**** BEGIN LOGGING AT Sun Nov 07 02:59:58 2010 Nov 07 05:14:49 can we use ccache? Nov 07 05:15:00 (or do we?) Nov 07 05:46:31 The recipe for glibc is broken now. Nov 07 05:47:30 which? Nov 07 05:48:31 glibc_2.9.bb Nov 07 05:50:30 did it change in the last couple days? i haven't updated to the HEAD of master in a couple days, but i've built that recipe several times for the BB xM using angstrom 2008 Nov 07 05:50:59 how is it failing for you? Nov 07 05:51:35 http://pastebin.com/LuQ9qUi3 Nov 07 05:52:41 hmm.. Nov 07 05:52:48 is there any additional info in config.log? Nov 07 05:52:59 DISTRO_VERSION = "2010.7-test-20101107" Nov 07 05:53:28 are you building angstrom? Nov 07 05:53:59 Yes Nov 07 05:54:00 i've tried 2010.7, but haven't heard of -test-20101107 yet. i guess that's a testing branch? Nov 07 05:54:21 No idea. It's what comes up when I build. Nov 07 05:54:53 oh ok Nov 07 05:55:10 hmm, were you able to find that config.log file it references towards the end of your pastebin? Nov 07 05:56:17 I found the log. Nov 07 05:57:02 anything interesting in there? Nov 07 05:57:30 it looks like it found your compiler binary just fine from the test where it sees if it will accept -g, but don't know why -E would fail a sanity check in that case Nov 07 05:58:08 I don't know. Didn't see anything else interesting in the log. Nov 07 05:58:15 you wouldn't happen to have had a previous build fail would you? a couple times if i ctrl+c out of a build halfway through, or my disk fills up, or whatever i have to bitbake -c clean for a bunch of stuff Nov 07 05:58:34 it didn't show the command it ran through gcc -E or why it failed the sanity check? bummer Nov 07 05:59:21 I really don't know. Nothing jumped out at me. I ctrl/c a build a while ago. Nov 07 05:59:49 i could take a look at the config.log if you pastebin'ed it Nov 07 06:00:04 I've had builds fail on glibc, yes. Nov 07 06:01:09 otherwise i'd say to wipe out as big a chunk of the build as you're comfortable with (with bitbake -c clean or bitbake -c distclean) and start over; a common failure mode is that a package gets to the staging process, copies some files over to your sysroots directories, then even if you remove that package or change versions or rebuild or whatever, it still screws with subsequent builds Nov 07 06:01:30 http://pastebin.com/BQi0n8ic Nov 07 06:01:48 things like the kernel, libc, and gcc are particularly touchy, lots of things need to go right for them all to compile correctly :P Nov 07 06:01:56 I'm no comfortable wiping any part of a build out. Nov 07 06:02:36 was that file called "config.log", or was it a bitbake log file (i.e., log.do_configure.)? Nov 07 06:02:49 I'm just trying to do builds I've done before. Nov 07 06:03:12 I don't know what it was called. Nov 07 06:03:29 this thing that you just pastebin'ed (the second one), where did you get it? Nov 07 06:03:37 From the log Nov 07 06:03:58 it looks like the same stuff as the first pastebin. on line 21 of the pastebin, it tells you "See 'config.log' for more details" Nov 07 06:04:10 that's the file i was asking for, it will have more details on exactly what your problem is Nov 07 06:04:31 glibc-2.9-r37.4/temp/log.do_configure.25078 is the name of the log Nov 07 06:04:38 oh ok Nov 07 06:04:53 instead, go into glibc-2.9-r37.4/ and type "find -name "config.log" Nov 07 06:05:05 hopefully there's only one Nov 07 06:05:27 yeah, the files in temp/ are bitbake's own log files which show a log of the shell commands bitbake is executing Nov 07 06:05:42 this config.log is a separate output file of one of those shell commands, ./configure Nov 07 06:06:23 I don't understand what you want. Nov 07 06:06:40 cd into glibc-2.9-r37.4/ Nov 07 06:06:48 type: Nov 07 06:06:53 find -name "config.log" Nov 07 06:07:02 it should hopefully give you a single filename Nov 07 06:07:15 open that file with a text editor, or output its contents to the terminal via cat, and pastebin it Nov 07 06:07:38 if it gives you no files or multiple files, we can debug further Nov 07 06:08:32 I found what you want but how to paste all that?? Nov 07 06:08:46 how big is it? Nov 07 06:09:48 It's several screen full. Nov 07 06:10:30 that's ok. if you can open it in a text editor that allows you to select all, (i.e. gedit, kate, jedit), do that, select all, and ctrl+v into a pastebin Nov 07 06:10:33 It's a config.log Nov 07 06:10:50 yes, i'm imagining that it is pretty verbose Nov 07 06:10:58 i'm guessing the part of interest will be towards the bottom Nov 07 06:11:14 if you feel confident picking out the part that matters (about it trying gcc -E and failing), just paste that part Nov 07 06:11:27 otherwise, shouldn't be too big a deal to find somewhere to copy/paste or upload it Nov 07 06:12:17 http://pastebin.ca/upload.php for instance allows upload of text files up to 150KB Nov 07 06:12:31 use ls -l to see how many bytes a file is Nov 07 06:12:33 I can't do this. Too much text. Nov 07 06:12:51 how big is that file? is it like 100MB? Nov 07 06:13:34 It's 34k Nov 07 06:13:49 is the file on your local machine or a remote server? Nov 07 06:14:04 It's on my machine, of course. Nov 07 06:14:39 i'm not understanding the problem. can you open the file in gedit, kate, jedit, soffice, nedit, etc. and select all? Nov 07 06:14:45 do you have an editor that can select all? Nov 07 06:14:51 No, I can't do that. Nov 07 06:15:13 can you help me understand why? no