#qi-hardware IRC log for Monday, 2010-07-12

xiangfuunclouded: Hi00:15
uncloudedxiangfu: hi00:15
xiangfuI search a little bit. the sys_imageblit will call : color_imageblit which is at :    * drivers/video/cfbimgblt.c, line 7500:16
xiangfu    * drivers/video/sysimgblt.c, line 5200:16
xiangfuunclouded: from those two file's comment: it says: Generic BitBLT function for frame buffer with packed pixels of any depth.00:16
xiangfuunclouded: another is:   2  *  Generic 1-bit or 8-bit source to 1-32 bit destination expansion00:17
xiangfu  3  *  for frame buffer located in system RAM with packed pixels of any depth.00:17
xiangfuunclouded: so I guess . it support any depth. :)00:17
unclouded1-bit or 8-bit source doesn't sound so hot though.00:18
uncloudedI've been having trouble using fb_imageblit to draw a 1x1 image00:19
xiangfuunclouded: oh. that make me confuse. it said "any depth", what is the different "1-bit or 8-bit source" and "depth"00:19
xiangfuunclouded: 256-color mean 8 bit depth ? right?00:21
uncloudedThis was me trying to draw a single pixel in the top-left hand corner:00:21
unclouded  u8  pixel[] = { 0xff, 0xff, 0xff, 0xff};00:21
unclouded  struct fb_image  im = { dx:0, dy:0, width:1, height:1, depth:32, data:pixel};00:21
unclouded  info->fbops->fb_imageblit( info, &im);00:21
unclouded( from the kernel)00:21
uncloudedso I'm trying to draw a 24-bit image using fb_imageblit ( which I think is a pointer to sys_imageblit)00:23
uncloudedRE: packed pixels of any depth: I think that refers to the framebuffer, not the source image00:24
uncloudedbut I'm new to the kernel so I might have missed something00:25
xiangfuunclouded: oh. me too. :) need learn more about IMAGE and FB.00:26
uncloudedlooking at sys_imageblit, it looks like it will use fast_imageblit or slow_imageblit for 1-bit ( mono) images and color_imageblit for "color" images, I wonder if color_imageblit only support 8-bit ( 256 colour paletted) images, which would explain why my example doesn't work00:30
xiangfuunclouded: Hi. when you try to run your program. it is total not display or color not correct?00:44
uncloudedboth.  In the "one pixel" example, nothing is displayed.  When I try a 4x8 image with depth:24 I get mostly black but some random colors00:45
xiangfuunclouded:  oh. check this issur: http://projects.qi-hardware.com/index.php/p/ben-nanonote/issues/87/00:51
xiangfuunclouded: Rikard have send a patch to mailing list. and I applied it today. fix the color palette.00:52
xiangfuunclouded: so we better try the new kernel image.00:52
uncloudedI don't think that's related.  text is all monochrome ( 1-bit) images00:55
xiangfuunclouded: ok.00:56
uncloudedxiangfu: I might be able to do what I want with depth:8 and a 256-color palette so I'll try that instead and see if that works for me01:06
xiangfuunclouded: sure.01:11
qi-commitsBas Wijnen: fix build process, thanks to Xiangfu Liu <xiangfu@sharism.cc> http://qi-hw.com/p/iris/662df3803:02
qi-commitsBas Wijnen: fix things http://qi-hw.com/p/iris/2122a7403:02
qi-commitsBas Wijnen: more fixes by Xiangfu Liu <xiangfu@sharism.cc> http://qi-hw.com/p/iris/a2add9a09:06
rafakristianpaul: thanks to pastebin that error10:00
rafakristianpaul: let me fix that10:01
kristianpaulrafa: tks10:08
kristianpaulrafa: btw is too hard have pyqt4 on jlime?10:09
kristianpaulor what do you recomend to do devel of grapgical apps that works both on fb and X10:10
luoHi everybody, gbraad introduced me here, I'd like to do something for Qi-hardware. And currently I'm do an open project OpenOMAP: http://code.google.com/p/openomap/10:11
rafakristianpaul: sdl10:11
rafakristianpaul: but if pyqt4 is in OE and it faild to build for jlime repo ( I mean, if it is not in jlime repo) i will check the problem10:12
rafaand if it is on OE I will build it after to fix the building problems10:12
rafaI also I will upload the packages missing for froto10:13
kristianpaulthanks :)10:13
rafakristianpaul: about xoscope: it is something of xoscope10:13
