#qi-hardware IRC log for Sunday, 2010-09-12

wolfspraulxiangfu: strange, every time I run scripts/feeds install -a, it installs a few more packages00:29
wolfspraulI can run it 3, 4 times, and every time a few more get installed :-)00:30
xiangfuwolfspraul: yes. there are new packages commit.00:30
wolfspraulno no00:30
wolfspraulit looks like a problem of the scripts, or some other cause like dates or so00:30
wolfspraulI make no changes at all. Just start with a git clone, the scripts/feeds update -a, then scripts/feeds install -a00:31
wolfspraulbut 'install -a' will not install all package, only some00:31
wolfsprauland more will come down every time I run 'install -a' again00:31
wolfspraulnot just from the qipackages feed, all feeds behave like that. I have tried on the tracking_trunk branch too, same behavior00:32
wolfspraulmaybe when going through the packages in a feed, something will make OpenWrt stop, and go to the next feed00:33
wolfspraulthen next time I run 'install -a' it will continue?00:33
xiangfuwolfspraul:  I am not sure.00:34
xiangfuwolfspraul: there must some wrong with you feeds/ folder. I guess.00:34
wolfspraulno the feeds/ folder is complete00:34
wolfspraulthe 'update -a' worked00:34
wolfspraulbut 'install -a' will only install them in batches, I keep running 'install -a' now and every time a few more pacakges get installed :-)00:35
wolfspraulwhen you run 'install -a' next time, try to run it again, see whether more is installed...00:35
wolfsprauldon't trust it to run only once00:35
xiangfuwolfspraul: just run five times, here is the output:00:36
xiangfuwolfspraul: http://pastebin.com/9usabzMP00:36
xiangfuno new package installed.00:37
wolfspraulgood for you :-)00:37
wolfspraulI ran about 10 times, then all packages were installed, now it's done and seems stable (nothing more to install).00:37
wolfspraulxiangfu: http://pastebin.com/eqkApukQ00:39
wolfsprauluntil I understand the root cause, I will always run 'install -a' several times until it runs without installins new packages00:41
xiangfuwolfspraul: I am looking into the ./scripts/feeds file.00:46
wolfspraulexperimenting more. it may have something to do with dependencies.00:49
wolfspraultry this: scripts/feeds uninstall -a -p xfce00:50
wolfspraulthen scripts/feeds install -a -p xfce00:50
wolfspraulthen another scripts/feeds install -a -p xfce00:50
wolfspraulin my case, the first install -a will not install all packages from xfce, the second, third, fourth, etc. will install more00:51
wolfspraulI need about 9 times, then they are all installed. always in the same order/sequence.00:52
xiangfuwolfspraul: the first line give an error: Unknown option: p00:53
xiangfuwolfspraul: seems works fine here. : http://pastebin.com/aFasWdf600:55
qi-bot[commit] jow: [package] base-files: add an experimental "-c" flag which tries to preserve *all* changed files in /overlay/etc minus some system files http://qi-hw.com/p/openwrt-xburst/feaf88f01:03
qi-bot[commit] jow: [imagebuilder] http://qi-hw.com/p/openwrt-xburst/67f09fa01:03
qi-bot[commit] jow: [imagebuilder] fix arch detection pattern to cope with underscores in board name, e.g. "adm5120_mips" http://qi-hw.com/p/openwrt-xburst/cbcb7e201:03
qi-bot[commit] jow: [package] util-linux-ng: make build depend on libncurses, cfdisk is compiled unconditionally and will fail with missing input files if ncurses was not detected during configure http://qi-hw.com/p/openwrt-xburst/ba5962501:03
qi-bot[commit] mb: iproute2: Fix compilation http://qi-hw.com/p/openwrt-xburst/9d4993c01:03
qi-bot[commit] jow: [package] uhttpd: break tight loop when receiving eof during header reading (#7904) http://qi-hw.com/p/openwrt-xburst/8c59fdd01:03
qi-bot[commit] jow: [include] pass opkg offline root as command line arg, solves caching issues when switching between multiple subtargets, thanks Ralph Hempel for reporting http://qi-hw.com/p/openwrt-xburst/5e3ed6701:03
