**** BEGIN LOGGING AT Thu Nov 04 02:59:57 2010 Nov 04 03:07:59 khem: working out the target override issue isn't going to be trivial :) need to check sanity, determine exactly when the overrides are being processed, they seem to be used way before they should be Nov 04 03:08:04 hrm Nov 04 03:21:09 hmmm...4hrs till .eu wakes up. Nov 04 04:11:18 anyone know why handhelds.org has gone down? Nov 04 07:34:40 gm Nov 04 07:37:05 morning eFfeM_work Nov 04 07:37:22 ka6sox-work: doing ok ? Nov 04 07:38:03 well..you have me pretty busy! Nov 04 07:38:05 :D Nov 04 07:40:06 good morning.. Nov 04 07:42:38 I, and my colleagues, are going to have a brieef about OEDEM. Nov 04 07:42:58 Can anybody give the final minute of that meeting? Nov 04 07:43:06 Can anybody give me the final minute of that meeting? Nov 04 07:57:10 recalcati, good morning. is this what you're looking for: http://www.mail-archive.com/openembedded-devel@lists.openembedded.org/msg12583.html Nov 04 08:17:39 Slyon: thx very much. Nov 04 08:22:55 yw Nov 04 08:24:20 ANNOUNCE: Patchwork cleanup weekend. Only 2 days to go! Nov 04 08:42:18 moin Nov 04 08:56:30 good morning guys Nov 04 09:03:58 i've problems compiling glib (not glibc..) http://pastebin.com/aDR05hd4 Nov 04 09:11:28 Jay7: ynezz, http://wiki.openembedded.net/index.php/TestingScript Nov 04 09:11:32 feel free to improve it Nov 04 09:14:49 gm Nov 04 09:15:15 HopsNBarley: Hi there Nov 04 09:34:03 Um, why does parted depend upon lvm2? Nov 04 09:34:05 :/ Nov 04 09:57:08 B_Lizzard: volume manager? Nov 04 09:57:30 B_Lizzard: is it an opt-in configure option? Nov 04 09:57:36 I think so Nov 04 09:58:24 Yeah, it is Nov 04 09:58:27 It's not required Nov 04 09:58:56 hi Nov 04 09:58:57 parted 2.3 was added 8 days ago along with the lvm2 dep Nov 04 10:03:52 morning florian Nov 04 10:05:21 good morning Nov 04 10:15:14 argh, I see now I've pushed patches from the wrong branch...still full of commented code :/ Nov 04 10:20:52 good morning Nov 04 10:54:21 eFfeM_work: thanks, but wouldn't it be better to put that script somewhere on git{hub,orious}? :] Nov 04 10:54:37 eFfeM_work: it's crazy to handle that in the wiki :) Nov 04 10:54:45 eFfeM_work: just my 5c Nov 04 10:58:24 hi all, is there a way how can I get path within OE build tree which belongs to linux kernel ? Nov 04 10:58:28 ynezz: eFfeM_work: maybe in contrib/testscripts/ Nov 04 10:58:47 ericben: yep Nov 04 10:59:56 yes, I can do that Nov 04 11:00:19 np Nov 04 11:10:45 morning Nov 04 11:11:05 eFfeM_work: thanks for sharing your script :) Nov 04 11:11:16 yw Nov 04 11:13:27 is there a command to only fetch source archives (and not unpack/configure them) ? Nov 04 11:14:11 ericben: bitbake -c fetch Nov 04 11:14:26 i think you can even do bitbake -c fetchall world or so Nov 04 11:14:31 I'm doing fetchall and it unpack/patch Nov 04 11:14:59 at least for coreutils-native Nov 04 11:15:21 but as I'm on a slow machine I would like it to only fetch the files Nov 04 11:16:09 it *might* require something from coreutils for other things. No idea on task deps Nov 04 11:16:18 you could try -cfetch Nov 04 11:16:39 ye smaybe it's needed to build few native deps before fetching more things Nov 04 11:17:04 trying now, takes a moment Nov 04 11:17:06 I'll let it run during meal and see after. I'll think to a (c downloadall option also Nov 04 11:17:29 which would only get the source archives to the download directory Nov 04 11:17:52 ericben, tried bitbake -c fetch urjtag (which is a target recipe) Nov 04 11:18:34 ericben: http://www.pastebin.ca/1981546 Nov 04 11:20:35 ok, I think he needs some native tools before being able to fetch because on a clean tmp it tries to build coreutils-native, quilt& co) Nov 04 11:20:50 so once this is done the other packages will only be fetched Nov 04 11:20:52 thanks Nov 04 11:21:28 guess so from the log Nov 04 11:21:32 (got interrupt) Nov 04 11:21:44 this is a celeron so it's slow .... Nov 04 11:22:02 but it has a good bandwitdh to become my mirror :) Nov 04 11:22:31 it would be interesting to know which tasks it need (I see here expat-native, but that is not on a clean build, so there might be a few more) Nov 04 11:22:34 and why they are needed Nov 04 11:22:55 technically speaking I'd say this is an issue with the task dependencies Nov 04 11:23:23 guess fetch should not depend on expat, as fetching is doen by bitbake if i am correct Nov 04 11:25:55 package linux-libc-headers-native-2.6.32-r5: task do_unpack: Started Nov 04 11:26:08 ... so it nead the headers ! Nov 04 11:26:12 it needs Nov 04 11:26:13 sorry Nov 04 11:26:42 yuk Nov 04 11:26:44 hm. Nov 04 11:27:07 florian: my rss reader (akregator) can't read planet linuxtogo feed Nov 04 11:27:10 ericben: i suggest you let it run then prepare a list of packages and then post to the list Nov 04 11:27:17 or even can't parse Nov 04 11:27:23 I'll see witht he stamps Nov 04 11:27:44 to see the ones which are configured/compiled/packaged Nov 04 11:27:45 I suspect it was affected by that out of disk space Nov 04 11:28:10 Jay7: my browser too Nov 04 11:35:45 Jay7: fixed Nov 04 11:35:54 florian: thank you Nov 04 11:36:24 yw Nov 04 11:53:38 INHERIT += oestats-client and no network connection -> parse error ! Nov 04 11:54:04 weird Nov 04 11:57:34 even without oestats-client -> parse error so this is somewhere else. if only bitbake was telling where is the parsing error ... Nov 04 12:01:29 ericben: hm.. interesting Nov 04 12:37:52 fun that came from a SRC_URI in an overlay ! Nov 04 12:38:20 because of a SRCREV= AUTOREV Nov 04 12:38:37 so i norder to create the PN, bitbake has to make a querry to the server Nov 04 12:39:12 so having SRCREV in the PN for a package with AUTOREV is not a good idea for offline builds ;) Nov 04 12:39:54 :) Nov 04 13:59:20 ericben: heh, might want to consider using srctree or similar with a local checkout rather than AUTOREV Nov 04 13:59:45 kergoth: yes, in the present case, I had an example for AUTOREV left