**** BEGIN LOGGING AT Sat May 24 02:59:56 2008 May 24 03:02:04 * * OE Bug 4300 has been created by  May 24 03:02:06 * * jpeg-6b-autobuild May 24 03:02:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4300 May 24 03:33:04 * * OE Bug 4301 has been created by  May 24 03:33:06 * * klibc-1.5-autobuild May 24 03:33:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4301 May 24 05:56:06 oe May 24 06:01:55 mithro: why May 24 06:04:46 oer May 24 06:04:49 oe May 24 06:14:55 oe May 24 07:00:12 Instant Stress Relief: http://www.therightfoot.net/mystuff/whatever/swf/bubblewrap.swf May 24 07:11:15 03khem 07org.oe.dev * r46d53145... 10/ (22 files in 7 dirs): Move eglibc to a separate package. Adios glibc. May 24 07:11:18 03khem 07org.oe.dev * r3cfaf01d... 10/ (6 files in 3 dirs): Add 2.6.25 kernel headers. May 24 07:11:23 03khem 07org.oe.dev * r1ca3d261... 10/ (4 files in 3 dirs): Do not use asm/page.h its not exported anymore. May 24 07:11:28 03khem 07org.oe.dev * r12b269ba... 10/ (1 conf/distro/include/sane-srcrevs.inc): Bump uclibc and eglibc sane revs. May 24 07:11:33 03khem 07org.oe.dev * r62a822ce... 10/ (8 files in 2 dirs): uclibc-dev should rdepend on kernel-headers-dev May 24 07:11:38 03khem 07org.oe.dev * r3889e112... 10/ (1 packages/uclibc/uclibc_svn.bb): Reduce preference level of uclibc_svn recipe. May 24 10:04:43 morning all May 24 10:04:46 hey guys, I've had a good look at the docs but can't seem to find out how to add a cpan module to my perl package (on gumstix oe), any tips? May 24 10:21:49 jcoxon, can you find an example of one in the metadata and copy it? May 24 10:22:40 i've copied a similar module .bb file May 24 10:22:49 and changed the variables appropriately May 24 10:23:22 but was wondering do i need to add it to any other files May 24 10:31:11 good morning May 24 10:31:42 Is there any sign that koen is going to fix the glibc-intermediate breakage he introduced? May 24 10:31:48 jcoxon, I'm not sure what you mean, its been a while since I built any perl modules, even on a desktop .. May 24 10:32:10 it may be fixed May 24 10:32:21 do you have an example? May 24 10:34:48 Crofton, on the oe setup there are a couple of modules that are built by default May 24 10:35:22 so i copied a .bb file and changed the contents to the module i want (Device::SerialPort) May 24 10:35:40 and? May 24 10:35:53 i was hoping that when i built perl it would therefore compile this module May 24 10:35:58 but it doesn't touch it May 24 10:36:14 ah May 24 10:36:17 so i'm now thinking i need to add the file path to a overal perl conf file May 24 10:36:32 but you can bitbake the module individually? May 24 10:36:52 with bitbake -b? May 24 10:36:54 i'll try May 24 10:37:22 just bitbake name May 24 10:40:44 it doesn't seem to work May 24 10:40:52 though the .bb file is in the perl subdir May 24 10:41:16 however it did show that i hadn't set up the libserialport-perl-native.bb file properly May 24 10:41:29 now that i've fixed that it might find it in a perl build May 24 10:45:35 jeremy_laine, ping May 24 11:25:24 hi al May 24 11:25:26 all May 24 11:26:25 Crofton: very late pong May 24 11:26:42 hey May 24 11:26:58 I've got some builds that are listed as started, but nothing gets uploaded May 24 11:27:23 do you have any ideas? May 24 11:27:51 also, on khem's build, it fails, but you have to hunt through 7 pages to find the failed package May 24 11:29:40 Crofton: you may be hitting limits on the development server ,it's quite limited May 24 11:29:49 look in your logs for "disabling stats" May 24 11:29:54 what sort of limits? May 24 11:30:02 ok May 24 11:30:24 not quite sure, but basically if a request to the build server fails, the current policy is to stop sending reports May 24 11:30:35 ok May 24 11:30:40 this is done to avoid massive slowdowns in case the stats server is done May 24 11:30:45 I really like this service :) May 24 11:30:47 on the downside it's a bit fragile.. May 24 11:30:55 any of you running bitbake concurrently and make concurrently? One of my builds seems to sometimes feel in concurrent mode. May 24 11:31:00 check out the new search for May 24 11:31:12 search form: http://oestats.jerryweb.org/search/ May 24 11:31:24 I'm posting to amethyst May 24 11:31:36 but I'll check that out May 24 11:32:07 Crofton: it's just a matter of doing an "svn up" on amethyst then :) May 24 11:32:11 jeremy_laine: nice. further ideas are the bitbake version and revision :-) May 24 11:32:12 yeah May 24 11:32:30 and have drop down lists for machine and distro :) May 24 11:32:35 and build arch ... May 24 11:32:46 don't feel