**** BEGIN LOGGING AT Fri Mar 16 02:59:57 2007 Mar 16 06:09:30 03lenehan 07org.oe.documentation * rb4a59a6f... 10/ (1 usermanual/reference/dirs_staging.xml): Mar 16 06:09:30 reference/dirs_staging: Update some paths to match recent change make to Mar 16 06:09:30 bitbake.conf. Mar 16 06:09:32 03lenehan 07org.oe.documentation * r3ed18fe4... 10/ (1 usermanual/chapters/recipes.xml): Mar 16 06:09:32 recipes: Update the note about the PR variable making it clear that it never Mar 16 06:09:33 goes backwards and that you should always increase it if you are unsure Mar 16 06:09:35 about what effect it has on the generated packages. Mar 16 06:09:39 03nail 07org.oe.dev * r8c849d9d... 10/ (1 packages/xorg-lib/diet-x11/makekeys.diff): diet-x11: Re-add the makekeys-patch, just don't add it to the diet-x11 SRC_URI, since it's already being applied from libx11 Mar 16 06:09:42 03nail 07org.oe.dev * r2f8432d4... 10/ (1 packages/xfsprogs/xfsprogs_2.8.16.bb): Mar 16 06:09:44 xfsprogs: Change STAGING_BINDIR to STAGING_BINDIR_NATIVE, closes Mar 16 06:09:46 !oebug 1954 Mar 16 06:09:47 * * Bug 1954, Status: RESOLVED, Created: 2007-03-08 09:16 Mar 16 06:09:48 * * nail(AT)nslu2-linux.org: xfsprogs fails to build for powerpc on x86_64 host Mar 16 06:09:48 03lenehan 07org.oe.dev * rf2326806... 10/ (10 files in 3 dirs): Mar 16 06:09:48 * * http://bugs.openembedded.org/show_bug.cgi?id=1954 Mar 16 06:09:52 radvd: Add init scripts and sample configurations to documentation. Mar 16 06:09:54 - The init script, radvd.init, is directly from the debian package. Mar 16 06:09:56 - The radvd.conf.empty is from the redhat specs dir in the 1.0 Mar 16 06:09:58 sources - added here rather then using it from the package because Mar 16 06:10:00 it wasn't present in the 0.7.2 and I want to avoid breaking that Mar 16 06:10:02 version. Mar 16 06:10:04 03nail 07org.oe.dev * re54ea8f0... 10/ (1 packages/linux/linux_2.6.20.bb): linux: Fix install on n2100 Mar 16 06:10:08 03rwhitby 07org.oe.dev * r35bc21cf... 10/ (1 packages/linux/ixp4xx-kernel_2.6.20.bb): ixp4xx-kernel: Updated to latest SVN, including moving nas100d microcode to a separate partition. Mar 16 07:06:38 good morning Mar 16 07:14:12 morning all Mar 16 07:14:32 NAiL : galery was failing at libexif ? Mar 16 07:39:29 psokolovsky_: I was able to build opie-image. Thanks - next time I'll recheck my scripts and make sure I do follow the wiki... :) Mar 16 08:46:42 morning Mar 16 09:05:55 <_law_> can i set GUI_MACHINE_CLASS_aximx50 ? or is there no machine specific function available Mar 16 09:07:35 where would you want to do that? Mar 16 09:08:33 <_law_> mickeyl, machine.conf but this doesnt work i think.. i habe to build a second machine.conf for the axim x50 (with bigscreen fuinction removed) Mar 16 09:08:45 <_law_> mickeyl, and own kernel defconfig... Mar 16 09:08:48 i'm afraid so Mar 16 09:08:59 unfortunately we don't have the concept of sub-architectures or machines yet Mar 16 09:09:05 at least no consistent one Mar 16 09:09:21 the whole GUI_MACHINE_CLASS is messy anyway Mar 16 09:09:35 sooner or later we will (hopefully) realize we need a full device profile Mar 16 09:10:08 and some logics to build stuff depending on that Mar 16 09:10:46 <_law_> ok so i´ll build 2 machine.confs... Mar 16 09:11:29 yeah. to minimize duplication you may build one include conf Mar 16 09:11:38 and just add the unique stuff to your final confs Mar 16 09:12:55 Ifaistos: Dunno where it failed. It's not exactly important :-P Mar 16 09:18:22 morning all Mar 16 09:22:38 morning Mar 16 09:40:14 RP: morning Mar 16 09:40:43 NAiL: Got it to build both for x86 and powerpc after twiking libexif Mar 16 10:08:29 03pH5 07org.oe.dev * r12b22f97... 10/ (1 packages/mplayer/mplayer_0.0+1.0rc1.bb): Mar 16 10:08:29 mplayer: disable the ffvorbis codec (ffmpeg) in mplayer on ARM machines Mar 16 10:08:29 - instead of this floating point codec, the fixed point vorbis codec Mar 16 10:08:30 (tremor) is used instead. Mar 16 10:47:27 I'm having an odd problem with dependencies of native packages Mar 16 10:48:14 trying to build mtools-native, which will not build on my system (ubuntu-edgy) because of the version of install-info that is installed here Mar 16 10:48:44 the fix is to build texinfo-native first, which provides the correct version of install-info Mar 16 10:49:13 however, I'm adding DEPENDS="texinfo-native" to mtools-native, but texinfo-native is not being built Mar 16 10:49:44 if i manually bitbake texinfo-native, then it's fine... just isn't happening through DEPENDS Mar 16 10:50:20 can I see the dependency tree for a package somehow? Mar 16 10:54:25 why does bb build opie-image with a crippled rootfs (akita on both angstrom and openzaurus). will not boot. there's no init and a lot of stuff is missing. Mar 16 11:26:12 hi Mar 16 11:35:01 psokolovsky: Does your "MACHINE=foo" trick work when the machine is set via auto.conf? Mar 16 11:37:41 Hi! Mar 16 11:38:04 Laibsch, I use local.conf only. for me, it works. Mar 16 11:38:45 mhulden, because you haven't read http://www.linuxtogo.org/gowiki/OpieWithAngstrom Mar 16 11:40:13 Bernardo|away, good news, thanks for report ;-) I assume it boots too ;-). as for wiki, there's "monitor page" feature to stay in loop Mar 16 11:42:45 psokolovsky: I ask because I just built for collie where auto.conf (just a split-up local.conf) specifies spitz. A couple of packages are being built for armv5te. Mar 16 11:44:39 Laibsch, what MACHINE it said in blueb on bitbake startup? Mar 16 11:44:50 Laibsch, also, what collie cpu is? Mar 16 11:49:50 psokolovsky: AFAIK collie is StrongARM SA11xx (but only 90% sure) Mar 16 11:50:19 schurig, thanks Mar 16 11:50:51 Laibsch, then, you'd need to change DISTRO (if you build Anstrom) Mar 16 11:51:06 and DISTRO can't be passed on command line, unfortunately Mar 16 11:54:14 can you use SD wifi cards? Mar 16 11:58:50 Laibsch: or fix gcc so armv4 can support EABI :-) Mar 16 12:00:13 good morning Mar 16 12:00:35 WOW Mar 16 12:00:47 wifi cards now come in MINI SD size Mar 16 12:01:46 hmm a micro sd wifi card for my motorola would rock ;) Mar 16 12:02:10 i thought SD was small Mar 16 12:02:14 mini sd is awesome Mar 16 12:02:19 micro sd.... i'd lose it Mar 16 12:02:41 * XorA has to keep his micro cards in the platic case Mar 16 12:02:57 heh Mar 16 12:02:59 look: http://cgi.ebay.com.au/802-11b-Mini-SD-WiFi-Wireless-Card-4-PDA-smartphone-NEW_W0QQitemZ220076194483QQcategoryZ45003QQtcZphotoQQcmdZViewItem Mar 16 12:03:04 can;t believe they exist :P Mar 16 12:03:09 I have enough problems holding on to my usbsticks Mar 16 12:03:29 daurnimator: http://www.spectec.com.tw/sdw823.htm Mar 16 12:03:45 daurnimator: http://www.expansys.com/p.aspx?i=140798 Mar 16 12:03:48 * daurnimator orgasms Mar 16 12:04:08 Our Price: Not announced Mar 16 12:04:08 Availability: Pre-order Now Mar 16 12:04:56 so wifi in a neo anyone? Mar 16 12:05:12 neo? Mar 16 12:05:32 daurnimator: www.openmoko.org Mar 16 12:05:45 heh Mar 16 12:05:52 it rang a bell todo with moko Mar 16 12:05:57 i dismissed the link though Mar 16 12:20:35 hi I'm trying to compile dillo and I 've got this problem with diet-x11 IOError: [Errno 2] No such file or directory: '/stuff/org.openembedded.dev/packages/xorg-lib/diet-x11/./makekeys.diff' Mar 16 12:20:41 Can anyone help me? Mar 16 12:21:49 I'll see if I can locate that file and place it in that dir and try recompiling, for some reason i'm having problems with that particular package Mar 16 12:22:20 darmou : it was fixed today. sync with OE to get the changes Mar 16 12:22:47 great thanks Ifaistos much appreciated! Mar 16 12:55:27 I'd say claims that mtn now