in an overlay ;-) Nov 04 13:59:55 hehe, oops Nov 04 14:00:03 but now the offline build is quite finished Nov 04 14:00:10 so the principle works Nov 04 14:00:31 it's cool to fetch at 100MB/s ;-) Nov 04 14:02:05 ericben: its even cooler on 1000 MB/s but normally no server can keep up with the speed Nov 04 14:02:46 yes, here it's not on the network it's on the internal HD so int's 100MB and not 100Mb ;) Nov 04 14:03:29 full offline build with ifdown br0, to test Nov 04 14:04:04 ericben: ah, thats nice Nov 04 14:05:25 this way no problem when a customer tries to rebuild the thing in its internal network protected by a ultra expensive blackbox that noone knows to configure :-D Nov 04 14:06:05 03Michael Lippautz  07master * r89362b3b55 10openembedded.git/recipes/ddclient/ddclient_3.6.3.bb: Nov 04 14:06:05 ddclient: Unbreak recipe Nov 04 14:06:05 * ddclient expects its default config at /etc/ddclient/ddclient.conf instead Nov 04 14:06:05 of /etc/ddclient.conf Nov 04 14:06:05 Signed-off-by: Michael Lippautz Nov 04 14:06:12 03Michael Lippautz  07master * r2c9d8c7b47 10openembedded.git/recipes/ddclient/ddclient_3.6.3.bb: ddclient: cosmetics Nov 04 14:06:13 03Michael Lippautz  07master * r6ee447abb7 10openembedded.git/recipes/usb-modeswitch/usb-modeswitch-data_20100826.bb: usb-modeswitch-data: Fix deps and add udev rules to package Nov 04 14:06:14 03Michael Lippautz  07master * rca4dfd001c 10openembedded.git/recipes/usb-modeswitch/usb-modeswitch_1.1.4.bb: usb-modeswitch: Add sbin dir to package Nov 04 14:07:12 ericben: nice, thats a good idea. you can always replace fetch.bbclass with a no-op task, but bringing downt he interface catches cases where the buildsystems we run actually try to get to the network, sick as it is :) Nov 04 14:23:55 ericben: so you just downloaded all the sources with "bitbake -c fetchall world"? Nov 04 14:24:15 tdebrouw: not all, the ones necessary for the 5 recipes I build Nov 04 14:24:53 ericben: k, thx, i'll try it later on Nov 04 14:29:05 eFfeM_work: ok to have the fetchall working from scratch, the following recipes are built before (all native) : autoconf, automake, coreutils, curl, expat, git, gnu-config, libtool, linux-libc-headers m4, openssl, pkgconfig, quilt, stagemanager, unifdef, zlib Nov 04 14:29:43 hm, ok, does not really seem ok Nov 04 14:30:05 guess curl is the bad guy Nov 04 14:30:53 this seems reasonable : curl & git to fetch, othere are here to compile curl & git I think Nov 04 14:31:55 curl does inherit autotools pkgconfig and binconfig Nov 04 14:32:31 actually I would expect bitbake to take care of all fetching Nov 04 14:33:23 curl shouldn't be necessary to fetch, we use wget, not curl for that Nov 04 14:33:58 ah interesting Nov 04 14:34:07 kergoth: ericben I think it is only git, that already depends on opensslg, curl zlib expat and tcl-native Nov 04 14:34:22 * kergoth nods Nov 04 14:34:24 that'd explain it Nov 04 14:34:26 (don't see tcl-native in your list) Nov 04 14:34:30 oh yes you're right Nov 04 14:34:35 so if you have git, be sure to add git-native to your ASSUME_PROVIDED :) Nov 04 14:34:53 will try for next run Nov 04 14:35:13 kergoth: why would git be needed to fetch only? bitbake has a git fetcher i thought Nov 04 14:35:20 should reduce the amount of crap to be built nicely -- we really do need a class to fill out ASSUME_PROVIDED based on whats on the build machine Nov 04 14:35:27 eFfeM_work: it does, and its git fetcher runs git :P Nov 04 14:35:29 or does taht rely on git being there Nov 04 14:35:53 kergoth: ah understood Nov 04 14:36:18 as far as i know, all the git python bindings rely on being able to run git -- git doesn't have a good library to bind to and bypass the utility, as far as i know Nov 04 14:36:27 er, that grammar was terrible, but you get the point Nov 04 14:36:28 :) Nov 04 14:36:33 for other things (svn, cvs, mercurial) we assume host provides it Nov 04 14:36:56 not mercurial Nov 04 14:37:05 kergoth: :-) I'm the last to criticize your grammar Nov 04 14:37:14 we should probably add svn and cvs to use -natives too, can always add to ASSUME_PROVIDED.. Nov 04 14:59:36 03Eric BENARD  07org.openembedded.dev * r0ef914b250 10openembedded.git/classes/own-mirrors.bbclass: (log message trimmed) Nov 04 14:59:36 own-mirrors.bbclass: allow mirroring of scm fetched packages Nov 04 14:59:36 this way, it's possible to setup a local webserver (for example Nov 04 14:59:36 using busybox httpd -p "8081" -h backuped_download_dir) serving Nov 04 14:59:36 a presiously fetched download directory and to build wihout the Nov 04 14:59:37 need for an internet access Nov 04 14:59:38 this can also be used when connected to know to know which packages Nov 04 15:04:08 hey I was just re writing the comments ! Nov 04 15:06:03 * kergoth chuckles Nov 04 15:07:03 * kergoth tests a TERMCMD/TERMCMDRUN for Terminal / iTerm Nov 04 15:14:12 ericben: about to push your graphviz fix to both 1.10 and master Nov 04 15:15:23 kergoth: I'm not the initial author,I just updated it to master Nov 04 15:15:50 ah Nov 04 15:15:52 so don't forget its credits to the initial author Nov 04 15:15:58 I think I added a From: line Nov 04 15:16:07 yeah, i see the From there Nov 04 15:16:19 added an s-o-b for me since its going through me into bitbake Nov 04 15:17:11 anyone using QT4 here on ARM > 5 here ? Nov 04 15:17:22 ARM > v5 ? Nov 04 15:22:06 ericben: running kde 4.5 on armv7a counts? Nov 04 15:22:32 hrw: :) Nov 04 15:22:37 on the i.MX51 ? Nov 04 15:22:39 yes Nov 04 15:22:47 i.mx515 to be exact Nov 04 15:23:40 I have several i.MX515 boards here, do you have a link to the image you are using to test on it ? Nov 04 15:24:15 in the present case if anyone can try to switch QT_ARCH to armv6 on a qt4 build for armv7 this would be great Nov 04 15:24:36 I think that's the way ubuntu is also configuring qt for armv7 Nov 04 15:28:42 ericben: ubuntu iirc builds for armv6 Nov 04 15:30:45 hrw: no it's armv7 now Nov 04 15:30:57 ah, ok