#qi-hardware IRC log for Wednesday, 2011-03-23

tuxbrainxiangfu kyak, any progress on the building release image issue?00:14
tuxbrainI would love to test this http://en.qi-hardware.com/wiki/File:Nanokoder_draft_demo.ogv on NN :)00:14
xiangfutuxbrain: (libgobject-2.0.so) let me check my build log now.00:19
xiangfutuxbrain: BTW, if you only want build package, you can try using the SDK: http://downloads.qi-hardware.com/software/images/Ben_NanoNote_2GB_NAND/latest/OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1.tar.bz200:21
tuxbrainxiangfu : thanks! I will try to buil with sdk tomorrow :) ,00:24
qi-bot[commit] Werner Almesberger: renamed gerber/ to fab/ http://qi-hw.com/p/eda-tools/0f2073900:24
qi-bot[commit] Werner Almesberger: bom/Makefile: added generation of an SMT fab-type BOM http://qi-hw.com/p/ben-wpan/377bbe900:24
qi-bot[commit] Werner Almesberger: makefiles/Makefile.kicad: changed eda-tools/gerber/ to eda-tools/fab/ http://qi-hw.com/p/ben-wpan/ee4c70b00:24
tuxbrainwow >600Mb SDK00:25
tuxbrainis the full openwrt libs/sources  there?00:26
tuxbrainbtw to all any comments about nanocoder will be wellcome, maybe ignored, but welcome anyway00:27
tuxbraingood night dudes00:27
wpwrakSDK = Sizable Development Kit00:27
tuxbrainsorry, nanocoder->nanokoder00:28
xiangfutuxbrain: this SDK include all build depends of release.00:34
tuxbrainonly of the release... wow :)00:34
xiangfu(in fact I am always using the SDK for compile package now. build images in buildhost, build release need a loooot of time in my notebook :)00:40
wpwrakxiangfu: oh ... your "workstation" is a notebook ? so that's why it's so incredibly slow !00:41
xiangfutuxbrain: (/usr/lib/libgobject-2.0.so: could not read symbols: File in wrong format) for debug this problem, I think I need full build log of pango and the pango 'config.log'00:41
xiangfuwpwrak: yes. a notebook. with an old CPU. core 2.0GHz. btw I update my memory to 4GB early this year  :)00:42
wpwrakxiangfu: more memory is always good :)00:43
xiangfutuxbrain: make package/pango/{clean,compile} V=99 > pango.log 2>&1  and the build_dir/target-mipsel_uClibc-0.9.30.1/pango-1.26.0/config.log00:43
xiangfutuxbrain: I have to apply those patches for make openwrt compile fine in buildhost: http://downloads.qi-hardware.com/software/images/NanoNote/Ben/2011-02-23/patches/00:46
Action: xiangfu updating wiki page00:46
xiangfutuxbrain: this time I will write the wiki page. at the same time test those instruction on my PC :)00:55
xiangfujust meet the first error.00:55
qi-bot[commit] Werner Almesberger: fab/pos2fab: condition a KiCAD .pos file for SMT fab use http://qi-hw.com/p/eda-tools/5cb1aed01:04
qi-bot[commit] Werner Almesberger: Makefile.kicad: added generation of a positioning for the SMT fab http://qi-hw.com/p/ben-wpan/888546701:05
wpwrakgrmbl.                                                                    ^---- file01:05
kristianpaulwpwrak: borg yes :p01:24
kristianpaulviric: you there?02:25
kristianpaulthis line [fid, message] = fopen(settings.fileName, 'rb');02:25
kristianpaulhmm and i have others  settings.somthing in other codes..02:27
xiangfukristianpaul: I want know more about how RTEMS using the NOR flash in MM1.02:30
xiangfukristianpaul: please give me some tips :)02:30
kristianpaulxiangfu: hi02:31
kristianpaulxiangfu: what you need to know? if i know i can tell if not point it where find the info02:32
kristianpaulrtems is linked to the yaffs2 port by kernel02:33
kristianpaulonce that is done, i think all the rest is just like using the standard lib02:33
kristianpaulfor example in flicernise code i'm aware all flicernoise settins are loaded from a file called sysconfig.bin02:37
kristianpaul(line 138 syscongfig.c)02:37
kristianpauldoing a fast ack-grep i found fd = fopen(SYSCONFIG_FILE, "w");02:38
kristianpaulso if you follo ansi c, rtems will be just another friendly os :-)02:38
kristianpaul posix too02:38
kristianpaulyou should do a quick read or take a look at http://www.rtems.org/onlinedocs/releases/rtemsdocs-4.10.0/share/rtems/html/02:39
kristianpaulwas the first thing i did when seriosully start looking at rtems02:39
kristianpaulyou can do easy tasks like:02:39
kristianpaulmaking a driver for thre second uart port i added to mm1 core ;-)02:40
kristianpaulporting a util from busybox like ping02:40
kristianpaul(of course leaning tasks)02:40
kristianpaulleraning*02:40
kristianpaulalso checkout examples-v2 from  http://www.rtems.com/wiki/index.php/RTEMS_CVS_Repository02:41
kristianpaulbut well, i lost the track with the NOR question02:41
xiangfuI will reading today.02:42
xiangfuthen see if I can ask some real question :)02:42
xiangfuthanks kristianpaul02:42
kristianpaulso basicaly flicernoise NOT rtems is linked with yaffs2 and lekernel also added a custom command to the rtemns shell in order to be able to do formating (line 227 main.c flickernoise source code)02:47
kristianpaulflickernoise**02:47
kristianpaulsrc/shelltextx.c02:48
xiangfuthe line 227 of my main.c is only one "}". my git last commit is : 79f88b0091c79b5ce2e70b4a70c15061d4bbc3f302:52
xiangfuMon Mar 14 12:54:23 2011 +010002:52
xiangfukristianpaul: I want build a wallpaper inside the flickernoise. like a default wall paper.02:52
xiangfukristianpaul: and I always want start learn at FLASH. is more simply then other device. :)02:53
