**** BEGIN LOGGING AT Sun Jan 08 02:59:56 2006 Jan 08 03:04:27 hi all Jan 08 03:12:11 good morning all Jan 08 03:12:31 hey Jan 08 03:12:39 koen: the cpu graph is funny Jan 08 03:12:51 there's a bit missing :) Jan 08 03:12:56 koen: from 0:20 to 0:40 no cpu was utilized at all Jan 08 03:13:04 koen: did you hibernate the system? Jan 08 03:13:19 make -j did Jan 08 03:13:29 koen: it was only -j5 on monotone Jan 08 03:13:33 * koen lacks physical access to that machine Jan 08 03:13:58 I now have 1006 darcs revisions Jan 08 03:14:08 and git? Jan 08 03:14:43 sadly it stopped... and I had to restart it now Jan 08 03:14:52 357 rev's Jan 08 03:15:11 RP: +PACKAGE_ARCH = "native" in native.bbclass, is that correct? Jan 08 03:15:13 tailor couldn't handle empty merges where diff prints #\n#\ no changes Jan 08 03:15:34 RP: shouldn't that be something like ${BUILD_ARCH}? Jan 08 03:16:56 koen: the good thing is: I replaced my OE.db with a current one and started tailor again Jan 08 03:17:04 koen: and it found the csets and continues converting Jan 08 03:17:15 nice Jan 08 03:19:53 http://david.student.utwente.nl/DailyToons/cartoons/user_friendly_06jan_xuf008708.gif Jan 08 03:20:11 wait, that won't work Jan 08 03:20:12 http://www.userfriendly.org/ Jan 08 03:20:40 glibc-intermediate (i386) does not build. Any ideas ? Jan 08 03:20:56 uv1: did that patch work? Jan 08 03:21:22 * wingel does the engineer dance Jan 08 03:22:10 NAbyss: It is already in, but did not help. Jan 08 03:22:16 Smeg. Jan 08 03:22:50 And thats clear so far, as bitbake takes glibc-intermediate_cvs2004... bb file ... Jan 08 03:23:35 So I thought intermediate_cvs possibly, but "arm-audit.patch" failed ;-) Jan 08 03:23:53 s/;-)/:-( Jan 08 03:23:56 I've finally got the Acer N30 and the Acer N35 working well enough to run Linux as the only operating system. It can now suspend and resume to ram and probably has about 180 hours of standby time. Jan 08 03:28:03 wingel: very good! Jan 08 03:28:40 wingel: Nice one Jan 08 03:28:57 The question is, why gcc-cross-initial build "glibc-intermediate_2.3.2+cvs20040726.bb". Is that intended ? Jan 08 03:29:34 Are you asking whether it is intended to build glibc-intermediate at all, or that particular version? Jan 08 03:30:00 If you are on an NPTL platform, yes, glibc-intermediate is required/ Jan 08 03:30:05 RP: I have applied your patch to bitbake. I will write an email about testing etc... but now I need to work on my assignments Jan 08 03:30:19 good morning zecke Jan 08 03:30:34 hi master Jan 08 03:34:16 pb_: That particular version! Would not 2.3.5 needed ? Jan 08 03:34:59 Why should it be? Jan 08 03:35:20 The particular version isn't very important. You can use 2.3.5 if you like. Jan 08 03:35:42 pb_: Just thought as the final target is 2.3.5. But any way, does not build, so I can not build from scratch ;-( Jan 08 03:36:01 No, there is no requirement for the versions to match. Jan 08 03:36:21 glibc-intermediate is just required to build gcc-cross. It doesn't serve any other purpose. Jan 08 03:37:04 pb_: Do you have any idea, why gcc-cross-4.0.2 does not build any more (because of broken glibc-intermediate build) ? ... Jan 08 03:37:20 no, I have no idea Jan 08 03:38:00 glibc-intermediate does on the first glimpse not build due to unknown compiler (4.x). That was fairly quick fixed. But also compile fails. Jan 08 03:38:51 just in a general way, or with some particular error? Jan 08 03:39:35 Wait a few minutes, I have to restart the build process ... Jan 08 03:41:03 pb_: Ok we have the compiler version error "checking version of ccache i586-linux-gcc... 4.0.2, bad" Jan 08 03:41:46 Didn't you just say you had fixed that? Jan 08 03:41:58 uv1: corrects configure.in;rm configure;autoconf;run_do.configure Jan 08 03:42:12 pb_: manually fixed, until it works really. Jan 08 03:42:13 okay Jan 08 03:42:49 * pb_ breakfast time now Jan 08 03:42:50 bbiab Jan 08 03:45:00 <_bob_w> hi .. how do i set a configure value to a package ? Jan 08 03:45:31 pb_: breakfast at PC ? Jan 08 03:45:39 EXTRA_OECONF = "--with-woerle" Jan 08 03:45:55 <_bob_w> with woerle .. aha ha ha Jan 08 03:45:56 for autotooled packages Jan 08 03:46:09 <_bob_w> thanks koen Jan 08 03:46:22 :) Jan 08 03:46:46 <_bob_w> thanks Jan 08 03:48:30 morning all Jan 08 03:48:37 zecke: thanks :) Jan 08 03:48:58 koen: It perhaps should use BUILD_ARCH if we have such a thing Jan 08 03:49:17 MOrning RP Jan 08 03:50:36 RP: http://ewi546.ewi.utwente.nl/tmp/viewmtn/getdiff.py?id1=5f40addfed07ed4b104416c2d8464fe20c5ba399&id2=61486dc9a1b23e7088c746a1a9d6c6199cc8f5a5&fname=classes/native.bbclass Jan 08 03:50:57 zecke: My builds failed last night due to errors to do with md5sum :-( Jan 08 03:51:50 zecke: http://pastebin.ca/36169 Jan 08 03:52:30 koen: ok, we can update that :) Jan 08 03:52:42 ok, commiting that Jan 08 03:53:04 RP: Any ideas, why CL860 sound does not any more work ? Jan 08 03:53:32 RP: well you could add md5sum to a lot of uri's ;) Jan 08 03:54:05 uv1: Could be any of a lot of reasons. Perhaps you'd like to be more specific? Jan 08 03:54:12 if fetch_uri(uri, basename, dl, md5, parm, localdata): Jan 08 03:54:21 shouldn't that be md5sum instead of md5? Jan 08 03:54:25 uv1: Which kernel for a start. Did you know we'd switched to alsa? Jan 08 03:55:03 03koen 07org.oe.dev * r2b7b664e... 10/classes/native.bbclass: native.bbclass: use BUILD_ARCH instead of 'native' Jan 08 03:55:23 RP: opie does not play any sounds any more, nor does alsa-player ;-( kernel is "Linux c7x0 2.6.15-rc7 #1 PREEMPT Sat Jan 7 08:45:06 CET 2006 armv5tel unknow" Jan 08 03:55:58 uv1: Load alsamixer and turn on the left and right mixers Jan 08 03:58:39 RP: Was already, tuned everything to MAX, no change. Sound == Silence ;-) Jan 08 03:59:14 RP: could you try another thing Jan 08 03:59:29 uv1: Sounds is known to work so it will just be your mixer settings. The only thing you need to change is to turn on the left and right mixers Jan 08 03:59:53 Turn on or tune to 100% ? Jan 08 04:00:13 RP: try again please Jan 08 04:00:20 Turning on (unmuting) is not "tune to 100%" Jan 08 04:01:04 RP: Sure, forgive me, I am not familiar with alsa stuff (especially the mixer). Jan 08 04:02:19 zecke: Trying, thanks Jan 08 04:02:45 uv1: There'll be four controls you need to touch.. 'Left Mixer' and 'Right Mixer' need to be unmuted. 'Left Out 1' and 'Left Out 2' need to be turned to the appropriate volume level. Jan 08 04:04:35 NAbyss: OK, I am in alsamixer now. "left" and "right" select the item, right ? I do not find any left or right mixer "Item" ! Jan 08 04:05:15 uv1: This is the c7x0, isn't it? Jan 08 04:05:32 uv1: yes Jan 08 04:05:37 uv1: Hmm, weird. On mine, there's a 'left mixer' control, select it and hit 'm' Jan 08 04:05:44 You could have mentioned that... Jan 08 04:06:06 In that case you won't have a left and right mixer Jan 08 04:06:06 RP: CL860 == c7x0 ! Jan 08 04:06:48 * RP can't remember what needs to be set in alsamixer for c7x0... Jan 08 04:16:10 PB : Found it "Output Mixer HiFi Playback Switch" was off! Jan 08 04:17:13 s/PB/RP/ Jan 08 04:17:13 uv1 meant: RP : Found it "Output Mixer HiFi Playback Switch" was off! Jan 08 04:18:20 RP: Do you know, whether the "JackFunction" option works for CL860 ? Jan 08 04:21:07 could someone pls tell me how to import a privkey from an older monotone DB into a recent one? I always get "monotone: warning: unknown packet type: 'privkey'". The pubkey worked just fine Jan 08 04:22:17 Where is the exc key on my keyboard ;-)) Jan 08 04:24:26 uv1: The cancel button? Jan 08 04:24:53 NAbyss: Oh :-) Jan 08 04:28:31 CoreDump|home: use monotone 0.25 and do a 'db migrate' on the old db Jan 08 04:28:48 zecke: That's working much better now, thanks Jan 08 04:29:05 i was sort of trying to avoid doing that as I would have to pull 3 month worth of changesets heh Jan 08 04:29:11 uv1: Yes, the jackfunction should work Jan 08 04:29:27 CoreDump|home: you can't pull 3 months worth, see topic Jan 08 04:29:28 ahh nm trying to rip the privkey that way Jan 08 04:30:40 hmmm, time to switch of the automerger Jan 08 04:30:52 * koen looks at the automerger battle between ewi and nslu2 Jan 08 04:31:08 http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=cf24301bc3d3b5d5c1e4e83d3299d7efbe001007 Jan 08 04:31:22 CoreDump|home: I don't think the server would hold the privkey.. Jan 08 04:31:33 That kinda defeats the