**** BEGIN LOGGING AT Tue Feb 21 10:59:56 2006 Feb 21 11:00:07 hi greentux Feb 21 11:00:18 hi all :) Feb 21 11:00:35 hrw|work: ? is it possible to see the cpu speed of the borzoi? Feb 21 11:01:57 greentux_alt: cat /proc/cpuinfo ? Feb 21 11:01:57 greentux: there are cpufreq patch for pxa27x but do not know do we have it Feb 21 11:02:09 a.. see.. /proc/cpuinfo Feb 21 11:02:39 ~angstrom Feb 21 11:02:49 greentux: You can see the speed mentioned as the device boots up. As we don't have cpufreq running on the pxa27x, it doesn't change from that Feb 21 11:02:51 for the patch http://marc.theaimsgroup.com/?t=113992576800005&r=1&w=2 Feb 21 11:03:07 koen|assignment: no usable speed info in cpuinfo Feb 21 11:03:48 RP: ok, and it boots with max speed i assume? wheteher using battery or not? Feb 21 11:03:51 hmm, since upgrading bitbake I always get /home/dp/zaurus/build-gcc/tmp/staging/x86_64-linux/bin/install-sh: ./-m755 does Feb 21 11:03:54 not exist. Feb 21 11:04:03 greentux_alt: correct Feb 21 11:04:41 while using ipkg the borzoi seems very slow... dont know why Feb 21 11:05:18 greentux_alt: your feed lists are big Feb 21 11:05:44 koen|assignment: I took your hint about grep -v Source:, that works nicely Feb 21 11:05:59 XorA: it's only twice as fast :) Feb 21 11:06:09 koen|assignment: any more fields ipkg doesnt need? Feb 21 11:06:33 XorA: maintainer, homepage, etc Feb 21 11:06:52 it only needs provides, depends, replaces and conflicts Feb 21 11:07:08 koen|assignment: ok, me off to make those lists smaller Feb 21 11:08:34 * koen|assignment spots a way to make linux-openzaurus_2.6.15.bb 5 lines smaller Feb 21 11:09:09 file://defconfig-foo -> file://defconfig and make linux-openzaurus-2.6.15// subdirs Feb 21 11:17:36 ping RP Feb 21 11:18:12 lrg|home: pong Feb 21 11:18:35 RP: have you seen the email from Jingqing ? Feb 21 11:18:55 lrg|home: Yes. I didn't entirely understand it Feb 21 11:19:46 RP: I reading through the presentation now. English is not the first language so I expect things not to translate well Feb 21 11:20:51 lrg|home: Yes. I think I see what he's getting at althought I'm not sure if having all these virtual scenarios doesn't complicate things more than it simplifies... Feb 21 11:21:40 lrg|home: zaurusd is now in OE. It just manipulates the alsa config file depending on which jack function a user selects Feb 21 11:21:50 ah Feb 21 11:22:16 I'll do a build for borzoi after I get back from fosdem Feb 21 11:22:54 RP: I can see why they need the complexity, as it's initially designed around cellular devices Feb 21 11:23:44 lrg|home: Yes, I can see they useage case being a bit more complex Feb 21 11:24:19 gah... sdl needs fixing... 1.2.7 failes with "SDL_cpuinfo.c:107: error: PIC register `%ebx' clobbered in `asm'", and if i upgrade to 1.2.8 libsdl-net dont want to produce a .so :P Feb 21 11:25:12 lrg|home: Mixer simplification and their need for virutal devices are only partially related as I see it. Feb 21 11:35:13 Hey Liam Feb 21 11:35:47 Hey Dirk Feb 21 11:46:50 <_law_> NOTE: package tinymail-0.0+svn20060221: completed Feb 21 11:53:02 _law_: great, does it work? Feb 21 11:53:23 <_law_> pH5, build feed one moment :-) Feb 21 11:53:38 * pH5 curious Feb 21 11:58:55 hi lrg|home Feb 21 11:59:05 hey greentux_alt Feb 21 11:59:58 lrg|home: checked a sipphone, but it sounds like a robot Feb 21 12:00:55 greentux_alt: I think RP found some good settings for his headset yesterday. iirc he had to set the alc to stereo for best sound. Feb 21 12:02:56 lrg|home: ok, think that there can be some problem with dsp sample rate etc Feb 21 12:04:14 greentux_alt: sample rate should be 8k for voip, can you record the audio from the headset and email it to moe Feb 21 12:04:27 me Feb 21 12:05:19 lrg|home: That was just a gain issue - setting ALC to left worked fine once I turned down the input source Feb 21 12:06:02 ok, 1 less thing to worry about :) Feb 21 12:08:01 i tried only recieving of speech, no input/mic issues in the moment... Feb 21 12:08:18 RP: can you mail me the alsa mix conf? Feb 21 12:08:51 ~lart phpmyadmin Feb 21 12:08:51 * ibot wallops phpmyadmin with a main rotation server that needs rehubbing. It won't take long Feb 21 12:09:11 <_law_> ERROR: Cannot satisfy the following dependencies for eds-dbus: dbus Feb 21 12:09:49 hrw|work: I hate that frontend Feb 21 12:10:09 <_law_> any ideas where i can change this to dbus-1 ? Feb 21 12:10:45 XorA: I hate to have only this available Feb 21 12:11:12 hrw|work: I took to running mysql-admin via ssh -X Feb 21 12:11:13 I also hate gpe for their non-device behaviuor Feb 21 12:11:15 _law_: OE should have done that for you Feb 21 12:11:46 XorA: no shell there Feb 21 12:13:05 hey Feb 21 12:13:18 hey zecke Feb 21 12:13:39 <_law_> koen|assignment, its an selfbuild package (tinymail) Feb 21 12:14:01 <_law_> ah same problem with eds-dbus Feb 21 12:14:10 _law_: fallen victim to the hype? Feb 21 12:14:26 _law_: This is in .dev? If so, you probably need to rebuild dbus as you haven't wiped tmp for a while Feb 21 12:16:10 pH5: sorry to hear about my continuing incompetence vis a vis kernel26. I'll check those files in soonest. Feb 21 12:16:52 pb__: no problem, you see I found the patch ;) Feb 21 12:17:01 <_law_> RP, yes its in .dev i have done some updates in you eds-dbus .bb push it to bugs.treke.net ok? Feb 21 12:17:07 NOTE: package snes9x-sdl-qpe-1.39: failed Feb 21 12:17:10 pH5: is x11/fonts legal? Feb 21 12:17:14 I vote for removing it Feb 21 12:17:28 zecke: we never took decisions which sections are legal Feb 21 12:17:49 hrw|work: hehe, psst Feb 21 12:18:09 hrw|work: this is why I'm asking ph5 and others if we think this section is sane Feb 21 12:18:11 zecke: it seems hard to imagine that x11 fonts are actually illegal even in the harshest of jurisdictions Feb 21 12:18:29 _law_: Can you pastebin that? I'd like to see what kind of updates we're talking about Feb 21 12:18:30 zecke: who defines this? in debian those font packages are in section x11 Feb 21 12:18:54 pb__: I imagined X11 is so old it predates the existince if glyphs/fonts Feb 21 12:19:12 pH5: We define this, I dumply copied oelint.bbclass Feb 21 12:20:11 <_law_> RP, http://bugs.treke.net/show_bug.cgi?id=700 Feb 21 12:21:01 zecke: it supports "bitmap fonts", which were some primitive and archaic form of writing from which modern fonts are descended. Feb 21 12:21:46 <_law_> zecke, the memory usage hype? :-) Feb 21 12:22:10 I can use OOPS design patterns with assembly language :) Feb 21 12:22:39 these two b's there are my class, and this manipulation of PC is my inheritence :) Feb 21 12:22:53 zecke: eleet Feb 21 12:23:18 pb__: kudos belong to others :) Feb 21 12:23:23 _law_: You need camel? Feb 21 12:23:35 <_law_> RP, yes for tinymail Feb 21 12:24:03 <_law_> RP, also done something in do_stage Feb 21 12:25:15 diffs are usually better so people can easily see what changed... Feb 21 12:26:41 <_law_> RP, should i add an diff also? Feb 21 12:26:45 <_law_> http://handhelds.org/scap/port.20280.png :-) Feb 21 12:27:09 _law_: A diff would be easier Feb 21 12:27:19 <_law_> RP, ok mom Feb 21 12:27:21 hmm... Are there any known problems with that libtool reporting bas that it cannot create shared libs? Feb 21 12:27:34 back even Feb 21 12:27:39 lrg|home: i will send you a recorded file... Feb 21 12:28:58 <_law_> RP, http://helmsdeep.homelinux.net/~pat/myeds.diff Feb 21 12:29:49 Speaking of .dev... is gnu-config-native still broken? Feb 21 12:30:25 NAbyss: Yes, until we devide on the right fix... Feb 21 12:30:45 Ah, okies.. is there any easy workaround? Feb 21 12:31:17 NAbyss: yes: http://handhelds.org/hypermail/oe/65/6559.html Feb 21 12:31:22 NAbyss: Add RDEPENDS_${PN}="" to the file Feb 21 12:31:41 Thankye :) Feb 21 12:33:51 <_law_> should i post tinymail.bb so someone could add it to OE ? Feb 21 12:34:02 lrg|home: its with a external mic recorded mp3 (from the headphone) Feb 21 12:57:55 lrg|home: recieved mail? Feb 21 13:01:48 greentux_alt: not yet, I'll check again after lunch brb Feb 21 13:10:51 morning Feb 21 13:12:11 03coredump 07org.oe.dev * rc05500f9... 10/packages/altboot/altboot_0.0.5-rc2.bb: altboot: Add altboot_0.0.5-rc2.bb. The versioned .bb's download altboot from CVS, this makes stable releases of altboot possible. Feb 21 13:24:33 greentux_alt: not yet, how big is it ? Feb 21 13:28:23 03coredump 07org.oe.dev * r3043e4c5... 10/packages/altboot/ (4 files in 2 dirs): altboot: mv altboot.bb altboot_0.0.0.bb Feb 21 13:42:51 lrg|home: small... which email? Feb 21 13:43:17 lrg|home: tried your work email Feb 21 13:45:17 lrg|home: sent again Feb 21 13:54:42 greentux_alt: got it now. :) Feb 21 13:55:44 lrg|home: fine :) Feb 21 13:56:50 greentux_alt: we've some email problems today, your first email is probably sitting in a queue somewhere... Feb 21 13:59:07 "collect2: ld returned 1 exit status" - someone can tell me how to get more info from linker? Feb 21 14:02:39 greentux_alt: sounds like the gain is wrong (maybe too high as I hear clipping), can you try adjusting the mic volume and alc settings. Feb 21 14:03:03 lrg|home: what you hear is recorded from the headphone! Feb 21 14:03:19 lrg|home: and the mic of the sending (real) voip phone is right. checked that. Feb 21 14:03:32 lrg|home: so there is something in the playback way of the sound Feb 21 14:04:27 greentux_alt: ok, this may be a bug. I'll be able to look at this after FOSDEM (or during) Feb 21 14:05:31 lrg|home: but RP said recording and playback (both local) is ok. do you have a idea where the bug may be? Feb 21 14:05:48 <_law_> could someone test if dbus-monitor works on his Z ? Feb 21 14:06:34 greentux_alt: recording with mic is ok, recording with headset - I have not managed to fully test. Feb 21 14:07:00 greentux_alt: do you mean the audio was recorded from the far end ? Feb 21 14:07:17 lrg|home: that is playback via voip app (kphone/PI). perhaps something in setting sample rate or so? Feb 21 14:07:29 lrg|home: yes, recorded far end (snom voip phone) Feb 21 14:07:57 greentux_alt: ah Feb 21 14:08:17 lrg|home: :) Feb 21 14:09:07 greentux_alt: play back at 8k on 0.10rc2 is ok, it may be an incorrect setting with the older version. can you try using