**** BEGIN LOGGING AT Sat Jan 12 02:59:57 2008 Jan 12 03:35:01 * Crofton observes he needs to figure out kill files and ignore commands Jan 12 05:04:14 03pfalcon 07org.oe.dev * rb14f8bb0... 10/ (1 packages/supertux/supertux_0.1.2.bb): Jan 12 05:04:14 supertux 0.1.2: Disable flaming opengl. Jan 12 05:04:14 * Gives whole 0.2 FPS on average ARM hardware. Jan 12 05:04:19 03pfalcon 07org.oe.dev * ra541f884... 10/ (1 packages/tea/tea_17.3.5.bb): tea 17.3.5: Typo fix in DESCRIPTION. Jan 12 09:36:12 morning all Jan 12 10:30:40 why does sanity.bbclass check for texi2html, and not for texinfo ? Jan 12 10:30:58 more globally, I think the RequiredSoftware page need to be in sync with what is in sanity.bbclass Jan 12 10:31:38 and split _required_ software (such as make, m4, etc.) from _recommended_ software (such as psyco) Jan 12 10:45:01 cyrilRomain: ACK on all points Jan 12 10:47:24 moin henrykl Jan 12 10:47:26 s/ll// Jan 12 10:47:34 Moin Jan 12 10:47:51 OT: Firefox is oh so flakey on 64bit Jan 12 10:48:03 crashes all the time *sigh* Jan 12 10:48:14 mickeyl: works fine for me, but the plugins are hell. nspluginwrapper is a godsend Jan 12 10:48:39 hmm, bad Karma here then :) Jan 12 10:48:47 pity galeon was abandoned Jan 12 10:48:53 epiphany just doesn't cut it for me Jan 12 10:49:30 gm Jan 12 10:49:41 yes, galeon rocks the world. pity that due to the availability of NoScript I find myself using Firefox more and more Jan 12 10:50:16 *nod* Jan 12 10:50:56 * ant tried midori and seems ok to him apart small glitches (missing icons) Jan 12 10:51:07 ya, looking forward to any webkit based browser Jan 12 10:51:12 mickeyl: why not switching to a webkit-based browse, like midori (gtk) or improving the qt demo ? ;) Jan 12 10:51:18 mickeyl i don't have about any problem now with firefox on 64bit ... (Fedora 8) Jan 12 10:51:37 cyrilRomain: well, if I start getting into browser development, i can right on stop all other activities ;) Jan 12 10:51:46 mickeyl: heh Jan 12 10:51:56 gremlin[it]: may be my ubuntu gutsy. I'm not satisfied with that anyways... i should have kept my Mandriva Jan 12 10:52:03 cyrilRomain: bjour Jan 12 10:52:13 hey ant|work Jan 12 10:52:31 Talking about these browsers, I recently tried them out on the spitz. Are they all dog-slow or is it me? Jan 12 10:52:49 Laibsch: tried on c7x0 Jan 12 10:52:54 Laibsch: not sooo slow Jan 12 10:52:57 Furthermore, I found no way to enter an URL to open, again just me? Jan 12 10:53:05 ant|work: which one? Jan 12 10:53:19 Laibsch: midori - the text-boxes need resize.... Jan 12 10:53:42 mickeyl Firefox crashed easily in past Fedora version specially with 32bit plugings ... now is good Jan 12 10:53:50 mickeyl news about hotel ? Jan 12 10:54:06 gremlin[it]: yeah. we decided on the Astrid again Jan 12 10:54:17 i offered to take over the additional breakfast fee :) Jan 12 10:54:27 cyrilRomain:mickeyl: about required software, why don't start telling people about gcc-native? Jan 12 10:54:48 cyrilRomain:mickeyl: RP was opting for a discussion on .dev mailing list Jan 12 10:55:21 cyrilRomain:mickeyl: so that one can avoid the gcc3 mess Jan 12 10:55:37 sure, lets talk about that on .dev Jan 12 10:55:40 mickeyl hahhah for all of us ??? :) Jan 12 10:56:02 gremlin[it]: if it comes to that, yeah. i sincerely hope we're not more than 10 though ;) Jan 12 10:56:14 * mickeyl poor freelancer Jan 12 10:56:18 but the offer stands Jan 12 10:56:22 hahah :) Jan 12 10:58:49 different topic: anyone know whether we can build n810 / chinook compatible packages w/ OE? If so, which machine configu? Jan 12 10:58:50 mickeyl: and what about Origyn web browser (http://www.sand-labs.org/owb): is it ready to use ? Jan 12 10:58:57 never heard about that one Jan 12 10:59:15 mickeyl: it is OpenMoko related though ;) Jan 12 10:59:26 oops Jan 12 10:59:41 yet another browser on webkit Jan 12 10:59:43 need to have a lok Jan 12 10:59:45 look, as well Jan 12 11:01:04 * cyrilRomain think webkit is the future due to the very good javascript support of webkit and the heavyweight of Gecko Jan 12 11:05:02 ya Jan 12 11:05:11 i'm glad some people find browser engine hacking sexy Jan 12 11:05:13 *cough* Jan 12 11:05:24 ant|work:midori takes about a minute to start here. Then it tries to load a page which for some reason cannot be interrupted -> totally unusable for me Jan 12 11:05:41 Laibsch: did you built it? Jan 12 11:05:47 Laibsch: or from feeds? Jan 12 11:05:48 ... software and sexy ... ?!?!?! Jan 12 11:06:06 Laibsch: I took one from the feeds 1 week ago Jan 12 11:07:22 gremlin[it]: sure. i at least find hacking fluid UI apps sexy Jan 12 11:09:07 mickeyl recently i find something trully sexy http://picasaweb.google.com/gremlin.it/SickGirl200801Vicenza/photo#5152410289572225474 Jan 12 11:09:18 Laibsch: probably I did "echo 1 > /proc/sys/vm/overcommit_memory" before (for navit) Jan 12 11:09:38 see later, i go to cook ... Jan 12 11:11:30 ant|work: what does that overcommit thing do? Jan 12 11:11:36 why is it not default? Jan 12 11:11:46 Its about the OOM killer Jan 12 11:12:19 actually is an hard hack... Jan 12 11:18:35 nothing killed yet, so I guess that won't be needed Jan 12 11:22:09 Laibsch: sure you got all dependencies? Jan 12 11:22:29 Laibsch: built and installed? Jan 12 11:23:01 Laibsch: I'll retry in one hour...