**** BEGIN LOGGING AT Wed Aug 11 02:59:57 2010 Aug 11 07:36:19 morning Aug 11 07:36:45 morning Aug 11 07:43:07 morning Aug 11 07:44:21 good morning! Aug 11 07:44:35 ncurses package doesn't have libncurses.so.5 any more Aug 11 07:44:45 (this morning upgrade) Aug 11 07:45:10 daniele_athome: it's in separate package Aug 11 07:45:17 oh... Aug 11 07:45:20 daniele_athome: if you notice it because bash, just reinstall bash Aug 11 07:45:24 it didn't get installed Aug 11 07:45:45 daniele_athome: because nothing is linked against it (newer bash is linked only to libtinfo Aug 11 07:45:45 actually because of htop Aug 11 07:46:02 -force-reinstall htop Aug 11 07:46:08 i'm already doing Aug 11 07:46:35 old Depends: libc6, ncurses Aug 11 07:46:36 JaMa, still doesn't work Aug 11 07:46:47 libncurses.so.5 => not found Aug 11 07:47:18 strange, mmt Aug 11 07:47:30 oops Aug 11 07:47:31 even bash now Aug 11 07:47:36 i can't logout :) Aug 11 07:48:12 reinstall bash Aug 11 07:48:16 ok... Aug 11 07:48:27 JaMa, which package libcurses is now in? Aug 11 07:48:53 new Depends: libc6, libncurses5, libtinfo5 Aug 11 07:49:03 if you reinstall htop it should work! Aug 11 07:49:07 of course Aug 11 07:49:18 neither bash worked Aug 11 07:49:18 and pull libncurses5 for you Aug 11 07:49:25 i'll install it manually Aug 11 07:49:31 if it doesn't then opkg remove htop -force-depends and install it again Aug 11 07:49:42 JaMa, there are problems with eglibc build Aug 11 07:49:45 regarding nptl Aug 11 07:49:47 did you know? Aug 11 07:49:58 | running configure fragment for nptl/sysdeps/pthread Aug 11 07:49:58 | configure: error: compiler support for __thread is required Aug 11 07:50:05 I don't have any problems with eglibc now Aug 11 07:50:12 ehm lol Aug 11 07:50:15 daniele_athome: seen that in eglibc-initial, right? Aug 11 07:50:19 JaMa, yes Aug 11 07:50:25 i guess clean and rebuild Aug 11 07:50:29 that's because it's using wrong gcc Aug 11 07:50:33 mmm Aug 11 07:50:37 when did you rebuild from scratch? Aug 11 07:50:53 JaMa, when you switched to 2.6.32 Aug 11 07:50:54 if before friday, than you have to do it again Aug 11 07:51:15 so the problem should be solved if I rebuild from scratch Aug 11 07:51:17 no problems Aug 11 07:51:17 :) Aug 11 07:51:31 i'll delete tmp directory Aug 11 07:51:33 daniele_athome: not possible.. OE layout changed since I switched to 2.6.32 and it forced rebuild from scratch Aug 11 07:52:00 JaMa: so all must be deleted? Aug 11 07:52:03 JaMa, indeed, i rebuilt from scratch a day after the switch to 2.6.32 Aug 11 07:52:05 daniele_athome: and then it was changed again without OE layout bump, but needed also rebuild from scratch (cross dir is gone and toolchain paths changed) Aug 11 07:52:11 ok Aug 11 07:52:12 very well Aug 11 07:52:15 rebuilding so Aug 11 07:52:18 vanous: yes, keep only cache Aug 11 07:53:30 JaMa: where ? shr-unstable/tmp_cache? Aug 11 07:53:39 tmp/cache? Aug 11 07:53:43 Y Aug 11 07:53:53 empty here, so i will just rm -rf all Aug 11 07:53:54 vanous: that's where git revisions are stored Aug 11 07:54:19 and keep only makefile, correct? Aug 11 07:54:24 vanous: you don't have tmpdir-dev-shr/cache/om-gta02/bb_persist_data.sqlite3 ? Aug 11 07:54:31 vanous: and OE checkouts.. Aug 11 07:54:34 oh, this one? Aug 11 07:54:35 i do :) Aug 11 07:54:37 vanous: only shr-unstable/tmp Aug 11 07:54:56 i thought there were more in there Aug 11 07:54:57 ok Aug 11 07:55:06 then only tmp, but will keep cache, good Aug 11 07:55:32 JaMa, is it possible the system or X hanging because of /tmp filesystem full? Aug 11 07:55:41 i