**** BEGIN LOGGING AT Sat Jan 07 02:59:58 2006 Jan 07 04:00:12 morning all Jan 07 04:00:20 good morning RP Jan 07 04:00:56 RP: I think ewi546.ewi.utwente.nl/OE/new-tasks.conf can be much simpler after reading your gpe-image example again Jan 07 04:01:31 koen: Yes. This is one reason I want to get this code in and start fixing things :) Jan 07 04:01:44 Lots of things can be massively simplified Jan 07 04:03:09 all the nslu2 stuff is in, so we're ready to tag/branch/whatever Jan 07 04:04:00 The sooner, the better as far as I'm concered - I have more time at weekends than during the week... Jan 07 04:04:15 we just need someone of ar,andyc,anoncvs,ayer,bavery,cterry,davep,duffekn,france,gitars,hrw,jamey,jg,kergoth,mickeyl,mithro,noda,pb to edit the OE frontpage Jan 07 04:04:42 RP: uni starts on monday :( Jan 07 04:05:08 Ah :-/ Jan 07 04:05:38 I don't think we'll break things too badly to be honest. Jan 07 04:05:50 The package manager for staging is probably the most invasive change Jan 07 04:05:51 better safe than sorry Jan 07 04:06:03 I agree, I'm just saying :) Jan 07 04:10:42 * koen puts his course schedule into iCal Jan 07 04:11:26 koen: FYI, I just did a big rename in the nslu2-linux stuff (only about 10 directories, "openslug" -> "slugos") Jan 07 04:11:39 it's pushed now at monotone.nslu2-linux.org, so you should get it soon here. Jan 07 04:13:06 rwhitby: I haven't seen much response to your 'where to go from here mail' Jan 07 04:13:43 koen: there was a bit on the nslu2-linux mailing list, which you wouldn't have seen unless subscribed. Jan 07 04:14:02 http://news.gmane.org/gmane.comp.misc.nslu2.linux ? Jan 07 04:15:47 rwhitby: gmane doesn't have any replies to http://article.gmane.org/gmane.comp.misc.nslu2.linux/10693 Jan 07 04:16:01 03rwhitby 07org.oe.dev * re748d25b... 10/ (30 files in 15 dirs): slugos: Major rename of common SlugOS files and directories from openslug to slugos. Jan 07 04:16:05 03rwhitby 07org.oe.dev * rc98db1ad... 10/.mt-attrs: slugos: Major rename of common SlugOS files and directories from openslug to slugos. Jan 07 04:16:26 Hmm - for some reason none of the replies have ended up there. You can see the thread here: http://groups.yahoo.com/group/nslu2-linux/message/10496 Jan 07 04:17:33 and the follow-up proposal is here: http://groups.yahoo.com/group/nslu2-linux/message/10557 Jan 07 04:17:58 kergoth: I guess you didn't have a chance to look at the DEPENDS/RDEPENDS patch yet? :-/ Jan 07 04:23:05 crumbs, it's snowing again Jan 07 04:25:32 Hmm, they said that was going to happen... Jan 07 04:25:41 just damp here atm Jan 07 04:25:46 oh, did they? Jan 07 04:25:49 I guess I missed that. Jan 07 04:30:30 koen: Do we have some text for the front page ready for someone with admin access to place there? I think we might as well start the ball rolling and tag (branch if we're going to although I still think we lack manpower for this) Jan 07 04:30:48 morning all Jan 07 04:30:53 Someone could always start a branch based on the tag if they're going to maintain it I guess Jan 07 04:30:55 hi Dirk Jan 07 04:31:12 RP: go ahead and tag :) Jan 07 04:31:30 koen: You know my track record with monotone ;-) Jan 07 04:31:37 heh Jan 07 04:31:48 you can't do much wrong with tagging Jan 07 04:32:25 Is that a challenge? :) Jan 07 04:33:01 the disapprove weirdness is caused by the fact that every revision is a product of sha1 summing it's ancesters, so you can't just 'delete' a cset in the middle Jan 07 04:33:54 koen: I know - I'd just assumed it would apply a reversion diff and then automatically know how to merge it Jan 07 04:34:10 but it has to have the merging explained to it :-/ Jan 07 04:34:25 yeah, that sucks Jan 07 04:34:35 monotone needs be less kde and more gnome ;) Jan 07 04:34:56 Any ideas for a tag name? Jan 07 04:35:16 jan2006prebreakage? jan2006stable? Jan 07 04:35:30 jan2006prebreakage Jan 07 04:35:44 since it isn't really stable Jan 07 04:39:03 0b9a2dfead1136f996280e9827949f83cb032537 is the revision to tag? Jan 07 04:40:28 yes Jan 07 04:40:35 good morning, btw. Jan 07 04:40:49 Its a merge which why it doesn't match the commits list - was just confused for a moment :) Jan 07 04:40:51 morning mickeyl Jan 07 04:40:57 hey mickeyl Jan 07 04:41:15 * mickeyl packing Jan 07 04:41:20 tagged Jan 07 04:42:29 when you push you should just have one 1 cert Jan 07 04:43:43 koen: I did Jan 07 04:43:58 monotone: bytes in | bytes out | certs out | revs out Jan 07 04:43:58 monotone: 637 | 2052 | 1 | 0 Jan 07 04:46:30 http://ewi546.ewi.utwente.nl/tmp/viewmtn/tags.psp Jan 07 04:46:33 and it's in Jan 07 04:46:50 I'm writing a mail to oe@ Jan 07 04:46:56 excellent. thanks for doing this Jan 07 04:48:08 Hmm. A yahoo group/listserv? Jan 07 04:50:32 mickeyl: good morning Jan 07 04:51:38 morning phil Jan 07 04:51:44 Alan_K: its on handhelds.org Jan 07 04:51:51 mail sent to oe@ Jan 07 04:52:36 Does someone want to add " use monotone update -rjan2006prebreakage to avoid breakage" to the topic (or similar)? Jan 07 04:53:10 RP: go ahead :) Jan 07 04:53:25 koen: I don't have chanserv privs Jan 07 04:53:42 RP: no need to Jan 07 04:53:52 kergoth changed that a few weeks ago Jan 07 04:54:01 ah, I didn't know that Jan 07 04:59:05 I've added a note about this at the top of getting started on the wiki so just the front page on OE remains Jan 07 04:59:50 If someone could post an editted version of the email to oe@ there, that would be good... Jan 07 04:59:53 back in a few mins... Jan 07 05:02:36 hi folks Jan 07 05:03:11 hey alan|laptop Jan 07 05:03:21 hey koen Jan 07 05:03:36 hi seem to have a bug in monotone 0.25. Jan 07 05:03:44 does anyone use this version ? Jan 07 05:05:47 I do Jan 07 05:34:53 koen: did you comile from sources ? Jan 07 05:35:25 no, I used the static version from venge.net Jan 07 05:36:13 static version, huh ? ... ok, i think i'm going to try this one. thanks koen Jan 07 05:42:27 morning all Jan 07 05:43:22 hey lrg|home Jan 07 06:55:32 hi Jan 07 06:56:22 pb_: I'm subcribed to fam-dev for whatever reason :} Jan 07 06:56:33 zecke: oh, right Jan 07 06:56:33 heh Jan 07 06:57:20 * zecke is tired... Jan 07 07:00:50 I have converted 176 revisions already :} Jan 07 07:01:01 very good Jan 07 07:01:05 how many revisions remain? Jan 07 07:01:36 pb_: dunno. 6 more month ;) Jan 07 07:01:47 no time to lose then Jan 07 07:04:04 pb_: I'm starting another convert on a P4 system Jan 07 07:04:15 ah, good idea Jan 07 07:04:21 what were you using before; an ipaq? Jan 07 07:04:59 pb_: qemu on my shark to emulate PPC ;) Jan 07 07:05:14 pb_: actually it is my G4 PPC Jan 07 07:05:51 oh right. I would have expected that to be pretty fast too. Jan 07 07:06:06 (the g4, not the shark) Jan 07 07:06:24 pb_: tailor looks inefficent Jan 07 07:06:40 pb_: monotone update to rev, rsync to the targetdir Jan 07 07:06:48 pb_: svn add/git add each file Jan 07 07:06:52 pb_: svn commit Jan 07 07:07:00 oh, I see Jan 07 07:07:07 pb_: one can imagine that svn add takes ages on OE Jan 07 07:07:13 so it doesn't actually understand the monotone database; it just checks out every revision individually? Jan 07 07:07:41 yeah, one could indeed imagine this being a slow process Jan 07 07:09:12 pb_: I had hoped it would better as well Jan 07 07:09:19 pb_: at least this is a generic approach Jan 07 07:09:22 yeah Jan 07 07:09:53 pb_: I wonder why it doesn't use the svn status Jan 07 07:11:07 hi zecke Jan 07 07:11:25 RP: the git database is still not complete Jan 07 07:12:00 zecke: I gather its being slow. Perhaps I'll create a version with just the current files for now. We can always add history later... Jan 07 07:12:13 zecke: I assume you've seen the DEPENDS/RDEPENDS patch? Jan 07 07:12:40 RP: I have not seen the patch yet Jan 07 07:13:04 zecke: I'd be interested in your comments on it Jan 07 07:14:12 and pb_ 's for that matter :) Jan 07 07:14:48 OE is now tagged and ready for breakage - if we can get that into bitbake, I can start fixing DEPENDS... Jan 07 07:16:16 RP: the reverse hash building doesn't look too slow?! Jan 07 07:16:33 zecke: No, I fixed all the speed issues Jan 07 07:16:40 There is no speed problem now Jan 07 07:20:28 RP: I will read up the discussion later Jan 07 07:20:42 zecke: ok, thanks Jan 07 07:20:49 RP: the patch looks correct though. What kind of regexp do you need for the dynamic_depends? Jan 07 07:21:06 packages_dynamic even Jan 07 07:21:17 Packages set things like PACKAGES_DYNAMIC = "kernel-module-*" Jan 07 07:28:47 RP: two issues I see: There is no test case for your code ;) Jan 07 07:29:01 RP: do you want to have that patch applied before or after the release Jan 07 07:29:22 zecke: No test case in what sense? Jan 07 07:30:41 I think it should be applied after the next release and people will just have to update to use the "broken" oe data. If it all goes well, we can release a new version of bitbake as we mark OE "stable" again Jan 07 07:30:59 RP: A small metadat repository like we have for the BB Collections Jan 07 07:31:41 zecke: I did run it against OE itself after making the changes I mention in the email Jan 07 07:32:17 zecke: As OE will be making a lot of use of this feature, its likely we'd know about future breakage in it very quickly :) Jan 07 07:33:06 but a test case should be easy to create - just need an image creation .bb with only an RDEPENDS - if you get an image, it