**** BEGIN LOGGING AT Fri Sep 08 02:59:56 2006 Sep 08 06:09:09 is there anybody slove this problem ? http://rafb.net/paste/results/yLUXHA13.html Sep 08 06:09:21 X11/Xregion.h No such file or directory ? Sep 08 06:25:58 grr, in a twisted maze of python code Sep 08 06:29:36 How's openhand? Sep 08 06:31:29 not bad. getting things done, though for various reasons i'm not getting things done as quickly as i'd like :\ Sep 08 06:31:46 once again seriously considering leaving the field. starting to dread coding Sep 08 06:31:48 heh Sep 08 06:33:20 ah Sep 08 06:33:24 Maybe you should Sep 08 06:34:17 You know what might interest you? Sep 08 06:34:30 Ever considered pbx work? Sep 08 06:36:10 like asterisk type stuff? Sep 08 06:36:31 hi, what is the simplest way to diff the head version of a file with an older version (and of course i know the PR I want to comapre with, but not the mtn id) Sep 08 06:36:50 * kergoth dunnos, is still a noob when it comes to monotone :\ Sep 08 06:37:00 kergoth: Well among other systems. Sep 08 06:37:04 ggilbert: whatve you been up to lately? haven't seen you around much Sep 08 06:37:27 I bought a treo and haven't turned my zaurus on in a year :p Sep 08 06:37:29 not a bad idea, i'll have to think about that. i remember learning t1/t3 and frame relay technologies when i was at digi Sep 08 06:38:18 Still working at the school Sep 08 06:38:33 otherwise generally unproductive Sep 08 06:39:36 actually need to wander off to sleep Sep 08 06:39:39 so later Sep 08 06:40:48 k, night Sep 08 07:12:18 ;/ahtx Sep 08 07:13:15 morning all Sep 08 07:21:15 morning Sep 08 07:22:16 RP: hi Sep 08 07:35:00 mornung Sep 08 07:35:08 RP: any luck with debuggng? Sep 08 07:37:00 hey Sep 08 07:38:57 if no one has any objections, i'm going to pick up maintainance of the toolchain mess and start cleaning it up Sep 08 07:39:04 sheesh, they still have gb2 as the maintainer Sep 08 07:39:18 kergoth: that would be most cool Sep 08 07:40:03 zecke: you have any advice on running automated oe builds? this box is fast enough that i'd like to always have at least one build going to check for regressions when i make changes Sep 08 07:40:24 think i'll set up a seperate chroot for them Sep 08 07:45:41 good morning all Sep 08 07:46:23 hi, folks. http://rafb.net/paste/results/f7BNkm32.html Sep 08 07:46:24 kergoth: just make sure that box reports to the tinderbox :) Sep 08 07:46:44 is there anybody who sloved this ? Sep 08 07:48:17 koen: :) Sep 08 07:48:34 * mickey|working waves Sep 08 07:48:35 morning folks Sep 08 07:48:44 hey mickey|working Sep 08 07:48:46 it likely wont be stock oe though, i think i'll have it building a specific local testing tree with whatever i'm working on Sep 08 07:49:17 * kergoth kicks python's handling of unicode Sep 08 07:49:36 RP: i was pleasantly surprised to see overlay in the latest cxx00 kernel patch. Congrats! How far is it from mailine? Sep 08 07:50:23 python's handling of unicode annoys me, but i guess its alright once you get a handle on it Sep 08 07:50:31 kergoth: *nod* that's it Sep 08 07:50:46 mickey|working: hey! Sep 08 07:50:51 mickey|working: you were in berlin? Sep 08 07:50:54 no no Sep 08 07:50:57 i would have called Sep 08 07:50:58 mickey|working: ;) Sep 08 07:51:04 the berlin guy was here instead Sep 08 07:51:12 i had severe problems with a stiff neck Sep 08 07:51:14 i couldn't travel Sep 08 07:51:18 mickey|working: python unicode, how does that work? Sep 08 07:51:28 mickey|working: I never was able to mix QString and python :( Sep 08 07:51:41 zecke: it's just prefixing u" and then messing with string() and unicode() constructors Sep 08 07:51:48 