**** BEGIN LOGGING AT Tue Jun 22 02:59:56 2010 Jun 22 05:10:49 Question about kernel builds: is the .configure file in the work dir supposed to be the same as the defconfig from the sources? Jun 22 05:11:21 I am trying to modify some options in linux-kexexboot and they don't seem to be taking Jun 22 06:29:45 Hi all, Can anyone help me with this error http://pastebin.org/350416 Jun 22 06:38:37 Nikunj: I believe that part of the log was cut from pastebin Jun 22 06:39:25 valhalla : Ohh.. let me post the remaining part..! Jun 22 06:40:04 good morning Jun 22 06:41:30 valhalla : http://pastebin.org/350426 kindly help.! Jun 22 06:46:28 Nikunj: you should have some log in /home/nikunj/ope/build/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.9-r36.3/temp Jun 22 06:57:28 valhalla : yes, log.do_compile,configure,install and log.packagestaging_fastpath and log.sysroot_stage_all. rest all are 0 bytes. Jun 22 06:58:54 Nikunj: the task that failed was do_package (line 49 in the second pastebin) Jun 22 06:59:32 valhalla : there is no log for do_package. Jun 22 07:00:25 Nikunj: no clue on what happened, then, if there is no error message Jun 22 07:05:06 valhalla : should I redownload OE and try building again.. I am not sure will that work..! I have tried with "git pull" already.. same results. Jun 22 07:05:54 Nikunj: try deleting tmp and rebuild Jun 22 07:08:37 mckoan : I have tried that also. ! Jun 22 07:13:35 valhalla : anything to do with file access permissions..? Jun 22 07:13:46 mckoan : anything to do with file access permissions..? Jun 22 07:13:51 Nikunj: I have faced issues with glibc_2.9 a lot, I think mckoan's suggestion is the most apt. Jun 22 07:14:04 Just remove tmp/ and build again Jun 22 07:15:07 screwgoth : Okay..let me try once more. Which is the most smallest package to build..? Jun 22 07:15:36 Nikunj: build nano , i.e., bitbake nano Jun 22 07:15:50 screwgoth : okay. Jun 22 07:22:33 morning Jun 22 07:22:47 hrw : morning Jun 22 07:24:29 screwgoth : deleting the "source" folder is also required..?? In which all the tar stuff downloaded is kept. Jun 22 07:25:52 Nikunj: no not delete sources Jun 22 07:26:21 mckoan : cool..I am trying with nano now. Jun 22 07:50:07 Question about kernel builds: is the .configure file in the work dir supposed to be the same as the defconfig from the sources? Jun 22 07:50:10 I am trying to modify some options in linux-kexexboot and they don't seem to be taking Jun 22 07:57:01 hvontres|home, you mean .config? Jun 22 07:57:21 Crofton: yes. Jun 22 07:57:21 you need to make sure the configure method runs when you do the rebuild Jun 22 07:57:41 I think that is the stage that copies the defconfig over Jun 22 07:58:09 Crofton: What about -c configure? I checked after running that and the two files were diferent ... Jun 22 07:58:33 are you sure you are changing the right defconfig Jun 22 07:58:45 see eFfeM_work's email about file paths ..... Jun 22 07:59:26 Crofton: do you have a link handy? Jun 22 07:59:34 not handy Jun 22 07:59:47 a quick check, rename the file you are mdoifying and see if it fails Jun 22 08:00:27 Crofton: hmmm, I checked, and the modified file made it to the tmp/work dir, just not the linux source dir. Jun 22 08:00:36 ok Jun 22 08:00:46 not sure then Jun 22 08:03:40 Crofton: thanks for the pointers, I'll try again in the morning Jun 22 08:03:43 ->bed Jun 22 08:03:54 gn Jun 22 08:13:06 When I am cross-compiling a package, I see that it makes a wrapper shell script on my laptop and when I do "make install" it install the cross-compiled ARM binary at the prefix path Jun 22 08:13:20 can anyone explain this behaviour Jun 22 08:26:53 kergoth: RP: is there already any idea when bitbake 1.10 will be tagged and will be made available (e.g. on