@/home Jan 12 11:24:40 ant|work: If I was lacking dependencies, I thinkit would crash Jan 12 11:24:49 Maybe I can cook up a qemu image Jan 12 11:25:54 Laibsch: I remember the render was very nice Jan 12 11:26:37 Laibsch: but the textbox for URL input was too small Jan 12 11:26:52 Laibsch: and two icons mising Jan 12 11:26:59 Laibsch: *missing Jan 12 11:27:18 Laibsch: then I zapped all the nand.... Jan 12 11:28:38 bbl Jan 12 11:28:50 dito Jan 12 12:45:11 mickeyl, gm Jan 12 12:45:19 hi crofton Jan 12 12:45:46 Hi - How can I see the dependency tree for package "XYZ" in OE? Jan 12 12:47:48 I think bitbake can spit out fot files you can make the graph from Jan 12 12:49:23 pto: it's in the FAQ Jan 12 12:49:56 pto: bitbake is also able to output a graph with graphviz Jan 12 12:51:00 pto: with '-g' option Jan 12 12:51:02 Nice Jan 12 12:51:10 Will try this - thanx cyrilRomain Jan 12 12:51:55 pto: It's gonna be huge Jan 12 12:52:15 and of course the files output by -g are just text files that you can less or grep. Jan 12 12:52:20 The graphical representation tends to be too complicated to be of any help, at least IMHO Jan 12 12:52:27 Henryk: Exactly Jan 12 12:52:33 Which is what I usually do Jan 12 12:52:40 I just grep for what I want Jan 12 12:52:51 Henryk: but they can be seen in a graph with dotty iirc (or neato) Jan 12 12:53:20 cyrilRomain: Have you tried looking at one? Jan 12 12:53:22 cyrilRomain: sure. but that tends to be useless, if you have 2000 packages arranged over 10.000x10.000 pixels Jan 12 12:53:34 Henryk: indeed Jan 12 12:54:05 :-) Jan 12 12:56:53 Can I run bitbake "some big thing" and bitbake --dry-run --verbose at the same time? Jan 12 13:00:44 building a distro that uses gcc-4.2.x for i686 is borked at glibc-intermediate. To fix, don't -mcpu=pentiumpro with tune-pentiumpro.inc you gotta -march=i686 instead and add this gentoo patch too 1130_all_glibc-2.4-undefine-__i686.patch. Explanation why is in the patch. Good luck. Jan 12 13:17:06 pto: of course Jan 12 13:17:17 you can even run bitbake more than once for real Jan 12 13:17:26 Usually Jan 12 13:17:59 ~lart xerces-c Jan 12 13:17:59 * ibot frags xerces-c with his BFG9000 Jan 12 13:18:02 ~autotools Jan 12 13:18:03 somebody said autotools was a package of GNU software development tools including automake, autoconf and libtool, or made by the devil, or according to mickeyl doing more harm than good Jan 12 13:18:04 When both instances try to build the same package, the build will break, but easy to fix by "-c rebuild $brokenpackage" Jan 12 13:18:45 ibot: autotools is also -- however, still better than handcooked scripts Jan 12 13:18:45 mickeyl: okay Jan 12 13:18:49 ~autotools Jan 12 13:18:49 i guess autotools is a package of GNU software development tools including automake, autoconf and libtool, or made by the devil, or according to mickeyl doing more harm than good, or -- however, still better than handcooked scripts Jan 12 13:19:31 mickeyl, what is up with xerces-c? Jan 12 13:19:42 its build system sucks a bit Jan 12 13:19:45 but i just tamed it Jan 12 13:19:48 ah Jan 12 13:19:57 I did the original bb file .. Jan 12 13:20:03 ah Jan 12 13:20:07 and you are correct, it sucks Jan 12 13:20:14 I no longer care about it though Jan 12 13:20:16 take a look at my 2.8.0 one Jan 12 13:20:18 it's a bit cleaner Jan 12 13:20:25 good Jan 12 13:20:27 will commit soon Jan 12 13:20:36 need it for some strange package someone requested Jan 12 13:20:42 I removed it from the sphere of software I care about Jan 12 13:20:57 * Crofton|work only needs to read xml files Jan 12 13:24:58 | configure: Cannot find a working xsd tool Jan 12 13:25:10 any idea which package it needs for that? Jan 12 13:25:45 xerces-c? Jan 12 13:25:49 no Jan 12 13:25:51 this is a new one ... Jan 12 13:25:52 something different Jan 12 13:26:25 maybe something that xerces should put in staging bin? Jan 12 13:26:46 hmm, i hope not. otherwise i would need to add xerces-native Jan 12 13:26:55 ~lart dtnrg Jan 12 13:26:55 * ibot hooks into a hydrant and hoses dtnrg down Jan 12 13:26:59 urg Jan 12 13:27:06 googling is not helpful Jan 12 13:29:35 anyone knowing offhand how to add a local aclocal directory to the aclocal paths? Jan 12 13:30:45 mickeyl: FYI, xerces-c has no xsd-related dependency on Gentoo Jan 12 13:30:56 yeah Jan 12 13:30:58 this is DTNRG Jan 12 13:31:01 not xerces-c Jan 12 13:31:05 oh, sorry Jan 12 13:31:09 xerces-c being another dependency of DTNRG Jan 12 13:31:14 i have xerces 2.8.0 now Jan 12 13:31:24 but I'm looking for this mysterios xsd tool :) Jan 12 13:39:58 <_diego__> hello Jan 12 13:40:32 Was it so that the openmoko dudes are using OE however made their own setup afterall? Jan 12 13:43:59 can you rephrase that question? Jan 12 13:45:14 the openmoko dudes are using OE more or less out of the box Jan 12 13:49:15 <_diego__> i see that "screen" package don't compile becouse a patch is missing from debian server Jan 12 13:49:44 <_diego__> should that patch be added to OE repository? Jan 12 13:49:50 probably Jan 12 13:53:05 _diego__: hmm if the dowload fails it should fall back to the angstrom distribution sources, which has that file Jan 12 13:54:26 <_diego__> cyrilRomain, i just trying that patch downloaded from an old mirror found with google. Jan 12 13:54:59 _diego__: do you have recent metadata ? (there was a fix to make angstrom mirror work again (2007/12/31)) Jan 12 13:56:03 <_diego__> i'm using org.openembedded.dev branch, and i've updated just a week ago Jan 12 13:56:21 :-/ Jan 12 13:56:54 * cyrilRomain tries to build screen right now Jan 12 13:57:25 <_diego__> but i can't be sure that i hadn't screw something just to avoid the glibc-intermediate build problem ^^ Jan 12 14:12:12 whoaho ! parsing files and building from qemu and through sshfs is awfully slow! Jan 12 14:12:28 one have to be really crazy to do that though ! ;-p Jan 12 14:20:04 <_diego__> cyrilRomain, does screen builds correctly for you? Jan 12 14:20:41 _diego__: I need to wait glibc-intermediate to build beforehand before I can tell you :-/ Jan 12 14:21:44 NOTE: package screen-4.0.2: completed Jan 12 14:21:50 builds for me ,,,, Jan 12 14:23:10 <_diego__> i wonder why when i tried with original .bb, it tried to download it from every damn debian mirror around the world... :-/ Jan 12 14:23:36 heh, I can simply run bitbake shell to test the download Jan 12 14:23:58 Crofton|work: did you removed the patch from your source directory beforehand ? Jan 12 14:25:37 _diego__: OE set a list of debian mirrors to get sources from Jan 12 14:26:46 _diego__: and fall back to angstrom sources if no one provides it Jan 12 14:27:10 _diego__: but if that screen patch has been removed from all the debian mirrors, the screen .bb file should be updated accordingly though Jan 12 14:27:12 <_diego__> can i see that list somewhere? Maybe there was some problem with the fall back.... Jan 12 14:27:26 _diego__: sure, in classes/base.bbclass Jan 12 14:27:51 <_diego__> i've modified .bb locally