kristianpaulxiangfu: line 241 sorry02:53
xiangfukristianpaul: this shelltextx.c is for reflash mm1. right?02:53
kristianpaulreflash NOR (wich is part of mm1) yes02:54
kristianpaulfrom _rtems_ shell02:54
rjeffrieskristianpauk is rtems the OS used my milkymist?02:54
rjeffriess/my/by02:54
kristianpaulrjeffries: Yes Sr02:54
rjeffriesinteresting02:55
kristianpaula lot :-)02:55
kristianpaulxiangfu: wallpaper.png, just do some gimp and thats it ! :-)02:56
xiangfukristianpaul: this shelltextx.c is only 'erase' for now.02:56
kristianpaulxiangfu: is harcoded in the code02:56
rjeffriesis linux launched as a task on top of rtems ? or instead of rtems02:56
kristianpaulxiangfu: fow now ;-)02:56
kristianpaulrjeffries: linux is not part of rtems or flickernoise02:56
kristianpaulalso is a bit complex to be a "task" :-)02:57
rjeffriesi thought someone said mm could boot linix so i guess that was bogus02:57
xiangfu:D02:57
kristianpaulrjeffries: it could02:57
rjeffriesis flickernoise an app on top of rtems02:57
kristianpaulrjeffries: but not related with rtems, is another way02:57
rjeffriesok02:57
kristianpaulrjeffries: lets said yes,02:57
kristianpaulfor me is not clear tha app concept in rtems02:58
kristianpauli could said flicernoise is a compilation of rtems + mtk + libs + patches02:58
kristianpaulmtl is the gui toolkit for the graphics02:58
kristianpaulmtk**02:58
Action: rjeffries nods understood02:59
kristianpaulgood :-)02:59
xiangfukristianpaul: now I know how to add a command in rtems. base on shellext.c :)03:00
kristianpaulxiangfu: yay :-)03:00
kristianpaulis prety easy, isnt? :_)03:00
kristianpaulbut also shell.pdf have quick explanation also03:01
xiangfukristianpaul: I will try to hardcode a wallpaper to flickernoise first.03:04
kristianpaulxiangfu: with other name you mean?03:05
kristianpaulwill be nice if you can choose at least from two harcoded wallpapers ;-)03:05
xiangfukristianpaul: from the "http://www.milkymist.org/wiki/index.php?title=Flashing_the_Milkymist_One#Flash_Memory_Distribution"03:07
xiangfuonly the REGULAR APP    is yaffs2 right?03:07
kristianpaulhmm03:08
xiangfuthe flickernoise.fbi is flashed to REGULAR_APP.03:08
kristianpauli thik after, i dont remener bios can boot the regular app from a yaffs2 "parition"03:08
kristianpaulafter regular app i mean03:08
kristianpaulwich have a limit btw03:08
kristianpaulxiangfu: the wiki is not clear but after regular follow the yaffs2 partition03:10
kristianpaul*I think*03:10
kristianpaulso there is a limit on the refular app size03:11
xiangfuok.03:11
kristianpaulthats why lekernel impleneted the lzma i think03:11
kristianpaulalso !!!03:11
kristianpaulthe flash NOR is not intednded as a data storage FS03:11
xiangfuso. if you right. the REGULAR_APP should be only 4MB. then the rest is for yaffs2 partition.03:11
wolfspraulI think the partition sizes are arbitrary. there is a large (24mb?) data partition right now.03:12
wolfspraulwe can change this, if necessary03:12
wolfspraultotal we have 32 mb in flash, right?03:12
kristianpaulwe can change this, if necessary <- yes03:12
wolfspraulof course changing will cause trouble in all sorts of tools etc. so it needs to be thought through, and communicated well. not right now.03:12
kristianpauljust change the bios03:12
kristianpauli dunno id erase command is harcoded and also rtems03:13
kristianpauland so on..03:13
kristianpaul;)03:13
wolfsprauloh sure, I can imagine those partitions are hardcoded in several places03:13
kristianpaulbut my point is03:13
wolfspraullet's be very careful about changing this. I'm just making the point that the flash is flat.03:13
kristianpaulyou can also boot from a external memory03:13
kristianpaulaka 8:10 here03:13
kristianpaulbut03:13
kristianpaulsoftware side acording to lekernel need some polish03:14
kristianpaulas NOT all memory cards03:14
kristianpaulworks03:14
kristianpaulbut in theory you can try boot from a memory card03:14
kristianpaulfsboot command in milkmist bios03:14
kristianpauli just have kingstone and cant boot or even mount it..03:15
kristianpaulbut i think you see the plan :-)03:15
kristianpaulwolfspraul: should be flat until you hit the yaffs part and bigh ouch  for some data saved there :-)03:16
kristianpauls/bugh/big03:16
Action: kristianpaul strugling with octave03:16
xiangfuthe 'reset' command not working. instead 'shutdown' make MM1 reboot :)03:21
kristianpaulreally?03:21
kristianpaulah rtems03:21
kristianpaulyes03:21
kristianpaulas Fallenou about that03:21
kristianpaulbtw write mails to rtmes list is good03:21
xiangfukristianpaul: what is the detail? I don't know03:21
kristianpauland also they are very helfull03:22
kristianpaulxiangfu: me eiher i dont do the port03:22
kristianpaulyou can track it ;_)03:22
xiangfukristianpaul: for now. I know the flickernoise is running memory. the flash(24M nor flash) is mount to /flash03:23
xiangfulike '/' is mount to a memory filesystem.03:23
xiangfuright?03:23
kristianpaulrtems-milkymist/c/src/lib/libbsp/lm32 <- mainlly code milkymist port/bsp03:23
kristianpaulyeah some part of the NOT is mount, but actually i dont have the details of this03:24
kristianpauls/NOT/NOR03:24
kristianpauldetails to confirm, but it should be like that i guess, (jus some part of the NOR is maped as flash for rtems)03:25
kristianpaulso i shall say, right xiangfu  :-)03:25
xiangfuok.03:25
