**** BEGIN LOGGING AT Sat Jul 25 02:59:57 2009 Jul 25 03:45:17 TAsn: sorry for the late comment -- I think there is lower-hanging fruit to be had with opkg :) Jul 25 03:45:48 Specifically, the patchset I referred to earlier should solve a lot of problems. Jul 25 03:46:12 oooh opkg Jul 25 03:46:16 speaking of that... Jul 25 03:46:37 oplkg is a four-letter word. Jul 25 03:46:45 (sigh - I can't type) Jul 25 03:46:51 does it have any snapshotting/consistency system - eg if i pull power in the middle of a libc upgrade... Jul 25 03:47:11 very minimal. Jul 25 03:47:14 can i reboot and have a statically linked "opkgfix" trya nd clean up the mess from a "snap" ? Jul 25 03:47:19 no Jul 25 03:47:24 hmm ok. needs adding Jul 25 03:47:40 * mwester_ thinks we should just use apt and dpkg Jul 25 03:47:49 eg make a list of old files (well have it in pkg - but list anywya Jul 25 03:47:55 ) and new files, tar up old files. Jul 25 03:48:03 why re-invent the wheel? I'd rather extend those tools for a small system, than fix opkg. Jul 25 03:48:20 on boot if file exists "am doing install" delete a;l new files and replace with old ones from tar. Jul 25 03:48:26 (simply stupid thing - but an example) Jul 25 03:48:41 well i dont think apt and dpkg have this either?> Jul 25 03:48:57 No, but they at least work. opkg doesn't. Jul 25 03:50:15 last i knew apt required things like perl Jul 25 03:50:40 Perl is wonderful. Jul 25 03:50:43 Perl is good. Jul 25 03:50:58 Perl should be on every gta0x, anyway. :) Jul 25 03:51:15 it's a matter of your packaging system requirng a (heavy) scripting lang beyond what you already have (sh) Jul 25 03:51:37 you wont convince most commercial users of that :) Jul 25 03:51:38 Have you looked at the opkg source code? Jul 25 03:53:18 no - but i guess it's spaghetti-hell? Jul 25 03:59:46 It's worse than spaghetti. It's totally inconsistent in architecture and style, poorly-thought-out, leaks memory all over, duplicates code --- it was clearly written by some very inexperienced software developers, and they were not given the guidance and help they needed from a senior dev or architect. :( Jul 25 04:12:53 thats bad Jul 25 04:13:08 thol i disagree on "just use dpkg+apt" Jul 25 04:13:23 i think opkg just needs work - possibly a rewrite Jul 25 05:14:43 morning Jul 25 05:14:46 want to boot the shr-image on an other device...runs OK but crashes after starting xserver...how can I change the config to a 320x240 display? Jul 25 05:18:24 it is a gta01 image on a s3c2449 device Jul 25 05:27:58 schasch: i think x server determines the resolution from framebuffer parameters, that are supplied by the kernel. Jul 25 05:37:51 PaulFertser:tankx ...so adding something to "fb.modules"? Jul 25 05:38:39 or menuconfig kernel? Jul 25 05:40:51 schasch: what's fbset output on your device? Jul 25 05:49:09 morning guys! Jul 25 05:50:35 PaulFertser: 320x240 ...or what do you mean? Jul 25 05:53:17 I see the black screen..than comes the X in the middel....then CRASH Jul 25 05:54:19 just edited the Xserver file with debug-haltpoint-inf in ${IMAGE_ROOTFS}/etc/X11/ Jul 25 05:56:11 PaulFertser:going shower now and eating...till later... Jul 25 06:05:21 schasch: probably it's E crashing, not xserver. I'd suggest trying to start xterm instead of any WM, then to start E via ssh so you'll see why it crashes. Remove ~/.e before that. Probably starting in gdb will help. Jul 25 06:34:59 Hmm, building as pointed out on http://trac.shr-project.org/trac/wiki/Building%20SHR fails, because libframeworkd-phonegui-efl_git is missing "configure.in or configure.ac" Jul 25 06:35:11 schasch: i highly sauggest going to oe-dev (oe dev git