to get the patch from the local screen_4.0.2 directory Jan 12 14:28:21 <_diego__> and know it's builded fine (at least :) ) Jan 12 14:29:59 <_diego__> btw i've open a ticket on bugzilla about screen (#3663) Jan 12 14:30:23 _diego__: ok Jan 12 14:37:49 cyrilRomain, I just typed bitbake screen Jan 12 14:38:37 Crofton: yeah so it succeed because you already had that patch in your sources directory :) Jan 12 14:39:17 very likely, so a patch has fallen off the internet? Jan 12 14:39:45 Crofton|work: it has actually been renamed by debian dudes Jan 12 14:42:16 _diego__: ok I just bitbake screen Jan 12 14:42:28 I can confirm the patch has been renamed by debian dudes Jan 12 14:42:53 and that bitbake fallback to http://www.angstrom-distribution.org/unstable/sources/screen_4.0.2-4.1.diff.gz Jan 12 14:43:26 03Laibsch 07org.oe.dev * r4989371c... 10/ (10 files in 3 dirs): (log message trimmed) Jan 12 14:43:26 conf/machine/: fix configuration for Strongarm devices Jan 12 14:43:26 * replace erroneous TARGET_CC_ARCH line optimising for xscale instead of strongarm Jan 12 14:43:26 * remove inclusion of tune-strongarm.inc from collie.conf and replace it with the Jan 12 14:43:26 correct TARGET_CC_ARCH. Collie is the only Strongarm 1110 device in OE and thus Jan 12 14:43:27 needs special treatment Jan 12 14:43:32 * rename tune-strongarm.inc to tune-strongarm1100.inc and leave a short note to Jan 12 14:44:25 <_diego__> ok Jan 12 14:47:36 <_diego__> thanks to all ^^ Jan 12 14:51:39 cyrilRomain, what does this patch do? Jan 12 14:51:45 Crofton: dunno Jan 12 14:57:22 Crofton: it fix some autotools thingy and add debian specific init script and conf files Jan 12 14:57:51 ok Jan 12 14:58:24 I'll go ahead and look at the updated/renamed patch today sometime Jan 12 14:58:31 I actually use screen :) Jan 12 14:58:37 me too Jan 12 14:59:00 do you have a url for the new patch location? Jan 12 14:59:04 but in addition to update the 4.0.2, one might try to add the recipe for 4.0.3 Jan 12 14:59:24 Crofton: I'm going to create a patch and post it to bug #3663 Jan 12 14:59:35 I don't see the source for 4.0.3 on the FSF mirror Jan 12 14:59:40 ok Jan 12 14:59:45 that would be great Jan 12 14:59:57 where is the 4.0.3 source? Jan 12 15:00:08 Crofton: you can already tweak locally the .bb by s/screen_4.0.2-4.1.diff.gz/screen_4.0.2-4.1sarge1.diff.gz Jan 12 15:00:11 I can run push the patches when you are ready Jan 12 15:00:42 I'm doing other work today, but I'd be glad to look over and push what you come up with Jan 12 15:00:42 ok but moment, I'm back in few minutes Jan 12 15:01:47 <_diego__> looks like 4.0.3 is not an official gnu release Jan 12 15:05:22 lets stay with what we have for now Jan 12 15:06:00 searching for screen in the bug tracker is not helpful Jan 12 15:06:28 ok, it's rather strange. bitbake fallback to angstrom sources and download it successfully but still returns: Jan 12 15:06:34 NOTE: Task failed: Unknown fetch Error: [Errno 2] No such file or directory: '/home/cyril/pocketsys/oe/sources/archives/screen_4.0.2-4.1.diff.gz.lock Jan 12 15:06:46 well, let's patch screen .bb file first Jan 12 15:07:11 I'll try to look why bitbake fail even after successfully download it from angstrom sources Jan 12 15:07:17 latter Jan 12 15:07:27 I see that too - and tsclient also burns Jan 12 15:14:48 glibc-intermediate fails to compile: http://pastebin.com/m5ffdb3b7 what should i do? Jan 12 15:15:31 this is a very recent glibc Jan 12 15:15:38 i recommend building a slightly older one Jan 12 15:21:13 should i type "bitbake opie-image" or should i restart from scratch moving the /tmp and then type "bitbake opie-image" ? Jan 12 15:27:50 Crofton|work: http://bugs.openembedded.org/show_bug.cgi?id=3663 Jan 12 15:29:05 thanks cyrilRomain , I will try it out today when I have a few minutes ... Jan 12 15:29:31 Crofton: thx Jan 12 15:36:55 hmm bitbake code is ok. I'm wondering how it is possible for bitbake to raise the fetch error whereas it just succeed to download the patch Jan 12 15:37:38 probably due to a thread locking mechanism .. Jan 12 15:45:11 Is there a way to figure out if a wifi card had firmware loaded for initialization? Jan 12 15:45:21 Something in /proc or so? Jan 12 15:46:09 Laibsch: all hostap have firmware in ram Jan 12 15:46:16 Laibsch: for angstrom Jan 12 15:46:25 how to test that? Jan 12 15:46:37 Laibsch: you dont need to test, we force it Jan 12 15:46:47 /lib/udev/prism-fw.sh Jan 12 15:46:50 thesing said he does not like even the updated "pccardctl eject" hack Jan 12 15:47:04 XorA|gone: Just a sec Jan 12 15:49:27 but he is right, finding out why hostap drivers dont suspend right would be best, but I dont expect you to do that, so the hack gets the OK from me Jan 12 15:49:57 also if you suspend more than 2 minutes all tcp connections are dead anyway, thats the maximum allowed time with no packets Jan 12 15:50:08 XorA|gone: Yes, I don't think of this as a fix, just a workaround Jan 12 15:50:20 because it dramatcially improves what we have Jan 12 15:50:37 Laibsch: and if it annoys a kernel hacker, all the better :-D Jan 12 15:50:48 yeah, I kind of thought the argument was thin albeit not totally without merit Jan 12 15:50:56 *if* Jan 12 15:51:09 But it would also distract him from the collie work Jan 12 15:51:16 which IMHO is more important right now Jan 12 15:51:27 if machine != collie :-) Jan 12 15:52:36 XorA|gone: So you think I should not bother trying to come up with another test to separate handling of different wifi cards? Jan 12 15:52:41 Laibsch: collie will soon be my only zaurus Jan 12 15:53:03 Laibsch: I think you should make it hostap only, Im pretty sure orinoco works over suspend Jan 12 15:53:15 OK Jan 12 15:53:23 that is fine with me and could easily be added Jan 12 15:53:32 if it doesnt, someone with orinoco card can always file new bug Jan 12 15:53:36 It was just to show a way to handle not only hostap Jan 12 15:53:40 Sure Jan 12 15:53:49 Easy to expand later on if need be Jan 12 15:54:11 Maybe I am lucky and thesing does not have a hostap card and thus he would not be bothered ;-) Jan 12 16:03:20 03c.romain 07org.oe.dev * rebeb2271... 