aplay to play an 8k sample Feb 21 14:10:14 i have an older version? Feb 21 14:10:19 lrg|home: i have older version? Feb 21 14:10:29 lrg|home: can u send me a right sample file? Feb 21 14:10:56 greentux_alt: yes, ASoC 0.10rc2 came out yesterday. RP will upgrade after FOSDEM i think. Feb 21 14:11:28 greentux_alt: although it's best to check at aplay @ 8k Feb 21 14:11:37 lrg|home: ok, sen dme sample. will try. would be nice, if that is the bug... :) Feb 21 14:13:30 greentux_alt: I don't have any 8k samples where I am at present. However, mplayer can generate an 8k wav easily from an mp3. iirc itr's something like mplayer -rate 8000 -c 2 -ao pcm file.pcm file.mp3 Feb 21 14:14:45 lrg|home: moment Feb 21 14:17:24 greentux_alt: I've found an 8k sine wave, sending.... Feb 21 14:17:31 fine Feb 21 14:19:08 someone know current status of kergoth dns work? Feb 21 14:20:03 lrg|home: aply 997.wav does: :) Feb 21 14:20:16 lrg|home: a tone with approx 300Hz Feb 21 14:20:30 greentux_alt: ah that's a bug then Feb 21 14:20:39 it should have 997Hz tone Feb 21 14:20:52 lrg|home: ok, do you think its fixable? Feb 21 14:21:04 greentux_alt: yes, already in 0.10rc2 Feb 21 14:21:15 lrg|home: puh you saved my job :) Feb 21 14:21:36 greentux_alt: is this for borzoi ? Feb 21 14:21:46 lrg|home: yes Feb 21 14:22:16 greentux_alt: 8,11,16,22,44.1,48k are all ok on borzoi in 0.10rc2 Feb 21 14:22:25 lrg|home: :) Feb 21 14:25:39 I can try and upgrade to the new ASoC later today if appropriate Feb 21 14:27:03 RP, you are hero Feb 21 14:32:09 ~hail RP Feb 21 14:37:24 lrg|home: ibot? Feb 21 14:37:50 * lrg|home thinks ibot is deaf Feb 21 14:40:56 hello Feb 21 14:41:05 lrg|home: Hello. Feb 21 14:41:32 hey swc Feb 21 14:41:52 This is stanley from SH. Nice to meet you in the channel Feb 21 14:42:08 swc: Hi Stanley, glad to hear from you Feb 21 14:42:31 SH as in SuperH? Feb 21 14:42:54 swc: I'm just writing a reply to Jingqing abot the scenario's for ASoC Feb 21 14:43:01 koen: Shanghai Feb 21 14:43:20 ah, I see Feb 21 14:43:34 koen: Stanley is the Linux BSP team leader at Intel Feb 21 14:43:39 cool Feb 21 14:44:04 koen: He's going to evaluate OE for some internal use Feb 21 14:44:04 http://handhelds.org Feb 21 14:44:39 RP: I'm having surprisingly few problems with gcc4 Feb 21 14:44:46 RP: none, actually Feb 21 14:44:51 RP: good work :)l Feb 21 14:44:55 s/l// Feb 21 14:44:56 koen meant: RP: good work :) Feb 21 14:45:00 ibot: shut up Feb 21 14:45:02 yes, Master koen Feb 21 14:46:41 Angstrom-bootstrap-image-test-20060220-h2200.rootfs.jffs2 Feb 21 14:46:41 Angstrom-gpe-image-test-20060220-h2200.rootfs.jffs2 Feb 21 14:46:42 Angstrom-opie-image-test-20060220-h2200.rootfs.jffs2 Feb 21 14:46:51 swc: must be late for you. Have you managed to install oe on a build machine yet ? Feb 21 14:46:52 hehe Feb 21 14:46:59 no idea if they work, but they compiled :) Feb 21 14:47:25 lrg|home: I am downloading oe. It is huge. Feb 21 14:47:46 60MB or so Feb 21 14:47:48 * koen checks Feb 21 14:48:19 43MB even (not counting bitbake) Feb 21 14:48:40 koen, swc: it's getting bigger Feb 21 14:48:41 koen: how they differ to familiar 0.8.3 except being built using gcc4? nearly nothing Feb 21 14:48:47 RP: did you fix anything yet? :} Feb 21 14:49:10 hrw|work: gpe differs a lot Feb 21 14:49:19 hrw|work: gtk 2.8, dbus 0.60 Feb 21 15:01:12 someone remember how much time it takes to propagate dns entries? Feb 21 15:01:22 lrg|home: some problem with your Z... Feb 21 15:02:13 greentux_alt: no, everthing is fine. It's just at home now and I've had to come into work. Feb 21 15:02:57 lrg|home: :) no... we need it back next week. thats th eproblem. but i think i can org another one next weeks Feb 21 15:03:27 greentux_alt: ah Feb 21 15:03:49 greentux_alt: I'll bring it to fosdem Feb 21 15:03:54 Can anyone view this video? Feb 21 15:03:56 http://video.google.com/videoplay?docid=4258374615945369026&q=night%20explosion Feb 21 15:03:58 hrw|work: it depends on the TTL (time to live) on the server Feb 21 15:04:18 lrg|home: tnx (its owned by another "division", so i have to org one for me... its a matter of time) Feb 21 15:04:25 hrw|work: on hh.org somethings are set to 10minutes others to 72 hours. Feb 21 15:04:33 greentux_alt: np Feb 21 15:04:42 lrg|home: we should talk on fosdem... Feb 21 15:04:44 Crofton: works over here Feb 21 15:04:48 grrr Feb 21 15:04:55 google sucks Feb 21 15:05:07 I get a free lunch if I find a viewable copy :) Feb 21 15:05:15 greentux_alt: fyi, I arrive in Brussels on Thursday Feb 21 15:06:24 Crofton: http://ewi546.ewi.utwente.nl/tmp/NightExplosion.avi (in ~ 3 mimutes) Feb 21 15:06:42 thanks koen :) Feb 21 15:06:51 enjoy the lunch :) Feb 21 15:06:57 lrg|home: will arrive friday 17 o clock Feb 21 15:07:07 koen: We (OH) tried gtk+2.8 and ended up reverting to 2.6 Feb 21 15:07:37 RP: it's fast enough for me Feb 21 15:07:46 RP: and it fixes a bug in gpe-today Feb 21 15:07:57 koen: Does xtscal work nicely? Feb 21 15:08:23 koen: Nice to see gcc4 might eventually work out though :) Feb 21 15:08:26 no idea, I save my /etc/pointercal files Feb 21 15:08:49 Crofton: upload complete Feb 21 15:08:55 try it - had real problems in our images Feb 21 15:10:32 building for c7x0... Feb 21 15:12:17 koen, you should draft shadows, he seems pretty keen on patching everything for gcc4.x Feb 21 15:12:32 koen, I have a local copy, thanks! Feb 21 15:12:42 greentux_alt: I'll meet you at the beer festival :) Feb 21 15:12:42 emte: that's on my todolist Feb 21 15:12:49 lol Feb 21 15:16:32 lrg|home: where? Feb 21 15:17:21 greentux_alt: I'm sure philippe mentioned it in an email. checking.... Feb 21 15:18:39 greentux_alt: details at bottom of the main fosdem page Feb 21 15:18:49 ok Feb 21 15:19:14 hopefully the others can make it there too. I'm sure there will be enough beer for all :) Feb 21 15:23:11 The Leffe Brun is mine, you can drink the other flavours Feb 21 15:23:31 btw., badges would be handy :D Feb 21 15:24:08 hey mickeyl Feb 21 15:24:12 hi lrg|home Feb 21 15:24:17 @check_leffe_brun Feb 21 15:24:17 hi mickeyl Feb 21 15:24:22 mickeyl: did you see my email Feb 21 15:24:23 def drink(self): Feb 21 15:24:39 def drink(self,drink): even Feb 21 15:24:44 lrg|home: I wasn't in the office today. let me check remotely Feb 21 15:24:59 mickeyl: what did I model with my class? (what is the classname?) Feb 21 15:25:06 hi mickeyl Feb 21 15:25:14 mickeyl: np Feb 21 15:25:15 zecke: class computerScientistOnFOSDEM: Feb 21 15:25:49 it's a problem though, your function is recursive and lacks a termination Feb 21 15:25:50 heh Feb 21 15:26:08 mickeyl: I had more thought of class mickeylAtFosdem Feb 21 15:26:17 hey mickeyl Feb 21 15:26:20 mickeyl: I think I fell in love with pythons decorator's Feb 21 15:26:39 mickeyl: now I would love to have a for loop to inject a decorator into each method of a module Feb 21 15:26:50 mickeyl: that would be an awesome tracing framework Feb 21 15:27:16 hi mickeyl Feb 21 15:28:20 hi mickeyl Feb 21 15:30:15 hi marcin, phil, richard Feb 21 15:30:36 zecke: yeah, the decorators are pretty handy, though I'm still a bit sad about the syntax. Feb 21 15:31:42 zecke: well, i need to control my beer consumption - i need all brain cells until I submit my Dissertation ;) Feb 21 15:31:54 FYI MIPS users - it looks like MIPS has been officially delegated to glibc-ports Feb 21 15:32:10 CosmicPenguin: ah, more breakage Feb 21 15:32:53 yep Feb 21 15:35:14 * koen prints http://www.fosdem.org/docs/plan-grey.pdf Feb 21 15:41:10 heh.. kmail and groupware functionality... Feb 21 15:45:10 hrw|work: have fun :) Feb 21 15:46:29 RP: I'm going to fix the metadata now Feb 21 15:46:31 zecke: looks like it is another place in kde where 'everything works' but only for one developer between 13:00 and 13:30 each friday (if it does not rain) Feb 21 15:46:56 hehe Feb 21 15:47:09 zecke: Are we going to try ASSUME_PROVIDED covering both DEPENDS and RDEPENDS? Feb 21 15:47:40 RP: I think we should try it :) Feb 21 15:47:51 zecke: ok, feel free to fix the metadata then :) Feb 21 15:48:26 I think the meta data needs fixing anyway :) Feb 21 15:48:28 mickeyl: finally you are about Feb 21 15:49:04 koen: did you ever use BibDesk? it is wonderful Feb 21 15:49:23 zecke: haven't even heard of it Feb 21 15:49:55 koen: bibdesk.sf.net Feb 21 15:49:57 ooh Feb 21 15:49:58 zecke: I agree. I just don't fancy being in here when people find OE doesn't work because of it ;-) Feb 21 15:50:01 looks nice Feb 21 15:50:31 RP: well the tinderbox shows it is fucked up :) Feb 21 15:50:33 *tired* Feb 21 15:50:46 mickeyl: I want to make slight changes to your python packages Feb 21 15:51:23 shadows: if you want commit access send your pubkey to koen@handhelds and mickeyl@handhelds.org Feb 21 15:51:47 okay, by pubkey you mean ssh key? Feb 21 15:52:24 monotone pubkey @openembedded.org Feb 21 15:52:28 ah Feb 21 15:52:38 i'll make a note of it, and do that later today Feb 21 15:52:42 need to go to work :) Feb 21 15:53:13 err - stupid question - whats the point of UDEV_DEVFS_RULES? Feb 21 15:53:32 backwards compatibility with devfs Feb 21 15:53:36 CosmicPenguin: hardcoded paths in old apps Feb 21 15:53:36 i think Feb 21 15:53:53 like some apps were written with the /dev/discs/discn/partn/ thing Feb 21 15:54:12 CosmicPenguin: Its only used by familar Feb 21 15:56:19 ahhh Feb 21 15:56:22 * CosmicPenguin sees Feb 21 15:56:48 These things happen when I only look at packages once a year Feb 21 15:57:03 gosh I'm firewalled Feb 21 16:02:26 hola pH5 Feb 21 16:03:55 zecke: did you manage to build a static monotone 0.26+ double root yet? Feb 21 16:04:07 koen: I did not try :) Feb 21 16:11:19 * chouimat is away: lunch Feb 21 16:11:39 hola zecke :) Feb 21 16:12:15 koen: the tinderbox looks more solid :) Feb 21 16:27:07 koen, I was beaten by about 8 minutes for the lunch competition :) Feb 21 16:29:54 http://labnol.blogspot.com/2006/01/download-watch-google-videos-in-any.html Feb 21 16:47:37 cu Feb 21 16:52:12 mickeyl: email sent, time for your marketing hat ;) Feb 21 16:56:21 * france is away: Away Feb 21 17:02:52 RP: no issues with gtk 2.8/gcc4/gpe yet Feb 21 17:03:04 RP: except for the keylaunch problem, of course Feb 21 17:06:44 lrg|home: heh, ok, i'll