#qi-hardware IRC log for Wednesday, 2012-04-11

pabs3more open hardware folks: https://lwn.net/Articles/491515/01:47
qi-botThe build was successful: http://fidelio.qi-hardware.com/~xiangfu/build-nanonote/openwrt-xburst.full_system-20120409-2355 04:42
kyakmth: thanks, i'll try the fix05:11
kyakjow_laptop: i agree with what you said about the toolchain name.. i added your comment to https://dev.openwrt.org/ticket/1126405:17
wolfspraulI regret a little that we stopped working on the atben/atusb casing05:28
wolfspraulthat asks for resurrection :-)05:28
wolfspraulI just want to embed the whole thing in a blob of epoxy or so. or a transparent plastic/glue, then you can still see the innards. it should not deteriorate the rf performance though.05:29
xiangfukyak, this build is for release 05:30
kyakwolfspraul: having atusb upstream would be so much nicer than a casing!05:30
kyakxiangfu: yea, this looks good for release.. most fallouts have been fixed05:31
xiangfukyak, Hi. since you have the directvnc. do you have the Makefile already? I would like try directvnc here. 05:32
kyakyes, i got the Makefile05:32
kyakshould i commit or rather upload it somewhere? it builds and runs, but the screen is mangled05:32
xiangfukyak, do you think if a normal openwrt router can running a vnc server?05:33
xiangfukyak, commit is better.05:33
xiangfukyak, it builds and runs. better then those @BROKEN packages. :-)05:33
kyakxiangfu: i didn't see any vnc servers ported to openwrt so far.. i also think that X11 is a requirement05:34
kyakwhich may be problematic for routers05:34
xiangfuok. 05:35
xiangfujust asking. 05:35
xiangfuif there is vnc server running on openwrt router. then Ben Nanonote can be used as 'Monitor' :-)05:35
xiangfube back later. 05:35
qi-bot[commit] kyak: directvnc: initial port (master) http://qi-hw.com/p/openwrt-packages/96cd14605:36
wolfspraulkyak: what can we do? [atusb upstream]05:36
wolfspraullet's break it down into smaller tasks05:36
wolfsprauldon't underestimate the casing though, I think the current bare-pcb makes the boards essentially unusable, confining them to a lab-only developer desk05:36
wolfspraulevery time I even carry them around I have to take special precautions05:37
kyakwolfspraul: ressurect the atusb software development :) i might be mistaken, but i think it is currently in a half-tested stage, and not able to compile as kernel module (only monolitically).. when it's nice and clean and working and tested, then it perhaps can be accepted by upstream05:38
kyakand helping with atusb driver development is beyond my limits/knowledge.. so i could only help by testing05:39
wolfspraulyes, ok. let's see.05:40
wolfspraulmy use case here would probably be that I have an atusb in an openwrt-based router, and atben in the Ben05:40
wolfspraul(for networking)05:40
wolfspraulthat should be super easy to setup and get to work05:41
wolfspraul*should* :-)05:41
wolfspraulagreed? is that the use case we are targeting?05:41
kyakyeah, putting atusb in openwrt router is the use case.. putting atusb in a laptop from time to time is also a use case05:43
kyakthe latest commit in ben-wpan-stefan is from 25 Sep 2011. .this is the problem05:44
wolfspraulthat's probably a little harder to achieve initially, so I'd rather target openwrt on the other side first05:44
wolfspraulsure, it's, let's say, "on hold" :-)05:44
kyakso you mean, upstream it to openwrt, then perhaps it will find its way to upstream kernel.. sounds good05:45
wolfspraulnot 'perhaps', but yes05:45
kyakok, 'yes' :)05:45
wolfspraulI am thinking about the shortest path to a practical use case05:45
wolfspraulI want to use this thing05:45
wolfspraul(other than looking at it, which is a start but not enough :-))05:46
wolfsprauland like I said I think I also want to experiment with some casing05:46
wolfspraulprobably in the form of an epoxy or glue blob05:46
wolfspraulyes it takes away 'hackability', but what exactly would one want to do?05:47
wolfspraulso it makes more sense to document some good materials and how to exactly get it sealed, and impact on RF performance05:47
xiangfukyak, the build host mplayer still have faad2. :( http://pastebin.com/W1JdwxtD06:33
xiangfukyak, I try to disable it under './configure' 06:33
xiangfure-generate the rootfs now. 06:33
wolfsprauldon't worry, be happy06:35
kyakmaybe just add faad2 to rootfs?06:35
wolfspraulkyak: have you found any time to play with your milkymist board?06:35
xiangfukyak, it's patent package.06:35
