**** BEGIN LOGGING AT Tue Feb 26 10:59:56 2008 Feb 26 11:00:02 mrt2k7: well, you mix bitbake versions? you use the 'bitake' binary from one version but your libs are from another? did you install it once and now use it svn directly? Feb 26 11:01:16 i use bitbake from debian 4.0 Feb 26 11:01:28 1.8.10-3 Feb 26 11:01:49 should I use bitbake from svn? Feb 26 11:02:19 hi zecke Feb 26 11:03:17 wooohha...looks good Feb 26 11:03:41 now bitbake actually seems to do something for the very first time :) Feb 26 11:11:28 has someone tried to use OE for building a minimal OS for st. like VMware? Feb 26 11:12:14 mrt2k7: like qemuarm MACHINE :-) Feb 26 11:12:31 I had a look at several frameworks like GNAP, Embdebian, T2, ... but OE is the first that looks actually usable Feb 26 11:13:02 XorA: i was wondering what qemu in the conf is about...what is it for? Feb 26 11:13:13 in this context Feb 26 11:13:25 do you simulate ARM with qemu? Feb 26 11:13:57 yes Feb 26 11:14:11 there is a qemux86 target as well I beleive Feb 26 11:14:34 hmm cant find it in the wiki Feb 26 11:15:01 ls conf/machine/qemu* Feb 26 11:15:27 aahh thx Feb 26 11:15:31 was looking in distro Feb 26 11:15:40 hey, theres vmware too Feb 26 11:15:42 thx :) Feb 26 11:16:51 so what do you get in the end... a diskimage or iso-image? Feb 26 11:17:37 hi, all! Feb 26 11:18:02 could anybody link me to autobuilder doc (mainly where to get logs)? Feb 26 11:20:07 mrt2k7: a tarball or ext2 image from memory, but I dont 100% know Feb 26 11:24:36 mickeyl: alive? Feb 26 11:25:02 slowly coming back to live, yeah Feb 26 11:25:13 like always, FOSDEM was incredibly exhausting Feb 26 11:25:38 mickeyl: g'day Feb 26 11:26:02 hi pb_ Feb 26 11:26:18 mickeyl: so we are e.V. now? Feb 26 11:26:21 the amazing thing is i did manage not to get sick (which is a first time after FOSDEM) Feb 26 11:27:02 heh, congratulations Feb 26 11:27:03 hrw: not yet, we did a good share, but there's some bureaucracy pending Feb 26 11:27:21 ok Feb 26 11:27:23 hrw: i expect us to be installed by April/May. Feb 26 11:27:27 will keep all of you posted Feb 26 11:28:15 mickeyl : Hi ! Feb 26 11:29:35 hi steliosk Feb 26 11:49:11 hi Feb 26 11:49:26 hi zecke hrw mickeyl steliosk Feb 26 11:50:41 i was trying to "Geting Started" with the OE, but my purpouse is to personalize a full image for a HW evaluation Board Feb 26 11:51:19 there is a lot of documentation, but i can not find any way to learn the dependencies and how to resolve it Feb 26 11:51:43 methril: what dependencies do you mean? the ones of your build system? Feb 26 11:52:07 yes, but first I'm trying to understand the other ones Feb 26 11:52:39 i'm looking into openmoko builds, for example Feb 26 11:54:33 I did not try building moko stuff, but in general all deps that are needed (well, maybe except for the latest bitbake) are in the distro repos Feb 26 11:54:49 in Fedora you can yum install all deps required for building Feb 26 11:55:02 I'm sure this is also the case for other distros Feb 26 11:55:33 it's not the host dependencies what i'm traying to understand Feb 26 11:55:39 are the target ones Feb 26 11:55:52 that's defined by the distro that you choose / image that you build Feb 26 11:56:03 k Feb 26 11:56:08 thks Feb 26 11:57:44 if you try to start a new image, do you have any clue where to start? :D Feb 26 11:57:54 hi chouimat|work Feb 26 11:58:04 methril: my approach was to start from angstrom Feb 26 11:58:56 but inside the conf you've the distros and the machine Feb 26 11:58:57 methril: I did something similar that Moko does - I pretended that my distro is angstrom and then added some custom dependencies and some custom image tasks Feb 26 11:59:03 methril: pick one image you want to copy, make sure it is working (try compiling the original), then copy and adjust? Feb 26 11:59:19 k Feb 26 11:59:20 yeah, I guess zecke explained it in an easier way :) Feb 26 11:59:28 :) Feb 26 12:00:21 zecke: Did you see the mail about the exec_task changes? Feb 26 12:01:25 i have it in my mailbox Feb 26 12:01:42 food => packing => flying => taipei is more urgent :} Feb 26 12:02:06 zecke: np ;-) Feb 26 12:02:33 zecke: RP search for someone who is able to understand that mail Feb 26 12:02:42 I failed Feb 26 12:02:44 you should blog and do podcasts Feb 26 12:02:59 if you try to add this to a new Hardware? You 've to find a similar one, copy and change some things? Feb 26 12:03:35 methril: mostly yes, find a machine that is similar to yours, adapt it to fit your needs Feb 26 12:04:19 actually same for the distro Feb 26 12:04:19 hrw, zecke: I think to move forward on several bugs/improvements we have to do this and I was just after to sanity check my reasoning. I'm sure it will be fine though ;-) Feb 26 12:04:41 RP: ;) Feb 26 12:04:49 * hrw