**** BEGIN LOGGING AT Wed Apr 08 02:59:57 2009 Apr 08 03:00:14 kergoth: But unfortunately I can't get anywhere since everything depends on that package, it's the first one that builds heh Apr 08 03:00:21 mekius: the only thing i can think of is environment. bitbake sanitizes its environment, and emits the vars from its env into the output, so there could be vars set in your current env that bitbake removes Apr 08 03:00:32 kergoth: hmm Apr 08 03:00:42 i don't see how that would cause that behavior, but... Apr 08 03:00:50 BBPATH would be the only one I can think of Apr 08 03:01:12 bitbake doesn't remove that Apr 08 03:01:20 it removes all but a certain set of whitelisted env vars Apr 08 03:02:27 is this shasum-native issue related to ccache? Apr 08 03:02:44 It is using ccache Apr 08 03:02:45 * kergoth wonders why we don't use hashlib in python or something Apr 08 03:02:47 I remember someone mentioning it before... Apr 08 03:02:56 but running the run.* scripts always works Apr 08 03:03:14 btw, how does it determine whether to use ccache? does it just use it if it's installed? Apr 08 03:03:30 yep Apr 08 03:03:38 * kergoth thinks that "feature" should go away Apr 08 03:04:03 ah Apr 08 03:04:08 yeah that's not very good heh Apr 08 03:04:11 any way to force it? Apr 08 03:04:21 set CCACHE = "" Apr 08 03:04:23 to disable it Apr 08 03:04:24 is there CCACHE="" Apr 08 03:04:25 but not sure that's the problem Apr 08 03:04:30 that var is referenced in the cc vars Apr 08 03:04:32 k, will give that a shot Apr 08 03:04:37 ccache is so nice though :) Apr 08 03:04:47 set that in local.conf? Apr 08 03:04:51 arguable. actual compilation is sometimes not as big a chunk of a build as you'd expect Apr 08 03:04:56 particularly with fetching, unpacking, patching, .. Apr 08 03:04:58 :) Apr 08 03:05:01 well sure Apr 08 03:05:14 someone was saying it is specific to some distro's ccache... Apr 08 03:05:15 but making the build as fast as possible is good when building thousands of packages :P Apr 08 03:05:20 people always ask that, "set it in local.conf?" Apr 08 03:05:23 denix: Archlinux? Apr 08 03:05:23 it doesn't matter where you put it Apr 08 03:05:27 oh ok Apr 08 03:05:27 as long as it gets parsed after bitbake.conf Apr 08 03:05:33 is it redhat based? Apr 08 03:05:34 which means it can be in any .conf that bitbake.conf loads Apr 08 03:05:38 distro, machine, local, site, auto, .. Apr 08 03:05:39 kergoth: sry, I'm new :D Apr 08 03:05:43 no worries Apr 08 03:05:49 I just got a palm centro so playing with it :) Apr 08 03:05:49 just shows that we need to improve our documentation Apr 08 03:05:55 denix: nup Apr 08 03:06:10 kergoth: I'll admit I haven't gotten through all of it Apr 08 03:06:20 Got stuck going through the getting started guide cause of this bug Apr 08 03:06:26 kergoth: nobody reads the documentation... :) Apr 08 03:06:35 * virk|sleep needs to knock his head against the wall or something, forgot to add a modified file to a commit Apr 08 03:06:40 Well I plan on getting to it once I get something built :D Apr 08 03:06:50 I'm sure random bug that makes no sense isn't covered though ;) Apr 08 03:07:01 bitbake finds conf/bitbake.conf in BBPATH and loads it. bitbake.conf loads in all the other .conf files (you can look through it and see the require/include lines). then it finds the recipes in BBFILES and parses them all in, produces a runqueue from the task/recipe deps, and runs the runqueue Apr 08 03:07:02 mekius: I didn't mean you, just in general, myself included :) Apr 08 03:07:08 denix: I know :) Apr 08 03:08:23 virk|sleep: don't operate heavy machinery while drowsy... :) Apr 08 03:09:05 don't even know what 'drowsy' means Apr 08 03:09:22 sleepy, tired Apr 08 03:09:38 but I guess it has something to do with it being 5am already here Apr 08 03:10:36 virk|sleep: wow, drop it then :) Apr 08 03:11:07 I mean until next day... Apr 08 03:12:11 yeah, I was thinking to commit this one for the 3th time, but this time for good, but I guess it'll indeed be for tomorrow Apr 08 03:12:26 alright, trying with ccache="" Apr 08 03:13:46 hrmph, i think I'm just gonna merge the collections bits, don't feel like waiting for more responses, given the positive feedback from rp, mickey and others Apr 08 03:15:20 ok, I'm really of to bed now Apr 08 03:15:25 kergoth: you've got my vote! Apr 08 03:15:30 virk|sleep: bye Apr 08 03:15:45 denix: thanks for the help Apr 08 03:15:57 * virk|sleep is zzZZzz Apr 08 03:17:00 kergoth: I promised I'd test it, but got stuck with that curl-sdk issue - now catching up on emails, stable branch testing and so on... Apr 08 03:17:13 * kergoth nods, know how that goes Apr 08 03:18:37 I guess I should go home too... Apr 08 03:19:15 * denix starts the clean build before leaving Apr 08 03:19:41 otavio: ping me when you have any results Apr 08 03:22:20 * denix|away is away: I'm not here Apr 08 03:24:09 kergoth: gah, that was it Apr 08 03:24:12 ccache ftl :/ Apr 08 03:24:24 haha, thats surprising Apr 08 03:24:28 what a weird bug Apr 08 03:24:47 yeah, not sure why it would matter unless the "ccache gcc" vs "gcc" is throwing it off somehow Apr 08 03:28:54 well, ccache can be buggy :( Apr 08 03:31:26 thats why it shouldnt be default. it should be easy to enable, yes, but not by default Apr 08 03:33:08 Tartarus: it works fine when I run the script myself Apr 08 03:33:16 But when it's run from the python it fails Apr 08 03:33:21 Sounds like a python problem to me :D Apr 08 03:34:56 I use ccache all the time, never had a problem until now Apr 08 03:35:59 ccache is a useful tool, but there's always a slight danger of cache corruption. not likely, but it can happen. still, useful for day to day builds, just not for the weeklies :) Apr 08 03:36:24 meh, I build EFL/E with it, spans of weeks between, no issues :) Apr 08 03:36:30 maybe i'm just lucky heh Apr 08 03:36:47 even so, just delete cache and move on :) Apr 08 03:36:54 well, from a configuration management perspective, you always want a completely fresh, pristine build on a regular basis Apr 08 03:37:25 * mwester comes from a config mgmnt background, and thinks ccache is a blight that should be stamped out. Apr 08 03:37:29 either way, it should still run :D Apr 08 03:38:15 meh, saves so much time when building certain things, if I need a fresh build it's easy enough to do that Apr 08 03:38:32 ccache: "I can build in 1/2 the time! (Nevermind that I never really can be certain that I'm building it correctly)" Apr 08 03:38:40 mwester: exactly Apr 08 03:38:42 sigh, this is going to take awhile, apparently it's going to build 4464 packages heh Apr 08 03:38:56 it can reduce your compile/test cycle during development though Apr 08 03:39:11 mwester: If issues arise, clear cache and rebuild Apr 08 03:39:47 If it works 99% of the time like it does for me it saves a lot of time :) Apr 08 03:39:56 If your compiler, linker, and host libs are immutable, then I agree that ccache is a good _development_ tool. one can never make that assumption in OE, so it fails even for developers. Apr 08 03:40:19 indeed, much better for the development perspective than in OpenEmbedded Apr 08 03:40:25 sure Apr 08 03:40:35 mekius: 1% of the time is too high a failure rate if the problem caused consumes 99% of the time. How long did this last ccache issue cost you? Apr 08 03:41:03 mwester: Maybe 4 hours :P Apr 08 03:41:07 over 3 days heh Apr 08 03:41:11 * mwester cannot quite understand why we default to having ccache enabled for OE. Apr 08 03:41:13 heh, OpenEmbedded has enough reproducability problems as is, without throwing ccache issues in too Apr 08 03:41:29 besides, using pstage to cache whole package builds is even better if you're careful Apr 08 03:41:32 well I don't think this was ccache, it didn't even seem to run it at all Apr 08 03:41:36 rather than doing it at the object level Apr 08 03:41:42 just went straight to fail Apr 08 03:41:54 when i ran the script by itself it worked fine Apr 08 03:41:55 * mwester agrees with kergoth on all comments! Apr 08 03:42:03 so ccache was not returning an error best I could tell Apr 08 03:42:12 otherwise I would have gotten some sort of error heh Apr 08 03:42:33 anyway Apr 08 03:42:35 anyone done a recent tool using event handlers in OpenEmbedded to log how long each of the tasks took, and emit that into a log file for the build? Apr 08 03:42:40 I can build stuff now so I'm happy :) Apr 08 03:42:41 i remember doing one like 4 years ago or something Apr 08 03:43:55 does 4000+ packages sound reasonable for my first bitbake run? Apr 08 03:44:25 what target did you try and bake? Apr 08 03:44:53 palm680, angstrom distro, illume-image Apr 08 03:47:16 not unreasonable Apr 08 03:47:42 kk Apr 08 03:47:58 figured there was a lot to build, just wanted to double check :) Apr 08 03:48:13 That's 4000 tasks, not packages, actually. Apr 08 03:49:17 oh ok Apr 08 03:55:24 159/4464 :P Apr 08 03:55:31 i'm still building native stuff heh Apr 08 03:55:52 hmmmmm Apr 08 03:56:04 * kergoth ponders possibilities for the recipe in the source tree model w/ git Apr 08 04:07:42 why is it that it builds all it's own gcc and glibc? Apr 08 04:08:05 or are these all for cross compiling? Apr 08 04:08:14 do you not see the "-cross" suffix? Apr 08 04:08:39 only been building -native stuff so far Apr 08 04:08:42 nothing -cross yet :) Apr 08 04:08:51 oh I take that back Apr 08 04:09:07 mostly native only 1 or 2 cross Apr 08 04:10:01 OE doesn't trust some of your native tools... I guess we should be thankful that it doesn't build "linux-native" and reboot your build system as part of the process. ;-) Apr 08 04:10:11 haha Apr 08 04:10:12 It's a bit extreme in my opinion. Apr 08 04:10:16 a bit :) Apr 08 04:10:25 at least it's a one time thing heh Apr 08 04:10:35 well I guess until one of them gets updated :P Apr 08 04:10:52 * mekius checks his available space Apr 08 04:11:10 The judicious use of "ASSUME_PROVIDED" can help, but use caution because the core devs pretty much live on the absolute bleeding edge latest versions of everything, so those of us who live with "stable" host distros can find that our host tools don't work with OE... Apr 08 04:11:24 gigs of tools to build 200MB of actual files :P Apr 08 04:11:37 ah Apr 08 04:12:13 SlugOS distro fits in 8MB of flash space... so