**** BEGIN LOGGING AT Fri Mar 05 02:59:58 2010 Mar 05 05:48:40 TAsn: phoneui-info.c:605: error: implicit declaration of function 'phoneui_phone_log_new_call' so libphone-ui failed and libphone-ui-shr wasn't built ;/ Mar 05 06:23:27 grr Mar 05 06:23:48 * mrmoku stupid Mar 05 06:25:37 hmm Mar 05 06:28:18 JaMa: which rev is that? Mar 05 06:30:51 JaMa: rebuilding from scratch now? Mar 05 06:48:07 freesmartphone.org: 03mickey 07msmcomm * rbfbc5a9d7873 10/ (55 files in 10 dirs): use top level autofoo w/ subdirs; the projects are closely coupled anyways Mar 05 06:58:58 morning folks Mar 05 07:01:10 mickey|preSprint: morning :) Mar 05 07:07:31 freesmartphone.org: 03mickey 07msmcomm * r53b5f4191425 10/configure.ac: fix configure.ac to not require msmcomm; this _is_ msmcomm Mar 05 07:07:38 morrning Mar 05 07:07:45 morning spaetz Mar 05 07:08:26 mrmoku: latest and yes Mar 05 07:09:06 NOTE: Running task 326 of 7231, only but sofar no segfault Mar 05 07:09:16 morning Mar 05 07:10:27 mrmoku: but lukhas from bearstech sait they will replace that ram next week, so not sure why it works sofar and if it's safe to rebuild already :/ Mar 05 07:12:08 * spaetz will try to get a fso-abyss backtrace today.... Mar 05 07:14:29 freesmartphone.org: 03morphis 07msmcomm * r5ea107026703 10/libmsmcomm/ (src/message.c src/msmcomm.h vapi/msmcomm-1.0.vapi): libmsmcomm: add message copy function to simplify the use of msmcomm in vala Mar 05 07:14:30 freesmartphone.org: 03morphis 07msmcomm * r235a5081c233 10/configure.ac: Merge branch 'master' of git.freesmartphone.org:msmcomm Mar 05 07:16:01 JaMa: hmm... latest should compile :/ Mar 05 07:16:18 freesmartphone.org: 03mickey 07msmcomm * r7d4b07d8ed0b 10/msmvterm/src/Makefile.am: msmcomm: fix msmvterm building Mar 05 07:23:02 mrmoku: is it autorev right? Mar 05 07:23:11 mrmoku: it failed also here locally Mar 05 07:24:36 DOH, fso-abyss.config!!!!!!!!!!!11!! Mar 05 07:24:48 pkg fs-abyss-config installs to /etc/fso-abyss.conf Mar 05 07:25:11 but we need /etc/freesmartphone/*/fso-abyss.conf, rigth? Mar 05 07:25:46 what is the correct abyss config package? Mar 05 07:28:08 fso-abyss is using the stock FSO system Mar 05 07:28:11 which means the search path is Mar 05 07:28:26 | /etc/.conf Mar 05 07:28:29 | /etc/freesmartphone/.conf Mar 05 07:28:35 | /etc/freesmartphone/conf//.conf Mar 05 07:28:49 first one found gets the t-shirt and the coffee machine Mar 05 07:29:54 ahh, so it should be picking up /etc/fso-abyss.conf k, thanks Mar 05 07:30:00 ohh coffee! Mar 05 07:30:04 * spaetz goes grab one Mar 05 07:30:27 shouldn't it look for the machine specific one first? Mar 05 07:30:37 not that it should matter though Mar 05 07:32:53 no, i want users to be able to override Mar 05 07:33:01 JaMa: duuh yeah... have to push :/ Mar 05 07:33:09 without having to dive into the conf/ dirs Mar 05 07:34:18 mrmoku: seems like the broken ram was stripped from our xen host 636/1792MB.. I guess we had 2G before Mar 05 07:34:40 mrmoku: so hopefully it will build ok now Mar 05 07:34:47 hehe :) Mar 05 07:34:59 SHR: 03mok 07libphone-ui * r97fda168d8d5 10/src/ (phoneui-info.c phoneui-info.h): phoneui-info: add callback list for call changes - right now only new calls are handled Mar 05 07:35:17 JaMa: ok now it *must* build :) Mar 05 07:35:23 * JaMa trying Mar 05 07:35:53 moin Mar 05 07:36:40 ok, you go first. Mar 05 07:36:51 and when you are done-I compile shr-testing there :) Mar 05 07:37:56 OK :) Mar 05 07:39:09 but IIRC Heinervdm reported yesterday that webkit-efl fails with newer EFL we have (and fails here too with newest EFL) and also python-edje.. so hopefully full image will be built soon too Mar 05 07:41:52 * graziano will be happy to test the new testing when ready, if that could help Mar 05 07:42:26 graziano: well, there is a testing candidate out there, but fso-abyss crashes, so no gsm yet... Mar 05 07:43:14 spaetz: as you know I am testing the already wonderful work made with the previous testing and reporting my little findings here: http://wiki.openmoko.org/wiki/User:Graziano Mar 05 07:44:04 if you know some way I could be of help that doesn't require coding, I am gald to help Mar 05 07:44:09 *glad Mar 05 07:44:56 graziano: I agree that mokonnect could go into the -lite image btw Mar 05 07:45:24 I read some thoughts on the shr settings page about having some wifi tool integrated there Mar 05 07:45:38 don't count on it any time soon. Mar 05 07:46:16 and once it's there it will probably just be mokonnect integrated into shr-settings or so :) Mar 05 07:46:32 yes that is exactely what I was thinking Mar 05 07:46:41 *exactly Mar 05 07:47:03 screen is also not in the l Mar 05 07:47:05 mrmoku: yes now it compiles fine, thanks Mar 05 07:47:06 ite image Mar 05 07:47:39 spaetz: what's screen? Mar 05 07:51:29 a nice utility that let's you "multiplex" a console. Mar 05 07:51:42 that is, you can disconnect and reconnect later and still use the same console Mar 05 07:51:47 spaetz: removed in -lite cleanup (moved to -full) Mar 05 07:51:58 or log in as many times as you want on one ssh connection Mar 05 07:52:05 are you shipping task-cli-tools in full? if not, you should probably Mar 05 07:52:07 and switch between those windows Mar 05 07:52:12 same for task-cli-tools-debug Mar 05 07:52:16 at least until we have a release ;) Mar 05 07:52:22 JaMa: ahh ok, that makes sense Mar 05 07:52:49 spaetz: ah THAT screen :D Mar 05 07:53:17 mickey|preSprint: that is my fso-abyss log Mar 05 07:53:20 http://shr.pastebin.com/UR0h0Pad Mar 05 07:53:22 strace. Mar 05 07:53:36 mickey|preSprint: no, but I'll add it.. good point Mar 05 07:53:40 spaetz: feel free to add to that page everything you want to Mar 05 07:53:40 why would it receive a SIGTERM? Mar 05 07:55:44 good q. Mar 05 07:58:17 * spaetz doesn't know ERESTART_RESTARTBLOCK Mar 05 07:59:25 fso-abyss doesn't die all the time. Now, I started it in gdb and I get the SIM PIN dialog. Mar 05 07:59:37 fun :/ Mar 05 08:00:47 spaetz: that looks like the race we have since some time... Mar 05 08:00:58 spaetz: try to set frameworkd loglevel to DEBUG and see if it works then Mar 05 08:01:29 can we enable coredumps? Mar 05 08:01:39 then we could try to attach with gdb later on Mar 05 08:01:40 and get a traceback Mar 05 08:01:53 without gdb tainting the process like Heisenberg Mar 05 08:09:18 hmm eglibc is already building and bears are sleeping so nobody to set minimal mmap in time :/ Mar 05 08:09:40 setting frameworkd logging to DEBUG made it work. rebooting and retrying Mar 05 08:10:13 mrmoku: thanks for the hint. How does shr-u solve this? Mar 05 08:10:27 and by solve I mean a) short-term and b) properly :) Mar 05 08:12:43 mrmoku, mickey|preSprint: right, rebooted, SIM PIN dialog comes up just fine (frameworkd.log =DEBUG) Mar 05 08:12:49 spaetz: no solution yet Mar 05 08:12:52 hmm, k Mar 05 08:13:18 time for fso-gsmd? :-) Mar 05 08:13:38 not necessarily Mar 05 08:13:50 if the problem is in libgsm0710mux, then fsogsmd will suffer as well Mar 05 08:14:05 can we get a traceback from a coredump? Mar 05 08:14:28 we need to enable core dumps in the kernel, right? Mar 05 08:15:08 ulimit -c unlimited Mar 05 08:17:59 spaetz: it is the problem of fso-abyss Mar 05 08:22:32 mickey|preSprint: tried: this is the debug output of fso-abyss: http://paste.in.ua/410/ Mar 05 08:22:36 no core file was created Mar 05 08:23:09 Q-Master: it seems to work when frameworkd logging is set to DEBUG, so it must be some race Mar 05 08:24:09 spaetz: it is the problem of fso-abyss not correctly