**** BEGIN LOGGING AT Tue Nov 22 02:59:58 2011 Nov 22 06:24:09 khem wrt the uclibc question: i am using r34.3+gitr0+162cfaea20d807f0ae329efe39292a9b22593b41 Nov 22 06:24:50 unfortunately I cannot sell moving to oe-core to mgmt; this product is to be completed in a few weeks and no additional potential risks will be accepted I fear Nov 22 06:27:02 this is what gdb gives me with busybox static: Nov 22 06:27:03 Program received signal SIGSEGV, Segmentation fault. Nov 22 06:27:03 0x00000000 in ?? () Nov 22 08:48:09 good morning Nov 22 09:05:50 hi mckoan Nov 22 09:17:49 good morning Nov 22 09:18:19 hi florian, all Nov 22 09:19:11 hi mckoan Nov 22 09:31:12 mornin Nov 22 09:47:01 hi Jin^eLD Nov 22 09:50:40 good morning Nov 22 09:51:09 for people going to Fosdem : is there a recommended hotel ? Nov 22 09:51:34 and are there people staying there on synday evening ? Nov 22 09:56:29 ericben: quite some people leave on sunday Nov 22 09:56:31 * eFfeM_work will Nov 22 09:57:08 and traditionally oe people stay in the Astrid (although I haven't, was booked by the time I wanted to book) Nov 22 09:58:09 ericben: I usually staid at Ibis next to Grand Place Nov 22 10:06:53 ok thanks I'll book the hotel this week Nov 22 10:07:48 astrid is still availabel for fosdem weekend Nov 22 10:49:32 hi Nov 22 10:51:42 khem: Hello; libgcc is failing to build here Nov 22 10:52:03 khem: it claims gcc-4.6-linaro-r106723.patch has already been applied Nov 22 10:52:24 otavio: using debian and make 2.82 ? Nov 22 10:52:34 khem: it seems meta-oe needs to be updated Nov 22 10:52:54 pwgen: 3.81-8 make Nov 22 10:53:05 pwgen: but yes, debian Nov 22 10:53:07 otavio: i vaguely recall having seen something about this on the ML yesterday or so Nov 22 10:53:45 eFfeM_work: I think meta-oe has not been synced with oe-core Nov 22 10:53:47 " gcc-4.6: Fix gcc ICE on qt4-x11-free/armv7-a " .. the last commit made oe-core fail .. Nov 22 10:53:58 could well be Nov 22 10:55:44 go back one revision. ( this helped me yesterday , git checkout f0e17d436b865edeb3dd113ac617433639982ded ) Nov 22 10:56:05 pwgen: are you using meta-oe too? Nov 22 10:56:17 yes .. Nov 22 10:57:05 pwgen: it seems I fixed it locally; will test it and if it works I send a pull request Nov 22 11:00:47 hi all: I have compiled oe-core for beagleboard (virtual/kernel) can I now change PARALLEL_MAKE and BB_NUMBER_THREADS from two to 8, as I have 8 cpus? should it work? Nov 22 11:03:54 : how many thrads your system han handle is depending on many things. harddrive.memory and so on. Nov 22 11:04:11 try a value of 4 . Nov 22 11:04:24 and increase the parrallel mage Nov 22 11:04:27 and increase the parrallel make Nov 22 11:09:14 5GB RAM, 8cpus, ssd very fast Nov 22 11:22:18 recalcati: I think setting to 8 would be ok, just give it a try, also use -j8 or os Nov 22 11:23:22 -j8 and 4 for threads number I'm going Nov 22 11:24:08 it happens that fetch has errors and stops the compilatio, is there any timeout that I can increase, maybe? Nov 22 11:58:54 : u can use the bitbake -k to continue building ( dependend packages will not be build ) Nov 22 12:00:07 : sometimes it helps googeling for missing packages and put them by hand in your downloads folder Nov 22 12:02:26 you shouldn't have to do that Nov 22 12:02:32 just fix the SRC_URI in the errant recipe Nov 22 12:07:37 ... and if this manual fix is working, inform someone to update the SRC url ...