its pretty painless, you just convert your regular python string into a unicode object, then encode it into whatever else you need from there Sep 08 07:52:01 u"\u0123".encode('utf-8') Sep 08 07:52:11 or whatever Sep 08 07:52:22 afaict anyway Sep 08 07:53:38 mickey|working: http://www.kdedevelopers.org/node/2334 Sep 08 07:53:44 haha Opie is just ahead of time Sep 08 07:53:53 morning Sep 08 07:53:56 zecke: heh! :)) Sep 08 07:53:59 hey hrw Sep 08 07:54:23 mickey|working: save 700 bucks, get a greenphone release the source code :) Sep 08 07:54:43 ljp: Am I right that with the greenphone you are going to release Qtopia Phone as GPL? Sep 08 07:55:12 zecke: Nah -- I rather work on this phone project here and then release the source :D Sep 08 07:57:00 hehe Sep 08 07:57:17 sadly Qtopia is just another proprietary system :( Sep 08 07:59:39 everything that happened with qtopia and opie saddens me Sep 08 07:59:41 such wasted potential Sep 08 08:05:04 kergoth: hey, on the hand you are absolutely right Sep 08 08:05:04 kergoth: on the other hand, I would not say wasted. We have not managed to stay relevant but personally I have learned a lot Sep 08 08:05:04 on hardware, software (on all levels), design, leading, planning Sep 08 08:05:04 which I assume is a big win for my next employer, which TT never cared to be Sep 08 08:05:26 that's right Sep 08 08:05:33 it was a great experience Sep 08 08:05:45 albeit a bit wasted at the same time Sep 08 08:05:46 zecke: that's a good point Sep 08 08:05:57 i learned about qt, which is a big part of what got me the job at CSC Sep 08 08:06:16 (though i still dont have the whole model-view-controller aspect of qt4 down yet :P) Sep 08 08:06:25 kergoth: e.g. thanks to you I know the Linux keyboard legacy crap, tslib Sep 08 08:06:29 CSC? Sep 08 08:06:38 computer sciences corporation, the folks that were doing the land warrior stuff Sep 08 08:07:00 ah right Sep 08 08:07:34 found a bug in the non-binary glibc locale handling btw Sep 08 08:07:37 damn I need to bring my laptop to repair... Sep 08 08:07:55 some of hte locales contain unicode codepoints encoded in a format Sep 08 08:07:56 kergoth: hehe, try to file one and you will get shut down Sep 08 08:08:14 i just added the necessary code to parse that out, convert to python unicode, then encode in utf8 and replace those instances Sep 08 08:08:27 hehe Sep 08 08:09:03 had to do it in multiple places too, since we pull the package names out of those files Sep 08 08:09:06 heh Sep 08 08:09:46 * mickey|working ponders whether to attend to the TT developer days in October Sep 08 08:09:58 (because of Qt, not because of Qtopie-core....) Sep 08 08:10:00 mickey|working: if you go there, I will join Sep 08 08:10:20 send the OE baseball bat heavies to get that GPL source :-) Sep 08 08:10:37 mickey|working: I think I could arrange a small tour around the lab as well Sep 08 08:10:49 zecke: it's in munich, right? Sep 08 08:10:58 mickey|working: yes Sep 08 08:10:59 * zecke considers moving to munich :} Sep 08 08:11:15 zecke: excellent. I'll reserving a hotel then. Sep 08 08:11:22 s/reserving/reserve/ Sep 08 08:11:26 s/reserve/book/ Sep 08 08:11:27 *sigh* Sep 08 08:11:35 mickey|working: I wanted to talk about Robert about hotels Sep 08 08:11:41 mickey|working: and wait a second :} Sep 08 08:11:58 np, i wanted to do it on weekend Sep 08 08:12:12 mickey|working: I need to check if my dutch relatives are in berlin Sep 08 08:12:12 I suck at dates Sep 08 08:12:54 hehe Sep 08 08:16:33 morning #oe Sep 08 08:16:43 hey mikearthur Sep 08 08:17:43 ical helps me a bit Sep 08 08:17:43 mickey|working: okay dutch folk arrives on the 