**** BEGIN LOGGING AT Mon Aug 15 07:19:38 2005 Aug 15 07:35:03 Hmm.... bitbake/OE availability is starting to become a barrier for casual users, I think Aug 15 07:35:58 it is? Aug 15 07:36:46 downloading a tar.gz and a db.bz2 shouldn't be too hard Aug 15 07:37:19 (and are 'casual users' a target for OE?) Aug 15 07:38:54 if I had the brains a good project would be to dump the files nightly into a cvs somewhere Aug 15 07:39:19 wget nslu2-linux.org/Makefile ; make setup Aug 15 07:39:29 that will download, install and setup OE Aug 15 07:39:36 koen: I was just about to say that :-) Aug 15 07:39:43 :) Aug 15 07:40:13 bloomin eck, thats cool Aug 15 07:40:17 * rwhitby-asleep thinks koen has been converted :-) Aug 15 07:41:02 XorA: and I did that on an OE built distro (openslug) Aug 15 07:41:09 * XorA just needs to walk down a street now and find slug in a skip Aug 15 07:44:50 * koen needs to find 20 slugs and switch for a compilefarm Aug 15 07:45:50 koen: connect them all with USB->USB and have some crazy routing Aug 15 07:45:59 koen: trust me, its a barrier Aug 15 07:46:19 And its going to get bitbake/OE pulled out from underneath me, and that pisses me off to no end Aug 15 07:49:12 XorA: usb<>usb doesn't work Aug 15 07:49:21 XorA: usb is too badly designed for that Aug 15 07:49:33 you could do it with firewire though Aug 15 07:50:52 CosmicPenguin: what does a 'casual user' or your management want? Aug 15 07:51:05 s/want/need/ Aug 15 07:57:29 koen: I have no idea, to tell you the honest truth - I can't get anything more then a "it sucks" out of anybody Aug 15 07:57:45 I hate when that happens Aug 15 07:57:50 koen: but I do know that I had three e-mails indicating that various people couldn't download the software this weekend Aug 15 07:58:02 I've never had a problem.. :) Aug 15 07:58:30 "it sucks" "why does it suck" "it just sucks" Aug 15 07:58:43 no win32 gui Aug 15 07:59:02 no telepathic AI Aug 15 07:59:09 We keep OE and bitbake in RCS here, we have a setup script, you just type it and go Aug 15 07:59:13 yet, people still hate it Aug 15 07:59:17 hate it with a passion, even Aug 15 07:59:40 click-a-zoids? Aug 15 07:59:47 typing is bad, etc.... Aug 15 07:59:47 no, penguins Aug 15 08:01:37 bizarre Aug 15 08:02:40 reminds me of the effort it takes to point out maemo uses X so using qt/e is a bad idea to certain companies Aug 15 08:02:54 *cough* theKompany *cough* Aug 15 08:03:04 I think they miss the makefile mentality Aug 15 08:03:24 alias make=bitbake :-) Aug 15 08:03:28 but i have been hesitant to introduce it, because I want people to learn how to build packages Aug 15 08:03:37 bah Aug 15 08:03:44 all those people suck Aug 15 08:04:40 bitbake/oe is easier than gentoo, easier than debian, far easier than rpm Aug 15 08:05:24 maybe I should set up a cronjob for daily snaps Aug 15 08:06:17 Well, luckily, its my decision, so we're using bitbake Aug 15 08:06:24 everybody else can cram it with walnuts.. :) Aug 15 08:06:37 * CosmicPenguin goes off to write e-mails to that effect Aug 15 08:08:47 bitbak and monotone are certainly not easier. you have to dl MUCH more informantion that specifically necessary to do what you want Aug 15 08:09:02 and the build process is regularly broken Aug 15 08:10:09 monotone using a monolithic 100mb+ db to store all information instead of integrating that into the meta data tree is a joke Aug 15 08:12:42 i think large projects like OE should stick with proven methods even if they have short comings, instead of jumping onto new programs that aren't even backwards compatible with theirselves *cough bitbake* Aug 15 08:13:36 slisk: your kidding right? Aug 15 08:13:49 plz, retort... Aug 15 08:14:23 slisk: the bitbake developers are right here, bitbake and oe have gone hand in hand since the beginning Aug 15 08:15:04 Heh - in fact, there was a time when bitbake and OE were one Aug 15 08:15:06 bitbake was created because the 'proven' methods don't scale Aug 15 08:15:38 try maintaining 2500 packages in buildroot Aug 15 08:15:52 i'm criticizing bitbake's lack of backwards compatibility...with itself! Aug 15 08:15:52 for >30 different platforms Aug 15 08:16:25 you mean OE isn't compatible with an old bitbake? Aug 15 08:16:35 slisk: the .bb format has changed little, what backwards compatibility do you mean? Aug 15 08:17:30 i was told while trying to build the OZ 3.5.3 archived metadata (from just April of this year) that i needed the same version of bitbake that was used then, cause bitbake isn't backwards compatible Aug 15 08:17:43 just 4 months and it's lost backwards compatibility!?!? Aug 15 08:18:44 slisk: There were improvements needed to bitbake to let oe do new things. The modern metadata therefore depends on a modern bitbake... Aug 15 08:19:04 a 95% reduction in RAM usage for example Aug 15 08:19:08 but why can't the modern bitbake work with older metadata? Aug 15 08:19:13 RP: he's talking about the other way around: a new bitbake won't work with old metadata. Aug 15 08:19:40 which is definitely bad news, but in this instance seems to be unavoidable. Aug 15 08:19:50 i don't even buy that Aug 15 08:20:07 i have greater faith in the source than that Aug 15 08:20:10 the cache changes mean that metadata needed to be updated when using a new bitbake? Aug 15 08:20:24 slisk: I don't understand that comment Aug 15 08:21:13 why could the new bitbake just convert older metadata to the new format, or just respect the old format regardless? (the source == the force == joke) Aug 15 08:21:23 mithro: zecke's and proti's work on making it use less than 1GB of ram require some changes on both parts Aug 15 08:21:42 the format didn't change Aug 15 08:21:50 the classes did Aug 15 08:22:06 define classes Aug 15 08:22:32 you talking about the source code or something in the metadata? Aug 15 08:22:56 * XorA doesnt see the problem in requiring bitbake VX.XX for OE tag VX.XX this is pretty standard in software world Aug 15 08:23:06 OE = classes + metadata Aug 15 08:23:31 look in the classes/ subdir Aug 15 08:24:10 slisk: Certain functionality was moved from the classes into bitbake. Retaining backwards compatibility was not something we could do with the number of developers we have. Now if you're volunteering... Aug 15 08:24:13 XorA: disallowing a newer version for use with materials created for use with an older version is NOT standard. ever hear of word2003 refusing to load a word95 file? Aug 15 08:24:51 slisk: ever head of postgres SQL not being able to read older versions of the database? Aug 15 08:25:06 slisk: stop confusing metadata and classes Aug 15 08:25:20 the metadata format hasn't changed Aug 15 08:25:27 slisk: word2003 does not read word2 files with 100% success Aug 15 08:25:40 slisk: An upgrade script is enirely possible, just nobody has taken the time to write it Aug 15 08:25:41 slisk: and word for DOS files it really mangles Aug 15 08:25:56 well, an update script ought really not to be necessary. Aug 15 08:26:05 slisk is right, it does suck that newer bitbakes will not work with old oe trees. Aug 15 08:26:34 pb_: not to the degree that bitbake should be abaondoned though Aug 15 08:26:36 with the exception of the OVERRIDES overrides thing, I don't know of any good reason why this should really be the case. Aug 15 08:26:49 pb_: I don't think you should stop the project moving forwards by sitpulating we have to have backwards compatibility Aug 15 08:27:36 when a major project like OZ 3.5.3 breaks in under 4 months, there are problems that need to be addressed Aug 15 08:27:39 whats armeb? the new arm binary format? Aug 15 08:27:44 especially in the situation we were (keeping 1GB worth of cruft in RAM) Aug 15 08:27:44 doesnt this all boil down to one of the oe classes knowing too much about bitbake internals Aug 15 08:27:53 mithro: bigendian arm Aug 15 08:28:16 slisk: stop talking crap Aug 15 08:28:21 slisk: It isn't broken - use an older bitbake with it and it is fine. Aug 15 08:28:24 slisk: my zaurus with oz353 still works Aug 15 08:28:34 * mickeyl yaws and refuses to scroll up to follow a pointless debate Aug 15 08:28:35 I can't see how bitbake breaks that Aug 15 08:28:50 yawns, even Aug 15 08:28:52 mickeyl: i now have 3 OE devices, a wrt54g, nslu2 and a c860 :P Aug 15 08:28:57 mithro: excellent! :) Aug 15 08:29:08 we could need some help for the wrt Aug 15 08:29:19 atm. it's dangerous though without a serial cable Aug 15 08:29:26 mickeyl: Tried spitz 2.6 yet? :) Aug 15 08:29:27 br1 seems to be missing in action Aug 15 08:29:33 mickeyl: that doesn't sound good Aug 15 08:29:46 mickeyl: maybe he's on holiday Aug 15 08:29:47 RP: not yet, I'm preparing my Spitz for a demo on weekend Aug 15 08:30:03 RP: although I hardly can wait to try Aug 15 08:30:14 guess I'll throw the current thing away and just do it :)) Aug 15 08:30:21 * mickeyl bitbake virtual/kernel Aug 15 08:30:29 [12:37:34] i get this http://pastebin.ca/19951 error trying to build a package using OZ 3.5.3 archived metadata. can any1 help? Aug 15 08:30:29 [12:37:53] slisk: with a recent bitbake? Aug 15 08:30:29 [12:38:06] if so, that won't work Aug 15 08:30:29 [12:38:10] BitBake Build Tool Core version 1.3.2, bitbake version 1.3.2 Aug 15 08:30:29 [12:38:27] you'll also need a bitbake version from the time 3.5.3 was released[12:39:32] bitbake isn't backwards compatible?!!? Aug 15 08:30:32 [12:39:40] no, it isn't Aug 15 08:30:32 mickeyl: just make a backup Aug 15 08:30:41 how can you say is still works? Aug 15 08:31:04 slisk: by opening my zaurus and tapping on the screen Aug 15 08:31:14 mickeyl: Don't destroy anything you need. I am keen to get opie running though :) Aug 15 08:31:22 whats Spitz again? the 4gig model? Aug 15 08:31:29 mithro: yes Aug 15 08:31:41 RP: Well, with the harddisk I can copy the stuff over to another partition Aug 15 08:31:44 are you saying you can follow the build instructions on the openzaurus webpage (today) and expect it to work??? Aug 15 08:31:52 RP: what state are we in? what doesn' work? Aug 15 08:31:57 (kernel wise) Aug 15 08:32:02 slisk: no I'm saying 'openzaurus 3.5.3' isn't broken Aug 15 08:32:19 mickeyl: Buggest thing missing is suspend/resume - I'm working on that as a priority Aug 15 08:32:22 oh, you misinterpretted, i meant the build process is broken Aug 15 08:32:40 mickeyl: Otherwise, USB, IR and Sound Aug 15 08:33:01 s/Buggest/Biggest/ :) Aug 15 08:33:02 slisk: I can build oz353 fine, since I have snapshots of both OE and bitbake Aug 15 08:33:17 slisk: get over it. foo-bar vX-Y needs lib-foo-bar vX.Y. There's no point in complaining here unless you are trying to piss us guys off. If it offends