overwhelmed :) May 24 11:32:52 Crofton: sure, that's next but down we basically have a search engine May 24 11:33:09 also "failed packages in last N days" May 24 11:33:10 and if you go to http://oestats.jerryweb.org/builders/oebuilder-waffle/ May 24 11:33:22 it's the same search form, just pre-set to the builder May 24 11:33:44 it gives you the comfort of a static url for your builds May 24 11:33:51 the really nice thing is we no longer need to pastbin failures May 24 11:34:13 Crofton: did you see the integration with seppuku? May 24 11:34:26 I did, I'm not using it yet May 24 11:34:36 since I not doing "real builds" May 24 11:34:40 just testing stuff May 24 11:34:52 I do not want to add bugs for stuff I am working on May 24 11:36:50 I'm doing a lot of bitbake dsplink atm May 24 11:58:04 * * OE Bug 4302 has been created by comte_fabien(AT)yahoo.fr May 24 11:58:06 * * php build error May 24 11:58:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4302 May 24 12:01:04 Crofton: ok have implemented multiple choice for distro / machine / build arch May 24 12:01:19 Crofton: see http://oestats.jerryweb.org/search/ May 24 12:02:53 looks great May 24 12:03:03 makes it much easier this way May 24 12:03:20 Thanks! May 24 12:05:52 hmm, dsplink built without errors, that's a firtst May 24 12:06:49 Crofton: dsplink is an alternative to what TI uses, or a clone, or... ? May 24 12:07:21 is what TI uses May 24 12:08:06 ok thanks. May 24 12:08:32 We need something that works May 24 12:08:48 We can fight the licensing and the sucky build system later May 24 12:09:02 the bb file should scare you :) May 24 12:15:01 Crofton: there are more bb files that scare me. May 24 12:19:17 Crofton: can you do an svn up on oestats on amethyst? May 24 12:48:35 jeremy_laine, I don't think I can May 24 12:50:21 jeremy_laine: can you still build angstrom for mpc8313e-rdb? My builds fail on glibc with bug #4303. May 24 12:51:07 * * OE Bug 4303 has been created by likewise(AT)gmx.net May 24 12:51:09 * * powerpc-angstrom-linux-gcc: error trying to exec May 24 12:51:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4303 May 24 13:19:02 03crofton 07org.oe.dev * r6ba34771... 10/ (3 files in 3 dirs): May 24 13:19:02 dsplink : More build fixes. bb file builds without errors at this point. May 24 13:19:02 More work needed. May 24 17:18:43 03rpurdie 07org.oe.dev * r4e6a3ba4... 10/ (6 files in 3 dirs): stagemanager-naitve: Sync with poky, add ipkg-build and ipkg scripts so packaged-staging dependencies can be removed May 24 17:18:47 03rpurdie 07org.oe.dev * r329b0550... 10/ (1 classes/patch.bbclass): patch.bbclass: Simplify patch task dependency handling and allow packages to override the dependency more easily (from poky) May 24 17:18:52 03rpurdie 07org.oe.dev * r0019f375... 10/ (1 classes/packaged-staging.bbclass): (log message trimmed) May 24 17:18:52 packaged-staging.bbclass: Sync with Poky for a number of updates. In summary: May 24 17:18:52 Support using opkg, ipkg and an internal shell script version of ipkg as the May 24 17:18:52 package manager allowing removal of a large number of build dependencies. May 24 17:18:52 Correctly install packages into staging creating the .list and .control files. May 24 17:18:54 Remove accidental install outside the locking in the build package function. May 24 17:18:56 Correctly handle error codes in the postamble function. Add code to work May 24 17:24:15 likewise: hi May 24 17:24:32 khem: hi khem May 24 17:24:48 khem, see #4303. That's what you saw yesterday, right? May 24 17:25:03 !oebug 4303 May 24 17:25:04 * * Bug 4303, Status: NEW, Created: 2008-05-24 14:50 May 24 17:25:05 * * likewise(AT)gmx.net: powerpc-angstrom-linux-gcc: error trying to exec 'cc1': execvp: No such file or directory May 24 17:25:07 * * http://bugs.openembedded.net/show_bug.cgi?id=4303 May 24 17:25:28 likewise: yeah May 24 17:25:40 khem: try with non-threaded bitbake. May 24 17:25:55 likewise: its strange though if I try next time same stuff works May 24 17:25:59 likewise ok May 24 17:26:04 khem: yes, same here May 24 17:26:20 likewise: I just updated 4303 May 24 17:26:23 khem: I hope you had BB_NUMBER_THREADS > 1 and = 1 will solve it May 24 17:26:31 I'm almost certain that its packaged-staging May 24 17:26:35 likewise: I am not sure if it has to do with build May 24 17:26:53 RP: extremely very much thanks! May 24 17:26:58 likewise: yep BB_NUMBER_THREADS=2 May 24 17:27:00 for me May 24 17:27:23 khem: yes, Richard can confirm it might be the combo of threading and packaged staging (used by Angstrom as well). May 24 17:27:36 it