**** BEGIN LOGGING AT Thu Jan 12 03:00:21 2006 Jan 12 03:53:58 gpe-image works with -r6 :) Jan 12 04:05:29 RP: Im upto gcc-cross :-) Jan 12 04:06:09 umm are there problems w/ ewi546.ewi.utwente.nl amt? I' getting "monotone: read from fd 9 (peer ewi546.ewi.utwente.nl) failed, disconnecting" when trying to push Jan 12 04:06:25 s/amt/atm Jan 12 04:13:44 CoreDump|home: Did you have your write access renabled after updating your database? Jan 12 04:15:45 ahh no Jan 12 04:16:24 who do I need to talk to for that as mickeyl is away? koen I guess? Jan 12 04:19:17 Just a question... Is it correct that bootstrap-image has /boot empty?? Jan 12 04:20:12 doppiaemme: That would depend on the machine Jan 12 04:21:26 RP: I bitbake bootstrap-image for VIA epia machine, it compiled ok but /boot is empty... Jan 12 04:24:35 doppiaemme: I'm not familar with the epia. Should it have a kernel in there or is the kernel loaded in some other way? Jan 12 04:26:31 RP: this is what I don't understand... for now I've a bootstrap-image, putting the rootfs on a CF and booting from it isn't enough... I know it's a newbie problem but I can't find any docs... Jan 12 04:27:46 doppiaemme: I don't know enough about the epia to offer any useful advice I'm afraid Jan 12 04:28:28 <_law_> is here someone, how has an clamshell where hostap drivers work without loading modules by hand once? Jan 12 04:28:53 <_law_> or g_ether Jan 12 04:29:09 RP: no problem, thanks :-) Jan 12 04:29:27 morning all Jan 12 04:29:27 hi liam Jan 12 04:29:43 RP thanks for that mail :) Jan 12 04:29:54 _law_: g_ether shouldn't autoload as you might want one of the other gadgets Jan 12 04:30:27 _law_: hostap should autoload but doesn't after a reboot - removing and inserting the card usually wakes it up Jan 12 04:30:44 lrg|home: np. I noticed you fell out the cc again :) Jan 12 04:30:47 _law_: hostap works fine here Jan 12 04:30:54 _law_: g-ether needs to be modprobed Jan 12 04:31:23 <_law_> RP, i always have to modprobe after reboot. i have got an netgear wlan card Jan 12 04:31:29 <_law_> will try my asus later Jan 12 04:31:53 RP: can you forward and I'll reply and remind him I'm the contact Jan 12 04:33:29 lrg|home: If it happens again, I will do Jan 12 04:34:13 RP: thanks, I guess thats another beer I'm due you a Fosdem. I'm losing count now :) Jan 12 04:35:20 lrg|home: Looks like I won't be very sober at fosdem either :) Jan 12 04:35:30 heh Jan 12 04:39:28 Hi Liam Jan 12 04:39:40 hi Dirk Jan 12 04:41:17 lrg|home: I haven't tested wm97xx_register_misc_dev. It was too late yesterday. Jan 12 04:42:07 hi Jan 12 04:42:12 CoreDump|home: lart you ;) Jan 12 04:42:23 do13: np, do you know Ole Reinhardt at kernelconcepts. He's doing some testing with the driver today using the wm9705 Jan 12 04:42:32 zecke: hey Jan 12 04:42:33 ? Jan 12 04:43:08 CoreDump|home: you broke the tailorization ;) Jan 12 04:43:22 I don't get it Jan 12 04:43:44 hi zecke Jan 12 04:43:58 CoreDump|home: "Akita/Spitz: Use OE f?r setting CMDLINE, not defconfig. Jan 12 04:43:59 " Jan 12 04:44:09 zecke: My builds all worked out with the bitbake patch so its looking good. I've created r7 due to a typo Jan 12 04:44:12 CoreDump|home: breaks importing the cset into a svn repository Jan 12 04:44:18 do13: he's testing wm9075 aux adc and touch at the same time on a custom platform. should be interesting to see if he has the pen down issue. Jan 12 04:44:19 lrg|home: I read a post on lakml Jan 12 04:44:27 http://www.rpsys.net/openzaurus/temp/bitbake-rproviders-r7.patch Jan 12 04:44:30 I don't even remember pushing that lol Jan 12 04:45:33 do13: my trip to Shanghai is now on the 3rd Feb, so I would hope we can have it all working before your trip to Spain Jan 12 04:46:01 CoreDump|home: I'm currently converting our monotone db to Git, Darcs and SVN Jan 12 04:46:09 lrg|home: Did you tosa arrived? Jan 12 04:46:11 i c Jan 12 04:46:45 RP: I will leave at 14:00 torwards the office. I will do a build as well and then commit the patch Jan 12 04:46:52 but what does OE care about that? Jan 12 04:47:14 do13: not yet, however snail mail is slow in the UK. :( Hopefully get it by the weekend. Jan 12 04:47:15 lrg|home: I'm next week on vacation Jan 12 04:47:40 lrg|home: you in the uk? Jan 12 04:47:54 lrg|home: So we have this weekend :) Jan 12 04:48:00 do13: np, I'll probably have it fixed on your return Jan 12 04:48:21 XorA: yes, Dunfermline Jan 12 04:48:33 lrg|home: we must get together, Im edin based Jan 12 04:49:34 anyway,gottagoto work, later Jan 12 04:49:43 XorA: do you go to edlug ? Jan 12 04:50:14 lrg|home: No, Ive been in holyrood on same night as it a few times 3-4 years ago Jan 12 04:51:32 XorA: ah, I recommend coming along one night. Incinerator also goes. Do you work in Edin ? Jan 12 04:51:45 lrg|home: upstairs from Virgin Megastore Jan 12 04:52:00 lrg|home: I shall have to try and make it sometime Jan 12 04:52:12 lrg|home: Ime sure I owe you a beer for the sound driver :-) Jan 12 04:52:35 XorA: I work at Wolfson next to Murrayfield. I'm probably due you a beer as well :) Jan 12 04:52:56 lrg|home: How far is Dunfermline away from Salisbury? Jan 12 04:53:13 XorA: if you were a student, I could have offered you a summer placement to linux audio Jan 12 04:53:28 lrg|home: been a long time since I was a student :-) Jan 12 04:53:47 do13: oh about 500 miles (english geog not very good) Jan 12 04:54:10 lrg|home: thats about right Jan 12 04:54:15 lrg|home: the distance that is Jan 12 04:54:27 lrg|home: Ok. It's too far away. Jan 12 04:55:01 do13: are you going to Salisbury ? Jan 12 04:55:19 lrg|home: We have an office in Salisbury Jan 12 04:55:43 do13: ah, pb_ lives closer to Salisbury. He lives in Cambridge Jan 12 04:56:44 lrg|home: Thanks Jan 12 04:57:04 lrg|home, RP you will visit FOSDEM? Jan 12 04:57:40 do13: yes, I've still to make my hotel arrangements. It would be nice if we were all booked into the same hotel Jan 12 04:59:05 do13_: Yes, that's the plan Jan 12 04:59:18 lrg|home: You wil arrive on 24th? Jan 12 05:00:06 do13: I've not booked any travel yet, so I'm flexible Jan 12 05:01:28 I'll look at a flight. Maybe we can met. Jan 12 05:02:21 do13: yes, that would be cool. I'll book my travel to be there on the 24th. Jan 12 05:02:41 I suspect I'll arrive midday on the 24th Jan 12 05:02:49 but nothing's booked yet Jan 12 05:03:12 RP: cool, let me know if your hotel and I'll try and book the same Jan 12 05:03:55 First of all I have to fight with my wife:) Jan 12 05:04:06 lrg|home: will do. Its in my boss' hands atm. Jan 12 05:10:25 RP: Let me know your hotel Jan 12 05:10:25 koen|away: ping! Jan 12 05:12:06 koen|away: dpkg-reconfigure locales and please add german! Jan 12 05:27:01 hi i know that i am still asking the same questions but i am still trying to compile php for my openslug and i always get this error : http://pastebin.ca/36581 Jan 12 05:30:09 kami22: Its trying to use libs in /usr/lib instead of staging. Jan 12 05:30:39 okay but how can i fix this? Jan 12 05:30:43 easily :) Jan 12 05:31:03 Find out why it passes -L/usr/lib into that commandline Jan 12 05:31:07 It shouldn't be Jan 12 05:31:39 but i didnt change anything to the standard bb files Jan 12 05:31:48 cant you try to compile php? Jan 12 05:32:15 maybe there is a bug in the bb file? Jan 12 05:32:53 kami22: There is a bug either in the .bb or in php itself which is breaking when cross compiling. Someone needs to work out why and fix it Jan 12 05:34:03 yes but i think i am not so intelligent to do this on my own Jan 12 05:35:02 kami22: You can learn ;-) Jan 12 05:35:34 darcs and git are up-to-date again Jan 12 05:36:00 RP: maybe i thousand years :) Jan 12 05:36:05 zecke: I'm still trying to