you, sends us a patch. Period Aug 15 08:33:25 but I don't want to build obsolete distros Aug 15 08:33:28 RP: excellent. that's hell of a progress Aug 15 08:33:58 slisk: If we add explicit instructions on how to get bitbake from the OZ release date, does that help? It will work then... Aug 15 08:34:09 mickeyl: I remember someone saying 'oe is dead' because we change from .oe to .bb Aug 15 08:34:10 someone on a 32bit system that could try a compilation-stage patch for xorg for me? Aug 15 08:34:17 keon: i'd appreciate a snapshot of bitbake if you can tell me how to get it (never did follow up the svn explanation) Aug 15 08:34:26 RP: that would be greatly appreciated Aug 15 08:34:57 koen: I can't count the number of valuable hours that I spent with people complaing and not giving _ONE_ fscking thing back. I'm pretty tired of it. Aug 15 08:35:12 mickeyl: actually i'm trying to transport the OZ kernel into debian right now, dump all these problems Aug 15 08:35:15 mickeyl: my /ignore list is growing too Aug 15 08:35:25 effort that could've gone to contributing instead Aug 15 08:35:30 mickeyl: lame complainers-not-doers Aug 15 08:35:34 slisk: sounds good. have fun. Aug 15 08:35:42 * mithro is currently adding rtorrent to OE :P Aug 15 08:35:50 slisk: Dump what problems exactly? Aug 15 08:35:55 one more of the debian-based-is-not-enough, i want the real-thing guys Aug 15 08:36:01 * mickeyl shrugs Aug 15 08:36:01 slisk: On second thoughts, don't tell me... Aug 15 08:36:04 and between i've creaetd a custom keymap and taught ppl how to backup and reflash custom images, that contributes to the community directly Aug 15 08:36:37 rtorrent kicks ass, its only uses about 6MB of ram for 400k/sec upload with 60 peers Aug 15 08:36:45 btw i've only had my zaurus for about a week now Aug 15 08:37:08 mithro: is it better as ctorrent? Aug 15 08:37:15 yay Aug 15 08:37:18 s/yay/ya Aug 15 08:37:19 mithro: I'll looking for a nice client to run on the slug Aug 15 08:37:29 koen: yeah it should work nicely on the slug Aug 15 08:37:42 i'm just fixing up some stuff which means it doesn't compile on uclibc Aug 15 08:38:34 mithro: you saved me doing it, I was planning to look at rtorrent Aug 15 08:38:48 XorA: i kind of know the guy who wrote it Aug 15 08:39:05 mithro: cool, one day I plan to try it instead of azureus Aug 15 08:39:27 XorA: not as featureful as azureus, but definately a hell of a lot lighter Aug 15 08:39:58 mithro: I dont need many features, mainly bandwidth limiting, azureus is sucking my CPU a lot though Aug 15 08:40:31 XorA: rtorrent is pretty light on the CPU as well :P Aug 15 08:40:41 the guy who wrote it is pretty good Aug 15 08:41:16 he hopes to reduce the binary footprint soon as well Aug 15 08:41:37 using some of the gcc 4.0 visibility stuff Aug 15 08:41:49 anyway :P Aug 15 08:42:08 koen: any idea why the slug takes to long to boot? Aug 15 08:42:41 it spends about 20 secs in redboot Aug 15 08:43:11 booting openslug is pretty fast Aug 15 08:43:14 is that the red light on the status bit? Aug 15 08:43:24 re Aug 15 08:43:39 mithro: yes the red light is used for bootstatus Aug 15 08:43:49 the nslu2 wiki has details on that Aug 15 08:47:32 The slug boots a lot faster when using apex :D Aug 15 08:47:39 makeinfo -I/home/tim/oss/nslu2/ucslugc/tmp/work/cpio-2.5-r3/cpio-2.5 cpio.texi Aug 15 08:47:47 make[2]: makeinfo: Command not found Aug 15 08:47:56 do we not have makeinfo as a native package? Aug 15 08:48:40 i don't think so Aug 15 08:48:43 wanna add one? Aug 15 08:49:02 hmm Aug 15 08:49:03 thinking about that Aug 15 08:49:06 isn't that part of tex? Aug 15 08:49:10 i already added tetex-native Aug 15 08:49:24 so perhaps it's just looking into its stage() method and depending on that Aug 15 08:49:55 ah yes Aug 15 08:49:56 already staged Aug 15 08:50:00 just depend on tetex-native Aug 15 08:50:07 and it should just work(tm) Aug 15 08:50:11 CosmicPenguin: you're using OE for the geode stuff? Aug 15 08:50:29 CosmicPenguin: or for the alchemy line? Aug 15 08:50:47 mickeyl: do we have a list of companies using OE? Aug 15 08:50:57 koen: only what's on the Success stories page Aug 15 08:51:04 from memory Aug 15 08:51:15 TI, AMD, Siemens, 3g-systems Aug 15 08:51:35 s/3g/4g/ Aug 15 08:51:37 Vanille Media, if that counts ;)) Aug 15 08:51:45 Wooo, Kernel Traffic :D Aug 15 08:52:29 mickeyl: University of Twente, University of San Diego Aug 15 08:52:37 University of Frankfurt Aug 15 08:52:43 FU-Berlin Aug 15 08:52:44 OpenedHand (or some members of it anyway :) ) Aug 15 08:53:08 "Nokia" :) Aug 15 08:53:15 somehow Aug 15 08:53:46 mickeyl: I'm not so sure about Nokia. I can't get oe generated programs to run on the 770 :-/ Aug 15 08:54:09 or not without a complete oe generated image so all the versions match... Aug 15 08:54:54 socket.error: (110, 'Connection timed out') Aug 15 08:54:56 ~lart CIAbot Aug 15 08:54:57 * ibot puts on some milking gloves. "All right, now, CIAbot, this won't hurt a bit...." Aug 15 08:55:30 mickeyl: I was wondering if we could put a list online of people using OE Aug 15 08:56:01 koen: ya. I've always wondered why our two-weeks-admin made a static site out of success stories :/ Aug 15 08:56:15 feel free to add a corresponding site on the oe wiki Aug 15 08:56:16 i'll add the ones i know then Aug 15 08:56:39 ~lart people talking a lot and then vanishing Aug 15 08:56:40 * ibot whips people talking a lot and then vanishing with a wet and grimy noodle just because Aug 15 08:58:13 mickeyl: when did you add tetex-native? Aug 15 08:58:22 two weeks ago Aug 15 08:58:23 ~ Aug 15 08:58:39 * mithro ponders why its not in ucslugc Aug 15 08:59:24 because the ucslugc package manager hasn't added it? Aug 15 09:00:53 mickeyl: http://oe.handhelds.org/cgi-bin/moin.cgi/SuccesStories Aug 15 09:01:45 k Aug 15 09:01:46 * mickeyl adds Aug 15 09:06:04 mickeyl: when is the next newsletter? Aug 15 09:06:49 mithro: this week Aug 15 09:06:57 when koen gets finished :) Aug 15 09:08:20 have something to add? Aug 15 09:10:02 what is the function of the kernel-image package for machines that have the real kernel image in a separate flash partition? Aug 15 09:10:16 mickeyl: did you seen my comments on angstrom yesterday? Aug 15 09:10:21 koen: I'm afraid not Aug 15 09:10:30 or the ones you sent me via /query? Aug 15 09:10:34 got those Aug 15 09:10:35 that ones Aug 15 09:10:38 yep Aug 15 09:10:47 will take them into account somehow Aug 15 09:10:53 although I'm slowly feeling kind of pushed Aug 15 09:11:03 yeah Aug 15 09:11:11 it's my newsletter and people should be careful about what they demand Aug 15 09:11:30 i can add a huge DISCLAIMER: I'm not speaking for the XXX project Aug 15 09:11:32 :/ Aug 15 09:13:46 mickeyl: unless you want information from a totally unrelated project nope Aug 15 09:13:55 heh Aug 15 09:13:55 k Aug 15 09:14:24 i think i need a way to use openembeded with my thousand parsec project to try and get more exposure :P Aug 15 09:14:35 :) Aug 15 09:14:54 maybe i should sue linux or something and get mentioned on slashdot Aug 15 09:15:13 ah i don't think you want _that_ kind of publicity Aug 15 09:15:14 heh Aug 15 09:15:48 Hi - I am pretty new to oe and after quite a bit of work I got it all set up on my machine. I try to build gpsd, but I get an error during the build process, due to the wrong version of the required file. Could some one please tell me where to change the old version to the new one. Aug 15 09:15:53 some slashdot attention for OE might be good Aug 15 09:16:08 NOTE: fetch http://www.oesources.org/source/current/flex-2.5.31.tar.bz2 Aug 15 09:16:09 --13:15:54-- http://www.oesources.org/source/current/flex-2.5.31.tar.bz2 Aug 15 09:16:09 => `/home/phm4/progs/oe2/sources/flex-2.5.31.tar.bz2' Aug 15 09:16:09 Resolving www.oesources.org... 192.216.230.225 Aug 15 09:16:09 Connecting to www.oesources.org[192.216.230.225]:80... connected. Aug 15 09:16:09 HTTP request sent, awaiting response... 404 Not Found Aug 15 09:16:11 13:15:55 ERROR 404: Not Found. Aug 15 09:16:13 NOTE: Task failed: Fetch failed: http://aleron.dl.sourceforge.net/sourceforge/lex/flex-2.5.31.tar.bz2 Aug 15 09:16:28 sorry aboutt the last message - really bad copy and paste job :-( Aug 15 09:16:58 why dont you talk to EETimes or the Register ... atleast someone with an ounce of clue and credibility? Aug 15 09:16:58 upstream sources sometimes vanish :/ Aug 15 09:17:14 chip and volts, etc Aug 15 09:18:17 or was it nuts and volts .. Aug 15 09:19:07 mickeyl: i dunno i think after the inital deluge of clueless newbies and people who complain about every i might pick up 1 or 2 developers Aug 15 09:20:25 not sure what the project quarter used here to build linux ... Aug 15 09:21:20 was for a voice activated x10 system using a pxa Aug 15 09:21:52 maybe i could sue linux for being too good? Linux is so stable it put me out of business, back when they ran windows i use to be roling in consolting fees, now people just have a linksys slug or wrt54g and i never have any work :P Aug 15 09:22:07 welp off to cook dinner Aug 15 09:22:46 btw if you wanna see a novel project ... http://cq.cx/pcb-router.pl Aug 15 09:22:50 non-linux ofcourse Aug 15 09:23:02 but interesting Aug 15 09:23:47 mithro: true enough. might help Aug 15 09:23:51 bbl Aug 15 09:24:11 * koen is a bit underwhelmed by the number of projects in http://oe.handhelds.org/cgi-bin/moin.cgi/SuccesStories Aug 15 09:25:20 shouldnt meshbox, uslu2, etc be there? Aug 15 09:25:51 and the other 10 or so projects? Aug 15 09:26:08 or is success just buildable? Aug 15 09:26:28 koen: both Aug 15 09:26:52 koen: this list looks biger Aug 15 09:26:54 http://oe.handhelds.org/success.php Aug 15 09:27:34 it is it is! Aug 15 09:29:49 * koen hints at the fact that it's a wiki and anyone can edit it Aug 15 09:37:58 damnit ... i need to find another dedicated server ... Aug 15 09:38:03 hmm Aug 15 09:41:39 Is there a place to put kernel modules in the OE tree? Looke like they go in packages together with any software package... Shouldn't that kinda be separated? Aug 15 10:06:12 mickey|sports: Then install the GPE image and install the E stuff on top of that (e-wm, entrance). Honestly people, how many times do I need to tell you this is how *I* got it working. Aug 15 10:06:38 CoreDump|afk: why are you not asking me about entrance. I have it working fine. *fine* on my Zaurus. Again, I installed it *on top of* GPE. Aug 15 10:07:28 Argh, I suppose I'll have to make a bunch of commits today to amek the e-image work since people don't seem to want to 1) ask me how I got it working and 2) use the GPE image as they should be if they