**** BEGIN LOGGING AT Sat Oct 21 02:59:57 2006 Oct 21 07:11:38 RP: 'bitbake foo-image bar-image' fails with multithreaded bitbake as both tasks are going to stomp on rootfs/ Oct 21 07:11:48 good morning all Oct 21 07:24:26 morning koen Oct 21 07:24:58 hey nicolasfr Oct 21 07:26:47 I have to push some small patches to build kf with gcc4.0, do I put them on the bugtracker? And also a small fix to build hostap drivers in wpa-supplicant-0.4.9 Oct 21 07:29:23 I need to find some time to build a rootfs + kernel for the Tatung Webpad (x86 National Semiconductor / AMD Geode SC3200). Do you think a distro=Generic + machine=geodegx will do? Oct 21 07:29:52 that should work Oct 21 07:31:54 I don't know what is the brand of touchscreen. I hope it will work "as is"... Oct 21 07:39:39 try dissecting one of the slackware installs people talked about Oct 21 07:45:37 kernel 2.6.18 has support for USB on sharp w-zero3. It means someone has booted linux on it? That would make a nice target for OE! Oct 21 07:45:56 client or host? Oct 21 07:46:16 I bet it's support for attaching the w03 to you workstation Oct 21 07:48:40 you're right :( too bad Oct 21 08:09:35 good night all Oct 21 08:35:35 any idea of what that error could mean: /opt/OE/build/tmp/staging/arm-linux/qt2/lib/libqte-mt.so.2: undefined reference to `vtable for __cxxabiv1::__si_class_type_info' ? Oct 21 08:37:12 you switched between gcc3 and gcc4? Oct 21 08:37:41 koen: I started from scratch with gcc3 Oct 21 08:40:25 suihkulokki: ping Oct 21 08:40:31 suihkulokki: any idea on http://rafb.net/paste/results/S8BpR488.html ? Oct 21 08:48:44 suihkulokki: I guess sbox 0.9.8.ancient.maemo would be a problem, right? Oct 21 08:57:39 koen: if you are trying my toolchain, yes Oct 21 09:00:08 koen: specifically, the ld.so the gcc uses isn't where it expects to find it Oct 21 09:03:01 koen, does oe use tslib? Oct 21 09:09:25 Marex: OE is a buildsystem Oct 21 09:10:03 ok Oct 21 09:10:09 so if I'm building opie image Oct 21 09:10:16 does it use tslib Oct 21 09:11:03 that depends on the machine you are building for Oct 21 09:11:16 suihkulokki: ok, installing apophis now Oct 21 09:13:01 koen, then tell me what else can it use Oct 21 09:13:20 ask the opie dudes Oct 21 09:14:39 Marex: What is your machine? Oct 21 09:14:40 you said recently, that if I use oe then my ts will work Oct 21 09:15:00 but now it looks like you dont know if it will or not ...hmm Oct 21 09:15:04 right Oct 21 09:15:06 den-ros, palm TT Oct 21 09:15:09 but I don't use opie Oct 21 09:15:22 opie is unmaintained Oct 21 09:15:23 koen, I told you I use opie Oct 21 09:15:33 so? Oct 21 09:15:48 Marex: I think, it is best to build GPE now. Oct 21 09:15:49 so all you said are excuses Oct 21 09:15:54 you said you had troubled with ts-detect Oct 21 09:16:02 den-ros, on system with 16Megs of ram? Oct 21 09:16:03 :) Oct 21 09:16:05 which I said were fixed by tslib and udev in OE Oct 21 09:16:21 koen, nah ... ts-detect worked fine Oct 21 09:16:31 I couldnt calibrate ts in opie Oct 21 09:16:50 den-ros, I must use opie, there is no other way for me Oct 21 09:17:03 Marex: Opie will be supported less further. Oct 21 09:17:25 den-ros: that's sad. Oct 21 09:17:44 goxboxlive: Do not agree :) . Oct 21 09:17:54 den-ros, I know Oct 21 09:17:55 :( Oct 21 09:18:08 even though opie is only choice for low mem systems Oct 21 09:18:46 I am trying to build Opera from .dev this days, but with no luck. And i know that there are no more Opie developers left in OE, so i guess i have build my last Opie image Oct 21 09:18:58 Marex: Maybe one yet variant to use only text mode. Oct 21 09:19:28 very funny Oct 21 09:22:44 goxboxlive: For why to build Opie based old Qt/E, if QtopiaCore libs much improved. For why use even new Qt at all if now, finally good possible to concentrate only on one environment - GPE. Oct 21 09:23:58 I don't think gpe is the answer Oct 21 09:24:03 I build qtopie-core yesterday, and tried to build an Opie-image, but it failes. Oct 21 09:24:12 good platform independent frameworks are Oct 21 09:24:29 kone: Wich GUI do you mean is bether to use? Oct 21 09:24:51 Is enlightment something to go for? Oct 21 09:25:04 koen: Me too :) - but more importance to _not_ depend from TT managment. Oct 21 09:25:49 goxboxlive: e17 is nice, but it depends on your needs and tastes Oct 21 09:26:23 I use e17 on my netbook (8 inch 800x600 screen + keyboard), but I wouldn't use it on an ipaq or zaurus Oct 21 09:26:29 Are you able to start e17 on hx4700? Oct 21 09:26:36 ok Oct 21 09:26:49 why? because of speed? Oct 21 09:27:03 usability Oct 21 09:27:12 ok Oct 21 09:28:26 suihkulokki: apophis seems to work! Oct 21 09:28:31 After i change some themes in GPE it looks alot bether. But, the speed is so slow. It takes ages to open a application. Oct 21 09:28:55 goxboxlive: I tracked that down to a dbus timeout Oct 21 09:29:28 after fixing that apps open instantaniously (as they should) Oct 21 09:30:08 ok, when did you update that in OE? Oct 21 09:30:51 I didn't do that yet Oct 21 09:30:58 the fix is too ugly to commit Oct 21 09:32:15 ok Oct 21 09:32:28 Yesterday my OE build new Angstrom for my hx4700. Mostly looks working nice. Oct 21 09:32:47 as long as you login as root ;) Oct 21 09:33:01 den-ros did you get your soundcard working? Oct 21 09:33:58 koen: :) Yes i login from root yet :) , but when i login several times from user - there is no meet any issues fornow. Oct 21 09:34:41 goxboxlive: Sound looks like work now - but i did not test it yet. I say You after test. Oct 21 09:35:40 ok, Oct 21 09:36:55 psokolovsky_ I can see that you have updated the htcuniversal.conf. But you ghave added 'pcmciautils' in it. We dont have any pcmcia stuff. Could you please remove it. It will break building images for the Universal. Oct 21 09:37:26 goxboxlive: sounds works on my hx Oct 21 09:37:54 koen: Ok, thene there is something with the htcuniversal kernel driver. Oct 21 09:38:18 goxboxlive: the pcmcia utils thing doesn't affect you Oct 21 09:38:19 goxboxlive: Oct 21 09:38:21 Hi! Oct 21 09:38:23 goxboxlive, I you sure you don't have it at all? Oct 21 09:38:30 goxboxlive: pcmcia isn't in MACHINE_FEATURES Oct 21 09:38:45 goxboxlive, it won't break anything, but may be empty load, of course Oct 21 09:39:17 psokolovsky_ Yes we dont have it, we even dont have support for it in the kernel Oct 21 09:39:32 There are only a SD card reader on Universal Oct 21 09:40:00 goxboxlive, well, any PXA device has pcmcia, as it's in CPU. the q is it is used for anything or not Oct 21 09:40:11 03koen 07org.oe.dev * r118a68d9... 10/ (3 files in 2 dirs): gpe-calendar: move over to gpe svn Oct 21 09:40:20 03koen 07org.oe.dev * r2317228d... 10/ (3 files in 2 dirs): libgpevtype: move _cvs.bb to gpe svn Oct 21 09:40:22 03koen 07org.oe.dev * r849c0d40... 