**** BEGIN LOGGING AT Fri May 09 02:59:56 2008 May 09 03:43:05 * * OE Bug 4237 has been RESOLVED (FIXED) by May 09 03:43:07 * * Outdated URI for 2.6.25-rc4 May 09 03:43:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4237 May 09 03:46:13 03Laibsch 07org.oe.dev * rb64b561e... 10/ (1 packages/irssi/files/autofoo.patch): irssi: autofoo.patch sent upstream. partly closes 3828. May 09 03:46:18 03Laibsch 07org.oe.dev * r007bea50... 10/ (2 files in 2 dirs): xorg: w100-autofoo.patch sent upstream May 09 03:46:23 03broonie 07org.oe.dev * r4232c252... 10/ (1 packages/linux/linux-mainstone_2.6.25.bb): May 09 03:46:23 linux: update SRC_URI for linux-mainstone patches. Closes 4237. May 09 03:46:23 * commit by Laibsch May 09 04:09:05 * * OE Bug 240 has been RESOLVED (LATER) by May 09 04:09:07 * * mISDN bitbake May 09 04:09:09 * * http://bugs.openembedded.net/show_bug.cgi?id=240 May 09 04:16:15 apparently Laibsch does not need this thing called sleep... either that or cdbot2 is very slow. May 09 04:17:31 I had some sleep May 09 04:17:44 But indeed I had trouble getting some tonight May 09 04:17:50 So might as well hack a bit ;-) May 09 04:20:15 ;-) May 09 04:22:34 :-) May 09 04:25:05 * * OE Bug 1098 has been RESOLVED (FIXED) by May 09 04:25:07 * * Hostap modules with rawtx patch applied May 09 04:25:09 * * http://bugs.openembedded.net/show_bug.cgi?id=1098 May 09 04:28:42 03Laibsch 07org.oe.dev * r83a7a57c... 10/ (1 packages/mISDN packages/nonworking/mISDN): misdn: moving to nonworking and resolving bug 240 as LATER May 09 04:28:47 03Laibsch 07org.oe.dev * r9baaea2c... 10/ (1 packages/hostap/hostap-modules-rawtx_0.4.7.bb): hostap: add hostap-modules-rawtx (untested). Closes 1098. May 09 05:16:23 rwhitby: wrt bug 822, if you are not using OE for wireless device anymore, does that mean May 09 05:16:41 a) you are not using any custom distro for wireless devices at all May 09 05:16:43 or May 09 05:16:54 b) you are using something other than OE May 09 05:18:53 Laibsch: none of my devices using distros built from IE are being used for wireless. May 09 05:19:04 s/IE/OE/ May 09 05:19:56 I have an FSG3 running SlugOS/BE, but am not currently using it for wireless access. If required, I can test madwifi/hostapd for ixp4xx for 2.6.24.4 and glibc. May 09 05:20:07 !oebug 882 May 09 05:20:09 * * Bug 882, Status: RESOLVED (FIXED), Created: 2006-04-20 03:26 May 09 05:20:11 * * c3c9lbc02(AT)sneakemail.com: Seagate Microdrive not recognized by kernel in 3.5.4.1 May 09 05:20:11 * * http://bugs.openembedded.net/show_bug.cgi?id=882 May 09 05:20:17 !oebug 822 May 09 05:20:18 * * Bug 822, Status: NEW, Created: 2006-04-08 07:57 May 09 05:20:19 * * evanchsa(AT)gmail.com: madwifi package was incomplete May 09 05:20:20 * * http://bugs.openembedded.net/show_bug.cgi?id=822 May 09 05:25:49 Laibsch: BTW, that bug report refers to madwifi-modules, not madwifi-ng-modules. I have only ever used the latter. May 09 05:26:29 if it refers to outdated software, we should either transfer the intention to the newer soft or close as wontfix May 09 05:26:47 I just think that on a lot of these old bugs, decisions should be taken May 09 05:26:54 in or out May 09 05:26:57 agreed. May 09 05:27:09 it keeps the mind clear and focused, too May 09 05:27:14 the hostapd staging part of the patch is already in madwifi-ng May 09 05:28:07 the rest is about building madwifi, not madwifi-ng, and madwifi-ng seems to be building fine in OE. May 09 05:28:14 I say "wontfix" May 09 05:28:25 actually, the one I really wanted to talk about is May 09 05:28:25 !oebug 823 May 09 05:28:25 * * Bug 823, Status: NEW, Created: 2006-04-08 09:11 May 09 05:28:26 ;-) May 09 05:28:26 * * evanchsa(AT)gmail.com: hostap-daemon package updates May 09 05:28:28 * * http://bugs.openembedded.net/show_bug.cgi?id=823 May 09 05:28:35 But those two are closely related May 09 05:28:54 OK, wontfix it is May 09 05:28:56 for 822 May 09 05:31:05 * * OE Bug 823 has been RESOLVED (WONTFIX) by May 09 05:31:07 * * hostap-daemon package updates May 09 05:31:09 * * http://bugs.openembedded.net/show_bug.cgi?id=823 May 09 05:31:45 Laibsch: my suggested fix seems to have been incorporated in hostap-daemon_0.5.10.bb May 09 05:32:45 OK May 09 05:32:46 how do i submit patches? May 09 05:33:05 fenfir: http://bugs.openembedded.net May 09 05:33:11 preferably "mtn diff" May 09 05:37:24 Laibsch: however when running do_configure on hostap-daemon on ubuntu, I get: May 09 05:37:37 [cc]smart: 633: ==: unexpected operator May 09 05:37:49 so that needs to be fixed in hostap-daemon too May 09 05:38:38 what's the replacement for '==' in '[' in dash shell? May 09 05:40:07 aha: '=' instead May 09 05:41:14 which explains why hostapd doesn't work with madwifi May 09 05:47:05 * * OE Bug 822 has been RESOLVED (WONTFIX) by May 09 05:47:07 * * madwifi package was incomplete May 09 05:47:09 * * http://bugs.openembedded.net/show_bug.cgi?id=822 May 09 05:51:42 Hmm - HAS_PCI doesn't get set correctly either in hostap-daemon May 09 05:55:45 morning May 09 06:01:09 03hrw 07org.oe.dev * rf74141b6... 10/ (1 MAINTAINERS): MAINTAINTERS: update my entry May 09 06:12:35 I be back in 15m May 09 06:22:04 * * OE Bug 1030 has been RESOLVED (WONTFIX) by May 09 06:22:07 * * // in path May 09 06:22:08 * * http://bugs.openembedded.net/show_bug.cgi?id=1030 May 09 06:41:49 hi all May 09 06:50:26 switching to target hdd takes time May 09 06:50:42 ~hail labels in fs May 09 06:50:43 * ibot bows down to labels in fs and chants, "I'M NOT WORTHY!!" May 09 06:51:18 LABEL=rootfs500 is easier to get working then remembering does it is /dev/sda1 or maybe /dev/sdb1 or /dev/sdj1 May 09 07:12:30 hi rschuster May 09 07:12:55 hi hrw May 09 07:15:27 * hrw -> shop May 09 07:31:06 * * OE Bug 4239 has been created by spam(AT)campana.vi.it May 09 07:31:08 * * linux-davinci_2.6.x+git-davinci patch outdated for davinci-sffsdr May 09 07:31:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4239 May 09 07:57:14 bonjour May 09 08:13:36 03koen 07org.oe.dev * re09221f0... 