has decent log performance are exaggerated (if there were such, of course) Mar 16 13:02:24 any idea why I'd be getting a "NOTE: preferred version 2.5 of glibc-intermediate not available", followed by an attempted build of glibc-intermediate-2.3.2+cvs20040726 (which dies) since yesterday? I don't think I changed anything here and it was using 2.5 just fine before. I tried touching conf/local.conf to get bitbake to rescan but that didn't help. I even got it to rebuild 2.5, but it still stays 2.5 is not available when I try to build anyth Mar 16 13:15:41 Laibsch, you touched glibc lately? ;-) Mar 16 13:16:11 ~lart Laibsch Mar 16 13:16:11 * ibot decapitates Laibsch conan the destroyer style Mar 16 13:16:51 * XorA chuckles, harsh but fair Mar 16 13:20:19 apaulsen: what distro? Mar 16 13:20:48 my own = generic + some extra packages Mar 16 13:21:50 psokolovsky: Why the bad treatment? Mar 16 13:22:30 Laibsch, see above and all other cases this issue was brought already (I counted 3 myself) Mar 16 13:28:00 slapin_nb, try to -c rebuild it Mar 16 13:29:40 psokolovsky, trying... Mar 16 13:38:21 slapin_nb, ok, so somehow it build for me now Mar 16 13:49:54 slapin_nb, confirmed? Mar 16 13:53:49 psokolovsky, works Mar 16 13:54:36 Laibsch, I gues you should post about this to ML ^ Mar 16 14:11:26 ~seen RP Mar 16 14:11:56 rp is currently on #gpe (1d 8h 38m 30s) #oe (1d 8h 38m 30s) #openmoko (1d 8h 38m 30s) #openzaurus (1d 8h 38m 30s) #handhelds.org (1d 8h 38m 30s). Has said a total of 1 messages. Is idling for 4h 53m 33s, last said: 'morning all'. Mar 16 14:13:00 RP: I have a quick qestion about the sharp bootloader. Aside from the cmdline, is the rest of the information (ie memory setup) correct? Mar 16 14:21:58 so are other people having problems like this then? http://www.pasteserver.net/541 Mar 16 14:32:17 Ok, I'll try rebuilding from scratch then. Mar 16 14:32:33 apaulsen, please post to ML, RFC reverting broken changes to glibc, unless their author doesn't fix them timely Mar 16 14:34:47 psokolovsky: I'm not sure if it is a broken change or just something stupid that I did so I wanted to see if there was anyone else with a similar problem. If my rebuild doesn't work, I'll send something to the ML. I looked at the glibc changes from yesterday and I can't see any reason why it wouldn't work now. Mar 16 14:39:50 that's the reason why changes to the core components should go thru the proper change control (RFC, approval from core mainatiners, notifications, etc.) Mar 16 14:46:23 psokolovsky: I moved packaging for some obviously debug stuff into debug package. Does this really break compilation? Mar 16 14:47:17 I would not have touched this if I had thought this was anything more important than moving files to the proper packages Mar 16 14:49:43 * Laibsch writes to ML Mar 16 14:53:24 psokolovsky: I think what apaulsen is seeing is that his angstrom build is building glibc-2.3 instead of 2.5, not necesarily that 2.5 is failing Mar 16 14:54:27 2.5 builds, but then bitbake can't find it and tries to build 2.3 Mar 16 14:57:15 apaulsen, psokolovsky, I confirm this bug (if anyone cares) Mar 16 15:00:08 slapin_nb: which distro are you using? Mar 16 15:02:23 apaulsen, angstrom Mar 16 15:02:43 apaulsen, build is from scratch Mar 16 15:03:30 apaulsen, http://builds.hackndev.com/builds/angstrom/daily-1173988621/build-palmz72.log.gz Mar 16 15:03:45 apaulsen, it is with bitbake -DDD Mar 16 15:07:16 Should I revert the commit? Mar 16 15:07:25 apaulsen: have you tried building angsrom-x11-image instead of gpe-image? Mar 16 15:07:37 I'm not using angstrom... Mar 16 15:07:57 I tried building task-base though and that also fails Mar 16 15:08:32 Ifaistos: try to build with seppuku so we get bug reports of failed packages Mar 16 15:08:42 btw, any chance to connect my builds to tinderbox? Mar 16 15:09:33 slapin_nb: sure Mar 16 15:10:00 zecke, how? Mar 16 15:10:00 slapin_nb: there should be some documentation, and use the angstrom tree to report to Mar 16 15:10:08 apaulsen: try blowoing out the cache in Mar 16 15:10:36 zecke, could you please supply link? Mar 16 15:10:38 apaulsen: try blowoing out the cache in /home/packet/OpenEmbedded/build/tmp/cache/compulab-pxa270' to see if there are any parsing errors in any of the glibc files Mar 16 15:10:48 slapin_nb: check conf/tinder.conf and paste the lines into the local.conf of your build Mar 16 15:12:30 slapin_nb: I just ran with -DDD as well and I have the same error starting on line 44089 Mar 16 15:13:31 hvontres|poodle: cache nuke done, parsing in progress Mar 16 15:15:01 * hvontres|poodle thinks there is a typo somewhre... Mar 16 15:15:47 /s/somewhre/somewhere :) Mar 16 15:16:09 apaulsen: hey, you found it :) Mar 16 15:16:25 ~lart lysdexia Mar 16 15:16:26 * ibot slaps a compatible dib on lysdexia's head Mar 16 15:16:51 slapin_nb : use this -> http://rafb.net/p/zehGHH15.html make a file called tinder.conf and included into your local.conf Mar 16 15:17:15 slapin_nb : just change the name of tinder machine Mar 16 15:17:30 hvontres|poodle: cache nuke & rebuild failed Mar 16 15:17:54 apaulsen: which file fails? Mar 16 15:17:55 zecke : I just inherit seppuku or does it need any configs ? Mar 16 15:18:45 NOTE: preferred version 2.5 of glibc-intermediate not available Mar 16 15:18:45 NOTE: multiple preferred providers are available (glibc, glibc, glibc-intermediate); Mar 16 15:18:45 NOTE: consider defining only one PREFERRED_PROVIDER to match runtime glibc-gconv-iso8859-1 Mar 16 15:19:25 then it tries to build glibc-intermediate 2.3.2 and dies Mar 16 15:19:44 apaulsen: but no parsing errors? Mar 16 15:20:01 hvontres|poodle: no parsing errors Mar 16 15:20:17 apaulsen: well, there went my theory...:( Mar 16 15:21:09 slapin_nb: I think http://www.openembedded.org/wiki/TinderBox might help you to understand what the entries actually do. Mar 16 15:24:57 Ifaistos: I assume you will at least need to configure the bugzilla to report to, your username and your password. Just inherit will not be sufficient looking at classes/seppuke.bbclass Mar 16 15:26:03 zecke: Any documentation apart from the source for seppuku yet? Mar 16 15:29:03 Laibsch: check if bitdoc still works, or go to conf/documentation.conf Mar 16 15:29:20 Laibsch: but there is no documentation on how to use it on a bugzilla != openmoko Mar 16 15:30:04 Thanks Mar 16 15:30:16 apaulsen: Are you building opie for angstrom by chance? Mar 16 15:30:39 gpe-image for a lightly modified generic Mar 16 15:32:59 Laibsch: thanks have not look at seppuku.bbclass at all Mar 16 15:33:30 Ifaistos: I will put the openmoko config on the wiki and explain what to do to make it report to the OE bugtracker Mar 16 15:33:57 zecke : ok Mar 16 15:39:29 Good afternoon Mar 16 15:39:36 zecke: That would be awesome. Mar 16 15:42:23 zecke : i am planning on doing a complete (as in 'possible') build on a standard base i.e every month Mar 16 15:42:24 ouch mtn messed up my db :} Mar 16 15:44:35 Ifaistos: Oh, so you are the Stelios on the ML? Just wrote an answer to inform you about seppuke ;-) Mar 16 15:44:42 seppuku Mar 16 15:45:00 zecke: BTW, why seppuku? Wanna slit your gut? Mar 16 15:45:05 yep Mar 16 15:45:33 i go after several aliases so ex-girlfriends can't track me easily :) Mar 16 15:50:18 Laibsch: hehe Mar 16 15:50:30 morning Mar 16 15:50:39 Laibsch: well, I think if a build knows it will fail soon, it is best to commit proper suicide Mar 16 15:51:09 Laibsch: while writing this class I browsed wikipedia for samurai... Mar 16 15:59:11 zecke: how to choose a propper mimetype and which mimetype would you suggest? what do you mean with "please select patch"? Mar 16 16:02:37 hmm Mar 16 16:02:52 what's the ${}-name for g++? Mar 16 16:10:04 mr_nice: click on details Mar 16 16:10:27 mr_nice: and there is a "[] patch" checkbox, your mimetype was application/octet-stream which is hard to view inline :) Mar 16 16:10:40 mr_nice: your patch contained whitespace changes again :) Mar 16 16:11:34 zecke: thx Mar 16 16:11:43 Laibsch: not knowing the surrounding, you might want to disapprove your cset Mar 16 16:11:52 03nail 07org.oe.dev * r2b3568b1... 