**** BEGIN LOGGING AT Thu May 12 23:59:56 2005 May 13 00:49:48 morning May 13 01:14:27 Hullo May 13 01:15:33 Say that I played with my openembedded/classes directory a bit too much -- how can I revert to latest BK? May 13 01:16:22 Did you bk commit the changes? May 13 01:16:39 no.. May 13 01:17:08 Did you bk ci them, or bk edit the files? May 13 01:17:32 no, just say the the entire classes directory is gone.. :) May 13 01:18:23 morning all May 13 01:18:25 That's more difficult, (if you used bk it's easy). Do (did) you have any local bk commits in there? May 13 01:18:38 I'm looking for 'svn update' without going through the BKmanual :) May 13 01:19:06 no, none.. May 13 01:19:11 good morning folks May 13 01:19:15 rm -f classes/*; bk co classes May 13 01:19:16 heh May 13 01:19:21 hah! May 13 01:19:31 let's try that May 13 01:19:32 note: NOT rm -rf May 13 01:19:40 you dont want to remove bk's metadata (SCCS) May 13 01:19:53 Yes, that might be wrobbie's problem... May 13 01:20:13 ... "say the entire classes directory is gone" May 13 01:20:45 that did the trick.. May 13 01:20:53 if you screw up bk's file metadata / history, you'll have to either reclone or use bk repair, which essentially pulls down all the changesets anywya May 13 01:21:03 jbowler: sorry for being confusing May 13 01:21:39 no problem, I did the metadata destruction thing once on a directory with local commits, that was bad. May 13 01:23:02 kergoth: Koen told me to ask you -- what makes my setup pick glibc_2.3.2+cvs20040726 when newer ones are available? May 13 01:23:57 I can't fine PREFERRED_VERSION or something that specifies that particular one May 13 01:24:51 wasn't cvs20040726that hardcoded some time ago in some class? May 13 01:25:23 See the settings of "DEFAULT_PREFERENCE" in packages/glibc/*.bb May 13 01:25:45 wrobbie: the DEFAULT_PREFERENCE inside of hte .bb's defines the deafult preferences. if you want to override that, then set your PREFERRED_VERSION for that package. May 13 01:26:26 but it's only things like DEFAULT_PREFERENCE_i686 = "0".. what does that mean? May 13 01:26:39 I should have known that is was done a bit smarter..:-) May 13 01:26:40 it means that the dfault preference is 0 for i686. May 13 01:26:52 zero what? :) May 13 01:27:01 its just a numeric. if one package has a default preference of 0, and one has a default preference of -1, the one with 0 is default. May 13 01:27:08 if it isnt set at all, its 0 May 13 01:27:26 in that package, its setting DEFAULT_PREERENCE to -1 by default, but overriding that back to 0 for certain packages for which its known to work. May 13 01:27:31 er May 13 01:27:33 certain archs May 13 01:27:34 heh May 13 01:31:08 kergoth: Have you had a chance to look over zecke's patches? May 13 01:31:26 was planning on it. didnt get to it today. i'll take a look tomorrow May 13 01:32:02 kergoth: ok, np May 13 01:37:21 kergoth: ok, thanks May 13 01:38:33 Morning! May 13 01:38:38 Is where a reason why nobody uses glibc > 2.3.3? (no .bb files for it at least) May 13 01:38:52 ? May 13 01:39:47 there is. there's the floating _cvs build that always goes from current cvs. just probably no really recent locked down dates fromc vs May 13 01:40:03 Question regarding BBMASK. BBMASK = "^gpe.*|^x.*" does not work. I want to mask all gpe and x files. What is wrong ? May 13 01:40:04 feel free to try current cvs and see how it is May 13 01:40:14 uv1: it matches against path+filename May 13 01:40:18 uv1: it odesnt _start_ with gpe May 13 01:40:27 ah, ok May 13 01:40:32 something like /(gpe|x).* May 13 01:40:36 would work more to your liking May 13 01:40:45 or what have you May 13 01:40:59 kergoth: Thanks, will try that ... May 13 01:41:26 course the .* is superfluous, since there's no need to match what comes after gpe or x May 13 01:41:30 heh May 13 01:42:57 I made a new bb file for FusionSound, which depends on DirectFB. But when making it can't find directfb headers.. What to do? Post .bb on mailing list? May 13 01:43:14 it use autoconf/automake/ May 13 01:43:16 ? May 13 01:43:20 configure stage seems to work May 13 01:43:28 no, configure is present May 13 01:43:32 uh? May 13 01:43:38 configure existing means absolutely nothing. May 13 01:43:48 i didnt ask if it _runs_ autoconf or automake May 13 01:43:54 i asked if it _used_ autoconf or automake May 13 01:44:08 is the configure script an autoconf generated configure script, or not? May 13 01:44:10 well, it means it doesn't need automake, autoconf, not? May 13 01:44:15 yes May 13 01:44:17 it is May 13 01:44:22 and you're inheriting autotools? May 13 01:44:44 yep, inherit autotools pkgconfig May 13 01:45:09 not much else i can do to diagnose from here. you need to ensure that the package's buildsystem is obeying our CFLAGS, LDFLAGS, etc May 13 01:45:18 those vars are already set up with the proper paths into staging to find our stuff May 13 01:45:23 just a matter of exerting control over their buildsystem May 13 01:45:59 hmm.. ok May 13 01:46:26 look for a -I that points into staging on the compile line that's failing May 13 01:47:04 if it isnt there, then you need to figure out how to get it to use our cflags. we export CFLAGS into the environment, which means its configure script should be picking it up and substituting it for @CFLAGS@. so you'd want to make sure their Makefile.in is using that May 13 01:47:08 or whatever May 13 01:47:09 * kergoth wanders off May 13 01:47:59 -I/usr/include/directfb, that can't be good.. May 13 01:48:49 wrobbie: okay, thats a seperate problem from it pointing into staging :) there's probably a configure option, i.e. --with-directfb-libraries= or what have you. you'll want to set those to the correct places in staging. there are lots of examples in other .bb files May 13 01:49:01 run ./configure --help, take a look May 13 01:49:24 *yawn* May 13 01:49:25 ok, thanks May 13 01:49:28 hey mickeyl May 13 01:49:32 morning kergoth May 13 01:49:35 didn't move yet ? May 13 01:50:05 nope. moving company finally got off their collective ass and sent me the paperwork though, so should be happening soon May 13 01:50:13 :) May 13 01:50:15 ah finally ;) May 13 01:50:41 will you be disconnected for a while? May 13 01:53:11 shouldnt be for long. a couple days at most, for travel time. my dad has cable at my townhouse, so once i'm unpacked, will be connected again May 13 01:53:21 good May 13 01:54:27 03mickeyl 07 * r1.3281 10openembedded/packages/pcmciautils/pcmciautils_003.bb: add pcmciautils 003 May 13 01:54:30 03mickeyl 07 * r1.3280 10openembedded/conf/machine/zaurus-clamshell-2.6.conf: add infrared and ppp modules to c7x0 bootstrap image (kernel 2.6) May 13 01:54:32 03mickeyl 07 * r1.3279 10openembedded/BitKeeper/deleted/ (2 files): remove meta-grant and gpe-image-grant: please don't commit private .bb files May 13 01:54:35 03mickeyl 07 * r1.3278 10openembedded/packages/orinoco/ (2 files in 2 dirs): spectrum_cs: catch up with linux 2.6.12 May 13 01:54:45 hey CIA-9 no longer lags. May 13 01:54:52 nice May 13 01:55:05 has anyone gotten the pcmciautils stuff working yet? /me wants to see cardmgr dead and buried May 13 01:55:28 not yet, but since we have 2.6.12-rc on c7x0 we are slowly experimenting with it May 13 01:55:37 cool May 13 01:55:51 kergoth: it seems fusionsound's only way of locating/check for directfb is via pkg-config.. does that break CFLAGS etc? May 13 01:55:59 hello May 13 01:56:04 i'm just not sure yet how expandable this in-kernel solution is, e.g. to have a new card recognized we added a simple entry to some pcmcia conf file. How will we do that without cardmgr ? May 13 01:56:12 did anyone ever get around to splitting out the common hotplug scripts between both linux-hotplug and diethotplug into a seperate package? May 13 01:56:28 wrobbie: check if directfb installs its pc config file into staging May 13 01:56:35 wrobbie: if so, check if fusion picks