many GB of disk just to create a 6MB flash jffs2 image. :) Apr 08 04:12:28 yeah heh Apr 08 04:12:49 got anything that will fit into 256k of flash? :) Apr 08 04:13:10 * mwester got burned because he did ASSUME_PROVIDED += "git-native" --- which worked until somebody added a feature that only works with very recent git versions. Apr 08 04:13:26 mekius: u-boot Apr 08 04:14:00 gah, failure building qemu-native Apr 08 04:14:17 sanity check? Apr 08 04:16:03 not sure, looking Apr 08 04:16:17 /usr/bin/ld: cannot find -lm Apr 08 04:19:00 I guess this means I'm removing all the ASSUME_PROVIDED from my conf :P Apr 08 04:19:40 hmm, nvm Apr 08 04:19:47 I only have preferred providers set Apr 08 04:19:48 hmm Apr 08 04:32:08 gah, what is this business about missing -lm :P Apr 08 04:35:15 anyone? Apr 08 04:45:56 mekius: do you have /lib/libm.so.6 Apr 08 04:46:01 on your host Apr 08 04:54:17 hi. who enable DSP of omap3 for media coding/decoding successfully? Apr 08 05:21:53 * kergoth ponders Apr 08 05:53:35 hi, I excute "bitbake gstreamer-ti", and I got an error, please see pastebin(http://www.pastebin.ca/1385705). what should I do to solve this issue? Apr 08 06:38:13 hi gremlin[it] , have you received your shhevaplug? Apr 08 06:38:27 no ... Apr 08 06:38:30 hi mckoan Apr 08 06:39:46 gremlin[it]: but they charged credit card immediately, bastards! Apr 08 06:39:49 i did the order on march 30 ... so it should be sell as soon as possible (not at the end of april) ... i hope al least ... but globalscale neither respond to my mails :( Apr 08 06:39:53 mckoan, yes ! Apr 08 06:40:15 good morning Apr 08 06:41:11 morning methril|work Apr 08 06:42:01 morning methril|work Apr 08 06:42:25 mckoan, when u did the order ? Apr 08 06:47:49 gremlin[it]: I did it on 31/3 Apr 08 06:48:23 shipping with FedEx intl priority Apr 08 06:48:26 about the same ... let's wait ... i send an email just now to have info about my shipment ... i hope to have good news Apr 08 06:48:34 mckoan, me too ... Apr 08 06:56:17 03Koen Kooi  07org.openembedded.dev * rc528cdb0ba 10openembedded.git/recipes/e17/ (e-wm/fix-profiles.diff e-wm_svn.bb): e-wm: fix profile selection in first-run wizard Apr 08 06:56:51 03Koen Kooi  07org.openembedded.dev * r8b680af5bc 10openembedded.git/recipes/e17/diskio_svn.bb: diskio: new e17 module Apr 08 07:00:43 03Koen Kooi  07org.openembedded.dev * ra6169a4840 10openembedded.git/recipes/angstrom/ (4 files in 2 dirs): e-wm-config-angstrom: refresh configs, add systray Apr 08 07:01:36 morning Apr 08 07:03:48 morning recalcati_oe Apr 08 07:04:03 * mckoan -> off for 40min Apr 08 07:26:17 i see some memory leak issue, i have built a gpe-image with gnome-mplayer, gxine, abiword and gnumeric and it stores in a sdcard. Apr 08 07:28:57 i have a 128MB ram, when it boots up i see 83MB free, when i start gnumeric and exit that app, i see memfree to be 63MB, now i open and close gnome-mplayer is see memfree to be 53MB Apr 08 07:29:42 when i close some app, memfree should be what it was before isnt it? Apr 08 07:34:55 Hello everybody Apr 08 07:36:02 I'm having a problem using bitbake Apr 08 07:38:03 I want to build wxbase-2.8.9. If I launch "bitbake wxbase-2.8.9", it works and it gets built Apr 08 07:38:41 now I want to create a custom image, and get wxbase-2.8.9 added to it, so I have created a new recipe with this code: http://paste.dprogramming.com/dpikccag Apr 08 07:40:01 But it fails, when I try to build it, I get this error: http://paste.dprogramming.com/dpjz9il7 Apr 08 07:40:21 Any suggestions? Apr 08 07:41:41 Also note, if I add to the recipe "wxbase" instead of "wxbase-2.8.9", my custom image builds, but I get wxbase version 2.6 instead of 2.8.9. I'm a bit lost... Apr 08 07:44:06 mckoan: bye Apr 08 07:49:31 good morning Apr 08 07:51:22 Browsing wxbase_2.8.9.bb file, I've noticed it has this line 'DEFAULT_PREFERENCE = "-1"'. Can my problem be because of this? Apr 08 07:52:14 morning Apr 08 07:52:26 would it be correct to change it to 'DEFAULT_PREFERENCE = "1"' and also change 'PR = "r0"' to 'PR = "r1"'? Apr 08 07:55:52 how to bitbake the oe DSPlink? I mean it need some setting by myself. Apr 08 07:56:48 morning hrw Apr 08 07:59:02 favor: bitbake ti-codec-engine ? but you need to install all the TI stuff first Apr 08 08:01:16 * mckoan back Apr 08 08:04:53 yeah, I had done that. but regarding The 'all', I don't know how many TI stuffs should I install first. Apr 08 08:06:29 favor: I've bios_5_33_03 / cgtools_6_0_21 / xdctools_3_10_03 Apr 08 08:07:58 03Koen Kooi  07org.openembedded.dev * rdb35931307 10openembedded.git/recipes/powervr-drivers/omap3-sgx-modules_1.3.13.1397.bb: omap3-sgx-modules: add compile fix for DSS2 changes in linux-omap 2.6.29 Apr 08 08:08:04 Good morning! Apr 08 08:08:06 favor: see oe/openembedded/recipes/dsplink/ti-paths.inc Apr 08 08:08:49 mpoullet|work: I have bios_5_33_02 cg6x_6_0_16 xdctools_3_10_03 dmai_1_20_00_06 Apr 08 08:09:11 03Jeremy Lainé  07org.openembedded.dev * r1c535e07c1 10openembedded.git/recipes/ppp/ (ppp-2.4.3/ldflags.patch ppp_2.4.3.bb): ppp: add patch to use LDFLAGS (fixes QA error) Apr 08 08:09:18 03Jeremy Lainé  07org.openembedded.dev * r5f00d873c0 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Apr 08 08:09:53 favor: so everything is ok ?? or do you get errors during bitbaking ? Apr 08 08:11:14 mpoullet|work: yeah, I have met with some errors. and I put it at beagleboard@google.com maillist. Apr 08 08:14:52 mpoullet|work: I would like you to take your time to see that thread, thank you.http://groups.google.com/group/beagleboard/browse_thread/thread/3d2560c087941100?pli=1 Apr 08 08:20:46 03Jeremy Lainé  07org.openembedded.dev * rb04da76f3d 10openembedded.git/recipes/irda-utils/ (irda-utils/ldflags.patch irda-utils_0.9.16.bb): irda-utils: add patch to use LDFLAGS (fixes QA error) Apr 08 08:25:54 03Koen Kooi  07org.openembedded.dev * r4ccabab2da 10openembedded.git/recipes/dsplink/ti-dmai_svn.bb: ti-dmai: add compile fix for DSS2 changes in linux-omap 2.6.29 Apr 08 08:28:40 03Jeremy Lainé  07org.openembedded.dev * r4b49e2ab31 10openembedded.git/recipes/obexpush/obexpush_1.0.0.bb: obexpush: use LDFLAGS (fixes QA error) Apr 08 08:28:41 03Jeremy Lainé  07org.openembedded.dev * r7ae24f3d57 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Apr 08 08:33:16 favor: first the minor issue with svn, you solved it the right way but you don't need to generate a package and put it in the recipes Apr 08 08:33:59 favor: you should simply restart and the second time svn remembers the certificate and the fetch work Apr 08 08:34:53 03Jeremy Lainé  07org.openembedded.dev * ra6d06c8abf 10openembedded.git/recipes/gsoap/gsoap_2.7.7.bb: gsoap: use autotools_stage Apr 08 08:35:59 favor: yous should put the log /home/user/myoe/tmp/work/beagleboard-angstrom-linux-gnueabi/ti-codec-combos-316-r15/temp/log.do_compile.14017 in the mail too, otherwise it's difficult to help you Apr 08 08:36:22 favor: and you should try the #gst_ti channel too Apr 08 08:36:46 favor: and/or try to find koen there or in #beagle Apr 08 08:38:03 does someone do builds for avr32? Apr 08 08:39:10 mpoullet|work: yeah, I tried restart the bitbake command the second time, but it remain those errors, so I generate a package and put it in the recipes, and modify the bb file of recipes SRC_URI. Apr 08 08:40:33 mpoullet|work: The compiling log is too long, should I put the log file in attachment of the mail? Apr 08 08:42:11 hrw: sounds like it doesn't work currently... Apr 08 08:42:55 03Jeremy Lainé  07org.openembedded.dev * r0377cdf5da 10openembedded.git/recipes/hostap/ (hostap-utils-0.4.7/ldflags.patch hostap-utils_0.4.7.bb): hostap-utils-0.4.7: use LDFLAGS (fixes QA error) Apr 08 08:43:06 florian: e2fsprogs fails Apr 08 08:43:07 favor: strange, for me the second rebuild works Apr 08 08:43:24 favor: or at least so a tail on the log Apr 08 08:43:34 s/so/do/ Apr 08 08:43:47 hrw: Its a long time ago I tried it but I remember some problem like this. Apr 08 08:44:05 03Jeremy Lainé  07org.openembedded.dev * r10e6229075 10openembedded.git/recipes/hostap/ (3 files in 2 dirs): Apr 08 08:44:05 hostap-utils: only keep version 0.4.7 Apr 08 08:44:05 * hostap-utils is obsolete, if we must keep one it should be version 0.4.7 Apr 08 08:44:08 But I can't remeber if its really e2fsprogs that fails. Apr 08 08:44:50 mpoullet|work: ok. the tail of the log.do_compile. I 'll update the main. Apr 08 08:45:50 s/its/it was Apr 08 08:46:03 ok. kernel booted Apr 08 08:47:51 03Jeremy Lainé  07stable/2009 * rd441ef73ac 10openembedded.git/recipes/gd/gd_2.0.33.bb: Apr 08 08:47:51 gd: use LDFLAGS (fixes QA error) Apr 08 08:47:51 Signed-off-by: Jeremy Lainé Apr 08 08:47:51 Acked-by: Denys Dmytriyenko Apr 08 09:03:34 i see that the memory consumption keeps reducing as and when i open and close x apps like gnumeric, gxine, gnome-mplayer.., once it exits i dont see memory size as it was before opening the app Apr 08 09:05:17 03Jeremy Lainé  07org.openembedded.dev * rdbe77279e4 10openembedded.git/recipes/hostap/ (hostap-daemon-0.6.9/ldflags.patch hostap-daemon_0.6.9.bb): hostap-daemon-0.6.9: add patch to use LDFLAGS (fixes QA error) Apr 08 09:08:39 Linux atngw100 2.6.28 #2 PREEMPT Wed Apr 8 10:59:44 CEST 2009 avr32 unknown Apr 08 09:15:21 http://www.flickr.com/photos/hrwandil/3423727862/ Apr 08 09:20:11 florian: good morning Apr 08 09:20:25 pb_: hi Apr 08 09:20:55 hrw: I like this :-) Apr 08 09:21:28 florian: I hate serial port on ngw100 Apr 08 09:21:49 it use female instead of male so I do not have cable for it Apr 08 09:22:02 will have to buy db9 gender changer Apr 08 09:22:52 I have quite a lot of dvices sufferening from strange serial ports so I have a pile of adapters here :) Apr 08 09:23:25 03Jeremy Lainé  07org.openembedded.dev * r846330c96c 10openembedded.git/classes/insane.bbclass: insane.bbclass: make "LDFLAGS ignored" QA errors fatal Apr 08 09:25:15 hrw: ok, just flipped the switch, LDFLAGS related QA errors are not longer accepted Apr 08 09:38:06 is a memory leak in xwindows, as and when i open/close the xapps, my memory goes low by 10000K to 20000K bytes Apr 08 09:45:14 http://marcin.juszkiewicz.com.pl/2009/04/08/new-architecture-on-my-desk/ Apr 08 09:57:29 vsr: try to wait some time Apr 08 10:02:13 03Jeremy Lainé  07org.openembedded.dev * rac0e1bd19c 