kristianpaulbut please check the drivers03:25
kristianpaulto confirm03:26
xiangfuyes.03:26
kristianpaulor ask sebastien ;-))03:26
kristianpauldrivers are also not so hard, even i can understand it :-)03:28
kristianpaul#define FLASH_BASE                      (0x80000000)03:29
xiangfukristianpaul: just ask you said:   { .start_address = 0x80920000, .length = 0x0400000 }, \03:29
kristianpaulrtems-milkymist/c/src/lib/libbsp/lm32/milkymist/include/system_conf.h03:30
kristianpaulgood :-)03:30
xiangfuthe REGULAR_APP is only 4M. then the rest is for /dev/flash503:30
xiangfuline 271 of system_conf.h03:30
kristianpaulhe i tought was five ;)03:30
xiangfuI will update wiki.03:30
xiangfuupdate this one: http://www.milkymist.org/wiki/index.php?title=Flashing_the_Milkymist_One#Flash_Memory_Distribution03:30
kristianpaulplease :-)03:31
wpwraktuxbrain: ah, and you should also ask the SMT fab what to do with the USB connector of atusb. this is a through-hole component and i have no idea how this sort of thing is normally mounted. unfortunately, it seems that through-hole is the only common form for this kind of connector. (probably because of mechanical properties)03:44
kristianpaulohh03:45
kristianpaulqtoctave show gplt wich wasnt when runnning octave from console :o03:45
wpwraktuxbrain: specifically, if the smt fab wants a stencil for the solder paste, they should tell you if the usb connector should have solder paste, too (or if it shouldn't have solder paste or if it doesn't matter)03:47
kristianpaulhmm, how i can show cotent of line 145 and column 32 for all files in my directory?04:03
kristianpaulcat -n *.m | grep -a1 14504:05
kristianpaulargg04:05
kristianpaulwhy octave tell me wich file have the error!!04:06
xiangfukristianpaul: find ./ -name "*.c" -exec sed -n '154p'  {} \; | cut -b35-3604:07
kristianpaulwow!04:07
xiangfufind ./ -name "*.c" -exec sed -n '145p'  {} \; | cut -b32-3304:08
xiangfufind ./ -name "*.c" -exec sed -n '145p'  {} \; | cut -b32-32 :D04:08
kristianpaulhmm.. not helping too much ://04:09
kristianpaulbetter i read how debug code in octave..04:09
Action: kristianpaul zzz04:10
kristianpaulgn804:10
xiangfunight04:11
tuxbrainwpwrak: the holes of USB are specified in the gerbers for the PCB maker drills it?06:27
wpwraktuxbrain: they're specified in the drill file (excellon, not gerber)06:30
wpwraktuxbrain: but yes, that's part of the package06:31
wpwraktuxbrain: another thing to check is the point where their pick and place tool would grab the usb connector07:34
qi-bot[commit] Werner Almesberger: fpdoc: new utility to generate a PDF with the footprints used in a project http://qi-hw.com/p/eda-tools/47b53c807:38
qi-bot[commit] Werner Almesberger: added generation of SMT fab file package (in progress) http://qi-hw.com/p/ben-wpan/d53c2c208:03
qi-bot[commit] Werner Almesberger: include overview images in PCB and SMT fab packages http://qi-hw.com/p/ben-wpan/2e5329008:03
qi-bot[commit] Werner Almesberger: makefiles/Makefile.kicad: fix dependency bugs http://qi-hw.com/p/ben-wpan/ef6dd2808:03
qi-bot[commit] Werner Almesberger: added project-specific footprint reference for SMT fab http://qi-hw.com/p/ben-wpan/507e6c308:03
qi-bot[commit] Werner Almesberger: generate SMT fab package http://qi-hw.com/p/ben-wpan/2aed51508:03
wpwraktuxbrain: new goodies in http://downloads.qi-hardware.com/people/werner/wpan/fab/08:04
wpwraktuxbrain: the *110323* files should be what you need08:04
wpwrakxiangfu: and there are a few updates for fped as well :)08:10
rohwpwrak: morning08:17
rohany clue what happened with the shipment?08:17
wpwrakroh: no idea. did anything show up on your side ?08:18
wpwrakroh: maybe the postman wasn't able to find the building of the space agency and it was returned ? the tracking on the .de side looks extremely untrustworthy, so it may never show up in that system (i also looked on DHL, without success)08:21
qi-bot[commit] werner: fped: don't output pad names consisting only of spaces http://qi-hw.com/p/fped/e1f4e0308:22
qi-bot[commit] werner: fped: added option  -1 package  to select a single package to output with -p/-P http://qi-hw.com/p/fped/f44accd08:22
qi-bot[commit] werner: hack: set errno to ENOENT when -1 fails, so that the error isn't too ugly http://qi-hw.com/p/fped/2de385e08:22
wpwrakxiangfu: whee, that was quick ! :) thanks !08:22
xiangfumaking debian package...08:22
rohwpwrak: nope. nothing showed up08:26
xiangfu  CC        y.tab.o08:26
xiangfufpd.y: In function ‘yyparse’:08:26
xiangfufpd.y:678: error: too few arguments to function ‘dump’08:26
xiangfumake: *** [y.tab.o] Error 108:26
rohand ive got all mail from the last year. nothing missing so far08:26
xiangfuwpwrak: ^08:26
wpwrakxiangfu: argh08:26
wpwrakxiangfu: fixed08:28
wpwrakroh: sometimes, mail between .de and .ar can take unusually long. that happens at least for journals. (e.g., c't)08:29
wpwrakroh: not sure if this parcel takes a similar route, though08:29
rohhm08:29
qi-bot[commit] werner: fpd.y: dump() requires a dummy "one" argument here too, oops http://qi-hw.com/p/fped/335ddb608:30
wpwrakroh: not much we can do but wait :-(08:30
tuxbrainwpwrak: two cuestion, the smt vendor tellsme is very important to have the "fiducial"(I don't know if it's called like that in english) marks on the PCB and in the panel08:32
