
eocanhaI get a 404 not found from my local fake obs server. In my case it's: osc meta prjconf mer:Core:i58614:04
lbtslaine: we'll split them fe, be, worker for OBS. I tend to run MDS on it's own machine because I use debian for that14:07
slainewhat's mds ?14:08
lbtdjszapi: yes, osc chroot is very good for that (I assume harmattan?)14:08
* slaine notes this stuff seems to change every time he looks at it ;)14:08
lbtfor mer we have the new SDK but it too would use osc chroot for now14:08
eocanhaslaine: MDS = http://wiki.merproject.org/wiki/FakeOBS14:08
lbtslaine: mds is fakeobs but better branding :)14:09
slaineAh14:09
slaineWell, I've got MDS as a VM already ;)14:09
lbtgood14:09
slaineAllocations you mentioned yesterday where, fe/api == 6G /  10G /var  2G swap  100G /srv14:10
lbtyes14:10
slaineSame for be, same for worker ?14:10
* lbt thinks ... does fe need a /srv14:10
lbtno, it doesn't14:11
slaineif it's not actually building anything and hosting anything, I shouldn't14:11
lbtso /srv just on be14:11
slaineI'm assuming fe is only the webui and control code14:11
lbtyes, webui, api, mysql14:11
djszapilbt: nope, mer.14:11
lbtdjszapi: so, for now, mainly osc chroot14:12
lbtdjszapi: we have the platform SDK too14:12
slaineWhats the difference between be and worker then ? I guess that's the next bit I'm not clear on14:12
djszapilbt: platform SDK as in scratchbox ?14:12
lbtdjszapi: I'm working on SDK SB2 right now14:12
djszapiliek on Harmattan ?14:12
lbtdjszapi: no, nothing like except the spelling14:13
lbtslaine: the be will do a few things: scheduling, dispatch jobs, serve binaries, serve sources, publishing14:14
slaineand the worker obviously then does the compilations14:15
*** veskuh has quit IRC14:15
slainewhats the benefit of separating out the be and the worker to different vm's then14:16
djszapilbt: my use case is a scratchbox/chroot alike environment where I can catch typos and other silly stuff without rebuilding everything from scratch like with osc. Osc chroot works to a certain extent. I have never seen anybody getting it work yet on Archlinux. :/14:16
lbtyes - the worker runs the same osc/build that you would on a desktop14:16
lbtslaine: you'll have lots of workers14:16
lbtdjszapi: I understand - that's my current task14:16
lbtthe platform SDK will enable that14:17
lbtfor now osc chroot is almost there14:17
djszapilbt: eta ?14:17
slainedoes the backend not spawn off workers as required ? I've got a single OBS VM at the moment that was able to rebuild from fakeOBS14:17
*** csslayer has quit IRC14:17
lbtslaine: no, think build farm14:17
lbthttps://build.pub.meego.com/monitor14:18
lbteach box is a 64Gb 4core machine14:18
lbteach row is a worker running on the box14:18
lbtdjszapi: days14:18
slaineI'm not really sure that would be optimal for a small company14:18
lbtslaine: that's why you're running a single VM14:19
slaineI've got access to A server which will have 18GB RAM and 8 Cores,14:19
lbtyou now allocate a 2nd box and run a couple more workers14:19
lbtOK - so you'll probably scale down the memory useage of the fe/be over time14:20
lbtand run some more worker VMs on it14:20
djszapilbt: could you please drop me a query or email ? I do not follow the Mer happenings too often, but I would come back with packaging some stuff.14:20
lbtyou'll probably run 2-3 workers on there using tmpfs14:21
lbtdjszapi: what for?14:21
lbtwhen we have SB2 ?14:21
lbtmaybe follow me on twitter if you do that - I rarely tweet anything non-Mer related14:22
djszapiI do not use twitter.14:22
*** raignarok has quit IRC14:22
lbtok - mer-general mailing list is fairly low volume14:23
lbtlow enough that we don't bother having an -announce yet :)14:24
djszapimmh, probably never mind since I need to get a package work before the PA sprint anyway14:24
djszapiwhich begins next Tuesday.14:24