editor? Nov 07 06:15:49 You would not understand my way of doing things. I can't get to the file with another editor. Nov 07 06:16:28 another editor besides what? and what way of doing things? are you blind or similar and using a specialized editor? Nov 07 06:17:13 Hang on, I just got an idea. Nov 07 06:18:47 You're just going to have to accept that I can't get to the file with another editor. Nov 07 06:19:26 another editor besides what? Nov 07 06:20:07 not trying to be a dick, trying to see if there is any way that i can actually find out what your error message is Nov 07 06:20:29 if you can't do a make clean and try again and i can't see your error message, this is a complete non-starter Nov 07 06:21:53 alright, i'm done. Nov 07 06:21:56 good luck dealing with him. Nov 07 06:27:11 http://pastebin.com/XMTL4vAF <-- This is the best I can do. There is no more information I have to give. Nov 07 06:28:59 Figures he would leave just when I figured out how to get him what he was asking for. :( Nov 07 08:18:07 Hi, I am just starting out on an embedded project. Currently using Angstrom dist. I am evaluating what graphics systems to use (frame buffer based). I have looked at Qte, EFL and Nano-X. Any others that I should be looking at? Nov 07 08:31:07 guyvdb_, now is in between the folks in North America going to bed and Europe waking up on a Sunday...wait around a bit. Nov 07 08:35:52 :) thx Nov 07 12:43:50 stefan_schmidt, how did it go yesterday? Nov 07 12:46:46 ka6sox: quite good. We are down from 260 patches to 180 and that includes all the new ones that came in yesterday Nov 07 12:47:31 stefan_schmidt, This one could be removed no ? http://patchwork.openembedded.org/patch/2250/ Nov 07 12:49:05 pix|home: there are more outstanding stable patches Nov 07 12:49:23 pix|home: I will mention them in my status mail this evening Nov 07 12:49:43 pix|home: please leave it for now. I would like to hear from the stable mainatiners about it Nov 07 12:49:49 hi Nov 07 12:50:02 hi ericben Nov 07 12:50:20 stefan_schmidt: I've pushed several patches. I don't see more that I can push for now Nov 07 12:50:22 stefan_schmidt, I don't have admin rights so it will probably stay... :) Nov 07 12:50:36 ericben: cool, thanks Nov 07 12:50:39 pix|home: :) Nov 07 12:51:06 many patches are about classes and must be discussed (or trashed as since the time no one took care of them) Nov 07 12:51:25 ericben: indeed Nov 07 12:51:37 many lead to discussions which didn't get answers so I think they also can be trashed Nov 07 12:51:53 ATM I have three kind of patches that I will mention in my status mail Nov 07 12:52:04 stable -> stable maintainers Nov 07 12:52:20 core/classes -> More feedback discuss Nov 07 12:52:24 after 4 weeks without care a patch in patchwork could get a ping, and after 1 week wihtout care -> trash Nov 07 12:52:24 and the rest Nov 07 12:52:42 ericben: thats a bit short imho Nov 07 12:52:51 on the other hand we have patches from 2009 Nov 07 12:52:54 a total of 5 weeks ? Nov 07 12:53:25 ericben: not all people reading the list on a regular base Nov 07 12:53:36 patch -> discussions -> 4 weeks -> ping -> 1 week -> answer : ok, no answer ->trash Nov 07 12:53:39 ericben: but yes, 5 weeks should still be enough Nov 07 12:53:51 :) Nov 07 12:54:19 ericben: Personally I would like to have patches in only for 2 weeks actually Nov 07 12:54:37 I don't think this is possible though Nov 07 12:54:52 in the patch I applied yesterday, nearly all needed rework : sqash 4 patches into one, manually apply especially in linux_*.bb so, th emore we wait, the more work we have to do manually Nov 07 12:55:48 ericben: normally I ask the submitter to update and resent for such things Nov 07 12:56:00 ericben: that also rules out that he maybe already lost interest Nov 07 12:56:15 yes, but for some patches, submitter was already asked for fixes and never answered Nov 07 12:56:31 so in that case either we loose the patch or take time to update it. Nov 07 12:56:55 ericben: hmm, I would ping give him a moment and then remove Nov 07 12:57:04 frankly speaking, yesterday evening I was tired enough to update the patches but UI don't think I'll do that everytime :-D Nov 07 12:57:14 if it is something you are personally interested to get fixed reworking it may be also an option Nov 07 12:57:34 Should I resend a patch which have been deleted from patchwork ? (During the attempt to activate my patchwork account) (A patch for rng-tool fixing the initscript) Nov 07 12:57:53 pix|home: if it was deleted by accident, yes Nov 07 12:58:21 stefan_schmidt, Yep, I'll do it Nov 07 12:58:26 I've found someone submitted qwt patches in june, Koen acked the patches but didn't push them, then I wrote a recipe in october, then Koen sent an other recipe he wrote : which means 3 persons worked on the same recipe in 5 months of time only because the initial patch was not applied Nov 07 12:58:48 ericben: that is bad Nov 07 12:59:07 if they got acked we need to make sure they will land in the repo and don't rot on patchwork Nov 07 12:59:46 yes, but having 2 or 3 pages of pathces in patchwork prevent to correctly use the tool as it doesn't have asearch engine Nov 07 12:59:56 to filter patches which were acked or not Nov 07 13:00:42 ericben: agreed, thats why I want to get the count down to use it more effectively. With a patchwork weekend like this. :) Nov 07 13:00:43 so having a policy to handle