(:-)) Nov 22 12:35:00 JaMa: I think I found a problem with vesa xorg driver Nov 22 12:35:22 JaMa: it seems to hard-depends on dgaproto Nov 22 12:48:37 pwgen: thx, I had forgotten it! Nov 22 13:08:40 03Richard Purdie  07master * r94c97534ad 10bitbake.git/lib/bb/runqueue.py: Nov 22 13:08:40 bitbake/runqueue.py: Fix incorrect task number reference in debug message Nov 22 13:08:40 Signed-off-by: Richard Purdie Nov 22 13:08:50 03Richard Purdie  07master * r78c02276c7 10bitbake.git/lib/bb/runqueue.py: Nov 22 13:08:50 bitbake/runqueue.py: Sort the list of skipped tasks as it makes searching the list easier when debugging Nov 22 13:08:50 Signed-off-by: Richard Purdie Nov 22 13:08:51 03Richard Purdie  07master * r2f6d987d99 10bitbake.git/lib/bb/runqueue.py: (log message trimmed) Nov 22 13:08:51 bitbake/runqueue.py: Add BB_SETSCENE_VERIFY_FUNCTION hook Nov 22 13:08:51 The OE metadata has cases where it needs to prevent some setscene Nov 22 13:08:51 functions from running. An example of this is where we know a task Nov 22 13:08:51 is going to run do_configure (which would clean out do_populate_sysroot) Nov 22 13:08:52 and hence we don't want do_populate_sysroot_setscene to run. Nov 22 13:08:52 This change adds in a hook so that the metadata can allow any Nov 22 13:32:47 damn, I always forget, what package in OE-core spits out the ldd utility? Nov 22 13:45:14 Jin^eLD: ldd is part of the (e)glibc Nov 22 13:45:31 (but I know almost nothing about OE - I just started to read the user manual) Nov 22 13:45:43 oops :) indeed, after you said that - grepped the deploy dir and it's already buil Nov 22 13:45:50 t Nov 22 13:46:42 thnks :> Nov 22 13:46:43 : ) Nov 22 15:26:31 otavio: yes koen should apply that patch soon Nov 22 15:27:01 khem: is there a place where I can get it? Nov 22 15:27:11 patchwork Nov 22 15:28:29 http://patchwork.openembedded.org/patch/15113/ Nov 22 15:28:40 apply it to meta-openemebedded Nov 22 15:29:09 khem: found Nov 22 15:29:14 ok Nov 22 15:29:34 khem: thx anyway :) Nov 22 15:29:34 there can be some sync up time between layers we can not avoid it Nov 22 15:29:43 different maintainers Nov 22 15:29:54 and one of them might be hit by bus or something Nov 22 15:30:06 khem: sure. I should have noticed this patch before Nov 22 15:30:09 please bear with lag Nov 22 15:30:18 khem: heh Nov 22 15:31:11 otavio: did u go to LinuxCon in Brazil Nov 22 15:50:14 khem: I did; I met sgw over there Nov 22 15:50:19 khem: it was great Nov 22 15:58:21 n\ Nov 22 15:59:18 t Nov 22 17:25:47 hi anarsoul Nov 22 17:27:21 hi bluelightning Nov 22 17:27:45 anarsoul: have you by any chance built recent kernels (3.1+) for h1940? Nov 22 17:28:10 built 3.0+ for rx1950, it's almost the same Nov 22 17:28:16 still have to unbrick my h1940 :) Nov 22 17:28:31 bluelightning: can try this weekend Nov 22 17:29:10 bluelightning: it does not work? Nov 22 17:29:30 anarsoul: I haven't tried either Nov 22 17:29:41 anarsoul: what happened to your h1940? Nov 22 17:30:19 flashed new version of u-boot with wrong nand timings, and then flashed newer version with it :) Nov 22 18:00:28 bluelightning, I think I am going to have questions about qt "soon" Nov 22 18:00:37 unless I give up working due to horrordays Nov 22 18:03:58 Crofton|work: horror days? Nov 22 18:04:39 anarsoul: doh Nov 22 18:05:28 holidays Nov 22 18:05:42 but so money of them compresse into the end of the year Nov 22 18:05:54 and travel and dealing with familiy members Nov 22 18:06:04 it feels like the horror days at times :) Nov 22 18:06:52 Crofton|work: ok; it sounded like you already made some progress with cmake though? Nov 22 18:09:39 well, I eleeted my qmake recipe Nov 22 18:09:43 deleted Nov 22 18:13:53 need to make sure qmake2 binary comes from the right place though Nov 22 18:19:58 bluelightning, do you know of any