works Jan 07 07:39:09 i'm having a monotone pull problem... here are the last two messages Jan 07 07:39:15 monotone: trying 3-way merge Jan 07 07:39:20 monotone: misuse: no unique private key for cert construction Jan 07 07:39:32 any idea? i haven't made any branches locally... Jan 07 07:40:28 abrewer: Did you commit anything locally? Jan 07 07:40:47 no, i'm not even sure how to yet :-) Jan 07 07:42:11 For some reason you're seeing that there are two heads and its wanting to merge. Its not letting you merge as you don't have a key Jan 07 07:42:22 However you shouldn't be seeing two heads Jan 07 07:42:45 Which branch is this? Jan 07 07:44:47 uh... org.openembedded.dev is that right? Jan 07 07:44:57 Yes, that should be fine Jan 07 07:45:21 i get this message prior: common ancestor 4f6c7596b5871a9e3eafaef263afe31a0239a07b koen@openembedded.org 2005-12-17T17:39:26 found Jan 07 07:45:54 17/12/05 is quite a while ago :-/ Jan 07 07:46:03 how do I tell monotone to update from a particular branch? Jan 07 07:46:09 i took some time off! (-; Jan 07 07:46:17 --branch= Jan 07 07:46:28 then give the key i just pasted? Jan 07 07:46:59 No, that's a revision but that won't solve your problem... Jan 07 07:47:05 i'm sorry, i'm a rcs/cvs guy, and don't quite yet have my head around monotone. Jan 07 07:47:59 abrewer: Have you replaced your database afer the corruption issues we saw at the beginning of Decemeber? Jan 07 07:49:40 no i have not. should i check on the mailing list archives for instructions? Jan 07 07:50:10 abrewer: It might be an idea. Basically, download a new snapshot of the database and try using that to replace your existing one Jan 07 07:51:27 I somehow suspect that the topic isn't getting the attention it deserves :( Jan 07 07:53:08 koen|away: It isn't obvious that email link contains info related to this I guess Jan 07 08:14:18 lol I have fucked up tailor... I need to restart Jan 07 08:34:49 lrg: hey Jan 07 08:35:05 lrg: may I message you? Jan 07 08:35:12 lrg|home Jan 07 08:36:40 zecke: doh Jan 07 08:38:18 pb_: not too bad. I'm already at rev 71 on my P4 workstation Jan 07 08:38:43 pb_: doing just one conversion at a time. no verbose, no debug output (no io) makes it go way faster Jan 07 08:41:06 ah, very good Jan 07 08:41:13 pb_: obviously converting is bound to disk io. my UDMA133 machine can do that hopefully better than my ibook Jan 07 08:41:42 pb_: I probably should have gotten a Thinkpad with Serial-ATA and PCI-Express Jan 07 08:42:14 talking about thinpads... my contract @ university ends and i need to give back my good-ole-trusty-T22 workhorse. Anyone experiences with the X41 ? Jan 07 08:42:28 * mickey|packing wants a thinkpad again Jan 07 08:42:36 mickey|packing: got a picture of that? Jan 07 08:42:45 the X41? yeah, wait a moment Jan 07 08:42:57 mickey|packing: yes, but actually all thinkpads are supported quite good Jan 07 08:43:51 my reqs are #1 keyboard, #2 trackpoint #3 duratio Jan 07 08:43:58 the x41 is the small one, right? Jan 07 08:44:09 right Jan 07 08:44:14 one of the guys I work with bought a t43 recently, which I gather rules, but it's quite a bit bigger Jan 07 08:44:29 mickey|packing: https://wiki.ubuntu.com/LaptopTestingTeam/ThinkpadX41-2527 Jan 07 08:44:54 woah, that's a cool site. thanks zecke Jan 07 08:45:24 i'm tempted to buy the tabletPC version, but since I would be running Linux mainly i guess it's not worth the additional fee Jan 07 08:45:30 hello everyone Jan 07 08:45:35 hi alan Jan 07 08:45:48 hi mickey|packing Jan 07 08:46:16 argh... Jan 07 08:46:18 ./monotone-0.25-linux-x86: /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.6' not found (required by ./monotone-0.25-linux-x86) Jan 07 08:46:23 I'm still quite fond of my sharp, too, which seems to work okay in linux. Jan 07 08:46:36 ibot: mail to hrw|gone: i have problems with bitbaking gpe-image and time is running out. I'll handle the release when I'm back. Nevertheless please do some tests, if you can. Jan 07 08:46:38 that's too long, mickey|packing Jan 07 08:46:39 does anyone know what i need to correct this ?* Jan 07 08:46:45 ~curse ibot Jan 07 08:46:47 May the fleas of a thousand camels infest your most sensitive regions, ibot ! Jan 07 08:46:53 ibot: mail to hrw|gone: i have problems with bitbaking gpe-image and time is running out. I'll handle the release when I'm back. Jan 07 08:46:55 mickey|packing: that's too long Jan 07 08:46:58 *sigh* Jan 07 08:47:24 ibot: do you know tcp? create multiple mails... *sigh* Jan 07 08:47:26 yes, I know tcp? create multiple mails... *sigh*. Jan 07 08:47:53 ibot has humour... Jan 07 08:47:56 ibot: botmail to hrw: i have problems with bitbaking gpe-image and time is running out. I'll handle the release when I'm back. Jan 07 08:48:04 aah! botmail vs. mail Jan 07 08:48:08 heh Jan 07 08:48:15 crazy ibot Jan 07 08:48:20 I wonder what "mail" does. Jan 07 08:48:23 heh, me too Jan 07 08:48:35 ~lart timriker Jan 07 08:48:35 * ibot sends a legion of lawyers after timriker's head Jan 07 08:48:42 perhaps a SMS Jan 07 08:48:55 hehe Jan 07 08:49:15 * mickey|packing thinks about reasons to convince his wife the X41 is necessary Jan 07 08:50:03 no reason to give... offer flowers, a good restaurant, and that's it :) Jan 07 08:50:05 maybe you should just convince your next employer to give you one Jan 07 08:50:18 "just this last toy, please" Jan 07 08:52:15 or, judging from your new nick, I guess you could stage some kind of heist against the local computer store Jan 07 08:54:45 mickey|packing: could you have a look at http://handhelds.org/hypermail/oe/54/5432.html ? Jan 07 08:55:04 pb_: hehe Jan 07 08:55:34 koen: ah yes. I'll do that in a minute. Jan 07 08:55:49 mickey|packing: thanks Jan 07 08:58:15 mickey|packing: while you have that page in your editor, could you change the 'open embedded' title to 'OpenEmbedded' ? Jan 07 08:59:27 Good evening. Jan 07 08:59:52 hey uv1 Jan 07 09:00:51 koen: done Jan 07 09:01:00 Have problem to get qmake working again, since I upgraded my Debian Sarge. Jan 07 09:01:48 "QMAKESPEC is not set" is complained. Do we do so with bitbake, or do I have to set that manually ? Jan 07 09:02:14 according to qmake.bbclass we are doing that Jan 07 09:03:09 mickey|packing: "NOTE: using qmake spec in /data/scdev2/build-boxer-mono/tmp/staging/i686-linux/share/qmake/linux-oe-g++, using profiles ''" Jan 07 09:03:14 on my system that is set anyway. dunno i it still appplies to make things work i had to unset it porior to starting the build Jan 07 09:03:29 but that was ages ago, in the days of oemake etc Jan 07 09:03:57 yeah, unsetting it in your top level env is best, but i think uv1's problem has nothing to do with that Jan 07 09:04:06 k Jan 07 09:04:06 uv1: try showdata on the said file Jan 07 09:04:24 uv1: are all your rpm patches in oe already? Jan 07 09:04:38 mickey|packing: showdata what ? Jan 07 09:04:49 sorry, i need to leave for the preflight checkin - leaving in the middle of the nite, so we can checkin a couple of hours before Jan 07 09:04:55 bbl Jan 07 09:04:57 koen: No change since last time. All patches in buglist. Jan 07 09:04:57 hehe, qmake has just starte dbuilding here Jan 07 09:06:29 Oh its different "make: Nothing to be done for `first'" Jan 07 09:07:05 RP: looks like you can start breaking stuff now ;) Jan 07 09:07:42 he my history is too small, does anubody else still use simpad at all? Jan 07 09:07:54 i wanna have another go at getting it running Jan 07 09:08:19 dan2003: like this http://familiar.handhelds.org/snapshots/v0.8.3/RC1/simpad/ ? Jan 07 09:08:27 last i tried it had been updated to use 2.6 but gfx driver was broken.. so i think ill try it with a 2.4 kernel Jan 07 09:08:37 koen: zecke: The fetch.py patch would be necessary for taht module problem from 2 days ago ... Jan 07 09:08:51 ohh, cool! Jan 07 09:09:06 that's with 2.4.27 Jan 07 09:09:16 ok, thanks.. ill give that a go Jan 07 09:09:37 Tried localdir="." yesterday. That build a cvs target tgz with ALL cvs dirs ;-( So that does not work as expected. Jan 07 09:09:37 im still on some brokem old versink hre all the onts are messed up! Jan 07 09:12:11 Somone here is using a 5500 with gpe.. ? I'm having trouble with usb.. the Zaurus reset himself the connection, this don't do it with opie.. the problem is where ? Jan 07 09:13:47 Sorry bad channel Jan 07 09:13:48 :S Jan 07 09:18:18 do_pobulate_staging does not find the libs any more. That worked since approx. a month ago. Any changes here ? Jan 07 09:18:33 s/pobulate/populate/ ;-) Jan 07 09:19:51 uv1: we broke the meta data Jan 07 09:20:06 zecke: Means what ? Jan 07 09:21:53 hm, is it possible to make a make an image of the flash in simpad, i appear not to have image file i am running anymore? just in case i need to go back Jan 07 09:22:31 dd if=/dev/mtblock of=/path/to/card/dump.jffs2 Jan 07 09:23:11 ok, i have 0, 1 and 2, i should do all of them? Jan 07 09:23:38 uv1: well we broke the OE tree and people are in the process of fixing it Jan 07 09:23:47 or i s0 th eboot, 1 the rootfs and 2 home ? Jan 07 09:23:51 uv1: though I have not followed the discussion to closely Jan 07 09:24:50 zecke: More important : Did you catch my fetch.py hint ? Jan 07 09:25:16 uv1: for cvs? I did not get whyt you try to do :} Jan 07 09:25:20 dan2003: check with 'cat /proc/mtd' Jan 07 09:25:55 zecke: Lets try to explain a