10/ (1 conf/machine/htcuniversal.conf): htcuniversal: pcmcia isn't in MACHINE_FEATURES, so setting PCMCIA_MANAGER is a bit strange Oct 21 09:40:24 koen: Even though there isent a machine features i dont think we shall have stuff in the file that dont belong in it. It will confuse new builders Oct 21 09:40:45 goxboxlive: already fixed :) Oct 21 09:40:51 goxboxlive, but as koen said, there's no pcmcia it features Oct 21 09:40:52 i saw it thx. Oct 21 09:41:25 koen, so we can't define PCMCIA_MANAGER if no pcmcia in features? Oct 21 09:42:05 psokolovsky_: you can, but there's no use Oct 21 09:43:29 koen, well, then task-base, being new, already needs some de-tangledness. PCMCIA_MANAGER should be the app to use *if* there's pcmcia support for machine. It should have zero importance if it doesn't have. Oct 21 09:43:55 koen: having it otherwise is exposing some internal (mis)assumption in task-base, i.e bug. Oct 21 09:44:22 ehm Oct 21 09:44:31 I think you are a bit halucinating Oct 21 09:45:08 I removed it from htcuniversal to stop newbies going "OMG PCMCIA!@!!!" when looking at the machine.conf Oct 21 09:45:19 koen, lol, why do you think so? ;-) Oct 21 09:45:21 for bitbake/OE/task-base it didn't change anything Oct 21 09:45:31 :wq Oct 21 09:45:32 lol Oct 21 09:46:01 koen, let's default it to ?= "pcmciautils" and remove it everywhere not needed Oct 21 09:46:25 I think it already defaults to that Oct 21 09:47:13 actually Oct 21 09:47:14 conf/bitbake.conf:PCMCIA_MANAGER ?= "pcmcia-cs Oct 21 09:47:29 it defaults to pcmcia-cs to protect legacy machines Oct 21 09:47:31 03koen 07org.oe.dev * r196dba7f... 10/ (3 files in 2 dirs): libgpewidget: move cvs version to svn Oct 21 09:47:37 03koen 07org.oe.dev * rde2cfb4a... 10/ (3 files in 2 dirs): gpe-conf: move _cvs to svn Oct 21 09:48:10 koen, ok, let it be that way for now. minor thing, considering refactors I have in q. Oct 21 09:54:32 03pfalcon 07org.oe.dev * rf42f6c6d... 10/ (6 files in 4 dirs): handhelds-pxa-2.6: Remove orphaned, stale, and outdated patches. Oct 21 09:58:23 psokolovsky_: http://www.openembedded.org/bonsai/query/author/pfalcon@openembedded.org/ Oct 21 10:00:21 koen: what's wrong there? Oct 21 10:00:28 nothing :) Oct 21 10:01:23 oh, cool ;-) Yep, I'm just waiting when I can go to http://www.openembedded.org/bonsai/ and see the form to qyery stuff ;-) Oct 21 10:04:07 http://www.openembedded.org/bonsai/query/{author,file,description}/ Oct 21 10:06:00 03pfalcon 07org.oe.dev * r8335ff00... 10/ (1 packages/gdb/gdb_6.4.bb packages/libxml/libxml2_2.6.26.bb): Oct 21 10:06:00 gdb 6.4, libxml2 2.6.26: Prefer .tar.bz2 over .tar.gz in SRC_URI. Oct 21 10:06:00 * Per the corresponding RFC in the mailing list. Oct 21 10:08:14 hey Oct 21 10:08:42 hey zecke Oct 21 10:08:56 ho Oct 21 10:09:04 Hi! Oct 21 10:09:49 03koen 07org.oe.dev * r09016de8... 10/ (3 files in 3 dirs): dbus: move fix-segfault to dbus/ so dbus 0.94 can find it Oct 21 10:28:43 03pH5 07org.oe.dev * r4b013376... 10/ (1 conf/machine/magician.conf): magician.conf: add machine configuration for htc magician Oct 21 10:32:40 hey, what is the best way to build for different pxa27x machines with bitbake? use two different bbpaths and have two caches but use the same work and staging dirs? Oct 21 10:33:20 pH5: multimachine.bbclass Oct 21 10:34:16 pH5: see http://rafb.net/paste/results/LKhq4O30.html Oct 21 10:35:04 koen: hm, but doesn't that mean I have to reparse everything every time I switch machines? Oct 21 10:35:24 pH5: right, that's the downside Oct 21 10:35:37 pH5: RP and I discussed a solution for that at OEDEM Oct 21 10:35:40 ~lart slow p4m Oct 21 10:35:41 * ibot duct-tapes slow p4m to the floor and drools on him Oct 21 10:36:11 goxboxlive: Yes, sound work. Oct 21 10:36:45 den-ros: Ok, good to hear, are you compiling the sounddriver in to the kernel or is it loaded as a modules? Oct 21 10:37:27 koen: ok, thanks. what is the solution? Oct 21 10:38:19 goxboxlive: As a module (fornow i did not change anything by myself, becouse learn OE and test Angstrom builds). Oct 21 10:38:53 pH5: (unwritten) patches to bitbake, putting $MACHINE in CACHE and env vars iirc Oct 21 10:38:59 den-ros: Ok, could you pastebin your content in /etc/modutils/ please? Oct 21 10:39:23 Can somebody push bug 1518 real quick? The SRC_URI for flex is wrong. Thanks. Oct 21 10:39:37 goxboxlive: Here stop charging after 100% seems stop to work. Ok one minute, please. Oct 21 10:40:38 den-ros: That is good. Power doesent work on the universal yet. Oct 21 10:41:50 goxboxlive: Thats bad :) ,becouse previous it work on my hx4700 in Angstrom and still work at Koen. Oct 21 10:42:05 ok Oct 21 10:42:57 Laibsch: done, thanks Oct 21 10:43:05 morning, btw Oct 21 10:43:09 hey mickeyl Oct 21 10:43:12 mickeyl: Thanks for pushing and good morning. Oct 21 10:43:17 hey mickeyl Oct 21 10:43:24 mickeyl: could you do s/ncurses/etk/ tomorrow? Oct 21 10:44:15 koen: that would be a huge task Oct 21 10:44:31 :) Oct 21 10:44:51 modules: g_ether hidp tty ircomm-tty pxa27x_uds rfcomm | snd-hx4700_audio snd-mixer-oss snd-pcm-oss Oct 21 10:44:57 speaking of that, i need to answer at the mailing list. someone asked why we don't autogenerate the efl++ bindings Oct 21 10:45:26 03mickeyl 07org.oe.dev * r4a8ed832... 10/ (1 packages/flex/flex.inc): flex fix SRC_URI close #1518 Oct 21 10:45:32 mickeyl: Hi :) Oct 21 10:45:50 hi den-ros Oct 21 10:45:57 mickeyl: why don't you autogenerate them? Oct 21 10:46:23 goxboxlive: See modules atop. Oct 21 10:46:34 zecke: because i'm a fan of highlevel, bindings. bindings that feel like a native API. autogenerated bindings can't do that most of the times Oct 21 10:46:52 zecke: http://blogs.qtdeveloper.net/ Oct 21 10:47:05 zecke: qt 4 doesn't have AA font support? Oct 21 10:47:05 mickeyl: ask raster to use GObject Oct 21 10:48:00 koen: It uses Xft, maybe someone thinks Xft is fluff? Oct 21 10:48:08 den-ros: I took a look in the hx4700.conf foile, and i saw that u had autoload both snd-pcm-oss and snd-pcm-mixer- I have added them to my conf file now, and i will try to build and test. thx Oct 21 10:49:31 koen: When i am building a angstrom-gpe-image or gpe-image from a fresh oe.dev it failes because it depends on detect-stylus. Why isent this included som it automaticly build detect-stylus when building a gpe-image? Oct 21 10:49:54 So i allways have to build detect-stylus before i start to build a pe-image Oct 21 10:51:37 weird Oct 21 10:53:01 koen: nice Oct 21 10:53:11 suihkulokki: does a 'rootstrap' have to contain anything special? Can I use a rootfs tarball for that? Oct 21 10:53:33 btw, sb-menu rocks Oct 21 10:53:53 koen: nothing special in rootstrap Oct 21 10:54:32 koen: I think basicly you shuld just need libs and headers from the rootstrap Oct 21 10:55:21 goxboxlive: I did not know, all this loaded modules needed. For example, there is kernel "ALSA driver loaded, but did not find any soundcards". Oct 21 10:56:39 suihkulokki: ok, I'll use a tarball later today Oct 21 10:57:19 koen: if your rootstrap has device nodes or other stuff that needs root permissions, you can use fakeroot to unpack it Oct 21 11:01:26 den-ros: Ok, thats the same message i get during boot. Oct 21 11:05:54 goxboxlive: I look into ALSA section kernel config - nothing useful while. Oct 21 11:09:36 goxboxlive: Looks like 100% charging control green light did not work, becouse some according it module did not load. Oct 21 11:11:09 ok Oct 21 11:12:50 koen, So, noone yet started on those bitbake get-rid-of-reparsing patches? Because I had the same idea in mind. Oct 21 11:13:20 morning all Oct 21 11:13:32 hey RP Oct 21 11:15:11 psokolovsky_: It should really be quite simple - change CACHE= to include ${MACHINE} then run bitbake as MACHINE=spitz bitbake foobar (and no MACHINE in local.conf) Oct 21 11:16:13 So from that point of view, its just something we