it up May 13 01:56:40 wrobbie: no, OE can handle pkgconfig. any package that provides a .pc (like directfb) needs to inherit pkgconfig. May 13 01:56:51 kergoth: no i don't think anyone did that yet May 13 01:56:54 k May 13 01:57:10 looking forward to being productive again, and doing some coding on some of my much delayed projects May 13 01:57:13 heh May 13 01:57:20 that'd be fine :) May 13 01:57:32 blow the dust off of some devices May 13 01:57:48 how was the user reaction to the most recent oz release? May 13 01:57:52 * kergoth 's been pretty out of it May 13 01:58:05 quite good May 13 01:58:14 kergoth: pcmciautils won't compile... May 13 01:58:14 overall pretty positive. there are some issues with sharp-compat libs May 13 01:58:24 kergoth: I haven't had a chance to work out why... May 13 01:58:27 RP: try 003, which i just added May 13 01:58:42 mickeyl: right :) May 13 01:58:46 RP: can you answer my question to how expandable the in-kernel pcmcia approach is? May 13 01:58:58 i have reviewed the recent pcmcia patches May 13 01:59:09 and to me it looks like adding a new card needs a kernel patch now May 13 01:59:24 mickeyl: I think that would be correct May 13 01:59:37 mickeyl: There seem to be a lot of deviceid tables in the kernel May 13 01:59:37 i don't know if I like that. May 13 01:59:50 morning zecke May 13 01:59:56 hi zecke May 13 02:00:10 eh, doesnt the module just have to use the right macros to define its ids? doesnt pcmciautils provide a map tool not unlike the usb and pci stuff? May 13 02:00:25 * kergoth hasnt reviewed the code, but that was the impression he'd gotten from the ml traffic May 13 02:00:33 at least, back when i was reading my email May 13 02:00:37 god, i'm like weeks behind on mailing lists May 13 02:00:42 its gonna take ages to get caught up May 13 02:00:43 zecke: hello May 13 02:00:56 Each module seems to supply a list of ids it supports May 13 02:01:00 hey May 13 02:01:15 RP: did you've a chanche to test the fetcher.py changes? May 13 02:01:19 RP: right, that would make sense May 13 02:01:32 zecke: Not yet. I'll do that soon though May 13 02:01:49 considering the driver probably knows much much better than the user what cards it can handle May 13 02:01:57 heh May 13 02:02:52 hmm, it got late May 13 02:03:03 I've just had a surprisingly polite and detailed email from an OZ user asking for help with his SD cards. I think I'm in shock... May 13 02:03:13 hehe May 13 02:04:07 schocking May 13 02:04:21 kergoth: Your BBMASK suggestion works, thanks. What happens, if you have masked a package file, but it is needed due May 13 02:04:36 uv1: whatever package needed it will fail with an unsatisfied dep May 13 02:04:43 so be careful with it May 13 02:04:50 kergoth: Your BBMASK suggestion works, thanks. What happens, if you have masked a package file, but it is needed due to dependency reasons. Will it be found / builded ? May 13 02:05:02 BBMASK excludes things from even being parsed May 13 02:05:06 its compeltely outside bitbake's knowledge May 13 02:05:19 (the info in those files, that is) May 13 02:05:50 kergoth: OK, will keep that in mind. May 13 02:09:03 3 May 13 02:10:53 ppl please apply my opie-calculator patch... May 13 02:12:52 "connect to host oe-devel.bkbits.net port 22: Network is unreachable" May 13 02:13:02 Is anyone else getting that? May 13 02:14:03 can't try now May 13 02:15:18 RP: Just documenting the SysRq things we did yesterday. What does EmergencySync exactly do ? Are there other KeyCombinations possible ? I assume it works only with 2.6 kernel ? May 13 02:15:20 gints|wrk: looking at it May 13 02:15:36 uv1: read src/linux/Documentation/sysrq.txt May 13 02:16:02 uv1: EmergencySync saves any data in memory to the flash. If you don't do that you risk file system corruption May 13 02:16:04 gints|wrk: in future remember also about _1.2.0.bb May 13 02:16:19 hrw|work: ok May 13 02:16:25 uv1: It also works under the 2.4 kernel but the keypress is different May 13 02:16:45 (I don't know what it is) May 13 02:16:49 gints|wrk: ohhh, *sorry* May 13 02:17:01 gints|wrk: btw - can opie-calculator use converstion table from zuc? it will got more units to convert.. May 13 02:17:02 FN+HOME May 13 02:17:04 hi bluelightning May 13 02:17:13 zecke: that's life... :) May 13 02:17:26 hrw|work: should test ... May 13 02:17:38 I have tested with that from opie cvs.. May 13 02:17:42 Will check src/linux/Documentation/sysrq.txt. May 13 02:17:53 if zuc will work we can update opie cvs May 13 02:18:40 gints|wrk: I'm about to leave now May 13 02:19:06 gints|wrk: if no one comitted it until tonite, I'll commit it, together with the RTL text rendering May 13 02:19:40 Error: Max pending log exceeded, commit aborted. May 13 02:19:41 :( May 13 02:19:46 zecke: hrw I think will take care... May 13 02:20:15 If no I will remaind :) May 13 02:20:36 if I get connect to openlogging then I will commit it May 13 02:20:43 hrw: I have an mmc_timeout patch sitting locally. Will commit as and when I can connect to oe-devel.bk.... May 13 02:21:01 RP: ok May 13 02:21:11 RP: now I have 0xb0000 kernel May 13 02:21:28 hrw|work: it is true also if you will get connect after two days? :) May 13 02:21:44 gints|wrk: sure May 13 02:21:47 hrw|work: mean do you need remainder :) May 13 02:22:09 gints|wrk: if not today then monday May 13 02:22:38 ~lart openlogging.org May 13 02:22:38 * ibot drops a humongous exploding nuke on openlogging.org May 13 02:24:25 cya later May 13 02:25:25 03RP 07 * r1.3283 10openembedded/packages/linux/linux-openzaurus_2.6.12-rc3-mm2.bb: bump May 13 02:25:29 03RP 07 * r1.3282 10openembedded/packages/linux/linux-openzaurus_2.6.12-rc3-mm2.bb: May 13 02:25:29 Merge bk://oe-devel@oe-devel.bkbits.net/openembedded May 13 02:25:29 into tim.rpsys.net:/usr/oe/openembedded May 13 02:25:32 03RP 07 * r1.3266.1.1 10openembedded/packages/linux/ (2 files): linux-oz-2.6: Fix suspend/resume issue + add mmc timeout fix May 13 02:25:56 Conenction came back May 13 02:26:24 hrw|work: That should "fix" your SD cards, hopefully May 13 02:27:51 Hmmm. zecke's just gone - I was about to tell him his patch works :) May 13 02:29:42 ok May 13 02:32:23 hi lardman May 13 02:32:27 hi all May 13 02:32:30 hi gints|wrk May 13 02:32:30 hi hrw|work May 13 02:32:40 hey bluelightning May 13 02:37:55 so is bitbake going to branch and be played with before work diverts back to bitbake-ng ? May 13 02:38:44 a testing/unstable bitbake ? May 13 02:38:47 ade|desk: Work isn't diverting as such, some of us just don't understand enough about bitbake to work on bitbake-ng May 13 02:39:11 Its like some "unstable" versions will be released... May 13 02:39:41 no need to branch May 13 02:39:54 bitbake-ng is a different entity May 13 02:39:59 meanwhile we improve bitbake 1.x May 13 02:40:03 Its not as if bitbake has any other direction I guess :) May 13 02:40:12 some of the concepts will end up in -ng, some won't May 13 02:40:36 Was just trying the different sysRq's documented and struggled over a ugly issue. If switched to vt1 (e.g.) a "dim display" and "display off" command also affects the virtual console. So after approx 40 secs your screen is off even if entering data. Should'nt working on the console disable the apm system ? May 13 02:40:45 let's call it bitbake NT May 13 02:40:48 or bitbake XP May 13 02:41:02 hmm.. my push doesnot call cia trigger.. May 13 02:41:02 or XD for our spanish users May 13 02:41:03 uv1: Yes it should. However, this is not yet implemented in Opie nor GPE. May 13 02:41:10 BitBake Enterprise Edition May 13 02:41:19 Bitbake Longhorn ? May 13 02:41:19 I like bitbake XP :) May 13 02:41:24 1.3261.3.1 opie-calculator: add files from etc/ to get unit conversations work - patch by Gints Polis May 13 02:41:26 bitbake CE, for the lowmem version May 13 02:41:43 bitbake ME for the shit version that crashes all the time May 13 02:43:07 hrw|work: thanks,I was