little bit more ... Jan 07 09:26:08 uhmm. i dont have that ;(! Jan 07 09:26:38 2 secs,.im being aplank Jan 07 09:26:44 uv1: I have -d:ext:foo@cvs.bmco:/cvs/uv1 als CVSROOT Jan 07 09:27:00 k, got it Jan 07 09:27:00 uv1: now my module is 'uv1'? Jan 07 09:27:12 uv1: and the file will clash with another file?! Jan 07 09:29:41 Shit Ctrl+Backspace instead of Ctrl+Return ;-) Jan 07 09:29:52 STep by step .. Jan 07 09:30:12 zecke: Assume you have that CVSROOT ... Jan 07 09:30:34 Now you checkout with module="modabc"... Jan 07 09:30:34 uv1: yes Jan 07 09:31:06 Result is no as expected cvs/abc-1.2.3/modabc/{bin,src,lib} ... Jan 07 09:31:27 but cvs/abc-1.2.3/{bin,src,lib} (note the missing modabc dir) Jan 07 09:31:29 uv1: where is abc-1.2.3 coming from? Jan 07 09:31:49 package name ? Jan 07 09:31:59 okay Jan 07 09:32:06 ah right Jan 07 09:32:13 fetch.py gives up because modabc is not there. Jan 07 09:32:46 My patch creates that dir, if necessary and just copies all other stuff inside before tar starts Jan 07 09:32:57 zecke: Got it ? Jan 07 09:33:40 uv1: yes. but why should it miss 'modabc' and then need it again? Jan 07 09:33:51 uv1: I do not get where it should be missing Jan 07 09:35:02 fetch.py wants to tar cvfz cvs/abc-1.2.3/modabc! All other stuff is leading to an error ... Jan 07 09:35:05 koen, howdo i use that simpad file? i just unpacked the tar.bz2 and its all the files for the rootfs rather than an image of it Jan 07 09:35:15 do i nned to just build an image of them and flash that? Jan 07 09:35:22 dan2003: upload the jffs2 file with bootldr? Jan 07 09:35:31 I'm not really familiar with simpad stuff Jan 07 09:35:33 And our CVS is not organized that way. If you checkout a module, you don't get it in a separate direktory ... Jan 07 09:35:57 uv1: ah well Jan 07 09:36:16 zecke: I thought , that possible localdir="." helps (found in fetch.py), but this leads to a tar archive containing the COMPLETE cvs content ;-) Jan 07 09:36:22 koen, hehe, ill go back to bed! i dled the tar.bz2 instead of the jff2. not awake yet Jan 07 09:38:15 zecke: Still alive ? Jan 07 09:38:16 zecke: I noticed there was a perforce fetcher path on the bitbake list a while back - we should probably merge that at some point now Jan 07 09:38:29 uv1: yes Jan 07 09:38:46 uv1: okay so there is no bb file that currently fails (in oe.org.dev) Jan 07 09:38:54 RP: yes. How can we test that fetcher? Jan 07 09:39:02 zecke: Maybe ... Jan 07 09:39:34 uv1: could you mail me your patch again Jan 07 09:39:41 zecke: I have no knowledge of perforce so testing is tricky. Its proably ok to take it on trust... Jan 07 09:39:57 RP: and remove the bk fetcher (which is not working anyway) Jan 07 09:40:11 zecke: sure, where to ? Jan 07 09:40:11 zecke: Yes, the bk fetcher is a joke :) Jan 07 09:40:14 RP: well FreeBSD uses preforce for some szuff Jan 07 09:40:26 uv1: zecke@handhelds.org Jan 07 09:40:34 don't the trolls use perforce as well? Jan 07 09:40:47 koen: yes... Jan 07 09:41:15 I think we add it and its then there is anyone wants to use it. If its broken, they'll just have to fix it :) Jan 07 09:41:31 RP: I will try to get the free perforce client Jan 07 09:41:55 * RP wonders what to break in oe first :) Jan 07 09:42:48 Are we all in agreement that BUILD_ALL_DEPS and the PACKAGES_DYNAMIC variables are needed? Jan 07 09:43:01 add multimachine stuff to fam + oz Jan 07 09:43:07 I can add them without changing any existing behaviour... Jan 07 09:43:30 http://www.perforce.com/perforce/loadprog.html Jan 07 09:43:51 I do not like the DYNAMIC Jan 07 09:44:01 name Jan 07 09:44:48 zecke: What do you suggest instead? Jan 07 09:45:05 RP: I don't know Jan 07 09:46:47 zecke: patch was sent. Jan 07 09:47:08 Different question again. The qmake stuff seems broken. Jan 07 09:47:19 I'll listen to suggestions but we need something. dynamic was chosen as the sub packages list is unknown and dynamically generated at packaging time... Jan 07 09:47:45 For some reason, oe_libinstall does not find libs any for for staging / installing. That worked before. Jan 07 09:48:27 Which part of bitbake / bbclass does handle that stuff. I assume before the *.pro file was checked to find out the targt lib dir ? Jan 07 09:48:28 PACKAGES_SURPRISE Jan 07 09:48:44 uv1: check the OE mailinglist archive about staging discussions Jan 07 09:52:07 zecke: Lots of stuff, but did not find any thing regarding qmake and changes already done Jan 07 09:54:04 uv1: oe_libinstall is likely to be in base.bbclass :} Jan 07 09:58:39 * RP adds CVS_TARBALL_STASH to the hit list of variables to rename (and in this case make it accept multiple items) Jan 07 10:06:04 zecke: rearranged my packages to store libs in the sources dir. That halp for now. Will try to find out, whats wrong tomorrow. Jan 07 10:06:17 Did you check my patch already ? Jan 07 10:08:02 BTW : I would have some packages, which possibly are helpfull for others as well ... Jan 07 10:08:49 Though they are not really opie-compliant, they are able to run on opie (since approx. 3 years ;-)) Jan 07 10:09:17 How could I get that packages in the ore.or.dev feed ? Jan 07 10:11:51 is blupdater.img the same file but with a different name as bootldr-simpad-2.21.12.bin ? Jan 07 10:17:34 uv1: Submit patches to the bugzilla and someone will hopefully look at them Jan 07 10:18:07 RP: You mean Open a bug / feature request ? Jan 07 10:19:04 uv1: Yes. Adding packages is a feature :) Jan 07 10:19:11 This way the patches don't get lost Jan 07 10:19:32 RP: Thanks ... and bye Jan 07 10:20:50 OT: do you guys know you can do MD5SUM = "sum"? Jan 07 10:22:32 zecke: No, what does that do? Jan 07 10:22:40 03eFfeM 07org.oe.dev * r37e4e397... 10/packages/pvrusb2-mci/pvrusb2-mci_20060103.bb: pvrusb2-mci: added 20060103 release Jan 07 10:23:08 RP: will check the md5sum of the to be downloaded file Jan 07 10:23:30 RP: MD5SUM_filename for multiple files Jan 07 10:23:53 RP: I'm going to add a SHA256 sum as well (now that I see freeBSD starts to use it) Jan 07 10:23:58 zecke: I always thought this should be part of the SRC_URI Jan 07 10:24:16 RP: wait a second Jan 07 10:24:23 MD5SUM = "sum" if you have only one file Jan 07 10:24:29 SRC_URI = "http://somefile;md5sum=xxx" Jan 07 10:25:21 RP: I have MD5SUM_basename for now Jan 07 10:26:02 zecke: I'd consider my suggestion above - it totally connects the file with the checksum and we can implement it nicely in the fetcher code Jan 07 10:26:17 RP: yes it makes sense Jan 07 10:26:56 It also stops us having massive namespace pollution with loads of MD5SUM and SHA256 etc Jan 07 10:27:42 RP: (looking at the wget fetcher): I would love to implement the (pre)mirror handling in the fetch module itself Jan 07 10:28:10 RP: this way you could use scp/rsync as premirror as well... Jan 07 10:28:42 zecke: Where is this currently handled? Somewhere in OE? Jan 07 10:29:12 RP: in the fetchers (wget.py) Jan 07 10:29:37 RP: same with the tarballstash Jan 07 10:29:44 zecke: I'm in favour of moving that into some common code Jan 07 10:30:02 It was always hoped we could refactor it - splitting it up was only the first step Jan 07 10:30:12 right Jan 07 10:30:36 I'd love to see multiple CVS_TARBALL_STASH sources support Jan 07 10:31:23 and renamed to SRC_SCM_STASH or something Jan 07 10:33:00 RP: Would you kill MD5SUM completely? Jan 07 10:33:14 zecke: given the lack of use of it, yes Jan 07 10:33:27 RP: qte uses it (if I pushed it) Jan 07 10:33:44 converting shouldn't be too painful though? Jan 07 10:34:27 Screw you, I'm going home. Jan 07 10:35:49 sorry :) Jan 07 10:38:02 03jbowler 07org.oe.dev * r0241a34e... 10/packages/fetchmail/fetchmail_6.2.5.bb: Jan 07 10:38:02 fetchmail: ensure fetchmail builds with USE_NLS=no in 6.2.5 Jan 07 10:38:02 - on systems without NLS (i.e. USE_NLS=no in the config) Jan 07 10:38:02 - fetchmail will still reference libintl.h unless Jan 07 10:38:02 - --disable-nls is used, inheriting gettext ensures that Jan 07 10:38:03 - the correct nls option is passed to config Jan 07 10:39:57 Which gpe developer thinks .bb files use tabs as white space? Jan 07 10:40:06 * RP shudders Jan 07 10:42:04 RP: okay m5sum is working now Jan 07 10:43:16 zecke: That sounds good :) Jan 07 10:44:38 We really need a lindent equivalent for .bb files... Jan 07 10:44:41 didn't we have md5sum support in qt/e for months? Jan 07 10:45:10 * koen wants oelint be be on by default Jan 07 10:45:22 We have one? Jan 07 10:45:44 koen: Do you know which developer has put the tabs into the gpe files? Jan 07 10:45:49 a bbclass Jan 07 10:45:52 probably me Jan 07 10:45:58 I don't need to know names but if someone could have a word... ;-) Jan 07 10:46:58 there was a discussion about that a while back on oe@ Jan 07 10:47:09 What was the concensus? Jan 07 10:47:18 I'm all for a bit of maintainer freedom but this is going to have to be a policy... Jan 07 10:47:56 I think the consensus was don't use tabs Jan 07 10:48:08 right :) Jan 07 10:48:18 * RP updates the wiki style guide Jan 07 10:48:18 I still thinks some aligning benefits readability Jan 07 10:48:55 sometimes maybe the odd extra space here or here but in general its something I don't like... Jan 07 10:49:04 you could override your tabs for spaces ... Jan 07 10:49:13 Every .bb file I just opened with tabs in looked dreadful :-/ Jan 07 10:49:32 Maybe that my tab spacing setting but