branch) Jul 25 06:35:12 that's one thing I can't fix myself, don't have the know how Jul 25 06:35:19 and starting a new machine target Jul 25 06:35:31 and rolling over to angstrom Jul 25 06:36:06 "Tasks Summary: Attempted 1812 tasks of which 1811 didn't need to be rerun and 1 failed." Jul 25 06:36:16 hi. i've got an error with ophonekitd. latest unstable Jul 25 06:36:19 Unknown internal dbus error: Process /usr/libexec/dbus-daemon-launch-helper received signal 6, 79 (dbus-glib-error-quark), code 26 Jul 25 06:36:38 anyone encountered that lately? Jul 25 06:40:38 my problem is probably that I'm using my local checkout/repo Jul 25 06:40:41 ceda_wrk: report to respective dev .maintainer of that. if those are missing nothing is going to start building if its an autofoo (autoconf) based project Jul 25 06:41:17 it could be that the build has changed mechanisms and the oe build expects the older style and is checking out a specific revision Jul 25 06:41:25 raster: when I look at it, the files are there. I just burped what the error log file said. Jul 25 06:41:30 in that case the problem is yours to fix :) Jul 25 06:41:36 yes :) Jul 25 06:41:49 well then.. you're going to have to hunt Jul 25 06:41:52 is it in the right dir Jul 25 06:41:58 does it go delete them during bb build Jul 25 06:41:58 etc. Jul 25 06:41:59 I think I just need to add an ext "shr/" in the path config Jul 25 06:42:10 i have my own errors to chase here in oe right now Jul 25 06:43:09 u can use a directory path in .bb's as src Jul 25 06:43:16 ie instead of svn:// ... Jul 25 06:43:18 building without local-builds.inc works so it's definately me screwing the paths up Jul 25 06:43:26 u can use /home/me/src/thing Jul 25 06:43:28 for example Jul 25 06:43:48 yup, that's what I'm attempting. I just must've configured it wrong Jul 25 06:59:13 yeay Jul 25 06:59:36 bew! Jul 25 07:00:25 now I just need to find the packaged file ... Jul 25 07:00:46 * raster awaits an updated build Jul 25 07:02:52 waiting sucks. I changed six strings yesterday about this time. Jul 25 07:03:50 once again, I'm amazed by the patience you possess Jul 25 07:03:58 hahahah Jul 25 07:04:02 i dont "Wait" Jul 25 07:04:06 i... do something else Jul 25 07:04:26 invariably i forget about it and come back days later to figure out what it is i wanted to do Jul 25 07:04:27 :) Jul 25 07:04:34 hehe Jul 25 07:04:49 ok. xorg still fucked Jul 25 07:05:41 yeay again! Jul 25 07:05:47 I actually fixed something :] Jul 25 07:06:09 and then a segfault :/ Jul 25 07:06:24 but that's because GSM isn't up Jul 25 07:14:05 * raster grumbles about how much pain it is to add just 1 patch... Jul 25 07:14:35 depends on if you want to test out the code change or not :) Jul 25 07:14:59 unlike some... i actually test my changes :) Jul 25 07:17:18 grr. ok. bigger problems Jul 25 07:17:31 no let's see if my code changes works too (and not just translation stuff) Jul 25 07:18:11 now* Jul 25 07:20:48 hmmm interesting Jul 25 07:21:05 different arm targets futz with xorg-server .pc file differently Jul 25 07:21:06 wtf... Jul 25 07:23:34 same xorg version too Jul 25 07:23:37 1.6.1 Jul 25 08:24:42 PaulFertser:start xterm instead of any WM: wich file? Jul 25 08:25:42 raster: schasch: i highly sauggest going to oe-dev (oe dev git branch).....do you have complete git-adress? Jul 25 08:26:40 that is ALL i use Jul 25 08:26:52 i dont want to use shr git trees or specific oe branches per device Jul 25 08:26:59 i want to build for N targets Jul 25 08:27:31 (gta02, nokia800, palmt650, rokre6, overo, c760) Jul 25 08:28:00 and a new git repo or branch every time i simply want to "spew out image for device X" is silly Jul 