should document in the OE documentation if it works. I dont want it enabled by default (although adding MACHINE to the cache path is harmless I guess) Oct 21 11:16:15 RP: Hi! I had an idea to track dependencies on *variables* vs *files* in bitbake cache at all. Oct 21 11:16:33 psokolovsky_: Ah, that is a long term aim :) Oct 21 11:16:48 RP, nice! Oct 21 11:17:03 psokolovsky_: The origianl plan was to do that in bitbake-ng but its more realistic to do it in the current code Oct 21 11:17:11 psokolovsky_: Its a lot of work though :-/ Oct 21 11:17:35 RP, another quick fix would be to add bb cache version to its filename. So, different BB version can be run on same workdir w/o reparsing with each BB version switch. Oct 21 11:18:17 RP: I don't think it's that much. But I wanted to ponder about it a bit more before sending mail to the list about it ;-) Oct 21 11:18:32 RP: could you bump bitbake trunk up a microversion? Oct 21 11:18:49 yipee, I will waste my day implementing stupid stuff for the uni assignment Oct 21 11:18:53 the term started :) Oct 21 11:19:02 psokolovsky_: You can do that locally - just alter CACHE Oct 21 11:19:30 psokolovsky_: Personally, I'd like to understand why people need to switch bitbake versions and fix that instead ;-) Oct 21 11:19:31 RP, locally, sure. suitable for mainline? ;-) Oct 21 11:20:04 RP: simple, QAing 1.7 ;-) I told you, it behaves differently than 1.6 for me. Oct 21 11:20:23 RP: putting $MACHINE in CACHE isn't harmless Oct 21 11:20:30 RP: kergoth will cluebat you :) Oct 21 11:20:30 RP: I think that's good enough reason to have that small unobtrusive fix in mailine ;-) YMMV ;-) Oct 21 11:21:06 psokolovsky_: I really don't want to have to actively support it Oct 21 11:21:41 koen: I'm sure we could find a way to make it work regardless of whether MACHINE was set or not Oct 21 11:22:22 ;-( but that's easy way to test new bitbake versions - compare how it works with older ones Oct 21 11:22:44 psokolovsky_: You can do it in local.conf Oct 21 11:22:57 ok Oct 21 11:23:27 psokolovsky_: Anyhow, the behaviour did change in 1.7 and comparing the two isn't entirely fair. 1.6 is broken IMO which is why I worked on 1.7 ;-) Oct 21 11:24:21 RP: Hope that doesn't mean that you say 1.7 works perfectly and doesn't need more testing and QAing ;-) Oct 21 11:24:45 psokolovsky_: That isn't what I said, far from it ;-) Oct 21 11:25:00 psokolovsky_: All I said is it works differently and not all differences are regressions Oct 21 11:25:10 ok. I got it - that's stuff for local use, no need to botehr to submit patch for it Oct 21 11:25:15 ok Oct 21 11:25:57 03rpurdie * r619 10bitbake/ (bin/bitbake lib/bb/__init__.py): bump 1.7 versions Oct 21 11:27:21 RP: thanks Oct 21 11:27:40 RP: did you look at this bug report send to the email? Oct 21 11:27:57 zecke: Which one? Oct 21 11:28:33 RP: the one with the diff attached? Oct 21 11:28:52 RP: BitBake crashes on 'which' Oct 21 11:29:21 oh it is the shell, so I have to ask mickeyl Oct 21 11:29:39 mickeyl: Do you have a clue why bb.providers would crash in lookup Oct 21 11:29:48 mickeyl: but importing it as Providers will work? Oct 21 11:29:50 zecke: can you give me a subject or date? Oct 21 11:30:13 zecke: I think there is an internal variable in the shell called providers which probably conflicts Oct 21 11:30:15 RP: 18th of october, BitBake-1.6 crashes in `which Oct 21 11:30:20 RP: do_rootfs needs locking as well for multithreaded bitbakes Oct 21 11:30:26 RP: but bb. qualifies it :) Oct 21 11:30:53 RP: What do you think, we backport this patch and release BitBake 1.6.2? Oct 21 11:31:01 RP: and maybe do a snapshot release of BitBake 1.7 as well? Oct 21 11:31:25 zecke: If I remember rightly, there was a good reason that patch only applied to 1.7. I need to look at it Oct 21 11:31:50 RP: sadly this guy didn't append the 'crash' Oct 21 11:32:07 zecke: Let me find a copy of 1.6... Oct 21 11:32:31 RP: I still find the word crash for an exception amusing. For me crash is something where fire and explosions are involved Oct 21 11:32:36 and PAIN Oct 21 11:33:02 zecke: do you want to store the diffs in postgres, or generate then on-demand? Oct 21 11:33:06 zecke: To me it can mean either... Oct 21 11:33:21 RP: hehe, okay so it is only me :) Oct 21 11:33:27 koen: I would store them in postgres :} Oct 21 11:33:30 zecke: or link to viewsvn/viewmtn? Oct 21 11:33:58 koen: doing that is another option. You might have noticed that this is planned Oct 21 11:34:50 gah. First thing I try in 1.6 gives an error as the command is invalid :} Oct 21 11:36:21 someone have try to build recently (a week) task-bootstrap for an ipaq ? ... the resulting image don't have the simlink zImage-2.6.xx-hhy+cvsyyyymmdd -> zImage ... so normal bootloadr setup fail :( Oct 21 11:37:18 pH5: this works: http://rafb.net/paste/results/ZWrrmV61.html Oct 21 11:37:35 gremlin[it]: task-bootstrap is deprecated Oct 21 11:38:01 so what else ? Oct 21 11:39:02 i already use task-base instead of task-bootstrap ... but i need an image :) Oct 21 11:45:21 hi all Oct 21 11:45:31 gremlin[it], what machine? Oct 21 11:45:36 Hi! Oct 21 11:46:15 zecke: The problem is shell.py never import bb :} Oct 21 11:46:17 h3600 Oct 21 11:46:22 hehe Oct 21 11:46:53 RP: hehe, yeah, now you say that Oct 21 11:47:00 it says from bb import :) Oct 21 11:47:03 gremlin[it], maybe it just misses need image postprocessor? Oct 21 11:47:55 gremlin[it], require conf/machine/include/LAB-settings.conf ? Oct 21 11:48:20 no i don't need Oct 21 11:48:25 03pH5 07org.oe.dev * r05f835c3... 10/ (1 packages/linux/handhelds-pxa-2.6_cvs.bb): handhelds-pxa-2.6-cvs: add magician to COMPATIBLE_MACHINEs Oct 21 11:49:02 gpe-image is weel done ? can i look at this ? Oct 21 11:49:30 gremlin[it], well, then bitrot on task-bootstrap started ;-) Oct 21 11:49:39 03rpurdie 07bitbake-1.6 * r620 10/lib/bb/shell.py: bitbake-1.6/lib/bb/shell.py: import bb to fix bb.providers references Oct 21 11:50:08 bitrot ? Oct 21 11:50:17 yep ;-) Oct 21 11:51:19 my dictionary don't have 'bitrot' :( ... Oct 21 11:51:40 gremlin[it]: try bit and rot Oct 21 11:51:59 03pfalcon 07org.oe.dev * ra4e8b376... 10/ (1 packages/linux/handhelds-pxa-2.6.inc): handhelds-pxa-2.6: This kernel source is not really limited to PXA devices. Oct 21 11:52:00 gremlin[it]: means decay of computer data Oct 21 11:52:41 thanks RP ... Oct 21 11:53:04 can you point me to a well done *-image.bb ??? Oct 21 11:54:10 gremlin[it]: http://svn.o-hand.com/view/poky/trunk/meta/packages/images/oh-image-pda.bb?rev=743&view=markup ? :) Oct 21 11:54:45 gremlin[it]: for .dev s/inherit image/inherit image_ipk/ Oct 21 11:54:54 ok thanks ;) Oct 21 11:55:27 gremlin[it]: The trick is a good task file - the image then just puts a few task files together Oct 21 11:55:51 so well i home task-base is good :) Oct 21 11:55:55 i hope Oct 21 11:56:40 gremlin[it]: yes, but you need something more like task-oh to actually create an image unless you just want something like bootstrap-image Oct 21 11:57:02 bootstrap-image was good for me ;) Oct 21 11:57:20 i need only a serial console ;) Oct 21 11:57:37 gremlin[it]: http://svn.o-hand.com/view/poky/trunk/meta/packages/tasks/task-oh.bb?rev=787&view=markup is realitvely clean Oct 21 12:08:49 i had breakfast.