working with debug IMO. There's no problem in frameworkd. I've checked this. Mar 05 08:26:06 I have some libgsm0710mux logs that show what happens http://pastebin.com/AtnGYCuB Mar 05 08:26:56 the reply comes just after the timeout Mar 05 08:28:29 hmm Mar 05 08:28:43 okay, if it just that, give me a second Mar 05 08:29:31 but I don't think it is just a matter of increasing the timeout, is it? Mar 05 08:29:51 the reply comes way too quick after it gives up Mar 05 08:32:18 I suspect the same , I were thinking to increase an decrease the timeout and see the log Mar 05 08:38:27 * spaetz tries to give the modem more time to wake up (device_wakeup_waitms = 300) Mar 05 08:40:01 no process may coredump on a missing reply, no? Mar 05 08:40:29 sounds utterly odd to me Mar 05 08:40:31 [09:22] spaetz | no core file was created Mar 05 08:40:43 It SIGTERMs Mar 05 08:40:52 aah, ok Mar 05 08:41:36 so it's a missing syslog print there, giving hint on the fact it does sigterm and why Mar 05 08:41:48 strace here: http://shr.pastebin.com/QtPCsKZa Mar 05 08:42:13 the last thing it does is write(6, "~\3\357#\17WAKEUP!p~", 14) Mar 05 08:42:16 :) Mar 05 08:42:42 what's 6? Mar 05 08:43:28 I think /dev/ttySAC0 Mar 05 08:43:37 not sure though Mar 05 08:43:39 :-P Mar 05 08:49:34 OK, must be around these 3 lines: Mar 05 08:49:36 write(6, "~\17\357AT%CUNS=2\r\n\357~", 16) = 16 Mar 05 08:49:36 poll([{fd=8, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN, revents=POLLIN}], 3, -1) = 1 Mar 05 08:49:39 read(6, "~\r\357\r\nOK\r\n5~", 4095) = 11 Mar 05 08:49:55 it writes to one channel (line 1), it hten hangs and times out in line 2) Mar 05 08:50:30 and immediately after it gives up waiting it reads from fd=6 (the channel we have written to in line 1), so it gets the reply then Mar 05 08:50:43 that seems to be consistent what the libmux log showed... Mar 05 08:53:01 freesmartphone.org: 03mickey 07libgsm0710mux * r861b234386de 10/gsm0710mux/ (consts.vala manager.vala multiplexer.vala): Mar 05 08:53:01 freesmartphone.org: provide configurable channel ACK timeout in seconds; default is: Mar 05 08:53:01 freesmartphone.org: [libgsm0710mux] Mar 05 08:53:01 freesmartphone.org: device_ack_timeout = 5 Mar 05 08:53:30 wtf is that "~" for 1. char of each read/write? Mar 05 08:53:50 frame header Mar 05 08:53:54 this is 07.10 Mar 05 08:54:02 aah Mar 05 08:54:04 thanks Mar 05 08:54:12 bbiab, moving to the train station Mar 05 08:56:45 spaetz: looks like a sane sequence of actions Mar 05 08:59:18 DocScrutinizer: wanna see a deformed usb socket on FR? Looks like some really bad socket was used. Mar 05 08:59:24 spaetz: nah, actually if it really timeouts on line2, rather than catching a reply on handle6, then it's wrong Mar 05 08:59:28 DocScrutinizer: http://brolin.be/graphics/FreeRunner_Mini_USB_F_5-pin.jpg Mar 05 09:00:02 oh, shiiiiit Mar 05 09:00:30 PaulFertser: or a really nasty plug was inserted Mar 05 09:01:02 DocScrutinizer: i'd say criminally nasty plug. Never seen one though. ;) Mar 05 09:03:11 PaulFertser: how'd you run into that? Mar 05 09:03:53 DocScrutinizer: no idea, it's from Brolin (now-famous) phone, he said he used it the usual way, no nasty falls, no nasty cables... Mar 05 09:04:32 who or what is Brolin Mar 05 09:05:20 DocScrutinizer: Brolin is the guy famous for his frank descriptions of his mental illness and difficulties in socialising on openmoko-community ML. Mar 05 09:05:39 uhuh Mar 05 09:06:08 seems there's more reasons than I've been aware of NOT to read [community] Mar 05 09:07:10 DocScrutinizer: hehe Mar 05 09:09:47 No wonder ar6000 authors remain anonymous, or else that'd earn them an eternal disrespect :E Mar 05 09:10:05 lol Mar 05 09:11:05 * DocScrutinizer suspects that to be major reasoning behind a lot of closed source Mar 05 09:11:39 wpwrak: In fact i suspect that wmi part of the driver was written by some other, more sane developer, and the linux part of the driver by some "less talented" "coder". It'd be interesting to evaluate what amount of effort it'd require to port the driver to cfg80211 and what the end result will be like. Mar 05 09:11:53 yay, 1st test worked with a different device_ack_timeout Mar 05 09:12:33 DocScrutinizer: i didn't have many chances to look at proprietary code but it looks very much like that. Mar 05 09:13:24 spaetz: I noticed that bitbake on shr buildhost is now using -j1 and 1 thread.. so maybe you can start shr-t in paralel if you really want.. because shr-u will take LONG (in last 4 hours => NOTE: Running task 436 of 7231) Mar 05 09:13:42 spaetz, mrmoku: what about switching bitbake branch to master? Mar 05 09:13:52 spaetz: would you agree on above sketched code sequence must not read from #6, after seeing timeout on poll() Mar 05 09:13:54 ok, thanks. Is it much different (master)? Mar 05 09:14:18 spaetz, mrmoku: I'm using it, it has few smaller issues (ie tinderbox-client.bbclass needs small hack) Mar 05 09:14:38 DocScrutinizer: I think it must not wait too long for reading from one line, so the other line has already timed out before we can read its reply Mar 05 09:14:54 spaetz, mrmoku: but RP is trying to make it new recommended version for oe.dev and we WANT that too, because better support for SRCPV Mar 05 09:15:32 spaetz, mrmoku: I've already switched that in shr-u checkout for current build (till next make update) Mar 05 09:15:40 OK, that solved it indeed. Rebooted, I get a SIM PIN dialog now. Mar 05 09:15:47 I reduced device_ack_timeout to 2 Mar 05 09:15:50 spaetz: why "wait on one line"? AIUI the poll is waiting for filehandles #8, #3 and #6 in parallel Mar 05 09:16:14 mmh, might actually be. true Mar 05 09:16:52 so if it sees a timeout, the following case construct must not try to read from any of these Mar 05 09:17:43 (case cunstruct a mere anticipation of mine ;-) Mar 05 09:18:45 if you share a URL, I might want to have a look at actual sourcecode Mar 05 09:22:19 spaetz: ^^^ Mar 05 09:29:50 DocScrutinizer: the stuff must be happening around here: http://git.freesmartphone.org/?p=libgsm0710mux.git;a=commit;h=861b234386de071822d5f7726205058858e1e668 Mar 05 09:30:04 but I think it works with a lower timeout. Mar 05 09:31:17 spaetz: "it works" doesn't mean it's correct. I bet there's a missing "else" catchall for the error case of poll timeout Mar 05 09:35:09 agreed that "works" != "correct" Mar 05 09:35:33 but I have never seen the code, so I can't say what would be correct here. Mar 05 09:38:31 spaetz: where can I get the new libgsm0710mux? my buildhost is on task 380 of 8447 Mar 05 09:39:19 maybe we should modify the logger a little bit to add file and line to the logs Mar 05 09:40:15 jl_gene: I just pushed the change to shr/unstable and shr/testing2010 Mar 05 09:40:30 spaetz: you just cherry-picked last 2 to shr/unstable? without merge? Mar 05 09:40:57 thanks Mar 05 09:40:59 I merged Mar 05 09:40:59 jl_gene: shr buildhost is Running task 437 of 7231 Mar 05 09:41:08 was that wrong? Mar 05 09:41:17 spaetz: strange.. I merged too and now it's twice there Mar 05 09:41:41 spaetz: which usually happen only when cherry-picked and then merged or rebased in shr/unstable and then merged Mar 05 09:42:31 ohh, actually I cherry picked from dev to shr/unstable Mar 05 09:42:35 you are right yes. Mar 05 09:42:51 because there were more conflicts when I tried to merge and I did not have time for that :) Mar 05 09:42:52 ok.. no problem I was just curious :) Mar 05 09:42:59 and then I merged unstable->testing Mar 05 09:43:22 5 out of 5 tests working now Mar 