try to come up with some good reasons asap Feb 21 17:07:17 mickeyl: I just booted my first angstrom image :) Feb 21 17:07:19 XorA|gone: sorry, gone for sports tonite. will be back tomorrow evening. in the meantime, you can just drop me an email with what you want to change, if you want to Feb 21 17:07:25 koen: congrats! :)) Feb 21 17:08:48 koen|food: Did you check xtscal? Feb 21 17:10:36 RP: that's run on first boot, isn't it? Feb 21 17:11:08 the thing with the blue crosshairs Feb 21 17:11:13 koen|food: to calibrate the screen, yes Feb 21 17:12:10 works like it should Feb 21 17:12:55 I'll have to work out why it didn't in our images then :-/ Feb 21 17:13:16 gtk+ 2.8.9-r0 Feb 21 17:13:34 This was about 2.8.6 time so perhaps a bug was fixed... Feb 21 17:13:43 RP: http://ewi546.ewi.utwente.nl/angstrom/c7x0/ Feb 21 17:14:09 you have to rm /usr/bin/keylaunch, but after that it should work Feb 21 17:14:26 koen|food: If I reflash at the moment, it will take some time to recover the device :-/ Feb 21 17:15:38 cu Feb 21 17:16:33 RP: it'd probably be easier for you to build poky with gcc4/gtk 2.8.9 Feb 21 17:17:17 * koen|food gets some more soup Feb 21 17:43:18 * france is back (gone 00:46:58) Feb 21 17:53:41 * chouimat is back. Feb 21 18:12:52 * Philippe is back (gone 19:41:55) Feb 21 18:27:33 evening Feb 21 18:49:06 re Feb 21 18:49:22 RP: I have now an updatet list of packages I'm going to fix Feb 21 18:49:35 I'm just home after a day in the office and visiting my mom in hospital Feb 21 18:50:21 reenoo: ping, *sorry* I suffered from the meld bug Feb 21 18:57:30 JustinP: ping Feb 21 18:57:31 pb_: ping Feb 21 18:57:41 ping -B OE.255.255.255 Feb 21 18:58:33 hi zecke Feb 21 18:59:29 hi zecke Feb 21 18:59:48 I'll look at enhancing the ASSUME_PROVIDED functionality Feb 21 19:00:02 zecke: I take it you're referring to the gtk-webcore changes you applied? a current checkout looks like you've reverted your own changes and everything's fine. is that correct? Feb 21 19:00:14 reenoo: I changed the osb-n* stuff myself to properly parse and then I had a merge conflict and it reverted... Feb 21 19:00:22 reenoo: well someone else fixed my crap Feb 21 19:00:29 but current co looks good Feb 21 19:00:49 zecke: I had checked in a fix a few hours before I think Feb 21 19:01:14 s/before/before you tried to check your changes in Feb 21 19:01:15 reenoo: I had it there since saturday (on ewi though) Feb 21 19:01:28 locally on ewi Feb 21 19:01:48 reenoo: well it looks fine now so there is nothing to worry about Feb 21 19:02:02 sure Feb 21 19:02:18 RP: http://ewi546.ewi.utwente.nl/OE/depends.log Feb 21 19:02:26 RP: these are the current issues Feb 21 19:03:11 RP: and I think I'm suffering the explode_deps bug (see kaffe) Feb 21 19:03:50 zecke: That should be fixed in svn now Feb 21 19:04:28 hmm.. 3.5 hours to implement, use for an application, test, and document a backpropagation algorithm... Feb 21 19:04:35 * reenoo heads off Feb 21 19:04:38 later all Feb 21 19:07:42 good luck Feb 21 19:23:19 zecke: pong? Feb 21 19:23:32 JustinP: http://ewi546.ewi.utwente.nl/OE/depends.log Feb 21 19:23:45 JustinP: see the e*-native dependencies, they are wrong :) Feb 21 19:26:11 zecke: and...what is the point fo RDEPENDS for native packages? Feb 21 19:26:32 JustinP: they have runtime dependencies on non-native packages Feb 21 19:26:36 I'm also assuming that this check only used the most recent version #'s... Feb 21 19:26:45 that's not what I'm asking Feb 21 19:26:56 I'm asking what RDEPENDS means on -native packages Feb 21 19:27:17 I haven't heard most of the recent discussion Feb 21 19:27:24 it seems that DEPENDS would be enough... Feb 21 19:27:40 JustinP: it is the same as any other RDEPENDS when BUILD_ALL_DEPS is set Feb 21 19:28:05 JustinP: DEPENDS are build time dependencies. RDEPENDS are runtime dependencies. RDEPENDS are only needed once built, before you run something. This applies to native packages, much the same as anyting else Feb 21 19:28:10 JustinP: and yes this is new, it was introduced with the BUILD_ALL_DEPS change Feb 21 19:28:21 .... Feb 21 19:29:01 well, I don't see how it could make a difference....I agree that the RDEPENDS should be fixed for these...but I don't understand what I'm supposed to put in there Feb 21 19:29:11 RDEPENDS = "" I assume? Feb 21 19:29:25 JustinP: No, things like patch-native instead of patch Feb 21 19:29:35 the build system doesn't know when it's "running" native packages so I don't see what use the information would be... Feb 21 19:29:36 cpp-native Feb 21 19:29:50 JustinP: We can then device whether we should use the system's versions of these tools or build our own Feb 21 19:29:55 I suppose....but that's provided by gcc-cross Feb 21 19:30:01 hmmm Feb 21 19:30:19 s/device/decide Feb 21 19:31:29 In theory we'll end up with a list of our prereq. software in ASSUME_PROVIDED Feb 21 19:35:01 I'll try to fix this later... Feb 21 19:35:12 JustinP: I just wanted to make you aware of it Feb 21 19:36:09 zecke: at your service Feb 21 19:36:23 pb_: wow, with batteries included? Feb 21 19:36:37 pb_: did you form an opinion on git vs. svn+svk already? Feb 21 19:36:54 zecke: I didn't actually try svn+svk yet. I guess I should give that a go. Feb 21 19:37:07 everyone seems to say it's very slow, though, which doesn't sound encouraging Feb 21 19:37:15 pb_: to give you a summary. git is the fastest :) Feb 21 19:37:30 yeah, it seems Feb 21 19:37:34 pb_: svk is replaying every commit one by one - so yes it is slow Feb 21 19:37:44 and, much as I fear git, it does seem to be the most attractive option at the moment Feb 21 19:37:54 pb_: issue with git: The only secure usage is using the git server Feb 21 19:38:35 pb_: and it has a custom protocol. I do not know if it will be changed (likely), I don't know if it is firewall friendly (unlikely), If the server is secure (unlikely) :) Feb 21 19:38:50 ah, you can't use git over ssh? Feb 21 19:38:54 that is a bit unfortunate Feb 21 19:40:02 pb_: well packing is racy with ssh,rsync,http Feb 21 19:40:22 pb_: you get the index, then soneone repacks, and you want to get a file and it vanished (as it is now packed) Feb 21 19:41:01 oh right, hm Feb 21 19:41:05 under what circumstances would you repack? Feb 21 19:41:32 pb_: I don't know if this happens automatically so I dunno Feb 21 19:41:45 pb_: I can only say it happened when I git pulled linus' tree Feb 21 19:41:57 if it's a rare maintenance operation then I guess you would just need to disable remote access during packing Feb 21 19:42:12 or, alternatively, patch git to do some kind of repository locking. doesn't seem like it can be that hard. Feb 21 19:43:15 pb_: right, git server is one solution, proper locking is another... Feb 21 19:43:54 pb_: well - the http case will always have a locking issue :) Feb 21 19:44:07 pb_: (violation of keithp's Feb 21 19:44:10 sorry Feb 21 19:44:22 packing is a violation of keithp's requirement :) Feb 21 19:45:19 yeah Feb 21 19:45:26 let's try to find out what causes packing to happen Feb 21 19:46:05 pb_: git repack, I think on ssh+push we will locally pack and unpack on the server Feb 21 19:46:26 pb_: so I would guess (slightly educated) that git push will not automatically pack things Feb 21 19:50:50 repacking the tree is a manual operation Feb 21 19:51:01 gitrepack doesn't cause a problem Feb 21 19:51:22 its only when you remove the files that you placed in the packs that you can upset people's transfers Feb 21 19:51:35 so prune* is guilty? Feb 21 19:51:52 zecke: git-prune-pack specifically Feb 21 19:52:05 hehe I didn't rember that name :) Feb 21 19:52:27 okay and this could be an operation we do once in a month? Feb 21 19:52:33 yes Feb 21 19:53:07 a repull will sort itself out so any user seeing an issue can just rerun the command that failed Feb 21 19:53:08 okay then this is not an issue at all (for us) I think Feb 21 19:53:28 RP: I'm thinking of automatic scripts that need a while_git loop :) Feb 21 19:53:29 and as long as we have an agreement over when the main repository would be repacked, it shouldn't impact automation Feb 21 19:54:04 i.e. the automation could know not to do things between certain times, or could check for a lock etc. Feb 21 19:54:07 RP: agreement on when to prune right? Feb 21 19:54:13 zecke: yes Feb 21 19:54:39 kernel.org prune extremely regularly. With our volume of changes, we shouldn't need to Feb 21 19:54:58 kernel.org do it as any tiny bandwidth saving is valuable to them Feb 21 19:55:37 and we are not dependant on git server - nice Feb 21 19:55:49 So we do not have a firewall issue Feb 21 19:56:20 correct Feb 21 19:57:14 RP: okay then git is slowly winning against svn+svk (from my p.o.v) Feb 21 19:57:26 RP: even if I think that all cli's are more mature than git ones Feb 21 19:57:41 git will be a love and hate relationship Feb 21 19:58:20 indeed. Its cli is insane :) Feb 21 19:58:41 I have grown to love git, even though it routinely abuses me Feb 21 19:59:14 I have similar feelings :) Feb 21 19:59:23 RP: e.g. it should be easy to write a 'meld' resolver Feb 21 19:59:39 but the thing is so underdocumented I didn't even bother to try it :} Feb 21 19:59:45 stgit is very useful though Feb 21 19:59:50 zecke: Try the latest git as I think they might have made merges easier Feb 21 19:59:52 worth the pain Feb 21 19:59:58 stgit? Feb 21 20:00:24 RP: it is quilt for git Feb 21 20:00:39 ah, yes. I must try that :) Feb 21 20:00:48 http://www.procode.org/stgit/ Feb 21 20:01:09 It has some cool features - I like how it can seemlessly handle empty patches Feb 21 20:01:38 I know I heavily abuse quilt so I suspect I might find it useful Feb 21 20:02:11 Half the time I manually edit quilts files rather than run the commands :-/ Feb 21 20:02:22 The only thing I'm really annoyed about is that I sometimes forget which command to use Feb 21 20:02:35 <_chronic> i work with cl who wrote svk. should i try to get him on this channel? Feb 21 20:02:42 I hate it's gnu'isms :) Feb 21 20:03:07 _chronic: hehe nice to know, I have no issues with svk :) Feb 21 20:03:28 _chronic: I do not think it can be made faster :) Feb 21 20:03:32 <_chronic> i tried it briefly, but i ain't no scm