use bitbake 1.6.6 now Feb 26 12:06:35 I tried modifying the local.conf file in org.embedded.snapshot/conf but my changes dont get applied...any ideas why? Feb 26 12:06:45 RP: read it, sounds sane Feb 26 12:07:23 pbne04: IIRC, it needs to go in build/conf/ Feb 26 12:07:33 zecke: thanks :) Feb 26 12:07:35 RP: I have to see why exec_task is broken after the removal of dirgraph and can't be fixed Feb 26 12:08:16 NAbyss: there is no such thing as 'build/conf' in OE - this is not Poky ;d Feb 26 12:08:35 NAbyss: ah thanks Feb 26 12:08:57 zecke: exec_task will continue to work and be used, it just won't check dependencies. Dependencies become the responsibility of runqueue Feb 26 12:08:57 hrw: There isn't? Bloody hell, I haven't built angstrom in a while then :) Feb 26 12:09:40 NAbyss: what about the conf files I added to distro and machine dirs? they stay there? Feb 26 12:09:47 NAbyss: I have ~/devel/build/{chinook,angstrom,angstrom-2007.12,angstrom-uclibc}/conf/ Feb 26 12:17:06 RP: and for StampUpdate we have this array of stamps in bitbake anyway? Feb 26 12:40:18 what do i do when bitbake cant download some files? Feb 26 12:40:42 -> Connecting to ftp.rpm.org|209.132.176.171|:21... failed: Connection timed out. Feb 26 12:41:14 (package popt) Feb 26 12:42:35 <_diego_> there is a way to use ipkg to install from a local repositary? Feb 26 12:43:30 hmm so i would have to get the source somewhere else and use ipkg before bitbake task-base? Feb 26 12:44:39 zecke: yes, its kept as part of the cache Feb 26 12:46:32 mrt2k7: use angstrom as distro so it will be fetched from source mirror Feb 26 12:46:42 i can't build python-edbus because a bb for dbus-glib-1 - i think mickeyl did something here, is this bb just not checked in or did i something wrong? Feb 26 12:48:09 emdete: s/dbus-glib-1/dbus-glib/ and check Feb 26 12:49:15 hrw: ? Feb 26 12:49:43 in python-edbus recipe Feb 26 12:49:57 hm, ich meine da waer eine abhaenggkeit, dass das genau nicht geht Feb 26 12:50:01 ...da war was Feb 26 12:50:23 ja, naturlish Feb 26 12:50:28 aber mickeyl muss ja kuchen essen! :D Feb 26 12:50:39 emdete: /join #oe-de please Feb 26 12:50:49 oups... sry ;) Feb 26 12:51:30 ich sprech Deutsch nicht (or something like that) Feb 26 12:52:21 hrw: i just switched from neo-germany here and forgot. sorry. i just said mickeyl has to have coffey & cake Feb 26 13:06:05 03mickeyl 07org.oe.dev * r70282542... 10/ (1 packages/s3c2410-utils packages/s3c24xx-utils): s3c2410-utils: rename to s3c24xx-utils Feb 26 13:06:15 03mickeyl 07org.oe.dev * r4338a15e... 10/ (3 files in 3 dirs): add gpio, a userspace utility to tweak the GPIOs on the s3c24xx platform Feb 26 13:06:49 mickeyl: that 'gpio' tool is s3c24xx only? Feb 26 13:07:20 yeah Feb 26 13:07:39 i don't know what would be needed to make it platform agnostic Feb 26 13:07:45 our kernel hackers did that Feb 26 13:07:55 kind of pxaregs for pxa? Feb 26 13:07:59 similar, yeah Feb 26 13:08:14 why not s3c-gpio then? Feb 26 13:08:25 you mean the name of the binary? Feb 26 13:08:30 emdete: what's your problem with python-edbus? it builds fine here Feb 26 13:08:31 and recipe Feb 26 13:09:18 yeah, might be good Feb 26 13:09:33 mickeyl: maybe a bb is missing in repo? Feb 26 13:09:50 emdete: always possible. what error do you get? Feb 26 13:10:08 dbus-glib-1 missing as far as i remember Feb 26 13:10:33 at package dependency time, build time, run time? Feb 26 13:10:41 bitbake time Feb 26 13:10:52 DEPENDS += "edbus python-dbus" Feb 26 13:10:59 one mom, checking Feb 26 13:11:15 k Feb 26 13:12:07 it's while configure: checking for DBUS_GLIB... configure: error: Package requirements (dbus-glib-1 >= 0.70) were not met: No package 'dbus-glib-1' found Feb 26 13:12:27 aaah Feb 26 13:12:28 build time Feb 26 13:12:46 ok, so this comes while you're building python-edbus? Feb 26 13:13:16 yes Feb 26 13:14:57 ok, so lets check the dependency tree Feb 26 13:15:16 python-edbus depends on edbus and python-dbus Feb 26 13:15:28 aah Feb 26 13:15:42 righto Feb 26 13:15:48 edbus only depends on dbus Feb 26 13:15:50 of course Feb 26 13:15:53 not on dbus-glib Feb 26 13:16:09 so there's a missing dependency Feb 26 13:17:13 can u fix that? is there a workaround for me? Feb 26 13:17:19 will fix in a minute Feb 26 13:17:26 okay. Feb 26 13:18:53 the workaround is to just build dbus-glib before python-dbus, but i'm fixing the python-dbus dependency as we speak Feb 26 13:20:18 gm Feb 26 13:20:28 heya Crofton ! Feb 26 13:20:29 anyone else that went to FOSDEM sick? Feb 26 13:20:39 not this time! Feb 26 13:20:41 * mickeyl dances a jig Feb 26 13:20:49 first time ever I return healthy after a fosdem Feb 26 13:20:51 I think I have the FOSDEM disease then Feb 26 13:20:55 bummer Feb 26 13:20:57 yeah Feb 26 13:20:59 just a