berlios) as a downloadable tarball ? Jun 22 08:27:18 .dev requires bitbake from git now I think .... Jun 22 08:28:00 eFfeM_work, I suspect kergoth is still asleep :) Jun 22 08:28:23 Crofton, dev head still works with .8.18, Jun 22 08:28:48 I wonder why I switched to git :) Jun 22 08:28:58 but guess it would be nice to kinda define a standard .10 version Jun 22 08:29:20 if I recall correctly .10 has much better performance Jun 22 08:29:26 I think that 1.10 should be required Jun 22 08:29:40 I think I had an issue 10 fixed Jun 22 08:29:59 I forget Jun 22 08:30:25 good morning Jun 22 08:30:25 i'm still on .8.18, with that I've build virtual/kernel, bash and busybox without problem, no idea about things like x and qt Jun 22 08:30:29 hi florian Jun 22 08:30:33 and GNUtoo|laptop Jun 22 08:31:48 i can imagine for problems it is easier if we have a fixed reference point for bb (like a tagged 1.10) with git you never know what version someone else is using (if investigating an issue) Jun 22 08:33:19 hi eFfeM_work Jun 22 08:43:36 screwgoth : same error with bitbake nano too. Jun 22 08:43:51 mckaon : same error with bitbake nano too. Jun 22 08:43:55 heya. if I want Packages.gz to be generated in my deploy dir I have to do what? Jun 22 08:45:55 mckaon : http://pastebin.org/350549 Jun 22 08:47:18 mckoan : http://pastebin.org/350549 Jun 22 08:47:31 mckoan : same error as before. Jun 22 08:47:58 screwgoth : http://pastebin.org/350549 Jun 22 08:49:09 Nikunj, is there a more descriptive error in that log? Jun 22 08:50:52 Crofton : nopes.. its failing at do_package and thrs no log for do_package in "build/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.9-r36.3/temp" Jun 22 08:52:00 Crofton : I think thrs something to do with permissions. Jun 22 08:52:45 yeah, I was hoping for an error message that told you what to do :) Jun 22 08:53:14 Nikunj: which distro are you using? Jun 22 08:53:17 Crofton : nopes, thrs nothing like that. Jun 22 08:53:20 Angstrom Jun 22 08:53:29 Nikunj: on the host Jun 22 08:53:40 CentOS 5.5 Jun 22 08:53:47 damn Jun 22 08:54:06 Nikunj: look at your locales settings Jun 22 08:54:17 mckoan : Other system with everything similar has got it all right... Jun 22 08:54:48 Nikunj: in your previous pastebin you said: # Jun 22 08:54:49 Nikunj: in your previous pastebin you said: NOTE: generating locale es_NI (UTF-8) Jun 22 08:54:49 # Jun 22 08:54:49 NOTE: Task failed: localedef returned an error Jun 22 08:54:58 http://pastebin.org/350426 Jun 22 08:55:02 start from there Jun 22 08:55:19 mckoan : yeahh.. how to change locale settings? Jun 22 08:55:53 Nikunj: I use only Debian and derivatives Jun 22 08:56:56 try google-ing "NOTE: Task failed: localedef returned an error" Jun 22 08:57:03 morning Jun 22 08:57:24 maybe this http://lists.openmoko.org/pipermail/community/2007-September/010567.html Jun 22 09:02:36 Nikunj : Oh, I know this error. Jun 22 09:03:11 Nikunj: In my case, I was setting up OE on Ubuntu 10.04 and some googling around told me that this is sue to the gcc and glibc versions of Ubuntu Jun 22 09:03:27 I shifted to Ubuntu 9.10 and I was able to compile. Jun 22 09:03:44 Do you have a machine with CentOS, say 5.2 or 5.3 Jun 22 09:04:11 screwgoth : I have CentOS 5.5 and is latest.. I did "yum update" too. Jun 22 09:04:48 host has CentOS 5.5 no other machine has 5.3 or less. Jun 22 09:06:35 Damn ... but loooks like mckoan has given a good link. Even I'm trying the hack it has suggested on Ubuntu 10.04 to see if that works Jun 22 09:09:42 screwgoth : please. Jun 22 09:29:36 screwgoth : got any clue? Jun 22 09:30:10 I'm bitbaking nano ... have not reached to a point where it builds glibc_2.9 Jun 22 09:30:14 Will let you know Jun 22 09:30:30 cool... Jun 22 09:53:23 screwgoth: I built