kyakreally? oh..06:36
xiangfuit's for AAC. mpeg-406:36
kyakwolfspraul: not yet really.. but it's on my desk in front of me, so i won't forget :) just need some time06:37
xiangfukyak, how to find out the ./configure option. ./configure --help give me nothing .06:37
kyaklemme have a look..06:38
wolfspraulthe fpga is really a new angle/dimension to think of, so having it there for a while is good06:40
kyakxiangfu: --disable-faad should do the trick06:43
kyaki think this can be put into CONFIGURE_ARGS under the else for CONFIG_BUILD_PATENTED06:45
xiangfukyak, v06:57
xiangfuunrecognized options: --disable-faad06:57
xiangfukyak, easy way is just './scripts/feeds uninstall faad2' :)07:08
xwalkCan anyone assist me in the process of hooking up the Ben Nanonote to a Fedora Machine through usb? I can ssh into the Nanonote no problem, but I can't seem to get a ping sent from the Nanonote to a web address (e.g. google.com).07:23
kyakxiangfu: some package pulls that faad2 as a dependency, i guess07:33
kyakxwalk: did you follow instructions in http://en.qi-hardware.com/wiki/Ethernet_over_USB ?07:35
kyakxiangfu: it doesn't complain for --disable-faad for me..07:39
xiangfukyak, ok. then we also add this to makefile. 07:39
xiangfukyak, I driect test this option under build_dir/.../mplayer.../07:40
kyakxiangfu: better add this to Makefile and make package/mplayer/{clean,compile} V=9907:41
kyakthen i see faad2 in disabled codecs07:42
kyakat the end of configure process07:42
xiangfugreat. 07:42
xiangfuI am re-generate the rootfs now... after (./scriptfs/feeds uninstall faad)07:43
kyakdo yo umean clean rebuild?07:43
xiangfuno. no.07:43
xiangfunot make clean. just make for re-generate the rootfs and packages.07:44
kyakthis might be not sufficient, as ./scriptfs/feeds uninstall faad doesn't remove the faad2 headers/libs07:44
kyakvery strange, as faad2 depends on BUILD_PATENDED07:44
kyakhow did it even get built?07:45
xiangfumake package/faad/clean && ./scripts/feeds uninstall faad. 07:45
kyakyes, this should work.. but then again, why is it even there07:45
xiangfubug somewhere.07:46
kyaki don't see faad anywhere in config as well07:46
kyakit should not have been built07:47
kyakit's not in resulting packages as well07:47
qi-bot[commit] kyak: mplayer: disable faad explicitely (master) http://qi-hw.com/p/openwrt-packages/e07e94107:51
kyakit seems that CONFIG_ALL makes it so that all packages are built; but only those packages not depending on BUILD_PATENTED are actually installed (i.e. actual ipk created).07:57
kyakby looking into build log i see that libfaad2 is built, but no ipk is created with "WARNING: skipping libfaad2 -- package not selected"07:58
kyakthis is pretty strange.. i would say that BUILD_PATENTED must be stronger than CONFIG_ALL07:59
xiangfumake sense. 07:59
xiangfureport bug again :)07:59
kyakso we need to disable shit very explicitely..07:59
kyakthere are other packages that exhibit the same behaviour08:01
kyaki think i get it now08:02
kyakpackage pianobar has libfaad2 as PKG_BUILD_DEPENDS08:02
xiangfuso BUILD_PATENTED must be stronger than  PKG_BUILD_DEPENDS and all others. 08:03
kyakthe problem in this particular case is that the package pianobar provides several packages, like 'pianobar' and 'libpiano'; and only 'pianobar' depends on libfaad2, but 'libpiano' doesn't depend on anything patented, so it gets build. And it leads to building of packages in PKG_BUILD_DEPENDS, which has faad208:08
kyakin principal, this particular package is just poorly packaged08:09
kyakbut there might be more packages like that08:09
kyakit could also mean that some other ipk's might require patented libraries as well, because they were present at build time, but (luckily) ipk's for these libraries are not built anyway08:12
kyakso we are not limited to mplayer problem08:13
xiangfukyak, please report a bug to upstream since you already dig such information 08:16
wolfspraulI think dependency issues are potentially and in reality infinite. so I would try to clean them up carefully on an as-needed basis08:17
wolfspraulrather than trying to fix "them all", which must be the ultimate job for sysiphos08:17
wolfspraulbut little by little we can make the dependencies both more accurate and flexible at the same time, sure08:18
kyakthat's quite true08:18
kyakfind the bug - fix the bug :)08:19
qi-bot[commit] Xiangfu: nanonote buils scripts file: add patch-openwrt.sh (master) http://qi-hw.com/p/openwrt-packages/caa7c0908:22