qi-bot[commit] mb: iproute2: Once the make calls were fixed up, parallel compile seems to work well. http://qi-hw.com/p/openwrt-xburst/177f4b101:03
qi-bot[commit] jow: [imagebuilder] fix a small typo in sed pattern http://qi-hw.com/p/openwrt-xburst/12bb9fe01:03
qi-bot[commit] mb: Allow receiving input events even if the X-server grabbed the device. http://qi-hw.com/p/openwrt-xburst/4aff85301:03
qi-bot[commit] mb: Refresh patch http://qi-hw.com/p/openwrt-xburst/261721101:03
qi-bot[commit] florian: [package] fix acx-mac80211 build failure with latest compat-wireless http://qi-hw.com/p/openwrt-xburst/eba6ff501:03
qi-bot[commit] florian: [package] update acx-mac80211 to 20100830 http://qi-hw.com/p/openwrt-xburst/60b11dc01:03
qi-bot[commit] mb: omap24xx: Add pointercal file http://qi-hw.com/p/openwrt-xburst/1d74d1e01:03
qi-bot[commit] mb: omap24xx: Add default network config http://qi-hw.com/p/openwrt-xburst/53687e501:03
qi-bot[commit] florian: [package] nvvram: no checksum error recording on WRT54G3GV2(-VF) http://qi-hw.com/p/openwrt-xburst/edc292c01:03
qi-bot[commit] florian: [buildroot] ccache TARGET_CXX http://qi-hw.com/p/openwrt-xburst/691fd2401:03
qi-bot[commit] jow: [package] 6in4: don't use abbreviated notation for, iproute apparently does not like it anymore http://qi-hw.com/p/openwrt-xburst/7b3bd1201:03
qi-bot[commit] jow: [package] firewall: http://qi-hw.com/p/openwrt-xburst/c65371301:03
qi-bot[commit] mb: Fix kernel crypto modules for >= 2.6.36 http://qi-hw.com/p/openwrt-xburst/8eeeaa501:03
wpwrakDocScrutinizer: another question: the thickness of my board differs indeed from that of ti's reference design that uses the same antenna. i have 0.8 mm, they have 1 mm.02:48
wpwrakDocScrutinizer: the application note says "To compensate for a thicker/thinner PCB the antenna could be made slightly shorter/longer."02:48
wpwrakDocScrutinizer: any idea how to interpret this ? since my board is 20% thinner, should I make my antenna 20% longer ? or maybe 10% ? also, which of the lengths would expand - only L2 ? L2 and L5 ? also L3 ? etc.02:51
wpwrakDocScrutinizer: lengths from page 4 of http://focus.ti.com/lit/an/swra117d/swra117d.pdf02:52
wpwrak DocScrutinizer i suppose i could somehow measure the antenna performance with my usrp2. things like reflection. hmm ...03:01
wpwrakhere's some guy who did characterize an antenna with an usrp: http://swigerco.com/gnuradio/phase/isotron.html03:11
qi-bot[commit] kyak: kbd: added kbd console maps http://qi-hw.com/p/openwrt-packages/b75443805:01
wolfspraulkyak: do you know whether we include cyrillic fonts now?05:08
bartbesso, does anyone know why i can't build gforth yet?05:20
wolfspraulI'm working my way through the build system, but not there yet...05:20
kyakwolfspraul: we do, i can read and type cyrillic in console (UTF-8).. with certain limitations right now - those keymaps still need to be adapted to Ben's keyboard layout05:29
wolfspraulkyak: are they included in our normal OpenWrt builds?05:36
wolfspraul(the cyrillic fonts)05:36
kyakthere is a new package kbd-console-maps including many standard Linux console fonts05:38
kyakif it is included on a next build, then we would have it05:38
wolfspraulah ok, cool05:38
kyakoh, no.. not exactly like this05:39
kyakthere is a "kbd" package, that provides tools for setting console fonts and manupulating keymap switching05:40
kyakthe kbd-console-fonts package includes the fonts05:40
kyakthe kbd-console-maps package includes those keymaps05:40
kyakthere are several problems: 1) the fonts provided with kbd are too big; 2) keymaps are only partial usable and 3) some user-space programs still don't work correcly in such console05:43
kyakfor example, sdcv (console version of stardict) is showing cyrillic correctly, but it doesn't accept cyrillic input05:44
wolfspraulok, but as a first step we should still include those 3 packages in the image?05:45