10/ (1 packages/postgresql/postgresql_8.1.4.bb): postgresql: changed url Mar 16 16:11:54 NAiL: CXX? Mar 16 16:11:56 03nail 07org.oe.dev * r41fe9ff5... 10/ (1 packages/postgresql/postgresql_8.1.8.bb): postgresql: Upstream upgrade to 8.1.8 Mar 16 16:11:58 zecke: is sysfs hard to understand? Mar 16 16:12:00 03nail 07org.oe.dev * rb9293118... 10/ (1 packages/meta/foonas-packages.bb): foonas: Add postgresql to feed Mar 16 16:12:07 mr_nice: no, very easy Mar 16 16:12:18 mr_nice: you can add files and define read/write methods Mar 16 16:12:30 zecke: Yeah, I spelled it CXx, and didn't see it ;) Mar 16 16:12:44 That's why I wondered Mar 16 16:12:46 zecke: ok, I will read the documentation Mar 16 16:13:19 mr_nice: there might be some more generic way to enable disable power (e.g. add devices to the platform tree, but I don't know which way is prefered) Mar 16 16:13:51 mr_nice: I'm sure /proc/cs3 is not the proper one :) Mar 16 16:14:04 esound is pulling in host libraries.... Mar 16 16:15:26 zecke: OK Mar 16 16:15:32 zecke: hm, maybee I will ask at #kernelnewbies Mar 16 16:45:41 Laibsch: you missed a space in your glibc commit Mar 16 16:45:56 Laibsch: += has to start with a space Mar 16 16:46:22 Laibsch: look at the removed line in the patch Mar 16 16:47:18 how is monotone 0.33? Mar 16 16:49:27 XorA: Thanks for looking into it. Mar 16 16:49:49 My slow computer has been calculating how to revert all this for the last 15 minutes or so. Mar 16 16:49:50 Laibsch: but I cant see how that has made glibc 2.3.2 become the default Mar 16 16:50:04 I cannot either how any of my commits did that Mar 16 16:50:12 I will revert all of them Mar 16 16:50:19 We can still apply them at a later time. Mar 16 16:51:44 Laibsch: found it Mar 16 16:51:56 Really? Mar 16 16:52:01 PROVIDES += "virtual/libintl virtual/libiconv" should have a leading space in glibc.inc Mar 16 16:52:13 OK Mar 16 16:52:18 otherwise it backs onto the previous PROVIDES= Mar 16 16:52:30 Alright. Mar 16 16:52:53 * XorA tests Mar 16 16:52:56 So you think that adding those two spaces should fix the mess? Mar 16 16:53:07 Let me know about the result. You were seeing this? Mar 16 16:53:18 nope still tries to build 2.3.2 Mar 16 16:54:12 OK, my computer is almost through. Mar 16 16:54:19 I will revert then. Mar 16 16:54:29 gimme a sec Mar 16 16:54:52 Sure Mar 16 16:54:52 nope, fixed all the missing spaces and still problems Mar 16 16:55:21 More than two? Mar 16 16:55:35 Laibsch: not all yours Mar 16 16:55:43 I thought I had read somewhere that spaces after the " were not needed. Mar 16 16:56:41 quotes are necessary (at least I want people to believe that) :) Mar 16 16:57:25 zecke: I meant that I thought I had read that PROVIDES = " foo" and PROVIDES = "foo" were equivalent. Mar 16 16:57:26 hmm, one of += or _append adds space, other doesnt, I can never remeber Mar 16 16:57:51 XorA: IIRC += adds space Mar 16 16:57:53 XorA: How about making them both add space? Mar 16 16:58:08 Laibsch: I was wrong in my initial assessment then Mar 16 16:58:10 Seems like this should not break anything if there are two or more spaces at times. Mar 16 16:58:19 btw, bisection method could be used here to find which line causes problem... (just IMHO). Mar 16 16:58:36 XorA: No. I am not sure about this being really equivalent. Mar 16 16:58:47 Laibsch: we don't want to break old metadata (well we sadly do sometimes). and e.g. some variables like OVERRIDES are sensitive to spaces Mar 16 16:59:02 OK, it asks me for the password to push the reversion upstream. Shall I do it? Mar 16 16:59:45 * Laibsch reverted all changes to packages/glibc/* done by me after the latest commit from Ifaistos Mar 16 17:02:54 ah well time to escape from work Mar 16 17:04:47 03Laibsch 07org.oe.dev * rbe6200db... 10/ (13 files in 2 dirs): disapproval of revision '2c516898480b1bc053c9cbf6da853a016059c4c6' Mar 16 17:04:54 03Laibsch 07org.oe.dev * rc3ff8bda... 10/ (13 files in 2 dirs): disapproval of revision '286930242424d10b2503eb290dbc31c7975bab74' Mar 16 17:05:00 03Laibsch 07org.oe.dev * r22194037... 10/ (1 packages/glibc/glibc-package.bbclass): disapproval of revision 'a96c9acac99b5599d2b0fb0a518603646414d166' Mar 16 17:07:47 psokolovsky: re MACHINE=foo, I was building for openzaurus: http://rafb.net/p/X3llQq70.html Just providing the MACHINE var on the command line does not seem to change a thing. Mar 16 17:12:03 Laibsch: are you inheriting multimachine in your local.conf? Mar 16 17:14:54 hvontres|poodle: I don't think so. But I thought multimachine was part of OZ? Mar 16 17:18:25 Laibsch: hmmm, your right. did you try commenting out the MACHINE setting in your local.conf? Mar 16 17:18:28 Laibsch, I have no idea. Using cache in '/export/interim/openembedded/tmp/openzaurus/cache/spitz' looks right Mar 16 17:19:17 Laibsch: or maybe using MACHINE ?= instead of MACHINE = ? Mar 16 17:20:56 psokolovsky: No, the cache in the collie dir should have been used, of course. Mar 16 17:21:12 hvontres|poodle: ?= sounds like a very good idea Mar 16 17:22:24 Laibsch, I mean, multimachine is in place Mar 16 17:24:20 psokolovsky: Ah, alright. Mar 16 17:25:48 XorA|gone: BTW, the PROVIDES lines I moved to .inc were all of the no-space-at-beginning type. I usually copy and paste when unifying to create as exact copies as possible. It seems the space is indeed not needed (at least it was not until now). Mar 16 17:35:36 re Mar 16 17:38:39 Hi Koen! When you're back? Mar 16 17:38:58 psokolovsky: sunday night Mar 16 17:39:15 anyone know if it's possible to compile static c++ binaries in OE? :) Mar 16 17:39:23 cool! hope you have great time in the meantime ;-) Mar 16 17:40:05 sure Mar 16 17:40:09 NAiL, tried to use {C/CXX/LD}FLAGS? Mar 16 17:40:16 there's this big summer festival tonight Mar 16 17:41:05 summer. ummm... ;-) Mar 16 17:41:20 (even if it's autumn over here) Mar 16 17:41:31 * slapin_nb read 'big' as bug.. oh my... Mar 16 17:41:48 psokolovsky: uh, no ;) Mar 16 17:42:15 hvontres|poodle: Thanks for the ?= hint. That did it! Mar 16 17:43:32 Laibsch, oops. I looked now and see that I use that too ;-) Mar 16 17:43:45 another distro testing OE: http://labs.vivi.eng.br/blog/ Mar 16 17:44:28 can someone bribe my teachers so I can work on getting LLVM to work in OE instead of doing mandatory courses? Mar 16 17:47:04 hey Mar 16 17:47:06 otw home? Mar 16 17:56:16 03nail 07org.oe.dev * r4b96e486... 10/ (1 packages/lvm2/lvm2_2.01.15.bb): lvm2: Add PR so the package doesn't end in PV- Mar 16 17:56:21 03nail 07org.oe.dev * r5a870b11... 10/ (1 packages/uboot-utils/files/fw_env.config): uboot-utils: Add a default fw_env.config so the build doesn't fail when built for a new machine Mar 16 18:03:54 XorA|gone, slapin_nb: Can you please update your OE tree and see if the reversion fixed the glibc problem for you? Mar 16 18:05:46 Laibsch, I'm building with tinderbox now but it is not full image, seems to be passed for now Mar 16 18:06:17 Laibsch, I'll retry that as next build starts Mar 16 18:06:43 Laibsch, it will update tree automatically. Mar 16 18:08:19 03nail 07org.oe.dev * r9b860048... 