**** BEGIN LOGGING AT Thu Aug 16 02:59:57 2007 Aug 16 05:41:26 exit Aug 16 05:41:50 RP: After your latest changes to bb I get these errors: http://www.acc.umu.se/~cm/bitbake_errors.txt Aug 16 05:42:15 It's on gentoo 64-bit, python 2.4.4 Aug 16 05:56:56 mwester: ping **** BEGIN LOGGING AT Thu Aug 16 06:01:56 2007 Aug 16 06:29:28 RP: Sorry, must have had some cruft around. Removed everything and reinstalled, and now it works fine. Aug 16 06:29:52 * CM hangs head in shame and compiles instead. Aug 16 07:29:55 CM: heh, I have the same problem.. And can't get rid of it :) Aug 16 07:30:18 wrobbie: Using the MokoMakefile? Aug 16 07:30:59 Glad I'm not the only one though. I couldn't get rid of it, had to clean it all out and restart from scratch. Aug 16 07:31:03 No, own distro.. After updating bb to head. Aug 16 07:31:19 Ok Aug 16 07:32:07 Guess something changed internally. All I did was updating bb to head too Aug 16 07:33:18 r899: bitbake-1.8: Backport fixes for bugs from trunk Aug 16 07:33:53 At r896 it works again. Aug 16 07:35:28 Strange that it works when you zap the whole directory and check out from scratch again. Aug 16 07:39:45 The bitbake dir or tmp? Aug 16 07:40:25 I wanted to build OpenMoko, so I cleaned out everything Aug 16 07:40:57 New bitbake and reset all tmp and build dirs Aug 16 07:42:07 I used a new tmp dir and killed *.pyc from bitbake (in case python uses old one, just a try) Aug 16 07:42:38 Still errors. Didn't try actually chucking the bitbake dir though :) Aug 16 07:43:53 I did try removing the build dir and the tmp, but that didn't help. And when I updated from r899 to r900 I started getting those type object 'C' errors too Aug 16 07:48:38 good morning Aug 16 07:59:19 florian : hi Aug 16 08:06:54 good morning all Aug 16 08:07:35 morning koen Aug 16 08:09:05 morning Aug 16 08:09:39 mickeyl: http://rafb.net/p/x8Mxo535.html Aug 16 08:10:32 nice. more bugs closed than opened :) Aug 16 08:11:06 hi Shoragan Aug 16 08:13:04 !ombug 2731 Aug 16 08:13:06 * * Bug 2731, Status: Unknown Error () Aug 16 08:13:09 !oebug 2731 Aug 16 08:13:10 * * Bug 2731, Status: NEW, Created: 2007-08-11 04:57 Aug 16 08:13:11 * * jstoefer(AT)schuett-elektronik.de: efl does not build Aug 16 08:13:12 * * http://bugs.openembedded.org/show_bug.cgi?id=2731 Aug 16 08:13:45 night all Aug 16 08:14:00 uhmm Aug 16 08:14:02 Enlightenment libraries do not buid in familiar tree Aug 16 08:14:04 heh Aug 16 08:15:50 *sigh* Aug 16 08:15:52 Sorry, there are 58 bugs outstanding for this version. You must move those bugs to another version before you can delete this one. Aug 16 08:16:07 how can i mass move bugs to another product version? Aug 16 08:21:02 mickeyl: What do you do with familiar?! Aug 16 08:21:13 aj Aug 16 08:21:20 2731 Aug 16 08:21:22 hehe Aug 16 08:21:43 wow is familiar still going ? Aug 16 08:22:02 ade|desk: slowly, but its svn does get work done Aug 16 08:22:08 florian: nothing. i wanted to revove the product version in bugzilla, but that's not easily possible Aug 16 08:22:12 remove, evne Aug 16 08:22:13 ade|desk: Its almost obsolete, but a well known "brand name". Aug 16 08:23:03 I haven't looked at it for a while, must be about 5 years since it was on my 3630 Aug 16 08:23:06 koen, mickeyl : good morning ! Got some good news today. I have submitted a proposal for an OE article to Linux journal and it was accepted :) Aug 16 08:23:33 steliosk: groovy Aug 16 08:23:33 mickeyl: I bet they refuse to update GPE ;) Aug 16 08:23:51 steliosk: cool Aug 16 08:23:55 congrats steliosk Aug 16 08:24:00 hi zecke Aug 16 08:24:18 paving the road to world domination by OE :) Aug 16 08:25:01 btw has anyone heard that IBM is to buy Windriver ? Aug 16 08:25:24 moin Aug 16 08:25:39 steliosk hm that would be huge deal Aug 16 08:26:42 hi! may I ask some opie-image build problems with bitbake here? Or should I ask at the distro's irc-channel for suppert? Aug 16 08:26:46 florian: why do I get bombed with gpe bugs once a week, shouldnt bugzilla send them out a bit more often? Aug 16 08:26:59 i was at a friends house yesterday and one of the people there is into financial services and said that.... Aug 16 08:30:07 http://pastebin.ca/659431 what happend to bitbake 1.8 branch ? Aug 16 08:32:58 * * OE Bug 2731 has been RESOLVED (WONTFIX) by mickeyl(AT)handhelds.org Aug 16 08:33:00 * *  efl does not build Aug 16 08:33:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2731 Aug 16 08:34:49 can anyone assist me with a bitbake problem on x86_64 architecture? Aug 16 08:36:15 cyberdeck : shoot Aug 16 08:37:10 thank you! while bulding the opie-image i constantly get a "glibc double free corruption". Aug 16 08:37:16 XorA: bugzilla mails end up in the moderation q Aug 16 08:37:35 XorA: so once in a while florian or I let them through Aug 16 08:37:45 koen: :-D Aug 16 08:37:57 i use debian etch under 64 bit and the bitbake supplied with the familiar 0.8.4 build. Aug 16 08:38:23 XorA: something with implicit destination and mailman sucking Aug 16 08:38:37 i wonder if that "double free" is a "well known 64-bit error"? Aug 16 08:39:17 currently i try to get an old 32-bit xen image to boot to run the build process with it. Aug 16 08:39:39 use real bitbake, works fine for me on amd64 on etch Aug 16 08:40:01 morning all Aug 16 08:40:12 morning RP Aug 16 08:40:45 ok. i'll give it a try. Aug 16 08:40:51 RP: any ideas on this bitbake issue with current 1.8 head http://pastebin.ca/659431 Aug 16 08:40:52 XorA: ty! Aug 16 08:41:26 cyberdeck : This message comes from bitbake or from the toolchain ? Aug 16 08:42:08 ade|desk: I got that too, http://www.acc.umu.se/~cm/bitbake_errors.txt Aug 16 08:43:03 ade|desk, CM: tried checking bugzilla? Aug 16 08:43:17 http://bugs.openembedded.org/show_bug.cgi?id=2773 Aug 16 08:43:29 ade|desk: Does that happen if you wipe the cache? Aug 16 08:43:40 RP yes Aug 16 08:43:45 steliosk: I pasted the error over here: http://pastebin.ca/659443 Aug 16 08:44:02 I cleaned out my whole dir, and restarted from scratch, and then it worked Aug 16 08:44:19 CM: Which directory did you clean? Aug 16 08:44:29 According to wrobbie earlier it works if you go back to r896 Aug 16 08:45:00 RP: All of it. I was using the MokoMakefile, so I started from scratch again Aug 16 08:45:18 I mean clean checkout of bitbake, new OE.mtn and rebuild all Aug 16 08:46:03 RP: FYI, there is an alpha MokoMakefile for OM-2007.2 which uses the tip of the 1.8 stable branch for bitbake. Aug 16 08:46:25 (and no patches applied to it) Aug 16 08:46:31 (and no openmoko svn overlay) Aug 16 08:46:45 drat, the bf538 board lacks ethernet Aug 16 08:47:04 CM is using that alpha MokoMakefile. Aug 16 08:47:07 XorA: I just tried to convince mailman to behave a litte bit nicer :-) Aug 16 08:47:11 cyberdeck : weird. Clean your tmp, get a new version of bitbake and try again Aug 16 08:47:38 the same bitbake (rev 900) is working fine here on Debian Etch 32 bit python 2.4.4 Aug 16 08:48:29 rwhitby : not sure if its a bitbake issue or one of the tools it uses Aug 16 08:48:30 RP: proof, removed cache and re-run bitbake http://pastebin.ca/659446 Aug 16 08:48:52 this is on debian sid Aug 16 08:49:19 steliosk: nod. just giving data points. Aug 16 08:49:21 python 2.5.1 Aug 16 08:49:37 rwhitby : if its a tool issue it should occur with the new bitbake also Aug 16 08:49:39 and 2.4.4 Aug 16 08:49:47 ade|desk: I believe you :) Aug 16 08:49:55 :) Aug 16 08:50:14 zecke: I added a renameVar function, it seems to break for some people though :-( Aug 16 08:50:53 This bitbake revision works fine for me too... Aug 16 08:51:03 RP: in which way does it break? Aug 16 08:51:08 the above pastebin? Aug 16 08:51:13 yup Aug 16 08:51:18 zecke: http://pastebin.ca/659431 Aug 16 08:51:44 ade|desk: which version of python? Aug 16 08:51:48 zecke, RP: http://bugs.openembedded.org/show_bug.cgi?id=2773 Aug 16 08:51:54 2.5.1 Aug 16 08:52:59 removed 2.5.1 for now , trying with 2.4.4 Aug 16 08:53:19 I think I take off from work today :} Aug 16 08:53:26 I have a 2.5.1c1 which works here... Aug 16 08:53:51 and a 2.4.1 Aug 16 08:54:11 I think this was mithro's code and I will need to read on metaclasses again Aug 16 08:55:32 same error with python 2.4.4 Aug 16 08:55:49 ade|desk: with psyco, without? Aug 16 08:56:02 (the -d option disables) Aug 16 08:56:09 I'm without psyco on 64 bit with python 2.4.4 Aug 16 08:56:36 * CM sneaks