25 08:28:02 not going to happen Jul 25 08:28:18 the probvlem is everyoen runs off to do stuff in their own git branch or repo Jul 25 08:28:39 once every how and again things are dumped into oe dev then not tested across multiple targets or distros Jul 25 08:28:50 and invariably i am cleaning up a bunch of these Jul 25 08:29:17 so.. already there :) Jul 25 09:42:47 blu3: oh, i haven't thought of that :) Jul 25 10:02:27 good morning Jul 25 10:11:09 morning ptitjes Jul 25 10:11:27 ptitjes: I saw you name in the contributors of Vala ;-) Jul 25 10:11:43 hi onen|openBmap Jul 25 10:11:51 where ? Jul 25 10:12:37 ptitjes: http://live.gnome.org/Vala/Release -> 0.7.2 Jul 25 10:12:51 ha yes Jul 25 10:12:52 :) Jul 25 10:13:25 onen|openBmap: http://git.gnome.org/cgit/vala/log/?qt=author&q=didier Jul 25 10:13:26 :) Jul 25 10:14:11 ptitjes: ;-) Jul 25 10:14:46 and there are still 20 patches on git.freesmartphone.org that did not yet move upstream Jul 25 10:30:43 TAsn: please... just fix opkg do not remove files... Jul 25 10:30:47 s/do/to/ Jul 25 10:30:47 dos1 meant: TAsn: please... just fix opkg to not remove files... Jul 25 10:31:07 sqlite stuff and others do later :x Jul 25 10:39:21 i just put in the SD and my phone wouldn't start anymore Jul 25 10:39:45 let me tell you that it helps having the battery in the phone Jul 25 10:41:47 dos1, huh? Jul 25 10:42:09 Meomer: ? Jul 25 10:42:10 TAsn: that problem with deleting random files on upgrade/install Jul 25 10:42:25 TAsn: which also causes overall slooowness of opkg Jul 25 10:42:26 I think it's just opkg dying in the middle without memory Jul 25 10:42:37 TAsn: and which was introduced by you Jul 25 10:42:43 TAsn: as i didn't saw it before Jul 25 10:42:43 I think reverting my patch and adding mewster's might help. Jul 25 10:42:47 dos1, aye. Jul 25 10:42:56 I think my patch added more memory usage Jul 25 10:42:59 TAsn: and now i see it everywhere :x Jul 25 10:42:59 I didn't think about that. Jul 25 10:43:00 ;\ Jul 25 10:47:28 whats the problem with opkg? Jul 25 10:48:06 it can't fork :) Jul 25 10:50:01 whats the problem with shr? Jul 25 10:52:32 just a joke Jul 25 10:52:53 i am happy with gsm in the last 1-2 weeks Jul 25 10:53:56 i'm happy with gsm in the last 4 months :P Jul 25 10:54:05 it seems to work properly, no ophonekitd crashes anymore Jul 25 10:54:18 but i didn't upgraded to latest opkg Jul 25 10:54:29 i did Jul 25 10:54:40 (before that i was always upgrading everything, without reflashing) Jul 25 10:55:01 i also did an upgrade Jul 25 10:55:01 TAsn: you did not meet this problem with opkg while testing your patch? Jul 25 10:56:21 TAsn: so... when you'll play with mwester_'s patches? :P Jul 25 10:56:48 onen|openBmap, no, I haven't. Jul 25 10:56:52 though I have now. Jul 25 10:57:04 it seems like it appears on a really big upgrade Jul 25 10:57:10 like all other opkg issues ;\ Jul 25 10:57:30 dos1, mwester_'s patch are well tested Jul 25 10:57:37 as he uses them on slugos for a long time. Jul 25 10:57:38 TAsn: pfff, why do we bother to make tests, as the issues always raise *after* the release! (it happened to me too) Jul 25 10:57:41 Hey there Jul 25 10:57:50 onen|openBmap, :) Jul 25 10:57:54 anyhow, I gtg. Jul 25 10:57:55 ciao. Jul 25 10:58:07 Ainulindale: good morning Jul 25 10:59:40 ohey onen|openBmap o/ Jul 25 11:04:30 larsc: ping Jul 25 11:07:24 PaulFertser: pong Jul 25 11:08:43 larsc: i hope you understand that i'm an occassional hacker, not a developer or kernel maintainer of any kind. I'm (unpleasantly) surprised Nelson didn't commit your bq27000 patch yet. Jul 25 11:09:22 larsc: what is your vision wrt kernel maintaining? Probably you want