10/ (5 files in 4 dirs): fis: Add utility for parsing/editing a Redboot FIS partition layout from userspace Mar 16 18:09:57 slapin_nb: Was that before or after the reversion? Mar 16 18:10:30 Laibsch, dunno, you could compare time Mar 16 18:10:43 * Laibsch still has a hard time to believe the trivial changes to packaging and the semantically equivalent unification changes triggered this. Mar 16 18:11:20 Laibsch, that could well be something else, Mar 16 18:11:27 slapin_nb: Well, the most important question is: When did you last do "mtn pull;mtn up"? Mar 16 18:12:06 More or less than an hour ago? Mar 16 18:12:33 slapin@slapin:~/OE/org.openembedded.dev$ mtn log |head Mar 16 18:12:33 ----------------------------------------------------------------- Mar 16 18:12:33 Revision: b9293118c194a01c44b71ffff72e461a3fe86807 Mar 16 18:12:33 Ancestor: 41fe9ff5e157a956581beb94d9aeed0e5ba535d9 Mar 16 18:12:33 Author: nail@nslu2-linux.org Mar 16 18:12:34 Date: 2007-03-16T15:27:37 Mar 16 18:12:36 Branch: org.openembedded.dev Mar 16 18:12:38 Modified files: Mar 16 18:12:40 packages/meta/foonas-packages.bb Mar 16 18:13:39 Date is in GMT Mar 16 18:14:13 IOW, you still have a tree with the glibc changes I made. Mar 16 18:14:22 Now I wonder why it suddenly builds for you. Mar 16 18:15:11 Laibsch, becaus I don't build image now, I build package libgpewidget Mar 16 18:15:20 *because Mar 16 18:15:42 seems to be something is broken next. Mar 16 18:16:10 as soon as this build finishes, new one will start with full image build. Mar 16 18:16:32 but you would need glibc nonetheless, wouldn't you? Mar 16 18:16:45 Let's see how the full build one goes. Mar 16 18:16:57 Laibsch, yes, so as I see, this is what is strange here. Mar 16 18:17:25 I tried to reproduce your problem with and without my changes and was unable to do so in both cases. glibc was the same version as always. Mar 16 18:21:49 Laibsch, last info from me is that glibc2.3.5 is being pulled by gtk+ %) Mar 16 18:23:37 bitbake gtk+ Mar 16 18:23:46 NOTE: preferred version 2.5 of glibc not available Mar 16 18:23:47 NOTE: preferred version 2.5 of glibc-initial not available Mar 16 18:23:47 NOTE: preferred version 2.5 of glibc not available Mar 16 18:23:47 NOTE: preferred version 2.5 of glibc-intermediate not available Mar 16 18:23:47 NOTE: multiple preferred providers are available (glibc, glibc, glibc); Mar 16 18:23:47 NOTE: consider defining only one PREFERRED_PROVIDER to match runtime glibc-gconv-iso8859-1 Mar 16 18:23:49 NOTE: package glibc-2.3.5+cvs20050627: started Mar 16 18:24:04 so maybe it's even not glibc changes %) Mar 16 18:26:05 psokolovsky, btw, seems that glibc-package.bbclass should be in the same dir as glibc, so I have to overide it only with whole package? Mar 16 18:26:30 slapin_nb, don't understand Mar 16 18:27:39 psokolovsky, as I wanted to test your patch I put it to mypkg/packages/glibc/glibc-package.bbclass. and either it is not taken or your patch doesn't work... Mar 16 18:28:15 slapin_nb, besides my patch, you also should set config vars Mar 16 18:28:27 psokolovsky, they are set. Mar 16 18:28:42 slapin_nb, what is mypkg/ ? Mar 16 18:28:52 psokolovsky, anyway - + '[' -f /home/slapin/OE/tmp/cross/arm-angstrom-linux-gnueabi/lib/libgcc_s.so /home/slapin/OE/tmp/cross/arm-angstrom-linux-gnueabi/lib/libgcc_s.so.1 ']' Mar 16 18:28:52 /home/slapin/OE/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.5-r5/temp/run.do_prep_locale_tree.7524: line 694: [: /home/slapin/OE/tmp/cross/arm-angstrom-linux-gnueabi/lib/libgcc_s.so: binary operator expected Mar 16 18:28:52 + install -m 0755 /home/slapin/OE/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.5-r5/image/usr/bin/localedef /home/slapin/OE/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.5-r5/locale-tree/bin Mar 16 18:28:52 is it ok? Mar 16 18:29:04 psokolovsky, mypkg is topmost BBPATH Mar 16 18:30:41 slapin_nb, it should work Mar 16 18:30:57 slapin_nb, don't spam channel here, or be kicked ;-) Mar 16 18:31:16 psokolovsky, sorry, will use pastebin Mar 16 18:35:34 psokolovsky, would you push that patch so to have only 1 problem to cope with? Mar 16 18:35:47 slapin_nb, which patch? Mar 16 18:36:04 locale generation limitation. Mar 16 18:36:36 of course, no! I don't wnat the current issues to be blamed on me. so, first they resolve, then, maybe Mar 16 18:37:13 psokolovsky, too bad Mar 16 18:37:39 psokolovsky, everybody avoids responsibility... Mar 16 18:38:08 slapin_nb, opposite - we do responsible engineering here Mar 16 18:38:21 slapin_nb: no one wants to break other people's code. Mar 16 18:38:42 hi psokolovsky :) Mar 16 18:38:49 hi Mar 16 18:40:10 mmp, I did not mean that anything will break. Mar 16 18:41:02 slapin_nb: yes, I just meant that any patch can possibly break other things... Mar 16 18:41:37 mmp, risk to break something always exists. Mar 16 18:42:24 yes... Mar 16 18:42:53 the idea is very simple: if some code is broken, don't commit anything to that code except fixes Mar 16 18:44:09 psokolovsky, I think that code is not broken. Mar 16 18:44:22 psokolovsky, it is somewhere else. Mar 16 18:44:34 psokolovsky: Calm down about the current issues. No need for any further fingerpointing. .dev will break at times and everybody is aware of that. The question is how you deal with situations like this and possibly how you try and minimize their occurrence that matters. Mar 16 18:45:21 Laibsch, you probably mix me with someone else. I fix issues, not fingerpoint. Mar 16 18:45:32 no I don't Mar 16 18:46:54 Laibsch, in the meantime, you reverts fixed the issue. So, I suggest you commit safe -dbg patch, and look carefully at the unif. refactors. Mar 16 18:49:17 Is there a general rule about positioning of CONFFILES lines in recipes? I've seen almost random cases of (for example) "*** Error: CONTROL/conffiles mentions conffile /etc/openpbx.org/musiconhold.conf which does not exist" depending on te ordering of statements in the recipe but I haven't been able to identify a pattern or indication of the proper ordering Mar 16 18:49:45 Laibsch, and - be prepared to take responsibility for your changes (which are much appreciated!), but don't take any phrase as point at you ;-). you jumped right at the talk of me and slapin_nb on rather different issue. Mar 16 18:49:50 03nail 07org.oe.dev * r6a7bea85... 10/ (1 packages/fis/fis_1.0.bb): fis: Make one shared and one static version, split into two packages Mar 16 18:50:14 hillct, we have script in contrib/ which knows how to format recipes Mar 16 18:50:22 Is there a "standard" way to install a file that isn't supposed to be in the rootfs, but is needed in an image? Mar 16 18:50:52 psokolovsky: I've been warned that script orders things in broken ways Mar 16 18:51:50 psokolovsky: I guess I'll run it through that script just for kicks but the last time I used it when building a different package, it causes more headaches than it was worth Mar 16 18:51:58 psokolovsky: I was taking into account the course of the day. Mar 16 18:52:43 hillct, I'm talking about oe-stylize.py . who told you it's broken? where's bug#? Mar 16 18:53:20 hillct: I consult the style guide in the wiki Mar 16 18:53:32 The script cannot know more and be more current than that doc Mar 16 18:54:08 I also felt uneasy about blindly accepting the output of that script. Mar 16 18:54:22 psokolovsky: I was told by NAiL and others not to rely upon the results of oe-stylize.py ad it's ordering was not reliable, and my experience with it causes me to agree Mar 16 18:54:25 Laibsch, again, suggestion: you should be prepared that someone will work alongside you for fixing your breakages. you shouldn't take taht as poking