about to add gints' patch when my cable connection went down last night May 13 02:43:16 and Bitbake Server for tinderboxes May 13 02:43:17 (which is still down) May 13 02:43:26 koen: np May 13 02:46:21 i prefer bitbake-nwomosdgpc (now works on most open source developers grade personal computers) , catchy no ? May 13 02:47:49 hehe May 13 02:48:06 RP: Can the kernel disable "opie" apm requests ? This would help in the above mentioned situation! May 13 02:48:25 nah May 13 02:48:27 that's not it May 13 02:48:37 brightness control isn't done via apm May 13 02:48:55 Qt/Embedded needs to be tought to disable PowerManagement when it becomes aware of a VT change, that's all May 13 02:49:12 mickeyl: So can we do that ? May 13 02:49:16 sure May 13 02:49:20 just needs someone doing it May 13 02:49:26 this is on my pile of things to do since ages May 13 02:49:27 :/ May 13 02:49:45 mickeyl: did your guitar arrive yet? May 13 02:50:00 koen: not yet :( looks like i can't play over the weekend May 13 02:50:08 ah, too bad May 13 02:51:37 * koen waits for his books to arrive May 13 02:51:39 hi May 13 02:51:41 hey Bernardo May 13 02:51:43 hi Bernardo May 13 02:51:53 Bernardo: did you got serial cable? May 13 02:52:22 how can I reconfigure a package (not re-fetch & unpack); something like 'bitbake -c configure directfb' doesn't work. May 13 02:52:35 hi Bernardo May 13 02:52:39 wrobbie: bitbake --help May 13 02:52:44 wrobbie: look at -f May 13 02:52:54 you can force a command May 13 02:53:05 hrw|work: nope, the one I was auctioning for in ebay got snipped from me... May 13 02:53:11 * mickeyl zecke-bitbake opie-image May 13 02:53:47 Bernardo: aha May 13 02:53:50 mickeyl: Anotherone for you pile ;-) If SysRq+E terminates all processes from vt1 and you press "Fn+Right", the system gets stuck. You have to switch it off (remove battery). May 13 02:54:01 mickeyl: trying now.. May 13 02:54:39 I have to get schematics, tcx3106 connectors and send it - so friend will make some serial cables on 'taiwan' May 13 02:54:50 uv1: sysrq+b == reboot May 13 02:55:38 meh.. it nags bout quilt: May 13 02:55:45 NOTE: package quilt-native-0.39: started May 13 02:55:45 ERROR: InvalidTask event exception, aborting May 13 02:56:15 What's generally a reliable way to check if a device supports SDIO in Linux if there aren't any specsheets available? May 13 02:56:39 check the sd driver May 13 02:56:51 hrw|work: sysrq+b does not work either. May 13 02:56:55 uv1: that's rather for RP. It's a kernel problem May 13 02:56:57 And what if the driver doesn't support it currently? :) May 13 02:57:04 pray :) May 13 02:57:07 SDIO is more than just software support, right? May 13 02:57:08 Heh May 13 02:57:24 iirc SDIO is a command subset May 13 02:57:26 the kernel is stuck in the VT layer and waits for the ACK signal to change the VT May 13 02:57:34 a signal that never comes since all processes are dead May 13 02:57:34 heh May 13 02:57:36 * koen looks at RP for some SDIO advice May 13 02:57:51 koen: it's a layered approach May 13 02:58:00 it's hardware, base software, and driver May 13 02:58:02 hardware - ok May 13 02:58:05 software - ok May 13 02:58:07 driver - no way May 13 02:58:14 to sum up the situation :D May 13 02:58:32 * koen looks a cguys for opensourcing their drivers May 13 02:58:49 ~lart sharp May 13 02:58:49 yeah, that'd be nice May 13 02:58:49 * ibot executes killall -HUP sharp May 13 02:59:41 I heard this one company released SDIO Wifi for the C860 and others, but I've never come across any products or drivers of theirs. May 13 02:59:45 ~herring koen May 13 02:59:47 * ibot whacks koen on the side of the head with a large red herring named alfred May 13 02:59:52 mmmm herring May 13 03:00:01 SuN: cguys May 13 03:00:07 Sour or salty? :) May 13 03:00:13 RP: Can you take the SysRQ+E thing onto your stack ? May 13 03:00:20 normail, with onions May 13 03:00:22 Ah, that's what they're called. May 13 03:00:44 koen: with oil or cream? May 13 03:01:16 s/normail/normal/ May 13 03:01:32 hrw|work: without oil or cream May 13 03:01:52 I was hoping to get something like that to work with my Motorola E680... but first I need to find out if it's SDio compatible in the first place, then I need to slap Motorola until they release kernel source, then it needs major hacking because large parts of the flash are read-only as it is now. May 13 03:02:52 ya that sucks May 13 03:03:04 ~emulate zautrix May 13 03:03:05 this is exactly what the GPL is about. grab the stuff and run away May 13 03:03:22 Heh May 13 03:03:25 to make it even worse, you would need to reprogram everything since neither the EZApp API nor the libraries are published May 13 03:03:36 I can hardly believe they're getting away with it/ May 13 03:04:04 Yeah, some people got Opie to run on it... so you basically have an Ipaq... but the phone/GPRS part is utterly useless without drivers and apps. May 13 03:04:14 hrw|work: I'll send you my contacts, then. A serial cable will help (if/when I get adsl again) to start working on kernel 2.6 May 13 03:05:17 But I could live with "stealing" their API and libraries... though they're probably about as good as Sharp's stuff. May 13 03:05:23 Or as bad, I should say. May 13 03:05:38 Bernardo: ok. I had to ask you about it May 13 03:05:41 * koen dreams about X over gprs May 13 03:06:03 bleh May 13 03:06:15 more sun cvm hacking todo this afternoon May 13 03:06:22 SuN: you will have a hard time stealing an API if not even a header file is available May 13 03:06:23 ~lart sun May 13 03:06:24 * ibot decapitates sun conan the destroyer style May 13 03:06:31 ~lart java May 13 03:06:32 * ibot blames java for all the evil in the world May 13 03:06:50 ~botsnack May 13 03:06:51 thanks, koen May 13 03:07:02 Well with "stealing" I meant just using it. May 13 03:07:17 THough that's probably going to get problematic sooner or later. May 13 03:07:34 I can't imagine what's touched by those API's right now. May 13 03:07:42 SuN: There is no SDIO support in Linux and no specifications to let us do so. Until that changes, we're stuck May 13 03:08:14 SuN: We know the Zaurus hardware is SDIO compatible as it has the appropriate irq line for 1 bit mode May 13 03:08:41 uv1: I'm not really sure what you expcet the kernel to do once you've killed all the processes... May 13 03:08:45 And the cguys driver isn't GPL, I'm guessing? May 13 03:09:15 SuN: Its not been released at all... May 13 03:09:27 Hm alright... just rumors then. May 13 03:09:30 Bah May 13 03:09:46 I guess SD can be considered an "evil" standard. May 13 03:09:50 indeed May 13 03:09:52 yes May 13 03:10:04 ibot: SD is also evil May 13 03:10:07 We can't even get plain SD support into the kernel at the moment May 13 03:10:07 koen: okay May 13 03:10:11 ~SD May 13 03:10:14 hmm... sd is http://www.zaurususergroup.com/modules.php?op=modload&name=phpWiki&file=index&pagename=SD%20And%20CF%20FAQ, or evil May 13 03:10:35 Kinda odd, seeing how open CF is. May 13 03:10:42 But I guess corporations don't give a fuck. May 13 03:10:56 SuN: CF is just 16 pcmcia May 13 03:11:09 ...which is also open :) May 13 03:11:46 hrw|work: thanks a lot May 13 03:11:59 RP: Where should one start looking to debug why SD is so slow? May 13 03:12:23 RP: It took about 3 hours to extract a 70MB tar file last night May 13 03:12:51 ouch. That is bad :-/ May 13 03:13:00 faster sd would mean faster swap on my corgi May 13 03:13:03 RP: Actually, is might have been cloer to 2 hours, but it was very slow anyway May 13 03:13:38 koen: are you experiencing any deterioration in the card by using it as swap? May 13 03:13:39 RP: The system should not lock up no circumstances. Thats all. May 13 03:14:02 SuN: not yet, but sd is evil anyway, so I want it to suffer :) May 13 03:14:05 03mickeyl 07 * r1.3281.1.1 10openembedded/packages/libqpe/libqpe-opie_cvs.bb: apply the libqpe fix-titleheight patch for the cvs