05 09:43:51 before it worked 2 out of 25 times or so. Mar 05 09:43:57 so *something* is different Mar 05 09:45:00 I really want to replace mokonnect with something that is more stable... sigh Mar 05 09:45:10 pressing "scan" wifi and no feedback at all... Mar 05 09:45:39 spaetz: have you tried that nwa? qt based? Mar 05 09:45:49 nwa? never heard of it. Mar 05 09:46:02 is it in the feed? Mar 05 09:46:39 ok, I feel comfortable in pushing this out to users soon. Mar 05 09:48:12 spaetz: can you try TAsn test? write some utf8 characters with numeric keyboard in phoneui-messages? Mar 05 09:48:51 mmh, setting an ffalarm shows "stereout.state: could not find alsa state" Mar 05 09:49:50 where are those alsa states supposed to be again? Mar 05 09:49:58 spaetz: do you have alsa-scenarii-shr? Mar 05 09:50:00 /etc/freesmartphone/? Mar 05 09:50:23 no Mar 05 09:50:35 spaetz: I've added provider for it to fsodeviced, but let compatibility-providers still pointing to alsa-scenarii-shr to pull both to image now (just because of ffalarms) Mar 05 09:50:49 as a matter of fact, I don't... Mar 05 09:51:08 spaetz: can you please try if installing that fixes it for you? Mar 05 09:51:42 spaetz: then we can put RDEPEND on alsa-scenarii-shr directly instead of provider and then remove it as soon as lupan says it's converted to fsodeviced api completely Mar 05 09:53:07 something is weird there. If I look at the depends of ffalarms it says: Mar 05 09:53:13 "ffalarms" -> "alsa-scenarios" [style=dashed] Mar 05 09:53:18 but I don't have that installed Mar 05 09:53:22 that's right Mar 05 09:53:30 alsa-scenarios is provider Mar 05 09:53:50 provider name, now provided also by fsodeviced Mar 05 09:54:07 but it works properly only for build dependencies IIRC Mar 05 09:54:24 so for runtime dep it pulled fsodeviced (not respecting compatibility-providers.conf) Mar 05 09:54:33 in fact, no "scenario|ii" package has been build at all. Mar 05 09:54:36 mmh Mar 05 09:55:22 spaetz: shit, I can't even locate the related lines of code :-/ Mar 05 09:55:58 Doc: nope, it's not that obvious :-( Mar 05 09:55:59 spaetz: which you wouldn't think should be too hard, given we got searchpatterns like "CUNS" Mar 05 09:56:15 JaMa: I'll try to install alsa-scenarii-shr directly Mar 05 09:57:18 right, I see that: fsodeviced_git.bb:RPROVIDES_${PN} = "openmoko-alsa-scenarios virtual/alsa-scenarios" Mar 05 09:58:13 JaMa: but fsodeviced does not pull in any scenario package, does it? Mar 05 09:58:23 although it claims to provide packages Mar 05 09:58:54 spaetz: and here is that missing part of that commit http://gitorious.org/~jama/angstrom/jama-shr-experimental/commit/63f2d2fd5b5b8139cf714e12488502ef662540ee Mar 05 09:59:11 spaetz: fsodeviced package contains all needed scenarios Mar 05 09:59:26 /etc/freesmartphone/conf/openmoko_gta/alsa-default/gsmheadset Mar 05 09:59:58 spaetz: but they are not compatible with ffalarms AFAIK (maybe just wrong location) Mar 05 10:00:08 right. Mar 05 10:02:12 so we need either to fix ffalarms to use the fsodeviced files, or to directly pull alsa-scenarii-shr Mar 05 10:03:07 10:52:23 < JaMa> spaetz: then we can put RDEPEND on alsa-scenarii-shr directly instead of provider and then remove it as soon as lupan says it's converted to fsodeviced api completely Mar 05 10:03:17 ahh :-) Mar 05 10:04:32 spaetz: I was expecting that alsa-scenarii-shr will be pulled even after setting that RPROVIDER.. Mar 05 10:05:17 spaetz: but as it fails I think that easiest sollution is to modify RDEPENDS and then remove it again with SRCREV bump for version setting scenario through fsodeviced Mar 05 10:06:34 I agree Mar 05 10:07:11 I'll install the scenarios and see if it fixes that, but explicitely RDEPENDing on it seems ok for now. Mar 05 10:08:23 spaetz: I just found this: http://wiki.openmoko.org/wiki/NWA Mar 05 10:10:21 mmh, and something needs to pull in connman-plugin-udhcp Mar 05 10:13:24 * spaetz makes mokonnect rdepend on udhcp plugin Mar 05 10:21:10 spaetz: mickey|zzZZzz: I wonder... Mar 05 10:21:14 462 public int read( void* data, int len ) Mar 05 10:21:15 463 { Mar 05 10:21:17 464 assert( logger.debug( "0710 -> should read max %d bytes to %p".printf( len, data ) ) ); Mar 05 10:21:18 465 var numread = transport.read( data, len ); Mar 05 10:21:20 466 hexdump( false, data, numread, logger ); Mar 05 10:21:21 467 return numread; Mar 05 10:21:23 468 } Mar 05 10:22:37 can't transport.read() ever fail? Mar 05 10:23:02 well, if it times out.... Mar 05 10:26:12 spaetz: exactly Mar 05 10:32:03 mrmoku, spaetz: btw what about moving some obsolete shr recipes to obsolete dir? Mar 05 10:32:03 JaMa: as for bitbake, I don't care which version we use as long as it works. Mar 05 10:32:17 or removing them. It's a versioning system :) Mar 05 10:32:50 +1 for doing away with them Mar 05 10:33:26 I just pushed a change to have 1) mokonnect depend on connman-plugin-udhcp and 2) fflarms depend on shr-alsa-scenarii Mar 05 10:34:02 moved already libframeworkd-phonegui-efl* shr-contacts_git.bb shr-dialer_git.bb shr-messages_git.bb, but didn't push it withour your ACKs :) and there is probably more obsoleted recipes.. Mar 05 10:34:54 ffalarms worked with shr-alsa-scenarii Mar 05 10:35:03 jl_gene: are you still waiting for that package? Mar 05 10:35:21 which package? Mar 05 10:36:16 jaMa yes Mar 05 10:37:39 jl_gene: you can use http://jama.homelinux.org/org.openembedded.shr/armv4t/libgsm0710mux0_0.9.0+gitr3+861b234386de071822d5f7726205058858e1e668-r0.4_armv4t.ipk Mar 05 10:37:52 don't forget to edit fso-abyss.conf Mar 05 10:38:03 jl_gene: fso-abyss-config_0.9.1+gitr1+b13b3b31975d5c921326d38a528627cff53b8bb9-r1.4_om-gta02.ipk fso-abyss_0.9.1+gitr1+b13b3b31975d5c921326d38a528627cff53b8bb9-r1.4_armv4t.ipk Mar 05 10:38:22 thanks Mar 05 10:38:31 jl_gene: fso-abyss-config is in om-gta02 directory instead armv4t Mar 05 10:40:28 I did edit the config when mikey pushed the patch Mar 05 10:40:48 ohh, those default dimming timeouts are a pain. ACK for increasing them? Mar 05 10:41:26 increasing timeout? (I don't care much - my FR is always connected :)) Mar 05 10:42:30 spaetz: mickey|zzZZzz: anyway, no matter if that's vala generic mainloop and opaque to the sourcecode, or if it's actually some verbatim code in libgsm0710mux and/or relatives: if a central wait timeouts, it mustn't read from an IO the wait was waiting for. I tend to think that's a bug in vala mainloop probably - quite uneducated guess Mar 05 10:46:36 actually s/wait/ (2) poll/ Mar 05 10:47:56 ahh, we use the timeouts directly from the cornucopia repository. Mar 05 10:50:45 ok, done for today. Compiling from scratch now. BBL Mar 05 10:52:06 spaetz: {fd=6, events=POLLIN, revents=POLLIN}] and return = 1 actually doesn't look like the poll timed out though Mar 05 10:52:41 * spaetz shrugs. I have no clue about that stuff. Mar 05 10:52:51 so basically there's nothing fishy at all with the 3 strace lines you pasted above Mar 05 10:54:34 spaetz: according to "man 2 poll" the revents=POLLIN means there's data to read for filedescriptor #6, so the subsequent read from #6 is perfectly sane Mar 05 10:56:34 mmh Mar 05 10:58:11 it sends "ATUNS=2" and gets "\r\nOK\r\n" for reply Mar 05 10:58:40 ok, so that might not have been the culprit then. Mar 05 10:58:46 no reason it sighups later on Mar 05 11:07:23 so I think this issue needs further investigation, as there's no good storry why the longer timeout does fix the issue - heissenbug still Mar 05 11:08:27 JaMa: (bitbak) fine for me if it works :) Mar 