expert ... Feb 21 20:03:50 * RP has never actually used svk... Feb 21 20:04:04 RP: it is very robust :) Feb 21 20:04:28 RP: and the merging is mature - Only it doesn't tell me which file is left and which is right Feb 21 20:04:41 * france is away: Away Feb 21 20:08:13 <_chronic> ok. just sent cl a message, don't want to help him flag wave, but i trust he would give good input none-the-less. Feb 21 20:09:40 <_chronic> invited him to join this channel, if he wishes. Feb 21 20:12:51 zecke: Checking ignored_dependencies against the rdepends list appears to work. I'll check it in? Feb 21 20:13:30 RP: let me take a look at the current code :) Feb 21 20:13:56 yes do so :) Feb 21 20:15:23 RP: http://ewi546.ewi.utwente.nl/OE/depends3.log this is the current list Feb 21 20:16:37 I have a patch for quilt if that helps start the ball rolling? Feb 21 20:17:05 Im still at a*.bb Feb 21 20:19:56 ok, bitbake updated Feb 21 20:21:33 hi, i was informed that someone mentioned svk? :) Feb 21 20:21:48 clkao: hey and hehe Feb 21 20:22:12 clkao: OE is currently the largest owner of monotone, and we have started test migrations Feb 21 20:22:31 clkao: one candidate is svn+svk and besides it is really mature I have two questions Feb 21 20:23:48 clkao: one wish: when smerge triggers manual merging it would be nice to be able to see from which repo the to be merged files are coming Feb 21 20:24:07 clkao: and the second question is, if the svk sync can be speed up somehow Feb 21 20:24:31 clkao: besides that I really like it, and my university assignment directory is under svk control Feb 21 20:25:07 svk sync will be speeded up soon, but it requires 1.4 on the server side. there's a new ra api doing proper reply, finally. Feb 21 20:25:38 clkao: so it will become a file copy only? Feb 21 20:26:06 no, but it will be less expensive in many cases Feb 21 20:26:14 however Feb 21 20:26:39 http://oe.handhelds.org/cgi-bin/moin.cgi/SvnScmTrial Feb 21 20:26:45 most of the people (pugs, parrot, etc) is using the bootstrap technique. so people can just download a boostrap dump and load it. rather than syncing from scratch Feb 21 20:27:02 http://svk.elixus.org/?SVKBootStrap Feb 21 20:27:28 <_chronic> sounds like some of the things done with oe and monotone recently .... Feb 21 20:27:29 clkao: okay, how reliable is signing of changes? Is it using GnuPG? Feb 21 20:28:08 yes. but it's just signing the list of files with md5 sum with gpg. md5 might not be so secure Feb 21 20:28:32 but no one has asked us to add sha1 support etc yet. so. Feb 21 20:29:42 clkao: skip SHA1 and go to the next generation :) Feb 21 20:30:18 clkao: besides that I'm still making up my mind if I prefer git or svn+svk Feb 21 20:31:14 what are the main annoyance of each? Feb 21 20:32:04 wow, svk also has the initial-pull-is-slow bug? I guess I feel a little better... ;-) Feb 21 20:32:05 clkao: git, the cli is insane, and it might be moving too fast (hacky). I'm scared to death if I need to merge files with git on OS X Feb 21 20:32:32 njs: hey, monotone is a couple of days slower :) Feb 21 20:32:34 heh Feb 21 20:32:42 zecke: oh, well, fair enough... Feb 21 20:32:47 njs: not really - but it's a nature thing when you get a large repository Feb 21 20:32:57 njs: but monotone 0.26 is way faster already Feb 21 20:33:48 zecke: yeah, and I know the guy who's playing with fixing the next round of bottlenecks had at least one prototype that pulled 3x faster than current 0.26 Feb 21 20:33:52 zecke: but, we shall see! Feb 21 20:33:58 clkao: and for svk. I'm not sure I like it very much. smerge was reliable, killing pull's, smerge's didn't lead to inconsistencies Feb 21 20:34:36 and you don't like a version control system being consistent? Feb 21 20:34:38 clkao: so I guess the main issue with svk is the slower initial sync :) Feb 21 20:34:58 using bootstrap should save lots of people's arse. gcc people are using that as well Feb 21 20:35:01 clkao: as I said I had no issue with svk yet Feb 21 20:35:07 oh ok :) Feb 21 20:35:35 they have insanely huge repository Feb 21 20:35:48 clkao: I had one issue with a cvs2svn converted repo that had a bogus branch but that can't be blamed on svk Feb 21 20:36:58 njs: will the double root handling land in 0.26? Feb 21 20:37:05 zecke: yes Feb 21 20:37:32 zecke: we just didn't want to hold up the testing of 99% of the code, because 1% of it wasn't finished yet :-) Feb 21 20:37:55 njs: good, I think it is no news but it successfully converted the OE repo (well I didn't verify the conversion to closely) Feb 21 20:38:10 yeah Feb 21 20:38:18 njs: I like the heavy use of invariants and such, I wish more Free Software projects would use them Feb 21 20:38:25 me too... Feb 21 20:38:38 hi njs Feb 21 20:38:42 though it's a bit of a pain in the butt. Feb 21 20:38:43 morning Feb 21 20:38:48 hrw: morning Feb 21 20:38:56 invariants? Feb 21 20:39:04 hmm, since I'm out anyway, I might as well mention for the people who haven't met me before, I'm the main monotone cat-herder :-) Feb 21 20:39:12 clkao: do you think it would be possible to call stuff like meld/kdiff3 for doing the merge? Feb 21 20:39:20 (been lurking here the last day or two to try and get a better idea what's been going wrong) Feb 21 20:39:46 clkao: ~1 in 9 lines in monotone code is for detecting or diagnosing bugs Feb 21 20:39:59 zecke: they are supported amongst the 13 tools that would be autodetect4ed and prompt you for launching Feb 21 20:40:03 (that collapses assertions, debug logging, data tracing, etc.) Feb 21 20:40:49 clkao: ah okay I do not have kdiff3 in my path :) Feb 21 20:40:49 it's like 1 in 20 for assertions alone, or something like that Feb 21 20:41:13 so just answer 'm' and it should list available, recognised merge tools Feb 21 20:41:33 defensive programming good Feb 21 20:41:55 how big is oe repository? Feb 21 20:41:58 yeah Feb 21 20:42:26 anyway, have to run to class, if anyone wants to rant about monotone's suckiness at me, please do :-) Feb 21 20:42:46 njs: will do later, hopefully with a good result :) Feb 21 20:42:53 njs: are you now in the complaint-driven-development camp as well? :D Feb 21 20:43:11 hehe Feb 21 20:43:56 clkao: zecke@bitbake:~/scm_convert/svnrepo/db/revs$ ls -la | wc -l Feb 21 20:43:57 3656 Feb 21 20:44:07 zecke@bitbake:~/scm_convert/svnrepo/db/revs$ du -m Feb 21 20:44:07 262 . Feb 21 20:44:49 clkao: dunno, I just figure, why would I be spending years of my life trying to help people work together better, if I didn't care about how it worked for them? :-) Feb 21 20:44:52 ciao Feb 21 20:45:39 njs: cya later Feb 21 20:46:27 clkao: OE might not end up using svn+svk but regardles you have written a very important tool Feb 21 20:46:41 clkao: I use it daily! Feb 21 20:46:59 * kergoth too Feb 21 20:47:05 * clkao bows Feb 21 20:47:27 svk sync -a /oe Feb 21 20:47:30 svk sync -a /road Feb 21 20:47:32 svk sync -a /uni Feb 21 20:47:33 :) Feb 21 20:47:46 kergoth: hi Feb 21 20:48:02 hey Feb 21 20:48:15 zecke: your uni is under revision control? how futuristic. Feb 21 20:48:17 clkao: is it important that I inform svk that I have deleted a co? Feb 21 20:48:19 g'day kergoth Feb 21 20:48:24 pb_: some folders :) Feb 21 20:48:35 03freyther 07org.oe.dev * r80577fdb... 10/packages/ (6 files in 3 dirs): (log message trimmed) Feb 21 20:48:35 packages/antlr/antlr-native: Feb 21 20:48:35 -Fix the RDEPENDS to depend on a native virtual Feb 21 20:48:35 machine. Feb 21 20:48:35 packages/autoconf/autoconf-native_2.59.bb: Feb 21 20:48:36 -Change the RDEPENDS Feb 21 20:48:37 pb_: basicly all my assignments are under svn/svk control Feb 21 20:48:37 no. you can do svk co --purge if you like, though Feb 21 20:48:38 packages/automake/automake-native*: Feb 21 20:48:48 i should use more than one depotmap. i tend to get lazy and just throw everything in //, unless its a depot for a local svn file:// path Feb 21 20:49:16 clkao: will I have performance issues if I don't do --purge ones in a while? Feb 21 20:49:48 RP: feel free to push your quilt change, maybe it already builds again Feb 21 20:50:29 it can be. but from profiling that part is usually less than 5% Feb 21 20:55:54 * RP ponders why ./tmp/work/tmp/work exists. Its no wonder the build wasn't working... Feb 21 21:02:18 kergoth: how goes dns things? Feb 21 21:05:30 hrw: paid, waiting for them to get me back going again Feb 21 21:05:57 hmm someone backed out everything from me :} Feb 21 21:07:27 ~lart matchbox-wm as usual Feb 21 21:07:27 * ibot accelerates a free AOL cd to 50,000 rpm and lets matchbox-wm as usual feel it Feb 21 21:13:54 kergoth: good, and I have a paypal account to plunder.. we can spend roughly ~300$ on OE stuff Feb 21 21:14:17 03rpurdie 07org.oe.dev * rcbd8d3ca... 10/packages/quilt/ (quilt-native_0.39.bb quilt-native_0.42.bb): quilt-native: Set RDEPENDS correctly Feb 21 21:14:21 03rpurdie 07org.oe.dev * r799e136c... 10/conf/bitbake.conf: bitbake.conf: Add some new ASSUME_PROVIDED targets Feb 21 21:15:20 RP: what about perl-native? Feb 21 21:15:48 looks like I should wait few weeks before start to use .dev again ;) Feb 21 21:16:32 hrw: no no, it will be sane soon (TM) and then I will mail a contract to oeAThh.org that I promise to improve the quality Feb 21 21:16:37 zecke: Does perl require its native counterpart in staging or anything like that? Feb 21 21:16:55 hrw: It will work as long as you have a current bitbake Feb 21 21:16:57 RP: automake claims RDEPENDS on perl :) Feb 21 21:17:28 zecke: I had a problem with python - it needs python-native Feb 21 21:18:39 zecke: So if I add ASSUME_PROVIDED = "python-native", we break python :-/ Feb 21 21:19:27 RP: then we should name these things *-required Feb 21 21:19:42 and leave -native for build tools? dunno on how to do the separation Feb 21 21:20:18 RP: http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=4&logfile=20060221222502.log Feb 21 21:21:42 zecke: or make python-native a special case and rename it? Feb 21 21:21:55 I'm not sure what to though... :-/ Feb 21 21:26:54 wait, why would you want to rename it? Feb 21 21:28:12 kergoth: I think bitbake requires python but our python packages Feb 21 21:28:23 kergoth: require a special python version? IIRC Feb 21 21:29:11 