remember how to use git again :) Jan 12 05:36:22 RP: cg-clone, cg-update ;) Jan 12 05:36:28 RP: cg-log Jan 12 05:36:35 cg- Jan 12 05:36:49 zecke: I have raw git installed :) Jan 12 05:37:00 that is disgusting Jan 12 05:37:11 git clone ? Jan 12 05:37:15 git update? Jan 12 05:37:17 git log? Jan 12 05:37:40 near enough. but then try switching branches :) Jan 12 05:37:51 kami22: do you fail turing tests? Jan 12 05:38:01 RP: cd ../other-branch ; cg-update? Jan 12 05:38:22 git branch xyz; git checkout -f abc Jan 12 05:38:40 The -f had me puzzled for a while :) Jan 12 05:38:45 kami22: if not I'm confident you can fix php's build system Jan 12 05:39:04 I should try cognito :) Jan 12 05:40:39 hehe Jan 12 05:43:55 kami22: I think unless you are willing to roll up your sleeves and look into problem yourself its going to be a long while before anyone here does Jan 12 05:48:42 sorry but i think i dont know what the problem is to compile php for the slug it cant be that difficult to get it working? Jan 12 05:49:29 kami22: the problem is zecke, RP etc are all working on other stuff, there isnt enough man power about to fix every problem in every bb, someone will eventually do it when they need php Jan 12 05:50:39 kami22: you are spending at least three days on that issue already. What about putting some energy/brain power into it? Jan 12 05:51:03 anyway cya later Jan 12 05:56:33 i am trying everything now but i cant get a solution and i am not a programmer Jan 12 06:13:05 RP: http://pastebin.com/502284 probably ny fault somewhere Jan 12 06:43:40 XorA: This is latest OE metadata with the -r7 bitbake patch and a bitbake gpe-image? Jan 12 06:47:18 What does the class debian do`? Jan 12 06:49:02 Cobelius: Renames packages in a debian like way Jan 12 07:06:57 RP: yes to all Jan 12 07:08:33 re Jan 12 07:10:43 XorA: Not good then. I don't understand it as everything worked ok here :-/ Jan 12 07:12:00 XorA: Try running "bitbake -D -n gpe-image" and I'll have a look for any problems Jan 12 07:18:05 RP: http://www.xora.org.uk/buildlog Jan 12 07:23:00 XorA: Was this from a clean tmp dir? Jan 12 07:23:05 RP: yes Jan 12 07:26:13 XorA: There is no mention of gpe-task-* in there and there should be :-/ Jan 12 07:27:54 RP: dropped too many depends somewhere? Jan 12 07:28:16 hi all Jan 12 07:29:25 hi florian_kc Jan 12 07:29:58 * florian_kc finally decided to repair his OE devel stuff Jan 12 07:30:04 XorA: looks a bit like it somehow. I get the same results from bitbake -n -D here which is somewhat puzzling as the build worked earlier... Jan 12 07:30:47 RP: well at least I can quickly test any fix Jan 12 07:33:36 probably down the side of the couch with all the other missing stuff Jan 12 07:34:16 XorA: Could you rerun that with -D -D please? The thngs I'm interested in don't appear unless you do that :) Jan 12 07:34:45 morning Jan 12 07:35:20 morning chouimat Jan 12 07:35:27 hi chouimat Jan 12 07:39:36 Oh well.. merging in ~450 revisions takes a while. Jan 12 07:47:12 florian_kc: good decision. If you want something only read only you might try git or darcs Jan 12 07:49:46 RP: running Jan 12 07:51:56 zecke: well... since i need to fix it anyway i give it another try. i think i have quite some stuff to commit. Jan 12 07:52:10 btw. did koen|away commit my quilt fixes? Jan 12 07:52:26 RP: http://www.xora.org.uk/buildlog2 Jan 12 07:53:51 http://pastebin.com/501092 <-- why is the bbclass not being included? Jan 12 07:55:54 Luke-Jr: -DDD ??? Jan 12 07:57:25 zecke: nothing suggests any bbclass for it Jan 12 07:58:16 http://pastebin.com/502426 is w/ debug Jan 12 07:58:24 about 7 Ds, IIRC Jan 12 07:59:13 Luke-Jr: well the logging stills of bitbake are missing. I fear you will need to take it on with the parser Jan 12 07:59:27 Luke-Jr: just make it print a line on any file it tries to include Jan 12 07:59:52 we should change inherit's semantic to use require instead of include Jan 12 08:00:03 I set up a new computer for compiling OE. I think I followed the GettingStarted page in the wiki but I still get an error "ERROR: Unable to open conf/bitbake.conf". I am in the build directory and build/conf/local.conf exists. Any clues? Jan 12 08:00:29 Laibsch: echo $BBPATH sais what? Jan 12 08:00:34 zecke: ./lib/bb/parse/parse_py/BBHandler.py: debug(2, "BB %s:%d: inheriting %s" % (fn, lineno, file)) Jan 12 08:00:59 zecke: /usr/src/openembedded/build:/usr/src/openembedded/org.openembedded.dev Jan 12 08:01:42 /usr/src/openembedded/build/conf/local.conf exists and is readable Jan 12 08:02:03 Luke-Jr: good then either you misspelled the inherit or ... Jan 12 08:02:16 :\ Jan 12 08:02:47 Laibsch: then /usr/src/... has no conf/bitbake.conf?! Jan 12 08:03:35 zecke: it's parse-time: DEBUG: BB :0: inheriting classes/kde_split.bbclass Jan 12 08:03:44 DEBUG: CONF file 'classes/kde_split.bbclass' not found Jan 12 08:03:45 :( Jan 12 08:04:33 zecke: Oops, I should read more carefully :-[ Jan 12 08:06:21 * Luke-Jr stabs self Jan 12 08:07:12 XorA: There is not one mention of the word runtime in that log which makes me wonder if it is the same bitbake/OE metadata you're using :-/ Jan 12 08:07:50 XorA: It never says things like "Additional runtime dependencies" Jan 12 08:08:30 RP: ah pish, I just realised, its my fault Jan 12 08:08:57 argh... why won't bitbake find my bbclass Jan 12 08:08:58 ? Jan 12 08:09:12 RP: we need to kill the usage of bb.encodeurl as well?! (on the fetcher front) Jan 12 08:09:55 zecke: We should make a list of things we'd like to do to bitbake Jan 12 08:10:25 I'm currently adding means to bitbake to have a special 'now' meaning Jan 12 08:10:47 zecke: excellent - I think there's a bit in the git fetcher that would like that Jan 12 08:10:48 After I did a "monotone update" can I go back to prebreakage? I get an error now. Jan 12 08:10:59 zecke: I'd like to see gitruncmd become a general fetcher function Jan 12 08:11:09 RP: totally forgot on build machine I use at work I have an installed bitbake, I shall now start again from afresh Jan 12 08:11:19 RP: sorry for the bother Jan 12 08:12:42 XorA: I thought it was strange :) Jan 12 08:12:55 * RP resolve to patch bitbake's version in future patches Jan 12 08:13:41 RP: I owe you another beer now, thats twice this week Ive done that Jan 12 08:14:23 * RP wonders if he can take all this beer :) Jan 12 08:15:07 RP: s/beer/pop so OE doesn't lose a kernel dev ;) Jan 12 08:20:00 anyone here with admin access to the Monotone Database? Jan 12 08:21:10 | /home/luke-jr/src/oe2005/build.alt/tmp/work/kdelibs-dcop-native-3.4.1-r0/temp/run.do_unpack.23432: line 381: `kde_split,_do_unpack': not a valid identifier Jan 12 08:21:20 WTF? O.o Jan 12 08:29:32 Luke-Jr: not what you wanted to see? Jan 12 08:29:45 what's adding commas? Jan 12 08:29:48 RP: now those errors that I think you were talking of earlier, ERROR: Nothing provides runtime dependency python-native NOTE: no buildable providers for ipkg-utils-native Jan 12 08:32:16 XorA: Change the RDEPENDS in ipkg-utils-native to "" Jan 12 08:35:58 RP: just scrolled back to find that Jan 12 08:40:54 erm Jan 12 08:41:20 Luke-Jr: sounds like a ','.join to me.. but I have no clue where it might come from Jan 12 08:41:27 why does qt-x11-free-3.5-r10 stage into tmp/staging/x86_64-linux/qte3/ ? Jan 12 08:41:44 zecke: I was doing 'inherit native, kde_split', sorry Jan 12 08:42:09 but now the issue with qte3 vs qt3 Jan 12 08:42:11 Luke-Jr: that is a bug Jan 12 08:42:21 Luke-Jr: either the qt3.bbclass is broken Jan 12 08:42:35 Luke-Jr: the STAGING_INCDIR var for Qt gets set in a bbclass Jan 12 08:49:46 aha, correction Jan 12 08:49:53 it's the -native variant that is qte3 Jan 12 08:50:57 tho I can't see why: Jan 12 08:51:01 