05 11:10:36 spaetz: increasing the timeout to 6 fails with the same problem and the eleapsed debug message is never printed http://shr.pastebin.com/Qn3AgUNv Mar 05 11:11:30 actually, and now this is weird. Mar 05 11:11:43 my patch was wrong. I patched in +#channel ACK timeout (in seconds) Mar 05 11:11:44 +device_ack_timeout = 5 Mar 05 11:11:57 which is supposed to be the default according to mickey. Mar 05 11:12:21 However, it works now when it would fail without the version one commit earlier Mar 05 11:12:39 It currently works, but now I am really puzzled. Mar 05 11:12:44 I need food now. BBL Mar 05 11:14:46 spaetz: that's because you do not listen to Doc and trying magic instead of proper bugfixing process ;{ Mar 05 11:14:54 s/{/(/ Mar 05 11:14:54 PaulFertser meant: spaetz: that's because you do not listen to Doc and trying magic instead of proper bugfixing process ;( Mar 05 11:15:08 hm Mar 05 11:15:23 ;{) Mar 05 11:15:53 8{)} Mar 05 11:16:29 PaulFertser, as I said, I have never seen that code, I donĀ“t really know vala and and I have no clue how the modem communication is even supposed to lioook like. Mar 05 11:16:42 PaulFertser: LOL Mar 05 11:16:54 So how and where should I properly debug? Mar 05 11:16:55 PaulFertser: :-) well prolly you're right Mar 05 11:18:08 * PaulFertser needs to read tricky canopen specs and is not able to participate in debugging :| Mar 05 11:18:43 that "I don't like that" loop looks really nasty. I'm missing a wait, a increment for elapsed time, and I miss thwe debug( @"elapsed: $(t.elapsed())" ) in actual debug log Mar 05 11:19:31 canopen specs? wierd canned food for lunch? :-P Mar 05 11:20:46 err s/wait/sleep()/ Mar 05 11:22:58 DocScrutinizer: damn weird. Multi-layered multi-object tricky specifications that many (even supriningly simple) devices use for communication over CAN bus.. Mar 05 11:23:15 and actually I also dunno if a "do while" loop is actually same like a "do until ! " loop. I.E if the condition is really evaluated *after* first iteration, or maybe before Mar 05 11:23:47 CAN bus - hehe Mar 05 11:25:17 At least in C, C++, Java etc do while loop is of post-condition type. I would be surprised if Vala had it different. Mar 05 11:25:59 * DocScrutinizer hates "do while" loops, for being completely missleading in wording Mar 05 11:27:41 * DocScrutinizer suggests "until () do {}" loops to maximize ambiguity and confusion Mar 05 11:29:10 or go use COBOL ftw :-P Mar 05 11:29:22 VBA Mar 05 11:32:56 * DocScrutinizer wonders where t.elapsed is set to start value... o.O Mar 05 11:36:13 if that's a local object that gets instanciated inside the relevant waiting loop, then I guess I don't like vala Mar 05 11:37:54 DocScrutinizer: My bad http://shr.pastebin.com/e4eanTxG (is in multiplexer.vala) Mar 05 11:37:55 aah, 263 t.start(); Mar 05 11:38:55 DocScrutinizer, show me the line and i tell you if you like vala or not :) Mar 05 11:39:33 263, and I think the jury is still out on it ;-) Mar 05 11:51:00 playya: what's MainContext ? Mar 05 11:51:26 MainContext.default(); Mar 05 11:52:18 or MainContext.iteration() ? Mar 05 11:55:14 searching in valadoc I think it only can be a Glib.MainContext Mar 05 11:55:20 yes Mar 05 11:55:57 so what's iteration() supposed to do? Mar 05 11:56:50 ehhh. glib interals. but i think it's a step in the MainLoop Mar 05 11:57:01 yep figured Mar 05 11:57:15 thinking sometimes helps :-P Mar 05 11:57:36 actually *very* cryptic code Mar 05 11:57:49 hard to understand - or is that just me? Mar 05 11:58:46 Runs a single iteration for the given main loop. This involves checking to see if any event sources are ready to be processed, then if no events sources are ready and may_block is TRUE, waiting for a source to become ready, then dispatching the highest priority events sources that are ready. Mar 05 11:59:14 aiui a usleep() in place of mc.iteration(false) would better meet my notion of a wait loop Mar 05 12:02:56 a MainContext.iteration(false) might easily eat up all cpu cycles, and it's not clear (to me) if concurrent threads are even processed at all (when calling mc.iteration() in a tight loop like done here) Mar 05 12:03:17 that's how I get it, at least Mar 05 12:03:34 i think every thread should have an own mainloop Mar 05 12:03:56 mainloop.iterate(false); is like polling Mar 05 12:04:05 s/thread/ Mar 05 12:04:42 busy polling (without any usleep() ) is *bad* anyway Mar 05 12:05:30 probably that's why DEBUG cures the issue Mar 05 12:05:34 i think it should be a async method which is chancelable Mar 05 12:05:42 274 debug( @"elapsed: $(t.elapsed())" ); Mar 05 12:07:36 playya: (please keep in mind I'm a poor C++ coder, and have no idea whatsoever about vala) Mar 05 12:08:34 and to do such hacks you schould know each damn detail about the guts of mainloop etc Mar 05 12:11:02 my guts feeling says adding a usleep(50000) after 274 debug(... should fix it Mar 05 12:12:43 i think it need some work to fix it Mar 05 12:12:47 5afb35f6 (Michael 'Mickey' Lauer 2009-02-22 19:10:09 +0100 268) // FIXME: Ok, I don't like that, but until Vala supports asnyc dbus Mar 05 12:12:48 16bcfbbe (Michael 'Mickey' Lauer 2009-04-01 02:08:57 +0200 269) // on manager side, we have to live with it. Mar 05 12:13:03 and vala now supports async dbus servers Mar 05 12:13:04 I seen that Mar 05 12:13:10 ooh Mar 05 12:13:21 which is the best solution Mar 05 12:13:25 so for sure we should gat rid of that nasty hack Mar 05 12:13:29 get Mar 05 12:13:33 yes Mar 05 12:14:17 maybe it requires my patch for async delegates, too Mar 05 12:14:22 at some points Mar 05 12:14:38 aha :o) Mar 05 12:15:06 see my comment about my vala experience Mar 05 12:15:39 also [2010-03-05 12:18:42] that "I don't like that" loop looks really nasty. I'm missing a wait, a increment for elapsed time, and I miss thwe debug( @"elapsed: $(t.elapsed())" ) in actual debug log Mar 05 12:16:28 scratch all after "nasty." Mar 05 12:16:51 let me check the generated c code Mar 05 12:17:28 have fun Mar 05 12:18:26 it think it's sth like while(true){...if(..)break;} Mar 05 12:19:37 the 274 debug is never executed Mar 05 12:20:16 I think it depends on DEBUG setting Mar 05 12:21:46 http://shr.pastebin.com/cvycmCUm Mar 05 12:22:31 I've the logglevel set to DEBUG and 285 gets executed Mar 05 12:22:51 yeah, but if you disable debugging, it probably never is called Mar 05 12:23:00 ooh Mar 05 12:23:51 285 assert( logger.debug Mar 05 12:24:00 274 debug( Mar 05 12:24:23 dunno if those are identical or not Mar 05 12:25:36 maybe debug() gets kicked out at compile time Mar 05 12:25:41 the assert stuff is there to remove it later with --disbale-asserts Mar 05 12:26:06 anyway, bbl Mar 05 12:26:10 time for lunch Mar 05 12:26:38 mrmoku: ACK for moving obsolete to obsolete? :) Mar 05 12:28:25 mrmoku: I'll continue using bitbake master at buildhost and after a while we could add it to Makefile for the rest of builders :) but it works ok.. hopefully nobody will commit something which will break it as it still gets commits (in contrast to 1.8 branch) Mar 05 13:13:45 hi all Mar 05 13:16:58 JaMa: did you notice problems with running GL programs on SHR-u? Mar 05 13:17:04 I think some more patching is neede Mar 05 13:17:06 +d Mar 05 13:17:32 any example of GL program you use on FR? Mar 05 13:17:50 glxinfo :) Mar 05 13:17:56 heh :) Mar 05 13:23:58 JaMa: gtk+ fails to compile Mar 05 13:24:08 spaetz: I know Mar 05 13:24:19 Fuck I should have cherry-picked fixes and not merging into /testing Mar 05 13:24:34 spaetz: you can disable