cold? Feb 26 13:21:04 yeah Feb 26 13:21:08 could be worse then Feb 26 13:21:11 flight was ok? Feb 26 13:21:12 but awful to travel with Feb 26 13:21:14 yeah Feb 26 13:21:15 i can imagine Feb 26 13:21:16 no problems Feb 26 13:21:29 I was poping decongestant all day Feb 26 13:22:26 another billion email about Android on OMAP .... Feb 26 13:22:31 heh, yeah Feb 26 13:22:36 ridicolous Feb 26 13:22:38 on the omap list Feb 26 13:23:04 I really need to stop working for money and do some OMAP work like getting openmoko running on the OSK Feb 26 13:23:13 can you run OM in 32M of flash and ram? Feb 26 13:23:49 Did anyone run into apaulsen at FOSDEM? Feb 26 13:24:29 i don't think so Feb 26 13:25:05 Crofton: 32MB flash should work if you throw away some things, but 32MB of RAM is impossible Feb 26 13:25:32 bummer Feb 26 13:25:32 you'll have a hard time with X on anything below 64MB Feb 26 13:25:37 so no OM on OSK Feb 26 13:26:25 I'm afraid so. at least not right now. my secret plan is to have a full EFL-based environment that allows us to use the framebuffer backend on low-end devices Feb 26 13:26:31 crofton: even 64MB RAM it doesnt work Feb 26 13:26:32 but that's probably a good number of months away Feb 26 13:26:40 ok Feb 26 13:26:41 good Feb 26 13:26:48 I can work on other cool things :) Feb 26 13:26:59 just got a package of crack from sakoman Feb 26 13:28:11 more gumstix goodness? Feb 26 13:28:54 another lcd board Feb 26 13:29:00 excellent Feb 26 13:29:01 likely with fixed touchscreen Feb 26 13:29:31 and most importantluy a jtag baord Feb 26 13:29:38 SO I can debrick stuff locally Feb 26 13:30:19 ~curse xserver 1.4 Feb 26 13:30:19 May the fleas of a thousand camels infest your most sensitive regions, xserver 1.4 ! Feb 26 13:30:35 hrw, gm, How is fatherhood? Feb 26 13:31:10 Crofton: my girls will be at home on Friday (if Mira results on Tuesday will be fine) Feb 26 13:32:09 thats good Feb 26 13:36:42 which filesystem most pda's that run linux use? Feb 26 13:36:50 jffs2 on flash Feb 26 13:36:53 jffs? Feb 26 13:37:54 i mean on sdmmc Feb 26 13:38:03 ext2/ext3 Feb 26 13:38:30 i thought they use vfat or fat32 Feb 26 13:39:16 it depends Feb 26 13:39:24 because most ppl are comfortable using it i think, they can connect it to a windows pc and transfer data Feb 26 13:39:39 if you want to use it for data then vfat (nevermind fat16/32 - it is side depend) Feb 26 13:39:45 if for rootfs then ext2/ext3 Feb 26 13:40:06 okay, i am facing a strange fs tools incompatability issue. Feb 26 13:40:24 which one? Feb 26 13:40:46 i have a 4G sdmmc card, i try to format using fdisk and mkfs.vfat i see only 1G Feb 26 13:41:05 * * OE Bug 3837 has been marked as DUPLICATE of bug 3611 by miska(AT)atrey.karlin.mff.cuni.cz Feb 26 13:41:06 * * linux-hackndev-2.6-2.6.21-hnd3-r13-do_fetch Feb 26 13:41:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3837 Feb 26 13:41:34 Crofton: I have stinking cold after FOSDEM :-( Feb 26 13:41:43 Crofton: I blame mickeyl :-D Feb 26 13:42:02 im taking of on the target, i use the busybox fdisk i am able to partition as a single partion of 4G and they use mkfs.vfat to format it, when i mount it and use df -h i see only 1G Feb 26 13:42:59 shri: which kernel? Feb 26 13:43:06 XorA, me too Feb 26 13:43:07 2.6.22 Feb 26 13:43:23 Crofton: maybe it was revenge of the american chicks Feb 26 13:43:30 could be Feb 26 13:43:45 shri: kernel recognize it as 4GB one? Feb 26 13:44:07 yes Feb 26 13:44:17 did you miss the -F 32 Feb 26 13:44:27 XorA: fat16 handle up to 4gb Feb 26 13:44:39 hrw: not in dosfsutils last I checked Feb 26 13:44:43 ah Feb 26 13:44:49 shri: format card on desktop? Feb 26 13:44:52 hrw: it went wrong somehow or other Feb 26 13:45:26 i see only 1G in desktop the same card Feb 26 13:45:56 if i format it on the target using ext2 fs, then it works, but i cant use it on my windows desktop Feb 26 13:48:24 install ext2 IFS for windows :-) Feb 26 13:48:31 XorA: hah Feb 26 13:48:58 there was a constant flow of air @ the booth Feb 26 13:49:09 Xora: i can install, my customers cant, they dont pay if i ask them to Feb 26 13:49:10 next time we need to get one that's not directly opposite the main entrance ;) Feb 26 13:49:15 03mickeyl 07org.oe.dev * rdfa12fcb... 10/ (1 packages/python/python-dbus_0.82.4.bb): python-dbus: add dependency to dbus-glib (thanks Emdete) Feb 26 13:49:25 mickeyl: next fosdem need better place Feb 26 13:49:55 yeah Feb 26 13:50:11 there's rumours they will look for something else Feb 26 13:50:19 ULB is too small for us Feb 26 13:50:26 I assume its grown in size since It started? Feb 26 13:50:34 a lot I think Feb 26 13:50:50 or they need to charge a token admission ..... Feb 26 13:50:58 XorA: constantly Feb 26 13:51:00 to run off the groupies :) Feb 26 13:51:05 heh Feb 