even so... Jan 07 10:49:38 stuff like SRC_URI really benefits from alignenment Jan 07 10:50:59 hmm Jan 07 10:51:07 i just broke familiar ... Jan 07 10:51:36 koen: I totally agree with that Jan 07 10:51:48 koen: Just not things like PV = "r4" Jan 07 10:52:31 me neither, I would put that in PR ;) Jan 07 10:52:55 koen: gah. You know what I mean :) Jan 07 10:53:07 :p Jan 07 10:53:08 Looking at the wiki we have a style guide but its totally confused with the details about bb file variables :-/ Jan 07 10:56:18 reenoo_, mickeyl: wb Jan 07 10:56:40 thanx Jan 07 10:56:42 mickeyl: all done with packin' and checkin, then? Jan 07 10:57:03 evening all Jan 07 10:57:04 pb_: yep. now to prepare the data on the laptop and the rest of hand luggage Jan 07 10:57:07 evening reenoo_ Jan 07 10:57:07 re pb_ Jan 07 10:57:31 mickeyl: very good Jan 07 10:57:38 i never was on a plane that takes off 4:00 in the middle of the nite Jan 07 10:58:03 heh Jan 07 10:58:27 hmm.. lanzarote... Jan 07 10:58:46 mickeyl: have fun! Jan 07 10:59:05 yeah 203 of 2800 revisions converted Jan 07 10:59:20 mickeyl: heh. somehow, almost every plane I fly on seems to take off at that sort of time Jan 07 10:59:39 zecke: rock Jan 07 10:59:50 sounds like tailor is really motoring now Jan 07 11:00:02 heh Jan 07 11:02:40 Hmm. oe.handhelds.org just stopped talking to me as I was editing the wiki :-/ Jan 07 11:21:08 RP: hh.org seems to have dropped of the net Jan 07 11:29:20 koen, have you guys started on fam 0.9? Jan 07 11:32:39 hmm guess i am still on ignore :( Jan 07 11:59:05 ibot, seen florian Jan 07 11:59:08 florian was last seen on IRC in channel #gpe, 18h 35m 13s ago, saying: 'good night'. Jan 07 12:01:52 fun git and tailor don't work on FreeBSD :} Jan 07 12:09:46 that's very sad Jan 07 12:10:33 that forces me to use this evil penguin system Jan 07 12:11:53 life is hard Jan 07 12:13:56 zecke, even with linux compat enabled? Jan 07 12:14:50 emte: not tried that. /usr/local/bin/git had #!/bin/sh in it Jan 07 12:15:05 emte: but uses bashism or such (shift doesn't like they way it is invoked) Jan 07 12:15:27 ah Jan 07 12:16:06 emte: and now I want to get something done ;) Jan 07 12:18:42 yeah, it must be annoying when no software is compatible with your system Jan 07 12:18:49 lol Jan 07 12:18:53 he isnt using windows Jan 07 12:18:56 can you even run opie on openbsd? Jan 07 12:19:06 er, freebsd Jan 07 12:19:44 pb_: QVFb works (on OS X and FreeBSD) Jan 07 12:20:31 pb_: it is quite tough to live in a binary world... where you get libraries for X86/Linux... ;) Jan 07 12:20:46 what would life without pain Jan 07 12:22:08 heh, indeed Jan 07 12:37:34 hmm ... has anyone successfully used monotone db migrate? Jan 07 12:43:20 does anybody know if florian has a snapshot of the buildsystem used to build the simapd image 0.8.3-rc1 anywhere for download? Jan 07 12:44:30 hard to say, buildsystem sanpshots are rare Jan 07 12:47:10 * france is back (gone 28:30:47) Jan 07 12:47:49 hey france you kick the hh.org server? Jan 07 12:48:59 load average up over 185 Jan 07 12:49:26 backing up? Jan 07 12:49:56 site is down is all Jan 07 12:52:33 the site appears fine now Jan 07 12:53:01 ah, thnks Jan 07 12:53:18 RP, will appreciate that Jan 07 12:54:14 from the logs, msn, google, yahoo and askjeeves all hit the site from 1:30pm to 2:15pm est time. Jan 07 12:54:34 bad them Jan 07 12:54:48 dan2003: http://familiar.handhelds.org/source/familiar-0.8.3-OE-tree.tar.bz2 or checkout org.openembedded.oz354fam083 Jan 07 12:54:59 Googlebot/2.1; +http://www.google.com/bot.html, etc.... Jan 07 12:55:25 Yahoo! Slurp; http://help.yahoo.com/help/us/ysearch/slurp Jan 07 12:55:31 france: is bugzilla shielded of from search engines? Jan 07 12:55:38 koen: nope Jan 07 12:55:58 everytime I access bugzilla apache and mysql take 100% cpu for a few seconds Jan 07 12:56:02 koen: parts of bugzilla, but much of it is exposed. Jan 07 12:56:35 koen: buzilla needs some tlc Jan 07 12:57:16 koen, thanks Jan 07 12:57:20 and we had over 20 cvs downloads at the time, and 3 rsyncs.... Jan 07 12:57:29 yeah, the script to add new stuff from the unstable feed needs to be taught to ignore locale stuff Jan 07 12:58:08 the machine needs more DRAM for heavy i/o like that, but I do not think even 8G would have helped. Jan 07 12:58:30 heh Jan 07 12:58:37 a ramdisk to hold all databases :) Jan 07 12:58:43 heh Jan 07 12:59:03 second server for load balancing perhaps ... Jan 07 12:59:30 emte: it is in progress.... Jan 07 12:59:51 emte: cvs is being moved to its own server, with a 2ndary for anoncvs Jan 07 13:00:21 ah, did not know that Jan 07 13:00:23 emte: we have two new severs with 4 time the HD space for Round robin ftp/http services. Jan 07 13:00:50 donation from someone? Jan 07 13:01:02 emte: e-mail is being moved to its own pair of servers, so we have enough cpu power to spam filter the individual user accounts as well as the mailing lists.... Jan 07 13:01:36 I recieved 11 servers from HP Labs when they shut down. 1/2 at MIT 1/2 at OSU. Jan 07 13:01:57 they are not the newest, but better than what we had. Jan 07 13:02:00 * koen tries to integer divide 11 by 2 Jan 07 13:02:04 lol Jan 07 13:02:11 the 11th is for the devcluster. Jan 07 13:02:37 the cluster was donated as well? Jan 07 13:03:02 yup.. only took me 3 months ... and many meetings... Jan 07 13:03:22 the dev cluster is at OSU ..... Jan 07 13:03:36 I will probably have to make one more trip out there to get it up and working. Jan 07 13:04:33 I also managed to get a rack of alpha systems....it is now at osu. Jan 07 13:05:14 it takes a long time to get everything setup. Jan 07 13:05:57 and it is after the 1st of the year, so on monday, I should start spending time looking for a project instead of doing hh.org stuff. Jan 07 13:06:11 Alpha *nice* Jan 07 13:06:36 I have a quad processor at home, with 16G DRAM and 16M cache on each cpu Jan 07 13:06:40 it is very fast. Jan 07 13:08:50 16mb l1? Jan 07 13:09:02 koen: yes Jan 07 13:09:06 or is the l1/l2 split a x86ism? Jan 07 13:10:20 alphas had l2 as well Jan 07 13:10:21 the l1/l2 split is an x86 thing but other arches do it as well Jan 07 13:10:32 emte: yes, some of the early ones did. Jan 07 13:10:43 atleast the microware systems did Jan 07 13:10:52 and the ones not made by DEC Jan 07 13:15:46 ahh Jan 07 13:32:07 http://linuxdevices.com/news/NS9808373276.html Jan 07 13:32:17 I guess we need to market OE better Jan 07 13:32:32 kergoth has been doing that kind of stuff with OE for years Jan 07 13:33:50 where is the build date cached? Jan 07 13:36:33 monotone: misuse: error opening file packages/initscripts/initscripts-1.0/akita/keymap: Datei oder Verzeichnis nicht gefunden Jan 07 13:36:45 the monotone db seems to be corrupt :} Jan 07 13:38:11 zecke: I've seen that error when something has been renamed, but monotone has not renamed the entry in the .mtattrs file. Jan 07 13:40:06 koen: the difference is that "andLinux" has a company behind it which can pay for the article on linuxdevices ... Jan 07 13:40:20 yeah Jan 07 13:40:32 monotone --db OE.db co --revision b30357b97c19f5314613214ef63f861fd7effd42 --branch org.openembedded.dev Jan 07 13:40:38 linuxdevices keeps misspelling gpe with a q in almost every article Jan 07 13:40:39 rwhitby: this is a fresh co Jan 07 13:41:14 zecke: it's possible for that to happen in a fresh co if the .mtattrs file is not in sync with the directory contents Jan 07 13:41:33 rwhitby: where is that .mtattrs coming from? Jan 07 13:41:41 (it happened to me yesterday when I did the big openslug rename) Jan 07 13:41:48 in the top-level of the co Jan 07 13:42:04 it stores any execute or manual_merge attributes Jan 07 13:42:18 rwhitby: isn't the mtrattrs created when I do a co? Jan 07 13:42:21 The problem is reproducible on my local db too, but only for a fresh co Jan 07 13:42:21 but it seems monotone does not modify it completely when you do directory renames Jan 07 13:42:43 rwhitby: again this is a co from a db Jan 07 13:42:50 rwhitby: no living monotone tree yet... Jan 07 13:42:53 rwhitby: monotone doesn't track dirs, only files Jan 07 13:43:10 zecke: the .mtattrs is just a special file in the db Jan 07 13:43:14 which is a huge source of all kinds of strange bugs Jan 07 13:43:37 fixed in 0.26 Jan 07 13:44:31 The .mt-attrs file has execute for packages/initscripts/initscripts-1.0/akita/keymap, but that file no longer exists Jan 07 13:44:41 bingo Jan 07 13:45:24 (sorry for the .mt-attrs misspelling if that confused anyone) Jan 07 13:45:39 rwhitby: not that part :} Jan 07 13:45:59 you can edit .mt-attrs manually and commit Jan 07 13:46:15 rwhitby: you miss the point ;) Jan 07 13:46:27 rwhitby: it is about converting the history Jan 07 13:46:52 rwhitby: anyway thanks for pointing the error out Jan 07 13:47:27 zecke: it's a serious problem, because it aborts the rest of the stuff in .mt-attrs, so the co is broken. Jan 07 13:47:44 Looks at, for example packages/initscripts/initscripts-1.0/banner, which should be execute. Jan 07 13:48:02 * koen gets remembered of the mplayer cvs mess a few years ago Jan 07 13:48:15 they edited rcs files by hand..... Jan 07 13:48:24 like you said, .mt-attrs is broken :-( Jan 07 13:48:50 koen: and with 0.26 these files stop beeing broken? Jan 07 13:49:12 zecke: that, or the migrate blows up Jan 07 13:50:43 zecke: .mt-attrs is