version as well, since I feel uncomfortable pushing it upstream May 13 03:14:15 You need to find out where the code is spending all its time. Sadly, as it probably queues requests and is interrupt driven, things like oprofile probably won't help May 13 03:14:35 uv1: Did the system actually completely lock or did it just get confused? May 13 03:14:42 (Did Sysrq+B reboot it) May 13 03:16:15 lardman|busy]: I'd start by checking the card is really running at the frequency you think its running at May 13 03:16:22 argh... fscking konq/emb May 13 03:16:54 RP: Frequency I think it is? May 13 03:17:27 RP: SysRQ+B did nothing May 13 03:19:32 lardman|busy: The system normally says the card is running at a frequency of 20Mhz. I've often wondered if it really is running at that frequency May 13 03:19:54 uv1: Ok, that is a problem. I might look into it but its a low priority May 13 03:23:43 speaking about sd.. I was thinking about buying 512M card but now with that timeout problems I will rather spend that cash on something else May 13 03:24:33 the 512mb kingston works with default timeout May 13 03:25:19 mickeyl: kingston is brand not manufacturer.. there are kingston with toshiba, hitachi cores... May 13 03:25:50 g'day all May 13 03:25:58 hi loial May 13 03:26:49 does anyone here have a 'remote control' thingie for the zaurus? May 13 03:27:07 hmm how long is the new bitbake supposed to hang after the "building the provider hash" ? May 13 03:27:12 loial: x2x? May 13 03:27:31 nm May 13 03:27:37 ehm, no, i mean the CE-RH1 May 13 03:27:56 mickeyl: quite long May 13 03:28:04 iirc lirc doesn't work with pxa ir devices May 13 03:28:16 sa should be fine May 13 03:28:35 CE-RH1 is that remote via headphones/mic? May 13 03:28:40 soryr, I guess I wasn't clear...it's a headphone remote control on a whire May 13 03:28:49 yep hrw May 13 03:29:16 http://www.nippon-export.com/catalog/images/PDA-CE-RH1.jpg May 13 03:29:35 yep, that's the one May 13 03:29:51 When I plug it in sound still comes from the zaurus speaker May 13 03:29:55 1.3284.1.1 added opie-icon-reload to opie-image - it is small and usefull May 13 03:30:09 I still need to figure out how to hook up a microphone to my corgi May 13 03:30:19 koen: left channel is mic May 13 03:30:35 only when i push a button on the remote thingie the sound comes from the headphones May 13 03:30:36 (and how to convince someone to swap my corgi for a husky) May 13 03:31:10 loial: and how it works with sharp'rom'? May 13 03:31:22 but when i release the button sound comes from the zaurus again May 13 03:32:14 with sharp rom what happens is: sound from zaurus internal speaker, plug headphone remote in, sound still from internal speaker, use remote once, sound switches to headphones and stays there May 13 03:32:42 moin zecke May 13 03:32:53 with OE 3.5.3 it doesn't stay at headphones but goes back to zaurus internal speaker May 13 03:32:55 loial: in any music player or only in sharp one? May 13 03:34:48 hrw|work: I don't understand your question, but in pdaXrom e.g. it worked with xmms the same way as on sharp ROM May 13 03:35:18 I guess the zaurus needs to be notified but OE 3.5.3 seems to forget it's notified May 13 03:36:33 loial: probably you have to trace it - dont know who in OZ team have ce-rh1 May 13 03:37:34 all apps in OE 3.5.3 have the same behaviour, i've tried mplayer, xmms and opieplayer May 13 03:38:34 loial: as I said - you have to trace it.. May 13 03:38:49 when I plug in 'normal' headphones it works okay:) May 13 03:39:08 dunno how much i can do, but i'll see May 13 03:40:08 loial: users use more hardware then OZ team has - so not all is supported in same way like in other distros May 13 03:40:48 loial: sometimes its easy to resolve (like it was with flexis keyboard), sometimes not May 13 03:40:50 I understand :) May 13 03:40:53 donation driven development :) May 13 03:41:15 loial: Is this with the 2.6 kernel? May 13 03:41:19 at least you've heard of it now:) May 13 03:41:23 yep RP on 860 May 13 03:41:36 I killed all the remote control support in 2.6 May 13 03:41:39 dmesg shows 'Insert HP' and 'Not Insert HP' May 13 03:41:51 When pressing the button? May 13 03:41:54 guess that means headphone May 13 03:42:02 It does May 13 03:42:06 when inserting headphones (normal, not the remote one) May 13 03:42:43 I don't have the remote ones, so I have no idea how they work or what the kernel should do when they're inserted May 13 03:42:48 koen: yes ;) and I was one step from getting that flexis keyboard... May 13 03:43:25 loial: The whole sounds driver is rather scary :-/ May 13 03:43:26 RP: well, I can see what's wrong May 13 03:43:27 koen: sometimes bugs are too easy to resolve... May 13 03:44:13 loial: What is wrong is that I removed all the remote control support from the kernel... May 13 03:44:32 Well, I never ported it accross to be more accurate May 13 03:45:02 mmh May 13 03:45:13 1.3284.1.1 won't be of any use soon :) May 13 03:45:24 mickeyl: on 2.6 kernel only... May 13 03:45:36 there is anything else than 2.6 ? ;) May 13 03:45:39 What is 1.3284.1.1? May 13 03:45:53 RP: when the remote headphone is inserted dmesg says some combinatin of 'Insert HP' and 'Not Insert HP', but it always ends with 'Not Insert HP'. Now if i touch a button on the remote headphones and keep it touched, dmesg says 'Insert HP' and sound plays okay. However when i release it, it shows 'Not Insert HP' again....all it should do is remember it's inserted I think. May 13 03:45:57 mickeyl: our lovely collie May 13 03:46:02 yah right May 13 03:46:10 RP: opie-icon-reload in opie-image May 13 03:46:43 smacznego mickeyl May 13 03:46:44 RP: is it a kernel thing? May 13 03:46:59 merci May 13 03:47:02 loial: Its a kernel thing May 13 03:47:14 now its time to add setserial and fuser into bootstrap May 13 03:47:17 loial: It reads a hardware gpio to decide if headphones are inserted or not. If that says they're not inserted it says "Not Insert HP" May 13 03:47:28 okay May 13 03:47:31 So your remote control tells the kernel its not inserted... May 13 03:47:55 Its not simple at all. Somehow, you have to detect the presence of the remote control and override the gpio May 13 03:48:52 If my remote is working okay.... I guess it should, but it's been a while since i used it:) Probably I should flash the sharp ROM to check May 13 03:49:23 loial: I think it will be working fine. The problem is I never ported support for the remote control accross May 13 03:49:44 The buttons will not work and it looks like it has some kind of mad detection scheme which is also broken May 13 03:50:42 I've never seen one, so I can't comment further. You need to fine someone kernel literate to port the code accross if you want it to work. That person is going to need hardware to test with. May 13 03:51:28 I understand. Thanks for the info:) May 13 03:52:14 ~lart savannah.gnu.org May 13 03:52:15 * ibot follow's savannah.gnu.org with a gauntlet and ... scratch ... HUMILIATION May 13 03:52:52 Something different, is there a way to type the ' character in a terminal? i sometimes like to type "grep bla 'find' " May 13 03:53:07 fn-7 ? May 13 03:53:11 i mean the ` character May 13 03:53:24 no idea May 13 03:53:35 Shift+7 ? May 13 03:53:35 I don't have my ZS in front of my at the moment May 13 03:54:34 shift 7 gives ', but i need `, that's the one below tilde at a normal keyboard, i don't know it's name :) May 13 03:55:29 it looks like the key next to space might give the correct one May 13 03:55:52 but i can't get it to produce it :( May 13 03:56:29 Please don't tell me i'm the only one that uses it :) May 13 03:57:56 backtick? May 13 03:58:14 yes, that might be it's name May 13 03:58:51 where does opie get it's TSLIB_TSDEVICE from? May 13 03:59:24 koen: tslib has scripts that declare it May 13 03:59:34 not on this rootfs :) May 13 03:59:41 RP: does your 860 reboot when you issue