**** BEGIN LOGGING AT Sat Jul 28 02:59:56 2007 Jul 28 07:08:00 how can i run angstron in qemu. I have downloaded files form the download section Jul 28 07:11:39 good morning all Jul 28 07:12:28 empty_mind: have a look at the qemu scripts here: http://svn.o-hand.com/view/poky/trunk/scripts/ Jul 28 07:50:22 anyone know where I can get internal PCMCIA interface compatible (as in pin compatible, not slot) wifi chips? Jul 28 07:52:27 good morning Jul 28 08:29:37 * Bernardo is away: Ausente por agora. Jul 28 08:33:36 xipietotec, you still awake? Jul 28 08:33:43 yep Paladine =) Jul 28 08:33:52 link me to that risc board again mate Jul 28 08:35:16 Paladine, http://www.applieddata.net/products_AGX.asp :) Jul 28 08:35:35 I think I can get it to operate at 2wh or less for what I want. Jul 28 08:36:00 :) Jul 28 08:36:01 then I just need to build about a 85wh cell array. Jul 28 08:36:25 that makes about 42 hours of continuous use =) Jul 28 08:37:21 which of course, is simply instane :D Jul 28 08:37:25 err, insane :) Jul 28 08:37:38 hehehe Jul 28 08:38:31 in fact...the heaviest part of the final product will probably be the battery. Jul 28 08:39:59 * * OE Bug 2695 has been created by jbs(AT)bandos.homelinux.com Jul 28 08:40:01 * * opie reader installs debug and dev versions by default Jul 28 08:40:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2695 Jul 28 08:42:59 * * OE Bug 2694 has been created by jbs(AT)bandos.homelinux.com Jul 28 08:43:01 * * Opie reader doesn't need unicode patch Jul 28 08:43:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2694 Jul 28 08:43:59 * * OE Bug 2696 has been created by jbs(AT)bandos.homelinux.com Jul 28 08:44:01 * * konqueror-embedded fails looking for oe_runconf Jul 28 08:44:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2696 Jul 28 08:44:30 Paladine, also, just for fun, check this out: http://www.youtube.com/watch?v=3bnL16rbmb0&mode=related&search= Jul 28 09:09:48 03koen 07org.oe.dev * r93ddc0ab... 10/ (4 files in 2 dirs): Jul 28 09:09:48 gnuradio: packaging fixes: Jul 28 09:09:48 * package all libs seperately Jul 28 09:09:48 * make the usrp python stuff have its own package Jul 28 09:09:48 * add 3.0.4 tarball release Jul 28 09:09:49 * create a .inc for common stuff Jul 28 09:09:53 03koen 07org.oe.dev * rfa64070c... 10/ (1 packages/gnuradio/gnuradio.inc): gnuradio: gnuradio-usrp also needs python-core Jul 28 09:36:43 morning all Jul 28 09:44:09 hey RP Jul 28 09:46:30 hey peeps :) Jul 28 09:46:49 I made you all famous last night ;) Jul 28 09:46:51 http://blog.paladine.org.uk/?p=26 Jul 28 09:47:15 itis dugg too :) Jul 28 09:47:35 my iPAQ has server 508 web pages in 12 hours without so much as a hint of a problem :) Jul 28 09:47:40 served* Jul 28 09:55:09 mickeyl: moin Jul 28 09:55:22 morning mickeyl Jul 28 09:57:02 morning guys Jul 28 10:03:53 how do I set the timezone in busybox? Jul 28 10:04:32 I need to change it from utc to GMT+1 Jul 28 10:08:51 mickeyl: have you got a few minutes? Jul 28 10:09:19 RP: sure Jul 28 10:09:49 mickeyl: I'm considering this svnnow problem and some of the problems in the fetcher in general Jul 28 10:10:12 RP: ah cool Jul 28 10:10:28 mickeyl: The underlying issue is we need some kind of secondary data store to store either processed data or cached data like the head svn revision Jul 28 10:10:51 mickeyl: The processed data could be the processed form of the SRC_URI variable for example Jul 28 10:11:41 These pieces of data have a determined lifetime - the processed SRC_URI lasts for as long as bitbake runs, no longer. The cached head svn revision's lifespan would be configurable Jul 28 10:12:05 Some users would want to wipe that cache manually for example Jul 28 10:12:59 The problem is how to add such a cache to bitbake. Obviously this is something bitbake needs to implement in the core but how do we do it? Should we try for some kind of generic cache that other things can possibly hook into too Jul 28 10:13:29 At present, the SRC_URI cache is an ugly global variable in the fetcher :/ Jul 28 10:14:13 hmm, i see. Jul 28 10:14:34 we have some more use cases for caching data between tasks Jul 28 10:14:47 so this may finally be the one use case that is huge enough to get us started on a generic cache Jul 28 10:14:49 When you add in threading issues it gets more tricky too Jul 28 10:15:25 i'm all for making it a generic hookable cache since I'm sure eventually we have a need for this flexibility Jul 28 10:16:03 mickeyl: What are the other use cases for data between tasks? Jul 28 10:16:41 RP: you remember previously we had some packages write stuff into staging for later packages to pick up? Jul 28 10:17:02 RP: IIRC we removed this since we found it to be very ugly, but still it was neat idea Jul 28 10:17:17 it may even be the foundation for giving information about configurations of packages Jul 28 10:17:40 think USE flags Jul 28 10:17:53 I'm not keen on that Jul 28 10:18:09 Really not keen. Builds should be deterministic Jul 28 10:18:45 i agree. but add per-package-staging and then we are deterministic again Jul 28 10:19:01 since we could even build multiple configurations at once Jul 28 10:19:08 but that's pie in the skie Jul 28 10:19:10 sky, even Jul 28 10:19:14 lets come back to the original problem Jul 28 10:19:17 * koen hears pie Jul 28 10:19:36 must be near lunch time :) Jul 28 10:19:46 The issue is looking at a libmatchbox package how do you know whether it was built with xml support or not Jul 28 10:20:17 pkgconfig, autoconf tests? Jul 28 10:20:41 koen: I know you can work it out, its just no longer obvious what a package is though Jul 28 10:20:52 