not broken in the current head. Jan 07 13:51:01 Sounds like that revision is inconsistent Jan 07 13:51:15 zecke: you can do the tailor stuff on ewi if you want to Jan 07 13:52:08 koen: thanks. I need to get at least one conversion done :} Jan 07 13:52:38 during the last monotone experiments in june it was ~20% faster a a dual 3.1GHz xeon :) Jan 07 13:53:42 koen: well mickeyl had a dozens of mencoder instances running for sure :} Jan 07 13:53:54 heh Jan 07 13:54:01 * france is away: Away Jan 07 13:54:03 koen: tailor is IO bound Jan 07 13:54:16 will u320 scsi do? Jan 07 13:54:34 koen: ist that better than IDE UDMA133? Jan 07 13:54:56 it should be Jan 07 13:55:40 /home and /data are on seperate disks, so yo could take advantage of that Jan 07 13:57:11 koen: how far is monotone 0.26? Jan 07 13:57:44 there going to run 0.26pre1 on venge.net asap Jan 07 13:57:59 s/there/they are/ Jan 07 13:58:01 koen meant: they are going to run 0.26pre1 on venge.net asap Jan 07 13:58:14 ooooh Jan 07 13:58:26 that's cute Jan 07 13:58:44 s/cute/nice/ Jan 07 13:58:44 koen meant: that's nice Jan 07 13:58:50 ibot: botsnack Jan 07 13:58:50 koen: aw, gee Jan 07 14:03:00 s/ibot/wiseass/ Jan 07 14:03:00 koen meant: wiseass: botsnack Jan 07 14:06:34 zecke: I find it a little surprising that tailor is io bound when working with monotone Jan 07 14:07:27 although checkout out stuff doesn't need all that much checksumming Jan 07 14:09:47 hacks on monotone Jan 07 14:17:04 rwhitby: execute in the .mt-attr is there to mark the file as executable? Jan 07 14:17:50 zecke: yes Jan 07 14:18:27 BTW, that inconsistent revision is the result of a merge, probably from one branch which deleted the file and another which made it 'execute' Jan 07 14:18:45 It's difficult to merge that kind of thing... Jan 07 14:41:50 and so it begins Jan 07 14:42:52 03rpurdie 07org.oe.dev * r84cacf03... 10/ (398 files in 329 dirs): Convert CVSDATE -> SRCDATE. Also standardise cvs and svn PVs to x.x.x+cvsYYYYMMDD format and some includes some minor whitespace cleanup. Jan 07 14:43:24 * lrg|home runs for cover Jan 07 14:46:03 koen: could you install cogito and darcs? Jan 07 14:46:07 please Jan 07 14:51:20 zecke: git and darcs should be available now Jan 07 14:51:36 I also installed the build-deps for those in case you need to build a newer version Jan 07 14:54:07 thanks Jan 07 14:55:38 hmmm Jan 07 14:55:44 that's the wrong git Jan 07 14:55:51 cogito ;) Jan 07 14:56:14 git, the filemanager with GNU Interactive Tools, is now called gitfm.If you are looking for git, Linus Torvald's content tracker, installthe cogito and git-core packages and see README.Debian and git(7). Jan 07 14:57:04 Package git-core is not available, but is referred to by another package.This may mean that the package is missing, has been obsoleted, oris only available from another source Jan 07 14:58:21 koen: opps :P Jan 07 14:59:20 ah, adding unstable seems to have fixed it Jan 07 14:59:59 zecke: you should have git-core cogito git-arch git-cvs git-svn git-email gitk available now Jan 07 15:00:22 see ya people Jan 07 15:00:25 cu Jan 07 15:00:27 I've updated the StyleGuide on the OE wiki a bit - if anyone wants to make a stab at the preferred variable order, they're free to do so Jan 07 15:00:31 of to bratsilava tommorrow! Jan 07 15:00:35 have fun Jan 07 15:00:44 currently in Prague Jan 07 15:00:44 cu mithro Jan 07 15:00:49 :P Jan 07 15:02:52 Would bitbake.conf be the correct place to add ASSUME_PROVIDED entries for our prerequisite software? Jan 07 15:03:18 what is our? Jan 07 15:03:21 local.conf.sample might be a better place Jan 07 15:03:48 there's site.conf, which was always intended for system specific things Jan 07 15:03:54 throw it in build/conf/ Jan 07 15:03:54 It's not variable on a local build basis - stuff like python, bitbake, must be there. Jan 07 15:04:02 or in ~/.oe/conf or whatever Jan 07 15:04:09 right, site.conf is more appropriate Jan 07 15:04:26 Ah, I'll create that :) Jan 07 15:04:35 An ideal use for it :) Jan 07 15:04:48 I already use it for site-specific paths Jan 07 15:04:53 Along with auto.conf Jan 07 15:04:59 :) Jan 07 15:05:09 i should play with oe some, havent touched it in ageas Jan 07 15:05:11 ages, too Jan 07 15:05:12 I.e. I assume it is local to the specific site Jan 07 15:05:40 kergoth: http://linuxdevices.com/news/NS9808373276.html Jan 07 15:05:41 So: local.conf = per-user, site.conf == per.site, auto.conf == per-build Jan 07 15:06:07 Why not just add "include required.conf" in bitbake.conf? Jan 07 15:06:14 ? Jan 07 15:06:24 local.conf was always intended to be per build, not per user Jan 07 15:06:27 technically Jan 07 15:06:34 auto.conf was intended to be generated by a frontend Jan 07 15:06:40 jbowler-away: It makes a good default site.conf. If you want to have your own, feel free :) Jan 07 15:06:40 i.e. a qt or gtk2 interface Jan 07 15:07:25 kergoth: Did you get a chance to look at that patch? Jan 07 15:07:37 It'd be a good way to play with oe ;-) Jan 07 15:07:45 RP: it means we have to rewrite the NSLU2 master makefile build system if you do that. Jan 07 15:08:52 jbowler-away: or just add the entries from the default to your existing file. I don't expect it to change much again for a while Jan 07 15:09:05 The file 'site.conf' is machine generated Jan 07 15:09:23 (sort of, it actually includes something else). Jan 07 15:09:56 I.e. the master makefile works out the absolute paths, as required by bitbake, and writes them in to site.conf Jan 07 15:10:25 how about appends to the default site.conf? Jan 07 15:10:27 As kergoth said, auto.conf is auto-generated to set DISTRO and MACHINE Jan 07 15:11:40 RP: I can change it to "include ${PKGDIR}/conf/site.conf" (PKGDIR is somewhat misnamed - it's the openembedded tree) Jan 07 15:12:16 The point is that I will have to change it, whereas you could use something with a less misleading name (site-config.conf?) Jan 07 15:12:22 jbowler-away: I think that's the sensible move. These entries do belong in the default site.conf, I'm quite convinced of that Jan 07 15:12:44 What is the definition of 'site.conf' then? Jan 07 15:13:05 I assumed "configuration specific to the site", not "things the site must do" Jan 07 15:13:21 jbowler-away: Site specific config data. We provide a sensible default, the site is free to change it just like any other conf file Jan 07 15:14:00 Some systems might want bitbake to build cvs itself for example rather than provide it Jan 07 15:14:06 It seems a bit weird to provide "default site specific" data. Jan 07 15:14:36 Conventionally defaults go in bitbake.conf and are overridden by the includes at the end Jan 07 15:14:36 Surely things are either defaults, or they're site specific, but they can't be both. Jan 07 15:14:45 jbowler-away: yeah Jan 07 15:15:10 The other alternative is I put it in bitbake.conf and force them on everyone - it is a prerequisite after all... Jan 07 15:15:17 E.g., from the current bitbake.conf: Jan 07 15:15:19 TMPDIR = "${TOPDIR}/tmp" Jan 07 15:15:19 CACHE = "${TMPDIR}/cache" Jan 07 15:16:19 Overrides are certainly compatible with what the NSLU2 stuff is doing at present with auto/local/site Jan 07 15:16:33 That sounds like a better alternative. You aren't really "forcing" it on anyone, since they can always override your defaults in their own config files. Jan 07 15:16:41 such as, for example, site.conf :-} Jan 07 15:17:05 The slight nasty problem is our local.conf.sample has ASSUME_PROVIDED = in it Jan 07 15:17:26 now if I add ASSUME_PROVIDED += to site.conf... Jan 07 15:17:33 Well... it has TMPDIR in it too IIRC Jan 07 15:17:47 What does local.conf.sample say is provided? Jan 07 15:18:04 local.conf.sample became the documentation Jan 07 15:18:07 hence that stuff Jan 07 15:18:11 users wouldnt read anything else Jan 07 15:18:22 even then, they often dont Jan 07 15:18:23 heh Jan 07 15:18:25 The current one says: # ASSUME_PROVIDED = "virtual/${TARGET_PREFIX}gcc virtual/libc" Jan 07 15:18:26 isn't that stuff mostly commented out, though? Jan 07 15:18:31 yep, it is Jan 07 15:18:36 Yes, ASSUME_PROVIDED is, only DL_DIR isn't Jan 07 15:18:38 right, so that's fine Jan 07 15:18:45 pb_: 2.95.3 toolchain for zaurus users Jan 07 15:19:15 RP: even that isn't in the sample (iirc); you have to add it manually Jan 07 15:19:21 zecke: how's the converting going? Jan 07 15:21:04 I think putting this in bitbake.conf with an = is asking for chaos and isn't right as is a site parameter... Jan 07 15:21:08 koen: cd /home/zecke/scm_convert/gitside Jan 07 15:21:11 koen: git log Jan 07 15:21:31 RP: ...? Jan 07 15:21:45 and I have a hacked monotone to ignore the mt-attr issue Jan 07 15:22:26 zecke: the log looks ok Jan 07 15:22:32 pb_: A large number of users will have copied or uncommented the ASSUME_PROVIDED line in some form or another. I know its in my local.conf. The bugreports for this will be chaos. Jan 07 15:23:02 Whether or not cvs, svn, python are present or not is site dependent. Jan 07 15:23:19 Eh... site.conf is included before local.conf Jan 07 15:23:24 heh. well, without python you will be pretty stuffed. Jan 07 15:23:36 if a user uncomments lines in local.conf without knowing what they're doing, they're going to have issues in the future regardless :P Jan 07 15:23:38 I.e. ASSUME_PROVIDED in local.conf will zap anything in site.conf or auto.conf Jan 