to take care of OM tree yourself? Jul 25 11:09:38 larsc: you seem to be pretty much capable of that. Jul 25 11:12:42 PaulFertser: ok. i CCed you cause you were the only one i knew who was active commiting to the om tree Jul 25 11:14:23 larsc: of course i can push this patch, it obviously does the right thing. I just waited for Nelson to do it because i assume he's the maintainer. Since we'll have great changes to come soon (soc support coming back from upstream) we need to decide on some working process now. Jul 25 11:15:39 i thought the idea was that there'd be multiple people committing, rather than one "gatekeeper"? Jul 25 11:16:08 PaulFertser: I would like to see a branch based on upstream 2.6.31. Jul 25 11:17:40 Weiss: those multiple people need to agree on some process first, at least wrt in-team communication. And there were no discussion about that. Just several people got commit access in case something (like Nelson going on vacation) happens. Jul 25 11:18:38 larsc: do you want to see it right now or after 2.6.31 release? Jul 25 11:18:55 PaulFertser: soon Jul 25 11:21:20 TAsn: any news ? Jul 25 11:24:05 dunno whether this came through (disconnected...) Jul 25 11:24:05 is there anything missing in your kernel tree? Jul 25 11:24:06 (hardware-support-wise) Jul 25 11:24:07 if not, we might as well declare your tree as the official then Jul 25 11:24:09 what about gta01 support, btw.? Jul 25 11:25:35 does Nelson Castillo do IRC? Jul 25 11:26:15 gismo_: yes, sometimes. He's Arhuaco. Jul 25 11:27:14 PaulFertser: andy-tracking is a mess. There are lots of obsolete files in there, like gta03 support and leftovers from the pcf* refactoring. Jul 25 11:27:23 larsc: i know Jul 25 11:28:13 What I would like to see is that each driver gets it's own branch. And then two master branches one for gta02 and one for gta01 which regularly merge on the drivers they need. Jul 25 11:28:36 makes patch management a lot easier Jul 25 11:28:56 mickey|train: was that question for me? Jul 25 11:29:18 larsc: i think yes Jul 25 11:29:30 PaulFertser: thx, we (= Debian) would like to know if it is possible to have a single kernel for both GTA devices Jul 25 11:29:52 gismo_: a single means single executable or single source tree? Jul 25 11:30:10 larsc: any comment about http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=529078#34 Jul 25 11:30:20 PaulFertser: single executable (even if not optimized) Jul 25 11:31:02 PaulFertser: according to the Linux documentation, the s3c2410 architecture should support both -> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=503292#10 Jul 25 11:33:01 gismo_: sorry, but that is a limitation of the xorg xrandr implementation. If I allow sizes bigger then the actual screen, there are rather unpleasant side effects. Jul 25 11:34:47 gismo_: afaik it is posible to create a kernel which runs on both devices Jul 25 11:34:48 larsc: can you clarify single executable for both gta's question? I lack knowledge about supporting two platforms in the same binary. Jul 25 11:34:56 Ah :) Jul 25 11:37:59 larsc: I do not understand why on Openmoko Trac you claimed the bug has been fixed (which is not) Jul 25 11:38:11 larsc: even with VirtualSize `xrandr --rotate` does not work Jul 25 11:38:48 larsc: about the kernel, do you know if there is any public documentation/howto? I mean, should I try the different combinations until the correct one? :-) Jul 25 11:38:50 mickey|train, so you would prefer to start off ms5.5? Jul 25 11:39:28 mrmoku: too late (mickey|train has left this server) :P Jul 25 11:39:49 dos1, really? Jul 25 11:40:05 no, just kidding ;p Jul 25 11:40:42 mickey|train, wouldn't it