xwalkkyak: Sorry for the delayed response. At work showing a new employee the ropes. Yes, I certainly did.08:22
xwalkkyak: I made sure I am using wicd as well.08:22
xiangfuFedora have some default rules at 'iptables'08:25
xiangfuxwalk, you may just disable all iptables. then add iptables -A POSTROUTING -t nat -j MASQUERADE -s
xiangfuand sysctl -w net.ipv4.ip_forward=108:26
xiangfuxwalk, then your nanonote should able to access Internet.08:26
xwalkIs removing the iptables service going to need a reboot to take effect?08:30
xiangfuxwalk, iptables -F should be ok.08:30
kyakxiangfu: https://dev.openwrt.org/ticket/11266 (lots of letters)08:31
xwalkxiangfu: Got it working. Thank you very much. =)08:31
xwalkkyak: Thank you as well.08:31
xiangfukyak, thanks 08:32
qi-bot[commit] Xiangfu: js28f256j3f105: add missing A8 pin (master) http://qi-hw.com/p/kicad-libs/102585910:43
cladamwxiangfu, nice now, you can back up me. ;-)10:48
DocScrutinizertsunami warning cancelled11:10
wpwrakDocScrutinizer: in nuernberg ? that must be a big one11:12
DocScrutinizerI raely spam IRC with nuernberg local news11:14
wpwrakso it's a global tsunami ? even scarier11:15
DocScrutinizerat least not the "global" channels of IRC11:15
DocScrutinizernah, only indian ocean, like 200411:15
DocScrutinizerso 'only' a few millions affected11:16
DocScrutinizernot exactly *global*11:16
DocScrutinizera global tsunami would mean either a) yellowstone exploded and took USA with it, or b) some meteorite size in km-range hit some ocean. In both cases I think you couldn't globally cancel the tsunami warning11:22
wpwrakwell, sometimes meteorite trajectories get mis-calculated11:23
DocScrutinizeraftershock quake 8.1, tsunami warning still extended by 2 hours11:44
qwebirc39331anyway of upgrading all openwrt packages from nanonote?12:15
xiangfuqwebirc39331, Hi12:18
xiangfuqwebirc39331, first you needs make sure the nanonote connect to Internet: openwrt-xburst.full_system-20120409-235512:19
xiangfuqwebirc39331, I think so far no one have tried to upgrading all packages. 12:20
qwebirc39331a new image? so I hace to use usbboot instead?12:20
xiangfuusbboot is faster.12:21
qwebirc39331I was trying to use a command I saw on a website, but didn't work.12:21
xiangfuqwebirc39331, you mean usbboot?12:22
qwebirc39331no, the command I saw was " opkg upgrade  'list-upgradable' "12:24
qwebirc39331also, where can I find a pygame package?12:24
DocScrutinizer"tsunami aftershock not causing any tsunami either" says German TV ZDF correspondent Peter Kunz13:01
DocScrutinizer"aftershock earthquake not causing any tsunami either" says German TV ZDF correspondent Peter Kunz13:02
xiangfuhmm... we have some base components on board-qi-avt2 and milkymist 13:02
xiangfulike testpoint.13:03
xiangfuNow we have kicad-libs. should add all the components we used to that repo. 13:03
wolfspraulsure you can clean that up, - BUT - only if we can keep the quality up13:04
wolfspraulif it's a blind copy/paste activity then maybe let's wait until we really use those parts13:04
xiangfuyes. 13:04
wolfspraulthe #1 problem of kicad libraries is low quality imho13:04
xiangfuagree. I will copy the testpoint first. since I start to need that.13:04
qi-bot[commit] Xiangfu: add testpoint (master) http://qi-hw.com/p/kicad-libs/a7b0a8b13:07
xiangfusorry. what you mean "low quality" ? leak feature?13:09
wolfspraulincomplete, buggy13:09
wolfspraulthe parts I mean13:09
Fallenoukyak: not sure if X11 is required by VNC server, I've seen it working on an Android plateform without X11. Android was just using the OMAP3 framebuffer (omapfb) in /dev/graphics/fb013:12
Fallenouso I guess there is a way of using VNC server with a framebuffer13:13
viricthe vnc server is an X11 server13:13
viric(at least the traditional Xvnc)13:14
qi-bot[commit] Xiangfu: add common.lib include R, C, LED (master) http://qi-hw.com/p/kicad-libs/75a4fba13:48
xiangfuwpwrak, adam please check the NORFlash sheet. I just finished it. 13:50
xiangfuI am not sure on junction. please give some advice. 13:51
xiangfucheck log tomorrow.13:51
wpwrakhmm, LibDir is a bit of a problem13:54
wpwrakif you have /home/qi/kicad-libs  and /home/qi/board-m1, then the components aren't found13:54
wpwrak(substitute /home/qi for anything you like)13:55
wpwrakand yes, shorting all the FLASH_{RESET*,etc.} lines together doesn't look right :)13:56
wpwrakyou also picked A3 paper. does everyone around here have an A3 printer ?13:57
wolfspraulI think we should merge those styles & units things into a file in the repo, no?13:57