on a fresh install of Ubuntu last night Jun 22 09:53:44 2 issues: a)dash symlink b)broken gzip Jun 22 09:53:59 fix those and bitbake happily Jun 22 09:56:19 is anyone using OE on FreeBSD? There is question about this in angstrom-devel@ ML Jun 22 09:56:48 * Jay7 can check this but only on customer's servers ;) Jun 22 09:57:20 03Koen Kooi  07org.openembedded.dev * rd2d2a6cb05 10openembedded.git/recipes/gcc/gcc-cross-sdk_4.5.bb: Jun 22 09:57:20 gcc-cross-sdk 4.5: sync configure options with cross version. Jun 22 09:57:20 Still fails with: Jun 22 09:57:20 | Jun 22 09:57:20 arm-angstrom-linux-gnueabi-strip:/OE/angstrom-setup-scripts/build/tmp-angstrom_2008_1/work/x86_64-armv7a-sdk-angstrom-linux-gnueabi/gcc-cross-sdk-4.5-r4+svnr160764/image/usr/local/angstrom/arm/arm-angstrom-linux-gnueabi/lib/libstdc++.so.6.0.14-gdb.py: Jun 22 09:58:32 hi , i got igep-v2 card , i tried to bitbake console-image for igep-v2 card but i m having problem , i got error , please set a valid MACHINE in your local.conf Jun 22 09:58:57 the machine name i tried : igep0020b Jun 22 10:05:10 ant_work: Wait, I'm confused. Are you saying that you built a Ubuntu distro for Beagleboard or that you built an Angstrom distro on Ubuntu ? Jun 22 10:05:40 Nikunj: It's compiling glibc_2.9 now .. We'll come to know in another minute or so if that fixes the issue Jun 22 10:06:00 screwgoth: angstrom on ubuntu host Jun 22 10:06:06 ant_work means this :) Jun 22 10:06:23 Jay7: Ahh... thought as much :-) ... and was it a Ubuntu 10.04 ? Jun 22 10:07:09 that is lucid Jun 22 10:07:18 not sure about verion number :) Jun 22 10:09:52 that's 10.04 Jun 22 10:10:21 Then he must have used the ENABLE_BINARY_LOCALE_GENERATION = "0" option Jun 22 10:12:51 Has anybody managed to build the angstrom-gnome-image ? **** BEGIN LOGGING AT Tue Jun 22 10:15:29 2010 Jun 22 10:56:32 Nikunj: Good news dude !! that worked !!! Jun 22 10:57:31 Nikunj: So just set the variable 'ENABLE_BINARY_LOCALE_GENERATION = "0" ' in the file .../openembedded/conf/local.conf and you should be good Jun 22 11:10:49 03Philip Balister  07org.openembedded.dev * r978dee1513 10openembedded.git/recipes/tasks/task-sdk-gnuradio-native.bb: task-sdk-gnuradio-native : Add a SDK for building GNU Radio natively on devices. Jun 22 11:15:08 screwgoth: I built Angstrom on Ubuntu 10.4 32bit. No gcc/binutils issues Jun 22 11:15:31 but I only built two images Jun 22 11:16:57 and no, I did indeed build the glinc locales Jun 22 11:19:42 screwgoth : okay.. let me try.. Will let you know. Jun 22 11:20:19 Nikunj: out of interest, are you building X86_64? Jun 22 11:20:39 I mean *on* 64 bit host? Jun 22 11:20:41 ant_work : nopes, OMAP5912OSK Jun 22 11:20:53 So ant_work or Nikunj or Jay7 , has anyone of you cross-compiled 3rd party autotools based packages ? Jun 22 11:21:04 ant_work : no..32 bit. Jun 22 11:22:08 screwgoth : not yet, btw, I am putting this variable in build/conf/local.conf file.. !! ? Jun 22 11:22:41 Nikunj: yes, there Jun 22 11:23:29 screwgoth: you mean outside of OE? Jun 22 11:23:46 yes... Jun 22 11:23:56 wel, yes, with gentoo's crosstool Jun 22 11:24:11 what do you mean "outside of OE" ... how do you cross-compile "inside of OE" ? Jun 22 11:24:26 ? Jun 22 11:24:32 OE is alla bout crosscompiling Jun 22 11:26:58 back to the locale generation issue, it is indeed bizarre it still fails Jun 22 11:27:39 iirc the mmap_min_addr have been solved Jun 22 11:28:03 ant_work: yeah yeah .. but when I tried to cross-compile telepathy-sofiasip , it's not working Jun 22 11:28:04 only with qemu-native from OE.. Jun 22 11:28:22 not with ASSUME_PROVIDED=qemu-native Jun 22 11:28:27 OMG Jun 22 11:28:40 why does someone still use those? Jun 22 11:28:58 (if not purposedly) Jun 22 11:29:30 I have never needed any