10/ (4 files in 3 dirs): linux-omap2 git: bump SRCREV, add patch for beagleboard RTC support and add (disabled) patch that aims to enable the DVI chip May 09 08:19:05 * * OE Bug 4240 has been created by  May 09 08:19:07 * * background in qemu May 09 08:19:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4240 May 09 08:32:36 ~lart Laibsch for random naming of bugs May 09 08:32:37 * ibot gets a hotmal account and SPAMs Laibsch for random naming of bugs May 09 08:32:40 !oebug 4240 May 09 08:32:41 * * Bug 4240, Status: UNCONFIRMED, Created: 2008-05-09 10:18 May 09 08:32:42 * * : OPIE background for qemu machines needs adapting May 09 08:32:43 * * http://bugs.openembedded.net/show_bug.cgi?id=4240 May 09 08:33:24 hrw: You're free to rename May 09 08:33:29 no need to lart May 09 08:33:46 I agree I've had better summaries ;-) May 09 08:33:56 if it would have 'opie' in name I would not even have to look May 09 08:36:00 * Laibsch learns to never include opie in the summary to get hrw to look May 09 08:36:02 ;-) May 09 08:37:21 ;D May 09 08:44:28 hi May 09 08:51:19 03rpurdie 07org.oe.dev * r453d96c1... 10/ (1 classes/sanity.bbclass): (log message trimmed) May 09 08:51:19 sanity bbclass: Add check for problematic /proc/sys/vm/mmap_min_addr values May 09 08:51:19 This patch informs people of strange vm setting before starting a build instead of having qemu choking on it during glibc compile. May 09 08:51:19 Author: rpurdie@openembedded.org May 09 08:51:19 Branch: org.openembedded.dev May 09 08:51:23 Changelog: sanity.bbclass: Add check for problematic /proc/sys/vm/mmap_min_addr values May 09 08:51:25 Date: 2008-05-07T09:39:08 May 09 08:52:47 hi rp May 09 08:52:58 hi woglinde May 09 08:53:41 * RP wonders where that commit got lost... May 09 08:53:52 *g* May 09 08:54:06 hm is there a channel from the autotools guys? May 09 08:55:19 hms savannah should get a sourcecode searchf ast May 09 08:55:26 like sourceforge May 09 08:56:32 ah git searchable May 09 08:58:15 hi all May 09 08:58:23 morning like2wise May 09 09:14:36 gm May 09 09:20:00 hi all, a question maybe a bit off-topic May 09 09:21:09 on a ep93xx board when i plug something on USB the related device file don't appear on /dev ... udev is running and lsusb correctly report the device plugged ... what can be missing ? May 09 09:22:16 gremlin[it]: are you sure the device does not appear with a name defined in one of the udev rules ? May 09 09:22:56 gremlin hm you can run udev with debug May 09 09:23:04 gremlin and see whats happen May 09 09:23:23 gremlin[it]: run udevmonitor and then plug May 09 09:23:43 hrw okay good to know May 09 09:23:52 about sure ... the udev.rule file is quite empty ... and the devices i test was a usb-key (storage) and a usb-serial ... May 09 09:24:42 what i suspect is that udev isn't signaled about the new device ... what i have to check on kernel parameters about that ? May 09 09:25:12 gremlin[it]: run udevmonitor and then plug - it will show reaction May 09 09:25:16 other platform devices seem to be correctly recognized by udev (such as ttyAM0 and so on) May 09 09:25:28 and if no reaction happen ? May 09 09:26:01 then your usb port is broken May 09 09:26:11 or usb driver May 09 09:26:18 no cuase lsusb correctly report the devices May 09 09:26:50 the latteris what i scare ! May 09 09:26:59 so run udevmonitor May 09 09:27:29 i don't have board here ... May 09 09:28:14 i remember i user udevinfo /class/tty/ttyUSB0 ... but /sysfs/class/tty/ttyUSB0 didn't exist ... May 09 09:34:19 03Laibsch 07org.oe.dev * rcda384a5... 10/ (5 files in 4 dirs): anthy: document unapplicable upstream-status for patches May 09 09:40:41 RP: to continue our conversation from April 29th, how is the git server coming along? May 09 09:40:59 Laibsch: I've just not had a chance to look at it :( May 09 09:41:32 * RP had hoped to over the bank holiday weekend but failed :/ May 09 09:42:23 The way I understood it, getting the bitkeeper history back was the only remaining major task May 09 09:42:30 The rest were important details May 09 09:42:33 RP: bank holiday weekend is .uk only, we don't have a clue :-) May 09 09:42:55 * Laibsch suspects it was sometime in the recent past May 09 09:43:06 It was on Monday... May 09 09:43:23 What tools are there for the bk conversion? May 09 09:43:46 Laibsch: There is a cvs importer in git with a bkcvs mode May 09 09:44:03 good May 09 09:44:08 Just found this May 09 09:44:09 http://kerneltrap.org/mailarchive/git/2007/7/9/251095 May 09 09:44:16 so, it looks doable May 09 09:44:27 need a hand? May 09 09:44:35 * Laibsch really wants this off his mind May 09 09:46:20 /root/git-tester.txt May 09 09:46:56 * Laibsch is stunned about the lack of /etc/git May 09 09:47:02 * Laibsch is stunned about the lack of /etc/git* even May 09 09:47:34 you want a global config of a per user usage software ? May 09 09:48:38 My thinking is that there will be some server running May 09 09:48:46 Those things need configuring May 09 09:48:52 That is where I'd expect it May 09 09:48:59 there is config for git-daemon May 09 09:49:03 oki May 09 09:50:03 i'm sure they'll do that when git-daemon could take more that 4 options on cmdline. May 09 09:50:33 RP: Do you already know what to use for commit messages to irc and mailing lists? May 09 09:55:19 Laibsch: I've not looked into commit messages to irc or mailing lists May 09 09:55:37 Laibsch: Let me mail you some stuff... May 09 10:01:48 good morning May 09 10:01:59 hi Dirk! May 09 10:02:06 hey Richard May 09 10:02:21 Laibsch: Most stuff is confined for the git-trial user which uses gitosis May 09 10:02:41 Laibsch: git-daemon is the only server process needed May 09 10:03:03 Dirk: hi May 09 10:03:15 hi Marcin May 09 10:09:01 03rpurdie * r1064 10/ (4 files in 4 dirs): providers.py: No longer weight providers on the basis of a package being "already staged". This leads to builds being non-deterministic. May 09 10:09:18 * RP has been plucking up courage to do that for a while... May 09 10:12:14 RP: Where is the bk repo for OE? May 09 10:12:33 03rpurdie * r1065 10/ (4 files in 4 dirs): runqueue.py: Flush stdout/stderr before forking to fix duplicate console output May 09 10:13:25 Laibsch: http://ewi/tmp/OE/ May 09 10:14:16 hrw: nothing