tuxbrainalso if they can change the 22 Pf  que es High Q, Low Loss by the one they have standarized 22 pf 50V 0402 COG  AVX04025A220JAT2A08:32
virickristianpaul: Now I'm here08:32
tuxbrainwe have such fiducials marks?08:32
rohtuxbrain: know a german word? /me doesnt understand what you mean08:34
rohor spanisch?08:34
tuxbrainfiduciales08:34
tuxbrainhttp://en.wikipedia.org/wiki/Fiduciary_marker see section PCB08:35
rohaaaah.. fitting marks? for the picknplace robot to align on?08:35
qi-bot[commit] werner: fped: don't output pad names consisting only of spaces http://qi-hw.com/p/fped/e1f4e0308:35
qi-bot[commit] werner: fped: added option  -1 package  to select a single package to output with -p/-P http://qi-hw.com/p/fped/f44accd08:35
qi-bot[commit] werner: hack: set errno to ENOENT when -1 fails, so that the error isn't too ugly http://qi-hw.com/p/fped/2de385e08:35
qi-bot[commit] werner: fpd.y: dump() requires a dummy "one" argument here too, oops http://qi-hw.com/p/fped/335ddb608:35
qi-bot[commit] Xiangfu Liu: add debian package stuff http://qi-hw.com/p/fped/a3b3e9408:35
qi-bot[commit] Xiangfu Liu: clean up the Build-Depends. http://qi-hw.com/p/fped/efa4fd508:35
qi-bot[commit] Xiangfu Liu: use the new version rules. http://qi-hw.com/p/fped/92b084108:35
qi-bot[commit] Xiangfu Liu: add debian/fped.manpages  for install manpage http://qi-hw.com/p/fped/01039f208:35
qi-bot[commit] Xiangfu Liu: update to svn rev 5982, enable dh_auto_test http://qi-hw.com/p/fped/222ef2608:35
qi-bot[commit] Xiangfu Liu: use usual name for orig tarball top-level directory http://qi-hw.com/p/fped/f5f48c208:35
qi-bot[commit] Xiangfu Liu: update take svn rev: 5983 http://qi-hw.com/p/fped/0e88d0e08:35
qi-bot[commit] Xiangfu Liu: remove the Build-Depends ttf-liberation http://qi-hw.com/p/fped/649c47f08:35
qi-bot[commit] Xiangfu Liu: update to svn rev 5986 http://qi-hw.com/p/fped/769d1e108:35
qi-bot[commit] Xiangfu Liu: override dh_auto_clean, use make spotless instread http://qi-hw.com/p/fped/ae5b87f08:35
qi-bot[commit] Xiangfu Liu: add ghostscript to Build-Depends http://qi-hw.com/p/fped/b639a2908:35
qi-bot[commit] Xiangfu Liu: update the homepage to help webpage http://qi-hw.com/p/fped/5630e9708:35
qi-bot[commit] Xiangfu Liu: update to r5997 http://qi-hw.com/p/fped/74852e108:35
qi-bot[commit] Xiangfu Liu: debian package update to 5999 http://qi-hw.com/p/fped/4f2ca0b08:35
qi-bot[commit] Xiangfu Liu: update to r6005 http://qi-hw.com/p/fped/3fb567b08:35
tuxbrainyeah roh :)08:36
rohyou mean the 'test dummy icons' ;)08:36
wpwraktuxbrain: (fiducial) i think that's a set of crosshairs, for alignment. i've never used that. do they need them per board or per panel ? in the latter case, they would come from the pcb fab08:36
rohnarf. qi-bot is annoying and worthless08:36
tuxbrainon both08:36
tuxbrainroh : I like to see the commits :)08:37
rohtuxbrain: something yes. but the flooding is not helping08:37
wpwraktuxbrain: how many do they need per board, two ? and how wide should the lines be ? 3 mil ?08:37
wpwraktuxbrain: (22 pF) maybe it works, maybe not. the problem is that regular caps don't have well-defined behaviour at such high frequencies08:38
tuxbrainwpwrak: then better use the one you specified, if it doesn't work it will be your fault and I will have some one to blame08:42
wpwrak;-))08:43
viricwpwrak: you either sleep according to CET time, or you take some heavy drugs :)08:43
kyakdid someone say drugs?!08:44
viricYou are always "in the foot of the canon"08:44
wpwrakviric: i take catnaps :)08:44
rohviric: timezones are for users08:44
kyakcat is napping ~18 hours a day08:45
wpwrakkyak: alright, what's colloquially called "catnaps" ;-)08:45
viricwpwrak: so you sleep periods no longer than half an hour?08:45
wpwrakkyak: and with only 18 hours, that's a hyper-active cat :)08:46
tuxbrainviric wpwrak is not real, is a virtual entity generated by the collective mind waves of qi-hw community and an asado08:46
roh .oO(wpwrak is being kept awake by caffeine and women i believe)08:46
kyakwpwrak: insomniac cat :)08:46
wpwrakviric: hmm, maybe 2-6h at a time. depends.08:47
viricwpwrak: do you end up being awake more than 16h every day?08:47
wpwrakroh: naw, have to banish the women from my lair. too distracting.08:47
viricwpwrak: I mean.. is your purpose to be more time awake, or to break the usual phase-locked-loop with the day and night?08:48
wpwrakviric: duh, no idea ;-)08:48
viric:)08:49
wpwrakviric: i just take a nap when i feel i need one. and i wake up when ... well, when i wake up. easy, no ? ;-)08:49
viricso you have no PLL with day and night :)08:49
wpwrakviric: i tend to drift into nighttime when i can. and then i get some sort of jet-lag when the outside world forces me to synchronize with solar time :(08:50
kyakwpwrak: do you consume any food? (i think wpwrak is failing the Turing test ;)08:50
wpwrak;-))08:50
viricaah08:50
wpwrakkyak: just beer :)08:50
wpwraktuxbrain: so .. how many marks ? and how thick shall the lines be ?08:53
tuxbrainasking08:53
wpwraktuxbrain: oh, and can they be obscured by components or do they have to be visible also after smt ? (in this case, the usb connector - there's a nice big empty space underneath it that i'd like use for one such mark)08:56
tuxbrainroh the boards didn't arrive yet? at this rhythmI will be able to send you production ones before the werner's ones arrives :P09:51