Nov 04 15:31:04 https://wiki.ubuntu.com/ARM/Thumb2 Nov 04 15:31:35 funny they drop the imx51 in maverick Nov 04 15:31:50 it's repalced by the omap4 Nov 04 15:32:48 maybe will return in 11.04 out of list Nov 04 15:33:06 I hope, with a fully accelerated graphic driver compatible with a 2.6.36 kernel Nov 04 15:33:09 :) Nov 04 15:39:34 ericben: there is such? :D Nov 04 15:40:04 this would allow to leave the fsl's 2.6.32 kernel which is quite old Nov 04 15:41:31 well, the omap 4 has a POWERVR SGX540 onboard Nov 04 15:41:41 no idea on the performance though Nov 04 15:41:52 yes but the ompa4 is impssible to puchase as a chip Nov 04 15:42:01 at this point Nov 04 15:42:04 it isn't Nov 04 15:42:06 :) Nov 04 15:42:15 this will last from what I've understood from TI Nov 04 15:42:19 but the pandaboard has it Nov 04 15:42:27 yes, I've one on order Nov 04 15:42:33 to see the real power of the beast Nov 04 15:42:38 I don't Nov 04 15:42:41 not yet Nov 04 15:42:42 :) Nov 04 15:42:51 i'll order 2 :P Nov 04 15:42:55 03Michael Smith  07master * r57388be76e 10openembedded.git/recipes/liboil/ (liboil-0.3.17/fix-unaligned-whitelist.patch liboil_0.3.17.bb): Nov 04 15:42:55 liboil 0.3.17: fix cross-compile on x86_64 Nov 04 15:42:55 Fix typo in whitelist of hosts allowing unaligned memory access. Nov 04 15:42:55 Signed-off-by: Michael Smith Nov 04 15:42:58 lol Nov 04 15:43:06 03Michael Smith  07master * rf3e06cb4d1 10openembedded.git/recipes/xcb/libxcb.inc: Nov 04 15:43:06 libxcb: DEPEND on xcb-proto-native Nov 04 15:43:06 libxcb needs the xcbgen Python modules staged by xcb-proto-native. Nov 04 15:43:06 Signed-off-by: Michael Smith Nov 04 15:43:11 should ship on the 21 of october if digikey don't fail Nov 04 15:43:16 of november Nov 04 15:43:34 but I suppose the omap 3 had this availability as well? Nov 04 15:43:46 first, only for wireless (huge market) Nov 04 15:43:51 then beagle board Nov 04 15:43:57 then available. Nov 04 15:43:58 omap3 was available quite soon after the beagle Nov 04 15:43:58 so Nov 04 15:44:03 well Nov 04 15:44:06 for industrial markets Nov 04 15:44:17 I expect something similar for the omap 4 Nov 04 15:44:43 but now they have split Omap and sitara ranges, we need to wait for the omap4 to be integrated in the sitara range Nov 04 15:45:04 as now omap is wireless and sitara is industrial Nov 04 15:45:14 at least from what I understood Nov 04 15:46:10 :( Nov 04 15:46:26 that would mean additional waiting time Nov 04 15:46:28 we will see Nov 04 15:46:29 :( Nov 04 15:46:38 but I think they will have to do something under presure Nov 04 15:47:01 they can't ship 180$ boards with omap4 through digikey and tell to industrial peoples : you can play with it but not design with it Nov 04 15:47:47 that's a nonsens except if the strategy is to use the pandaboard community as testers for a product that only phone makers can use ! Nov 04 15:48:10 but the sitara only has an arm? Nov 04 15:48:13 no dsp? Nov 04 15:48:26 euh Nov 04 15:48:32 i'm not sure about this Nov 04 15:49:33 tdebrouw: yes Nov 04 15:49:37 yes, the latest ones ven have a gigabit MAC and a PCI express Nov 04 15:49:52 tdebrouw: sitara is with arm only, integra is arm+dsp Nov 04 15:50:13 i have 2 integra;s at my desk :) Nov 04 15:50:14 and ompa is wireless Nov 04 15:50:35 tdebrouw: send the second one to me pls :-) Nov 04 15:50:43 i'm sorry Nov 04 15:50:47 no can't do Nov 04 15:50:54 I need to test the PCIe interface with my FPGA's :-) Nov 04 15:51:32 tdebrouw: which integra series is on your desk? Nov 04 15:52:01 euh? Nov 04 15:52:11 what do you mean exactly? Nov 04 15:52:23 http://focus.ti.com/docs/toolsw/folders/print/tmdxevm8168ddr2.html Nov 04 15:52:55 ericben: how do you know they drop imx51 from ubuntu? Did I miss something on that page? (probable answer: yes) Nov 04 15:53:28 tdebrouw: ok, the lastest goodies Nov 04 15:53:29 i have a DDR2 version Nov 04 15:53:31 -s Nov 04 15:53:35 likewise: https://wiki.ubuntu.com/ARM What kernel flavours are available? Nov 04 15:55:32 the ddr3 version is on order Nov 04 15:55:35 currently Nov 04 16:07:00 hi Nov 04 16:07:14 khem are you there? Nov 04 16:26:30 03Chris Larson  07master * r83cdb23f8b 10bitbake.git/lib/bb/cooker.py: Nov 04 16:26:30 Fix GraphViz .dot output for rdepends and rrecs Nov 04 16:26:30 Use bb.utils.explode_deps to break up the rdepends and rrecommends strings. Nov 04 16:26:30 This fixes the same issue which was fixed by a number of patches floating Nov 04 16:26:30 around, but uses explode_deps rather than regular expressions. Nov 04 16:26:30 Signed-off-by: Chris Larson Nov 04 16:40:01 03Michael Smith  07master * r12de407d31 10openembedded.git/recipes/xorg-driver/xf86-video-openchrome_0.2.904.bb: Nov 04 16:40:02 xf86-video-openchrome: DEPEND on libxvmc Nov 04 16:40:02 configure fails without it. Nov 04 16:40:02 Signed-off-by: Michael Smith Nov 04 17:14:45 tdebrouw: you .be based? Nov 04 17:19:37 trying to install and use bitbake but get the following error: Nov 04 17:19:55 NameError: global name 'msg' is not defined Nov 04 17:22:45 ~pastebin Nov 04 17:22:45 [~pastebin] A "pastebin" is a web-based service where you should paste anything over 3 lines so you don't flood the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://bin.cakephp.org/ , http://asterisk.pastey.net/ , or install pastebinit with yum or aptitude. Nov 04 17:23:00 Please use that and put the whole error, from $ bitbake ... to the next $ and give us the link Nov 04 17:23:00 alkopop79: sounds like you're on an old oe, that was fixed Nov 04 17:23:12 Or listen to kergoth :) Nov 04 17:23:30 :) Nov 04 17:23:44 :D Nov 04 17:23:51 thx! Nov 04 17:24:03 * kergoth_ busted out the crystal ball Nov 04 17:24:05 :P Nov 04 17:24:05 tried to update Nov 04 17:24:10 with git pull Nov 04 17:24:19 but it says it's the letest update Nov 04 17:24:24 check it anyway Nov 04 17:24:52 what branch? Nov 04 17:25:02 that issue is fixed in master. Nov 04 17:25:58 the error: 