purpose of public key crypto. Jan 08 04:31:46 koen: That is a bit of a battle :-/ Jan 08 04:32:01 koen: Shouldn't the automerger have its own key? Jan 08 04:32:04 RP: good. and on a sidenote the git convert is at rev 400 Jan 08 04:32:23 RP: it has on nslu2, on ewi it's just a shell script on my own db Jan 08 04:32:31 NAbyss: you misunderstood. I have dev access and my own privkey which I'm currently trying to extract from an old DB and import it into a new one Jan 08 04:32:38 zecke: so it might finish today? Jan 08 04:32:48 zecke: Slowly getting there :) Jan 08 04:33:06 RP: it stopped at nite... Jan 08 04:33:16 RP: darcs is at rev 1000 Jan 08 04:34:05 koen: thanks, "db migrate" on the old DB did what I was looking for Jan 08 04:34:38 hi folks Jan 08 04:34:46 hey alan|laptop Jan 08 04:34:54 hey koen Jan 08 04:35:29 "I saw that apparently OZ doesn't come with Samba installed" Jan 08 04:35:53 * alan|laptop bbl Jan 08 04:35:54 from the same people that use qt/e 'because it's embedded' Jan 08 04:36:45 koen: well Sharp.ROm has Jan 08 04:37:30 CoreDump|home: Ahh, okies.. missed that bit Jan 08 04:37:37 It's getting late here =) Jan 08 04:39:12 =) Jan 08 04:39:27 zecke: http://lists.freedesktop.org/archives/xorg/2006-January/012047.html Jan 08 04:42:22 koen: summary: X is old? Jan 08 04:43:04 sort of Jan 08 04:45:30 his statement about double buffering is interesting Jan 08 04:45:52 in what way? Jan 08 04:46:11 that things can look faster when they are actually slower Jan 08 04:47:34 * koen is of to work now Jan 08 05:12:18 I'm having a problem building "quilt". As their CVS is b0rked, I changed SRC_URI to a http download. Now "fetch" bails out with "KeyError: 'md5sum'". Any idea what this is about? Jan 08 05:15:47 CoreDump|home: a) svn up your bitbake. I sadly broke it yesterday Jan 08 05:16:09 CoreDump|home: b) savannah changed their CVSROOT Jan 08 05:16:18 did that an hour ago trying again Jan 08 05:16:27 CoreDump|home: you will need to rm -rf cvs/ for packages from savannah Jan 08 05:16:45 jep, could get the new CVSROOT to work for "anonymous" Jan 08 05:17:46 how often is http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 generated ? Jan 08 05:19:34 koen can answer that =) Seems quite often tho. It was only 2 changestes behind last time I tried Jan 08 05:19:34 alan|laptop: look at the dates Jan 08 05:19:39 alan|laptop: also there is a snapshot.db Jan 08 05:20:51 zecke: ok. found. thanks. Jan 08 05:21:42 * alan|laptop is trying hard to understand how everything work. Jan 08 05:21:55 * CoreDump|home so knows the feeling Jan 08 05:24:56 i'm really in troubles with monotone... the binary 0.25 version requires GLIBCXX_3.4.6, which i do not have on my mandriva 2006. the src.rpm sucks and can't be rebuilt... the cooker version crashes... and i'm not sure compiling it from sources is the best solution when time to update comes... Jan 08 05:25:06 any advice would be welcome... Jan 08 05:30:28 I think we're now in a position to start correcting DEPENDS... Jan 08 05:38:08 alan|laptop: compile it from source :} Jan 08 05:47:20 meh, anyone got working quilt*.bb's? Jan 08 05:47:45 yep. Jan 08 05:48:08 see bugs.openembedded.org ..... one mo... Jan 08 05:50:00 just trying to look up a ?number for you ... try ?540 Jan 08 05:50:20 bugs.openembedded.org is a _bit_ busy (!) Jan 08 05:50:29 indeed =) Jan 08 05:50:43 _drat_. Jan 08 05:50:45 okay. Jan 08 05:50:50 heh Jan 08 05:50:54 http://hands.com/~lkcl/blueangel Jan 08 05:51:29 gpe.monotone.diff Jan 08 05:51:44 excellent, thanks! Jan 08 05:51:52 you need to reverse a couple of arguments in the SRC_URI that does the cvs getting stuff. Jan 08 05:52:08 it's a bug in bitbake (or whatever, don't know, don't care) Jan 08 05:52:27 oh. that one. Jan 08 05:52:50 are you getting the stupid warning from cvs.savannah.nongnu.org that you must use cvs.savannah.nongnu.org? Jan 08 05:53:12 i fixed it by using cvs.sv.nongnu.org Jan 08 05:53:22 nope, no warning. A pefect pull here Jan 08 06:08:50 koen|away: could we serve the darcs tree from ewi as well? Jan 08 06:10:05 cool. Jan 08 06:12:54 i just make a 'monotone update' and doing a 'bitbake task-bootstrap' i got a lot of error about missing org.openembedded.dev/conf/machine/handheld-common.conf .. it's normal ? how to correct ? Jan 08 06:14:05 umm did you set MACHINE in local.conf? Jan 08 06:15:21 yes h3600 Jan 08 06:15:33 before last 'monotone update' all was fine ... Jan 08 06:15:40 well, dunno then,sry Jan 08 06:15:52 might want to svn up bitbake Jan 08 06:15:52 i read on ml about big changes in oe so i'm asking ... Jan 08 06:16:32 mhh it's a chance ... thanks Jan 08 06:16:48 np Jan 08 06:33:15 gremlin[it]: RTFT Jan 08 06:35:47 gremlin[it]: That file has moved to conf/machine/include. You'll need to see where its being included from incorrectly Jan 08 06:42:34 the parsing of all .bb files give me such error :( Jan 08 06:42:53 "use monotone update -rjan2006prebreakage to avoid breakage" Jan 08 06:43:23 Luke i want to test :) Jan 08 06:44:30 gremlin[it]: Do you have some custom .conf files? Jan 08 06:44:42 RP : yes in fact i'm checking ;) Jan 08 06:44:57 gremlin[it]: Its probably in one of them... Jan 08 06:44:58 gremlin[it]: well, if you're purposely testing breakage, don't complain when breakage is broken =p Jan 08 06:48:18 ok now parsing of all .bb file seem to work :) Jan 08 06:48:59 Luke-Jr: I don't see this as a complaint, more a request for assistance which is ok Jan 08 06:49:37 gremlin[it]: excellent. Also watch out for CVSDATE's in your own .conf files (should be SRCDATE) Jan 08 06:51:03 The changeset I just commited is perhaps going to be a bit more painful - removal of incorrect DEPENDS from packages/meta... Jan 08 06:53:24 ok RP ... but mainly i have (badly) hack oe to build kernel 2.6 for h3600 ... all other is unchenged Jan 08 07:04:44 zecke: if it doesn't require any extra free ports we could serve the darcs tree on ewi Jan 08 07:10:14 koen: it would be plain http Jan 08 07:11:29 mhh i have problem connecting to www.oesources.org :( Jan 08 07:12:46 zecke: iirc /var/www/OE should be writeable for you Jan 08 07:13:00 gremlin[it]: Its down atm Jan 08 07:16:20 RP ok ... Jan 08 07:19:10 Hmmm. Is a machine setting BOOTSTRAP_EXTRA_DEPENDS = "zaurus-updater" the right thing to do? Certainly not, but what would the right thing to do be? Make is an RDEPENDS? Jan 08 07:21:31 Sounds more like it should be in IMAGE_DEPENDS, but I suppose it would depend (!) on exactly what the machine uses that file for. Jan 08 07:22:08 BOOTSTRAP_EXTRA_RDEPENDS would only be appropriate if you actually wanted the file installed as part of task-bootstrap, which seems a bit unlikely. Jan 08 07:22:09 pb_: It just triggers some extra files to end up in deploy (zaurus updater.sh script for example) Jan 08 07:22:27 pb_: But it doesn't generate any packages Jan 08 07:22:44 or won't when I've changed it so having it as a runtime dependency is probably ok Jan 08 07:22:54 okay. well, if it doesn't generate any packages, putting it in any RDEPENDS will surely be bad news. Jan 08 07:23:07 true :) Jan 08 07:23:29 What are the files that it puts in deploy actually used for? Jan 08 07:23:38 flashing the images Jan 08 07:24:08 currently it generates empty packages... Jan 08 07:24:16 okay. in that case, I think IMAGE_DEPENDS is the place. Jan 08 07:24:41 We have an IMAGE_DEPENDS? Jan 08 07:24:49 RP: and a += would give you less surprises Jan 08 07:26:08 RP: heh. yes, we do. Jan 08 07:26:18 though, actually, EXTRA_IMAGEDEPENDS might be the variable you actually want. Jan 08 07:26:37 IMAGE_DEPENDS itself is subject to overrides, which makes assigning to it a bit unpredictable. Jan 08 07:29:22 I'll have to create EXTRA_IMAGE_DEPENDS but that looks very much like the right thing to do Jan 08 07:30:11 What would be the difference between EXTRA_IMAGE_DEPENDS and EXTRA_IMAGEDEPENDS? Jan 08 07:30:21 It seems like it would be a bit confusing having two variables with such similar name. Jan 08 07:31:31 Yes, I'm just confused - ignore me ;-) Jan 08 07:31:54 I typed EXTRAIMAGE_DEPENDS into grep and thought it didn't exist Jan 08 07:45:00 zecke: I'll be away for work next week, so if you need anything on ewi, send me an email Jan 08 07:45:52 * koen starts packing Jan 08 08:20:29 morning Jan 08 08:22:57 darcs get http://ewi546.ewi.utwente.nl/OE/other_scms/darcs/org.openembedded.dev/ Jan 08 08:23:04 in two or more hours Jan 08 08:44:18 anyone feel like testing? Jan 08 08:47:53 zecke: hmmm ? is darcs faster? Jan 08 08:49:47 chouimat: faster than what? Jan 08 08:49:59 monotone Jan 08 08:50:12 well compared to monotone darcs is insecure Jan 08 08:50:22 but way faster as it does not have merkle tree Jan 08 08:50:28 k Jan 08 08:51:20 zecke: will probably test later ... once I got my subversion and track server running on my slug ... Jan 08 08:59:06 bootstrap-images still appear to work :) Jan 08 08:59:15 RP: cg-clone http://ewi546.ewi.utwente.nl/OE/other_scms/git/git-tree/ Jan 08 08:59:46 a couple of month are still missing Jan 08 09:00:12 zecke: Thanks :). I'll experiement with that and hopefully produce a test repository for people to experiement on Jan 08 09:03:41 I will rename that tree Jan 08 09:28:09 re all Jan 08 09:28:33 dumb question: is linux-openzaurus_2. Jan 08 09:28:33 6.14+2.6.15-rc7.bb supposed to compile? Jan 08 09:31:03 RP: git-tree is now named org.openembedded.dev Jan 08 09:34:19 OE's using git now? Jan 08 09:41:45 can anyone give me a clue how "run.do_patchcleancmd" is generated? I have _bash_ function containing python code here heh Jan 08 09:44:32 Luke-Jr: not quite yet - we're just planning to try it Jan 08 09:47:57 Luke-Jr: there is a bitbake git fetcher Jan 08 09:48:09 Luke-Jr: OE itself is not known for using git Jan 08 09:48:25 CoreDump|home: it is generated by bitbake Jan 08 09:49:52 there might be a problem w/ BB then. python in a bash function can't be right =) Jan 08 09:50:32 CoreDump|home: pastebin that file please Jan 08 09:52:36 hi Jan 08 09:52:45 hi Jan 08 09:53:07 just to be sure... is bugs.openembedded.org down ? Jan 08 09:53:11 zecke: http://pastebin.ca/36174 Jan 08 09:53:15 obergix[home]: yes Jan 08 09:53:24 RP: ok ;) Jan 08 09:53:35 bug is at line 509 Jan 08 09:53:43 CoreDump|home: you can end up with python in a bash function wth incorrect syntax :-/ Jan 08 09:54:11 CoreDump|home: where is that python code? :} Jan 08 09:54:29 in do_package Jan 08 09:54:32 # Jan 08 09:54:32 if (bb.data.getVar('INHIBIT_PACKAGE_STRIP', d, 1) != '1'): Jan 08 09:54:32 # Jan 08 09:54:32 bb.build.exec_func('do_strip_modules', d) Jan 08 09:54:32 # Jan 08 09:54:35 err sry Jan 08 09:54:58 line 509 Jan 08 09:54:58 ah thanks Jan 08 09:55:08 bitbake -i parse please Jan 08 09:55:29 package should be a python function then... Jan 08 09:55:40 umm opens mickeys shell Jan 08 09:55:50 * RP suspects something did do_package_append with some python code... Jan 08 09:56:11 * zecke feels dizzy Jan 08 09:56:14 RP: ahh I didn't think of that Jan 08 09:57:45 zecke: that "-i parse" opens up the BB shell and doesn't do anything after that? Jan 08 09:57:48 yet another reason to hate _append :) Jan 08 09:58:03 type parse into the shell... Jan 08 09:58:06 CoreDump|home: bitbake -i and then type parse ;) Jan 08 09:58:13 oh lol Jan 08 09:58:27 :) Jan 08 09:58:42 guess I've been gone to long heh Jan 08 09:59:50 NOTE: Parsing finished. 2901 cached, 0 parsed, 64 skipped, 0 masked Jan 08 09:59:55 no errors Jan 08 10:00:27 now look into the package var Jan 08 10:00:33 and specially the flags Jan 08 10:02:27 hmm how do I do that? Jan 08 10:04:44 getvar sais there ain't no package or PACKAGE var Jan 08 10:07:21 zecke: fwiw, that INHIBIT_PACKAGE_STRIP lossage happens if your DISTRO is faulty Jan 08 10:07:39 hehe Jan 08 10:08:11 pb_: You can start cloning using git and darcs Jan 08 10:08:16 iirc, module_strip.bbclass does that, expecting that package.bbclass will have declared "python do_package() {...}". Jan 08 10:08:20 pb_: svn will be available within days Jan 08 10:08:28 zecke: rock Jan 08 10:08:46 meehhhhh Jan 08 10:08:47 pb_: not all information is available yet Jan 08 10:09:03 I don't think I have the 1337 k3rn31 ski11z to use git, but I might have a go with darcs Jan 08 10:09:08 looks like someone removed/renamed openzaurus-3.5.4.... Jan 08 10:09:34 pb_: thx Jan 08 10:09:35 pb_: cg-clone http://ewi546.ewi.utwente.nl/OE/other_scms/git/org.openembedded.dev Jan 08 10:10:31 so to build on OZ 3.5.4 image I'd set openzaurus-unstable as my DISTRO? Jan 08 10:11:27 can anyone explain me briefly what the "breakage" from the topic is ? Jan 08 10:11:36 zecke: I'll try that directly Jan 08 10:11:56 do I need any particular version of git? Jan 08 10:14:17 Hmm. I appear to have broken gpe-images... Jan 08 10:14:31 That can't be good. Jan 08 10:14:35 Not too badly. Just a question of what the correct fix is... Jan 08 10:14:52 gpe-image.bb DEPENDS on meta-gpe which is fair enough Jan 08 10:15:10 but meta-gpe doesn't have any DEPENDS or RDEPENDS Jan 08 10:15:33 Perhaps it should have RDEPENDS on the packages it creates? Jan 08 10:15:56 or alternatively should gpe-image have REPENDS on the specific packaes from meta-gpe it needs? Jan 08 10:16:29 mm? I don't really understand what you mean by "it should have RDEPENDS on the packages it creates". Jan 08 10:16:47 oh, was the first "it" referring to gpe-image? Jan 08 10:17:16 no, meta-gpe... Jan 08 10:17:37 meta-gpe isn't a package in its own right but creates a number of things like gpe-task-base etc. Jan 08 10:17:44 right Jan 08 10:17:56 Somehow we need to get gpe-task-base into the RDEPENDS list... Jan 08 10:18:07 RDEPENDS list of what? Jan 08 10:18:13 pb_: no it must support the right netsync protocol though :} Jan 08 10:18:28 zecke: okay. I'll do "apt-get install cogito" and take pot luck. Jan 08 10:18:34 pb_: the image I guess Jan 08 10:18:44 yeah Jan 08 10:18:56 I think adding those task packages to the RDEPENDS in gpe-image.bb would be okay Jan 08 10:19:07 pb_: darcs get http://ewi546.ewi.utwente.nl/OE/other_scms/darcs/org.openembedded.dev Jan 08 10:19:14 pb_: the darcs get is faster than git Jan 08 10:19:27 but it would be nice if a DEPENDS = "meta-gpe" was enough to make bitbake do the right thing... Jan 08 10:19:30 pb_: but darcs pull is some-how funny Jan 08 10:19:45 RP: yeah. why isn't that sufficient? Jan 08 10:20:17 pb_: because it builds meta-gpe but meta-gpe doesn't have any depends or rdepends Jan 08 10:20:35 Some of the subpackages meta-gpe generates do but it doesn't care about that... Jan 08 10:20:37 RP: right, but surely that shouldn't matter. Jan 08 10:20:42 once meta-gpe is built, you should be all set Jan 08 10:21:10 woglinde: hey Jan 08 10:21:17 pb_: The meta packages are purely a set of packages that point at other packages Jan 08 10:21:35 pb_: They shouldn't have DEPENDS as they don't need them to build the meta packages Jan 08 10:21:49 RP: right, but they have RDEPENDS. Jan 08 10:22:06 pb_: but meta-gpe itself does not Jan 08 10:22:24 hence the question should meta-gpe have RDEPENDS or the image? Jan 08 10:22:27 indeed not. meta-gpe isn't an output package, and it wouldn't make any sense for it to have RDEPENDS. Jan 08 10:23:05 so DEPENDS = "meta-gpe isn't enough Jan 08 10:23:07 I still don't quite understand what the problem is. Jan 08 10:24:30 bitbake currently never sees gpe-task-base added to the RDEPENDS list of anything so doesn't build its DEPENDS/RDEPENDS Jan 08 10:25:21 The image has the RDEPEND so in theory it is where the metadata is wrong Jan 08 10:25:39 but DEPENDS = "meta-gpe" no longer makes any sense Jan 08 10:26:00 maybe I should just remove that Jan 08 10:26:05 cat anyone help me with a python error I get in bitbake? --> http://pof.eslack.org/~pof/tmp/bitbake-error.txt Jan 08 10:26:22 ah, okay, I see Jan 08 10:26:28 pof: svn up Jan 08 10:26:35 yes, it sounds like the right thing would be to remove that and add RDEPENDS for the actual packages it needs. Jan 08 10:26:36 jup svn up Jan 08 10:26:57 zecke: cogito appears to be... cogitating. Jan 08 10:27:09 do you know how many "objects" there are? Jan 08 10:27:34 pb_: IIRC more than 6.000 Jan 08 10:27:35 RP: svn up? where? Jan 08 10:27:49 pof: from bitbake's svn repo? Jan 08 10:27:53 zecke: okay. it seems to do about ten per second, so that isn't too frightening. Jan 08 10:27:55 pb_: some guy broke bitbake... Jan 08 10:28:07 zecke: doh, that was careless Jan 08 10:28:29 RP: you mean in /stuff/org.openembedded.dev/ ? Jan 08 10:28:31 pb_: co depends on your network bandwidth Jan 08 10:28:36 pb_: bitbake Jan 08 10:28:52 pof: No, I mean update bitbake to current svn Jan 08 10:28:56 does his name start with l, end with n, and rhyme with larson? Jan 08 10:28:57 :p Jan 08 10:29:03 