the time a patch can stay in patchwork could be a good solution Nov 07 13:01:07 I'm going to talk to the server people about installing pw at osuosl Nov 07 13:01:09 ericben: And then people should use it on a regular base when doing reviews. Like khem is doing it Nov 07 13:01:24 http://pastebin.com/XMTL4vAF <--- Problem building glibc 2.9 Nov 07 13:01:29 Crofton|work: is it still on amethyest? Nov 07 13:01:29 and at the same time upatiing to bleeding edge registration and pw code Nov 07 13:01:32 yeah Nov 07 13:01:43 stefan_schmidt: yes that's a good solution, but I think we also nee da policy not to have < n months patches left in patchwork Nov 07 13:01:48 Crofton|work: oh, good to know. Thatnk for the admin work Nov 07 13:01:59 Crofton|work, I love those sarcasms Nov 07 13:02:20 I am hopeful we can find a registration module that needs emails conf to cut down on bs acconts Nov 07 13:02:21 ericben: agreed. I will mention this in my mail about the weekend. Or do you want to send out a mail about it? Nov 07 13:02:53 stefan_schmidt: go with your email, I can answer to it to provide more details if needed Nov 07 13:02:59 doing this a a regular basis should solve some of the problems you notice Nov 07 13:03:04 ericben: ok, in the evening Nov 07 13:03:12 winter is a good time to get people in the habit also Nov 07 13:03:56 Crofton|work: yup, doing a review round and get into the habit of updating patchwork directly Nov 07 13:04:05 Crofton|work: ye swe need to get the most heat from our CPU to warm the houses so it's a good time to test patches :-D Nov 07 13:04:05 yeah Nov 07 13:04:21 and people have more time in fornt of computers :) Nov 07 13:05:22 one habit to recall it that people who push patches mark them as applied in patchwork: this would save lot of work later trying to check if the patch was already applied Nov 07 13:05:48 this suppose all comiters have the rights in patchwork Nov 07 13:06:09 ericben: that and the submitter should also have a look Nov 07 13:06:12 Does anyone even care that the build for glibc is broken? Nov 07 13:06:27 ericben: I would say > 90% have it already Nov 07 13:08:14 RobotGuy: without further details nobody can comment on such a pastebin. What distro? For which machine did you build? Did you build from lastest master? etc Nov 07 13:08:19 RobotGuy: http://tinderbox.openembedded.net/packages/?status=Failed glibc seems not that broken Nov 07 13:08:26 RobotGuy: It did build fine for me yesterday Nov 07 13:08:41 stefan_schmidt: What other details are needed? I went through this last night. Nov 07 13:09:06 RobotGuy: And what was the outcome? Nov 07 13:09:17 ok have a nice afternoon Nov 07 13:09:22 RobotGuy: I don't have galss ball to know such things :) Nov 07 13:09:26 bye ericben|away Nov 07 13:10:12 I'm not going through this all over again. I'll figure it out myself or find another way to build software for my beagle. Nov 07 13:13:45 hi mickeyl Nov 07 13:14:03 mickeyl: I shoudl great you from the other two Nov 07 13:14:29 cool, thanks. everyone still alive? ;) Nov 07 13:14:32 stefan_schmidt, we are going to get pw running at osuosl Nov 07 13:14:48 what do you think of leaving the amethyts version availabel as a seperate ip Nov 07 13:14:58 and not moving over the old patches? Nov 07 13:15:08 mickeyl: yeah, we were sitting together yesterday in daniels new apartment Nov 07 13:15:20 stefan_schmidt: ah, cool. where is it? Nov 07 13:15:44 Crofton|work: thanks, sound like a good plan to me to clean up the old patches :) Nov 07 13:15:57 :) Nov 07 13:16:03 Crofton|work: we just need to make sure that the new patches are only reported to the new installation Nov 07 13:16:03 they will still be around Nov 07 13:16:29 Crofton|work: give it a 3 months period to get all applied that are of interest and then remove the installation Nov 07 13:16:39 mickeyl: ~ 700m from the old location Nov 07 13:16:54 mickeyl: directly over a dentist :) Nov 07 13:17:13 excellent, always healthy teeth Nov 07 13:17:19 heh Nov 07 13:17:21 hmm, can you hear screams? Nov 07 13:17:27 And the screams from the patients Nov 07 13:17:31 :) Nov 07 13:17:37 I don't think so :) Nov 07 13:23:33 Crofton|work: who are the active stable maintainers? You and koen? hrw|gone is no longer working on it Nov 07 13:25:10 I try to read patches for it and ack ones I understand Nov 07 13:25:26 ask on the list how uses it :) Nov 07 13:25:46 my work is all bleeding edge at the moment Nov 07 13:26:04 the users should really do the acks Nov 07 13:27:08 Crofton|work: ok, I will also bring this up Nov 07 13:27:31 yeah, it would be good to talk about Nov 07 13:27:38 * stefan_schmidt needs to do some other work the next hours. Will be back later today Nov 07 13:27:38 the basic principle is good though Nov 07 13:28:12 one of the goals is to avoid acks from random people who have no interest in wheter or not stbale actually works Nov 07 13:29:45 Crofton|work: I understand. But normally stable patches do not get acks at all over the last months. The java stuff was an exception Nov 07 13:30:27 I already stopped to get out BugLabs stuff from the overlay into stable as it is not worth the effort. We will hopefully go with future releases anyway :) Nov 07 13:30:59 Not a real problem if patches get rejected for stable. Just no reaction is bad. But thats not stable specific. Nov 07 13:31:18 OE needs to learn better to welcome new contributors :) Nov 07 13:32:36 