off to a meeting Aug 16 08:56:43 I'm only on 32 bit. I wonder if its 64 related... Aug 16 08:56:52 i'm only 32 bit Aug 16 08:57:05 trying with -d now Aug 16 08:57:09 RP: Same error here with Python 2.4.4 from Debian Etch Aug 16 08:57:40 RP: is ssh for berlios working for you? Aug 16 08:57:52 psyco does not work on 64bit (last time i checked) Aug 16 08:57:54 zecke: I've never used ssh with them I don't think Aug 16 08:58:23 bbl after finished calling mickey Aug 16 08:58:47 RP: Same effect both with and without psyco Aug 16 09:01:49 * RP doesn't know why this isn't working :-( Aug 16 09:03:19 yup -d still has the same effect Aug 16 09:03:30 damn thats slow Aug 16 09:12:43 hi Aug 16 09:13:05 I'm trying to build a few packages, but it errors out on eglibc with the following message: Aug 16 09:13:39 yukito oh you are testing eglibc? Aug 16 09:13:55 NOTE: :name 'get_glibc_fpu_setting' is not defined while evaluating: Aug 16 09:14:13 i'm building gcc and it seems to depend on eglibc Aug 16 09:14:31 my config clearly has TARGET_FPU = "hard" Aug 16 09:15:42 what is eglibc, and why is it so strange that I'm testing it? is it some experimental glibc for embedded devices or something? Aug 16 09:16:22 pleemans, ugent? Aug 16 09:16:41 Yes. Aug 16 09:17:08 coincidence Aug 16 09:17:37 Why? Are you also here? Aug 16 09:18:12 no, but I'm a student there Aug 16 09:18:39 I work here as consultant in the IT department, not OE related. Aug 16 09:19:07 so what brings you to OE then? Aug 16 09:19:17 woglinde, are you still there? Aug 16 09:20:03 I allready following the project grom the start, I own a zaurus cl860. Aug 16 09:20:06 yukito: Try TARGET_FPU="" Aug 16 09:20:24 RP: thanks Aug 16 09:20:47 pleemans: I've probably owned 50% of all handhelds supported, but never a zaurus :( Aug 16 09:20:54 well 50% of the conventional ones :P Aug 16 09:21:45 Don't want to spend that much money on handhelds, wife wouldn't be happy. Aug 16 09:22:29 yeah, I know that feeling, although its not my wife :P Aug 16 09:23:10 mine are all second hand Aug 16 09:24:35 mickeyl: I don't suppose you have any idea on this python error? http://pastebin.ca/659431 Aug 16 09:24:40 I bought also a moto a780 from ebay, and am waiting for the fic 1973/openmoko to go to the broader public. Aug 16 09:25:18 i'm waiting for the public release with wifi Aug 16 09:26:35 RP: It seems to be working Aug 16 09:27:39 ade|desk, florian: Did the stack trace change with psyco disabled? Aug 16 09:28:41 * florian checks Aug 16 09:28:42 The "chaos theory" in action. A forest fire on the north side of the town, the whole city without electricity Aug 16 09:29:07 RP: still didn't work Aug 16 09:29:14 RP: no Aug 16 09:29:37 The ups in the office are going up and down like crazy due to voltage spikes Aug 16 09:32:43 re Aug 16 09:33:45 RP: never seen that. looks strange. which python version is that? i don't have this problem here Aug 16 09:33:54 (python 2.5.1 here) Aug 16 09:34:10 okay ssh works again Aug 16 09:34:12 mickeyl: Seems to be on a variety of platforms... Aug 16 09:34:26 koen: is ewi exposing this issue? Aug 16 09:35:36 RP: might be a psyco problem Aug 16 09:35:50 mickeyl: We've ruled that out Aug 16 09:36:16 RP: can you pinpoint which revision is to blame? Aug 16 09:36:24 i.e. which change lead to the problems Aug 16 09:36:38 mickeyl: yes, renameVar introduced it :) Aug 16 09:36:44 mickeyl: Its from when I added in that code last night (which is a bugfix) Aug 16 09:37:12 RP http://pastebin.ca/659482 Aug 16 09:37:41 RP: did you check how i iterate at other positions? Aug 16 09:38:01 ade|desk: Can you try: http://pastebin.ca/659484 ? Aug 16 09:38:02 RP: it looks like "for key in self._special_values[i]" is failing? Aug 16 09:38:17 zecke: See above :} Aug 16 09:39:06 ah yeah, use has_key only :) Aug 16 09:39:10 like in data.py Aug 16 09:40:08 RP, what should I do when TARGET_FPU = "" doesn't work? Aug 16 09:40:12 Rp trying now Aug 16 09:40:35 yukito: I don't know, I just know that "hard" isn't a valid value Aug 16 09:41:09 ah ok :P I just saw soft and I know i didn't need that one :P Aug 16 09:41:23 * ade|desk taps finger on desk , half way Aug 16 09:42:25 * zecke finally looks at runqueue Aug 16 09:43:06 RP: that worked Aug 16 09:43:13 * ade|desk does the happy dance Aug 16 09:43:23 ade|desk: great :) Aug 16 09:43:47 RP: move the assignment of key into the if branch Aug 16 09:47:12 zecke: I don't understand what you mean? Aug 16 09:47:55 RP: has_key(keyname):\n key = _special_value[i]\n remove(key)\n add(key) Aug 16 09:49:21 zecke: Read that patch again ;-) Aug 16 09:54:58 03rpurdie * r901 10/ (2 files in 2 dirs): data_smart.py: Add has_key() check before accessing data Aug 16 09:56:09 03rpurdie * r902 10/ (2 files in 2 dirs): git.py: Handle tag=master correctly (from poky) Aug 16 09:57:45 * RP crosses fingers :) Aug 16 09:58:13 I know it can still be None and you check for that inside :) Aug 16 09:59:02 ah okay :) Aug 16 09:59:09 damn, i should get food Aug 16 10:00:38 zecke: I did change the API slightly since I changed the way expandKeys works when passed a second option readdata Aug 16 10:01:05 I can't find a single use of that second option and I can't think of good usecases so I'm not going to loose sleep over it Aug 16 10:01:18 lose Aug 16 10:34:22 someone said my name? Aug 16 10:46:05 mithro: I did, mentioning you wrote the COW.py Aug 16 10:50:45 COW pie, yummies Aug 16 10:51:16 oehm Aug 16 10:51:25 this scottish right? Aug 16 10:53:12 steak pie by its normal name Aug 16 10:53:19 * florian wonders how to implement COW for pie ;) Aug 16 10:53:53 XorA: hehe Aug 16 10:55:33 Cow pies in Australia are things from which Dung Beetles break off small pieces and roll them away ... Aug 16 10:55:43 rwhitby lol Aug 16 10:55:46 rwhitby: same here :) Aug 16 10:55:53 he koen Aug 16 11:04:12 * Bernardo is away: Ausente por agora. Aug 16 11:12:37 zecke: oh Aug 16 11:12:40 zecke: cool Aug 16 11:12:53 I didn't remeber writing that - but it sounds like something I would do Aug 16 11:13:10 mithro *g* Aug 16 11:13:25 you used metaclasses for that Aug 16 11:14:18 XorA: is posble create a X especific driver for pxa27x ? Aug 16 11:15:29 03koen 07org.oe.dev * r84189660... 10/ (1 conf/machine/adsp-bf538-ezkit.conf): adsp-bf538-ezkit: add machine file Aug 16 11:17:27 03koen 07org.oe.dev * rfb352e88... 10/ (3 files in 2 dirs): adzs-bf538f-ezlite: rename from adsp-bf538-ezkit Aug 16 11:25:01 psycoman: I think OH looked into this and it wasnt efficient Aug 16 11:26:28 XorA: For the mmx instructions, its not, for xv it would be a good idea Aug 16 11:30:17 zecke: How soon do you think we release? koen/mickeyl voted for ASAP and I agree. Is today too soon? Aug 16 11:31:09 rp new bitbake version? Aug 16 11:31:35 woglinde: yes Aug 16 11:32:24 lrg: the blackfin board is funny Aug 16 11:32:43 lrg: it has an *extra* bf535p and xilinx spartan sc3 Aug 16 11:32:57 lrg: neither are in the schematics or BOM Aug 16 11:33:46 lrg: I'm trying to find out is the blackfin linux tree supports the board, since they are focused on the bf537 board (which has ethernet) Aug 16 11:36:21 koen, the gnu radio guy is using a gumstix Aug 16 11:36:33 sounds like he is fooling with buildroot Aug 16 11:39:15 I'm going to try and fix the gumstix build this morning Aug 16 11:40:09 zecke: if it uses metaclasses then is was probably me :P Aug 16 11:41:45 koen: ah, interesting. I purchased it based on it's audio. I thought, they would all have been supported :-/ (fingers crossed it's supported) Aug 16 11:43:57 koen: hi, what distro should I build for ts72xx? What's recomended/supported? Aug 16 11:43:58 RP: it is, I took off from work and want to get PkgStarted/Failed firing changed :) Aug 16 11:54:01 RP: http://rafb.net/p/BGq62x98.html, r=? (do you want to review it? r=me means you agree with the change) Aug 16 12:04:49 ynezz depends on what you want Aug 16 12:05:01 qt or X11/gtk apps Aug 16 12:06:19 im tring test my images in qemuarm with machine set as spitz, but i git error in partition table Aug 16 12:06:39 i have used a ext2 image in hda Aug 16 12:08:18 woglinde: it's display-less, so just something default Aug 16 12:09:20 ynezz hm Aug 16 12:09:21 and 32MB onboard flash Aug 16 12:09:52 ynezz: DISTRO=angstrom-2007.1 or DISTRO=angstrom-2008.1 Aug 16 12:10:08 yeah and than angstrom-minimal-image Aug 16 12:10:19 RP: I will