but as you say, back to the original question :) Jul 28 10:20:53 cbrake: if that gstreamer changes works, could you commit that? Jul 28 10:21:14 mickeyl: You'd agree there is a case for some kind of new data store? Jul 28 10:21:20 RP: absolutely Jul 28 10:21:36 RP: which storage format are you thinking of and which way of accessing it? Jul 28 10:21:36 mickeyl: Any ideas how that data store should work? Jul 28 10:21:42 ah, heh Jul 28 10:21:47 mickeyl: I was about to ask you :) Jul 28 10:22:01 would sqlite3 make sense? Jul 28 10:22:15 it would give you basic locking Jul 28 10:22:51 hmm... it would help for the threading, however I'd rather not introduce pysqlite as a dependenc Jul 28 10:22:54 y Jul 28 10:23:10 mickeyl: henryk's patch just used pickle for a very simple state file Jul 28 10:23:24 I suspect performance would suck if we use it for much more though Jul 28 10:23:54 RP: well, pickle is certainly overkill Jul 28 10:23:57 mickeyl: Is there any db format supported by python "out the box" Jul 28 10:24:11 i'd rather vote for something like anydb Jul 28 10:24:20 since we are not going to store objects, aren't we? Jul 28 10:24:46 I suspect the fetcher code is doing that atm Jul 28 10:24:52 mickeyl: is it safe to delete *.pyo and *.pyc from a package? Jul 28 10:25:12 koen: yes for pyc at least, they should not be distributred Jul 28 10:25:24 insane should warn about them in fact Jul 28 10:25:32 ok, since some of them contain paths to my buildserver Jul 28 10:25:40 koen: yes, both can be deleted, however it increases startup time then Jul 28 10:25:55 mickeyl: iirc python recreates some of them Jul 28 10:26:05 mickeyl: Since they won't match the target system, they're pointless anyway Jul 28 10:26:12 RP: ? Jul 28 10:26:15 RP: welcome to bytecode Jul 28 10:26:17 :D Jul 28 10:26:23 (but I just overfilled jffs2, so can't do testing for a while) Jul 28 10:26:27 mickeyl: They're system independent? Jul 28 10:26:33 RP: i sincerely hope so! Jul 28 10:26:47 but you make me think about that Jul 28 10:26:56 i'm going to research athat Jul 28 10:27:04 if not, i'll remove all using distutils Jul 28 10:27:10 if we can remove them, a lot of python packages can become arch neutral Jul 28 10:27:29 mickeyl: not everything uses distutils :( Jul 28 10:27:31 agreed. basically everything that doesn't ship c extensions Jul 28 10:27:39 koen: sad enough :D Jul 28 10:27:53 koen: but most are using distutils-base Jul 28 10:27:59 e.g. gnuradio uses autotools (and even gets it right!) Jul 28 10:28:01 so we might add an OPT-in task there Jul 28 10:28:10 mickeyl: ok. I'd always assumed there was something arch specific about them. If they have build server paths in that's probably a bad sign Jul 28 10:29:01 RP: you may be right. I'll put that on my agenda to look into. back to the database. Jul 28 10:29:31 RP: anyways, python always ships with anydbm and this tries some of the installed db's. at the last resort, it uses dumbdbm Jul 28 10:29:49 this might be just enough of a database for us to use Jul 28 10:30:20 mickeyl: ok, that sounds reasonable. I guess for the fetcher, I could pickle objects in/out of it? Jul 28 10:31:11 RP: yes, if you really need to store objects, pickle them to binary and dump them there Jul 28 10:37:10 hrw|gone: ping Jul 28 10:37:42 koen, I see more gnu radio changes Jul 28 10:38:25 Crofton: I think dialtone.py needs shmfs in the kernel Jul 28 10:38:33 mickeyl: offhand do you know how the locking works with anydbm? Jul 28 10:38:54 Crofton: if you enable that and build the svn version, you should have a working gnuradion on the efika and osk Jul 28 10:39:13 ok Jul 28 10:39:36 I think you have to hack at things to get sound going on the Efika Jul 28 10:39:47 likewise has made it work Jul 28 10:39:49 RP: no. it also depends on which actual dbm is sitting behind. might take a look at dumbdbm Jul 28 10:40:18 I'll add this to my list of things to look at next week Jul 28 10:40:35 It wuold be really funny to have it run on the neo though Jul 28 10:43:02 * koen tries mount devshm /dev/shm/ -t tmpfs Jul 28 10:44:47 Crofton: I hear a dialtone out of the neo's speaker :) Jul 28 10:44:59 * * OE Bug 2694 has been RESOLVED (FIXED) by Jul 28 10:45:01 * *  Opie reader doesn't need unicode patch Jul 28 10:45:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2694 Jul 28 10:45:29 mickeyl: None of the formats implement any kind of locking unless we open/close the file for every transaction :/ Jul 28 10:45:58 RP: sigh nasty. what do you think about pysqlite3 then? Jul 28 10:46:15 mickeyl: It starts to look tempting Jul 28 10:46:18 cool Jul 28 10:46:34 mickeyl: It would save use having to worry about locking Jul 28 10:47:31 I'll mention this on the gnu radio list. Jul 28 10:47:37 RP: hmm, in that case I'd say yes to this additional dependency Jul 28 10:47:52 There was alot of misunderstanding how openmoko related to SDR Jul 28 10:48:34 morning ! Jul 28 10:49:11 Crofton: ipkg install gnuradio-usrp to get the usrp stuff Jul 28 10:50:28 http://lists.gnu.org/archive/html/discuss-gnuradio/2007-07/msg00108.html Jul 28 10:50:30 mickeyl: do the moko2 libs use proper libtool versionsing? Jul 28 10:50:53 Crofton: I read about that yesterday :) Jul 28 10:51:09 console image for simpad, 2.4 kernel. has /sbin/modprobe linked to /sbin/modprobe.26 - is this o.k. ? Jul 28 10:51:17 koen: dunno offhand. Jul 28 10:51:32 koen: they will use 2 in the name anyways Jul 28 10:51:33 mickeyl: if so, they are parallel installable and buildable Jul 28 10:51:48 I need a few lackeys to help me with all the things I want to do Jul 28 10:52:00 graduate students