I'm a newbie, i've just followed oe's gettingstarted steps Jan 08 10:29:14 pof: That bug you're seeing is fixed in the latest svn Jan 08 10:29:22 treke: sadly not. It starts with 'z' and with 'ecke' Jan 08 10:29:36 * RP is surprised it was me who broke bitbake :) Jan 08 10:29:45 s/was/wasn't/ Jan 08 10:29:46 RP meant: pof: That bug you're seeing is fixed in the latest svn Jan 08 10:30:02 heh, stupid ibot Jan 08 10:30:04 you can break it next weekend Jan 08 10:30:07 keep up ibot :) Jan 08 10:30:07 hehe Jan 08 10:30:11 ok Jan 08 10:31:39 pb_: I think the neatest answer might be to split gpe into individual task-gpe-xxx.bb files eventually (but we can fix it as it stands too) Jan 08 10:31:58 zecke: 7400 objects so far Jan 08 10:32:47 pb_: than add a '0' to my number Jan 08 10:33:01 fear Jan 08 10:33:07 pb_: it is still faster than monotone ;) Jan 08 10:33:14 heh, indeed Jan 08 10:33:28 pb_: and darcs has twice as much information but would be done with getting by now Jan 08 10:33:46 I'll start a darcs checkout on another machine Jan 08 10:34:52 mmm... in my local.conf, can i leave the PREFFERED_PROVIDERS lines by default ? Jan 08 10:35:01 ah yeah, darcs seems to be zooming along Jan 08 10:35:06 "patch 200 of 1477" Jan 08 10:35:24 pb_: will add some new rev's so you can do a pull after your get Jan 08 10:35:32 git is still grinding away, just passed 10000 objects Jan 08 10:35:41 zecke: very good Jan 08 10:36:27 zecke: You know that git shared via http: has locking issues? Jan 08 10:37:09 RP: no ;) Jan 08 10:37:38 RP: at least the tailor converting should be safe :} Jan 08 10:37:52 zecke: If you just add some new revs it might be ok. Just don't repack it or anything Jan 08 10:37:55 RP: darcs writes to a file 'pending' and then does a rename(2) Jan 08 10:38:07 THinking about it, its probably not packed which is why this is taking as long... Jan 08 10:38:08 RP: ah okay... Jan 08 10:38:31 RP: what is packing? Jan 08 10:38:43 pb_: your darcs get is done? Jan 08 10:38:48 nearly Jan 08 10:38:52 patch 1214 of 1477 Jan 08 10:39:05 pb_: with applying or getting? Jan 08 10:39:09 "copying" Jan 08 10:39:11 whatever that means Jan 08 10:39:38 pb_: your network is slower than mine ;) Jan 08 10:39:44 heh Jan 08 10:39:51 yeah, my network is almost infinitely slow Jan 08 10:40:20 thinking about it, it probably doesn't help that I was checking out into an nfs directory Jan 08 10:41:23 hm, now it's just sitting there Jan 08 10:41:31 "Copying patch 1477 of 1477..." Jan 08 10:41:32 zecke: packing is where all the indvidual commits get combined into one big pack file. Makes git more compact Jan 08 10:41:33 Hi, is it OK to have the openembedded directory on an NFS server and have two or more computers build from the same files? The tmp and all other directories are on the NFS server as well. Jan 08 10:41:34 pb_: it will start aplying soon Jan 08 10:41:39 okay Jan 08 10:41:42 * pb_ waits patiently Jan 08 10:41:47 pb_: the first diff is 35mb ;) Jan 08 10:41:54 pb_: 16mb compressed or so Jan 08 10:41:58 meantime, back at the git ranch... Jan 08 10:42:09 Failed to find remote refs Jan 08 10:42:09 unable to get tags list (non-fatal) Jan 08 10:42:09 New branch: a0e8c5dbdefb5b681f03140f11e4036fb0ada11a Jan 08 10:42:10 is that good? Jan 08 10:42:21 Laibsch: Sharing the openembedded dirs is fine. I'd not like to try tmp over nfs though... Jan 08 10:42:30 pb_: well Jan 08 10:42:39 Laibsch: yeah, that should be fine. I used to do that all the time. Jan 08 10:42:44 RP: Why? Jan 08 10:42:47 pb_: this is what it says here (I'm no git instance though) Jan 08 10:43:07 heh Jan 08 10:43:10 pb_: OK, thanks. Jan 08 10:43:28 ah yeah, darcs is applying now Jan 08 10:43:41 pb_: now you can do cg-fetch to get new objects Jan 08 10:43:56 okay Jan 08 10:44:07 * pb_ does Jan 08 10:44:35 Laibsch: Just the large amounts of data it works over would be faster on a local disk. It shouldn't break anything doing it over nfs though Jan 08 10:44:36 pb_: and RTFM for the rest... as I do not have a clue on how to use git ;) Jan 08 10:44:57 heh Jan 08 10:45:18 well, that seemed to work, I have lots of "D", "A" and "M" lines Jan 08 10:45:41 pb_: ah you could use gitk as a gui Jan 08 10:46:02 pb_: you would be the first to see the converted history Jan 08 10:46:19 tk? jeez, what is it with these version control people. Jan 08 10:46:23 pb_: and you could do darcs pull now as well (I've applied all patches) Jan 08 10:46:36 zecke: rock. darcs just finished applying the patches. Jan 08 10:46:59 * pb_ grit teeth and apt-get install gitk Jan 08 10:47:07 pb_: git log will give you the commit log Jan 08 10:47:16 pb_: you can see how many month are missing in the convert Jan 08 10:48:06 <[lala]> hi Jan 08 10:48:10 <[lala]> ~seen greentux Jan 08 10:48:20 greentux was last seen on IRC in channel #oe, 3d 8h 36m 4s ago, saying: 'morning'. Jan 08 10:48:21 mm, gitk is quite cute Jan 08 10:48:29 apart from the being written in tk part, of course Jan 08 10:49:01 pb_: is the history full of viagra ads? or does it look okay? Jan 08 10:49:43 looks ok to me, though it seems to end at Aug 8. Jan 08 10:50:24 pb_: do a pull and you might be on the 9th ;) Jan 08 10:50:37 heh Jan 08 10:50:59 what does the "[project @ 2719afbdce99244805e624d2ed5519342779c936]" stuff mean? Jan 08 10:51:16 pb_: it is inserted by tailor. It is the monotone revision Jan 08 10:51:21 ah right Jan 08 10:52:19 pb_: we could remove that Jan 08 10:52:31 what's the darcs equivalent to git log? Jan 08 10:52:32 pb_: we can automatically rewrite the logs on the fly Jan 08 10:52:48 "darcs obliterate" sounds amusing, but I doubt it is what I need Jan 08 10:53:00 sounds like fun Jan 08 10:53:04 zecke: crumbs, it sounds like we are invincible Jan 08 10:54:34 ah, darcs changes Jan 08 10:54:38 pb_: I have rsynced another 10 darcs patches Jan 08 10:54:47 pb_: the git convert is running on ewi Jan 08 10:54:54 no viagra ads here either Jan 08 10:54:57 pb_: darcs is running on my linux workstation Jan 08 10:55:16 zecke: cool Jan 08 10:55:34 pb_: I had to 'patch' monotonte to work as there were some issues with the database Jan 08 10:55:59 :-} Jan 08 10:56:21 pb_: I find it weird that darcs asks me to apply patches Jan 08 10:56:48 yeah, I was surprised by that too Jan 08 10:56:54 still, you can say "a", so it doesn't seem too bad Jan 08 10:57:05 maybe there's some commandline switch to tell it to just apply everything Jan 08 10:57:33 pb_: maybe we can convince the darcs people to use monotone's merkle trees and signature handling Jan 08 10:57:47 pb_: pressing 'm' will give you the secure monotone behaviour Jan 08 10:59:00 zecke: by "the secure monotone behaviour", presumably you mean something along the lines of "destroy your local database and have koen revoke your access to the server". Jan 08 10:59:40 that's probably a bit more security than we need just at the moment. Jan 08 10:59:48 hehe Jan 08 11:00:07 that stupid zd wlan dongle... Jan 08 11:00:12 one rsync, ifdown, ifup... Jan 08 11:00:26 heh, zd1211? Jan 08 11:00:30 pb_: yeah Jan 08 11:01:06 yah, that driver is a bit of a nightmare. we have about this many patches for it in our local kernel tree at the office. Jan 08 11:01:30 it does seem to work fairly well for us nowadays, though. I guess I should dig out the patches and try to send them to the crazy upstream guys. Jan 08 11:02:48 fink install igit... Jan 08 11:03:46 We have a slight problem. Jan 08 11:03:56 RP: yes? Jan 08 11:04:26 Take a file like meta-gpe.bb. Currently we ignore all the RDEPENDS of subpackages of that .bb :-/ Jan 08 11:05:01 I think bitbake needs to iterate over PACKAGES, adding all the RDEPENDS_${PACKAGE} to the depencency list Jan 08 11:05:20 doesn't sound wrong Jan 08 11:05:27 but slightly inefficient Jan 08 11:05:49 * RP can't see an alternative Jan 08 11:06:02 RP: right. not with the current bitbake Jan 08 11:06:21 RP: we should sit down and recheck all requirements for bitbake Jan 08 11:06:27 RP: and then use kergoths data model Jan 08 11:06:29 My fixes to gpe-image.bb only partially solved the problem - it still doesn't build due to this Jan 08 11:07:06 zecke: Agreed, we need a new data model. Jan 08 11:07:43 My only concern with kergoth's propsal is that recalcing dependencies if you change any of the overrides variable would be painful Jan 08 11:17:41 Hmm. This creates some very strange interdepencies