10openembedded.git/recipes/flite/ (flite-1.3/flite-1.3-Makefile.patch flite_1.3.bb): flite-1.3: fix LDFLAGS breakage Apr 08 10:06:46 jay7: i wait for 10 or more minutes, but i see the same memory usage Apr 08 10:07:02 vsr: have you swap? Apr 08 10:07:32 no Apr 08 10:09:06 that's strange.. I think that x11 app uses shm memory or mmap Apr 08 10:09:18 then garbage collector frees it after some time.. Apr 08 10:09:38 but I think that it should not be longer that 10 mins Apr 08 10:12:38 03Michael 'Mickey' Lauer  07org.openembedded.dev * r2b4bf8dbde 10openembedded.git/recipes/vala/vala-bootstrap-native_0.6.0.bb: vala-bootstrap-native: bring back timestamp hack; still necessary Apr 08 10:18:44 hrw: you either need a null-modem cable or an extension (1-on-1) cable for boards. Does gender changer work to change between null-modem and extension (1-on-1) cable? Apr 08 10:20:45 right.. Apr 08 10:21:01 anyway need cables and serial ports Apr 08 10:25:01 mwester: ping? Apr 08 11:05:07 hrw: what I wanted to say: a gender changer might not convert null-modem into 1:1. I'm not sure though Apr 08 11:08:32 jay7: please see http://pastebin.com/m53395f3f and http://pastebin.com/mb993448 Apr 08 11:08:51 if i open and close gnome-mplayer - it takes 10MB of memory away Apr 08 11:09:12 i dont get back even after 10 minutes Apr 08 11:09:41 is this garbage collector part of the kernel/user space or both Apr 08 11:11:15 probably the python overhead :-) Apr 08 11:12:20 what do you suggest Apr 08 11:16:38 see if it started any processes that are still running Apr 08 11:17:06 especially ones taking 16% of memory Apr 08 11:17:19 for 64M machine anyway Apr 08 11:26:38 i dont see any new processes Apr 08 11:46:58 vsr: there isn't really anything to worry about in the output you pasted. Apr 08 11:47:45 you have plenty of ram free, and as long as this happy situation persists the kernel will not bother to trim its caches. Apr 08 11:52:13 denix0: hi Apr 08 11:52:18 denix0: works Apr 08 11:52:41 * broonie wonders what's going on with getting his key added. Apr 08 11:52:53 broonie: mailed it to mickeyl? Apr 08 11:53:09 pb_: i have a 500MHZ processor and it wont play video file without break Apr 08 11:53:12 booxter: it takes few days usually; mine took mostly two weeks Apr 08 11:53:15 oops Apr 08 11:53:18 broonie: ^ Apr 08 11:53:30 vsr: and you do that without any video acceleration? Apr 08 11:53:36 vsr: plain fbdev or plain x11? Apr 08 11:53:48 vsr: that's very sad, but it isn't obviously related to your "missing" 10MB of memory Apr 08 11:53:56 hrw: Twice. Apr 08 11:53:58 i have a x11 file system and i have gnome-mplayer Apr 08 11:54:09 hrw: He denied getting it the first time :( Apr 08 11:54:51 I'm tempted to mail it via the list. Apr 08 11:55:26 pb_: why memory keeps reducing as and when i open and close the app. it is supposed to be the same before it was opened isnt it? Apr 08 11:55:49 vsr: no, because more data will be in cache afterwards. Apr 08 11:57:02 vsr: x11 != x11... Apr 08 11:57:10 vsr: do you have XVideo atleast? Apr 08 11:57:11 unless you are actually having problems with lack of memory, I suggest you just ignore those numbers. Apr 08 11:58:25 hrw: i build a file system using x11-gpe-image Apr 08 11:58:44 i added gnome-mplayer, gnumeric to it Apr 08 11:59:15 vsr: I do not care what you have in rootfs. I asked about your HW not rootfs Apr 08 12:00:19 i dont think im using hardware acceleration Apr 08 12:00:32 vsr: zaurus c7x0 has XVideo so plays videos quite good on pxa255/400MHz. BUG lacks xvideo so it is not so nice with playing videos even on its imx31/533MHz Apr 08 12:00:40 i use omap3 processor Apr 08 12:00:54 vsr: so you should have xvideo support Apr 08 12:01:02 khem: I have libm-2.9.so and libm.so.6, seems there is no symlink, what library is this? Apr 08 12:01:33 hrw: it might just put it on the frame buffer Apr 08 12:02:39 on omap usually decoding is done on dsp. i dont have dsp codecs up yet Apr 08 12:03:30 vsr: use omapfbplay Apr 08 12:09:58 ant_work: hiya Apr 08 12:10:57 hello :) Apr 08 12:11:30 ant_work: LDFLAGS QA errors are now fatal :) Apr 08 12:12:03 sounds good... Apr 08 12:12:27 but I did not build world...just console- opie- x11- images Apr 08 12:12:43 possibly some other is floating around, let wait and see Apr 08 12:12:52 libsdl-gfx and libsdl-ttf have issues Apr 08 12:13:04 but so far I haven't found got a working fix Apr 08 12:14:05 about other QA wrt .so, eh..dunno Apr 08 12:20:51 is there commercial support available for OpenEmbedded? Apr 08 12:24:34 koobe_: yes of course Apr 08 12:25:21 mckoan: where should I look? Apr 08 12:26:24 koobe_: ask me: http://www.kaeilos.com or to somobody else ;-) Apr 08 12:30:07 mckoan: great, thanks. Apr 08 12:33:07 03Koen Kooi  07org.openembedded.dev * r579fcb8a6c 10openembedded.git/recipes/powervr-drivers/libgles-omap3.inc: libgles-omap3: these are evil binaries, so skip insane.bbclass tests Apr 08 12:50:46 03Mark Brown  07org.openembedded.dev * r26865dffa4 10openembedded.git/ (conf/checksums.ini recipes/watchdog/watchdog_5.6.bb): Apr 08 12:50:46 watchdog: Update to version 5.6 Apr 08 12:50:46 No packaging changes, just the new upstream. Apr 08 12:50:46 Acked-by: Koen Kooi Apr 08 13:00:55 welcome to the party broonie Apr 08 13:02:35 koobe_: there are companies ;) Apr 08 13:05:45 btw. OE e.V. got tax exemption status Apr 08 13:05:51 see oe-priv Apr 08 13:14:37 morning all Apr 08 13:14:52 good morning RP Apr 08 13:15:09 * RP is trying to spend more time online in various places again :) Apr 08 13:16:05 hey RP Apr 08 13:16:06 RP i have a GUI installer for OE and Poky I'll be releasing shortly. Apr 08 13:16:22 kgilmer: Sounds cool :) Apr 08 13:16:36 hi XorA Apr 08 13:16:38 kgilmer: hi, what is " GUI installer" Apr 08 13:18:22 I'm wrapping this up in a wizard mckoan: http://wiki.openembedded.net/index.php/Getting_started Apr 08 13:19:40 RP: wb Apr 08 13:20:22 kgilmer: something related to your SDK v. 1.3.0 for Eclipse ? Apr 08 13:22:42 yes you got it mckoan Apr 08 13:22:53 if my resolution of lcd changes do i need to recompile x? Apr 08 13:24:59 mckoan, it could be easily adapted to work with your distro as well Apr 08 13:25:05 kgilmer: we are working on it and we solved the problems we got with your plugin Apr 08 13:25:16 is it safe to share DL_DIR's for downloaded sources? Apr 08 13:25:20 great! so it's working? Apr 08 13:25:45 i emailed your java developer mckoan but he did not respond to me. Apr 08 13:25:55 kgilmer: it's working, but my 'vision' will require a lot of work since now :-D Apr 08 13:26:56 haha Apr 08 13:27:02 that's good to hear. Apr 08 13:27:35 there is a new project inside of eclipse starting, basically for linux distro tool projects. Apr 08 13:27:41 kgilmer: I still wonder if Elipse is the right tool for that rather than Anjuta for example Apr 08 13:27:45 i am considering integrating my code with them. Apr 08 13:28:31 kgilmer: do you have a link? Apr 08 13:28:36 I cannot say about that mckoan :) I don't know much about Anjuta. Apr 08 13:29:53 mckoan: http://www.eclipse.org/linuxtools/ Apr 08 13:31:30 they have some pretty ambitious goals Apr 08 13:31:33 I'll take a look thx Apr 08 13:31:50 hi everyone Apr 08 13:33:38 just a stupid question from a newbie: I've submitted a patch for gsoap, and sent it to the -devel mailing list for review. Koen repliad that it is good enough to go in. What now? Apr 08 13:34:10 again without really knowing anything about ajunta, eclipse primary strength is in it's popularity and large variety of plugins available. if you're looking for specific fixed functionality that may not be important to you. Apr 08 13:34:20 my 2 cents mckoan Apr 08 13:37:18 virk, give us a day or so Apr 08 13:37:21 then ping us Apr 08 13:37:34 some times when i boot, i dont get gpe-login screen, my screen becomes blank with a x in the middle Apr 08 13:38:04 what could be going wrong? Apr 08 13:38:05 Crofton: I was in fact just wondering If I still needed to do something Apr 08 13:38:14 no, wait :) Apr 08 13:38:22 ok, that's fine :) Apr 08 13:38:31 thanks for your patience and taking care of comments promptly Apr 08 13:39:17 glad to help Apr 08 13:39:44 kgilmer: OTOH eclipse primary weak is java :-D Apr 08 13:40:06 OTOH? Apr 08 13:40:55 kgilmer: On The Other Hand Apr 08 13:41:05 base_set_filespath usage is deprecated, glibc-initial-2.6.1 should be fixed Apr 08 13:41:24 haha. ok. how does the language something is written matter so much to be the primary weakness? Apr 08 13:41:27 I get a lot of warning like these when I build the stable branch Apr 08 13:43:00 kgilmer: in fact that's the main weakness, not the primary :-P Apr 08 13:43:36 kgilmer: I know you love java, I'm joking :-D Apr 08 13:43:44 i salute your objectivity mckoan :) Apr 08 13:43:46 when i try to clock touch pas multiple times i get the error glibc: xorg free : invalid pointer Apr 08 13:44:16 actually it's the only language i'm really efficient in ATM, so I tend to use it. Apr 08 13:44:54 exact message is glibc detected: xorg: free (): invalid pointer: Apr 08 13:45:13 kgilmer: I see, I can say the same for C Apr 08 13:45:35 RP, btw we got connman running on bug (poky) and it works very nicely. Apr 08 13:46:15 i see mckoan. in that case if you are doing the develoment I strongly suggest Anjuta :) Apr 08 13:48:32 03Koen Kooi  07stable/2009 * rf6e2171acc 10openembedded.git/recipes/angstrom/ (4 files in 2 dirs): Apr 08 13:48:32 e-wm-config-angstrom: refresh configs, add systray Apr 08 13:48:32 Signed-off-by: Koen Kooi Apr 08 13:48:32 Acked-by: Philip Balister Apr 08 13:48:38 03Koen Kooi  07stable/2009 * r4d766d84fe 10openembedded.git/recipes/e17/systray_svn.bb: Apr 08 13:48:38 systray: as the name implies, a systray module for e17 Apr 08 13:48:39 Signed-off-by: Koen Kooi Apr 08 13:48:41 Acked-by: Philip Balister Apr 08 13:52:59 it's almost 11 pm here Apr 08 13:53:04 my day is about done Apr 08 13:53:40 Crofton: Is a patch needed when a newer version (of for example pcsc-lite) did not change at all on how it is 'bitbaked'. I mean, pcsc-lite 1.5.2 builds perfectly the same as version 1.4.102, just by changing the filename of the recipe Apr 08 14:04:38 virk, still need a patch Apr 08 14:04:51 it is the easiet for us to process Apr 08 14:05:30 sometimes you can point out