want it working right. Aug 15 10:13:27 JustinP, e-image should not depend on any aspect of gpe Aug 15 10:13:42 perhaps that is why they keep asking Aug 15 10:14:15 emte: damn it, the e-image is not freaking ready! Aug 15 10:14:36 emte: I am making the e apps work one by one and I'm doing on the GPE image because it already has all of the X stuff on it. Aug 15 10:15:06 emte: not only this, but e is not really too useful without some little bits of GPE around, like the matchbox panel in order to use xmonobut Aug 15 10:15:34 I keep saying that people shoudl installing E on top of GPE for now but not only is no one listening, they're not even askign me how I got it done. Aug 15 10:16:09 i am aware Aug 15 10:16:13 Do you have *any* idea how frustrating it is to be the "official" maintainer for this stuff and have everyone going *around* me trying to get this to work? Aug 15 10:16:24 hence i never attempted to fix the e-image Aug 15 10:16:43 I already have it working. I've had it working for a few weeks now. Yet no-one even thinks to ask me how I got it done. Aug 15 10:16:47 there really is not any point until some apps are written to work on the scale of an embedded device Aug 15 10:17:02 yes i am aware Aug 15 10:17:15 'embedded device' Aug 15 10:17:16 i was actually appointed maintiner of it before :P Aug 15 10:17:30 a 400MHz 64MB ram device with a VGA screen Aug 15 10:17:33 hardly embedded Aug 15 10:18:06 yes but your not going to be able to run a 1024x968 locked screen on that device Aug 15 10:18:31 unfort that is how a lot of programmers write things Aug 15 10:18:33 koen: I agree. The Zaurii are handheld computers, not embedded devices. Well, at least the C3000+ are like this. The 5500 and such are closer to embedded... Aug 15 10:18:37 * koen first ran E on a 133MHZ/48MB computer Aug 15 10:18:58 but there is very little stuff written for e17 Aug 15 10:19:17 emte: we will delete all E stuff till you say it's ready Aug 15 10:19:20 happy now? Aug 15 10:19:31 and the touchscreen issue for e17 will have to be resolved... by me or someone else Aug 15 10:19:52 i wouldnt delete everything Aug 15 10:20:01 koen: no, it's perfectly fine that people are using it Aug 15 10:20:05 but i dont think its right to have e-image either tho Aug 15 10:20:20 koen: I just wish people would ask me how *I* got it working instead of hacking it up even more. Aug 15 10:20:27 not untill it can be built standalone Aug 15 10:20:44 change it to e+gpe-image or something Aug 15 10:20:47 emte: I agree....does BROKEN workf or image bb's? Aug 15 10:20:57 no Aug 15 10:21:01 damn Aug 15 10:21:16 and BROKEN is against our policy anyway Aug 15 10:21:27 if something doesn't work it should go into nonworking/ Aug 15 10:21:41 everything outside nonworking/ is assumed to be working Aug 15 10:21:51 well, it's not working and I wish people would focus on fixing the individual programs before using the e-image Aug 15 10:21:56 it sortof works, but not as it is named Aug 15 10:22:14 * JustinP moved e-image to nonworking, then Aug 15 10:22:17 moves Aug 15 10:22:25 after a pull.... Aug 15 10:22:29 lol Aug 15 10:22:55 koen: fwiw, there is no such policy Aug 15 10:23:17 Well anything to make people realize that they should not be relying on e alone for now Aug 15 10:23:19 and BROKEN just excludes stuff from a world build Aug 15 10:23:27 ehm Aug 15 10:23:32 which is legitimate Aug 15 10:23:39 isn't EXCLUDE_FROM_WORLD suppossed to do that? Aug 15 10:24:54 what the...? Aug 15 10:25:09 ummm...koen...how did you manage to push *two* head revisions? Aug 15 10:25:30 koen: could you please merge the heads Aug 15 10:25:54 which heads? Aug 15 10:26:03 http://monotone.vanille.de/viewmtn/headofbranch.psp?branch=org.openembedded.dev Aug 15 10:26:11 both have your signature Aug 15 10:26:15 (cert) Aug 15 10:26:30 ewi546.ewi.utwente.nl/tmp/viewmtn/headofbranch.psp?branch=org.openembedded.dev only lists one Aug 15 10:26:58 ah...it hasn't been pushed to vanille yet...any reason for that? Aug 15 10:27:21 I guess the 10 minutes for the cron job aren't finished Aug 15 10:27:25 I see Aug 15 10:28:09 the two head were caused by working on two machines and some bad timing Aug 15 10:29:57 it's in vanille now Aug 15 10:30:06 I se Aug 15 10:30:07 e Aug 15 10:30:08 thanks :-) Aug 15 10:31:10 OE is a 17MB bz2 tarball already... Aug 15 10:32:34 koen: well. technically the effect is similar. but that wasn't my point anyway. if you want to point people to a policy you'll have to write/establish one first. we don't have anything like that at present Aug 15 10:35:53 there, maybe this commit will clue people in Aug 15 10:36:53 reenoo_: isn't the 'don't push broken stuff' written down somewhere? Aug 15 10:37:22 Its been in effect since 1973 Aug 15 10:38:01 koen: not that I'm aware of Aug 15 10:41:55 it's not that the e-image is really "broken"...it just doesn't work without lots of love. ;-) Aug 15 10:42:38 * emte conciders it not building standalone pretty broken Aug 15 10:43:30 emte: eh? e will build fine Aug 15 10:43:43 emte: if it doesn't people need to tell me. It builds fine for me Aug 15 10:43:57 emte: and the packages install and work just fine...on a GPE image Aug 15 10:44:00 you havent built it standalone Aug 15 10:44:08 I don't know what you mean by standalone Aug 15 10:44:17 JustinP: ewm doesn't build, do_install fails Aug 15 10:44:18 I built e-wm from a clean tmp dir and it builds fine Aug 15 10:44:27 without GPE Aug 15 10:44:35 koen: is it's failing file a bug or send me an error personally Aug 15 10:44:41 X and the e17 libs Aug 15 10:44:44 emte: what do you mean? e needs X.... Aug 15 10:44:52 I did rm -rf tmp ; bitbake e-wm Aug 15 10:44:58 koen: when? Aug 15 10:45:06 yesterday Aug 15 10:45:06 koen: and why don't you file a bug or send me the error? Aug 15 10:45:13 working on it.... Aug 15 10:45:19 koen: there was an e.h file missing. I fixed that yesterday Aug 15 10:46:04 yep, I ran a completely new build yesterday and it went all the way through Aug 15 10:46:16 I'll run *another* one just to make sure... Aug 15 10:48:39 grr Aug 15 10:48:57 i dont want to rent more servers to test this ... Aug 15 10:50:49 emte: why would you need to? Aug 15 10:51:34 because i havent been able to get ahold of my admin Aug 15 10:52:12 emte: ah... Aug 15 10:52:40 emte: well what do you want to test? I have a pretty fast machine at my disposal that I can use to create testing packages for whichever device you want to test on Aug 15 10:52:45 whooo...just sent my first patch to the oe-mailinglist :) Aug 15 10:52:59 no, its for the monotone->hg stuff Aug 15 10:53:25 i havent decided what to do about emacs yet ... Aug 15 10:53:27 ldc: can you attach it to bugzilla too? That will make sure it won't get lost Aug 15 10:54:32 we know there are two packages that depend on it ... but how many more Aug 15 10:55:39 emte: would adding an emacs-native build perhaps fix the problem? Aug 15 10:55:45 koen, yea, but it needs someone with a x86 system to test it before... Aug 15 10:56:08 JustinP, no Aug 15 10:56:19 its a change in the way emacs handles -batch Aug 15 10:56:38 so when its testing its opening the emacs shell and hanging Aug 15 10:56:49 instead of parsing it Aug 15 10:57:18 emte: ok, I'll leave it alone, then... Aug 15 10:57:57 yeah autoconf and gettext are the current problems with it Aug 15 11:06:08 hmm i might have found the issue ... Aug 15 11:06:49 well i'm stuffed Aug 15 11:07:25 mithro: your router stopped working? Aug 15 11:08:29 nope i ate to much food Aug 15 11:08:56 I did that yesterday Aug 15 11:09:03 or today Aug 15 11:09:09 sometine around midnight Aug 15 11:09:15 bad idea... Aug 15 11:09:35 hmm Aug 15 11:09:46 my parse exclusiosns arnt working ... Aug 15 11:09:52 it's only 8 pm here Aug 15 11:09:54 regex must be wrong Aug 15 11:11:16 mithro: same here! Aug 15 11:11:51 koen: could we split the success stories like the non-dynamic page one? Aug 15 11:12:09 sure go ahead Aug 15 11:13:01 Hello, could some one please help me. bitbake always failes on NOTE: fetch http://www.oesources.org/source/current/flex-2.5.31.tar.bz2 - it tries to load the old flex package, the latest one is pre>flex-2.5.4a.tar.gz I did an monotone update Aug 15 11:13:41 oho: try to change the SOURCEFORGE_MIRROR in openembedded/conf/bitbake.conf Aug 15 11:17:36 should lsmod be in /sbin/lsmod or in /bin/lsmod? there is a conflict between busybox and module-init-tools here. Aug 15 11:18:40 JustinP: http://pastebin.com/337600 Aug 15 11:19:04 according to my debian it should be /bin/lsmod Aug 15 11:19:53 is there any command that will tell you what device a file exists upon? Aug 15 11:20:25 koen: my bitbake.conf is in ./bitbake/conf/bitbake.conf and hasn't got SOURCEFORGE_MIRROR , but /build/conf/local.conf has: export SOURCEFORGE_MIRROR = "http://aleron.dl.sourceforge.net/sourceforge" is this enough or should I also add it to the bitbake.conf ? also is the bitbake.conf in the right directory ? Aug 15 11:20:34 koen: should busybox be patched to reflect that in its /etc/busybox.links? otherwise i get an update-alternatives error when configuring module-init-tools Aug 15 11:21:03 i've tried "find file -printf %D" but it returns a number i don't know how to translate into a /dev file Aug 15 11:21:08 pH5: I think so, could you file a bug on it? Aug 15 11:21:28 oho: bitbake/conf/bitbake.conf doesn't get parsed Aug 15 11:21:37 koen: i'll do. Aug 15 11:21:48 pH5: DaKa or M Aug 15 11:21:49 hmm Aug 15 11:22:00 whats your regex package exclusion line look like? Aug 15 11:22:01 DaKa or NAiL is oour u-a guru Aug 15 11:22:34 * pH5 notes that on his guru list Aug 15 11:22:38 :) Aug 15 11:24:44 hmm maybe it doesnt work anymore .. Aug 15 11:25:26 ;) Aug 15 11:27:14 koen: well that's a strange error.... Aug 15 11:28:18 you have a corrupt file i think ... Aug 15 11:29:09 or somehow you tried to insert an non jpg into edje_cc Aug 15 11:29:19 "you" being...who? Aug 15 11:29:31 thats between you three Aug 15 11:29:37 This is e-wm, it has nothing to do with any changes I've made Aug 15 11:29:43 bitbake, JustinP , and koen Aug 15 11:30:08 but i'd guess bitbake probably did a bad substitution Aug 15 11:30:19 koen: what kind of system are yuo running this on? Aug 15 11:30:37 32bit debian on a dual opteron Aug 15 11:31:00 familiar-0.8.3/h3900 Aug 15 11:31:02 you may not beable to compile eets in bitbake Aug 15 11:31:15 hmmm....well it worked fine yesterday on my little-old Athlon Gentoo system Aug 15 11:31:23 or you may need to write an exception class Aug 15 11:31:37 and e-wm always builds fine on the amd64 system I'm using... Aug 15 11:31:41 