move this somewhere else Aug 16 12:12:50 lrg: I suspect bf537 vs bf538 is like pxa255 vs pxa263 Aug 16 12:13:05 lrg: lots of emphasis on 'builtin flash' Aug 16 12:14:00 koen: yeah, I'm emailing my contact now. Aug 16 12:14:10 lrg: robin getz? Aug 16 12:14:21 koen: yes and Roy Haung Aug 16 12:14:40 koen: do you know Robin ? Aug 16 12:14:47 lrg: I've been hitting a wall with robin about patches Aug 16 12:15:06 lrg: his only reply is something like "use our svn, nothing else can work" Aug 16 12:15:56 koen: what's prefered to use? I simply want to update recipe for kernel, check the ditro and make something like MokoMakefile available for those asking for EABI on mailing lists... Aug 16 12:16:33 zecke: If they're not using SRCREV, it doesn't need to be set does it? Aug 16 12:16:48 if people can't read, they shouldn't be working with ARM cpus Aug 16 12:17:22 * XorA quits OE Aug 16 12:17:23 RP: well, we call len(None) which will raise an exception. So it must be set in the bitbake.conf Aug 16 12:17:31 XorA: ? Aug 16 12:17:42 zecke: Im known for my inability to read Aug 16 12:17:57 zecke: It checks tag exists first? Aug 16 12:18:24 zecke: Ah, that code is wrong Aug 16 12:18:37 RP: All I know is I get tons of exceptions when parsing my old tree Aug 16 12:19:12 RP: E.g. I can think of BitBake moving forward and not working with old configurations. That is sad but acceptable when the error message is clean Aug 16 12:19:14 woglinde: FYI, it's this board http://www.embeddedarm.com/epc/ts7250-spec-h.html Aug 16 12:20:02 03rpurdie * r903 10/ (2 files in 2 dirs): git.py: Fix reference to tag Aug 16 12:20:07 zecke: Try that ;-) Aug 16 12:20:18 koen: well true, but still better to help them use something working and up2date, than 2.4.26 with some old debian... Aug 16 12:21:01 ynezz: MACHINE=ep93xx DISTRO=angstrom=2007.1 TMPDIR=/path/to/somewhere Aug 16 12:21:13 why do you need a makefile for that? Aug 16 12:21:13 RP: now fix the svn fetcher as well :) Aug 16 12:21:23 koen: i don't Aug 16 12:21:27 RP: Exception: Message:argument of type 'NoneType' is not iterable Aug 16 12:22:35 RP: line 77, on svn.py Aug 16 12:22:56 zecke: yes, I'm thinking about that one :} Aug 16 12:23:58 RP: I think it is okay to require updated bitbake.conf (fallback) Aug 16 12:24:44 zecke: Maybe. The git problem was a definite bug in the code. I'm just trying to think about whether we can change the svn code so it doesn't error Aug 16 12:25:14 RP: leave ud.date set in the case rev is None? Aug 16 12:26:29 lrg: I hope robin can shed some light on the cpu family and linux support Aug 16 12:28:11 koen: me too, email sent. I may need to order other boards if he doesn't support this board :-/ Aug 16 12:28:50 zecke: I think so, yes Aug 16 12:28:52 lrg: if so, the 537-stamp seems to be the best bet Aug 16 12:29:04 RP: BTW: I was too busy working but does the SRCREV change mean I know which revision of the software is used to build? Aug 16 12:29:09 lrg: I'll try booting a kernel when I get back next week Aug 16 12:29:16 RP: okay, let me try then Aug 16 12:29:18 03rpurdie * r904 10/ (2 files in 2 dirs): svn.py: Handle case where SRCREV=None Aug 16 12:29:34 zecke: ^^^^^ ;-) Aug 16 12:30:01 zecke: Yes, SRCREV means the revision is injected into PV Aug 16 12:30:18 bastard :) Aug 16 12:30:37 koen: ok cool thanks :) Aug 16 12:31:28 looks like my change Aug 16 12:31:44 no, it doesn't ;) Aug 16 12:31:48 else if => elif Aug 16 12:32:19 zecke: You'd better fix that ;-) Aug 16 12:32:37 wait then :) Aug 16 12:35:26 03zecke123 * r905 10/ (2 files in 2 dirs): Fix syntax and indention Aug 16 12:35:32 long time no commit Aug 16 12:42:46 RP: which method is runqueue.* calling on the cooker to execute a task? Aug 16 12:45:30 I wonder how far vivijm is with building source packages Aug 16 12:47:45 zecke: tryBuild Aug 16 12:48:32 zecke: At a guess you want to hook into task_complete Aug 16 12:49:30 RP: I fail to see the that maps from filename to the number of to be completed tasks (FileName -> #to run tasks) Aug 16 12:50:08 zecke: There isn't one, there is just a list of tasks Aug 16 12:51:32 zecke: A "task" is a fn, taskname pair of values Aug 16 12:52:29 I really need to rewrite runqueue and taskdata so they code is understandable. It grew too organically :/ Aug 16 12:53:02 OT: Anyone wants to share The Smith, Cure, ... music with me? Aug 16 12:54:39 RP: did you guys look into creating debian/{changelog,rules,*.install} files automatically? Aug 16 12:55:59 * * OE Bug 2775 has been created by autobuild(AT)openembedded.org Aug 16 12:56:00 * * python-formencode-0.4-r0-do_fetch Aug 16 12:56:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2775 Aug 16 12:56:11 * * OE Bug 2774 has been created by autobuild(AT)openembedded.org Aug 16 12:56:12 * * python-egenix-mx-base-2.0.6-r1-do_patch Aug 16 12:56:14 * * http://bugs.openembedded.org/show_bug.cgi?id=2774 Aug 16 12:57:36 zecke: No, not us Aug 16 12:58:26 zecke: I have some Cure albums.. Aug 16 12:59:44 * zecke was forced to package some software, and wants to use bitbake for that now Aug 16 13:04:02 RP: first iteration http://rafb.net/p/NDBNes73.html. Rename it to tryBuildPackage to tryBuildPackageTask, use this method instead of tryBuild from runqueue Aug 16 13:06:03 RP: now PkgStarted, PkgFailed and PkgSucceed is not fired at all :) Aug 16 13:09:41 hm Aug 16 13:10:39 where bitbake genereate image files ? Aug 16 13:13:05 tmp/deploy/*/ipk/* Aug 16 13:13:14 tmp/deploy/*/iages/* Aug 16 13:13:19 crappit I give up spelling Aug 16 13:13:58 sorry i want my wrong question Aug 16 13:14:08 i want see the sourcecoude off process Aug 16 13:14:18 s/sourcecoude/sourcecode Aug 16 13:14:32 org.oe.dev/classes/image.bbclass Aug 16 13:14:46 thnkx Aug 16 13:15:46 XorA: ext2 images i genareted with this code ? Aug 16 13:18:54 RP: no pressing need but have you considered updating bitbake to GPLv3? Aug 16 13:19:18 psycoman: yes Aug 16 13:19:58 * * OE Bug 2776 has been created by autobuild(AT)openembedded.org Aug 16 13:20:00 * * python-gammu-0.21-ml0-do_compile Aug 16 13:20:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2776 Aug 16 13:20:57 does anyone know what the fw_printenv and setenv tools from u-boot do/are? Aug 16 13:21:24 crofton I could onlyguess Aug 16 13:22:11 crofton but set/read env variables beside the include files Aug 16 13:22:33 well, I think I lost the patches when I moved u-boot-utils into ubooy, that seem to be needed by these tools Aug 16 13:23:03 I am thinking of making uboot-utils only buils mkimage and stage it Aug 16 13:23:10 since this is what we badly need Aug 16 13:23:35 crofton yes Aug 16 13:23:59 and since I do not think the package has built in a while, I do not think this will annoy anyone :) Aug 16 13:25:25 I have stumbled upon a known problem with nmap. Aug 16 13:25:49 When compiling nmap with gcc4, it doesn't work as user root, only un-priveleged accounts work. Aug 16 13:26:13 Are there any way to use an older gcc version for a single package? Aug 16 13:26:29 is there a patch for nmap? Aug 16 13:26:58 I have not been able to find a patch for this, but I feel rather confident that it is fixed in newer versions. Aug 16 13:26:58 * * OE Bug 2777 has been created by jstoefer(AT)schuett-elektronik.de Aug 16 13:27:00 * * packages eflpp resp. edje fail to build Aug 16 13:27:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2777 Aug 16 13:27:38 On my ubuntu with nmap-4.20 the bug is gone. Aug 16 13:27:55 But building nmap-4.20 does not seem that easy. Aug 16 13:28:17 I don't have the time to continue working on it for now :-( Aug 16 13:28:38 03xora 07org.oe.dev * r9026ff8d... 10/ (1 packages/gsm/libgsmd_svn.bb): gsm/libgsmd_svn.bb : remove patches that have been superceded upstream Aug 16 13:29:19 RP: runqueue.py: if executing a task fails you sometimes sys.exit(1) and otherwise just forward the exception? Aug 16 13:30:54 Crofton|home: should I file a bug on it? Aug 16 13:31:07 sure Aug 16 13:31:23 please link to info documenting it as a known issue Aug 16 13:31:42 zecke: That's in the child, I'm not sure why it doesn't always exit 1. There would have been a good reason... :/ Aug 16 13:32:58 * * OE Bug 2778 has been created by autobuild(AT)openembedded.org Aug 16 13:32:59 * * python-m2crypto-0.13.1-r0-do_compile Aug 16 13:33:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2778 Aug 16 13:38:31 zecke: Two things, I'm not sure moving the loadDataFull out of cooker is a good idea and I'm not sure moving the fire calls to the call sites is good either. Ultimately I'd like some functions in cooker to do these things through a single call Aug 16 13:39:10 zecke: Maybe we need tryBuildPackage and tryBuildPackageTask ? Aug 16 13:39:44 RP: maybe, at least runqueue shouldn't call tryBuild Aug 16 13:40:47 zecke: right :) Aug 16 13:42:14 on further reading, gumtstix uses u-boot, so we should just make machine depend on u-boot and build u-boot image for completeness Aug 16 13:42:19 * zecke tries to find his way through the for loops of runqueue and taskdata Aug 16 13:42:48 zecke: I feel ashamed of that code... Aug 16 13:42:55 don't worry Aug 16 13:43:00 zecke: The only good thing I can say about it is it works... Aug 16 13:43:23 I feel ashamed for myself. "I'm just fixing other's code" is my excuse (and you could use that as well) Aug 16 13:43:58 zecke: taskdata and runqueue are mostly all my own code now though ;-) Aug 16 13:43:58 * * OE Bug 2779 has been created by eha(AT)doredevelopment.dk Aug 16 13:44:00 * * nmap does not work as root when built with gcc4 Aug 16 13:44:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2779 Aug 16 13:44:09 I should just forget that... :) Aug 16 13:44:49 *tootache* Aug 16 13:47:58 * * OE Bug 2780 has been created by autobuild(AT)openembedded.org Aug 16 13:48:00 * * mysql-4.1.18-r2-do_compile Aug 16 13:48:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2780 Aug 16 13:54:24 esben, I took a quick look at the nmap bb file and it won't be a five minute fix Aug 16 13:54:33 I'll try and come back to it at some point Aug 16 13:54:39 hehe Aug 16 13:57:23 Crofton|home: Nope, I spent a bit more than 5 minuttes, finding a patch which fixed some of the autoconf problems only to open up for even more :-/ Aug 16 13:58:01 you could try making a clean bb file for the latest version and see if the hacks are still needed Aug 16 13:58:06 Crofton|home: you might want to start out with looking at http://seclists.org/nmap-dev/2006/q4/att-0282/nmap-4_20_diff Aug 16 13:58:21 Crofton|home: I tried that, but no luck. Aug 16 13:58:22 esben, not now Aug 16 13:58:31 I have to leave for today as well. Aug 16 13:58:33 I am looking into fixing gumstix build atm Aug 16 13:58:43 cya Aug 16 14:03:59 * * OE Bug 2781 has been created by autobuild(AT)openembedded.org Aug 16 14:04:01 * * python-soappy-0.11.3-r0-do_compile Aug 16 14:04:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2781 Aug 16 14:04:59 03mickeyl 07org.oe.dev * r34526254... 10/ (6 files in 3 dirs): python-pycairo: update to 1.4.0 Aug 16 14:05:04 03mickeyl 07org.oe.dev * r22ab6b46... 10/ (3 files in 2 dirs): python-pyfits: update to 1.1 Aug 16 14:05:09 03mickeyl 07org.oe.dev * rce6afe15... 10/ (1 packages/python/python-numarray_1.1.1.bb): python-numarray: fix packaging debug files Aug 16 14:05:14 03mickeyl 07org.oe.dev * rbaa4171c... 10/ (3 files in 2 dirs): python-pyraf: update to 1.4 Aug 16 14:05:21 03mickeyl 07org.oe.dev * r799f65eb... 10/ (6 files in 3 dirs): python-soappy: update to 0.11.6 Aug 16 14:05:25 03mickeyl 07org.oe.dev * r836b0ed6... 10/ (1 packages/python/python-pygoogle_0.6.bb): python-pygoogle: add HOMEPAGE Aug 16 14:05:32 03mickeyl 07org.oe.dev * ra1141fe7... 10/ (3 files in 2 dirs): python-pyflakes: update to 0.2.1 Aug 16 14:05:36 03mickeyl 07org.oe.dev * re8206e46... 10/ (3 files in 2 dirs): python-pygame: remove 1.6 Aug 16 14:05:41 03mickeyl 07org.oe.dev * rf2fd60bb... 10/ (1 packages/python/python-pylinda_0.6.bb): python-pylinda: add HOMEPAGE Aug 16 14:05:46 03mickeyl 07org.oe.dev * r4278f3cd... 10/ (3 files in 2 dirs): python-pylint: update to 0.13.2 Aug 16 14:05:51 03mickeyl 07org.oe.dev * r6e1b6a76... 10/ (12 files in 2 dirs): python-pygtk[2]: change naming, python-pygtk is now 2.x, update to 2.10.4 Aug 16 14:05:56 03mickeyl 07org.oe.dev * re6ba4be1... 10/ (6 files in 3 dirs): Aug 16 14:05:56 python-pyreverse: update to 0.5.2 Aug 16 14:05:56 pyxml: touch Aug 16 14:06:01 03mickeyl 07org.oe.dev * r5bcaa008... 10/ (3 files in 2 dirs): python-pyro: update to 3.7 Aug 16 14:06:06 03mickeyl 07org.oe.dev * rbf2b7453... 10/ (3 files in 2 dirs): python-pyweather: update to 0.7.0 Aug 16 14:06:12 03mickeyl 07org.oe.dev * rdfb4e433... 10/ (3 files in 2 dirs): python-pytest: update to 0.6.0 Aug 16 14:06:18 03mickeyl 07org.oe.dev * re0ac0e1c... 10/ (3 files in 2 dirs): python-pyvisa: update to 1.1 Aug 16 14:06:23 03mickeyl 07org.oe.dev * rca9071da... 10/ (3 files in 2 dirs): Aug 16 14:06:23 remove pyserial 2.0 Aug 16 14:06:23 remove python-quicklauncher 0.0.1 Aug 16 14:06:33 03mickeyl 07org.oe.dev * r042519f1... 10/ (3 files in 2 dirs): python-scapy: update to 1.1.1 Aug 16 14:06:45 03mickeyl 07org.oe.dev * r22a64554... 10/ (5 files in 2 dirs): python-scons[-native]: udpate to 0.97 Aug 16 14:06:50 03mickeyl 07org.oe.dev * r7b07ebfd... 10/ (17 files in 2 dirs): python-pyqt: promote qt4 to being the default, rename other packages accordingly Aug 16 14:06:59 03mickeyl 07org.oe.dev * r44702518... 10/ (3 files in 2 dirs): python-sip: update to 4.7 Aug 16 14:07:06 03mickeyl 07org.oe.dev * r1a41c577... 10/ (3 files in 2 dirs): python-spydi: update to 0.9.7 Aug 16 14:07:13 03mickeyl 07org.oe.dev * rb16bd834... 10/ (3 files in 2 dirs): python-spyro: update to 0.9.22 Aug 16 14:07:23 03mickeyl 07org.oe.dev * rc0f79b6d... 10/ (3 files in 2 dirs): swig[-native]: remove 1.3.29 Aug 16 14:07:31 03mickeyl 07org.oe.dev * rf557fb5d... 10/ (4 files in 2 dirs): sword: inherit lib_package Aug 16 14:07:38 03mickeyl 07org.oe.dev * rbc9c1233... 10/ (3 files in 2 dirs): python-formencode: update to 0.4+svn Aug 16 14:07:48 03mickeyl 07org.oe.dev * r8f03732d... 10/ (3 files in 2 dirs): python-imaging: update to 1.1.6 Aug 16 14:07:58 03mickeyl 07org.oe.dev * re5f326fb... 10/ (3 files in 3 dirs): move python-egenix-mx-base to nonworking, source is no longer available Aug 16 14:08:06 03mickeyl 07org.oe.dev * r8e2b9fd6... 10/ (3 files in 2 dirs): python-imdbpy: update to 3.1 Aug 16 14:08:15 03mickeyl 07org.oe.dev * r840d3024... 10/ (1 packages/python/python-imdbpy_3.1.bb): python-imdbpy: fix debug packaging and SRC_URI Aug 16 14:08:22 yeah, distributed systems! Aug 16 14:08:24 03mickeyl 07org.oe.dev * rc73d88f5... 10/ (6 files in 3 dirs): move python-m2crypto to nonworking, it needs to catch up w/ swig Aug 16 14:08:33 03mickeyl 07org.oe.dev * rede20708... 10/ (5 files in 2 dirs): python-pyiw: update to 0.3.3 Aug 16 14:08:33 7 hours of python work Aug 16 14:08:40 03mickeyl 07org.oe.dev * rb17c1ea9... 10/ (3 files in 2 dirs): python-gst: update to 0.10.8 Aug 16 14:08:48 03mickeyl 07org.oe.dev * r1c4cb646... 10/ (5 files in 3 dirs): python-pysqlite: remove. this is now included in the python distribution Aug 16 14:08:56 03mickeyl 07org.oe.dev * r2a44f628... 10/ (3 files in 2 dirs): python-sqlobject: update to 0.9.1 Aug 16 14:09:12 still some tiny things to fix, but the overall state of python is pretty good in OE Aug 16 14:09:31 damn dude. Aug 16 14:09:36 nicely done Aug 16 14:09:40 :) Aug 16 14:09:51 that was a bit of an uber update, I would accuse you of flooding ;-) Aug 16 14:09:55 hehe Aug 16 14:10:00 RP: http://paste.lisp.org/display/46213 next base Aug 16 14:11:07 03mickeyl 07org.oe.dev * r1b815ac5... 10/ (3 files in 2 dirs): python-tlslite: update to 0.3.8 Aug 16 14:11:13 03mickeyl 07org.oe.dev * reb7be243... 10/ (3 files in 2 dirs): python-urwid: update to 0.9.8.1 Aug 16 14:11:21 oops, the rest :) Aug 16 14:11:21 03mickeyl 07org.oe.dev * r3e3c84a5... 