in glibc :-/ Jan 08 11:20:21 pb_: Do glibc 2.3.5 have a libsegfault? Jan 08 11:20:47 RP: yes Jan 08 11:21:10 yeah, all versions of glibc (in living memory, anyway) have libSegFault Jan 08 11:22:55 Both me and bitbake are very confused. It seems to think the solution is to build glibc-intermediate... Jan 08 11:23:26 glibc-intermediate shouldn't be building any PACKAGES. If it is, that's a bug. Jan 08 11:25:36 pb_: I think bitbake is more interested in its PROVIDES having ruled out glibc provding something due to circular dependencies... Jan 08 11:25:56 If only it would tell me what the circular dependency was... Jan 08 11:28:24 RP: hm, I don't think it should be PROVIDING libSegFault either. Jan 08 11:28:50 afaik, the only thing that glibc-intermediate PROVIDES is libc-for-gcc. Jan 08 11:32:05 I think glibc-intermediate is ok and bitbake is upset about something else... Jan 08 11:32:50 * RP curses his ADSL dropping out and killing a load of ssh sessions Jan 08 11:35:53 has cout been removed from the c++ libraries or summin? Jan 08 11:38:01 Ah, in bitbake's mind glibc now both depends and provides on libsegfault Jan 08 11:40:38 im getting thse when im trying to build an old app i had in an really old simap image.. main.cpp:(.text+0xb60): undefined reference to `std::cout' Jan 08 11:40:38 do i need to -l something in my .pro file? Jan 08 11:46:27 sounds like a question for #opie Jan 08 11:54:04 man you've *got* to love NFS booting Jan 08 12:10:55 When building libstdc++-v3 i get errors about cannot convert 'const __ctype_touplow_t*' to 'const int*', have anyone here encountered this error before? Jan 08 12:11:28 I have the build log at http://pastebin.com/496710 Jan 08 12:17:25 Do RDEPENDS = the name in PACKAGES or the name in PKG? Jan 08 12:18:16 RP: re Jan 08 12:18:32 dan2003: -lstdc++ Jan 08 12:18:46 rehi zecke :) Jan 08 12:18:56 zecke, yeah, i have that already, i thought that would be the case hto! Jan 08 12:19:08 little bit confused! Jan 08 12:19:15 dan2003: there is a stupidity in the qmake.bbclass Jan 08 12:19:40 i have LIBS += -lqpe -lpthread -lopiecore2 -lopieui2 -lstdc++ Jan 08 12:19:41 pb_: how do you like darcs/git so far? Jan 08 12:19:58 dan2003: qmake.bbclass does LIBS-=-lstdc++ :} Jan 08 12:20:16 oh! how do i get round that? Jan 08 12:20:30 put it in twice?! Jan 08 12:21:28 dan2003: no make qmake.bbclass less stupid :} Jan 08 12:21:37 zecke: both of them seem to rule Jan 08 12:21:46 k Jan 08 12:21:53 pb_: you really prefer them over monotone? Jan 08 12:22:06 zecke: well, of course, I haven't actually tried checking anything in yet. Jan 08 12:22:07 Being more specific, why would kernel-image get renamed kernel-image-2.6.15-rc7 as it confuses the handling of RDEPENDS no end :-/ Jan 08 12:22:32 dan2003: proper fix: fix qmake.bbclass fast workaround: do not use palmtop.bbclass (it wasn't qmake.bbclass) Jan 08 12:23:05 pb_: if you could spare your zd1* patches... I would really appreciate it Jan 08 12:23:33 zecke: it will take me a while to collect the patches, but i could send you a tarball of our zd1211 directory Jan 08 12:24:18 BTW: 83 new revisions for darcs Jan 08 12:25:06 evening Jan 08 12:25:09 reenoo_: hey Jan 08 12:25:20 hi zecke Jan 08 12:25:29 * RP could be evil and put a PACKAGES_DYNAMIC = "kernel-image-*" in kernel.bbclass... Jan 08 12:25:31 hi reenoo_ Jan 08 12:25:35 reenoo_: darcs get http://ewi546.ewi.utwente.bl/OE/other_scms/darcs/org.openembedded.dev Jan 08 12:25:56 reenoo_: cg-clone ... git/org.openembedded.dev Jan 08 12:25:57 reenoo_: Can you tell me why kernel-image packages get renamed to kernel-image-${KERNEL_VERSION} ? Jan 08 12:26:47 zecke: heh. what happened to the svn plans? Jan 08 12:27:00 reenoo_: still current. But conversions takes soo long Jan 08 12:27:01 zecke: svn+svk that is Jan 08 12:27:19 reenoo_: and the machine I did the svn conversion dropped from the network :} Jan 08 12:27:31 I did the svn conversion on... even Jan 08 12:27:59 zecke: the point being that I'm not convinced that switching to yet another rather experimental SCM would buy us anything Jan 08 12:28:17 reenoo_: agree Jan 08 12:28:26 reenoo_: I have your ghc issues in ming Jan 08 12:28:34 in mind even Jan 08 12:29:16 only kergoth is using athe svk expotort isnt it? Jan 08 12:29:19 export* Jan 08 12:29:57 emte: if we switch to svn. I think the main repository will be on hh.org Jan 08 12:30:06 emte: svk will be used by people who want to or need to Jan 08 12:30:17 i ment currently Jan 08 12:32:19 reenoo_: can I offer a svk repository just like that? Jan 08 12:33:18 zecke: not sure what you mean by offer. but you can use svk locally with a remote svn repository Jan 08 12:33:54 zecke: to add offline features you'd create a branch of a local mirror Jan 08 12:33:55 reenoo_: after the sonversion I have a svn repository Jan 08 12:34:09 reenoo_: can I do ln -s my_repo /var/www/repo Jan 08 12:34:14 reenoo_: and svk mirror will just work? Jan 08 12:35:32 * RP discovers this is due to PARALELL_INSTALL_MODULES and was added by none other than pb_... :) Jan 08 12:35:42 * RP -> food Jan 08 12:35:44 zecke: well. to serve an svn repository over http you'll need to use the apache module Jan 08 12:36:52 reenoo_: well that sucks ;) Jan 08 12:41:34 I guess you could lock down one ssh account to just a wrapper around svnserve which sets $USER and enable access by adding ssh keys like the bitmover people did with bk Jan 08 12:42:02 but I haven't done anything like that with svn before Jan 08 12:43:04 obviously that doesn't allow anonymous read access Jan 08 12:43:17 reenoo_: the issue is not locking down but having a box to toy with Jan 08 12:43:50 * reenoo_ nods Jan 08 12:45:16 I would have offered a box at university but it appears to have been "opened up" via physical access Jan 08 12:49:26 hopefully we will have server rooms with restricted access in the building we're moving to at the end of january... Jan 08 12:51:07 reenoo_: some rooms were cleared here as well... Jan 08 12:51:23 reenoo_: really cleared... complete removal of all electronica devices in certain rooms Jan 08 12:51:49 heh Jan 08 12:53:13 I guess I wouldn't mind if they got rid of all the file sharing/rooted/virus infected boxes here Jan 08 12:53:59 instead they're upgrading the link to the outside world Jan 08 12:54:45 yay! gigabit access for all the worms! Jan 08 12:54:51 Jan 08 12:55:06 <[lala]> if a ipk package contains directories which don't alread exist on the filesystem - do i have to create those dirs in postinst or what am i missing? Jan 08 12:55:10 reenoo_: hehe Jan 08 12:56:16 reenoo_: as long as FTP World Warez does not point to your uni everything should be fine Jan 08 12:56:20 <[lala]> at least ipkt install doesn't seem to create the missing dirs automatically Jan 08 12:56:53 [lala]: it should do. if that isn't happening, it would be a bug in ipkg. Jan 08 12:57:14 kind of a bizarre bug, though. might be worth checking with strace to see if it makes any attempt to create them. Jan 08 12:57:36 <[lala]> pb_: ok, i try again with strace then Jan 08 12:58:59 got me app built, but have a ofnt problem.. absoulutly know labels, text show up wotsoever in cluding button labels, tab labels, comboboxes etc.. no textx at all.. the font is helvectica and i have this installed, what could i be missing? Jan 08 12:59:08 zecke: hehe. dunno. being a free software user I don't have a need to visit sites like that. Jan 08 13:00:15 hehe Jan 08 13:00:35 I guess zecke doesn't have an actual need, as such, he just visits them for fun. Jan 08 13:00:48 I thought it was an official GNU mirror Jan 08 13:00:51 ;) Jan 08 13:01:01 <[lala]> pb_: i found the bug ... sitting in front of the keyboard ;-) Jan 08 13:01:15 <[lala]> of course it tries to create the dirs Jan 08 13:01:16 [lala]: heh Jan 08 13:10:20 good night Jan 08 13:21:02 hi all Jan 08 13:21:10 reenoo_: ping Jan 08 13:21:19 hi theBohemian Jan 08 13:21:51 reenoo_: I want to build classpath from cvs for OE. is this possible? Jan 08 13:22:42 theBohemian: should be, yeah Jan 08 13:23:39 reenoo_: theres no comprehensive doc about the .bb format. do you know of a package which is built from CVS too where I can steal some ideas? Jan 08 13:25:18 find packages -name '*_cvs.bb' Jan 08 13:27:22 long story short: copy classpath_0.18.bb to classpath_cvs.bb. adjust SRC_URI to point to cvs and override PV Jan 08 