demos (remove _virtclass_native from SRC_URI in last gtk+ recipe) Mar 05 13:24:59 or 14:01:39 < XorA> JaMa: the one made from gdk-pixbus-csource-native seems to work Mar 05 13:25:09 let's just do it in shr/unstable, should we? Mar 05 13:25:40 spaetz: but then you need to rebuild gtk+ again because of http://bugs.openembedded.org/show_bug.cgi?id=5405 Mar 05 13:25:49 s/gtk+/gtk+-native/ Mar 05 13:25:50 JaMa meant: spaetz: but then you need to rebuild gtk+-native again because of http://bugs.openembedded.org/show_bug.cgi?id=5405 Mar 05 13:26:32 sigh, ok Mar 05 13:27:13 spaetz: I hope this gets resolved soon properly in oe.dev Mar 05 13:27:29 spaetz: because we can fix it now.. but problem with gdk-pixbus-csource-native would stay Mar 05 13:38:04 spaetz: now you break gtk+-native as well as let gtk+ broken... Mar 05 13:38:21 I do? Mar 05 13:38:28 14:25:16 < JaMa> spaetz: you can disable demos (remove _virtclass_native from SRC_URI in last gtk+ recipe) Mar 05 13:38:34 only _virtclass_native Mar 05 13:39:23 this is what I have done in shr/testing: http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?h=shr/testing2010&id=9affebf996c30f641ce6f01d2dc7cc75dc38c7f4 Mar 05 13:39:24 I want to compile a C program on my freerunner with shr but get the following message: cannot find /usr/lib/libc_nonshared.a Any one knows how to solve this? Mar 05 13:40:19 spaetz: no it's not :) I mean sed -i 's/_virtclass_native//g' gtk+.inc Mar 05 13:41:02 ohh Mar 05 13:41:18 spaetz: and better to do this without PR bump Mar 05 13:41:36 spaetz: because if it's compile issue than nobody was able to build it with old PR Mar 05 13:42:04 spaetz: and now with next PR bump in oe.dev (when it's fixed properly) you'll have that PR already used for this "fix" Mar 05 13:42:23 JaMa: true, I usually only bump by 1/2 Mar 05 13:42:28 spaetz: so if you're going to fix it, please revert PR back to ${INC_PR}.3 Mar 05 13:42:45 But I am not going to merge anymore anyway, but just cherry pick things :) Mar 05 13:42:52 and least for a while. Mar 05 13:43:04 will do Mar 05 13:43:05 lets see how long "a while" can be :) Mar 05 13:43:29 just checking if gtk+ compile really breaks. Mar 05 13:44:07 I know it's damn hard to keep track of old "right" git commits with all their dependencies to keep them "right" Mar 05 13:44:18 s/old/all/g Mar 05 13:44:19 JaMa meant: I know it's damn hard to keep track of all "right" git commits with all their dependencies to keep them "right" Mar 05 13:44:19 it is Mar 05 13:45:11 spaetz: once I proposed here to merge all shr-u in some semi-regular basis (like a month window) and keeping track of all new issues in shr-u between which need to be fixed before next merge Mar 05 13:46:06 spaetz: this seems like easier way to maintain shr-t, but you're the boss here.. so it's just an idea Mar 05 13:48:58 JaMa: this is basically what I planned (and what I did between testing2009 and testing2010). I am just annoyed that I merge because of a tiny fix and a whole bunch of new problems comes in at the same time. Mar 05 13:49:15 I just want to get this thing to compile to test usability issues :) Mar 05 13:49:40 hopefully pushed the right fix now. Mar 05 13:49:56 seems right, thanks Mar 05 13:50:20 JaMa, spaetz: can you enter utf8? Mar 05 13:50:25 I've fixed it manually in shr-u on buildhost and still expecting oe.dev fix Mar 05 13:50:38 TAsn: I cannot :/ Mar 05 13:51:19 TAsn: I'll test on angstrom/spitz today to cut out eglibc/images changes Mar 05 13:51:24 and spaetz ? Mar 05 13:51:27 okie Mar 05 13:51:27 Umlaut stuff worked fine. At least receiving Mar 05 13:51:33 You mean typing? Mar 05 13:51:53 testing Mar 05 13:54:15 JaMa: now that you mention it, I don't have a non-english keyboard installed :) Mar 05 13:56:17 spaetz, just use the numeric keyboard Mar 05 13:57:13 TAsn: your alt-numberic keyboard wasn't shown in illume1 for me, no idea why Mar 05 13:57:25 TAsn: but maybe it was some fault of my local build.. Mar 05 13:57:54 TAsn: so to test it I installed default-numeric and tried with it Mar 05 14:00:07 JaMa: compiling gtk+ still fails even with latest Mar 05 14:00:16 I removed gtk+-native before Mar 05 14:01:08 did you -c clean it before? Mar 05 14:01:27 err, DOH Mar 05 14:01:31 only -native :) Mar 05 14:01:34 -c clean for gtk+ should be enough, as -native works OK Mar 05 14:01:43 ok, thanks Mar 05 14:02:02 spaetz: btw what happend with shr-t build on shr buildhost? Mar 05 14:02:14 spaetz: seen Keyboard interrupt and then eglibc build fail Mar 05 14:02:21 I aborted it, pulled in the latest fixes Mar 05 14:02:24 and restarted it now Mar 05 14:03:43 ok good Mar 05 14:38:10 hmm.. what still needs to happen to make 2.6.32 ready for general use? (I think it'd good to ask this question regularly) Mar 05 14:38:19 still modem race conditions? ..anything else? Mar 05 14:40:17 Weiss: does suspend/resume work now? Mar 05 14:40:28 sorry, I've been missing out a lot lately Mar 05 14:41:01 with KMS, there are no more WSoDs Mar 05 14:41:13 that is great to hear. Good job Mar 05 14:41:35 (there is a little more to check/fix to banish them FOREVER, but so far I haven't heard of any WSoDs at all, even from people who found 2.6.29+KMS didn't work when it worked for me) Mar 05 14:42:12 +1 for switching :) Mar 05 14:42:25 no xrandr yet.. working on that now Mar 05 14:51:06 JaMa: bb complains about fsousaged using "legacy staging". Is that because it inherits "autotools_stage"? Mar 05 14:51:22 otherwise, I can't see anything that could cause that, but I know very little about this. Mar 05 14:51:58 spaetz: no probably do_stage() from vala.bbclass Mar 05 14:55:31 ahh, that contains a do_stage()... Mar 05 14:59:43 freesmartphone.org: 03mickey 07msmcomm * r32de84daf4b2 10/msmvterm/src/ (commands.vala terminal.vala): msmvterm: improve help output Mar 05 14:59:44 freesmartphone.org: 03mickey 07msmcomm * ra253bf6a992d 10/ (3 files in 2 dirs): convert first bunch of accessors to properties Mar 05 14:59:48 freesmartphone.org: 03mickey 07msmcomm * r22520d0bc4a4 10/ (MAINTAINERS README): add README Mar 05 15:23:22 DocScrutinizer: the canopen specs're so unexciting, that i literally fall asleep, yay, it won. Mar 05 15:24:04 hehehe Mar 05 15:24:27 oh well, Friday afternoon Mar 05 15:24:36 WEEKEND! Mar 05 15:41:51 spaetz: you still have DISTRO_TYPE = "debug" in auto.conf which overwrittes value set in shr.conf (I've already removed that line from shr-u auto.conf) Mar 05 15:42:25 spaetz: ah Makefile also sets it to debug by default :/ Mar 05 15:46:28 let me fix that :P Mar 05 15:46:33 what is the correct value? Mar 05 15:46:41 or just omit it? Mar 05 15:47:18 omit Mar 05 15:47:27 ok Mar 05 15:47:55 it's set to release in shr.conf, but it doesn't make big difference in resulting image :) IIRC task-cli-tools-debug included in image by default Mar 05 15:48:47 and quiet on kernel cmdline Mar 05 15:49:01 SHR: 03mok 07shr-makefile * re90dda129ffc 10/Makefile: Makefile: do not set DISTRO_TYPE in auto.conf Mar 05 15:49:37 JaMa: you have to do a make update-common on the buildhost for users freshly getting it Mar 05 15:59:17 JaMa, I'm away from my laptop. can you puushh to testing? Mar 05 15:59:32 otherwise, i do it tomorrow Mar 05 16:00:05 ahh, on buildhost you mean, rightt. Mar 05 16:02:31 spaetz: yeah, just to update the downloadable Makefile Mar 05 16:02:42 for your auto.conf in testing you have to do it manually Mar 05 16:02:56 ok Mar 05 16:18:28 I'm so tired :) Mar 05 16:22:14 hi Mar 05 16:22:28 mrmoku: have you been able to fix