10/ (3 files in 2 dirs): python-webpy: update to 0.21 Aug 16 14:11:26 03mickeyl 07org.oe.dev * r85c472fc... 10/ (1 packages/moin/moin_1.2.2.bb packages/moin/moin_1.5.8.bb): moin: update to 1.5.8 Aug 16 14:11:32 03mickeyl 07org.oe.dev * r9fa0c99d... 10/ (3 files in 2 dirs): task-python-everything: update Aug 16 14:11:32 geez Aug 16 14:11:38 still? Aug 16 14:11:41 no, that's about it Aug 16 14:11:49 One, more thing :) Aug 16 14:11:52 * mickeyl rm -rf tmp and rebuilds everything. lets see where it breaks now Aug 16 14:12:55 zecke: runqueue doesn't need is_current, thats handled before the fork for efficiency Aug 16 14:13:19 RP: how do you load the data? Aug 16 14:13:25 well, I find it myself Aug 16 14:13:57 zecke: We cache the stamp in cache.py so runqueue can do this Aug 16 14:14:09 self.dataCache Aug 16 14:16:04 zecke: tryBuildPackage can call tryBuildPackageTask too if I was being picky ;-) Aug 16 14:16:32 RP: and catch the exceptions twice :) Aug 16 14:16:58 * * OE Bug 2782 has been created by autobuild(AT)openembedded.org Aug 16 14:17:00 * * python-pyode-1.0.0-r0-do_compile Aug 16 14:17:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2782 Aug 16 14:17:10 * * OE Bug 2783 has been created by autobuild(AT)openembedded.org Aug 16 14:17:12 * * python-sip4-4.4.5-ml0-do_compile Aug 16 14:17:14 * * http://bugs.openembedded.org/show_bug.cgi?id=2783 Aug 16 14:17:22 * * OE Bug 2784 has been created by autobuild(AT)openembedded.org Aug 16 14:17:24 * * python-pyqt4-4.2-ml0-do_generate Aug 16 14:17:25 zecke: Once should be enough Aug 16 14:17:26 * * http://bugs.openembedded.org/show_bug.cgi?id=2784 Aug 16 14:17:58 RP: how? tryBuildPackagetask will catch the exception, print and raise it again Aug 16 14:18:12 RP: tryBuildPackage would need to catch it again and send the PkgFailed event Aug 16 14:18:55 zecke: ah, I just mean you can catch that with one exception handler in trybuild package rather than the current two with duplicate fire lines... Aug 16 14:18:58 if I have a .bb file for a shared library, and the file is named "name_version.bb", when I write other bb files and want to specify dependencies, or when I specify RDEPENDS, do I have to use "name" or "libname", because ultimately a libname is produced from a library package .bb? Aug 16 14:19:34 Jin^eLD: You use the name of the package for an RDEPENDS Aug 16 14:19:35 I am asking because sometimes it finds stuff specified in RDEPENDS but sometimes it doesnot and I am not quite sure what I am doing wrong Aug 16 14:19:55 RP: right Aug 16 14:19:59 Jin^eLD: Note this is the name of the package as it is in PACKAGES, not after things like debian renaming Aug 16 14:20:17 RP: and if do_rootfs bombs out, telling its not found, why can that happen (the package was built) Aug 16 14:20:44 Jin^eLD: Does the name its looking for match the name in the rootfs error log? Aug 16 14:21:05 I will have to check Aug 16 14:21:39 Jin^eLD: Sorry, I mean does the name of the ipk match the name its looking for in the rootfs log? Aug 16 14:21:58 * * OE Bug 2748 has been RESOLVED (INVALID) by philippedeswert(AT)scarlet.be Aug 16 14:22:00 * *  librsvg fails to build Aug 16 14:22:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2748 Aug 16 14:22:53 crap, dont have the log anymore, changed some other stuff and doing a rebuild, but I will look at that Aug 16 14:23:04 RP: thanks Aug 16 14:27:58 * * OE Bug 1991 has been RESOLVED (WORKSFORME) by philippedeswert(AT)scarlet.be Aug 16 14:28:00 * *  librsvg-2.16.0 build fails: QA Issue: non debug package contains .debug directory Aug 16 14:28:02 * * http://bugs.openembedded.org/show_bug.cgi?id=1991 Aug 16 14:36:55 morning Aug 16 14:39:57 hey chouimat Aug 16 14:40:08 afternoon ! Aug 16 14:40:58 * * OE Bug 2785 has been created by autobuild(AT)openembedded.org Aug 16 14:41:00 * * evas-0.9.9+cvs20070721-r0-do_configure Aug 16 14:41:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2785 Aug 16 14:41:58 * * OE Bug 2786 has been created by autobuild(AT)openembedded.org Aug 16 14:42:00 * * directfb-1.0.0-r5-do_rm_work Aug 16 14:42:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2786 Aug 16 14:43:31 mickeyl: you had a busy week: http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-August/002808.html :) Aug 16 14:44:32 koen: indeed. worked 60 hours or so Aug 16 14:44:38 perhaps even more Aug 16 14:44:40 i lost track Aug 16 14:45:18 koen: did you know OE can only build ROM images? Aug 16 14:45:36 zecke: I didn't Aug 16 14:46:05 mickeyl: You noticed I found the bitbake dependency glitch btw? Aug 16 14:46:17 RP: yeah. some off by one error? Aug 16 14:46:33 RP: how come it differed on your/my setup? Aug 16 14:46:48 mickeyl: if foo: instead of if foo not None: where foo could be int(0) Aug 16 14:46:55 RP: aaah :) Aug 16 14:47:14 mickeyl: I typed bitbake x y, you typed bitbake y, x Aug 16 14:47:27 koen: http://t2-project.org/handbook/html/t2.intro.different.distribution.html :) Aug 16 14:47:27 mickeyl: The first option gets a task id of zero Aug 16 14:48:43 mickeyl: It might explain a few "random" dependency issues people reported :) Aug 16 14:48:54 right, cool! Aug 16 14:49:03 yay for more predictability Aug 16 14:49:37 RP: http://paste.lisp.org/display/46217, besides the runqueue garbage it should be okay Aug 16 14:49:39 mickeyl: DId you see the one about the order of files on the disk affecting the order of providers? Aug 16 14:49:50 hell, no! Aug 16 14:49:53 RP: what'S that about? Aug 16 14:50:42 mickeyl: When we work out a list of possible providers if was intially sorted by the order we read the files from disk. If none was set, the one chosen therefore depended on the machine :} Aug 16 14:51:12 RP: imho this was okay :) Aug 16 14:51:36 zecke: In a way I wanted to leave it but we need to be a bit more predictable Aug 16 14:52:20 mickeyl: I added in a sort to make things more reproducible... Aug 16 14:52:26 RP: yes, I want to see us moving into a direction of locking stuff down with one bitbake command (emitting config files), selecting features of packages, etc Aug 16 14:52:34 RP: phew. "fun" Aug 16 14:53:26 mickeyl: The thing is on devs machines, "new" providers always ended up at the end of the list so they didn't see problems. People making fresh checkouts saw a different order though Aug 16 14:53:39 right Aug 16 14:53:46 good catch Aug 16 14:53:58 I was amused when I worked it out :) Aug 16 14:54:40 zecke: How do you mean "locking down" ? You mean locking certain versions? Aug 16 14:55:06 RP: bitbake tell me your PREFERRED_* of this build, and it would write _VERSION, _PROVIDERS, etc. Aug 16 14:55:37 zecke: A kind of modified version of bitbake -s ? Aug 16 14:55:52 RP: yes Aug 16 14:56:01 RP: so it can be directly used by distributions, etc. Aug 16 14:56:55 zecke: agreed, something like that would be good Aug 16 14:57:18 RP: and I want something like make menuconfig back :) Aug 16 14:57:20 zecke: I'd also like a mode to see how much my locked down SRCREV versions have changed upstream :} Aug 16 14:57:41 zecke: To select versions? Aug 16 15:00:22 RP: versions, flags, virtual providers, and features as well. I think it could be a distro decision to build certain things with less/more features Aug 16 15:00:28 zecke: I haven't forgotten about UIs, I'm slowly trying to merge my UI changes into trunk Aug 16 15:01:00 hello all Aug 16 15:01:03 I've just been giving bitbake 1.8 priority for a while... Aug 16 15:01:05 hi likewise Aug 16 15:01:37 RP: you are my mvp :) Aug 16 15:03:55 in what channel can I ask for mmap() expertise? Aug 16 15:05:47 zecke: I very much like your idea of locking down -- I'll happily volunteer to test such a feature. Aug 16 15:06:20 mwester: we don't need testers, we need crack and meth :) Aug 16 15:06:58 * * OE Bug 2708 has been RESOLVED (WORKSFORME) by philippedeswert(AT)scarlet.be Aug 16 15:07:00 * *  Package gtk+-2.10.14-r1 fails to compile Aug 16 15:07:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2708 Aug 16 15:07:19 The best I can provide is beer... ;) Aug 16 15:07:42 zecke : hmmm is that for OEDEM ? Aug 16 15:08:02 steliosk: what? hacking that? Aug 16 15:08:28 zecke : no. The supplies of crack and meth :) Aug 16 15:08:46 you have to bring your own, I'm bad at sharing Aug 16 15:09:37 RP: runq_fnid[task] will give me the fnid to be used with taskData? Aug 16 15:09:46 * steliosk makes a note Aug 16 15:10:31 zecke: yes Aug 16 15:12:56 RP: as a definition for now. PkgStarted is called when the first task of a package is going to be executed Aug 16 15:13:17 zecke: sounds reasonable Aug 16 15:13:21 RP: PkgFailed (in worst case might be emitted more than once), is emitted whenever a task fails Aug 16 15:13:42 RP: and PkgSuceeded when all tasks for a given fnid were executed Aug 16 15:14:10 zecke: I'd still like to know is these are useful in the real world? :) Aug 16 15:14:22 s/is/if/ Aug 16 15:14:59 RP: yes, for the varpoware tinderbox this is useful. This means no more tasks are going to be executed during this build for this given package Aug 16 15:15:18 RP: and probably my version of rm_work will use it (Yes, I am danger mouse) Aug 16 15:15:39 danger mouse on crack and meth? Scary indeed! Aug 16 