the command? Mine often stops at Rebooting... Restarting system. May 13 04:00:27 Issue which command? May 13 04:00:39 reboot command May 13 04:01:19 also pressing reboot in OPIE often makes the zaurus stop responding to anything May 13 04:01:51 That sounds like a rootfs issue. I always use the sysrq keys... May 13 04:02:20 what keys are they? May 13 04:02:40 uv1: Did you finish that writeup and have you a link? :) May 13 04:02:52 Ah, he's gone :-/ May 13 04:03:49 loial scroll back and see if you can see when I explained it to uv1 May 13 04:04:02 RP: http://www.uv-ac.de/openembedded/ May 13 04:04:33 sorry i wasn't online then May 13 04:05:23 loial: See http://www.uv-ac.de/openembedded/openembedded-8.html#ss8.1 May 13 04:05:32 uv1|cleaningMyBi: thanks :) May 13 04:05:33 uv1|cleaningMyBi: thanks for the link:) u to RP:) May 13 04:06:27 RP: it should read the env vars for tslib, right? May 13 04:07:18 later all May 13 04:07:25 bye koen May 13 04:09:31 RP: I always see Flash instructions mention removing the battery and reinserting. I have found that this isn't necessary. When I want to reflash I just reboot the zaurus and simply keep the OK button pressed jjust before it performs the actual reboot. It's a lot easier and has never given me any problems. May 13 04:30:30 loial: what model of zaurus? May 13 04:30:44 loial: C700? May 13 04:31:32 Each of the Zaurii models is a bit different. I've found I can reflash both the C700 and the C760 without removing the battery May 13 04:32:03 mithro: C860 May 13 04:32:17 The C760 just needs the Sysrq+B, the C700 needs the battery switch flipping before you SysRq+B May 13 04:32:54 ~seen pb_ May 13 04:33:44 RP: I just tried the sysrq+b, it said it was rebooting, but nothing happened, became unresponsive. Didn't you release a second kernel update? I'll try that one in a bit I guess May 13 04:34:05 pb_ <~pb@2002:5168:d330:1:a00:1fff:fe06:93c> was last seen on IRC in channel #gpe, 12h 23m 56s ago, saying: 'florian: heh'. May 13 04:35:02 loial: You'll probably have to press the power button to allow it to restart May 13 04:35:20 I tried, nothing worked, had to remove battery May 13 04:35:27 did you have 860? or 760? May 13 04:35:40 ah, 700 and 760 i guess May 13 04:35:41 760. It could be a difference between them :-/ May 13 04:36:02 The 700 doesn't need the power button press... May 13 04:36:25 hehe, guess higher numbers doesn't always mean better :) May 13 04:36:45 The 760 is better - the power button press makes sense May 13 04:36:54 nod May 13 04:37:32 I suspect that as sysrq+b fails, that's probably why your reboots fail as well May 13 04:37:50 my thoughts exactly May 13 04:37:53 It might not be, but I'd suspect it... May 13 04:45:49 how to remove -frename-registers from CFLAGS in just one .bb? May 13 04:52:01 debian finally starts to be more up-to-date when sarge is frozen... glibc 2.3.5 in experimental May 13 05:15:13 just tried to build opie-image for the first time.... bitbake failed trying to download http://www.oesources.org/source/current/config_savannah.gnu.org__20050513.tar.gz May 13 05:15:24 which does not exist on the website May 13 05:15:30 yeah, savannah is giving us crap again May 13 05:15:39 couldn't download it either May 13 05:15:44 20050512 should be present May 13 05:16:51 koen: Did you sort out your tslib problem? May 13 05:16:52 the most recent one in the file listing is http://www.oesources.org/source/current/config_savannah.gnu.org__20050504.tar.gz May 13 05:17:09 how would I manually plug that into bitbake? May 13 05:18:54 RP: maybe, but we tried a recent rootfs where qpe is giving us crap May 13 05:32:56 oh good i'll shoot all OZ users May 13 05:33:00 s/good/god/ May 13 05:33:12 why are they so braindead May 13 05:33:18 Dare I ask what they've done now? May 13 05:33:38 of course, you remember I put up a big fat WARNING for possible kernel mismatches stating that "running this combination is unsupported", right ? May 13 05:33:42 mickeyl: I wanna live May 13 05:33:56 with that knowledge in mind, read http://www.oesf.org/forums/index.php?showtopic=12388&hl= May 13 05:33:58 mickeyl: probably in 3.5.4 it have to shutdown May 13 05:34:00 i want to shoot this guy May 13 05:34:09 "everything is running fine" May 13 05:34:12 grrrrrrrrrr May 13 05:34:55 shoot it twice - once for me May 13 05:35:47 How can you ignore that warning :-(* May 13 05:36:04 That warning drives me *nuts* :) May 13 05:36:15 heh May 13 05:36:20 it's helpful May 13 05:36:21 then again May 13 05:36:26 it doesn't matter May 13 05:36:30 people will query us anyway May 13 05:36:34 i can safely remove that thing :/ May 13 05:36:39 mickeyl: make it shutdown in 3.5.4 May 13 05:36:40 support load won't be decreased May 13 05:36:53 hrw|work: yeah, i should consider that for the release images May 13 05:37:01 mickeyl: yes, for 3.5.4 it hangs :) May 13 05:37:22 mickeyl: Just make sure devs can set it so it doesn't appear at all for debug imges May 13 05:37:36 3.5.4 put kernel in initrd then dd if=kernel of=/dev/mtdblockblah on boot :-) May 13 05:37:57 XorA: yeah, or something like that. May 13 05:37:58 RP: ctrl C still get through it :-) May 13 05:38:35 XorA: How many users know where Ctrl+C is? May 13 05:38:42 RP: none May 13 05:38:50 XorA: It depends if the keymap gets loaded first anyway ;-) May 13 05:39:12 RP: well none that are going to fall over the flashing problem anyway, you need to read to find out where ctrl c is May 13 05:43:30 or better still, just display a message saying there machine has a virus, please return to sharp for removal May 13 05:44:00 now that's an idea May 13 05:44:24 You are too dumb to flash a kernel, please revert to a SharpROM (clone) May 13 05:44:28 *cough* May 13 05:44:30 ;) May 13 05:44:30 ;)) May 13 05:44:37 It gets my vote :) May 13 05:45:48 we have one updater.sh for all machines? May 13 05:45:48 pity we cant install PPC2003 :-) May 13 05:45:53 rofl, May 13 05:46:54 hrw|work: unfortunately not May 13 05:47:02 hrw|work: c7x0 and 5600 are sharing one May 13 05:47:21 seperate for all others May 13 05:47:28 could be merged May 13 05:47:28 ah May 13 05:47:35 I'm sure someone could merge them May 13 05:47:38 its shell code? May 13 05:47:45 Its just coorrdinating the testing that's tricky May 13 05:47:48 we already used the pdaxrom based once, but that one failed to work on everything except 6000 May 13 05:47:49 hrw|work: yes May 13 05:47:52 ya, it's shell May 13 05:48:01 hrw|work: do some voodoo and make it into a shar archive :-) May 13 05:48:11 and sharp gives plain busybox only? May 13 05:48:38 and is it possible to dump to the shell in updater.sh? May 13 05:49:37 hrw|work: I've never actually gone to a shell within it but I have wondered about it. It should be possible May 13 05:50:25 I go eat May 13 05:51:11 bon appetit May 13 05:51:17 or like the germans say May 13 05:51:18 "Guten" May 13 05:51:41 schmecklich May 13 05:51:44 :P May 13 05:51:53 yeah, it's pretty easy to drop off to a shell May 13 05:51:54 did that May 13 05:52:01 very handy to inspect what's on the initrd May 13 05:52:08 (the real initrd, this time :) May 13 05:53:27 RP: I just flashed original sharp ROM, the headphone remote is working correctly, so it's not broken :) May 13 05:54:14 loial: As I said, the remote control code is missing from 2.6 May 13 06:05:39 hrw|work: trying to get updater.sh to work with the C1000 I just added a call to drop off to a shell May 13 06:26:01 * emte sends my email server some electricity "clear" May 13 06:49:02 anyone here have azuerus? May 13 06:49:30 jk2: it doesnt run on OE May 13 06:49:36 jk2: but I have it May 13 06:49:41 I need someone to test out a download of BitBake on a QEMU image May 13 06:49:52 XorA: can you test this magnet:?xt=urn:btih:7YPQB3DWM3ARTZE3JL4FMLKL7FT7MT5H May 13 06:50:00 see if it downloads May 13 06:50:13 jk2: gimme a sec May 13 06:51:37 jk2: one seed, zero