be better to base on oedev and merge in the stuff needed from ms5.5 and shr/import? Jul 25 11:40:44 gismo_: I though VirtualSize would fix it. It doesn't. I did some further investigation how to get it supported and came to the conclusion that there is no clean way to get it done. Jul 25 11:40:55 mrmoku: hmm... i was kidding about kidding :x Jul 25 11:41:01 mrmoku: there is only mickey_away :P Jul 25 11:41:32 (my sense of humour is kind of specific, i know ;D) Jul 25 11:41:36 larsc: which means that xf86-video-glamo does not support RandR-2.0, right? Jul 25 11:41:57 larsc: I am fine with that AFAIK the driver does not claim something wrong ;-) Jul 25 11:42:01 gismo_: kernel: I havn't tested it myself yet, since it dont have a gta01, but shouldn't it work if you select support for both devices in the kernel config. Jul 25 11:42:23 dos1, mickey|train tab-completes for me... Jul 25 11:42:58 gismo_: It does. It just doesn't support rotating a display. You can rotate the output though by using -o {left,right,normal,inverted} Jul 25 11:43:20 mrmoku: hm? for me not Jul 25 11:43:35 mrmoku: there is only mickey_away :x Jul 25 11:43:38 larsc: I know about -o (which I reported in the Debian bug) Jul 25 11:44:09 dos1, interesting... but as he does not respond... you might be right :P Jul 25 11:44:19 larsc: [kernel] if I compile the GTA01 stuff into the GTA02 .config, the boot stops at "Starting kernel", I need to cherry-pick the various options Jul 25 11:45:01 hm ok Jul 25 11:47:19 hi F4t, you joined at the same time than me :) Jul 25 11:47:45 hi Mwyann :), so hows the stesting going? Jul 25 11:48:07 wait, I'm just awake :p Jul 25 11:48:14 lol so am i Jul 25 11:48:34 well, the test did show me that I'm suffering a big, very fat bug that's been submitted some time ago Jul 25 11:48:37 tadam, /me is working at regexps in opimd Jul 25 11:48:42 that prevents me to set up my wifi anyway Jul 25 11:48:52 http://docs.openmoko.org/trac/ticket/1939 Jul 25 11:48:52 i was trying to start working on it for about week :DDD Jul 25 11:48:55 this one Jul 25 11:49:11 * dos1 really has to do something with his laziness ;x Jul 25 11:49:12 if it's been fixed somehow, please tell me because it's a pain... Jul 25 11:50:03 (and FYI I flashed shr-shr-image-glibc-ipk--20090721-om-gta02.rootfs.jffs2 and uImage-2.6.28-stable+gitr0e5fe639e234cdeb11d8441f19c5b3109a8b6a17-r2-om-gta02.bin then upgraded, but still nothing) Jul 25 11:50:24 gismo_: -o rotates the whole screen. --output --rotate rotates a single output. As the glamo only has a single output one could think both are equivalent, but they are handled different by randr. Jul 25 11:50:50 larsc: can you sum up your proposal and post it on kernel ML please? Probably with approximate examples of git commands to help git newbies see how cool your proposal is wrt maintaining patches? I like the idea of driver topic branches, it makes submission upstream much easier and clean. Jul 25 11:50:50 mrmoku: what UI for git do you recommend? Jul 25 11:51:26 PaulFertser: sure Jul 25 11:51:58 dos1, the included git gui and gitk commands Jul 25 11:52:07 Mwyann, fully upgraded? opkg update, and then opkg upgrade (that is buggy, and can show errors).. Jul 25 11:52:18 mrmoku: included git gui? Jul 25 11:52:21 yes, with the swapon workaround Jul 25 11:52:49 mrmoku: are you sure "git gui" is something that is really useful? I've looked at it once but couldn't understand why i need that. Jul 25 11:52:50 i definitely agree about driver branches... so much better than the current situation, with more patches than anyone could ever manage Jul 25 11:52:57 it did not showed any error, just asked for one