10/ (1 packages/screen/screen_4.0.2.bb): screen_4.0.2.bb : Resolve bugg 3663 by updating patch location. Jan 12 16:03:25 03mickeyl 07org.oe.dev * r3aaa4985... 10/ (1 classes/insane.bbclass): insane.bbclass: fix formatting of fatal include error Jan 12 16:03:32 03mickeyl 07org.oe.dev * r9ee92e85... 10/ (4 files in 2 dirs): xerces-c 2.8.0 add, remove non-working 2.6.0 and 2.7.0 Jan 12 16:03:37 03mickeyl 07org.oe.dev * r94380cc6... 10/ (1 packages/xerces-c/xerces-c_2.8.0.bb): xerces-c 2.8.0 make it compile for the target architecture *cough* Jan 12 16:03:43 03mickeyl 07org.oe.dev * reaee5612... 10/ (1 packages/dtnrg packages/dtnrg/dtn_2.5.0.bb): Jan 12 16:03:43 add dtn, from the Delay Tolerant Networking Research Group Jan 12 16:03:43 yes, this includes python-dtn (of course :) Jan 12 16:03:44 cyrilRomain, can you check I got it right :) Jan 12 16:03:48 03mickeyl 07org.oe.dev * r623fafc1... 10/ (1 packages/dtnrg/dtn_2.5.0.bb): dtn 2.5.0 change package name for python bindings Jan 12 16:03:58 Crofton|work: thank you :) Jan 12 16:06:07 Crofton|work: what about angstrom stable branch ? Jan 12 16:06:30 can you propose it? Jan 12 16:06:32 Crofton|work: I probably should send a mail on the angstrom devel ML Jan 12 16:06:40 yeah Jan 12 16:07:01 Even though it is pretty obvious, we should still show restraint Jan 12 16:07:10 otherwise stable becomes not so stable :) Jan 12 16:22:13 any bitbake hacker ? I'm trying to understand how http://rafb.net/p/C0mjSl64.html is possible Jan 12 16:30:02 I am not a bitbake hacker. but I saw what sometimes the fetch task keep running when I try to control-c out of bitbake. Jan 12 16:30:30 keesj: it is because bitbake use thread ;) Jan 12 16:32:49 keesj: but your comment is valid though. There is probably good reasons for bitbake to not kill all the thread it started .. Jan 12 16:33:29 (actually bitbake doesn't use threads but processes :) Jan 12 16:33:56 Henryk: indeed, thanks to correct me :) Jan 12 16:35:15 perhaps it is hard to do Jan 12 16:37:00 ah, I probably found what's wrong Jan 12 16:38:25 RP: ping Jan 12 16:47:58 Does anyone know how to change classes dir from relative path to absolute? Jan 12 16:48:03 (in bitbake) Jan 12 16:52:47 SyNko: why would you do that ? Jan 12 16:53:25 cyrilRomain: Hey, isnt correct to answer whit a qwestion Jan 12 16:53:36 SyNko: in python you can os.path.abspath(aRelativPath), but I wouldn't change how bbclasses are handled Jan 12 16:53:57 cyrilRomain: but nvm, i need that because without i cannot continue with bb compiling Jan 12 16:54:11 SyNko: set the BBPATH correctly? Jan 12 16:54:19 cyrilRomain: i got bb working, but no at all Jan 12 16:54:24 cyrilRomain: yes Jan 12 16:55:27 cyrilRomain: i need to set all absolute path , for my environment ;) so ive changed all the paths but the last is where is located the classes path, it still in relative like "classess/.." Jan 12 16:56:06 ) zecke: BBPATH=$BASAL_PATH/OE/:$BASAL_PATH/OE/build/:$BASAL_PATH/OE/org.openembedded.angstrom-2007.12-stable/ Jan 12 16:56:50 zecke: i think yes, cause al works now , only classes are wrong Jan 12 16:56:58 SyNko: Without access to your machine I can't verify...... Jan 12 16:57:05 SyNko: did you export the variable? Jan 12 16:57:35 zecke: would you like to make a ssh conn? Jan 12 16:57:46 no Jan 12 16:58:03 zecke: yes, all exported weel Jan 12 16:58:37 SyNko: what bitbake is doing is checking if: $BASAL_PATH/OE/classes/foo.bbclass or $BASAL_PATH/OE/build/classes/foo.bbclass or ... exists Jan 12 16:59:01 SyNko: so if you want to make loading the classes use absolute paths, you do something wrong Jan 12 16:59:03 zecke: if i use the bb shell i see only SRC_URI with a rel path(but i think it doesnt count Jan 12 17:00:23 SyNko: if I look out of the window I see stars, but this has nothing todo with loading of bbclasses ;) Jan 12 17:00:44 SyNko: you didn't export BBPATH, or it is wrong. Use strace to proof me wrong Jan 12 17:01:33 zecke: k , im strace. But i post the recursive error that exit for all the packages :: ERROR: Could not inherit file classes/autotools.bbclass while parsing /home/SyNko/slugga_comp/anstrom/OE/org.openembedded.angstrom-2007.12-stable/packages/automake/automake_1.9.6.bb Jan 12 17:02:25 zecke: Hey wait a mo', wtf is strace? :) Jan 12 17:02:47 zecke: i dont have it, neither on binary repo Jan 12 17:03:07 SyNko: is your BBFILES set correctly ? i.e. looking into org.openembedded.angstrom-2007.12-stable/packages/*/*.bb and not org.openembedded.dev/packages/*/*.bb ? Jan 12 17:03:28 i pastebin again all Jan 12 17:03:33 wait a ssec Jan 12 17:03:42 SyNko: boy, believe me your BBPATH is wrong or not exported Jan 12 17:06:16 zecke: but bb finds the bb files itself, how is it possible if the BBPATH is wrong ? Jan 12 17:07:29 SyNko: magic? Or because BBFILES != BBPATH? Jan 12 17:08:01 SyNko: BBFILES is a "glob" and you point it to files Jan 12 17:08:13 SyNko: BBPATH is a PATH, yeah like the $PATH variable in your shell... Jan 12 17:08:35 pls, check this out :: http://pastebin.com/d3b6268e2 Jan 12 17:09:28 zecke: imcheck b=the bbfat agai Jan 12 17:11:21 SyNko: why do you set SRC_URI to point to bitbake.conf? But my last sentence on this topic. Your BBPATH is wrong or not exported ;) Jan 12 17:12:20 SyNko: or you don't have a autotools.bbclass... or rm -rf your classes/ directory but all of these issues can only be found out with access to your system and you happen to be in this position Jan 12 17:13:11 SyNko: no idea, you might some more "user" orientated help in #OpenSlug but I'm sure I'm right ;) Jan 12 17:13:38 zecke: XD oK ! Jan 12 17:14:21 SyNko: source your script and do echo $BBPATH Jan 12 17:14:54 SyNko: then use the mouse and select everything until the first ':', ls this path? does it exist? then go to the next? Jan 12 17:15:21 SyNko: and if every directory exists and at least one directory has a classes/autotools.bbclass then talk to me again Jan 12 17:16:32 ) zecke: wow .o/ im gonna check the path issues Jan 12 17:20:33 zecke: all the paths exist like i exist Jan 12 17:21:46 zecke: but if i make echo BBPATH after source dont out nothing!! Jan 12 17:21:58 hi zecke Jan 12 17:22:26 * SyNko start feeling dumb Jan 12 17:23:13 SyNko: maybe do you have a tcsh shell or another shell that dislike the export command, and need to use setenv instead ? (but 'export' would raise an error then) Jan 12 17:24:13 SyNko: how to you call your source.sh ? 