kyaki'd like to map the "Ben" key for layout switch instead of "Ctrl" :)05:45
kyakwolfspraul: yes, they can be included (in fact, kbd and kbd-console-fonts were already included i think)05:45
DocScrutinizerwpwrak: moinmoin06:56
DocScrutinizerwpwrak: wasn't there some pointer to details about "longer/shorter" in appnotes of the whole device?06:56
wpwrakDocScrutinizer: good morning :)  i found only the sentence i quoted08:49
wpwrakDocScrutinizer: but nothing specific. what i'm loooking for is at least an educated guess for how much i should add08:50
DocScrutinizerI'm sure I've read "for details see application notes of WX5836-4711foobar" OWTTE08:51
wpwrakyes, but that are the details of the reference design board. so there you find out that it's 1.0 mm. an043 doesn't tell you even that.08:53
DocScrutinizerbtw: http://www.google.com/webhp?#q=swr-meter+selbstbau09:00
wpwrakoh, nice. thanks !09:02
wpwrak(swr) meter. hmm, they need a lot of power. but i'm beginning to get the concept09:29
wpwrakDocScrutinizer: so ... any suggesting for a first try antenna resizing for the 1 mm to 0.8 mm change ?09:37
DocScrutinizerI'd not care and simply go for the "tunig mod" I suggested (the L shaped end at 'top' of antenna)09:42
DocScrutinizerthis "increase/reduce length" is rubbish, nobody gets the real meaning09:43
DocScrutinizeralso it unclear to me how exactly PCB thickness influences ant tuning at all09:44
wpwrakokay, so i'll just add a little "tail"09:45
DocScrutinizerhowever I can see huge impact of even plastic case covering antenna area - of course nothing metal must cover antenna area either side09:45
wpwrakthe "description" of the changes necessary is indeed rather vague. ti tend to be obscure about these things in general, it seems. it's also quite a bit of work to find out how to do a balun, assuming you don't want their "just copy the gerbers with all the microstrip inductors" design09:47
wpwrakthe effect of plastic sucks a bit indeed. i think this will be fun. hence my idea to build my own SWR meter, so that I can at least see where things go wrong.09:48
DocScrutinizerIn TPE legendary bookstore I bought a nice book about antenae only ~600+ pp. Gave it to Tully, maybe should have sent to you instead :-D09:49
wpwrakheh, i miss that bookstore09:50
DocScrutinizerme too09:50
DocScrutinizernever thought I'd say that but actually I miss whole TPE experience every once in a while :-D09:51
DocScrutinizermodulo the flights09:51
wpwrakthere had been good times amidst the bad ones, yes.09:52
DocScrutinizeryeah actually09:52
wpwraki find it fascinating how often germans complain about little flights of 12 hours or such :)09:52
DocScrutinizerI bet tampere FI couldn't compete, if ever that'd become an option for me09:53
DocScrutinizeronly since they forbid smoking09:54
DocScrutinizerflight to New Orleans 1985 was nice. Smokers, whole row for me so I could stretch over 3 seats...09:56
wpwraki once had a flight geneva-jfk the week after 9/11. that was nice. almost empty ;-)09:57
wolfsprauld :w09:57
DocScrutinizerhi wolfgang09:58
wolfspraulDocScrutinizer: I think I may have seen that book at Adam's place. Tully gave it to Adam.09:59
DocScrutinizerwolfspraul: any idea if om.com postmaster could be prodded to forward my .com email address?10:01
DocScrutinizeror at least implemet an autoresponder telling about the address should be .org now10:03
wolfspraulI massively doubt this will ever happen.10:03
Action: DocScrutinizer as well :-D10:03
wpwrakDocScrutinizer: here's an interesting SWR meter: http://ludens.cl/Electron/swr/swr.html10:36
viricAnyone with a Neo Freerunner?10:37
wolfspraulviric: just shoot your FreeRunner question...10:43
viricI'm about to buy one10:44
viricI wanted to know if there is anything bad I should know...10:44
viriccan it properly receive/make calls, receive/send sms?10:44
wolfspraulare you serious?10:45
viricI heard it had troubles10:45