file to be replaced (and nothing to do with wifi) Jul 25 11:53:13 mrmoku: i would like to work to do branch, but i'm too lazy to learn, how to do that from command line ;P Jul 25 11:53:54 Weiss: i think current development model was justified at some point of time, it's just that now the situation is changed so we need to change too. Jul 25 11:55:10 and the trac link does not shows any sort of fixing, but I think it should have been, as the trac has been created 10 months ago. Jul 25 11:55:35 I searched the web a bit, but it was late,so I didn't went that far Jul 25 11:56:49 did you know about such a problem, F4t ? Jul 25 11:57:15 Mwyann, nop, i heard about some problem n gta01 with the wifi, but never heard of this... Jul 25 11:57:15 PaulFertser, wow, it is *very* usefull :) Jul 25 11:57:29 PaulFertser, you can commit only hunks of files or even single lines Jul 25 11:57:42 and nicely see what changed (including new files) Jul 25 11:57:57 mrmoku: i thought git add -i allows the same? Jul 25 11:57:57 dos1, use git gui then ;) Jul 25 11:58:01 Mwyann, run: "opkg list | grep connman" Jul 25 11:58:05 well, I think I'll just have to flash a good-working version... I'm on the unstable branch now, I can go to testing, but I hope it has same functionnalities... Jul 25 11:58:12 Mwyann, and: "opkg list | grep mokonnect" Jul 25 11:58:29 Mwyann, the unstable is the better version anyway Jul 25 11:58:31 PaulFertser, did not try that for a long time... Jul 25 11:58:38 can you commit single lines with it? Jul 25 11:58:47 yeah, that's what I've rea here and there Jul 25 11:59:13 mrmoku: i'm not sure, never needed that Jul 25 11:59:36 mmmh, I'm installing samba right now, so I need to wait it finishes Jul 25 11:59:51 btw opkg doesn't output his errors in stderr... Jul 25 11:59:53 Can anybody show me "grep sysfs /proc/mounts" output for any OE-based distro? Jul 25 12:00:21 sysfs /sys sysfs rw 0 0 Jul 25 12:01:03 mrmoku: strange thing changing chgmode worked for gta01 owners with SHR despite it's 0444... Jul 25 12:01:17 gismo_: --rotate only works if there is enough space available on the screen for the roteted output. which is not the case for 480x640 but for 240x320 --rotate works perfectly fine. Jul 25 12:02:56 * mrmoku wonders if he should spend some time to look at opkg.... Jul 25 12:03:45 mrmoku, carefull not to feed it... Jul 25 12:04:19 mrmoku: i wonder why you still consider using opkg despite everybody agrees that it's utter crap and needs rewriting. Since nobody's rewriting it why don't you just use apt+dpkg for the time being? Jul 25 12:06:05 eolson: ping Jul 25 12:06:18 PaulFertser, thats what I was thinking about... look at it to find out what would be needed for a rewrite :P Jul 25 12:06:19 F4t, now smaba is installed (took it's time...) can I output the opkg list results in your MP ? Jul 25 12:06:21 PM Jul 25 12:06:30 eolson: ls -l /sys/class/i2c-adapter/i2c-0/0-0008/pcf50606-mbc/chgmode Jul 25 12:06:32 eolson: please Jul 25 12:06:40 Mwyann, sure... Jul 25 12:06:45 maybe its better to not even look at it ;) Jul 25 12:06:53 if its long, use pastebin.com Jul 25 12:07:26 mrmoku: why rewrite it at all? Is it urgent? i use apt+dpkg on debian and it's ok. Many people say it's in fact as fast as current opkg. Why just not use it then? Jul 25 12:07:51 well, quite 20 lines Jul 25 12:10:32 wich /etc/init.d/*.script is responsable for starting e17? Jul 25 12:11:32 larsc: but is not RandR-1.0 deprecated now? Jul 25 12:11:52 schasch: xserver-nodm Jul 25 12:13:26 gismo_: the glamo driver implements the xrandr-1.2 hooks Jul 25 12:14:01 IT WORKS! Jul 25 12:14:05 but now eating :D Jul 25 12:14:20 dos1|away, first