at you. I finished on this matter. Mar 16 18:54:48 hillct: huh? I told you what? I don't think that was me ;) Mar 16 18:55:02 hillct, you guys should brought that issue on the ML, because it was RFCed as official way to reformat recipes Mar 16 18:55:41 pretty sure it was, but in any case, I was told only to use it as a guide and not gospel Mar 16 18:55:59 I'm not sure I can accurately articulate the bounds of the problem Mar 16 18:56:19 just seems that it reorders some recipes in ways that cause builds to fail Mar 16 18:56:31 cyrilRomain|zZz, see above about oe-stylize.py Mar 16 18:58:03 I can't find the earlier discussion in my IRC logs but I don't log OE Mar 16 18:58:14 ~logs Mar 16 18:58:16 rumour has it, logs is apt/ibot/infobot/jbot/purl all log daily to http://ibot.rikers.org// where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/.html.gz Mar 16 19:00:15 the reverted patches solve my problems too Mar 16 19:05:28 hi Mar 16 19:15:58 hmm Mar 16 19:16:10 Ifaistos: http://www.openembedded.org/wiki/AutoBug but I need to tweak it for bugs.openembedded.org Mar 16 19:16:29 koen|recife: ping? Mar 16 19:16:32 just went looking for oe-stylize.py and I can't find it. Where does it exist within the directory tree? Mar 16 19:16:39 I've forgotten Mar 16 19:17:09 hillct, org.openembedded.dev/contrib/oe-stylize.py Mar 16 19:17:42 yeah, I thougt it was in contrib... wonder if my structure is partial Mar 16 19:17:45 thanks Mar 16 19:21:38 * Laibsch tells hillct about the wonders of "locate" Mar 16 19:22:37 Laibsch: the wonders of 'find' were not useful suprisingly... Mar 16 19:23:38 isn't this valid? Mar 16 19:23:40 RECOMMENDS_${PN}-ogi = "perl perl-module-strict" Mar 16 19:24:48 oe-stylize says it isn't. That doesn't seem right... Mar 16 19:26:41 hillct: RRECOMMENDS? Mar 16 19:27:05 oh Mar 16 19:27:08 thanks Mar 16 19:27:18 duh Mar 16 19:32:30 apaulsen, slapin_nb: Can you tell me the distro/machine combination you were using when you ran into problems building glibc? I cannot reproduce your problem yet (and thus cannot fix it) Mar 16 19:33:50 Laibsch: I'm using a custom distro based on generic and my machine is a compulab-pxa270 Mar 16 19:34:01 Was the compilation failing half-way through or right at the start? I guess if version 2.3.5 was chosen when 2.5 should have been, that would qualify as failing at the start in the sense that I am looking at. Mar 16 19:34:48 Laibsch, DISTRO=Angstrom MACHINE=qemuarm Mar 16 19:35:07 slapin_nb: Nice. Saves me from setting up a separate distro. Mar 16 19:35:24 I will try to reproduce for angstrom/qemuarm Mar 16 19:35:44 I'd be happy to run tests too Mar 16 19:37:07 Laibsch, at current tree state I was unable to reproduce. Mar 16 19:37:25 apaulsen: Thank you for the offer. I will come back to it if I still have problems to reproduce. Mar 16 19:38:28 Laibsch, http://builds.hackndev.com/builds/angstrom/daily-1173988621/build-qemuarm.log.gz it was previous times build. Mar 16 19:41:14 morning all Mar 16 19:44:07 psokolovsky: thanks for your ping. Does it mean oe-stylize.py need update ? Mar 16 19:44:54 cyrilRomain, don't ask me, ask folks who said they have issues with it. otherwise, it's just FYI Mar 16 19:44:56 it is quite clear oe-stylize.py cannot handle correctly all bitbake files, especially since its lake of semantic analysis Mar 16 19:45:32 hillct, NAiL : can you please tell me (or giving me examples) where oe-stylize.py fails ? Mar 16 19:47:22 cyrilRomain: it was several weeks ago that I encountered this Mar 16 19:48:28 I forget what what package I was working on, but I ran the recipe through the script and it reordered the statements in a way that caused the build to fail. I'll poke around and see if I can find the package in question Mar 16 19:48:53 hillct: thanks, it would help to see what's wrong Mar 16 19:49:04 probably either ntpclient or logrotate Mar 16 19:49:18 those were the two I was working on at the time Mar 16 19:49:48 * cyrilRomain is looking at those recipes Mar 16 19:51:19 it moved some variable to the bottom of the recipe as I recall Mar 16 19:51:30 hillct: right Mar 16 19:51:42 hillct: this is CFLAGS_prepend and LDFLAGS_prepend Mar 16 19:52:04 the script indeed but unknown variable at the end of the script Mar 16 19:52:31 oh. those are unknown? Mar 16 19:52:40 docs say they're fine Mar 16 19:52:41 hillct: although I can fix it to take AKNOWNVARIABLE_prepend or _append into account Mar 16 19:53:19 the script can not take for example variable dependence into account, like a = "t", b = "${a} g" Mar 16 19:53:21 those were my first packages. They wre reviewed by others but it's possibly i just screwed them up Mar 16 19:53:36 yeah Mar 16 19:53:52 hillct: if the docs say those variable are ok, they just have to be included into oe-stylize.py Mar 16 19:54:10 it also reports that setting PN manually is not allowed Mar 16 19:54:36 and indeed, it may not be. I don't know for sure Mar 16 19:54:55 hillct: oe-stylize.py was written month ago, and need update apparently Mar 16 19:55:10 ah Mar 16 19:55:45 might it be able to take advantage of the main recipe parser to determine allowed elements? Mar 16 19:56:36 * hillct 's python-foo is not strong so I can't evaluate the code structure in a meaningful way Mar 16 19:57:10 hillct: no, unfortunately it is quite separate from the parser :-( Mar 16 19:57:53 hi :) Mar 16 19:58:05 hillct: when I wrote it, it seemed to me not really easy and possible to include into the bitbake code. Mar 16 19:59:36 now that I better understand oe and bitbake, I might take a deeper look into bitbake code (with RP and zecke) help to see how it could be done within bitbake Mar 16 19:59:47 cyrilRomain: and the current parser is quite horrible Mar 16 19:59:55 zecke: shame :-/ Mar 16 20:00:06 cyrilRomain: you can extract the regexp's it is using to find keywords and such :) Mar 16 20:00:30 the good thing with stupid things inside BitBake is I can say "It wasn't me" :) Mar 16 20:01:15 always nice to be able to say things like that Mar 16 20:01:22 the good thing with stupid things inside oe-stylize.py is I can say ".. " heu no I can't ;) Mar 16 20:01:31 cyrilRomain: you can use the parser from trunk/bitbake-parser Mar 16 20:01:54 zecke: yep I'll try to get some time to take a deeper look into it Mar 16 20:01:59 cyrilRomain: it requires prior compilation so it is not as easy to use as oe-sanitize Mar 16 20:02:28 zecke: well, c, c++ ? Mar 16 20:03:03 zecke: ok, not a big problem Mar 16 20:04:01 cyrilRomain: it is wrapped into python, we might add better error reporting to it Mar 16 20:04:19 cyrilRomain: but still it is lex/lemon based parser and needs ahead of time compilation Mar 16 20:04:52 zecke: do you think I should try to use a better parser than lex ? like cdiggins YADP ? Mar 16 20:05:24 Yet Another Decent Parser, written in c++ metaprogramming Mar 16 20:05:53 cyrilRomain: no :) Mar 16 20:06:03 well, the code is really tiny (at least enough so that I can understand how to use it ;) Mar 16 20:06:19 zecke: ok, I'll try to just use lex then :) Mar 16 20:06:23 cyrilRomain: I say use bitbake-parser but it will require compilation :) Mar 16 20:06:39 zecke: ok :) Mar 16 20:06:58 cyrilRomain: one thing stylize should definately check for is duplication of file listings. I've screwed that up several times where I have a specific file listed multiple times, perhaps in subpackages, in FILES variables Mar 16 20:07:35 and if stylize doesn't catch it, it only throws an error durring the do_package phase Mar 16 20:08:17 and in my case at least, the package I'm working on takes 20 minutes to compile each time so getting errors