install -d ${STAGING_DIR}/${HOST_SYS}/qt3/lib Jan 12 08:54:05 err! Jan 12 08:54:20 I've never built the native variant... where does the qte3 dir come from? :x Jan 12 08:56:50 pH5: your shell patch has been applied Jan 12 08:57:03 Luke-Jr: uicmoc maybe? Jan 12 08:57:27 zecke: hrm, but shouldn't that be in qt3 instead? Jan 12 08:57:48 Luke-Jr: uicmoc3-native uses QtE3 ;) Jan 12 08:58:12 Luke-Jr: dunno why people insist on installing a static qte3 :} Jan 12 08:58:35 zecke: I wonder if that will be a problem :\ Jan 12 08:59:17 Luke-Jr: opie-lupdate uses that lib... Jan 12 09:01:59 x11-native is broken :\ Jan 12 09:05:04 RP: I've benn out for some weeks and I just read the changelog with your commits, I can only say congratulations!! Jan 12 09:05:54 http://pastebin.com/502518 <-- x11-native b0rked Jan 12 09:06:37 hi everyone Jan 12 09:06:41 zecke: thanks! Jan 12 09:10:56 katossi_uni: thanks :) Jan 12 09:10:58 hi pH5 Jan 12 09:12:21 hi RP Jan 12 09:15:20 Hmm. Are the monotone servers ok? Jan 12 09:15:50 vanille is refusing connections and ewi connects but doesn't do anything :-/ Jan 12 09:16:09 same here Jan 12 09:16:16 as soon as I type that, ewi wakes up :) Jan 12 09:16:32 it does? O.o Jan 12 09:17:45 RP: vanille is down for several hours now... Jan 12 09:18:38 I can't even connect to ewi now Jan 12 09:21:43 * florian_kc feels lucky ;-) Jan 12 09:21:50 is monotone.vanille.de down? Jan 12 09:22:05 as usual, yes Jan 12 09:22:06 ewi restarts automatically after some minutes. Jan 12 09:22:24 well mickeyl's irc bot dropped off as well Jan 12 09:22:25 bah, 0 revs in Jan 12 09:22:38 this could mean the whole Uni Frankfurt/DFG-Net could have ussues Jan 12 09:23:17 heh right... seems to be a similar situation like in Siegen :-/ Jan 12 09:23:57 I wonder how long it takes for the first department to drop DFN and use DSL ;-) Jan 12 09:24:20 DFN? O.o Jan 12 09:24:27 Luke-Jr: German Research Network Jan 12 09:24:44 Luke-Jr: Network in the sense of ISP Jan 12 09:26:40 RP: I feel like breaking bitbake fetcher tonight Jan 12 09:26:52 fdv: ping Jan 12 09:27:07 bedboi: ping Jan 12 09:27:09 zecke: I think I have some good ideas on how we could break it :) Jan 12 09:27:28 ugh, x11-native has been broken for almost a year? Jan 12 09:27:39 RP: I will move bb.encodeurl,bb.decodeurl to bb.fetch Jan 12 09:27:46 hi! Jan 12 09:27:54 RP: and then I will put mirroring into Fetch Jan 12 09:28:17 zecke: Sounds good. we could move gitruncmd there as well Jan 12 09:28:23 suitably renamed Jan 12 09:28:48 and make the svn/cvs fetchers use it instead of os.system Jan 12 09:28:56 Any Zaurus-Owners and -Builders here? Jan 12 09:29:31 RP: is it secure? that would be nice... Jan 12 09:29:51 zecke: Not so much secure but at least it correctly sets PATH... Jan 12 09:30:09 zecke: But making is secure would be a lot easier with all the os.system calls in one place Jan 12 09:30:25 Hrm... what did i do wrong here: | checking target system type... Invalid configuration `INVALID-linux': machine `INVALID' not recognized ? Jan 12 09:30:47 florian_kc: Change your DISTRO line to something valid Jan 12 09:30:55 probably familiar-unstable Jan 12 09:31:55 hwtechnik: yep Jan 12 09:32:40 RP: doh... let me see Jan 12 09:32:58 CoreDump: what model? Jan 12 09:33:12 Akita and Collie Jan 12 09:33:18 hwtechnik: There are a few of us around Jan 12 09:34:08 RP: Cool! At last. Jan 12 09:34:49 RP: Does anybody play around with kernel 2.6 on collie? Jan 12 09:35:56 hwtechnik: Its lagging behind the other 2.6 kernels :-(. John Lenz did a bit of work and more recently, Pavel Machek (from LKML) has been doing some. Jan 12 09:36:49 RP: hrm... familiar-0.8.3 should be valid for the familiar/oz stable branch. Jan 12 09:37:30 florian_kc: For the banch, yes. I assumed you were using .dev to see that. Its caused by TARGET_OS not being set... Jan 12 09:37:49 which usually means DISTRO isn't right as they set TARGET_OS Jan 12 09:39:00 RP: well... i'll invalidate the cache and try again, if it doesn't work then i'll continue to wonder ;-) Jan 12 09:39:07 RP: So, apart from them, nobody really works with 2.6? I try to get OE to build me a collie2.6 machine-config Jan 12 09:39:26 hwtechnik: correct Jan 12 09:39:54 hwtechnik: Pavel Machek's git tree is probably the most uptodate kernel for collie Jan 12 09:40:48 RP: And how would I get a working and compatible initial filesystem with his kernel? Jan 12 09:42:31 hwtechnik: I think the standard collie one is passable but we relly need someone to develope a proper 2.6 one (with udev and things) Jan 12 09:44:14 um, but of what use is a collie w/o SD card support? To many people next to none Jan 12 09:44:53 collie lacks CF? Jan 12 09:45:04 CoreDump: My current one is not used, so I dontīt need the SD-support anyway. Jan 12 09:45:20 I think I've seen a SD card once Jan 12 09:45:25 CoreDump|work: get the SanDisk Connect + working :-) Jan 12 09:45:28 that's about all my SD experience ;) Jan 12 09:46:09 XorA: Is there a problem with the SanDisk? Jan 12 09:47:50 Um, I was talking about a Collie running a 2.6 kernel. No SD support there Jan 12 09:48:05 RP: any way to change how the touchscreen works? Jan 12 09:48:28 RP: eg, have it only move the cursor and use OK/Cancel for left/right buttons Jan 12 09:48:44 hwtechnik: No-one seems to have got IDE and Wireless working at same time Jan 12 09:49:41 XorA: And whatīs with network-cards, CF? billionton, in my case. Jan 12 09:50:06 hwtechnik: all hostap supported ones work fine, the Sandisk works fine in wireless or IDE, just not both Jan 12 09:50:15 both would rock Jan 12 09:50:30 that would indeed rock Jan 12 09:51:45 XorA: Ahhh! So youīre talking about the one "special" dual-CF card from SanDisk. Jan 12 09:51:57 hwtechnik: yeah the Connect Plus :-) Jan 12 09:52:32 XorA: Finally gotcha ;) Jan 12 09:58:42 ~lart XorA_ for being on dodgy connection Jan 12 09:58:42 * ibot whacks XorA_ with a giant beaver's tail for being on dodgy connection Jan 12 10:02:33 generel question: Does openembedded support downloading directly from a git tree ? Jan 12 10:05:38 hwtechnik: bitbake does support that, yes Jan 12 10:06:26 CosmicPenguin: Sorry for mixing the two up. Am just curios. Jan 12 10:08:18 hwtechnik: thats ok, its easy to mix up the two - but the actual fetching is done by bitbake, so we need to give credit where credit is due Jan 12 10:09:03 CosmicPenguin: Roger that :) Jan 12 10:10:08 Has anybody succeeded in building a 2.6 kernel for the collie? I can definitely see the definition-files in there. Jan 12 10:10:25 But the build process always switches to 2.4. Jan 12 10:23:53 ahh now how did that work again. In theory the kernel is set with DISTRO and / or MACHINE. I believe there was a var you could set in local.conf to use the 2.6 kernel. Jan 12 10:24:15 like KERNEL = "2.6" or something like that Jan 12 10:24:37 sadly I can't check that ATM, Jan 12 10:26:49 check conf/distro/$your_distro.conf and conf/machine/$your_machine.conf for clues Jan 12 10:29:49 CoreDump: Thnx. Will look around there. Jan 12 10:53:35 re Jan 12 10:53:40 RP: at your order Jan 12 10:53:45 koen|away: ping Jan 12 10:56:13 zecke: ping Jan 12 10:57:11 wb zecke Jan 12 10:58:18 mithro: pong Jan 12 10:59:16 zecke: still in Budapest Jan 12 10:59:47 mithro: do you like Buda or Pest more? Jan 12 11:03:18 <[g2]> how many non NSLU2 devs have write access to the OE repo ? Jan 12 11:03:34 <[g2]> guesstimate is what I'm looking for Jan 12 11:04:06 zecke: i think i like Pest a little bit more Jan 12 11:04:19 hello Jan 12 11:04:56 where i find info starting with open embedded Jan 12 11:05:10 where i can find info starting with open embedded Jan 12 11:05:25 clorabit: oe.handhelds.org? Jan 