07 15:23:46 jbowler-away: yes, so local.conf can override it Jan 07 15:23:52 jbowler-away: That's correct Jan 07 15:24:11 kergoth: but we could do without explaining that 100 times over Jan 07 15:24:32 That doesn't really seem any worse than any of the other incompatible changes that are going on. Users with, say, CVSDATE in their local.conf are going to lose just as badly. Jan 07 15:25:19 RP: if I have an existing, working, ASSUME_PROVIDED, and stuff gets changed to need a default, why isn't my local.conf ASSUME_PROVIDED going to require a change? Jan 07 15:25:52 local should always override site. by definition, a build is more specific than the system it's on Jan 07 15:25:57 heh Jan 07 15:25:59 Quite apart from that, it doesn't matter if it is in site.conf or bitbake.conf - they both happen first. Jan 07 15:26:00 jbowler-away: if I use += in the site.conf, it will just append Jan 07 15:26:10 No it won't Jan 07 15:26:38 RP: no, it won't. It will still get zapped by local.conf. Jan 07 15:26:45 local.conf is after site.conf - that overrides doesn't it? You would need ASSUME_PROVIDED_append Jan 07 15:26:58 yes, that is flawed reasoning :) Jan 07 15:27:03 right, and ASSUME_PROVIDED_append would be ghastly. Jan 07 15:27:17 all _appends are ghastly Jan 07 15:27:31 well, yeah, but this one would be particularly so Jan 07 15:27:37 ;-) Jan 07 15:27:50 ok, so we're going to break things. It is a site decision so it still belongs in site.conf though :) Jan 07 15:27:55 _append.. guh.. there's no way to prevent that from occurring farther down the line Jan 07 15:27:58 Anything in 'site.conf' can also be put in bitbake.conf. Jan 07 15:27:59 evil Jan 07 15:28:38 RP: do you know how git rev-list works? Jan 07 15:28:43 all that said, it seems fairly clear that the fundamental problem is that the ASSUME_PROVIDED thing is a broken design. Jan 07 15:29:09 so, probably, the right answer is just to do the same as we did for PREFERRED_PROVIDER, and break it apart. Jan 07 15:29:10 ? maybe - it's not a list I guess. gcc is either provided or not Jan 07 15:29:31 pb_: Broken in what way? Jan 07 15:30:00 zecke: I might have done at one point. I'd have to stare at git for a while again before remembering though :-/ Jan 07 15:30:15 The git fetcher encompasses most of my git knowledge ;-) Jan 07 15:30:23 RP: in that, like the original PREFERRED_PROVIDERS, you have this one variable with a pile of stuff in, and it is very hard to manipulate in a useful way. Jan 07 15:30:47 so, for example, if you set a default in bitbake.conf, it is hard for a site to remove one particular thing from it. Jan 07 15:31:15 (in absence of the -= operator which, iirc, bitbake doesn't have yet) Jan 07 15:31:45 likewise, as you observed, careless assignments in local.conf will blow away all the defaults. Jan 07 15:32:14 So what's the alternative? Jan 07 15:32:36 ASSUME_PROVIDED_kergoth = "you bet" Jan 07 15:33:27 Setting a default in bitbake.conf (like TMPDIR) and commenting out an override in local.conf.sample seems fine to me, if people want to override in site.conf that's fine too. Jan 07 15:33:59 pb_: This would mean we'd have to have a way to unset a variable or are you thinking of using true/false values or something? Jan 07 15:34:43 yeah, that would be fine. You could signify "not provided" by setting it to the empty string, or "hell no", or some such value. Jan 07 15:35:48 and this of course is going to need more changes to bitbake although it is a good time to make them Jan 07 15:35:55 There's plenty of precedent for variables that are empty for false and contain something else (usually "1") for true. Jan 07 15:36:08 PREFERRED_PROVIDER_kergoth = "", or PREFERRED_PROVIDER_kergoth = "site" Jan 07 15:36:11 I know, I just added one :) Jan 07 15:36:15 admittedly there's slightly less precedent for "you bet" and "hell no", but I don't regard that as an obstacle. Jan 07 15:36:37 You would if I suggested it ;-) Jan 07 15:36:40 I.e. ASSUME_PROVIDED is not necessary if PREFERRED_PROVIDER can say "it's there already" Jan 07 15:37:23 thats ugly, pollutes the namespace. what if i have a package providing site? Jan 07 15:37:35 jbowler-away: I don't think overloading PREFERRED_PROVIDER like that is a good idea. Jan 07 15:37:37 * kergoth gets caffeine Jan 07 15:37:45 it doesn't really buy anything, and it's a bit confusing Jan 07 15:38:06 kergoth: that's what the first form meant (i.e. empty string means no provider required) Jan 07 15:38:22 pb_: it's not overloading it - the site (or build system) is providing the package Jan 07 15:38:50 yes, and you're polluting the namespace. what happens if i create site-1.0.bb? Jan 07 15:38:58 its ugly and gains you nothing, as pb said Jan 07 15:39:00 Right, bad idea. Jan 07 15:39:12 restricting the names due to a "magic" value is silly Jan 07 15:39:18 I think we stick with ASSUME_PROVIDED as its more obvious what's going on. For now I'll add the appropriate string to bitbake.conf Jan 07 15:39:52 If someone wants to tweak bitbake to make ASSUME_PROVIDED work better. I'm all for it but there are more important patches I'd like to see in first Jan 07 15:40:33 RP: don't forget to highlight the changes on the mailinglist Jan 07 15:41:05 kergoth: the problem is that in PREFERRED_PROVIDER_foo = "bar" foo is, effectively, virtual whereas bar is either virtual or a real package Jan 07 15:41:46 how exactly is this a problem? Jan 07 15:41:52 I don't think bar is allowed to be virtual there, actually. **** BEGIN LOGGING AT Sat Jan 07 16:04:11 2006 Jan 07 16:04:23 pb_: Especially if you share the deploy directory between builds even just locally Jan 07 16:13:42 03rpurdie 07org.oe.dev * r46d1a20c... 10/conf/bitbake.conf: Add ASSUME_PROVIDED line for prerequisite software to bitbake.conf. People may like to check their local.conf files append to ASSUME_PROVIDED... Jan 07 16:13:44 if this all works out we could even release OE 1.0 Jan 07 16:13:46 03rpurdie 07org.oe.dev * r68a67db6... 10/conf/local.conf.sample: Update local.conf.sample to append to ASSUME_PROVIDED as a better example of usage. Jan 07 16:13:50 03rpurdie 07org.oe.dev * rbb0ef2cf... 10/packages/ipkg-utils/ (ipkg-utils-native_1.6cvs.bb ipkg-utils_1.6cvs.bb): ipkg-utils: Correct the dependencies and specifically set PACKAGES = for ipkg-utils-native. Jan 07 16:13:55 03rpurdie 07org.oe.dev * r451d8c07... 10/classes/image_ipk.bbclass: image_ipk.bbclass: Set BUILD_ALL_DEPS = "1" for images in readyness of bitbake updates Jan 07 16:13:56 and pimp it at the appropriate news channels Jan 07 16:14:29 RP: i would apply your change right away?! and do a release of bitbake Jan 07 16:14:46 zecke: I'd like that :) Jan 07 16:15:09 * RP can break more things then :) Jan 07 16:15:19 zecke: http://ewi546.ewi.utwente.nl/rrd/graph/cpu_1_7200.png Jan 07 16:15:54 koen: not even 50%... Jan 07 16:16:12 indeed Jan 07 16:17:07 when does it estimate to have finished? Jan 07 16:17:29 koen: one or two days Jan 07 16:17:30 zecke: and can it do incremental adds? Jan 07 16:17:41 koen: yes there is a state file Jan 07 16:17:50 koen: this theoretically allows incremental adds Jan 07 16:18:02 koen: and you even ask it to start at a specific revision Jan 07 16:18:34 i saw maemo in distro, is the nokia 770 suite based on oe ? Jan 07 16:18:44 or have i mixed this up altogtehr! Jan 07 16:19:16 dan2003: it is not based on it (yet) Jan 07 16:20:33 k, to build for simpad i should still specify opensimapd dor distro or not? Jan 07 16:20:46 or familiar Jan 07 16:21:59 and i get th eline about unversiond distro buiiding, should i coment out the line that genrate sthe warning? i should i be specifying a version? Jan 07 16:23:18 you should be specifying a version Jan 07 16:24:38 k, which leads me on to th next question ;) in conf/distro there is only opensimapd-0.9.0 yet florians release calls itself 0.8.3-rc1.. so wot version should i say to build ? Jan 07 16:25:06 ahh. i just saw familiar 0.8.3, i quess its that then Jan 07 16:25:59 Why is multimachine in classes? Jan 07 16:26:22 Should it not be conf/includes or something? Jan 07 16:26:56 zecke: how can I check in opensolaris how many cpus it thinks it has? Jan 07 16:27:10 probably. people got in the habit of doing everything as a class, even things that arent for a .bb, back before .inc existed Jan 07 16:27:48 koen: dunno :} Jan 07 16:28:00 kergoth: This has a .inc extension so its very confused... Jan 07 16:28:12 * zecke just found another error in or OE.db Jan 07 16:28:33 zecke: alvaro talked me into trying opensolaris and nexenta and I wonder if it would do the right thing Jan 07 16:29:24 koen: Solaris should cope with two AMD64 cpu's Jan 07 16:29:35 oops I think I killed ewi Jan 07 16:29:44 I only have those pauper amd32 cpus at home Jan 07 16:30:39 I wonder how you killed it Jan 07 16:30:49 make -j NUMBER Jan 07 16:31:21 in theory that should just lock it solid for a long time Jan 07 16:31:31 right Jan 07 16:31:40 we'll see in the morning :) Jan 07 16:32:20 hmm, some weird message whilst processing the bb files http://pastebin.com/495641 Jan 07 16:32:37 rm tmp/cache -rf Jan 07 16:32:43 k Jan 07 16:36:07 im writing a music player/jukbox app, and i want to keeo the frontend remote to the decoding side, so it can use it both as a music player on zaurus, but also as a media player for the lving room where the music