16th Sep 08 08:17:57 zecke: when will they depart? Sep 08 08:19:17 hey mikearthur, you were absent last night :-) Sep 08 08:19:42 zecke: ok, I'm writing a mail to opensource@trolltech.com to register myself. please do the same Sep 08 08:19:54 XorA: unfortunately, I have 3 things on wednesday nights Sep 08 08:20:20 XorA: EdLug, band rehearsal for a teenagers christian outreach project, band rehearsal for church Sep 08 08:22:15 mikearthur: simple, get Edlug to learn teenage christian pop at the church rehearsal place Sep 08 08:22:57 ade|desk: as long as I get to keep playing my bass, that sounds good :D Sep 08 08:24:11 mikearthur: luckilly edlug is thursdays :-) Sep 08 08:24:45 hmm are there open source synths etc that are any good Sep 08 08:25:01 * XorA misses ReBirth Sep 08 08:25:03 ade|desk: I hear fluidsynth and timidity are the main ones Sep 08 08:25:10 ade|desk: what are you wanting to do Sep 08 08:25:34 play Sep 08 08:25:40 MIDI/MOD? Sep 08 08:25:45 midi Sep 08 08:26:01 what sound card do you have Sep 08 08:26:05 presumably one without midi Sep 08 08:26:20 sb live Sep 08 08:27:01 mikearthur: I used to have an AWE64 with all my friends midi instruments sampled in the wavefonts, that was awesome Sep 08 08:27:20 that is pretty cool Sep 08 08:27:26 ade|desk: then you want soundfonts Sep 08 08:27:32 XorA: I remember the AWE64, nice card that Sep 08 08:27:59 ade|desk: there are some really good ones, just google around, can't remember any names off the top of my head Sep 08 08:28:42 i will thanks for the head up Sep 08 08:29:22 ade|desk: nice apart from the horrible white noise, lousy analogue on it Sep 08 08:30:34 mickey|bbl: If I rip the 18,19bpp bits out, it should get accepted Sep 08 08:31:13 I'm still wondering the best way to handle it - I could also try and fix up the 18,19bpp handling... Sep 08 08:34:43 I used to use the internet 4Meg soundfont on AWE64, it was very good, ll bet with SB live even bigger ram there are even better ones Sep 08 08:35:58 RP: your overlay patch doesn't use sram for qvga speedup? Sep 08 08:36:29 03clarson 07org.oe.dev * r1d300011... 10/ (1 classes/package.bbclass): Sep 08 08:36:29 package.bbclass: Sep 08 08:36:29 Make legitimize_package_name also convert style encoding of Sep 08 08:36:29 unicode codepoints into their utf-8 representation, as in glibc locale Sep 08 08:36:29 files. Sep 08 08:36:46 RP: oh I meant to say, found a way to hard kill X on corgi in a grphically cool manner, just run mplayer inside Xw100 :-) Sep 08 08:38:38 XorA: it doesn't like that? Sep 08 08:39:18 RP: nope, it gives a trange syncless screen looking like its in overlay mode and looking quite cool Sep 08 08:39:30 erm I meant that for koen Sep 08 08:39:56 unfortuneately that crashed X and I wasnt ssh'd in to see if it killed whole Z Sep 08 08:40:05 mickey|bbl: that overlay patch should work for a780 as well Sep 08 08:42:46 anyone know canadian law ? Sep 08 08:42:54 koen: and other pxa27x fb devices Sep 08 08:47:29 hmm, is there anybody want review this patch for opie-eye ? Sep 08 08:47:30 http://rafb.net/paste/results/y47MsI91.html Sep 08 08:51:54 good morning Sep 08 08:53:08 leoncamel: good patch Sep 08 08:53:43 hrw|work, hmm, the opie-image still can't build here. :( Sep 08 08:58:05 03Leon 07org.oe.dev * r49a6cfc2... 