ASSUME_PROVIDED Jun 22 11:30:20 if you have sane host and want to save some cpu time rebuilding some native stuff you already have... Jun 22 11:30:36 pls define 'sane host' Jun 22 11:30:39 :D Jun 22 11:30:52 but now with qemu-native new enough (and even better) then what gentoo have I'm not using ASSUME_PROVIDED too Jun 22 11:31:37 ant_work: well ie for git-native OE doesn't expect anything extra (AFAIK) so any git-native from host will be OK Jun 22 11:32:17 I see Jun 22 11:32:19 screwgoth: it's easier for me to write recipe and compile software by bitbake :) Jun 22 11:32:58 Jay7: show-off !! :-) Jun 22 11:33:16 JaMa: still dubious warnings wrt debian-distros: "and select /usr/bin/git-scm to provide git instead of /usr/bin/git.transition. This is not necessary in sid " Jun 22 11:35:50 ant_work: well it looks like your git still points to git.transition instead newer git-scm? Jun 22 11:36:11 it is in OE wiki Jun 22 12:04:02 khem, wrt the 64 bit problem I had, this is for gcc-cross-initial when compiling crtstuff.c using xgcc Jun 22 12:04:09 log is at http://filebin.ca/kgvddm/log.do_compile.7054 Jun 22 12:04:46 this is a gcc with nios2 patches, can make them available if desired Jun 22 12:06:56 anyone with strong opinion agains killing old mesa recipes? (older than 7.2)? I'm going to convert it to new staging, but with 3 levels of include files it's a bit messy.. Jun 22 12:16:48 hi - i have a system build with DBus for an ARM based platform (AT91RM9200) - when i try to start the dbus-daemon or a test program using DBus, i get a segmentation fault often with an alignment trap - here is a backtrace for the test program: http://pastebin.com/raw.php?i=C0qjLqGR - and here for the dbus-daemon: http://pastebin.com/raw.php?i=W814WxT4 - does anyone have any ideas or a similar problem? Jun 22 12:32:51 hi , i ve igepv2 card , i tried to use "bitbake console-image" on igepv2 and set machine=igepv2 , i got error "please set a valid machine in your local.conf , then i change it to igep0020 and tried igep0020b , whats wrong ? someone help ? Jun 22 12:41:54 JDuke128: looks like conf/machine/igep0020.conf exists Jun 22 13:20:29 yet another virtual keyboard http://florence.sourceforge.net/english.html Jun 22 15:17:33 zecke: hi. finnally i see you online. I hope you have some time Jun 22 15:18:01 i tried to build qtopia for x11 like you did 2 years ago and it sort of worked Jun 22 15:18:30 but i'd like to know your thoughts about how to proceed. also what i should have been looking at. Jun 22 15:18:43 apps worked. but there was no qpe or homescreen Jun 22 15:42:19 Hey anyone know the magic to have a recipe ignore the DL_DIR entirely and simply pull from source? Jun 22 15:43:02 darknighte, the question is, do you want to keep the "unpack/patch" process, but based on a local tree, or do you want to actually *build* in a local source tree? Jun 22 15:43:21 the former you can just use a file:///home/foo/bar/baz path in SRC_URI, it'll copy it into workdir in unpack and build there Jun 22 15:43:35 the latter, you can utilize the 'srctree' class, but it needs some TLC Jun 22 15:44:38 03Khem Raj  07org.openembedded.dev * r89f4040953 10openembedded.git/recipes/gcc/ (gcc-4.5.inc gcc-4.5/arm-bswapsi2.patch): Jun 22 15:44:38 gcc-4.5: Fix the bswapsi libgcc infinite loop at -Os Jun 22 15:44:38 Signed-off-by: Khem Raj Jun 22 15:44:52 kergoth_, not quite either. We are actively developing a new module and want to be able to pull the latest from trunk without having to manually remove the tar balls and local checkout Jun 22 15:45:03 ah Jun 22 15:45:05 hmm Jun 22 15:45:15 okay, here's an idea Jun 22 15:45:24 provide a custom do_unpack Jun 22 15:45:32 which just does a git clone directly into the workdir Jun 22 15:45:37 don't