plays on a headless machien (epia) and the controls are on my zaurs/simpad etc.. can anybody coment on best ways to transfer the playlist info between the two? ill be using tcp sockets i think.. and will tell the playback s Jan 07 16:36:07 ide which track to play by sending the id of a track in a sqlite db Jan 07 16:36:58 i have tables for albums, artists etc so selctions can be bemade by artists/album/genre etc Jan 07 16:37:27 xml, or would that be overkill? Jan 07 16:37:44 if you are lazy you could just send over the sql query :) Jan 07 16:37:55 but am trying to decide on the best methid of keeping the fornt end in sync with wots on the player.. Jan 07 16:37:56 hehe Jan 07 16:38:41 i wish to be able t add tracks to the collection as well, and not have to send the whole list of tracks back to the fron end each time.. im eally not sure wot to do! maybe transfer the sqlite database file ? Jan 07 16:39:08 think xml would be quite resourse intensive? Jan 07 16:43:51 'night all Jan 07 16:44:03 koen|sleep: sleep well Jan 07 16:44:23 night Jan 07 16:50:17 wot causes this? | checking target system type... Invalid configuration `arm-INVALID': system `INVALID' not recognized Jan 07 16:50:45 dan2003: local.conf ;) TARGET_OS needs to be set Jan 07 16:51:40 k, ill give that ago, i build a zaurus image, and chnage the distro to familiar-0.8.3 and machien to simpad, to try and build for it.. and that sprung up ! Jan 07 16:51:57 'night koen|sleep Jan 07 16:52:56 dan2003: OZ sets TARGET_OS to linux... but see local.conf.sample Jan 07 16:53:09 ewi lives again Jan 07 16:53:28 k, Jan 07 17:29:40 zecke: I think monotone.vanille is sulking Jan 07 17:32:04 RP: should I kill it? Jan 07 17:38:48 zecke: I suspect it needs restarting, yes Jan 07 17:39:10 RP: okay let me get the password from my old laptop Jan 07 17:47:29 Can I build a snapshot release of gpe using OE ? Jan 07 17:51:39 paxl_: I think so Jan 07 17:51:48 RP: grepping now through gigabytes of emails Jan 07 17:53:19 zecke: Its not just me who has an archive problem then? :) (thanks) Jan 07 17:57:41 hmm it won't let me log in Jan 07 17:58:17 zecke: ah :-/ Jan 07 17:58:43 RP: it is running (it is serving webpages) but it doesn't want my pw Jan 07 17:59:08 zecke: Perhaps koen can sort it tomorrow then... Jan 07 17:59:52 maybe I should try telnet instead ssh Jan 07 18:00:26 heh Jan 07 18:01:33 RP: sorry it won't let me in Jan 07 18:01:42 zecke: wouldn't recommend that unless you're using kerberos for authentication ;) Jan 07 18:01:50 zecke: ok, np. thanks for trying Jan 07 18:02:52 reenoo_: hehe Jan 07 18:03:46 I've set my autobuilder off against the current changes so tomorrow, I'll know how much I've broken :) Jan 07 18:06:00 anyways. 'night all Jan 07 18:06:42 'night reenoo_ Jan 07 18:06:55 RP: git rev-list --all | wc -l Jan 07 18:06:55 171 Jan 07 18:07:28 RP: now to your latest bitbake patch. Should I apply it now. Wait for some feedback and release a new version of bitbake? Jan 07 18:08:57 zecke: I think that would be the best approach Jan 07 18:09:23 RP: okay. Do you have something newer than revision two? Jan 07 18:09:38 zecke: No, that's the latest Jan 07 18:15:15 Time for me to sleep. Too many late nights recently :-/ Jan 07 18:15:19 'night all Jan 07 18:15:36 hehe sleep well Jan 07 18:56:05 good nite Jan 07 20:28:23 gpe-dm doesn't seem to work on latest openzaurus-unstable :\ Jan 07 22:44:16 is www.oesources.org down? Jan 07 22:46:19 seems so or the search spiders hit it all at once like they did hh.org Jan 07 22:46:57 at this rate search engines/spiders are going to become a DOS method Jan 07 22:47:42 robots.txt? Jan 07 22:47:58 yes.. that would help... disallow:all :D Jan 07 22:48:07 nah, I was thinking it might allow rate limits Jan 07 22:48:36 not all spiders actually respect the robots.txt file Jan 07 22:48:42 most ignore it now Jan 07 22:48:50 most don't ignore it, last I checked Jan 07 22:49:28 how do you think google doubled it's url/search list overnight? Jan 07 22:49:47 why does Google lack indexing of certain webpages? Jan 07 22:50:10 i think it respects metatags that disallow robots Jan 07 22:50:44 hard to say tho Jan 07 22:50:47 if a company ignores robots.txt, then they can be guilty of copyright infringement, since robots.txt is seen as explicit denial of the copying/indexing Jan 07 22:51:18 anyway... Jan 07 22:51:27 64.68.82.164 - - [05/Sep/2004:13:06:30 +0000] "GET /robots.txt HTTP/1.0" 404 2002 "-" "Googlebot/2.1 (+http://www.google.com/bot.html)" Jan 07 22:51:39 i think the only one that holds in court is if you actually use the digital copyright certificate Jan 07 22:52:04 technically, copyright is automatic and you don't need to do anything to enforce it Jan 07 22:52:25 yes and no Jan 07 22:52:33 you have to prove you had it first Jan 07 22:52:37 only by the standard of robots.txt does implicit approval exist Jan 07 22:52:44 timestamps are not a valid method Jan 07 22:53:11 no, you only need to prove you created it Jan 07 22:55:14 http://www.gocopyright.com/ Jan 07 22:55:28 is one of many companies that esist now to prove copyrights Jan 07 22:55:32 exist* Jan 07 22:56:26 " Your copyright must be registered in order to take legal action against an act of infringement. " <-- false Jan 07 22:57:14 Copyright exists immediately when an original work is created and fixed into a writing, tape, or electronic medium. However, while legally correct, the right immediately created through original creation does not entitle you to bring legal action for infringement. Jan 07 22:57:59 its not exactly false Jan 07 22:58:14 you have to prove somehow that you had it first Jan 07 22:58:26 with written media you can register a letter to yourself Jan 07 22:58:36 you cannot do that with digital media Jan 07 22:59:48 anyway this is moot since we were talking about spiders and not actual copyright Jan 07 23:00:56 aha finally Jan 07 23:01:03 image reflashed Jan 07 23:02:06 what outputs libcairo2? Jan 07 23:05:20 cairo2 perhapse Jan 07 23:05:31 no such package Jan 07 23:08:43 cairo_cvs.bb maybe Jan 07 23:13:03 hrm Jan 07 23:13:10 why does bitbake insist on building cairo 0.5.2? Jan 07 23:14:15 thats the default for your distro Jan 07 23:15:38 I don't see any such default for openzaurus-unstable Jan 07 23:16:02 aha, there it is Jan 07 23:16:07 check your wm choice Jan 07 23:16:38 that leaves the question of where to get libcairo2 from again Jan 07 23:18:05 what do you need ciro2 for anyway? Jan 07 23:18:09 cairo* Jan 07 23:18:11 gpe-login Jan 07 23:18:25 ... Jan 07 23:18:27 ... Jan 07 23:18:41 * emte looks at my feed list Jan 07 23:18:53 "gpe-login: error while loading shared libraries: libcairo.so.2: cannot open shared object file: No such file or directory" Jan 07 23:19:09 wait! Jan 07 23:19:11 bad you Jan 07 23:19:13 ... Jan 07 23:19:24 what?\ Jan 07 23:19:25 libcairo.so.2 is NOT libcairo2 Jan 07 23:19:34 yes it is... o.O Jan 07 23:19:43 no, not at all Jan 07 23:19:44 or at least, that's what ipkg removed to break me Jan 07 23:19:59 libcairo.so.2 comes from cairo Jan 07 23:20:05 not 0.5.2 Jan 07 23:20:37 and 'cairo' outputs libcairo1 Jan 07 23:20:41 i cant say which one makes it Jan 07 23:21:02 yes but do you understand binary naming? Jan 07 23:21:19 not totally, but I was obviously referring to the binary package by libcairo2 Jan 07 23:21:25 libcairo.so is not libcairo2.so Jan 07 23:21:28 if I knew the bb package for it, I wouldn't be searching ;) Jan 07 23:21:35 libcairo2-...ipk Jan 07 23:21:43 libcairo.so.2 is not libcairo2.so.2 Jan 07 23:22:07 libcairo1_0.5.2-r0_armv5te.ipk Jan 07 23:22:20 I need libcairo2_???_armv5te.ipk Jan 07 23:22:25 NO Jan 07 23:22:36 you need nothing to do with cairo2 Jan 07 23:22:40 ....... Jan 07 23:22:50 then why did 'ipkg remove libcairo2' give the error? Jan 07 23:22:59 because it doesnt exist Jan 07 23:23:06 no, it removed Jan 07 23:23:10 and now gpe-login errors Jan 07 23:23:18 ipkg remove libcairo Jan 07 23:23:19 or Jan 07 23:23:23 ipkg remove libcairo.so.2 Jan 07 23:23:23 there is no libcairo Jan 07 23:23:24 i get an ugly python error when i use bitbake, something related to md5sum... anyone can help me with this ? Jan 07 23:23:26 would work Jan 07 23:23:44 'ipkg remove libcairo2' worked and caused gpe-login to break Jan 07 23:24:20 libcairo1_0.5.2-r0_armv5te.ipk lacks any .so.2 Jan 07 23:24:53 did you check libsvg-cairo ? Jan 07 23:25:32 yes, I installed that too Jan 07 23:25:48 i find it strange you need cairo at all tho Jan 07 23:25:56 oh? Jan 07 23:26:02 its not showing up on fam 0.8.3 Jan 07 23:26:07 GPE isn't supposed to require cairo? Jan 07 23:26:21 i am not seeing it in my pkg list Jan 07 23:26:27 O.o Jan 07 23:26:45 note the pkgs are all libcairo* Jan 07 23:27:38 mine doesnt list cairo Jan 07 23:27:48 is yours public? Jan 07 23:28:04 gtk, glib, pang, and x11 Jan 07 23:28:24 ? Jan 07 23:28:35 its the latest familiar 0.8.3 gpe image ... Jan 07 23:29:48 libs? Jan 07 23:32:36 http://familiar.handhelds.org/releases/v0.8.3/ Jan 07 23:32:37 bit slow Jan 07 23:38:30 'bit'? Jan 07 23:41:23 Packaged contents of cairo into /home/luke-jr/src/oe2005/build.alt/tmp/deploy/ipk/libcairo2_1.0.2-r0_armv5te.ipk Jan 07 23:43:23 world/libcairo1_0.5.2-r0_armv5te.ipk Jan 07 23:43:25 on your feed Jan 07 23:46:06 its not ON my device tho Jan 07 23:47:55 root@h3600:~# ipkg info gpe-login Jan 