virickristianpaul: you are aware of modules like this, right? http://www.cooking-hacks.com/index.php/gps-module-for-arduino.html11:23
tuxbrainany one with qt compiling knowledge?11:29
tuxbrainld is bitching me11:29
tuxbrainhere is more detailed info about my qt fight http://pastebin.com/xr5D6wUY11:39
rohtuxbrain: i fear something like that (shipping time)11:50
tuxbrainwhere is this damn ld loocking at!!!, the gcc_s thing is on -L /home/mutrox/Proyectos/tuxbrain/productos/nanonote/SDK/OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1/staging_dir/target-mipsel_uClibc-0.9.30.1/root-xburst/usr/lib also12:01
virictuxbrain: if you use '-v' on the g++ call, you will get the '-L' passed to ld12:02
viricyou sohuld also paste the command line of the g++ linking call issued12:03
tuxbrainviric http://pastebin.com/nWwCKdzy12:06
viricm'ho miro en acabat de dinra!12:07
viricdinar12:07
tuxbraingracies nen :)12:07
tuxbrainI'm so tired of build trial and error.... :(13:08
tuxbrainand error and error and error...13:09
virictuxbrain: where is libgcc_s.so* ?13:15
viricit looks as if you have it in multiple places13:15
tuxbrain/home/mutrox/Proyectos/tuxbrain/productos/nanonote/SDK/OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1/staging_dir/target-mipsel_uClibc-0.9.30.1/root-xburst/lib13:15
tuxbrain/home/mutrox/Proyectos/tuxbrain/productos/nanonote/SDK/OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1/staging_dir/target-mipsel_uClibc-0.9.30.1/root-xburst/usr/lib13:15
tuxbrain/home/mutrox/Proyectos/tuxbrain/productos/nanonote/SDK/OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1/staging_dir/target-mipsel_uClibc-0.9.30.1/root-xburst/usr/lib/gcc/mipsel-openwrt-linux-uclibc/4.3.313:15
tuxbrain/home/mutrox/Proyectos/tuxbrain/productos/nanonote/SDK/OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1/staging_dir/toolchain-mipsel_gcc-4.3.3+cs_uClibc-0.9.30.1/lib13:15
xiangfu_tuxbrain: Hi13:20
virictuxbrain: hm I don't see it at once.13:20
xiangfu_tuxbrain: I am also try to re-compile the whole openwrt in my PC.13:20
tuxbrainhi xianfu13:21
xiangfu_tuxbrain: how about your pango compile problem?13:21
xiangfu_tuxbrain: for now. I meet some alsa-utils problem. as kyak said every PC give you different errors :)13:21
tuxbrainuff I have changed them for qt compile problems13:21
kyakxiangfu_: there is a patch for alsa-utils problem13:22
xiangfu_kyak: then I am try to collect those patches put them at : http://downloads.qi-hardware.com/software/images/NanoNote/Ben/latest/patches/13:23
xiangfu_kyak: :)13:24
tuxbrainviric I have rid off the the gcc_s problem, now I have a new one13:24
viricahh13:24
virictuxbrain: how did you do for the gcc_s?13:24
tuxbrainI changed the path of libs to /home/mutrox/Proyectos/tuxbrain/productos/nanonote/SDK/OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1/staging_dir/target-mipsel_uClibc-0.9.30.1/root-xburst/usr/lib13:25
tuxbrainbut now http://pastebin.com/WXnLAuEk13:27
kyakxiangfu_: i think these patches are related to release build, not latest build.. some of them are not needed anymore, and some more patches are needed13:27
tuxbrainbut again the exasperate thing is that the libs are in same path!!!13:29
roh   13:34
rohyay. werners package arrived. finally13:34
wpwraktuxbrain: any update on the fiduciary marks ?13:34
tuxbrainroh: yay!13:34
tuxbrainwpwrak: nop13:34
wpwrakroh: heeey hooo !! :)13:34
rohyay.13:37
tuxbraingcc hates me :(13:37
rohgerman post fucked up. didnt meet me on the 16th.. wrote a card. which i got in the mail _today_.. a full week later13:37
rohif i wouldnt have fetched the package today it would have been sent back. *sigh*13:37
wpwrakroh: wow ;-) didn't they think of just leaving a note on the 1st attempt ?13:38
wpwrakroh: nice timing :)13:38
rohwpwrak: the dhl guy doesnt have a doorkey so he cannot get to the mailbox13:38
rohonly the letter guy13:38
wpwrakroh: oh, complicated mailbox access13:39
wpwrakroh: did customs charge anything ?13:39
rohnope13:40
wpwrakroh: excellent13:40
kyakxMff: ping13:44
rohack. did you really pay 30$ for sending it?13:46
wpwrakroh: 30 pesos, a bit less than EUR 6. these things are something like 2-3 days of work, so even at USD 30, the shipping cost would be insignificant ;-)13:54
rohwow is that stuff tiny13:55
tuxbrainvotes for motto for 8:10 thingys: "wow, that stuff is tiny" :)13:57
wpwrak;-))13:58
wpwraktuxbrain: btw, did the smt fab give you a cost estimate at least ?13:59
kristianpaulviric: sure13:59
tuxbrainwpwrak: np14:00
tuxbrainnop14:00
virictuxbrain: post about the gcc errors14:00
virictuxbrain: you have such a long paths that it is almost impossible to understand anything at glance14:00
tuxbrainviric I have already posted (well really not gcc, is ld that really hates me)14:01
viricld does what gcc tells it to do14:01
tuxbrainviric: (paths) sorry dude14:01
virichey don't call me dude :)14:01
tuxbrainperdona nen :)14:02
viric;)14:02
viricaixí m'agrada14:02
virictuxbrain: et cal -rpath-link14:02
wpwraktuxbrain: ah well, there are still a few work hours left in the day. let's hope they'll get back to you so that we can finalize things this week.14:02
tuxbrainwpwrak: hope so14:03
virictuxbrain: when a dynamic library depends on OTHER dynamic libraries, you need to tell ld "-rpath-link"14:03
virictuxbrain: it means "if any library depends on more libraries, find them at this -rpath-link path"14:03