something that someone is interesting in and they can do it right away, but best to send a pach Apr 08 14:05:56 some comments about hy the new version is "better" can be helpful also Apr 08 14:06:54 hello,I've wesnoth,boost and an assembly part for atomic_count in boost that I stolen and adapted from lwn...the assembly part seems to work(++num increment and --num decrement but num++ and num-- don't increment or decrement) but the most strange part is that one: http://rafb.net/p/SXbw4p66.html gdb fails... Apr 08 14:07:10 (I was helped for the adaptation and the stealing) Apr 08 14:07:50 is it a distro problem? is it an oe problem? Apr 08 14:08:15 I also had that in dmesg: [ 2199.335000] send sigkill to 1803 (wesnoth), adj 0, size 11921 Apr 08 14:08:29 what did you get from lwn? Apr 08 14:08:42 url? Apr 08 14:08:47 that: http://lwn.net/Articles/314144/ Apr 08 14:08:49 I need to file things like this :) Apr 08 14:08:59 Posted Jan 12, 2009 19:58 UTC (Mon) by npitre (subscriber, #5680) [Link] Apr 08 14:09:28 it's woglinde that pointed me to the article... Apr 08 14:09:30 any idea why is the GCC-4.3.x broken in OE ? Apr 08 14:09:37 it sigsegvs while compiling ncurses Apr 08 14:10:05 maybe wesnoth+gdb uses too much memory? Apr 08 14:10:20 does remote gdb uses less memory Apr 08 14:10:21 ? Apr 08 14:11:09 any idea why we get message like glibc detected: xorg: free (): invalid pointer: , i see a blank screen, no gpe-login Apr 08 14:11:35 marex: thumb is broken, turn off thumb Apr 08 14:12:18 vsr: memory corruption Apr 08 14:12:46 XorA, in 4.3 only, right ? Apr 08 14:12:50 xora: what could cause it, i just boot my board, it was working all while and suddenly it stopped working Apr 08 14:12:57 marex: yes, 4.2.X worked Apr 08 14:13:04 I see Apr 08 14:13:18 vsr: no idea, but sounds like something corrupted that pointer Apr 08 14:13:30 is this related to compilers in any way? Apr 08 14:13:34 a bad file system? Apr 08 14:13:58 vsr: what platform? Apr 08 14:14:09 omap3 Apr 08 14:14:44 vsr: well as Crofton and koen seem to make reliable FS with geat regularity Id suspect your board Apr 08 14:16:20 vsr what image are you building? Apr 08 14:16:48 XorA, is anyone going to fix that bug ? Apr 08 14:17:18 crofton: gpe-image Apr 08 14:17:44 if I run wesnoth alone...it segfault but I don't see a segfault in the kernel Apr 08 14:17:47 dmesg Apr 08 14:17:55 vsr, I do not run gpe Apr 08 14:18:00 mostly I do conolse work Apr 08 14:18:02 s/kernel kernel's dmesg/ Apr 08 14:18:07 which omap3 board? Apr 08 14:18:44 marex: don't know, ask the gcc guys Apr 08 14:19:08 marex: it seems gcc/binutils guys never test thumb, it breaks every other version Apr 08 14:20:40 crofton: custom board Apr 08 14:21:15 do you have evm/beagle/overo you can test on? Apr 08 14:22:01 XorA, sucks :/ what about fixing it ? ;-) Apr 08 14:23:12 crofton: it was working all while, i suspect a file system corruption Apr 08 14:24:23 crofton: you run mplayer on omap and it plays video without any issue, like video/audio breaking Apr 08 14:24:55 no, 99% of my work has no screen :) Apr 08 14:26:29 okay Apr 08 14:26:54 marex: patches are welcome Apr 08 14:27:25 pb_, hehe ;-) Apr 08 14:27:44 03Koen Kooi  07stable/2009 * r399a19118b 10openembedded.git/recipes/dsplink/ (gstreamer-ti_svn.bb ti-codec-combos_3.16.bb): Apr 08 14:27:44 gstreamer-ti: fix path to codec combos Apr 08 14:27:45 Acked-by: Philip Balister Apr 08 14:27:53 pb_, it wasnt actually a request Apr 08 14:29:28 marex: why do you so desperately require 4.3.3? Apr 08 14:29:58 XorA, since 4.4.0 is near and it'd be good to keep stuff up-to-date Apr 08 14:31:24 you could try the latest 4.4 snapshot, that might well fix your issue anyway. Apr 08 14:32:09 again OE patches welcome :-) Apr 08 14:43:09 mmm...wesnoth works fine with ssh -Y(needed for having at least 400x300 resolution) and armv6 Apr 08 14:43:17 the same doesn't work with armv4 Apr 08 14:43:25 I've only a segfault Apr 08 14:44:02 I'll look at the assembly again Apr 08 14:45:42 mmm...there is for armv6 or superior that: "ldrex %0, %1 \n\t" and I did a "ldr r0, [%1] Apr 08 14:45:55 with %1 beeing r2 Apr 08 14:49:10 what exactly are you trying to do? Apr 08 14:55:40 morning Apr 08 14:56:01 g'day kergoth Apr 08 15:00:14 kergoth: hi Apr 08 15:05:23 good afternoon :) Apr 08 15:13:29 03Dirk 'Virk' Van Haerenborgh  07org.openembedded.dev * rff30a9ba2d 10openembedded.git/ (5 files in 3 dirs): Apr 08 15:13:29 add gsoap v2.7.13, generates C/C++ source codes for Web service and client applications Apr 08 15:13:29 gsoap-native: unbreak self-dependency patch included Apr 08 15:13:30 FILES_${PN} includes {datadir} as soapcpp2 is some sort of compiler which needs the headers in datadir in order to be able to compile its Apr 08 15:13:32 soap stuff properly Apr 08 15:13:34 (third time's the charm) Apr 08 15:13:36 03Dirk 'Virk' Van Haerenborgh  07org.openembedded.dev * r47ef4977b2 10openembedded.git/ (conf/checksums.ini recipes/pcsc-lite/pcsc-lite_1.5.2.bb): Apr 08 15:13:39 pcsc-lite: Update to version 1.5.2 Apr 08 15:13:41 No packaging changes. Apr 08 15:13:49 pb_, do you know how exactly to set up the gcc-4.4.0 snapshot for OE ? Apr 08 15:14:06 pb_, do I need to write .bb file for it ? Apr 08 15:14:19 interesting that using dolt breaks gettext-native somehow, even though it isn't regenerating its autotools bits.. and it breaks in a very odd way.. wonder why Apr 08 15:14:24 or can I somehow use the gcc_svn.bb Apr 08 15:15:36 marex: you should be able to use gcc_svn and adjust the src_uri to pull from the 4.4 branch rather than the head Apr 08 15:16:48 ok, I'll look into it once I get back home Apr 08 15:21:30 hi, i've just recently begun seeing 'ImportError: No module named _md5' error messages during my final do_rootfs task, are there new requirements for python or..? Apr 08 15:30:47 rebuild openssl-native and then python-native Apr 08 15:38:56 hmmmmm Apr 08 15:39:21 indeed Apr 08 16:13:31 03Dirk 'Virk' Van Haerenborgh  07org.openembedded.dev * r647623726b 10openembedded.git/recipes/pcsc-lite/pcsc-lite_1.5.2.bb: pcsc-lite_1.5.2: wrong download link Apr 08 16:22:26 hi, what does ldrex %0, %1 do? (not ldrex %0,[%1]) Apr 08 16:24:51 that's just an instruction pattern, it doesn't "do" anything directly. it would all depend on what operands are substituted into it. Apr 08 16:25:02 what does your resulting assembly output look like? Apr 08 16:26:30 the problem is that I'm trying to fix boost for armv4...here the source of the assemlbly file: http://rafb.net/p/4duwO956.html Apr 08 16:26:57 I'm using boost with the new wesnoth Apr 08 16:27:18 with armv6 it works(if your screen is 400x300 at minimum) Apr 08 16:27:23 but not with armv4 Apr 08 16:27:28 but it compiles Apr 08 16:28:21 pb_, maybe I should change the register int val asm("r2") = (value_); to & or * ? Apr 08 16:28:34 so I could keep the assembly syntax as is Apr 08 16:28:47 because without the [] it doesn't compile Apr 08 16:29:09 even if I change the input/output list at the end Apr 08 16:30:20 I've made a small test program but I can't reproduce all the compilation process with it...in fact it doesn't fail where the whole compilation fails Apr 08 16:30:37 what's the call at 0xffff0fc0, is that cmpxchg or something? Apr 08 16:31:08 pb_, it's a call to the kenrel Apr 08 16:31:16 http://lwn.net/Articles/314144/ Apr 08 16:31:18 right, but which call? Apr 08 16:31:26 see Posted Jan 12, 2009 19:58 UTC (Mon) by npitre (subscriber, #5680) [Link] Apr 08 16:31:45 okay, yeah, I see, it is indeed cmpxchg Apr 08 16:31:59 yes Apr 08 16:32:14 and what exactly is the problem that you're having with this code? Apr 08 16:32:22 segmentation fault Apr 08 16:32:30 on which instruction? Apr 08 16:32:33 and gdb don't work Apr 08 16:32:43 maybe I should try remote gdb? Apr 08 16:33:06 http://rafb.net/p/SXbw4p66.html Apr 08 16:33:19 here's gdb on my device trough ssh Apr 08 16:33:32 or a different version of gdb, I guess. or some other tool. Apr 08 16:33:41 how did you establish that this piece of code was causing the segfault? Apr 08 16:33:43 ok I'll bitbake gdb Apr 08 16:33:49 I didn't Apr 08 16:34:07 It was that it worked on armv6 and not armv4 that made me think that... Apr 08 16:34:12 ah Apr 08 16:34:18 but I'm not shure it's the cause Apr 08 16:34:26 well, the first thing to do is to find out where your program is actually crashing. Apr 08 16:34:46 it's silly to try to debug that code without even knowing which instruction is causing the problem, the actual bug might be somewhere completely unrelated. Apr 08 16:35:07 if gdb doesn't work then you can use catchsegv, or in extremis CONFIG_DEBUG_USER or something. Apr 08 16:35:11 03Chris Larson  07org.openembedded.dev * r7ab3964056 10openembedded.git/conf/collections.inc: Apr 08 16:35:11 collections.inc: add support for collection archives. Apr 08 16:35:11 Adds the ability to specify an archive in COLLECTIONS. Any archives of a Apr 08 16:35:11 supported format (default tar.gz/tgz/tar.X, tar.bz2/tbz/tbz2, zip/jar) will be Apr 08 16:35:11 unpacked into COLLECTIONS_UNPACKDIR and used from there. The default value of COLLECTIONS_UNPACKDIR is ${TMPDIR}/collections. Apr 08 16:35:13 yes that's a big problem.... Apr 08 16:35:14 Signed-off-by: Chris Larson Apr 08 16:35:17 ok thanks i'll try Apr 08 16:35:21 03Chris Larson  07org.openembedded.dev * r47ff456098 10openembedded.git/conf/ (bitbake.conf collections.inc): (log message trimmed) Apr 08 16:35:21 bitbake.conf, collections.inc: Add COLLECTIONS mechanism w/ a sane default. Apr 08 16:35:23 Added code which takes a list of directories in COLLECTIONS, in priority Apr 08 16:35:25 order (highest to lowest), and uses those to populate BBFILES, Apr 08 16:35:27 BBFILE_COLLECTIONS, BBFILE_PATTERN_*, and BBFILE_PRIORITY_*. Apr 08 16:35:28 ah there are debugging in wesnoth...i'll try that Apr 08 16:35:29 The default COLLECTIONS is based on BBPATH, so you can now specify a sane Apr 08 16:35:31 suppose i shouldve added some acked-by's, but oh well Apr 08 16:35:31 BBPATH, not bother setting up the collections variables or BBFILES, and Apr 08 16:35:31 but it shouldn't be too hard to find a version of gdb that works well enough to at least give you a register dump. Apr 08 16:36:19 kergoth: good new, hope collections will go in stable too! Apr 08 16:36:36 would have to talk to one of the stable maintainers about that Apr 08 16:37:50 if someone wouldn't mind, please do a bit of testing using it, both relying on BBPATH (don't set COLLECTIONS at all yourself), and setting COLLECTIONS explicitly. just remove the BBFILES and BBFILE_* vars from your configs, because it won't overwrite yours Apr 08 16:38:34 03Koen Kooi  07org.openembedded.dev * reda32159da 10openembedded.git/ (2 files in 2 dirs): linux-omap 2.6.29: reenable BT for overo as well Apr 08 16:38:37 03Koen Kooi  07org.openembedded.dev * re5f45da5c7 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 08 16:38:49 ok i'll bitbake gdb Apr 08 16:38:49 thanks a lot Apr 08 16:39:42 03Chris Larson  07org.openembedded.dev * r9bf9f5e64e 10openembedded.git/conf/collections.inc: Apr 08 16:39:42 collections.inc: add note about not overwriting existing vars. Apr 08 16:39:42 Signed-off-by: Chris Larson Apr 08 16:41:16 kergoth: I'll definitely test that, but after my week holidays :)! Apr 08 16:41:22 :) have fun Apr 08 16:43:28 the worse is that I don't even know which program segfault...