07 23:47:55 Package: gpe-login Jan 07 23:47:55 Version: 0.83-r1 Jan 07 23:47:55 Depends: xkbd, gpe-login-locale*, libgpewidget1 (>= 0.105), gtk+ (>= 2.6.10),)Provides: gpe-session-starter Jan 07 23:47:55 Status: install ok installed Jan 07 23:47:56 Architecture: arm Jan 07 23:48:41 root@h3600:~# ipkg whatdepends -A gpe-login Jan 07 23:48:41 Root set: Jan 07 23:48:41 gpe-login Jan 07 23:48:41 What depends on root set Jan 07 23:48:41 gpe-conf 0.1.27-r0 depends on gpe-login Jan 07 23:48:46 gpe-task-base 1.0-r42 depends on gpe-login Jan 07 23:48:48 gpe-task-settings 1.0-r42 depends on gpe-conf Jan 07 23:48:50 suspend-desktop 1.0-r0 depends on gpe-conf Jan 07 23:48:52 whoops Jan 07 23:48:54 wrong direction Jan 07 23:49:44 weird Jan 07 23:49:58 should I clean cairo and rebuild gpe-login? =p Jan 07 23:50:24 * emte shrugges Jan 07 23:51:24 i do have cairo Jan 07 23:51:31 but not for gpelogin Jan 07 23:51:34 root@h3600:~# ipkg list |grep cairo Jan 07 23:51:34 libcairo1 - 0.5.2-r0 - Jan 07 23:51:34 libsvg-cairo1 - 0.0cvs20050601-r0 - Jan 07 23:51:54 well, missing deps hardly mean much other than a bug report ;) Jan 07 23:52:16 aha Jan 07 23:52:24 gpe-bootsplash 1.14-r5 depends on libcairo1 (>= 0.5.2) Jan 07 23:53:14 curious... Jan 07 23:53:25 cairo isn't in gpe-login src Jan 07 23:53:34 yet the segfault appears in libcairo... Jan 07 23:53:50 gpe-task-base 1.0-r42 depends on gpe-bootsplash Jan 07 23:54:02 gpe-task-base 1.0-r42 depends on gpe-login Jan 07 23:54:27 so your problem is with gpe-task-base Jan 07 23:54:33 not gpe-login Jan 07 23:56:23 gpe-login segfaults Jan 07 23:56:30 gdb shows the segfault in libcairo Jan 08 00:46:49 emte: gtk+ depends on libcairo2, it seems Jan 08 00:46:58 as does pango Jan 08 00:48:49 Morning all Jan 08 00:49:54 Installed fresh opie-image (.dev pull from yesterday morning) on my CL860. Have sound any more! Any info's ? Jan 08 00:51:39 uv1: learn English Jan 08 00:51:47 do you or do you not have sound? Jan 08 00:55:07 Luke-Jr, gtk2 does Jan 08 00:55:12 not 1 tho Jan 08 00:55:25 actually i should correct that Jan 08 00:55:26 Luke-Jr: no sound, its still too early to write full sentence ! Jan 08 00:55:34 gtk2 supports cario Jan 08 00:57:17 emte: my gtk+ packages are depending on libcairo2 Jan 08 00:57:24 and nobody uses gtk1... Jan 08 00:58:31 i've no listing for libcairo2 anywhere Jan 08 00:58:52 i dont think i even have a listing for gtk2 Jan 08 00:59:21 nope i lied Jan 08 00:59:26 it lists gtk2 Jan 08 00:59:33 but not cairo2 Jan 08 01:00:44 I can't figure out where the name libcairo2 is coming from, either Jan 08 01:01:51 aside from you insisting it exists, i dont either Jan 08 01:03:06 as i showed you above: libcairo1 - 0.5.2-r0 - Jan 08 01:03:30 is the only cairo that exists in this image Jan 08 01:06:03 is there a simple way to list what BitBake considers as runtime dependencies? Jan 08 01:06:13 and/or why it considers them? Jan 08 01:09:50 * Luke-Jr wipes his tmp dir Jan 08 01:10:35 let's see libcairo2 survive that Jan 08 01:10:37 =p Jan 08 01:14:16 lol Jan 08 01:14:35 the ones i listed were working runtime deps Jan 08 01:15:06 I know Jan 08 01:15:19 but if it never gets built, nothing can depend on it, right? ;) Jan 08 01:21:07 correct Jan 08 02:03:59 You guys probably noticed bugs.openembedded.org is gone atm Jan 08 02:04:06 I uh fucked something up :) Jan 08 02:04:48 but it's one of two things Jan 08 02:04:58 and both would be fairly easy to fix just as soon as I get there Jan 08 02:05:02 love and devotion? Jan 08 02:05:09 :) Jan 08 02:05:13 either I installed a broken kernel Jan 08 02:05:20 or I firewalled off the outside world Jan 08 02:05:53 * emte votes for the second option Jan 08 02:06:06 well I installed a new kernel right after making the iptables changes Jan 08 02:06:11 * NAbyss votes for the former Jan 08 02:06:12 so I don't entirely know which it i Jan 08 02:06:13 is Jan 08 02:06:25 dang namit, spamers have gotten a new trick Jan 08 02:06:35 treke: No serial console access via ssh to another box? Jan 08 02:06:43 nope. Jan 08 02:07:06 Smeg. Jan 08 02:07:17 mithro, you talking about the IE/outlook issue? Jan 08 02:07:21 The place is closed up for the weekend Jan 08 02:07:33 so monday I can get someone to take a look at it Jan 08 02:07:53 they send you an email which could be perfect legitimate "re: Going out tonight, Hi, I'll swing by and pick you up after work" and then have the spam as the email you replied tooo Jan 08 02:08:01 which means bayes gets very confused Jan 08 02:08:16 Interesting.. Jan 08 02:09:17 because the first part looks very much like a real email (infact I have legitimate emails which are almost identical) but the second part looks like spam Jan 08 02:09:19 ah Jan 08 02:09:47 i thought you were talking about MS's newest auto-update problem Jan 08 02:10:19 where it exicutes and installes thinsg that look like updates without asking Jan 08 02:11:04 hmm Jan 08 02:11:11 * emte needs a spelling lesson Jan 08 02:11:22 i don't use windows - those problems don't worry me Jan 08 02:11:28 anyway I have to go! Jan 08 02:13:46 spamers are using it to instal trojan mailers Jan 08 02:13:51 emte: I gave up on keeping local mail long ago. Currently use gmail as the Inbox 'Preview' shows the first few chars of the email after the subject, so I don't have to look at it to immediately know its spam. ;) Jan 08 02:14:06 lol Jan 08 02:14:32 yeah its more that you can get an email or visit a website and it will install the virus/torjan/worm Jan 08 02:14:39 trojan* Jan 08 02:17:35 I'd love to drop into pure Linux (or perhaps Solaris 10 or somesuch), but there are varying issues with requirements in connecting and using the network at work. Such as Exchange. Jan 08 02:18:45 yeah... the open version isnt quite ready yet Jan 08 02:19:04 There's also game support to consider. Jan 08 02:19:15 depends on the game i suppose Jan 08 02:19:25 In my case, Uru. Jan 08 02:21:20 Alan_K, you play with Brutus yet? Jan 08 02:21:24 http://www.omesc.com/modules/main_module/ Jan 08 02:22:12 That won't help. Corporate requirement for a Windows environment. Jan 08 02:22:27 ah Jan 08 02:22:38 wasnt sure if you had remote access from home Jan 08 02:22:54 Sort of. Using their webmail. Which is not too bad. Jan 08 02:23:46 better than noting anyway :) Jan 08 02:23:49 nothing* Jan 08 02:24:26 I'm a Solaris admin by experience. Just now starting to poke more into Linux, since there are a few I'm needing to support. Jan 08 02:25:11 Plus I'm experimenting with ultra-low-power systems for anything persistent/100% uptime at home. Hence the Slug. ;) Jan 08 02:25:14 ah Jan 08 02:25:44 i dont mind using solaris, but not really my favorite by any means Jan 08 02:26:35 Well, I started with the old SunOS 4.1.3, so the current gen Solaris like 8 and 10 are several orders of magnitude better. Jan 08 02:27:10 There's also that ZFS thing that will be an open-source contender to Veritas. Jan 08 02:27:31 hmm Jan 08 02:28:10 ZFS is still in the works. They just got a system booting off an all-ZFS system in dev, and it was quite a fight. Jan 08 02:29:36 i would imagine Jan 08 02:30:15 Hmm, glibc-internediate does not build any more (i386). Configure stops (because 4.x compiler). If this is fixed, the build process stops. Any idea ? Jan 08 02:30:23 the first is always the hardest Jan 08 02:30:41 dont use gcc 4.x Jan 08 02:30:48 uv1: There's a patch in place to get around that Jan 08 02:31:05 NAbyss: Where and since when ? Jan 08 02:31:15 do any images even build with gcc 4 yet? Jan 08 02:31:18 http://zoo.weinigel.se/n30/downloads/glibc.patch Jan 08 02:31:23 emte: I'm using gcc4 here Jan 08 02:31:34 uv1: Not sure how long.. only picked up OE last week here Jan 08 02:31:36 you've built a full working image? Jan 08 02:31:37 emte: opie-image does (pxa) Jan 08 02:31:51 emte: Yes, opie-kdepim-image Jan 08 02:31:57 It's running on my zaurus at present Jan 08 02:32:16 wow Jan 08 02:32:25 i am supprised Jan 08 02:32:43 someoen must have faught a lot with it Jan 08 02:32:51 fought* Jan 08 02:34:26 I thought RP had checked in my glibc patch Jan 08 02:35:41 might depend on where your using the db from Jan 08 02:36:15 ah, RP only checked it in on the .dev branch. Jan 08 02:36:37 i was thinking time, but branch would matter too Jan 08 02:36:52 anyway sleep :P Jan 08 02:38:41 Strange, bitbake wants to build "glibc-intermediate-2.3.2+cvs20040726". Thats wrong I suppose ? Jan 08 02:38:51 Strange, bitbake wants to build "glibc-intermediate-2.3.2+cvs20040726". Thats wrong I suppose ? Jan 08 02:39:01 Sorry 2 times ... Jan 08 02:39:28 wingel: Thats ok, I am working on .dev. Jan 08 02:56:55 03rpurdie 07org.oe.dev * r72409a63... 10/conf/machine/ (71 files in 2 dirs): Move machine include files (files that are not machines in their own right) into conf/machine/include. Jan 08 02:57:00 03rpurdie 07org.oe.dev * r5f40addf... 10/ (4 files in 3 dirs): Move multimachine.inc to conf/distro/include/multimachine.conf and include it by default for familiar/openzaurus Jan 08 02:57:05 03rpurdie 07org.oe.dev * r61486dc9... 10/ (7 files in 5 dirs): Set some packages/classes PACKAGE_ARCH correctly and also correct some PACKAGE_ARCH= to . Jan 08 02:57:09 03rpurdie 07org.oe.dev * ra39e3722... 10/packages/gchub/gchub-module_svn.bb: gchub-module: fix PV typo **** ENDING LOGGING AT Sun Jan 08 02:59:56 2006