are so unreliable Jul 28 11:03:18 Crofton: the cpu is 75% idle when running dialtone (200MHz arm920t core) Jul 28 11:07:20 not to bad Jul 28 11:07:31 no dsp instructions on teh T right? Jul 28 11:07:47 and no signal processing assembly for ARM in gnu radio anyway Jul 28 11:08:33 how comes your 920T is only 200MHz? Jul 28 11:08:37 mineis 400MHz Jul 28 11:09:24 Paladine: because they come in different flavours Jul 28 11:09:55 ahh Jul 28 11:10:11 dude lighttpd is running so sweet on this iPAQ Jul 28 11:10:44 546 pages served in about 14 hours Jul 28 11:11:39 I wonder what the record uptime is on an iPAQ :) Jul 28 11:12:37 Crofton: dialtone takes 5-12% cpu on a 624MHz pxa270 Jul 28 11:13:04 grr, I need to go drink beer in a cow field today Jul 28 11:13:21 if I stay home and do this, my friends will call me a nerd Jul 28 11:13:22 beats drinking cows in a beer field Jul 28 11:13:29 yes Jul 28 11:13:38 funny to have multiple devices emit a dialtone in a small room :) Jul 28 11:14:56 Iwonder how well thiswill handle running my blog with the document root mounted over nfs and using mysql-client to connect to a remote database Jul 28 11:21:57 koen, I owe you a beer for the gnuradio work Jul 28 11:32:07 mickeyl: One further dilemma - do we have a persistent cache per machine or a more global one? :/ Jul 28 11:33:09 Crofton: I owe you a beer for the initial recipe :) Jul 28 11:33:25 RP: per machine needs to be enough. global opens a can of worms Jul 28 11:34:04 mickeyl: By global I mean shared between machines in a given build, not system wide Jul 28 11:34:28 oh Jul 28 11:34:28 mickeyl: but yes, I think it will have to make do with per machine... Jul 28 11:34:42 i was afraid of a system wide Jul 28 11:35:09 mickeyl: The thought of system wide makes me very scared :} Jul 28 11:35:45 hehe. but even the build wide is kind of frighting giving the concurrency issues, so i'd vote for a per-machine one as a start Jul 28 11:36:52 mickeyl: Yes, we can always change it later I guess Jul 28 11:39:15 03koen 07org.oe.dev * r65bc8067... 10/ (1 packages/xorg-xserver/xserver-kdrive_1.2.0.bb): xserver-kdrive 1.2.0: bump PR to get Xw100 rebuilt against a more recent xrandr Jul 28 11:39:20 03polyonymous 07org.oe.dev * rceaf972d... 10/ (4 files in 3 dirs): opie-reader: unicode patch went upstream. Close #2694 Jul 28 11:39:36 koen, | checking for jack >= 0.8... gr-audio-jack requires package jack, not found. Jul 28 11:39:36 | configure: error: Component gr-audio-jack has errors, stopping. Jul 28 11:39:57 hmmm Jul 28 11:40:03 jack is in DEPENDS... Jul 28 11:50:43 does jack use float? Jul 28 11:51:07 no idea, all I know it that it is a low-latency soundserver thingy Jul 28 11:51:28 I looked once and it liked floats, I suppose I will have to check again Jul 28 11:51:53 yeah jack is in my DEPEnDS Jul 28 11:52:14 but I don not think it is built? Jul 28 11:55:01 Are you using newer autoconf? Jul 28 11:55:14 2.61 Jul 28 11:55:59 I'm still confused why I do not build jack Jul 28 12:01:10 I am also building for Efika Jul 28 12:01:59 I will be gone soon, so maybe time to nuke tmp and start over Jul 28 12:04:02 Crofton: efika build is queued after the i586 build Jul 28 12:04:08 should be finished today Jul 28 12:04:11 heh Jul 28 12:04:18 I'll be back later Sunday Jul 28 12:04:31 sadly our subversion server dies, so we need to rebuild it Jul 28 12:04:44 and the hw purchased was an AMD with integrated video Jul 28 12:04:49 the video does not work Jul 28 12:05:01 I am trying to stay out of that can of worms Jul 28 12:13:42 03koen 07org.oe.dev * r89643518... 10/ (6 files in 3 dirs): gtk 2.10.14: add some patches from poky to have some widgets behave better on high-dpi screens Jul 28 12:19:18 RP, mickeyl: we'll also need to think about how to handle switching from SRCDATE to revision in PV for svn packages while maintaining an upgrade path Jul 28 12:28:59 I bought a Ps@ type KVM switch, but I have no instructions on hwo to switch PC's .... Jul 28 12:35:51 mickey|NeoMeetin, be sure to erport GNU radio runs on the neo :) Jul 28 12:36:50 hi Jul 28 12:37:03 he crofton ;) Jul 28 12:37:04 hey woglinde Jul 28 12:37:08 je koen Jul 28 12:37:10 ups he Jul 28 12:37:49 hm speaks some against that I check in STAGING_LIBDIR_NATVIE in bitbake.conf? Jul 28 12:37:59 * * OE Bug 2696 has been RESOLVED (FIXED) by Jul 28 12:38:01 * *  konqueror-embedded fails looking for oe_runconf Jul 28 12:38:01 hi all Jul 28 12:38:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2696 Jul 28 12:38:20 woglinde: why would that be needed? Jul 28 12:38:24 hey steliosk Jul 28 12:38:26 hi steliosk Jul 28 12:38:31 koen for java jar Jul 28 12:38:32 koen: Have any ideas on how to do that? Jul 28 12:38:42 which are arch idenpened Jul 28 12:38:50 koen : is this an libintl issue that needs to be fixed in the package or something in angstrom definitions ? -> http://bugs.openembedded.org/show_bug.cgi?id=2348 Jul 28 12:38:50 so we need no extra native package Jul 28 12:38:57 woglinde: just use STAGING_LIBDIR for that Jul 28 12:39:07 koen woglinde : hi Jul 28 12:39:27 RP: if we go from 0.0+svn to 0.0+svnr is should be ok Jul 28 12:40:00 koen: That seems reasonable Jul 28 12:40:31 steliosk: I'm with comment #1 from leon Jul 28 12:40:34 koen: git will be a bigger problem Jul 28 12:41:10 RP: I think we should just stick to SRCDATE with git, that gives the least problems at install time Jul 28 12:41:37 today i found sometime to go over bugtrucker and fix a couple of things :) Jul 28 12:42:01 it's nice to be on vacations :) Jul 28 12:42:09 koen: It makes no sense with regard to the scm though :/ Jul 28 12:42:42 