'source source.sh' or './source.sh' ? Jan 12 17:24:35 SyNko: you need to source it Jan 12 17:26:52 zecke: btw since you're here :) I got a strange fetch failure with bitbake : http://rafb.net/p/BvwqQV39.html Jan 12 17:27:19 zecke: I tracked down the problem: in fetch/wget.py:70 Jan 12 17:27:39 zecke: of course s/not// Jan 12 17:28:35 zecke: No such file or directory: '/home/cyril/pocketsys/oe/sources/archives/screen_4.0.2-4.1.diff.gz.lock' Jan 12 17:28:52 zecke: soory , remove the last Jan 12 17:29:11 cyrilRomain: No such file or directory: '/home/cyril/pocketsys/oe/sources/archives/screen_4.0.2-4.1.diff.gz.lock' Jan 12 17:29:34 cyrilRomain: hey , check the papappaaatthh Jan 12 17:30:32 SyNko: I think you started me laughing Jan 12 17:33:24 cyrilRomain: :) Jan 12 17:34:09 SyNko: I found what's wrong with your export Jan 12 17:34:30 SyNko: you have to 'export VARIABLE=VALUE' Jan 12 17:35:03 SyNko: setting `TOTO=bla` `TATA=tyu` and then 'exporting TOTO TATA' does work Jan 12 17:35:59 SyNko: I think command called with `` gets a new variable environement Jan 12 17:36:34 SyNko: so use 'export BASAL=...' 'export BBPATH=...' instead Jan 12 17:38:29 SyNko: tell us if it helped Jan 12 17:42:39 cyrilRomain: im using a korn hell, so i prefer to (empiric think) use both the ` and normal command, but it's not a prob.. in the final echo i can see the var pointing in the right path Jan 12 17:43:10 cyrilRomain: but im putting the ' too ;) Jan 12 17:43:28 cyrilRomain: megaman script path Jan 12 17:49:58 SyNko: all the line with `` can be safely removed, and btw your angstrom directory is called 'anstrom' Jan 12 17:50:59 SyNko: eheh, sorry but im on the run.. and what bout the bitbake.conf file, i never touch it Jan 12 17:51:22 cyrilRomain: eheh, sorry but im on the run.. and what bout the bitbake.conf file, i never touch it Jan 12 17:52:40 SyNko: if you start to s/anstrom/angstrom/ do it every where it is needed, including your local.conf file Jan 12 17:56:06 cyrilRomain: this is the correct path :: /home/SyNko/slugga_comp/anstrom/OE/org.openembedded.angstrom-2007.12-stable this :: imnot an idiot(i hope so) so dont give me that useless advices Jan 12 17:59:26 cyrilRomain: ive added BBPATH it to local.conf, it works!!! humm but make some mistake with the cache path Jan 12 18:03:07 03mickeyl 07org.oe.dev * re18d2142... 10/ (3 files in 3 dirs): add task-openmoko-python-devel and corresponding openmoko-python-devel-image Jan 12 18:07:47 cyrilRomain: i discover that my BB doesnt work without CACHE var, in the local.conf is setted with ="" for me is like to keep it unsetted, if i point that var to the root distro is it ok? Jan 12 18:10:49 hello i have a packaging faillure...http://pastebin.com/m19d578da Jan 12 18:11:10 s/packaging/do_package:/ Jan 12 18:13:36 SyNko: it is not recommended to set CACHE = "", until your BBPATH is wrongly set (as in your bitbake -i environement stamp where BBPATH="/usr/local/share/bitbake"), you will have problem to build things Jan 12 18:16:15 GNUtoo: you set the glibc build binary locale option (don't remember the exact name). and wrongly so. Jan 12 18:16:48 Henryk, ah "en" is wrong? Jan 12 18:17:40 yes. that's a real unfortunate thing. the format is not clearly documented and all google will get you is people yelling "stop using the wrong format". Jan 12 18:17:51 GNUtoo: I just looked up what I'm using: GLIBC_GENERATE_LOCALES = "en_GB.UTF-8 en_GB.ISO-8859-1 en_GB.ISO-8859-15" Jan 12 18:18:12 Henryk,ok thanks a lot Jan 12 18:19:50 zecke: what do you think about fetch/wget.py:70,71 ? Jan 12 18:21:57 cyrilRomain: Have not looked into it, need to fix a webkit bug first :) Jan 12 18:22:08 zecke: ok Jan 12 18:23:07 cyrilRomain: pong Jan 12 18:23:38 RP: hi RP ! I'm trying to understand how http://rafb.net/p/BvwqQV39.html is possible Jan 12 18:24:54 I thought I found the problem in wget.py:70,71, but now I think the problem is somewhere else Jan 12 18:25:16 cyrilRomain: I think the bb.utils.lockfile and unlockfile have a race somehow Jan 12 18:25:50 RP: maybe the wget.py:71 should not remove .lock file ! Jan 12 18:27:22 cyrilRomain: That rm one line 71 is not to do with the lockfile Jan 12 18:27:31 s/one/on/ Jan 12 18:28:59 hi Jan 12 18:29:15 ok, I thought it could be because lockfile is nammed localpath + '.lock' Jan 12 18:29:46 at the same time, I get 'cannot remove ..screen_4.0.2-4.1.diff.gz*': No such file or directory, so you're right Jan 12 18:29:53 hi likewise Jan 12 18:30:30 cyrilRomain: ah, hmm, I didn't think about it like that Jan 12 18:32:51 Do we still get the sourceforge mirror errors? Jan 12 18:33:18 no idea Jan 12 18:33:53 RP: remove that code if it is causing troubles Jan 12 18:34:15 zecke: I think that is going to be the saftest option Jan 12 18:34:25 The return code from wget must have been wrong too though Jan 12 18:34:52 returning success when it blatently didn't get the file :/ Jan 12 18:35:05 loser :( Jan 12 18:39:51 returning ret==0 should be enough, or even 'return os.path.exists(ud.localpath)' Jan 12 18:40:35 * cyrilRomain don't know about the sourceforge error though Jan 12 18:40:55 03rpurdie * r997 10/ (4 files in 4 dirs): wget.py: Change the wget fetcher failure handling to avoid lockfile problems Jan 12 18:41:05 cyrilRomain: Try that ;-) Jan 12 18:41:06 ERROR: Unable to parse conf/bitbake.conf (Could not include required file /home/SyNko/slugga_comp/anstrom/cache/conf/distro/include/sane-srcdates.inc) Jan 12 18:41:11 cyrilRomain: back in 2006 they had a nice game. Instead of serving you a file they served you an advertising site :) Jan 12 18:41:33 zecke: ;-/ Jan 12 18:41:40 RP: thanks Jan 12 18:43:11 SyNko: a small hint. Instead of trying to jump through rings of fire, do it simple and by hand for the first time? Jan 12 18:43:19 zecke: With the checksumming we now do, corrupt files should be easier to spot :) Jan 12 18:43:35 SyNko: set