emte: who? Aug 15 11:31:48 emte: eet builds fine... Aug 15 11:31:59 .eet files Aug 15 11:32:05 that is what is failing Aug 15 11:32:14 edje_cc is trying to compile one Aug 15 11:32:27 and it was given invalid image data Aug 15 11:32:56 most likly because bitbake did a path substitution in the .eet file Aug 15 11:33:03 would be my guess Aug 15 11:33:05 .... Aug 15 11:33:18 koen: are you sure your snapshot is good? Aug 15 11:33:33 koen: have you tried deleting the tgz its using and having it re-pull form CVS? Aug 15 11:33:53 I can try that Aug 15 11:34:01 koen: http://oe.handhelds.org/cgi-bin/moin.cgi/SuccesStories?action=show Aug 15 11:34:14 to tell you the truth, I haven't deleted my sources directory in a long while....I'm trying that as well myself Aug 15 11:34:38 anyway that is a common error Aug 15 11:34:38 mithro: nice work! Aug 15 11:35:06 i'm to bloated to do anything else Aug 15 11:35:33 I love the dreambox guys for stating qt/e is too bloated :) Aug 15 11:36:09 how so? Aug 15 11:36:25 * treke doesnt see how it's any more or less bloated than x and gtk Aug 15 11:36:33 They're all bloated Aug 15 11:36:41 but bloat is a relative term Aug 15 11:36:59 What's wrong with a console? :) Aug 15 11:37:20 lets see if i fixed emacs ... Aug 15 11:37:37 RP: those whining users want graphics, for some reason.. :) Aug 15 11:38:11 all you need is a console with high enough resolution and aalib Aug 15 11:38:22 treke: they use directfb Aug 15 11:38:37 sdl comes in handy if you want to look at an image .. or svgalib Aug 15 11:38:39 because X is too bloated as well? Aug 15 11:39:54 damn you emacs Aug 15 11:40:29 sdl is not a solution for bloat - its a "solution" for cross platform development Aug 15 11:40:36 * CosmicPenguin doesn't think highly of SDL Aug 15 11:40:56 it's good at what it was designed for Aug 15 11:41:46 koen: does removing the tarball fix the problem? Aug 15 11:42:26 guess its time to learn how to use emacs :( Aug 15 11:42:55 it was meant to basically be a replacement for parts of directx, not to be an embedded gui base :) Aug 15 11:45:28 OE is giving me nothing but problems recently :-( Right now I cannot connect to monotone.vanille.de, so I figured I'd use koen's mirror instead. That crashes monotone :-( I was wondering if it might have to do with identification keys since monotone claims it is the first time I connect to that server which is not the case. Did the key change for koen's mirror? Aug 15 11:49:09 koen: I did copy bitbake.conf to /openembedded/conf and added the mirrors, but I still get the problem that OE is looking for flex-native-2.5.31 - which is outdated. Any ideas ? as I said I did update monotone --- could this be related to the fact that I downloaded the files from http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 and updated it ? there where no errors during the update Aug 15 11:50:23 oho: OE.db.bz2 is a snapshot and fine. Aug 15 11:50:52 Can you do a monotone pull and get revisions written? Aug 15 11:52:02 oho: Can you connect to monotone.vanille.de Aug 15 11:52:04 ? Aug 15 11:52:37 Laibsch, i can Aug 15 11:52:55 Laibsch, you using monotone 0.22/ Aug 15 11:52:56 emte? oho? Aug 15 11:52:57 ? Aug 15 11:53:01 is CIA down? Aug 15 11:53:05 I use .21 Aug 15 11:53:24 hmm Aug 15 11:53:32 monotone is a bitch :/ Aug 15 11:53:39 its not connecting to vanille now ... Aug 15 11:53:45 I can connect fine to dominion.kabel.utwente.nl Aug 15 11:53:50 Just found that out. Aug 15 11:54:06 to bad its the only tool with somewhat close to the features OE needs Aug 15 11:54:06 I think vanille and koen's server have some problem Aug 15 11:54:48 i think koen upgraded his server to 0.22 Aug 15 11:55:09 emte: that doesn't matter Aug 15 11:55:15 Well, I think I rejoiced too early. monotone just crashed with dominion too. Aug 15 11:55:24 monotone 0.22 and 0.2x are compatible Aug 15 11:55:34 okay, thought it could be a reason to crash it Aug 15 11:56:55 it kind of makes me long for the good old days with bitkeeper :( Aug 15 11:57:01 I am getting farther than with either vanille.de (no connection) or koen's server (crash before the merkle trees) Aug 15 11:57:37 koen: Anything strange going on with the servers? Aug 15 11:57:51 re Aug 15 11:58:16 Laibsch: working perfectly over here Aug 15 11:58:30 ??? strange Aug 15 12:00:37 I'll just pull OE.db.bz2 again and rebuild. Not that it got me compiling so far. Aug 15 12:02:53 Laibasch: I can't connect to monotone.vanille.de at the moment - it worked a hour ago Aug 15 12:04:30 but I still got the problem that bitbake tries to pull the wrong flex-native file Aug 15 12:18:24 I just hit https://savannah.nongnu.org/bugs/?func=detailitem&item_id=12595 Aug 15 12:19:02 I had two simultaneous pulls, one from the cronjob (which has taken over an hour now) plus the manual one from my initialisation script. Aug 15 12:27:57 oho: At what stage is it "pulling" the wrong flex-native file? What is the command you use when you experience that problem? Aug 15 12:32:41 I get the problem wenn I do: bitbake gpsd or bitbake nano - everything works fine till it hits flex: Aug 15 12:33:06 NOTE: package gettext-native-0.14.1-r3: task do_build: completed Aug 15 12:33:06 NOTE: package gettext-native-0.14.1: completed Aug 15 12:33:06 NOTE: package flex-native-2.5.31: started Aug 15 12:33:06 NOTE: package flex-native-2.5.31-r2: task do_fetch: started Aug 15 12:33:06 NOTE: fetch http://aleron.dl.sourceforge.net/sourceforge/lex/flex-2.5.31.tar.bz2 Aug 15 12:50:49 Laibsch: are you sure your monotone and boost work properly? How is it crashing? Aug 15 12:52:24 koen, emte: I've passed evas-native in my re-build of e-wm from absolute scratc (including fetching from scratch). I'll let you know how it goes. Aug 15 12:53:04 thats good Aug 15 12:54:08 crap Aug 15 12:54:13 I forgot to restart it Aug 15 12:54:18 * koen does that now Aug 15 12:56:24 damn it.... Aug 15 12:56:27 oh, I know why.. Aug 15 12:56:46 stupid...nothing to do with the checked in code.. Aug 15 12:57:55 ~lart RP Aug 15 12:57:56 * ibot beats RP over the head with a microkernel Aug 15 12:58:21 I think I know why pcmcia has been playing up for a while on c7x0 :-/ Aug 15 12:59:38 why? Aug 15 13:00:50 treke: Someone made a silly error which meant the card detect interupt wasn't getting registered... Aug 15 13:01:03 ha Aug 15 13:01:10 I suspect anyway... Aug 15 13:09:05 JustinP: boost? Aug 15 13:09:05 JustinP: I think I found the reason for the immediate trouble which is bug 12595 -> https://savannah.nongnu.org/bugs/?func=detailitem&item_id=12595 which I pasted here earlier. It was my cron pull job taking longer than expected (almost two hours but done now). Aug 15 13:09:43 BTW, I have created a new page on tips for dealing with slow monotone at http://oe.handhelds.org/cgi-bin/moin.cgi/SpeedUpMonotone Aug 15 13:10:01 If anybody has any hints for speeding up monotone, feel free to share. Aug 15 13:11:19 Laibsch, yeh Aug 15 13:11:25 but its from yesterday Aug 15 13:11:32 and i need to adjust it a bit Aug 15 13:11:39 you have hg installed? Aug 15 13:12:06 hg pull -u http://uhtc.net/cgi-bin/openembedded Aug 15 13:12:13 What is from yesterday? I do not think I have hg installed Aug 15 13:12:16 takes ~ 10 min for the OE repo Aug 15 13:12:31 for initial pull -u Aug 15 13:12:49 you can view the repo at the same url Aug 15 13:12:57 emte: and using a db snapshots takes even less time Aug 15 13:13:20 not really Aug 15 13:13:30 you still have to update Aug 15 13:13:38 so? Aug 15 13:13:40 I do not even claim to understand the content on that URL. Aug 15 13:13:59 emte: monotone is the official SCM Aug 15 13:14:06 i am aware Aug 15 13:14:22 I don't want to get bugreports because someone used something else Aug 15 13:14:28 i am working on the translation ... mostly for non devs/non push people Aug 15 13:14:38 emte: monotone is the official SCM Aug 15 13:15:02 I think it's good that people are experimenting with other version control systems. Aug 15 13:15:33 Laibsch: well, I suggest you add somehting to your cronjob to stop it from happening simultaneously...a lock file perhaps Aug 15 13:15:33 no doubt Aug 15 13:15:49 but they shouldn't market is as 'recommended' Aug 15 13:16:00 i didnt say recommended Aug 15 13:16:19 he just asked for faster Aug 15 13:16:32 JustinP: It checks for the pid of monotone. I think that should be enough, no? Aug 15 13:17:13 if your a dev i'd use monotone Aug 15 13:17:23 Laibsch: yep, that looks like it should work Aug 15 13:17:31 emte: What I meant is to share tips on getting monotone to be a bit faster and put that info on the wiki. But thanks for sharing your info, too. Aug 15 13:17:51 koen, emte: e-wm has fully compiled from scratch just fine Aug 15 13:17:52 buy a server farm? Aug 15 13:18:00 emte: everyone should use monotone Aug 15 13:18:18 emte: monotone is the official SCM Aug 15 13:18:25 yes i am aware Aug 15 13:18:42 so don't go pointing people to your hg repo Aug 15 13:19:01 I'd like to know what problems they encounter Aug 15 13:19:16 the only two as far as i know who are interested are CoreDump|afk and zeke Aug 15 13:19:25 instead of ignoring the problem and sending them on a wild goose chase around various SCM's Aug 15 13:20:12 anyway the intent is for transparent integration Aug 15 13:20:21 which so far seems to be working Aug 15 13:20:35 integration into what? Aug 15 13:20:56 they can both be used at the same time on the same files Aug 15 13:21:12 OE will only have one (1) SCM Aug 15 13:21:25 yup Aug 15 13:22:33 so stop pretending your thing will be 'integrated' Aug 15 13:22:54 it can 'replace' monotone Aug 15 13:23:01 but it would nb Aug 15 13:23:12 need more than being faster than monotone Aug 15 13:25:01 i think your misunderstanding but thats okay Aug 15 13:25:08 koen: have you tried removing that sources tarball yet? I'd like to know if it's your machine / OE checkout or tarball causing the problems. Aug 15 13:25:19 i am only testing it with OE Aug 15 13:25:24 JustinP: removed all cvs stuff and TMPDIR Aug 15 13:25:39 I see...so this is going to take forever again Aug 15 13:25:43 OK, pull and updare work fine now that I have only one process fiddling with the db ;-) Aug 15 13:27:34 btw the emacs people's only suggestion was taht it might be built from cvs ... Aug 15 13:29:49 can someone explain why the docs say "Do not build OE as root", but when I try to run 'bitbake bootstrap-image' it's trying to create /usr/lib/ipkg, which requires root privs ? Aug 15 13:30:10 it shouldnt be. Aug 15 13:30:24 there is either a bug in the bb or your config is wrong Aug 15 13:30:26 the creation of /usr/lib/ipkg