**** BEGIN LOGGING AT Sun Mar 25 03:01:03 2007 Mar 25 05:01:59 03rwhitby 07org.oe.dev * r921c450e... 10/ (5 files in 4 dirs): gpstk: Open source libraries and applications to facilitate working with GPS data for research and high accuracy uses. Mar 25 07:19:23 morn :) Mar 25 08:24:49 hi alll i have a zaurus 3000 now :) ... Mar 25 08:25:34 gremlin[it]: I bet is faster as your h3600 :) Mar 25 08:26:05 somebody know how to open the powersupply of an compaq ipaq ... Mar 25 08:26:46 it have a pxa270 ... it have to be fast with OE !!! with the original rom as now seem a bit slow ;) Mar 25 08:26:57 in the ipaq or the powerbrick? Mar 25 08:27:29 i need to open the powerbrick ... Mar 25 08:29:08 it seem to be closed by fusision or glue :( Mar 25 08:34:07 gremlin[it]: power packs like that are usually glued or moulded shut and quite difficult to open :/ Mar 25 08:34:31 gremlin[it]: btw, you can charge the Z with your ipaq brick and vice-versa Mar 25 08:34:52 one way I've used is to use a hammer to crack the seal, but it usually ends up cracking the case as well Mar 25 08:38:06 yes yes ... but now i have two (one form the ipaq h3800 and one from zaurus) power supply with US plug shape ... but no one seem to work :( Mar 25 08:59:52 morning Mar 25 09:01:07 gremlin[it]: I'm using my h3600 AC adaptor for my zaurus c760, if you are interested. The AC adaptor packaged with the zaurus was only for 120 V, AFAIR. Mar 25 09:05:29 no i have already the multivoltage one for zaurus ... Mar 25 09:05:57 gremlin[it]: ok Mar 25 09:08:03 Hmm, I think that jffs2 IMAGE_CMD in zaurus-clamshell.conf is buggy. It doesn't pass eraseblock option to mkfs.jffs2. I expect that to be the cause of my husky hunging while booting Mar 25 09:22:36 03koen 07org.oe.dev * r1889b23e... 10/ (1 packages/gpstk/gpstk_1.2.bb): Mar 25 09:22:36 gpstk: add a custom do_install Mar 25 09:22:36 * the autotools based system doesn't build everything Mar 25 09:22:36 * only a static libgpstk is created Mar 25 09:22:36 * somehow strip fails, leaving us with a 9.5MB big gpstk-bin package Mar 25 09:41:17 * sirfred flashing again Mar 25 09:41:45 sirfred: using an image with -e ? Mar 25 09:42:27 koen: Yes, I've changed zaurus-clamshell.conf to pass EXTRA_IMAGECMD to the mkfs.jffs2 command Mar 25 09:42:49 I'm firs trying with the summary image, as 2.6.20 should have support for that. Mar 25 09:43:28 Great. It mounted and found init Mar 25 09:44:02 so zaurus-clamshell.conf really is buggy Mar 25 09:44:25 I felt a little nervous, as with the summary image, the progress dots in the update screen surpassed the 100% mark. Mar 25 09:44:27 koen: I think so. Mar 25 09:44:38 koen: I can fill a bug report and send the patch. Mar 25 09:45:09 First time I got an openzaurus-3.5.5 image up. Mar 25 09:47:03 Curious, it died after calibrate gpe screen Mar 25 09:48:41 Humm, again. The new user screen shows for a second, and it goes black inmediatly. Mar 25 09:52:11 hi all Mar 25 09:53:18 mrdata: hi Mar 25 09:53:28 Hmm, this summary image I've generated is 31Mb size. Perhaps I've overwritten something Mar 25 09:53:46 mr_nice: hi Mar 25 09:54:11 mrdata: do you know if I am able to give a compiled in module an argument on kernel init? Mar 25 09:55:04 * rwhitby notices that openmoko-today seems to be missing a build-time depends on libebook/libecal ... Mar 25 09:55:06 mr_nice: over the kernel command line, i think Mar 25 09:56:02 mrdata: I will try it thx Mar 25 09:56:42 mr_nice: i use this to overwrite the default flash partition schema Mar 25 09:57:42 good ... as i suppose .. keeping the zaurus in japanese help me a lot to practice my japanese ... in half a day i'm re-able to read hiragana and katakana :) ... Mar 25 09:59:11 mrdata: do you know what the max_charge is good for? http://pastebin.ca/408987 Mar 25 10:00:50 rwhitby: eds-dbus Mar 25 10:00:59 yep, testing that now Mar 25 10:03:00 Hmm, it seems I'm broken something with that 31Mb jffs2 image Mar 25 10:05:29 mr_nice: i have no real idea, for what we should use max_charge Mar 25 10:06:04 03rwhitby 07org.oe.dev * ree9cf209... 10/ (1 packages/openmoko-pim/openmoko-today_svn.bb): openmoko-today: Added missing eds-dbus build time dependency. Mar 25 10:06:09 03rwhitby 07org.oe.dev * r5df32e09... 10/ (1 conf/distro/include/openmoko.inc): openmoko.inc: Added PREFERRED_PROVIDER_gtk+ like in angstrom. Mar 25 10:06:24 mr_nice: we have only three values -> vbatt;vcharger;icharger Mar 25 10:06:25 rwhitby: when I added -today it didn't depend on eds-dbus Mar 25 10:06:41 koen: nod Mar 25 10:07:02 the svn dependencies may change fairly rapidly I expect Mar 25 10:07:50 mrdata: also I didn't see any function which uses the CALIBRATE_BATTERY(a) makro? Mar 25 10:07:54 RP: sorry, I should have some more time today... Mar 25 10:10:13 mr__nice: not true; have look of simpad_get_voltage() -> = vbatt = li-ion accu voltage Mar 25 10:10:59 rwhitby: any idea why strip would fail on the gpstk binaries? Mar 25 10:11:33 koen: no idea - I'm still pondering why simply changing a noinst to a lib in automake