13:29:42 theBohemian: where's classpath cvs located these days? Jan 08 13:30:27 reenoo_: so you would feel most confortable with svn? Jan 08 13:31:27 svn is certainly the one with the most real-world deployments. Jan 08 13:31:42 if it does the job, I think everybody would be more comfortable with that rather than something more exotic. Jan 08 13:31:50 slashdot:pb said Linux is not ready for Real World ;) Jan 08 13:31:53 reenoo_: cvs.gnu.org/cvsroot/classpath module classpath Jan 08 13:32:14 pb_: right. Mono and KDE trust svn Jan 08 13:32:43 zecke: personally, yes. together with svk it should cover most relevant use cases for us and is rather unlikely to blow up every other month or force adventurous upgrade procedures on users/developers. Jan 08 13:33:04 reenoo_: I have a bb file for jamvm 1.4.1 and classpath 0.19 here. however since my device has no X I disabled Gtk & Qt AWT (and all the sound stuff) Jan 08 13:33:13 reenoo_: HTTP access should be firewall friendly... Jan 08 13:33:23 reenoo_: push access should be through ssh though Jan 08 13:33:33 zecke: indeed. likewise gcc, and no doubt many other large projects with many users. Jan 08 13:33:50 reenoo_: it might be an idea to introduce this as classpath minimal or should I override the configure options from my distro configuration file? Jan 08 13:33:55 amusingly, though, uli drepper seems to plan on switching to monotone for glibc. Jan 08 13:34:16 pb_: good idea Jan 08 13:34:22 * zecke still likes monotone Jan 08 13:34:49 does bitbake have an equivelant coomand to oebuild compile, that does scan all the .bb data every time? Jan 08 13:35:01 s\does\doesnt Jan 08 13:35:06 dan2003: -b Jan 08 13:35:44 theBohemian: ideally, you'd split it up at packaging time. i.e. build with everything enabled and ship the X/gtk bits in a separate package. Jan 08 13:36:09 k. Jan 08 13:36:24 theBohemian: I know that may be problematic with classpath but would be worth fixing (if possible) Jan 08 13:36:41 is there a proper way to insert build time and data into a binary? i had been using comand line stuff to insert date into version.h Jan 08 13:36:43 bbl Jan 08 13:38:09 * pb_ stabs evolution Jan 08 13:38:17 how can it need 650MB of memory just to read my email? Jan 08 13:38:45 pb_: Artificial Intelligence Jan 08 13:38:58 oh, that explains it Jan 08 13:38:58 pb_: it loads all the mails and news sites and links it expects you to look at Jan 08 13:39:28 oh, that would explain why my screen keeps filling up with viagra websites Jan 08 13:39:41 I did wonder about that Jan 08 13:54:48 mithro: hey Jan 08 14:16:38 hey zecke Jan 08 14:16:44 in bratislavia now Jan 08 14:17:38 mithro: darcs get http://ewi546.ewi.utwente.bl/OE/other_scms/darcs/org.openembedded.dev Jan 08 14:17:41 mithro: ;) Jan 08 14:33:35 theBohemian: handhelds.org/~rw/cp-cvs.diff Jan 08 14:35:19 theBohemian: that patch assumes you're using a tree that doesn't contain the CVSDATE -> SRCDATE changes yet Jan 08 14:49:29 reenoo_afk: thank Jan 08 14:49:33 s Jan 08 14:51:54 * RP wonders about the legality of having PACKAGES = "kernel-image" when it isn't called that at all... Jan 08 14:53:11 zecke: hows it working? Jan 08 14:53:36 you remebered to tag the repository after import? (otherwise download will take forever) Jan 08 14:53:52 darcs get --partial http://ewi546.ewi.utwente.bl/OE/other_scms/darcs/org.openembedded.dev Jan 08 14:54:03 thats a better and considerablly faster Jan 08 14:54:09 (if you have done the tag) Jan 08 14:54:11 mithro: after which convert? Jan 08 14:54:15 mithro: it is still going on Jan 08 14:54:18 zecke: yeah Jan 08 14:55:00 I think we all prefer svn for now... (if we move away from monotone) Jan 08 14:56:56 svn isn't distributed? Jan 08 14:57:00 you mean svk? Jan 08 14:57:17 mithro: well, I will use svk on top of svn Jan 08 14:57:22 svk is just a layer on top of svn afaik Jan 08 14:57:25 mithro: if others will do it, I don't know Jan 08 14:57:37 treke: svk can do git, perforce and cvs as well Jan 08 14:57:39 you can choose to use it yourself and others can just use straight svn if they prefer it Jan 08 14:57:44 zecke: Oh cool Jan 08 14:58:08 zecke: it can? Jan 08 14:58:22 mithro: git is experimental but perforce and cvs should be solid Jan 08 14:58:22 I havent looked at version control tools since the bk issue Jan 08 14:58:36 At the moment the SVK CVS connector doesn't work properly, as a work around you'll need to keep an active CVS repository and manually import patches. Jan 08 14:58:46 :P Jan 08 14:59:09 * treke resists the urge to make a cvs joke Jan 08 14:59:09 :) Jan 08 14:59:09 darcs is likely to be ruled out due ghc Jan 08 15:00:15 ghc? Jan 08 15:00:47 mithro: it is claimed to have portability issues Jan 08 15:01:11 what is ghc? Jan 08 15:01:16 what is ghc? Jan 08 15:01:22 haskell compiler? Jan 08 15:01:26 Glasgow Haskell Compiler Jan 08 15:01:43 zecke: i'm using the bins on Linux and Windows :P Jan 08 15:02:09 mithro: heh, that's hardly evidence of wide portability. Jan 08 15:02:13 mithro: well I can only repeat what I have heard ;) (and believe it is true) Jan 08 15:02:43 GHC works on PPC/Darwin Jan 08 15:03:07 darcs does seem to be built for all the arches in debian, though, which I guess is something. Jan 08 15:03:37 that is good Jan 08 15:03:40 pb_: if the debs exist then it most probably means it passed the tests too Jan 08 15:03:45 http://www.haskell.org/ghc/survey2005-summary.html Jan 08 15:04:18 svn: Can't convert string from native encoding to 'UTF-8': Jan 08 15:04:20 haha Jan 08 15:04:41 but, well, the freebsd and solaris users amongst us might still be out of luck Jan 08 15:04:52 zecke: doh, svn sucks Jan 08 15:04:57 pb_: why? Jan 08 15:05:21 well it is not proprietary so it must suck... Jan 08 15:05:47 pb_: actually it's in freebsd ports? Jan 08 15:06:52 ah, well then. sounds like it's probably portable enough for the purpose. Jan 08 15:07:08 the problem with GHC is that it is itself written in Haskell Jan 08 15:07:20 so you have to bootstrap it someway if you don't have a GHC binary Jan 08 15:07:39 mithro: gch compiled yesterday Jan 08 15:07:50 mithro: darcs did work as well (on FreeBSD) Jan 08 15:08:28 http://www.haskell.org/ghc/download_ghc_641.html Jan 08 15:09:05 no solaris thought :P Jan 08 15:09:16 o wait I lie Jan 08 15:09:21 http://www.haskell.org/ghc/download_ghc_641.html#sparcsolaris Jan 08 15:09:45 Linux Sparc as well? Jan 08 15:10:22 I'm moving to bed now Jan 08 15:10:53 sweet dreams Jan 08 15:11:21 pb_: I will fail an exam tomorrow... Jan 08 15:11:58 zecke: sounds like a bad plan Jan 08 15:12:00 zecke: yeah Linux Space is supported Jan 08 15:12:15 pb_: well I have failed it once already... Jan 08 15:12:33 zecke: give up on Uni and become a Life Long Thousand Parsec hacker :P Jan 08 15:12:41 zecke: heh, oh dear Jan 08 15:12:46 pb_: for this guy an oral exam is... letting me write down everything Jan 08 15:13:12 alternatively he writes down half of the stuff I said.. Jan 08 15:13:30 zecke: well, let's hope you break your failing habit eventually Jan 08 15:14:23 pb_: next term another professor is holding this lecturer... Jan 08 15:15:42 pb_: what was wrong with subversion? Jan 08 15:15:42 luckily I have no exams tomorrow, though our CFO will be back from holiday so I must be on my best behaviour Jan 08 15:16:06 zecke: kicking him in the nads wouldn't have helped i'm assuming? Jan 08 15:16:30 mithro: well it was his first oral exam... and if he is not rwading this log... Jan 08 15:16:36 * mithro doesn't like subversion :P Jan 08 15:16:40 oral exam == written_exam/3 Jan 08 15:17:09 what don't you like about svn? Jan 08 15:17:10 zecke: aside from not being able to convert strings to utf-8, nothing I know of Jan 08 15:17:15 * pb_ dinner time now Jan 08 15:17:16 bbiab Jan 08 15:17:25 but it's more of a the same way I don't like raw tomatos rather then a real valid rational dislike Jan 08 15:17:43 heh Jan 08 15:17:46 plus i prefer real distributed SCMs Jan 08 15:17:52 I get a long fine with svn Jan 08 15:18:11 for what it is Jan 08 15:18:27 pb_: ah right... Jan 08 15:18:33 SVN == CVS++ and after using distributed SCMs I no longer feel the need for centrailsed ones ;) Jan 08 15:18:37 * zecke hates his body for physical disabilities... Jan 08 15:18:40 heh Jan 08 15:18:47 svn is good at what it tries to do Jan 08 15:19:30 it seems like the only distributed system that