you just want a floating AUTOREV recipe? Jun 22 15:45:40 don't add the url to SRC_URI, so fetch does nothing Jun 22 15:45:44 heh, that'd work too :) Jun 22 15:45:45 that's a good thought. Jun 22 15:45:47 good point Jun 22 15:45:51 heh Jun 22 15:46:01 pb_, what's an floating AUTOREV recipe? :) Jun 22 15:46:02 * kergoth_ tries to avoid srcrev/autorev madness, for his sanity Jun 22 15:47:24 darknighte, SRCREV is how you specify the version you want when using an SCM URI in SRC_URI. if you set it to ${AUTOREV}. it'll use the latest Jun 22 15:49:33 kergoth_, I think I tried that, but I will take another look at it. If that doesn't work, I can always try the do_unpack idea. Jun 22 15:49:37 anyone want to take a look at the varref stuff I've been working on? I'm a bit stuck at the moment Jun 22 15:50:01 darknighte, the unpack method would be pretty straightforward too, i expect. one liner shell function :) Jun 22 15:50:57 grr Jun 22 15:52:48 kergoth_, I'm happy to help, but my experience is limited. Of course, explaining it to a "novice" might help.... ;) Jun 22 15:53:45 I'm working on code which tracks what variables reference what other variables, both at a high level for things like ${FOO}, and for things like shell functions running other shell functions, and python functions running getVar/expand to get at variables Jun 22 15:54:31 right now I'm trying a first pass integration with bitbake, which patches getvar/expand when a task starts so that it errors if the task tries to use a variable we didn't pick up during the parsing/analysis, so I can spot problems or missing explicitly marked refs Jun 22 15:55:47 kergoth_, sounds reasonable so far, but tricky due to multiple levels of indirection, perhaps? Jun 22 15:57:42 it does get tricky. I've added some recursion checking to find and report on cycles in variable references, but I keep running into ones that shouldn't exist, and which I can't seem to find the cause of Jun 22 15:57:48 it thinks things are recursing that aren't Jun 22 15:57:55 basically, this shit is starting to hurt my head :) Jun 22 15:58:19 hey, if it makes you feel better, it's making my head hurt thinking about it already. Jun 22 15:59:33 hehe Jun 22 16:00:18 as with most recursion problems, the trick is to detect and terminate the recursion appropriately. I've used forcible termination after "n" recurses, to help me in the past find issues. Jun 22 16:00:28 * kergoth_ nods Jun 22 16:00:59 other than that, I don't have much more for you... Jun 22 16:01:03 at the moment, i'm just maintaining a 'path' stack when the tree is traversed Jun 22 16:01:16 if the value you want is already there, it breaks out and reports that path as a cycle Jun 22 16:03:55 but clearly something is still hokey, if its hitting recursions that don't exist Jun 22 16:04:13 the odd thing is that the checking isn't actually finding this particular recursion. this is hitting a stack blown RuntimeError in python :| Jun 22 16:05:26 hmmm, two areas to focus on would be the detection and the path building. Either could throw you off pretty badly. Jun 22 16:07:48 yeah.. the difficulty is at the moment I have no idea how this is happening. i wonder how useful pdb would be with bitbake.. Jun 22 16:07:50 * kergoth_ ponders Jun 22 16:08:34 a coworker of mine used it to find an issue in the bowels of bitbake the other day.... YMMV Jun 22 16:08:46 ah, nice Jun 22 16:09:05 * kergoth_ normally doesn't bother with debuggers much, since usually a thorough knowledge of the code + prints will get you there, but sometimes its just not enough Jun 22 16:09:30 especially when following the twisty paths gives me a headache :) Jun 22 16:09:40 well, good luck. sorry I couldn't help more Jun 22 16:09:46 ttfn. Jun 22 16:09:52 