RP: git tags aren't monotonically increasing and revisions can't be tracked down to a date Jul 28 12:42:56 koen hm the problem is with cacao Jul 28 12:43:11 cacao needs for crosscompiling cacaoh-native Jul 28 12:43:15 koen: I know. For gitnow, we'll just have to use a locally generated "localrevision" number Jul 28 12:43:18 RP: can we set a BASE_SRCDATE that changes only if the fetcher finds a change? Jul 28 12:43:35 or a localrevision Jul 28 12:43:37 that needs cldc.jar in NATVIVE STAGING_LIBDIR Jul 28 12:43:53 koen: I guess distro maintainers will just have to ban such packages from their distro feeds i.e. they're for development only Jul 28 12:44:07 and than cacao for arm needs that .jar too Jul 28 12:44:09 woglinde: 'inherit native' sets up STAGING_LIBDIR correctly Jul 28 12:44:34 yes but midpath can provide this for both packages Jul 28 12:44:38 koen: I will create something that only rebuilds if the head changes Jul 28 12:44:47 and midpath isnt arch depend Jul 28 12:44:48 koen: Storing the data to do that is why we need this data store Jul 28 12:45:03 RP: you and I and know how to ban packages, but how many users will shoot themselves in the foot? Jul 28 12:45:58 koen: I can't see a way to avoid it Jul 28 12:46:06 woglinde: just build install cldc.jar twice, no need to complicate OE to save a few 100kB of diskspace Jul 28 12:46:31 RP: setting SRCDATE=now is a manual action, so they deserve what they get Jul 28 12:46:48 koen okay Jul 28 12:47:03 koen: For git packages we can just force people to have to do that Jul 28 12:47:49 koen: We should lock down core tools like mtdutils and xcalibrate anyway Jul 28 12:48:09 RP: OE has xcalibrate and friend locked down already Jul 28 12:48:24 koen: right, its just mtdutils that isn't then Jul 28 12:48:32 RP: atmel has a sucky firewall, so they sent me some patches for the X stuff Jul 28 12:48:51 RP: afaik only mtd-utils, yes Jul 28 12:48:58 They're not the only ones :/ Jul 28 13:05:06 koen: does angstrom glibc-intermediate build for ixp4xxbe for you? Jul 28 13:06:10 rwhitby: yes Jul 28 13:07:25 Hmm. I get "WARNING: You should not compile GNU libc without versioning. Not using versioning will introduce incompatibilities so that old binaries will not run anymore. For versioning you need recent binutils (binutils-2.8.1.0.23 or newer)." Jul 28 13:07:33 (this is building on Debian Etch) Jul 28 13:07:49 armv5teb-angstrom-linux-gnueabi/glibc-intermediate-2.5-r6 Jul 28 13:08:18 a build from scratch 3 days ago didn't have any problems Jul 28 13:08:40 what host OS binutils version? Jul 28 13:09:09 GNU assembler (GNU Binutils for Debian) 2.17.50.20070426 Jul 28 13:09:37 2.17-3 here. Jul 28 13:10:00 (probably the same) Jul 28 13:11:48 very weird. This is a freshly installed Debian Etch system (the new nslu2-linux autobuilder), with no local packages, nothing other than what's in debian etch, and nothing local.conf for a MACHINE="ixp4xxbe" DISTRO="angstrom-2007.1" angstrom-console-image build Jul 28 13:12:41 freshly updated monotone database, complete rebuild from scratch after removing tmp Jul 28 13:18:17 koen: do I need to have 'gas' installed? Jul 28 13:18:30 (this machine currently doesn't) Jul 28 13:19:18 nm - it's installed as 'as' of course, not 'gas' Jul 28 13:27:31 koen: what MACHINE do you use? Jul 28 13:28:51 When I run the configure test by hand, I get /home/slug/angstrom/tmp/cross/lib/gcc/armeb-angstrom-linux-gnueabi/4.1.2/../../../../armeb-angstrom-linux-gnueabi/bin/ld: cannot represent machine `arm:ixp4xxbe' Jul 28 13:30:16 Hmm - I wonder if it's related to me setting MACHINE='ixp4xxbe' on the make command line, instead of in the auto.conf ... Jul 28 13:51:53 rwhitby: binutils can play games when you do that. In theory OE works around that but people keep saying its not working Jul 28 13:52:11 It does work on in Poky and I wish I knew what the difference was Jul 28 13:55:22 RP: do we unset MACHINE in every binutils recipe? or where do we have this logic again? Jul 28 14:00:06 zecke: btw, your update script hangs at mtn automate Jul 28 14:00:24 koen: that is bad :) Jul 28 14:04:05 thats a subjective opinion ;-) Jul 28 14:10:04 zecke: In base.bbclass we mark it as unexport Jul 28 14:10:46 zecke: Did you see my conversation with mickeyl earlier? Jul 28 14:10:51 no Jul 28 14:11:04 * zecke joined this channel 40 minutes ago Jul 28 14:12:28 how do i enable rm_work with bitbake 1.8.6, adding it to conf/local.conf doesn't seem to work like in 1.6 Jul 28 14:12:48 Shoragan: it should work the same Jul 28 14:13:33 hmm, but after building i.e. bootstrap-image, all sources are still left in tmp/work Jul 28 14:13:45 and i don't see it running do_rm_work Jul 28 14:14:20 zecke: http://www.rpsys.net/openzaurus/temp/bb_persist.txt Jul 28 14:16:01 RP: I'm obsessed with WebKit ATM :) Jul 28 14:16:22 zecke: you have no objection to needing sqlite though? Jul 28 14:17:05 RP: I would like to have some sort of 'prerun' for OE. It validates the configs and then tries to select 'variants' of all packages according to the configuration Jul 28 14:17:23 RP: no objections. I think berkeley db support is installed by default but sqlite (3) sounds like a sane option Jul 28 14:18:01 zecke: The attraction of sqlite is its locking. We can do what we need in 58 lines of code (including comments) Jul 28 14:18:20 zecke: The idea of a prerun scares me a lot :/ Jul 28 14:18:39 RP: sqlite: ack Jul 28 14:18:59 RP: prerun the validation part is not scary at all and could/should be part of BitBake :) Jul 28 14:19:16 zecke: http://www.rpsys.net/openzaurus/temp/persist_data.py is what I have atm Jul 28 