and we need metadata that creates usable images Nov 07 13:32:45 and doesn't break package upgrade paths Nov 07 13:32:50 etc Nov 07 13:32:52 heh, that as well Nov 07 13:32:56 it is a hard problem Nov 07 13:33:03 I know Nov 07 13:33:11 Nothing to fix in a wekk Nov 07 13:33:13 week Nov 07 13:33:21 we just need to set the right route Nov 07 13:33:22 right :) Nov 07 13:33:58 Anyway, have to work on a compiler for my studies. Getting my grammar parsed. cu later Nov 07 13:35:48 awesome Nov 07 13:35:51 good luck Nov 07 15:06:34 03Martin Jansa  07master * r1a841b207a 10openembedded.git/recipes/xorg-xserver/ (3 files in 3 dirs): Nov 07 15:06:34 xserver-xorg-conf: om-gta0* enable right-click emulation and disable AutoAddDevices Nov 07 15:06:34 Signed-off-by: Martin Jansa Nov 07 16:14:26 khem: http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=5142b50dab9731fb0c508f8e24666865e2b84e4c ? :-D Nov 07 16:15:26 khem: just hint, subject is saying it's vlc patch, but you've commited my lua-lanes patch Nov 07 16:17:21 khem: and pls don't push that lua-lanes recipes yet, because it's not 100% working, thanks Nov 07 16:35:14 Hi all Nov 07 16:35:29 i send now a patch to openembedded on distro metano Nov 07 17:14:41 Question: I am writing recipe which source has local includes of apache2 headers. But it fails on make because can not find them. Nov 07 17:15:08 I added this line: CFLAGS += "-I${STAGING_INCDIR_NATIVE}/apache2" Nov 07 17:15:50 But I still get errors that those includes are not found. Nov 07 17:15:53 Any hints? Nov 07 17:16:02 did you tried include on recipe a deppends parameter? Nov 07 17:16:06 that depends apache Nov 07 17:16:17 DEPENDS = "apache2-native" Nov 07 17:16:58 DEPENDS += "apache2" Nov 07 17:17:06 DEPENDS += "apache2-native" Nov 07 17:17:26 try with apache2 Nov 07 17:17:32 Give me a minute Nov 07 17:19:22 ok Nov 07 17:19:26 Same Nov 07 17:19:55 All files are in required directories Nov 07 17:20:01 But it still can not see them. Nov 07 17:20:08 !paste Nov 07 17:20:09 can you give the output of errors? Nov 07 17:20:15 Oh, Nov 07 17:20:18 Give a sec Nov 07 17:21:56 http://pastebin.com/3qevfq0w Nov 07 17:22:00 recipie and errors Nov 07 17:22:57 fcgi.h has includes like this: #include "httpd.h" Nov 07 17:23:03 give-me a sec Nov 07 17:24:58 hum... Nov 07 17:25:06 i think is necessary do_compile Nov 07 17:25:49 isn't compilation done by oe_runmake? Nov 07 17:26:50 NOTE: package modfastcgi-2.4.6-r0: task do_compile: Succeeded Nov 07 17:27:05 I think it's inherited from autotools Nov 07 17:27:12 * davidlt new to this thing Nov 07 17:27:40 i'm too Nov 07 17:27:41 but Nov 07 17:27:44 see Nov 07 17:27:49 do_compile success Nov 07 17:27:57 that's because is nothing to compile Nov 07 17:28:08 if you see the log of log.do_compile_balbalbal Nov 07 17:28:18 there's nothing on file Nov 07 17:28:18 It has Nothing to compile: Nov 07 17:28:45 davidlt, you're writing a recipe that include native haaders? Nov 07 17:28:54 that doesn't sound right Nov 07 17:29:03 try this :http://angelox_123.pastebin.com/Xx9Aj9LR Nov 07 17:29:11 hi GNUtoo|laptop Nov 07 17:29:37 angelox_123, hi Nov 07 17:29:49 angelox_123: it does not have configure script, this is why copying of Makefile is done. Nov 07 17:30:06 davidlt: i'll see the source :) Nov 07 17:30:07 davidlt, try without native stuff, try target INCDIR Nov 07 17:30:13 GNUtoo|laptop: This is what I found on other recipes. Nov 07 17:30:21 ah? Nov 07 17:30:37 what's the software? Nov 07 17:30:39 I found in on modphp I think. Nov 07 17:30:46 fastcgi Nov 07 17:30:47 ok Nov 07 17:31:12 And I am trying to write modfastcgi, because I can not make modphp build as I want. Nov 07 17:31:48 It depends on libxml native and still has on include armv version of libxml2. Nov 07 17:31:58 And I can not make that to happen. Nov 07 17:32:43 bbl Nov 07 17:37:20 Trying renamking, making do_configure() and do_compile() Nov 07 17:37:51 Also changed dependence to apache2 and including not native headers. Nov 07 17:38:07 Same result. Nov 07 17:40:52 Any suggestions? Nov 07 17:43:39 If I change to include still not found. Nov 07 17:43:42 hi, which package name do I have to set as a DEPENDS in a recipe that needs fortran? Nov 07 17:44:21 I mean, which packages provides a fortran compiler? Nov 07 17:47:11 I don't use fortran, but I think gnu compiler provides that? Nov 07 17:47:34 03Simon Busch  07org.openembedded.dev * r4e97bc336d 10openembedded.git/recipes/linux/ (linux-palmpre_2.6.24.bb linux-palmpre_git.bb): Nov 07 17:47:34 linux-palmpre: add kbd-native to DEPENDS Nov 07 17:47:34 Signed-off-by: Simon Busch Nov 07 17:48:20 davidlt, I'm getting this: configure: error: cannot compile a simple Fortran program Nov 07 17:48:35 btw, I'm trying to adda recipe for R Nov 07 17:49:54 and here's the log: http://paste.pocoo.org/show/287319/ Nov 07 17:52:24 pespin: can you give recipe? Nov 07 17:52:38 angelox_123, http://paste.pocoo.org/show/287319/ Nov 07 17:52:44 thanks Nov 07 17:52:50 angelox_123, ops sorry Nov 07 17:52:53 this one: http://paste.pocoo.org/show/287321/ Nov 07 17:53:00 angelox_123: have you checked modfastcgi? Nov 07 17:53:02 ok Nov 07 17:53:16 davidlt: no sorry,i was seeing another thing on my home Nov 07 17:53:19 i'll see soon Nov 07 17:53:24 angelox_123, r_2.12.0.bb <- file name :) Nov 07 17:53:33 set it in recipes/r Nov 07 17:54:03 ok pespin.... Nov 07 17:54:06 i need go Nov 07 17:54:08 later i back