the shr-u regressions? Mar 05 16:24:44 bumbl: I pushed the fix... and probably JaMa|GoNe built it Mar 05 16:24:49 well he started a build from scratch Mar 05 16:24:54 don't know how the status is Mar 05 16:25:09 ok Mar 05 16:29:16 spaetz: do you know if shr-u build finished? Mar 05 16:30:50 mrmoku: no opkg upgrade available Mar 05 16:31:24 bumbl: NOTE: Running task 2407 of 7231 Mar 05 16:31:28 ok Mar 05 16:31:31 :/ Mar 05 17:06:03 mrmoku: no clue Mar 05 17:06:19 but it should really. My local build of shr-t finished Mar 05 17:06:27 and that was merged a few hours ago Mar 05 17:06:57 but he started a build from scratch a few h ago Mar 05 17:07:06 so I would expect nothing today Mar 05 17:43:12 freesmartphone.org: 03morphis 07msmcomm * rc6f89ff87841 10/libmsmcomm/src/ (11 files): libmsmcomm: refactor message descriptor handling so we can calc the size of every message/response/event Mar 05 17:43:13 freesmartphone.org: 03morphis 07msmcomm * r2ef7d5bf4130 10/TODO: msmcomm: Update TODO list Mar 05 17:43:14 freesmartphone.org: 03morphis 07msmcomm * r6b74023cebbd 10/ (5 files in 3 dirs): Merge branch 'master' of git.freesmartphone.org:msmcomm Mar 05 17:46:43 freesmartphone.org: 03morphis 07msmcomm * r61f1dc77206e 10/libmsmcomm/src/internal.h: libmsmcomm: remove unused macro (it's ugly by the way ...) Mar 05 17:47:59 heyho Mar 05 18:12:38 mrmoku: thanks for Makefile Mar 05 18:33:56 JaMa: well... it was not *that* much work :P Mar 05 19:09:21 spaetz: please play with the new setting; i doubt that it helps, but lets see Mar 05 19:09:26 (in libgsm0710mux) Mar 05 19:15:00 freesmartphone.org: 03mickey 07cornucopia * r157b8f309438 10/docs/ (7 files): Mar 05 19:15:00 freesmartphone.org: docs: remove sample configurations, since we ship default configurations now Mar 05 19:15:00 freesmartphone.org: TODO++ Mar 05 19:44:12 TAsn: hello, i have a problem building shr-unstable "make setup" fails ! Mar 05 19:45:38 fatal: git checkout: branch master already exists Mar 05 19:46:19 fatal: git checkout: branch master already exists Mar 05 19:46:51 TAsn: fatal: git checkout: branch master already exists Mar 05 19:48:37 nschle85: calling make setup in an empty directory? Mar 05 19:49:52 not empty a helper script is the to wget the makefile, the rest of old sandbox is purged Mar 05 19:50:07 is there to wget Mar 05 19:50:37 mrmoku: sorry my apple keyboard is too slow for my fingers Mar 05 19:51:25 mrmoku: are you interested in console output ? Mar 05 19:52:01 nschle85: yeah, pastebin it please Mar 05 19:52:32 mrmoku: do i need an account for that ? Mar 05 19:54:07 nschle85: no, just go to shr.pastebin.com and paste it Mar 05 19:54:18 (or any other pastebin service for that) Mar 05 19:55:53 mrmoku: i will create a ticket on shr Mar 05 19:56:05 mrmoku: mom Mar 05 19:56:56 mickeyl, weirdly enough, with the latest revision it worked 9 out of 10 times, although I used the default of 5 seconds. Mar 05 19:57:35 I will try out some settings, yes. Mar 05 19:57:52 k Mar 05 19:59:05 mrmoku: I guess it's because master branch is already created with checkout Mar 05 19:59:30 mrmoku: so when it wants to checkout master branch to master it fails Mar 05 20:05:57 mrmoku: http://www.shr-project.org/trac/ticket/987 Mar 05 20:06:10 JaMa: http://www.shr-project.org/trac/ticket/987 Mar 05 20:08:54 JaMa: oh, it is the setup of bitbake that fails Mar 05 20:09:13 JaMa: have you manually changed the Makefile on the buildhost? Mar 05 20:10:22 mrmoku: yep.. sorry reverted to 1.8 there.. Mar 05 20:10:31 nschle85: try again Mar 05 20:11:03 mrmoku: checkouting right version in already created setup is enough.. so I put 1.8 back to Makefile Mar 05 20:11:20 JaMa: though that is still a bug in the Makefile Mar 05 20:11:42 NOTE: Running task 3940 of 7231, seems like we can expect lite image in the morning :) Mar 05 20:11:43 hi all! Mar 05 20:11:56 it's me or opkg is broken on last shr-u ? Mar 05 20:12:08 mrmoku: yup, you have some logic for not-checkouting already checkouted branch Mar 05 20:12:13 JaMa: now it works Mar 05 20:12:19 johnnymnemonic: did you read my email? Mar 05 20:12:29 JaMa: where? Mar 05 20:12:35 it's completely broken without Mar 05 20:12:49 partialy broken with my libopkg + reinstall from feeds Mar 05 20:13:02 and there is even newer revision in repo (not in feed) Mar 05 20:13:07 JaMa: why didnt you check your changes ? Mar 05 20:13:12 so I have no idea in which state you're Mar 05 20:13:24 nschle85: which changes? Mar 05 20:14:14 JaMa: the change in Makefile Mar 05 20:14:57 nschle85: it happen only when setting new environment.. so I was checking more important things in the morning while trying to start new rebuild asap (sorry for that) Mar 05 20:15:34 JaMa: ok Mar 05 20:17:42 JaMa: i would like to support shr, i am interested in setting up a build server which always creates clean builds Mar 05 20:18:57 nschle85: good, thanks Mar 05 20:19:43 JaMa: i dont know howto setup tinderbox so i would like to use hudson Mar 05 20:20:12 JaMa: but there is a problem to provide the logfiles for failed buils Mar 05 20:20:23 mrmoku: are you checking that Makefile bug or should I? before we forget about it when we'll switch all to master for real? Mar 05 20:20:50 nschle85: why don't you report to OE tinderbox? Mar 05 20:22:20 JaMa: howto report ? i thought tinderbox is a buildserver. Mar 05 20:22:44 why don't you check google first? :) Mar 05 20:23:10 it's really easy to setup (2 lines in config) Mar 05 20:24:27 nschle85: can you please retry modified Makefile? Mar 05 20:25:06 JaMa: ups shy Mar 05 20:25:28 JaMa: yes i retry Mar 05 20:25:51 JaMa: if you have time... I can't right now :/ Mar 05 20:26:04 JaMa: know what... I'll give you access to the makefile repo :P Mar 05 20:26:45 JaMa: done Mar 05 20:27:28 :) Mar 05 20:27:29 done Mar 05 20:27:43 will push as soon as nschle85 confirms Mar 05 20:28:58 walk 3841c2a3cda1936ce5a539938b469422ab0231cb Mar 05 20:28:58 /bin/sh: -c: line 4: syntax error: unexpected end of file Mar 05 20:28:58 make: *** [setup-bitbake] Error 2 Mar 05 20:28:58 make: Leaving directory `/home/nschle85/shr-build' Mar 05 20:29:36 JaMa: the version before was better :-) Mar 05 20:30:28 nschle85: try again Mar 05 20:33:11 JaMa: better but "make setup" is still running, i inform you in a few minutes about the result Mar 05 20:34:33 SHR: 03shr 07shr-makefile * rf685f849d91e 10/Makefile: Makefile: checkout bitbake branch only it it doesn't exist Mar 05 20:34:43 SHR: 03shr 07shr-makefile * r95df45e4ff3f 10/conf/site.conf: site.conf: remove SRCPV as it's already in bitbake.conf Mar 05 20:35:16 who is shr Mar 05 20:35:17 ? Mar 05 20:35:17 :P Mar 05 20:35:28 :) Mar 05 20:35:34 :P Mar 05 20:35:44 TAsn: its the always not compiling thing :-) Mar 05 20:35:57 nschle85, oh, I know him (it) :P Mar 05 20:36:10 It always compiles fine! Mar 05 20:36:22 sometimes it just needs few patches.. Mar 05 20:36:36 JaMa: no ! thats the reason i deleted my whole sandbox Mar 05 20:36:40 but that's our purpose.. without it our lives will be meaning less Mar 05 20:37:13 * JaMa off Mar 05 20:38:03 JaMa: "make setup" finished ok :-) Mar 05 20:41:15 JaMa: Ticket #987 (nice number) set to fixed Mar 05 20:41:30 huh, GSM seems to be somehow not reliable (SHR unstable) Mar 05 20:42:47 <[Rui]> hi Mar 05 20:42:53 I was sent a few SMS, but I wasn't aware of it until rebooting Mar 05 20:43:28 furthermore, a SMS I've sent did not reach its destination, it seems Mar 05 20:44:06 TAsn: who /what is SHR: ? Mar 05 20:44:17 what do you mean? Mar 05 20:44:50 