makes my builds work again. May 24 17:27:54 khem: I'll start off a eglibc build right now May 24 17:28:09 likewise: with BB_NUMBER_THREADS=1 ? May 24 17:28:13 khem: yes May 24 17:28:21 mine is running atm May 24 17:28:24 khem: k May 24 17:28:28 man I love this oestats May 24 17:28:46 :-) May 24 17:28:47 It tells me what all crap is going on May 24 17:28:59 My builds somehow do not report. May 24 17:29:22 jeremy_laine, said if a report fails to upload, it stops trying May 24 17:29:38 RP: what does packaged staging May 24 17:29:43 Crofton|work: I know, but why would it fail with me all the time. May 24 17:30:08 khem: http://oe.linuxtogo.org/wiki/PackagedStaging May 24 17:30:13 I think we need to talk to jer May 24 17:31:05 Crofton|work: are the stats supposed to be uploaded on each task being finished - i.e. "live" ? May 24 17:31:10 yeah May 24 17:31:29 Khem, do you have any experience with TI's dsplink? May 24 17:32:18 Crofton: not much May 24 17:32:31 Crofton: I care about the ARM parts May 24 17:32:47 I'm trying to get a bb file put togehter for it May 24 17:33:47 it runs, but does not build the .lib's and kernel modules May 24 17:34:18 Crofton: I guess it has to be tied to kernel May 24 17:34:33 yeah May 24 17:34:38 or something May 24 17:34:43 the build system is nuts May 24 17:35:12 Hmm so packaged staging in layman language is that u create a packge ipk and then install into a temporary root file system (staging) and you operate on packages May 24 17:35:24 Crofton|work: modules are buildable against a already-built kernel, provided the right Makefile is used. See Linux Device Drivers 3rd edition online May 24 17:36:27 re May 24 17:37:41 khem: sorry, this mail is much more up-to-date than is the wiki: http://lists.linuxtogo.org/pipermail/openembedded-devel/2008-March/004833.html May 24 17:38:01 hi flo_lap! how's life? May 24 17:38:53 Khem: It puts staging under control of a package manager so we can remove things from it and are able to rebuild it May 24 17:39:22 RP: OK yeah I have experience in doing same with rpm May 24 17:39:58 RP: Its difficult for packages which have bootstrap deps like toolchains has May 24 17:40:04 Khem: In theory you can share the staging packages and shorten build times if suitable staging packages are available May 24 17:40:32 RP: are staging packages differrent than deployed ones May 24 17:40:32 and "-c clean" removes files from staging May 24 17:40:39 Khem: yes May 24 17:40:42 likewise: hi... this weekend is hard. Tax stuff, preparing LinuxTag booth and a short talk and the garden (which means cutting grass a whole day) May 24 17:40:47 RP: hmmm I see May 24 17:40:59 RP: why not use same ipks May 24 17:41:09 Khem: The staging package includes the deploy/ and stamps/ output too May 24 17:41:12 RP: how should I rebuild my staging ? Suppose I want to clear my staging area completely and repopulate it? May 24 17:41:28 Khem: and our staging output is not the same as our deploy output May 24 17:41:48 RP: hmmm ok May 24 17:41:51 likewise: wipe tmp apart from the deploy/pstage directory May 24 17:42:13 RP: ok tnx May 24 17:42:54 RP: I think it should be possible to have same package used what are extra requirements for staging than deploying May 24 17:43:37 Khem: We need the data from deploy/ and stamps/ for a start May 24 17:43:45 RP: I have had trouble rebuilding glibc-intermediate because I think it removed parts of glibc when I did -c clean May 24 17:44:03 from staging May 24 17:44:55 Khem: This has been done to death on the mailing lists in the past - there are two ways to approach the problem, one is to reuse the -dev packages, the other way is separate staging packages. The latter way is the way packaged-staging works atm, if someone else wants to make it do the other, good luck to them ;-) May 24 17:45:23 Khem: There is the problem that if packages overlap, cleaning one might damage another May 24 17:45:55 Khem: gcc-cross and glibc are a problem in that regard with their -initial and -intermediate varieties... May 24 18:04:17 Stupid question May 24 18:04:55 If you have foo_1.0.bb, PR=r1 in both say openembedded/packages/foo/ and override/packages/foo/, shouldn't you be able to set the BBFILES path so that overrides/packages/foo/foo_1.0.bb is used? May 24 18:06:22 Tartarus: yes, thats what collections should do May 24 18:07:20 and it'd be openembedded/packages/*/*.bb override/packages/*/*.bb, right? May 24 18:07:25 (Same order for BBPATH stuff too?) May 24 18:10:00 Tartarus: You have to set the priority of the collections May 24 18:11:23 ok. higher priority wins? May 