12 11:05:56 I'm trying to figure out where to go next Jan 12 11:08:09 my choices are Zagreb, Ljubljana or Belgrade Jan 12 11:08:16 zecke, where i can look OE tree for online trough html ? Jan 12 11:09:25 clorabit: http://ewi546.ewi.utwente.nl/tmp/viewmtn/branch.psp?branch=org.openembedded.dev Jan 12 11:10:07 mithro: I have not traveled to any of these cities yet :( Jan 12 11:18:21 Zagreb is a 5.5 hour journey, Sarajevo is a 12 hour (overnight) journey, Belgrade is a 7 hour journey Jan 12 11:20:17 anyone know what a Schnellzug means? Jan 12 11:20:28 fast train Jan 12 11:21:12 doesnīt stop in all stations and is therefore faster at your destination Jan 12 11:21:16 ahh Jan 12 11:21:34 it's still pretty slow :P Jan 12 11:21:57 well. there is always a faster type of Schnellzug Jan 12 11:22:26 EC's are faster Jan 12 11:22:56 zecke, have u tried run dgen ? Jan 12 11:23:13 clorabit: no what is that? Jan 12 11:25:10 dgen is sega genesis emulator Jan 12 11:27:40 zecke, do u have experience compiling with crosstool ? Jan 12 11:30:50 clorabit: no it does not work on OS X (crosstools) Jan 12 11:31:12 what is athens in german? Jan 12 11:32:09 Athen Jan 12 11:32:19 zecke, do u know anyone expert with crosstool ? Jan 12 11:33:00 what is Greece in German? Jan 12 11:33:20 mithro: Griechenland Jan 12 11:33:38 mithro: or do you mean the language? Jan 12 11:33:40 whats Wurzburg ? Jan 12 11:33:49 mithro: dunno Jan 12 11:34:06 mithro: where is that? Jan 12 11:34:10 db.de can't figure out what I mean when I try in "Athen" or "Athens" Jan 12 11:34:15 http://reiseauskunft.bahn.de/bin/query.exe/en?ld=212.60&seqnr=1&ident=8a.02996018.1137089995&OK#focus Jan 12 11:34:31 mithro: just a sec Jan 12 11:34:53 Trying to go from Istanbul to Athens Jan 12 11:35:36 athen Jan 12 11:35:52 should work Jan 12 11:35:57 mithro: how do you change start and destination? Jan 12 11:37:25 hit the change button? Jan 12 11:37:50 http://reiseauskunft.bahn.de/bin/query.exe/en?ld=212.60&seqnr=8&ident=6k.02747718.1137089744&OK#focus Jan 12 11:37:53 try that one Jan 12 11:38:26 mithro: seems that this is only possible through germany. so you must choose a point in germany and from thereon get to Athene. Jan 12 11:38:55 [g2]: my guess would be all of them Jan 12 11:39:02 hrm.... Jan 12 11:39:10 I know you can go Istanbul -> Athens Jan 12 11:40:20 yeah. but not from that site. itīs supposed to be for travelers coming from or going to germany. Jan 12 11:40:42 hwtechnik: i've used it for every other trip Jan 12 11:41:03 ie Zagreb -> Belgrade etc Jan 12 11:42:14 hmm... obviously it doesnīt know Athen Jan 12 11:42:29 try Frankfurt -> Athens? Jan 12 11:44:04 I always get back Würzburg Jan 12 11:44:50 I wonder what that is Jan 12 11:45:27 both the Greece and Turkey rail lines websites are borked Jan 12 11:46:15 welp it nows about Thessaloniki Jan 12 11:46:48 funny. wait, trying to figure that one out. Jan 12 11:49:48 cu all Jan 12 11:49:54 RP: with bitbake -r7 I've got the behaviour again whereby slugos-image.bb RDEPENDS don't get build , but IPKG_INSTALL = "${RDEPENDS}" does have the correct packages, however only for BE builds... Jan 12 11:56:33 mithro: check this out: http://www.ose.gr/%287216627309326357%29/ecPage.asp?page=dromologia.asp&lang=2 Jan 12 11:57:45 mithro: even the greece site says: There is not direct line Jan 12 11:58:46 mithro: oh, and donīt forget, they call it Athina. Jan 12 11:59:29 mithro: I tried it with the Austrian Railway too, and they donīt know Athene as well. Jan 12 12:04:23 interactive bitbake dies if it has a parse error on a bb Jan 12 12:04:29 well, a bbclass anyway Jan 12 12:06:31 Luke-Jr: it really dies? Jan 12 12:09:19 zecke: drops to bash Jan 12 12:09:28 zecke: after dumping error msgs Jan 12 12:09:49 Luke-Jr: okay Jan 12 12:10:31 * Luke-Jr was kinda hoping it'd error and drop to a BB prompt Jan 12 12:11:23 !archive Jan 12 12:11:24 Laibsch: Don't know, don't care Jan 12 12:11:37 bbl Jan 12 12:12:22 !log Jan 12 12:12:58 !logs Jan 12 12:12:59 Laibsch: Dunno? Jan 12 12:13:13 Is there an archive somewhere? I don't want to ask the same question I got answered a couple of months ago but I do not have access to my personal logs right now. Jan 12 12:15:18 hwtechnik: it turns out there is a direct line Jan 12 12:15:25 http://www.seat61.com/Turkey2.htm#Turkey%20-%20Greece Jan 12 12:15:56 Laibsch: ibot.rikers.org IIRC Jan 12 12:16:39 mithro: funny. maybee all the train-railways havenīt updated their websites? Jan 12 12:16:51 hwtechnik: it could be a private line Jan 12 12:16:58 db.de only has "national" lines Jan 12 12:17:40 mitro: true. but this sentence "As from July 2005" ? Iīm not sure ;) Jan 12 12:18:02 true Jan 12 12:19:27 mithro: I could phone them :) Jan 12 12:21:57 [g2]: have a look in the .monotone/monotonerc on nudi - that has all the OE devs that currently have write access. Koen mailed out a list of all who *used* to have write access last week I think. Jan 12 12:24:01 zecke: yes, that is the one. Thank you. Jan 12 12:26:35 currently I'm thinking along these lines Jan 12 12:26:50 koen|away: Could you please re-enable write access to the mt db for coredump@handhelds.org? Jan 12 12:27:32 http://pastebin.com/502782 Jan 12 12:28:41 I guess i could cut out Zagreb - having some more days to play with in Athens/Istanbul might be nicer Jan 12 12:29:46 I'm having problems with bitbake, if someone could help me... Jan 12 12:30:05 zekoZeko: if you ask a question, maybe ;) Jan 12 12:30:06 i've done everything according to gettingstarted page on the wiki Jan 12 12:30:18 zecke: of course, that was my introduction only :) Jan 12 12:30:49 zekoZeko: you don't need one ;) Just one rule. If you post error messages and they are longer than two lines consider using pastebin Jan 12 12:30:51 * Luke-Jr waits for a question :) Jan 12 12:30:57 i copied local.conf.sample to build/conf/local.conf, and set BBFILES correctly. Jan 12 12:31:01 zecke: only consider? Jan 12 12:31:13 Luke-Jr: that is a turing test ;) Jan 12 12:31:14 zecke: what do you think? Jan 12 12:31:17 zekoZeko: and DISTRO, MACHINE, etc, I hope? Jan 12 12:31:26 of course. Jan 12 12:31:30 zecke: I don't know turing o.o Jan 12 12:31:32 but this is what i get: Jan 12 12:31:53 ERROR: no files to build. Jan 12 12:32:00 mithro: Zagreb? Nice town, historical. How is the climate in the winter? Jan 12 12:32:00 for bitbake task-bootstrap Jan 12 12:32:31 zekoZeko: did you set bbpath? Jan 12 12:32:39 Zagreb is not too cold now, around 0 in the morning, a few degrees above during the day. Jan 12 12:32:43 yes Jan 12 12:33:02 zekoZeko: what are you trying to build? Jan 12 12:33:12 ~pastebin Jan 12 12:33:14 rumour has it, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.com/, or http://pastebin.ca Jan 12 12:33:14 Luke-Jr: i'm trying to create a firmware for soekris 4826 Jan 12 12:33:29 i'll paste it right now, just a sec Jan 12 12:33:30 zekoZeko: yes, but what providee are you building for it? Jan 12 12:33:40 zekoZeko: could you do env, pwd, and put the output of bitbake task-bootstrap to pastebin? Jan 12 12:33:55 Luke-Jr: as he said ;) Jan 12 12:34:17 ah, missed that line, sorry Jan 12 12:34:39 RP: here's the bug (it's in bitbake I think), slugos-image.bb PROVIDES "${SLUGOS_IMAGENAME}-image" which, for DISTRO="ucslugc" is ucslugc-image, if I bitbake ucslugc-image the RDEPENDS are not build, if I bitbake slugos-image they are. Jan 12 12:34:54 zecke: i've got no idea - but I'm assuming it would be slightly better then here in Budapest Jan 12 12:35:19 the problem is going to be getting a reservation on the early train tommorrow Jan 12 12:35:46 * mithro ponders Jan 12 12:36:18 mithro: I would have chosen between Sarajevo and Zagreb