http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=4&logfile=20060221223356.log Feb 21 21:29:19 okay this is the log of the active build Feb 21 21:30:36 only python should depend on python-native. if you're using bitbake, your machine already has python, so i dont see a need for it, do you? Feb 21 21:30:53 I agree with you Feb 21 21:31:06 so just kill off the unnecessary deps and call it a day Feb 21 21:31:09 I might overlook something... Feb 21 21:31:11 * kergoth needs to play with oe some more nowadays Feb 21 21:31:35 our happy oe site files have saved my ass a number of times crosscompiling tools at work :) Feb 21 21:32:19 kergoth: If someone rewrites bitbake in C, that argument becomes flawed. Admittedly half of OE is python as well mind :) Feb 21 21:32:44 chicken depends on egg Feb 21 21:32:46 oh wait Feb 21 21:33:01 RP: Well if you would use PyPy you would end up having Java, C, C++, C# code :) Feb 21 21:33:22 RP: http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=73ffcb18024618ab0eefb4fa030b0d95485544d1 this needs to be disapproved as well Feb 21 21:34:21 kergoth: In theory, as we depend on a provided python-native, we should be able to set ASSUME_PROVIDED = "python-native" but as python needs python-native, we can't. Its probably best resolved by just not referencing python-native anywhere but it is an inconsitency in the metadata... Feb 21 21:34:28 RP: well, in that case you'd want version specific dependencies on the native python for the python package, and do an assume_provided of the version installed on the machine, and make the other python packages just depends on python-native without the version requirement Feb 21 21:34:35 cu - have a good night/day/whatever you have in next hours Feb 21 21:34:42 coursey ou'd need version specific assume_provided for that Feb 21 21:35:25 python-oe-native, python-native, python ;) Feb 21 21:37:04 kergoth: and depends that handles versioing which we don't have :) Feb 21 21:37:46 03freyther 07org.oe.dev * r10801c13... 10/classes/tinderclient.bbclass: Feb 21 21:37:46 classes/tinderclient.bbclass: Feb 21 21:37:46 Redo the changes... they were magically lost. Feb 21 21:37:46 readlines is a sequence, we join them (we might Feb 21 21:37:46 need to remove the \n) Feb 21 21:40:29 wb pb_ Feb 21 21:40:34 re Feb 21 21:40:55 hm, I seem to be in ipv4 world all of a sudden Feb 21 21:40:57 how very backwards Feb 21 21:42:01 hehe Feb 21 21:42:08 you have been downgraded Feb 21 21:42:18 it seems Feb 21 21:50:43 RP: http://ewi546.ewi.utwente.nl/OE/depends4.log <- this is the current output of depends_checker Feb 21 21:54:45 RP: heh, yeah, that too Feb 21 22:03:16 zecke: The native packages needing update-rc.d are scary :) (can just be removed) Feb 21 22:03:47 its looking much better though Feb 21 22:09:36 RP: let us hope init.d/PN restart is not called :) Feb 21 22:15:33 and I'm happy with the tinderbox :) Feb 21 22:16:00 I like the logview, now it should summarize the issues it finds (gcc warnings etc...) Feb 21 22:25:41 openembedded.org and openzaurus.org are back Feb 21 22:25:44 fyi Feb 21 22:25:53 sorry for the delay, been busy Feb 21 22:26:31 yay! Feb 21 22:26:48 now I can pimp OE again ;) Feb 21 22:28:34 'night all Feb 21 22:29:10 * Philippe is away: visual contact - melancholic dreams Feb 21 22:33:42 kergoth: Thanks :) Feb 21 22:34:02 np Feb 21 22:34:16 easydns is raping me. $70 for two domains for a year for registration + dns Feb 21 22:34:20 need to transfer asap Feb 21 22:34:45 ouch Feb 21 22:35:00 I guess you had to renew before transfering Feb 21 22:35:40 yeah Feb 21 22:35:58 if the hosing is over DNS, use zoneedit for DNS Feb 21 22:38:39 i'll just move them over to godaddy for both registrar and dns. have kergoth.com there, works well, cheap Feb 21 22:39:49 hi Feb 21 22:40:56 kergoth: You said this last year so transfer them now ;-) Feb 21 22:42:17 haha, yeah, i did Feb 21 22:42:20 forgot about that Feb 21 22:50:24 kergoth: could you do a s/coredump/CoreDump|home in #openzaurus' access list please? The nick coredump isn't registered to me heh Feb 21 22:50:36 heh, k Feb 21 22:50:46 thanks Feb 21 22:52:39 CoreDump|home: done Feb 21 22:52:46 thanks =) Feb 21 23:03:13 RP: it is still building - building that shitty openzaurus 2.6 kernel ;) Feb 21 23:05:49 ~lart zecke Feb 21 23:05:49 * ibot whips out his power stapler and staples zecke's foot to the floor Feb 21 23:05:54 * CoreDump|home hugs 2.6 Feb 21 23:05:54 hehe :) Feb 21 23:06:01 =) Feb 21 23:06:20 no doubt RP is doing a fantastic job - from any p.o.v. Feb 21 23:09:29 * CoreDump|home rogers that Feb 21 23:15:40 rwhitby: hey Feb 21 23:17:04 rwhitby: if you still do regular builds put http://oe.pastebin.com/565916 Feb 21 23:17:15 into your local.conf, we have a tinderbox again Feb 21 23:17:49 set CLOBBER = "1" if you do not rebuild from scratch Feb 21 23:18:01 zecke: yep, we still do regular builds. every half hour Feb 21 23:18:16 (incremental, not from scratch) Feb 21 23:18:40 would love to see your reports soon :) Feb 21 23:18:51 http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=OpenEmbeddedBuilds Feb 21 23:19:00 this version of the tbox looks more reliable Feb 21 23:19:35 can I change TINDER_MACHINE? i.e. we have Unslung, OpenSlug, UcSlugC and DebianSlug builds now Feb 21 23:20:08 rwhitby: yes, this determines the header of the column Feb 21 23:20:19 you could place your viagra ads there if you want to :) Feb 21 23:20:23 ok, so I can just use those names? Feb 21 23:20:32 yes Feb 21 23:20:50 I'll set it up now Feb 21 23:20:52 I think we can use spaces as well :) Feb 21 23:20:57 wow thanks :) Feb 21 23:21:07 I'm addicted to view the compile log :) Feb 21 23:21:16 is timezones fixed now? Feb 21 23:21:55 it is completely different code, and yes it doesn't suffer the timezone issue Feb 21 23:22:12 it doesn't suffer database corruption as it is using postgres Feb 21 23:22:28 it has less features though and some other issues but I'm fixing these one by one Feb 21 23:26:29 ok, conf files are in, and starting an autobuild run now Feb 21 23:26:46 let us hope it is working :) Feb 21 23:27:38 hmm - three way merge of tinderclient.bbclass Feb 21 23:28:22 hehe, be careful *please*, these fixes were reverted already Feb 21 23:28:57 we never to *any* merges on the autobuild machine. Feb 21 23:29:13 so you have to fix it at your end, and then push a single head to monotone.nslu2-linux.org Feb 21 23:29:53 rwhitby: hmm we only have one head Feb 21 23:30:26 hmm Feb 21 23:30:41 I'm pulling to my home machine, this can take some time Feb 21 23:37:09 rwhitby: hmm you have an old version of the tinderclient.bbclass Feb 21 23:37:19 gosh who is playing ping pong with my tinderclient :} Feb 21 23:37:39 zecke: ok, monotone.nslu2-linux.org has a single head, it's just the autobuild machine with multiple heads Feb 21 23:37:53 so I'll blow that one away and just get the db from the main server Feb 21 23:37:59 (no merging) Feb 21 23:38:51 hmm both have an outdatet tinderclient.bbclass Feb 21 23:39:00 http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=6&logfile=20060222004339.log Feb 21 23:39:16 no new lines and only '<--' and not '<---' in the separation Feb 21 23:40:11 http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=10801c1345a1ed51c3f9e884c31ac5a7073880d6 Feb 21 23:41:46 yep, the next autobuild will be using that one after the merge problems disappear Feb 21 23:46:15 nice :) Feb 21 23:49:58 * france is back (gone 03:45:20) Feb 21 23:53:04 zecke: starting new autobuild with a fresh checkout of openembedded Feb 21 23:54:04 monotone: already up to date at 10801c1345a1ed51c3f9e884c31ac5a7073880d6 Feb 21 23:55:34 hmm Feb 21 23:56:06 bitbake starting now Feb 21 23:59:25 rwhitby: okay it is looking better now, and there is some bug with my tinderclient. It doesn't report success Feb 22 00:01:11 getting lots of tinderbox server responses in the build log too Feb 22 00:01:36 rwhitby: is it too noisy already? Feb 22 00:01:52 yes Feb 22 00:02:24 11 lines of server response to every bitbake NOTE: line Feb 22 00:02:31 okay, will fix this soon then Feb 22 00:04:45 hi, i'm newbie on linux/gnu tools and i need some guidance Feb 22 00:05:23 plese be pacients, what's the difference between scratchbox and openembedded? Feb 22 00:09:12 linitrofe: I'm not too familiar with scratchbox so this could be wrong Feb 22 00:09:39 linitrofe: scratchbox is doing native compilation using tools like qemu and real devices to run configure and such Feb 22 00:09:54 linitrofe: OpenEmbedded can be used to do native and cross compilation Feb 22 00:10:09 linitrofe: scratcbox can only produce debian packages, OpenEmbedded is way more flexible Feb 22 00:10:30 linitrofe: we can generate tar.gz, ipk, deb, rpm and it is easy to add other systems as well Feb 22 00:11:04 linitrofe: OpenEmbedded is flexible, it is easy to build Distributions with OE and truely customozing the content/shipped packages and also the configuration of these packages Feb 22 00:11:56 OpenEmbedded can generate all kind of filesystems, jffs2, ext2 and can be customized Feb 22 00:13:29 ok, thanks for that, i'll focus on OE Feb 22 00:13:52 on the main page said you are a opie hacker Feb 22 00:14:02 linitrofe: OE has many downsides as well Feb 22 00:14:17 linitrofe: the lack of good documentation is surely one - so you need the help of this channel Feb 22 00:14:26 03freyther 07org.oe.dev * r9bb97623... 