no worries, helps just to talk about it Jun 22 16:10:15 * kergoth_ reads the code again in the hopes something will jump out at him Jun 22 16:32:34 Anyone try bitbaking angstrom-gnome-image ? I'm getting errors for libbonoboui packages saying "libgnomecanvas-2.0 missing" although I ensure libgnomecanvas is present by bitbaking it separately Jun 22 17:03:39 PythonExpansionError: 'RuntimeError: maximum recursion depth exceeded while calling a Python object' while resolving 'PythonSnippet(['compute_os_portion_of_target_triplet(d)'])' via: Jun 22 17:03:40 PythonSnippet(["bb.data.expand('staging-", VariableRef(['PN']), '-', VariableRef(['MULTIMACH_ARCH']), VariableRef(['TARGET_VENDOR']), '-', VariableRef(['TARGET_OS']), "', d).replace('_', '-')"]) -> Jun 22 17:03:40 Value([PythonSnippet(['compute_os_portion_of_target_triplet(d)'])]) Jun 22 17:03:43 * kergoth_ scratches head Jun 22 17:13:47 screwgoth: are you the same person as kergoth_ ? Jun 22 17:14:03 no Jun 22 17:14:21 screwgoth is a kergoth with screw ;) Jun 22 17:32:15 Jay7: Lol !! Jun 22 17:32:25 dcordes: No Jun 22 17:32:58 Jay7: that was what I was thinking Jun 22 17:33:06 glad I know better now Jun 22 17:33:11 dcordes: you are not alone with this :) Jun 22 17:33:34 ^^ Jun 22 17:35:49 * kergoth_ 's never anything but kergoth{,_,-,`}, generally Jun 22 17:52:07 03Martin Jansa  07org.openembedded.dev * r034c2cd49a 10openembedded.git/recipes/unixbench/unixbench_5.1.2.bb: Jun 22 17:52:07 unixbench: add make to RRECOMMENDS Jun 22 17:52:07 Signed-off-by: Martin Jansa Jun 22 17:52:18 03Martin Jansa  07org.openembedded.dev * r5cb3bdc4a3 10openembedded.git/ (30 files in 10 dirs): Jun 22 17:52:18 xorg: bump packages 2010-06-22 Jun 22 17:52:18 Signed-off-by: Martin Jansa Jun 22 17:59:47 eFfeM_work: hmm thats an ICE in cross-gcc Jun 22 18:00:07 JaMa: I committed a fix for -Os to work wirh gcc 4.5 Jun 22 18:00:20 JaMa: you can retry the benchmark with -Os Jun 22 18:01:25 khem: so should I enable OPTSPACE again and rebuild whole image? Jun 22 18:01:47 khem: have you seen those benchmarks from 5.1.2 I pasted here? Jun 22 18:07:36 yes Jun 22 18:07:45 I have seen the benchmarks Jun 22 18:07:51 they are neck to neck Jun 22 18:08:09 JaMa: so was the whole rootfile sytem compiled with same compiler too ? Jun 22 18:08:13 or just the benchmarks Jun 22 18:09:50 03Martin Jansa  07org.openembedded.dev * r51142306bf 10openembedded.git/recipes/xorg-driver/ (2 files in 2 dirs): Jun 22 18:09:50 xf86-video-glamo: add patch removing alloc calls deprecated in xserver 1.9 RC1 Jun 22 18:09:50 Signed-off-by: Martin Jansa Jun 22 18:09:57 03Martin Jansa  07org.openembedded.dev * rd6bf243f65 10openembedded.git/recipes/xorg-driver/ (4 files): xf86-input-(evdev,keyboard,mouse,tslib): bump PR of drivers used in SHR for rebuild with xserver 1.9 ABI Jun 22 18:10:06 khem: everything with same compiler Jun 22 18:10:38 khem: last time recompiled after you bumped SRCREV in gcc.. Jun 22 18:11:13 JaMa: ok Jun 22 18:11:14 thx Jun 22 18:31:54 03Koen Kooi  07org.openembedded.dev * re1a2447f2a 10openembedded.git/recipes/xorg-lib/ (3 files in 2 dirs): pixman 0.18.2: add 2 more neon patches Jun 22 18:49:01 can someone enlighten me about the usage of FILES_${PN} and (for example) installing things in the directory "/root" ? Jun 22 18:51:42 no packages should ever install anything into a home directory, including /root Jun 22 18:51:44 first of all Jun 22 18:52:43 kergoth_: It's an embedded device. Please answer the question. Telling me I shouldn't do it is not helpful. Jun 22 18:55:10 i'd suggest you read the docs. PACKAGES and FILES are straightforward and documented Jun 22 18:55:28 for each package in PACKAGES, a FILES_ specifies what files are included in that package, from the output