Jan 12 12:36:57 zecke: and skip belgrade? Jan 12 12:37:19 mithro: no, I just know too little about belgrade ;) Jan 12 12:37:37 unbelievable the iron curtain is in my mind :( Jan 12 12:37:44 zecke: http://pastebin.com/502796 Jan 12 12:38:15 mithro: all cities must have this kind of post-war spirit Jan 12 12:38:28 mithro: progressing into a new age/epoch Jan 12 12:38:49 Sarajevo is the coldest of all, in case you were wondering :) Jan 12 12:39:01 I live in Ljubljana and I've been to all of these places :) Jan 12 12:39:03 the trip from Zagreb/Sarajevo -> Sofia sucks however :P Jan 12 12:39:11 zekoZeko: recomendations? Jan 12 12:39:46 zekoZeko: it does not load your local.conf ;) Jan 12 12:40:02 zekoZeko: I don't see a DISTRO set ;) Jan 12 12:40:22 zekoZeko: it is not able to find it. Your BBPATH would be guilty but it looks correctly Jan 12 12:40:53 where is local.conf? Jan 12 12:40:53 I did another step that is not in the GettingStarted Jan 12 12:41:02 I copied bitbake.conf to build/conf/bitbake.conf Jan 12 12:41:09 ..... Jan 12 12:41:13 otherwise it spat another error at me, I con't remember which Jan 12 12:41:25 can't Jan 12 12:41:43 zekoZeko: rm conf/bitbake.conf ;) Jan 12 12:41:50 zekoZeko: it shadows your error only Jan 12 12:41:55 i see Jan 12 12:42:22 choices choices Jan 12 12:42:32 ooh Jan 12 12:42:39 and now it works Jan 12 12:42:44 stupid me Jan 12 12:42:50 XD Jan 12 12:42:56 at least it's parsing the *.bb files now. Jan 12 12:43:02 jbowler: are you using RP's latest patches? Jan 12 12:43:13 mithro: Sarajevo is the nicest to see of them all I think Jan 12 12:43:23 mithro: the most ecclectic, although less so after the war Jan 12 12:43:39 mithro: and the "spirit of the war" that you were talking about is most present there Jan 12 12:44:04 actually, none of the ex-yu capitols has been hit by the war nearly as much as Sarajevo has. Jan 12 12:44:24 Are packages being uncompressed into ./tmp/work/ or somewhere else? Jan 12 12:44:44 Laibsch: uncompressed in what form? Jan 12 12:45:10 Laibsch: they are placed in the deploy dir e.g tmp/deploy/ipk Jan 12 12:46:26 zekoZeko: Sarajevo sounds cool, however it's the furthest out of my way and I'm already running out of time... Jan 12 12:46:42 so the questions are, should I goto or skip either Zagreb and/or Belgrade Jan 12 12:46:56 mithro: where are you and where are you going? Jan 12 12:47:20 mithro: of those two, I'd say skip Zagreb. Jan 12 12:48:27 I've been to Prague, Budapest, Vienna, Krakow, so it places are similar to them then I can skip em Jan 12 12:48:52 yeah, go to belgrade Jan 12 12:49:24 means I don't have to get up so early too ;) Jan 12 12:50:08 been to Istanbul? I think that city might (hopefully) be diffrent to all the Braque stuff I've seen already Jan 12 12:50:21 no, not yet Jan 12 12:50:39 It's supposed to be nice. Jan 12 12:50:57 just don't hang out with any birds :) Jan 12 12:51:08 birds? Jan 12 12:51:17 mithro: bird flu broke out in turkey ;) Jan 12 12:51:26 I think I'll head to Belgrade tommorrow then Jan 12 12:51:32 mithro: i think 16 people already died Jan 12 12:51:47 yesterday it was something like 5, but I wasn't watching the news today. Jan 12 12:52:23 okay then at least some people died ;) Jan 12 12:52:35 I saw that on the news a few days ago Jan 12 12:52:38 mithro: Youīve been here (Vienna) ? Damn. Jan 12 12:53:14 zecke: I meant into what dir is source code decompressed before compilation. I found a different way to get to a solution of the real problem I had. Jan 12 12:53:33 mithro: If you decide not going to Turkey, you might still consider Sarajevo, Mostar and Dubrovnik. Jan 12 12:53:40 Laibsch: what ever dir you configure ;) Jan 12 12:54:10 hwtechnik: yeah about 6 months ago Jan 12 12:54:34 mithro: adriatic is quite warm now, you might get +10C there if the weather is any good. Jan 12 12:54:41 mithro: Uh.Oh. Thatīs a while ago :) Jan 12 12:55:31 welp, I'm getting kicked out Jan 12 12:55:36 this place closes in 5 minutes Jan 12 12:56:02 mithro: lol. have a nice trip. Jan 12 12:56:12 have a good time, wherever the path takes you :) Jan 12 12:57:22 see ya people! Jan 12 12:57:27 ok, gotta bathe the kid... bbl. Jan 12 12:58:44 I have read http://oe.handhelds.org/cgi-bin/moin.cgi/bb_20file?action=highlight&value=cvs but I am not a CVS guru. Would somebody be kind enough to tell me what the SRC_URI line should look like for checking out something like this: cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/gakusei co -P nunome Jan 12 12:59:18 SRC_URI = "cvs.sourceforge.net;module=nunome;tag=$TAG;method=$METHOD" Jan 12 12:59:27 But what are tag and method? Jan 12 12:59:52 How do I specify the project, in this case gakusei? Jan 12 12:59:59 If you don't want a tag, you can omit it. Jan 12 13:00:11 OK Jan 12 13:00:13 For the example you mentioned, you need method=pserver Jan 12 13:01:43 pb_: Thanks a lot. But SRC_URI = "cvs.sourceforge.net;module=nunome;method=pserver" does not quite look like it has got all the required information. Jan 12 13:01:50 SRC_URI="cvs://anonymous@cvs.sourceforge.net/cvsroot/gakusei;module=nunome;method=pserver" Jan 12 13:02:12 right, what ph5 said Jan 12 13:04:35 cool, thanks guys. Jan 12 13:05:06 hi to all :) Jan 12 13:05:13 hi gremlin[it] Jan 12 13:05:19 hi gremlin[it] Jan 12 13:05:36 monotone.vanilla.de is down ? Jan 12 13:05:44 hi pH5, pb_ :) Jan 12 13:05:45 gremlin[it]: as usual Jan 12 13:07:24 RP: Wouldnīt it be good to have 2 seperate collie.conf files to include into the existing one, depending on the kernel choosen? Jan 12 13:07:58 RP: Or am I talking to the wrong one here? Jan 12 13:08:49 RP: wrong person, I mean. Jan 12 13:13:40 grr Jan 12 13:13:45 BASH syntax isn't valid in OE? Jan 12 13:14:23 ${PN/-native//} would be nice Jan 12 13:18:49 Luke-Jr: Maybe you want to have a look at http://pfaffe3.dyndns.org/openembedded/org.openembedded.dev/packages/nunome/nunome_1.0.bb and see if we can get it to compile? Jan 12 13:18:49 http://www.google.de/search?hs=Q3m&hl=en&client=firefox&rls=org.mozilla%3Aen-US%3Aunofficial&q=+site%3Apurl.rikers.org+nunome+laibsch+mickeyl&btnG=Search lists some of the previous discussions about problems I had. Jan 12 13:18:58 zecke: I'm using the -r7 patch Jan 12 13:19:07 hwtechnik: well, when I last checked (a few month back) they were doing some magic in $DISTRO.conf which selects certain ipk's depending on kernel type (2.4 / 2.6) Jan 12 13:19:20 worked perfecty well =) Jan 12 13:20:31 CoreDump: Thanks again for your info :) Iīm trying to get a 2.6 to get to build and work right now. Jan 12 13:21:13 CoreDump: After that I will try to get Pavels-git tree to compile and from there on I shall see... Jan 12 13:21:43 CoreDump: Kinda step by step approach... A little slow, I admit. Jan 12 13:21:57 sounds interesting. Didn't hear about anyone trying 2.6 on Collie for some time. Jan 12 13:22:53 The required adjustments to the root image weren't that complicated. IIRC they had to use different modutils or some such. Jan 12 13:22:54 CoreDump: Well I tried to start 3 weeks ago. So it wasnīt really a kickstart :) Jan 12 13:23:10 hehe =) Jan 12 13:23:31 CoreDump: Oh. So there are gotchas. Well I will resolve them step by step anyway. Jan 12 13:24:04 I'm suprised that the original code is gone regarding 2.4 / 2.6 adjustments. Jan 12 13:24:08 cp: cannot stat `/mnt/old/b-old/tmp/OE/org.openembedded.dev/packages/uclibc/uclibc-0.9.28/uClibc.config': No such file or directory Jan 12 13:24:27 should that be packages/uclibc/uclibc-0.9.28/i386/uClibc.config ? Jan 12 13:28:07 zekoZeko: no Jan 12 13:28:24 zekoZeko: 'i386/uClibc.config' would be used if TARGET_ARCH is i386 Jan 12 13:28:33 zekoZeko: this is the OVERRIDES