would cause the compile to fail ... Mar 25 10:11:37 mr_nice: you should also change the macro to CALIBRATE_BATTERY(a) ((((a + 3)*12610)/860) + 170) Mar 25 10:12:02 (as that's the automake way to generate install targets ...) Mar 25 10:12:06 mr_nice: the result values are more possible Mar 25 10:12:23 rwhitby: comparing the jam and autofoo built stuff shows more strange differences Mar 25 10:13:24 koen: because gpstk is a shell script? Mar 25 10:13:40 zecke: no, lots of c++ Mar 25 10:13:58 zecke: parallels coherence mode rocks Mar 25 10:14:36 koen: does it work for linux as well? Mar 25 10:14:52 probably not Mar 25 10:15:00 mrdata: ok for both battery types? Mar 25 10:16:22 koen: currently parallels USB implementation is ... broken Mar 25 10:16:24 mr_nice: yes, measurement of voltage is independently from capacity of the accu Mar 25 10:16:51 koen: to connect a Serial<->USB link I need to plug a USB 1.1 HUB first, then plug an active USB2.0 HUB and then I can connect the serial :) Mar 25 10:38:55 mrdata: can you upload your gpe-image on opensimpad.org? I am not able to create one which is functional. I want to give it to martin on the oesf.org forums. Mar 25 10:40:49 mr_nice: i could, but my gpe_image is in developer state, kernel with debug messages ... Mar 25 10:42:03 mr_nice: i searching the suspend bug, the kernel tell me a Internal error: Oops - bad syscall: 5f5353 [#1] Mar 25 10:42:21 mr_nice: after i do apm --suspend Mar 25 10:43:30 mrdata: yes no problem he knows about that. but mine is even less useful because you can not log in to gpe Mar 25 10:44:18 mr_nice: okay, where should i upload the image Mar 25 10:45:29 mrdata: best to do that on opensimpad.org just log in and use upload. else you can choose anyplace on the web. Mar 25 10:47:15 mr_nice: okay, but it will take some time, must make backup from running image, decompress, build new one Mar 25 10:48:09 mrdata: no problem. it even can take a few days or so. I just don't know when I will be able to create a funktional Image. Mar 25 10:48:19 mr_nice: have a seperate flash kernel partition and no home Mar 25 10:50:20 mrdata: th Mar 25 10:50:26 mrdata: thx Mar 25 10:50:38 mr_nice: okay, the loog time will take the new kernel compile for my developer machine Mar 25 10:51:32 mr_nice: because no mtdcore, mtdblock and mtdram support for now Mar 25 10:53:54 mr_nice: i use sidux, and the mtdram support is not on by default Mar 25 10:54:37 mrdata: you can use the one from oe. Mar 25 10:55:16 mr_nice: thx!, i will look to it Mar 25 10:55:21 It seems that there's a problem with the corgi backlight. But only a second or two after starting a gpe-program. Mar 25 10:55:26 mrdata: or do you mount the jffs2 image and extract the stuff? Mar 25 10:55:41 Is there any kernel sys interface to change the backlight value/state ? Mar 25 10:57:49 mr_nice: yes, i use mtdram 32MB, copy the image to /dev/mtdblock0, mount -t jffs2 /dev/mtdblock0 /mnt/simpad Mar 25 10:59:09 mr_nice: than i have the full access, copy to local system, make modification, build a new image Mar 25 11:00:21 mrdata: there is allway a .tar.gz or tar.bz file this one contains the filesystem as well. or you can copy the tmp deploy rootfs directory Mar 25 11:00:22 mr_nice: split the image in rootfs part and homefs part for flashing from cf-card with bootloader Mar 25 11:00:45 sirfred: sys/class/backlight//brightness Mar 25 11:00:59 Humm, I've catched a gpe-conf task_backlight off Mar 25 11:01:15 It seems it's spammed by gpe-dm, I don't know why. Mar 25 11:01:17 mr_nice: i know, but i have done modification direct on simpad Mar 25 11:01:29 mrdata: ah, ok Mar 25 11:01:34 koen: Thanks. I've runned gpe-conf task_backlight on, and it's right now. Mar 25 11:02:02 The bad thing is that I've NAND restored and flashed again thinking that the machine was hunged. Mar 25 11:02:16 I have to consider buying a serial cable Mar 25 11:05:46 mrdata: build it only if you have time for it. it is not that important. Mar 25 11:08:25 03nail 07org.oe.dev * r84e15076... 10/ (1 packages/haserl packages/haserl/haserl_0.8.0.bb): haserl: Add package haserl Mar 25 11:12:52 mr_nice: i have looked at my modules tree and found that i have done the work on 2007/03/01 Mar 25 11:13:26 mr_nice: so backup and rebuild should not be that problem now Mar 25 11:14:07 zecke: no problem, I was a bt tired last night anyhow. I can probably think clearer today :) Mar 25 11:16:52 RP: Hello. Could I ask you a pair of questions? Mar 25 11:17:13 sirfred: certainly Mar 25 11:18:06 RP: First one is: Is SD/MMC supported for c7x0 on 2.6.20 ? Mar 25 11:18:37 sirfred: It should be although I keep hearing of problems. It works for me :/ Mar 25 11:18:50 RP: Is CONFIG_MMC_PXA=y the option to set it? Mar 25 11:18:56 sirfred: yes Mar 25 11:19:04 RP: It doesn't work for me either. :( Mar 25 11:19:26 RP: Any idea to diagnose it? Mar 25 11:19:38 sirfred: Someone with a machine/card showing the problem is going to need to track down the change in mainline that broke it... Mar 25 11:19:54 RP: Should something