about bitkeeper ther May 09 10:14:17 e May 09 10:14:58 ops - s/OE/cvsexport/ May 09 10:17:41 RP: good move! May 09 10:17:44 RP: I think that weighting explained my undeterministic builds (you ask bitbake to build foo-0.8.2 and it starts doing tasks for both foo-0.8.2 and foo-0.8.0 of which the latter was already there). May 09 10:18:27 like2wise: No, thats a different problem and there is a bug open about that :( May 09 10:20:18 like2wise: If you're interested, see the last emails on the bitbake list May 09 10:20:26 yup, will read. May 09 10:23:27 03rpurdie * r1066 10/ (6 files in 4 dirs): runqueue/taskdata.py: Make sure recrdeps tasks include all inter-task dependencies of a given fn May 09 10:24:45 * like2wise is doing the bisect approach now on a uclibc problem. May 09 10:25:51 RP: git-cvsimport is not suitable? May 09 10:26:04 Does it loose the history? May 09 10:26:26 Just reached that part in your mail ;-) May 09 10:31:01 Laibsch: It the best I've found but it seems to corrupt the history May 09 10:31:21 In what way? May 09 10:31:37 Have you gathered enough information for a bug report? May 09 10:31:44 Laibsch: no, thats the problem May 09 10:31:55 Laibsch: I can't even easily describe it May 09 10:32:24 What did you do after the import to verify that something was corrupted? May 09 10:33:14 There were two issues, one is that commits were being merged in the wrong order, the other is that big rename of .oe files -> .bb files doesn't show well May 09 10:33:24 Laibsch: I viewed it with gitweb and it looked wrong May 09 10:34:12 Is it OK if I add myself as a user to the git install? May 09 10:34:30 I guess, I'd need to copy an ssh key May 09 10:34:31 Laibsch: You mean in gitosis? May 09 10:34:41 yep May 09 10:34:58 Laibsch: With the way it works I'll need to do it - mail me a key May 09 10:35:24 * cyrilRomain finally found why uicmoc4-native compilation is in an infinite loop. It is a Qt implementation bug ... or a gcc 4.3.0 one May 09 10:37:03 RP: for the records, I just setup gitosis and gitweb on git.freesmartphone.org, both are running very good. May 09 10:37:28 i don't know whether gitosis would handle our use case, but it makes hosting git server very easy and secure May 09 10:37:37 mickey|sofa: You should have said, I have some instructions for that... May 09 10:37:52 but yes, gitosis should work well May 09 10:37:56 ya, but i wanted to go the hard way May 09 10:38:01 and it was all pretty easy May 09 10:38:04 :) May 09 10:38:18 so you must be disappointed May 09 10:38:22 :-D May 09 10:38:31 not at all May 09 10:38:50 hey mickey|sofa May 09 10:38:50 * Laibsch adjusts mickey|sofa's irony detector May 09 10:38:52 I'm always pleased if I'm only cut away from coding for one day because of admin stuff May 09 10:38:59 hey dirk! May 09 10:39:05 do13: how's the family? May 09 10:39:12 RP: cgit and gitweb subdomains created. May 09 10:40:02 RP: "grep rolf /root/.ssh/authorized_keys2"@amethyst May 09 10:40:18 mickey|sofa: thanks, I just need to find time to sort them out now :) May 09 10:40:23 mickey|sofa: good, sometimes stressful May 09 10:41:03 mickey|sofa: the first teeths :) May 09 10:41:34 um cranky May 09 10:43:43 Laibsch: done, you should have access now May 09 10:43:48 RP: cool May 09 10:43:52 Argument "7092 has collisions" isn't numeric in addition (+) at /usr/bin/git-cvsimport line 796, line 1. May 09 10:43:56 during conversion May 09 10:44:11 Laibsch: I've not got as far as adding repos yet May 09 10:44:27 I should do so and start a push of zecke's tree May 09 10:44:43 please do May 09 10:45:08 now that I'm actually using git, i can't believe the speed difference May 09 10:45:16 it's as it's not doing anything :D May 09 10:45:37 Yes, its much nicer... May 09 10:48:26 is git the new vcs for oe? May 09 10:49:15 do13: under strong consideration :-) May 09 10:49:16 nobody has officially called the shots May 09 10:49:25 * like2wise_ shoots May 09 10:49:32 But I think it will be a clear case May 09 10:49:35 * like2wise_ calls May 09 10:49:45 * mickey|sofa joins May 09 10:49:45 Please, not ClearCase. May 09 10:49:47 Yes clearcase would be awesome May 09 10:49:50 :-) May 09 10:49:56 rofl May 09 10:50:08 at least we are now having fun with the transition blues May 09 10:51:53 what is the location of the oe git repository if i were to try it out? May 09 10:51:55 do13: yes May 09 10:55:01 kario: nowhere, yet May 09 10:55:20 RP: Are you aware of the location for the bugtracker for git and friends? May 09 10:55:55 Laibsch: I think they use a mailing list May 09 10:56:04 yucky! May 09 10:56:04 its on vger May 09 10:56:12 Now how backward is that? May 09 10:57:47 RP: Did you ever try to inspect that CVS-from-BK-tarball and see if the data displays correctly in CVS? May 09 10:57:48 hhm, we are already looking for the git bugtracker? May 09 10:57:59 Laibsch: yes, it does May 09 10:58:13 Crofton: yes, the conversion seems to have some issues May 09 10:58:24 Maybe somebody already has a solution May 09 11:00:03 How can I stop bitbake from rebuilding a package whenever I try to rebuild it? May 09 11:00:33 I mean, bitbake rebuilds packages even when they're properly and completely built already. May 09 11:00:54 I found some leads in base.bbclass, but nothing seemed to help. May 09 11:02:17 ? May 09 11:02:34 Eh May 09 11:02:41 Say, I build rxvt. May 09 11:02:42 You tell bitbake to rebuild a package and are then disappointed it rebuilds the package? May 09 11:02:46 No May 09 11:02:49 Um May 09 11:02:59 Say, I build rxvt with "bitbake rxvt" May 09 11:03:04 It builds nicely May 09 11:03:48 I type "bitbake rxvt" again, and instead of just reporting that the package is already built (the older behaviour), it rebuilds the package. May 09 11:04:06 It assumes I want it rebuilt just because I called it again. May 09 11:04:17 It makes building "world" a bit harder. May 09 11:04:35 doesn't happen here May 09 11:04:38 rm_work? May 09 11:04:43 Hmm May 09 11:04:46 Could be that May 09 11:05:12 Though the previous behaviour was rm_work as a cleanup operation at the end of the building process. May 09 11:05:23 Just that. May 09 11:05:36 