system of bitbake Jan 12 13:28:41 CoreDump: well it isnīt exactly gone. But I made an error following the Wiki. Had the crosscompiler installed. Jan 12 13:30:01 zecke: my TARGET_ARCH is i486, so I guess that is why it's failing (there's no i486 dir in there). What could I do? Jan 12 13:30:47 CoreDump: Right now Iīm slowly trying to understand the BB-build process step by step. From that I will try my own BB file. Jan 12 13:33:06 right, but It might be a good idea to just steal current .bb files and adjust them to your liking =) Replacing kernel 2.4 w/ 2.6 and correcting the module dependencies should be quite doable Jan 12 13:33:29 bbl Jan 12 13:34:04 zekoZeko: create one ;) Jan 12 13:34:10 zekoZeko: and copy the file Jan 12 13:37:07 should I change anything in uClibc.config? I think it would make sense, looking at it, but I'm not familiar with it... Jan 12 13:38:10 zekoZeko: you can start with the i486 one Jan 12 13:38:10 apparently not.. the i686 one looks the same :) Jan 12 13:38:24 zekoZeko: and then you can do make menuconfig in the uclibc dir ;) Jan 12 13:39:19 I'll do that later, thanks. Now I just wanna make something bootable on the Soekris board as soon as possible, to start playing with it :) Jan 12 14:15:21 RP: ping Jan 12 14:35:34 Ciao all Jan 12 14:35:51 hi Pigi Jan 12 14:36:14 ciao Pigi Jan 12 14:36:33 ciao gremlin[it] Jan 12 14:36:43 Hi pH5 ! Jan 12 14:37:30 NOTE: package pkgconfig-native-0.15.0-r1: task do_build: completed Jan 12 14:37:30 NOTE: package pkgconfig-native-0.15.0: completed Jan 12 14:37:33 hmm Jan 12 14:39:33 zecke: can you repeat the trick with 0.20? Jan 12 14:39:45 hi pigi Jan 12 14:39:52 hi pb_ ! Jan 12 14:39:56 everithing ok ? Jan 12 14:40:04 pb_: I have deleted my 0.20 copy, but let me try Jan 12 14:40:10 Pigi: yep, pretty good Jan 12 14:40:15 zecke: thanks Jan 12 14:40:25 cool. Jan 12 14:40:39 did some one had the time to test the new ipkg revision ? Jan 12 14:41:23 zecke: pong Jan 12 14:41:35 RP: will you be around another hour? Jan 12 14:41:45 zecke: Yes, I'm here for a while now Jan 12 14:42:10 zecke: Got called away earlier and ended up 20 miles away unexpectedly :) Jan 12 14:42:14 Pigi: not me, sorry Jan 12 14:42:26 np, It was just to know. Jan 12 14:42:34 RP: hehe Jan 12 14:42:48 RP: I need to a) compile qemu, b) compile pkg-config Jan 12 14:45:31 jbowler-away: Did you get any further with this image problem you're seeing? I don't think I fully understand why the problem is occuring yet... Jan 12 14:47:22 jbowler-away: Ah, I understand. Its not bitbake's fault :) Jan 12 14:48:49 jbowler-away: When you build ucslugc-image, it has no RDEPENDS. Only slugos-image has RDEPENDS. If you add RDEPENDS_ucslugc-image = RDEPENDS to the image .bb, I think it will work Jan 12 14:55:08 jbowler-away: The other alternative would be to set PACKAGES ="${SLUGOS_IMAGENAME}-image" but this might cause issues trying to package an image :) Jan 12 14:55:36 I'm beginning to think we need a different way to inhibit package generation... Jan 12 14:57:02 At the moment, PACKAGES="" means you can't RPROVIDE anything... Jan 12 14:58:33 jbowler-away: actually, RPROVIDES = "${SLUGOS_IMAGENAME}-image" might help. Jan 12 14:59:00 I'd be interested to see the debug logs from the failed image generation in a dry run mode (bitbake -D -D -n) Jan 12 15:00:34 I'm toot tired . Nite all Jan 12 15:04:47 evening Jan 12 15:05:41 RP: having build slugos-image in both of the problem builds I can't repro the failure, but I'll try for the symptom with bitbake -v openslug-image (I know when this is wrong because it doesn't output stuff about the RDEPENDS) Jan 12 15:08:44 jbowler: If doing a dry run, it won't matter if you've already built things Jan 12 15:09:25 jbowler: See my comments above - this is all about the file being called one thing but providing something else. As far as bitbake is concerned, you only renamed the build time name, not the run time one Jan 12 15:09:54 If it did package, it would always be called slugos-image Jan 12 15:10:09 Therefore the RDEPENDS only apply to slugos-image Jan 12 15:11:05 hi reenoo_ Jan 12 15:19:14 RP: That log output is 1873 lines and it seems to be too big for pastebin.ca Jan 12 15:19:52 Nope, pasted it into the description... trying again ;-) Jan 12 15:20:37 RP: http://pastebin.ca/36634 Jan 12 15:21:18 And I'll try with the RPROVIDES, since that seems more natural to me. Jan 12 15:23:08 jbowler: The last line is the key one: "DEBUG: Additional runtime dependencies for openslug-image are:" Jan 12 15:23:31 There is no package called openslug-image so it can't find any dependencies Jan 12 15:24:08 hey RP Jan 12 15:25:02 RP: okay what to apply now to bitbake Jan 12 15:25:26 zecke: Is the runtime patch applied? Jan 12 15:25:46 RP: at least not -r7 Jan 12 15:25:54 RP: I guess I want openslug-image to be virtual Jan 12 15:26:13 The RPROVIDES makes no difference. Jan 12 15:26:13 zecke: ok :) Jan 12 15:27:12 jbowler: That might or might not be called a bug :-/ Jan 12 15:27:30 What, me wanting it or it not working? Jan 12 15:27:42 jbowler: It not working... Jan 12 15:28:15 I wonder what might happen if I put in a PREFERRED_PROVIDER_${SLUGOS_IMAGENAME} = "slugos-image" Jan 12 15:28:29 Rather: PREFERRED_PROVIDER_${SLUGOS_IMAGENAME}-image = "slugos-image" Jan 12 15:29:00 jbowler: that makes no sense. The problem is in the runtime namespace, not the buildtime one Jan 12 15:29:07 k Jan 12 15:30:13 RDEPENDS_${SLUGOS_IMAGENAME}-image = "${RDEPENDS}" would make more sense perhaps, if bitbake will parse that :) Jan 12 15:30:49 but even that is going to fail Jan 12 15:30:58 $SLUGOS_IMAGENAME isn't in PACKAGES Jan 12 15:31:47 Yes, RDEPENDS_openslug-image fails too... So I will add PACKAGES += "openslug-image", that's fine, we don't mind having a package for it. Jan 12 15:32:45 jbowler: The other solution is create a openslug-image.bb and have "include slugos-image.bb" in it Jan 12 15:32:47 Anyway, it will be empty. Jan 12 15:34:26 RP: that's what we had before, it would result in a proliferation of packages/meta/nslu2-poo-image.bb files - we have a lot of DISTROs (and we don't care about the .ipk from slugos-image because the real result is in deploy/images) Jan 12 15:34:55 We can't actually install a slugos-image ipk Jan 12 15:35:49 jbowler: ok, I think this is probably the best solution. I'm beginning to think we need an INHIBIT_PACKAGING flag instead of setting PACKAGES="" anyway (think of the problems in native.bbclass) Jan 12 15:36:06 *cry* arm-user only works on linux Jan 12 15:36:13 You might have to set PACKAGES after you inherit image_ipk Jan 12 15:36:23 * zecke wonders how fast a qemu will be that gets emulated by a qemu... Jan 12 15:36:39 RP: any changes to your patch? otherwise I will apply and commit it now Jan 12 15:37:06 zecke: I think the patch is standing up to testing ok, so it can be applied Jan 12 15:37:26 We have odd little issues but nothing that makes me think its doing anything wrong :) Jan 12 15:38:11 RP: r5 is the latest? or should I get it off from sam? Jan 12 15:38:22 zecke: you want -r7 from tim Jan 12 15:38:32 http://www.rpsys.net/openzaurus/temp/bitbake-rproviders-r7.patch Jan 12 15:39:02 RP: that one looks almost perfect ;) Jan 12 15:39:06 koen|away: please re-enable monotone write access for mgoebl@openembedded.org Jan 12 15:39:09 only a doctest for explode_deps is missing Jan 12 15:39:35 zecke: I have no experience with doctests Jan 12 15:39:51 RP: could you do a rediff please? the lib/bb/shell.py bit was applied this afternoon :} Jan 12 15:40:09 (only if it is an easy svn diff call) Jan 12 15:40:46 Oops, yes, rootfs_ipk.bbclass... still no joy. It really only works if slugos-image itself is the build target. Jan 12 15:41:20 zecke: http://www.rpsys.net/openzaurus/temp/bitbake-rproviders-r8.patch Jan 12 15:41:22 + if not self.addRunDeps(p, item , buildAllDeps): Jan 12 15:41:22 + return False Jan 12 15:41:23 + return True Jan 12 15:41:42 why not a simple return self.addRunDeps (no request for change though) Jan 12 15:42:35 zecke: I think I borrowed that code from the call inside a while loop :) Jan 12 15:42:51 zecke: it can be simplified in that case (I can rediff easily again if you like) Jan 12 15:43:42 no no this is something for later Jan 12 15:43:52 http://www.rpsys.net/openzaurus/temp/bitbake-rproviders-r9.patch Jan 12 15:43:54 :) Jan 12 15:43:57 RP: argh r8 has a conflict now Jan 12 15:44:17 oops :-/ Jan 12 15:44:18 I've used other parenthesis :} Jan 12 15:45:19 * CosmicPenguin hugs xorg7.0 Jan 12 15:45:29 err yeah Jan 12 15:45:32 you know what I mean Jan 12 15:45:49 ok, http://www.rpsys.net/openzaurus/temp/bitbake-rproviders-r10.patch Jan 12 15:47:23 http://pastebin.com/503174 Jan 12 15:47:42 zecke: it has a bug in the change I just made - I left the colon in :-. Jan 12 15:47:45 for my relocation of bb.