24 18:14:55 Tartarus: I can't remember which way around ;-) May 24 18:15:50 RP: I think we could do something like this. gcc-cross obsoletes gcc-cross-initial May 24 18:16:24 so when gcc-cross is installed into staging gcc-cross-initial will be punted automatically May 24 18:16:47 and when we do gcc-cross-initial rebuild then it wont find it installed May 24 18:18:51 Tartarus: yes higher wins or you can have a higher version number on the package you wish to include in your package or you can also add PREFERRED_VERSION_pacakge="x.y.z" May 24 18:19:15 Tartarus: yes, higher priority number go first May 24 18:19:30 Tartarus: I can pastebin a working example, but it's in the user manual somehwere May 24 18:19:44 khem already anwered that - whooops May 24 18:23:51 likewise: my build succeeded May 24 18:23:57 :) yippie May 24 18:24:12 khem: mine did not May 24 18:24:34 likewise: hmmm whats error May 24 18:25:01 likewise: I see ur name but no reports from you on oestats May 24 18:25:08 khem: fails somewhere in gcc-cross, gcc/config/rs6000/darwin-ldouble.c: In function 'fmsub'. version node not found for symbol __longjmp@@GLIBC_2.3.4 May 24 18:25:31 khem: I know, my reports are not getting through... don't know why, I just switched to amethyst again, but to no avail yet. May 24 18:25:33 likewise: mac struff May 24 18:25:57 khem: I'm building for powerpc (mpc8313e0rdb on x86 32-bit ubuntu 8.04) May 24 18:26:12 likewise: OK May 24 18:26:18 khem: which stats server are you suing? May 24 18:26:26 s/suing/using/ May 24 18:26:35 likewise: amethyst May 24 18:27:26 likewise: so you are buildiing for powerpc-*-*-gnuspe May 24 18:27:31 right May 24 18:27:45 mpc sounds a freescale thing May 24 18:28:34 khem: Freescale, yes, e300 (follow-up to 603e core). Don't know about the -gnuspe part... May 24 18:28:54 likewise: yeah thats the triplet for ppc with SPE cores May 24 18:28:58 khem: older eglibc's went fine (I have kept a few known-good build revisions) May 24 18:29:08 likewise: hmmm May 24 18:29:10 khem: no SPE core in the E300, it's not a Cell May 24 18:29:26 likewise: I will soon add 2.5 2.6 2.7 recipes May 24 18:29:30 for eglibc May 24 18:29:55 so you do not need to be using bleeeding stuff always May 24 18:30:11 khem: its seems my build is now reporting to amethyst. i hope it remains working May 24 18:30:27 khem: saves me a lot of copy-pasting May 24 18:34:50 yeah May 24 18:36:16 likewise: your error does not relate to eglibc though May 24 18:36:39 khem: is there anything substantially different between 2.5 2.6 2.7 and 2.8 btw? May 24 18:36:44 likewise: Its to do with your host glibc its cross-compile May 24 18:36:52 khem: I mean, bb recipe wise May 24 18:37:09 likewise: I dont think so except few patches may be May 24 18:37:15 for enabling builds May 24 18:37:17 and backports May 24 18:37:34 likewise: eglibc does not release tar balls May 24 18:37:44 so all these releases are actually branches May 24 18:37:49 in svn repo May 24 18:38:05 now I think even glibc has started doing that May 24 18:38:08 khem: that's nice May 24 18:38:09 no more tar balls May 24 18:38:31 because with tar balls there were always following .x releases May 24 18:38:56 likewise: Can you paste your error somehwere May 24 18:39:39 khem: yup, will do, hold on May 24 18:40:18 likewise: if its a gcc thing then you can try to use a different version may be 4.3 :) May 24 18:40:32 khem: pastebin.ca/1028116 May 24 18:41:03 khem: well, I had a perfect glibc build earlier, and this toolchain has been working for me ok so far :-) May 24 18:41:20 khem: change *one* item at a time :-) May 24 18:41:37 hmmm ok May 24 18:41:40 hi likewise, hi khem! May 24 18:41:48 HopsNBarley: howdy May 24 18:42:12 HopsNBarley: hi! May 24 18:42:52 HopsNBarley: you must certainly catch-up on the oestats feature we have now. May 24 18:42:58 what's the state of gcc 4.3 in OE? May 24 18:43:08 i was just looking at the logs here, about to try it (-; May 24 18:43:14 HopsNBarley: flakey, under test, depends on your target, etc... May 24 18:43:18 mips? May 24 18:43:30 HopsNBarley: Dunno. You going to mips? May 24 18:43:31 (new contract, possibly OE :-) May 24 18:43:37 HopsNBarley: :-) May 24 18:43:49 some board bringup stuff - pretty cool. May 24 18:43:50 HopsNBarley: on a different note I have built a mips toolchain last week outside OE with gcc 4.3 all went file May 24 18:44:02 HopsNBarley: I have only tried build May 24 18:44:06 crosstool? May 24 18:44:22 khem: no weird patches needed? we can just copy paste any toolchain that's out there and copy their build May 24 18:44:23 