10/ (1 packages/opie-eye/opie-eye.inc): opie-eye: fix install Sep 08 09:13:01 XorA: was liam at edlug? Sep 08 09:13:57 mikearthur: no, he was working Sep 08 09:14:38 i have a theory about this gconf-dbus problem Sep 08 09:14:45 03koen 07org.oe.dev * r39edb7cb... 10/ (9 files in 4 dirs): minimo: check in stuff from #1391 (various authors) to make it easier for people to test Sep 08 09:19:28 dang, stupid irc, won't let me change my nick into eFfeM Sep 08 09:19:33 ~lart irc Sep 08 09:19:33 * ibot DoSes irc Sep 08 09:20:12 but not without the trailing _ Sep 08 09:20:20 _eFfeM: '/msg nickserv ghost effem ' Sep 08 09:21:54 hm, /msg returned ghost effem password then redid an identify and it worked Sep 08 09:34:23 can anyone explain me when do_stage is ran? Sep 08 09:34:56 after do_package, before do_build Sep 08 09:34:57 and is there a stamp associate with it (and entries in temp) ? Sep 08 09:35:20 I think you want do_populate_staging Sep 08 09:36:31 actually I don't really want anything, but bug 1380 complained that for uclibc crt1.o was not installed; for me this worked in the past but with a fresh build it now also fails for me Sep 08 09:36:50 and i guess that do_stage in uclibc.inc is actually installing this (but I cannot find that it is executed) Sep 08 09:37:23 or should someone somewhere have build uclibc-initial? Sep 08 09:37:26 florian_kc: tell me more Sep 08 09:37:50 as uclibc-initial is not build .... Sep 08 09:37:53 * koen boots workstation to do some work on merging packaged-staging.bbclass Sep 08 09:38:45 hrw|work: the differenct between your setup and my one is the dbus version. i built with a newer one. Sep 08 09:40:19 florian_kc: default OZ use dbus 0.23.4 and I do not know which version can be used to upgrade Sep 08 09:42:05 koen, shoudl I interpret the line "boots workstation" as "looking into this staging issue" or is this just a coincidence Sep 08 09:42:35 hrw|work: i used 0.62 but koen suggested to some 0.9x because of the stable api Sep 08 09:43:06 florian_kc: how much stuff would need to be rebuilt/upgraded? Sep 08 09:43:54 koen: does dbus_0.92 probide glib bindings or do we just lack the matching dbus-glim bb? Sep 08 09:44:59 hrw|work: its not that much... gpe-contacts, gpe-calendar, gpe-bluetooth, gpe-filemanager, gpe-beam Sep 08 09:45:00 eFfeM: I'm not looking into the staging issue atm Sep 08 09:45:13 i can't type today :-/ Sep 08 09:45:18 florian_kc: you'll need dbus-glib 0.71 Sep 08 09:45:53 koen, clear Sep 08 09:45:54 koen: hmm... sounds strange, why this? Sep 08 09:46:02 should uclibc-initial have been build first ? Sep 08 09:47:27 florian_kc: because the glib bindings are now in a seperate package Sep 08 09:47:34 basically the problem is that crt1.o exists but is only in the install dir, but not copied to staging or cross Sep 08 09:49:36 koen: sure, but why should i use dbus bindings from a different version of dbus? Sep 08 09:49:46 eFfeM: yes uclibc-initial used to be built, i don't see why that should have changed Sep 08 09:50:05 florian_kc: different version? Sep 08 09:50:25 eFfeM: you need initial so that the first pass of gcc is built Sep 08 09:50:48 florian_kc: "why do I have to use libgpewidget 0.114 with gpe-calendar 0.72 and not libgpewidget 0.72?" Sep 08 09:50:52 *back* Sep 08 09:50:55 koen: i'm going to use dbus 0.92 and glib bindings version 0.71 Sep 08 09:50:57 ~lart broken internet Sep 08 09:50:58 * ibot hereby declares broken internet a troll Sep 08 09:51:46 florian_kc: You can ask the dbus-glib people to bump the version number if that's more clear Sep 08 09:52:01 florian_kc: but since there were no changes, such version inflation is bogus Sep 08 09:52:03 mickey|bbl: still there? Sep 08 09:52:03 mickey|bbl: I have mailed TT as well Sep 08 09:52:09 koen: I still dont get any luck with minomo, same bug Sep 08 09:52:31 koen: ah, so it acually is a upstream issue :-) Sep 08 09:52:41 eFfeM: need to review the uclibc stuff as 0.9.29 is out soon(ish) before NPTL is