mean /var/volatile Aug 11 08:02:00 daniele_athome: possible.. I've seen ie some fso daemon eating 100% when it cannot log to /var/log (volatile) Aug 11 08:02:19 mmm Aug 11 08:02:25 so it could be this the cause of my hangs Aug 11 08:02:31 JaMa, my suite logs a lot :D Aug 11 08:02:39 i'm redirecting to /dev/null now... let's see Aug 11 08:03:03 especially my window manager is pretty a log eater Aug 11 09:17:55 mrmoku: I'm already building Angstrom with -Os (as we have) instead of -O2 Aug 11 09:18:02 mrmoku: to check alignment traps Aug 11 09:18:06 JaMa: interesting :) Aug 11 09:30:04 hi guys! I read yesterday that you spoke about nano. so perhaps you have my answer to this :Error opening terminal: xterm. Aug 11 09:31:07 nano build was failing because of ncurses.. fixed in http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=db4786b7bef7abbefdd561fb3a5e42377da3a547 no idea about nano opening xterm.. Aug 11 09:32:21 GarthPS: but you're right, I have it too in reflashed image (the same version from my build works ok in 2nd image) Aug 11 09:35:02 JaMa: ok. I though it was fix in repo Aug 11 09:35:10 but not yet commited :) Aug 11 09:35:23 JaMa: by bro reported the same last night Aug 11 09:36:20 maybe opkg install ncurses-terminfo ? Aug 11 09:37:02 hmm, will have to ask him a i haven't reflashed Aug 11 09:37:30 yup that worked Aug 11 09:39:56 JaMa: yop too :) Aug 11 09:40:00 thx! Aug 11 09:40:09 yw Aug 11 09:45:52 Testing building can somebody suggest a component package which don't have many dependencies. libphone-ui-shr results in over 4,000 packages being built. Aug 11 09:47:00 Arigead: stagemanager-native Aug 11 09:47:19 JaMa: thanks a million that might save me a few hours ;-) Aug 11 09:47:46 it depends on what you're trying to accomplish :) Aug 11 09:50:30 I'm just trying to see if I can get local builds to work. mrmoku said that there's a bit in bitbake, well I assume that it's in bitbake. Anyhow was going to try a different version of bitbake to see if I can get around that problem Aug 11 09:51:37 Arigead: no it's in one of bbclasses AFAIK Aug 11 09:52:05 openembedded then Aug 11 09:52:10 :-( Aug 11 09:53:07 Arigead: maybe mrmoku pointed to this commit? http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=5ac52371b09a25518b60e23e80c8c812cdea6e36 Aug 11 09:56:10 JaMa: was that a commit that broke things or fixed things? Aug 11 09:56:34 Arigead: I think broke Aug 11 09:56:49 but wait for mrmoku to confirm it or just debug yourself :) Aug 11 09:56:49 Arigead: it fixed other things... but broke local builds Aug 11 09:57:12 that commit and another one restoring backwards compat for older bitbake Aug 11 09:57:39 JaMa: problem is that dest is pointing to the source file Aug 11 09:57:44 and thus it does not get copied Aug 11 09:58:27 but we wanted to investigate the correct way to do local builds anyway :) Aug 11 09:58:53 mrmoku: can I use an older version of the openembedded tree? It sounds like a simple enough problem to fix locally? can bbclasses be edited Aug 11 09:59:32 Arigead: don't use older OE if you want resulting packages compatible with current shr Aug 11 09:59:50 JaMa: well that answers that question ;-) Aug 11 10:00:07 and you can edit bbclasses :) Aug 11 10:00:32 but make update will drop your changes (so just cd openembedded; git pull --rebase; instead Aug 11 10:00:47 and ofc commit your changes before git pull Aug 11 10:01:07 I'm surprised as OE is used on quite a few projects and local builds seems like a fairly important feature to break. Aug 11 10:01:51 JaMa: cheers for