tell us what works ;) Jul 25 12:14:33 larsc: OK, can I quote you on the Debian BTS, thus I can mark the bug as upstream/wontfix? Jul 25 12:14:43 anyone knows if the wifi wep crashing (http://docs.openmoko.org/trac/ticket/1939) has been fixed, and if yes how to fix it ? Jul 25 12:14:47 gismo_: yes Jul 25 12:15:00 blindcoder:thankX Jul 25 12:15:07 larsc: thx Jul 25 12:23:21 how can I know the hardware revision from a running system? Jul 25 12:27:38 gismo_: a5/a6 difference is shown with cat /proc/cpuinfo Jul 25 12:28:33 mrmoku: regexps in opimd ;) Jul 25 12:28:36 PaulFertser: is there any difference WRT a7? Jul 25 12:28:40 Mwyann: to avoid wifi crashes don't do ifconfig up/down manually at all. Rely on wpa_supplicant, it should work. Jul 25 12:29:26 gismo_: to see a6/a7 you need cat /sys/bus/platform/drivers/neo1973-version/neo1973-version.0/pcb Jul 25 12:29:35 gismo_: for a7 it's 0x101 iirc Jul 25 12:30:13 Mwyann: at least i personally used wep network that way. Jul 25 12:33:16 PaulFertser: thx Jul 25 12:33:25 PaulFertser, using iwconfig it crashes at least Jul 25 12:33:44 if I setup the essid, then the key (or both on the same line) the essid disappears Jul 25 12:33:54 when I try to set it up again, crash Jul 25 12:34:31 Mwyann: are you sure you're using andy-tracking? Jul 25 12:34:43 andy-tracking ? Jul 25 12:34:45 Mwyann: and once more: i didn't use iwconfig, i used wpa_supplicant. Jul 25 12:35:02 Mwyann: yes the kernel branch. What's your kernel version? Jul 25 12:35:15 root@amoko ~ $ uname -a Jul 25 12:35:15 Linux amoko.mwyann.net 2.6.29-rc3 #1 PREEMPT Wed Jul 8 21:33:42 CEST 2009 armv4tl unknown Jul 25 12:35:26 uImage-2.6.28-stable+gitr0e5fe639e234cdeb11d8441f19c5b3109a8b6a17-r2-om-gta02.bin Jul 25 12:35:35 hm no Jul 25 12:35:44 uImage-2.6.29-oe10+gitr119823+d1c828aa7073b356207b967e394defb26ab4da82-r3.5-om-gta02.bin Jul 25 12:35:55 Mwyann: ok, looks like it's it. Try wpa_supplicant without touching iwconfig. Jul 25 12:36:18 ok, gonna try it... Jul 25 12:40:06 well eth0 dissapeared, great... Jul 25 12:40:16 rebooting... Jul 25 12:42:31 PaulFertser: you are right, a6 -> 0x001, a7 -> 0x101 Jul 25 12:43:19 and, as expected : wpa_supplicant -ieth0 -c .wifi/Mwyann ioctl[SIOCSIWENCODEEXT]: Operation not supported and crash Jul 25 12:44:13 anyway, the fact eth0 disappears with no reason is not normal too Jul 25 12:44:16 Mwyann: i can't see why it doesn't crash here and crashes for you... Probably different wep flavours... Jul 25 12:45:07 you're using the same kernel ? Jul 25 12:48:44 Mwyann: andy-tracking HEAD Jul 25 12:49:12 Mwyann, you are aware of the fact that SHR turns off the wifi at boot time, and u should turn it on from shr-settings if u want eth0 to appear Jul 25 12:49:39 well, it appears at boot time, and after a while it disappears Jul 25 12:49:50 if it's just a power down issue then it's okay Jul 25 12:49:54 Mwyann, thats the time it takes frameworkd to come up Jul 25 12:50:12 is it possible to power it up from commandline ? Jul 25 12:50:18 before u try stuff on it, wait for it to disapear, go to shr-settings, and turn it on from connectivy there Jul 25 12:50:32 possible, but not recommended Jul 25 12:50:46 well, then is it possible to NOT shut it down ? Jul 25 12:50:55 at boot time Jul 25 12:51:13 maybe, but what do u care? just wait for full boot, the goto srttings and power it on Jul 25 12:51:51 mrmoku, hi Jul 25 12:51:58 mrmoku, can I pm you? Jul 25 12:52:30 F4t, the fact is, I would need to set up wifi automagically at boot time, unattended Jul 25 12:52:41 Mwyann: oh Jul 25 12:52:50 Mwyann, we are trying to isolate crash problem first.... Jul 25 12:52:51 Mwyann: then