wpwrak(long paths) OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1   wow, and not a single slash in there :)14:03
virictuxbrain: ld looks at -L directories only for "-l" libraries14:04
virictuxbrain: for dependencies of the libraries, it looks at -rpath-link14:04
viric(if I recall well)14:04
rohwpwrak: did you really solder all these gnd layer connecting wires by hand?14:05
tuxbrainrpath-link, for each file reclaimed?14:05
tuxbrain/home/mutrox/Proyectos/tuxbrain/productos/nanonote/SDK/OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1/staging_dir/target-mipsel_uClibc-0.9.30.1/root-xburst/usr/lib: file not recognized: Is a directory14:05
rohbtw.. http://www.youtube.com/watch?v=KZrP3KlmHto&hd=1 deshakers are cool stuff. (no worries.. it cant be rods.. they would be different sized and _hot_ when there14:09
wpwrakroh: (vias) every one of them ...14:13
rohwow14:13
wpwrakroh: the main problem are the large ground areas in atusb. in atben, the vias weren't so much trouble. i just put solder paste on both sides, stick a wire through each hole, and go around with the soldering iron. then check visually and fix whatever doesn't look right. takes maybe 0.5-1 hours.14:16
wpwrakroh: i had a lot of trouble with atusb, a) because the large ground areas suck up so much heat, and b) because i had chemical surface degradation on some of the boards14:17
tuxbrainmipsel-openwrt-linux-uclibc-g++: unrecognized option '-rpath-link' :(14:23
rohi see14:28
wpwraktuxbrain: something like -Wl,-rpath-link=path ?14:29
tuxbrainmipsel-openwrt-linux-uclibc-g++: unrecognized option '-rpath-link=/home/mutrox/Proyectos/tuxbrain/productos/nanonote/SDK/OpenWrt-SDK-xburst-for-Linux-x86_64-gcc-4.3.3+cs_uClibc-0.9.30.1/staging_dir/target-mipsel_uClibc-0.9.30.1/root-xburst/usr/lib'14:31
wpwrakwith -Wl,... ?14:31
tuxbrainok now did'n compain but thows same error like withou rpath-link or rpath14:34
wpwraktuxbrain: if all else fails, run the build with  strace -f -o log  and then search from the end for gcc_s. that may give you a hint where it looked and what it found.14:39
tuxbrainwpwrak, you understand spanish?14:54
kristianpaultuxbrain: he *should*14:56
wpwraktuxbrain: as long as it's not overly poetic ;-)14:57
kristianpaulcomo la roza que derramo lagrimas sobre la nube? ;-)14:57
tuxbrainkristianpaul eso no lo he entendido ni yo14:58
kristianpaultuxbrain: jajaj ;)14:58
tuxbrainwpwrak: I have resend you the smt vendor answer.14:59
tuxbrainif there is something you don't understand let me know14:59
wpwraktuxbrain: easy enough :) okay, i understand now what they mean. i was thinking of a different type of marking.15:04
tuxbrainwpwrak: any more question to fordward?15:04
kristianpaulviric: (sirf gps + arduino) But the idea is dig on the IC part wich you'll amaze all the software involved there, and most people asume it like "hardware"15:07
viricclear15:09
kristianpaul:-)15:09
kristianpaulbtw i as trying some mathlab code in octave15:09
kristianpaulbut15:09
kristianpaulit uses catch for catch errors15:10
kristianpaulit said something like:15:11
kristianpaulnumber of rows must match (0 != 1) near line 145, column 3215:11
kristianpaulbut no tellme in wich file too look..15:11
wpwraktuxbrain: two open issues: 1) what about the stencil ? does the smt fab need one ? if yes, you need to ask the pcb fab if they can make it and how much it would cost.15:11
wpwraktuxbrain: 2) how is the usb connector soldered ? specifically, if we need a stencil, should the usb connector have solder paste or not ?15:12
kristianpaulviric: if you are curios you can grab the code from http://kristianpaul.org/~paul/tmp/GNSS_SDR.tar.gz15:12
viricI'm not curious enough.. sorry15:12
virickristianpaul: I don't know 'catch' in octave15:13
wpwraktuxbrain: regarding 1), they may also have a dispenser that doesn't require a stencil. in this case, the interpretation/application of the solder paste layer would pass from the pcb fab to the smt fab15:13
kristianpaulviric: i also have to modify temporaly a pwelch fuction was not compatible with same values as mathlab..15:13
kristianpaulok viric15:14
GNUtoo|bug20tuxbrain, hi I'm seeing a lot of activity about smt vendor etc... a new board is in preparation?15:24
tuxbrainGNUtoo|bug20: yep the atben and atusb WPAN RF comm in the ben finallly15:25
GNUtoo|bug20ok nice15:25
GNUtoo|bug20btw I saw that:15:25
GNUtoo|bug20let me launch midori15:25
GNUtoo|bug20it takes time on a armv7@500MHz15:25
GNUtoo|bug20http://hackaday.com/2011/02/20/rf-control-from-just-about-any-device/15:25
GNUtoo|bug20I've the requirements15:26
viricyou can run midori in the nanonote?15:26
GNUtoo|bug20I've some gpio15:26
GNUtoo|bug20no15:26
GNUtoo|bug20on bug2015:26
GNUtoo|bug20and I run a linux kernel 2.6.3115:26
viricI don't know any bug2015:26
GNUtoo|bug20what does this rf control permit me to do?15:27
tuxbrainviric search for buglabs15:27
GNUtoo|bug20viric, imagine an arduino running GNU/Linux and java15:27
GNUtoo|bug20tuxbrain, btw the bug 2.0 is usable15:27
GNUtoo|bug20unlike the 1.x15:27
GNUtoo|bug20software is not totally perfect15:27
GNUtoo|bug20but hardware seem way more solid15:27
GNUtoo|bug20I can comment on it if you want15:27
tuxbrainGNUtoo|bug20: still being so pricy?15:28
GNUtoo|bug20it's not copyleft hardware but MIT like btw15:28
GNUtoo|bug20yes15:28