of do_install Jun 22 18:55:35 it's that simple Jun 22 18:55:39 see conf/bitbake.conf for the default values Jun 22 18:56:11 this is openembedded, no shit it's an embedded device Jun 22 18:56:19 It seems counterintuitive to me that I didn't need a FILES_ line before, and now I do because I'm installing to a directory it doesn't expect. How do I know the list of "standard" directories? Why doesn't it do everything? Jun 22 18:56:31 once again, conf/bitbake.conf specifies the default values Jun 22 18:56:35 what part of that do you fail to grasp? Jun 22 18:56:38 you're too lazy to read it? Jun 22 18:57:43 The part I fail to grasp is your general bad attitude and profanity. Jun 22 18:58:27 it's not that hard to understand. you're too lazy to read the documentation, hence the bad attitude Jun 22 18:58:32 pretty straightforward Jun 22 18:58:59 You don't know me. Jun 22 18:59:09 i know you don't bother to read, that's enough Jun 22 19:01:34 I've read the docs. Doesn't mean my recall is perfect. I read a lot, actually. Jun 22 19:02:16 pb_/kergoth_, could either of you point me to a reference for an AUTOREV recipe? Jun 22 19:02:31 I've tried to get one to work and am so far not having much luck. Jun 22 19:08:50 how 'bout anyone else? Does anyone know of a good example of how to use AUTOREV properly? Jun 22 19:10:15 Anyone want to comment / ack the stage-manager-ipkg rfc I posted? Jun 22 19:10:55 or the pstage package arch stuff Jun 22 19:14:00 i will in a few, knee deep in this code Jun 22 19:25:35 pb__, were you talking about creating an AUTOREV recipe earlier with me? Jun 22 19:38:54 better yet, where can I find more out about SRCPV? Jun 22 19:56:12 darknighte, ask JaMa Jun 22 19:56:30 he's involved in SRCPV Jun 22 19:58:24 GNUtoo, thanks! Jun 22 19:58:46 darknighte, altough you may have to wait for him Jun 22 20:00:01 I may have found the issue on my own in any case. Jun 22 20:00:10 ok Jun 22 20:00:11 nice Jun 22 20:00:20 silly mistake in the SRC_URI.... Jun 22 20:04:37 ok Jun 22 20:14:56 03Koen Kooi  07org.openembedded.dev * r2ecd529f30 10openembedded.git/recipes/opencv/gst-opencv_git.bb: gst-opencv: bump SRCREV, fix versioning Jun 22 20:15:04 03Koen Kooi  07org.openembedded.dev * r34a12af48d 10openembedded.git/recipes/opencv/opencv_svn.bb: opencv: bump SRCREV and fix compat header Jun 22 21:50:59 * dcordes is wondering which asshat terminated his irc client Jun 22 21:54:32 GNUtoo: do you get this same error compiling gpm ? http://bugs.openembedded.org/show_bug.cgi?id=2114#c9 Jun 22 21:54:41 hmmm Jun 22 21:55:02 | gpm.c:813: error: storage size of 'sucred' isn't known Jun 22 21:55:05 same error Jun 22 21:55:26 what are the deps? Jun 22 21:55:52 PARALLEL_MAKE = "" Jun 22 21:56:07 hmmm that sound not good Jun 22 21:56:37 DEPENDS = "ncurses" Jun 22 21:56:53 GNUtoo: you think accidentally deps are not provided ? Jun 22 21:57:00 no Jun 22 21:57:15 I think deps<->gpm are not in sync Jun 22 21:57:32 ah that makes sense Jun 22 21:58:00 | gpm.c:813: warning: unused variable 'sucred' Jun 22 21:58:15 | gpm.c:814: error: invalid application of 'sizeof' to incomplete type 'struct ucred' Jun 22 21:58:19 I'll look Jun 22 21:58:57 struct ucred sucred; Jun 22 21:59:51 grep ucred -r * returns no declaration Jun 22 21:59:54 => deps problem Jun 22 22:00:23 http://bugs.gentoo.org/show_bug.cgi?id=241144 Jun 22 22:01:33 so there are 2 options Jun 22 22:01:35 or patch Jun 22 22:01:36 http://recipes.gobolinux.org/r/?list=OpenLDAP&ver=2.3.39-r2&file=01-ucred.patch Jun 22 22:01:39 or bump Jun 22 22:01:45 I prefer bump Jun 22 22:02:16 me too. can you provide a patch to test ? Jun 22 22:02:46 http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/sys-libs/gpm/gpm-1.20.4.ebuild?r1=1.2&r2=1.3 Jun 22 22:02:53 maybe that