15:16:22 ;) Aug 16 15:16:23 zecke: Please don't check that into .dev ;-) Aug 16 15:16:46 RP: it is less broken than the current version of rm_work, should I revive the zecke branch? Aug 16 15:17:18 zecke: I will fix rm_work as I described previously which won't do dangerous things Aug 16 15:17:19 RP: you could beat me and fix the deps of rm_work with bitbake 1.8 Aug 16 15:20:29 RP: well, I see how PkgSucceed will be fired for the -b -cbla build :) Aug 16 15:20:41 did anyone test drive gcc 4.2.1 under oe yet? Aug 16 15:20:50 RP: but for the normal build (by name), it should be the same Aug 16 15:20:58 * * OE Bug 2787 has been created by autobuild(AT)openembedded.org Aug 16 15:21:00 * * python-ecore-0.0.0+cvs20070816-r0-do_compile Aug 16 15:21:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2787 Aug 16 15:25:26 zecke: My concern is it will cause packages to be multiple times... Aug 16 15:25:55 RP: no, it will just break due the files being removed :) Aug 16 15:26:49 zecke: That too Aug 16 15:27:56 than prevent me doing crazy things :) Aug 16 15:28:17 zecke: by fixing rm_work? Aug 16 15:28:26 yes Aug 16 15:28:28 Building for guinness/angstrom, this is not picked up: SRC_URI_append_i686-generic = "file://i686-defconfig" (as it is set to i686-angstrom). What's the proper fix for this? Aug 16 15:28:50 RP: I need PkgStarted/Succeeded for the tinderbox though but it is 'sane' in this case Aug 16 15:28:58 * * OE Bug 2788 has been created by autobuild(AT)openembedded.org Aug 16 15:29:00 * * python-evas-0.0.0+cvs20070816-r0-do_compile Aug 16 15:29:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2788 Aug 16 15:29:10 * * OE Bug 2789 has been created by autobuild(AT)openembedded.org Aug 16 15:29:10 zecke: Agreed, its just using them for rm_work I disagree with Aug 16 15:29:12 * * python-pytest-0.6.0-ml0-do_fetch Aug 16 15:29:14 * * http://bugs.openembedded.org/show_bug.cgi?id=2789 Aug 16 15:29:28 zecke: and I do have a patch to test for rm_work Aug 16 15:29:43 * RP just hopes we call expand in the right places for it to work Aug 16 15:32:05 I'm reading runqeue.py and I don't feel anger Aug 16 15:32:47 zecke: You surprise me :) Aug 16 15:33:16 I just like you Aug 16 15:33:22 Does code often make zecke angry? Aug 16 15:33:49 mwester: people not code :) Aug 16 15:33:49 zecke: Seriously, it is quite bad code to understand though? :/ Aug 16 15:34:05 * mwester resolves to not let zecke ever see any of his code... :) Aug 16 15:34:16 RP: yes, but you will be the one to review Aug 16 15:34:31 mwester: need to invent GPL + zecke exclusion clause Aug 16 15:35:25 "linux-rp.do_rm_work" -> "linux-rp.do_${RMWORK_ORIG_TASK}" is task-depends.dot suggests we do miss some expand calls Aug 16 15:35:31 * florian does not fear publishing bad code - he always knows there are worse examples ;) Aug 16 15:38:11 hi peeps Aug 16 15:38:28 hi philippe_work Aug 16 15:38:56 would it be ok to remove the gtk-webcore fixed svn dates for 20060212? As they fail to build and gpe these days uses the svn versions anyway? Aug 16 15:38:59 hi RP! Aug 16 15:39:29 this would also fix bug #2358 Aug 16 15:43:58 * * OE Bug 2790 has been created by autobuild(AT)openembedded.org Aug 16 15:44:00 * * python-sword-1.5.9-r0-do_compile Aug 16 15:44:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2790 Aug 16 15:49:58 * * OE Bug 2755 has been RESOLVED (FIXED) by likewise(AT)gmx.net Aug 16 15:50:00 * *  Perl's Cross/Makefile re-uses TOPDIR from environment Aug 16 15:50:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2755 Aug 16 15:53:24 zecke: http://rafb.net/p/b6rLId11.html Aug 16 15:54:21 RP: btw, rm foo/* leaves .svn dirs (== big space waste) Aug 16 15:54:58 RP: and you commented removing out Aug 16 15:55:12 RP: but doesn't runqueue completely ignore BB_DEFAUL_TASK? Aug 16 15:55:18 zecke: I would uncomment that before committing ;-) Aug 16 15:55:38 zecke: It gets setup in cooker, runqueue doesn't care about details like that Aug 16 15:55:50 koen: That should be fixed then Aug 16 15:57:27 RP: I miss something. I assume it is correct. But by default do_build is the default task? Aug 16 15:57:42 RP: but someone adds package_write as default task? Aug 16 15:57:57 zecke: Yes, do_build is the default task Aug 16 15:58:11 zecke: Why would someone do that? Aug 16 15:58:32 I will look at the task graph... Aug 16 15:58:50 The only point of BB_DEFAULT_TASK is not to hardcode "build" into bitbake Aug 16 15:58:58 * * OE Bug 2791 has been created by autobuild(AT)openembedded.org Aug 16 15:59:00 * * python24-native-2.4.0-r2-do_compile Aug 16 15:59:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2791 Aug 16 15:59:42 RP: man it is just too loud here to focus... Aug 16 16:00:41 zecke: If you run bitbake foo -c package_write it will not run rm_work afterwards Aug 16 16:00:49 zecke: I'd consider that correct though Aug 16 16:01:14 RP: let us start way earlier Aug 16 16:01:37 bitbake foo Aug 16 16:01:43 we call the default task on foo Aug 16 16:01:48 which is now rm_work Aug 16 16:02:12 which will be called after do_build Aug 16 16:02:27 zecke: yes Aug 16 16:02:33 but if foo DEPENDS on base Aug 16 16:02:48 foo.do_rmwork -> base.do_package_write Aug 16 16:03:06 (I have not seen the place where this dep is added, but I assume it is) Aug 16 16:03:42 It will add foo.do_rm_work -> base.do_rm_work Aug 16 16:04:08 base.do_rm_work -> base.do_build Aug 16 16:04:09 why? in the current graph I do not see deps on do_build Aug 16 16:04:55 zecke: rm_work is added after build Aug 16 16:05:26 zecke: Are you inheriting rm_work in a single .bb file or in your local.conf? Aug 16 16:05:50 RP: okay. Again. In the current (as in not applying your patch), if see no do_build on the right side/target Aug 16 16:06:18 zecke: Right, because nothing depends on it Aug 16 16:06:33 so how will do_rmwork turnup then? Aug 16 16:06:39 zecke: This patch adds the dependency Aug 16 16:06:48 zecke: rm_work needs the dependency, we add it Aug 16 16:06:51 by the recrdeps Aug 16 16:06:55 right Aug 16 16:07:21 ack then Aug 16 16:08:07 zecke: It will cause more to get built but it will operate safely. Anyone not using rm_work doesn't get the uneeded dependencies, everyone is happy :) Aug 16 16:08:49 zecke: You agree with the added expands in build.py too? Aug 16 16:09:19 yupp Aug 16 16:09:31 hi guys Aug 16 16:10:15 I getting a problem with bitbake nano Aug 16 16:10:29 its giving me the following error Aug 16 16:10:35 NOTE: Handling BitBake files: - (2816/4555) [61 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Aug 16 16:10:38 ${@['', 'opie-help-en (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Aug 16 16:10:47 any idea where to look for ? Aug 16 16:11:27 update your bitbake and monotone Aug 16 16:11:51 that would take care of it ? Aug 16 16:11:59 should, yes Aug 16 16:16:15 http://www.floydfest.com/2007/dir.php?dir_id=1&page=1 Aug 16 16:16:28 root____2, pastebin Aug 16 16:18:34 root@ubuntu:/home/ubuntu/MyDesk/oe/build# bitbake nano Aug 16 16:18:34 NOTE: Handling BitBake files: \ (2816/4555) [61 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Aug 16 16:18:37 ${@['', 'opie-help-en (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Aug 16 16:18:40 ERROR: EOL while scanning single-quoted string (line 1) while parsing /home/ubuntu/MyDesk/oe/org.openembedded.dev/packages/opie-help-en/opie-help-en_cvs.bb Aug 16 16:18:43 NOTE: Handling BitBake files: / (2855/4555) [62 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Aug 16 16:18:46 ${@['', 'opie-keytabs (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Aug 16 16:18:49 ERROR: EOL while scanning single-quoted string (line 1) while parsing /home/ubuntu/MyDesk/oe/org.openembedded.dev/packages/opie-keytabs/opie-keytabs_cvs.bb Aug 16 16:18:52 NOTE: Handling BitBake files: \ (2904/4555) [63 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Aug 16 16:18:55 root____2, again, use pastbin Aug 16 16:18:55 ${@['', 'opie-mediaplayer2-skin-default-landscape (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Aug 16 16:18:58 ~pastebin Aug 16 16:18:58 ERROR: EOL while scanning single-quoted string (line 1) while parsing /home/ubuntu/MyDesk/oe/org.openembedded.dev/packages/opie-mediaplayer2/opie-mediaplayer2-skin-default-landscape_cvs.bb Aug 16 16:18:59 ~pastebin Aug 16 16:19:02 03pH5 07org.oe.dev * r051681fc... 10/ (1 packages/gtk+/gtk+_2.10.14.bb