# Oct 21 12:09:02 ah, sounds like you already fixed it without me Oct 21 12:09:03 excellent :D Oct 21 12:09:33 morning RP, btw. Oct 21 12:09:57 hi mickeyl Oct 21 12:10:08 mickeyl: hey, yeah the MVP fixed it :) Oct 21 12:10:19 MVP? Oct 21 12:10:19 MontaVistaPartner? Oct 21 12:10:24 hehe Oct 21 12:10:26 ModelViewPontroller? Oct 21 12:10:39 MickeyVicePresident? Oct 21 12:10:45 yeah Oct 21 12:11:10 OT: suspend-to-disk works great on my dual-xeon workstation. Oct 21 12:11:19 ~hug swsusp2 Oct 21 12:11:40 * ibot jumps into swsusp2's lap and huggles and *hugs*! Oct 21 12:12:11 I should really see if I can get that working... Oct 21 12:14:04 mickeyl: good news, Ill have to give that a go as well Oct 21 12:14:18 mickeyl: its been years since I last reied Oct 21 12:14:20 tried Oct 21 12:14:37 yeah, it looks like it just works (tm) now Oct 21 12:14:57 on Mandriva all i had to do was 'urpmi hibernate' and then it was there and could be called as 'hibernate' Oct 21 12:15:11 Suspend to ram usually works on my laptop which is handy Oct 21 12:15:12 ah well i had to add resume2= Oct 21 12:15:28 RP: that doesn't work yet on my thinkpad x60s Oct 21 12:15:58 mickeyl: I had to compile my own recent kernel for my Z60m but with that, it works Oct 21 12:16:08 mickeyl: I thought hibernate was that Java Database abstraction thing? Oct 21 12:19:44 Linux mhcln01 2.6.19-rc2-git3-mh01 #1 SMP Thu Oct 19 17:57:40 CEST 2006 i686 GNU/Linux, tell me about recent kernels Oct 21 12:21:28 hehe Oct 21 12:21:33 * koen hails the new reparseless overlords Oct 21 12:22:01 * zecke hails XNU Oct 21 12:25:34 mickeyl would be an improvement as V :) Oct 21 12:25:37 er VP Oct 21 12:27:28 zecke: so much for a unique namespace :D Oct 21 12:27:35 * mickeyl bows to Crofton|home Oct 21 12:27:42 mickeyl: ? Oct 21 12:27:45 ah Oct 21 12:28:35 [sbox-SDK-angstrom: /etc] > cat angstrom-version Oct 21 12:28:35 Angstrom test-20061012 Oct 21 12:28:53 suihkulokki: 'Error: f0005' mean I need a better qemu? Oct 21 12:30:12 [g2]: considering and testing several options for FPGA (pre-linux, post-linux, 1-1 from flash, compressed/uncompressed, etc) Oct 21 12:30:27 whoops, wrong channel Oct 21 12:34:41 koen: uh Oct 21 12:35:56 adding alsa to machine and distro features will load the alsa libs on task-base? Oct 21 12:36:44 florian: FYI, I'm going to commit xserver-common_cvs.bb, which was part of my initial h4000 support patches ;-). Need it now to support hh.org kernel button codes change refactor. Oct 21 12:37:15 florian: Of course, patches for xserver-common will follow once it's in hh.org cvs trunk. Oct 21 12:37:51 psokolovsky_: won't that break all my machines? Oct 21 12:38:02 psokolovsky_: since I hh7 doesn't have the refactor? Oct 21 12:38:09 koen, nothing will break. DEFAULT_PREFERNCE=-1 Oct 21 12:39:49 psokolovsky_: oh nice... Oct 21 12:40:37 florian, maybe you could make a new release of xserver-common... Oct 21 12:40:49 psokolovsky_: sure Oct 21 12:41:49 florian, thanks! in its current state, I mean. so, there was a release before such changes. Oct 21 12:44:06 florian: I converted some gpe stuf to svn this morning Oct 21 12:44:37 koen: oh cool Oct 21 12:45:06 that reminds me that there is quite a lot of work to do Oct 21 12:45:28 http://devmand.home.linuxtogo.org/ ?? Oct 21 12:45:35 index.php of size 0 :) Oct 21 12:45:48 florian, I also posted report of rendering issue with LTG's viewcvs, hope you got it. Oct 21 12:46:41 zecke: that's the super clever gforge... in a directory where php isn't allowed Oct 21 12:46:53 gforge rocks! Oct 21 12:46:58 psokolovsky_: yeah, i just answered... Oct 21 12:47:08 ok, thanks Oct 21 12:47:17 zecke: sure - if you have a full time admin to fix it Oct 21 12:49:33 florian, the issues is that mainstream viewcvs (now ViewVC) is in python, I have no idea where the seemingly-PHP version in GForge is from. The UI is quite similar though, so it's either direct rewrite, or clever proxying happens behind the scene. Oct 21 12:50:15 psokolovsky_: gforge to 'wget http://foo/viewcvs' and puts in a page Oct 21 12:50:20 s/to/does/ Oct 21 12:50:29 weird ;-) Oct 21 12:52:23 so far about "clever" ;) Oct 21 13:05:13 03pH5 07org.oe.dev * rdb67a4d3... 10/ (3 files in 3 dirs): handhelds-pxa-2.6: add magician defconfig Oct 21 13:17:36 bbl Oct 21 13:31:06 morning Oct 21 13:33:13 pH5: Are you interested in study software for your handheld? Can I motivate you to take a look at http://oz.leggewie.org/wip/jards/jards_1.0.bb and http://oz.leggewie.org/wip/jards/log.txt Johnny and I worked a bit on it, but I am stuck again. Oct 21 13:36:41 03coredump 07org.oe.oz354x * r98f6aefa... 10/ (1 conf/distro/openzaurus-3.5.4.2.conf): openzaurus-3.5.4.2.conf: Add CVSDATE for gnome-vfs-dbus Oct 21 13:36:46 03coredump 07org.oe.oz354x * r18cc8bbf... 10/ (7 files in 3 dirs): gnome: Add svn version of gnome-vfs-dbus as earlier versions do not compile with recent dbus packages.Thanks to hrw for tracking this down Oct 21 13:37:13 Laibsch: that looks like opie magic to me, which I don't know anything about Oct 21 13:44:23 03pH5 07org.oe.dev * ra01743ea... 10/ (1 conf/machine/magician.conf): magician.conf: correct recommended kernel modules Oct 21 13:44:28 03pH5 07org.oe.dev * r250d4a39... 10/ (1 packages/linux/handhelds-pxa-2.6/magician/defconfig): handhelds-pxa-2.6-cvs: fix magician defconfig Oct 21 14:04:31 koen: yes Oct 21 14:05:37 koen: check the cpu transparency devkit from: http://scratchbox.org/~rvoipio/armel/ and use select qemu-arm-0.8 Oct 21 14:06:23 suihkulokki: I tried using a different qemu, but that didn't help much Oct 21 14:06:41 hi koen Oct 21 14:06:48 hey chouimat Oct 21 14:07:50 grmmmbbl ... 1935/11141 .... better than the 5000/11141 on my other computer Oct 21 14:07:50 koen: what version of cputransp devkit do you have installed? Oct 21 14:08:01 the apophis one Oct 21 14:08:01 * koen checks Oct 21 14:08:12 http://scratchbox.org/download/files/sbox-releases/apophis/tarball/scratchbox-devkit-cputransp-1.0-i386.tar.gz Oct 21 14:08:37 cputransp devkit ? Oct 21 14:09:02 chouimat: trying to get OE toolchains working under scratchbox Oct 21 14:09:10 nice .. Oct 21 14:09:45 koen I'm currently waiting for the update of my OE.db file to finish Oct 21 14:10:02 errr I mean migration Oct 21 14:10:24 there's also a migrated snapshot available Oct 21 14:10:34 to 0.30? Oct 21 14:10:41 yes Oct 21 14:10:53 http://www.openembedded.org/snapshots/ Oct 21 14:11:31 as it is said in the user manual: http://www.openembedded.org/user-manual&dpage=ch03s02 :) Oct 21 14:11:52 koen manual are for wimps ;p Oct 21 14:11:55 * chouimat ducks Oct 21 14:12:08 Windows Icons Mouse Pointers? Oct 21 14:12:14 haha Oct 21 14:16:01 koen: it might be missing the eabi patch, this one should work better: http://scratchbox.org/~rvoipio/armel/scratchbox-devkit-cputransp-1.0.6~20060717armel-i386.tar.gz Oct 21 14:16:33 * koen boots workstation Oct 21 14:17:26 * koen wonders why users can't seem to grasp the notion of 'incompatible' Oct 21 14:18:07 * chouimat is listening to Land Of The Pig (The Butcher Is King) by Meat Loaf on Bat Out Of Hell III - The Monster is Loose [Amarok] Oct 21 14:18:23 koen because it said PC Compatible somewhare ;) Oct 21 14:24:33 koen: Do you have access to http://linuxtogo.org/mirror/openzaurus.org/ ? Oct 21 14:25:13 CoreDump|home: I think so Oct 21 14:25:41 It's the last mirror missing the 3.5.4.2 feeds ;) Oct 21 14:26:14 that's good to know Oct 21 14:27:11 suihkulokki: still Error: f0005 Oct 21 14:28:33 koen: how do the last lines of /tmp/cputransp_$user.log look like? Oct 21 14:29:03 [2006-10-21 14:46:58 5402] method: "/scratchbox/devkits/cputransp/bin/qemu-arm-0.8.1-sb1" pwd: "/home/koen" cmd: "/usr/bin/ipkg" Oct 21 14:29:03 [2006-10-21 14:46:58 5402] rc: 250 time: 0.007120 Oct 21 14:30:19 hrw|gone: Do you have any plane to release .2 for SL-Cxx00 as well? Oct 21 14:31:19 koen: Like I said, 0.8.0 works better :) Oct 21 14:31:44 hmm.. actually I didn't say that. Oct 21 14:33:45 03coredump 07org.oe.oz354x * r5a329154... 