10/classes/tinderclient.bbclass: Feb 22 00:14:26 classes/tinderclient.bbclass: Feb 22 00:14:26 Be less verbose do not print the results for now. Feb 22 00:14:26 A proper fix/configurable one is pending. Feb 22 00:14:53 rwhitby: it should be less verbose now Feb 22 00:15:00 if i had an arm cpu running a linux kernel and a VGA driver on the board... can i get running opie without many workaround? Feb 22 00:15:25 (scratchbox has the same problems as i see at this moment) Feb 22 00:16:07 linitrofe: arm is well supported/tested in OE Feb 22 00:16:10 linitrofe: it shouldn't be too dificult Feb 22 00:17:22 perfect... i'll read the documentation on OE... the OE developers are all the time here if i need help? Feb 22 00:17:39 There is usually someone around Feb 22 00:17:40 linitrofe: most of us live in Europe Feb 22 00:17:59 linitrofe: some insane people are awake at insane hours Feb 22 00:18:08 documentation is something we know we need to improve as there's not a lot of it... Feb 22 00:18:09 wow Feb 22 00:18:36 zecke: you and me, insane? figures :) Feb 22 00:19:08 yes we take documentation and other quality things seriously - but we need to improve Feb 22 00:19:23 RP: hehe I have an excuse. Today I'm worried Feb 22 00:20:12 thanks for your support... i work with uC for many years, now i bought an arm HW and i have experience with linux/gnu, but not with embedded OS ans tools so i'm investigating the bases Feb 22 00:21:17 (sorry for my english) Feb 22 00:21:31 I hope to be able to answer your questions soon, and maybe you have some super cow powers at writing documentation Feb 22 00:21:56 aaaaaaaahahahhaha Feb 22 00:22:19 emmmm... my tesis teacher hopes the same thing Feb 22 00:23:03 linitrofe: within 7 hours today, I wrote about 40 lines of text Feb 22 00:23:19 you are my man! Feb 22 00:24:01 it's so easy write code for hours... but that damned tesis... Feb 22 00:25:11 linitrofe: I have to finish a paper until Friday/Tuesday and everything I write sounds like crap Feb 22 00:26:36 xD Feb 22 00:33:46 rwhitby: did you stop something? did I break something? Feb 22 00:34:54 'night all Feb 22 00:35:01 RP: cya Feb 22 00:35:26 RP: I think it works now (again) Feb 22 00:37:00 zecke: ? Feb 22 00:37:19 http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=OpenEmbeddedBuilds Feb 22 00:37:40 I'm not seeing new builds, but I guess you sleep some time before rebuilding? Feb 22 00:37:57 the autobuilds restart every 30 mins Feb 22 00:38:18 ah good to know, it should be less verbose by now Feb 22 00:38:33 ah, actually, every 15 minutes Feb 22 00:38:41 should be another starting in 5 Feb 22 00:38:58 I can kick it off early Feb 22 00:39:25 no no :) Feb 22 00:39:43 I was simply scared that my untested commit+push didn't work Feb 22 00:46:09 oE uses bitbake or monotone? Feb 22 00:46:15 it's not very clear on the wiki Feb 22 00:46:56 linitrofe: both Feb 22 00:47:04 linitrofe: monotone is a SCM tool Feb 22 00:47:39 linitrofe: bitbake is a task executor to execute the 'MetaData' of OpenEmbedded. Well the MetaData are descriptions on how to build package XYZ Feb 22 00:48:08 like .deb on debian? Feb 22 00:48:16 linitrofe: e.g. 1.) get the source 2.) unpack it 3.) patch it 4.) do some configuration 5.)... Feb 22 00:48:43 ok, like gentoo (as mentionated on the page) Feb 22 00:48:56 some parts are like the debian/rules script somes are like gentoo Feb 22 00:52:03 i read what means SCM on wikipedia... it's like subversion? Feb 22 00:52:49 linitrofe: yes, but it is distributed so it is more like svk Feb 22 00:54:09 ok Feb 22 00:56:51 i'm getting the db... i'm going to check my girlfriend cat (push a pill inside her mouth and get scratched) Feb 22 00:57:09 I'm off to bed now :} Feb 22 00:57:23 ok, thanks for your help... see you tomorrow Feb 22 00:57:34 yeah cya later Feb 22 01:02:33 rwhitby: okay now builds should turn green as well :) Feb 22 01:03:00 03freyther 07org.oe.dev * rbae5019e... 10/classes/tinderclient.bbclass: Feb 22 01:03:00 classes/tinderclient.bbclass: Feb 22 01:03:00 -Have a none empty log messages when the build is Feb 22 01:03:00 completed otherwise the log is not transfered Feb 22 01:03:00 to the server. Feb 22 01:03:00 Now the build turns green :) Feb 22 03:31:29 RP: i should have a serial cable by tomorrow Feb 22 04:23:26 * france is away: Away Feb 22 05:05:17 ~lart Feb 22 05:23:50 koen|sleep: mickey|sports: i have sent to you both an email, persuant to requesting commit access in OE Feb 22 05:26:00 hey shadows Feb 22 05:26:29 you join Ångström yet? Feb 22 05:27:50 http://handhelds.org/hypermail/angstrom-dev/0/0024.html Feb 22 05:28:07 emte: i should do that, thanks Feb 22 05:28:45 shadows, wasnt sure if koen had finally talked to you or not ... the gcc4.x part should be significantly interesting to you Feb 22 05:29:05 yes Feb 22 05:29:22 i'm confused about what exactly Angstrom is Feb 22 05:29:42 and how it applies to my target hardware (Sharp Zaurus C3000, soon also C3100) Feb 22 05:29:47 http://handhelds.org/hypermail/angstrom-dev/0/index.html#2 Feb 22 05:30:18 generic OS for handhelds, or what is the idea Feb 22 05:30:22 it is a distro Feb 22 05:30:31 i'm reading, and it's like presuming something already that i don't know about Feb 22 05:30:36 well by the looks it will be more then that Feb 22 05:30:53 http://handhelds.org/hypermail/angstrom-dev/0/0000.html <- first post of concept Feb 22 05:32:49 yes i still fail to grok what the point of Angstrom is Feb 22 05:32:54 believe me, i'm reading it Feb 22 05:35:50 there's an explicit reason we have nslu2-linux, familiar, and openzaurus distros Feb 22 05:35:58 the distros have different goals Feb 22 05:36:13 yes and no Feb 22 05:37:06 as far as i understand Ångström is a much larger picture and closer to how Debian, Gentoo, etc are run Feb 22 05:37:13 hmm Feb 22 05:37:24 shadows: it was decided a while ago that we don't have the resources to maintain multiple different distros when we're all basically doing the same things Feb 22 05:37:33 JustinP: i agree to that Feb 22 05:37:51 shadows: it makes far more sense to make things more general and have machine-based images or builds Feb 22 05:38:00 shadows: as emte said, like Debian/Gentoo/etc. Feb 22 05:38:01 though, tipping my hat to a line from hitchhikers, i think "Angstrom" is not a solution, it is a different name for the problem Feb 22 05:38:08 all the distros in OE dont really have different goals, they have different hardware Feb 22 05:38:34 we're all using the same metadata and all, the only difference really is how things are built and what's installed on different machines Feb 22 05:39:23 this was we have closer cooperation between all of the developers, less overlap, and a way to centralize things like bugs Feb 22 05:39:37 heh, I would come in just now Feb 22 05:39:42 :) Feb 22 05:39:43 centralizing the bugtracker would be great Feb 22 05:40:22 * JustinP needs to simplify efl bbs.... Feb 22 05:40:23 it is/was/kinda sorta Feb 22 05:40:26 more .inc files Feb 22 05:40:42 ach Feb 22 05:40:57 I also want my remote to work....darn max1111 isn't giving me the values I need... Feb 22 05:40:59 there is no clear definitions about what should be on which bugtracker Feb 22 05:41:07 that's also a problem, yes Feb 22 05:41:15 each distro has one as does OE Feb 22 05:41:21 for good reasons Feb 22 05:41:33 but the problem is where do some bugs belong Feb 22 05:41:41 yes Feb 22 05:41:46 and users really don't know Feb 22 05:41:46 you can readily split embedded devices into sectors based on their user interface Feb 22 05:42:11 not really Feb 22 05:42:20 some are web interface, others full touchscreen Opie, or Gtk+ and matchbox WM Feb 22 05:42:34 that is how the device interacts with a human Feb 22 05:42:53 so where do you put devices that fall into all categories? Feb 22 05:42:54 but you can run X prog exported from an nslu2 if you want Feb 22 05:43:08 and I can run a web interface in my Z if I want Feb 22 05:43:18 yes of course Feb 22 05:43:24 you can do many things Feb 22 05:43:35 most devices will run all the interfaces in some state of usability Feb 22 05:43:37 which one are you likely to do in most cases of day-to-day use? Feb 22 05:43:59 familair always has two choices currently Feb 22 05:44:03 tho you could have 4 Feb 22 05:44:07 for an PDA target, the device needs to suspend and resume properly, and provide useful scheduler software Feb 22 05:44:26 for a portable laptop replacement, the device needs to be easily flexible to run common desktop software Feb 22 05:44:27 suspend/resume are not GUI Feb 22 05:44:31 yes they are Feb 22 05:44:44 you push a button on the interface (physical) and the device reacts Feb 22 05:44:46 no, they are physical UI (kernel) Feb 22 05:45:05 i consider that to be integral to a PDA device's function Feb 22 05:45:07 you dont need a GUI to suspend a device Feb 22 05:45:16 thats just semantics Feb 22 05:45:24 mmhm Feb 22 05:45:37 but you're right, emte, shadows made it too simple saying it was GUIs Feb 22 05:45:38 shadows, i am curious, are you aware not all the devices in OE are pdas? Feb 22 05:45:39 i said the user interface, not GUI Feb 22 05:45:40 it's classes of devices Feb 22 05:45:47 ok, then Feb 22 05:45:56 how does this device interact with the user Feb 22 05:46:04 meshcube is a good example of a nonpda Feb 22 05:46:09 sure Feb 22 05:46:13 well, nslu2 as well Feb 22 05:46:20 otherwise there would be gabillion mhz CPUs in everything (okay maybe not!?) Feb 22 05:46:41 yeah Feb 22 05:46:52 * emte would still like a meshcube Feb 22 05:46:58 http://www.meshcube.org/ Feb 22 05:47:11 if the device is going to be specific to say, routing network traffic, why would you need dual CPU support or X11 support? Feb 22 05:47:38 * emte wonders how much shadows knows about network hw ... Feb 22 05:47:40 OpenBSD have this issue, the emphasis is rightly on security versus support for new hardware with SMP Feb 22 05:48:05 there are MANY routers and switches with multiple CPUs Feb 22 05:48:06 FreeBSD, arguably not as secure, does support better the SMP, but it doesn't actually get used Feb 22 05:49:01 i'm bringing up the issue of what is Angstrom going to be Feb 22 05:49:02 anyway this is getting carried away Feb 22 05:49:24 devices cannot be classified by UI Feb 22 05:49:27 if you want it to do everything, it will end up like Slackware or Gentoo Feb 22 05:49:51 emte: i think i have provided examples that devices can and should be classified by the user interface Feb 22 05:49:58 * JustinP loves Gentoo Feb 22 05:50:01 how the human interacts with the hardware via software Feb 22 05:50:06 or physical interface Feb 22 05:50:10 shadows, and i provided an equal amount that violate your definitiona Feb 22 05:50:21 no? Feb 22 05:50:29 you provided examples that fit my description Feb 22 05:50:41 such as? Feb 22 05:50:53 the meshcube and nslu2-linux distros/related hardware interact with humans differently than PDA devices Feb 22 05:50:54 emte: he was