(could be a dep such as an X library or similar) Apr 08 16:43:34 I'm bitbaking gdb Apr 08 16:52:24 morning all Apr 08 16:54:48 ~curse xkb Apr 08 16:54:50 May the fleas of a thousand camels infest your most sensitive regions, xkb ! Apr 08 16:54:57 ~curse x11 too Apr 08 16:54:58 May you be reincarnated as a Windows XP administrator, x11 too ! Apr 08 16:55:29 x11 sucks when it comes to keyboards ;( Apr 08 17:14:24 03Chris Larson  07org.openembedded.dev * r3b215f7a41 10openembedded.git/conf/collections.inc: Apr 08 17:14:24 collections.inc: trim off trailing slashes for the entries in COLLECTIONS. Apr 08 17:14:24 Signed-off-by: Chris Larson Apr 08 17:21:55 bye Apr 08 17:23:21 hmm, i should push some of these little bitbake fixes Apr 08 17:23:58 03Koen Kooi  07org.openembedded.dev * r6f89098a07 10openembedded.git/recipes/bluez/ (bluez4_4.31.bb files/bluetooth.conf): Apr 08 17:23:58 bluez4: add less restrictive (and hence less secure) bluetooth dbus config Apr 08 17:23:58 * this makes bluetooth actually work for non-root users and when using ssh/vnc/etc Apr 08 17:24:08 03Koen Kooi  07org.openembedded.dev * r02655ea9a9 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 08 17:35:12 03kergoth * r1153 10bitbake/lib/bb/cooker.py: Apr 08 17:35:12 BBPKGS: act as a default that can be overridden via the cmdline. Apr 08 17:35:12 Signed-off-by: Chris Larson Apr 08 17:35:24 03kergoth * r1154 10bitbake/lib/bb/cooker.py: Apr 08 17:35:24 BBMASK: survive a non-existant path and don't keep the dir after globbing. Apr 08 17:35:24 Signed-off-by: Chris Larson Apr 08 17:35:44 03kergoth * r1155 10bitbake/lib/bb/parse/parse_py/ (BBHandler.py ConfHandler.py): Apr 08 17:35:44 Make the bbpath change for includes temporary rather than permanent. Apr 08 17:35:44 We currently modify BBPATH so that an include/require looks in the current Apr 08 17:35:44 working directory before it looks elsewhere. Modifying it permanently Apr 08 17:35:45 doesn't gain anything, just polluting it with locations that don't need to Apr 08 17:35:47 stay there. Apr 08 17:35:49 Signed-off-by: Chris Larson Apr 08 17:36:45 03kergoth 07bitbake-1.8 * r1156 10/lib/bb/cooker.py: Apr 08 17:36:45 BBPKGS: act as a default that can be overridden via the cmdline. Apr 08 17:36:45 Signed-off-by: Chris Larson Apr 08 17:37:01 03kergoth 07bitbake-1.8 * r1157 10/lib/bb/cooker.py: Apr 08 17:37:01 BBMASK: survive a non-existant path and don't keep the dir after globbing. Apr 08 17:37:01 Signed-off-by: Chris Larson Apr 08 17:37:37 03kergoth 07bitbake-1.8 * r1158 10/lib/bb/parse/parse_py/ (BBHandler.py ConfHandler.py): Apr 08 17:37:37 Make the bbpath change for includes temporary rather than permanent. Apr 08 17:37:37 We currently modify BBPATH so that an include/require looks in the current Apr 08 17:37:37 working directory before it looks elsewhere. Modifying it permanently Apr 08 17:37:37 doesn't gain anything, just polluting it with locations that don't need to Apr 08 17:37:40 stay there. Apr 08 17:37:42 Signed-off-by: Chris Larson Apr 08 17:42:53 I tried an older gdb and it gave that: Program exited with code 0140. Apr 08 17:43:09 => it doesn't work Apr 08 17:53:35 03Tom Rini  07stable/2009 * r9bf1072440 10openembedded.git/ (11 files in 2 dirs): (log message trimmed) Apr 08 17:53:35 gcc: Fix 4.1.[012], 4.2.[1234] and 4.3.[123] SRC_URI to use ${GNU_MIRROR} and update checksums.ini Apr 08 17:53:35 This fixes a big problem with 4.3.3 as it wasn't using ftp.gnu.org but Apr 08 17:53:35 an alternate mirror that's gone away. Since I had to fix one I noticed Apr 08 17:53:35 others not calling ${GNU_MIRROR} but ftp.gnu.org. Also a few weren't Apr 08 17:53:39 using ${PV}, so use that too. Apr 08 17:53:41 Acked-by: Denys Dmytriyenko Apr 08 17:55:52 oh good I'm glad that patch made it in :) Apr 08 17:56:35 kernel 2.6.29 needs a patch in order to be able to load modules on ARM, by the way Apr 08 17:56:54 not sure if this is something that goes in OE metadata though - it's fixed in 2.6.29.1 Apr 08 18:05:18 when i boot a gpe-image, i see flashes of the login screen on my display, i am not sure what could be wrong, it alternates between login screen and blank screen. all while i see a 'x' mark at the centre of lcd Apr 08 18:07:13 03Tom Rini  07org.openembedded.dev * r1caaca70e6 10openembedded.git/ (6 files in 2 dirs): Apr 08 18:07:13 mingw-binutils: Add 2.19.1 recipe and split common parts out into mingw-binutils.inc Apr 08 18:07:13 Signed-off-by: Esben Haabendal Apr 08 18:07:13 Signed-off-by: Tom Rini Apr 08 18:07:54 i think my xserver is getting killed and respawning again Apr 08 18:08:26 the pids of xserver keep changing Apr 08 18:08:53 what would make xserver to behave in such a fashion Apr 08 18:10:37 it's crashing Apr 08 18:10:41 or exiting Apr 08 18:10:52 do you really want us to list every possible cause of X not starting up? Apr 08 18:10:56 it'd be an awfully long list. Apr 08 18:11:26 i would be interested to know why? is there a way to find out? Apr 08 18:12:40 hey - is there a bitbake recipe for compat-wireless or the wireless-testing kernel tree? Apr 08 18:16:16 Hmmm Apr 08 18:16:29 Did the requirement for stuff to be in .dev first then stable get dropped? Apr 08 18:16:37 I thought it was on the Stable page but I don't see it Apr 08 18:17:17 Tartarus: I think everybody just assumed it was there ;-) Apr 08 18:17:38 heh, hi Esben :) Apr 08 18:17:43 we would like to make sure stuff goes in to .dev :) Apr 08 18:17:54 to avoid divergence as long as possible Apr 08 18:18:01 well serious divergence Apr 08 18:19:04 yes, but I hope to see divergence over time. I would like to be able to support products based on stable/2009 in let's say 5-10 years from now, and at that time I hope there will be almost nothing in common between stable/2009 and .dev ;-) Apr 08 18:19:32 which is why I don't see that requirement for going to .dev as a sustainable policy. Apr 08 18:19:47 but as of now, it is quite possible and sane. Apr 08 18:19:51 so you want us to ignore it and diverge right now, leaving dev without the fixes, instead of doing that for now? Apr 08 18:19:58 no. Apr 08 18:20:26 well i don't think anyone expects current policy to last for 10 years Apr 08 18:20:48 but I think we lack a good strategy for how to remember which commits on the .dev should be applied to stable later on. Apr 08 18:20:48 Esben, yes Apr 08 18:21:04 we are learning Apr 08 18:21:31 * Crofton|work hates Verizon Apr 08 18:21:47 perhaps we could use a strategy where we: 1) submit with [STABLE][PATCH] to list for review, 2) have it reviewed, 3) applied to .dev, 4) have it tested on .dev, 5) cherry-pick it to stable Apr 08 18:22:02 that's exactly what koen is doing today, esben Apr 08 18:22:17 I'm not sure what you think is actually happening right now, if not that Apr 08 18:22:35 maybe we could add some states to patchwork that helps control which patches are now in .dev, and which have been sufficiently tested, and needs to be moved to stable. Apr 08 18:23:02 kergoth, stuff from Koen in .dev doesn't usually first goto the list Apr 08 18:23:13 well, my mistake, it seems like 3) happens first, prior to the stable email Apr 08 18:23:20 (A short write up of the patchwork workflow would help, I'm unsure on how to help the proper way) Apr 08 18:23:23 But jeremy has been post for .dev the post for stable Apr 08 18:23:23 which maeks much more sense than esben's proposal, imo Apr 08 18:23:38 kergoth: you are getting quite close to what my concern is: I don't really see a clear picture of what is happening now. Apr 08 18:23:49 the problem is eventually, .dev and .stable will diverge Apr 08 18:24:13 i get the message whenever x crashes "*** glibc detected **** Xorg: free():invalid pointer:0x184e84 ***" on my lcd Apr 08 18:24:20 fixes go into dev, if they're seen as good and tested, they get cherry picked to stable. that's what's happenign today, not rocket science. when it diverges, it diverges, no point freaking out about it today Apr 08 18:24:56 But to be honest, I strongly believe that an agile/efficient procedure, know to everyone, usable to everyone, is a better strategy than "what koen is doing today".... Apr 08 18:25:25 * Esben is not freaking out Apr 08 18:25:25 * mickeyl agrees with esben Apr 08 18:25:38 fwiw, i have been proposing testing -> devel -> stable for a while, with only cherry-picking allowed Apr 08 18:25:54 but that doesn't seem to be attractive to many Apr 08 18:26:13 like in FreeBSD! :) Apr 08 18:26:16 * Jay7 hides Apr 08 18:28:16 Esben: I'm not sure what you seem to think is so difficult about getting it into dev, sending an email proposing its inclusion in stable, and then doing the cherry pick assuming it's approved. Apr 08 18:28:30 mickeyl, the problem is there is lots of stuff flowing into .dev that does not make .dev stable Apr 08 18:28:31 maintaining long lived bugfix branches isn't rocket science, CM people do it every day in a wide variety of scms Apr 08 18:29:32 Crofton|work: exactly. that's why there's no autoflow, but cherry-picking Apr 08 18:30:31 kergoth: did I say it was difficult? but I do dare to say that it is error prone to commit to .dev, have to wait a while, and then (without every