slainelbt, so on the above monitor page, each worker is a VM14:24
lbtOK - not sure what you're trying to actually do djszapi ?14:25
lbtslaine: on that deployment yes14:25
lbtslaine: on that deployment, no14:25
lbt:)14:25
lbteach worker is a physical host there14:25
slaineso on that deployment, each worker is a 64GB 4core machine ?14:25
lbtyes14:25
slainewowza14:25
lbtmeego wasn't poor14:26
djszapilbt: package a dependency for KDE Mobile Application development.14:26
djszapiwith zero rpm knowledge.14:26
slainelbt, unfortunately, I am :)14:26
lbtslaine: Mer too14:26
lbthence we spend our money on these http://www.hetzner.de/en/hosting/produkte_rootserver/ex4s14:26
slaineSo, for me, and my single server, and I'll likely never get more than a single server, I need to figure out the optimal setup14:27
lbtnow that's a stonking good deal14:27
lbtyou run one of them with openvpn back to the office and you get a fair few high speed workers for peanuts14:28
slainepretty nice deal alright14:28
lbtjust sayin :)14:28
slaineI wouldn't really need all that building power though14:28
lbtno, indeed - one worker is a fine start14:28
lbtdjszapi: I think you'll be fine using osc chroot14:28
lbtwhat you may want to do is set one up14:28
djszapiyup, for the time being.14:29
lbtthen cp -a it somewhere else14:29
lbtthen mount --bind your home into it14:29
djszapibut osc does not work for Archlinux14:29
djszapiand I do not have access to the ubuntu server right now14:29
lbtOK, not a problem14:29
djszapiso I am kinda out of luck.14:29
lbtyou download the SDK14:29
lbt80Mb14:29
lbtand install that as a chroot14:29
lbtand do it in there14:29
*** pohly has quit IRC14:29
lbtit's a little more fiddly but I'll talk you through it14:30
ali1234did you compare that against AWS starting builders on demand?14:30
lbtali1234: no14:30
ali1234ok, fair enough14:30
ali1234i have no idea if it would be cheaper either14:30
lbtAWS had issues doing nested virt last time I looked14:30
lbtwe need that for security14:30
ali1234ah ok14:30
lbtHetzner give us raw metal so no nesting needed14:31
slaineWondering if for the optimal deployment, I should have 1 fakeOBS vm, 1 fe/be vm and 1 worker vm14:31
lbtslaine: nag14:31
lbtnah14:31
slaineI could then in theory add other workers if other servers became available14:31
lbteven14:31
ali1234i guess getting bare metal will always be a problem for cloud services14:31
lbtali1234: yep - nested kvm may help one day14:32
lbtslaine: the setup I described will give you the best flexibility14:32
*** norayr has quit IRC14:32
lbtand won't be improved by consolidating vms14:32
slaineI'm just wary that each VM will allocate resources away from the worker and that's the one that needs the most resources14:33
lbtthe main drawback is that memory isn't dynamically allocated14:33
lbtdon't overthink it14:33
lbtwhen you're talking $60/month for ^^^  your time is by far and away more important14:34
*** jukkaeklund has joined #mer14:34
slainetrue14:34
slaineSo, to recap14:34
slaine1 MDS vm, 1 OBS fe/api VM, 1 OBS be/scheduler VM, 1 OBS Worker VM14:34
lbtyes14:35
lbtyou only need a single large /srv on the be machine14:35
slaineMDS and fe/api should be pretty light weight, 1 cpu, 512MB14:36
lbtacceptable start - MDS will go smaller, fe too probably14:36
*** bergie has quit IRC14:36
slainecool, yeah, actually, I think I've my MDS one on my desktop down at 25614:36
lbtbut the good thing is that you can tune that later14:36
slaineanyway, can change that14:36
*** HazardousRooster has joined #mer14:36
slainesnap14:36
lbtI wrote this ages ago: http://en.opensuse.org/openSUSE:Build_Service_Backend_Key_Components14:37
slaineso, be is the big one14:37
lbtyes14:37
slainewhat about the worker14:37
*** HazardousWaster has quit IRC14:38
lbtsetup as fe but more memory - 6Gb14:38
lbtyou'll probably run 2 of them using tmpfs14:38