peers May 13 06:53:14 jk2: its downloading May 13 06:55:27 XorA: cool.... I wasn't sure if it would go through the firewall at my house.... looks like it's good through May 13 07:14:43 XorA: I posted my thing about using BitBake in QEMU with that download here http://www.oesf.org/forums/index.php?showtopic=12564 May 13 07:14:53 thanks again for testing it out for me May 13 07:18:17 jk2: your link doesn't look right May 13 07:18:22 jk2: itsnt really slow? May 13 07:19:16 jk2: pretty interesting though... nice idea May 13 07:19:40 jk2: www.colinux.org is probably more suitable for that trick May 13 07:24:03 jh_ what link? May 13 07:24:54 XorA: it's pretty slow yeah.... I don't know how long it would actually take to build as bitbake crashed due to missing some config tarball from today May 13 07:25:27 but I expect it should be fast enough to do a build overnight on my AMD64 3500 May 13 07:26:01 jk2: the link in the forum post May 13 07:26:18 jk2: the link on oesf forum has become corrupted, its become http/magnet/..... May 13 07:26:52 oooh that link May 13 07:29:50 k, I fixed the post.... the forum software prefixes all urls with http:// May 13 07:30:00 so I made it plain-text May 13 07:32:50 jk2: how do you open that in azureus? :) May 13 07:33:19 loial: File->Open Torrent->URL May 13 07:34:42 URL? the link isn't an URL... May 13 07:35:29 I was thinking about building oz 3.5.3 in /usr/local/openzaurus-3.5.3/ and then packing and sharing tmp of it May 13 07:37:17 loial: it's a magnet URI.... new protocol supported in the latest Azuerus (released beginning of this month) May 13 07:38:13 ah...i see...that version isn't stable on gentoo yet May 13 07:42:42 bummer HyperThreading is defective May 13 07:43:42 hehe May 13 07:43:49 * CosmicPenguin calms himself down May 13 07:43:56 shouldn't gloat - shouldn't gloat May 13 07:44:08 heh May 13 07:49:05 morning May 13 07:49:35 zecke: presumably it's easily fixed with a syscall to temporarily block other threads while a secure section is running. May 13 07:50:04 jamie: how do you know if something is secure? May 13 07:50:15 jamie: if a page of root is in memory? May 13 07:56:18 I'm off, bye all May 13 08:25:04 Has anyone looked at the patches I sent yesterday for omap osk board support (and omni*) yet? May 13 08:25:53 Crofton: I'm too thin for it May 13 08:27:16 to thin? May 13 08:30:49 not enough knowledge to check it May 13 08:31:25 ok. time to go shopping and home May 13 08:31:31 ah May 13 08:31:42 have fun! May 13 08:31:46 have a nice weekend May 13 08:31:55 you too! May 13 08:45:08 Good evening May 13 09:06:05 hrw|gone,RP: ping May 13 09:07:33 morning May 13 09:08:05 'night May 13 09:08:14 morning treke May 13 09:08:15 ~ugt May 13 09:08:17 ugt is, like, Universial Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. May 13 09:08:34 :) May 13 09:08:59 found a couple broken cases in the mirror code last night :p May 13 09:09:09 note to self, test with more than a dozen bbs May 13 09:11:43 mickeyl: hello May 13 09:12:31 mickeyl: Your package has just arrived which means I might get a chance to look at it this weekend :) May 13 09:15:59 cool May 13 09:16:12 RP: could you verify something for me when you have a chance? i have a problem with zecke's patches May 13 09:16:26 bug description follows: May 13 09:16:38 say we build an image for spitz May 13 09:16:42 set all configs May 13 09:16:46 bitbake virtual/kernel May 13 09:16:52 bakes openzaurus-pxa27x May 13 09:16:54 fine May 13 09:16:57 then i apply all patches May 13 09:17:06 rm -rf tmp May 13 09:17:10 bitbake virtual/kernel May 13 09:17:12 ... May 13 09:17:24 dependency arm-linux-gcc-2.95 not found May 13 09:17:33 although bbread shows it May 13 09:18:04 mickeyl: I'd guess nobody has used this code with the 2.95 compiler as yet... May 13 09:18:18 Anybody with a good knowledge of qte internals? May 13 09:18:20 well yeah, the problem is that a variable is missing May 13 09:18:27 ASSUME_PROVIDED is not taken into account May 13 09:18:44 that indicates a major problem in the variable logic somewhere May 13 09:19:10 I'm a little confused about the offscreen memory management. May 13 09:19:18 mickeyl: I'd suspect one of the bbclass files makes asumptions which are no longer valid May 13 09:19:37 s/little/very/ May 13 09:19:37 mickeyl: Does any of the bbclass files specifically handle ASSUME_PROVIDED? May 13 09:20:01 no May 13 09:20:08 assume_provided is only dealt with in bitbake May 13 09:20:17 which is hardly touched May 13 09:20:17 SirFred: I'm afraid I know nothing about qte... May 13 09:20:21 this is why i'm worried May 13 09:20:26 SirFred: i don't think anyone of us can help you there. best, ask that on qt-embedded-interest list May 13 09:20:56 RP: Thanks. I'm starting to think that I know nothing too May 13 09:21:01 mickeyl: Offhand, do you know where in bitbake assume_provided is handled? May 13 09:21:07 ya May 13 09:21:10 mickeyl: Thanks. I'll try May 13 09:21:24 ignore = bb.data.getVar("ASSUME_PROVIDED", make.cfg, 1) or "" May 13 09:21:26 __ignored_dependencies = ignore.split() May 13 09:21:32 bitbake line 642 May 13 09:22:09 mickeyl: And that list belongs to...? May 13 09:22:18 SirFred: lists.trolltech.com May 13 09:22:22 mickeyl: What project/webpage/whatever? May 13 09:22:26 mickeyl: Thanks again. May 13 09:22:29 np May 13 09:23:34 mickeyl: Which of zeckes patches do you have applied? May 13 09:23:43 all May 13 09:24:03 those posted on the list in the past two mails May 13 09:24:11 mickeyl: Do you have no_default_data.patch ? May 13 09:24:20 yes, applied May 13 09:24:24 should i try to revert that ? May 13 09:25:12 mickeyl: No, it should be fine May 13 09:25:16 ok May 13 09:25:29 let me see if the ASSUME_PROVIDED value makes it out of the data struct May 13 09:25:31 mickeyl: I doubt its anything really major as I managed an opie-image earlier May 13 09:27:01 ok, i hope so. still worried though. May 13 09:28:24 mickeyl: I share the concern, particularly given that code looks like it should work... May 13 09:30:13 we have way too many debian mirrors :) May 13 09:30:22 mickeyl: It would be interesting to have it print what it thinks ignore is... May 13 09:30:31 already inserted May 13 09:30:35 and now i'm confused May 13 09:30:42 because nothing gets printed, not even my "ignore is = " May 13 09:30:49 looks like this code path is skipped May 13 09:30:50 wtf ? May 13 09:31:18 Its in the middle on main() ?! May 13 09:31:23 yeah May 13 09:31:40 oh now May 13 09:31:46 ignore is virtual/arm-linux-gcc-2.95 virtual/arm-linux-gcc-2.95 May 13 09:32:01 ok, so the variables is extracted out of ASSUME_PROVIDED at last May 13 09:32:09 now i'll add some prints into the check dependency loop May 13 09:32:37 This is where I get lost in the code :) May 13 09:33:28 ok, must leave for some things. will pursuit this lateron May 13 09:33:30 cya May 13 09:34:01 cya mickey_away May 13 09:39:43 bye guys, see you monday (if I'm lucky) May 13 10:29:12 re May 13 10:29:22 mickey_away: could you send me your local.conf? May 13 10:44:32 evening May 13 10:52:08 mickey_away: did Hilbert only had 23 problems? May 13 10:52:42 RP: thanks for the mail May 13 10:52:57 RP: mickeyl has some problem with ASSUME_PROVIDED though May 13 11:27:11 * kergoth mutters about the moving company taking too long to email him necessary forms May 13 11:35:16 Thats not the last muttering you'll do about the moving company May 13 11:35:21 I still mutter about mine a year later May 13 11:36:50 Can somone look at the osk patches I posted yesterday? May 13 11:36:54 two years and still muttering May 13 11:37:15 Heh, Max had a moving company pick up his stuff on monday and we haven't heard from him since .... May 13 11:37:55 Crofton: what was the right url again? May 13 11:38:57 http://omap-dev.balister.org/~balister/p1.patch May 13 11:39:04 though p5.patch May 13 11:44:00 Crofton: they're not readable (use export -tpatch) May 13 11:45:29 hmm May 13 11:46:20 k makepatch -d -r1.2930.119.1 > p1.patch May 13 11:50:28 bk export -tpatch -r1.2390.119.1 >~/public_html/p1.patch ? May 13 11:50:47 it claims it can't find revision 1.2390.119.1 May 13 11:50:53 oops May 13 11:51:19 OK how is the p1.patch file now? May 13 11:51:28 what's the best way to restart pcmcia on a package install? Is it okay if I put this into .bb: pkg_postinst() { /etc/init.d/pcmcia restart } ?? May 13 11:55:12 zecke, is the p1.patch better now? May 13 12:02:51 # site/i686-linux May 13 12:02:51 # 2005/03/28 10:43:59-05:00 balister@max.mprg.org +1 -0 May 13 12:02:52 # sizeof float hack, this should probably fixed elsewhere May 13 12:04:18 Crofton: linux-omap_bk.bb can be changed May 13 12:04:31 It can actually be left out May 13 12:04:44 you can set COMPATIBLE_HOST instead of that anonymouse method May 13 12:04:54 I screwed up and leaked it into a change set I shouldn't have May 13 12:05:24 later I have a bb file that grabs a patch against a specific kernel version May 13 12:05:42 when I started, I needed to use the bk kernel version May 13 12:06:29 Crofton: p1 is readable now :) May 13 12:06:37 OK, I'll convert the rest May 13 12:07:01 hmm strange May 13 12:08:26 treke: what? May 13 12:09:45 File "/usr/lib/python2.4/site-packages/bb/parse/ConfHandler.py", line 33, in init May 13 12:09:45 bb.data.setVar('TOPDIR', os.getcwd(), data) May 13 12:10:12 OSError: [Errno 2] No such file or directory May 13 12:10:37 so the current workign directory doesnt exist? did you remove it out from under your shell? May 13 12:10:40 heh May 13 12:10:47 oh May 13 12:10:50 I wonder May 13 12:11:41 odd thing is it only happens on one bb so far May 13 12:14:51 zecke, they are all converted now May 13 12:20:59 Crofton: we keep patches normally in the unified diff format May 13 12:21:06 kergoth: did you've time to look at the patches? May 13 12:21:09 must be something about the gd bb May 13 12:21:09 not yet May 13 12:21:23 kergoth: don't hurry May 13 12:21:29 today i'm going to be running errands and stuff. faxing paperwork to the movers, packing, cleaning the car, etc May 13 12:21:34 but probably this evening May 13 12:21:44 cleaning the car? May 13 12:21:50 you could do that in minnesota :p May 13 12:22:01 not when a bunch of my shit is going with me in it, for the drive May 13 12:22:11 need to be able to fix my bags with the clothes and stuff May 13 12:22:14 er, fit May 13 12:22:16 heh May 13 12:22:36 kergoth: you moving to minnesota? May 13 12:23:04 yep. quit the job at ti. moving back to mn since i own a house there. can cut my expenses a fair bit while i figure out what i'm doing next. May 13 12:23:07 heh May 13 12:23:34 I just moved to st. paul from san francisco... where is your house? May 13 12:23:42 suburb of minneapolis May 13 12:23:53 cool May 13 12:24:25 i like minnesota in general. good quality of life, good cost of living. the snow drives me insane, and there isnt exactly much of a nightlife May 13 12:24:29 heh May 13 12:24:31 ah well May 13 12:25:10 well I've only been here less than 3 months, but it's been nice so far May 13 12:25:19 ok really strange May 13 12:25:24 if I start on gd it works May 13 12:25:26 at least you picked a good time of year to move May 13 12:25:29 if gd isn't first it fails May 13 12:25:52 treke: weird May 13 12:25:54 i guess.. I didn't really pick it... the layoff from work did ;) May 13 12:26:00 hehe May 13 12:26:18 why MN? land a job there? May 13 12:26:39 no, the in-laws are here, and my other half has been begging to move to mn for years May 13 12:26:41 kergoth: aha May 13 12:26:43 ah May 13 12:26:53 gchub looks like it's fucking things up May 13 12:27:12 gonna bet it's something in the svn fetcher May 13 12:29:12 yup May 13 12:29:24 the svn fetcher is a bit of a mess May 13 12:30:19 if I comment out that uri in gchub it keeps going. so I wonder where it's chdiring into May 13 12:32:33 ahahaha May 13 12:32:37 I think I see it May 13 12:32:46 kinkos charges a buck a page to send a fax May 13 12:32:48 what a ripoff May 13 12:33:00 the module changes into a bunch of directories May 13 12:33:20 which is fine, if it saves the original cwd and changes back.. May 13 12:33:21 heh May 13 12:33:22 but if the class raises a FetchError it never backs out May 13 12:33:25 aha May 13 12:34:34 that hurts May 13 12:36:57 bitbake does that alot. changing the cwd around. should really have a pushd/popd just for convenience May 13 12:37:00 heh May 13 12:37:21 well it actually looks like that isn't my problem, but it's still probably wrong :p May 13 12:37:26 hehe May 13 12:38:02 really, we should fix the code to stop making assumptions about the cwd when it gets called. make everything chdir() explicitly, so where you are isnt a factor May 13 12:38:05 heh May 13 12:38:25 the annoying bit is that it doesn't fail until bb.parse.handle() :p May 13 12:38:40 and at that point, I'm sure whatever is failing is long done May 13 12:48:54 lunch calls May 13 13:08:27 Ciao all May 13 13:09:08 hi May 13 13:09:18 you follow cycling at all? May 13 13:11:49 kergoth: should EXTRA_DEFINES_collie = "Foo" May 13 13:12:02 kergoth: EXTRA_DEFINES_append = " FOO" May 13 13:12:10 result in EXTRA_DEFINES = "Foo FOO"? May 13 13:20:55 I use cd `dirname $0` in a lot of my shell scripts May 13 13:21:47 hi all May 13 13:22:14 offroadgeek: I'm going to fix up my QEMU image and then I'll PM you and you can host it if you like.... May 13 13:22:19 that would be cool May 13 13:25:37 jk2: cool... I'm going to try that setup you posted in the forums May 13 13:26:26 hi pb_ May 13 13:26:36 Pigi: good evening May 13 13:39:53 03zecke 07 * r1.3287 10openembedded/packages/qte/ (qte-2.3.10/bidimetrics.patch qte_2.3.10.bb): May 13 13:39:53 Qt 2.3.10: May 13 13:39:53 Patch Qt to support Right To Left scripting of text. May 13 13:39:53 Patch is courtsey ALynch, the Makefile.in changes are May 13 13:39:55 done by myself May 13 13:40:17 mickey_away: I hope Qt doesn't get slower due that change May 13 13:46:29 zecke: I guess that if you can make Qt 2 slower, there will be no point in upgrading to Qt 4. May 13 13:47:24 pb_: right May 13 13:47:53 one could build a gui around cairo *shouder* May 13 13:48:03 yah, one could call it GTK May 13 13:48:11 *shouder* even more May 13 13:48:32 we have many compatibility problems with qt2 May 13 13:48:33 * zecke forks GTK and calls it GTK+ May 13 13:48:53 like many projects requests at least qt3 May 13 13:49:26 mean projects that could be ported to opie... May 13 13:49:58 i need a windows irc client that doesnt suck ass May 13 13:50:21 kergoth: You don't like xchat ? May 13 13:50:30 kergoth: irssi? May 13 13:50:35 SirFred: the xchat windows builds require registration to use after 30 days. May 13 13:50:46 kergoth: No older versions. May 13 13:51:04 zecke: yea, running an irssi in a remote screen atm. just prefer running a local client to the proxy. guess i could try irssi in cygwin May 13 13:51:13 kergoth: can't you use a cygwin build of oe to compile your own xchat? May 13 13:51:25 heh, if i were so inclined, sure May 13 13:51:29 am i? no, not really May 13 13:51:32 heh May 13 13:51:38 fair enough May 13 13:51:48 there are no doubt still some cygwin build quirks, and i'm not really in the mood to work through them May 13 13:51:59 right May 13 13:52:02 could always run xchat in colinux displayed on the cygwin x server May 13 13:52:03 hehe May 13 13:52:10 heh, yeah May 13 13:52:30 * zecke is too stupid to open his Überraschungs Ei May 13 13:52:35 I guess you could run GAIM under windows. May 13 13:52:54 thats true. not a fan of gaim's irc, but itd be better than nothing May 13 13:53:17 kergoth: I use a xchat on my windows machine at job. May 13 13:53:31 kergoth: It's modern enough for me, and don't require registering. May 13 13:53:45 * SirFred looking the