be created in /proc or /sys that I should check? Mar 25 11:20:22 RP: It used to work for me, but in 2.6.20 Mar 25 11:21:14 sirfred: I think the only way will be to run through the changes, bisecting them, or look at the changelogs and guess at the changes that might have caused it Mar 25 11:21:26 RP: The other question is what I've tried to explain in bug 2039. Mar 25 11:21:36 RP: ok. I wil try Mar 25 11:21:38 I will try Mar 25 11:21:55 NAiL: Just a hint (or two) after reading your haserl commit. PR = "r0" is the default, no need to specify it. And using the SOURCEFORGE_MIRROR variable instead of hard paths makes fetching more robust. Mar 25 11:22:16 RP: jffs2 images are generated without eraseblock arg for zaurus-clamshells. I wonder if that is correct (it doesn't work for me) Mar 25 11:23:55 sirfred: I wonder if you were using a kernel which didn't have lzo compression support? Mar 25 11:24:56 RP: What's the configuration option to look for ? Mar 25 11:25:53 sirfred: LZO options in the defconfig Mar 25 11:25:54 CONFIG_JFFS2_LZO=y in my .config Mar 25 11:26:06 sirfred: ok, that should work Mar 25 11:26:15 RP: It seems to be set. It's a linux-rp_2.6.20 without modifications. Mar 25 11:26:42 sirfred: Right, I just had to check you were using a modified kernel Mar 25 11:26:43 RP: So, should be legal to build a jffs2 image without specifying the eraseblock ? Mar 25 11:27:15 sirfred: No, we should be able to set that without problems Mar 25 11:27:55 sirfred: Is this .dev or .oz345x? Mar 25 11:28:00 RP: .dev Mar 25 11:30:15 RP: Thanks for your answers. :) Mar 25 11:30:20 I have to leave now. bbl Mar 25 11:32:15 Laibsch, NAiL: and ${PV} in SRC_URI Mar 25 11:34:35 koen: Yes, absolutely. Mar 25 11:34:50 use variables when you can ;-) Mar 25 11:40:45 Laibsch: I specify PR = "r0" because it's easier to bump the PR next time, and it's easier to remember ;) Mar 25 11:42:14 anyway, fixed. Mar 25 11:44:38 03nail 07org.oe.dev * rb113220b... 10/ (1 packages/haserl/haserl_0.8.0.bb): haserl: Fix SRC_URI Mar 25 12:13:37 03nail 07org.oe.dev * r6f4b6bf3... 10/ (1 packages/haserl/haserl_0.8.0.bb): haserl: Correct my SRC_URI-commit Mar 25 12:56:03 mrdata: ping Mar 25 13:07:54 mr_nice: pong Mar 25 13:10:20 mr_nice: sorry, my girlfriend got me work in kitchen Mar 25 13:11:27 mrdata: hehe no problem I am knowing that situation. just ping me if you find some time Mar 25 13:11:50 mr_nice: work was done, i am ready now Mar 25 13:12:42 mrdata: I think I am complete with the battery driver Mar 25 13:13:29 mrdata: here is the source (http://pastebin.ca/409094). It would be nice if you can check if the values and calculation formulas are correct Mar 25 13:13:47 mr_nice: okay, few minutes Mar 25 13:14:40 mrdata: thx Mar 25 13:22:27 mr_nice: the logic in simpad_get_status() is not correct Mar 25 13:24:37 mr_nice: vcharger > battery->min_supply tells you only, is power supply connected or not Mar 25 13:28:10 mrdata: yes, but if it is not charging and power is not connected the battery will discharge? Mar 25 13:31:52 mr_nice: yes Mar 25 13:33:25 mr_nice: not_charging is when power supply connected and icharger< charging_led_label Mar 25 13:34:26 mr_nice: you must also use different charging_led_label values Mar 25 13:35:22 mr_nice: small battery = 12 , big battery = 27 or 28 (i use better 27) Mar 25 13:36:26 mrdata: I am doing that Mar 25 13:36:32 mr_nice: also charging_max_label had different values (120 - 265) Mar 25 13:37:03 mr_nice: but charging_max_label was never be used for calculation Mar 25 13:39:22 mr_nice: stop, it is useable for calculating percentage of accu fill Mar 25 13:40:57 mr_nice: while charging: 100-100*(icharger-CHARGING_LED_LEVEL)/(CHARGING_MAX_LEVEL-CHARGING_LED_LEVEL) Mar 25 13:47:12 mrdata: I have to think a bit about the stuff. where do you want to export the percentage of the accu? Mar 25 13:50:10 mrdata: I could extend http://pastebin.ca/409125 for something like file percentage Mar 25 13:50:51 mr_nice: have a look at http://pastebin.ca/409126 Mar 25 13:55:28 zecke: ping Mar 25 13:56:31 zecke: have you some time for me, to solve a kernel26 problem on simpad while suspending Mar 25 14:39:45 morning Mar 25 14:51:38 wth? Why does uboot have 13mb of tar.bz2 files? Mar 25 15:05:10 k Mar 25 15:14:13 03nail 07org.oe.dev * r37b1c3b6... 10/ (6 files in 3 dirs): uboot: Update linkstation uboot Mar 25 15:26:06 NAiL: because I didn't pay attention while merging the openmoko stuff... Mar 25 15:26:46 ~lart koen for not paying attention Mar 25 15:26:47 * ibot squeezes koen till koen turns blue like papa smurf for not paying attention Mar 25 16:47:28 03rpurdie 07org.oe.dev * rb4f34bb5... 