added Sep 08 09:52:42 florian_kc: no, you're missing the point Sep 08 09:53:26 koen: i thought this split was introduced by oe, but in fact its an upstream change. Sep 08 09:53:37 ade|desk: uclibc initial is not build for me, also was not build in the old tree I have floating around Sep 08 09:53:44 but I did get a gcc cross initial Sep 08 09:53:53 florian_kc: yes, dbus-glib is a seperate project, with seperate numbering :) Sep 08 09:54:05 eFfeM: wow thats weird Sep 08 09:54:08 manually bb-ing uclibc-initial does not resolve things Sep 08 09:54:33 koen: yeah... that was the missing piece of information i just found :-) Sep 08 09:54:34 ade|desk: there is a do_stage in uclibc. inc which I assumed to give me that stuff Sep 08 09:55:24 hrw|work: i'll test 0.92/0.71 Sep 08 09:56:26 eFfeM: i have in my i386-linux-uclibc stamps uclibc-initial-0.9.28+svn20060713-r2.do_populate_staging etc Sep 08 09:56:43 hey, folks. where is libX11 package in OE ? Sep 08 09:57:10 i use multi machine as this seems to work better for when host ~== target Sep 08 09:57:55 eFfeM: what is your target? Sep 08 09:58:02 florian_kc: ok Sep 08 09:58:14 ade|desk: very odd Sep 08 09:58:33 ade|desk: my target is nslu2 (xscale) Sep 08 09:58:51 if I manually bb uclibc-initial nothing really happens at least not with crt1.o Sep 08 09:58:53 hmm so target is very much not host Sep 08 09:59:03 leoncamel: x11 or diet-x11 iirc Sep 08 09:59:23 no, host is 686 (intel D930, dual core, running ubuntu 6.06) Sep 08 09:59:26 packages/xorg-lib Sep 08 10:02:08 eFfeM: from gcc-cross-initial you need libc-initial which should pull in uclibc-initial Sep 08 10:02:20 eFfeM: do you have gcc-cross-initial ? Sep 08 10:02:24 hrw|work, hmm, here is a problem. http://rafb.net/paste/results/TDM4QS27.html Sep 08 10:02:33 hrw|work, it can't find X11/Xregion.h. Sep 08 10:03:08 ade|desk: yes Sep 08 10:03:39 leoncamel: probably not staged or diet-x11 lack it Sep 08 10:04:02 hrw|work, OK. I am checking it now. Sep 08 10:04:14 hrw|work: my guess is not staged Sep 08 10:04:31 ade|desk: is this ok: uclibc-initial_0.9.28.bb:PROVIDES = "virtual/${TARGET_PREFIX}libc-initial" Sep 08 10:04:48 especially the target_prefix part Sep 08 10:04:56 hrw|work, it seems that there isn't any Xregion.h in diet-x11. Sep 08 10:05:03 hrw|work, hmm, where is wrong ? Sep 08 10:05:51 eFfeM: yes that is correct as that is what gcc-cross-initial is looking for , but there is some weird shit with targets arm, armeb , mips and mipsel Sep 08 10:06:25 i'm on armeb Sep 08 10:06:29 in the gcc-cross-initial bb Sep 08 10:06:51 leoncamel: I'm not x11 expert Sep 08 10:07:41 hrw|work, OK. Sep 08 10:08:08 eFfeM: anyway thats all nothing to do with crt1.o but being in the correct place Sep 08 10:08:16 ade|desk: found this but don't really understand: DEPENDS += "${@['virtual/${TARGET_PREFIX}libc-initial',''][bb.data.getVar('TARGET_ARCH', d, 1) in ['arm', 'armeb', 'mips', 'mipsel']]}" Sep 08 10:08:48 hrw|work, hmm, the OE is changed a lot recently. why ? Sep 08 10:08:58 eFfeM: like i said weird shit Sep 08 10:09:12 ade|desk: I removed tmp and gcc-cross-initial gets build then uclibc, but crt1.o is not installed properly, guessed from uclibc.inc that do_stage should have done this Sep 08 10:09:58 btw this is gcc 3.4.4, gcc 4.1.1. will not build uclibc with thumb Sep 08 10:10:08 compiler error Sep 08 10:10:15 bad :( Sep 08 10:10:25 eFfeM: really? Sep 08 10:10:25 can't allocate register where one should be free Sep 08 10:10:36 eFfeM: bugger Sep 08 10:11:20 crt1.o should be put in coss/lib/ and staging/lib Sep 08 10:11:58 ade|desk: I know, but it isn't. that is why i started asking