TAsn: is it an automatic information about git commits ? Mar 05 20:44:55 yes Mar 05 20:45:03 you are talking about CIA-45 I assume Mar 05 20:45:06 and yes, it is. Mar 05 20:45:27 TAsn: yes and what is CIA-45 ? Mar 05 20:45:37 cia is a commit bot Mar 05 20:45:47 exactly what we just said Mar 05 20:46:16 TAsn: wow its so nice waht compuerts can do :-) Mar 05 20:46:31 :) Mar 05 20:46:46 It's even more amazing what humans can do Mar 05 20:47:01 TAsn: updated my dedicated server, hope it will reboot :-) Mar 05 20:47:06 :) Mar 05 20:49:06 TAsn: i have a general question Mar 05 20:49:23 I hope I'll be able to provide a specific answer Mar 05 20:49:25 ask ahead. Mar 05 20:49:25 :P Mar 05 20:50:32 TAsn: i purge very often my shr sandbox because it does not compile after an update, how can i imagine what to do to repair the sandbox ? Mar 05 20:50:57 TAsn: today my problem was : sanity check of gcc fails Mar 05 20:51:00 nschle85, to tell you the truth Mar 05 20:51:04 here it works Mar 05 20:51:07 even after updates Mar 05 20:51:11 maybe you are doing something wrong. Mar 05 20:51:37 nschle85, anyhow, JaMa/mrmoku know more than me about this topic, I'm not an OE expert. Mar 05 20:51:38 iam doing the following steps: make update && make image Mar 05 20:51:48 I never do a make image Mar 05 20:51:50 but I also do Mar 05 20:51:56 make update && bitbake whatever Mar 05 20:52:41 TAsn: and how do you install ? Mar 05 20:53:08 scp packagename.ipk Mar 05 20:53:16 I didn't even build myself a local repository Mar 05 20:53:25 as I rarely care about self built packages Mar 05 20:53:34 Whatever I develop I test on my box Mar 05 20:53:43 and only after X time of testing Mar 05 20:53:51 I build a package and test it on the device Mar 05 20:55:05 TAsn: how do you test gsm or dialer without device ? Mar 05 20:55:22 TAsn: what is you test environment ? Mar 05 20:55:58 I don't test gsm/dialer Mar 05 20:56:05 I just run frameworkd with trimmed down features Mar 05 20:56:07 phoneuid Mar 05 20:56:12 and mdbus Mar 05 20:56:15 this is my test env Mar 05 20:56:19 which is all I need. Mar 05 20:57:18 TAsn: ok you compile native and test local ? Mar 05 20:57:39 yes Mar 05 20:57:45 and only after proper testing Mar 05 20:57:49 I cross compile Mar 05 20:58:07 And when I do UI design, I just do relative design Mar 05 20:58:18 so when scaled down, everything looks the same Mar 05 20:58:38 so I rarely need anything Mar 05 21:01:05 TAsn: so i understand, why nobody has the current illume-keyboard problem like me :-( Mar 05 21:01:16 what issue do you have? Mar 05 21:02:23 Btw, we don't tend to fix issues we don't know of :P Mar 05 21:02:29 TAsn: mom Mar 05 21:02:34 mom? Mar 05 21:02:54 TAsn: mom its the short form of jut a moment please Mar 05 21:03:00 just Mar 05 21:03:11 (in german) Mar 05 21:03:12 if you say :P Mar 05 21:03:16 oh, in german Mar 05 21:03:17 okie. Mar 05 21:03:48 shortening of "moment" Mar 05 21:03:54 (so could be english too" Mar 05 21:04:45 Yeah, but it's not Mar 05 21:04:52 have heard of cases where someone said "yes, mom", meaning "yes, one moment please"... which was misunderstood as "yes, mommy"... Mar 05 21:04:53 (doesn't exist in english) Mar 05 21:05:13 Wonka, I thought he was calling me mom :P Mar 05 21:05:30 the misunderstanding one was quite pissed of... Mar 05 21:05:39 :) Mar 05 21:08:07 mompl Mar 05 21:08:37 TAsn: could not find ticket Mar 05 21:08:46 TAsn: i ll explain Mar 05 21:09:03 DocScrutinizer, this one makes sense. Mar 05 21:09:32 or momplz Mar 05 21:10:49 TAsn: 1.open terminal 2. try to type some text, the letter grows but a line feed is sent to the console, multiple cklicks on a letter may lead in typing it in the console Mar 05 21:11:06 self built image? Mar 05 21:11:12 SHR: 03mok 07libphone-ui-shr * rbbbe468394b2 10/src/phoneui-phone-log.c: phoneui-phone-log: remove obsolete phoneui_backend_phone_log_new Mar 05 21:11:14 SHR: 03mok 07libphone-ui-shr * r4a9a7098c442 10/src/view/phone-log-view.c: phone-log-view: set fixed limit of 25 calls... needs to be configurable though Mar 05 21:11:15 SHR: 03mok 07libphone-ui-shr * rbbcfaeab4826 10/src/view/ (phone-log-view.c phone-log-view.h): phone-log-view: register for call and contact changes Mar 05 21:11:16 SHR: 03mok 07libphone-ui-shr * r2e596f2ffbd5 10/src/view/phone-log-view.c: phone-log-view: activate missed calls toolbar item on init and hide Mar 05 21:11:17 SHR: 03mok 07libphone-ui-shr * r11991a258d75 10/src/view/phone-log-view.c: phone-log-view: actually add call and contact changed handlers Mar 05 21:11:22 mrmoku, nice spam :P Mar 05 21:11:22 SHR: 03mok 07libphone-ui-shr * r6a91d620ac2d 10/data/quick-settings.edc: Merge remote branch 'origin/contacts' into contacts Mar 05 21:11:23 SHR: 03mok 07libphone-ui-shr * r152e4e2becea 10/src/view/phone-log-view.c: phone-log-view: use common_utils_timestamp_to date Mar 05 21:11:24 SHR: 03mok 07libphone-ui-shr * rdb151d5ea0b7 10/data/phonelog.edc: phonelog.edc: add the parts for date and duration to labels Mar 05 21:11:43 mrmokospamattack Mar 05 21:11:45 TAsn: don't get confused by the quick-settings thing in there :P Mar 05 21:11:54 TAsn: yep from scratch (multiple times, error since "several" days/weeks Mar 05 21:12:11 mrmoku, no changes, only a merge. Mar 05 21:12:20 nschle85, no idea then. :| Mar 05 21:12:39 nschle85, There isn't really a way we can debug self built images Mar 05 21:12:57 makes me realize I need certain exceptions when installing the general spammer kickbot Mar 05 21:13:15 TAsn: build one and you will see Mar 05 21:13:28 that's not a sane way :P Mar 05 21:13:41 My pc sucks, I can't afford building images all day long Mar 05 21:13:46 but we do build them on the build host Mar 05 21:13:49 and there it works Mar 05 21:14:00 (and on people's computers, where it work as well) Mar 05 21:14:17 spaetz: hey mate! How are things going with the new testing release? Mar 05 21:14:22 TAsn: does the build host e clean build ? Mar 05 21:14:25 for proper testing I need a lot of time and full access to your box, both are not available Mar 05 21:14:42 nschle85, yes, and once the ram issue is fixed, jama will do another one Mar 05 21:16:00 TAsn: is it possible to sent you my image ? I would like to know that this error happens also on other phones. Mar 05 21:16:19 I bet it will happen on my phone as well Mar 05 21:16:32 and sorry, no time to start reflashing and testing :| Mar 05 21:16:33 but i did noch change any source Mar 05 21:16:47 only build Mar 05 21:16:55 nschle85, I bet you didn't but sometimes things go wrong in OE :P Mar 05 21:17:27 sorry but i think nobody cares about this problem Mar 05 21:18:04 [2010-03-05 19:14:57] the GSM problem is usualy resolved with loglevel DEBUG on ogsmd Mar 05 21:18:08 :-S Mar 05 21:18:42 nschle85, kind of Mar 05 21:19:07 TAsn: kind of ? (what does that mean ?) Mar 05 21:19:37 DocScrutinizer: The issue I am having (to which that was a reply) is GSM not working on boot Mar 05 21:19:52 nschle85, To tell you the truth, images and buildhost are not my field of interest/responsibility and it takes too much time to investigate such a rare case (I know notihng about) so at least I don't care enough about it :P Mar 05 21:20:00 DocScrutinizer: I am able to start it if I kill a few things and start manually, slowly. Mar 05 21:20:00 try your luck with people who know more about OE Mar 05 21:20:16 DocScrutinizer: I believe I read on SHR trac that it's an FSO issue of some kind. Mar 05 21:20:22 krayon: we had that topic 10h ago Mar 05 21:21:01 just