that. I'm still only a very basic git user but I'll take a note of that. Aug 11 10:03:41 * mrmoku is trying srctree now... Aug 11 10:04:02 Arigead: you can still rsync your source between -c unpack and -c patch (as I do now) Aug 11 10:04:35 Arigead: or "proper" way of doing local builds (srctree+giver) as mrmoku is now trying Aug 11 10:05:03 Arigead: local builds as described in SHR wiki are not really a feature of OE Aug 11 10:05:47 yeah, a convenient hack though... which was working nicely :) Aug 11 10:16:08 any idea why xalan-j is failing for me? Here is log http://wklej.org/id/375221/ Aug 11 10:18:17 soltys: yes.. [oe] problem building openjdk-6 on oe-devel ML Aug 11 10:18:33 soltys: same here and shr buildhost already reported to java guys on #jalimo Aug 11 10:21:27 JaMa: thanks for info ;) Aug 11 10:28:24 JaMa: heh Aug 11 10:28:25 FATAL: Unable to start to 'knotty' UI: 545 Aug 11 10:28:54 WTH is knotty? Aug 11 10:56:23 mrmoku: /usr/lib/python2.6/site-packages/bb/ui/knotty.py Aug 11 10:56:53 # BitBake (No)TTY UI Implementation Aug 11 11:08:43 JaMa: yep, but why does it fail? Aug 11 11:09:02 hmm Aug 11 11:09:15 on second run it did not fail Aug 11 11:10:19 JaMa: ok, srctree and gitver seem to work if I put a copy of the bb to the srctree Aug 11 11:10:30 * mrmoku looking into that amend thing now Aug 11 11:10:46 mrmoku: it doesn't fail here Aug 11 11:11:24 JaMa: nvm, just failed once Aug 11 11:12:26 JaMa: what do you think is preferable? Have a copy of the bb for local builds in the repo of the project? Aug 11 11:12:40 or try to make it work with amend.bbclass Aug 11 11:12:43 ? Aug 11 11:14:07 the first one would probably be easier for everyone... Aug 11 11:14:39 hmm no strong opinion about that, we can sync local build bb with real one when real one is changed Aug 11 11:14:50 with amend it won't be needed so ofter I think Aug 11 11:14:58 s/ofter/often/ Aug 11 11:14:58 JaMa meant: with amend it won't be needed so often I think Aug 11 11:15:18 but I never used srctree :) Aug 11 11:15:25 or amend Aug 11 11:15:31 IIUC amend means we have inherit amend in the original recipe Aug 11 11:15:47 and user must stuff an amend.inc somewhere with the correct content... Aug 11 11:15:54 or in shr.conf maybe Aug 11 11:16:39 I think I prefer the 'duplicate recipe' way :-) Aug 11 11:16:56 Arigead: I will commit a recipe for local building to libphone-ui... in case you want to try Aug 11 11:17:10 ok, it's probably easier to read/understand Aug 11 11:19:11 SHR: 03mok 07libphone-ui * raa6b9ac4ed7a 10/libphone-ui.bb: add a bitbake recipe for local building Aug 11 11:19:40 mrmoku: that would be most agreeable. And Can I say that I really like the duplicate recipes method in bitbake collections for local builds. OK it don't work a the moment Aug 11 11:20:54 Arigead: bitbake -b /path/to/your/libphone-ui/libphone-ui.bb should just work now :) Aug 11 11:20:55 if you override a recipe you can easily still build the origional recipe or find out which version or recipe you're actually building. Putting things into files like local-builds.inc Aug 11 11:21:57 just to a git pull and it should work. That sounds just too easy. Thanks a millio Aug 11 11:22:00 million Aug 11 11:22:43 yw Aug 11 11:23:07 there was a bit need to solve this issue on my side too... otherwise Q-Master will spank me someday for not fixing that segfault ;) Aug 11 11:23:17 s/bit/big/ Aug 11 11:23:18 mrmoku meant: there was a big need to solve this issue on my side too... otherwise Q-Master will spank me someday for not fixing that segfault ;) Aug 11 