Nov 07 17:54:10 ok? Nov 07 17:54:13 sorry peoples Nov 07 17:54:16 ok :) Nov 07 17:58:17 Compilcated Nov 07 18:12:25 Crofton or Crofton|work how did you fix the "usrp_prims.cc:244 : internal error!" for native GNU Radio build? i'm using lsusb-0.1.12 but that ain't helping Nov 07 18:31:06 how to define recipe type? coss / native? Nov 07 18:38:08 davidlt: Nov 07 18:38:20 see the example of zip recipe Nov 07 18:38:29 that shows how to use FLAGS Nov 07 18:39:20 Gonna try, thanks for input Nov 07 18:39:43 np Nov 07 19:04:42 ynezz: re lua stuff. Its fine to let it in there if you are going to fix it soon enough as there may not be larger user base yet. But if you wont then I will revert it Nov 07 19:05:36 morning khem Nov 07 19:06:41 ka6sox: hi Nov 07 19:07:08 bbiaf Nov 07 19:12:52 khem: but the commit is misleading, isn't it? Nov 07 19:13:06 khem: you've commited it as vlc, not as lua-lanes, or am I mistaken? Nov 07 19:14:32 khem: it should've been lua-lanes: add new recipe instead of 'vlc_1.1.4.1: Why is `lua5.1-native` not added to `DEPENDS` automatically?' :) Nov 07 19:14:47 khem: that's what I've meant Nov 07 19:14:54 anyway, beer time, bbl Nov 07 19:28:43 re Nov 07 19:30:00 angelox_123: mod_fastcgi almost working! Nov 07 19:33:10 good davidlt :) Nov 07 19:33:19 angelox_123: It compiles but prblems with install Nov 07 19:33:25 hum.... Nov 07 19:33:31 oe_runmake -f Makefile.AP2 top_dir=${STAGING_DIR}/${MULTIMACH_TARGET_SYS}${exec_prefix} DESTDIR=${D} Nov 07 19:33:34 fails Nov 07 19:33:45 can you pastebin the error and actual recipe? Nov 07 19:33:57 Oh at the end should be "install" Nov 07 19:34:28 | make[1]: *** No rule to make target `local-install'. Stop. Nov 07 19:34:28 | make[1]: Leaving directory `/home/gum2011/overo-oe/tmp/work/armv7a-angstrom-linux-gnueabi/modfastcgi-2.4.6-r1/mod_fastcgi-2.4.6' Nov 07 19:34:38 I tried it manually and it worked Nov 07 19:35:35 can you pastebin the actual recipe? Nov 07 19:37:08 So far looks like this: http://pastebin.com/4EbeHLdz Nov 07 19:37:28 But it requires apache2 modification (which I didn't do yet) Nov 07 19:37:43 hum... i think i know the error Nov 07 19:37:46 Just copy i686-linux/usr/build to armv7.... Nov 07 19:38:52 mod_fastcgi includes some apache2 *.mk files, which are ir ../usr/build directory and later uses them. Nov 07 19:39:05 So local-install should be defined in those files. Nov 07 19:39:18 rules.mk maybe. Nov 07 19:39:47 strange Nov 07 19:39:52 it likes ok the file Nov 07 19:40:34 am? Nov 07 19:41:06 i've submited a correction Nov 07 19:41:08 try that Nov 07 19:41:39 Hmm.. I just added to DESTDIR to EXTRA_OEMAKE and it failed with the same error on compile. Nov 07 19:43:13 What was the correction? Nov 07 19:46:44 at the same link that you gave for me Nov 07 19:46:57 ops Nov 07 19:46:57 http://pastebin.com/6US55RBU Nov 07 19:47:00 sorry Nov 07 19:47:00 :) Nov 07 19:48:37 same Nov 07 19:50:34 strange Nov 07 19:50:45 wait more help while i see more recipes Nov 07 19:54:57 I could do install myself, still there is only one .so file generated. Nov 07 19:57:04 angelox_123: I know the problem. Nov 07 19:59:08 No I don't :( Nov 07 20:00:19 It's something with make files. Nov 07 20:00:35 If I rename Makefile.A2 to Makefile and relaunch, it works. Nov 07 20:00:50 even thought that make runs with -f Makefile.A2 Nov 07 20:18:59 hehe.. Nov 07 20:19:33 bitbake goes crazy when specifying wrong DISTRO Nov 07 20:19:47 no user-readable diagnostic Nov 07 20:22:02 Jay7, I blew it when I was building recently and was puzzled for a few minutes Nov 07 20:22:07 same issue Nov 07 20:24:45 well.. seems non-absolute TMPDIR does alike thing Nov 07 20:30:14 al_fayez, I haven't done a native build in a while Nov 07 20:30:31 I should do n one on Tuesday Nov 07 20:30:43 but I will be using an overo tide with 512M of RAM Nov 07 20:34:11 GNUtoo|laptop: Hi. Nov 07 20:34:20 angelox_123, did you take a loook at the fortan issue? :) Nov 07 20:34:31 I was pinged for you 2 day ago Nov 07 20:35:20 GNUtoo|laptop: ^ Nov 07 20:41:44 may I place my testbuilder script into contrib too? Nov 07 20:42:06 what is our policy for this? Nov 07 20:45:06 Jay7 if they don't want it I do. Nov 07 20:45:09 :D Nov 07 20:45:40 as I'm trying to work on testing and autobuilder stuffs. Nov 07 20:46:10 well.. I'll push it then after some code commenting and cleanup Nov 07 20:47:53 03Roman Khimov  07org.openembedded.dev * r971a58b486 10openembedded.git/recipes/libprelude/ (libprelude-1.0.0/fix-ltdl-hack.patch libprelude_1.0.0.bb): (log message trimmed) Nov 07 20:47:53 libprelude: workaround build failures with libtool 2.4 Nov 07 20:47:53 Like: Nov 07 20:47:53 In file included from prelude-plugin.c:56:0: Nov 07 20:47:53 /include/prelude-plugin.h:66:20: error: conflicting types for 'lt__PROGRAM__LTX_preloaded_symbols' Nov 07 20:47:54 /oe/build-angstrom-next/angstrom-dev/sysroots/armv7a-angstrom-linux-gnueabi/usr/include/ltdl.h:106:36: note: previous declaration of 'lt__PROGRAM__LTX_preloaded_symbols' was here Nov 07 20:47:55 make[4]: *** [prelude-plugin.lo] Error 1 Nov 07 20:50:25 Writing script for mod_fastcgi (not native) Nov 07 20:50:32 using oe_runmake Nov 07 20:50:38 At the end I get: modules/mod_fastcgi.so: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked, not stripped Nov 07 20:50:47 My architecture is ARMv7 Nov 07 20:50:52 fidencio, hi Nov 07 20:51:04 This made native, why is that? Nov 07 20:51:16 fidencio, you