is good seems to be bitkeeper ... Jan 08 15:19:43 and monotone ;) Jan 08 15:20:07 well cya tomorrow Jan 08 15:20:10 heh Jan 08 15:20:26 monotone has its rough edges it seems Jan 08 15:21:06 No Rev's lost since 365 one day(s)... Jan 08 15:21:11 just like in Futurama Jan 08 15:21:49 I'm quite enjoying darcs Jan 08 15:22:24 mithro: I have messed it up once already. killing it will pulling Jan 08 15:22:34 but I'm not really sure if it was really messed up Jan 08 15:22:46 cya Jan 08 15:22:54 mithro: and don't get lost in europe Jan 08 15:22:55 ! Jan 08 15:37:08 In python how does one add one list to the end of another to form one larger list? Jan 08 15:37:45 RP: + Jan 08 15:37:59 who ever would have thought :p Jan 08 15:38:47 ~herring treke Jan 08 15:38:48 * ibot whacks treke on the side of the head with a large red herring named alfred Jan 08 15:39:31 treke: Having watched it create a list entry containing the second list when I didn't expect it, I decided it easier to ask :) Jan 08 15:39:36 pb_: thanks Jan 08 15:40:28 I do not get on with python... Jan 08 15:41:43 I've also found the killer bug in bitbakes new found rdepends handling :-( Jan 08 15:42:20 rdepends of dynamically generated packages are near impossible to obtain Jan 08 15:46:02 I'm hopeful this won't affect our current metadata but we'll see... Jan 08 15:57:52 pb_: Can you define what you mean by an epoch bump please? Jan 08 15:58:41 RP: incrementing the epoch part of the version number Jan 08 15:59:07 i.e. from 0.5 to 1:0.4 Jan 08 15:59:28 ah, ok. thanks Jan 08 16:00:03 * RP has a few thousand PV's changes to review Jan 08 16:00:49 or what will feel like a few thousand anyway... Jan 08 16:01:23 heh Jan 08 16:02:54 * reenoo_ puzzled Jan 08 16:03:48 RP: you really want to change PV = "0.0cvs..." into PV = "1:0.0+cvs..." then? Jan 08 16:04:33 reenoo_: Where does "+" come in the sort order? Jan 08 16:05:03 well. this is ascii isn't it? Jan 08 16:05:14 reenoo_: That's what I'm asking... Jan 08 16:06:33 If it is ascii, I have a lot of work to do... Jan 08 16:07:23 pb_: happen to know about that offhand? I'd have to look it up in the debian docs or read the code in bitbake/ipkg Jan 08 16:07:46 I think it is ascii, but a moment's experimentation with "ipkg compare-versions" should reveal the truth. Jan 08 16:08:33 RP: I don't know who suggested a common standard would be absolutely neccessary. I for one would know better things to spend my time with Jan 08 16:09:29 PV will differ anyway either due to the epoch or the additional/missing '+' Jan 08 16:09:47 reenoo_: The problem is at the moment upgrades involving cvs versions are invariably broken. This was an attempt to fix that Jan 08 16:10:07 broken? how so? Jan 08 16:11:06 I was lead to believe a number of the version numbers formats we've been using don't work cleanly with upgrades Jan 08 16:11:52 Picture having cvs20050607 installed and wanting to install version 2.3.4 Jan 08 16:12:21 that's a different story Jan 08 16:12:29 in what way? Jan 08 16:12:39 i.e. it's not a cosmetic thing Jan 08 16:12:57 no. which is why some of the changes were made Jan 08 16:14:32 right. but it's rather tedious to fish them out of the hundreds of 0.0cvs -> 0.0+cvs changes, not to speak of the CVSDATE -> SRCDATE ones which don't modify the format Jan 08 16:15:12 reenoo_: Do I have any choice in that matter? Jan 08 16:15:37 you could have checked that stuff in as separate changesets Jan 08 16:16:24 in retrospect, yes Jan 08 16:16:51 in general that's how you use revision control systems Jan 08 16:17:22 i.e. don't touch misc stuff as part of mass changes Jan 08 16:17:38 thanks for telling me, I'd never have known otherwise Jan 08 16:21:13 to be fair, both were mass changes of one thing i.e. PV. In retrospect making them together was wrong Jan 08 16:23:26 RP: sorry, no offense intended. it's the mere size of that changeset that's a bit frustrating ;) Jan 08 16:23:52 reenoo_: Believe me, I'm more fustrated :-( Jan 08 16:24:12 anyways. I have an appointment at 8am. off to bed for now Jan 08 16:24:15 'night all Jan 08 16:24:22 'night reenoo_ Jan 08 16:24:28 and its my fault so I will fix. Probably needs doing ASAP as well :-/ Jan 08 16:26:20 NOTE: dependencies for glibc are: autoconf-native automake-native libtool-native gnu-config-native ipkg-utils-native quilt-native virtual/arm-linux-gcc-initial linux-libc-headers l i n u x - l i b c - h e a d e r s - d e v Jan 08 16:26:29 ERROR: dependency l (for glibc) not satisfied Jan 08 16:26:37 It made me smile anyway :) Jan 08 16:28:39 heh Jan 08 17:38:39 There are 111 problematic PV changes as far as I can see... Jan 08 17:40:59 Would now be a good time to mention that ipkg doesn't understand epochs? Jan 08 17:44:48 Yeah, probably better now than after it's done. Jan 08 17:47:58 Anyhow, according to debian version policy, 0.0cvs20050607 sorts lower that 0.0+cvs20050607 Jan 08 17:48:08 Letters sort lower than non letters Jan 08 18:10:50 Actually there's no problem. ipkg works and "+" does sort higher than letters so my changes don't break things Jan 08 21:38:19 hrm, something has degraded in openzaurus-unstable since July: Jan 08 21:38:30 modules for my wifi card are no longer autoloaded Jan 08 22:22:45 I want to compile a driver for my zaurus 5500 running OZ 3.5.3, how do I cross compile? I'm running linux on my PC right now if that helps Jan 08 22:23:03 mnm: install OpenEmbedded Jan 08 22:23:11 and BitBake Jan 08 22:23:13 ? whats that Jan 08 22:23:29 http://openembedded.org Jan 08 22:23:51 oh ok... um thanks Jan 08 22:24:18 you can use it to build an entire OS, if you want =p Jan 08 22:24:25 or just compile a kernel module or app Jan 08 22:24:59 how easy is this is it just install and go or is there a lot of confing Jan 08 22:25:19 a bit of configuration Jan 08 22:25:54 no messy gcc version match-ups and such? Jan 08 22:26:08 all-in-one right? Jan 08 22:26:23 it compiles its own GCC Jan 08 22:26:31 nice.. Jan 08 22:26:53 so its ALL intergrated then Jan 08 22:27:12 sort of. Jan 08 22:27:16 you just configure what output you want, and run 'bitbake gpe-image' or such Jan 08 22:27:22 it will build all the compilers, etc it needs Jan 08 22:27:31 very cool Jan 08 22:27:32 not to mention all the apps Jan 08 22:27:44 gpe-image will result in a flashable firmware Jan 08 22:27:53 if you use the org.openembedded.oz354fam083 branch it will mostly work. Jan 08 22:27:55 so it just spits out the correct binarys for the kernel ready to go!? Jan 08 22:28:04 wingel: it will mostly work w/ .dev Jan 08 22:28:05 if you're lucky... Jan 08 22:28:18 mnm: it builds the kernel too, yes Jan 08 22:28:27 mnm: if you're looking to change kernel config, you might need to poke at it Jan 08 22:28:28 very very nice Jan 08 22:28:33 nah Jan 08 22:28:50 I just gonna compile a patched hostap driver module Jan 08 22:29:23 you'll need to poke to patch, too Jan 08 22:29:45 ok well thanks for your help I'm gonna mess around with it tommorow then... g'night Jan 08 22:29:49 in particular, you'll need to add the patch to the kernel bb rules Jan 08 22:30:15 would i have to recompile the kernel? Jan 08 22:30:48 you'd need to edit OE to include your patch Jan 08 22:30:53 and bitbake will recompile it Jan 08 22:31:08 yes, you need to use the same version of gcc to compile the kernel and the modules. Jan 08 22:31:10 but i already patched and just got the C files ready too go Jan 08 22:31:34 do i still have too Jan 08 22:32:00 mnm: you have the other Zaurus-required patches? Jan 08 22:32:17 bitbake will only work with bb rules Jan 08 22:32:20 what "other Zaurus patches" Jan 08 22:32:28 display drivers, etc Jan 08 22:32:32 oh yeah Jan 08 22:32:40 I installed the OZ distro Jan 08 22:32:59 but you'll need to recompile the entire kernel Jan 08 22:33:04 oh no! Jan 08 22:33:06 even if you don't install the recompiled one Jan 08 22:33:12 thats what i was afraid Jan 08 22:33:13 of Jan 08 22:33:18 kernel compiles are quick Jan 08 22:33:30 btu daunting Jan 08 22:33:34 no Jan 08 22:33:39 bitbake linux-openembedded Jan 08 22:33:40 *wait* Jan 08 22:33:41 done! Jan 08 22:33:45 hmm Jan 08 22:33:53 I'll have to check it out tommorow Jan 08 22:33:53 err linux-openzaurus Jan 08 22:34:07 thnx buh-bye Jan 08 22:53:08 03pH5 07org.oe.dev * r304de750... 