viricI heard little about it, and I heard little good...10:45
viricMaybe those telling me that were iphone fans, I don't know10:45
viricI'm reading http://wiki.openmoko.org/wiki/Neo_FreeRunner_Hardware_Issues10:45
wolfsprauldo you want to have a phone you can use as a normal daily phone?10:46
wpwrakwolfspraul: rafa seems to be quite happy with his as a daily phone:)10:46
wpwrakDocScrutinizer: a little tail like this ? http://downloads.qi-hardware.com/people/werner/wpan/antenna-with-tail.ps10:46
viricI thought your "are you serious?" was more about "of course it works". But maybe you mean "of course it still does not work" ?10:47
wolfspraulviric: ah no, that's just because I was laughing since yes, there are people in this channel who know an awful lot about FreeRunners... :-)10:47
wpwrakviric: maybe he as referring to the fact that openmoko stopped making phones more than a year ago ...10:48
wolfspraullong story short, if you are a real true hardcore free software lover, go buy your FreeRunner and (most likely) start the suffering10:48
wpwrakviric: ... and of course, you find a surprisingly large number of ex-openmokoers around here :)10:48
wolfspraulout of 100 people, I think maybe 2 or 3 successfully continue to use their FreeRunners as a daily phone after let's say 1 week10:49
wolfspraulmost use it as a GPS device, tangopos was by far the most successful application10:49
viricWhat should I buy, then, if I want a not-expensive quite-free phone?10:49
wolfspraulviric: do you like Android? is that free enough?10:49
DocScrutinizerwpwrak: yoh. can mill away until antenna tuned10:49
qi-bot[commit] kyak: kbd unicode cyrillic map for Ben http://qi-hw.com/p/openwrt-packages/d7307ad10:50
viricaren't android-running phones quite expensive?10:50
DocScrutinizercompare tuning of tha monster ionon or whatever, you linked me to10:50
wolfspraulviric: I have no good advice for you. I switched to the cheapest possible China phones. totally non-free.10:51
wpwrakDocScrutinizer: i'll just try RSSI for now :)10:51
wolfspraulmost former FreeRunner fans seem to have moved on to Nokia N900 or some Android phones from HTC10:51
wolfspraulbut they are all expensive10:51
viricwolfspraul: maybe if they can run j2me programs fine, I'd be happy, if it is cheap enough :)10:51
wpwrakDocScrutinizer: received signal strength10:51
viricwolfspraul: are you happy with your phone?10:51
wolfspraulit's off most of the time10:51
viricmaybe I should ask someone else :)10:52
DocScrutinizerwpwrak: yepyep. Anyway the tuning of that monster was by "moving the lower lever", which is kinda exactly what you'll do by milling away the lil foot10:52
wolfspraulviric: what you are really looking for doesn't exist right now. cheap & free.10:53
wolfspraulimho it doesn't exist10:53
wolfspraulif you have an upper price limit, well, maybe you should just either get the cheapest Android phone, if you think that's free enough, or forget about the freedom thing altogether until another attempt at a free phone comes along.10:54
DocScrutinizernot even N900 with stock maemo is really a fullfeatured everyday phone and still free10:54
viricI planned to do that10:54
wolfspraulmany people still work on free phones, I consider myself one of them. But admittedly if you look at what i do today, it's quite a stretch to see how that's part of a phone :-)10:54
DocScrutinizerviric: or you reduce expectations about "everyday flawless usage" and actually go for an A7 FR10:55
DocScrutinizerwith SHR10:55
wolfspraulbut I have nothing for you I would dare to offer you as a 'phone', and will not have it for a while I guess10:55
viricDocScrutinizer: what is that?10:55
DocScrutinizeror maybe QtMoko (heard it's also getting nice lately)10:55
viricDocScrutinizer: a7 fr?10:56
DocScrutinizerFreRunner revision A710:56
wolfspraulDocScrutinizer: but he needs to be quite generous in terms of lowering his expectations, and also a bit lucky in terms of his network/area.10:56