HopsNBarley: no its my own scripy May 24 18:44:37 Khem, okay. May 24 18:44:59 HopsNBarley: I will put this script on my site soon :) when the site is ready May 24 18:45:12 I can send it your way if u want too May 24 18:45:19 that would be awesome. May 24 18:45:27 wanna meet this week? May 24 18:45:30 Crosstool is an overkill May 24 18:45:35 HopsNBarley: sure May 24 18:45:59 what area do you work? May 24 18:46:03 khem: if you think crosstool is an overkill, than OE certainly is as well? :-) May 24 18:46:03 likewise: I see the error is in libgcc so there seems to be some eglibc factor May 24 18:46:26 likewise: OE is a complete distro May 24 18:46:51 Crosstool has so much of legacy/old cruft that it get confusing May 24 18:46:58 ct-ng is better though May 24 18:47:03 khem: true. crosstool-ng - -yup May 24 18:47:09 :-) May 24 18:47:29 likewise: you are still using old bitbake 1.8.10 ? May 24 18:47:36 Khem, where do you work here in the valley? May 24 18:47:48 Khem: yes, 1.8.10 rev 1074 May 24 18:47:51 location wise, i mean. May 24 18:47:58 HopsNBarley: Milpitas May 24 18:48:14 I live in Santa Clara near university May 24 18:48:21 cool - new contract is just north of airport - how about rivermark? May 24 18:48:27 cool May 24 18:48:35 rivermark is ok May 24 18:48:38 piatti ? May 24 18:48:54 haven't tried it. May 24 18:49:03 ok they also have others May 24 18:49:15 yeah, lots of stuff there. i like the japanese place. May 24 18:49:22 lets meet outside blockbuster then decide May 24 18:49:27 yoyo sushi May 24 18:49:28 tuesday? May 24 18:49:37 tuesday is not good May 24 18:49:47 thursday. May 24 18:49:55 thursday is ok May 24 18:50:05 monday/wednesday not good for me. May 24 18:50:20 RP, ok, higher priority collections win May 24 18:50:51 likewise: your triplet should be linux-gnuspe May 24 18:51:10 likewise: I mean TARGET_OS May 24 18:51:15 for ppc with spe May 24 18:51:32 spe as in synergetic processing engine?? May 24 18:51:50 likewise: :) May 24 18:54:12 khem: where did you find this? May 24 18:54:39 likewise: in your log May 24 18:57:33 Tartarus: howdy May 24 18:59:09 hey khem May 24 18:59:40 i'm off to lunch, see ya Khem, likewise! May 24 18:59:58 khem: I'm not fully knowledgable of the tripplet . You suggest that I should set TARGET_OS to linux-*-*-gnuspe? How did you find I need -gnuspe? May 24 19:00:02 HopsNBarley: bon appetit May 24 19:00:19 HopsNBarley: guten appetit May 24 19:00:39 likewise: you are using mpc with spe May 24 19:00:48 and thats the right one May 24 19:01:04 for you processor May 24 19:02:44 Khem: strange that I never needed that tripplet before... May 24 19:02:54 Khem: eglibc is more strict? May 24 19:03:19 likewise: generally if you want to use spe that triplet will select stuff for you May 24 19:03:39 likewise: what doubles are you using May 24 19:03:43 format I mean May 24 19:04:03 likewise: this error you are getting is strange I have not seen it so far May 24 19:04:27 but it looks like new glibc removed some symbols may be May 24 19:05:39 khem: I have no clue. I used to build for ppc603e (machine=efika), and then for mpc8313e-rdb (e300c3 core). I never touched TARGET_OS when I brought this machine in. Worked so far. I am not very familiar with the FPU double format representations under PPC architecture. I have been involved with PPC 15 years ago, then much ARM, now back to PPC again :-) May 24 19:09:59 likewise: builds ok for me http://oestats.jerryweb.org/packages/886/ May 24 19:11:13 likewise: if you built two different powerpc machines I've seen oddities May 24 19:11:28 jeremy_laine: are you using multithreaded bitbake?' May 24 19:11:44 likewise: I'm using -j3 for all my builds May 24 19:11:56 jeremy_laine: that's multithreaded make, not bitake May 24 19:12:10 likewise: I'd try killing build/tmp and rebuilding May 24 19:12:24 jeremy_laine: I always do that. May 24 19:12:28 jeremy_laine: BB_NUMBER_THREADS? May 24 19:12:37 I think the staging goes wrong somewhere and you end up with leftoftvers May 24 19:12:53 jeremy_laine: I always do builds from scratch. I have had 80% fail over the last weeks. May 24 19:12:53 likewise: ok May 24 19:13:14 jeremy_laine: what is your BB_NUMBER_THREADS setting? May 24 19:13:19 likewise: ah no I don't set BB_NUMBER_THREADS May 24 19:13:26 jeremy_laine: that explains then May 24 19:13:39 jeremy_laine: thanks. May 24 19:13:40 likewise: its you and me only :) May 24 19:13:53 Khem: well, it really speeds up the builds here :-) May 24 19:14:09 likewise: do u have multi cpu machine ? May 24 19:14:16 Khem: I'm getting a dual quad-core Xeon workstattion