BBPATH by hand, fill in the conf/local.conf (four lines) and execute bitbake, done Jan 12 18:44:07 RP: yeah, I had the same thought. I think with wget failing the sf.net way we just tried another mirror Jan 12 18:44:28 RP: now we will see a user joining this channel, but I want to wait until this happens :) Jan 12 18:47:31 RP: I'll let you know if that fix the issue in a couple of minutes (waiting for debian mirror failure and timeouts ...) Jan 12 18:47:38 zecke: ok , im follow you Jan 12 18:47:55 zecke: * i follow you now Jan 12 18:48:50 SyNko: the info is in the wiki... GettingStarted or such Jan 12 18:49:07 SyNko: that page has been written for: http://www.angstrom-distribution.org/building-angstrom Jan 12 18:49:51 zecke: hey, do you still thinking shht that ive not read that page? Jan 12 18:50:18 zecke: if there was something to read ive allready read all Jan 12 18:50:34 SyNko: then read it again Jan 12 18:51:01 SyNko: if you want to work with .dev, look at http://www.openembedded.org/wiki/GettingStarted Jan 12 18:51:16 SyNko: if you want to work woth angstrom stable branch, look at http://www.angstrom-distribution.org/building-angstrom Jan 12 18:51:19 zecke: i eat the infos, but i cannot undestand very well the BB's fashion src of the client, so im stuck here waiting Jan 12 18:51:24 SyNko: my local.conf is four lines, BBFILES, MACHINE, DISTRO and disabling of binary locales Jan 12 18:51:52 zecke: k , i take a deep breath and read again Jan 12 18:52:08 SyNko: the only environment variables I set are PATH for bitbake/bin and BBPATH Jan 12 18:52:43 zecke: instead ive the BB istalled as a normal command in /usr/.. Jan 12 18:53:25 SyNko: so, only set BBPATH Jan 12 18:54:00 SyNko: write your local.conf from scracth Jan 12 18:54:04 SyNko: BBFILES := "/home/SyNko/slugga_comp/anstrom/OE/org.openembedded.angstrom-2007.12-stable/packages/*/*.bb" Jan 12 18:54:09 SyNko: make this your first line Jan 12 18:54:24 DISTRO = "angstrom-2007.1" Jan 12 18:54:33 MACHINE = "nslu2le" Jan 12 18:55:08 set the BBPATH=MY_BUILD_DIR:MY_OE_DIR, MY_BUILD_DIR/conf/local.conf should exist, MY_OE_DIR/conf/bitbake.conf should exist as well Jan 12 18:55:18 bitbake nano should parse and build stuff Jan 12 18:57:39 zecke: k Jan 12 18:57:47 no idea if your machine is correct :) Jan 12 19:01:31 zecke: the MACHINE means the target machine, right? Jan 12 19:02:55 yes Jan 12 19:03:11 SyNko: I'm just not up-to-date regarding the slug devices Jan 12 19:04:00 zecke: do you read the source of the bitbake? Jan 12 19:04:10 zecke: (client) Jan 12 19:04:29 sadly not recently Jan 12 19:05:07 zecke: cause i spent a hour recently, but im blocked on some steps Jan 12 19:05:19 zecke, have you built qtopia in OE? Jan 12 19:05:59 SyNko: It is unlikely that you have to read bitbake source for your first build Jan 12 19:06:10 Crofton: QtopiaCore yes, Qtopia no Jan 12 19:06:29 Crofton: but it is certainly on my list of things todo Jan 12 19:06:34 the gumstix crowd seems interested Jan 12 19:08:13 RP: problem solved, thanks :) Jan 12 19:08:50 zecke: all the unlikely or not dont touch me, im an I.T. technician / h4x ready to collaborate and dev my future with no fkn walls.. today ive discovered that on bsd works only the absolute paths, hey no probs, where is also the classes path var? (im thinking to read the code again) Jan 12 19:09:54 SyNko: you don't make any sense Jan 12 19:10:09 zecke: because your faulty system damage my brain and my methods Jan 12 19:10:44 lol Jan 12 19:11:05 zecke: im out only for the classes path, with my first sets i get the bb works Jan 12 19:11:56 zecke: it finds all the bb files, but when it search for classess path an error occour Jan 12 19:12:01 SyNko: no idea what you means with classes path? reread this conversation and the analogy of BBPATH with PATH (known from your favorite BSD shell) Jan 12 19:13:10 SyNko: You can properly export/set your BBPATH, I will not collaborate in stupidity Jan 12 19:13:20 17:58:45) SyNko: zecke: k , im strace. But i post the recursive error that exit for all the packages :: ERROR: Could not inherit file classes/autotools.bbclass while parsing /home/SyNko/slugga_comp/anstrom/OE/org.openembedded.angstrom-2007.12-stable/packages/automake/automake_1.9.6.bb Jan 12 19:13:37 zecke: k Jan 12 19:19:08 zecke: im again on the code, definitly because your changes to my local.conf give me more problems, tnx, bye Jan 12 19:25:26 Hrrm I am at the OE stage where I ca build angstrom 2008.1 for arm using "bitbake vim" or alike Jan 12 19:26:06 however I dont get an arm disk image with standard Linux utils df, du, cp mv etc. which I can deploy Jan 12 19:26:42 is that done by bitbake gpe-image? Jan 12 19:27:00 pto: yes for instance Jan 12 19:27:27 What is bitbake opie-image compared to gpe-image? Jan 12 19:27:41 pto: you can take a look at packages/images/ to see what are the available image Jan 12 19:32:38 cyrilRomain: found the dir, but how do I understand the content of each target? Jan 12 19:33:16 e.g. maemo - that I get (I was at Nokia for 9,5 years)... Jan 12 19:33:57 pto: an image includes a set of .bb if you like Jan 12 19:34:01 nylon-image-* - no clue :) Jan 12 19:34:23 so gpe-image is an image with tools based on the gtk GUI toolkit Jan 12 19:34:40 cyrilRomain: get that idea, but how do I get a clue to the contents of e.g. maemo Jan 12 19:34:45 console-image is an minimaliste image with only console tools Jan 12 19:34:55 x11-image is an image with an X server Jan 12 19:34:58 Guess that there is a command line command to see this Jan 12 19:34:59 etc. Jan 12 19:35:07 or a webpage Jan 12 19:35:34 and a clue on the target image size would be nice Jan 12 19:35:46 order of magnitude is fine Jan 12 19:37:43 pto: I don't know if there is documentation about the content of the available image Jan 12 19:38:18 at least I didn't find such documentation Jan 12 19:38:35 pto: I usually look into the images .bb files to know what is included Jan 12 19:38:46 any ideas about this OE crash ? http://www.pastebin.ca/852329 Jan 12 19:39:06 pto: especially look at the IMAGE_INSTALL variable Jan 12 19:39:07 the machine froze, when doing a pull /update Jan 12 