it's obviously the problem of FSO turning off wifi Jul 25 12:52:56 or if you're not there maybe Ainulindale is Jul 25 12:53:03 Mwyann: just wrap wpa_supplicant invocation in fsoraw -r WiFi Jul 25 12:53:10 PaulFertser, the punishment is quite hard ;) Jul 25 12:53:42 .. how PaulFertser ? Jul 25 12:54:10 mrmoku: thanks for tip with git gui, i have dos/opimd/regexps branch now ;) Jul 25 12:54:17 What? Jul 25 12:54:58 what what? Jul 25 12:56:08 Mwyann: it's possible, but only for all resources (so bluetooth will be also on on boot) Jul 25 12:56:09 Mwyann: prepend fsoraw -r WiFi to wpa_supplicant command Jul 25 12:56:50 I need to install fsoraw first, I guess... Jul 25 12:57:21 hopefully there's a package:) Jul 25 13:01:38 root@om-gta02 ~ $ fsoraw -r WiFi -- wpa_supplicant -ieth0 -c .wifi/Mwyann Jul 25 13:01:39 Using **pending_return in dbus_connection_send_with_reply_setup() without pending_setup is deprecated and strongly discouraged Jul 25 13:01:39 Using **pending_return in dbus_connection_send_with_reply_setup() without pending_setup is deprecated and strongly discouraged Jul 25 13:01:39 ioctl[SIOCSIWENCODEEXT]: Operation not supported Jul 25 13:01:43 and... crash. Jul 25 13:02:28 I'm gona try toenable wifi in the settings now Jul 25 13:04:36 * dos1 already knows how to implement phone number normalization in opimd :) Jul 25 13:07:15 still crashes. weird. Jul 25 13:09:06 Mwyann, did u wait for it to boot up? (ie wth0 disappeared) Jul 25 13:09:11 wth0=eth Jul 25 13:09:21 yes, I checked iwconfig to see eth0 Jul 25 13:09:24 it was there Jul 25 13:09:40 it shouldnt be there if the phone finished booting Jul 25 13:09:49 I enabled it in the settings Jul 25 13:10:50 *after it dissapeared* Jul 25 13:10:57 freesmartphone.org: 03seba.dos1 07dos/opimd/regexps * rbf79d23bcc6d 10framework/framework/subsystems/opimd/ (pimd_contacts.py pimd_messages.py): Jul 25 13:10:57 freesmartphone.org: opimd: Messages: use regexps instead of difflib Jul 25 13:10:57 freesmartphone.org: Signed-off-by: Sebastian Krzyszkowiak Jul 25 13:10:59 freesmartphone.org: 03seba.dos1 07dos/opimd/regexps * r5813d689af8c 10framework/framework/subsystems/opimd/ (pimd_calls.py pimd_messages.py): Jul 25 13:11:00 freesmartphone.org: opimd: Calls: use regexps instead of difflib Jul 25 13:11:02 freesmartphone.org: Signed-off-by: Sebastian Krzyszkowiak Jul 25 13:11:26 contacts were already commited ;) Jul 25 13:11:27 yes, after Jul 25 13:11:46 Mwyann, well, im out of ideas :) Jul 25 13:11:52 me too Jul 25 13:12:10 I tried to boot with Set console to USB, but I cannot open USB then Jul 25 13:12:38 and now it seems to have enabled cdc_ether Jul 25 13:13:02 I would like to see what are the last kernel words... Jul 25 13:13:08 Mwyann: AFAIK Set console to USB sets only u-boot console to USB, not kernel Jul 25 13:13:25 Mwyann: to last kernel words you should probably use debug board Jul 25 13:14:31 yeah, but it's 69� and I think (and hope) it should be possible to get kernel messages somehow, by USB.... Jul 25 13:15:26 dos1: what's new? 8) Jul 25 13:15:46 Q-Master: multiple values of one field is already implemented Jul 25 13:16:07 Mwyann, theres lindi's ramconsole patch.... u'll have to compile ur own kernel with the patch to see the panic message Jul 25 13:16:13 dos1: really? you're not joking? how? 8) Jul 25 13:16:17 Q-Master: (but field names as "Work phone" are still recommended :P) Jul 25 13:16:32 Q-Master: {'Phone': ['233', '456', '643']} Jul 25 13:16:59 yep, yep yep.... great Jul 25 13:17:08 * Q-Master is drunk Jul 25 13:17:40 Q-Master: to specify types use just {'Home phone': '233', 'Work phone': '456'} Jul 25 13:18:05 � Jul 25 13:18:43