lbtthis we'll have to suck and see as it depends on the apps you build14:39
slainecool14:39
* lbt -> coffee14:39
*** eman` has left #mer14:39
slaineAnd then, be, big disk as discussed previously14:39
slainewhat about RAM/CPU14:40
slaineI could give it all 8 CPU's as I think kvm is clever about how it distributes load14:40
*** jstaniek_QFridge has quit IRC14:40
slainebut I'm not sure about how much RAM to allocate it14:40
slaine(also, I'm still waiting on my RAM delivery, so that might happen yet anyways)14:40
slaineI'll get the other VM's up and running14:41
slaineThanks for your patience lbt14:41
Stskeepseocanha: going out for shop, bbl 20m14:41
eocanhadon't worry14:41
lbtnp - don't forget that all this is adjustable - making lvs bigger is trivial, changing ram/cpu, the same14:41
eocanhaI've located the physical path of the file in my MDS setup and I've copied the contents of the file by hand14:42
*** MerBot has joined #mer14:57
*** JoseLuis has joined #mer14:58
*** jluisn has quit IRC15:01
*** HazardousWaster has joined #mer15:02
*** HazardousRooster has quit IRC15:03
*** phaeron has joined #mer15:04
*** JoseLuis has quit IRC15:06
*** jluisn has joined #mer15:06
*** eocanha has quit IRC15:08
*** eocanha has joined #mer15:08
*** tpn has joined #mer15:18
* slaine -> coffee15:24
djszapianybody having on idea for this spectacle issue ? http://paste.kde.org/432344/15:25
*** mlfoster has joined #mer15:27
slainelbt, back to the specs of the be and worker15:29
Stskeepsdjszapi: OS?15:30
djszapiStskeeps: Linux15:30
Stskeepsdjszapi: okay, distribution..15:31
djszapiArchlinux.15:31
Stskeepsno clue then15:31
Stskeepsgrab platform sdk, it has a working spectacle I believe15:32
*** MerBot has joined #mer15:51
*** dm8tbr has joined #mer15:52
Stskeepsphaeron: btw, any news on the dawati artwork stuff?15:53
*** lilstevie has joined #mer15:53
djszapiStskeeps: I see.15:54
djszapithanks.15:55
phaeronStskeeps: the git repos returned to normal afair and I built packages some time back15:59
*** kurtul has joined #mer16:00
Stskeepsphaeron: alright16:00
Stskeepsphaeron: i'm planning a total revamp of my home IT and putting mer (almost) everywhere, so :)16:00
phaeronStskeeps: I still have a bunch of interesting ux projects to try on top of mer16:02
phaeronjust I don't have enough real life time right now16:03
phaeron:D16:03
djszapimake a time creator application xD16:04
*** niqt has quit IRC16:04
djszapiStskeeps: the problem seems to be that this spectacle is not compatible with python3...16:06
Stskeepsdjszapi: install python2.6 then16:06
djszapiit is "print()" in there.16:06
Stskeepsor 2.716:06
djszapiI have.16:07
djszapiit picks up the wrong one, I believe.16:07
Stskeepswell, edit the #! then16:08
slainevgrade: I was contemplating looking into what it would take to get b2g up on nitroid on the n90016:08
slaineso many ideas, so little time16:08
Stskeepshopefully we'll turn mer into a playground for projects, where we have time to work on the actual projects instead of the base systems it needs16:09
phaeronStskeeps: that is a very true statement16:10
djszapiStskeeps: python2.7 setup.py build -> did the trick.16:10
phaeronimho the time it takes to try new stuff is low due to the small adequate core16:10
Stskeepsyeah16:10
phaeronand once you build the stuff once you can try many things with it16:11
phaeronbut I feel  the need to organize a middle ware project to share effort16:11
djszapi specify libkdeedu.yaml16:11
djszapiError: Cannot open "/usr/share/spectacle/pkgconfig-provides.csv", maybe the package was not installed correctly.16:11
djszapimeh16:11
Stskeepsphaeron: yeah.. my own future idea is to have the toolbox for new ui types16:12
phaeronright now everyone copypacs stuff around16:12
Stskeepsyeah16:12