12http://pastebin.ca/1981990 Nov 04 17:26:33 yep, thats almost certainly the one i'm referring to Nov 04 17:26:59 5efecae8b7f4b814f2d055daf6dc7ae1ce05e41a fixed it. Nov 04 17:32:35 how can i update oe? Nov 04 17:32:45 i tried git pull Nov 04 17:32:52 with no success Nov 04 17:32:59 how can i check the version? Nov 04 17:33:11 git pull is how you update. i asked you which branch you're on, you never responded. Nov 04 17:33:17 run git branch, look at the one marked with * Nov 04 17:34:01 http://progit.org/book/ Nov 04 17:39:57 apologies kergoth Nov 04 17:40:14 i have zero experience with linux Nov 04 17:40:19 just learning Nov 04 17:40:26 anyway Nov 04 17:40:41 downloaded and installed the latest version of oe Nov 04 17:40:47 1.10.1 Nov 04 17:40:57 but the problem still persists Nov 04 17:41:03 no. Nov 04 17:41:06 that's bitbake Nov 04 17:41:09 bitbake is not oe Nov 04 17:41:11 :D Nov 04 17:41:15 oke Nov 04 17:41:17 so Nov 04 17:41:18 oe Nov 04 17:41:24 (cd openembedded;git pull) Nov 04 17:41:28 cd openembedded; git checkout master; git pull Nov 04 17:41:28 thta's emberrasing Nov 04 17:41:40 awesome! Nov 04 17:41:41 We were all newbies once :) Nov 04 17:42:30 that's a nightmare Nov 04 17:42:34 cd openembedded Nov 04 17:42:44 no such file or directory... Nov 04 17:42:50 where did i go wrong? Nov 04 17:43:00 i clearly remember i have an openembedded foler Nov 04 17:43:01 ] Nov 04 17:43:04 folder Nov 04 17:44:23 kill AUTOREV? Nov 04 17:45:04 03Chris Larson  07master * r72a44c7959 10openembedded.git/classes/base.bbclass: Nov 04 17:45:04 base.bbclass: rename base_eventhandler to build_summary Nov 04 17:45:04 This better reflects what the event handler is doing -- it's no longer a pile Nov 04 17:45:04 of random base event handling bits, it now just displays a build summary at Nov 04 17:45:04 build start. Nov 04 17:45:05 Signed-off-by: Chris Larson Nov 04 17:45:14 03Chris Larson  07master * r9cd6a9dcdb 10openembedded.git/recipes/db/ (db-5.0.21/cp.patch db_5.0.21.bb): Nov 04 17:45:14 db 5.0.21: fix cp argument portability Nov 04 17:45:14 Signed-off-by: Chris Larson Nov 04 17:45:16 alkopop79: we aren't psychic. we can't tell you what you did. Nov 04 17:46:34 alkopop: type "pwd" and paste the result here Nov 04 17:46:37 and "ls" Nov 04 17:46:44 it's workin fine Nov 04 17:46:51 thank you very much Nov 04 17:47:02 i've been struggling with it for days Nov 04 17:47:15 thx! Nov 04 17:47:39 np Nov 04 18:09:13 alkopop79: if you post a question, then find the answer yourself, be so kind to post it in IRC. It will get picked up by the logs so that other people having the same issue can find the answers. (For future, more daunting problems :-) ) Nov 04 18:12:15 03Chris Larson  07master * r3620e5ee5e 10openembedded.git/classes/srctree.bbclass: Nov 04 18:12:15 srctree: don't merge tasks if using B Nov 04 18:12:15 Signed-off-by: Chris Larson Nov 04 18:18:31 kergoth_, what do you mean using B? Nov 04 18:19:05 B is builddir Nov 04 18:22:55 kergoth_: around ? Nov 04 18:30:35 03Chris Larson  07master * rc35b4a0846 10openembedded.git/bin/ (cp sed wrapper.sh): Nov 04 18:30:35 bin/{cp,sed}: simplify, consolidate Nov 04 18:30:35 Signed-off-by: Chris Larson Nov 04 18:35:29 kergoth_: I wonder if we should exclude meta recipes from bitbake world Nov 04 18:35:51 kergoth_: I am thinking of adding the exclusion in meta.bbclass Nov 04 18:36:01 bad ? ok ? Nov 04 18:36:26 Crofton|work: like khem says, B is for separating build dir from source dir -- if the build dirs aren't shared between target/cross/native, then the task merging crap is unnecessary, since they can't step on one anothers toes Nov 04 18:36:50 kergoth_: I have an idea Nov 04 18:37:01 kergoth_: lets make poky as core layer of OE :)) Nov 04 18:37:12 you guys confuse me :) Nov 04 18:37:17 khem: probably a good idea, metas aren't unlike images, really, both should be excluded from world Nov 04 18:37:21 and whatever we have extra in OE stick it around as layers Nov 04 18:37:30 khem, discuss on list :) Nov 04 18:37:46 I see value Nov 04 18:37:58 Crofton|work: poky is much more polished Nov 04 18:38:08 Crofton|work: if you have both, say, m4-native and m4, and both are using srctree with the same S, there's a danger of task interleaving -- ./configjure from native and make from cross, etc, so srctree merges the tasks which operate there into a single populate_sysroot task. this is the step that can be skipped if B is used Nov 04 18:38:22 does that make it slightly less confusing? :) Nov 04 18:38:30 :) Nov 04 18:38:37 madness! Nov 04 18:38:40 kergoth_: we should make unique B for all recipes Nov 04 18:38:46 all I care is can I ship and will I have SW support Nov 04 18:38:51 may be append the -native cross etc Nov 04 18:39:13 khem: it should work, actually -- if you set B = and B isn't set in the recipe, it'll be obeyed everywhere autotools is used -- and the recipe can always set B = "${S}" to opt out Nov 04 18:39:23 we should try it Nov 04 18:40:15 kergoth_: so I will exclude images and metas both Nov 04 18:40:47 khem, the larger OE community needs to agree on a way forward Nov 04 18:41:16 and we need to make sure if the yocto project folds, we are not screwed Nov 04 18:41:25 Crofton|work: one thing we all want is that OE should develop into a better software Nov 04 18:41:32 well yocto is yocto its about OE Nov 04 18:41:43 03Chris Larson  07org.openembedded.dev * ra3d8c448c2 10openembedded.git/classes/metadata_scm.bbclass: (log message trimmed) Nov 04 18:41:43 metadata_scm: use rev-parse rather than show-ref Nov 04 18:41:43 show-ref will show all matching refs, so a "show-ref HEAD" will show not just Nov 04 18:41:43 the local HEAD, but the fetched remote ones as well. This isn't what we want Nov 04 18:41:43 for this function, so change it to use rev-parse with --verify, and also Nov 04 18:41:43 change it to use --short, to shorten the hash to a more palatable length. Nov 04 18:41:44 Signed-off-by: Chris Larson Nov 04 