about do_stage (as I thought that would have done it) Sep 08 10:12:13 eFfeM: did you try uclibc_svn with thumb while removing the thumb patches that have been applied upstream Sep 08 10:13:43 nope Sep 08 10:14:33 i assume you mean with 3.4.4 not with 4.1.1. Sep 08 10:15:17 with 4.1.1 Sep 08 10:18:37 does thumb with glibc work on gcc 4.1.1 ? Sep 08 10:20:34 eFfeM: does your uclibc do a install_dev and install_runtime as this is where it copies crt1.o as part of do_stage() in my logs Sep 08 10:23:33 ade|desk: it does do the install_dev, but only as part of do_install, not as part of do_stage, and in do_install it copies to image not to cross Sep 08 10:23:40 (or staging) Sep 08 10:24:09 hmm thats not right Sep 08 10:24:13 koen: No, we don't use sram to speedup qvga Sep 08 10:24:33 XorA: Cool. The lockup doesn't surprise me ;-) Sep 08 10:25:05 ade|desk: there is also no staging in my temp dir , just configure, compile and install (and the latter fails) Sep 08 10:26:22 gone for lunch Sep 08 10:34:35 aargh Sep 08 10:34:53 I overwrote stuff before committing Sep 08 10:42:42 good morning folks Sep 08 10:42:49 hi chris144 Sep 08 10:42:57 hi eFfeM Sep 08 10:47:57 ade|desk: manually tried to build uclibc-initial, but that one does not build crt1.o etc so it fails with install: cannot stat `lib/*.[ao]': No such file or directory Sep 08 10:48:52 03koen 07org.oe.packaged-staging * r79e1ebe0... 10/ (5 files in 4 dirs): sync local changes Sep 08 10:49:00 03koen 07org.oe.packaged-staging * r7439fd03... 10/ (7 files in 4 dirs): ipkg-utils: sync with .dev Sep 08 10:49:04 03koen 07org.oe.packaged-staging * r48f9d0d4... 10/ (1 classes/packaged-staging.bbclass): classes/packaged-staging.bbclass: save list of installed packages Sep 08 10:52:36 ade|desk: uclibc-initial can Sep 08 10:52:57 ade|desk: uclibc-initial can't do the trick because its do_compile is empty, so it won't produce a crt1.o Sep 08 10:53:50 and this is in uclibc.inc but does noever get executed when uclibc is build: Sep 08 10:53:51 do_stage() { Sep 08 10:53:51 # Install into the cross dir (this MUST be done first because we Sep 08 10:53:51 # will install crt1.o in the install_dev stage and gcc needs it) Sep 08 10:55:14 koen, ade|desk should this actually be do_populate_staging ??????? Sep 08 10:56:15 there is this in uclibc.inc: EXTRA_OEMAKE_task_do_populate_staging = "${OEMAKE_NO_CC}" Sep 08 10:56:26 * eFfeM is completely lost on how this works ... Sep 08 10:56:55 03koen 07org.oe.packaged-staging * re8b10c54... 10/ (1 classes/packaged-staging.bbclass): classes/packaged-staging.bbclass: fix thinko in handling cross-dir Sep 08 10:56:59 eFfeM: do_stage is a wrapper around do_populate_staging Sep 08 10:57:46 eFfeM: and do_stage runs after do_install now Sep 08 10:57:57 eFfeM: which is probably why you are seeing those problems Sep 08 10:57:58 koen, was it different in the past ? Sep 08 10:58:12 yes, do_stage ran before do_install Sep 08 10:59:03 ok, and how do I best fix this ? paste the code from do_stage at the beginning of do_install (don't really like that) Sep 08 11:00:44 or s/do_stage/do_install_prepend/ Sep 08 11:00:53 koen: udev 0.100 released Sep 08 11:01:11 ow, and I just build 097 Sep 08 11:01:52 koen, your solution seems more neat Sep 08 11:03:43 koen also any idea why initial is not build (and do I need it) Sep 08 11:04:45 no idea on the initial stuff Sep 08 11:05:05 hi koen Sep 08 11:05:17 hey chris144 Sep 08 11:06:07 03koen 07org.oe.dev * r7fe7f8c6... 