packages/gtk+/gtk-2.10.inc): gtk+-2.10.14: package printbackends -dev and -dbg files Aug 16 16:19:02 NOTE: Handling BitBake files: | (2906/4555) [63 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Aug 16 16:19:06 ${@['', 'opie-mediaplayer2-skin-default (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Aug 16 16:19:07 hmm... pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Aug 16 16:19:08 pastebin is probably a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Aug 16 16:19:08 ERROR: EOL while scanning single-quoted string (line 1) while parsing /home/ubuntu/MyDesk/oe/org.openembedded.dev/packages/opie-mediaplayer2/opie-mediaplayer2-skin-default_cvs.bb Aug 16 16:19:12 NOTE: Handling BitBake files: / (2908/4555) [63 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Aug 16 16:19:15 ${@['', 'opie-mediaplayer2-skin-pod (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Aug 16 16:19:17 ERROR: EOL while scanning single-quoted string (line 1) while parsing /home/ubuntu/MyDesk/oe/org.openembedded.dev/packages/opie-mediaplayer2/opie-mediaplayer2-skin-pod_cvs.bb Aug 16 16:19:20 NOTE: Handling BitBake files: - (2910/4555) [63 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Aug 16 16:19:21 #root____2: like Crofton already said, use pastebin.ca Aug 16 16:19:24 ${@['', 'opie-mediaplayer2-skin-techno (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Aug 16 16:19:27 ERROR: EOL while scanning single-quoted string (line 1) while parsing /home/ubuntu/MyDesk/oe/org.openembedded.dev/packages/opie-mediaplayer2/opie-mediaplayer2-skin-techno_cvs.bb Aug 16 16:19:30 NOTE: Handling BitBake files: - (2966/4555) [65 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Aug 16 16:19:33 ${@['', 'opie-pics (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Aug 16 16:19:36 ERROR: EOL while scanning single-quoted string (line 1) while parsing /home/ubuntu/MyDesk/oe/org.openembedded.dev/packages/opie-pics/opie-pics_cvs.bb Aug 16 16:19:39 NOTE: Handling BitBake files: | (3050/4555) [66 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Aug 16 16:19:42 ${@['', 'opie-sounds (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Aug 16 16:19:45 ERROR: EOL while scanning single-quoted string (line 1) while parsing /home/ubuntu/MyDesk/oe/org.openembedded.dev/packages/opie-sounds/opie-sounds_cvs.bb Aug 16 16:19:48 NOTE: Handling BitBake files: \ (4555/4555) [100 %] Aug 16 16:19:49 SPAM!!!!!!!!!!!!!!!!!1 Aug 16 16:19:50 ok guys .. sorry abt it Aug 16 16:19:51 I love /igonore Aug 16 16:19:53 I love /ignore Aug 16 16:20:09 that's four underscores in his nick, btw :-) Aug 16 16:20:11 quiet Aug 16 16:20:28 likewise: I love tab completion :) Aug 16 16:20:52 chouimat: lol :-) Aug 16 16:22:40 guys . dnt know this thing .. Aug 16 16:22:53 http://pastebin.ca/659867 Aug 16 16:26:22 root____2: sounds like a bug that was fixed yesterday or so. mtn pull and mtn update Aug 16 16:27:00 thanks likewise will be right on it Aug 16 16:30:31 RP: we can do a bitbake release without the pkgstarted/succeeded changes Aug 16 16:31:32 zecke: I should add the build.py ones though? Aug 16 16:31:41 sure Aug 16 16:31:53 rm_work should work for the moko boyscout team Aug 16 16:32:02 thanks for fixing Aug 16 16:32:26 zecke: np, its been on the todo list for an age :/ Aug 16 16:32:37 RP: we should ask people to start a build and then do a release Aug 16 16:33:30 RP: and I have issues with runq_fnids. You use task numbers range(len(fnids)) but this means the last task number has no file attached? Aug 16 16:34:59 zecke: hmm. That doesn't sounds healthy Aug 16 16:35:34 03rpurdie * r906 10/ (4 files in 4 dirs): build.py: Expand data to addtask Aug 16 16:36:40 zecke: I agree about a release soon Aug 16 16:37:11 what is in runq_revdeps? Aug 16 16:37:19 We can always do a further release to sort out any bugs Aug 16 16:37:51 zecke: reverse depenencies (things that depend on a given task) iirc Aug 16 16:40:58 rb Aug 16 16:40:59 brb Aug 16 16:41:35 * RP -> back later Aug 16 16:53:17 03philippe 07org.oe.dev * r5e20d633... 10/ (7 files in 2 dirs): (log message trimmed) Aug 16 16:53:17 gtk-webcore: move fixed svn date to a newer and compiling date Aug 16 16:53:17 * rename "packages/gtk-webcore/osb-jscore_20060212.bb" Aug 16 16:53:17 to "packages/gtk-webcore/osb-jscore_20070816.bb" Aug 16 16:53:17 * rename "packages/gtk-webcore/osb-nrcit_20060212.bb" Aug 16 16:53:17 to "packages/gtk-webcore/osb-nrcit_20070816.bb" Aug 16 16:53:19 * rename "packages/gtk-webcore/osb-nrcore_20060212.bb" Aug 16 16:59:58 * * OE Bug 2358 has been RESOLVED (FIXED) by philippedeswert(AT)scarlet.be Aug 16 17:00:00 * *  osb-nrcit-0.5.0+cvs20060212-r0 fails to compile Aug 16 17:00:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2358 Aug 16 17:01:58 * * OE Bug 1432 has been RESOLVED (WORKSFORME) by philippedeswert(AT)scarlet.be Aug 16 17:02:00 * *  gpe-image fails to find glibc-localedata-i18n Aug 16 17:02:02 * * http://bugs.openembedded.org/show_bug.cgi?id=1432 Aug 16 17:05:20 Hi zecke, in http://www.honigmond-berlin.de/ there are 2 hotels (2 address)... which one should I choose? Aug 16 17:05:49 honigmond or gargen? Aug 16 17:07:21 asking my colleague Aug 16 17:08:16 all slept in honigmond. if you get a room torwards the garden you should try to sleep there Aug 16 17:08:28 but both are close to the office Aug 16 17:08:44 vivijim: how far are you with creating debian source packages? Aug 16 17:09:16 zecke: thanks, I'll try that... Aug 16 17:10:03 zecke: about the sources I'm having some problems right now because I changed to use oe_unpack function... Aug 16 17:29:04 good evening Aug 16 17:47:58 * * OE Bug 1993 has been RESOLVED (FIXED) by philippedeswert(AT)scarlet.be Aug 16 17:48:00 * *  ntop 3.0 fails do_configure Aug 16 17:48:02 * * http://bugs.openembedded.org/show_bug.cgi?id=1993 Aug 16 17:48:10 re Aug 16 17:49:58 * * OE Bug 2038 has been RESOLVED (WORKSFORME) by philippedeswert(AT)scarlet.be Aug 16 17:50:00 * *  gtk-webcore (osb-nrcore_0.5.0.bb) doesn't compile against gtk+2. 6.10 Aug 16 17:50:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2038 Aug 16 17:59:15 03lenehan 07org.oe.dev * rd2ff8556... 10/ (3 files in 3 dirs): Aug 16 17:59:15 ntop: fix compilation issue (bug #1993) Aug 16 17:59:15 * applied patch by Jamie Lenehan Aug 16 18:05:15 re Aug 16 18:05:40 woglinde: wb Aug 16 18:18:59 * * OE Bug 2040 has been RESOLVED (WORKSFORME) by philippedeswert(AT)scarlet.be Aug 16 18:19:01 * *  gnome-doc-utils_0.7.1.bb build failed (depend on xlstproc) Aug 16 18:19:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2040 Aug 16 18:27:09 03likewise 07org.oe.dev * rae5744fa... 10/ (1 packages/linux/linux-x86_2.6.20.bb): linux-x86-2.6.20: Use ARCH override, not MACHINE override to bring in correct defconfig. Should be safe for current i?86-generic machines. Aug 16 18:28:35 ^^ ehrm, will this commit still work w/ the current bb issues? Aug 16 18:29:29 whahahaaa???? Aug 16 18:29:49 why the heck we have an extra -x86 recpie for the kernel? Aug 16 18:35:30 does ${DATETIME} change during the build, or is it set once and used for the entire build? Aug 16 18:37:09 cbrake: documentation.conf:DATETIME[doc] = "The date and time the build was started at" Aug 16 18:37:34 pH5: excellent, thanks! Aug 16 18:39:41 RP: any idea why bbimage could show me this nice "python-pysqlite2' missing error? Aug 16 18:42:00 woglinde: good question. shall we drop it? Aug 16 18:43:11 likewise hm look at linux_2.6.20 linux_2.6.21 linux_2.6.22 Aug 16 18:44:41 I dont see any special Aug 16 18:44:50 woglinde: me neither, it seems like split effort Aug 16 18:44:51 for x86 to have his own .bb Aug 16 18:44:53 split=dual Aug 16 18:45:10 koen: I think I have fixed bonsai, need to update the db now... Aug 16 18:46:11 koen is supposedly away to the Lowlands music fest with unlimited beer to his availability. Aug 16 18:46:27 likewise hm he was here Aug 16 18:46:32 at afternoon Aug 16 18:47:23 I know, he heard him say bye on angstrom. Aug 16 18:47:24 hmm... Aug 16 18:47:29 student life Aug 16 18:47:42 he'll be back tuesday, he'll alive on thursday Aug 16 18:47:48 ^be Aug 16 18:48:18 zecke: yeah, been there 15 years or so but they wouldn't let me stretch it any longerr Aug 16 18:49:05 likewise lol Aug 16 18:54:34 umm beer Aug 16 18:54:59 people wonder why I do not want to actually get a real job Aug 16 18:55:54 