slaineI'm looking forward to getting a middleware and ux project setup for our internal hardware and then being able to port that to different hardware targets16:12
Stskeepsphaeron: Mer was originally meant to be for http://projectgrande.wikispaces.com/16:13
phaeronStskeeps: I think the day made of glass video is a very strong depiction of that page16:14
Stskeepsyeah16:14
phaeronoh its linked there16:14
phaeron:D16:14
bigbluehatStskeeps: what happened to Grande?16:15
bigbluehatnever "made it"?16:15
bigbluehator just spun off into things like Mer?16:15
Stskeepsbigbluehat: well, not exactly true.. hang on16:15
bigbluehatk…sorry16:16
bigbluehat(and thanks)16:16
Stskeepshttp://gitorious.org/project-grande16:16
* Stskeeps tries to find alterego's video of state transfer16:16
*** pohly has quit IRC16:17
*** pohly1 has joined #mer16:17
djszapiso this was the final solution then: DESTDIR=/ PYTHON=/usr/bin/python2.7 make install16:18
Stskeepsbigbluehat: http://hacking-mobility.blogspot.com/2011/10/qmlhtml5-application-state-transfer.html16:19
djszapiis this the expected output for specify ? http://paste.kde.org/432386/16:19
Stskeepsdjszapi: depends, sometimes people do customizations in .specs16:20
Stskeepsdon't try to download ${} stuff16:20
djszapiwhy16:21
*** bigbluehat has quit IRC16:21
Stskeepsspectacle doesn't interpret rpm macros16:21
djszapimmh, that is not so good ..16:22
Stskeepssend a patch16:22
djszapino time, sorry :)16:22
djszapino knowledge either.16:23
Stskeepsthen a bug report works16:23
Stskeepsand we'll prioritise it in triage16:23
djszapihttp://paste.kde.org/432392/16:25
*** pohly has joined #mer16:25
djszapigrep -rn bz2 libkdeedu.yaml16:25
djszapi9:    - ${name}_${version}.orig.tar.bz216:25
djszapiis this an acceptable schema ?16:25
Stskeepsyeah, check the .spec output16:25
djszapiit would be consistent with the debian way this form.16:25
djszapihttp://paste.xinu.at/NusZi/16:26
*** alexxy has quit IRC16:26
Stskeepslooks fine, though you probably need to specify files16:27
djszapidoes not look too bad, whereas I, admittedly, have zero spec knowledge.16:27
Stskeepsthat's why spectacle is cool16:27
Stskeepsbrb, something screwed up in builder..16:27
djszapispecify files ?16:28
Stskeepsyou'll see an error with "unpackaged files" when done building the package, use Files: section in yaml16:29
*** raignarok has joined #mer16:29
Stskeepsto indicate what files are part of the package16:30
djszapiinteresting, other PA packages do not use that.16:30
Stskeepsthey probably have a .spec where its in >> files << files instead16:30
Stskeepssee the .spec16:30
djszapiright, so it is a post-modification after the specify run then ?16:31
djszapitheir spec files indeed contain such files :)16:31
djszapiis there way to pass everything into one package ?16:33
djszapiperhaps with wildcards or other way around ?16:33
Stskeepsyou can use wildcards but i would advise against "/*"16:34
djszapiso something similar to debian packaging when you do not have an install file, and just one "subpackage".16:34
*** bigbluehat_ has joined #mer16:35
djszapiAlthough, I guess I can just redirect cat ./debian/*.install to this file16:35
bigbluehat_Stskeeps: sorry I dropped off16:35
bigbluehat_ma's being dumb today16:36
Stskeepssec, ordering a noodle box16:36
bigbluehat_mac's16:36
bigbluehat_that is16:36
bigbluehat_:-)16:36
bigbluehat_no worries16:37
djszapimmmh, there are specific variables, like libdir.16:37
djszapibigbluehat_: do you develop on Mac ?16:37
bigbluehat_yep16:37
bigbluehat_like it most days16:38
Stskeepsbigbluehat: http://hacking-mobility.blogspot.com/2011/10/qmlhtml5-application-state-transfer.html16:38
djszapicool :)16:39
*** Alison_Chaiken has joined #mer16:39
bigbluehat_djszapi: you?16:40
Stskeepsdjszapi: %{_libdir} expands to /usr/lib16:40
Stskeepsphaeron: any idea why 'the-boss' quits/rejoins every 30 minutes?16:42
Stskeepson the spot16:42