10/ (1 conf/distro/openzaurus-3.5.4.2.conf): Oct 21 14:33:45 openzaurus.conf: Prepare for ewi shutdown: http://ewi546.ewi.utwente.nl -> http://feeds.openzaurus.org/ Oct 21 14:33:45 distro: Prepare for ewi shutdown: http://ewi546.ewi.utwente.nl -> http://feeds.openzaurus.org/ Oct 21 14:34:09 oops, sorry for the messed up message Oct 21 14:38:53 suihkulokki: D'oh, yes Oct 21 14:41:06 suihkulokki: it seems to work fine now, although it complains about fakeroot not being preloadable Oct 21 14:42:57 RP: OE also needs to remove -dbg entries from RDEPENDS generated by the shlib code Oct 21 14:48:49 koen: sb-conf install --fakeroot (or same from sb-menu) Oct 21 14:49:13 already tried that ;) Oct 21 14:49:14 sb-conf: No fakeroot versions found for compiler: arm-angstrom-linux-gnueabi-4.1 Oct 21 14:49:29 hmmm Oct 21 14:49:31 ~lart qemu for qemu: Unsupported syscall: 256 Oct 21 14:49:31 * ibot puts qemu into a headlock and administers a mighty noogie, rubbing half of qemu's hair of for qemu: Unsupported syscall: 256 Oct 21 14:54:50 re Oct 21 14:55:14 wb florian Oct 21 14:55:45 koen: personally, I'd not lart qemu for fully supporting NPTL threads Oct 21 14:56:02 right Oct 21 14:56:06 ~lart ARM Oct 21 14:56:06 * ibot whacks ARM upside the head Oct 21 14:57:44 the easy part is to implement the thread related syscalls in qemu Oct 21 14:58:00 the hard part is that after that you have to be sure qemu is internally thread-safe Oct 21 14:58:06 ~lart my life for being boring Oct 21 14:58:06 * ibot DoSes my life for being boring Oct 21 15:02:13 Laibsch: this is an upstream error. Oct 21 15:03:01 probably just needs a simple #include "editlistswidget.h" Oct 21 15:07:00 anyone fool with a powerpc405 and OE? Oct 21 15:08:17 Crofton|home: not yet, still waiting for the efika :) Oct 21 15:08:36 the only powerpc I have is a 7450 Oct 21 15:09:01 http://www.digilentinc.com/Products/Detail.cfm?Nav1=Products&Nav2=Programmable&Prod=XUPV2P Oct 21 15:09:07 I want to play with one of these Oct 21 15:09:16 Crofton|home: I have a 405gp, but I'm chronical out of time :) Oct 21 15:09:28 Laibsch: heh, no it's even more easy Oct 21 15:09:37 - _itemDeleteButton->setEnabled(false); Oct 21 15:09:37 - _itemDetails->setEnabled(false); Oct 21 15:09:37 + __itemDeleteButton->setEnabled(false); Oct 21 15:09:37 + __itemDetailsButton->setEnabled(false); Oct 21 15:09:57 simple typos. i recommend fixing that upstream, not patching in oe Oct 21 15:10:53 it compiles through then :) Oct 21 15:12:40 * bhughes_ has a STB04xxx ppc sat tuner (ppc), dunno if it's a 405e though Oct 21 15:12:46 mickey|reading: Great. Thanks a lot. I do have access to upstream CVS ;-) Oct 21 15:13:02 np Oct 21 15:15:17 it looks like you working on Opie is the only reason atm. holding us back to move opie to packages/obsolete/ ... Oct 21 15:15:34 s/only/major/ Oct 21 15:16:05 mickey|reading: for me it is the only :) Oct 21 15:20:51 bitbake opie-image doesn't even start here Oct 21 15:20:56 because of some circular dependency Oct 21 15:20:58 *shrug* Oct 21 15:24:18 mickey|reading: *I* am the reason? Oct 21 15:24:30 If we have Japanese input on GPE, I am happy to switch. Oct 21 15:24:48 The rest can be moved over somehow I assume. Oct 21 15:25:10 Don't want to be a stepstone here. Oct 21 15:25:22 Laibsch: Not working on Opie doesn't mean advocating GPE :) Oct 21 15:25:53 Laibsch: not removing Opie now is all about not having wasted your time! :) Oct 21 15:27:22 Well, I just want my stuff working. Whether on opie or GPE or elsewhere. Oct 21 15:28:01 What is the direction, if neither Opie nor GPE? Oct 21 15:28:28 Laibsch: direction for OE? Mobile Linux? Oct 21 15:28:46 If keeping opie is an effort for no other reason than my stuff I will seek to move over to some other WM Oct 21 15:29:26 zecke: direction for OE. I mean if I try to get Japanese input on some other platform now I'd rather pick the one that has the highest chance of long-term support in OE. Oct 21 15:29:27 Laibsch: one really needs to differentiate here. Opie has one paid developer left, besides that no one is working on Opie upstream Oct 21 15:29:55 Well, but that is a paid one. Some projects have nothing but volunteers. Oct 21 15:30:08 Laibsch: the call for Opie OE maintainer has not yet resulted in any volunteers, so everything Gtk+/X11 based has the best chanche of surviving Oct 21 15:30:10 So what is the probable direction things are going to take? Oct 21 15:30:21 we are doing mosaic work now. when we have a certain level of pieces together, we need to step back a bit, look at all the pieces, and use them to build something homogenous, something concrete. from scratch, but leveraging existing pieces. we _need_ a common framework. Oct 21 15:30:38 Laibsch: the issue is that the paid dev works at Trolltech and doesn't commit too much, let alone fixing bugs :) Oct 21 15:30:53 I see. Oct 21 15:31:19 I thought you guys were quite enthusiastic about the TT days in MUC. Oct 21 15:31:21 Laibsch: so anything Gtk/X11 has currently the best chanche to survive :) Oct 21 15:31:40 Laibsch: we are, one really needs to differentiate here. It is quite tough though Oct 21 15:31:53 * Laibsch tried to understand. Oct 21 15:31:57 tries Oct 21 15:32:14 Laibsch: Qt4.2 is awesome. If performance tuning really starts it will kick ass Oct 21 15:32:48 So what is the problem? If it is a good platform it should attract devs? Oct 21 15:33:23 Laibsch: for Opie, well I'm tired of giving TT an alibi that their platform is Free. I'm no longer interested in being associated with Qtopia. And until they don't present a clear strategy regarding Qtopia I will not show/write any code that is using Qt4 on a portable device Oct 21 15:47:23 There are unfree bits in Qtopia? Oct 21 15:47:45 Laibsch: lol, for the last 5 releases there are no free bits at all :) Oct 21 15:48:35 Wow. I thought you did receive source code and had a dual-license on this stuff. The free one being GPL. I guess I was mistaken. Oct 21 15:49:06 Qtopia != Qtopia Core Oct 21 15:49:13 Laibsch: First it was fun, but basicly we have created a Wine/Samba :) Oct 21 15:50:03 zecke: ? Oct 21 15:50:05 Laibsch: Qtopia, or back then QPE, started a Free Software (by Trolltech) Oct 21 15:50:20 chouimat: trying to be compatible to a proprietray platform :) Oct 21 15:50:47 zecke: ok ... I think I will annoy ljp a little this weekend :) Oct 21 15:50:49 Laibsch: don't get me wrong. I have hopes Trolltech will free Qtopia again but until that day :) Oct 21 15:50:57 chouimat: don't do that! Oct 21 15:51:12 zecke: :) Oct 21 15:51:13 Hm, I was always under the impression that all of Qtopia was free after there initial opening up. Oct 21 15:51:35 their Oct 21 15:51:49 Laibsch: a lot of people think this, I think Opie is partly to blame for this believe Oct 21 15:52:27 zecke: then we just use kde4 as base ;) Oct 21 15:53:49 OK Oct 21 15:54:59 What is the benefit of moving it to obsolete? I mean I am all for clarity and also support not supporting non-free stuff. But is there indeed a tangible benefit moving it to obsolete. Is there a cost involved with having it under the normal .dev tree? Oct 21 15:55:35 Laibsch: yes, people will try to compile it, it will break or not work :) Oct 21 15:55:52 kill those suckers ;-) Oct 21 15:56:01 Laibsch: no, there are too many Oct 21 15:56:02 on sight Oct 21 15:56:31 zecke: Then you need a bigger gun ;-) Oct 21 15:56:56 Laibsch: why use a gun when you can use a nuke :) Oct 21 15:58:20 mickey|dinner: yes! frameworks! Oct 21 15:58:58 OpenEmbedded Portable Linux Specification :) Oct 21 15:59:36 haha Oct 21 16:00:04 zecke: you forgot 'award winning' Oct 21 16:01:09 NOTE: package fluxbox-1.0svn20061020: completed =) Oct 21 16:02:53 ah, a violation of our versioning policy Oct 21 16:07:30 ok, i thought i can do it when i have more time, but it is really necessary Oct 21 16:07:45 guys, i'll write up something about the state of linux on mobile Oct 21 16:08:12 i'll send it to a couple of you to enhance, add up, clarify, add your own points Oct 21 16:08:12 hehe Oct 21 16:08:18 mickey|dinner: and release it in ~3 weeks? Oct 21 16:09:07 koen: no, I'm wearing the FOSS hat now. this has to go out asap, monday or so Oct 21 16:09:16 FOSS hat? Oct 21 16:09:19 free and open source software Oct 21 16:09:19 * koen is wearing OE shirt Oct 21 16:09:38 heh, me too! Oct 21 16:10:02 was the first thing that fell out of my cupboard this morning Oct 21 16:10:25 garderobe, actually Oct 21 16:10:51 bbiab, have to clean dishes :/ Oct 21 16:10:55 koen: actually I stole that versioning string from a matchbox package. What would be the correct way? Oct 21 16:11:37 1.0+svn${SRCDATE} Oct 21 16:11:43 check the polices wiki page Oct 21 16:11:52 thanks Oct 21 16:16:20 NOTE: package fluxbox-1.0+svn20061020: completed ;) Oct 21 16:24:46 hmm Oct 21 16:24:57 anyone by chance build the dreambox branch lately? :) Oct 21 16:29:53 bhughes_: oh I have not tried, do you expierence an error? Oct 21 16:35:20 zecke, got a url to your opie-stay-the-hell-away-from-my-keymap patch? Oct 21 16:35:49 Kristoffer: he? Oct 21 16:35:57 zecke: yeh, dmutils doesn't link (can't find libqte, which is stange, i didn't think dreambox stuff linked with qt) Oct 21 16:36:13 zecke, We are having some issues with opie changing the keymap (so for instance arrows doesnt work properly) Oct 21 16:36:24 zecke, talked about it awhile back and you said you had a patch for it Oct 21 16:36:34 Kristoffer: http://www.openembedded.org/repo/org.openembedded.dev/packages/qte/qte-2.3.10/kernel-keymap.patch Oct 21 16:36:45 zecke, great thx Oct 21 16:36:58 Kristoffer: make sure you apply this patch :) Oct 21 16:37:05 bhughes_: oh Oct 21 16:37:17 zecke: can bonsai and tinderbox already talk to eachother? Oct 21 16:38:13 koen: well, I either go to a pub or start implementing tinderbox using django Oct 21 16:38:39 * koen would choose pub Oct 21 16:40:25 brb, reboot Oct 21 16:42:06 zecke, what do I need to recompile in order for it to be implemented? libopie, qte..? Oct 21 16:42:18 Kristoffer: libqte Oct 21 16:42:24 Kristoffer: the rest should be binary compatible Oct 21 16:42:39 oki roger Oct 21 16:48:48 lol: root@h6563:..ar/lib/amavis/virusmails >ls -l | wc -l Oct 21 16:48:49 47336 Oct 21 16:48:58 koen: hmm no response yet Oct 21 16:49:08 CoreDump|home: try wine and see how compatible it is :) Oct 21 16:49:20 =D Oct 21 16:49:31 zecke: response to bonsai? Oct 21 16:49:47 koen: response if we go to a pub :) Oct 21 16:50:11 I think I'll split the difference Oct 21 16:50:24 * koen gets some beer from the fridge Oct 21 16:52:59 "grolsch legioen" Oct 21 16:54:06 * koen reads http://www.ratebeer.com/beer/grolsch-legioen/47800/ Oct 21 16:59:41 hmmhttp://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=210742 Oct 21 17:02:53 zecke: easy, only use floats :) Oct 21 17:04:07 RP: should I do a bitbake release now? Oct 21 17:04:08 libxml2 ia failing building. Here is pastbin: http://pastebin.ca/213820 Oct 21 17:04:11 zecke: btw, I set Debug to False in the settings.py Oct 21 17:04:36 koen: I set it to true to find out about the broken RSS feeds Oct 21 17:04:42 goxboxlive: bzip2: /usr/src/universal/sources/libxml2-2.6.26.tar.bz2: trailing garbage after EOF ignored Oct 21 17:04:49 goxboxlive: it looks the download failed Oct 21 17:05:21 goxboxlive: 'rm /usr/src/universal/sources/libxml2-2.6.26.tar.bz2*' and rebuild Oct 21 17:27:20 koen: pub ;) Oct 21 17:27:49 uh oh Oct 21 17:28:12 quick warm all the single women in berlin that zecke is out there ;) Oct 21 17:28:23 um Oct 21 17:28:27 you meean warn Oct 21 17:28:29 ehm Oct 21 17:28:30 yes :) Oct 21 17:29:50 * CoreDump|home kicks off a poodle build Oct 21 17:32:24 warm works too :) Oct 21 17:37:25 koen: The problem was. In the libxml2_2.6.26.bb the source points to libxml2**.tar.bz2. But bz2 doesent exist ( i manually downloaded the source) so the source link shoud point to tar.gz instead. Oct 21 17:38:28 koen: Why should the shlibs RDEPENDS be ignored? Oct 21 17:39:34 RP: foo_2.3.2.bb shouldn't RDEPENDS on bla-dbg_5.6.5.bb via shlibs Oct 21 17:40:01 I did 'ipkg install gtk+-dev' and got ~15 -dbg packages Oct 21 17:42:36 koen: Last time I looked, it was broken gtk packaging, not an shlibs problem Oct 21 17:43:06 koen: I was thinking you meant something else btw (that foo-dbg shouldn't RRECOMMEND bla-dbg) Oct 21 17:45:09 koen: Can you be more specific about the packages with the incorrect RRECOMMENDS? Oct 21 17:45:32 I'm trying to find them via http://www.angstrom-distribution.org/repo/?action=details&pnm=gtk%2B-dev Oct 21 17:48:46 aha Oct 21 17:49:00 a lot of foo-locale packages have: Oct 21 17:49:01 Package: gconf-dbus-locale-amVersion: 2.10.0+svn20060719-r0 Oct 21 17:49:02 Depends: gconf-dbus-dbg, virtual-locale-am Oct 21 17:49:47 Package: gdk-pixbuf-loader-ani Oct 21 17:49:47 Depends: gtk+-dbg, gtk+ (>= 2.8.9), libglib-2.0-0 (>= 2.12.3), libc6 (>= 2.5) Oct 21 17:49:54 koen: Right, those packages have broken split packaging Oct 21 17:50:10 I did ask kergoth to look at this :-/ Oct 21 17:50:30 it seems that do_split_packages is indeed broken Oct 21 17:50:59 maximal instead of minimal matching Oct 21 17:50:59 gnumeric-plugin-foo, libgoffice-plugin-foo, etc all depends on a -dbg package Oct 21 17:53:08 I guess there isn't an easy way to exclude -dbg stuff from shlibs? Oct 21 17:53:47 shlibs isn't the problem Oct 21 17:56:52 http://bugs.openembedded.org/show_bug.cgi?id=1519 for future reference Oct 21 18:02:45 koen: I think the solution is to catch the debug packages in FILES and add them to the debug package before spliting Oct 21 18:02:53 * RP back later Oct 21 18:05:29 hmm Oct 21 18:08:03 ca i upgrade on my devel box to monotone 0.30 or it broke with oe database ? Oct 21 18:16:43 Wow, mtn is clever about updates to renamed dirs! if you have renamed dir, and someone adds new file to it/modifies existing, that will go to new dir. not so bad! Oct 21 18:19:41 03coredump 07org.oe.oz354x * r95bc3261... 