exact version May 13 13:54:13 pb_: off hand, do you've an idea how to improve buildDepgraph from bin/bitbake May 13 13:54:19 pb_: without needing much more memory? May 13 13:54:36 hmm, windows gaim's irc works. the weird thing is it doesnt seem to pick up the list of users in the channel properly from the proxy May 13 13:55:06 kergoth: It's 2.0.10c. You can get it at http://www.xchat.org/files/binary/win32/xchat-2.0.10c.exe May 13 13:56:32 zecke: not really. One obvious win would be to not bother collating the stuff for "world" unless a world build was actually requested. May 13 13:56:45 SirFred: ah, thanks May 13 13:56:58 kergoth: np May 13 13:58:48 zecke: that would also allow you to avoid building the reverse provider hash, I think. May 13 13:59:47 I don't think there's any way to get out of building the forward hash, though you might be able to combine that with some other place that iterates over all the packages. May 13 14:00:09 pb_: Yes I plan to combine the steps May 13 14:04:02 it's time to sleep. May 13 14:04:04 Nite all May 13 14:06:15 kergoth: please add zecke123 to the bitbake group ;) May 13 14:11:35 03zecke 07 * r1.3288 10openembedded/packages/libsdl/libsdl-qpe-1.2.7/ipaq.patch: May 13 14:11:35 libsdl-qpe: May 13 14:11:35 Patch by Gints Polis to fix the rotation for h38xx iPAQ May 13 14:11:35 models. May 13 14:13:00 woot, got the paperwork faxed to the movers, and called and confirmed they got it May 13 14:13:05 * kergoth breathes a sigh of relief May 13 14:16:04 *puh* May 13 14:21:17 kergoth: rock May 13 14:21:29 have they given you an actual date for moving? May 13 14:21:46 nope. just know its between the 13th and the 30th. given the 13th is today, i assume it isnt today May 13 14:21:48 heh May 13 14:21:55 heh May 13 14:22:42 so, they just show up sometime in the next two weeks and start hauling your stuff away? May 13 14:23:10 mickey_away: I guess I've found your assume_provided problem May 13 14:23:13 nah, their dispatcher is supposed to call a few days before or something to let people know when the truck is doing the run May 13 14:23:31 ah, right May 13 14:28:40 a wise decision May 13 14:29:00 just wandering into a texans apartment and loading stuff into a truck isn't the safest thing to do May 13 14:29:14 hehe May 13 14:32:07 yeah, I always forget that everyone is armed to the teeth down there. May 13 14:33:03 hi May 13 14:33:43 i get 'Patch config-guess-uclibc.patch does not apply (enforce with -f)' when i try to bitbake nano for collie (openzaurus) May 13 14:34:11 pb_: I think putting buildDepgraph and execiItem into one execution path sounds good May 13 14:34:26 pb_: and making meta-world a special case as well May 13 14:34:32 zap-zero: read oe@handhelds.org May 13 14:35:02 zecke: what's execiItem? May 13 14:35:12 yeah, the world stuff should all be a special case. May 13 14:35:29 buildPackage(item) May 13 14:35:34 kergoth: is there any particular reason I wouldnt be able to print to the stdout from within the bb.fetch module? May 13 14:35:47 at least the 'best' provider for item can be put into the common executing path May 13 14:35:54 kergoth: fwiw, there are xchat windows builds that have the registration stuff removed May 13 14:36:31 treke: not that i know of May 13 14:36:34 reenoo_: url? May 13 14:36:40 dammit May 13 14:36:43 dammit May 13 14:36:44 dammit May 13 14:36:47 and what does fwiw mean, anyway May 13 14:37:18 I have bitbake installed in the system python directory May 13 14:37:27 so the debugging code I'm patching isn't really being used May 13 14:37:40 doh May 13 14:37:43 hehe May 13 14:37:47 * zecke asks the mover to move me into bed May 13 14:37:50 time to adjust PYTHONPATH, eh? May 13 14:38:04 or run the bitbake binary thats in its source tree May 13 14:38:19 i'm not using the bitbake binary May 13 14:38:21 sys.path.append(os.path.join(self.dirs["bitbake_dir"],'scripts', 'bitbake','lib')) May 13 14:38:26 so it's going to the end of the path :p May 13 14:38:36 heh May 13 14:39:44 kergoth: IIRC, http://www.psyon.org/projects/xchat-win32/ is what I suggested TyggerBob to use and it worked for him. haven't tried it myself May 13 14:39:58 so you know how two hours ago I noticed that we weren't chdiring back to the olddir before raising an error? May 13 14:40:02 that was the problem May 13 14:40:08 Sad, the only answer I got on qt-embedded-interest was from one guy from Scitech recomending me to try their drivers. May 13 14:40:48 I'll mail a patch off to the list May 13 14:46:05 I wonder how much we're doing s += "" instead of "".join() May 13 14:47:26 i'm amazed at just how slow += is, in python May 13 14:47:26 heh May 13 14:47:58 * CosmicPenguin is reading a wiki page on python performance May 13 14:52:25 We should try the profile function some time May 13 14:53:48 i think python is generally slow May 13 14:57:29 so anyone think it would hurt if we pulled some of these debian mirrors of the MIRRORS list? May 13 14:57:40 once one mirror loses the file they all do in pretty short order May 13 14:57:52 yeah, all those should really be _possibilities_ for PREMIRRORS May 13 14:58:04 since they exist more for local perforamnce than anything else May 13 14:58:06 and it's annoying sitting while it tries all of them :p May 13 14:58:09 yea May 13 15:00:01 zecke: works with the patch from the ml, thanks May 13 15:12:08 does the 2.6 Cxx0 guy lurk here at all? May 13 15:23:20 yes May 13 15:27:01 good nite May 13 15:47:22 Luke-Jr: that would probably be me... May 13 16:45:21 RP: any idea where the problem w/ the microphone is and how complex it might be to fix it? May 13 16:46:41 in particular, sometimes dsp can't be opened for reading at all and sometimes just not when it's opened for writing May 13 16:46:55 I think the latter is 2.4 tho, actually May 13 16:47:25 I don't recall getting dsp open for writing under 2.6 at all May 13 16:47:28 err... reading May 13 20:30:33 well my flights to europe are booked May 13 20:30:43 yay! May 13 21:42:59 hi May 13 21:43:25 in which channel can I find help about hacking a kernel together for a PDA? May 13 22:01:08 maybe #handhelds.org May 13 22:36:50 thanks! May 13 22:39:29 hmm May 13 22:48:03 boo May 13 22:48:30 * france is away: Away May 13 23:02:36 aaaaaaaaaaaaaaaaaaaaaassssdddddddddd May 13 23:13:57 hmm May 13 23:16:42 yeah May 13 23:16:59 hmmm... NOTE: package gnu-config-native-0.1cvs20050513-r3: task do_patch: failed May 13 23:17:13 bitbake's not liking my qemu image May 13 23:17:54 what doe nt work May 13 23:18:01 is this config okay? May 13 23:18:13 look good to me May 13 23:18:35 OE Build Configuration: TARGET_ARCH = "arm" TARGET_OS = "linux" MACHINE = "collie" DISTRO = "" TARGET_FPU = "" May 13 23:18:53 or is there supposed to be something in distro and target_fpu also May 13 23:19:48 do I build familar? I thought OZ and familiar were different May 13 23:21:37 do other folks use DISTRO="familiar" for OZ? May 13 23:22:12 > May 13 23:22:29 ? May 13 23:22:44 for the horde! May 13 23:23:08 um yeah... May 13 23:23:10 we can May 13 23:24:28 can someone post a collie /stuff/build/conf/local.conf config file that is good to go (i.e. doesn't crash with ulibc patch failures) to http://www.oesf.org/forums/index.php?showtopic=12564 so I can add it to the image? May 13 23:25:05 heh ... I'm clueless as to what options to put in there May 13 23:26:42 they are documented in the file... and I since people actually have bitbake going to build OZ and hengtes images that someone's either got a better config then what I put in the QEMU image or they are building off a better release that doesn't have ulibc patch failures.... May 13 23:28:30 there's the error message http://www.oesf.org/forums/index.php?showtopic=12564&st=0&p=79710&#entry79710 May 13 23:37:02 aha May 13 23:37:17 my email server is back ...( wonder where it was ) **** ENDING LOGGING AT Fri May 13 23:59:56 2005