Not an automatic rebuild as per the above. May 09 11:06:36 Its rm_work May 09 11:06:59 We improved it so if you ran things like bitbake rxvt -c compile it would work May 09 11:07:07 looks like we improved it a bit too much though :/ May 09 11:07:11 OK May 09 11:07:31 I'm guessing it's the function in line 34 May 09 11:07:36 addhandler rmwork_stampfixing_eventhandler May 09 11:09:27 yes May 09 11:09:41 you probably want that to ignore build/buildall May 09 11:11:25 thanks a lot dude May 09 11:12:20 B_Lizzard: if you fix it, send a patch :) May 09 11:12:40 push in progress for the git repo... May 09 11:12:41 :) May 09 11:14:28 I'm not sure I can fix this, my python knowledge is a bit lacking... May 09 11:14:45 Should I first post a bug report? May 09 11:20:06 B_Lizzard: I guess so May 09 11:28:08 re May 09 11:32:07 XorA: did you built qtopia successfully with gcc 4.3.0 on x86 ? May 09 11:34:11 XorA: I meant uicmoc4-native May 09 11:38:04 * * OE Bug 4241 has been created by C.Linuz(AT)gmail.com May 09 11:38:06 * * Building an already completed package with "bitbake " automatically cleans and rebuilds. May 09 11:38:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4241 May 09 11:43:46 cyrilRomain: never tried May 09 11:44:07 hi zecke May 09 11:44:35 XorA, zecke: without this patch http://pastebin.com/d73f2ad82 it build endlessly here (when rcc'ing qrc_network.cpp) May 09 11:44:44 XorA, zecke: I'm tempted to say it is a Qt bug, but the implementation is the same in both qt and qtopia >= 4.3.x and 4.4.0 (and probably even before) :-/ May 09 11:45:28 moreover uicmoc4-native.inc unset all CFLAGS, LDFLAGS, etc. so it cannot be du to 'geeky' flags May 09 11:49:53 anyway, it look strange to compare iterator with a const_iterator .. May 09 12:06:16 cyrilRomain: msg tronical about it :) May 09 12:06:52 zecke: yep. I'll do some test outside OE before :) May 09 12:07:18 cyrilRomain: too late, I gave the pastebin url to him, follow up now :) May 09 12:07:41 hey zecke May 09 12:09:38 zecke: hehe. should I know tronical ? Or it was a contraction of 'Trolltech' and 'technical' ? ;-} May 09 12:10:12 hey cyrilRomain will you go at RMLL ? May 09 12:10:23 cyrilRomain: /msg tronical "What you have to say" May 09 12:10:26 * Genesis 'll be pleased to meet some french OE devel May 09 12:10:35 Genesis: yes I think May 09 12:10:41 great :) May 09 12:10:47 cyrilRomain: If I would start with an introduction I would have to talk two days.... May 09 12:10:57 Genesis: it is too near where I leave not to go to RMLL ;) May 09 12:11:09 hehe lucky boy :) May 09 12:11:37 i'll have a conference with a friend on embedded stuff May 09 12:11:50 but i'll not speak about OE since there is already a conference about it May 09 12:12:37 03dirk 07org.oe.dev * re49ebd52... 10/ (3 files in 2 dirs): cairo: upgrade to 1.6.4 May 09 12:12:41 cyrilRomain http://2008.rmll.info/Tech-stories-Aspects-des-projets.html <- i'll assist my friend mathieu deschamps May 09 12:12:42 03dirk 07org.oe.dev * rfe11e603... 10/ (1 packages/openipmi): openipmi: add 2.0.14 a set of tools to access the IPMI subsystem May 09 12:12:46 03dirk 07org.oe.dev * raf90841f... 10/ (3 files in 2 dirs): netperf: upgrade to 2.4.4 May 09 12:15:12 zecke: so yes I should know him ;) May 09 12:15:34 Genesis: cool! May 09 12:16:18 OE should move packages/(obsolete,nonworking) from packages/ May 09 12:16:23 cyrilRomain: tronical is working at Trolltech, if you get him interested in your bug he will take a look. It is something that will interest him May 09 12:16:28 Genesis: do you know who will talk about OE ? May 09 12:16:51 zecke: ok thank you so much! May 09 12:17:38 do13: tnx for the cairo commit, is pixmap up-to-date for it? May 09 12:17:56 cyrilRomain Ahmed Hajji - http://2008.rmll.info/Les-BSP-libres-dans-l-embarque.html May 09 12:18:00 i don't know him May 09 12:18:04 like2wise: 0.10.0 should be ok May 09 12:19:19 Genesis: me neither. Thanks for the info anyway :) May 09 12:20:32 M. Ahmed Haji - Directeur Open Ways - Tunisie May 09 12:20:55 seems to be network focused company May 09 12:21:23 for which domain of application do you use OE cyrilRomain ? May 09 12:21:29 Genesis: yep. There are lots of network related companies in Tunisie .. May 09 12:22:05 Genesis: I use OE for my personal webserver, which is a C3 + a tosa May 09 12:22:21 oki May 09 12:24:32 Genesis: but I hope I could get angstrom with latest qtopia (OpieII ?) on the tosa I always have in my pocket ;) May 09 12:25:28 Genesis: and definitively replace the sharpROM with a new working qtopia-based application stack May 09 12:25:41 i'd prefer gtk based desktop as GPE :) May 09 12:25:57 i hate trolltech stuff :/ May 09 12:26:07 Genesis: heh. As a user I could understand. May 09 12:26:30 Genesis: As a coder, I'm pretty sure you'd prefer Qt over gtk+ ;-p May 09 12:26:33 Genesis: hate never makes sense May 09 12:26:37 i've already integrate GPE as a system ingeeneer on this : http://www.pipublish.com/fr/index.html May 09 12:28:09 ( my first stuff with OE ) May 09 12:29:43 my english is so poor, i feel better with GTK if i could say that May 09 12:29:49 Genesis: nice! May 09 12:30:05 is there a do_prepend_package? or similar? May 09 12:30:32 ScaredyCat: for? May 09 12:30:44 Genesis: will the PiBook be available in public store ? May 09 12:30:45 removid .debug dir May 09 12:30:50 removing May 09 12:31:01 ScaredyCat: fix FILES vars then May 09 12:31:16 cyrilRomain : i worked on this for a prototype company , i've no news since last year, seems to be a dead project May 09 12:31:38 i've change 2 times of work since :) May 09 12:31:41 Genesis: seems they underevaluated the risks then .. :( May 09 12:31:59 03hrw 07org.oe.dev * r6fa9ef4e... 10/ (1 conf/distro/angstrom-2008.1.conf): angstrom-2008.1: prefer bluez-utils for bluez-utils-dbus May 09 12:32:21 my point of view is that is was a bad idea . May 09 12:32:49 bad idea are often marketing success, but all people are not apple or MS. May 09 12:34:46 i'd some fun to make kdrive work with 2 screens :} ugly hack -_- May 09 12:35:43 Genesis: hehe. It has probably been a nice job for you anyway