forgetting) remember to ask for it to be cherry-picked an undefinite amount of time later on. Apr 08 18:30:36 but eventually, someone will submit something into .dev that makes .dev recipes stop building in .stable Apr 08 18:30:43 this isgood btw Apr 08 18:30:48 Esben: if you want it in stable, its your responsibility to push for it going into stable. Apr 08 18:31:03 if you don't, then let it be forgotten, and move on with your day Apr 08 18:31:06 again, not difficult Apr 08 18:31:12 or "error prone" Apr 08 18:31:14 can you explain what the mingw binutils stuff is? Apr 08 18:31:17 maintaining stable software is not rocket sciences, it is quite simple if efficient and not to error-prone procedures are used. Apr 08 18:32:08 Crofton|work: it is just a version of GNU binutils, hacked up a bit so it fit's MinGW. Apr 08 18:32:26 why do we need it? Apr 08 18:32:35 if you're too lazy to push for a commit going to stable, thats your problem. error-prone my ass. Apr 08 18:32:44 I'm not trying to be difficult, just trying to understand why it should go in Apr 08 18:32:45 kergoth: if things start to be forgotten, then stable will slowly die. I guess that is not what anyone wants. Apr 08 18:33:22 no matter what the workflow, stuff won't get commited if we do not understand it Apr 08 18:33:42 Ok, I see it is kind of difficult to discuss these things without some people getting personal. Sad. Apr 08 18:33:43 * Crofton|work really hates Verizon now Apr 08 18:34:45 Crofton|work: what is really easy to explain, is that mingw-binutils 2.19.1 is a newer version than mingw-binutils 2.17.xxxx which is in our metadata right now, and is needed to build MinGW canadian-sdk. Apr 08 18:35:06 ah Apr 08 18:35:17 * Crofton|work barely understands canadiam sdk's :) Apr 08 18:35:50 so there is no impact on anyone other than people using canadian sdk's? Apr 08 18:35:59 but it really is just a special branch of GNU binutils that happens to be able to link stuff so it works on win32 :-) Apr 08 18:36:22 I'm trying to understand how badly I would mess up stable if i push it :) Apr 08 18:36:42 nope, that is why it is now DEFAULT_PREFERENCE=-1 as it should not be used for anything else than mingw32 canadian sdk stuff. Apr 08 18:36:48 ok Apr 08 18:37:16 I need to do some paying work now, but I am trying to set aside some time for stable work in the next day Apr 08 18:37:20 I feel quite confident that it would be a bad idea to use it to link Linux binaries. Apr 08 18:37:36 if there are no outstanding comments and the acks are there I will push Apr 08 18:37:48 but, I need to do some real work first Apr 08 18:37:51 :) Apr 08 18:37:52 thanks. Apr 08 18:39:00 re Apr 08 18:39:07 thumb interworking is screwed in 2.19 it does not emit thumb to thumb trampolines correctly Apr 08 18:39:35 as OE seems to be more and more thumb than arm be careful when using it Apr 08 18:39:37 Crofton|work: question on patchwork Apr 08 18:39:47 yikes Apr 08 18:40:01 Crofton|work: should processed patches be archived? Apr 08 18:40:16 I think koen added a state for applied? Apr 08 18:40:25 denix0: koen's message said they should be changed to applied Apr 08 18:40:30 heh Apr 08 18:40:40 we should review the available states Apr 08 18:40:48 but I really need to do some work :) Apr 08 18:41:02 yeah, available states are not clear Apr 08 18:41:03 I sweep sometimes Apr 08 18:41:10 03Koen Kooi  07org.openembedded.dev * r3bd414d3c7 10openembedded.git/recipes/bluez/ (bluez4_4.31.bb files/bluetooth.conf): bluez4: fix dbus service file Apr 08 18:41:40 plus, archived is a flag, not a state. my assumption was once a patch is processed, change its state and set the archive flag Apr 08 18:42:37 but I see there are many patches in applied/accepted/rejected state, but not archived... so I'm not sure Apr 08 18:51:02 otavio: ping Apr 08 19:09:10 denix0: hi Apr 08 19:09:21 otavio: saw your message Apr 08 19:09:45 otavio: what do you want to do with it? Apr 08 19:10:05 denix0: well; revert it? Apr 08 19:10:10 :) Apr 08 19:10:32 it is kind of a good fix overall, but it breaks other things... Apr 08 19:10:38 hrw: ping! Apr 08 19:10:48 lets ask hrw about it... Apr 08 19:11:03 pong Apr 08 19:11:16 remember RPM commits? Apr 08 19:11:31 one of the commits breaks other stuff in oe... Apr 08 19:11:37 ok, revert that commit if it breaks stuff Apr 08 19:12:27 ah, I don't like reverting good commits just because it unmasks problems elsewhere... :) Apr 08 19:13:45 hrw: I take you've read this thread http://thread.gmane.org/gmane.comp.handhelds.openembedded/22376/focus=22556 Apr 08 19:14:01 curl 7.19 curl-sdk 7.18 one? Apr 08 19:14:07 yeah Apr 08 19:14:20 it also break something for otavio Apr 08 19:14:29 want me to write to ML that I Ack reverting? Apr 08 19:15:00 if you think reverting is best... Apr 08 19:15:17 I do not have a time to check Apr 08 19:15:28 ah, sorry Apr 08 19:15:32 revert and then Tartarus or me will check it one day Apr 08 19:16:03 denix0: how about fixing the real problem I havent read your post yet Apr 08 19:16:05 ok, thanks Apr 08 19:16:15 i want to install some good games which i can play with touch screen, are the games on oe heavy in size? Apr 08 19:16:20 khem: I'd like to fix it properly too Apr 08 19:16:54 denix0: I would suggest that we try to fix it before reverting Apr 08 19:17:08 khem: I agree Apr 08 19:17:23 the recipe for came_1.7 shows a dependency for giblib, but fails when building because of a missing giblib header file (i guess there is no -dev pkg made); is this a problem i can correct is it a problem with the came package maintainer ? Apr 08 19:17:57 hrw: how about we look for a proper fix and if not, then revert it later... Apr 08 19:18:07 assem, it sounds like do_stage isn't copying all of the headers Apr 08 19:18:21 Fix and post to the ML :) Apr 08 19:18:58 otavio: sorry for the delay. hope you don't mind waiting a little longer for a fix. I may need your help reproducing the issue... Apr 08 19:20:17 denix0: I believe your fix for curl is fixing the root cause in a way Apr 08 19:20:37 it only fixes curl Apr 08 19:20:59 but otavio is having a similar issue with other packages Apr 08 19:22:04 Tartarus: i'm pretty new at this - only made it through the first tutorials :) but i would like to help.. Apr 08 19:22:07 now, can there be more than 1 pkgdata/runtime files with the same name in different directories? Apr 08 19:24:47 are there any gurus in pkgdata stuff? :) Apr 08 19:27:07 i checkout the stable branch and i get the following error: ERROR: maximum recursion depth exceeded while calling a Python object while parsing /stuff/openembedded/recipes/sato-icon-theme/sato-icon-theme_0.1.bb Apr 08 19:27:17 bye Apr 08 19:27:20 is the repository broken? Apr 08 19:27:31 hrw|gone: bye Apr 08 19:27:48 hrw|gone: bye Apr 08 19:29:03 kante: which bitbake version do you use? Apr 08 19:29:35 denix0: let me build and I will try to see if I can find something Apr 08 19:29:52 khem: thanks Apr 08 19:29:58 1.8.13 Apr 08 19:30:26 otavio's problem is with something I've never built before... Apr 08 19:30:57 Esben: something wrong with my bitbake Apr 08 19:32:27 Unless you have a very good reason, I suggest you use the tags/bitbake-1.8.12 version. Apr 08 19:33:00 so if you are using branches/bitbake-1.8 or maybe even trunk, you should try 1.8.12 Apr 08 19:33:56 Esben: if i do change my bitbake version now, do i need to recompile all the packages i have until now? Apr 08 19:34:32 denix0: please explain Apr 08 19:34:46 otavio: what exactly? Apr 08 19:34:54 [16:30] < denix0> otavio's problem is with something I've never built before... Apr 08 19:35:48 otavio: lxde, openbox? I've never built those Apr 08 19:36:09 what's the target for those? Apr 08 19:36:10 denix0: it fails with gtk-engine-clearlooks too Apr 08 19:37:54 are those part of one target? Apr 08 19:50:55 denix0: tell me more about these filters on the output messages... :) Apr 08 19:55:12 kergoth: basically, some regexps to colorify/highlight the output and sometimes send notifications on specific events... Apr 08 19:56:36 * kergoth wonders how the frontend stuff in bitbake trunk is coming along Apr 08 19:57:25 so, I need to look closely to your changes for breakage to my regexps and if any event messages I'm interested in are changed/removed Apr 08 19:59:09 i think i only changed 2 messages, the rest were just switched from note to debug Apr 08 19:59:38 * kergoth didn't mean to slip the couple minor message format changes into that commit, shouldve been a separate one :P Apr 08 19:59:52 ah, ok. so if I need something I can get it back with -DD Apr 08 19:59:58 yup Apr 08 20:00:05 i figured it was best not to remove anything completely, yet Apr 08 20:00:21 fair enough Apr 08 20:00:34 sir, i want ro install a simple ftp server, ssh client, could anyone help me regarding the packages to select? Apr 08 20:00:51 ssh client - dropbear Apr 08 20:01:10 it is also ssh server Apr 08 20:02:09 openssh works better ;) Apr 08 20:02:20 but bigger Apr 08 20:02:22 i dont need a ssh server, i need a ftp server Apr 08 20:02:35 inetutils or inet-utils Apr 08 20:02:42 one of the two is the right name Apr 08 20:02:48 there are plenty of them Apr 08 20:02:54 proftpd Apr 08 20:02:56 vsftpd, proftpd? Apr 08 20:02:58 vsftpd Apr 08 20:03:10 bftpd Apr 08 20:03:34 Somebody built bftpd, but I don't know that they ever submitted the recipe to OE... I shoiuld go check that and dig it up, i guess. Apr 08 20:03:39 does any of net-kit/net-tools have one? Apr 08 20:03:55 there is bftpd in recipes Apr 08 20:04:20 inetutils has the original ftpd Apr 08 20:04:58 shoot, we're drowining in networking utils! :D Apr 08 20:04:59 i have a 64mb nand flash, i cant fit in all Apr 08 20:05:13 ;-) Apr 08 20:05:14 64mb or 64MB? Apr 08 20:06:03 64 mega bytes Apr 08 20:06:16 SlugOS fits in 8MB total flash space, with kernel and bootloader, and has the full openssh... another few MBytes and it will easily support the whole inetutils package. Apr 08 20:06:26 64MB is lots of space :) Apr 08 20:06:46 yep Apr 08 20:08:00 been doing 2MB flash / 8MB RAM for years with quite complete linux router/gateway system Apr 08 20:08:53 so, bootloader, kernel and filesystem were all fitting in 2MB flash (used dropbear for ssh though) Apr 08 20:08:59 03Jeremy Lainé  