11:23:43 8) Aug 11 11:25:42 Arigead: btw. you first have to build the 'normal' libphone-ui Aug 11 11:25:49 because -b does not build the dependencies Aug 11 11:26:58 mrmoku: deadly cheers Aug 11 11:28:15 mrmoku: confirmed, O2 doesn't do alignment traps, Os does Aug 11 11:28:23 JaMa: wow Aug 11 11:28:27 ~lart gcc Aug 11 11:28:27 * apt pulls out a ClueBat (tm) and thwaps gcc Aug 11 11:28:50 JaMa: rebuild from scratch for changed CFLAGS? Aug 11 11:29:26 * JaMa trying to rebuild just eglibc Aug 11 11:29:33 mrmoku: that ain't sane! Aug 11 11:29:41 mrmoku: gcc is most probably not guilty ;) Aug 11 11:29:59 mrmoku: hmm well maybe I'm not right.. ignore that last message Aug 11 11:29:59 PaulFertser: no? who then if O2 works and Os not? Aug 11 11:30:27 mrmoku: almost all optimisation-related issues were found to be bugs in the software, not in the compiler. Aug 11 11:31:17 mrmoku: actually it's vice-versa.. Angstrom build even with Os (instead of O2) doesn't have those traps Aug 11 11:32:31 PaulFertser: point taken... Aug 11 11:32:34 ~lart eglibc ;) Aug 11 11:32:35 * apt pours gasoline all over eglibc ;), ignites the fire, and then enjoys some toasty marshmallows with the glorious blaze Aug 11 11:33:02 JaMa: but shr built with O2 is fine? Aug 11 11:33:34 mrmoku: haven't tried yet.. Aug 11 11:33:52 mrmoku: is it really that hard to determine the place those alignment traps come from? You all folks have a running system with plenty of them, and you can use gdb right on the target... Aug 11 11:33:54 mrmoku: now only diff between CFLAGS is -ggdb3 Aug 11 11:35:35 rebuilding Angstrom without -ggdb3 and with Os now Aug 11 11:42:13 PaulFertser: it's in here: Aug 11 11:42:14 4004b000-4015f000 r-xp 00000000 b3:01 9096 /lib/libc-2.12.1.so Aug 11 11:42:25 how to check the exact place with gdb? Aug 11 11:43:00 [22223.020000] Alignment trap: opkg (617) PC=0x4000a7dc Instr=0xe7952008 Address=0x401565eb FSR 0x013 Aug 11 11:43:45 err Aug 11 11:43:48 here 40000000-4001b000 r-xp 00000000 b3:01 9102 /lib/ld-2.12.1.so Aug 11 11:44:22 mrmoku: i'd to start the process in gdb placing a breakpoint beforehand. Aug 11 11:44:32 mrmoku: lindi- definitely knows better how to use gdb Aug 11 11:44:58 ok, thanks :) Aug 11 11:46:58 PaulFertser: khem already said that those are from lib loader and that he fixed similar bugs before.. so we were waiting for him to fix it faster than we could.. Aug 11 11:50:43 JaMa: earlier you mentioned some (unrelated?) futex alignment bugs. Aug 11 11:52:09 hmm... if I tell the kernel to not fix things gdb gives me just SIGBUS without bt :/ Aug 11 11:53:01 PaulFertser: I've noticed fix for futex alignment traps in eglibc svn log.. so I was hoping it's realated to our traps.. but after building it I can confirm it's not Aug 11 11:56:06 mrmoku: without bt? what do you get with "x/16i $pc"? Aug 11 11:57:52 (gdb) x/16i $pc Aug 11 11:57:52 No registers. Aug 11 11:58:18 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=548842 sounds interesting Aug 11 12:00:16 mrmoku: ok at least the address of 0x401565eb is a clear alignment problem Aug 11 12:02:15 mrmoku: I think I'm just silently letting kernel do software emulation for those Aug 11 12:04:00 lindi-: ahh, you have them too and just warning turned off? Aug 11 12:04:30 hi all Aug 11 12:04:31 JaMa: do you know that you have broken bash in feed? Aug 11 12:04:47 mrmoku: I don't think I have them often Aug 11 12:05:31 mrmoku: cat /proc/cpu/alignment? Aug 11 12:05:32 alexxy: no, it's not broken in feed, just reinstall it from feed to get right version Aug 11 12:06:12 JaMa: i cannot reinstall it Aug 11 12:06:20 because i alredy rebooted phone Aug 11 12:06:25 and i cannot login Aug 11 12:07:04 lindi-: yep Aug 11 12:07:12 mrmoku: i mean, what do you have there :) Aug 11 12:07:20 ahh :P Aug 11 12:07:38 is there an acronym for "please run the command X and put the output online somewhere, e.g. paste.debian.net?"