wolfspraulthen I agree, he could try, jump into FreeRunner land :-)10:56
wolfspraulviric: if you do it, DocScrutinizer told you so, not me, ok?10:56
viricok :)10:56
Action: DocScrutinizer hides10:57
DocScrutinizerviric: you might want to /join #openmoko, #openmoko-cdevel10:58
viricI tried10:58
viricbut they remain silen10:58
DocScrutinizerviric: ask on -cdevel about status of latest SHR10:58
viricbut the problems there are in userland software?10:58
DocScrutinizerwith A7 revision, yes mostly10:59
viricis a7 a *hardware* revision?10:59
viricIsn't there any website listing the freedomness of phones?10:59
DocScrutinizerthe one where I finilly managed to push a semi-decent fix for buzz issue10:59
viricI'm at linuxfordevices.com taking a look10:59
DocScrutinizerviric: btw A6 with buzzfix alegedly is even a tad better than A711:00
wolfspraulviric: the freedomness for phones is in such a bad state that any list is mostly a rhetoric exercise anyway. who gets away with what claims of 'openess'.11:01
Action: bartbes tries compiling gforth again11:03
viricI have a journal explaining a bit about the mobile phone manufacturers, and how they behave...11:03
viricI'll take a look11:03
viric(troubles with worker conditions, where they produce, ...)11:04
bartbesokay, seriously11:07
bartbeshow come I am the only one with problems when compiling gforth?11:07
wolfspraulyou are most likely not11:09
wolfspraulbut I am still stuck at the boot issues I am debugging, so can't try out gforth yet... maybe tomorrow...11:10
wolfspraulif you can fix it first, great!11:10
DocScrutinizerviric: [2010-09-12 17:09:19] <mrmoku> basic calls work fine... GPRS and GPS too11:12
bartbeswell, it just seems to be a missing file11:12
DocScrutinizerviric: http://logs.nslu2-linux.org/livelogs/openmoko-cdevel.txt /join #openmoko-cdevel !11:13
qi-bot[commit] Werner Almesberger: Added antenna tuning "tail" suggested by Joerg. Made board slimmer. http://qi-hw.com/p/ben-wpan/b3374dd11:45
qi-bot[commit] Werner Almesberger: Moved cam/ to atrf/cam/ http://qi-hw.com/p/ben-wpan/f8ae35111:45
qi-bot[commit] Werner Almesberger: atusd/cam/pcb.pl: updated for new board shape http://qi-hw.com/p/ben-wpan/843a0bc11:45
qi-bot[commit] kyak: fbterm depends on fontconfig http://qi-hw.com/p/openwrt-packages/a86327812:54
qi-bot[commit] Werner Almesberger: BOM work in progress. Completed generation of capacitor characteristics. http://qi-hw.com/p/ben-wpan/c0052fa16:05
qi-bot[commit] Werner Almesberger: Rename bom/yageo-cap.* to bom/yageo.* to keep things simple. http://qi-hw.com/p/ben-wpan/66e5aa516:05
qi-bot[commit] Werner Almesberger: Adjusted parsing of component values. Now all capacitors get properly selected. http://qi-hw.com/p/ben-wpan/636011c16:05
qi-bot[commit] Werner Almesberger: Moved all capacitor tolerance and _E12_P translations to shared includes. http://qi-hw.com/p/ben-wpan/f2824f116:05
qi-bot[commit] Werner Almesberger: Added generation of Yageo RC0402 resistors. We can now produce the full BOM. http://qi-hw.com/p/ben-wpan/499df0e16:05
Ornotermesit would be cool if the actually can make that thing for that price18:25
zombi-_whats this hcnnale for?18:44
Ornotermeszombi-: http://en.qi-hardware.com/wiki/Main_Page18:47
zombi-ahh, totally not what i expected, although interesting :)18:48
Ornotermeszombi-: dare i ask what you expected? :P18:49
CongoZombiehas anyone tried the 2.6.35 patches for OpenWRT?18:50
zombi- hardware support channel :(18:50
Ornotermeszombi-: ah, but it's kind of support for qi-hardware18:56
qi-bot[commit] Carlos Camargo: aaa http://qi-hw.com/p/nn-usb-fpga/0975cd720:59
wpwrakrafa: nice subtle advertizing in your prompt ;-)23:31
wpwrakrafa: btw, is it still really necessary to pre-fill a swap file with dd ? i thought mkswap ... size does the trick as well ?23:33
--- Mon Sep 13 201000:00

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