GNUtoo|bug20still not cheap15:28
GNUtoo|bug20since I was the best contributor of their community...15:29
GNUtoo|bug20you imagine what happened....15:29
GNUtoo|bug20*you caan15:29
GNUtoo|bug20*can15:29
GNUtoo|bug20I didn't test the java part yet on the 2.0 but the hardware seem mostly fine15:30
GNUtoo|bug20basically so far I found the following issues15:30
GNUtoo|bug20*I had to recalibrate the screen => software issue15:30
GNUtoo|bug20the screen is so great now, top precision15:31
GNUtoo|bug20but the borders are still a bit problematic15:31
GNUtoo|bug20it's usable tough15:31
GNUtoo|bug20I had to recalibrate to be able to scroll15:31
GNUtoo|bug20the VGA display is not very usable but with an hdmi->vga or hdmi->dvi adapter it15:31
GNUtoo|bug20's ok15:32
GNUtoo|bug20there is a little ffff with sound15:32
GNUtoo|bug20not earable when using a hifi set15:32
GNUtoo|bug20only trough headphones15:32
GNUtoo|bug20also there is some pop at volume change15:32
GNUtoo|bug20that is totally softwaRE15:32
GNUtoo|bug20I know even why15:32
GNUtoo|bug20that's described in pop and click in the kernel documentation in asoc15:33
GNUtoo|bug20basically the incorrect power sequence was applied15:33
GNUtoo|bug20so that's software15:33
GNUtoo|bug20apart that it's so great15:33
tuxbrainGNUtoo|bug20: regarding RF modules we work bitbanging gpios in c so I suppose is only matter to asign to the righ gpios and done15:33
GNUtoo|bug20ok15:33
GNUtoo|bug20the bug1.x was a nightmare harware wise15:34
GNUtoo|bug20the 2.0 is fine15:34
GNUtoo|bug20so if you want I can put you in touch with them15:34
GNUtoo|bug20or do some test for you15:34
GNUtoo|bug20basically it's the dream device for hardware/software hackers15:35
GNUtoo|bug20pre-requisite: knowing how to program in java ,or GNU/Linux, or any other programming language15:35
tuxbrainI was once in touch with them... but we stop comunicating (no aparenting reason, things that happens) I'm fine to retake contact with them15:36
GNUtoo|bug20ok15:36
GNUtoo|bug20maybe they were too busy trying to get the bug 2.0 out15:36
GNUtoo|bug20so I was wondering what you can do with the RF modules15:37
qi-bot[commit] Werner Almesberger: modules/INFO, modules/Makefile, modules/fiducial.fpd, AUTHORS: added fiducial http://qi-hw.com/p/kicad-libs/de2174d15:38
GNUtoo|bug20and is bitbanging done in userspace15:38
GNUtoo|bug20such as using /sys/class/gpio15:38
GNUtoo|bug20what frequency can you use with the rf module from hackday/community news15:38
GNUtoo|bug20hmmm the source code seem to contain a kenrel module source15:40
GNUtoo|bug20can't it be done totally in userapace?15:40
GNUtoo|bug20because else it requires porting15:40
GNUtoo|bug20for each kenrel version15:40
tuxbrainyou should ask mirko the author about this, is quite silence lately here15:42
tuxbrainI must leave15:42
GNUtoo|bug20ok15:42
wpwraktuxbrain: if they're still around, can you check if this looks okay ? http://downloads.qi-hardware.com/people/werner/wpan/tmp/atben-110323-overview.png15:42
GNUtoo|bug20see you later then15:42
GNUtoo|bug20ah nice 80215.415:44
GNUtoo|bug20altough I don't know the compatibility with zigbee15:44
wpwrakGNUtoo|bug20: the RF is ZigBee-compatible. you just need the protocol stack :)15:45
GNUtoo|bug20ok15:46
Action: GNUtoo|bug20 is a bit confused between zigbee and 80215.415:46
GNUtoo|bug2080215.4 is like the MAC layer15:46
GNUtoo|bug20and zigbee is on top of it?15:46
wpwrakyes15:46
GNUtoo|bug20ok15:46
wpwrakor you can put 6LoWPAN on top of 802.15.4. unlike ZigBee, 6LoWPAN is an open standard (IETF)15:47
GNUtoo|bug20yes I know zigbee has some issues with freedom15:47
GNUtoo|bug20for instance can you make a microcontroller talk 6LoWPAN?15:48
GNUtoo|bug20or are the zigbee stuff from sparkfun zigbee-only?15:48
rjeffrieswpwrak have all of your atben/atusb packages now arrived at destinations?15:49
GNUtoo|bug20because I bet their stack are some kind of black box15:49
wpwrakrjeffries: yes. all three have reached their destination.15:50
wpwrakGNUtoo|bug20: 6LoWPAN is designed for low resources, yes15:51
wpwrakGNUtoo|bug20: no idea what sparkfun have15:51
GNUtoo|bug20ok15:51
rjeffrieswpwrak now the race is on to see who confirms a transmission across the ether (outside of Argentina) FIRST! My money is on... roh?15:55
wpwrakmine too :)15:56
wpwraki'm actually a little surprised that he hasn't reported any success yet :)15:57
GNUtoo|bug20ahh I understand16:01
GNUtoo|bug20basically it uses cheap radios like the sparkfun ones @433 Mhz16:01
GNUtoo|bug20and bitbang the protocol16:01
GNUtoo|bug20so all is in softwaree16:01
GNUtoo|bug20so it ccan talks to everything16:02
GNUtoo|bug20that is in 433 Mhz16:02
GNUtoo|bug20right?16:02
GNUtoo|bug20or at least in theory16:02
wpwrakno no ... very different16:03
wpwrakit's in the 2.4 GHz band16:03
wpwrakalso, bitbang is only the communication with the transceiver's buffer. the transceiver takes care of the bit timing.16:03
GNUtoo|bug20ok16:06
GNUtoo|bug20but....16:06
GNUtoo|bug20I was talking not about the 80215.4 one16:07
GNUtoo|bug20but this one:16:07
GNUtoo|bug20http://nanl.de/blog/2011/02/ben-nanonote-able-to-control-radio-power-sockets/16:07
wpwrakah, that one's different, yes. i think the chip also controls the bit timing, though16:09
GNUtoo|bug20ok16:10
wpwraktuxbrain: and the same for atusb: http://downloads.qi-hardware.com/people/werner/wpan/tmp/atusb-110323-overview.png16:13