are the one who wants to port emotion+lightmediascanner+enjoy? Nov 07 20:51:47 GNUtoo|laptop: yeap, I'm. Nov 07 20:52:07 fidencio, I've done some recipes for that locally but they are not quality recipe so they are not commitable Nov 07 20:52:09 Enjoy? hehehe. I don't have port to enjoey yet, but I can do. Nov 07 20:52:19 but I already ran enjoy on my device Nov 07 20:52:38 GNUtoo|laptop: Which device? Nov 07 20:52:44 fidencio, what's the status of lightmediascanner? do you have a recent recipe? Nov 07 20:52:57 fidencio, I've many devices, I tried on n900 and nexusone Nov 07 20:53:08 altough I failed at making it output to sdl with nexusone Nov 07 20:53:19 GNUtoo|laptop: can you make a video video or something like this? :-) Nov 07 20:53:50 fidencio, yes if you want, a bit slow because I've a lot of music with a cover each time Nov 07 20:54:08 maybe 16/32/16+32 G Nov 07 20:54:15 GNUtoo|laptop: I'm using a recipe for lms 0.4. But i think that profusion is the only maintainer of lms, for now. or i'm wrong Nov 07 20:54:27 ok Nov 07 20:54:31 let me find my recipes Nov 07 20:55:14 fidencio, btw is there a way to have a progressbar, and normal play controls? Nov 07 20:55:20 in enjoy Nov 07 20:55:31 because it's so great Nov 07 20:55:34 but it lacks that Nov 07 20:55:46 also it uses too much CPU Nov 07 20:56:51 http://pastebin.com/MiJ15LXV Nov 07 20:57:15 so for lightmediascanner I've an issue Nov 07 20:57:19 there is a test binary Nov 07 20:57:22 named test Nov 07 20:57:29 which I'd like to package Nov 07 20:57:32 and compile Nov 07 20:57:33 etc... Nov 07 20:57:42 I didn't see a makefile target for that Nov 07 20:57:48 but I don't remember well Nov 07 20:58:02 the test binary is necessary for enjoy Nov 07 20:58:20 else you can't populate the sqlite library Nov 07 20:58:25 Happy patchwork beer :) Nov 07 20:58:32 also I had to use a special command Nov 07 20:58:39 because the defult one didn't work Nov 07 20:58:44 I had to use only 1 thread Nov 07 20:58:46 Question: what do to if make uses libtool from native, not for armv? Nov 07 20:58:47 for the test binary Nov 07 20:59:11 GNUtoo|laptop: let me see Nov 07 20:59:24 ok Nov 07 20:59:29 feel free to use my recipes Nov 07 20:59:49 just put my name somewhere(like taken from ,improved from or something like that) Nov 07 20:59:58 A laptop, A beer and a kitten, What else ? :) Nov 07 21:00:05 (in the commit message) Nov 07 21:01:02 GNUtoo|laptop: Why emotions depends of webkit-efl? Nov 07 21:01:10 -emotions +emotion Nov 07 21:01:14 fidencio, I don't remember Nov 07 21:01:23 I can look Nov 07 21:01:31 GNUtoo|laptop: not depends Nov 07 21:01:42 ok Nov 07 21:01:49 I think that you c&p and forgot the webkit-efl :-) Nov 07 21:03:13 hrrr hate bitbake Nov 07 21:03:13 GNUtoo|laptop: please, force emotion and lms as DEPENDS to enjoy Nov 07 21:03:40 ok Nov 07 21:04:29 GNUtoo|laptop: which backend are you using with emotion? Nov 07 21:04:36 fidencio, gstreamer Nov 07 21:04:41 maybe that's why it's a bit slow Nov 07 21:04:51 but I think it's rather pixbuf loading Nov 07 21:04:53 GNUtoo|laptop: cool. vlc was disabled last week Nov 07 21:04:53 or something like that Nov 07 21:05:14 let me build it again Nov 07 21:05:20 GNUtoo|laptop: do you know Eve? Nov 07 21:05:30 yes Nov 07 21:05:32 the browser Nov 07 21:05:59 yeap :-). And I'm working on a pdf reader ... "Envision" Nov 07 21:06:20 I think that in some months we'll have a lot of efl apps working Nov 07 21:06:46 How to force oe_makerun use libtool from armv7, not native? Nov 07 21:08:37 GNUtoo|laptop: Dude, I need take a shower and go to dinner with my gf! Nov 07 21:08:53 ok Nov 07 21:09:07 fidencio, then let's talk later, ping me on irc later Nov 07 21:09:16 GNUtoo|laptop: Please, if you can make a video running EFLs apps in your devices ... send me: fidencio@profusion.mobi or fabianofidencio@gmail.com Nov 07 21:10:06 fidencio, ok I build and do a video Nov 07 21:10:10 GNUtoo|laptop: about the recipes ... are ok to me (less than what I sayd) Nov 07 21:10:17 note that I've a huge jamendo music collection Nov 07 21:10:26 each album has a cover Nov 07 21:10:31 so it's usable Nov 07 21:10:45 but sometimes it freeze while scrolling trough the huge quantity of albums Nov 07 21:10:52 I bet it wait for loading the cover Nov 07 21:10:53 Ok. I really wanna see it! Nov 07 21:11:30 GNUtoo|laptop: are you in #edevelop? If you found something wrong/strange with enjoy, ping acidx or otaviobp in #edevelop Nov 07 21:12:07 no, but I'm here Nov 07 21:30:43 hi folks - am creating .bb files for pythonwebkit and pywebkitgtk - am running into a teeeeny problem with the correct paths etc. for patch files. is there anyone with time to advise? Nov 07 21:31:12 hi, it seems something's wrong with libsdl-x11-1.2.14 recipe - it disables alsa support at all Nov 07 21:31:22 other than "cheat: create your own git clone and patch it there..." Nov 07 21:40:31 What kind of paths? Nov 07 21:40:48 Hi mrj10 Nov 07 21:42:10 ohhh Nov 07 21:42:15 there is a play button Nov 07 21:43:55 davidlt: quilt's bombing out saying it can't find webkit/webkit.defs so i modify the patch to add pywebkitgtk-1.1.8/ in front of "--- webkit/webkit.defs" Nov 07 21:43:58 and still no difference. Nov 07 21:44:55 i'm inclined to say f*** it and do a git clone, patch it there, but i'd really like to know how you're supposed to do patches of packages. looking at python-cairo.bb i "seem to be