in the packaging phase is frustrating as hell Mar 16 20:08:23 like the recent diet-x11 makekeys.diff patch whih is already applied on libx11 ? yep Mar 16 20:09:37 hillct: I see. But to do a better oe-stylize.py, we have to integrate it into bitbake, so it need really more work and understanding than the time I wrote oe-stylize.py Mar 16 20:10:05 Is there any gpe frontend in oe for mplayer? Mar 16 20:10:19 zecke: so I guess as soon as I have a question, I'll bother you ? ;) Mar 16 20:10:31 sure Mar 16 20:15:09 sirfred: I don't think so Mar 16 20:15:38 Any one for gtk that could be ported? Mar 16 20:20:21 you could look through these: http://www.mplayerhq.hu/design7/projects.html Mar 16 20:21:18 hello Mar 16 20:21:46 hi Mar 16 20:21:50 mickey_away: ping Mar 16 20:22:08 i am trying to build oe on os x tiger Mar 16 20:22:21 but, I am have a bit of a problem. Mar 16 20:22:27 sirfred: gtk mplayer frontend or gtk application which use mplayer? Mar 16 20:22:36 is there someone who can help me? Mar 16 20:23:14 zecke: pong Mar 16 20:23:22 hrw: Hmm, what's the difference? Mar 16 20:24:09 hrw: I want something embedding the video window of mplayer, I suppose that launching mplayer as a different process. Mar 16 20:24:13 zecke, are you available? Mar 16 20:24:19 sirfred: 1st need change to mplayer or mplayer-gtk. 2nd does not need change to mplayer or small change Mar 16 20:24:25 koen|recife: should Angstrom autobuild failures go to Build component? Mar 16 20:24:45 dave_c: I should study for an exam, if you can help me to avoid this go ahead :) Mar 16 20:25:09 hrw: Hm, but AFAIR gmplayer has a control windows and a video window. I supposed that was not suitable for gpe Mar 16 20:25:13 hrw: hi. Do you have any experience with the Sharp bootloader? I am trying to figure out which parameters we use and which we ignore (like cmdline) Mar 16 20:25:30 hvontres|poodle: we ignore cmdline totally Mar 16 20:25:37 hrw: I'm looking at http://dekorte.homeip.net/download/gnome-mplayer, that was my idea, more or less. Mar 16 20:25:43 sirfred: that depends on theme Mar 16 20:26:00 hrw: So, perhaps that's the best solution. Mar 16 20:26:09 hrw: what about the memory configureaton? I am looking at adapting arm-kernel-shim for sharp machines Mar 16 20:26:10 sirfred: mplayer-gtk then Mar 16 20:26:40 hvontres|poodle: we check it in kernel. if c7x0 and pxa250 then 32MB ram. 64MB for rest Mar 16 20:26:40 hrw: But I suppose that there should be problems or it should be available as a package already, didn't it? Mar 16 20:26:57 sirfred: no one wanted mplayer with gtk yet Mar 16 20:27:15 I'm the first one, so. Mar 16 20:27:29 I want it. It's the only thing I'm missing. Mar 16 20:27:53 I have games with gpe-snes, the gpe PIM applications, I only need to be able to watch videos Mar 16 20:28:17 sirfred: then create mplayer-gtk_VER.bb which will include mplayer_VER.bb and will add few DEPENDS and few EXTRA_OECONF probably Mar 16 20:28:42 hrw: ok. I'm going to try Mar 16 20:29:08 sirfred: probably mplayer.inc will need to be created which will keep EXTRA_OECONF common so nogtk will add --without-gtk and -get will add --with-gtk Mar 16 20:29:10 Once the non gtk version compilation I've launched finishes, this machine is getting old. Mar 16 20:29:31 hrw: thanks Mar 16 20:29:44 well Mar 16 20:30:04 what kind of an exam if you don't mind me asking? Mar 16 20:30:44 dave_c: Operating System class, but go ahead :) Mar 16 20:31:02 zecke: probably Mar 16 20:31:02 this is the error I get: Mar 16 20:31:11 ERROR: Nothing provides dependency quilt-native Mar 16 20:31:24 dave_c: on which configuration? Mar 16 20:31:53 i don't know how to answer that correctly, Mar 16 20:32:09 i am going to assume its what I want to build ultimately Mar 16 20:32:17 dave_c: what are you trying to build, why do you think I can answer it? Mar 16 20:32:42 i read your wiki entry on Mar 16 20:32:49 openembedded website Mar 16 20:32:55 I am using os x Mar 16 20:33:04 to build oe Mar 16 20:33:14 for to purpose of building a dreambox image Mar 16 20:34:11 i don't understand why quilt-native does not build Mar 16 20:34:20 does it need diffstat Mar 16 20:34:23 ? Mar 16 20:34:36 dave_c: first of all your basic configurations looks wrong, then building on OSX is not really supported. You need to fix some issues a long the way. Mar 16 20:34:51 ok Mar 16 20:34:59 dave_c: bitbake -vvv or -DDD and see which dependencies quilt-native Mar 16 20:35:09 dave_c: alternatively use Parallel or VMware :) Mar 16 20:36:07 it says nothing provides dependency for quilt-native Mar 16 20:36:32 dave_c: how much packages it parsed? Mar 16 20:36:49 if (!dave_c->isThinking() ) dave_c->enableThinking() :) Mar 16 20:36:52 0 parsed Mar 16 20:37:03 dave_c: this looks wrong right? Mar 16 20:37:09 oh yeah Mar 16 20:37:36 dave_c: e.g. fix BBFILES? set BBPATH? Mar 16 20:38:00 BBFILES are wrong Mar 16 20:41:03 dave_c: have you read the http://www.openembedded.org/wiki/GettingStarted page ? (it also includes required software) Mar 16 20:41:03 the opendreambox make should set those two Mar 16 20:41:18 yes I read it Mar 16 20:41:35 the only piece of software I don't have is diffstat Mar 16 20:41:45 can't download it Mar 16 20:43:30 hrw: Just curiosity. Any new about that toshiba? Mar 16 20:43:37 nope Mar 16 20:43:54 sirfred: I lost faith. will mail that guy in monday Mar 16 20:44:03 hrw: Perhaps he changed his mind Mar 16 20:44:14 hrw: Did you pay it already? Mar 16 20:44:25 I should not have 0 parsed files , correct? Mar 16 20:44:39 sirfred: I proposed pay-on-arrival and he accepted it. Mar 16 20:44:42 dave_c: yes Mar 16 20:44:51 dave_c: you have BBFILES or BBPATH wrong Mar 16 20:45:09 hrw: Well, at least you didn't lost money. Mar 16 20:45:10 ok, let me try and sort those two out Mar 16 20:45:15 sirfred: yes. Mar 16 20:45:21 NOTE: package gmplayer-0.0+1.0rc1: completed Mar 16 20:45:29 Let's see how does it look Mar 16 20:47:34 It needs a skin Mar 16 20:49:23 * Ifaistos wonders if people you write bb files actually use them... Mar 16 20:49:37 what fills in the "Depends:" section in ipk/Packages? Ever since I tried building Kaffe yesterday, classpath-gtk has an incorrect dependency for kaffe-gtk and I can't seem to get rid of it Mar 16 20:50:09 hrw: I'm looking at the official skins, and it seems all them have two windows: controls and video. Do you know of any skin with one window? Mar 16 20:50:47 no Mar 16 20:50:58 I do not remember when last time I used it Mar 16 20:55:25 apaulsen, RDEPENDS fill that, and that can be set explicitly or implicitly Mar 16 20:56:10 I grepped everywhere for references to kaffe and there aren't any. I wiped the bb cache several times today as well Mar 16 20:57:04 apaulsen, tried to wipe Packages file itself? that's very helpful oftentimes ;-) Mar 16 20:57:31 I did earlier, but that was before I wiped the cache, I'll try doing both Mar 16 21:00:39 cu Mar 16 21:04:33 rwhitby NAiL : I think i figured out why NAiL's (and mine) gcc-cross were showing r11 instead of r12 that gcc (not cross) is. Mar 16 21:04:53 rwhitby NAiL : The PR number in gcc-cross was not increased after the changes Mar 16 21:07:37 Ifaistos: http://www.openembedded.org/autobug and wait for my commit :) Mar 16 21:08:04 zecke : i have bookmarked it :) Mar 16 21:08:52 * Ifaistos hopes people won't think i am spamming them :) Mar 16 21:09:17 Ifaistos: create a new bug report account :) Mar 16 21:09:26 Ifaistos: icognito-spamer@hotmail.com Mar 16 21:10:18 viagra-bugs@free-mail.com Mar 16 21:10:19 I removed all the Packages* files and the bb cache and I'm still getting an error that classpath-gtk