*codeurl to bb.fetch (only base.bbclass in OE needs to change) Jan 12 15:48:36 http://www.rpsys.net/openzaurus/temp/bitbake-rproviders-r11.patch :-/ Jan 12 15:49:39 http://www.chucknorrisfacts.com/ <- spotted by woglinde Jan 12 15:49:42 svn revert * Jan 12 15:50:42 Chuck Norris responds: http://www.chucknorris.com/html/events.aspx Jan 12 15:50:42 sorry :-/. Jan 12 15:51:52 now it is my time to screw things up Jan 12 15:52:07 zecke: go for it Jan 12 15:54:42 zecke: Your patch looks good to me Jan 12 15:55:03 once applied, I can tweak the git fetcher to use some of this Jan 12 15:57:44 RP: there is one issue. First I want to put the SRCDATE and 'now' handling into the parameters one gets from decodeurl Jan 12 15:58:10 that looks right to me. One down side is, if the package is set to 'now' and the package is build Jan 12 15:58:30 sources will always be fetched, regardless of if they will be copied to the work dir... Jan 12 15:59:01 RP: but the goal is to move the 'MIRROR' handling to Fetch, running of commands as well Jan 12 15:59:10 zecke: Also, what happens if a user includes a srcdate= in their url? Jan 12 15:59:33 RP: with the current patch, it gets overwritten Jan 12 15:59:57 zecke: I know. I'm wondering if that should be the correct behaviour Jan 12 16:00:42 If a package is built, I'm let to wonder if the fetcher should be being called? Jan 12 16:01:16 s/let/left/ Jan 12 16:01:16 RP meant: If a package is built, I'm left to wonder if the fetcher should be being called? Jan 12 16:01:23 RP: well base.bbclass turns off stamping Jan 12 16:01:45 zecke: but should it? Jan 12 16:02:02 probably not Jan 12 16:02:39 zecke: I'd also say not. Jan 12 16:02:50 I'll ask on oe@ about that perhaps Jan 12 16:03:24 I like the idea of being able to set "now" as the date and then being able to control it via a stamp Jan 12 16:03:43 if i want a source refresh, I rm the stamp... Jan 12 16:03:53 RP: -cfetch -f ;) Jan 12 16:04:04 zecke: well, yes :) Jan 12 16:04:11 RP: http://pastebin.ca/36636 works Jan 12 16:05:43 (With PROVIDES removed from slugos-image.bb) Jan 12 16:05:49 jbowler: Actually, can you get away with setting PN in the slugos-image.bb? Jan 12 16:05:54 is there a reason that kernel.bbclass doesn't stage the files in include/sound? Jan 12 16:06:11 just an oversight, I think Jan 12 16:06:20 RP: well, bitbake built it fine, PN just gets defaulted from the file name doesn't it? Jan 12 16:06:55 jbowler: Yes, but you're free to override it Jan 12 16:06:57 pb_: do you know why stamping of fetch is turned off? It has to do with the .md5 thing but how remains unclear to me Jan 12 16:07:37 RP: sorry, I misread your question. No, both slugos-image and openslug-image are used in different contexts. Jan 12 16:08:03 jbowler: In the same build? Jan 12 16:08:04 We could fix that, since openslug-image comes from outside OE (from the 'master makefile'), but that would cause user confusion. Jan 12 16:08:16 zecke: no reason I can think of offhand. might just have been disabled because it isn't much use: you can always tell whether the fetch has happened by the presence of the .md5 file, so an extra stamp would be redundant. Jan 12 16:08:36 RP: no, nothing in OE should be depending on 'openslug-image' - only on slugos-image, I'll check but I'm pretty sure I cleaned that up. Jan 12 16:08:44 zecke: does the lack of stamping for fetch displease you? Jan 12 16:08:58 jbowler: I think we're talking cross purposes here Jan 12 16:09:18 jbowler: I'm asking if you could set PN = "${SLUGOS_IMAGENAME}-image" in slugos-image.bb? Jan 12 16:10:09 jbowler: bitbake looks for RDEPENDS beloging to ${PN} so it might be enough to make things work Jan 12 16:10:13 I'm saying 'openslug-image' can only come because someone typed it to a bitbake command, but 'slugos-image' is a DEPENDS of slugos-packages.bb, so while only one is ever required I don't know which. Jan 12 16:10:57 I.e bitbake slugos-packages will builds slugos-image, but bitbake openslug-image is also (currently) valid. Jan 12 16:11:28 How does SLUGOS_IMAGENAME get set? Jan 12 16:11:44 In conf/distro/... Jan 12 16:12:45 By default to DISTRO in conf/distro/slugos.conf (which probably should be in conf/distro/include...) Jan 12 16:12:53 So you can build openslug-image whilst SLUGOS_IMAGENAME is set to slugos? Jan 12 16:13:34 Well, you can try, but it won't find a provider... Jan 12 16:14:04 pb_: thanks. the btsco kernel module needs those. Jan 12 16:14:11 I can't see how setting PN = as above can cause a problem then? Jan 12 16:15:07 pb_: I think when cvs is considered, having the stamp might be useful Jan 12 16:15:14 RP: why? Jan 12 16:15:41 RP: If I do that then try "bitbake slugos-packages" I get: Jan 12 16:15:43 ERROR: Nothing provides dependency slugos-image Jan 12 16:15:43 ERROR: dependency slugos-image (for slugos-packages) not satisfied Jan 12 16:16:07 I want to use "slugos-image" in OE, but I'm provided backward compatibility for the makefile and all the users out there. Jan 12 16:17:05 pb_: set srcdate="now", build. when you rebuild it won't get repulled unless you remove the stamp (-c fetch -f). It would provide a convinient way of working with cvs="now" Jan 12 16:17:47 jbowler: Your still have the PROVIDES = "${SLUGOS_IMAGENAME}-image" line? Jan 12 16:18:33 Not in the file I modified - but the PN was ${SLUGOS_IMAGENAME}-image so it would provide it anyway. Jan 12 16:18:57 jbowler: Set it to PROVIDES = "slugos-image openslug-image" Jan 12 16:19:10 That should be enough to fool it :) Jan 12 16:19:16 RP: oh, does the fetcher understand "now"? Jan 12 16:19:23 pb_: as RP said, with SRCDATE="now" you would fetch source over and over again just to create tar.gz's that will never be used Jan 12 16:19:28 pb_: We're about to patch it to understand that Jan 12 16:19:42 Yes, but then when bitbake builds slugos-image it will still have the RDEPENDS problem - just it will now be for slugos-image... Jan 12 16:20:21 jbowler: Maybe not as I think if the PN is set as above, it should get the RDEPENDS right Jan 12 16:20:24 slugos-imagename.bb on the other hand is always going to work, since it looks just like slugos-packages.bb (i.e. it DEPENDS on slugos-image), so if it doesn't work neither will slugos-packages. Jan 12 16:20:28 jbowler: I might be wrong mind... Jan 12 16:20:31 zecke: mm, I am not entirely convinced that this "now" thing is good news. What is the use case for that? Jan 12 16:21:13 or, to put it another way, what does "now" buy you compared to ${DATETIME}? Jan 12 16:21:38 pb_: ever wanted to check