rafaperhaps it needs some special config or file that it does not find?10:13
kristianpaulwich line of code..?10:13
kristianpaulah ok10:14
kristianpauli'll hceck10:14
rafakristianpaul: for xoscope?10:14
rafakristianpaul: try to install strace and run xoscope10:14
rafastrace -f -o /tmp/log xoscope10:14
rafathen read /tmp/log to realize which is the problem10:14
kristianpaulrafa: how i can takea  screen shot of an app in Jlime?10:24
zearkristianpaul, fbgrab10:30
zearthough i don't know if it's in the repo10:30
kristianpaulzear: even for Jlime X?10:31
zearkristianpaul, yes, it just makes a screenshot of what's running on the framebuffer10:31
zearin jlime case it's X1110:31
kristianpauland it will work same, i understand..?10:32
kristianpauloh great10:34
zearthough if you don't want to have a terminal with fbgrab on screen, launch it via telnet/ssh rather than directly on the nn10:34
xiangfuluo: Hi do you use KiCAD to design hardware?10:54
luoSorry, no10:56
luoI'm using ORCAD and PADS10:56
xiangfuluo: just curious. :)11:28
luoxiangfu: it's ok, I'll try KiCAD11:30
mthkristianpaul: PySide might be a more future proof alternative to PyQt12:06
kristianpaulmth: oh, well i just wan try run a PyQT based app on the nano, no more12:07
mthah, I thought you wanted to develop new apps12:08
kristianpaulnah i don like QT :P12:08
wolfspraulmth: why is PySide more future proof than PyQt?12:09
kristianpauli just dont write rewrite stuff  now12:09
rafakristianpaul: X, so I guess that there are plenty of X screenshots tools around :)12:09
mthwolfspraul: more developers working on it, as far as I can see12:09
mthand it will be in the SDK for MeeGo devices, so more exposure than PyQt which always has to be downloaded separately12:10
wolfspraulhah, just read a bit12:12
wolfspraulseems the main reason to start PySide was the GPL license of PyQt - whereas PySide is LGPL licensed12:12
wolfspraulcannot believe that that is still the reason behind such big moves...12:13
rafawolfspraul: where will the screenshots go (in qi wiki)? I have read a couple of mails about people uploading screenshots but I do not find those12:13
wolfspraulrafa: for now I just dump them12:13
wolfspraulthen they need to be categorized (along with hundreds of other uncategorized uploaded files)12:14
wolfspraulthen maybe the Applications page?12:14
wolfspraulI think this could evolve into an illustrated app catalogue12:14
wolfspraulit's quite openwrt-centric though12:15
wolfspraulbut feel free to throw in screenshots, any screenshot is cool12:15
rafawolfspraul: ah.. okey :)12:15
kristianpauloh nice dump about electrician/startdic :)12:15
wolfspraulrafa: how did it go with that professor you gave SAKC to?14:10
rafawolfspraul: hey.. I traveled and it was great. I had a short meeting with him, so we was talking for a while. He knows the sakc a bit and he said me that there are some couple of researches ideas to do. He is going to contact to carlos, or he already did perhaps14:12
rafawolfspraul: I gave this professor, last friday, the SAKC, and he said me that he will take care of it with a proper academic work. SO I think that he is starting to move right now, since that sakc is a great stuff  ;)14:13
rafawolfspraul: I just arrived to BA again14:14
rafaHe knows the sakc a bit better than me*14:16
wolfspraulok all sounds good I think14:28
wolfspraullet's see whether he moves :-)14:28
rafayeah, and I know where he lives :)14:29
larscwejp: openwrt now has the lastest libflac15:57
wejpnice :)15:58
Ornotermesnice indeed :)16:47
qi-commitsLars-Peter Clausen: MMC: jz4740: drop max_clock field http://qi-hw.com/p/qi-kernel/9a7ae7116:58
qi-commitsLars-Peter Clausen: MMC: jz4740: Add some style cleanups http://qi-hw.com/p/qi-kernel/e0afb2516:58
qi-commitsLars-Peter Clausen: MMC: jz4740: Rework state handling http://qi-hw.com/p/qi-kernel/ef5286916:58
urandom_this http://hnb.sourceforge.net might be usefull on the ben17:14