18:41:52 OE remains where it is and what it is Nov 04 18:42:02 we just make it better Nov 04 18:42:05 yocto folding wouldnt' be a problem, its just something built upon poky. if OE became based on poky, and poky died, it still wouldnt' be a problem, as OE could maintain it Nov 04 18:42:15 I don't see an issue, its just something we need to discuss thoroughly Nov 04 18:42:21 kergoth_, exactly Nov 04 18:42:22 right Nov 04 18:42:36 we will needs processes in place for disaster recovery Nov 04 18:42:45 kergoth_: I tried to see if we could merge all poky changes into OE Nov 04 18:43:04 and it become s huge task given the number of recipe and targets and versions we have Nov 04 18:43:23 so porting all the poky stuff into OE is not something easy Nov 04 18:43:57 Crofton|work: we need to design around it Nov 04 18:44:02 vice versa probably wouldn't be trivial either, unfortunately Nov 04 18:44:10 kergoth_: yep Nov 04 18:44:17 but breaking into layers is an option Nov 04 18:44:53 also, I do not want to see duplicate work Nov 04 18:45:09 hi guys Nov 04 18:45:15 that wont happen, I guess if we keep OE/core and poky in sync Nov 04 18:45:19 the metadata split is a big part of why poky exists, i think -- if OE split, that would definitely reduce the deviation between the two Nov 04 18:45:22 hy khem :) Nov 04 18:45:36 lynx_deb: hello Nov 04 18:45:55 * kergoth_ ponders Nov 04 18:45:56 and we will also unify the communities Nov 04 18:46:10 03Chris Larson  07master * r7c05f4e4c4 10openembedded.git/recipes/termcap/ (termcap_1.3.1.bb termcap_11.0.1.bb): Nov 04 18:46:10 termcap, termcapfile: add GNU termcap 1.3.1 Nov 04 18:46:10 Signed-off-by: Chris Larson Nov 04 18:46:11 which would mean much bigger and more contributions Nov 04 18:46:16 i still have problem compile this damn wxwidgets.. after your tricks i've successfully built gcc and glibc, but no glib doesn't work :\ Nov 04 18:46:18 http://pastebin.com/zc1qty7u Nov 04 18:46:58 03Chris Larson  07master * rf3138434b4 10openembedded.git/MAINTAINERS: Nov 04 18:46:58 MAINTAINERS: add myself for termcap Nov 04 18:46:58 Signed-off-by: Chris Larson Nov 04 18:47:23 lynx_deb: error: too many arguments to function 'splice' Nov 04 18:48:03 lynx_deb: you might have to choose an older version of glib Nov 04 18:48:31 also.. why if i set PREFERRED_VERSION_glib-2.0 = "2.20.4" still try to compile 2.24? Nov 04 18:48:47 lynx_deb: you can stomp it into local.conf with PREFERRED_VERSION_glib-2.0_local = "x.y.z" Nov 04 18:48:55 damnit, where is the upstream for termcap 2.0.8 Nov 04 18:48:58 you need that local override Nov 04 18:48:58 yes i'm trying to compile older version.. but preferred doesn't seems to work Nov 04 18:49:08 lynx_deb: also set PREFERRED_VERSION_glib-2.0-native Nov 04 18:49:23 kergoth_: its gnu thing right Nov 04 18:49:36 i thought so, but only 1.3 and 1.3.1 are there Nov 04 18:49:48 the major distros seem to have heavily patched 2.0.8, including gentoo and fedora Nov 04 18:49:48 hrm Nov 04 18:49:51 ok.. i try Nov 04 18:49:54 course most people use ncurses for it, but still Nov 04 18:50:06 lynx_deb: the problem is that when you use older version of core libs and toolchain you have to hang out with older versions Nov 04 18:50:26 yes this was my first idea Nov 04 18:50:40 also i'm try to compile 2.20.4 to match the ambient on my nas Nov 04 18:50:49 but still have problems to understand preffered switch.. Nov 04 18:50:59 (or maybe bitbake ignore them..) Nov 04 18:51:23 lynx_deb: as I said use local override Nov 04 18:51:24 no i'm cleaning glibc and try also the native preferred Nov 04 18:51:59 what do yo mean for local override? local.conf or put the preferred var before bitbake command? Nov 04 18:52:10 Crofton|work: I think it will be a good thing for OE to layer'ize in general and while we are at it we should unify with poky Nov 04 18:52:47 Crofton|work: I see a good potential and enlarged base for OE Nov 04 18:53:06 lynx_deb: I even typed the command for you above just read it Nov 04 18:54:08 PREFERRED_VERSION_glib-2.0 = "2.20.4" Nov 04 18:54:08 PREFERRED_VERSION_glib-2.0-native = "2.20.4" Nov 04 18:54:09 right? Nov 04 18:54:14 Crofton|work: I will send an email to the ml and lets see if I get flares Nov 04 18:54:31 lynx_deb: arggs 11:44 < khem> lynx_deb: you can stomp it into local.conf with PREFERRED_VERSION_glib-2.0_local = "x.y.z" Nov 04 18:54:59 ops.. sorry, i miss the word "local" ^^ Nov 04 18:55:56 khem, I want to make sure everyone has a voice Nov 04 18:56:01 and voices their thoughts Nov 04 18:56:23 Crofton|work: yes thats thre right thing Nov 04 18:56:35 khem: I'm all for that plan, so you get my ack. Note that Yocto dropped 90% of our recipes and machines; we need to get the most useful ones to co-exist in Yocto, was has moved in a different direction a bit. Nov 04 18:56:51 We will adopt the stuff into OE its not that we are abandoning it or anything Nov 04 18:57:36 Our aim should be to minimize bidirectional merging efforts Nov 04 18:57:41 I am hopeful that if poky is so sound in OE then yocto will just derive out of OE Nov 04 18:57:58 and poky can contribute into OE Nov 04 18:58:14 likewise: yes thats true Nov 04 18:58:36 I think we will get a large community together as well which is a good thing Nov 04 18:59:20 I am hopeful that if we do that then poky will work like angstrom does today in OE Nov 04 19:00:17 interesting ideas, hopefully something will work out Nov 04 19:00:37 cbrake: its a lot of effort though Nov 04 19:02:23 morning Nov 04 19:02:49 ka6sox-work: hello Nov 04 19:03:07 hiya cbrake...got a few hours anyways. Nov 04 19:03:44 ka6sox-work: :-) Nov 04 19:04:00 khem can we set rm_work as the default when we are done with builds? Nov 04 19:04:22 or do we need things for failure analysis? Nov 04 19:05:12 ka6sox-work: I think it should be, but others disagree Nov 04 19:05:19 probably need an RFC on the list Nov 04 19:05:54 ka6sox-work: https://github.com/kergoth/openembedded/commit/72f3fd1 Nov 04 19:07:16 kergoth, as the maintainer of Infra and seeing the results of build pieces laying around forever I would agree. Nov 04 19:08:34 ka6sox-work: yes we should do that Nov 04 19:09:04 I'm of the opinion that the common case is what we should optimize for. debugging a particular problem with a particular recipe shouldn't be something we try to anticipate in advance -- ask for it when you need it Nov 04 19:09:20 kergoth_: yes I agree Nov 04 19:09:35 and for pack rats we should provide the option to disable it Nov 04 19:09:44 yep Nov 04 19:10:04 kergoth_: RFC it Nov 04 19:10:13 ya, good idea..RFC it. Nov 04 19:10:28 will do Nov 04 19:16:19 and… doesn't work, now with 2.20.4 stop to do_compile -.