10/ (1 conf/machine/include/zaurus-clamshell.conf): zaurus-clamshell: Fix jffs2 commandline, enable lzo compression Mar 25 16:54:34 It is like summer outside, and I must work :( Mar 25 16:55:25 join the club Mar 25 16:57:21 crap, I need to return a book to the library ... Mar 25 16:57:29 overdue, and I have already renewed Mar 25 17:02:40 it is 21-22C outside now Mar 25 17:12:11 hmmm Mar 25 17:12:27 I need to find time to get beryl/compiz/whatever running in the efika Mar 25 17:12:40 what are these? Mar 25 17:12:48 3d eyecandy things Mar 25 17:12:54 ah Mar 25 17:13:03 I have no graphics card on mine :) Mar 25 17:13:26 I have :) Mar 25 17:14:15 Hi! Mar 25 17:14:24 !oebug 1966 Mar 25 17:14:25 * * Bug 1966, Status: NEW, Created: 2007-03-10 17:59 Mar 25 17:14:26 * * pmiscml(AT)gmail.com: Allow to limit list of binary-generated GLIBC locales Mar 25 17:14:27 * * http://bugs.openembedded.org/show_bug.cgi?id=1966 Mar 25 17:14:40 koen, hrw|gone: I'd like to commit ^, any issues? Mar 25 17:18:25 looks ok to me Mar 25 17:18:43 ok, thanks Mar 25 17:34:19 why did we add the symling from bootstrap-image to Angstrom bootstrap-image ? Mar 25 17:35:20 anyone able to set openembedded on ubuntu edgy ? Mar 25 17:35:26 I'm not really complaining, but I was confused when I found two images for the price of one Mar 25 17:36:13 Crofton: the symlink should always point to the most recently built image Mar 25 17:36:19 ah Mar 25 17:36:21 Crofton: for e.g. lazy people Mar 25 17:36:22 convenient Mar 25 17:36:24 yeah Mar 25 17:36:27 empty_mind: no ubuntu, I've got it on debian etch Mar 25 17:36:39 I just blow away tmp alot Mar 25 17:36:47 summatusmentis, i got it working on debian etch Mar 25 17:37:01 summatusmentis, just cant figure out which packages to install Mar 25 17:37:42 empty_mind: read getting started, that points you to openembedded-essential.deb Mar 25 17:38:38 koen, i included the repository pointed in getting started, but all attempts to do apt-get update to result in repository not found Mar 25 17:39:29 koen, added this line in source.list deb http://www.openembedded.org/dl/ packages/ Mar 25 17:40:24 works over here Mar 25 17:40:32 and worked at the indt office last week Mar 25 17:41:28 koen, works for dapper, but giving error in edgy Mar 25 17:42:01 ah, dapper here Mar 25 17:42:35 * empty_mind goes on to debug further Mar 25 17:48:07 RP: Your changes to zaurus-clamshell.conf seems right to me. Unfortunately, I feel I've reached the flash quota of this week, ;-) Mar 25 17:49:41 RP: Also, my problem with the SD card seem to be limited to an old 64Mb SD card I used for Ximageon testing. I've tried with a newer Kingstom 512Mb card, and it seem to work fine. Also, changes in kernel mmc code from 2.6.17 to 2.6.20 are a lot to spend time trying to fix support for that old SD card. Mar 25 17:49:58 sirfred: ok, please do close the bug when you've tested (or close now if you think they will work) Mar 25 17:50:13 RP: I think it will work. I'm going to close it. Thanks<1 Mar 25 17:50:41 sirfred: It would still be nice to know why those old SD cards stopped working. You could try the 2.6.19 kernel to try and narrow it down Mar 25 17:51:02 ~lart toshiba for having '<' key near right shift Mar 25 17:51:02 * ibot holds toshiba to the floor and spanks him with a cat-o-nine-tails for having '<' key near right shift Mar 25 17:51:20 any hardcore ARM kernel hackers about? Mar 25 17:51:23 RP: I suspect mmc.c could be the source of 'problems' Mar 25 17:51:30 RP: 2.6.19-git ? Mar 25 17:51:50 RP: It seems to be the only linux-rp 2.6.19 version now. Mar 25 17:51:55 sirfred: There was a 2.6.19 although I removed it since I wanted to get 2.6.20 concentrated on - perhaps find an older checkout Mar 25 17:52:24 ok, i have IMAGE_PREPROCESS_COMMAND set in a distro config and IMAGE_PREPROCESS_COMMAND_platform in a class. it seems IMAGE_PREPROCESS_COMMAND_platform now ensures that the stuff in IMAGE_PREPROCESS_COMMAND no longer works Mar 25 17:52:25 Crofton: That would depend ;-) Mar 25 17:52:29 any suggestions? Mar 25 17:52:47 /proc//smaps output seems funny Mar 25 17:53:07 timtimred: You're overridden it so yes, the original wouldn't work Mar 25 17:53:14 Hello all Mar 25 17:53:20 hey likewise Mar 25 17:53:24 hi likewise Mar 25 17:53:48 I get sections with mode ---p Mar 25 17:53:56 Crofton: I've little direct experience with that although I keep meaning to look... Mar 25 17:54:05 right, i thought the behavior was that it inherited IMAGE_PREPROCESS_COMMAND Mar 25 17:54:37 timtimred: No, IMAGE_PREPROCESS_COMMAND_platform is a platform specific override Mar 25 17:54:42 I think I need to send an email to linux-omap list Mar 25 17:54:45 hmmm Mar 25 17:54:52 IMAGE_PREPROCESS_COMMAND += might be more what you want Mar 25 17:55:02 anyone got an arm that is not an omap kernel handy? Mar 25 17:55:10 * koen raises hand Mar 25 17:55:14 * RP too Mar 25 17:55:21 can you do a pmap 1 Mar 25 17:55:29 er, with procps installed ... Mar 25 17:55:35 done Mar 25 17:55:44 ah, that bit is harder for me Mar 25 17:55:52 http://rafb.net/p/HgfufV48.html Mar 25 17:56:00 try cat /proc/1/smaps Mar 25 17:56:09 right, but all the platform classes are inherited so they would all mess each other up Mar 25 17:56:10 and look at the sections mapped to libc Mar 25 17:56:14 * RP 's build environment is in shreds :-( Mar 25 17:56:23 * timtimred prods NAbyss Mar 25 17:56:27 oops NAiL Mar 25 17:56:31 looking