26 13:51:15 that reminds me... booth babes next year! Feb 26 13:51:17 or charge for booths and devrooms Feb 26 13:51:20 only 1 or 2 beer girls allowed :-) Feb 26 13:51:24 it has to be tough to find a facility that you can get for free/low cost Feb 26 13:51:25 mickeyl: babes or babies? :D Feb 26 13:51:31 hrw: babes, sorry ;) Feb 26 13:51:32 :) Feb 26 13:51:44 mickeyl: I like idea of booth babes ;D Feb 26 13:51:48 hi zecke Feb 26 13:51:48 hehe Feb 26 13:52:06 zecke: is your mtn2git working? Feb 26 13:52:21 vivijim: hi vivi. zecke is packing or already travelling towards Taipei Feb 26 13:52:25 we don't really need one, people stop by the booth already, but it would still be fun to introduce FOSDEM to the oncept Feb 26 13:52:56 *nod* Feb 26 13:53:04 mickeyl, who will handle letting people know about the ev status? Feb 26 13:53:04 hi mickeyl, thanks.. Do you know about the OE git repository? Feb 26 13:53:41 vivijim: I think zecke runs the script by hand and as he was at FOSDEM probably didnt happen Feb 26 13:55:17 Crofton: zecke and me will update openembedded-devel regularly once he's back from Taipei (early march) Feb 26 13:55:30 Crofton: i will send a first it's-progressing mail soon Feb 26 13:56:28 Any chance we can apply to SoC ? Feb 26 13:56:47 XorA: I'm using monotone yet, but we are having to many strange problems with monotone and I'm thinking to migrate to git... I'd like to know if the oe git tree is working... Feb 26 13:57:06 vivijim: it works Feb 26 13:57:08 morning Feb 26 13:57:34 vivijim: It works well enough I hope we can move it to somewhere in oe.org domain Feb 26 13:57:34 Crofton: i was thinking about that. we can try of course, but I'm afraid we're too domain specific to really raise interest among the committee. Feb 26 13:57:57 makes more sense that Angstrom apply for SoC not OE Feb 26 13:58:21 most of the interesting OE projects are too hard for a student in 3 months Feb 26 13:58:30 I'll try and read up on what is needed to apply Feb 26 13:58:37 worst that can happen is they say no :) Feb 26 13:59:16 agreed Feb 26 13:59:24 is the application open yet? Feb 26 13:59:33 i handled this last year for OpenMoko Feb 26 13:59:37 second week next month I think Feb 26 13:59:41 excellent Feb 26 14:01:10 sounds like we do not need a legal entity to apply Feb 26 14:01:17 so we could apply as Angstrom also Feb 26 14:01:40 I can handle money since it might be easier that way Feb 26 14:02:05 there will be tax implications for me though, but I could pretaned I was donating that to the effort .... Feb 26 14:04:37 yep Feb 26 14:04:56 That would be a good problem to have Feb 26 14:30:34 vivijim: there is one issue with root renaming but it is minor Feb 26 14:31:25 so it's possible to use git instead that slowass mtn? Feb 26 14:31:53 ynezz: well Feb 26 14:34:24 zecke: what is the problem with root renaming? Feb 26 14:34:46 zecke: what is the git tree address? I forgot Feb 26 14:34:53 vivijim: one file in the OE repository is placed at the wrong path Feb 26 14:35:06 vivijim: search for openembedded on repo.or.cz, you will see the root renaming bug there Feb 26 14:36:52 zecke: nice, it works. thanks Feb 26 15:08:27 03mickeyl 07org.oe.dev * rbb3a350c... 10/ (3 files in 2 dirs): naming tweaks for gpio on s3c24xx Feb 26 15:39:39 comming back to my sdmmc issue; if i format it using mkdosfs ; i am able to see a 4G mmc disk when i mount it. but when i format it using mkfs.vfat i see only a 1G disk mounted Feb 26 15:40:08 is there any issue with mkfs.vfat? Feb 26 15:41:41 shri: obviously yes Feb 26 15:43:52 lrwxrwxrwx 1 root root 7 wrz 13 09:30 /sbin/mkfs.vfat -> mkdosfs Feb 26 15:47:34 are both same? vfat and msdos fs. there is a difference between the 2 files in terms of size. for some reason mkdosfs works and mkfs.vfat doesnt Feb 26 15:49:05 is there any utility which will output the format of the filesystem stored? Feb 26 15:50:46 i think mkdosfs is different from mkfs.vfat Feb 26 15:50:57 both seem to be different format Feb 26 16:16:41 03Laibsch 07org.oe.documentation * r64b95f18... 10/ (1 usermanual/chapters/common_use_cases.xml): common_use_cases.xml: add a chapter about how to use svn/cvs in a bb file properly. Feb 26 16:28:05 03xora 07org.oe.dev * rff93c9e7... 