10/ (9 files in 3 dirs): udev: add 0.100, default preference = -1 Sep 08 11:10:55 koen, ade|desk with the change koen suggested things compile for me, it is now doing gcc-cross Sep 08 11:12:44 eFfeM: cool Sep 08 11:12:50 sorry was afk Sep 08 11:13:04 np Sep 08 11:13:08 now pushing the fix Sep 08 11:14:38 will also close bug 1380 Sep 08 11:14:56 03eFfeM 07org.oe.dev * r31da046f... 10/ (3 files in 2 dirs): uclibc: chaned do_stage to do_install_prepend to resolve order issue Sep 08 11:14:58 * koen tries to untangle his brain Sep 08 11:15:21 i need the slug people to review the thumb stuff in uclibc_svn w.r.t svn dates and then make svn more neutral and not so arm-ish. there is also the change in config vars such as HAS_FPU is now UCLIBC_HAS_FPU Sep 08 11:17:13 ade|desk: I am not really using the uclibc stuff, but at one point accepted the challenge to get this compiling again (as some things were severly broken) Sep 08 11:17:50 I can build the svn version, but I'd rather not have it in the release as svn versions are not really traceable (or can I use things llike SRCDATE ?) Sep 08 11:17:53 eFfeM: well done Sep 08 11:18:38 you can indeed, but with the srcdate the patches need to be for the srcdate too Sep 08 11:19:08 that's when I also stumbled into the gcc 4.1.1 problems, added several thumb patches, but at some point in time I got this iinternal error and gave up, probably should try to file a problem report somewhere for this Sep 08 11:19:34 else we end up with what we have now, broken svn cos upstream have included some of the patches but tweaked Sep 08 11:20:38 would be good to get 0.9.29 working as soon as it is released cos .28 is way too old Sep 08 11:20:56 current svn is much much more useful Sep 08 11:23:15 koen: I'll try something weird now - building OZ 3.5.4.x from .dev Sep 08 11:23:29 ade|desk: any idea on eta of .29 Sep 08 11:24:22 couple of weeks ish so says erik / mjn3 Sep 08 11:25:10 does it then make sense to spent much time on patching the svn version ? Sep 08 11:31:27 once .29 is out then i think we should clean svn patches out and keep it clean, have specific bb drops for known svn revs that people want /need Sep 08 11:31:33 we now have mindate/maxdate to mark patches with Sep 08 11:31:51 so if a patch gets applied at 20060707 you say ;maxdate=20060707 Sep 08 11:32:48 a grand idea but does seem silly to have the patches in the svn bb if a newer stable has them included Sep 08 11:33:39 people might have SRCDATE_uclibc = "somedate requiring those patches" Sep 08 11:33:43 uclibc_0.2.29-r3589.bb etc? Sep 08 11:33:54 else you end up with a svn fork in OE Sep 08 11:34:14 we have zd1211 with svn snapshots iirc Sep 08 11:36:36 koen: i think the srcdate_uclibc is only useful if after a date something major for the arch is broken and not likely to be fixed. the ucslug is the current only svn date fixer that i know of. most likely because the patch author does want to follow head beyond that date Sep 08 11:36:58 s/does/doesn't Sep 08 11:38:57 ade|desk: we don't know what people have in their private repo, that's why mindate/maxdate makes sense Sep 08 11:39:59 NOTE: multiple providers are available (tslib, tslib-maemo); Sep 08 11:39:59 NOTE: consider defining a PREFERRED_PROVIDER to match runtime tslib-conf Sep 08 11:40:16 PREFERRED_PROVIDER_tslib = tslib # is not enough now in .dev? Sep 08 11:40:48 if they have stuff in their private repos for the case of uclibc, then why is it that the stuff is in .dev svn bb Sep 08 11:42:17 mindate/maxdate is great for other packages i do agree Sep 08 11:44:14 03koen 07org.oe.dev * r350557ed... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: clean up preferred_providers a bit Sep 08 11:47:12 glibc does version-ed cvs drops with the patches for that date. Sep 08 11:47:55 if people want to add a 'snapshot', they should make a tarball Sep 08 11:48:09 that makes our users behind a fascist firewall much happier Sep 08 11:48:22 and avoid the cvs b0rkage we had with glibc Sep 08 11:51:55 hrw|work: Do you think a meta-bug for your images would make sense? Just testing out your 3.5.4.1 rootfs on my collie. Sep 08 11:52:54 Laibsch: does it boot at all? Sep 08 11:53:04 yes Sep 08 11:53:11 but no GUI Sep 08 11:53:27 hrw|work: Don't you have a collie anymore? Sep 08 11:53:28 Laibsch: good to know Sep 08 11:53:39 Laibsch: I have but do not have time for it Sep 08 11:54:07 Laibsch: 5500, 5600, 6000, c760 are on desk Sep 08 11:54:14 What about the meta-bug? I think the dependencies would clear up relationships and "responsibilities". Sep 08 11:54:25 hrw|work: Wow, quite a collection. Sep 08 11:58:02 1401 it is. Sep 08 12:02:53 Laibsch: http://bugs.openembedded.org/show_bug.cgi?id=1060 is same as your 1399 iirc Sep 08 12:03:18 no Sep 08 12:03:28 They are of course related. Sep 08 12:04:01 But the stuff about ENTER is completely different and only occurs at the first creation of the loop image apparently. Sep 08 12:05:03 in the SRC_URI section of a bb file getting something with https://... fails with "error checking the certitifcate" Sep 08 12:05:25 use --no-check-certificate .... Sep 08 12:05:39 anybody had a similar issue? Sep 08 12:09:19 hrw|work: What is the default IP for the Z on your 3.5.4.1 images? Sep 08 12:09:44 chris144: what bb file exactly? Sep 08 12:11:25 Laibsch: usb0? do not remember Sep 08 12:11:37 Laibsch: a custom made, kernel 2.6.11, downloading https://www.rtai.org/RTAI/rtai-3.4-test2.tar.bz2 Sep 08 12:11:47 OK, I will mount the image and check. Sep 08 12:12:29 192.168.0.201 iirc Sep 08 12:12:30 chris144: So you are asking about https support in bitbake in general? Sorry, cannot answer that. Sep 08 12:12:55 Laibsch: no prob...not a big issuse Sep 08 12:13:34 chris144: fix bitbake fetcher to add --no-check-certificate for https Sep 08 12:14:31 hrw|work: thanks Sep 08 12:16:27 hrw|work: do_commit_append_pavel = " HACK " in oe monotonerc? Sep 08 12:18:10 haha Sep 08 12:19:31 03MartinDietze 07org.oe.dev * r53ac04f0... 10/ (4 files in 3 dirs): Sep 08 12:19:31 configuration: merged in changes to nylon/mtx from 4G repository Sep 08 12:19:31 * added mtx-2 machine (4G Surfbox II) Sep 08 12:19:31 * updated several defaults Sep 08 12:21:24 hrw|work: This is of course still 2.4 kernel. I have not tried out 2.6. But bug 1400 says it won't boot at all. Sep 08 12:24:00 Laibsch: it boot but machine is unusable after boot Sep 08 12:25:50 NOTE: package dbus-0.92: completed Sep 08 12:25:59 NOTE: package dbus-native-0.92: completed Sep 08 12:26:02 NOTE: package dbus-0.62: completed Sep 08 12:26:42 koen: prev_profs += dbus-glib:dbus" gave it Sep 08 12:28:23 bizarre Sep 08 12:28:51 03MartinDietze 07org.oe.dev * r9faa5f9b... 10/ (48 files in 4 dirs): Sep 08 12:28:51 linux kernels: updated kernels plus patches for nylon/mtx Sep 08 12:28:51 * linux-mtx-1-2.4.27: the standard kernel used for the 4G access cube Sep 08 12:28:51 * linux-mtx-1u-2.4.27: the customized kernel for the 4G Surfbox I Sep 08 12:28:51 * linux-mtx-2-2.4.27: the customized kernel for the 4G Surfbox II Sep 08 12:31:35 03MartinDietze 07org.oe.dev * r68e43a30... 10/ (7 files in 4 dirs): libccaudio2: added new port. Sep 08 12:31:47 03MartinDietze 07org.oe.dev * r3c169f9b... 10/ (4 files in 3 dirs): libpcre: added new port. Sep 08 12:38:02 hrw|work: the https fetch can be enabled by adding FETCHCOMMAND_wget = "/usr/bin/env wget --no-check-certificate... to the .bb file.... Sep 08 12:38:30