07stable/2009 * r138407b114 10openembedded.git/recipes/postgresql/ (4 files): Apr 08 20:08:59 postgresql: use LDFLAGS for shared libs (fixes QA error) Apr 08 20:08:59 Signed-off-by: Jeremy Lainé Apr 08 20:08:59 Acked-by: Denys Dmytriyenko Apr 08 20:16:50 03Jeremy Lainé  07stable/2009 * r0fe5da9271 10openembedded.git/recipes/gsoap/ (gsoap/rename_bogus_ldflags.patch gsoap_2.7.7.bb): Apr 08 20:16:50 gsoap: use LDFLAGS (fixes QA error) Apr 08 20:16:50 Signed-off-by: Jeremy Lainé Apr 08 20:16:50 Acked-by: Denys Dmytriyenko Apr 08 20:49:15 * kergoth wonders where zecke's at, silly people.. logging off of irc Apr 08 20:49:51 kergoth: He is still in .tw Apr 08 20:50:06 ahh Apr 08 20:50:10 cool Apr 08 20:59:37 03Michael 'Mickey' Lauer  07fso/milestone5.5 * rb524e4fb6d 10openembedded.git/packages/vala/ (4 files in 2 dirs): vala[-native]: add patch to make ByteArray usable (pending upstream inclusion) Apr 08 20:59:38 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r348b715757 10openembedded.git/packages/vala/ (5 files in 2 dirs): Apr 08 20:59:38 vala[-native]: 0.5.6 -> 0.5.7 Apr 08 20:59:38 remove patches applied upstream Apr 08 20:59:41 03Michael 'Mickey' Lauer  07fso/milestone5.5 * re48b7b0f35 10openembedded.git/packages/vala-terminal/vala-terminal_git.bb: vala-terminal: bump version number Apr 08 20:59:44 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r7c424a82c5 10openembedded.git/packages/vala/ (7 files in 2 dirs): vala: add another patch against glib-2.0.vapi Apr 08 20:59:47 03Michael 'Mickey' Lauer  07fso/milestone5.5 * rc209f3740a 10openembedded.git/conf/distro/include/preferred-om-2008-versions.inc: preferred-om-2008-versions.inc: bump vala and libtiff Apr 08 20:59:50 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r099f346b5d 10openembedded.git/packages/vala/vala.inc: vala: touch gobject.vala.stamp before compiling; apparantly there's a timeglitch in the tarball. Closes OE Apr 08 20:59:56 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r624b08487e 10openembedded.git/classes/ (fso-plugin.bbclass vala.bbclass): Apr 08 20:59:59 fso-plugin.bbclass: new class for fso plugin packaging Apr 08 21:00:01 vala.bbclass: new class for vala-based packages Apr 08 21:00:03 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r5e4119af28 10openembedded.git/conf/distro/include/sane-srcrevs.inc: sane-srcrevs: bump EFL and add known-good for vala-native-git Apr 08 21:00:08 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r6850d2c9e1 10openembedded.git/packages/vala/ (vala-native_0.5.7.bb vala-native_git.bb vala.inc): vala-native: add git version (DP=-1) Apr 08 21:00:11 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r92fef63724 10openembedded.git/conf/distro/include/fso-autorev.inc: fso-autorev: add libeflvala Apr 08 21:00:14 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r2753c1d7ff 10openembedded.git/packages/vala/vala-native_git.bb: Apr 08 21:00:19 vala-native: build git version from freesmartphone.org which has some additional patches on top Apr 08 21:00:21 (always sent upstream for inclusion) Apr 08 21:00:23 03Michael 'Mickey' Lauer  07fso/milestone5.5 * rc692c88e63 10openembedded.git/ (3 files in 2 dirs): vala-native: bump to 0.6.0 and make git version the default; we need a bootstrap version now though. Apr 08 21:00:27 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r666af4b426 10openembedded.git/ (2 files in 2 dirs): libeflvala: new recipe; Vala bindings for the Enlightenment Foundation Libraries Apr 08 21:00:32 03Michael 'Mickey' Lauer  07fso/milestone5.5 * rb64eb17349 10openembedded.git/packages/vala/vala-bootstrap-native_0.6.0.bb: vala-bootstrap-native: bring back timestamp hack; still necessary Apr 08 21:00:35 03Tom Rini  07fso/milestone5.5 * r6f959af351 10openembedded.git/conf/checksums.ini: checksums.ini: Add vala-0.6.0 Apr 08 21:10:10 03Koen Kooi  07fso/milestone5.5 * reb90a0063a 10openembedded.git/packages/bluez/bluez4_4.33.bb: bluez4: also depend on libusb-compat Apr 08 21:19:19 kergoth: about cooker.py..likewise remebered recently http://bugs.openembedded.net/attachment.cgi?id=13797 . Is already fixed? Apr 08 21:20:12 don't know Apr 08 21:20:38 go :) Apr 08 21:29:54 in a gpe-image when i click on the settings->screen setup->touch calibration, nothing happens, is this the default behaviour? Apr 08 21:31:24 vsr|work: ehm..no Apr 08 21:32:47 may be just appropriate package is not installed? Apr 08 21:33:00 is there a way i can debug this, it is supposed to call xtscal Apr 08 21:35:04 when i run it on console gpe-conf screen, i get the message: XCALIBRATE extension missing: Success Apr 08 21:35:29 what does it mean? i see xtscal app on my system Apr 08 21:38:28 vsr|work: are you using kdrive? Apr 08 21:38:40 seems x-server is build w/o this extension Apr 08 21:38:45 or something like Apr 08 21:38:56 but I'm not sure Apr 08 21:39:33 jay7: i am using x.org Apr 08 21:41:51 try to remove /etc/pointercal and reboot Apr 08 21:42:26 xtscal does not workl i get the same error when i run xtscal, XCALIBRATE extension missing: Success Apr 08 21:43:39 Jay7: btw Zaurus keyboard is still broken in GPE... Apr 08 21:43:52 ant__: xmodmap? Apr 08 21:44:01 strikes again Apr 08 21:44:05 or keymap-2.6? Apr 08 21:44:22 first Apr 08 21:44:37 I've read about intermittent bug..funny Apr 08 21:44:40 i use the same gpe-image.bb file in oe repo. should i change it to take xserver-kdrive Apr 08 21:45:59 Jay7: http://www.oesf.org/forum/index.php?showtopic=26598 Apr 08 21:46:04 utx = Stanislav Apr 08 21:46:24 * Jay7 is still recompiling, debugging, fixing and recompiling againg :) Apr 08 21:47:23 ant__: yeah, I have noted this too Apr 08 21:47:57 I'm thinking already about implementing native zaurus touchscreen support in kdrive Apr 08 21:48:09 just to drop tslib dependency Apr 08 21:48:21 but I'm unsure why no one done this already.. Apr 08 21:49:03 why would you want to? you'd just have to reimplement the event filtering to remove noise yourself Apr 08 21:49:10 seems pointless Apr 08 21:50:00 or so :) Apr 08 21:51:56 hello Apr 08 21:51:56 jay7: are you suggesting me to select xserver-kdrive? Apr 08 21:52:11 vsr|work: I don't know :( Apr 08 21:52:20 but you can try at least Apr 08 21:52:48 jay7: i have given a build Apr 08 21:52:51 can someone help me on this ? I have openssl-native making parse errors: Apr 08 21:52:54 http://pastebin.com/m61de5070 Apr 08 21:54:09 the recipes are from oe Apr 08 21:58:26 i386-darwin hmm do you use a mac Apr 08 22:00:00 what does "?=" mean. does it mean. if not already defined use it? Apr 08 22:00:02 ptitjes: I dont know if many of us use macs here so not many have seen such an error Apr 08 22:00:16 vsr|work: yes its a weak assignment Apr 08 22:00:43 how do i know what is variable XSERVER set to? Apr 08 22:00:54 khem: I use Gentoo-Prefix above a Mac Apr 08 22:01:03 khem: so this should not be a problem Apr 08 22:01:17 khem: the only thing is don't have a linux kernel Apr 08 22:01:22 i know which .bb file it comes from Apr 08 22:01:42 khem: anyway I don't see why this should parse differently Apr 08 22:02:12 ptitjes: ERROR: Information not available for target 'i386-darwin' Apr 08 22:02:18 ptitjes: http://cgit.openembedded.net/cgit.cgi?url=openembedded/log/&h=koen/darwin-host-fixes Apr 08 22:02:38 probably you already know Apr 08 22:03:29 ant__: what is that ? I must use a different oe branch ? Apr 08 22:04:52 khem: what information does it want ? Apr 08 22:05:28 do I miss some commands ? Apr 08 22:06:06 just cherry-pick the patches Apr 08 22:06:08 ;; Apr 08 22:06:08 + darwin-i386) Apr 08 22:06:08 + target=darwin-i386 Apr 08 22:06:08 + ;; Apr 08 22:06:08 esac Apr 08 22:06:16 btw.. Apr 08 22:06:36 I'm too lazy and actually did not yet test on LEopard Apr 08 22:06:41 ptitjes: I think openssl needs to know your triplet Apr 08 22:06:56 ant__: couldn't I fool oe making it think I'm not mac ? Apr 08 22:06:59 yep...openssl: add i386-darwin to known buildhosts Apr 08 22:07:01 ptitjes: it seems it does not know i386-darwin Apr 08 22:07:10 ant__: because in fact I just don't have the kernel Apr 08 22:07:25 ant__: the whole stack above is based on coreutils Apr 08 22:07:35 ptitjes: I'm sorry...iirc just Koen is building on Mac Apr 08 22:07:58 I can't give you direct info Apr 08 22:08:08 but why do oe needs to know about the kernel. I don't get it Apr 08 22:08:18 ptitjes: its not oe Apr 08 22:08:23 its a package openssl Apr 08 22:08:36 khem: so it is not a parsing error ? Apr 08 22:09:09 ptitjes: its a parsing error, I think it could have let it go and let openssl-native fail Apr 08 22:09:35 khem: could I provide openssl myself ? Apr 08 22:09:53 there are lots of issues with building OpenEmbedded on darwin Apr 08 22:09:55 *lots* Apr 08 22:09:57 in fact I already got it on my Gentoo Apr 08 22:09:59 koen has a branch thats a starting point Apr 08 22:10:03 but there are quite a few other issues Apr 08 22:10:06 i have a list of them here.. Apr 08 22:10:14 kergoth: I don't use Darwin Apr 08 22:10:22 kergoth: I did build once like ages back Apr 08 22:10:24 kergoth: I use Gentoo-Prefix Apr 08 22:10:47 ah, well, it chooses the OS from uname Apr 08 22:10:54 so why is your uname showing darwin? Apr 08 22:11:07 the triplet is not changed because it is the kernel is not changed Apr 08 22:11:33 either deal with the fact that it can't build on darwin, or override the BUILD_OS to Linux Apr 08 22:11:41 er, linux Apr 08 22:11:55 kergoth: cool in what conf file must I do this ? Apr 08 22:12:02 who cares? Apr 08 22:12:05 local.conf ? Apr 08 22:12:26 if that's where you want it, sure Apr 08 22:12:30 bitbake doesn't care Apr 08 22:12:36 * ptitjes is oe newbie Apr 08 22:12:42 why dont you want this patch http://cgit.openembedded.net/cgit.cgi?url=openembedded/commit/&h=koen/darwin-host-fixes&id=1462ede86c53ec7bf8858369467aeb57e68b0577 Apr 08 22:12:46 bitbake parses conf/bitbake.conf, which includes all the other .conf files in a specific order Apr 08 22:13:30 khem: I don't know how to do it properly Apr 08 22:13:42 khem: I use the SHR Makefile Apr 08 22:15:27 ptitjes: ugh, do you have openssl.inc somewhere Apr 08 22:16:38 * ant__ is reading how big the hack is (Gentoo Prefix) Apr 08 22:16:43 khem: you mean doing it by hand but then I'll have problems while updating... Apr 08 22:17:19 