qi-bot[commit] Werner Almesberger: makefiles/Makefile.kicad: added target "overview" http://qi-hw.com/p/ben-wpan/b93494e16:13
qi-bot[commit] Werner Almesberger: atben: added two fiducial marks http://qi-hw.com/p/ben-wpan/0923d2416:13
qi-bot[commit] Werner Almesberger: atusb/atusb.brd: some more silk screen cleanup http://qi-hw.com/p/ben-wpan/c33572116:13
qi-bot[commit] Werner Almesberger: atusb: added two fiducial marks http://qi-hw.com/p/ben-wpan/7a9e51816:13
wpwrak(having automated the process is starting to pay back. doing all these things manually several times would really suck :)16:14
wpwraktuxbrain: erm, will you buy the components they don't have in stock or will they buy them on your behalf ?16:15
rohrjeffries: wpwrak ? the package is here now16:19
rohor what do you mean?16:19
rjeffriesroh when will you see if it transmit/receives? ;)16:22
rohheh. dunno. i only have a very old nanonote prototype.. no clue what i need to patch to make it work16:23
rohi think it only has half the ram or flash. havent used it a while16:23
xMffkyak: pong17:17
kyakxMff: i had a look at that perl @INC problem, and proposed a patch in #9111. Coudl you have a look?17:19
xMffkyak: already seen it but I avoid openwrt stuff at work altely, too distracting :)17:19
xMffwill check now17:19
kyakthanks :)17:19
xMffafter defining otherlibdirs, the @INC is free of /usr/local stuff?17:23
kyaknope, not really17:23
kyaklet me explain...17:24
kyakthere are siteprefix and vendorprefix that can be defined (together with otherlibdirs, but otherlibdirs just gets appended to the end)17:25
kyakthe problem is, that siteprefix and vendorprefix expand liek this:17:25
kyak/usr/local/lib/perl5/site_perl/$version/$archname17:26
tuxbrainwpwrak: at fisrt glance they will buy and put in the quote, then I will check out if in farnell, Mouser, RSelectronics or in DigiKey would be more convenient.17:26
kyakxMff: in this case we could define siteprefix and get paths in INC like staging_dir../usr/lib/perl5/$version/$archname17:27
xMffkyak: I've seen some discussion about making perl relocatable, it involved passing very long pathes during compilation and then later patch them in the binary17:27
kyakinstead of just /usr/local17:27
kyakhm..17:27
kyakand $version is 5.10.X, but modules get installed into 5.10 (which is defined in perlmod.mk)17:28
kyakso i decided just to use otherlibdirs for that (i was reading INSTALL file, all these desribed there)17:28
xMffI wonder how perl reacts if we just replace /usr/local/foo with foo17:28
kyakis it done be sed?17:29
kyak*by17:29
xMffI just wonder what happens if some stuff using perl that is crosscompiled does not find required stuff in the otherlibdirs17:29
xMffand then skips to /usr/local17:30
kyakthis case is not worse than it is now :)17:30
xMffif there are perl modules there and they're using XS they mightn ot fit to the staging perl and crash17:30
kyakwhen there are only /usr/local17:30
xMfftrue17:30
kyaki didn't test perl on target17:31
kyakbut as it seems, it is build in a separate directory17:31
xMffah yes, that the other issue. it will have full paths embedded17:31
kyaknot host-perl17:31
kyakso it should not be a problem17:31
xMffah, right17:31
xMffwell I'll just apply your patch for now17:31
xMffit looks like an easy enough workaround17:31
xMffand no distro is shipping perl in /usr/local17:32
kyakno distro shoudl ship anything in /usr/local...17:32
xMffhm one thing in your fix... the 5.10 path component17:32
kyakyep17:32
xMffit should be inferred from the version17:33
kyakthe problem is, that it is hardcoded in other place17:33
kyakand all modules go into 5.10 directory17:33
xMffthere is PKG_VERSION17:33
kyakwhile actually the $version is 5.10.x17:33
kyakPKG_VERSION:=5.10.017:34
kyaki can't use this17:34
kyakPERL_LIB := $(STAGING_DIR)/usr/lib/perl5/5.10 in perlmod.mk17:34
kyaktherefore modules install there17:35
xMffwhy don't you jsut use this then?17:35
kyakhm..17:35
kyaki need to include this Makefile then, right?17:35
xMffit already is17:36
kyaki'm testing now17:37
xMffokay17:37
xMffI am waiting then17:37
kyakshouldn't take long.. after i replaced my SATA cable the HDD is UDMA-133 again, before it always fallbacked to DMA-33 -\17:38
xMff;)17:40
kyakxMff: here goes : http://dpaste.com/524921/17:46
kyakbuilt fine, the directory is in @INC17:46
xMffgreat17:51
xMffcommitting now17:51
kyakthanks again, and nice hint :)17:52
xMffokay, done17:54
qi-bot[commit] Werner Almesberger: fpd2pdf/fpdoc: added option  -a module  to manually add modules http://qi-hw.com/p/eda-tools/e92f45820:47
qi-bot[commit] Werner Almesberger: miscellaneous integration of fiducials http://qi-hw.com/p/ben-wpan/16ac7e520:50
qi-bot[commit] Werner Almesberger: makefiles/Makefile.kicad: replaced a few $(NAME) with patterns (%, $^, $<) http://qi-hw.com/p/ben-wpan/33f320e21:00
LunohoDlars_: jz47xx_i2c_xfer shouldn't always return ret21:23
LunohoDmaster_xfer should return the number of messages successfully21:23
LunohoDprocessed, or a negative value on error21:24
LunohoDif(ret) return ret; else return i;21:24
LunohoDsomething like that21:24
rjeffrieswpwrak re fiducial marks. some type of cross mark would make it easier to locate center of the mark23:41
--- Thu Mar 24 201100:00

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