phaeronhmm that's new16:43
bigbluehat_brb16:43
*** bigbluehat_ has quit IRC16:47
*** kurtul has quit IRC16:51
*** leinir has joined #mer16:52
*** djszapiN9 has joined #mer16:53
*** toscalix has joined #mer17:07
*** gimli has joined #mer17:10
*** bigbluehat has joined #mer17:12
*** NPX has quit IRC17:16
*** noopman has joined #mer17:17
*** toscalix has quit IRC17:17
*** toscalix has joined #mer17:18
*** NIN101 has joined #mer17:22
*** bigbluehat1 has joined #mer17:24
*** jarkko^ has joined #mer17:26
Alison_ChaikenAre Mer folks going to attend the May Tizen Conf?17:27
Alison_ChaikenMight there be a Mer satellite meeting?17:27
Stskeepsi submitted a talk at least17:27
Stskeepson cross compilation17:27
Alison_ChaikenI can arrange meeting space here in SF area.17:27
*** bigbluehat has left #mer17:27
sledgesStskeeps: did you take a look at the pastebin eglinfo of mine ?17:28
Alison_ChaikenCross-compilation in that a given Makefile has Mer and Tizen targets?17:28
Alison_ChaikenLike different linking?17:28
*** slaine has quit IRC17:28
*** phaeron has quit IRC17:31
StskeepsAlison_Chaiken: "In this talk, the various challenges, current, previous and under development approaches to cross compiling whole distributions such as Tizen, MeeGo, Mer and others will be discussed and analyzed. Lessons learnt from MeeGo cross compilation within OBS will be presented and put in contrast to approaches within other open source projects. Ideas about doing different kinds of mobile systems utilizing cross compilation ...17:32
Stskeeps... methods will be discussed as well as the day-to-day impact on platform developers and application developers by the cross-compilation approaches chosen."17:32
Stskeepssledges: try again17:33
sledgesStskeeps: eglinfo: http://pastebin.com/b6DYNE2417:33
Stskeepssledges: looks ok17:35
*** tomeff has quit IRC17:35
Stskeepsbut nothing shows on screen?17:35
sledgesnope17:35
sledgesi will try the kernel from the same place where i took all of the graphics things17:35
sledgesas i see stuff getting intertwined17:35
sledges/dev/gsl_mod17:36
sledgesfiles under /usr/lib/{*GLES*,*Q*}17:36
sledgeskernel modules17:36
sledgesso will fight17:37
sledgeson17:37
sledgesnice weekend!17:37
*** djszapiN9 has quit IRC17:37
Alison_ChaikenSuppose the Tizen money people said, "We want to have a relation with Mer like that Redhat has with Fedora"?17:38
Stskeepsfrankly, i'd be happy enough with tizen understanding their benefit in this world, ie, to make it a stamp for working web runtime with WAC, HTML5, Tizen apis17:38
Stskeepsthe answer on that seems a bit unclear by now17:39
*** tomeff has joined #mer17:39
Stskeepsmy own opinion is that tizen is more valuable to the general open source community alive than dead, though17:40
Alison_ChaikenFWIW, I agree.    Someone has to pay developers, who will do on their own time whatever they please, just as they always have done.17:41
Stskeepsbut let's see.. meanwhile, mer as a lifeboat doesn't sound like a bad idea, we're making great progress17:41
*** bigbluehat1 has quit IRC17:43
*** bigbluehat has joined #mer17:43
Alison_Chaikenmer is already going to be a success.17:43
Alison_ChaikenHey, even Maemo is going to ship in new Volkswagens according to a guy I talked to at Embedded Linux Conference!17:44
*** phaeron has joined #mer17:44
*** sledges has quit IRC17:44
Stskeepsmy deepest fear is still that someone found old Mer and reusing it for production purposes..17:45
Stskeeps:P17:45
*** eocanha has quit IRC17:46
bigbluehatStskeeps has skeletons in his closet…that people want :)17:51
*** beford has joined #mer17:51
* Stskeeps ponders idly what to work on tonight17:52
bigbluehatStskeeps: you were going to say or send more about Project Grande, I think ;)17:53
Stskeepsyeah, i gave you an url above17:53
bigbluehatand also some photoshop related fodder from a couple days back17:53