out the HEAD/trunk version of your software package? Jan 12 16:21:46 pb_: Picture working on some package from a scm. You run bitbake, it checks it out and breaks. You update the scm, you rerun bitbake and you'd like it to use the new changes Jan 12 16:21:48 setting CVSADATE="now" did work exactly once Jan 12 16:22:16 it creates a cvs_foo_now.tar.gz ;) Jan 12 16:23:48 zecke: and you can't wait until the next minute when ${DATETIME} would change? I guess the pace of opie development must be faster than I thought. Jan 12 16:24:08 * reenoo_ rotfl Jan 12 16:24:37 pb_: well you will then have thousands of thousands src tarballs then? Jan 12 16:25:06 yeah. is that a big hardship? Jan 12 16:25:24 I kind of feel I'd prefer to have thousands of tarballs rather than one tarball whose name stays the same but contents change. Jan 12 16:25:44 zecke: fwiw, SRCDATE=now isn't going to produce pleasant results. PV = 0.0+cvsnow is rather useless Jan 12 16:26:52 pb_: well if you put DATETIME in the $PV as well you will 2*thousands of thousands of dirs Jan 12 16:27:30 I think having stamps on fetching would be beneficial - picture the case where your build with some SRCDATE="${DATETIME}" in it fails half way through 50 packages. Restarting the build would be fun... Jan 12 16:28:01 RP: stamps wouldn't make any difference in that case. Jan 12 16:28:37 assuming PV is defined in terms of ${SRCDATE}, which it almost always is, the stamps would be invalidated anyway when SRCDATE changes. Jan 12 16:29:04 and, if SRCDATE doesn't change, the existing .md5 check will stop the source from being re-fetched anyway. Jan 12 16:29:42 yes, the link to PV would cause a problem Jan 12 16:32:05 'night all Jan 12 16:32:48 Hi All, can someone please have a look at http://bugs.openembedded.org/show_bug.cgi?id=399, its been around for ages now, and I tested last week and the prism2 still exibits the problem Jan 12 16:33:44 If we cant get this fixed before LCA I will have to go back to OZ 3.5.3 Jan 12 16:35:19 Geo_KM: Sadly, I think only a certain fraction of our developers can reproduce that bug and most of them are away at the moment :-/ Jan 12 16:37:15 RP: Ah! I see. Well if anyone needs me to test it out I am happy to do it. Its just I'm not really a coder/developer so I am not in a pos to actually fix the bug.... Wish I was 8-) Jan 12 16:38:49 Geo_KM: The problem is we don't actually have that many coders/developers with that set of hardware and its the kind of problem where on device debugging is often more productive. Jan 12 16:39:15 Is monotone.vanilla.de dead at the moment. It seems to be refusing connections. Jan 12 16:40:08 RP: Yeah I see what you mean. 8-) Jan 12 16:40:24 Geo_KM: yes, try one of the others Jan 12 16:40:35 OK thanks Jan 12 16:48:26 pb_: I had occassions (in development cycles) where this one minute would have been too much Jan 12 16:48:42 and in this case the PV did not contain a time stamp Jan 12 16:49:14 pb_: I recognize the problems that special treatment of now can give, but I find it useful and will construct a use case once I can think again Jan 12 16:49:23 okay. but, in those cases, presumably you wouldn't have wanted a stamp for do_fetch either. Jan 12 16:50:30 pb_: my personal use is. If I do bitbake -b foo.bb -cfetch I want it to contain the latest revision Jan 12 16:54:08 zecke: okay Jan 12 16:56:41 anyway sleep well guys Jan 12 16:57:01 hmm I should have dropped koen a mail Jan 12 16:57:20 zecke: good night Jan 12 17:21:05 Is the repo still in an unsafe state to do an update? Jan 12 23:03:41 Ahhh Silence Is Golden...... (Dead Channel RIP) Jan 12 23:08:24 mnm: it comes alive only when you sleep Jan 12 23:08:53 where are you people..? Jan 12 23:15:39 asleep Jan 13 00:29:40 good morning all ! Jan 13 01:06:09 OPIE_CVS_PV needs to have a default, because at present I get lots of messages of this general form: Jan 13 01:06:23 NOTE: Handling BitBake files: / (2032/2958) [68 %]NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Jan 13 01:06:23 ${@base_set_filespath([ "/home/nslu2/oe/org.openembedded.nslu2-linux/packages/opie-securityplugins/opie-securityplugins-${OPIE_CVS_PV}-r0", "/home/nslu2/oe/org.openembedded.nslu2-linux/packages/opie-securityplugins/opie-securityplugins-${OPIE_CVS_PV}", "/home/nslu2/oe/org.openembedded.nslu2-linux/packages/opie-securityplugins/opie-securityplugins", "/home/nslu2/oe/org.openembedded.nslu2-linux/packages/opie-securityplugins/files", "/home/nslu2/oe/org.open Jan 13 02:04:29 morning Jan 13 02:04:38 RP: 100% success on gpe-image Jan 13 02:08:58 morning all Jan 13 02:09:03 XorA: Excellent :) Jan 13 02:11:53 * XorA will try to apply brain before build in future Jan 13 02:13:14 XorA: the main thing is it works :) Jan 13 02:21:58 jbowler-zzz: It means those .bb files aren't inheriting opie :-/ Jan 13 02:22:06 I thought I'd caught them all... Jan 13 02:22:45 RP: you talking about the EOL errors in some opie-*files? Jan 13 02:24:12 XorA: yes Jan 13 02:24:43 RP: ok, I can keep quiet then Jan 13 02:26:36 Is it just opiesecurityplugins? Jan 13 02:29:13 RP: slugos-image, opie-help, opie-keytabs, opie-mediaplayer2-skin-default-landscape, opie-mediaplayer2-skin-pod, opie-mediaplayer2-skin-techno, opie-pics-, opie-sounds Ithink Jan 13 02:36:47 XorA: Ok, thanks. I'll attempt to fix... Jan 13 02:37:24 hrw|gone: Does opie-securityplugins itself actually do anything? Jan 13 02:40:00 morning all Jan 13 02:43:01 morning Jan 13 02:43:19 ~seen RP Jan 13 02:43:23 morning dirk, greentux Jan 13 02:43:25 rp was last seen on IRC in channel #oe, 6m 1s ago, saying: 'hrw|gone: Does opie-securityplugins itself actually do anything?'. Jan 13 02:43:36 RP: morning. do you remember the serial_cs bug? Jan 13 02:43:53 RP: it writes everytime ttyS0 even when it use ttyS3... Jan 13 02:43:58 greentux: yes Jan 13 02:43:59 morning Richard, greentux Jan 13 02:44:22 RP: who can fix that? maintainer of serial in the kernel? Jan 13 02:45:03 greentux: Someone needs to write a proper patch. Unforunately its not a simple issue and I'm not sure what a proper patch would consist of... Jan 13 02:45:41 greentux: The problem is the pxa shares the same device number space as the 8520 uart driver Jan 13 02:46:00 RP: oh, doesnt expect that is not simple :) Jan 13 02:46:42 One solution would be to get pxa its own device numbers. But then every pxa machine would have to update its references to /dev/ttyS0 to /dev/ttySP0 or something... Jan 13 02:47:28 The patch in the openzaurus tree is a hack and the serial maintainer (Russell King) will explode if people ask him about it again Jan 13 02:47:37 * XorA chuckles Jan 13 02:48:28 RP: ok, understand. for the usb client mode...: i think there is a opie app for switching the mode. Jan 13 02:48:54 RP: we will try to fix that Jan 13 02:49:18 greentux: ok, I'm sure one would be appreciated :) Jan 13 02:49:22 RP: Random idea re that.. have an OE-local patch to renumber the major for 8250? Jan 13 02:49:52 NAbyss: and break all the references to /dev/ttyS0? Jan 13 02:49:57 no thanks... Jan 13 02:50:10 NAbyss: cant do that, that would kill BT cards Jan 13 02:51:15 and from what I remeber about reading that flamewar, the reasons for PXA and 8250 sitting on same namespace are gone with udev based systems anyway, so I guess correct fix is use the PXAmajor/minors Jan 13 02:51:47 Ugh.. sounds rather tricky. Jan 13 02:51:57 * RP isn't so sure its as simple a that Jan 13 02:56:05 * XorA ponders the bizareness that tightvnc is worse at transmitting monochrome images than colour Jan 13 02:58:58 after loading serial_cs: detected caps 00000700 should be 00000010 Jan 13 02:59:16 some experience? **** ENDING LOGGING AT Fri Jan 13 02:59:57 2006