? ;) Aug 11 12:07:44 lindi-: I set it to 4 to get the SIGBUS Aug 11 12:07:58 default we have is 3 I think (warn+fixup) Aug 11 12:08:05 mrmoku: I had 0 Aug 11 12:08:08 mrmoku: I set it to 1 now Aug 11 12:08:35 mrmoku: what does 0 really do? Aug 11 12:08:39 ignore Aug 11 12:08:54 no warn, no fixup, no sibus Aug 11 12:08:55 +g Aug 11 12:09:00 mrmoku: so it reads random data? Aug 11 12:09:54 mrmoku: i guess it does fixups with 0, or else the system would crash? Aug 11 12:10:03 http://www.mjmwired.net/kernel/Documentation/arm/mem_alignment Aug 11 12:10:19 PaulFertser: could crash probably, yes Aug 11 12:10:28 mrmoku: yes I did read that from kernel source Aug 11 12:10:47 mrmoku: it does not really say what happens if it is zero Aug 11 12:10:49 mrmoku: ok, i see Aug 11 12:12:34 mrmoku: 0 is "ignored" Aug 11 12:13:40 dos1: what does "lux" mean? :) Aug 11 12:13:42 mrmoku: 0 causes set_cr(cr_no_alignment); Aug 11 12:23:10 PaulFertser: lux? Aug 11 12:23:45 dos1: yes, some polish guy uses it as an acknolegment or smth Aug 11 12:24:23 PaulFertser: dunno, i have never seen that (nor use :x) Aug 11 12:24:55 dos1: ok, thnx :) Aug 11 12:46:30 PaulFertser, lindi-: via core file I get some backtrace Aug 11 12:46:38 #0 0x4000a7dc in _dl_relocate_object () from /lib/ld-linux.so.3 Aug 11 12:46:38 #1 0x4001c938 in ?? () Aug 11 12:46:38 #2 0x4001c938 in ?? () Aug 11 12:47:19 mrmoku: btw, that patch in the ticket looks partially reasonable Aug 11 12:48:55 * mrmoku unpacks eglibc to see if it would somehow fit... Aug 11 12:52:19 PaulFertser: looks like it might even apply :) Aug 11 12:52:51 mrmoku: i do not think it's exactly correct Aug 11 12:53:24 mrmoku: can you then do x/16i $pc? Aug 11 12:54:23 lindi-: yep :) http://shr.pastebin.com/FR5LkCX5 Aug 11 12:55:04 mrmoku: ok and info register? Aug 11 12:55:56 lindi-: http://shr.pastebin.com/9tZxrw0m Aug 11 13:01:00 mrmoku: ok hmm. r8 is clearly odd Aug 11 13:01:32 mrmoku: (the patch by that very clever guy) i think the idea is that gcc will generate code that avoids unaligned access for the variables that are ((packed) Aug 11 13:02:45 mrmoku: probably it even does, but then there might be some other place that accesses the same address inappropriately. Aug 11 13:05:35 mrmoku: did the problem start when you upgraded ld-linux or when you upgraded opkg? Aug 11 13:07:28 lindi-: _any_ app gives two alignmentraps when it's running Aug 11 13:07:43 don't remember when it started Aug 11 13:08:07 mrmoku: ok Aug 11 13:08:10 JaMa: do you have better memory? ^^^ Aug 11 13:09:25 yes.. when we first built eglibc-2.12 with gcc-4.5.0 Aug 11 13:09:47 and I have eglibc-2.12 built with same gcc which doesn't do it Aug 11 13:10:04 in few minutes I'll have it also without -ggdb3 to check that Aug 11 13:12:12 JaMa: point is, is the bug in ld-linux or opkg binary or both? Aug 11 13:14:06 lindi-: afaik only ld-linux and then shown from every binary http://paste.pocoo.org/show/247924/ Aug 11 13:15:08 and see this http://paste.pocoo.org/show/248873/ Aug 11 13:15:32 only difference (I can think of) between OsAngstrom and OsSHR is -ggdb3 in CFLAGS Aug 11 13:15:45 both build with -Os now, same toolchain, same version of eglibc Aug 11 13:16:19 JaMa: how can