10/packages/avahi/avahi_0.6.3.bb: Jan 08 22:53:08 avahi: bugfix release 0.6.3 Jan 08 22:53:08 - http://lists.freedesktop.org/archives/avahi/2006-January/000407.html Jan 08 22:53:13 03rpurdie 07org.oe.dev * r56c5a634... 10/ (56 files in 19 dirs): Add PACKAGES_DYNAMIC to the appropriate .bb files. Jan 08 22:53:18 03rpurdie 07org.oe.dev * r391c051d... 10/packages/meta/ (20 files): Jan 08 22:53:18 DEPENDS cleanup of packages/meta. This doesn't fix all DEPENDS here but does Jan 08 22:53:18 remove a large chunk of the incorrect ones. Some entries that shouldn't be Jan 08 22:53:18 present have been marked FIXME as they indicate problems elsewhere in the Jan 08 22:53:18 metadata. Help from the opie/gpe gurus needed. I've not touched the NSLU code as Jan 08 22:53:19 I suspect they'd rather do this themselves? Jan 08 22:55:54 03rpurdie 07org.oe.dev * r8ecfb383... 10/ (48 files in 7 dirs): Jan 08 22:55:54 Following from the meta/packages changes, cleanup conf/machine for incorrect Jan 08 22:55:54 DEPENDS usage. Move certain things into EXTRA_IMAGEDEPENDS. No valid users of Jan 08 22:55:54 EXTRA_BOOTSTRAP_DEPENDS were found so it was removed entirely. Change Jan 08 22:55:54 zaurus-updater not to package as it doesn't provide any. Jan 08 23:24:42 hi, I want to get madwifi from svn, but its uri is "http://svn.madwifi.org/truck", what should i write for SRC_URI in a .bb file to tell bb that i don't want to use wget but svn? Jan 08 23:26:53 i am guessing the same way you do for cvs Jan 08 23:28:59 packages/uclibc/uclibc_svn.bb:SRC_URI += "svn://uclibc.org/trunk;module=uClibc" Jan 08 23:29:43 courtesy of the magic of "grep -R svn packages/ " Jan 08 23:32:59 emte: so what should I do? I'm confused with how bitbake knows what tool I want to use to get source from Internet. Jan 08 23:35:02 emte: if I use SRC_URI="http://svn.madwifi.org/trunk;module=madwifi-ng", it looks like bb treat it as a http protocol-based url and use wget to get the file? Jan 08 23:35:08 the URI tells it what to do Jan 08 23:35:18 yes because your telling it to Jan 08 23:35:44 http is hypertext Jan 08 23:36:23 SRC_URI="svn://svn.madwifi.org/trunk;module=madwifi-ng" Jan 08 23:36:32 as per the abouve example Jan 08 23:37:53 emte: http://madwifi.org/wiki/UserDocs/GettingMadwifi metioned that I should use "svn http://svn.madwifi.org/trunk madwifi-ng" to get it. Jan 08 23:38:53 i am not sure bitbake is smart enough to parse that properly Jan 08 23:39:06 i do not see any examples like that Jan 08 23:41:16 so I think I should use "http://svn.madwifi.org/trunk" not "svn://svn.madwifi.org/trunk", since if I use later as SRC_URI, bb use wget to get a index.html file, :-x Jan 08 23:42:17 you do understand hoe bitbake parses the SRC_URI right? Jan 08 23:42:25 how Jan 08 23:42:45 http means hypertext so bitbake uses wget Jan 08 23:43:08 cvs:// menas cvs so bitbake uses cvs Jan 08 23:43:11 means* Jan 08 23:44:24 * emte guesses not Jan 08 23:45:27 damn you familiar image Jan 08 23:45:40 why do you give options for modules that dont exist ... Jan 08 23:47:29 grr Jan 08 23:47:34 * emte shouts at ipaq Jan 08 23:48:10 emte: not sure. Jan 08 23:49:30 cmon stupid thing talk to the network Jan 08 23:50:37 emte: so my question is how to tell bb that I want to use svn when I use SRC_URI="http://svn.madwifi.org/trunk;module=madwifi-ng" Jan 08 23:50:43 * emte wonders if i can rollback monotone to the 8.2 release Jan 08 23:50:58 SRC_URI="svn://svn.madwifi.org/trunk;module=madwifi-ng" Jan 08 23:51:36 or you can try what they suggested but i think it will give you an error Jan 08 23:51:41 SRC_URI="svn http://svn.madwifi.org/trunk;module=madwifi-ng" Jan 08 23:51:59 emte: svn: Can't connect to host 'svn.madwifi.org': Connection refused Jan 08 23:52:40 is it working? Jan 08 23:52:51 and does it accept anonymous connections Jan 08 23:52:55 emte: trying... Jan 08 23:53:48 emte: SRC_URI="svn http://svn.madwifi.org/trunk;module=madwifi-ng" --> bb parse error :( Jan 08 23:54:33 SRC_URI = "svn+ssh://svn.berlios.de/bitbake/trunk;module=bitbake" Jan 08 23:54:34 hmm Jan 08 23:54:42 SRC_URI="svn+http://svn.madwifi.org/trunk;module=madwifi-ng" Jan 08 23:54:47 might work ... Jan 08 23:55:57 otherwise a new override will have to be written i guess Jan 08 23:58:00 emte: failed again. Jan 08 23:59:11 connection or BB Jan 09 00:06:58 emte: BB do_unpack failed. Jan 09 00:06:59 emte: but BB don't get anything from network. Jan 09 00:07:22 what do your logs say Jan 09 00:17:08 emte: SRC_URI = "svn://svn.o-hand.com/repos/matchbox/trunk;module=${PN};proto=http" maybe this is the right format Jan 09 00:17:35 looks as good as any of the others Jan 09 01:42:08 good morning all Jan 09 01:44:02 morning alan Jan 09 01:44:23 somebody knows how i can switch from gpe to text console? Jan 09 01:44:30 chvt 1 Jan 09 01:44:49 and chvt 2 to come back to gpe Jan 09 01:46:00 mmh some keys? FN right/left doenst work Jan 09 01:46:59 from text console to GPE, Fn+ right works. Jan 09 01:47:12 ah ok. fine Jan 09 01:47:27 and is it possible to start more gettys for more consoles? Jan 09 01:48:07 houlaaa... *cough*... i'm sure it is possible... Jan 09 01:48:23 but i don't know how... Jan 09 01:50:26 * france is back (gone 35:56:27) Jan 09 01:51:58 IIRC that's a kernel thing. You can configure the number of VT's with "make menuconfig" if I'm not totally mistaken Jan 09 01:52:30 morning Jan 09 01:52:40 hi XorA Jan 09 01:56:13 morning XorA france CoreDump|home Jan 09 01:57:24 Alan_K: good morning Jan 09 01:58:38 ok Jan 09 01:58:45 hi have an "unable to open conf/bitbake.conf" error. why that ? should i copy the bitbake.conf from .../OE/org.openembedded.dev/conf/ to .../OE/build/conf ? Jan 09 01:59:14 or did i forget to export something ? Jan 09 02:01:42 ho! ho! XorA, by the way ! you told me i should add PREFERRED_PROVIDER_virtual/libsdl = "libsdl-x11" to my local.conf. Should i write it instead of others PREFERRED_PROVIDER setings ? Or should that be one more line ? Jan 09 02:02:47 alan|laptop: one more line Jan 09 02:02:54 and yes to local.conf Jan 09 02:03:08 ha. Ok, thanks a lot XorA Jan 09 02:04:00 and do you know about the bitbake.conf error ? Jan 09 02:04:19 morning all Jan 09 02:04:25 alan|laptop: sounds like you forgot BBOATH Jan 09 02:04:32 BBPATH Jan 09 02:04:55 XorA: i exported that... i may have done an error in this process, then... Jan 09 02:05:03 thanks. That will help me Jan 09 02:05:18 morning RP Jan 09 02:05:45 greentux: To switch from gpe using FN left/right, switch to the raw mode keyboard first using SysRq+R, then Fn+Left/Right will work Jan 09 02:06:41 morning RP Jan 09 02:06:41 tnx RP Jan 09 02:08:04 XorA: thank you. I think i misspelled something, even though i don't know why... ^^' Jan 09 02:08:55 s/why/what Jan 09 02:26:04 <[lala]> morning Jan 09 02:26:26 <[lala]> RP: which is the sysrq key on zaurus keyboard? :) Jan 09 02:27:20 [lala]: on c7x9 Home/Power On key Jan 09 02:27:24 c7x0 Jan 09 02:27:39 morning all Jan 09 02:27:52 <[lala]> XorA: i have a cx00 Jan 09 02:28:05 <[lala]> aem, borzoi i mean Jan 09 02:28:25 <[lala]> cxx00 that is Jan 09 02:40:47 [lala]: http://openzaurus.org/wordpress/howto/c7x0-and-cxx00-26-kernel-information/ - I just wrote this as I get tired of explaining it :) Jan 09 02:43:32 RP: ok, one more thing to translate ^^ Jan 09 02:43:35 <[lala]> excellent :) Jan 09 02:44:57 I get the feeling there is something else I need to document but I can't think what it is at the moment... Jan 09 02:47:32 <[lala]> RP: i tried hooking up an usb standard keyboard on borzoi but i couldn't load a sensible keymapping with loadkeys. doesn't it work this way? Jan 09 02:48:30 [lala]: The standard keyboard's keymap will certainly be incompatible with that of an external keyboard Jan 09 02:48:50 [lala]: To be honest, I've never tried a USB keyboard with the device Jan 09 02:49:05 [lala]: Do you need a special cable for that? Jan 09 02:49:16 <[lala]> RP: yeah, that's why i tried loading a different keymap with loadkeys Jan 09 02:49:57 <[lala]> NAbyss: only the usual usb adapter (mini-usb -> normal) Jan 09 02:50:06 [lala]: ah, okies Jan 09 02:50:40 [lala]: We you trying to use the keyboard under opie/gpe or on the console? Jan 09 02:50:51 <[lala]> RP: console Jan 09 02:51:21 [lala]: I'd have expected that to work :-/... Jan 09 02:51:24 <[lala]> but i better try again cause it's already some weeks ago i tried last Jan 09 02:51:30 <[lala]> just a moment **** ENDING LOGGING AT Mon Jan 09 02:59:57 2006