- Nov 04 19:16:24 *do_configure Nov 04 19:20:59 hmmm there are tasks like beagleboard-demo Nov 04 19:21:16 which seems like should be disabled for non beagleboard machines Nov 04 19:22:17 kergoth_: I wonder if tasks should also be excluded from world Nov 04 19:22:31 http://pastebin.com/7A6Gbhw8 http://pastebin.com/6VTf9iTM Nov 04 19:22:40 hi, if someone can help me with a recipe/thing i'm trying to do i'd appreciate it Nov 04 19:23:23 i have shared library which I build in a tool external to OE, basically TI code for the C64x+ DSP Nov 04 19:24:02 lynx_deb: the automake and autoconf may be too new Nov 04 19:24:09 I'm trying to find a neat way to properly load the library on my beagleboard and still have it recognized in OE for other packages that I'm building to know it exists Nov 04 19:24:17 lynx_deb: I think you should try to use stable/2009 branch Nov 04 19:24:36 or you could keep juggling with the versions of glib and see if anyone works for you Nov 04 19:25:07 is better to write a recipe that locally loads the library and packages it or to move the beagleboard and deal with linking it to other packages from there Nov 04 19:25:22 ? Nov 04 19:25:42 what's the name of stable 2009 branch? or i can use 2008? Nov 04 19:25:45 al_fayez: write a recipe Nov 04 19:26:03 lynx_deb: look usin git branch Nov 04 19:26:34 ok so the recipe it I wrote is fairly simple i'm not sure if it's good enough or not can I paste the 3 lines here or should i find a place to post it at? Nov 04 19:26:57 paste it at some place Nov 04 19:27:42 http://pastebin.com/fSp69yph Nov 04 19:27:50 thanks for your help khem Nov 04 19:28:12 i've use this command to clone git on local: git clone git://git.openembedded.org/openembedded Nov 04 19:28:54 i managed to put the library in usr/lib on the beagleboard, I'm just having trouble making another recipe this that this shared library exists which is causing a do_compile Nov 04 19:29:39 al_fayez: thats a wrong recipe Nov 04 19:29:49 you are poking into builddir of some other recipe Nov 04 19:30:08 why dont you package the .so using that *some other recipe* Nov 04 19:30:48 to be honest I've worked with recipes which use the autotools but never a bear bones recipe like this. Nov 04 19:31:09 I essentially tried to follow the helloworld recipe format and I probably messed it up Nov 04 19:31:29 oooh sorry i miss understood you Nov 04 19:31:58 your saying to put it in the source of the other code and from there include it and package into the other recipe Nov 04 19:32:16 what a neat simple solution i'll try it out now Nov 04 19:32:50 sorry i've been up for a while now figuring out how to force the TI tools to generate shared libraries instead of static Nov 04 19:46:03 03Michael Smith  07master * r803372aca4 10openembedded.git/recipes/gnome/metacity_2.30.0.bb: Nov 04 19:46:03 metacity: depend on gnome-doc-utils to fix build Nov 04 19:46:03 Otherwise autocruft complains about missing HAVE_GNOME_DOC_UTILS Nov 04 19:46:03 conditional in gnome-doc-utils.make. Nov 04 19:46:03 Signed-off-by: Michael Smith Nov 04 19:50:34 so angstrom-task-gnome uses pulseaudio, but afaict most of the packages that use audio aren't pulse-enabled :P Nov 04 19:50:53 e.g. mplayer_svn.bb explicitly uses --disable-pulse in ./configure Nov 04 19:51:15 i get audio just fine if i kill all the pulse-related processes after boot Nov 04 19:51:19 and just use alsa Nov 04 19:55:23 kergoth_: whats a sane environment for bitbake 1.10.0? python version wise. Im having issues with bitbake directly after invocation, complaining about cooker.py/command.py and some unboundlocalerror Nov 04 19:55:49 Yep, that's like the bluez-related stuff -- we need a global "switch" so that a distro can select if they want packages to use a specific underlying library or not. Nov 04 19:56:12 kristoffer: i use 2.6.5 just fine w/ BB 1.10.0 Nov 04 19:56:43 (none of the systems I work with have enough RAM or CPU to do pulse, so it wouldn't be good for me if we enabled pulse in configure for those packages, of course!) Nov 04 19:57:39 kristoffer, 2.6.2 on 64-bit host is working ok with BB 1.10.1 Nov 04 19:57:52 kristoffer: 2.5-2.7 should be fine, as far as i know Nov 04 19:58:22 mwester: yeah, this thing has 512MB RAM and a 1GHz cortex-a8, so it should be able to do it. i don't really care if i use pulse or not, but it seems weird that the angstrom tasks would use it if e.g. mplayer has it disabled Nov 04 19:59:20 kergoth_: any special modules required? Im currently on 2.6 and its not working Nov 04 19:59:24 mwester: also, in the case of mplayer, --enable-pulse just adds the support code for pulseaudio output, and you can select between that or ESD or ALSA or Jack at runtime, it doesn't force you one way or the other. but i know what you mean Nov 04 19:59:37 not my machine, so dont know exactly whats installed (devserver) Nov 04 19:59:57 kristoffer: are all the bitbake-related environment variables set up? Nov 04 20:00:08 BBPATH yes Nov 04 20:00:24 PYTHONPATH not exported, but not sure where to point it Nov 04 20:03:37 kergoth_: mrj10 my bad. I had set BBFILES with typo Nov 04 20:03:40 works now Nov 04 20:05:36 cool! Nov 04 20:06:23 if i have library which is in my working directory