would be simplier Jun 22 22:03:29 I'll try new gpm Jun 22 22:03:50 hmmm I would have to rebase the patches Jun 22 22:03:59 and test everything Jun 22 22:04:08 maybe you could do it Jun 22 22:04:16 it's quite simple Jun 22 22:04:18 just bump gpm version up ? Jun 22 22:04:23 yes like this: Jun 22 22:04:26 I think I can manage that Jun 22 22:04:56 although I'm the biggest noob with org.oe.dev access : Jun 22 22:05:07 cp gpm_1.20.1.bb gpm_1.20.6.bb Jun 22 22:05:12 modify the checksum Jun 22 22:05:15 remove all the patches Jun 22 22:05:38 and look if all the patces have been included Jun 22 22:05:47 s/remove all the patches/ Jun 22 22:06:17 rebase the patch if they don't apply Jun 22 22:06:34 and finally test it and send a patch to the oe mailing list if you don't have commit access Jun 22 22:06:51 if it still don't work Jun 22 22:06:55 keep the new recipe Jun 22 22:06:58 and add a patch Jun 22 22:07:02 that is to say: Jun 22 22:07:13 file://my_patch;apply=yes \ Jun 22 22:07:25 note that the apply=yes is new Jun 22 22:07:38 the my_patch.patch is done with quilt Jun 22 22:07:41 in: Jun 22 22:07:52 /home/gnutoo/embedded/oe/oetmps/htcdream/work/armv6-novfp-oe-linux-gnueabi/gpm-1.20.1-r3/gpm-1.20.1/ Jun 22 22:08:14 I think that's all Jun 22 22:34:08 OK, more $ORIGIN stuff to the ML Jun 22 22:39:20 it's trickier than I thought Jun 22 22:39:43 anybody willing to bump gpm to http://www.ar.linux.it/pub/gpm/gpm-1.99.7.tar.gz ? Jun 22 22:40:09 old configure patch doesn't apply Jun 22 23:58:07 dcordes: patch the old recipe Jun 22 23:58:16 if thats less work Jun 22 23:58:33 khem: started with copy of old recipe Jun 22 23:59:03 khem: put the correct checksums.. but I'm not good at handling non-git patch stuff Jun 23 00:00:38 ok Jun 23 01:16:37 quitte: Qtopia is not supported by Nokia anymore Jun 23 01:35:45 dcordes: ok I have a patch to add new gpm recipe could you test it ? Jun 23 01:35:55 zecke: hi Jun 23 01:36:25 zecke: if not qtopia then do they supporte qte instead Jun 23 01:36:33 khem: nice. if you put it somewhere I will try it later toeday Jun 23 01:36:48 dcordes: I will send it mailing list Jun 23 01:37:07 its completely untested, so I will wait for you tests Jun 23 01:43:01 dcordes: check the ml I just posted it out Jun 23 01:43:46 ok Jun 23 01:44:03 note to self: really subscribe to oe ml Jun 23 01:45:22 dcordes: http://article.gmane.org/gmane.comp.handhelds.openembedded/33743 Jun 23 01:46:10 hmm, did anyone have any thoughts on the siteinfo patch? Jun 23 01:49:38 khem: Qt Embedded/Linux is supported Jun 23 01:53:17 kergoth: I liked it sort of Jun 23 01:55:16 zecke: cool ok Jun 23 01:56:05 zecke: so suppose If I want to have a nice gui on top of my qemu what would you recommend :) Jun 23 01:56:32 hehe. sort of. anything in particular you disliked? I think splitting the arch/os/target makes it easier to support, and not setting site files for native avoids the need to add darwin crap when darwin is only the build machine, not the target Jun 23 01:57:04 i thought about switching the content of hte info dictionaries to be lists, rather than space separated strings, but didn't bother Jun 23 01:58:45 khem: it depends, Qtopia had support for virtual keyboards, and had some implementations, QtE only provides key dispatching... but yes if you only want to have a shiny demo QtE is fine Jun 23 01:58:52 kergoth: sort of means I really did not review it fully yet :) Jun 23 01:59:16 zecke: demo is good Jun 23 01:59:18 ah, right ;) Jun 23 02:00:03 zecke: do we already have an image target for flashy qte Jun 23 02:00:05 * khem checks Jun 23 02:02:21 * kergoth_ has a pile of mails in his inbox awaiting his review / action as well, never seems to get caught up **** ENDING LOGGING AT Wed Jun 23 02:59:56 2010