should fail Aug 15 13:30:32 it sure does! Aug 15 13:30:43 it doesn't get any further than that Aug 15 13:30:44 OE shouldn't touch anything outside its sandbox Aug 15 13:30:50 ispeedy1: this is why you shouldn't run it as root ;-) Aug 15 13:30:56 heh :) Aug 15 13:30:57 ispeedy1: those things *should* fail Aug 15 13:31:25 or else even scarier things will succeed Aug 15 13:31:29 * koen wouldn't want any ARM binaries replacing system tools Aug 15 13:31:49 "ok look. I have an arm init" Aug 15 13:31:58 koen, well I wouldnt mind much Aug 15 13:32:04 koen, if I had an arm desktop Aug 15 13:32:49 Is it a bad idea to have a "monotone update" and a "bitbake nano" process running at the same time? Aug 15 13:33:08 treke: :) Aug 15 13:33:25 Laibsch: not recommended Aug 15 13:33:36 you may be fine Aug 15 13:33:36 OK, I'll stop the update process. Aug 15 13:33:44 but necessary files could disappear out from under it Aug 15 13:34:02 No, I do not want any trouble. I'll do the update later. Aug 15 13:34:04 Laibsch: I'd stop the bitbake process Aug 15 13:34:21 * koen is scared by interrupting monotone processes Aug 15 13:35:00 koen: That is of course a valid point. Aug 15 13:35:18 bitbake is way more stable than monotone. Aug 15 13:37:00 A byte walks into a bar and orders a pint. Bartender asks him "What's wrong?" Byte says "Parity error." Bartender nods and says "Yeah, I thought you looked a bit off." Aug 15 13:37:22 Laibsch: monotone is stable. I have never had it fail on me (well, I have had some strange pull problems locally but the main OE stuff works fine) Aug 15 13:37:38 mithro: heh Aug 15 13:37:53 JustinP: Well, I have had stuff mysteriously disappear just yesterday. Aug 15 13:38:34 after running bitbake bootstrap-image, I get the error message about /usr/lib/ipkg, and there's no image/ directory under build/tmp, is this something an update might fix ? Aug 15 13:39:04 the /usr/lib/ipkg is a red herring Aug 15 13:40:18 ok to paste several lines from the bitbake output here ? Or is there a channel for that? Aug 15 13:40:38 there's more after /usr/lib/ipkg error Aug 15 13:40:45 ispeedy1: rafb.net/paste Aug 15 13:40:46 ispeedy1: use a pastebin Aug 15 13:42:11 emte: I guess you are the author of hg? Is the idea behind hg to create a pull only client to monotone that will be faster because of it? Sounds like an interesting approach but of course, I'd wait until its stable to use it for anything but testing. Aug 15 13:42:26 author? no Aug 15 13:42:33 who is? Aug 15 13:42:38 working with the tailro files Aug 15 13:42:42 tailor* Aug 15 13:42:44 what is the idea behind it? Aug 15 13:42:45 ~pastebin Aug 15 13:42:46 methinks pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca Aug 15 13:43:23 hg itself is just a fast cross platform repo Aug 15 13:43:31 not much more Aug 15 13:43:48 hmm Aug 15 13:44:23 how does monotone come into the picture? You maintain a conversion server? Aug 15 13:45:17 Laibsch: he's trying to test hg as a possible replacement for monotone Aug 15 13:45:33 tailor is a set of scripts to provide translation to and from different systems Aug 15 13:45:45 no, not as a replacement Aug 15 13:46:03 The stepping stone has always been decentralized versioning. Does it support it? Aug 15 13:46:25 all i am doing is trying to finish the tailor monotone script Aug 15 13:47:06 Laibsch, if your interested you can read about it here: http://www.selenic.com/mercurial/ Aug 15 13:47:09 pb_: ping? Aug 15 13:47:15 NAiL: yo Aug 15 13:47:26 PARALLEL_INSTALL_MODULES... how does it work? ;) Aug 15 13:47:45 It's a toggle, but how does it figure out what is modules? inherit module? Aug 15 13:48:21 yeah, iirc it operates on the stuff that comes from kernel.bbclass and module.bbclass Aug 15 13:49:12 Ok, because I want to find a way to force rebuild of a certain module when the kernel has bumped PR Aug 15 13:49:13 emte: Not too interested, just curious. does it support decentralized versioning or is it really nothing but a (read-only?) frontend to monotone? Aug 15 13:49:30 (Actually, three or four of 'em) Aug 15 13:50:06 hmmm Aug 15 13:50:08 Laibsch: yes, it claims it does. Aug 15 13:51:44 emte: if you set up a server for OE sources, I'd be willing to play around with it by setting up a second OE tree and see if I can run into the same problems as in the original tree ;-) Aug 15 13:52:42 emte: So far I only do pulls. And I will probably never become much of a pusher even in the best of times. Aug 15 13:52:53 ok, I've pasted bitbake bootstrap-image output here: http://rafb.net/paste/results/NBvkRg87.html .. Aug 15 13:54:12 Laibsch: *NO* hg has nothing at all to do with monotone Aug 15 13:54:23 Laibsch: it's a completely different SCM Aug 15 13:55:19 emte: if it's not going to be a replacement then koen and others are going to have issues. Patching another system into the current one is *not* a good solution and could quite easily cause corrupted changesets. Aug 15 13:55:44 JustinP: I understood it that emte is thinking about proposing Aug 15 13:55:57 a replacement - I was just about to write Aug 15 13:56:10 I ran bitbake task-bootstrap beforehand, built without errors Aug 15 13:56:43 I guess not seeing it as a replacement is a bit of - hmm, pointless? At least I do not see the point if it not related to monotone in any way. Aug 15 13:57:45 though I won't be surprised if my config is wrong, I'm trying to make it build for a target I haven't seen any machine or distro files for (custom Motorola 860 board) Aug 15 13:58:21 *any similar machine ... Aug 15 13:59:22 there's no wiki docs covering that yet, operating in the dark here Aug 15 13:59:38 Laibsch: it's an alternate system which is supposedly faster. Some people have problems with the speed of monotone. Aug 15 14:00:54 JustinP: it's at diet-x11 right now Aug 15 14:01:09 how is it everyone is jumping to this comcept of me proposing anything? Aug 15 14:01:13 concept* Aug 15 14:01:39 i am doing nothing more than finishing a script and testing it Aug 15 14:01:48 JustinP: I am just discussing the question of speed on the monotone channel. The devs there claim that speed is a bit of a non-issue except for certain circumstances such as many big files. I was trying to get an opinion there if a read-only client could be made faster than the full binary. Aug 15 14:03:23 Laibsch: no, not unless you remove the cert/changeset validation Aug 15 14:03:44 Laibsch: the speed isn't that bad. It just takes a little while to pull lots of revisions...oh well... Aug 15 14:04:14 pb_: Ok, what am I not getting now... It changed the default kernel modules names from kernel-module-nls-base_2.6.12.2 to kernel-module-nls-base-2.6_2.6.12.2 (Adding -2.6). Aug 15 14:05:08 JustinP: a new system has to do a lot more than 'be faster' to justify a switch Aug 15 14:05:23 koen: I completely understand that Aug 15 14:05:30 koen: I think monotone is working out ok for us Aug 15 14:05:37 But it doesn't seem to rebuild stuff when I bump the PR. Only the kernel modules (obviously) are recompiled Aug 15 14:05:45 JustinP: the --lca thing has improved it a lot Aug 15 14:05:47 koen: I'm just suggesting that that seems to be the main point about monotone that people want to "fix" Aug 15 14:05:54 JustinP: true Aug 15 14:06:08 (other modules, like zd1211 and that at7csomething doesn't get reompiled) Aug 15 14:07:04 big feet eat meat Aug 15 14:07:19 NAiL: I don't think OE supports that kind of behaviour Aug 15 14:07:28 hmm Aug 15 14:07:37 Guess I'll have to hack it a bit then ;) Aug 15 14:07:43 :) Aug 15 14:08:10 koen: Is it possible to have an "include_openslug"? Aug 15 14:08:20 no idea Aug 15 14:09:53 nope, does not appear to be possible :( Aug 15 14:10:39 If it was I could put the openslug-kernel PR in a separate .inc, and include it in modules that should be rebuilt automatically... Aug 15 14:10:49 but that is an ugly hack anyway Aug 15 14:11:34 NAiL: maybe something in the kernel .bb could remove the stamps from the modules :-) Aug 15 14:11:47 that's probably an even uglier hack :-) Aug 15 14:12:00 well, it'd be more openslug-specific ;) Aug 15 14:12:20 Just hit this error (OE continued) in the "handling bitbake files" stage: ERROR: stat() argument 1 must be (encoded string without NULL bytes), not str while parsing /mnt/pfaffe/src/openembedded/org.openembedded.dev/packages/linux/handhelds-sa_2.4.19-rmk6-pxa1-hh36.12.bb Aug 15 14:18:39 pb_: is there a way of getting the ${PV} for virtual/kernel somehow out of the metadata whilst building a different kernel module (like the ixp modules that NAiL is talking about) ? Aug 15 14:18:54 JustinP: still the same error Aug 15 14:21:22 rwhitby: I bumped the ixp modules earlier today, since there had been kernel changes. It'd be nice to not have to do that ;) Aug 15 14:21:39 NAiL: that's why I'm asking the question. Aug 15 14:21:42 JustinP: http://pastebin.com/337743 Aug 15 14:22:18 rwhitby: you can get the kernel-abiversion, but (iirc) not the whole ${PV}. Aug 15 14:22:30 NAiL: actually, you're right. it would need to get ${PR} too. Aug 15 14:22:32 could add a mechanism to stash that in staging and pull it out again later, I guess, if it's important. Aug 15 14:22:53 pb_: which class does the kernel-abiversion stuff? Aug 15 14:23:21 NAiL: fwiw, if you are finding that a PR bump causes a kernel ABI incompatibility, you probably ought to rethink your versioning scheme Aug 15 14:23:29 rwhitby: kernel.bbclass, I guess Aug 15 14:23:48 pb_: actually, if we make these other packages depend on virtual/kernel, then would they recompile if the kernel PR changes? Aug 15 14:24:16 What is a likely problem if the do_stage phase of the first package already fails? More info on http://rafb.net/paste/results/ZEd7Or76.html Aug 15 14:24:17 rwhitby: no; OE doesn't (currently) have any mechanism to do inter-package versioned dependencies like that. Aug 15 14:25:13 NAiL: looks like it remains manual for the time being ... Aug 15 14:25:45 rwhitby: hmm... oh well Aug 15 14:26:24 NAiL: the ixp modules shouldn't need to be recompiled unless the kernel ${PV} changes. changes to the kernel PR shouldn't require a recompilation of those modules. Aug 15 14:26:57 pb_: if the other packages depend on kernel/virtual, would they recompile if the PV (not the PR) changes? Aug 15 14:27:20 no Aug 15 14:27:41 heya raster Aug 15 14:27:53 that's arguably a bug, but isn't obvious to me how to address it Aug 15 14:27:55 bling bling Aug 15 14:27:57 you still on strike? Aug 15 14:28:20 yup Aug 15 14:28:25 lol Aug 15 14:28:39 hell no, we won't go! Aug 15 14:28:48 more like accept Aug 15 14:28:58 he's on strike from email