talking about UI, not GUI Feb 22 05:51:31 i am aware Feb 22 05:51:52 yes that is the user interface Feb 22 05:51:56 now i am curious where he places a generic x86 or arm board Feb 22 05:52:18 the software would be different, depending on the intended use Feb 22 05:52:49 so you are back to the OE model Feb 22 05:53:37 for the same hardware, i would think that picking a different "distro" in OE will yield the applicable software distro to the use of the hardware Feb 22 05:54:28 bootstrap vs opie vs gpe vs e vs netappliance Feb 22 05:54:33 ok, time for me to go to bed Feb 22 05:54:34 use a Geode AMD board as an car GPS + music platform, your software configuration and user interface will certainly be different than using a Geode AMD board as a traffic router Feb 22 05:54:35 night Feb 22 05:54:39 'nite JP Feb 22 05:54:52 correct Feb 22 05:55:36 yes. therefore you can classify the emedded device into sectors based on the user interface Feb 22 05:56:33 when i said "split" i meant split up with the possibility a device may fall into multiple sectors, which does not defeat my point with regard to "Angstrom" or OE Feb 22 05:57:04 if a hardware device is that flexible, it does not change the use of that hardware from POV of an OE build Feb 22 05:57:19 you're only likely to use that device ultimately for one purpose Feb 22 05:58:10 and thus, the software distro you ultimately build will be tailored for that user interface and purpose Feb 22 05:58:57 JustinP grokked my meaning, so I hope by clarifying these points you will also grasp what I meant to say Feb 22 05:58:58 i see what your saying, so now that you have a distro how do you get outside users to support/add onto it? Feb 22 05:59:17 * shadows grumbles "you f!@#$%ng don't" Feb 22 05:59:21 sorry Feb 22 05:59:23 OZ has support mainly due to Sharp releasing SDKs Feb 22 05:59:28 Gentoo devel flashbacks Feb 22 06:00:13 i understand what you mean now with sectoring the hardware Feb 22 06:00:26 and yes i was viewing it differently Feb 22 06:00:53 since the hardware is mainly segmented by vendor currently Feb 22 06:00:55 at the embeddedlinux level, hardware ultimately has a task assigned to it in the scope of how a human interacts with that hardware+software comprising the total of the device Feb 22 06:01:22 now, are we going to task this that we also take on *bsd on embedded devices? Feb 22 06:01:35 what are our technical limits? ... then we ask, what are our project goals? Feb 22 06:02:15 koen's email to the angstrom devel list does not address these points to my satisfaction Feb 22 06:03:33 i think saying something like "pulling from live CVS is bullshit, we need to freeze our sources and provide a Q+A'd devel kit" is more realistic, without having to redefine the goals of any particular distro Feb 22 06:04:23 koen did not exactly say what problem he is trying to fix, or why it is a problem; it just gives a new name for the problem and that is Angstrom Feb 22 06:05:45 i dont think its a problem trying to be fixed Feb 22 06:06:45 hard to say Feb 22 06:07:43 i am curious again tho where your pulling from a live cvs? Feb 22 06:07:57 all over the damn place, in .dev Feb 22 06:08:02 ;) Feb 22 06:08:12 ah taht explains it Feb 22 06:08:16 anything without a SRCDATE Feb 22 06:08:34 only release or release destined stuff gets locked down Feb 22 06:08:53 right. that makes it annoying to be a dev and contribute stuff IMO Feb 22 06:09:11 it confused me as well when they pushed oz-fam into a branch out of dev Feb 22 06:10:07 realistically we should not need seperate branches like there are, except for the last-before-currently-stable releases Feb 22 06:10:36 things that are depreciated get their own legacy or archival branch Feb 22 06:10:38 quite a lot of work has gone into preventing peopel from compileing "unversioned" distros Feb 22 06:12:46 yeah i am still not sure why it was branched Feb 22 06:13:11 there is a large thread about it somewhere Feb 22 06:15:28 some of the sectors i mentioned for interfaces: full laptop replacements, information display or gateway device (nokia 770-ish), traditional PDA (trim and efficient input methods), enhanced user interactive interface for a task (GPS-in-car, music-in-car, airport ticket printing terminal, digital photo printing kiosk) Feb 22 06:17:04 ...routers (web interface optional), backend server variants of user interfaces (backend for mythtv enhanced user interactive interface) Feb 22 06:17:57 all these sectors require tailored distro considerations Feb 22 06:18:18 yeah Feb 22 06:18:32 sure you could do any one of those tasks on top of one of the other tasks Feb 22 06:19:00 if you really wanted to. Feb 22 06:19:04 will it gain any advantage over the vendor system used now? Feb 22 06:19:18 without a doubt. Feb 22 06:19:26 (atleast it seems vendor based) Feb 22 06:20:15 we (mythical company "we, Inc.") should not concern ourselves with what hardware target our voting system distro will run on Feb 22 06:20:48 build the software, and patch it to meet the constraints of a particular hardware target when necessary Feb 22 06:21:02 but the key would be to build the software with a rough guideline in mind of what is expected Feb 22 06:21:20 and whatever OE distro system we work out takes care of the details, with information provided from people who support the hardware Feb 22 06:21:34 that way a hardware person is not locked down into a particular distro Feb 22 06:21:36 but who writes the hardware specific aspect? Feb 22 06:21:42 people who love hardware Feb 22 06:21:55 companies who want their devices supported Feb 22 06:22:04 rabid F/OSS developers Feb 22 06:22:29 that would be nice Feb 22 06:23:07 my proposal solves the underlying problem *-I-* see which is, the hardware devs are being asked to follow a billion different distros and software revisions and other shit Feb 22 06:23:43 yeah that has more to do with the fact that distros are mainly vendor based again Feb 22 06:23:52 exactly Feb 22 06:23:59 familar=ipaq Zaurus=sharp etc Feb 22 06:25:26 if you love the (hypothetical) XABZY301 dsp IC, and have a hankering to write code for supporting it on your Me, Inc. Magnificent2xK handheld, well, it may just be that there's an identical chip on devices from 6 other vendors Feb 22 06:25:32 since that means 6 other distros Feb 22 06:25:43 fat freaking chance of them seeing the code in a timely manner Feb 22 06:26:20 i can't reasonably expect a hardware devel to follow mailing lists and code commits and code bases and kernel revisions and patches from 6 other distros Feb 22 06:26:56 even though the hardware dev may readily recognize how GPIO registers or whatever is involved, interact with hardware he/she does not posess, in order to make their favorite dsp chip supported Feb 22 06:27:26 but its hardware combinations that are the problem ... Feb 22 06:27:36 so there Feb 22 06:27:39 you've defined the problem Feb 22 06:27:49 Angstrom is not solving that problem ;) Feb 22 06:27:54 it's just the other name for the problem Feb 22 06:28:48 but you can approach the problem from other directions Feb 22 06:28:57 upwards, diagonally Feb 22 06:29:01 this way Feb 22 06:29:02 * shadows points Feb 22 06:29:45 are you aware the current licensing restrictions enforced in OE? Feb 22 06:29:54 actually, no Feb 22 06:30:02 that would be the first i've heard of such a thing Feb 22 06:30:12 there are many non-open tools and libraries we cannot use Feb 22 06:30:19 due to licensing Feb 22 06:30:26 where is this defined? Feb 22 06:30:35 i was hoping for an addition of *BSD to the build system Feb 22 06:30:57 or potentially something that implements the win32 API Feb 22 06:31:22 this was put in place a while ago when some very closed souce was discovered in the meta Feb 22 06:31:26 source* Feb 22 06:31:28 maybe an FreeDOS distro Feb 22 06:31:31 oh Feb 22 06:31:43 hence the LICENSE tags now Feb 22 06:32:45 anyway i agree with the assesment of the problem Feb 22 06:33:01 developers for the hardware have always been short in supply Feb 22 06:33:37 short in supply and also are asked to do a lot, when they themselves are short on time Feb 22 06:33:50 yes Feb 22 06:34:31 even if there were 2 developers for hardware, with infinite time to spend hacking, only a certain fraction would be spent hacking on things that yield (what i may consider to be) successful wins for the other non-hardware devels Feb 22 06:35:32 well gremlin is a prime example of this Feb 22 06:35:53 mmhm Feb 22 06:35:57 since he is the one doing all the h3600 port to the 2.6 kernel Feb 22 06:36:09 * shadows laughs Feb 22 06:36:11 what a stud! Feb 22 06:36:18 been at it 6-9 months now he said Feb 22 06:42:00 whose project primarily is OE? Feb 22 06:42:26 i should ask that differently, whom is the person that currently is the big boss for OE Feb 22 06:42:30 no ones everyones Feb 22 06:42:31 currently there isnt one Feb 22 06:42:34 okay Feb 22 06:42:53 i wanted to ask about supporting kernel sources other than Linux Feb 22 06:43:02 like as a prime example, FreeDOS Feb 22 06:43:20 feel free to add it, as long you set LICENSE correctly so that source distribution isnt done for licenses that dont allow it Feb 22 06:43:28 okay Feb 22 06:43:42 is there any limit to this, licenses permitting? Feb 22 06:43:52 i'm trying to get a feel for the scope of OpenEmbedded goals Feb 22 06:44:18 OE is a set of package metadata. the goals are more in the scope of the distribution policy Feb 22 06:44:24 okay Feb 22 06:44:45 i'd be cautious about adding anything whose license doesnt permit _binary_ redistribution, since that could easily end up in feeds of distros unknowingly Feb 22 06:45:01 we have a bbclass that does source distribution obeying license, but dont have anything like that for binaries. its assumed that thats allowed Feb 22 06:47:01 i should reactivate my safari account ... Feb 22 06:48:34 * emte passes kergoth a beer Feb 22 06:48:40 shadows: that is, if binary distribution is subject to possessing a license, which you have but others using oe may not, then keep your .bb files local and dont put them in the upstream oe repo. if binary distribution is allowed for everyone, then it can go into