May 09 12:37:14 that lauch my career as linux architect yeap May 09 12:38:03 as i take my holiday to go on RMLL , i hope to make some contact in the buisness. May 09 12:39:33 i dream to found a such company as RP done. May 09 12:40:02 Genesis: OH is hiring May 09 12:40:12 Genesis: but you would need to move to London May 09 12:41:20 hrw when are you moving to London :) May 09 12:41:42 i've a look at that in 7 month May 09 12:41:57 i need to make a longer exp in a company more that 6 month ... May 09 12:42:22 Crofton: ;DDD May 09 12:46:38 "Junior Kernel Developer" <- miam :) May 09 12:49:25 * trickie_ wishes he lived in london sometimes... always exciting jobs May 09 12:50:38 trickie_: .25....? May 09 12:50:48 * Genesis lives in Paris May 09 12:51:20 Genesis: there is direct train to London - you can travel to/from work each day May 09 12:51:36 hrw: this weekend is a long weekend in .nl so i plan on having heaps of time to geek it up! May 09 12:51:44 crappy week this week May 09 12:55:13 hrw: sorry, been long crappy week :( May 09 12:56:00 zecke: http://www.rpsys.net/openzaurus/temp/stamp.patch - is acceptable do you think? May 09 12:56:46 zecke: The idea being we can then call check_stamp externally which is useful for packaged-staging May 09 12:59:21 RP: yes, I can't think of a better way to export the runqueue though May 09 13:03:40 zecke: In due course I may change it to use a stamp class or something and cut down the amount we export... May 09 13:03:44 trickie_: happens May 09 13:04:16 This fixes an immediate problem whilst hopefully not cutting down on our future options though... May 09 13:05:17 RP: well, I assume that we can have more than one runqueue in the future? May 09 13:05:46 zecke: There is no technical reason why not May 09 13:06:11 We could add static variables but IMO that would be worse than this... May 09 13:06:45 yup May 09 13:07:30 morning May 09 13:07:33 RP: we could add _runqueue to the globals (when we add 'd') but I don't know if we have access to it there May 09 13:08:30 zecke: We don't. very little actually sees the runqueue at present May 09 13:08:55 and in a way thats good, I'd much prefer it was wrapped in access funtions rather than used directly May 09 13:09:23 morning mickeyl May 09 13:09:28 RP: nothing to say against it May 09 13:10:28 hi ho May 09 13:10:30 morning May 09 13:11:05 * * OE Bug 4242 has been created by cliff.brake(AT)gmail.com May 09 13:11:07 * * strace does not build with uclibc_svn May 09 13:11:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4242 May 09 13:11:25 conserver seems to work well as a remplacement of minicom to have timestamp and multiple access to tty? May 09 13:11:31 s/?/. May 09 13:11:40 May 09 13:12:50 ( it's far better than a screen+minicom setup ) May 09 13:12:56 03rpurdie * r1067 10/ (4 files in 4 dirs): runqueue.py: Add bb.runqueue.check_stamp_fn() for use by packaged-staging, also improve check_stamp_task debug output May 09 13:13:02 zecke: decision made :) May 09 13:13:30 Genesis: I use picocom mostly May 09 13:13:48 mickeyl: not from sofa today? May 09 13:14:19 no, it's getting uncomfortable after a while May 09 13:14:22 * mickeyl hugs the Herman Miller Aeron May 09 13:14:46 hmmm, people really like those chairs May 09 13:14:55 my cheap chair from office max sucks May 09 13:16:39 Ikea stopped making the chair I use, probably as they break after a while. I've welded it back together though... May 09 13:19:53 I use ikea one May 09 13:20:12 * cbrake_away tried a Aeron for a week and sent it back -- guess people are made different, so there is no one chair that is perfect for everyone. I really like the Steelcase Leap. May 09 13:20:29 ikea also for me May 09 13:21:22 but do not ask for model - I do not remember May 09 13:59:16 * RP wonders how we can cheer zecke|sad up May 09 14:00:47 We now have a happy zecke and a sad one? May 09 14:43:06 * * OE Bug 4243 has been created by elena.valhalla(AT)gmail.com May 09 14:43:08 * * twisted_2.5.0.bb creates empty packages May 09 14:43:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4243 May 09 14:52:37 hey I am getting this May 09 14:52:40 /proc/sys/vm/mmap_min_addr is not 0. This will cause problems with qemu so please fix the value (as root). May 09 14:53:02 on my ubuntu (hardy) May 09 14:53:10 is this check new ? May 09 14:53:58 it's new :) May 09 14:54:03 yes May 09 14:54:23 Commit by rpurdie on org.oe.dev :: r453d96c1... / (1 classes/sanity.bbclass): May 09 14:54:27 hmmm May 09 14:54:45 so I have to now change it manually on my host ? May 09 14:54:47 khem yes May 09 14:55:02 khem yes or via sysctl.conf May 09 14:55:11 is it safe ? May 09 14:55:20 seems so May 09 14:55:51 woglinde: what should I do to make this change permanent May 09 14:56:04 so far I tried overwritin the file May 09 14:56:32 echo "0" >/proc/sys/vm/mmap_min_addr May 09 14:57:43 is something other than 0 also accepted ? May 09 14:57:50 check is for 0 May 09 14:57:58 or does it have to be zero always May 09 14:58:24 I have this as 65536 and did not have qemu issues May 09 14:58:34 so far May 09 15:00:08 how do you force a recipe to generate a machine arch package instead of the default? May 09 15:01:09 must be PACKAGE_ARCH = "${MACHINE_ARCH}" May 09 15:01:42 # protect bottom 64k of memory from mmap to prevent NULL-dereference May 09 15:01:42 # attacks against potential future kernel security vulnerabilities. May 09 15:01:42 # (Added in kernel 2.6.23.) May 09 15:04:09 I give up and modify this to 0 in sysctl.conf May 09 15:07:38 minimal-image.bb does not consider ANGSTROM_EXTRA_INSTALL for IMAGE_INSTALL. Should we add it ? May 09 15:08:00 * Crofton thinks so May 09 15:08:22 also thinks we need to rename that var to IMAGE_EXTRA_INSTALL May 09 15:08:33 but needs to email the dev list about it May 09 15:12:37 bye guys May 09 15:14:24 gm May 09 15:14:26 er gn May 09 15:14:53 khem, email sent May 09 15:14:59 k May 09 15:16:12 I was trying hard to use openssh instead of dropbear. It would get built but not installed. Installed one is always dropbear May 09 15:16:25 tried it on console-image May 09 15:16:41 I tried to set DISTRO_SSH_DAEMON="openssh" May 09 15:16:48 hey khem May 09 15:17:00 Tartarus: moin moin May 09 15:28:22 RP: Have you had a look at tailor? May 09 15:28:31 Laibsch: no May 09 15:29:48 03xora 07org.oe.dev * r1669dd83... 