bigbluehatyeah, saw that…thought there was more17:54
bigbluehatno worries :)17:54
Stskeepsand a gitorious url with actual code17:54
* bigbluehat signed up for gitorious just for this ;)17:54
bigbluehatright :)17:54
bigbluehatare there screenshots on the wiki?17:54
* bigbluehat is digging back in17:54
bigbluehatalso, I tossed the wiki URL into #webos and swisstomcat (an HP dev engagement guy) said it looked like exciting/interesting stuff :)17:55
Stskeepsah, there's an actual #webos ?17:55
bigbluehat:)17:55
bigbluehatyep, I troll there too ;)17:55
*** phaeron has quit IRC17:55
bigbluehatand an #enyojs for their new framework17:55
Stskeepsi am fairly sure the reason HP bought Autonomy was for the 'a day made of glass' type of world17:55
bigbluehatwell, rewritten framework17:55
Stskeepsyeah, i'm in enyo17:55
Stskeepshttp://stage.rubyx.co.uk/grande/screencasts/20111222_003.mp4 is one of the (gpl) qml UIs we mocked up17:55
bigbluehatStskeeps: you've used webOS yes?17:56
bigbluehatparticularlly on the touchpad?17:56
Stskeepsi haven't actually tried it in person17:56
bigbluehatdude!17:56
bigbluehatyou need to :)17:56
bigbluehatlots of similar ideas…at least from a conceptual level17:56
Stskeeps:nod:17:56
bigbluehatthey have this Synergy account thing which integrates disparate info sources (mail, contacts, calendering sources) into a single "db"17:57
bigbluehatand from that into apps…Email, Calendar, Contacts :)17:57
Stskeepsgood idea, didn't work so terribly well performance wise in n917:57
Stskeepsi am a fan of json dbs though17:57
bigbluehatyeah, they're everywhere :)17:58
bigbluehatQT5 from what you (iirc) said recently17:58
bigbluehatI heart #couchdb17:58
bigbluehatMojoDB is #webos's17:58
bigbluehatand it's being open sourced this month17:58
Stskeeps:nod:17:58
bigbluehatreplicates with #couchdb, etc17:58
bigbluehatsorry…I'm keeping you from code17:58
* bigbluehat shuts up ;)17:58
Stskeepsnah, you're ot17:58
Stskeepsnot17:58
Stskeepsi'm out of work at the moment so my 'workday' is kinda over by now, think i should just enjoy a nice evening and look a bit around the web17:59
bigbluehatout of work like "unemployed"18:00
bigbluehator out of work, like you finished everything :)18:00
Stskeepswell, consultant temporarily without contracts18:00
bigbluehatah, been there18:00
Stskeepsso not a terribly bad situation18:01
*** toscalix has quit IRC18:02
*** toscalix has joined #mer18:02
*** jonnor_work has quit IRC18:04
*** Free-MG has joined #mer18:07
lbtBeen quiet - real world got in the way this afternoon - and tango tonight. I'll be around tomorrow morning though.18:08
Stskeepshave to enjoy weekends once in a while18:10
* Stskeeps doesn't always work either18:10
lbtDenise spotted a "1920s market" tomorrow - so we're going to look for Art Deco stuff18:11
Alison_ChaikenI'm going on a long bike ride in the morning, and then in the afternoon I'll try to stay awak so I can pretend to work.18:14
lbt:)18:14
bigbluehatlbt: market sounds nice :) where are you at?18:15
lbtReading, UK - it's at the hexagon18:15
Stskeepsbigbluehat: so, the idea i'm having is that if you really want to have a world like in the A day made of glass videos, you have to have frameworks and methods that enable people to quickly mock up and do these kind of environments18:15
bigbluehatlbt: nice…I'll check it out next time I'm across the pond ;)18:16
Stskeepswithout terribly much need for learning platform specifics18:16
lbtbigbluehat: it's a one-off, probably not worth a zero-notice flight :)18:16
bigbluehatlbt: k.18:16
* bigbluehat closes united.com18:17
bigbluehat;)18:17
bigbluehatStskeeps: good plan18:17
lbthehe18:17
Stskeepsbigbluehat: one step on the way is dragging designers and coders closer to eachother and collaborate (i'm from a research center in my uni studies that was big on cross-faculty collaboration)18:17