Aug 15 14:29:24 too much bloody email Aug 15 14:29:29 i answer some Aug 15 14:29:34 and there's more to answer Aug 15 14:29:39 i do that and more has turne dup Aug 15 14:29:47 These things happen when you are a open source demigod Aug 15 14:29:52 i dont get time to code because i'm always bloody doing email Aug 15 14:30:03 and iuts pissing me off Aug 15 14:30:10 i want to code Aug 15 14:30:14 not write bloody email Aug 15 14:30:18 or read it Aug 15 14:30:49 thus i'm on strike from email Aug 15 14:30:53 for a while Aug 15 14:31:00 to try get time to do other things Aug 15 14:31:02 like... code Aug 15 14:31:11 so if you get email from me at all Aug 15 14:31:15 it's a miracle Aug 15 14:31:17 atm. Aug 15 14:31:18 :) Aug 15 14:31:40 well you know i email you all the time Aug 15 14:34:17 hmm Aug 15 14:34:20 i wonder Aug 15 14:34:39 ... Aug 15 14:34:42 grrrr Aug 15 14:34:50 email is the spawn of satan Aug 15 14:43:47 koen: I'm not sure what to say...it's building fine on my two machines Aug 15 14:43:50 re Aug 15 14:43:52 JustinP: ping Aug 15 14:44:14 CoreDump|home: PONG! Aug 15 14:44:18 JustinP: not on my 3 machines Aug 15 14:44:40 (debian, debian and ubuntu) Aug 15 14:44:56 hmmm....my 2 are gentoo Aug 15 14:45:05 perhaps this is another edje error that's not quite right... Aug 15 14:45:27 raster: could you take a quick look at the error here: http://pastebin.com/337743 and let us know if you know what it might be? Aug 15 14:46:45 koen: thanks for submitting a bug report :-) Aug 15 14:46:55 NOTE: package ipkg-utils-native-1.6cvs20050815-r4: task do_patch: started Aug 15 14:46:55 JustinP: CoreDump|afk: why are you not asking me about entrance. Aug 15 14:47:01 ERROR: function do_patchcleancmd failed Aug 15 14:47:19 because obviously I try to make it work w/o GPE Aug 15 14:47:37 DISTRO=familiar-0.8.3 Aug 15 14:48:26 I'm trying to do bitbake bootsrap-image Aug 15 14:48:29 CoreDump|home: yes, but I could have given you the correct session files and such. Aug 15 14:48:44 CoreDump|home: there isn't much sense in getting it to run without GPE until it runs with GPE Aug 15 14:48:52 JustinP: were off when I'm online and vice versa# Aug 15 14:48:57 how does that make sense? Aug 15 14:49:06 I was trying to determine what is all needed Aug 15 14:49:45 JustinP: sounds like cpp inst working on the target system Aug 15 14:49:51 CoreDump|home: I understand but it *has* been a couple of days...and I do have e-mail and I usually read back on #oe, especially if someone msgs me Aug 15 14:49:58 edje_cc puts all .edc files thru cpp Aug 15 14:50:10 so ui can to #define macro()'s #ifdef etc. Aug 15 14:50:27 does any1 know the miminal required to get a cf slot working? Aug 15 14:50:58 IMO (and that's just _my_ opinion) e-image shouldn't rely on GPE at all. IMO it should work perfectly fine by itself (that is, no GPE installed) and in combination with GPE if the user wishes so Aug 15 14:51:02 emte: 1) e doesn't have all the programs it needs currently. 2) Get it to run on a known working X image, *then* make it use its own. If entrance fails how do you know if it's because it's broken or becaus eyou forgot to include some package? Better to make it work first, then try the image. Aug 15 14:51:27 raster: thanks, we'll look into that then Aug 15 14:51:37 JustinP, because it builds fine on other systems? Aug 15 14:52:14 CoreDump|home: ok. makes sense. *MY* point is that the various programs should be gotten to work before we try to build a system out of it. Building a system out of broken parts is a bad idea. Aug 15 14:52:38 CoreDump|home: for now it's best to test it on another system. *Then when they work* we make them work on their own. Aug 15 14:52:40 JustinP: I agree, we just have different ways of achieving the same goal Aug 15 14:52:44 putting icing on a cake before you build the cake is a bad idea as well Aug 15 14:53:32 JustinP: did you manage to login into GPE or E w/ entrance? Aug 15 14:53:40 CoreDump|home: yes Aug 15 14:53:45 CoreDump|home: *and* Enlightenment Aug 15 14:54:12 ok, doesn't work for me. entrance always executes /etc/X11/Xsession Aug 15 14:54:26 which it shouldn't as by the README Aug 15 14:55:13 entrance eciciute Xsession? Aug 15 14:55:19 emte: yeah Aug 15 14:55:20 exiciutes* Aug 15 14:55:25 odd Aug 15 14:55:28 CoreDump|home: check my patches Aug 15 14:55:34 * CoreDump|home pulls Aug 15 14:56:08 emte: you can choose to use Xsession to launch your session or define custom session scripts Aug 15 14:56:25 the latter wouldn't work for me :) Aug 15 14:56:25 CoreDump|home: the entrance code runs Xserver, not Xsession Aug 15 14:56:35 JustinP: during login Aug 15 14:56:39 I changed it to allow us to still use the rotation stuff Aug 15 14:56:40 not during startup Aug 15 14:56:45 CoreDump|home: all that is in the DB Aug 15 14:56:50 right Aug 15 14:57:06 but the setting in the db does _not_ work for me. Aug 15 14:57:41 session = /some/path -> entrance executes "/some/path" during login Aug 15 14:58:04 session = "JustAName" -> entrance runs Xsession JustAName Aug 15 14:58:32 session = "Default" -> entrance just runs Xsession w/o parameter Aug 15 14:58:43 my entrance does always the latter Aug 15 14:59:48 could you do me a favor and post entrance's logmessages after a successful login on your side pls? Aug 15 15:02:46 * CoreDump|home didn't get any new patches for entrance after pulling Aug 15 15:05:59 CoreDump|home: these aren't new patches Aug 15 15:06:09 CoreDump|home: I was talking about Xserver, not Xsession... Aug 15 15:06:16 'night all Aug 15 15:08:54 CoreDump|home: you mean the output that "entranced -nodaemon" shows? Aug 15 15:09:26 * CoreDump|home tries Aug 15 15:15:32 03koen 07org.oe.dev * rf9a24259... 10/packages/libcroco/libcroco_0.6.0.bb: Add libcroco, a css parsing an manipulation library Aug 15 15:15:36 03rwhitby 07org.oe.nslu2-linux * ra64c34c3... 10/packages/linux/unslung-kernel_2.4.22.l2.3r63.bb: Fixed the unslung binary kernel to build from source as well as using the binary (so that our additional kernel modules make it into the rootfs). Aug 15 15:15:40 03koen 07org.oe.dev * rdba8bfab... 10/packages/glib-2.0/glib-2.0_2.8.0.bb: Add glib 2.8.0, ABI and API compatible with 2.6.x Aug 15 15:15:44 03koen 07org.oe.dev * re0e76426... 10/packages/libcroco/libcroco_0.6.0.bb: Fix staging Aug 15 15:15:48 03nail 07org.oe.nslu2-linux * r43d115fb... 10/packages/linux/ (nslu2-kernel_2.6.12.2.bb openslug-kernel-2.6.12.2/defconfig): Add a bunch of iptables/bridging-related modules Aug 15 15:15:52 03koen 07org.oe.dev * rf8deecc6... 10/packages/librsvg/librsvg_2.9.5.bb: Add librsvg 2.9.5 Aug 15 15:15:56 03nail 07org.oe.nslu2-linux * r3eb93977... 10/packages/ (ixp425-eth/ixp425-eth_1.1.bb ixp4xx/ixp4xx-csr_1.4.bb): Bump PR Aug 15 15:16:00 03ccsmart 07org.oe.nslu2-linux * r33f54261... 10/packages/nail/nail_11.21.bb: Removed bogus copy and paste INITSCRIPT definition. Aug 15 15:16:04 03justinp 07org.oe.dev * r6de4d6cd... 10/packages/ (meta/e-image.bb nonworking/meta/e-image.bb): Move e-image to nonworking. Please use the GPE image and install e packages on top of it. e is not ready to run standalone. Aug 15 15:16:08 03ccsmart 07org.oe.nslu2-linux * r38a44153... 10/packages/initscripts/ (initscripts-1.0/bootmisc.sh initscripts_1.0.bb): Add support for /etc/sysctl.conf . Aug 15 15:16:12 03koen 07org.oe.dev * r08fd242b... 10/packages/pango/pango_1.9.1.bb: Add pango 1.9.1 Aug 15 15:16:16 03koen 07org.oe.dev * reffc35e7... 10/packages/atk/atk_1.10.1.bb: Add atk 0.10.1 Aug 15 15:16:20 03koen 07org.oe.dev * re7dc72e4... 10/packages/libpixman/libpixman_0.1.6.bb: Add libpixman 0.1.6 Aug 15 15:16:24 03koen 07org.oe.dev * rb8bc4287... 10/packages/cairo/ (cairo_0.9.2.bb libsvg-cairo_0.1.6.bb): update cairo and libsvg-cairo to the latest snapshot Aug 15 15:16:27 03koen 07org.oe.dev * r4fc219a6... 10/packages/treecc/ (3 files in 2 dirs): Add treecc, closes #226 Aug 15 15:16:31 03koen 07org.oe.dev * r84064808... 10/packages/treecc/ (treecc-native_0.3.6.bb treecc_0.3.6.bb): remove PR='' from treecc Aug 15 15:16:35 03justinp 07org.oe.dev * rbe25e7a6... 10/packages/e17/ (entrance/longer-sleep.patch entrance_20050810.bb): Removing unneeded longer-sleep patch Aug 15 15:18:12 hmm Aug 15 15:18:19 * mickeyl likes to remove hotplug-ng but it's used by one distro Aug 15 15:19:48 gah Aug 15 15:19:56 'hotplug-ng is dead. Forget that it exists.' Aug 15 15:20:07 righto. hotplug-ng is already being obsoleted before it was really in serious use and before people think it's a replacement and we have to support dozens of hotplugging schemes I'd rather remove it and point out the "real" successor. Aug 15 15:20:32 ibot: hotplug-ng is dead Aug 15 15:20:33 mickeyl: okay Aug 15 15:20:40 ibot: hotplug-ng is also read http://sourceforge.net/mailarchive/forum.php?thread_id=7936580&forum_id=3157 Aug 15 15:20:41 okay, mickeyl Aug 15 15:20:45 ~hotplug-ng Aug 15 15:20:45 hotplug-ng is probably dead, or read http://sourceforge.net/mailarchive/forum.php?thread_id=7936580&forum_id=3157 Aug 15 15:20:50 :) Aug 15 15:20:53 yeppa :) Aug 15 15:21:39 and I *hate* the fact that the kernel people are pushing initrds as the solution to everything Aug 15 15:21:46 ACK Aug 15 15:21:58 some of us have sucky bootloaders Aug 15 15:22:07 or value our flash space Aug 15 15:22:27 mickeyl: is the hotplug-ng using distro an nslu2-linux one? Aug 15 15:22:32 rwhitby-away: yes Aug 15 15:22:38 rwhitby-web: openslug is the only distro Aug 15 15:22:38 OpenSlug ? Aug 15 15:22:51 ok, what do we need to do? Aug 15 15:23:03 rwhitby-away: use udev :) Aug 15 15:23:10 'Combine udev with HAL and you pretty much have the full functionality' Aug 15 15:23:22 HAL is frikking huge and overkill Aug 15 15:23:29 udev now listens on netlink which is quite cool Aug 15 15:23:50 mickeyl: ok, udev is next on the list of development for OpenSlug - NAiL is leading that. Aug 15 15:24:09 mickeyl: but since you have OpenSlug running now, you can help us too :-) Aug 15 15:24:12 rwhitby-web: excellent. we can team up to get udev working as a fully replacement of hotplug then Aug 15 15:24:15 righto Aug 15 15:24:33 mickeyl: if you know how to do it, feel free to go right ahead. Aug 15 15:24:45 * rwhitby-web has no udev-foo Aug 15 15:25:15 'The bad side-effect is that on bootup your hook got called 600 times for Aug 15 15:25:16 all the /dev/tty* mess the kernel provides.' Aug 15 15:25:26 nah, you can disable that already Aug 15 15:25:33 yeah Aug 15 15:25:43 nail had some fun with that already yesterday Aug 15 15:26:24 mickeyl: anyway, consider yourself and koen to have license to make whatever changes are necessary to