you get different binaries when you recompile with same compiler? Aug 11 13:16:48 maybe -ggdb3 is not stripped later exactly the same as without it? Aug 11 13:17:08 I can give you both binaries if you want to diff them Aug 11 13:18:54 JaMa: I can try Aug 11 13:19:07 dumped to hex and diff is quite big Aug 11 13:19:23 JaMa: disassembly is probably what you want to diff Aug 11 13:20:30 http://jama.homelinux.org/Others/libc.ang http://jama.homelinux.org/Others/libc.shr Aug 11 13:25:41 JaMa: libc? Aug 11 13:25:51 JaMa: weren't we talking about ld-linux? Aug 11 13:34:35 this was only for test that 2 same builds are not exactly the same and running ie libc.so.6 emits it.. check that dir, there are also ld-2.12.1.so + *.dis Aug 11 13:36:33 and disassembly is different from beginning Aug 11 13:58:35 JaMa: .text starts at different addresses Aug 11 13:59:47 some first ~ 100B of text look similar, then it diverges a lot Aug 11 14:00:13 are you sure the only difference is -ggdb? Aug 11 14:00:20 ti looks like different code and/or compiler Aug 11 14:01:04 zub: .plt starts 20 later .text 30 later Aug 11 14:01:10 err also 20 later Aug 11 14:01:57 zub: the real difference is Angstrom/SHR, but I have P_V for gcc/binutils/eglibc the same in Angstrom and SHR Aug 11 14:02:22 zub: so remaining part is diff distro/include/angstrom-eglibc.inc distro/include/sane-toolchain-eglibc.inc Aug 11 14:02:29 I think -ggdb should not affect text; let me try with some sample code Aug 11 14:02:40 zub: in the morning I unified Os/O2 Aug 11 14:02:49 yeah, I read Aug 11 14:02:50 looks pretty identical through diff -u <(sed 's/[0-9a-f]//g' < ld-2.12.1.so.ang.dis) <(sed 's/[0-9a-f]//g' < ld-2.12.1.so.shr.dis)|less -S Aug 11 14:02:54 so only -ggdb3 is only diff now in this Aug 11 14:03:52 zub: and after "NOTE: Running task 1577 of 2087" I'll try Angstrom with -Os and also without -ggdb3 to minimize the diff Aug 11 14:03:56 even more if you add -ab to diff Aug 11 14:04:14 EH_FRAME is then about the only difference Aug 11 14:04:52 strange you don't have: Aug 11 14:04:54 74c: 000218b4 undefined instruction 0x000218b4 Aug 11 14:05:03 76c: 00022894 muleq r2, r4, r8 Aug 11 14:05:04 Aug 11 14:05:06 ? Aug 11 14:05:21 JaMa: but that's just a memory address Aug 11 14:05:59 JaMa: there's a load from that address to $pc just before it Aug 11 14:06:53 ah I see "unin instrution x" :) Aug 11 14:07:10 lindi-: I was surprised that it's not in "your" simplified diff Aug 11 14:07:14 lindi-: but I read it wrong Aug 11 14:12:55 just tested generating a hello world with -ggdb3 vs. w/o, then comparing .text and I get the same thing Aug 11 14:14:08 gNOTE: Running task 1992 of 2087 Aug 11 14:16:22 and done! :) Aug 11 14:17:43 and without -ggdb it does traps! http://paste.pocoo.org/show/248896/ Aug 11 14:21:20 shr build ang Angstrom build withtout ggdb3 are exactly the same from disassembly pov Aug 11 14:21:38 so it indeed is -ggdb? Aug 11 14:21:41 strange Aug 11 14:22:10 yes Aug 11 14:22:25 my ^ -ggdb test was with 4.3.3 Aug 11 14:22:39 (just something I have here) Aug 11 14:23:59 so the culprit is positively identified as gcc? Aug 11 14:24:06 zub: alignment traps started with gcc-4.5 Aug 11 14:24:27 zub: even before 4.5.0 release and now in 4.5.2 pre Aug 11 14:25:08 I can compare unstripped binaries (with and without ggdb3), but not today.. I have to go Aug 11 14:25:25 well yes you should find the minimal case that reproduces this Aug 11 14:25:38 JaMa: good work Aug 11 14:27:29 lindi-: I'll talk with Khem Raj (OE toolchain master) maybe he will know right