10/ (1 conf/distro/openzaurus-3.5.4.2.conf): distro: Add CVSDATE for fluxbox Oct 21 18:19:46 03coredump 07org.oe.oz354x * r0d0629de... 10/ (1 packages/fluxbox/fluxbox_svn.bb): fluxbox: Add fluxbox_svn since the rc snaps are not updated very often Oct 21 18:20:58 gremlin[it]: Maybe it will be usefull: you must have converted OE db with new monotone. You can downlosd this db, or convert current by yourself. Oct 21 18:21:59 now i'm using 0.28 ... Oct 21 18:22:55 gremlin[it]: Did not remember net adress to monotone-0.30.x (?) converted db, sorry. Oct 21 18:24:28 ~oe.mtn Oct 21 18:24:30 methinks oe.mtn is snapshot of OpenEmbedded metadata available on http://www.openembedded.org/snapshots/ as files for monotone <0.30 and for >=0.30 Oct 21 18:24:33 ;D Oct 21 18:24:41 go to get some beers - cu Oct 21 18:24:43 :) Oct 21 18:25:39 CoreDump|home: and no - I do not plan to release 3.5.4.2 for !(poodle, tosa) - even collie/2.6 will get skipped because too many users think that collie lack sd support in OZ after reading about 3.5.4.2-rc2/collie/2.6 Oct 21 18:25:40 thanks all !!! Oct 21 18:25:57 CoreDump|home: please don't fork OE Oct 21 18:26:25 :)) Oct 21 18:26:26 who is forking? Oct 21 18:26:40 CoreDump|home: Oct 21 18:26:56 when were Oct 21 18:27:03 phe4r my mad ph0rking skillz Oct 21 18:27:08 what? :) Oct 21 18:27:15 ;) Oct 21 18:27:36 why on earth would I do that Oct 21 18:28:09 hrw|gone: no problem, just checking Oct 21 18:28:22 XorA|gon1: try finding fluxbox_svn.bb in http://www.openembedded.org/filebrowser/org.openembedded.dev/packages/fluxbox Oct 21 18:45:10 03coredump 07org.oe.dev * rdea0b7b3... 10/ (1 packages/fluxbox/fluxbox_svn.bb): fluxbox: Note to self: Do Not Fork OE when koen is looking Oct 21 18:47:27 lol Oct 21 18:48:00 :) Oct 21 18:53:14 Hi all. I'm trying to compile for sharp-compatible, but bitbake says "unable to open conf/bitbake.conf". strace reveals that it tries to open a certain "org.openembedded.dev/classes/sharprom.bbclass" file?! Oct 21 18:53:51 no such file exists Oct 21 18:54:53 conf/bitbake.conf does exist, however,, and works for openzaurus compiles Oct 21 18:57:41 johnny_a: Does you make all like here http://www.openembedded.org/wiki/GettingStarted ? Oct 21 18:58:58 den-ros:yes, as far as I can tell. Oct 21 19:04:33 Just read the file conf/distro/sharprom-compatible.conf, there's some info at the top... let's see if that helps Oct 21 19:07:54 having some issues starting udev as it requires hotplug, but hotplug is added to the userland Oct 21 19:15:08 hmmm Oct 21 19:15:16 why is everything rebuilding? Oct 21 19:19:35 re Oct 21 19:20:38 koen, why is everything rebuilding on me? Oct 21 19:20:55 because of herbs Oct 21 19:21:28 i guess you don't read the mailinglist? Oct 21 19:21:55 I do, but I left my laptop at school last night Oct 21 19:22:07 Crofton: http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-October/000655.html Oct 21 19:22:09 so I when I finally got back to it, I skimmed the list Oct 21 19:22:34 ah Oct 21 19:22:42 I tend to not notice rebuild triggers Oct 21 19:22:43 bother Oct 21 19:22:47 I wipe tmp daily :) Oct 21 19:22:50 yeah Oct 21 19:23:01 I was hoping to make a quick test imge Oct 21 19:23:09 then rebuild the everything Oct 21 19:23:38 * Crofton kicks himself for not starting this earlier Oct 21 19:23:38 "House Defense Chair Asks Pentagon to Remove Embedded CNN Reporters" Oct 21 19:23:45 * koen wonders if embedded cnn runs on ARM Oct 21 19:23:48 heh Oct 21 19:24:00 I wonder if they are "ARMED" Oct 21 19:25:03 just cnn reporters? Oct 21 19:27:25 anyone building sharprom binaries with bitbake? I can't get it to work... Oct 21 19:29:02 * koen dislikes closed source Oct 21 19:29:10 so, no, I don't build for sharprom Oct 21 19:29:43 thanks, now I know that. now, anyone who _does_ build for sharprom? Oct 21 19:31:41 * koen bows to overlay master XorA Oct 21 19:33:49 koen: whats happening Oct 21 19:34:51 XorA: double episode of CSI:LV and a pint of warm milk Oct 21 19:34:58 koen: wierd Oct 21 19:37:57 ah, gcc 4.2 release branch has been started Oct 21 19:38:42 warm milk? Oct 21 19:39:43 Is udev actually working for anyone.. the script looks faulty to me Oct 21 19:39:52 seems to work for me Oct 21 19:39:55 Crofton: 10pm over here, warm milk is great for that time of day Oct 21 19:40:11 Crofton, keeps complaining about hotplug, when it shouldnt need it Oct 21 19:40:20 hmmm Oct 21 19:40:21 and running a simple udevd & works perfectly Oct 21 19:41:18 PREFERRED_PROVIDER_hotplug = "udev" Oct 21 19:41:23 maybe Oct 21 19:42:03 you certain about that? Oct 21 19:42:10 no :) Oct 21 19:42:30 but it seems ok with openomap distro and omap5912osk machine .... Oct 21 19:42:37 and they got that set? Oct 21 19:42:44 yes Oct 21 19:42:57 but I don't really know Oct 21 19:43:08 koen, any hints? Oct 21 19:43:15 nope Oct 21 19:43:20 mickeyl, Oct 21 19:44:51 XorA: watch the next commit Oct 21 19:45:07 * XorA hides Oct 21 19:45:57 what kernels are you using? Oct 21 19:46:00 03koen 07org.oe.dev * r46b07c78... 10/ (3 files in 3 dirs): initscripts: add an extra check to checkroot.sh, inspired by the logicpd cset earlier Oct 21 19:46:15 XorA: I hope that fixes your RO issue Oct 21 19:46:27 XorA: in combination with passing 'rw' to CMDLINE Oct 21 19:46:43 koen: cool Oct 21 19:46:56 ~hail cbrake_away Oct 21 19:46:58 * ibot bows down to cbrake_away and chants, "I'M NOT WORTHY!!" Oct 21 19:46:59 2.6.18 or so Oct 21 19:48:19 My problem seems to come from that I lack the /proc/sys/kernel/hotplug file . But the kernel is configured for hotplug Oct 21 19:48:30 and udevd works as I said well without it Oct 21 19:49:01 I have it Oct 21 19:49:01 I could just remove some parts of the script that looks for hotplug file, but feels like a bad solution Oct 21 19:49:21 linux-omap1_2.6.18+git kernel Oct 21 19:49:52 yeah, no idea why i dont got it. But should work without it Oct 21 20:10:23 Does OE no longer support building for Sharp ROM? Oct 21 20:10:58 actually it wasn't really supported. Oct 21 20:11:18 I hacked this in to get my SharpROM python binaries compiled Oct 21 20:11:23 and it should still work IMO Oct 21 20:12:02 and despite me bringing it on a golden plate to them, no one of the SharpROM zealots or their cloners were interested Oct 21 20:12:03 *shrug* Oct 21 20:12:37 * koen give mickeyl a 'OMG LOL' to simulate a cloner response Oct 21 20:12:38 Well, the guy who does jards and who I recruited for OE is interested and has problems. Oct 21 20:12:46 He asked here about an hour ago. Oct 21 20:12:52 I am chatting with him Oct 21 20:13:02 Seems like the sharprom.bbclass is gone. Oct 21 20:13:04 Where is it? Oct 21 20:13:11 There he is ;-) Oct 21 20:13:15 hehe... hiya Oct 21 20:13:16 Hi johnny_a Oct 21 20:13:20 hello dude Oct 21 20:13:44 Let me take a closer look at sharprom-compatible.conf Oct 21 20:14:54 hmm.. I just saw that there's something called "sharprom-toolchain" in packages... Oct 21 20:15:26 The sharprom-compatible.conf seems to suggest you should have built that before doing anything else. Oct 21 20:15:34 Or provide a tool-chain of your own Oct 21 20:16:05 But I wonder if that is going to give you your bitbake.conf where you need it or your sharprom.bbclass Oct 21 20:16:19 But please go ahead and compile that first. Oct 21 20:17:14 I'm not sure what that package will do... and anyway, I can't bitbake it anyway (same error as