at work because they noticed I did a lot of compile jobs :-) May 24 19:14:29 I have a dual-core E6400 at home May 24 19:14:33 likewise: wow May 24 19:14:40 we have a quad-core here but we usually run several builds concurrently so the cores are nicely used May 24 19:14:53 likewise: may be -jx will be ok to a certain extent May 24 19:14:56 Khem: indeed :-) They bought it for the FPGA guys but they don't want to switch machine :-) May 24 19:15:03 Khem: -jX is ok yes May 24 19:15:07 heh May 24 19:15:15 khem: not completed, the "configure" step takes a while and is single thread May 24 19:15:27 Khem: BB_NUMBER_THREADS is ok too, RP just reported in (see ml) that he is testing some breakage with it. May 24 19:15:36 yeah May 24 19:15:48 likewise: ok is there some fix ? May 24 19:16:16 likewise: feel free to include your builds on oestats.jerryweb.org May 24 19:16:16 khem: RP is testing fixes he did in their Poky distribution, after that he looks into bringing them over to .dev . May 24 19:16:27 likewise: by the way there is now a build search tool May 24 19:16:31 jeremy_laine: my build reports do not arrive at your place... May 24 19:16:52 likewise: http://oestats.jerryweb.org/search/?machine=mpc8313e-rdb May 24 19:17:14 jeremy_laine: did you ever build against eglibc (in the past)? May 24 19:17:16 likewise: libc { May 24 19:17:16 GLIBC_2.3.4 { May 24 19:17:16 _longjmp; __sigsetjmp; _setjmp; May 24 19:17:16 longjmp; setjmp; May 24 19:17:17 } May 24 19:17:51 likewise: this is from libc/sysdeps/powerpc/Versions May 24 19:18:00 * likewise is noting down May 24 19:18:07 so it should work May 24 19:18:29 khem: you mean, with -gnuspe or without? May 24 19:18:50 no its for all powerpc May 24 19:19:17 khem: argh. well let's see what the current build does. it's in eglibc intermediate now May 24 19:19:23 likewise: hmmm but I dont see longjmp with two underscores(__ May 24 19:19:29 ok May 24 19:19:49 likewise: I'd say possibly a threading issue May 24 19:20:40 jeremy_laine: must be -j then, because my BB_NUMBER_THREADS has been set to 1 on this build May 24 19:20:40 jeremy_laine: how does BB_NUMBER_THREADS handle dependencies May 24 19:21:01 khem: the intertask dependencies are still checked, so it's like a CPU pipeline. May 24 19:22:02 khem: however, with the arrival of packaged-staging, the interdependency descriptions seem broken - ask RP for details, I'm only a user of his magic :-) May 24 19:22:26 khem: eglibc-intermediate breaks again, same spot with 1 bb thread. May 24 19:22:38 likewise: OK so its better to be conservative for now May 24 19:22:53 hmmm May 24 19:23:00 execp May 24 19:23:02 khem: yes, indeed. see openembedded-devel list last email thread May 24 19:26:03 likewise: I can try one ppc build May 24 19:26:51 likewise: I hope this issue is not in eglibc's parallel build May 24 19:26:59 Khem: I use MACHINE="mpc8313e-rdb" ANGSTROM_MODE=eglibc May 24 19:27:05 khem: OK, I'll try with -j1 May 24 19:27:58 likewise: ok May 24 19:29:32 lol, clean tasks are also reported in via oestats :-) May 24 19:31:16 likewise: have you used seppuku May 24 19:31:22 khem: no never May 24 19:31:49 khem: I think it's useful for regression reports, not for development work May 24 19:31:55 like is oe login associated to oe bugzilla ? May 24 19:32:08 khem: no, need to create a new account AFAIK May 24 19:32:16 ok May 24 19:32:16 you need to ... May 24 19:33:29 hmm May 24 19:33:45 Should I file about about xserver-kdrive-xomap possibly trying to build other servers and failing? May 24 19:33:58 (This isn't in a stock setup really) May 24 19:34:22 Tartarus: if we can improve, yes report in May 24 19:35:35 likewise: what image do u try to build May 24 19:35:40 console-image May 24 19:35:45 likewise will helloworld do May 24 19:35:50 yes May 24 19:35:50 to reproduce this May 24 19:35:59 lemme try it now May 24 19:36:02 tnx May 24 19:36:31 my -j1 build is chewing away now, will take a bit longer of course May 24 19:37:30 likewise: this oestats will it ever delete the build reports too May 24 19:38:06 ask jeremy_laine he created this piece of art May 24 19:38:39 * Khem turns towards jeremy_laine__ May 24 19:39:09 Khem: for now the build reports are not automatically expired, but there is an admin interface May 24 19:40:03 khem: with -j1 I'm getting a different error. May 24 19:41:06 khem: http://amethyst.openembedded.net:8000/public/logs/13840.txt May 24 19:41:38 RP: are you about? (I have encountered a strange bitbake issue that I wonder if it has been fixed or noted before) May 24 19:41:44 khem: home/leon/sandbox/mpc8313e/openembedded/build/tmp/cross/lib/gcc/powerpc-angstrom-linux/4.2.2/../../../../powerpc-angstrom-linux/bin/ld: cannot find -lgcc_eh May 24 19:41:44 collect2: ld returned 1 exit status May 24 19:42:44 likewise: hmm well known error May 24 19:42:59 likewise: let me try it out May 24 19:43:05 and see what I find May 24 19:43:19 libgcc_eh is not there May 24 19:47:23 khem: http://svn.exactcode.de/t2/trunk/package/base/glibc/gcc_eh.patch.cross May 24 19:47:57 like2wise: yeah but I think we got rid of this May 24 19:49:16 exit May 24 19:49:32 like2wise: I think it needs to be looked at May 24 19:49:58 Khem: ok, I'll see if my glibc builds will work now that I have concurrency disabled. May 24 19:50:29 khem: I'll also look into he known-good revisions of OE for this machine. May 24 19:50:45 (i.e. were eglibc worked for me) May 24 21:26:27 is there a list of virtual packages defined somewhere ? or is it ok to add a new virtual/foo dependency, as long as there is at least one provider for virtual/foo ? May 24 21:29:19 cyrilRomain: no virtual packages list harcoded because oe rocks! so you are free to define new virtual packages May 24 21:30:37 cyrilRomain: yeah, cool! I asked because I'm afraid there is a need for virtual/mta in oe. May 24 21:31:13 Indeed some packages like 'at' does not build if you don't have a mail transfert agent :-/ May 24 21:34:18 cyrilRomain: I guess it's OK as long as there is enough similarity in fucntionality. May 24 21:35:29 likewise: unless you have code like that: http://rafb.net/p/BBhNPy37.html May 24 21:37:40 so I could a) install sendmail and don't care about that. b) create a dirty patch s,#error,perr(), c) add virtual/mta in oe May 24 21:39:01 cyrilRomain: eek May 24 21:39:05 * * OE Bug 4304 has been created by andrea.adami(AT)gmail.com May 24 21:39:07 * * c7x0 wont resume using linux-rp_2.6.23 May 24 21:39:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4304 May 24 21:39:54 hey! likewise and like2wise are likewise ;) May 24 21:40:25 not anymore :( May 24 21:43:28 well, debian and i bet other distros get away with virtual mta functionality May 24 21:43:41 since it's generall exec("sendmail %s", args) May 24 21:43:50 and lots of mtas will take fake sendmail args May 24 21:44:35 Tartarus: hmm actually Gentoo added virtual/mta in their ebuilds May 24 21:46:39 cyrilRomain: yes, they are 1 meter apart and the same person types it :-) May 24 21:47:25 cyrilRomain: it's just that one machine is now being devoted to COD4 after a full day OE. May 24 21:48:08 likewise: heh :) May 24 21:53:29 good morning May 24 21:54:01 hi ant May 24 21:54:24 cyrilRomain: bonsoir (entre nous pas de UGT) May 24 21:55:50 cyrilRomain: have you tried http://amethyst.openembedded.net:8000/ May 24 21:55:55 ant: UGT ? May 24 21:57:09 ant: nop, but it looks really nice May 24 21:57:45 err.. Exception Type: ViewDoesNotExist now... May 24 21:57:46 ant: hm well not _now_ but .. May 24 21:58:00 ;-> May 24 21:58:29 ~UGT May 24 21:58:30 i heard ugt is Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html May 24 21:59:05 arf, thanks mwester-laptop May 24 21:59:32 * cyrilRomain knows UGMT .. May 24 22:00:40 cyrilRomain: I did not respond before cause could not find Tilde (~) on the keyboard...too busy with OE and I let my poor Gentoo in a semi-finished state May 24 22:01:19 ~curse xkeyboard May 24 22:01:20 May the fleas of a thousand camels infest your most sensitive regions, xkeyboard ! May 24 22:01:46 seems like M in UGMT is for 'mean' May 24 22:02:12 Universal Getting Mean Time? May 24 22:03:05 haha lol ! May 24 22:03:12 I really need some sleep May 24 22:03:17 That would be the start of a soccer match, then. May 24 22:03:32 I read 'Universal Greenwich Time' :D May 24 22:15:29 cyrilRomain: I'm checking the differences between two virtuals May 24 22:16:23 is bitbake -g [pkg] similar to equery depgraph [pkg] ? May 24 22:16:56 * cyrilRomain lists packages that may depends on a mail transfert agent May 24 22:17:08 similar output I mean.... May 24 22:17:17 ant: hmm probably May 24 22:18:05 cyrilRomain: I have to find a way to see the diffs btw "xserver-kdrive" and "xserver-kdrive-imageon" May 24 22:18:13 .dot are too long... May 24 22:18:24 for human eye May 24 22:19:11 ant: hrw|gone should have a oneliner for that ;) May 24 22:21:14 yah, and on the Koen's blog too was something... May 24 22:22:27 http://dominion.thruhere.net/koen/cms/node?page=3 May 24 22:24:39 ant: hmm 'equery d virtual/mta' list build dependencies, but it cannot list runtime dependencies May 24 22:25:38 ant: strangely, at does RDEPEND virtual/mta, whereas