Alison_ChaikenWhen I worked for HP Labs, we had a video made in Bristol lab that was much like "world made of glass."18:17
Alison_ChaikenAnd when I worked for Nokia, the San Diego folks who worked on CDMA had made one too, although I don't know if it was ever released.18:17
lbtAlison_Chaiken: ah, I've had interviews there...18:17
Stskeepswhere we had architects, graphics designers, coders, antropologists working together towards solutions for healthcare18:17
Alison_ChaikenLots of things were never released from Nokia . . .18:17
bigbluehatStskeeps: I'm both designer and dev, so I understand the need to keep the brain's hemispheres on the same page :)18:18
bigbluehatfollowing so far18:18
StskeepsAlison_Chaiken: http://www.youtube.com/watch?v=A4pDf7m2UPE was one18:18
Alison_Chaikenlbt, I had planned to relocate to Bristol to work with Anthony Sowden's team on ubiquitous computing, but Hugh Whatsit whole lab got shutdown at the moment I was about to accept the transfer.18:18
Alison_ChaikenStskeeps, the video from SD was different: it was the cliched "day in a typical life in the future," assuming that you were a very wealthy white man in a Western country.18:20
Stskeeps:nod:18:20
Stskeepsbigbluehat: and one idea is simply that of being able to have a 'blank' mer device and quite easily deploy ideas and concepts to it18:21
Stskeepsand have guides to do that18:21
lbtAlison_Chaiken: name rings a bell18:21
Stskeepsbigbluehat: so, that's where things like photoshop to qml fits in18:22
Alison_ChaikenAnthony Sowden led HP's Nova e-book project, which was impressive for its time.   The technical leads were Keir Shepherd, now at the Beeb, and Ben Wynne, whom I've lost track of.   Those guys were ace!   And also Lon Barfield, a smart UI guy from academia.18:23
Alison_ChaikenOh well, so many bigcos, so many cancelled projects.18:23
lbtheh18:23
* Stskeeps tries to recall a research project he heard about some time back..18:24
Stskeepsah, there: http://oxygen.csail.mit.edu/18:27
Stskeepsthey used to have code available18:27
Alison_Chaikencsail!   The former home (as in sleeping quarters) of RMS.   Maybe still is for all I know.18:28
Alison_ChaikenI'll be back at MIT in a couple of weeks for a visit.18:28
djszapibigbluehat: cool18:32
*** phaeron has joined #mer18:33
bigbluehatStskeeps: k. back…and read up :)18:33
* bigbluehat changed a diaper somewhere in there :)18:33
bigbluehatStskeeps: if I'm tracking all this, you're looking for a super easy way to get designer/devs into a "production" (or at least prototype) stage with Mer apps18:34
bigbluehatcorrect?18:34
*** tarantism has joined #mer18:35
djszapibigbluehat: I might ask for help on Mac, if you do not mind.18:35
Stskeepsyeah18:35
djszapipardon me, if you mind.18:35
Stskeepsbigbluehat: not so much app but involving them easier in the process18:35
Stskeepsbigbluehat: so it's darn easy to for example do digital signage18:35
Stskeepsi feel sick to my stomach when seeing they're using full-grade windows xp machines in the metro here to run info banners :)18:36
*** npm_ has joined #mer18:38
*** npm has quit IRC18:38
bigbluehatdjszapi: I'lld o what I can but I'm not up to speed on Mer dev on a Mac18:41
bigbluehatI can get you the hookup on building #couchbase and #couchdb on it tho ;)18:42
bigbluehatStskeeps: yeah. detecting the OS on those signage installs are usually findable by the error messages displayed ;)18:42
bigbluehatStskeeps: so you're driving toward what the next "first step" toward simple designer-as-app-builder would be?18:43
bigbluehatPhotoshop 2 QML being a good obvious first step18:44
Stskeepsyeah, there is already an export script but it probably has deficiences18:44
Stskeepslike matrixx said the gimp one had18:44
Stskeepsie, it was png exports18:44
bigbluehatah, gotcha18:45