14:19:26 Shoragan: -c rmall Jul 28 14:19:39 Shoragan: provided you have an up to date OE checkout Jul 28 14:19:43 zecke: Validating the configuration is a job for OE, not for bitbake Jul 28 14:20:01 koen, i'm using that now, and it works Jul 28 14:20:05 RP: well, parsing 5000 files just because a spelling error in DISTRO sucks :) Jul 28 14:20:29 but rm_work should work, too, right? Jul 28 14:20:42 zecke: There is a cunning way to fix this. An event at the end of parsing the base configuration for sanity.bbclass Jul 28 14:20:43 zecke: I'll do a buid of webkit now to see what the error is Jul 28 14:20:46 RP: prerun: I like macport's variants a lot. But as each package can influence each other you would to have such variants in stable way Jul 28 14:21:26 RP: an event == builtin in bitbake :) Jul 28 14:21:36 koen: you might end up with an error on moc (eek) :) Jul 28 14:22:19 zecke: No, triggers the checks in OE. bitbake itself doesn't care about the DISTRO variable and shouldn't ever do so Jul 28 14:23:01 If I could find a way to add the event without breaking backwards compatibility, I would... Jul 28 14:23:07 I need to try again Jul 28 14:23:31 RP: how is that breaking? eventHandler in OE break? Jul 28 14:23:47 eventHandlers of the bbclasses break? Jul 28 14:24:03 zecke: We need to trigger the standard sanity checks only for certain versions of bb Jul 28 14:24:17 and use the new hook for others Jul 28 14:24:24 zecke: Can't exec "/data/build/koen/OE/build/tmp/angstrom/staging/i686-linux/bin/moc": No such file or directory at /data/build/koen/OE/build/tmp/angstrom/work/armv4t-angstrom-linux-gnueabi/webkit-0.0+svn20070728-r1/WebCore/ksvg2/scripts/make_names.pl line 91. Jul 28 14:24:40 Last time I tried it, it went badly wrong. Like I said, I should try again Jul 28 14:24:52 koen: DEPENDS += "uicmoc4-native" Jul 28 14:25:17 zecke: webkit_svn.bb:DEPENDS = "flex-native gperf-native gperf perl-native curl icu uicmoc4-native qmake2-native libxml2 sqlite3 cairo libxslt libidn gnutls gtk+" Jul 28 14:25:34 koen: then QMAKE_MOC of the makespec is broken :) Jul 28 14:25:43 koen: it should point to moc4? Jul 28 14:25:55 koen: ls /data/build/koen/OE/build/tmp/angstrom/staging/i686-linux/bin/moc* Jul 28 14:26:13 /data/build/koen/OE/build/tmp/angstrom/staging/i686-linux/bin/moc4 Jul 28 14:26:40 * zecke searches for a OE tree but is too lazy to open the vpn Jul 28 14:27:27 koen: what version of qmake* do we inherit? Jul 28 14:28:17 * Bernardo is back. Jul 28 14:28:19 re Jul 28 14:28:27 hi bernardo Jul 28 14:29:41 zecke: 'inherit qmake.bbclass' Jul 28 14:30:11 koen: cheat http://www.openembedded.org/repo/org.openembedded.dev/classes/qt4x11.bbclass and set QMAKE_MOC Jul 28 14:31:37 zecke: http://rafb.net/p/oQgPc186.html Jul 28 14:32:15 koen: moc is enough Jul 28 14:35:46 zecke: http://rafb.net/p/jAOAS288.html seems to do the trick Jul 28 14:35:51 I'd think your build process uses wrong mkspecs? Jul 28 14:36:13 ah, didn't you have it before? Jul 28 14:36:28 koen: I would move that out of this function below below the qmake inherit Jul 28 14:36:34 polyonymous: as I said cheating :) Jul 28 14:36:41 koen, I think this qt4 machine should be well straightened up... Jul 28 14:37:10 polyonymous: qt4x11.bbclass should be split into two. qt4.bbclass setting all qt4 vars and qt4x11.bbclass inheriting it and setting the x11 libs and adds the depends on qt4-x11 Jul 28 14:37:22 yes. Jul 28 14:37:49 http://rafb.net/p/GnTxSw48.html Jul 28 14:37:49 But why do you need thi export, anyway? qt4x11 does exporting? Or do you have to avoid inheriting it in webkit_svn? Jul 28 14:38:05 ah, got it. Jul 28 14:38:14 polyonymous: we don't use qt4-x11 :) Jul 28 14:38:23 Yeah, that's what I figured. Jul 28 14:38:32 What do we use then? :) Jul 28 14:38:33 koen: remove the export :) Jul 28 14:38:43 polyonymous: qmake2 for building and Gtk for the GUI Jul 28 14:38:49 ah Jul 28 14:38:50 I see. Jul 28 14:38:57 polyonymous: it started to use the moc as preprocessor (for the windows port) Jul 28 14:39:02 polyonymous: http://trac.webkit.org/projects/webkit/wiki/BuildingGdk Jul 28 14:39:03 weird setup. Jul 28 14:39:15 polyonymous: for some reasons windows doesn't ship with gcc installed Jul 28 14:39:21 but well, yes, we need to splitup qt4xx then. Jul 28 14:39:28 zecke, that sucks. Jul 28 14:39:54 polyonymous: well, the Qt and Gdk port share the same buildsystem. So apple folks only have to fix one buildsystem when they do changes Jul 28 14:40:27 zecke, understandable, but surely using qt buildsystem for gdk does sound strange. Jul 28 14:40:40 but then again, it's their choice. Jul 28 14:41:03 polyonymous: it is my choice :) Jul 28 14:41:11 zecke, even worse :) Jul 28 14:41:21 polyonymous: and if you get fixes for free, you should take them Jul 28 14:41:54 zecke, I'd disagree with it as a general rule. Jul 28 14:42:43 polyonymous: sure you can spend your time on keeping your autoconf system compiling, or you can have apple fix compile issues :) Jul 28 14:43:39 zecke, I said as a general rule - here it applies, I'd think. And btw, I could never understand these talks about incomprehensibility of autotools... Jul 28 14:45:38 anyway, gotta go eat something sweet :) Jul 28 14:48:26 hi psokolovsky Jul 28 14:49:31 Good Morning Jul 28 14:49:41 hi mwester Jul 28 14:58:00 zecke: ../../JavaScriptCore/wtf/TCSystemAlloc.cpp../../../../JavaScriptCore/wtf/TCSpinLock.h: In function 'void* TCMalloc_SystemAlloc(size_t, size_t)':../../../../JavaScriptCore/wtf/TCSpinLock.h:98: error: 'asm' operand requires impossible reloadmake[1]: *** [TCSystemAlloc.o] Error 1 Jul 28 14:59:04 koen: interesting