packages finding qt ok using cmake? Nov 22 18:20:22 I think the cmake macro runs some qmake commands that produce host paths for inccludes Nov 22 18:20:42 I just blew away tmp to make sure I have the right binaries Nov 22 18:21:20 Crofton|work: I did have some of my own (well, I say my... they were the rudimentary kde recipes we had in oe classic ported over) Nov 22 18:26:10 Crofton|work: can we pick up this discussion later? I'd like to finally sort all of this stuff out but unfortunately I have to head home... Nov 22 18:27:06 no problem Nov 22 18:27:17 It will be a bit before I am organized Nov 22 18:42:49 I'm feeling dumb today. Can someone explain what this does: PACKAGES =+ "${PN}-route ${PN}-nf ${PN}-genl ${PN}-cli" Nov 22 18:43:11 adds packages to the recipe Nov 22 18:43:32 wher ${PN} is the package name usually set from the recipe name Nov 22 18:43:37 But when I try to bitbake foo-genl, it says nothing provides that. Nov 22 18:43:43 so you can create addtional packages Nov 22 18:44:05 make sense? Nov 22 18:44:21 But when I try to bitbake foo-genl, it says nothing provides that. Nov 22 18:45:12 ${PN} is not foo Nov 22 18:45:25 build libnl and you will get libnl-genl for free Nov 22 18:46:17 sure, s/foo/libnl/g Nov 22 18:46:17 awozniak: you cannot build only one PACKAGE from it, it alwasy builds all and then splits them to runtime packages according to PACKAGES Nov 22 18:46:33 does it make seperate ipks ? Nov 22 18:46:46 awozniak: so if you want to build them all just do "bitbake libnl" Nov 22 18:46:49 yes Nov 22 18:47:39 but some other recipe can RDEPENDS="libnl-genl" or whatever? Nov 22 18:49:33 R in the beginning means runtime package Nov 22 18:49:39 so this is ok Nov 22 18:52:33 so is this done mostly just to save space on the target? Nov 22 18:53:47 yes if you don't need them all then shlibs will pull only whatever subset is neede Nov 22 18:54:36 you usually don't need to set RDEPENDS="libnl-genl" if the recipe produces something linked to libnl-genl (shlibs will add it for you automagically) Nov 22 19:46:19 gm Nov 22 19:56:56 http://www.linuxfordevices.com/c/a/News/Intel-Embedded-Software-Development-Tool-Suite-23-/ Nov 22 19:57:10 eFfeM: gm Nov 22 19:57:45 nice link, thanks Nov 22 19:58:56 what's the thing called oe-lite? Nov 22 19:58:59 khem, do you by any chance have a hint for my uclibc problem? I tried stepping into the code with gdb, seems I get a NULL deref somewhere in stdio_init (if I recall the name properly) Nov 22 19:59:29 eFfeM: hmmm, what version of uclibc is it ? Nov 22 20:00:42 khem, originally I had a git version of about a year ago, then tried git head, but that did not compile, then tried 0.9.32 (actually the git version with the 0.9.32 hash) and that also had the problem, gcc 4.5 iirc Nov 22 20:01:40 if your version does not have this patch http://pastebin.com/K4Yd3eBR Nov 22 20:01:44 please try it Nov 22 20:01:50 it might help Nov 22 20:02:11 xxiao: at the risk of being imprecise: a build system similar to oe; with some different twists and changes, with some different design ideas; see also #oe-lite here on freenode and oe-lite.org Nov 22 20:03:21 khem, will verify, but the patch is from before my checkout, so guess it is there Nov 22 20:03:41 hmmm Nov 22 20:03:54 can you use oe-core by any chance Nov 22 20:04:09 since I know uclibc works on all architectures there Nov 22 20:04:42 as you describe the above patch comes to mind Nov 22 20:04:52 thats why I suggested that Nov 22 20:05:07 checking for the patch now, I can try to give oe-core a spin, but I don't think I can sell to mgmt to move to a new toolchain