oe, whether source distribution is allowed or not, as long as you set LICENSE appropriately Feb 22 06:48:45 mmm, beer Feb 22 06:50:54 i wish i knew enough about the 2.6 kernel to poke at the battery ... Feb 22 06:53:31 shadows, what would you do with freedos btw? Feb 22 07:01:26 emte: maybe a customized OS distro for embedded games, in like a tavern or bar setting Feb 22 07:01:56 i think OpenEmbedded should remain friendly to projects like Voting Kiosks or other embedded uses Feb 22 07:02:42 something non-PDA, but definitely embedded Feb 22 07:03:05 people use oe for all sorts of things, lots of which arent pda related, and a number of which use internal package repositories with things that never go upstream Feb 22 07:03:13 sj Feb 22 07:03:15 they exist like the meshcube, developers are not too active tho Feb 22 07:03:16 i meant, ah Feb 22 07:03:37 anywho, i'm done soapboxing about Angstrom and issues for the moment Feb 22 07:03:44 lol Feb 22 07:04:06 * emte was lookting at FreeRTOS Feb 22 07:04:26 would you guys mind having a look at http://bugs.treke.net/show_bug.cgi?id=705 and tell me what from libmatchbox svn (see bug's newly attached file for reference) needs to go in? Feb 22 07:04:31 more my type of stuff ... even has a 18F pot i should try Feb 22 07:04:32 i have like zero clue about automake/conf Feb 22 07:05:12 does anyone really understand the auto tools? :P Feb 22 07:06:11 you'd think not, and then there is Chris White, who claims after a drunken night and some rough hangover downtime, the light became evident and his auto-fu skill increased by 1 point Feb 22 07:06:41 which makes him clearly 1 point ahead of everyone else in the known universe Feb 22 07:06:43 * shadows [/joke] Feb 22 07:07:34 lol Feb 22 07:10:05 i think i am missing why you want to split the changes? Feb 22 07:10:34 your jsut doing a full diff between the two repos are you not? Feb 22 07:11:58 emte: it's a very specialized diff, and yes it is essentially a diff of just the changes Feb 22 07:12:01 odd what is this? Feb 22 07:12:03 @XSET_LIBS@ Feb 22 07:12:14 so you can tell me what the heck I need to make fonts work again ;) Feb 22 07:12:52 i heard a rumor that it was a missing xsettings lib dependency in the linking phase that makes font settings not work Feb 22 07:13:57 will bitbake properly handle it tho? Feb 22 07:14:07 uhh Feb 22 07:14:30 the point is that someone with knowhow review that diff, and split it out into seperate patches Feb 22 07:14:41 so that i can try 'em out and see which one fixes the font issue Feb 22 07:14:42 yeah Feb 22 07:15:02 i already split out the "explicit types" patch Feb 22 07:15:06 there are quite a few tag changes Feb 22 07:15:27 @MB_EXTRA_LIBS@ etc Feb 22 07:15:32 mmhm, more than i feel comfortable parting out Feb 22 07:16:43 without knowing (and i dont) my guess is that these changes maybe why it wont build Feb 22 07:17:08 btw if you havent figured it out, i dont know most things :) Feb 22 07:17:24 it's good to ask questions, sir Feb 22 07:18:45 yeah ... i generally end up talking to myself when i ask questions ... Feb 22 07:18:56 and reading a lot Feb 22 08:29:46 morning Feb 22 08:49:33 morning Feb 22 09:11:26 morning all Feb 22 09:12:22 hey pH5 Feb 22 09:15:08 hey koen Feb 22 09:26:45 morning Feb 22 09:26:53 hey hrw|work Feb 22 09:29:15 loooks like tosa does not need apmd killing to get suspend/resume working Feb 22 09:37:19 morning all Feb 22 09:37:51 shadows: still around? Feb 22 09:42:28 was there a change in install-native recently, mine seems to be buggered Feb 22 09:44:24 XorA: I was having fun with it but it was qemu at fault in my case. No recent changes I know of... Feb 22 09:45:07 D'oh Feb 22 09:45:25 ASSUME_PROVIDED = stupid-gcc-2.95 Feb 22 09:45:31 don't put that in local.conf Feb 22 09:45:39 ah. That is going to catch people out... Feb 22 09:46:02 local.conf.sample has +=, so it should be safe Feb 22 09:46:09 RP: /home/dp/zaurus/build-gcc/tmp/staging/x86_64-linux/bin/install-sh: ./-Ds does no Feb 22 09:46:10 t exist. Feb 22 09:46:21 RP: and other such wierd messages Feb 22 09:46:22 I did change local.conf,sample a while back so people don't get caught out in future... Feb 22 09:47:20 XorA: Run strace against it and try to find out what the weierdness is? Feb 22 09:47:44 * RP watches his zaurus oops badly :-/ Feb 22 09:49:13 I think I need to talk to Liam... Feb 22 09:49:37 hi. can i get the patches for tslib and kdrive without downloading all oe somehow? Feb 22 09:50:48 i'd be happy if someone could get them for me,too. all i want is use them as an example for building debian packages Feb 22 09:51:20 quitte> etc/console/ Feb 22 09:51:20 10:11 < quitte> this is rather painful. but i can send you mine thats quite useful for german Feb 22 09:51:22 RP: the install-sh is working fine, so something in bitbake is mismunging the arguments Feb 22 09:51:23 10:14 < quitte> or you find the keymap jlime uses Feb 22 09:51:28 sorry Feb 22 09:52:10 quitte: get http://ewi546.ewi.utwente.nl/OE/OE-20060222.tar.bz2 and extract the patches Feb 22 09:52:42 didnt find any info about the beer event friday evening... somebody knows where and when? Feb 22 09:52:57 greentux_alt2: Roy d'Espagne Feb 22 09:53:10 like last year, tnx koen Feb 22 09:54:09 see http://www.fosdem.org/index/news/0/en/5 Feb 22 09:54:23 koen: thanks Feb 22 10:01:00 hmm. seems i misunderstood how ipkg works. i dont see anything like debian/rules or debian/control files. can you give me a hint what i should read to find out how ipkg packages are built? Feb 22 10:02:38 quitte: http://handhelds.org/moin/moin.cgi/BuildingIpkgs Feb 22 10:02:50 quitte: in the OE case, OE will write the control files for you Feb 22 10:03:04 RP: ok, Ive found the fault, install-sh is falty, the case statement that strips out the commandline doesnt all -Dblah -mblah, only these with spaces Feb 22 10:04:08 unfortuneately this isnt compatible with install command Feb 22 10:05:04 koen: ok. so oe won't help me. i have to do it myself :( thanks Feb 22 10:05:16 morning all Feb 22 10:05:54 morning ade|desk Feb 22 10:08:19 RP: I suspect you haven't tried building zaurusd from a clean .dev yet Feb 22 10:08:32 Image you want to have a User with a password in your image. How you solve this problem Feb 22 10:08:47 there are many different ways but how this would fit into the whole system? Feb 22 10:08:49 in the best way Feb 22 10:10:07 Cobelius: look at the base-passwd .bb for one way Feb 22 10:10:22 Cobelius: you could also make a package that adds a user, like ppp-dialin Feb 22 10:10:23 03koen 07org.oe.dev * r95868ff1... 10/packages/tslib/tslib_cvs.bb: tslib: also inherit pkgconfig Feb 22 10:10:41 okay there i can add the user in the passwd file but where can i describe that the home directory will be created? Feb 22 10:11:09 ahh okay this with the package would be a good idea Feb 22 10:11:50 ahhh Feb 22 10:12:10 is it shell scripting i can use in pkg_postinst? Feb 22 10:13:26 RP: why does zaurusd insist on killing mbinputmgr, mb-keyboard and mb-stroke? Feb 22 10:14:05 koen: In theory you don't need them when the keyboard is available... Feb 22 10:14:39 That logic should be made configurable... Feb 22 10:14:46 RP: that doesn't answer why they are being killed Feb 22 10:15:08 koen: How else would it remove them? Feb 22 10:15:35 why remove them? They aren't doing any harm Feb 22 10:16:19 I guess I need to stop using the virtual keyboard on landscape ;) Feb 22 10:16:20 The reasoning was they aren't needed when you have a full keyboard available Feb 22 10:17:12 koen: If you're on a c7x0, could you dump an alsa mixer config for me please? :) Feb 22 10:17:24 sure Feb 22 10:17:44 You'll note that its missing for c7x0 atm. Feb 22 10:18:10 RP: I also unbroke the zaurusd build with the above commit Feb 22 10:18:37 koen: ah, yes, thanks. I meant to do that... Feb 22 10:18:59 Now I have a working headset I might get the sound input tested on c7x0 at last :) Feb 22 10:19:19 Hmmm Feb 22 10:19:58 wehen pkg_postinst ist called during the image creation is the rootfs in the work directory the root or can i write in my build system? Feb 22 10:26:31 morning all Feb 22 10:26:38 hi liam Feb 22 10:26:52 RP: reading your mail Feb 22 10:27:06 lrg|home: I'm hoping its something simple... Feb 22 10:27:13 RP I think it is. Feb 22 10:38:26 koen: With multimachine, have you noticed problems with libtool? Feb 22 10:39:44 I am totally stumped as to why my install stuff is no longer working when it used to and as to what the fix is Feb 22 10:46:10 how can I say that a homedirectory is part of the package? Feb 22 10:46:41 RP: any chance for asoc update today? Feb 22 10:47:22 greentux_alt2: It doesn't work - lrg|home is looking into it but has gone very quiet... Feb 22 10:48:34 RP: I'm just adding some debug... Feb 22 10:49:38 ~seen jbowler Feb 22 10:49:48 RP: I haven't noticed any problems Feb 22 10:49:49 jbowler was last seen on IRC in channel #oe, 16d 16h 37m 15s ago, saying: 'k'. Feb 22 10:49:52 lrg|home: ok. I just worry when someone says simple and then goes very quiet :) Feb 22 10:50:14 hrw|work: which email addr does mickeyl read for OE stuff? Feb 22 10:50:17 RP: np, snail mail also arrived... Feb 22 10:50:48 RP: what kernel are you using ? Feb 22 10:50:58 morning Feb 22 10:51:01 lrg|home: 2.6.15 Feb 22 10:51:35 lrg|home: If that's at fault, I can change... Feb 22 10:51:38 RP: I have 2.6.16-rc2 on mainstone Feb 22 10:51:52 RP: I don't think its at fault Feb 22 10:52:08 XorA: mickeyl@hh Feb 22 10:52:34 RP: I'm going to try wm8750 on mainstone as its quicker for testing Feb 22 10:52:44 RP: what do you think about http://bugs.openembedded.org/show_bug.cgi?id=700 ? Feb 22 10:53:53 koen: I'll have it applied locally - I can push Feb 22 10:54:03 RP: same here :) Feb 22 10:54:13 koen: feel free to push then Feb 22 10:55:38 ok, pushed Feb 22 10:56:23 http://pastebin.de/4668 Feb 22 10:56:29 what's wrong? Feb 22 10:56:34 the package is not createt Feb 22 10:56:48 I subconciously avoid using monotone atm which is which certain things are just sitting locally :-/ Feb 22 10:57:28 6ALLOW_EMPTY = 1 Feb 22 10:57:43 without the 6 Feb 22 10:57:48 ahhh **** ENDING LOGGING AT Wed Feb 22 10:59:57 2006