Jul 28 15:00:42 koen: #if (PLATFORM(X86) || PLATFORM(PPC)) && COMPILER(GCC) Jul 28 15:00:59 yep, powerpc :) Jul 28 15:01:11 make[1]: Leaving directory `/data/build/koen/OE/build/tmp/angstrom/work/ppc603e-angstrom-linux/webkit-0.0+svn20070728-r1/WebKitBuilds/Debug/JavaScriptCore/kjs' Jul 28 15:01:21 koen: which ppc are you doing this on? Jul 28 15:01:29 MACHINE=efika Jul 28 15:02:04 ~seen pb_ Jul 28 15:02:07 pb_ was last seen on IRC in channel #handhelds.org, 107d 21h 26m 38s ago, saying: 'bd2: hi'. Jul 28 15:04:03 hmm, I forgot everything about inline assembly Jul 28 15:04:12 change the r to g, but I don't know why :) Jul 28 15:04:49 me wonders why we need so many different types of cpu's Jul 28 15:04:59 * steliosk wonders why we need so many different types of cpu's Jul 28 15:06:45 steliosk: NIH Jul 28 15:07:54 heh Jul 28 15:07:59 tmbinc: koen has a ppc assembly problem, we would be glad if you could help :) Jul 28 15:08:01 zecke *g* Jul 28 15:08:05 koen : true Jul 28 15:08:47 zecke : what kind of assembly problem ? Jul 28 15:09:51 steliosk: http://rafb.net/p/9joU8Y16.html Jul 28 15:10:18 steliosk: koen claim's a impossible reload Jul 28 15:10:42 hmmm Jul 28 15:10:44 steliosk: I wonder if one should put a register to the clobber list? but which one? what is it trying to reload anyway? Jul 28 15:11:58 http://www.cs.uaf.edu/2005/fall/cs301/support/ppc/index.html Jul 28 15:13:15 03polyonymous 07org.oe.dev * r577216ce... 10/ (1 conf/distro/include/preferred-opie-versions-1.2.3.inc): preferred-opie-versions-1.2.3: bump preferred konqueror version. Closes #2696 Jul 28 15:16:01 re Jul 28 15:16:27 can I have a bb file include another bb file? Jul 28 15:16:51 I want to try building an image based off, say angstrom-console-image.bb Jul 28 15:17:22 but making gnuradio.bb that includes A-I-C.bb and adds ANGSTROM_EXTRA_INSTALL to include packages Jul 28 15:19:01 sure Jul 28 15:19:19 good :) Jul 28 15:19:35 I have several sets of images I want to work with Jul 28 15:19:57 and dorking with A_E_I setting in local.conf annoys me :) Jul 28 15:20:00 zecke : its trying to store the contents of r1 to the location in r0 Jul 28 15:21:13 zecke : btw efika is powerpc not ppc Jul 28 15:21:28 Crofton: you could also make a task for all the sdr stuff Jul 28 15:21:28 where is the difference? Jul 28 15:21:45 zecke : for kernel there is Jul 28 15:21:49 zecke: between 'p' and 'r' Jul 28 15:22:00 p and p even Jul 28 15:22:25 I could Jul 28 15:22:37 but I like the idea of extending existing angstrom images Jul 28 15:23:01 Crofton|laptop: and I meant extending them with a task-sdr :) Jul 28 15:23:23 We'll have to keep thinking about this Jul 28 15:23:46 we need a general way for people to extand basic distro images and create specializing Jul 28 15:23:55 um Jul 28 15:23:56 the 2 first instruction try to resync the caches with the commands isync and eieio Jul 28 15:23:59 Go Levi! Jul 28 15:24:35 god I hope he isn't doping Jul 28 15:24:54 they all are doping Jul 28 15:25:28 Crofton|laptop: gnuradio for efika is in the feeds Jul 28 15:25:41 mine is at school Jul 28 15:25:54 as soon as tour finishes, I head for cow field Jul 28 15:26:02 moooooo Jul 28 15:26:08 hm sdl is another candidate not to run autoheader Jul 28 15:26:09 I'll get some pictures Jul 28 15:26:36 I'm with Boonen, dopers should be shot Jul 28 15:26:56 of course a doper is about to break the baseball home run record, and no one seems to care Jul 28 15:28:03 zecke : are you using -fPIC as flags to compile ? Jul 28 15:29:37 steliosk: ask koen :) Jul 28 15:29:58 koen : are you using -fPIC as flags to compile ? Jul 28 15:30:16 koen : if yes try to disable them Jul 28 15:30:23 I have no idea Jul 28 15:30:55 Dirk demol in team car is screaming at Levi Jul 28 15:38:49 Crofton|laptop: No-one cares about the baseball home run record *because* everyone knows it's not deserved by Bonds. Jul 28 15:39:14 :) Jul 28 15:39:16 yeah Jul 28 15:39:47 it is annoying to see sports that are at least trying to elimnate doping get bad press Jul 28 15:39:53 And we have a star football player who runs illegal dog-fights. Jul 28 15:40:06 he Jul 28 15:40:15 I am at Virginia Tech Jul 28 15:40:18 And basketball officials who are gambling on the games they officiate. Jul 28 15:40:41 I think we should all watch some other country's sports, our own are in shambles! Jul 28 15:40:43 yeah, but everyone hates the refs :) Jul 28 15:40:57 well, cycling is a very Euro thing :) Jul 28 15:41:36 True. Who was just sent home for drug use the other day? Jul 28 15:41:43 Chicken Jul 28 15:41:46 well Jul 28 15:41:57 suspected drug use, he was sent home for lying Jul 28 15:42:06 Crofton|laptop: http://www.angstrom-distribution.org/repo/?action=details&pnm=gnuradio Jul 28 15:42:54 very cool Jul 28 15:45:51 arrg seconds are counting Jul 28 15:46:49 03koen 07org.oe.dev * rdee4672e... 10/ (1 packages/jack/jack_0.103.0.bb): jack: add -ldl, needed for some archs Jul 28 15:50:21 RP: any news on the rebuild-dependants-on-change bitbake feature? Jul 28 15:52:04 koen: not yet Jul 28 16:00:56 There is some Levi fan who waves a giant american flag Jul 28 16:01:05 and wears enormous antlers Jul 28 16:01:24 and with that, it is time to get going :) Jul 28 16:10:54 morning Jul 28 16:12:32 hmm, we go from cycling to strong man competition Jul 28 16:12:41 I can smel the steroids over the air .... Jul 28 16:13:20 hi chouimat Jul 28 16:13:21 hm Jul 28 16:13:23 sdl Jul 28 16:13:36 I updated it a little bit locally Jul 28 16:14:01 zecke: I made the mistake of looking at the bitbake parser code. Do we really want to support inheriting conf files from