ant__: As I feel it the whole oe thing should have been based on Gentoo-Prefix Apr 08 22:17:32 put down the crack pipe Apr 08 22:17:46 but I don't want to launch a flame war ;) Apr 08 22:17:54 ptitjes: so what, you dont want any patches Apr 08 22:18:18 khem: how can I put them somewhere so that when I update it is reapplied transparently ? Apr 08 22:18:24 ptitjes: fyi, we *did* base on portage, and then removed a ton of its aspects that sucked ass for cross development Apr 08 22:18:43 kergoth: crossdev do not suck Apr 08 22:18:54 kergoth: it is rock solid Apr 08 22:19:20 ptitjes: what brings you to OE Apr 08 22:19:31 anyway that is not the point... I'm forced to use oe for its bad and goods. so I accept it :) Apr 08 22:19:37 using portage to build for embedded is like building a car engine with some string and a drinking straw Apr 08 22:19:39 weak. Apr 08 22:20:39 kergoth: hum... Apr 08 22:20:47 i respect the gentoo embedded guys, solar is one of the most clueful embedded people i know, but i wouldnt' touch that shit with a ten foot pole Apr 08 22:21:04 khem: so ok I apply the patch by hand Apr 08 22:22:49 anyone looked at e2factory yet? Apr 08 22:23:25 kergoth: I never had so much problems I had with 6 months using oe that I had using portage for 6 years Apr 08 22:24:20 good for you Apr 08 22:24:40 ptitjes: while portage is very well-suited for the build-host itself, it cannot cope as well with crosscompiling Apr 08 22:25:08 ptitjes: I use Gentoo since it's origins Apr 08 22:25:20 ant__: I don't see your point here. why is it not suited for it ? Apr 08 22:25:26 right. i don't have anything against portage, though i do think the old python codebase is a pile of crap, and I'm glad to see alternative implementations now, but i think trying to use it for what you use OpenEmbedded for is badly stretching it beyond its intended use and capabilities. Apr 08 22:25:53 nod Apr 08 22:25:56 not to mention gentoo's package database isn't suited to our highly granular binary packaging to save space on disk, they just hack it by rm -rf'ing man pages and crap.. guh Apr 08 22:26:21 kergoth: it's ok for a server and a good sysadm ;-) Apr 08 22:26:26 hehe Apr 08 22:26:57 I can even develop OE hassle-free on it ... Apr 08 22:27:01 :] Apr 08 22:27:02 OpenEmbedded has its issues, believe me, you should see my list of concerns with the current implementation and design, but i still think it's better than any of the current alternatives Apr 08 22:28:34 http://docs.google.com/Doc?id=dzb388z_67cb57qmgq exported from my omnioutliner doc Apr 08 22:28:41 well the patch doesn't fix it. I tried darwin-i386, i386-darwin. nothing does. do I miss something ? Apr 08 22:28:59 do I need to trigger a checksum regeneration or something ? Apr 08 22:29:25 aare you trying to build for OS X? Apr 08 22:29:45 good luck :) it has ooo some issues :) Apr 08 22:30:03 timtimred: no my target is gta02 Apr 08 22:30:09 darin? Apr 08 22:30:11 darwin? Apr 08 22:30:28 my host's host is darwin Apr 08 22:31:53 ok :) Apr 08 22:32:11 (wow peopel actually run darwin...( Apr 08 22:32:13 :) Apr 08 22:32:34 Jay: selecting xserver-kdrive worked, thank you Apr 08 22:33:26 Jay7: im happy, i can move to the next problem now Apr 08 22:33:31 vsr|work: :) Apr 08 22:33:34 khem, kergoth: setting BUILD_OS=linux made the parse error go away Apr 08 22:33:41 vsr|work: so you see xserver-xorg is broken? Me and Jay7 are using kdrive too... Apr 08 22:33:54 hm.. Apr 08 22:34:09 ant_, Jay7: rotation also works Apr 08 22:34:18 so we can't use xserver-xorg.. Apr 08 22:34:35 eh..there were proposals for moving to it Apr 08 22:34:45 ant__: I remember that :) Apr 08 22:34:52 also could someone point me to the sato-icon-theme fix. I heard it is a know problem, am I wrong ? Apr 08 22:35:15 Jay7: probably, i did not supply arguments to x.org like xf86-input-evdev Apr 08 22:35:17 xmodmap troubles on one side and non-working calibration and rotation on other Apr 08 22:36:05 jay7: should i have supplied arguments to xserver as shown in images/beagleboard-demo-image.bb? Apr 08 22:37:51 vsr|work: I'm not an xserver-xorg guru :) Apr 08 22:38:30 jay7: sorry Apr 08 22:38:42 kergoth: depends.dot is still 'broken' (bitbake -g xyz) Apr 08 22:38:52 try and look - better strategy :) Apr 08 22:38:55 thats not good. there a bug open for it? Apr 08 22:38:59 yes Apr 08 22:39:00 * kergoth hasn't looked at that code at all Apr 08 22:39:18 ant_: it works for me Apr 08 22:39:26 how many lines??? Apr 08 22:39:43 i see a huge file, really huge file Apr 08 22:39:51 he Apr 08 22:40:35 kergoth: http://bugs.openembedded.net/show_bug.cgi?id=5002 Apr 08 22:40:35 ant_: let me check again, i did that 3 days ago, i will let you know once i finish this build Apr 08 22:40:44 vsr|work: ^^ Apr 08 22:41:28 * kergoth wonders if gerritt or reviewboard would be useful in the context of OpenEmbedded Apr 08 22:41:34 and check the fun out of contrib/dependsgraph/dependsgraph.sh Apr 08 22:43:43 has gnome-games got any relation with gpe? can i have a image with both gpe and gnome-games? will it fill my rootfs with unwanted libraries and executables? Apr 08 22:45:22 \o/ yahooo Apr 08 22:52:17 i tired to build gnome-mplayer and it says package requirements not met it is not just one package: please see http://pastebin.com/mb496f2d Apr 08 22:52:40 i checkout stable git repo. arent all the package versions aligned in that? Apr 08 23:00:45 ant_, jay7: could you help? Apr 08 23:03:06 vsr|work: sorry canno test here...rebuilding gcc-4.3.2 on host... Apr 08 23:03:25 Requested 'libecal-1.2 >= 1.6.0' but version of libecal is 1.4.2.1 Apr 08 23:03:27 ^^^ Apr 08 23:03:55 yes, but the libraries versions should be aligned isnt it? Apr 08 23:04:00 in the stable tree Apr 08 23:04:01 wel.. Apr 08 23:04:20 perhaps nobody yet built that recipe Apr 08 23:04:54 gnome-panels? Apr 08 23:05:15 hm..http://tinderbox.openembedded.net/packages/ Apr 08 23:05:28 i dont find the recipe of libecal Apr 08 23:05:53 2009-04-08 22:59:56.697227 angstrom-autobuilder 2009-04-08 15:09:00.097553 gnome-mplayer 0.5.3+cvs20080101 r5 Succeeded Apr 08 23:06:00 ant__: i cant access that site, damn company firewall Apr 08 23:06:40 ant__: what do you suggest? Apr 08 23:10:58 vsr|work: try to check eds-dbus_svn Apr 08 23:12:08 what does mean "env: wget: No such file or directory" ? I have wget... I also get some "sh: git: command not found" but I also have git... Apr 08 23:12:36 I added --verbose but this does not add much inormations Apr 08 23:15:06 'nite all Apr 08 23:19:10 no idea ? Apr 08 23:21:27 sorry, newbee here Apr 08 23:22:01 ptijes: have you install all pre-requisites mentioned in getting started document? Apr 08 23:22:30 err what getting started document ? Apr 08 23:23:05 http://wiki.openembedded.net/index.php/Getting_Started Apr 08 23:23:26 vsr|work: you mean I need other things that could make it complain it does not have wget and git Apr 08 23:23:54 ptijes: what are you trying to do? Apr 08 23:24:05 03Angus Ainslie  07fso/milestone5.5 * rc884619ae6 10openembedded.git/packages/vala/vala-native_0.5.7.bb: vala-native : remove old version Apr 08 23:25:05 vsr|work: a simple bitbake -c build libfso-glib Apr 08 23:25:53 ptijes: sorry i dont know Apr 08 23:26:11 ptijes: i dont want to do any damage to your setup Apr 08 23:26:19 wait for an expert to turn up Apr 09 00:03:10 can i fit claws-mail in 640x480? Apr 09 00:09:26 why does kernel.bbclass do_deploy() create a tarball of kernel modules in DEPLOY_DIR_IMAGE? shouldn't all your kernel modules be in your image by virtue of image/task configuration? Apr 09 00:10:38 tharvey: convenience I think. Apr 09 00:11:01 And, no not all kernel mods will be in the final image; some may be isntallable via opkg later on. Apr 09 00:11:34 if i dont want any kernel modules, what do i need to do? Apr 09 00:11:47 well right, but my point is you shouldn't untar the modules tarball into your root, you should use ipkg Apr 09 00:11:55 so I was just curious why this was an output Apr 09 00:11:58 whenever i build a image, i get modules of all the kernels .26 .27 .28 Apr 09 00:12:14 But I certainly appreciate having a separate kernel , and a tarball of modules to install in (for example) a rootfs for nfs mounting -- good for setting up a rootfs to debug startup/init problems, or debug a new kernel. Apr 09 00:12:27 vsr|work, if you want to minimize build time disable them from your kernel good Apr 09 00:12:49 vsr|work: then something is very much wrong with your OE build env, because it should only build a single kernel, and should only install the modules you have instructed it to install into the rootfs/image. Apr 09 00:13:11 how do i disable oe from building/putting the kernel images of other versions which i might have built earlier into final image Apr 09 00:13:40 vsr|work: My question would be how is it that you are managing to get older versions into your final image?!! Apr 09 00:13:59 i dont know Apr 09 00:14:39 The do_rootfs step should remove any existing rootfs dir, and proceed to assemble a new one, so the only way old stuff would get in is if your recipes are explicitly including them -- and it is not possible, as far as I know, to reference more than one version of any package, and that includes the kernel! Apr 09 00:15:56 mwester:i just build gpe-image, all the 3 kernels which i have build at different points of time and thier modules get into rootfs Apr 09 00:16:45 vsr|work: quite borken, I think. You might like to ask on the mailing list to see what might be wrong. Apr 09 00:18:35 vsr|work what bitbake recipe are you building? Apr 09 00:19:12 perhaps your seeing previously built kernels - gpe-image will just have 1 kernel built and output Apr 09 00:19:51 kernel's get rev'd over time so if you have been keeping in sync with git and not doing cleans you could be accumulating old kernel builds Apr 09 00:24:10 i build gpe-image Apr 09 00:39:38 arg Apr 09 00:39:43 OpenEmbedded hates me today Apr 09 02:56:19 So when building qemu-native I'm getting the following error: Apr 09 02:56:28 /usr/bin/ld: cannot find -lm Apr 09 02:57:12 <~/Misc/phone/oe/build>$ ls -l /usr/lib/libm.so Apr 09 02:57:12 lrwxrwxrwx 1 root root 19 2009-01-29 11:06 /usr/lib/libm.so -> ../../lib/libm.so.6 **** ENDING LOGGING AT Thu Apr 09 02:59:57 2009