the code clearly indicates a build dependency :-/ May 24 22:26:06 * cyrilRomain suspects 'at' does not build on Gentoo without a mail transfert agent installed May 24 22:30:42 ok, if I didn't missed one there are 3 packages in oe that can provide a mail transfert agent: esmtp, msmtp, and of course postfix May 24 22:31:39 masqmail? May 24 22:32:21 Tartarus: i missed that one May 24 22:32:54 clamsmtp May 24 22:33:03 * Tartarus does an apt-cache search mail-transport-agent and checks for oe packages May 24 22:33:09 * cyrilRomain perform a cross-check May 24 22:33:32 ssmtp May 24 22:33:46 that's it from me :) May 24 22:35:50 Can anybody else access ftp.prm.org, as I can't May 24 22:36:03 likewise, Khem: I pushed the packaged-staging fixes into OE.dev this morning May 24 22:36:21 mwester: I'm about very briefly now... May 24 22:37:54 Tartarus: i did that search in Gentoo and checks for oe, but we also need to search the other way around May 24 22:38:52 RP: this morning? May 24 22:39:02 RP: so after we tested even? May 24 22:41:02 RP: I will do some concurrent builds to test. May 24 22:41:38 cyrilRomain, k. I'd say you've probably got a good enough list now, too :) May 24 22:41:42 * Tartarus heads our May 24 22:41:45 s/our/out/ May 24 22:43:09 Tartarus: yes thanks for your help :) May 24 22:53:33 Tartarus: hmm maybe tinymail and uucp too May 24 22:56:13 well May 24 22:56:14 $ apt-cache show uucp May 24 22:56:25 Recommends: postfix | exim4 | exim | smail | courier-mta | zmailer | rmail May 24 22:57:09 Tartarus: ah ok it run upon a real mail transfert agent May 24 22:57:14 yeah May 24 22:57:26 UUCP could be setup to move mail around May 24 22:57:34 but it's not an mta in the sense we think of now May 24 22:57:42 yeah May 24 22:58:12 and re tinymail, tmut is a tiny email client that uses libtinymail May 24 22:58:45 libtinymail could be used by programs that want a built-in mta May 24 22:58:59 but not used in the sendmail/etc sense, without another program in front May 24 22:59:06 and now really, back to heading out :) May 24 22:59:31 Tartarus: ok, thanks for your checks :) May 24 23:00:42 cyrilRomain: for embedded systems postfix would be overkil... May 24 23:01:15 ant: what package would you use ? May 24 23:01:35 I use postfix in production but it eats up almost 200 mb on my server (ok, with amavisd-new, spamassassin, etc) May 24 23:01:52 and before-queue filtering.. :-) May 24 23:02:39 I see Gentoo standard uses ssmtp May 24 23:03:33 ant: yeah in binary size: ssmpt < esmtp < msmtp May 24 23:03:39 er May 24 23:03:53 ssmpt < msmtp < esmtp May 24 23:04:09 if you don't need a MDA (local delivery) would be fine then May 24 23:04:16 ssmtp is extremy small! May 24 23:04:48 it's not a demon, it is called by a pipe May 24 23:05:12 and you need external SMTP server May 24 23:07:32 ah .. it is the same with esmtp and msmtp ? May 24 23:07:44 cyrilRomain: this link (italian) talks about installation on FOX BOARD (Linux 4MB) http://wiki.matteocarli.com/tutorial/linux/ssmtp May 24 23:08:10 try on Google for more info May 24 23:08:40 ant: ok thx May 24 23:09:07 I'll look at that tomorrow (I'm too tired now) May 24 23:09:11 nite all May 24 23:09:23 g'nite May 24 23:14:48 guys, that's small...http://www.acmesystems.it/?id=4 May 24 23:15:48 OX Board runs a real Linux operating system (not a uC Linux) on an ETRAX 100LX microprocessor, a 100MIPS RISC CPU made by Axis May 25 00:49:55 so i found that the portaudio package has entirely changed its file layout since the bb file was written May 25 00:50:23 i managed to point it at all the right files, and now i get QA errors: debug files found in non-debug packages May 25 00:50:45 the documentation on qa stuff seems to be a skeleton May 25 00:50:57 i can rm the debug files in the install step, but what is the Right Way? May 25 01:07:21 oooor i can't rm the debug files in the install step...? May 25 01:07:59 sakoman, do you know the Right Way to banish debug files elsewhere so that there aren't fatal QA errors in the package step? May 25 01:08:59 i got portaudio to build and install but May 25 01:09:00 ERROR: QA Issue: non debug package contains .debug directory: portaudio-examples path /work/armv5te-angstrom-linux-gnueabi/portaudio-1:0.0.19+cvs20060814-r1/install/portaudio-examples/usr/bin/.debug/patest_write_sine May 25 01:09:22 er, i mean, hi sakoman :) May 25 02:45:31 03mwester 07org.oe.dev * r9d5560df... 10/ (1 packages/linux/unslung-kernel_2.4.22.l2.3r63.bb): Unslung: kernel - fix double suffix on kernel image name (.bin.bin) **** ENDING LOGGING AT Sun May 25 02:59:56 2008