things like svn? Jul 28 16:15:06 RP: ah, no! Jul 28 16:15:11 zecke: The new data store needs to know what CACHE is. Picture what happens if CACHE is set in some file we have to fetch first and the fetcher code needs the data store Jul 28 16:15:36 zecke: What about remote .bb files? Jul 28 16:15:47 RP: no! Jul 28 16:16:22 zecke: Currently is appears you can do BBFILES = "http://somewhere/somefile.bb" Jul 28 16:16:29 right Jul 28 16:16:53 I've very against conf files, not sure about bb files... Jul 28 16:17:25 Probably a big can of worms... Jul 28 16:17:41 * koen mentions corporate proxies Jul 28 16:17:58 * chouimat is scared Jul 28 16:27:27 hm when a .bb has requir bla.inc then the do_stage() {} of the .inc dont get in the .bb? Jul 28 16:27:31 require Jul 28 16:27:49 woglinde: It should do Jul 28 16:29:43 ah Jul 28 16:29:46 hm Jul 28 16:29:46 * RP actually has no idea how the current fetcher works at all. Its making some really strange assumptions Jul 28 16:30:11 my require should came after inherit Jul 28 16:33:48 okay yes the order is important Jul 28 16:34:05 I should notice this somehow in my head and never forget Jul 28 16:36:46 hi psycoman Jul 28 16:36:50 hi woglinde Jul 28 16:39:39 desk? Jul 28 16:39:44 ah desktop Jul 28 16:39:45 okay Jul 28 16:41:51 03koen 07org.oe.dev * r9d2379ad... 10/ (1 packages/bluez/bluez-libs_3.13.bb): bluez-libs: update to 3.13 Jul 28 16:41:55 03koen 07org.oe.dev * r484f4908... 10/ (1 packages/bluez/bluez-utils_3.13.bb): bluez-utils: update to 3.13 Jul 28 16:41:59 03koen 07org.oe.dev * r08765357... 10/ (1 packages/bluez/bluez-hcidump_1.38.bb): bluez-hcidump: update to 1.38 Jul 28 16:52:07 03woglinde2 07org.oe.dev * rc9de0582... 10/ (4 files in 2 dirs): Jul 28 16:52:07 -introduce new recipes for classpath-minimal Jul 28 16:52:07 -classpath-minimal now at 0.95 Jul 28 16:52:07 -seperate install location so no confusion with normal-classpath Jul 28 16:52:07 -chosse a sane glibj.zip-location, but we have no java policy yet Jul 28 16:52:07 -provide no packages, because it is only for building java virtual machines Jul 28 16:52:09 -second step to get MIDP2 support in OE Jul 28 16:58:04 yeap desk, desktop xD Jul 28 17:16:34 03rpurdie * r847 10bitbake/lib/bb/ (msg.py persist_data.py): Add persistent data store module Jul 28 17:22:56 03rpurdie * r848 10bitbake/lib/bb/ (cooker.py fetch/__init__.py): Convert fetcher to use persist_data to remove global urldata variable and cleanup Jul 28 18:02:06 * koen curses at AC_CHECK_FILE Jul 28 18:30:57 * ljp is sick of working on n800 Jul 28 18:31:42 don't you just love how nokia uses binary stuff for just about everything? Jul 28 18:32:22 i hate it Jul 28 18:32:41 cant find proc entry for battery level. Jul 28 18:33:02 there isn't any Jul 28 18:33:07 exactly Jul 28 18:33:17 remember those exploding nokia batteries in the news a few years ago? Jul 28 18:33:23 i dont understand why vendors dont just use defauilt stuff like htat Jul 28 18:33:33 03koen 07org.oe.dev * r505a28e9... 10/ (1 site/common): site: we all have /dev/random, so put it in common Jul 28 18:33:37 03koen 07org.oe.dev * rb3780b3b... 10/ (1 packages/gnome/gnome-keyring_0.6.0.bb): gnome-keyring: update to 0.6.0 Jul 28 18:33:44 03koen 07org.oe.dev * rf4ee8a63... 10/ (1 packages/gnome/gnome-desktop_2.18.3.bb): gnome-desktop: add more DEPENDS Jul 28 18:33:55 cant seem to get audio working Jul 28 18:35:27 i think i am going to have to hack a dbus-send to get battery level Jul 28 18:36:28 i think i can get it that way Jul 28 18:36:42 I should poke rishi about the opensource gstreamer dsp codecs Jul 28 18:36:58 that should solve a lot audio problems on the nokia tablets Jul 28 18:39:14 is there an easy way to install things into cross/ like dbus, bluetooth and friends, so I can deploy a toolchain? Jul 28 18:40:53 ljp: finish packaged staging, and there will be Jul 28 18:51:53 Crofton : Well got Xilinx ise 8,2 running on amd64 Ubuntu :) Jul 28 18:52:25 * steliosk is happy that no windows machine is need it Jul 28 18:59:21 hhm uclibc dont compiles osb-jscore nptl is need Jul 28 18:59:27 ed Jul 28 19:05:05 03koen 07org.oe.dev * re444b83e... 10/ (5 files in 3 dirs): gnome-vfs: update to 2.18.1 Jul 28 19:05:10 03koen 07org.oe.dev * r0c329f0e... 10/ (1 classes/pkgconfig.bbclass): pkgconfig.bbclass: more munging Jul 28 19:06:23 03koen 07org.oe.dev * r0dec9369... 10/ (1 packages/gnuradio/gnuradio.inc): gnuradio.inc: fix quoting Jul 28 19:06:31 03koen 07org.oe.dev * rd513468c... 10/ (1 packages/gnome/gnome-menus_2.16.1.bb): gnome-menus 2.16.1: fix packaging Jul 28 19:30:27 RP: what do you think of http://rafb.net/p/kutPVw23.html ? Jul 28 19:39:02 03koen 07org.oe.dev * rca3aead4... 10/ (1 packages/gnome/libwnck_2.13.5.bb): libwnck: update to 2.13.5 Jul 28 19:39:07 03koen 07org.oe.dev * r07ab3835... 