crofton you should Aug 16 18:56:38 do_fetch beer Aug 16 18:56:44 task finished Aug 16 19:04:57 well, you can always pay me to hack on OE Aug 16 19:05:26 * Crofton does seek consutling Aug 16 19:14:22 later Aug 16 19:14:33 ok, I have hacked at uboot-utils so that it only builds mkimage Aug 16 19:14:39 and that it really does this Aug 16 19:16:58 hey zecke Aug 16 19:17:15 crofton hehe Aug 16 19:17:32 ok, lets see if the gumstix build completes now Aug 16 19:17:41 without having to build for OSK first Aug 16 19:18:05 long term, I need to have it build u-boot for the gumstix, which also solves the mkimage problem Aug 16 19:22:23 re khem Aug 16 19:22:33 yo Aug 16 19:25:38 * RzME I'm trying to build angstrom-console-image for collie and I'm getting this error: Aug 16 19:25:39 | ERROR: Cannot satisfy the following dependencies for task-base-extended: Aug 16 19:25:40 | bluez-dtl1-workaround Aug 16 19:25:56 Does anyone know how to fix this? :P Aug 16 19:28:12 RzME: bitbake bluez-dtl1-workaround Aug 16 19:29:13 likewise, ping Aug 16 19:29:20 Philippe: building in progress... Aug 16 19:29:37 Philippe: It worked! Thank you! :D Aug 16 19:30:22 Crofton: pong Aug 16 19:30:57 I have been callign for libtool-1.5.24 in my local.conf Aug 16 19:31:34 and can build for omap5912osk, omap2430sdp, gumstix, davinci-dvevm Aug 16 19:31:43 just Efika fails Aug 16 19:31:55 I still need to try and exxternal tool chain with Efika Aug 16 19:32:36 how to handle AC_CHECK_PYMOD in OE? I tried to build avahi's pygtk programs, but configure fails because the build env's python can't import gtk (there is no python-pygtk-native) Aug 16 19:32:37 RzME: you're welcome Aug 16 19:41:56 * khem sharpens his axe to hack eglibc cross locale generation Aug 16 19:42:50 ph5 kick mickeyl that python-pygtk-native is build Aug 16 19:48:02 woglinde: Thinking about it, pygtk is not really a build dependency at all, it's a runtime dependency. Aug 16 19:48:51 hm ah Aug 16 19:48:55 http://en.pastebin.ca/660135 should do the trick. I can make sure the modules are there via RDEPENDS. Aug 16 19:49:13 yes Aug 16 19:49:17 RDEPENDS Aug 16 19:54:00 Khem: did you work with powerpc toolchains recently? we need a good reference toolchain to test against. Aug 16 19:54:31 whee, that means I could probably enable the avahi pygtk stuff in the standard build without hurting anyone. Aug 16 19:54:51 likewise: yes infact you name an arch :) Aug 16 19:56:54 Khem: mpc5200 specific in this case, is that powerpc e300 arch? Aug 16 19:59:14 likewise: it is 603e based Aug 16 19:59:24 likewise: has double precision floating point Aug 16 20:00:16 khem, we have a specific problem in our Efika build Aug 16 20:00:19 mpc is not that well supported on glibc Aug 16 20:00:30 What kind of problem Aug 16 20:00:38 C++ apps have trouble linking with the commands from libtool Aug 16 20:00:57 we have a hack for 1.5.10 add LDFLAGS=-lstdc++ Aug 16 20:01:06 that gets that flag passed through libtool Aug 16 20:01:07 ah hmmm what toolchain combination do you use Aug 16 20:01:08 but Aug 16 20:01:22 when we got to libtool 1.5.24, the hack stops working Aug 16 20:01:30 now without the hack, Aug 16 20:01:38 1.5.24 works fine on ARM builds Aug 16 20:01:50 gcc-4.1.2 I think Aug 16 20:02:06 standard angstrom version Aug 16 20:03:11 hmmm ok Aug 16 20:03:29 I would say libtool problem Aug 16 20:03:34 well Aug 16 20:03:43 that is what I am trying to disprove Aug 16 20:03:44 are you using cross-libtool Aug 16 20:03:53 libtool-cross Aug 16 20:04:04 yeah libtool might have issue with ppc Aug 16 20:05:37 Actually I did not use latest and greatest libtool Aug 16 20:05:53 old libtool works, with hack Aug 16 20:05:58 new libtool does not Aug 16 20:06:53 it is one of these magic combination of versions and hw Aug 16 20:06:54 lemme see if I have something in my notes Aug 16 20:07:00 thanks Aug 16 20:15:14 Crofton: what is exact thing you are trying Aug 16 20:15:25 commandline options Aug 16 20:15:38 khem Aug 16 20:15:57 I compare the command lines for OSK and Efika and they seem the same Aug 16 20:16:08 gcc is used to link Aug 16 20:16:33 and they specify the stdc++ lib by giving path to libstdc++.so Aug 16 20:16:39 actually libtool decides how it should invoke gcc Aug 16 20:16:44 not by using -lstdc++ Aug 16 20:16:51 hmmm Aug 16 20:17:02 libtool does nostdlib Aug 16 20:17:17 right to gcc is not allowed to look in standard paths Aug 16 20:17:25 (s/to/so/) correction: right so gcc is not allowed to look in standard paths Aug 16 20:18:36 basically, I need to install something like the Denx tool chain and see what I get from that Aug 16 20:19:18 Can you give a commandline I can try here Aug 16 20:19:30 but I use gcc 4.2.1 Aug 16 20:19:46 * Crofton wonders if we have bb files for 4.2.1 Aug 16 20:21:08 we have Aug 16 20:21:10 koen Aug 16 20:21:19 does it all Aug 16 20:21:42 khem, http://pastebin.ca/660185 Aug 16 20:21:46 is a problem line Aug 16 20:21:57 same bb file builds for ARM on OSK Aug 16 20:22:45 ah, angstrom-2008.1 Aug 16 20:23:02 I will try that with Efika tonight Aug 16 20:23:09 just in case Aug 16 20:49:14 Crofton: Can you add -t to the g++ commandline and paste output Aug 16 20:49:54 Not right now, I have some other builds running Aug 16 20:50:07 this is to the one that comes from libtool? Aug 16 20:51:27 I can arrange for -t to go to the g++ lines Aug 16 20:52:59 What does -t do? Aug 16 21:08:49 Crofton: it shows trace of where compiler picked libraries and objects frm Aug 16 21:09:05 while linking Aug 16 21:09:12 ok Aug 16 21:09:34 this will clearly tell us what gcc is doing Aug 16 21:09:40 I need to finish a test on a gumstix build, then I am going to fire off another Efika build Aug 16 21:09:42 with the stuff it is being fed Aug 16 21:09:50 ok Aug 16 21:10:02 what is efika btw. Aug 16 21:10:28 oh its a board name Aug 16 21:10:32 They had a program where they handed them out for free Aug 16 21:10:33 http://www.genesippc.com/efika.php Aug 16 21:11:45 do they do it still Aug 16 21:11:58 (the program) Aug 16 21:12:13 no, but the board is now $99 Aug 16 21:18:28 birunko: ? Aug 16 21:54:32 03crofton 07org.oe.dev * r166c7981... 10/ (1 packages/uboot/uboot-utils_1.2.0.bb): uboot-utils : Eliminate non-working builds. Package now builds and installs mkimage. Aug 16 21:54:39 03crofton 07org.oe.dev * re1d538a4... 10/ (1 packages/linux/linux-gumstix_2.6.15.bb): linux-gumstix : Add dependency on uboot-utils. This builds mkimage so kernel build completes. Aug 16 22:10:05 ~botmail for zecke You're using python from staging and need the fixes in OE.dev Aug 16 22:27:23 bother glibc-initial dies on some patches with angstrom-2008 Aug 16 22:27:37 fine time for Koen to go to a festival Aug 16 22:36:58 good nite all Aug 16 22:48:58 * * OE Bug 2746 has been RESOLVED (FIXED) by philipp.zabel(AT)gmail.com Aug 16 22:49:00 * *  libgsmd_svn.bb uses SRCREV instead of SRCDATE Aug 16 22:49:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2746 Aug 16 22:50:07 Hmm. We have another core bitbake function getting called 28,000 times :} Aug 16 22:54:23 well, if it was only called a few times, would it really be a function? Aug 16 22:55:28 Crofton|work: Each time this function gets calls it loads state and then saves state and its called in a double for loop though Aug 16 22:56:39 :) Aug 16 22:56:54 I suspect being called that much makes a candidate for a lot of study Aug 16 22:57:31 Crofton|work: Yes, I know it can be rewritten more efficiently Aug 16 23:00:24 03stefan 07org.oe.dev * r533bcb71... 10/ (5 files in 4 dirs): Aug 16 23:00:24 matchbox-keyboard-inputmethod: Add matchbox-keyboard variant with gtk Aug 16 23:00:24 input-method support. (modified version from poky) Aug 16 23:00:29 03stefan 07org.oe.dev * r84bfe56f... 10/ (1 packages/tasks/task-openmoko.bb): task-openmoko: Use matchbox-keyboard-inputmethod Aug 16 23:25:13 6% speedup and a reduction in function calls from 8.8 million to 8.2 million... Aug 17 00:10:15 03rpurdie * r907 10bitbake/lib/bb/ (build.py parse/parse_py/BBHandler.py): BBHandler.py/build.py: Rework add_task to be more efficient (6% speedup, 7% number of function calls reduction) Aug 17 02:05:58 * * OE Bug 2793 has been created by jjpfeifferjr(AT)comcast.net Aug 17 02:06:00 * * change SRC_URI Aug 17 02:06:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2793 **** ENDING LOGGING AT Fri Aug 17 02:59:56 2007