19:39:25 looks like a new install is the only way out here Jan 12 19:39:55 ERROR: Could not inherit file classes/rootfs_ Jan 12 19:42:38 cyrilRomain: In reality I need to build a normal mini-image for arm where I have the basics - kdrive xserver + qt4 Jan 12 19:42:48 and tslib Jan 12 19:43:24 pto: which is not so mini ;) Jan 12 19:43:43 pto: customize either console-image or x11-minimal-image (I would go with the console-image) Jan 12 19:44:10 cyrilRomain: he he - I agree :) Jan 12 19:44:20 interfaith: you probably forgot to uncomment 'INHERIT = "package_ipk"' or 'INHERIT = "package_tar"' in your local.conf file Jan 12 19:44:32 zecke: customize - is there some description of that? Jan 12 19:44:56 interfaith: or 'INHERIT = "package_deb"' Jan 12 19:46:14 so I dont get the coupling between "bitbake qt4-core" -> which builds 1000+ packages Jan 12 19:46:29 pto: Looks like you have an oppurtunity to contribute to our manual (http://www.openembedded.org/user-manual&dpage=commonuse_new_image) Jan 12 19:46:31 and the bitbake x11-minimal-image Jan 12 19:47:46 pto: Best thing is to look into the BitBake manual and findout what tasks and providers are Jan 12 19:47:52 zecke: given that I have written 1000+ pages so far at www.linuxbog.dk, if I understand how to do it - then I would happily document it Jan 12 19:48:24 pto: I'm searching for easy words. In OE every .bb file is a provider for something Jan 12 19:48:56 pto: this something depends on the classes it inherits and which tasks it has (e.g. bitbake -clisttasks -b /path/to/x11-minimal-image.bb) Jan 12 19:49:22 pto: e.g. witll bitbake nano try to create packages for nano and on its way compile all dependencies Jan 12 19:50:03 pto: bitbake x11-minimal-image will try to create an image (by installing packages to a dir and then creating a filesystem), and on its way it needs to make sure that the packages it want to install got created Jan 12 19:50:42 pto: all of our images should DEPEND on one task (e.g. task-boot) and then install this task Jan 12 19:51:27 pto: so customising is happening on two points. 1) create your task on where you define build dependencies and say which packages you want to install (one .bb file/Provider can create zero to many packages) Jan 12 19:52:19 pto: the 2nd step is to install your task. And the best thing is to look at an example, hence the reference to console-image or x11-minimal-image found in the packages/images directory. x11-minimal-image is a bit misleading as it has a visual login program requiring Gtk+ (at least the last time I checked) Jan 12 19:53:21 it worked before, i will try that thx ! Jan 12 20:00:22 that change is working , amazing ~! Jan 12 20:01:45 anyone had luck with snmp module development under oe ? Jan 12 20:06:07 asterisk shows in the tree, but not zaptel ,which is the driver for asterisk ? Jan 12 20:06:56 perhaps someone has done zaptel now ? Jan 12 20:08:52 interfaith: for now in OE asterisk is built without the zaptel support Jan 12 20:40:18 any idea about this error Jan 12 20:40:38 ]ERROR: Error Method already seen: python_dir in' distutils-base.bbclass' now in 'dtn_2.5.0.bb' Jan 12 20:40:54 must some bb file be removed ? Jan 12 22:16:09 error ?ERROR: Error Method already seen: python_dir in' distutils-base.bbclass' now in 'dtn_2.5.0.bb' Jan 12 22:16:20 some bb file must be removed ? Jan 12 22:17:59 hello, i have a problem to report with the popt package...its repository can't be reached... Jan 12 22:23:48 GNUtoo: someone posted popt-1.10.4 and popt-native-1.10.4 recipe http://bugs.openembedded.org/show_bug.cgi?id=3408 (seemingly because of the non reliable SRC_URI) Jan 12 22:24:00 GNUtoo: didn't tested though Jan 12 22:33:35 sorry for beeing away but i had an electricity problem Jan 12 22:35:17 23:29 < cyrilRomain> GNUtoo: someone posted popt-1.10.4 and popt-native-1.10.4 recipe http://bugs.openembedded.org/show_bug.cgi?id=3408 (seemingly because of the non reliable SRC_URI) Jan 12 22:35:20 23:29 < cyrilRomain> GNUtoo: didn't tested though Jan 12 22:36:51 mickey|dinner: ERROR: /home/slug/slugos/openembedded/packages/dtnrg/dtn_2.5.0.bb:40: unparsed line: 'export BUILD_SYS' while parsing /home/slug/slugos/openembedded/packages/dtnrg/dtn_2.5.0.bb Jan 12 22:37:33 interfaith: I'm having trouble parsing that package too - I've pinged mickey|dinner Jan 12 22:38:39 mickey|dinner: looks like the problem is the line above that one: PYTHON_DIR = "${@python_dir(d)}" Jan 12 22:39:11 perhaps then an mtn update will fix it soon Jan 12 22:39:30 interfaith: only if you see CIA-35 report a fix by mickey|dinner Jan 12 22:46:01 ? is snmp doing ok on oe Jan 12 22:46:41 rwhitby: btw is there on-going effort to makes nslu2 distros building with OE ? Jan 12 22:47:17 what do i do if ftp://ftp.rpm.org is down and a package depends on it's availability in order to be downloaded Jan 12 22:47:44 interfaith: if you delete lines 30 to 39 from packages/dtnrg/dtn_2.5.0.bb, it should parse - those lines are already in distutils-base.bbclass, so I don't know why mickey|dinner has them in the recipe as well. Jan 12 22:47:45 rwhitby: last time I wanted to build the last monotone for a Synology (ds101) and I fail to build one, even after tweaking a bit the .mk :-/ Jan 12 22:47:54 cyrilRomain: they already do. what's the question? Jan 12 22:48:28 (Unslung, SlugOS and Angstrom all support the NSLU2 and are built with OE) Jan 12 22:48:41 rwhitby: cool :) Jan 12 22:48:55 that's been the case for almost 2.5 years now ... Jan 12 22:48:57 rwhitby: should I use the org.nslu2-linux.dev branch for that ? Jan 12 22:49:32 amazing the latest snmp bb file fails Jan 12 22:49:39 Build of 'net-snmp' failed Jan 12 22:50:19 cyrilRomain: Depends whether you want to use the nslu2-linux Master Makefile, or set up your OE build area manually. That branch you mentioned *not* related org.openembedded.dev, it is the nslu2-linux Master Makefile files. Jan 12 22:51:12 FATAL: oe_runconf failed , has anyone had snmp build ? Jan 12 22:51:17 rwhitby: I tried the MasterMakefile way, without much success Jan 12 22:51:45 cyrilRomain: if you want help with that, then go to #nslu2-linux - it works for