10/ (1 packages/gnome/gnome-panel_2.18.1.bb): gnome-panel: add 2.18.1 Jul 28 20:12:46 03koen 07org.oe.dev * rca25b3ea... 10/ (1 packages/gnome/gnome-power-manager_2.18.3.bb): gnome-power-manager: add 2.18.3 Jul 28 20:14:36 http://scap.linuxtogo.org/files/b91821e1f85831ddb8b7e42c5da1d478.png :) Jul 28 20:15:44 koen: hmm, pretty ... short battery life:-P Jul 28 20:15:45 hi slapin Jul 28 20:16:08 mmp: wifi + backlight at 100% Jul 28 20:16:27 ahaa, that's another story:) but it's still nice -- what is the memory footprint of GPM? Jul 28 20:16:47 mmp: I have no idea Jul 28 20:17:06 but it drags in a few megs of gnome due to the retarted design of gnome panel applets Jul 28 20:17:36 :-/ Jul 28 20:20:23 Since the main happiness of EABI is mixing hard and soft float, can anyone tell me how to force hardfloat instead of softfloat for a reciepe? Jul 28 20:20:47 why would you want hardfloat? Jul 28 20:20:59 have you hardware for it? Jul 28 20:21:50 Is there anyway I can convince you that I know what I'm doing and that I need hardfloat? Because if I explain it you will all refuse to tell me how to do it... Jul 28 20:21:53 :-) Jul 28 20:22:09 so you don't need it Jul 28 20:22:32 No, the usual response is "That's a bad idea" Jul 28 20:22:36 zero_chaos grep for softfloat and adapt it Jul 28 20:23:01 thats what I would do Jul 28 20:23:05 the only reason to have hardfloat is when you have an FPU Jul 28 20:23:57 I have a kernel module that uses float and double math that causes missing symbols when compiled with softfloat... and since no one can seem to help me resolve the missing symbols "because floating point is not allowed in the kernel" I figured I would use hard float and let it do that kernel trap thing so the damn code at least works Jul 28 20:24:32 the kernel doesn't do floating point Jul 28 20:24:44 you will always keep having unresolved symbols Jul 28 20:24:47 yeah yeah, I noticed Jul 28 20:24:56 so how can I use hardfloat? Jul 28 20:25:06 ~lart c-developer for lazyness and using intp_t u_intp_t Jul 28 20:25:06 * ibot throws a AN/M-8 smoke grenade at c-developer for lazyness and using intp_t u_intp_t Jul 28 20:25:16 [22:24] koen: the kernel doesn't do floating point Jul 28 20:25:17 [22:24] koen: you will always keep having unresolved symbols Jul 28 20:25:17 okay Jul 28 20:25:18 tv now Jul 28 20:25:34 till later Jul 28 20:25:54 koen: okay, but despite poor programming practice I know it works, so.... is there a way to resolve this or not? Jul 28 20:26:07 Zero_Chaos: the kernel can't trap itself Jul 28 20:26:20 do even if it compiles it will never work Jul 28 20:26:29 s/do/so/ Jul 28 20:26:50 which is why floating point isn't allowed inside the kernel Jul 28 20:27:11 "will never work on a system without FPU", that is Jul 28 20:27:26 koen: okay... I know it works when there is an FPU (i.e. my laptop), but you are saying there is no way at all with no FPU? Jul 28 20:27:33 koen: dirty or not Jul 28 20:27:36 right Jul 28 20:27:47 well that just sucks Jul 28 20:27:55 and use clue-by-4 on the authors head Jul 28 20:28:45 floating point isn't allowed in the kernel, not matter how 1337 that guy is Jul 28 20:29:23 oh the code is shit, I know that Jul 28 20:29:50 it's terrible code but until kernel 2.6.22 I can't use any other driver for the hardware, so I'm kinda up a creek Jul 28 20:30:02 why not use 2.6.22? Jul 28 20:30:45 I don't want to try and update angstrom's kernel and last I checked it was 2.6.21-RP Jul 28 20:30:54 I'm sure there are a few things I do not know that would cause me issue Jul 28 20:31:09 http://svn.o-hand.com/view/poky/trunk/meta/packages/linux/linux-rp_2.6.22.bb?rev=2205&view=auto Jul 28 20:31:14 linux-rp 2.6.22 Jul 28 20:31:32 ahh, so this is the famous "poky" that I've heard about Jul 28 20:32:28 and why are poky packages not in sync with monotone? just poky is unstable and monotone is testing? Jul 28 20:33:10 ~poky Jul 28 20:33:11 well, poky is http://projects.o-hand.com/poky, or a subset of OpenEmbedded, created by o-hand.com to testdrive some of their technology Jul 28 20:36:14 ~o-hand Jul 28 20:36:20 ~o-hand.com Jul 28 20:37:17 LoL, and of course, I need .23 not .22 Jul 28 20:37:27 I guess it is time to start making myself more useful :-) Jul 28 20:46:04 'night all Jul 28 22:06:27 <_diego__> uhm: checking how to run the C++ preprocessor... /lib/cpp Jul 28 22:06:38 <_diego__> configure: error: C++ preprocessor "/lib/cpp" fails sanity check Jul 28 22:06:57 <_diego__> i got this error building libtool-native Jul 28 22:16:15 _diego__: yes? check config.log? Jul 28 22:17:57 <_diego__> don't say anything interesting, but looking on google, i found out it was a problem that g++ was not installed Jul 28 22:18:24 <_diego__> weird error message ^^ Jul 28 22:20:37 koen|away: The answer to c) is that we need to install all packaged in DEPENDS Jul 28 22:20:52 koen|away: Other than that, feel free to post it Jul 28 22:21:28 zecke: Given the invasive nature of the fetcher changes, should we run a 1.10 "stable" release series for them? Jul 28 22:24:56 if we say 2.0 is allocated for the GUI changes we should do that? Jul 28 22:25:02 s/?/!/ Jul 28 22:25:48 zecke: 1.10 would be 1,8 with trunk fetcher changes and 1,9 would become 1.11 Jul 28 22:26:12 zecke: UI would still be 2.0 and the changes towards to UI would not hit 1.10 Jul 28 22:26:27 Not entirely clean but... :} Jul 28 22:26:59 RP: ah okay, poky is already using the fetcher changes? Jul 28 22:27:12 zecke: no Jul 28 22:27:24 zecke: I would branch for 1.10 and request testing Jul 28 22:27:30 note not tag Jul 28 22:27:50 RP: you would feel more secure to not touch 1.8? We can do it that way then Jul 28 22:28:04 zecke: I'm happy enough to change 1,8 Jul 28 22:28:22 zecke: I just wondered if we should have a 1.10... Jul 28 22:28:31 I have the commit ready for 1.8 as it happens Jul 28 22:30:58 * zecke is a great help and says do whatever you want :) Jul 28 22:44:40 03rpurdie 07bitbake-1.8 * r849 10/ (7 files in 3 dirs): Add persitent data store from trunk, sync the fetcher changes to use the persistent store Jul 28 22:53:50 03slapin 07org.oe.dev * r7794e8dc... 10/ (1 conf/machine/palmt650.conf): conf/machine/palmt650.conf: Adding pxa2xx-ac97 module to autoload Jul 28 22:59:31 03rpurdie * r850 10/ (4 files in 2 dirs): Update parser to work optimally with the new fetcher code **** BEGIN LOGGING AT Sat Jul 28 23:27:25 2007 **** ENDING LOGGING AT Sun Jul 29 02:59:56 2007