10/ (1 packages/u-boot/u-boot-openmoko_1.3.1+git+svn.bb): Feb 26 16:28:05 u-boot-openmoko_1.3.1+git+svn.bb : update svn and make it generate Feb 26 16:28:05 gta02v5 u-boots Feb 26 16:36:10 what exactly is SRCREV doing? Feb 26 16:36:51 Jin^eLD: picks the right SRCREV Feb 26 16:37:18 what is a "right srcrev"? Feb 26 16:37:35 where is it defined? Feb 26 16:38:20 you mean - the SVN fetcher sets SRCREV and then it can be used in the package name? Feb 26 16:39:23 Jin^eLD: look at sane-srcrevs.inc Feb 26 16:39:58 just found it... I think now I get the idea Feb 26 16:40:04 that explains where the weird number was coming from :) Feb 26 16:40:06 thanks Feb 26 16:40:28 still struggling with uclinux ;) but at least I found a more recent elf2flt version onw Feb 26 16:40:34 s/onw/now/ Feb 26 16:42:29 anyone around that wants to talk samba bugs? Feb 26 16:45:36 Tartarus: did you find more Feb 26 16:45:52 nope, but there's a handful open Feb 26 16:46:20 and we might wanna move to latest stable being an option at least, more security fixes Feb 26 16:46:34 (and maybe delete some of the ancient versions) Feb 26 16:47:19 Tartarus: I will boot my target soon with latest .dev then I can verify some of them Feb 26 16:47:44 Tartarus: yes updating might be a good thing to do atm Feb 26 16:48:07 k Feb 26 16:51:02 But take it easy on deleting ancient versions; that sometimes causes problems for distros that require ancient stuff. Feb 26 16:51:35 mwester-road, true enough. The reporter on 2900 suggested maybe we see what versions of samba we can get rid of Feb 26 16:52:10 Did they say why? It's not like it causes any real harm to have a few older versions about. Feb 26 16:52:24 (that one has a maybe some older version might be broken by the change note, but tbh, we should probably change things around a tiny bit so older versions would inherit the fixes anyhow, it's easy enough) Feb 26 16:52:41 and, er 3146, not 2900 Feb 26 16:52:49 bye Feb 26 16:53:15 and it's 7 versions, but yeah, 3-4 are just using samba.inc for everything Feb 26 16:54:06 Again we need to be careful, if a given recipe is used by some other distro and they expect it to be stable, there's no need or desire to be changing old stuff. I don't think anything I'm responsible cares about Samba, specifically, but I certainly have a distro that's old (still uses a 2.4 kernel) Feb 26 16:54:12 i really really just wanna get the mount.smb{,fs} problem fixed :) Feb 26 16:54:42 (and my patch really should work on all of the possible versions) Feb 26 16:55:44 mwester-road: I think addin a new version would not hurt if we don't delete old ones Feb 26 16:57:13 03rpurdie * r1017 10/ (2 files in 2 dirs): utils.py: Tweak explode_deps function to remove unused manipulation Feb 26 17:06:41 KhemHome: exactly my point - add a new one, don't change the old ones unless necessary. :) Feb 26 17:07:57 why is ANGSTROM_MODE distro specific and not a generic variable for selecting between eglibc/glibc/uclibc? Feb 26 17:08:59 mwester-road, don't delete or don't change? Feb 26 17:09:05 Jin^eLD: there have been discussions on this see mailin list archive Feb 26 17:09:15 they're all busted for mount.smb/smbfs i bet Feb 26 17:09:25 03rpurdie * r1018 10/ (5 files in 4 dirs): shell.py: Update to use cooker.buildFile() Feb 26 17:09:27 Tartarus: We can change if there are bugs Feb 26 17:09:40 I disagree. Feb 26 17:09:41 Tartarus: probably that counts as bug Feb 26 17:10:24 would it be the right approach to modify TARGET_OS if I want to add "uclinux-uclibc"? or is TARGET_OS only used for naming stuff and nothing else? Feb 26 17:10:35 would be nice if someone could point me to a place where to start hacking Feb 26 17:11:00 Jin^eLD: start at your local.conf/site.conf file Feb 26 17:11:03 If you were to go back to build a copy of Red Hat 9 from scratch, would you expect that suddenly patches would appear? Similarly, if users go to build an older version of a package, they do not expect to see patches in their version of that same package that weren't in the original. Feb 26 17:11:04 basically it seems that I need to add global linker flags to trigger the elf2flt utility and also use -static Feb 26 17:11:04 Jin^eLD: Its used as part of the triplet passes to autotools Feb 26 17:11:10 Historical accuracy counts. Feb 26 17:11:38 But I shall not argue the point, if it breaks anything I'm responsible for, I'll just quietly put it right :) Have at it, folks! Feb 26 17:11:39 Jin^eLD: What triplet is used for uclinux for autotools in general? Feb 26 17:11:41 mwester-road, isn't that why OE exists in an scm of sorts? Feb 26 17:11:57 RP: hmm, let me check Feb 26 17:12:00 "of sorts" is the operative term. Feb 26 17:12:10 heh Feb 26 17:12:22 There's reproducible keys :) Feb 26 17:12:40 Sure, just try giving that to a user who's trying to rebuild something. Feb 26 17:12:49 it's say RedHat 9 vs the community driven / updated fedora version Feb 26 17:13:02 RP: --target=arm-uclinuxeabi --host=arm-uclinuxeabi Feb 26 17:13:07 RP: this stuff or did you mean something else? Feb 26 17:13:46 mwester-road, I do agree there's no reason older distros shouldn't be kept working :) Bugs are Bad. Feb 26 17:14:12 Jin^eLD: That looks like what I mean - "uclinuxeabi" would be your TARGET_OS Feb 26 17:15:03 RP: how would I add