at smaps should help me Mar 25 17:56:54 I have three sections r-xp, ---p, and r--p Mar 25 17:57:13 Crofton: http://dominion.kabel.utwente.nl/koen/smaps Mar 25 17:57:32 Crofton: I don't see any ---p, all have r Mar 25 17:57:38 ok Mar 25 17:57:40 Linux fic-gta01 2.6.20-moko8 #1 PREEMPT Mon Mar 5 19:16:16 CET 2007 armv4tl unknown Mar 25 17:57:49 I am beginning to suspect omap kernel specific problem Mar 25 17:58:12 hmmm, koen has a ---p in his libc Mar 25 17:58:27 RP, what kernel? Mar 25 17:58:35 Crofton: 2.6.20 Mar 25 17:58:38 thx, i will go annoy NAiL into fixing it Mar 25 17:58:47 :) Mar 25 17:58:49 vanilla? Mar 25 17:58:54 Crofton: linux-rp ;-) Mar 25 17:59:11 what's the syntax for defining a sub-package postinst script? Mar 25 17:59:17 Crofton: Nothing touching memory mapping stuff Mar 25 17:59:54 ok Mar 25 18:04:16 I wonder if libsdl shouldn't install an sdl-config script into the staging/$target/bin directory Mar 25 18:04:26 Is that a correct approach? Mar 25 18:07:09 RP: Hi, what about kernel-image cleanup patch? ;-) Mar 25 18:09:11 Humm, I see there's already an sdl-config script in staging/$HOST/bin/$TARGET Mar 25 18:09:36 03pfalcon 07org.oe.dev * r1476457b... 10/ (3 files in 3 dirs): (log message trimmed) Mar 25 18:09:36 glibc, local.conf.sample: Introduce and handle new OE variable, GLIBC_GENERATE_LOCALES Mar 25 18:09:36 to limit set of generated binary locales. Mar 25 18:09:36 * If ENABLE_BINARY_LOCALE_GENERATION is set to "1", you can limit locales Mar 25 18:09:36 generated to the list provided by GLIBC_GENERATE_LOCALES. This is huge Mar 25 18:09:36 time-savior for developmental builds. Format: list of locale.encoding pairs Mar 25 18:09:40 with spaces as separators, e.g.: Mar 25 18:20:35 A question about bugs.openembedded.org policy. Mar 25 18:20:55 If I find what I think is an existing bug solution, how should I proceed? Mar 25 18:21:11 Just submit the patch as an attach and leave the bug state unchanged? Mar 25 18:21:19 sirfred: Append it to the bug Mar 25 18:21:30 RP: And nothing else. right? Mar 25 18:21:52 ~lart the linux-omap list for not allowing digitally signed emails Mar 25 18:21:58 RP: Is the bug submitter who changes it to FIXED and VERIFIED latter ? Mar 25 18:22:21 * ibot DoSes the linux-omap list for not allowing digitally signed emails Mar 25 18:22:22 sirfred: it depends, the committer can also change it depending on how confident they are about a fix Mar 25 18:22:33 ~lart ibot for being slow Mar 25 18:22:34 * ibot stuffs crofton into a shiny new tin can and vacuum seals it for being slow Mar 25 18:22:58 RP: Well, I've hit the same problem, I've made some changes and it works for me. Mar 25 18:23:21 RP: My only doubt is that it could, perhaps, break other .bb, as it's a change in a .inc file. Mar 25 18:23:58 sirfred: Right, so just add the patch and maybe highlight it here in case someone with commit access can pick it up Mar 25 18:24:29 RP: ok Mar 25 18:28:25 Is this legal (note the =+)? : PACKAGES =+ "avahi-daemon" Mar 25 18:29:27 likewise: =+ means 'prepend' Mar 25 18:29:56 I've added a possible fix to bug 2000, if someone is interested. Mar 25 18:30:41 pH5: tnx Mar 25 18:31:30 sirfred: 'FIXED' is only valid when the fix has been committed/applied/etc Mar 25 18:31:58 sirfred: I can confirm your proposed fix is correct and should be checked in... Mar 25 18:32:22 RP: Thanks. So we only need someone that want to check it in. Mar 25 18:32:57 There is another bug with scummvm, it seems to only appear when using -O options with gcc/g++ Mar 25 18:33:10 Is there any way I can generate two images automatically? I need a tar.gz image and a jffs2 image... Mar 25 18:33:29 NAiL: yes, IMAGE_FSTYPES = "jffs2 tar.gz" Mar 25 18:33:44 NAiL: you should check that variable name, this might be incorrect. Mar 25 18:34:01 ok, so it handles more than one Mar 25 18:34:03 likewise: It sounds right to me Mar 25 18:34:46 NAiL: Yes, that's how I found out some bugs :-) IMAGE_FSTYPES = "jffs2 squashfs squashfs-lzma ext2" I have here. Mar 25 18:35:23 likewise: Thanks a bunch Mar 25 18:35:45 NAiL: np Mar 25 18:35:48 And now that we're talking about bugs, I have another one related with gtk-webcore: 2038. If you are so kind to take a look at it... :) Mar 25 18:36:02 I have a proposed patch, also Mar 25 18:36:03 koen@bitbake:/data/build/koen/OE/build/tmp/angstrom/deploy/glibc/ipk$ find . -name "*ipk" | wc -l Mar 25 18:36:03 26340 Mar 25 18:36:04 hmmmm Mar 25 18:38:41 koen: does angstrom-console-image build for you? (I'm hitting a dependency on glib-2.0 that cannot be resolved somehow) Mar 25 18:38:59 'evening all Mar 25 18:39:05 likewise: I just built one Mar 25 18:39:50 koen: hmm, I tried to build one for EFIKA twice, but even the clean build bails out. Mar 25 18:40:14 bitbake -s | grep glib-2.0 Mar 25 18:40:14 glib-2.0 0:2.12.10-r0 Mar 25 18:40:14 glib-2.0-native 0:2.12.4-r1 Mar 25 18:40:34 zecke: ping Mar 25 18:42:53 cyrilRomain: bonsoir Mar 25 18:43:39 03likewise 07org.oe.dev * re77f9fae... 10/ (1 conf/bitbake.conf): bitbake.conf: Make