of a recipe what would be the easiest way to pass that information to a Makefile.am ? I tried using -L the absolute address of the working directory but that doesn't seem to work Nov 04 20:07:11 mwester: probably we could make pulse a distro feature ? Nov 04 20:28:26 if i have library which is in my working directory of a recipe what would be the easiest way to pass that information to a Makefile.am ? I tried using -L the absolute address of the working directory but that doesn't seem to work Nov 04 20:32:45 jkridner: RE: the patch you submitted to OpenEmbedded regarding updating beagleboard-demo-image.bb and task-beagleboard-demo.bb to reflect packages used at ESC -- that image uses gnome, which in turn uses pulseaudio. were you actually able to get audio out of mplayer, aplay, omapfbplay, etc. without killing pulseaudio? Nov 04 21:50:32 How to find out in which commit PHP bug was fixed? Nov 04 21:50:52 Would it work: git log | grep php ? Nov 04 21:57:36 git log recipes/php Nov 04 21:57:42 If it was a fix in that dir Nov 04 22:04:05 al_fayez, maybe you should put the library somewhere configure can find it? Nov 04 22:06:09 hey Crofton, I put it in the tmp_beagle/rootfs/usr/lib Nov 04 22:06:20 Er, do you have to register with tinderbox in some way? Nov 04 22:06:23 heh Nov 04 22:06:26 Or just configure the nickname free-form? Nov 04 22:06:30 I've done that hack :) Nov 04 22:06:35 that worked right? Nov 04 22:06:39 also in the staging directory Nov 04 22:06:44 it compiled Nov 04 22:06:47 heh Nov 04 22:06:55 you are hacking like mad you know? Nov 04 22:06:57 but the library didn't contain the proper links Nov 04 22:07:01 urg Nov 04 22:07:04 oooh yeah Nov 04 22:07:17 i did objdump and all the links weren't there Nov 04 22:07:18 good luck Nov 04 22:07:33 i was actually looking at a post of yours from a while Nov 04 22:07:43 this is some dsplink code? Nov 04 22:07:49 i'm trying to put the gr-custom blocks on the beagleboard and compile it there Nov 04 22:08:09 i just read a HUGE warning from the SWIG website Nov 04 22:08:15 rofl Nov 04 22:08:23 saying avoid static library linking like the plague Nov 04 22:08:28 what did they say Nov 04 22:08:29 ah Nov 04 22:08:45 that's what i was doing and I was using STATIC buffers to maintain a consitent link to the DSP Nov 04 22:09:09 so apparently python was not the issue as much as the scope of the static library linked in Nov 04 22:09:14 ah Nov 04 22:09:15 cool Nov 04 22:09:22 I am glad you are understanding all this Nov 04 22:09:32 so I hacked the dsplink makefiles and made it build me a shared library Nov 04 22:09:42 it's been a learning curve :) Nov 04 22:10:20 so now i have a shared library that should do the trick i just can't seem to properly link it as a shared library to the custom blocks using Bitbake Nov 04 22:10:29 Tartarus: you can have whatever nickname you want Nov 04 22:11:20 so now I'm trying to copy the shared library over to the BB and do it locally. but i'm running into a Make issue ... missing "swig.swg" and "python.swg" which from a previous log a Mckenzie student ran into Nov 04 22:11:52 i built swig-native and opkg all the swig packages even the doc ones but i can't find it Nov 04 22:13:42 al_fayez: what's the high-level goal? i missed that part Nov 04 22:14:15 mrj10, as I understand it, al_fayez is building gnuradio blocks that use the dsp in an omap3 Nov 04 22:14:18 i'm using the DSP in python blocks with all of the DSPLink stuff handled in a library Nov 04 22:14:43 sounds awesome Nov 04 22:14:50 the library used to be static with with static variable handleing buffers and handlers Nov 04 22:14:56 yeah, gotcha Nov 04 22:15:14 the problem the scope is getting messed up as the blocks are used in a hierarchical fashion Nov 04 22:15:41 al_fayez, you should look for Mackenzies student Nov 04 22:15:45 he is on irc some Nov 04 22:15:45 so i hacked the dsplink makefiles to generate shared libraries in hopes of having seamlesss DSP operation from python Nov 04 22:15:48 not now though Nov 04 22:16:16 also, you might ask denix for some help Nov 04 22:16:29 which one? the navy guy ... can't remember his name Nov 04 22:16:49 Frank someone or other Nov 04 22:16:56 ask Dr MAcKenzie Nov 04 22:17:23 havng a framework to do this in OE will be really nice ... Nov 04 22:17:23 cool, I helped Frank a few times so he owes me :) Nov 04 22:17:39 having someone to bounce ideas off Nov 04 22:17:51 I will be in CA until Thanksgiing Nov 04 22:18:12 yeah as soon as this stupid demo deadline is over Nov 04 22:18:47 thanks for all your insight through out this ... i'm hoping that this code will be ready for beta release early in the year Nov 04 22:19:11 cool Nov 04 22:19:33 i still hack a few things instead of going the OE path completely just because it's difficult to figure out the simplest things in it Nov 04 22:19:56 heh, you have a hard problem on a good day Nov 04 22:30:46 hi there, can anyone review/merge these patches?: http://patchwork.openembedded.org/patch/3490/ http://patchwork.openembedded.org/patch/3492/ Nov 04 22:32:17 who takes care of patchwork? Nov 04 22:35:20 I mess with user permissions Nov 04 22:43:41 anarsoul: khem needs to look at / do 3490 Nov 04 22:44:04 And does 3492 still work w/ eric's clean out of versions? Nov 04 22:49:36 Tartarus: eFfeM_work : for you test scripts one interesting option to git clone is --reference path_to_already_fetched_tree : this saves time & bandwidth Nov 04 23:00:24 I just updated the page to use bitbake 1.10.1 and mention python SimpleHTTPServer. Go update with more comments please :) Nov 04 23:01:49 time to sleep, will do when I finish to setup my test build server which now seems to be stable (hope it's still stable tomorrow as it has a low cost motherboard ;-) Nov 04 23:07:26 Building apache2: | checking for /dev/zero... configure: error: cannot check for file existence when cross compiling Nov 04 23:07:30 Suggestions? Nov 04 23:08:06 Maybe