ezdagorAnyone know where I can get a kernel.bin (Non-JLime) that has working sound and the ks7010 compiled in that works with the Debian-Lenny.ubi? I've tried compiling the source.. but int he "mkimage" phase of it, it will not boot.18:32
larscwhat kind of error do you get?18:33
ezdagorIt says it's booting the kernel, then the LCD gets real bright, and then goes into a continuous boot cycle.18:34
calamarzezdagor: these notes might be useful if you haven't spotted them already --> http://en.qi-hardware.com/wiki/Debian/Kernel18:34
ezdagorAwesome. Thanks.18:35
calamarzezdagor: here i have a kernel  (I could upload modules too) --> https://rhizomatik.net/nanonote/kernel/deb-kern-mouse-uinput/18:37
calamarzhm... no, but I disabled ks7010 :/18:38
ezdagorWhich config file does the kernel.bin use?18:40
ezdagornn-kconfig-generic, nn-kconfig-xburst or nn-openwrt-kern?18:44
mthlarsc: does set_timer_slack exist in 2.6.34?21:13
mthI can't find any header that defines it in the likely locations21:14
ezdagorOk... I have the cross compile kit... where would be the proper place to put it?21:20
ezdagorI remember there being a HOWTO.. but I lost the URL.21:20
larscmth: no. did i commit it?21:24
mthyes, it was added on qi-kernel/jz-2.6.3421:24
mthlast delta of this commit: http://projects.qi-hardware.com/index.php/p/qi-kernel/source/commit/ef52869/21:24
xiangfuezdagor: I would like /opt or ~/21:24
qi-commitsLars-Peter Clausen: MMC: jz4740: fix merge error http://qi-hw.com/p/qi-kernel/d2d5cee21:26
qi-commitsLars-Peter Clausen: jz4740: qi_lb60: Update mmc include file location http://qi-hw.com/p/qi-kernel/90d2c2921:26
qi-commitsLars-Peter Clausen: MMC: jz4740: Move header file to arch/mips/include/... http://qi-hw.com/p/qi-kernel/e59921b21:26
mththanks, that fixes it21:36
mthI didn't find any other issues in a quick test21:36
ezdagorHmm.. when doing a "make ARCH=mips menuconfig" the NN's processor isn't in the list.22:14
ezdagor.config - Linux Kernel v2.6.32.10 Configuration22:14
ezdagorI first did a "make kernel_menuconfig" from OpenWRT build system, it downloaded the kernel, did some shell scripting, then gave an error.22:15
ezdagornanodev@wolf:~/dev/openwrt-xburst-xburst> make kernel_menuconfig make[1]: Entering directory `/home/nanodev/dev/openwrt-xburst-xburst/target/linux' make[2]: Entering directory `/home/nanodev/dev/openwrt-xburst-xburst/target/linux/xburst' if [ -s "/home/nanodev/dev/openwrt-xburst-xburst/build_dir/linux-xburst_qi_lb60/linux-" ]; then (cd "/home/nanodev/dev/openwrt-xburst-xburst/build_dir/linux-xburst_qi_lb22:16
ezdagorrc=- push -a; else quilt --quiltrc=- top >/dev/null 2>&1; fi ); fi make[2]: *** [/home/nanodev/dev/openwrt-xburst-xburst/build_dir/linux-xburst_qi_lb60/linux-] Error 127 make[2]: Leaving directory `/home/nanodev/dev/openwrt-xburst-xburst/target/linux/xburst' make[1]: *** [menuconfig] Error 2 make[1]: Leaving directory `/home/nanodev/dev/openwrt-xburst-xburst/target/linux' make: *** [kernel_menuconfig] Er22:17
ezdagorI also added the patch from http://sdschulze.dnsalias.org/~soeren/content/debian-nanonote-kernel.txt22:22
ezdagorCould that be the problem?22:22
xiangfuezdagor: what the "git branch" output? and the first line of "git log"?23:46
ezdagorgit branch?23:49
xiangfuezdagor: yes.23:53
xiangfuezdagor: and the first line of git log23:53
ezdagorYou mean version of git?23:54
ezdagorBranch of what? The kernel?23:54
xiangfuezdagor: I want make sure you are the the [xburst] branch of git repos.23:55
ezdagorHold on a sec..23:56
ezdagorI have absolutely no idea... I cannot remember.. :/23:57
ezdagorwolf:~/nano/kernel/1 # l total 366184 drwxr-xr-x  7 root root      4096 Jul 10 21:14 ./ drwxr-xr-x  3 root root      4096 Jul 12 12:27 ../ drwxr-xr-x 10 root root      4096 Jul 10 15:00 openwrt-xburst-xburst/ -rw-r--r--  1 root root  12611443 Jul 10 14:52 openwrt-xburst-xburst.zip drwxr-xr-x 24 root root      4096 Jul 10 17:14 qi-kernel- drwxr-xr-x 24 root root      4096 Jul 11 20:38 qi-kernel-5396a9238205f20f811ea57823:58
ezdagorIf that helps..23:58
--- Tue Jul 13 201000:00

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