away Aug 11 14:27:44 cya Aug 11 14:29:37 maybe they dropped support for ARMv4? ;) Aug 11 14:29:39 * lindi- hides... Aug 11 14:29:48 heh Aug 11 15:07:37 hi mickeyl leviathan Aug 11 15:07:50 GNUtoo|laptop: hi Aug 11 15:07:50 I've bad wifi+vpn connection so I'll try to be brief Aug 11 15:07:56 ok Aug 11 15:07:58 I can't answer a call Aug 11 15:08:00 I could before Aug 11 15:08:04 hmm Aug 11 15:08:15 are the alsa state files guilty? Aug 11 15:08:21 dunno Aug 11 15:08:30 what happening exactly? Aug 11 15:08:30 note that I've not yet looked at the logs Aug 11 15:08:41 I answer Aug 11 15:08:43 and no sound... Aug 11 15:08:51 I talk Aug 11 15:08:53 you do hear no sound? Aug 11 15:08:56 and don't ear the contact Aug 11 15:09:06 hmm Aug 11 15:09:09 I'm not even shure the contact can ear me Aug 11 15:09:11 I know this problem Aug 11 15:09:22 the contact can hear you Aug 11 15:09:23 ok Aug 11 15:09:23 ;-) Aug 11 15:09:28 ah ok Aug 11 15:09:28 only you can't hear them Aug 11 15:09:35 what's the cause then? Aug 11 15:09:36 try to make a phonecall outside Aug 11 15:09:41 e.g. combox Aug 11 15:09:45 then it works again Aug 11 15:09:49 something with routing Aug 11 15:09:50 dunno Aug 11 15:09:55 happens sometimes Aug 11 15:10:03 it happened 2 times Aug 11 15:10:10 first time when I got 2 calls at once Aug 11 15:10:18 then second time when I got 1 call Aug 11 15:10:25 hmm Aug 11 15:10:33 try the call out fix Aug 11 15:10:36 should work Aug 11 15:50:02 hi Defiant Aug 11 15:50:03 oops Aug 11 15:50:05 sorry Aug 11 15:50:09 hi Deubeuliou Aug 11 15:50:15 hi :P Aug 11 15:50:34 sorry I wanted to talk to Deubeuliou Aug 11 15:50:43 but if you want to talk to me it's ok Aug 11 16:11:15 anybody of you .DE ppl from around Erlangen/Nurnberg? Aug 11 16:37:45 * mrmoku is located a little bit more south Aug 11 16:44:54 zub: planning some trip to the uni in erlangen? Aug 11 16:46:27 hi mrmoku Aug 11 16:46:27 are the new htcdream images built? Aug 11 16:47:27 GNUtoo|laptop: dunno... let me check Aug 11 16:47:54 ok Aug 11 16:48:52 GNUtoo|laptop: latest is from around 6:30 this morning Aug 11 16:48:56 so I would say yes :-) Aug 11 16:49:40 ok thanks a lot Aug 11 16:49:40 any tester yet? Aug 11 16:56:25 mickeyl: hey... Aug 11 16:56:31 mickeyl: i'm still struggling with stability Aug 11 16:56:39 under om-gta02 Aug 11 16:56:40 me too Aug 11 16:56:51 I've huge stability issues Aug 11 16:57:00 sometime I've the impression that it freeze Aug 11 16:57:05 but that's kernel related Aug 11 16:57:11 mickeyl: wanna see my latest logs? Basically one command times out (+CREG=2 on resume) and then immediately the modem turnes to CLOSING -> CLOSED and i'm fucked up because i have no idea that happened. Aug 11 16:57:43 is it sage to upgrade om-gta02 Aug 11 16:57:45 ? Aug 11 17:05:51 GNUtoo|laptop, ping. Aug 11 17:06:04 GNUtoo|laptop: did not hear about anyone testing it Aug 11 17:08:09 I believe he's not here. Aug 11 17:09:24 TAsn: he disliked your ping ;) Aug 11 17:10:09 My pings usually affect people in extreme ways. Aug 11 17:10:40 PaulFertser: you still have signal strength and provider when that happens? Aug 11 17:11:00 TAsn: yeah :P Aug 11 17:11:02 mrmoku: what do you mean? fsogsmd closes the modem resource etc Aug 11 17:11:28 PaulFertser: I'm refering to the unnoticed... Aug 11 17:11:41 mrmoku: yes, i think no signal is emitted Aug 11 17:12:02 mrmoku: but even if it is, it can't help if my phone is auto-suspended short after: i'd not see the indicator anyway. Aug 11 17:12:10 mrmoku, people barely remember the last guy I pinged. Aug 11 17:12:14