so close to release Nov 22 20:05:31 sure Nov 22 20:06:04 are you on maintainance branch ? Nov 22 20:07:02 hi khem Nov 22 20:07:05 and are you using nptl or lt Nov 22 20:07:09 hi woglinde Nov 22 20:07:32 khem maintenance branch of uclibc or oe ? Nov 22 20:07:40 eFfeM: of oe. Nov 22 20:08:12 hi effem Nov 22 20:08:15 no I think we froze somewhere before don't really recall, btw distro is minimal (and minimal-uclibc) Nov 22 20:08:16 hi woglinde Nov 22 20:08:22 only 2 oral exams left Nov 22 20:08:57 we are using nptl Nov 22 20:09:00 eFfeM: thanks. Nov 22 20:09:10 than I am finished Nov 22 20:09:49 xxiao: ask EsbenH for details, he is the lead oe-lite developer, but it is evening in his timezone now so he might not be at his terminal Nov 22 20:09:55 woglinde:nice Nov 22 20:10:04 effem hm after 8 years Nov 22 20:10:05 khem, the patch is there Nov 22 20:10:12 woglinde: :-) Nov 22 20:10:20 thats not so good Nov 22 20:11:21 eFfeM: I would suggest to try tip of maintainance branch Nov 22 20:11:40 since that will be closer to what you have and have a better shot at working Nov 22 20:12:01 woglinde: getting there eh ? Nov 22 20:12:13 khem yes Nov 22 20:12:14 woglinde: what do u plan to do after that Nov 22 20:12:25 working fulltime Nov 22 20:12:32 woglinde: cool. Nov 22 20:12:37 woglinde: on oe and stuff ? Nov 22 20:12:39 or something else Nov 22 20:12:52 khem, will kick off a build for it, not sure if I can do all from home (and actually I should be preparing my thursday class, otherwise my students will be disappointed) will also give oe-core a try Nov 22 20:12:53 hm only when we get some offers Nov 22 20:13:07 effem students? Nov 22 20:13:08 khem, could it be that -Os is causing damage? noticed that this was in the uclibc .config Nov 22 20:13:15 eFfeM: sounds ok Nov 22 20:13:30 effem did you test runtime? Nov 22 20:13:39 I did only building Nov 22 20:13:42 the last time Nov 22 20:13:49 woglinde: I teach one evening a week at a nearby technical university Nov 22 20:13:54 eFfeM: hmmm could be, I have fixed it on oe.dev long time ago but I dont know if you are using that or not Nov 22 20:14:06 effem hehe about what? Nov 22 20:14:08 eFfeM: as I said try latest maintance branch Nov 22 20:14:10 embedded linux? Nov 22 20:14:29 used to teach network management, now had to move to introducing computer systems, need to explain basics of IP Nov 22 20:14:43 khem, will do, will also try w/o -Os Nov 22 20:14:49 ok Nov 22 20:14:55 sufficient ideas for a day Nov 22 20:15:02 btw I am using an 8313 Nov 22 20:15:17 effem uhm Nov 22 20:15:21 woglinde: it should not matter as far as uclibc is concerned Nov 22 20:15:26 at a technical university= Nov 22 20:15:29 that sounds strange Nov 22 20:15:31 err that was for eFfeM Nov 22 20:15:32 khem this is what it says: UCLIBC_EXTRA_CFLAGS="-fexpensive-optimizations -fomit-frame-pointer -frename-registers -Os" Nov 22 20:15:42 eFfeM: it should be ok Nov 22 20:16:01 woglinde: first term Nov 22 20:16:11 hm Nov 22 20:16:13 SIGFOOD Nov 22 20:16:13 okay Nov 22 20:16:15 bbiab Nov 22 20:17:00 khem, enjoy Nov 22 20:17:06 woglinde: http://fontys.edu Nov 22 20:18:12 hehe Nov 22 20:18:22 bsc in food and flower management Nov 22 20:20:35 hm 40k student Nov 22 20:20:37 thats a lot Nov 22 21:40:51 calling it a day, khem I'll let you know how I proceeded with uclibc Nov 22 21:40:54 nite all Nov 23 02:38:09 Anyone know where I can get the source for psplash? Nov 23 02:40:34 cbrake: how are you liking submodules, out of curiosity? Nov 23 02:52:11 It looks like it has disappeared since Intel took over. **** ENDING LOGGING AT Wed Nov 23 02:59:57 2011