10/ (17 files in 5 dirs): alsa-state.bb : use openmoko-alsa-scenarios.bb from OM mtn May 09 15:29:53 03xora 07org.oe.dev * r471eef0d... 10/ (3 files in 3 dirs): illume_svn.bb : update from OM mtn May 09 15:29:57 03xora 07org.oe.dev * r3607c000... 10/ (1 packages/classpath/classpath-native_0.97.1.bb): classpath-native_0.97.1.bb : add from OM mtn May 09 15:30:02 03xora 07org.oe.dev * r1f573668... 10/ (4 files in 3 dirs): dvb-apps_1.1.1.bb : add from OM mtn May 09 15:30:07 03xora 07org.oe.dev * rf76bdb54... 10/ (5 files in 3 dirs): e-wm_cvs.bb : fixes to e-wm from OM mtn May 09 15:30:12 03xora 07org.oe.dev * r105c1f46... 10/ (1 packages/e17/exquisite_cvs.bb): exquisite_cvs.bb : new recipe from OM mtn May 09 15:30:17 03xora 07org.oe.dev * r4b3c7844... 10/ (5 files in 3 dirs): imlib2_cvs.bb : fixes from OM mtn May 09 15:30:22 03xora 07org.oe.dev * r120b8a58... 10/ (4 files in 2 dirs): May 09 15:30:22 git_1.5.3.bb : add from OM mtn May 09 15:30:22 git.inc : fixes from OM mtn May 09 15:30:27 03xora 07org.oe.dev * ra8ec9b29... 10/ (1 packages/hal/hal-info_20080313.bb): hal-info_20080313.bb : new version from OM mtn May 09 15:30:32 03xora 07org.oe.dev * ra7a011cf... 10/ (4 files in 3 dirs): libnl_1.0-pre8.bb : new recipce from OM mtn May 09 15:30:37 03xora 07org.oe.dev * r80944ae2... 10/ (3 files in 3 dirs): notification-daemon_0.3.7.bb : add a fix for text color from OM mtn. May 09 15:30:41 03xora 07org.oe.dev * rf619709f... 10/ (9 files in 3 dirs): openssl_0.9.7m.bb : new version from OM mtn May 09 15:30:46 03xora 07org.oe.dev * r30d0de3b... 10/ (5 files in 3 dirs): openssl_0.9.8g.bb : new version from OM mtn May 09 15:30:51 03xora 07org.oe.dev * rea4dfda5... 10/ (1 packages/openssl/openssl-native_0.9.7m.bb): openssl-native_0.9.7m.bb : new version from OM mtn May 09 15:30:56 03xora 07org.oe.dev * r9246f347... 10/ (7 files in 3 dirs): packages/openmoko-projects/diversity* : diversity updates from OM mtn May 09 15:31:01 03xora 07org.oe.dev * r139be51d... 10/ (3 files in 3 dirs): illume-theme_svn.bb, illume/keyboard.patch : illume fixes and theme from OM mtn May 09 15:31:06 03xora 07org.oe.dev * rb72cbdbf... 10/ (1 packages/openmoko-projects/enlazar_svn.bb): enlazar_svn.bb : add new from OM mtn May 09 15:31:11 03xora 07org.oe.dev * r40c1c8d6... 10/ (1 packages/openmoko-tools/pty-forward-native.bb): pty-forward-native.bb : add new from OM mtn May 09 15:31:15 03xora 07org.oe.dev * rd5556dd9... 10/ (1 packages/openmoko-tools/serial-forward.bb): serial-forward.bb : add new from OM mtn May 09 15:31:20 03xora 07org.oe.dev * r7c772145... 10/ (1 packages/python/python-etk_git.bb): python-etk_git.bb : add new from OM mtn May 09 15:31:25 03xora 07org.oe.dev * r89645b04... 10/ (8 files in 2 dirs): xorg-xserver/xserver-* : sgml path fix from OM mtn May 09 15:31:30 03xora 07org.oe.dev * r5720426d... 10/ (1 packages/xorg-doc/xorg-doc-common.inc): xorg-doc-common.inc : fix SRC_URI from OM mtn May 09 15:31:34 03xora 07org.oe.dev * ref0c03a9... 10/ (1 packages/tasks/task-openmoko-games.bb): task-openmoko-games.bb : add more games from OM mtn May 09 15:31:39 03xora 07org.oe.dev * rb75fe194... 10/ (1 packages/tasks/task-openmoko-feed.bb): task-openmoko-feed.bb : add more stuff to feeds, from OM mtn May 09 15:31:44 03xora 07org.oe.dev * r8d85ca0f... 10/ (1 packages/python/python-gsmd_svn.bb): python-gsmd_svn.bb : SRCREV doesnt belong here May 09 15:31:49 03xora 07org.oe.dev * rc23a6629... 10/ (1 packages/openmoko2/openmoko-browser2_svn.bb): openmoko-browser2_svn.bb : add fix for rpath from OM mtn May 09 15:31:54 03xora 07org.oe.dev * r47e31130... 10/ (1 packages/openmoko2/openmoko-alsa-scenarios.bb): openmoko-alsa-scenarios.bb : spell Openmoko right, from OM mtn May 09 15:31:59 03xora 07org.oe.dev * rcb4ea4ed... 10/ (1 packages/openmoko-projects/assassin_svn.bb): assassin_svn.bb : fix DEPENDS from OM mtn May 09 15:32:04 03xora 07org.oe.dev * r3f43406b... 10/ (1 packages/networkmanager/networkmanager_svn.bb): networkmanager_svn.bb : an update from Om mtn May 09 15:32:08 03xora 07org.oe.dev * rbaab714d... 10/ (3 files in 2 dirs): networkmanager_0.6.5.bb, networkmanager.inc : an update from OM mtn May 09 15:32:13 03xora 07org.oe.dev * raa1461fa... 10/ (1 packages/efl1/edbus_cvs.bb): edbus_cvs.bb : fixes from OM mtn May 09 15:32:18 03xora 07org.oe.dev * rb34a1eba... 10/ (1 packages/efl1/ecore.inc): ecore.inc : fix DEPENDS from OM mtn May 09 15:32:23 03xora 07org.oe.dev * rb3abeaa1... 10/ (1 packages/e17/exhibit_cvs.bb): exhibit_cvs.bb : cleanups from OM mtn May 09 15:32:27 03xora 07org.oe.dev * r7b8c7186... 10/ (1 packages/e17/examine_cvs.bb): examine_cvs.bb : cleanups from OM mtn May 09 15:32:32 03xora 07org.oe.dev * rc683f9c1... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs.inc : update some revs from OM mtn May 09 15:32:37 03xora 07org.oe.dev * r80ffebba... 10/ (1 conf/distro/include/sane-srcdates.inc): sane-srcdates.inc : some updated dates from OM mtn May 09 15:32:42 03xora 07org.oe.dev * r3b463c49... 10/ (1 conf/distro/include/moko-autorev.inc): moko-autorev.inc : add some new entries from OM mtn May 09 15:32:47 03xora 07org.oe.dev * r85a5e185... 10/ (1 classes/base.bbclass): May 09 15:32:47 base.bbclass : add check from zecke/OM mtn that errors if libname.lai is not May 09 15:32:47 unique. Acked by RP May 09 15:32:52 03xora 07org.oe.dev * rc5233e45... 10/ (1 conf/machine/om-gta02.conf): om-gta02.conf : prefer glamo xserver from OM mtn May 09 15:32:56 03xora 07org.oe.dev * r88288139... 