global CFLAGS for all apps, depending on the setting of TARGET_OS? Feb 26 17:15:35 just using CFLAGS_uclinuxeabi_local or something? Feb 26 17:15:41 Jin^eLD: CFLAGS_uclinuxeabi = "" Feb 26 17:15:53 Jin^eLD: TARGET_OS is in overrides iirc Feb 26 17:16:31 aha.. ok.. just CFLAGS_uclinuxeabi not _append or something? Feb 26 17:16:48 Jin^eLD: You're after to change it entirely? Feb 26 17:17:02 Jin^eLD: or you want to add to it? Feb 26 17:17:06 I want to append certain cflags to everything that I build Feb 26 17:17:15 at least I see that they did it like that when building uclinux with buildroot Feb 26 17:17:23 Jin^eLD: then you need _append Feb 26 17:17:38 ok, let's put that in local.conf and see what happens :> Feb 26 17:22:58 RP: this CFLAGS_uclinuxeabi_append= thing will not be used for cross/native stuff, right? Feb 26 17:23:08 Jin^eLD: no Feb 26 17:23:27 thx Feb 26 17:23:37 03rpurdie * r1019 10/ (6 files in 4 dirs): runqueue.py: Add StampUpdate event Feb 26 17:25:38 I need to build elf2flt after building the toolchain and before building any other apps, what would be the correct way to specify such a dependency? Feb 26 17:26:17 Jin^eLD: See the magic in base.bbclass Feb 26 17:26:27 thx Feb 26 17:30:12 is it the INHIBIT_DEFAULT_DEPS stuff? Feb 26 17:30:22 Jin^eLD: yes Feb 26 17:31:18 03xora 07org.oe.dev * r7de41b6c... 10/ (4 files in 3 dirs): Feb 26 17:31:18 gsmd : flip an extra bit on gta02 machines that stops hideous noise in Feb 26 17:31:18 sound system Feb 26 17:32:12 bye! Feb 26 17:32:15 03rpurdie * r1020 10/ (4 files in 4 dirs): Convert -b option to use taskdata/runqueue Feb 26 17:44:52 <_diego_> hi, there is a way to rebuild Packages.gz, after building a package? Feb 26 17:47:15 dijenerate: bitbake package-index Feb 26 17:48:55 <_diego_> RP: thanks Feb 26 17:59:18 03rpurdie * r1021 10/ (16 files in 6 dirs): Remove digraph and switch to new stamp checking code. Feb 26 18:01:27 * RP wonders whats going to break :/ Feb 26 18:02:48 :> Feb 26 18:06:40 mhm, something overrides my local.conf TARGET_OS setting... Feb 26 18:07:03 OVERRIDES = "local:${MACHINE}:${DISTRO}:angstrom:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}" Feb 26 18:07:07 what is that doing exactly? Feb 26 18:07:25 Jin^eLD: Its not the problem Feb 26 18:07:37 Jin^eLD: Its setting up a list of things that can be used as variable overrides Feb 26 18:07:38 I grepped ang* for TARGET_OS but there was nothing Feb 26 18:07:52 (since my distro is based on angstrom) Feb 26 18:07:59 Jin^eLD: include/angstrom-eglibc.inc Feb 26 18:08:24 eh, -glibc/-uclibc.inc Feb 26 18:08:29 mhm, I did a grep -r Feb 26 18:08:35 not sure why I missed it hrhr, thanks Feb 26 18:09:35 uuh that one has some weird stuff in there Feb 26 18:09:56 it seems that Koen was preparing something for blackfin Feb 26 18:10:07 TARGET_OS = "${@['${TARGET_OS_UC}', 'uclinux-uclibc'][bb.data.getVar('TARGET_ARCH',d) in ['bfin']]}" Feb 26 18:10:44 hm, well, I'll try adding my machine and see what will happen :> Feb 26 18:11:52 * RP -> afk Feb 26 18:17:55 can someone please help me to decypher this? Feb 26 18:17:56 ARGET_OS_UC = "linux${@['-uclibc','-uclibcgnueabi'][bb.data.getVar('TARGET_ARCH',d,1) in ['arm', 'armeb'] and bb.data.getVar('MACHINE',d) not in ['collie','h3600', 'h3800', 'simpad', 'htcwallaby']]}" Feb 26 18:18:08 where is linux$... coming from? Feb 26 18:18:30 I understand the TARGET_ARCH and MACHINE comparisons Feb 26 18:18:54 but not the beginning of the line Feb 26 18:22:02 ['this if x false', 'this if x true'][x] # x is a boolean value used as index into the string list Feb 26 18:22:22 Jin^eLD: if machine is collie','h3600', 'h3800', 'simpad', 'htcwallaby' use dont use EABI Feb 26 18:23:29 XorA|gone: what variable is linux-uclibc or linux--uclibcgnueabi taken from? Feb 26 18:24:24 TARGET_OS_UC Feb 26 18:24:47 I thought TARGET_OS_UC is being set to the result of the expression? Feb 26 18:25:38 yes, so in the end TARGET_OS_UC is either "linux-uclibc" (for collie, etc) or "linux-uclibcgnueabi" (for the EABI capable machines) Feb 26 18:26:23 ahaa Feb 26 18:27:02 ok and the next following line Feb 26 18:27:03 TARGET_OS = "${@['${TARGET_OS_UC}', 'uclinux-uclibc'][bb.data.getVar('TARGET_ARCH',d) in ['bfin']]}" Feb 26 18:27:24 what is being done with TARGET_OS_UC here? Feb 26 18:27:28 blackfin isnt arm, so previous expression doesnt work Feb 26 18:27:59 so if bfin then force TARGET_OS to uclinux-uclib otherwise use previous calculated value Feb 26 18:28:08 ok now I get it.. Feb 26 18:28:10 thanks a lot! Feb 26 18:28:28 damn python programmers :-) Feb 26 18:28:42 :) Feb 26 18:29:05 I generally do like python, but this stuff is a little over my head Feb 26 18:29:52 this intermix with shell like variables often confuses me Feb 26 19:03:18 hmmm.. is