getting it right" but obviously am not... Nov 07 21:48:41 did you try cleaning the package? it may think it already applied that patch Nov 07 21:48:50 try bitbake -c distclean Nov 07 22:03:37 mrj10: no, it definitely hasn't been applied: this is a very specific patch to add a totally different set of python webkit bindings - http://www.gnu.org/software/pythonwebkit Nov 07 22:05:39 lkclz: maybe it has not in fact been applied, but bitbake may think it has. when you run bitbake, does it run the do_patch task? Nov 07 22:08:30 mrj10: i don't know how to tell.... y'know what? sod it - i'm converting to a pre-patched git repo :) Nov 07 22:09:07 i have to get this up-and-running as quickly as possible. bypass the problem *sigh*... Nov 07 22:09:33 ok. if you have modified the patch, running the command i suggested and then running bitbake again should fix it Nov 07 22:09:39 but whatever works for you Nov 07 22:09:51 but for future reference: how do i tell if do_patch was run? if i get an error "quilt" would it work? Nov 07 22:10:01 ahh, you can run individial packages, by name?? Nov 07 22:10:07 you're running "bitbake " right? Nov 07 22:10:20 it will say "running task X of Y... running task Z of Y.. etc." Nov 07 22:10:27 all of those tasks have names Nov 07 22:10:54 e.g.: Nov 07 22:11:01 NOTE: Running task 5472 of 10421 (ID: 2451, /media/d2/beagleboard/oebb/sources/openembedded/recipes/gnome/epiphany_2.30.2.bb, do_patch) Nov 07 22:11:39 look for a line that says "(ID: , /path/to/your/recipe.bb, do_patch)" Nov 07 22:12:12 if it's not running that, then bitbake's cache is telling it that it has already run do_patch for that package, and won't do it again Nov 07 22:13:26 so to force it to do it again, you need to do some combination of: 1) bumping the package revision (PR) variable in the recipe, 2) clearing the cache by running 'touch /path/to/local.conf', and 3) running bitbake -c patch or bitbake -c distclean Nov 07 22:13:53 yes, you can do individual recipes Nov 07 22:14:19 yaay. Nov 07 22:14:32 ok, but what if i get an error reported by quilt, each time? Nov 07 22:14:58 would that mean that it's trying to run do_patch and failing each time? Nov 07 22:15:42 usually yes, a quilt error means an error in do_patch. when a task fails, bitbake prints out " failed" and then the log file for that task Nov 07 22:15:49 so if you scroll up you should be able to see what task it was Nov 07 22:16:20 can you pastebin the error? Nov 07 22:16:30 ack. Nov 07 22:17:11 erk - not any more, but it basically said quilt err cannot patch webkit/webkit.defs it was really obtuse and not particularly obvious. Nov 07 22:17:28 the SRC was http://...../pywebkitgtk-1.1.8.tar.bz2 Nov 07 22:17:40 are you familiar with how the 'patch' utility works and how to read unified diffs? Nov 07 22:17:59 can you look at the patch it was puking on and figure out why that might not be matching up with the tarball you have? Nov 07 22:18:08 the patch was a diff to webkit/webkit.defs which i modified ... yes of course - been at this for 12 years :) ... to be pywebkitgtk-1.1.8/webkit/webkit.defs Nov 07 22:18:13 _still_ no joy. Nov 07 22:18:46 this: Nov 07 22:18:59 SRC_URI = "https://github.com/downloads/jmalonzo/pywebkitgtk/pywebkitgtk-${PV}.tar.bz2;name=pywebkitgtk \ Nov 07 22:18:59 http://www.gnu.org/software/pythonwebkit/pywebkitgtk-1.1.8.patch;name=pywebkitgtkpatch\ Nov 07 22:18:59 " Nov 07 22:18:59 S = "${WORKDIR}/${SRCNAME}-${PV}" Nov 07 22:24:08 so.... the code is getting unpacked iiiin.... ${WORKDIR}/python-pywebkitgtk-1.1.8/pywebkitgtk-1.1.8 Nov 07 22:24:28 i _assume_ that the patch is applied in ${WORKDIR}/python-pywebkitgtk-1.1.8 not ${WORKDIR} Nov 07 22:24:31 is your recipe in the openembedded repo? Nov 07 22:24:35 which could be the source of the problem Nov 07 22:24:46 mrj10: no. Nov 07 22:24:53 i only just wrote it 2 hours ago! :) Nov 07 22:25:04 ok Nov 07 22:25:36 yeegods, bitbake startup takes forever... (debian amd64 - no python-psyco *sigh*) Nov 07 22:26:14 lkcl: switch to bitbake 1.10 or master :) Nov 07 22:26:34 (if not yet) Nov 07 22:26:39 Jay7: ahh joy! speeed freedooom.... Nov 07 22:26:55 lkcl@teenymac:~/oe/build$ bitbake --version Nov 07 22:26:55 BitBake Build Tool Core version 1.8.18, bitbake version 1.8.18 Nov 07 22:26:56 it should be little faster afaik Nov 07 22:27:01 ok taaa Nov 07 22:27:38 loovely: debian packages only 1.8.18, deep joy. Nov 07 22:28:12 I'm always using git version Nov 07 22:29:23 wark! what's the bitbake git repo url? Nov 07 22:29:37 bitbake.berlios.de doesn't have one! Nov 07 22:30:10 mom Nov 07 22:30:33 git://git.openembedded.org/bitbake Nov 07 22:30:36 only a manual - no sensible web site saying "get here" Nov 07 22:30:38 taaaaa Nov 07 22:45:21 oooo! that's niiice and quiiick Nov 07 22:45:29 Jay7: thank youuu Nov 07 22:45:43 you are welcome :) Nov 07 22:46:08 btw please don't be offended if i ask this: anyone want a patch to remove that f****g stupid ezzzh_setup? Nov 07 22:46:37 the reason i took so long to install it was because i had to hack out ez_setup, replace it with distutils setup! Nov 07 22:47:07 _no_ way i'm letting that stupid unnecessary library crap all over my debian system :) Nov 07 22:47:26 I have no opinion here, sorry :) Nov 07 22:47:44 you may ask on maillist Nov 07 22:48:21 it's utterly, utterly unnecessary! bitbake has zero dependencies, so it's utterly pointless using ez_setup. Nov 07 22:48:22 ok. Nov 07 22:48:37 hmm... my weekend bitbake world has left me a 1gb log file... Nov 07 22:48:45 :( Nov 07 22:49:17 grg: pass log output to gzip ;) Nov 07 22:49:33 Jay7, yeah, ill just post it to the mailing list, shall i? Nov 07 22:49:59 grg: did you mean log file? :) Nov 07 22:50:51 i run bitbake -k world | tee world.log.