IMAGE_CMD_tar build uncompressed archives. Regression fix. Mar 25 18:43:58 koen: You have a good cairo knowledge, haven't you? Mar 25 18:44:13 sirfred: egypt Mar 25 18:44:23 lol Mar 25 18:44:32 likewise: :) Mar 25 18:44:39 likewise: The other cairo, in gtk land, you know? Mar 25 18:44:49 sirfred: yes, been programming in cairo all weekend :-) Mar 25 18:45:01 sirfred: I thought the gtk land was a myth Mar 25 18:45:06 sirfred: the cairographics.org one. Mar 25 18:45:12 likewise: What about the weather ? Mar 25 18:45:27 chouimat: no, gtk land is a fata morgana Mar 25 18:45:27 chouimat::) Mar 25 18:45:45 chouimat: iirc the trolls pilaged it Mar 25 18:46:05 koen cool ;) Mar 25 18:46:11 koen: :) Mar 25 18:46:50 Well, I wonder if the gpe-clock not showing (neither the taskbar applet nor the clock app) could be caused by some cairo issue ? Mar 25 18:47:24 I always relate cairo with that kind of shapes, like curves and so Mar 25 18:47:41 And as the clock applet works when in digital mode... Mar 25 18:47:51 leon@witty:~/sandbox/efika/openembedded$ bitbake -s | grep glib-2.0 Mar 25 18:47:51 glib-2.0-native 2.12.4-r1 Mar 25 18:47:54 * chouimat hates cairo Mar 25 18:48:17 chouimat: why? Mar 25 18:48:29 chouimat: You don't like hot and sand ? Mar 25 18:48:49 likewise: because I hate gtk+ ... it's just my personal tastes Mar 25 18:48:59 it pains me to look at c code that would be prettier with c++ namespaces .... Mar 25 18:49:03 I hardly know it, only a tutorial about making a clock widget. Mar 25 18:49:03 sirfred: no I'm fat and ugly :) Mar 25 18:49:10 likewise: it seems bitbake doesn't pick up your glib-2.0 recipes Mar 25 18:49:11 chouimat: :) Mar 25 18:49:42 It seems that I'm using 1.4.2, perhaps trying with an older release... Mar 25 18:50:12 are there developers here who use gentoo as base system? Mar 25 18:50:31 koen: looking into it Mar 25 18:50:34 keesj: yep Mar 25 18:50:40 * chouimat hides Mar 25 18:51:20 * likewise points at his colleague and then notices I am not at the office. Mar 25 18:51:31 Humm, it seems that openzaurus-3.5.5.conf wants cairo 1.3.12, and that version is not available now. So it goes for the higher version. Mar 25 18:51:57 sirfred: you'd want 1.4.x anyway Mar 25 18:52:02 1.3.x is the unstable series Mar 25 18:52:28 cyrilRomain and for what platform are you programming (if I may ask)? Mar 25 18:52:31 03koen 07org.oe.dev * r33dc2a58... 10/ (1 packages/matchbox2/matchbox-panel-2_svn.bb): matchbox-panel2: don't make the applets replace mb-panel, that breaks image generation if you build both Mar 25 18:52:42 koen: I will ask hrw or XorA if there's any reason to lock cairo to that version in openzaurus-3.5.5 Mar 25 18:52:52 keesj: tosa and i686 Mar 25 18:53:13 koen: But I'm going to try with 1.3.14 to see if the problem is still present. Thanks. Mar 25 18:53:23 likewise: how much power does that h264 thingy consume? Mar 25 18:54:22 koen: 15 Watts per chip, so at least 8*15. Probably 200+ though due to lower efficiencies in all parts. Mar 25 18:54:49 likewise: so the same as an average servefr Mar 25 18:54:52 server* Mar 25 18:55:57 koen: yes. in fact, there is competition that actually uses a server platform for real-time HD encoding. It's just that broadcasters often do not like PC based solutions for 24/7 operation. Mar 25 18:56:36 koen: but I would not be scared away by it. Use a redundant PSU and you're off. Mar 25 18:58:51 cyrilRomain: I have so for not much luck in setting up oe on gentoo . perhaps I can first try the tosa configuration Mar 25 18:59:41 keesj: what is the problem ? Mar 25 19:00:12 the gcc won't compile Mar 25 19:00:57 gcc ? Do you have updated metadata ? Mar 25 19:01:15 this is what I get http://box.mmapps.net/~keesj/test.txt Mar 25 19:01:15 chouimat|away: cairo is quite usable apart from GTK+, for what it's worth. Mar 25 19:02:03 cworth: when I need a gui I use Qt4 so I don't need cairo :) Mar 25 19:02:06 cyrilRomain Yes I have tried just about everything I could think off , even swichted to different gcc hosts Mar 25 19:02:29 chouimat|away: yup, I'm using cairo without GUI and I can concur with cworth. Mar 25 19:02:32 keesj: maybe you just discovered a bug Mar 25 19:02:58 have a ARM kernel developer some time for me, help to find the reason for a kernel-2.6.20 Oops - bad syscall problem Mar 25 19:03:00 chouimat|away: Yeah, the trolls decided to re-invent cairo instead of using it. Not much I can do about that of course. Mar 25 19:03:08 * cyrilRomain wonders why he didn't encountered it.. Mar 25 19:03:13 cworth: We plan to demo a HDTV quad split system that has Cairo overlaid graphics at NAB2007 Las Vegas. Mar 25 19:03:26 likewise: fancy Mar 25 19:04:13 cworth: I once contacted you about it being my graduate project, but I could not get a coach at the uni for it. So I decided to work on it afterwards. Mar 25 19:04:24 cworth: their stuff was started before cairo iirc but if you prefers it it's find for me I just find it's API ugly so I don't use it Mar 25 19:04:25 I would love to , this error message has been there