there a way to say "append certain CFLAGS to everything except the toolchain"? Feb 26 19:03:46 or do I have to add something to the toolchain .bb files to make sure that the global setting is not taken? Feb 26 19:04:08 it seems that elf2flt postprocessing should not be done for gcc :> Feb 26 19:04:20 ...which is a guess, so if anyone is familiar with uclinux - correct me :> Feb 26 20:15:05 * * OE Bug 2410 has been RESOLVED (INVALID) by rodrigo.vivi(AT)gmail.com Feb 26 20:15:07 * * apt package: build breaking Feb 26 20:15:09 * * http://bugs.openembedded.org/show_bug.cgi?id=2410 Feb 26 20:45:10 hi all! Feb 26 21:04:00 gm Feb 26 21:10:56 hi people Feb 26 21:11:02 all rested from FOSDEM? Feb 26 21:11:14 sick from FOSDEM ... Feb 26 21:11:32 Crofton: -ETOOMUCHBEER? Feb 26 21:11:32 I caught mickey|shopping cold instead of him Feb 26 21:11:52 no, I recovered from that Feb 26 21:12:10 So mickey|shopping is not sick after FOSDEM for a change Feb 26 21:12:16 yes Feb 26 21:13:06 Crofton: well get well soon Feb 26 21:14:33 thanks Feb 26 21:26:37 re Feb 26 21:26:55 hi flo_lap Feb 26 22:31:44 g'night all ! Feb 26 22:31:51 night Feb 26 22:33:00 anyone else start getting a top level error when trying to build madwifi? Feb 26 23:21:36 Can anyone provide some assistance with Cross Compile Badness? Feb 26 23:21:42 i am building icecast (streaming audio server) Feb 26 23:21:59 I can clearly see '-I /usr/include' in the config.log and this is wrong Feb 26 23:22:03 the problem is I am not sure how to identify what is actually inserting the '-I /usr/include' Feb 26 23:22:11 any thoughts? Feb 26 23:22:16 here is the output from config.log Feb 26 23:22:24 http://pastebin.com/m47dbff15 Feb 26 23:23:04 dmfrey: probably it is hardcoded in the Makefil Feb 26 23:23:41 dmfrey: just gotta search through the .bb files and then the module specific makefiles Feb 26 23:24:12 dmfrey: check all the .inc files too Feb 26 23:24:58 the recipe is one that i have written, so i know that there are none in there Feb 26 23:25:41 also, the Makefile gets generated, however, I do have a Makefile.in and Makefile.am Feb 26 23:25:41 dmfrey: then most likely Phillipe's suggestion is a good place to look, it might be hardcoded in one of Icecast's makefiles or the makefile.inc Feb 26 23:26:30 k, i will find them all and look through them Feb 26 23:26:34 dmfrey: check Makefile,am then Feb 26 23:26:48 ok Feb 26 23:27:10 usually, are there just one makefile.in and makefile.am? Feb 26 23:27:22 i see makefiles in subdirectories Feb 26 23:28:14 dmfrey: every subdirectory usually has a MAkefile,am Feb 26 23:28:16 nevermind, i see there are lots of them Feb 26 23:28:56 from them the autotools make lots of Makefiles Feb 26 23:29:23 k Feb 26 23:29:34 does anyone else build Madwifi through OE? after updating the tree yesterday I have been getting an ubuildable dependency chain and as far as I can tell nothing changed Feb 26 23:30:00 but... I guess obviously something did Feb 26 23:32:03 http://pastebin.com/m1d32a59c Feb 26 23:32:42 dmfrey: just do a find . -type f -name "Makefile.am" | xargs grep "/usr/include" Feb 26 23:32:58 you should find the offending file pretty quickly Feb 27 00:05:18 03likewise 07org.oe.dev * r4f9025ff... 10/ (3 files in 2 dirs): libzvbi: inherit from pkgconfig so that .pc is staged. Feb 27 00:05:23 03rpurdie 07org.oe.dev * rfe54e422... 10/ (1 classes/seppuku.bbclass classes/tinderclient.bbclass): seppuku.bbclass/tinderclient.bbclass: Use exec_func, not exec_task for clean function Feb 27 00:05:28 03rpurdie 07org.oe.dev * r24fc5bcb... 10/ (1 classes/rm_work.bbclass): rm_work.bbclass: Add event handler to take care of removing stamps when trying to rebuild packages after sources have been removed. Needs bitbake 1.8 svn to have any effect. Feb 27 00:05:33 03rpurdie 07org.oe.dev * r91782f0e... 10/ (1 classes/base.bbclass): Feb 27 00:05:33 base.bbclass: Add code to make 'rebuild' continue to work after changes in Feb 27 00:05:33 bitbake. Bitbake 1.8.10 users should be unaffected, bitbake 1.8.11 users will Feb 27 00:05:33 need the latest svn stable branch bitbake. Only the rebuild task is affected by Feb 27 00:05:34 the changes Feb 27 00:05:45 03xora 07org.oe.dev * rbde94cf8... 10/ (4 files in 2 dirs): Feb 27 00:05:45 pulseaudio* : fix the adding of pulse to audio group, busybox does this Feb 27 00:05:45 arse about tit. Feb 27 00:37:10 thesing: are you lurking? Feb 27 00:37:19 thesing: klibc-config-eabi.patch needs first line removed Feb 27 01:02:51 03tmbinc 07org.oe.dreambox * rcfb3fb43... 10/ (3 files in 2 dirs): libxml2: fix URIs Feb 27 01:02:56 03tmbinc 07org.oe.dreambox * r582866c5... 10/ (1 packages/dreambox/dreambox-dvb-modules.bb): dreambox-dvb-modules: dm800 support Feb 27 01:03:00 03tmbinc 07org.oe.dreambox * rbaa17179... 