just remove line from configure script Nov 04 23:17:19 Anyone? Nov 04 23:19:26 add it to the site files Nov 04 23:19:40 check config.status/config.log for the variable name Nov 04 23:19:49 Then add to openembedded/site/common-linux Nov 04 23:23:55 There is already: ac_cv_file__dev_zero=${ac_cv_file__dev_zero=yes} Nov 04 23:26:51 Compiling for armv7, maybe should be added into arm-common or arm-linux ? Nov 04 23:36:42 Other suggestions? Nov 04 23:47:41 Where is it today? Nov 04 23:48:07 And are you sure that's the variable your program is checking? Nov 04 23:54:53 anarsoul: eglibc patch looks ok Nov 04 23:55:04 I was asking to apply this patch to other eglibc versions too Nov 04 23:55:14 so I am sort of waiting on that one Nov 04 23:55:43 Crofton|work: hey if you are in CA bay area Nov 04 23:55:47 lets meet sometime Nov 04 23:55:52 yes Nov 04 23:55:59 when will you be here Nov 04 23:56:04 Tartarus: I had to back port fix for recipe from newer release Nov 04 23:56:08 Nov 8-22 or so Nov 04 23:56:19 Crofton|work: cool, Nov 04 23:57:15 give me a couple days to get my clock in order Nov 05 00:02:50 Is it safe to use Bitbake 1.10.1 with 8 months old OE release? Nov 05 00:05:36 davidlt: should be Nov 05 00:06:38 thanks Nov 05 00:07:25 Had to take 8 months release, backport some fixed to get apache2/php Nov 05 00:09:43 davidlt: you can use older apache 2.2.3 Nov 05 00:09:59 Yeah I know. Tried it :) Nov 05 00:10:06 But made 2.2.14 compile too. Nov 05 00:10:22 On the latest release PHP, mod_php doesn't compile :( Nov 05 00:10:33 you can export that var in do_configure too Nov 05 00:13:16 with mod_php problem part apache2, when apxs is launched it tries accessing some files which are not staged (fixed that). After that axps is looking for shared libexpat. Which is not located in i686-linux/lib. (So was was apache2-native build if doesn't even exist? took from host OS? Shouldn't it be isolated?) Nov 05 00:13:43 With php it can not find libcrypto (which is in armv7 and i686 lib dirs) Nov 05 00:14:05 And if you remove SSL support from configure, it can't find mysql, later another package. Nov 05 00:14:09 Something is messed. Nov 05 00:14:39 But the same 5.2.13 (PHP) compiles on 8 months old release. Nov 05 00:43:42 hello :) Nov 05 00:46:01 i just installed oe / bitbake following http://wiki.openembedded.org/index.php/Getting_started Nov 05 00:46:21 but when i try to run : bitbake nano i get : Nov 05 00:46:48 bash: /home/franck/POKY/bitbake/bin/bitbake: Permission denied Nov 05 00:47:13 -rwxr-x--x 1 franck franck 9059 15 oct. 02:45 bitbake Nov 05 00:47:42 do you know where i should go to check where i made an error ? Nov 05 00:48:32 spir, permissions on your python binary maybe? Nov 05 00:48:57 -rwxr-xr-x 1 root root 2288272 9 oct. 14:18 /usr/bin/python2.6 Nov 05 01:00:51 PATH=/home/franck/POKY/bitbake/bin:/home/franck/bin:/usr/local/bin:/usr/bin:/bin:/usr/games Nov 05 01:01:02 BBPATH=/home/franck/POKY/build:/home/franck/POKY/openembedded Nov 05 01:03:58 spir: dont use root Nov 05 01:04:12 you can damage the rfs Nov 05 01:04:29 some time OE might try to install into / by mistake Nov 05 01:04:35 or symlink Nov 05 01:04:40 etc. Nov 05 01:04:48 so always build using normal user Nov 05 01:08:01 i'm loged in as normal user, and there is no symlink in my path... Nov 05 01:08:16 spir are you using Fedora by any chance Nov 05 01:08:25 debian sid Nov 05 01:08:40 and i wget + tar zxvf bitbake Nov 05 01:08:43 in Fedora SSE Linux occassionally messes up OE Nov 05 01:09:14 when SSE prevents it from running a command it comes back with a permission denied message i'm not sure if Debia has something similar to that Nov 05 01:10:40 don't know neither Nov 05 01:12:06 I've never used debian so check if it has an extra layer of security to enforce user permissions Nov 05 01:12:27 al_fayez, I run full se linux Nov 05 01:12:40 if you do not need BINARY LOCALES you can disbale that Nov 05 01:13:11 ENABLE_BINARY_LOCALE_GENERATION = "0" Nov 05 01:13:21 that avoids needing to run qemu Nov 05 01:13:29 aah cool, I finally know why Nov 05 01:16:44 qemu does some weird things that can cuase issues with selinux Nov 05 01:16:59 I avoid binary locales becasue I am a native speaker Nov 05 01:26:20 hum... cd ~/bitbake/bin ; python bitbake Nov 05 01:26:24 ImportError: No module named bb Nov 05 01:26:38 maybe all the problem come from here ? Nov 05 01:27:02 ARGH Nov 05 01:27:44 (did i specify before that even i have in my PATH /home/franck/POKY/bitbake/bin ... i didn't get completion on "bitbake" ... Nov 05 01:28:05 oe makes me angry Nov 05 01:28:36 why it so wants to give me dbus, cups and other sucks-more crap? Nov 05 01:29:18 zomfg, started to build orbit2 now Nov 05 01:29:41 i don't want half of gnome or other DE shit Nov 05 01:30:01 just telled bitbake xserver-kdrive bash Nov 05 01:31:09 does everybody here uses all this insane stuff? Nov 05 01:34:18 ok, now going to sleep hoping i won't get some kind of ubuntu-arm in the morning Nov 05 02:09:12 definitly... completion on bash NEVER works when it comes to bitbake... Nov 05 02:09:47 even : cd bitbake/bin ; python bitba .... doesn't want to do the completion... Nov 05 02:10:05 i really can't understand what's going on :( Nov 05 02:34:17 hmm Nov 05 02:34:36 wonder if itd be worth using Dulwich to clone git repositories for the new fetch module rather than subprocess+git Nov 05 02:36:17 but doesn't it make you feel powerful to run 'top' and see processes with 5-digit PIDs? Nov 05 02:36:59 hehe Nov 05 02:41:50 Crofton do you have to modify "libtool" to prevent if from find *.la libraries even though you don't have them? Nov 05 02:42:05 remove the .la files :) Nov 05 02:42:18 I don't think you need to modify it Nov 05 02:42:31 from all of the beagleboard? Nov 05 02:43:04 or just gnuradio ones? Nov 05 02:43:21 hmm, dulwich could use some convenience functions, will have to write some **** ENDING LOGGING AT Fri Nov 05 02:59:58 2010