before. the home page also mentions CROSS COMPILE Badness as source of all miserry Mar 25 19:05:27 keesj: as the message say, it tries to include a header from your gentoo instead of including from the staging directory Mar 25 19:05:31 chouimat|away: you are talking to the architect (cworth) so you might want to explain why you think it is ugly :-) Mar 25 19:06:06 likewise: I kind of guessed it ... like I said before it a matter of personal tastes ... Mar 25 19:06:50 chouimat|away: of course, and mostly a matter of language low/high-level-ness also, I think. Mar 25 19:07:18 likewise: it a matter of feeling confortable using it ... Mar 25 19:07:43 I would be happy to try a different distro if it might help Mar 25 19:08:43 likewise: also I spend more time doing desktop apps that connect to the devices remotly so I rarely need graphics capabilities ... Mar 25 19:09:10 keesj: you can run ./configure --help in the gcc-cross-initial directory. It seems a wrong path is determined by configure script, and hopefully a configure option is enough to get rid of the cross compile badness error Mar 25 19:09:35 chouimat|away: For that purpose, there is cairo for java. (OK, I will shut up now). Mar 25 19:09:43 keesj: what distro are you trying to compile ? Mar 25 19:10:10 likewise: I prefers using Qt ... too much investment in this :) Mar 25 19:12:16 cyrilRomain I am trying generic on with machine set to smdk2440 Mar 25 19:14:44 keesj: you could try angstrom, but as smdk2440 is armv4t, I'm not sure it is a good idea... Mar 25 19:14:51 cyrilRomain I also tried to just move the /usr/include away . that I how deperate I am :p Mar 25 19:15:07 keesj: don't be deseperate Mar 25 19:15:34 cross compile badness is error that even me can fix, so I guess you can :) Mar 25 19:16:55 keesj: jump into /home/keesj/oe/tmp/work/armv4t-linux/gcc-cross-initial-4.1.1-r11/ and take a look at the output of the configure Mar 25 19:20:16 keesj: if you don't plan to compile java, you can try to disable gcj (which include libffi apparently) Mar 25 19:22:04 that is a great tip. http://box.mmapps.net/~keesj/temp/ here are my logs I will try a bit longer and oterwize try to compile for angstrom-2007.1 Mar 25 19:27:11 keesj: looking at your logs, a regexp fails Mar 25 19:27:22 in http://box.mmapps.net/~keesj/temp/log.do_compile.28666 Mar 25 19:27:58 I guess this can be the reason Mar 25 19:28:32 I added --disable-languages=java in gcc-cross-initial_4.1.1.bb . I hope it works! Mar 25 19:29:28 keesj: well it might, but the question is why the regexp failed ? Mar 25 19:30:01 the MPFR thing? Mar 25 19:30:41 MPFR ? Mar 25 19:32:00 keesj: which sed version do you use on your host? Mar 25 19:32:44 hmm what "sed -n 's/^#define \([^_][a-zA-Z0-9_]*\).*/\1/p ;" is trying to do ? Mar 25 19:33:03 removing '#define' ? Mar 25 19:33:29 cyrilRomain: I think the expression continues on the next line, until the end quote. Mar 25 19:33:39 likewise: yes Mar 25 19:34:14 GNU sed version 4.1.5 Mar 25 19:34:53 hmm same version here Mar 25 19:35:36 sorry I was reffering to the configure output MPFR http://box.mmapps.net/~keesj/temp/log.do_configure.27466 Mar 25 19:35:38 same here (Ubuntu Feisty Beta something) Mar 25 19:36:37 likewise makes /me happy :p Mar 25 19:37:18 * likewise is confused. why? Mar 25 19:37:46 You don't have the same error message ? Mar 25 19:38:00 keesj: no :-) Mar 25 19:38:45 keesj: same sed version Mar 25 19:39:34 This seems related to the make version: http://mail.cs.uu.nl/pipermail/stratego/2006q4/000658.html Mar 25 19:39:47 keesj: I can't find where does this sed comes from. At least not in oe metadata. Mar 25 19:40:32 GNU Make 3.81 here Mar 25 19:41:00 GNU Make 3.81 | gcc (GCC) 4.1.1 (Gentoo 4.1.1-r3) Mar 25 19:41:04 Here is a supposed fix. The ";" is split over multiple -e(xpressions) http://mail.cs.uu.nl/pipermail/stratego/2006q4/000657.html Mar 25 19:41:25 make 3.81 here as well... but the fix approach is worth a try Mar 25 19:43:37 where do you see this sed command? Mar 25 19:43:53 keesj: in the Makefile Mar 25 19:44:39 keesj: in your /home/keesj/oe/tmp/work/armv4t-linux/gcc-cross-initial-4.1.1-r11/gcc-4.1.1/build.i686-linux.arm-linux/gcc directory Mar 25 19:48:57 thanks for help Mar 25 19:49:54 I'm trying to build the angstrom-console-image and it fails while configuring glib-2.0 Mar 25 19:50:27 i checked the config.log but the information is overwhelming Mar 25 19:50:51 it is for hx2000 Mar 25 19:52:32 http://pallid.bsnet.se/~dalurka/oe/ Mar 25 20:07:02 dalurka: __cplusplus is not defined Mar 25 20:07:17 the question is why? Mar 25 20:07:25 its a clean build also Mar 25 20:07:42 dalurka: isn't it obvious? Mar 25 20:07:53 zecke: what? Mar 25 20:08:16 "checking for growing stack pointer... configure: error: cannot run test program while cross compiling" Mar 25 20:08:34 error == something fundamentally wrong Mar 25 20:08:46 "cannot run test program while cross compiling" Mar 25 20:08:49 <- the reason Mar 25 20:08:58 Ah you can't run tests when you cross compile Mar 25 20:09:10 now check