10/ (5 files in 5 dirs): base-files: fix usbfs entry Feb 27 01:03:04 03tmbinc 07org.oe.dreambox * r5d4bb6b6... 10/ (1 packages/dreambox/dreambox-dvbincludes.bb): dreambox-dvbincludes: dm800,dm8000 support Feb 27 01:03:08 03tmbinc 07org.oe.dreambox * r0605c809... 10/ (1 packages/dreambox/dreambox-secondstage.bb): dreambox-secondstage: dm8000,dm800 support Feb 27 01:03:12 03tmbinc 07org.oe.dreambox * r09973bd8... 10/ (12 files in 12 dirs): add dm800 support Feb 27 01:03:16 03tmbinc 07org.oe.dreambox * r4fbb8333... 10/ (15 files in 3 dirs): linux-dm800: add support for dreambox dm800 Feb 27 01:03:19 03tmbinc 07org.oe.dreambox * r00dee604... 10/ (1 packages/enigma2/enigma2.bb): enigma2: add dm800,dm8000 support Feb 27 01:03:23 03tmbinc 07org.oe.dreambox * rd0574b34... 10/ (1 packages/images/dreambox-image.bb): dreambox-image: add dm800 support Feb 27 01:03:27 03tmbinc 07org.oe.dreambox * r8ad83d52... 10/ (1 packages/tuxbox/tuxbox-libtuxtxt.bb): tuxbox-libtuxtxt: support for dm800 Feb 27 01:03:30 03tmbinc 07org.oe.dreambox * rb81ee91d... 10/ (1 conf/machine/dm800.conf): conf/machine/dm800.conf: add BOOTSTRAP_EXTRA_DEPENDS Feb 27 01:03:35 03Laibsch 07org.oe.dev * r0ed9d9e6... 10/ (1 packages/libmrss/files/better-parse-url-r0.patch): libmrss: document the acceptance of better-parse-url-r0.patch into upstream cvs Feb 27 01:06:44 03tmbinc 07org.oe.dreambox * r76e80b5a... 10/ (3 files in 2 dirs): file: add 4.20 Feb 27 01:11:51 03tmbinc 07org.oe.dreambox * rb0c91a58... 10/ (1 packages/hotplug-ng/hotplug-ng.bb): hotplug-ng: add hotplug-ng from http://hotplug-ng.schwerkraft.elitedvb.net/ Feb 27 01:11:55 03tmbinc 07org.oe.dreambox * r195e59f1... 10/ (5 files in 2 dirs): file,file-native: move to 4.23 Feb 27 02:01:44 philippe: thanks for your suggestion before (find . -type f -name "Makefile.am" | xargs grep "/usr/include") Feb 27 02:01:49 i had to step a way for a bit Feb 27 02:02:18 dmfrey: ok. I have been looking into it myself, but it seems to be more subtle. Unless you found something I missed Feb 27 02:02:36 nope, didn't find the problem Feb 27 02:03:04 when I review my config.log, i get 3 locations where the Badness occurs Feb 27 02:03:17 they are odd programs too, like fgrep Feb 27 02:04:30 would it be beneficial to remove all of the Makefile* files and regenerate them again? Feb 27 02:05:12 dmfrey: I already tried. Feb 27 02:05:38 philippe; ok Feb 27 02:07:54 dmfrey: to be honest I do not know what happened Feb 27 02:08:39 i have been looking at this package for a few weeks now trying to understand what is wrong with it Feb 27 02:08:51 what about pulling from cvs? Feb 27 02:09:24 dmfrey: you could try that, Don't know what it is going to do Feb 27 02:09:39 but I guess there is a reason why icecast is in nonworking Feb 27 02:09:56 philippe: just drawing at straws :( Feb 27 02:17:09 philippe: i was looking for any file that could possibly be inserting a host-based reference Feb 27 02:17:21 but nothing Feb 27 02:18:22 dmfrey: same here Feb 27 02:18:58 because looking at other config.log files the configure_qa does not complain, while /usr/include is in those too Feb 27 02:20:12 i just did a quick recipe for ices as well, same issue Feb 27 02:20:37 something is up with xiph packages Feb 27 02:21:36 dmfrey: I'll look into it some more tomorrow Feb 27 02:21:42 I need to go to sleep now Feb 27 02:23:09 philippe: thanks for your help. Have a good night Feb 27 02:25:32 philippe: one more thing...this only appears to be an issue with building these two packages with OE. In my old buildroot environment, these issues built without issue. Just an FYI Feb 27 08:42:14 morning Feb 27 08:43:42 morning XorA Feb 27 09:09:11 morning Feb 27 09:11:46 hm when I bitbake an image recipe, it doesnt produce a uImage file Feb 27 09:11:47 what am I missing? Feb 27 09:20:55 03xora 07org.oe.dev * r79453456... 10/ (9 files in 4 dirs): Feb 27 09:20:55 xserver-kdrive_1.3.0.0.bb : remove glamo from xserver-kdrive, next commit Feb 27 09:20:55 will create new recipe for it. Feb 27 09:21:00 03xora 07org.oe.dev * rb2173b2f... 10/ (1 packages/xorg-xserver/xserver-kdrive-glamo_1.3.0.0+git.bb): Feb 27 09:21:00 xserver-kdrive-glamo_1.3.0.0+git.bb : an Xglamo recipe on its own pulling Feb 27 09:21:00 from the developers git tree directly. Feb 27 09:33:52 morning Feb 27 09:33:55 XorA: http://www.akihabaranews.com/en/news_details.php?id=15572 Feb 27 09:34:29 hrw: GREAT!!!!! Feb 27 09:35:06 ;) Feb 27 10:09:14 where do I find a list of the values I can use for TARGET_OS and others when I wanna change local.conf in build/conf? Feb 27 10:38:32 pbne04: you probably want to put DISTRO="angstrom-2008.1" and not mess with TARGET_OS at all Feb 27 10:43:04 morning there! **** ENDING LOGGING AT Wed Feb 27 10:59:57 2008