encouraging some devels to fix that issue for good Mar 05 21:21:11 DocScrutinizer: I see, I wasn't here 10 hours ago :P Mar 05 21:21:15 TAsn: ok, but i think that the current shr image (build from official sources) does not work Mar 05 21:21:32 DocScrutinizer: So is it just due to startup being too fast or something? Mar 05 21:21:50 TAsn: who is responsible for such problems ? Mar 05 21:21:55 nschle85, perhaps, but I seriously doubt it. Mar 05 21:22:04 nschle85, I think JaMa, but don't tell him I said his name :P Mar 05 21:22:10 it seems it's a race in a obsolete piece of code that needs fixing urgently Mar 05 21:22:29 nschle85: TAsn: I have not, as yet, been able to do much with it either :P I can't even get it started :/ Mar 05 21:22:33 JaMa = martin jansa ? Mar 05 21:22:39 yes. Mar 05 21:22:41 TAsn: How did you setup your build env btw? Mar 05 21:22:49 make install Mar 05 21:22:58 iirc Mar 05 21:23:25 TAsn: Did you follow any of this: http://trac.shr-project.org/trac/wiki/Building%20SHR Mar 05 21:23:31 yes Mar 05 21:23:40 oops Mar 05 21:23:42 make setup Mar 05 21:23:45 not make install Mar 05 21:24:01 krayon: what does not start ? Mar 05 21:24:02 though this page looks a bit outdated Mar 05 21:24:41 nschle85: When I try to do the make image Mar 05 21:24:58 er, WHAT? mc_4.6.2.bb fails to build, only line in log is "rm: cannot remove `config/config.guess': No such file or directory"... *facepalm* Mar 05 21:25:11 nschle85: It reports that nothing PROVIDES shr-image-lite (I also tried shr-lite-image) Mar 05 21:25:21 LOOOOOL Mar 05 21:25:43 I get the same error when I try doing: . ./setup-env && bitbake -c clean openmoko-dialer3 etc Mar 05 21:26:01 krayon: try http://www.shr-project.org/trac/wiki/Building%20SHR Mar 05 21:26:20 nschle85: Yes, they are the steps as per that page Mar 05 21:26:42 the Makefile had have a problem today Mar 05 21:26:51 get it again and retry Mar 05 21:27:17 nschle85: I've tried several times last week, and again about 5 minutes ago :P Mar 05 21:27:29 krayon, openmoko-dialer3 doesn't exist Mar 05 21:27:35 nschle85: I just rm -Rf 'd and am make setup again :'( Mar 05 21:27:38 i do "make setup" Mar 05 21:27:39 as I said, this page is outdated Mar 05 21:28:08 TAsn: Do you know of any more up to date documentation? Mar 05 21:28:24 krayon, nope, but it's a pretty simple topic Mar 05 21:28:25 krayon: then make image Mar 05 21:28:26 make setup Mar 05 21:28:32 make setup && have fun Mar 05 21:28:42 What's wrong with that page? There are some examples on how to do it Mar 05 21:29:02 Heinervdm: It doesn't work for me. Didn't last week anyway. Mar 05 21:29:11 bitbake -c clean openmoko-dialer3 Mar 05 21:29:30 krayon: lets try it together Mar 05 21:30:17 I'm just tarring the results of make setup to save me having to d/l it all again. I feel sorry for the servers :S Mar 05 21:30:23 TAsn: krayon: That section is an example on how to make changes to a random package Mar 05 21:30:38 Heinervdm, a random outdated package :P Mar 05 21:30:41 if you want to change a package you usally know it's name Mar 05 21:30:44 Heinervdm: I realise that but I only tried that after the other stuff failed :P Mar 05 21:30:46 anyhow, ok, so maybe taht' still relevant Mar 05 21:31:09 btw nschle85, you can try your luck with Heinervdm, he's also an OE expert. Mar 05 21:31:10 krayon: i opened a channel for you Mar 05 21:31:11 Just copying that text is sensless Mar 05 21:31:11 TAsn: My main objective is creating a dev env and the dev page says do all this stuff first. Mar 05 21:31:21 krayon, if you want a dev env Mar 05 21:31:30 *dev env for what? Mar 05 21:31:34 that page is for a build env Mar 05 21:31:40 dev env is just linux :P Mar 05 21:31:43 anyhow, sorry, gtg. Mar 05 21:31:46 ciao. Mar 05 21:32:32 krayon: please invite me in a separate channel so we can setup your environment Mar 05 21:34:15 Heinervdm: the page is ok, i dont know the problem of krayon: but we should help him Mar 05 21:34:29 TAsn: I should say a build env I guess :P Mar 05 21:35:20 nschle85: i think that he just has forgotten to change to the right directory Mar 05 21:35:41 Heinervdm: Hahah, nothing that bad :P Mar 05 21:36:14 Heinervdm: Incidentally the process seemed to work in the shr-testing dir, just not unstable :S Mar 05 21:36:15 krayon: you said, that it can't find shr-lite-image Mar 05 21:36:43 that should be impossible Mar 05 21:36:47 i think he did not setup the environment iam asking him .... Mar 05 21:37:29 krayon: so what did you do? Mar 05 21:41:27 Heinervdm: All I can think is that my initial checkout of that directory didn't finish/succeed but I don't recall any errors after the make setup. I'm starting from scratch now anyway :/ Mar 05 21:42:16 krayon: ok, if the inital checkout fails, because of some git server error, then everything is possible Mar 05 21:42:55 and sth like that can always occoure because no connection is 100% stable Mar 05 21:45:59 Heinervdm: krayon changed the local.conf he should work with a vanilla version Mar 05 21:46:57 nschle85: i think, if he starts from scratch now, it will work. I will go to bed now :) Mar 05 21:47:11 Heinervdm: nschle85: So you cannot set IMAGE_FSTYPES etc in that file? Mar 05 21:47:40 krayon: you can do it Mar 05 21:47:43 They were there, and they were what I wanted so I set them, I thought thta was reasonable as the wiki page said I can :P Mar 05 21:48:06 but you should know what those variables do Mar 05 21:49:25 I am going off the comments in the file. Mar 05 21:56:36 <[Rui]> elmdentica should have some speedups now it depends less on glib (I think only does for the nicely convenient regex functions, right now). Mar 05 21:59:46 <[Rui]> yup, launch is now a tiny bit faster. Mar 05 22:04:43 [Rui]: coool Mar 05 22:06:51 <[Rui]> And I'm now adding more timelines beyond the friends, one... Mar 05 22:16:29 <}mat> hi, where do i find a package for shr-u which provides libc6 (>= 2.11)? opkg only finds libc6 (2.10-r8.6.4) and therefore can't statisfy various dependencies Mar 05 22:18:13 }mat: SHR-u version is 2.11-r9.1.4 Mar 05 22:18:55 <}mat> yeah but I am in the process of updating an older shr-u and it doesn't find it Mar 05 22:19:50 <}mat> i saw that there is an eglibc-pic but that also doesn't statisfy the dependencies Mar 05 22:21:10 <[Rui]> I don't know if it's already ok to do upgrades, right now... Mar 05 22:21:43 }mat: install package manually perhaps ? Mar 05 22:21:46 <[Rui]> there was some hw problem, IIRC, at the buildhost and somestuff was really borked. Intervention was expected for next week, again, IIRC. Mar 05 22:21:51 opkg install http://build.shr-project.org/shr-unstable/ipk/armv4t/libc6_2.11-r9.1.4_armv4t.ipk Mar 05 22:24:33 <}mat> valos: seems to work maybe I have not all feeds configured, will have to check, thanks Mar 05 22:26:20 }mat: [Rui] is right, I think repository is broken and Packages.gz is false Mar 05 22:47:56 SHR: 03mok 07libphone-ui-shr * r9a8d1247ab2e 10/src/view/message-new-view.c: message-new-view: try to resolve the phone number passed in via options Mar 05 22:48:06 SHR: 03mok 07libphone-ui-shr * r714f1af75489 10/src/view/message-show-view.c: message-show-view: create a correct hashtable when answering Mar 05 22:48:25 TAsn: fixed :) Mar 05 22:48:42 still want to do some small things tomorrow Mar 05 22:48:49 then we can merge :D Mar 05 22:48:51 gnight all Mar 05 22:49:43 mrmoku|away, thanks, night. :P Mar 05 23:40:39 <[Rui]> Wohoo! Loading public_timeline works in elmdentica! **** ENDING LOGGING AT Sat Mar 06 02:59:57 2010