`date +%Y%m%d` each weekend Nov 07 22:51:04 grg: is bitbake world usable now? Nov 07 22:51:23 Jay7, more or less. Today seems to be one of the "less" days. Nov 07 22:51:32 hehe Nov 07 22:52:23 http://pastebin.com/sdNDTE3W <-- tail -n 500 world.log.20101105.1 Nov 07 22:52:30 I've thought about storing bb output to logfile but.. I don't want to see that NOTE lines Nov 07 22:52:39 raise ValueError, "insecure string pickle" Nov 07 22:53:02 bb should have better output classification imho.. Nov 07 22:53:08 "normal" world.log files seem to be about 50mb for me Nov 07 22:53:25 great for greping Nov 07 22:53:38 grg: pass it to syslog and use newsyslog/logrotate :) Nov 07 22:54:25 i'm comfortable with my current system Nov 07 22:54:36 its not perfect, but it works for me Nov 07 22:55:21 I suspect that bb pass warnings and errors to stdout instead of stderr.. Nov 07 22:55:51 1.10 at least.. Nov 07 22:57:34 it does, yes Nov 07 23:01:39 hm.. Nov 07 23:01:46 ant__: hey Nov 07 23:14:15 hi Jay7 Nov 07 23:14:39 ant__: do you have some tmpdir w/o rm_work? Nov 07 23:14:55 ah, no Nov 07 23:15:07 np Nov 07 23:15:17 what's failing? Nov 07 23:15:52 I just wish to know how big may be $TMPDIR/work directory :) Nov 07 23:16:32 considering about mount tmpfs there Nov 07 23:16:40 Jay7, I've heard numbers from 16GB to 100GB tossed out...depending on whats being built. Nov 07 23:17:06 I'm not sure when rm_work is started.. Nov 07 23:17:14 and whether rm_work is being used Nov 07 23:17:42 so.. I'll add disk usage measurer into testbuilder :) Nov 07 23:17:59 (Good Plan) Nov 07 23:18:00 df -Pk . | tail -1 every second Nov 07 23:18:19 even s/\./${TMPDIR}/ Nov 07 23:18:45 that'll only give you the entire partition Nov 07 23:18:53 (that $TMPDIR is on) Nov 07 23:18:54 there is nothing more Nov 07 23:19:05 only OE Nov 07 23:19:09 kk Nov 07 23:19:29 same here, is useful to test build from different linuxes Nov 07 23:19:31 I'll have to do du -sh $TMPDIR every 10 seconds or so. Nov 07 23:19:34 anyway no other good ways :) Nov 07 23:19:42 du is ugly though Nov 07 23:19:45 ka6sox: too expensive Nov 07 23:20:01 ya Nov 07 23:20:40 I guess do a diff of the starting and max deviation values from your df Nov 07 23:21:45 Jay7, what are you going to use to drive the builder script? Nov 07 23:22:30 mm.. what do you mean? distro or 'starter'? Nov 07 23:22:48 cron job, hudson, ??? Nov 07 23:22:56 ah, /dev/hands :) Nov 07 23:23:15 but I hope one time it will go into cron job Nov 07 23:23:17 okay... Nov 07 23:23:41 hudson is good for 'permutations' Nov 07 23:23:55 i.e. for build all combinations of MACHINE and DISTRO e.g. Nov 07 23:24:40 effem's script is more suitable to use with hudson Nov 07 23:25:06 I do want some of the features you have been talking about. Nov 07 23:25:23 and I am using effem's script Nov 07 23:25:44 well.. it's easy to add there :) Nov 07 23:26:43 do this at start: (while true; do df -Pk ${TMPDIR} | tail -1 >> $LOG; done) & Nov 07 23:26:51 then kill %1 before exit Nov 07 23:28:45 if there is multiple exits then better to replace exit invocation with function that will call kill %1 then exit Nov 07 23:30:44 ah.. Nov 07 23:30:56 well..an abend would leave it hanging Nov 07 23:30:58 (while true; do df -Pk ${TMPDIR} | tail -1 >> $LOG; sleep 1; done) & Nov 07 23:31:20 thats simpler Nov 07 23:31:27 sleep is needed :) Nov 07 23:31:48 and may be gzip before writing to LOG Nov 07 23:31:57 save space.. Nov 07 23:31:59 okay nn Nov 07 23:32:14 feel free to play with this :) Nov 07 23:32:26 yep...will do. Nov 07 23:42:22 hehe.. Nov 07 23:42:41 ^C will leave that process unkilled Nov 07 23:48:20 * lkcl yee gooods, webkit is large. Nov 07 23:49:52 Jay7, the problem is that you can't just watch some master thread and when it dies then you kill the logging thread. Nov 07 23:50:21 hm.. Nov 07 23:51:04 it may be good idea to have some start & watch process but this is not for shell language :) Nov 07 23:54:40 there are traps Nov 07 23:57:18 yaay! building pythonwebkit! Nov 07 23:57:58 ouch, some nasty cross-compile warnings, but i can live with that Nov 07 23:59:59 ./JavaScriptCore/wtf/ListHashSet.h:174: warning: cast from 'char*' to 'WTF::ListHashSetNode*' increases required alignment of target type Nov 08 00:00:13 anyone got any ideas what that's about? should i care? :) Nov 08 00:01:33 ka6sox: use trap "kill %1" EXIT Nov 08 00:01:49 this will kill job %1 on exit from script Nov 08 00:02:14 at least in bash :) Nov 08 00:04:05 well.. time to run test build and go sleep Nov 08 00:07:47 sleep 1 is too fast Nov 08 00:08:11 better to use sleep 5-10-15 Nov 08 00:13:57 03Michael Smith  07master * r8d4ec7688c 10openembedded.git/recipes/iproute2/iproute2.inc: Nov 08 00:13:57 iproute2: nuke /var/lib/arpd Nov 08 00:13:57 It belongs in volatiles, if we build arpd - but we don't. Nov 08 00:13:57 Signed-off-by: Michael Smith Nov 08 00:34:16 sleep() Nov 08 00:37:27 Hi there, has anyone ever had to deal with GPIO's? Nov 08 00:54:55 strange Nov 08 00:55:01 OPENEZX git are offline Nov 08 02:05:01 Bye Everybody **** ENDING LOGGING AT Mon Nov 08 02:59:57 2010