wolfsprauland remove it from the wiki13:57
wpwrakyes, there should be a proper manual13:58
wpwraknot sure how to deal with the current flurry of changes, though. my current plan is to wait until adam runs out of steam and then start fixing things :)13:59
wolfspraulwell it's better to discuss early14:00
wolfspraulor wait until the current status has reached consensus quality level14:00
wpwrakLibDir has an easy solution. good :)14:01
wolfspraulare any of the things you see being done a waste of time and need to be redone?14:01
wolfspraulas long as we are lifting each other up that's ok. but if someone goes in a wrong direction we should notify early.14:01
wpwrakthere are several font size issues14:02
wpwraki'm also not very happy with naming negated signals FOO#14:02
wpwrakthe pin number in TPxx is useless. there's only one anyway.14:03
wpwraksheet size is probably the worst time-wasting risk14:04
kyakFallenou: are you sure what you've seen wasn't vnc client14:56
Fallenoukyak: it was a VNC server running on an Android gadget15:14
kyakFallenou: probably something like that: http://code.google.com/p/fastdroid-vnc/15:34
kyakdon't see how it can be done on a router though.. first need to implement some kind of virtual framebuffer device?15:35
Fallenouwell yes you need what you want to share through VNC to be drawned in a framebuffer15:38
Fallenouthe vnc server I've seen was some proprietary official vnc server15:39
Fallenouwhich needed a license15:39
Fallenoubut it seems you have found a project which support android framebuffer which is nice :)15:41
Fallenouwill it support other framebuffers ... for your router ... dunno15:41
Fallenouthe code seems to be rather generic, it should work with any other framebuffer I think15:45
Fallenousee http://code.google.com/p/fastdroid-vnc/source/browse/trunk/fbvncserver.c15:45
kyakmth: the bug in gmenu2x is no more. Thank you!17:11
mthgood to have that confirmed, thanks17:12
qi-bot[commit] kyak: gmenu2x: update to latest git (master) http://qi-hw.com/p/openwrt-packages/7671b2b17:12
qi-bot[commit] Werner Almesberger: kicad-patches/README: added quilt-based update process for KiCad (master) http://qi-hw.com/p/eda-tools/1a72c9720:32
qi-bot[commit] Werner Almesberger: kicad-patches/README: bump KiCad version to bzr 3494 (master) http://qi-hw.com/p/eda-tools/508397320:32
wpwrakwolfspraul: btw, i couldn't build kicad bzr 3493 (pcbnew didn't link). so i bumped it to 3494.23:00
wolfsprauland that worked?23:00
wpwrakit compiled fine. there was a change in just the area where the problem appeared, which i take as a good omen. (didn't try to dig deeper)23:01
wolfspraullet me know if you need upleveling23:02
wolfspraulI think next time I can do it in 1h23:02
wolfspraulI have a simple test bench now, the uplevel steps are documented, etc.23:02
wpwrakyour patches applied fine to 349423:02
wpwrakafk for a bit23:04
wolfspraulit seems kicad is moving to metric here and there23:38
wolfspraulso in the modules position file for example, I changed the default (and only) to metric23:38
wolfsprauland front and back in the same file _all.pos (before it was _Front.pos and _Back.pos)23:38
wolfsprauljust fyi23:39
wolfspraulI don't think we are using that anyway23:39
wpwrakthe .pos file ? of yes, we're using that for fab output23:41
wolfspraulthen my upleveling certainly broke that23:41
wolfspraulkicad now has a dialog with some options there, metric or inches, 'force smd', front and back in separate files or in one file23:42
wolfspraulwhat do you want?23:42
wpwrakright now, pos2fab (from eda-tools/fab/) assumes inches. not sure what "force smd" means. not sure whether fabs are happier with front and back split or together.23:44
wpwrakpos2fab takes the .pos file and turns it into an excel-friendly .csv table23:44
wolfsprauljust tell me what you want :-)23:45
wolfspraulyou can look at the dialog now, see the options there23:45
wpwraksee http://downloads.qi-hardware.com/people/werner/wpan/fab/atusb-smt-110330.tar.gz23:45
wpwrakfile atusb/atusb-bom.csv23:45
wpwrakerr, wsorry23:46
wpwrakthat's the one23:46
wpwrakthe defaults look nice: inches, one file per side, with INSERT attribute set23:48
wpwraki don't care much about the inches. just pos2fab needs to agree with what we extract from kicad23:49
wpwrakperhaps we should have a switch23:49
wpwrakand afk again. may take a bit. proprietor's reunion ...23:58
--- Thu Apr 12 201200:00

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