get udev running on OpenSlug to replace out hotplug-ng. Aug 15 15:26:35 rwhitby-web: i'm still wrapping my head around it. at first i want to get it working on our 2.6.13 Zaurus. it works there (needs udev scripts for SD/MMC etc.) it works everywhere. Aug 15 15:26:40 rwhitby-web: thanks. Aug 15 15:27:10 last time i tried i got slapped by alignment errors Aug 15 15:27:17 i hope the current release behaves better Aug 15 15:27:22 (udev, that is) Aug 15 15:27:48 it doesn't look like greg K-H is having an arm testing board Aug 15 15:28:02 and rmk still seems to live in userland stoneage Aug 15 15:28:03 heh Aug 15 15:28:19 that makes us the guinea pigs Aug 15 15:28:39 * mickeyl grunts Aug 15 15:29:18 mickeyl: rmk will remain in the stoneage just to be able to say "ow, if you change that I'll have to shutdown arm-linux for 6 months to rebuild stuff" Aug 15 15:29:22 OT: auto_stage_libraries works like a charm. I think I'll leave it for now, but after the next batch of releases we might touch a lot of packages to use that Aug 15 15:29:40 koen: heh. sounds like you read alkml, too :) Aug 15 15:31:10 * CoreDump|home yawns Aug 15 15:32:20 * mickeyl considers ordering http://www.thinkgeek.com/computing/input/7727/ Aug 15 15:32:34 heh Aug 15 15:32:52 then again, nothing beats my natural keyboard Aug 15 15:33:03 ~praise microsoft hardware Aug 15 15:33:03 All hail microsoft hardware! Aug 15 15:33:08 ~lart microsoft software Aug 15 15:33:09 * ibot gives microsoft software an extra strength ACME sleeping pill, sending microsoft software to sleep for 150 years, and awakening to seven strange dwarfs and a large apple Aug 15 15:33:14 how l33t Aug 15 15:33:46 I want this one: http://www.thinkgeek.com/gadgets/electronic/6984/ Aug 15 15:33:47 a couple of years more usage of my keyboard and it'll look the same Aug 15 15:33:53 mickeyl: should we send greg K-H a slug ? Aug 15 15:34:11 rwhitby-away: that would help :) I'm afraid he already has enough projects thogh Aug 15 15:34:14 though, even Aug 15 15:34:29 i'm sure we will work it out without him Aug 15 15:34:47 I'll send a mail if the alignment errors are still there Aug 15 15:34:55 koen: heh, that's crazy Aug 15 15:35:00 well, if someone who has a direct line to him wants to offer, the nslu2-linux project will cover the cost of sending him a slug. Aug 15 15:35:14 is there an appropriate forum for build questions? Aug 15 15:35:35 koen: hmm... some Zaurii have usb host.... that would work with the aquarium Aug 15 15:35:39 forums are never appropriate Aug 15 15:35:42 mickeyl: :) Aug 15 15:36:03 http://www.thinkgeek.com/computing/input/785b/zoom/ hehe Aug 15 15:36:31 uaah Aug 15 15:36:33 crazy Aug 15 15:37:39 does that come with recipes for sessions at your local orthopaedist ? Aug 15 15:38:23 * mickeyl thinks prescriptions is probably a better synonym here Aug 15 15:39:03 ;) Aug 15 15:39:27 yay! Aug 15 15:39:56 I found another reason why I don't like zaurus using whiners: http://www.thinkgeek.com/computing/input/785b/zoom/ Aug 15 15:40:02 ehm Aug 15 15:40:13 http://www.oesf.org/forums/index.php?showtopic=14421&view=findpost&p=91942 Aug 15 15:40:14 cut'n'paste error in line 1 Aug 15 15:40:30 osx stinks at copy/pasting Aug 15 15:40:36 linux is much better in that regard Aug 15 15:40:39 heheh Aug 15 15:40:58 complainers Aug 15 15:41:04 * mickeyl swallows a valium and forgets Aug 15 15:43:31 poor guys. the truth is they are so angry because they really would like to use OZ but is has some quirks for them and it looks we're not interested to remove those for them Aug 15 15:43:40 * mickeyl shrugs Aug 15 15:44:08 and that they are not interested in reporting them through the proper channels Aug 15 15:44:24 the sad truth is that when you get into it enough to fix those quirks you forget about them because they're fairly easy to solve or work around Aug 15 15:45:25 true Aug 15 15:45:45 * mickeyl refrains from even answering on that thread Aug 15 15:45:54 smart choice Aug 15 15:46:02 like, who on earth needs a package-manager GUI? :) Aug 15 15:47:52 exactly Aug 15 15:48:10 depends if you have a kbd ... Aug 15 15:48:34 i cant tell you how annoying it is to try and enter an ipkg on an ipaq is ... Aug 15 15:48:50 especially if you misspell it Aug 15 15:50:01 emte: heh right Aug 15 15:50:20 all hail the TAB key Aug 15 15:50:38 the tab works on familiar again? Aug 15 15:50:49 * CoreDump|home dunnos Aug 15 15:50:50 maybe i need to change my image Aug 15 15:51:04 * koen hopes he used the fancy words right this time Aug 15 15:51:07 TAB complete is b0rked on fam? ouch Aug 15 15:51:21 tab key works for me Aug 15 15:51:26 alt-tab too Aug 15 15:51:38 works fin if i use minicom ... just not via touch Aug 15 15:51:44 only the '=' key doesn't work in rxvt Aug 15 15:51:58 hmm Aug 15 15:52:00 all with xkbd Aug 15 15:52:05 time for a new image i guess Aug 15 15:53:09 hmmm Aug 15 15:53:25 *sigh* Aug 15 15:53:36 looks like no patch from libxine 1.0.0 is applying to 1.0.2 Aug 15 15:53:44 we really should send more stuff upstream Aug 15 15:54:16 03koen 07org.oe.dev * r2d7197ee... 10/packages/cairo/cairo_0.9.2.bb: Remove libpixman from cairo depends, since it was usurped into cairo Aug 15 15:54:19 hmm Aug 15 15:54:26 multiple update candidates Aug 15 15:54:28 mickeyl: http://oe.handhelds.org/cgi-bin/moin.cgi/SuccesStories Aug 15 15:54:39 woot Aug 15 15:54:41 looks nice Aug 15 15:54:48 btwn JustinP and koen Aug 15 15:55:19 mickeyl: mithro styled it Aug 15 15:56:44 * koen works some more on the newsletter Aug 15 15:59:04 emte: blame koen, he just committed Aug 15 15:59:15 got fixed after he commited Aug 15 16:00:21 ah Aug 15 16:00:24 .... Aug 15 16:00:35 I know I checked after I committed last... Aug 15 16:01:03 you committed just after I did monotone pull and update Aug 15 16:03:40 I see :-) Aug 15 16:05:23 having mythfrontend on my pda would be cool Aug 15 16:05:46 slooooooow, but cool Aug 15 16:11:27 g'night Aug 15 16:11:47 'night mickey|zzZZzz Aug 15 16:12:08 please send the latest newsletter entry to me before leaving Aug 15 16:12:18 release schedule is wednesday Aug 15 16:12:26 * mickey|zzZZzz wanders off Aug 15 16:13:59 will do Aug 15 16:14:06 * koen leaves tomorrow evening Aug 15 16:14:21 n8 all Aug 15 16:21:10 03rpurdie 07org.oe.dev * ra83285fb... 10/packages/linux/linux-openzaurus_2.6.13-rc5-mm1.bb: linux-oz-2.6: Update the pcmcia to scoop interface fixing a card detect irq problem for both spitz and c7x0 Aug 15 16:37:29 03dyoung 07org.oe.nslu2-linux * r73e24017... 10/packages/linux/ (unslung-kernel/defconfig unslung-kernel_2.4.22.l2.3r29.bb): Added the missing NetFilter options Aug 15 16:46:19 JustinP: I've straced it and put a link to the output in bugzilla Aug 15 16:46:48 * koen zzz now Aug 15 16:46:53 good night all Aug 15 16:50:32 thanks koen.... Aug 15 16:59:06 ~lart Pierre A.J. Aug 15 16:59:07 * ibot hits Pierre A.J. with an anvil and laughs with a contralto voice ... Haha Ha HA Ha Aug 15 16:59:10 ~lart PHP Aug 15 16:59:11 * ibot puts PHP into a headlock and administers a mighty noogie, rubbing half of PHP's hair of Aug 15 16:59:14 ~lart all this crap Aug 15 16:59:15 * ibot blasts all this crap with a huge firehose then strangles all this crap with it Aug 15 17:29:53 oho: Are you still there? Aug 15 17:31:08 give me love baby Aug 15 17:32:34 hey CoreDump|afk you there by chance? Aug 15 17:35:00 that damn monotone .db file is evil i tell ya Aug 15 17:35:26 parsing it takes some serious juice Aug 15 17:57:04 has any1 got epenembedded to work on OSX yet?? Aug 15 17:58:36 you might want to rephrase that question Aug 15 17:58:51 sorry... Aug 15 17:59:04 are you asking about building for an apple or running the build system in osx? Aug 15 17:59:46 if you mean in osx, i am not aware of it currently working in any BSD Aug 15 18:00:25 i wanted to get the OE env on my powermac so i can start building stuff for the 5500 & 5600 and was woundering if it was possable from what i have read it should be? thought i would ask first to save me heeps of time stuffing around Aug 15 18:01:09 if OSX contains the linux compatability code you "might" be able to ... otherwise you would need soemthing like vmware Aug 15 18:01:57 alternativly you would install blacklab or yellowdog on your mac ... Aug 15 18:02:02 could* Aug 15 18:02:22 i think there are other mac distros that exist now as well Aug 15 18:02:23 ok, vmware dosn't run on ppc's :( only qemu and virtual pc andd this is only a 500mhz so virtual pc is out of the Q.... i could always use ubuntu from my duel boot but i will try and get it working on osx first i think Aug 15 18:02:51 find out if anyone has gotten it to work on a BSD Aug 15 18:03:03 if not, you will have quite a bit of work to do Aug 15 18:03:42 tell me about it been compiling stuff for months on my 5500... so used to doing it the hard/slow way :P Aug 15 18:04:26 i'll start installing it and see how i go Aug 15 18:04:48 ThanX Aug 15 18:07:17 np Aug 15 18:08:55 looks like fink has all the stuff we need so maybe in luck so far :) Aug 15 18:21:33 nasty nasty bug Aug 15 18:24:09 who what where what??? Aug 15 18:25:20 nick ZonE-HaKin9|AFK Aug 15 18:29:35 morning kergoth Aug 15 18:30:24 heh Aug 15 18:30:38 was home sick today, woke up with a fever and messed up stomach, and today was my third day at the new job Aug 15 18:30:41 sucks Aug 15 18:30:43 :( Aug 15 18:32:38 :( Aug 15 18:32:52 * emte sends kergoth more penicilin Aug 15 18:40:32 i wouldve gone into work if it was just hte messed up stomach, but i cant work worth a shit when i've got a fever, it fucks up my concentration Aug 15 18:40:34 heh Aug 15 18:40:43 yup Aug 15 18:40:48 well atleast its valid Aug 15 18:40:53 yea Aug 15 18:44:18 well just incase anyone is curious dont use hg with apache's caching enabled Aug 15 18:44:29 unless you want to crash your server Aug 15 19:36:01 evening kergoth Aug 15 19:36:16 heh, hey Aug 15 19:36:19 how goes? Aug 15 19:36:35 eh Aug 15 19:36:38 same as usual Aug 15 19:36:59 Sorry 2 interprupt, but...... any1 else here willing to help with 5600 support? Aug 15 19:37:20 this is not a distribution or device support channel. Aug 15 19:37:35 this channel is for developer discussion, and support of those using the development tools OE and bitbake to build things Aug 15 19:38:15 ok... i was trying to keep support for the 5600 going... dats kool i'll ask in the openz chan Aug 15 19:39:42 btw i'm trying to also get OE & bitbake running on OSX so will report back if all things go well....den i can start dev'ing more stuff for the Z's Aug 15 19:47:05 nick ZonE-HaKin9|AFK Aug 15 19:47:12 Opps Aug 15 19:52:47 in case