depends on kaffe-gtk. I have no idea where this dependency is, other than I tried to build kaffe-gtk yesterday. Mar 16 21:10:43 apaulsen: bitbake -vvv ? Mar 16 21:11:01 good idea Mar 16 21:11:29 btw has anyone *actually* used mysql build from OE ? Mar 16 21:11:40 not just build the package Mar 16 21:11:46 Ifaistos: I'm sure schurig has used it... Mar 16 21:13:20 apaulsen, maybe it indeed depends, after all? Mar 16 21:13:45 there is no reference to kaffe in the classpath directory Mar 16 21:13:50 * Ifaistos wonders how -> ../../../tmp/libc/staging/x86_64-linux/bin/powerpc-linux/mysql_config Mar 16 21:14:14 and this problem only appeared after I tried to build kaffe, I've been including jamvm and classpath-gtk in my images for a month now Mar 16 21:14:19 apaulsen, ?? in which directory and how that relates to package dependencies? Mar 16 21:14:49 sorry, in the packages/classpath/ directory Mar 16 21:15:05 apaulsen, easy way that could happen is that classpath detected kaffe, and build something which depends on kaffe's .so's Mar 16 21:15:51 ok, I've rebuilt classpath about five times already... Mar 16 21:16:49 I have a package which I'm building but within /contrib in this tarball there's a perl module which could use the CPAN build process. Is there a way I can inherit CPAN.class for purposes of building a sub-package? how do I specify the sub-package root (where I'd want CPAN.class to operate on contribb/openpbx-perl) where Makefile.PL can be found? Mar 16 21:16:52 apaulsen, rebuilding and magic dances won't help ;-) try to figure out where RDEPENDS came from. .so dependencies is common case Mar 16 21:17:45 could someone take a look at that -> http://rafb.net/p/Bm12zE60.html and tell me why the files don't get package correctly Mar 16 21:18:14 i am trying to split php to a number of packages main php, pear, mysql etc Mar 16 21:18:31 main php regardless of what gets all the files Mar 16 21:18:34 Ifaistos: this is an abuse of comment Mar 16 21:18:49 Ifaistos: and the order of PACKAGES matter Mar 16 21:19:29 zecke : you mean that -> # ${PN}-mysql \ Mar 16 21:19:47 Ifaistos: yes (but that is not the error) Mar 16 21:19:59 Ifaistos: but you might want to move the order in PACKAGES Mar 16 21:20:21 Ifaistos: FILES_${PN} +="/usr/bin" is pretty bogus as ${bindir} will get packaged :) Mar 16 21:20:38 Ifaistos: so you might want to change the order of PACKAGES Mar 16 21:21:11 zecke : so it should be from "specific" (i.e pear, mysql) to "generic" (main php) and not the other way around Mar 16 21:21:39 zecke : i thought what mattered was the order of FILES Mar 16 21:21:46 Ifaistos: yes, it will start with PACKAGES[0] Mar 16 21:21:54 Ifaistos: and then package everything in FILES Mar 16 21:22:06 Ifaistos: then it goes to PACKAGES[1] but at that point /usr/bin/* is empty :) Mar 16 21:22:11 Ifaistos: php is split in thus way in the freebsd ports system Mar 16 21:22:41 zecke : got it. makes sense Mar 16 21:23:39 mn__ : Well it seems the only way to get php and pear without having two package files Mar 16 21:24:25 mn__ : or stuff everything into php and make it huge ipkg Mar 16 21:26:41 Ifaistos: i thought you wanted to have separate packages for e.g. pear php-mysql php-base and so on? Mar 16 21:27:02 mn__ : yes, i meant 2 .bb files Mar 16 21:27:43 mn__ : because the way it's setup now in OE, you can get either php or php with pear support Mar 16 21:28:09 mn__ : and no mysql or gd etc Mar 16 21:29:25 it also seems it needs a buildpear-packages.bbclass in order to be able to create pear based packages Mar 16 21:29:41 any takers ? :) Mar 16 21:29:47 Ifaistos: what happens if you move ${PN} to the end of the packages list? Mar 16 21:30:58 hvontres|poodle: seems things are getting packaged correctly :) Mar 16 21:32:16 psokolovsky: thanks, there were a bunch of references to kaffe-gtk left in build/tmp/staging/ even after I -cclean'd it. Asside: I didn't build classpath-gtk five times because I expected different results, I was testing different versions :) Mar 16 21:33:44 apaulsen, yeah, staging is never cleaned at all (AFAIK?). bottom line: please submit this as bug, if you're not going to fix it Mar 16 21:34:25 psokolovsky: who do I submit this to? Mar 16 21:34:44 apaulsen, http://bugs.openembedded.org Mar 16 21:34:50 ok Mar 16 21:36:01 first task accomplished ;) now it's mysql's turn Mar 16 21:37:03 mn__ : do you have a list of the php packages in freebsd ? Mar 16 21:38:53 Ifaistos: not really it?s split into several categories - http://www.freebsd.org/ports/(devel|lang|www).html and few more Mar 16 21:39:38 mn__ : thanks.i'll check debian packages also to see how they split php Mar 16 21:39:46 Ifaistos: the used Makefiles might be useful for finding depenedecies Mar 16 21:40:57 mn__ : you are right. the install part can be usefull Mar 16 21:42:25 Ifaistos: you might find these usefull: Mar 16 21:42:36 Ifaistos: http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/ports-using.html Mar 16 21:42:49 Ifaistos:http://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/index.html Mar 16 21:48:05 zecke: Is there a way to inherit one class (autotools) for building of a primary package then inherit another class(cpan) for building a sub-package housed in contrib within the main tarball source for a build? Seems to me the main issue is defining the sub-package for which a given inherited class should act. Or should I simply use a separate recipe Mar 16 21:49:17 hillct: we currently can't do this, use two recipes Mar 16 21:49:27 K Mar 16 21:49:30 thanks Mar 16 21:52:05 what are the first steps to get linux on a netbook pro? Mar 16 21:52:56 I'm trying to compile mplayer and get this error: "dsputil.c:3575: internal compiler error: " Mar 16 21:53:23 Do I need to use a different compiler to fix this problem? Mar 16 21:54:17 mn__, http://linux-7110.sourceforge.net/ Mar 16 21:55:01 darmou: this project for the netbook - not for the nbpro Mar 16 21:55:56 darmou: i?d try to compile that file w/o optimization Mar 16 21:56:55 darmou: fix the compiler and attach the patch to gcc's bugtracker Mar 16 21:57:09 heh Mar 16 21:57:22 this is Free Software :) Mar 16 21:58:38 zecke: Does this make any sense to you -> ../../../tmp/libc/staging/x86_64-linux/bin/powerpc-linux/mysql_config Mar 16 21:59:33 zecke: the native mysql_config seems ok -> ../../../tmp/libc/staging/x86_64-linux/bin/mysql_config Mar 16 21:59:35 Ifaistos: yes, you have multimachine? Mar 16 21:59:42 yes Mar 16 22:00:25 Ifaistos: what do you think is wrong? One installs to STAGING_BINDIR and the other to STAGING_BINDIR_NATIVE Mar 16 22:01:55 zecke : The second one is at STAGING_BINDIR_NATIVE which is correct Mar 16 22:03:10 zecke : The first one i believe it should have been at ../../../tmp/libc/staging/powerpc-linux/mysql_config to be STAGING_BINDIR Mar 16 22:03:16 Ifaistos: what do you think is wrong? Mar 16 22:03:36 :) Mar 16 22:03:45 the other $BUILD_SYS inside the dir name? Mar 16 22:04:01 yes Mar 16 22:04:15 it does not make sense Mar 16 22:04:30 (at least to me) Mar 16 22:05:31 hmmm unless mysql_config gets executed both for native and cross at the build system Mar 16 22:06:38 03freyther 07org.oe.dev * r2a553ef0... 10/ (1 contrib/qa/bugzilla.py): Mar 16 22:06:38 contrib/qa/bugzilla.py: Extract all bug reports, even the ones from bugs.openembedded.org Mar 16 22:06:38 Abuse the class attribute even further and check for severities. We are Mar 16 22:06:38 able to find all 2006 bugs of OpenEmbedded and all 282 bugs of OpenMoko Mar 16 22:06:42 03freyther 07org.oe.dev * r8c1a9173... 