10/ (1 packages/e17/e-wm/applications.menu): e17/e-wm/applications.menu : fix menu from OM mtn May 09 15:33:37 happy merging :) May 09 15:34:26 * Genesis wait after git switch :) May 09 15:34:29 now do I commit the PR changes to make OM/OE totally in sync :-) May 09 15:35:38 XorA: I've done that with Poky before ;-) May 09 15:36:10 * mwester decides to wait an hour or so before he syncs for the next rebuild. May 09 15:39:07 03xora 07org.oe.dev * rf239711f... 10/ (14 files in 7 dirs): packages/* : PR changes from OM mtn to bring two back into sync May 09 15:55:42 RP: ah, now i see the issue with the sortable versions May 09 15:56:32 RP: I would need a local tree, but I don't have one when making up PV May 09 15:56:44 zecke: ah, yes :/ May 09 16:00:31 RP: would abusing the ref format help? May 09 16:00:55 zecke: I'm not sure what you mean? May 09 16:01:15 RP: ignore me :) May 09 16:05:16 re May 09 16:19:33 zecke: Can you give us a hand with converting the bk repo to git? May 09 16:20:00 I tried with very limited success in http://gitweb.openembedded.net/ May 09 16:28:34 Laibsch: hey May 09 16:28:46 ho May 09 16:28:47 Laibsch: what did you use for that? May 09 16:28:55 git-cvsimport May 09 16:29:05 But maybe I got the command line wrong May 09 16:29:16 and apparently cvsps could be more recent, too May 09 16:29:26 It's only a first shot ;-) May 09 16:29:40 Laibsch: didn't RP tried the import? otherwise you can convert koen's hg tree :) May 09 16:31:03 Koen's hg tree is bust too :( May 09 16:31:23 RP: latest cvsps does not compile May 09 16:32:08 Laibsch: I can't help atm. If I get a chance I will look at it again but not now, sorry :( May 09 16:32:15 sure May 09 16:32:33 we have 2.1-2 May 09 16:32:44 latest stable tar ball is 2.1 May 09 16:33:05 latest dev tarball is 2.2b1 May 09 16:33:11 but that does not compile on amethyst May 09 17:07:37 I'm going to try the hildon-input-method with it's keyboard, do you know if anyone already made the bb files for it? May 09 17:20:37 find packages -name *hild* May 09 17:20:45 That's what I'd do to find out May 09 17:26:41 Laibsch: did this but found only hildon related stuff and not the input-method implementation May 09 17:26:54 will try to build it, let's see how it goes May 09 19:20:49 hi everybody May 09 19:21:41 03koen 07org.oe.dev * r8816a193... 10/ (5 files in 3 dirs): linux-omap2 git: add working DVI patch (sent upstream for review), update defconfig May 09 19:23:08 hi thesing May 09 19:23:16 thesing will you be at linuxtag? May 09 19:24:15 probably May 09 19:27:38 i need a python magician... May 09 19:27:41 do_split_packages(d, engines_root, '^lib(.*)\.so$', 'gtk-engine-%s', 'GTK %s theme engine', extra_depends='') May 09 19:28:05 how can we change 'gtk-engine-%s' to be a variable ${Foo} ? May 09 19:28:29 hi steliosk May 09 19:28:41 i.e 'gtk-${foo}-%s' does not work May 09 19:28:56 steliosk why? May 09 19:29:01 woglinde : hi ! May 09 19:29:04 foo = "%s" % bb.data.getVar("BLA", d, True) + "-%s" May 09 19:29:09 it should work in this context May 09 19:29:23 hm May 09 19:29:26 okay May 09 19:29:33 zecke : thanks ! May 09 19:29:36 I thought it was shell stuff May 09 19:29:37 sorry May 09 19:56:08 RP: zecke suggested cvs2svn -> git-svn instead of straight git-cvs May 09 19:56:18 and indeed it seems to be working much better May 09 19:56:37 I guess I should have that BK repo soon and upload it and ping you then May 09 19:56:46 You can then take a look May 09 19:58:24 zecke: I am getting a couple of empty dir warnings May 09 19:59:28 Maybe I should retry but first do "find . -empty -type d -exec touch {}/.emptydir \;" first? May 09 19:59:52 Laibsch: that is fine May 09 20:00:01 the warnings you mean? May 09 20:00:09 That won't create problems? May 09 20:00:44 r2000 of ~30000 May 09 20:20:15 hi mikeyl May 09 20:21:05 mickeyl: whats the reason for not blogging about openmoko? May 09 20:21:31 i was considering to leave May 09 20:21:59 03dirk 07org.oe.dev * rbcb4d050... 10/ (1 packages/openipmi/openipmi_2.0.14.bb): openipmi: forgot the recipe May 09 20:22:04 03dirk 07org.oe.dev * r5bd5703a... 10/ (1 packages/fuse/curlftpfs_0.9.1.bb): fuse: add curlftpfs, filesystem client based on the FTP May 09 20:22:11 you too? whats wrong with it? May 09 20:23:00 hard to say. i thought we were not getting anywhere and the decisions made would not bring us forward to the goal May 09 20:24:16 and now you overcame you doubts? May 09 20:26:38 yes May 09 20:26:58 :) May 09 20:27:33 mickeyl: no idea what you currently work on. suspend/resume handling would be cool to have :) May 09 20:28:09 mhhh really a shame a OSS project will fail ... May 09 20:28:26 zecke: in how far? May 09 20:29:50 * mwester also thinks suspend/resume handling would be cool to have. ;) May 09 20:30:18 zecke: care to help get some rudimentary suspend/resume support in the qtopia image? May 09 20:30:30 (for GTA01, anyway?) May 09 20:31:38 mwester: I fear I mostly care for GTA02 atm, but there is nothing that wouldn't work for GTA01. And we have suspend/resume in the qtopia image May 09 20:31:52 mickeyl: to the end? beyong the limitiations of apm bios, policy, notifications :) May 09 20:31:53 mwester hm the greephone has suspend/resume May 09 20:32:43 There is certain kernel support not present yet, so I am not certain that the suspend/resume in qtopia is really working very well. May 09 20:33:50 zecke: hmm... i don't know what you mean. please be more specific. The odevice API will Suspend() calls to prepare devices for suspend, and there will be a global Suspend() call that just does what you would expect. May 09 20:38:07 mickeyl: might be my ignorance for the deviced. Basicly we exchanged emails on my usecases already (I just remembered) May 09 20:38:32 yeah May 09 20:38:41 you said you want to have the device fall into suspend, when you want to May 09 20:38:48 we can do that May 09 20:38:55 odeviced and usaged will cooperate on that May 09 20:39:06 mickeyl: cancel suspend, make apps (like the media player) inform the (p) manager about what they are doing May 09 20:39:54 mickeyl: e.g. if you press suspend but listen to music it is up to product manager to decide if the device should suspend or just dim the backlight. once the music finished the device should suspend though :) May 09 20:41:02 Has anyone spec'd what the interface to the kernel is -- the one to cause the GSM to be flow-controlled by means of GPIO (so that it will interrupt rather than just toss data into the UART fifo)? May 09 20:41:04 s/product manager/sane defaults or user