bigbluehatso what about starting with a simple "stencil" based approach?18:45
bigbluehator something like Ares (a webOS thing...again)18:46
bigbluehatAres 2 is supposed to be open sourced by summer, iirc18:46
Stskeepslooks like what qt creator can do with component18:46
Stskeepss18:46
* Stskeeps ponders how well the integration is18:47
bigbluehatyeah, I think I get what you're after :)18:48
bigbluehathow can I help?18:48
Stskeepswell, you're already helping18:48
Stskeepsi don't get to tap into designer brains every day ;)18:48
bigbluehatam I?18:48
bigbluehatah. :)18:48
* bigbluehat plugs his in via USB 3.018:48
bigbluehat;)18:48
Stskeepshttp://www.youtube.com/watch?v=G5Q4mrq2hl8 is qt designer with symbian components18:48
Stskeepsthat's something like ares?18:54
bigbluehatquite similar18:55
bigbluehatthese things are sadly a dime a dozen18:55
bigbluehatand none(?) of them interop18:55
Stskeepsyeah18:56
bigbluehatno reason they couldn't tho :)18:56
bigbluehathttp://ares.palm.com/Ares/about.html18:56
bigbluehatlook familiar? :)18:56
*** raignarok has quit IRC18:57
*** raignarok has joined #mer18:57
*** Free-MG has quit IRC18:57
bigbluehatit generates enyo, iirc18:57
bigbluehator Ares 2 will18:57
Stskeeps:nod:18:57
* Stskeeps looks at his $HOME18:59
Stskeepstime to clean up that beast..18:59
*** KaiRo_Mozilla has quit IRC19:01
bigbluehatStskeeps: rm -rf .19:02
bigbluehat;)19:02
*** Eismann has quit IRC19:03
*** notmart has quit IRC19:07
*** djszapi has left #mer19:07
*** jukkaeklund has joined #mer19:10
*** raignarok has quit IRC19:14
*** raignarok has joined #mer19:14
*** ALoGeNo has quit IRC19:16
*** jukkaeklund has quit IRC19:16
*** ALoGeNo has joined #mer19:17
*** ALoGeNo has joined #mer19:17
*** raignarok has quit IRC19:19
*** trbs has joined #mer19:33
*** bigbluehat has quit IRC19:47
*** mdavey has quit IRC19:53
*** Alison_Chaiken has quit IRC19:57
*** tarantism has quit IRC19:59
*** tarantism has joined #mer20:04
*** bigbluehat has joined #mer20:09
*** vivijim has quit IRC20:27
*** vivijim has joined #mer20:28
*** vivijim has joined #mer20:28
*** vivijim has joined #mer20:28
*** vivijim has joined #mer20:28
*** vivijim has joined #mer20:28
*** vivijim has joined #mer20:29
*** vivijim has joined #mer20:29
*** vivijim has joined #mer20:29
*** vivijim has joined #mer20:29
*** vivijim has joined #mer20:30
*** vivijim has joined #mer20:30
*** phdeswer has joined #mer20:31
*** harbaum has joined #mer20:58
*** norayr has joined #mer20:59
*** jstaniek has joined #mer20:59
*** swerden has joined #mer21:00
*** raignarok has joined #mer21:01
*** lizardo has quit IRC21:03
*** swer has quit IRC21:04
*** phdeswer has quit IRC21:04
*** ssirkia has left #mer21:07
*** jluisn has quit IRC21:15
*** harbaum has quit IRC21:24
*** NIN102 has joined #mer21:24
*** NIN101 has quit IRC21:26
*** harbaum has joined #mer21:29
*** M4rtinK has joined #mer21:29
*** pohly has quit IRC21:34
*** harbaum has quit IRC21:40
*** harbaum has joined #mer21:42
*** harbaum has quit IRC21:47
*** toscalix has quit IRC21:50
*** talavis has quit IRC21:56
*** talavis_ has joined #mer21:56
*** raignarok has quit IRC22:08
*** mlfoster has quit IRC22:16
*** bigbluehat has quit IRC22:20
*** norayr has quit IRC22:27
*** mlfoster has joined #mer22:30
*** phaeron has quit IRC22:31
*** mlfoster has quit IRC22:35
*** mlfoster has joined #mer22:36
*** Eren has joined #mer22:48
*** mlfoster has quit IRC22:50
*** NIN102 has quit IRC23:02
*** tpn has quit IRC23:08
*** buckle has quit IRC23:28
*** buckle has joined #mer23:31
*** berndhs has joined #mer23:31
*** mlfoster has joined #mer23:34
*** Eren has quit IRC23:36
*** mlfoster has quit IRC23:36
*** kavurt has joined #mer23:52
*** gimli has quit IRC23:57
*** berndhs has quit IRC23:59
*** berndhs has joined #mer23:59

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!