many other people. Jan 12 22:51:55 rwhitby: ok thanks Jan 12 22:52:05 rwhitby: I'll have a deeper look into that =) Jan 12 22:52:06 (including the autobuilder which builds all the binary images and packages) Jan 12 22:53:03 nslu2-linux has a policy of not releasing a binary image until it can be built by a third part from publicly accessible sources and build scripts. Jan 12 22:53:18 s/part/party/ - i.e. not one of the core team members Jan 12 22:53:50 rwhitby: interesting :) Jan 12 22:54:20 this guarantees buildability, and GPL compliance at all times. Jan 12 22:55:39 and we snapshot the OE metadata into an svn repository for each binary release, so rebuilding those is impervious to OE metadata changes. Jan 12 22:56:28 mickey|dinner: should the line be something like: export BUILD_SYS = "${BUILD_SYS}" Jan 12 22:56:28 ? Jan 12 22:56:52 hmm - no, that gives an error too. Jan 12 22:57:30 where are the details for the make ./configure commands ? in bb files ? Jan 12 23:10:30 i guess one can glean details of the ./configure command to build packages on other distro's ? Jan 12 23:11:17 as the bb files have all the details for the critical .configure Jan 12 23:13:57 interfaith: in .bb files you can either inherit class that help you building the package (e.g. inheriting autotools will call ./configure, make, and make install) Jan 12 23:14:29 interfaith: or do it yourself (for instance if the package do not use autotools but only hand-made makefiles) Jan 12 23:15:08 interfaith: ./configure and make logs are found in tmp/work/blabla/package/temp/ directories Jan 12 23:15:44 interfaith: hope it helped because I'm not sure I understood your question Jan 12 23:16:39 interfaith: removing the two export lines from dtn_2.5.0.bb will allow the build to continue - dunno whether the dtn package will build, but 99.99% of people won't care about that compared to not having all their OE builds broken. Jan 12 23:18:47 ok thx ! Jan 12 23:18:58 the autotools class also runs autoconf, automake, etc Jan 12 23:19:22 this is useful if you need to edit configure.ac in big ways, but also easily breaks the build Jan 12 23:19:32 RP: is there a technical reason why bitbake completely stops when it gets an error in a single recipe, rather than just ignoring that recipe and doing everything else? Jan 12 23:22:28 rwhitby: IMHO it would be safe to ignore bad 'leaf', but it is wise to stop the build for .bb or .inc that are included by other one Jan 12 23:23:34 needs more trickery to collect all the failure outputs and show them at the end, I guess Jan 12 23:23:44 the current behaviour makes OE very brittle in the face of errors in random .bb recipe files Jan 12 23:23:57 failures could also depend on each other in ways that bitbake cannot understand Jan 12 23:24:32 freebsd's portupgrade has the behaviour you describe, though, and it's useful there Jan 12 23:24:58 it causes it to upgrade as much as possible in the face of errors Jan 12 23:26:09 bitbake knows the dependency graph, so an error in one package could easily stop the build of dependent packages Jan 12 23:26:17 (and probably will by default anyway) Jan 12 23:27:01 portupgrade does that too Jan 12 23:31:01 rwhitby: The old logic meant the error message would scroll away and users would ignore it Jan 12 23:31:26 rwhitby: bitbake might then take a different path to the one expected and people get upset when that happens... Jan 12 23:31:52 RP: hmm - pros and cons as always. c'est la vie. Jan 12 23:32:16 rwhitby: I got sick of people not reading error messages... Jan 12 23:34:17 wow ! oe really helps a lot to cross compile by gleaning the details from the bb files @! Jan 12 23:34:47 try cross compiling snmp from scratch ! for another distro in my case montavista Jan 12 23:35:30 it is not a simple thing to build a new package for snapgear or montavista..but using the bb files gives most of it ! Jan 12 23:35:53 interfaith: We try our best ;-) Jan 12 23:35:59 thx ! Jan 12 23:36:19 really amazing ! Jan 12 23:36:38 it floors me :) Jan 12 23:36:44 oe has POWER Jan 12 23:37:07 hallelujah! Jan 12 23:39:01 u got that right ! Jan 12 23:39:36 zaptel from asterisk is no easy trick to cross compile Jan 12 23:39:50 but with all the info from bb files , it may happen ! Jan 12 23:40:17 zaptel here we come ! tdm over ethernet ya! Jan 12 23:40:24 I'm not sure I've ever seen OE fail at cross compiling a package if its cross compilable! Jan 12 23:40:57 that sounds good Jan 12 23:41:20 isn't that kinda true by definition? :p Jan 12 23:41:52 any url for newbies to add a package ? Jan 12 23:42:14 jilles: Not all build systems can build all software ;-) Jan 12 23:42:52 it all depends how much you want to hack it Jan 12 23:43:54 jilles: OE doesn't usually end up hacking software too much... Jan 12 23:47:21 * RP -> Zzzz Jan 13 00:02:32 03osas 07org.oe.dev * rf7c4009c... 10/ (1 packages/meta/slugos-packages.bb): enable tcpdump for sluos Jan 13 00:26:12 what is wrong Jan 13 00:26:14 NOTE: Resolving missing task queue dependencies Jan 13 00:26:14 NOTE: multiple providers are available for virtual/libc (external-toolchain, eglibc, glibc); Jan 13 00:26:14 N Jan 13 00:27:01 NOTE: consider defining PREFERRED_PROVIDER_virtual/libc Jan 13 00:27:08 a problem in local.conf ? Jan 13 00:44:10 here oe builds slugos ok but fails angstrom ? Jan 13 00:56:23 03osas 07org.oe.dev * r6ba62f12... 10/ (8 files in 3 dirs): asterisk: 1.4.11 -> 1.4.17 Jan 13 00:56:28 03osas 07org.oe.dev * re487928e... 10/ (25 files in 14 dirs): add asterisk sound files for 1.4 version Jan 13 01:10:14 wow using the smarts from oe ,the new snmp builds and runs on other distros like snapgear ,..ho ho Jan 13 01:22:43 now perhaps oe can cross compile zaptel the gears of asterisk Jan 13 02:07:51 hi all - i'm just doing a fresh build from scratch of openprotium, and for some reason, update-rc.d is pulling in update-rc.d-dev, and thus libc-dev, and so forth. yet, update-rc.d hasn't changed in a while. Jan 13 02:07:56 any idea what gives? Jan 13 02:33:13 does anyone knows what means empty string in http://toh.openwrt.org/, column "Status" ? **** ENDING LOGGING AT Sun Jan 13 02:59:57 2008