preferences/, but yes, i agree May 09 20:41:18 (if nobody has, I'm just going to invent one) May 09 20:42:50 sorry, no idea. May 09 20:42:59 * mickeyl only minimal kernel knowledge May 09 20:43:17 mwester: no idea, that is private to the serial driver? May 09 20:43:37 I was thinking it should be in the driver that handles power on/off for the GSM May 09 20:45:05 We basically need for userspace to signal to the kernel that the GSM should be forcibly flow-controlled when the software has completed communications with the GSM preparing for suspend. May 09 20:45:12 And the reverse on resume. May 09 20:45:31 mwester: why do you need userspace for that? May 09 20:45:45 I think it would be nice if there was a way that the user-space software could query to see if the reason for the resume was the GSM, or other. May 09 20:45:56 mwester: power on/off of the GSM? you are utterly confused :) May 09 20:46:21 I may be confusing you with my description, but the need is sound. May 09 20:46:36 No, I am not meaning to power on/off the GSM. May 09 20:46:52 mwester: that is a totally different topic for the gta0x. We don't wake up from flow control/ring but have a separate line May 09 20:46:52 I am merely saying that is a more logical place for this control, in terms of kernel code, than in the serial driver. May 09 20:47:12 Ok, let's step back here. May 09 20:47:20 Everyone is misunderstanding what I'm saying., May 09 20:47:21 mwester: right, you mix up two issues... May 09 20:47:39 mwester: notification of the wakeup reason: sure, send a kevent or such... May 09 20:47:44 Consider this use case: May 09 20:47:51 Phone begins to enter suspend. May 09 20:48:02 software completes comms with GSM to setup to suspend. May 09 20:48:08 Kernel begins suspend. May 09 20:48:11 Incoming call. May 09 20:48:13 GSM wakes May 09 20:48:22 Serial driver not yet suspended, recieves message. May 09 20:48:28 Serial driver suspends. May 09 20:48:33 Device sleeps. May 09 20:48:36 Call lost. May 09 20:49:08 Hence, we need user space to signal kernel to flow-control GSM before suspend begins. May 09 20:50:17 mwester: two things: suspend+resume is < 1s. You get multiple +RING outputs from the modem. So even if we would lose one... BUT May 09 20:50:46 mwester: we don't need a interface for that. You got the apm_bios (or whatever deviced will provide) and can use termios to do the flow control handling May 09 20:51:09 03likewise 07org.oe.dev * r5917d0f1... 10/ (1 site/avr32-common): avr32-common: Add ac_cv_type_error_t=yes for gnash-0.8.2 built with uclibc 0.9.29. May 09 20:51:31 mwester: A driver should remember state when suspending, right? May 09 20:51:39 The serial driver does not. May 09 20:51:55 mwester: ok, then I rest my case :-) May 09 20:52:10 Nor is there anyway to force a flow-control with termios. May 09 20:52:12 like2wise_: well, if one would have the source to a driver :) May 09 20:52:30 no source, no project May 09 20:52:32 ;-) May 09 20:52:42 You can hack up the serial driver if you wish, but that's only arguing *where* the code gets put in, not if its needed or not. May 09 20:52:51 * mwester has already hacked up the serial driver. May 09 20:53:19 * mwester has also observed that it is not sufficient, but understands also that's why it's *OPEN* source. :D May 09 20:54:10 So what code would you add in userspace that would issue the extended termio operation to force flow-control? May 09 20:54:48 mwester: you are utterly confused! May 09 20:54:57 Not on this one, zecke. May 09 20:55:16 But please defend your case. May 09 20:55:29 mwester: grep -rn TIOCM_DTR /usr/include/* May 09 20:55:45 grep -rn TIOCM_CTS /usr/include/* May 09 20:55:51 Yes, yes. May 09 20:56:04 Please stop treating me like an idiot, ok? May 09 20:56:26 Are you claiming, then, that there is code in the userspace that issues these operations when it suspends? May 09 20:56:40 And then you are aware that these operations don't do anything in the current driver? May 09 20:57:06 mwester: at road our qtopia copy did that and the look at uli luckas's mails about this issue at the arm kernel ml May 09 20:58:38 http://ussg.iu.edu/hypermail/linux/kernel/0604.3/0872.html May 09 20:59:19 Ah, those emails. May 09 20:59:51 hms why we need libxml2-pyhton foo May 09 21:00:14 I'm afraid that doesn't fix the problem. It solves part of it. May 09 21:00:44 mwester: i will ask uli to talk to you on tuesday May 09 21:01:00 The UART state is not saved across suspend/resume, unless the low-level debug is turned on in the kernel defconfig, which until recently was the default for gta0x kernels. May 09 21:01:29 Upon resume, the serial drivers reset all the termios info (and flush the fifos), which may well disturb any saved info as well. May 09 21:02:19 mwester: the serial driver is the s3c urart of your gta01? but I think we flush it on s3c2442 as well May 09 21:02:34 I guess I'll not bother with the proposal I was going to send to the om kernel list at this point. :( May 09 21:03:21 why not? May 09 21:03:22 please do May 09 21:03:26 The serial driver would be the combination of serial_core and the s3c24xx device-specific code. May 09 21:05:08 mickeyl: if someone on the project is working on the issue, it seems best to not confuse things by a proposal that might be viewed as a counter-proposal, and not as I envisioned it to be: a proposal to fill a void that was going unadressed. May 09 21:05:26 er, unaddressed, even. May 09 21:05:29 mwester: without any doubt I recognize that you have spent a lot of time and thought and more than anyone inside openmoko May 09 21:06:20 mwester: oh, I just remember these things from ROAD, they are not yet on my radar and appreciate what you are doing. Send a proposal, please, I will get uli to talk to you on tuesday, they should be on 2.6.24 as welll... May 09 21:46:02 some have used ep93xx USB host ??? May 09 22:50:05 * * OE Bug 4213 has been RESOLVED (FIXED) by likewise(AT)gmx.net May 09 22:50:07 * *  May 09 22:50:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4213 May 10 00:11:28 #clug +R *!ninex@do-ninex-v6.cdp.pl **** ENDING LOGGING AT Sat May 10 02:59:57 2008