the Autoconf manual on how to work around such issues... Mar 25 20:09:59 will look in to it but thats for tomorrow Mar 25 20:10:24 bed time over here Mar 25 20:11:05 zecke, have a config.site or some sort of cache somewhere where you can stash values? Mar 25 20:11:46 JoeSchmo: we have. Add the test itself to our m4 directory and uodate the sites file Mar 25 20:12:05 JoeSchmo: http://www.openembedded.org/filebrowser/org.openembedded.dev/site Mar 25 20:12:43 useful Mar 25 20:15:25 nite Mar 25 20:21:54 argh : NOTE: Couldn't find shared library provider for libc.so.6 Mar 25 20:28:47 Hi all... I'm certain this question has been asked quite often already (but neither google nor your documentation provided me with much insight): how do I get oe on x86 to support older kernels then 2.6.20? With 2.6.16, I get "Fatal: kernel too old" on attempting to boot. Mar 25 20:29:48 I guess it would work by setting the appropriate preferred_provider, but the documentation is not exactly transparent as to what I'd have to put in there. Mar 25 20:39:02 florian, hi Mar 25 20:39:03 hmm, even after a build from scratch, I get "Couldn't find shared library provider for libc.so.6" in gcc-cross-4.1.1 do_package ... Mar 25 20:39:27 hi meister Mar 25 20:39:38 argh Mar 25 20:39:39 does this mean the package will be broken ? Mar 25 20:39:54 hi florian Mar 25 20:40:03 cyrilRomain: i don't think so Mar 25 20:40:05 florian, do you have some news about the netbook pro stuff ? :) Mar 25 20:40:10 meister: oops Mar 25 20:40:19 * florian resumes his laptop :-} Mar 25 20:47:29 cyrilRomain: I am having similar issues with glib-2.0 Mar 25 20:47:49 bitbake -s | grep glib-2.0 did not list the cross-compile packages for me Mar 25 20:52:03 likewise: hmm no idea ... I guess you already updated to bitbake 1.8.1 Mar 25 20:52:05 florian: hmmm, netbook stuff. Sounds interesting... Mar 25 20:52:20 As I do not seem to have the kernel tools... Mar 25 20:57:24 something that is not obvious to me: why using qemu for glibc binary local generation ? Mar 25 20:59:43 Can somebody spot what might be happening here: http://www.pastebin.ca/409640 Mar 25 21:00:18 Somehow, bitbake 'skips' the package I want to build Mar 25 21:01:12 cyrilRomain: in oe IRC log, hvontres|poodle wrote "If I understand this correctly, the glibc locale information needs to be compiled into binary form. This can be done on the target(ususlly slow) or in qemu" Mar 25 21:04:16 gotta go, good luck in further bug hunting Mar 25 21:11:19 Good night. Mar 25 21:14:17 florian, many thanks, i will be happy to start hacking nb pro ;) Mar 25 21:14:37 i lacked a documentation about the boot process (: Mar 25 21:15:23 i have 9 nbpro on a shelve :) Mar 25 21:15:56 meister: its a start written from memory... but i'll try to answer questions. Mar 25 21:16:04 ok Mar 25 21:16:28 meister: uh, cool.... ok that's worth spending some time with them :-) Mar 25 21:16:45 it should be possible to back up the original flash Mar 25 21:20:25 Does OE work fully on FreeBSD, I see there's a section in OEAndYourDistro, but it didn't seem too detailed Mar 25 21:23:47 florian, "NETBOOK PRO-404 OS Image - English language version" Mar 25 21:23:53 on http://www.psionteklogix.com/public.aspx?s=fr&p=SoftwareDownloads Mar 25 21:24:05 i will have a try with this image Mar 25 21:24:37 meister: ok, i'll dig out the necessary kernel image and scripts tomorrow Mar 25 21:24:52 florian, ok thx Mar 25 21:25:14 SonicvanaJr: I see no reason why it shouldn't work, as soon as you have the required software Mar 25 21:25:16 florian, what about this : http://linuxtogo.org/~florian/index.php?path=netbook/ Mar 25 21:25:26 cyrilRomain: thanks Mar 25 21:28:05 nite all Mar 25 21:28:43 meister: the public part - the 3.5.4.1 image is one of the files you want. the others are old stuff, some extra packages and some screenshots Mar 25 21:29:09 * florian just starts to appreciate his workstation Mar 25 21:29:44 ok ;) Mar 25 23:32:18 urs, sounds like the machine you are building for has OLDEST_KERNEL set to something newer than what you are trying to build. Mar 25 23:32:57 v8jlene: no, same symptom if I set OLDEST_KERNEL to the appropriate kernel version. Mar 25 23:34:39 urs: Ok, no real idea then. 2.6.17 is the oldest kernel I build at the moment (machine = wrap). Mar 25 23:36:46 it does work with an older image that I found lying around here... I'll try to find out which change caused it to break. Mar 25 23:37:31 With the older version, some other packages are broken though, so I'll see if I can find a version where both things work. Mar 26 01:17:52 http://pastebin.ca/409995 < Anyone know why bitbake is crapping itself ? Mar 26 01:19:10 SonicvanaJr, I lopked, but have no idea what is happeing Mar 26 01:19:24 Thanks Mar 26 01:40:32 SonicvanaJr: the code says: you don't have the tool called md5sum in your PATH Mar 26 01:59:37 Henryk: Thanks Mar 26 02:25:05 is there a bitbake package for ubuntu ? or should i use debian repository? Mar 26 02:58:36 night **** ENDING LOGGING AT Mon Mar 26 02:59:57 2007