10/ (1 classes/seppuku.bbclass): Mar 16 22:06:42 classes/seppuku.bbclass: Make parsing of OpenEmbedded's bugtracker work Mar 16 22:06:43 OE's bugtracker has different classes for the buglist. I try hard Mar 16 22:06:45 to parse them all. All severities should be handled. Mar 16 22:06:49 03freyther 07org.oe.dev * r084b301b... 10/ (1 classes/seppuku.bbclass): Mar 16 22:06:51 classes/seppuku.bbclass: Make it work with OpenEmbedded bugtracker Mar 16 22:06:53 Do not error out on python2.5 when splitting the url. We had Mar 16 22:06:55 an empty key and a none value. Mar 16 22:06:57 If there was no output file, do not error. This check is the same Mar 16 22:06:59 as in tinderclient.bbclass Mar 16 22:07:01 03nail 07org.oe.dev * r48c496f0... 10/ (1 packages/fis/fis_1.0.bb): fis: Forgot boost dependency Mar 16 22:09:46 zecke: heh... one more bug and we match the year :) Mar 16 22:10:16 hvontres|poodle: hehe, and then we close the bug tracker until next year Mar 16 22:10:52 hmmm, only one bug per year.. that should help us keep up :) Mar 16 22:11:30 hvontres|poodle: we can apply DDR/Trolltech thinking: If there are no bugreports we must be perfect Mar 16 22:12:08 zecke: DDR? I thought they went out of buisness in 1989 :) Mar 16 22:12:20 or nobody is actually using our stuff.... Mar 16 22:12:57 hvontres|poodle: some people refuse to believe that :) Mar 16 22:13:12 Ifaistos: no, if you apply this thinking this thought will never come up Mar 16 22:14:52 zecke: you have got to admit that they were prety good at building "Firewalls", even if they pointed them the wrong way :) Mar 16 22:16:44 hvontres|poodle: besides not having large scale fires, so a classic example of bad overengineering Mar 16 22:20:37 good night Mar 16 22:20:38 zecke: so, how's the studying coming along ? ;) Mar 16 22:20:50 slowly, but I'm progressing Mar 16 22:20:55 I'm mostly bored :) Mar 16 22:26:10 mn__, how do you set the compile options in oe, is it in the bb file? Mar 16 22:28:48 zecke: heh..well, 30 more minutes and it will be weekend here :) good luck. Mar 16 22:29:34 It's already saturday morning at 9.30am here Mar 16 22:30:42 darmou: are you down in OZ? Mar 16 22:31:58 yep Melbourne Mar 16 22:32:21 where are you from hvontres|poodle Mar 16 22:33:23 darmou: I'm in SoCal... right now its 15:30 on Friday and I leave work at 4 Mar 16 22:34:48 hi bluelightning Mar 16 22:34:49 darmou: i have absolutely no idea about all this - i?d copy/run the compilation from stdout and would remove any -01,2 options ( Mar 16 22:35:55 mn__, thanks for your help, hopefully there will be an mplayer update that will compile properly Mar 16 22:36:30 wtf? Mar 16 22:36:43 repvik@apollo:~/slug/openembedded/packages/coreutils$ cat /home/repvik/slug/foonas/tmp/work/lrzsz-0.12.20-r3/temp/log.do_install.4658 Mar 16 22:36:46 mkdir: cannot create directory `/usr/lib/ipkg': Permission denied Mar 16 22:37:00 When is lrzsz.bb touching /usr/lib/ipkg? Mar 16 22:44:22 hi darmou Mar 16 22:44:30 NAiL: every recipe does that Mar 16 22:44:56 koen|recife: They should never touch /usr on the host ;-) Mar 16 22:45:15 anyway, I misplaced update-alternatives, that's what caused it :-P Mar 16 22:46:23 NAiL: everytime you invoke ipkg it wants to create /usr/lib/ipkg Mar 16 22:46:49 yeah Mar 16 22:49:45 03nail 07org.oe.dev * r1057a7dd... 10/ (1 packages/lrzsz/lrzsz_0.12.20.bb): lrzsz: Use update-alternatives to avoid conflicts with busybox. Closes bug 1970 Mar 16 22:53:36 CoreDump|cold: heater busted? Mar 16 22:54:18 not quite. I'm sick :\ Mar 16 22:58:13 re Mar 16 22:59:13 hey florian Mar 16 22:59:54 ~lart gcc-cross-sdk Mar 16 22:59:54 * ibot whips out a sword and chops gcc-cross-sdk in half Mar 16 23:01:25 Laibsch, rehi, any news? Mar 16 23:06:38 ibot: seen pb_? Mar 16 23:07:13 pb_ was last seen on IRC in channel #handhelds.org, 16d 4h 21m 49s ago, saying: 'hi zap'. Mar 16 23:14:44 night Mar 16 23:34:28 Is there any reason I should *not* add a SRC_URI in foonas-image.bb, to fetch a couple of binaries that need to be in the firmware-image (not rootfs)? Mar 16 23:34:46 * NAiL noted that none of the existing *-image.bb's have a SRC_URI Mar 16 23:35:51 nothing I'm aware of Mar 16 23:36:07 ok, thanks Mar 16 23:39:21 03nail 07org.oe.dev * ree278dc3... 10/ (1 packages/fis/fis_1.0.bb): fis: Remove fis-static, update shared fis Mar 16 23:42:02 03nail 07org.oe.dev * r0113d92c... 10/ (5 files in 4 dirs): foonas: Clean up distro/machines/image Mar 16 23:58:17 Hi all Mar 16 23:58:52 hi GStepper Mar 16 23:59:26 slapin_nb: I was watching a movie. Mar 16 23:59:51 I was able to reproduce your error so I should be able to fix things by adding stuff one by one until something breaks. Mar 17 00:08:03 Strange. Now I've lost the ability to reproduce. Mar 17 00:08:06 :-/ Mar 17 00:10:25 Ouch. Mar 17 00:11:11 Laibsch, it was reproducible building complete image (actually, gtk+), and not repr. building glibc only Mar 17 00:20:11 Laibsch: I hear that operation is reversible (losing the ability to reproduce) Mar 17 00:20:57 rwhitby: hehe. Let's ask my girlfriend on this topic ;-) Mar 17 00:21:03 psokolovsky: Do you ever sleep? ;-) Mar 17 00:21:24 dunno ;-) Mar 17 00:22:20 psokolovsky: Building gtk+ also chooses 2.5 (I reapplied my patches locally): http://rafb.net/p/EEdVPp26.html Mar 17 00:22:46 But let's wait until glibc is being build and not glibc-intermediate Mar 17 00:24:28 Laibsch, well, I don't know, it was pulled up for some glibc-iconv-iso8859-1 or something. I also could rebuild gtk+ from scratch, in normal way. but when baking it again, glibc2.3.5 was again pulled Mar 17 00:24:35 it's pretty weird, of course Mar 17 00:25:01 03lenehan 07org.oe.documentation * r70c52dee... 10/ (1 usermanual/chapters/recipes.xml): Mar 17 00:25:01 recipes: Fill in some details on the debian library policy in the Mar 17 00:25:01 packing section. Mar 17 00:41:00 Do we have an official version to show a dependency tree of packages now? Mar 17 00:41:36 I know about grepping the output of bitbake -vvv Mar 17 00:42:05 I'm working on a bb recipe for a package and I have it working pretty well but the .pc files end up having /usr/include in them.. If I add --includedir=${STAGING_INCDIR} to EXTRA_OECONF the .pc files are correct but now all of the dev files are put in WORKDIR/image/STAGING_INCDIR/ instead of just WORKDIR/image messing up the packaging of -dev Mar 17 00:42:34 dev files being the header files.. Mar 17 00:47:14 svolpe_gerrath: looks like either the build process for the pc files or to install the headers is wrong. Mar 17 01:08:37 03coredump2 07org.oe.dev * r8fbc5385... 10/ (6 files in 4 dirs): gspcav1: Add gspcav1 webcam driver Mar 17 01:31:35 03nail 07org.oe.dev * r77568721... 10/ (12 files in 5 dirs): Mar 17 01:31:35 netatalk: Add netatalk package, provided by Mar 17 01:31:35 timtimred. Thanks ;-) Mar 17 01:47:56 Could anyone explain to me how libtool managed to dig up this toolchain? Mar 17 01:47:57 /home/repvik/slug/foonas/tmp/staging/x86_64-linux/bin/powerpc-linux-libtool: line 380: /home/repvik/slug/optware/ts101/toolchain/powerpc-linux/gcc-3.4.3-uclibc-0.9.27/bin/sed: No such file or directory Mar 17 01:48:31 It's looking outside TOPDIR, and finding a completely incompatible toolchain Mar 17 01:48:52 (and of course, in true libtool tradition, hardcoding paths) Mar 17 01:49:09 ~lart libtool Mar 17 01:49:09 * ibot DoSes libtool Mar 17 01:52:58 ~lart libtool some more Mar 17 01:52:59 * ibot duct-tapes libtool some more to the floor and drools on him Mar 17 01:53:16 that kinda didn't make as much sense as I hoped it would. Mar 17 02:05:11 * mwester observes that the larts are getting old and stale... **** ENDING LOGGING AT Sat Mar 17 02:59:56 2007