anyone /away hacking on e17 in OE Aug 15 19:52:51 argh Aug 15 19:53:07 I'm still around but may not respond unless PMed... Aug 15 19:58:09 aha....ok....I'm passing int eh Xsession to use for entrance in EXTRA_OECONF....why didn't CoreDump see this..? Aug 15 20:17:09 hehe lol Aug 15 20:37:23 hey kergoth his questions were probably about bitbake/oe and using it in OSX Aug 15 20:37:44 he was trying to set it up Aug 15 20:37:59 well, i explained what the channel is for, and left it to him to decide whether his question was appropriate, given that. if it was, he shoudlve asked it Aug 15 20:38:05 heh Aug 15 20:38:08 lol Aug 15 20:38:32 he was thinking fink might have enough jam to do it Aug 15 20:38:46 * emte shrugges Aug 15 20:38:55 could be. main issue would be getting the gnu tools in place, given all teh assumptions made, i'd think Aug 15 20:38:58 anyone gotten a build to work in any BSD yet? Aug 15 20:39:10 people have played with it on occasion in the past, used to be notes about it in the wiki Aug 15 20:39:40 as long as Linux Compatibility is enabled i'd think it should be possible Aug 15 20:40:26 maybe one day i'll reinstall netbsd or try dragonfly as take it for a spin Aug 15 20:40:35 WTF! googlebot was spidering my viewmtn....that's not a public URL AFAIK Aug 15 20:41:12 i think google removed the "ploiteness" check Aug 15 20:41:26 hence the sudden doubling of thier db Aug 15 20:41:36 politeness* Aug 15 20:41:39 it's also on my cable modem connection and was absolutely killing everything Aug 15 20:41:42 damn them! Aug 15 20:42:18 you could allways trap it :P Aug 15 20:42:22 thats fun Aug 15 20:42:23 heh Aug 15 20:42:34 looping recursive links :) Aug 15 21:56:39 * france is away: Away Aug 15 22:17:28 blech Aug 15 22:17:37 bye bye bad code Aug 15 22:17:48 * emte restarts biohazard again Aug 15 22:17:55 man i feel like raster Aug 15 22:18:19 fell like the code is bad so you toss it out the window Aug 15 22:18:21 feel* Aug 15 22:18:37 lets draw a new picture ... Aug 15 22:18:52 what do i want it to do ... Aug 15 22:19:06 write itself Aug 15 22:19:11 taht would be great Aug 15 22:19:25 now realistically what do i want it to do? Aug 15 22:21:39 hahahahahahahahhahhahahahahhahaha Aug 15 22:21:41 aaaah Aug 15 22:21:43 * raster sighs Aug 15 22:21:46 :) Aug 15 22:22:53 lol Aug 15 22:23:03 i thought you mihgt like that rant Aug 15 22:23:05 might Aug 15 23:01:40 I have compiled gqview on oe. The executable gqview is getting created and i am able to run that successfully. on the GUI i am going to the directory which contains the image files (1.gif, 1.jpeg, 1.jpg). But on the Gui- Gqview does not list these files. i dont know why ? Aug 15 23:02:12 As, a result i cannot select these image files ? Aug 15 23:16:29 gqview -- I dont know what libraries are needed to list the image file format and what libraries are needed to actually load them ? Aug 15 23:17:07 Just wanted to know if any one has tried gqview with oe Aug 15 23:51:23 03justinp 07org.oe.dev * rd02f04fd... 10/packages/e17/ (7 files in 4 dirs): Aug 15 23:51:25 Use bash in entrance_session to get around "entrance quitting right after login" bug Aug 15 23:51:27 Alter entrance_wrapper to call Xinit scripts (forcefully disable 99gpe-login for now) Aug 15 23:51:29 Add a Failsafe Session file Aug 15 23:53:41 ummm...CIA-1 needs to be removed from that Excess flood list... Aug 16 00:08:04 Justin: have u worked on gqview ?. i am facing some issues ... Aug 16 00:08:16 gqview does not list image files ? Aug 16 00:48:03 morning Aug 16 00:48:29 eJumbo: I maintain gqview Aug 16 00:48:53 Xora: i am having some problems ? Aug 16 00:49:06 gqview does not list any files at all Aug 16 00:49:32 eJumbo: wierd, I use it all time here, what size screen you got? Aug 16 00:49:49 it lists directories. if i go into the directories. nothing is visible except . and .. Aug 16 00:49:57 i am using osk 5912 Aug 16 00:50:39 i dont think it is a problem with screen size Aug 16 00:51:08 no errors on the console? Aug 16 00:51:27 The menu works... no issues. but my image files are not getting listed. hence i am unable to select the my image files Aug 16 00:51:29 1 sec Aug 16 00:52:08 i get errors like : Aug 16 00:52:11 (gqview:750): GdkPixbuf-WARNING **: Cannot open pixbuf loader module file '/etc/gtk-2.0/gdk-pixbuf.loaders': No such file Aug 16 00:53:44 eJumbo: ah, youll need to install gdk-pixbuf loaders, Ill have to add them to RRECOMENDS Aug 16 00:53:49 or RDEPENDS Aug 16 00:54:38 gdk-pixbuf-loader-jpeg gdk-pixbuf-loader-gif gdk-pixbuf-loader-png probably Aug 16 00:56:22 but i dont find the update-gdk-pixbuf-loaders in the Makefile in gtk Aug 16 00:56:34 at all. i was trying to find it in the work directory Aug 16 00:57:24 Xora: can paste the latest (or updated) .bb file in the pastebin Aug 16 00:58:38 eJumbo: the latest one is the one you have Aug 16 01:05:11 morning all Aug 16 01:05:24 morning RP Aug 16 01:06:37 morning Aug 16 01:06:57 hey CoreDump|home Aug 16 01:08:18 /dev/mmcblk0p1 239.2M 36.6M 190.5M 1.6% / Aug 16 01:09:27 root on SD? Aug 16 01:09:59 CoreDump|home: yeah Aug 16 01:10:36 XorA: sorry, no l33t enough :D Aug 16 01:10:55 /dev/loop0 198337 92864 105473 47% / (image-file loop-mounted from an NFS drive) hehe Aug 16 01:11:10 * CoreDump|home hugs NFS Aug 16 01:11:10 CoreDump|home: root on SD without pivoting Aug 16 01:11:16 doh Aug 16 01:11:32 which device has support for _that_? Aug 16 01:11:38 very kool d00d ;) Aug 16 01:11:44 CoreDump|home: c7x0 Aug 16 01:11:53 hey ZonE-HaKin9 Aug 16 01:12:07 XorA: now that's sweet. Standard bootloader? Aug 16 01:12:16 hey buddy... i thnk is time to have a break from my study Aug 16 01:12:46 * CoreDump|home just got up Aug 16 01:13:14 CoreDump|home: its about 6:15pm here just about dinner time :P Aug 16 01:13:17 CoreDump|home: standard everything, one change in linux-openzaurus.bb Aug 16 01:13:19 CMDLINE_ROOT = "root=/dev/mmcblk0p1 rootfstype=ext2 " Aug 16 01:13:49 ZonE-HaKin9: 10:15am :) Aug 16 01:13:55 XorA: do u think that work also work for the 5600's & 5500's? Aug 16 01:14:06 ZonE-HaKin9: only if someone makes a 2.6 kernel Aug 16 01:14:12 XorA: very nice Aug 16 01:14:21 CoreDump|home: with new pcmcia layer, you can also make / on CF Aug 16 01:14:35 Arrr... damn it monotone compile just broke :( Aug 16 01:14:52 pivot_root of course has the advantage that you can choose during boot what to do Aug 16 01:15:17 CoreDump|home: I just wanted to see if it could be done Aug 16 01:15:26 XorA: Nice :) Aug 16 01:15:39 CoreDump|home: and also gives very quick method to switch kernels for debugging Aug 16 01:15:53 RP: its all due to your work Aug 16 01:16:27 RP: hey RP :) I've got the photos btw, will upload them to hentges.net/tmp before I go to work (~20MB) Aug 16 01:16:38 XorA: SD wasn't really mine, I just helped it along ;-) Aug 16 01:16:42 CoreDump|home: Excellent, thanks :) Aug 16 01:17:26 XorA: I fixed a nasty bug in CF handling last night. The card detect interrupt wasn't getting registered on c7x0 or anything with a scoop chip :-/ Aug 16 01:17:54 bbl got to figure why monotone won't build under fink Aug 16 01:18:12 RP: oh, didnt notice any breakage :-D Aug 16 01:18:38 XorA: hotunplug with cf memory even looks to work now ;-) Aug 16 01:18:59 It should have done after the ide changes but my bug was masking that... Aug 16 01:19:05 RP: cool Aug 16 01:19:38 I'm sure it now becomes a timing issue and will still oops, just much more unlikely ;-) Aug 16 01:24:25 now I need to buy a 4G SD card Aug 16 01:25:43 Xora: can i add these to RDEPENDS (gdk-pixbuf-loader-jpeg gdk-pixbuf-loader-gif gdk-pixbuf-loader-png) and try Aug 16 01:25:56 Xora: should i add anything else ? Aug 16 01:26:25 eJumbo: that should be enough I think Aug 16 01:27:47 i think these is one shell script/exe update-gdk-pixbuf-loaders cant find out that ?. Aug 16 01:27:59 is it required ? Aug 16 01:28:20 good morning all Aug 16 01:28:36 eJumbo: I dont know much about the gdk stuff, but if update-gdk-pixbuf-loaders is not found by the loaders then they are missing is depend somewhere Aug 16 01:30:54 hi koen Aug 16 01:31:01 hey CoreDump|home Aug 16 01:34:20 XorA: having fun at dnscon? Aug 16 01:40:11 koen: yes, lots of beer :-) Aug 16 01:40:17 koen: and talking shite about computers Aug 16 01:40:54 sounds like fun Aug 16 01:44:08 managed to show off Z with GPE a lot Aug 16 01:44:25 do they like it? Aug 16 01:44:28 yes Aug 16 01:44:35 it has high eye candy value Aug 16 01:45:06 http://dominion.kabel.utwente.nl/koen/pda/files/entrance.avi <-- even more (moving!) eye candy Aug 16 01:46:45 and Im sure I had one guy looking over my shoulder trying to work out how I was running firefox Aug 16 01:47:12 koen: /dev/mmcblk0p1 239.2M 37.6M 189.4M 17% / Aug 16 01:47:16 koen: with no pivot Aug 16 01:47:20 nice Aug 16 01:48:19 with /usr on flash? Aug 16 01:53:02 koen: no, everything is on SD Aug 16 01:53:38 and it's only 37MB? Aug 16 01:53:39 wow Aug 16 01:54:52 koen: GPE image as of yesterday Aug 16 01:55:00 koen: Im now adding all the other apps Aug 16 01:58:50 Xora: adding the loader libs does not work. gqview is not able to find them ? Aug 16 01:59:59 koen: can you help with this, Im not knowledgeable about gtk magic Aug 16 02:00:25 the loaders postinst should update the loader info Aug 16 02:00:35 try running the postinsts by hand Aug 16 02:00:58 it could be a mime issue too Aug 16 02:08:52 hi Aug 16 02:13:53 koen: did not get u. Aug 16 02:13:55 one of these days I will alias apt-get to ipkg Aug 16 02:14:23 actually, i copied the /etc/gtk-2.0/gdk-pixbuf-loaders onto the rootfs of osk and i did not get any messages on the console. Aug 16 02:14:58 the problem is gqview is not listing the files. unless it lists files i can't select the image file i want Aug 16 02:15:32 koen: whats the style for bring in gdk-pixbuf-loaders, RDEPENDS or RRECOMENDS Aug 16 02:15:39 it lists only directories and if i go inside the directory. it shows as if it is an empty directory. Aug 16 02:17:50 XorA: RRECOMMENDS I'd say Aug 16 02:31:29 ok. is there any image view (light weight) anyone knows ?. i cant make gqview work on my board Aug 16 02:43:59 what's the kernel module ipk name for v4l? Aug 16 02:44:01 oops, wrong window, sorry. Aug 16 02:54:29 03koen 07org.oe.dev * r9d72b7e4... 10/conf/distro/ (3 files): Add first cut at gpe 2.7 and make familiar + openzaurus use it Aug 16 02:54:33 03koen 07org.oe.dev * r9a4ae3ee... 10/conf/distro/preferred-gpe-versions-2.7.inc: Update preferred-gpe-versions-2.7.inc to comply with http://handhelds.org/hypermail/gpe/44/4456.html **** ENDING LOGGING AT Tue Aug 16 02:59:56 2005