before). Oct 21 20:17:35 maybe if I change DISTRO, and then bitbake it Oct 21 20:17:39 you need a prebuilt sharprom toolchain Oct 21 20:17:49 then set DISTRO = sharprom-compatible Oct 21 20:17:55 mickeyl: I got reparseless multimachine working today Oct 21 20:17:56 then try bitbake nano Oct 21 20:18:02 koen: woot. how that? Oct 21 20:18:12 mickeyl: listening to RP :) Oct 21 20:18:15 as I've said I didn't try for 12 months or so, it could be broken Oct 21 20:18:22 if it's easy, i might fix it Oct 21 20:18:32 but my motivation to do anything for SharpROM is _very_ low Oct 21 20:18:37 CACHE = "${TMPDIR}/cache/${MACHINE}" Oct 21 20:18:43 i don't want this wound to break up again Oct 21 20:18:44 *sigh* Oct 21 20:18:45 MACHINE=foo bitbake bar Oct 21 20:18:46 koen: cool Oct 21 20:19:20 mickeyl: and the angstrom sdk for scratchbox works as well Oct 21 20:19:56 next on the list is to try running an image under qemu system emulation Oct 21 20:21:03 mickeyl: I have the Embedix toolchain., and I've set DISTRO=sharprom-compatible. Still, bitbake says "ERROR: Unable to open conf/bitbake.conf", while strace tells me that it tries to open sharprom.bbclass (which doesn't exist Oct 21 20:21:33 sharprom.bbclass? Oct 21 20:21:40 hmm Oct 21 20:21:48 let me have a look at my environment Oct 21 20:22:15 koen: reparseless. Awesome! Another time of less wait. Oct 21 20:23:23 oh fsck! Oct 21 20:23:28 mickeyl: I'm drowning in SDKs for angstrom (OE, scratchbox, qemu-arm, task-sdk) :) Oct 21 20:23:40 koen: this _will_ come in handy :) Oct 21 20:23:53 darn, you guys are right. there's obviously a sharprom.bbclass which I never commited :/ Oct 21 20:24:00 wonderful Oct 21 20:24:04 depends Oct 21 20:24:09 mickeyl: I bet people will complain about having too much choice with sdks... Oct 21 20:24:13 * mickeyl digs in his old backups Oct 21 20:25:09 you guys need to use more closed software. :) (runs away) Oct 21 20:25:29 koen: I think I found the problem - do_split_locales assumes the first package in PACKAGES is the principle package Oct 21 20:25:43 RP: aaaah, that would do it Oct 21 20:25:54 I had thought of that, but didn't check it Oct 21 20:25:59 'it can't be that stupid' Oct 21 20:26:22 hmm, why can't I find init at boot anymore? Oct 21 20:26:39 Crofton: permission problems? Oct 21 20:26:54 koen: http://handhelds.org/scap/port.6594.png Oct 21 20:27:03 I untarred the image and it looks ok Oct 21 20:27:11 /sbin/init soft linked to sysvinit Oct 21 20:27:39 XorA: w00t (pardon my leetspeak) Oct 21 20:28:03 koen: It is that stupid :-( Oct 21 20:28:04 what is the deafult location for init? Oct 21 20:28:07 koen: I assume you grep the meaning :-) Oct 21 20:29:06 hmm Oct 21 20:29:30 setting init=/sbin/init got mme "Failed to execute /sbin/init" Oct 21 20:29:33 I'm afraid you're out of luck Oct 21 20:29:39 i don't have that class anymore Oct 21 20:29:46 nor can i remember what it did Oct 21 20:29:53 XorA: I'm still slightly worried about mixing libhal-nm and proper hal Oct 21 20:29:54 try removing it from INHERIT Oct 21 20:29:56 and see what happens Oct 21 20:30:13 hmm not much in /lib .... Oct 21 20:32:28 koen: we dont have hal in any distro yet though do we Oct 21 20:32:57 I dont think it's in INHERIT (Laibsch?)... Oct 21 20:33:46 XorA: we don't (yet) :) Oct 21 20:34:37 * Laibsch scolds mickeyl for being so careless Oct 21 20:34:45 ;-) Oct 21 20:34:58 weird Oct 21 20:35:03 now I have libs Oct 21 20:35:21 Laibsch: if we stumble over the first error, we probably know what sharprom.bbclass was supposed to do :D Oct 21 20:35:25 XorA: that's nm straight from OE + your patches, or does it require on-device editing? Oct 21 20:35:31 * mickeyl starts a build out of curiosity Oct 21 20:35:43 johnny_a: It does inherit sharprom. Oct 21 20:35:50 Try to remove that. Oct 21 20:36:05 INHERIT += " package_ipk debian" Oct 21 20:36:12 is what you need. Oct 21 20:37:01 hmm... sorry for being so clueless, but in what file? Oct 21 20:37:44 koen: that NM + wpa-supplicant 0.5.1 plus commit Im about to make, you need to edit a couple of files in /etc/dbus/system.d as we dont support at_console and you need to do /etc/init.d/wpa stop Oct 21 20:38:48 03xora 07org.oe.dev * rc50b6f0c... 10/ (3 files in 3 dirs): Oct 21 20:38:48 dhcp_3.0.2.bb : patch to add dbus support needed for NetworkManager/dhcdbd Oct 21 20:38:48 usage (shouldnt interfere with normal opps) thanks to gentoo. Oct 21 20:38:48 Add RDEPENDS_dhcp-client = "bash" as dhclient-script is bash only until Oct 21 20:38:48 someone wants to clean it for busybox Oct 21 20:38:58 and now I shall go Oct 21 20:41:36 Gotta sleep, see you tomorrow. Oct 21 20:43:27 hmmm Oct 21 20:43:43 it seems like I am not getting libs properly transfered to the image Oct 21 20:56:34 XorA: iirc bash isn't symlinked to /bin/sh automagically after install Oct 21 20:58:19 EOL errors in fluxbox Oct 21 20:58:38 koen: no it isnt Oct 21 20:58:54 koen: dhclient-script specifically calls /bin/bash Oct 21 20:59:05 ah, cool no probs then Oct 21 20:59:12 koen: way ahead :-) Oct 21 20:59:20 [ foo -a bar ] is a bashism, so it should :) Oct 21 21:04:44 03koen 07org.oe.dev * r21cef873... 10/ (1 packages/fluxbox/fluxbox_svn.bb): Oct 21 21:04:44 fluxbox svn : MAINTAINER isn't used in .dev anymore Oct 21 21:04:44 CVSDATE has been away for quite a while as well Oct 21 21:06:19 03koen 07org.oe.dev * r14bd8d9b... 10/ (1 packages/dhcp/dhcp_3.0.2.bb): dhcp: also package dhcp-dbg Oct 21 21:19:23 03mickeyl 07org.oe.dev * rdd230f15... 10/ (3 files in 3 dirs): Oct 21 21:19:23 make DISTRO = "sharprom-compatible" (sort of) work again Oct 21 21:19:23 no, I don't know why I am doing this... probably being too nice... Oct 21 21:21:10 lol Oct 21 21:38:32 mickeyl: you're too nice ;) Oct 21 21:38:36 grrr Oct 21 21:38:40 s/;/:/ Oct 21 21:39:03 i know... Oct 21 21:49:13 time to update the feeds with the latest fixes Oct 21 21:49:28 * koen crosses fingers for the initscripts fix Oct 21 21:49:46 03mwester 07org.oe.dev * r82b5799f... 10/ (1 packages/zd1211/zd1211-r83/makefile-unslung.patch): zd1211: makefile patch refresh for Unslung Oct 21 22:03:13 03rpurdie 07org.oe.dev * r3211e97a... 10/ (1 classes/package.bbclass): Oct 21 22:03:13 package.bbclass: Fix split_locales to remove spurious references to -dbg Oct 21 22:03:13 packages. The way it selects the principle package name needs totally reworking Oct 21 22:03:13 as the current code is flawed but apply a workaround for now (this is unrelated Oct 21 22:03:13 to other changes to package.bbclass) Oct 21 22:04:42 hey folks =) Oct 21 22:05:50 koen|away: is monotone the one that keeps track of version control? Oct 21 22:21:02 sweet: http://lists.freedesktop.org/archives/hal/2006-October/006317.html Oct 21 22:27:41 CSMan: we use monotone as our VCS, yes Oct 22 01:21:59 'evening gang - got through all the toolchain and even a kernel for my ppc device. Oct 22 01:22:19 now trying to make a bootstrap image via, well, bootstrap-image.bb. Oct 22 01:22:49 i think i'm running into an architecture naming problem. the packages are all ppc603e, which is "true". Oct 22 01:23:51 however, when i bake bootstrap image, all the dependencies (such as base-files, udev, netbase, etc) come up as unsatisfied. Oct 22 01:23:59 yet they have been built. Oct 22 01:24:04 any ideas? **** ENDING LOGGING AT Sun Oct 22 02:59:56 2006