**** BEGIN LOGGING AT Sat Mar 04 10:59:57 2006 Mar 04 11:27:25 Does anyone know if bugs.treke.net/bugs.openembedded.org and/or oesources.org will come alive again any time soon? Mar 04 11:28:05 Hoi Koen. Mar 04 11:28:49 johnX: Its out of our hands - someone needs to mention the problem to the owner of that machine (if someone hasn't already done so) Mar 04 11:28:50 Apparently not Mar 04 11:30:14 Well, I have no idea who this mysterious "owner" is, so don't count on me to tell him/her Mar 04 11:33:43 johnX: I've sent the owner an email Mar 04 11:34:34 ah, thanks Mar 04 11:35:52 I believe this is the man who runs the webserver: http://en.wikipedia.org/wiki/Paul_Allen Mar 04 11:59:13 greetings to all ! Mar 04 11:59:46 hi Ifaistos Mar 04 12:00:07 hi! how was FODSEM ? Mar 04 12:01:14 Ifaistos: I enjoyed it. Got to meet quite a few people. It was a shame you couldn't make it Mar 04 12:02:18 RP: Didn't do any work that weekend as i was expecting :( Got a bad case of cold and spend 3 days at bed Mar 04 12:03:18 Ifaistos: I hate it when that happens :-( Mar 04 12:41:37 good morning Mar 04 12:44:07 hi florian Mar 04 12:44:34 RP: Hi, good job on the sdk compilers! Mar 04 12:46:27 florian: Hopefully they'll work a bit better now :) Mar 04 12:46:42 RP: I ran into a minor packaging problem building gcc 3.4.4 for the sdk, but i think i have a fix. Mar 04 12:47:30 florian: sounds good :) Mar 04 12:48:41 RP: yeah - i tried to create a new 3.4.4 bb from the old sdk compilers and new bits, but obviously i missed some important thing. Mar 04 12:49:44 Getting it to work did drive me slightly mad as there are some subtle issues... Mar 04 12:50:15 RP: does this include building libstdc++? ;) Mar 04 12:50:40 I didn't test that... Mar 04 12:52:13 :-) Mar 04 12:53:03 I have a complete SDK with 3.4.4 now... will try this. Mar 04 12:57:56 arm-linux-gcc (GCC) 3.4.4 Mar 04 12:58:22 RP: Looks good, i just built libgpewidget. Mar 04 13:01:00 RP: I'm going to verify if my packaging fix causes any obvious problem with 3.4.3 now - if it works i'll push my change and 3.4.4 Mar 04 13:01:40 hi there. I'm trying to bitbake gtkmm for a h2200 machine and I'm getting the following error: http://rafb.net/paste/results/JcpGQM88.html Could anyone take a look at it, please? Mar 04 13:01:45 Hi there, Florian Mar 04 13:03:18 ArthurMaciel: looks like either your oe tree or bitbake is outdated. Mar 04 13:08:54 I'll try to update it. I do it with pull, right? Mar 04 13:09:20 ArthurMaciel: right, pull and update Mar 04 13:10:14 morning folks Mar 04 13:10:16 hi mickeyl Mar 04 13:11:08 morning, mickeyl Mar 04 13:16:02 hi mickeyl Mar 04 13:16:28 hi mickey Mar 04 13:20:02 grr ? Mar 04 13:20:09 what's wrong? Mar 04 13:25:57 doh, my water pipes are frozen again. how tedious. Mar 04 13:26:01 * pb_ stabs the winter Mar 04 13:26:08 pb_: ! Mar 04 13:26:38 Hey is the gnutls .bb broken? Mar 04 13:26:47 pb: darn Mar 04 13:27:28 Or is it just me? Mar 04 13:27:56 RP: hmm looks like libstdc++ and libg2c are located in different places in 3.4.3 and 3.4.4 :-/ Mar 04 13:28:00 Eddy: hard to say. you might consider describing the actual problem you're having. Mar 04 13:28:31 Well, I was just hoping somebody resolved the problem in the last say day Mar 04 13:28:47 I'll breaking in the populate stage Mar 04 13:28:53 Its breaking I mean Mar 04 13:29:14 just in a general kind of way, or is there a specific error message? Mar 04 13:29:21 florian: not good. How do the normal cross packages work around this? Mar 04 13:29:50 RP: not sure, need to check. Mar 04 13:30:04 Its hard to tell I get about 20 screens of crap flying by before it dumps to the shell Mar 04 13:30:20 I'm updating my repository and will check if the problem is still present Mar 04 13:33:41 Crap kernel upgrade, it'll be a bit before I can see if the problem is still present Mar 04 13:34:00 But it looked like it broke because while it was installing it couldn't find a specific file Mar 04 13:34:24 NOTE: package gnutls-1.3.4: completed Mar 04 13:34:29 works for me Mar 04 13:35:40 Maybe it'll work on the new repository Mar 04 13:36:21 I won't know for a while since the kernel has been upgraded Mar 04 13:36:25 :-P Mar 04 13:36:48 Building gpe-image Mar 04 13:37:42 bbl, going to do some light shopping Mar 04 13:38:40 have fun, pb_ Mar 04 13:42:31 morning Mar 04 13:43:56 mornign Mar 04 13:43:58 man gnutls failed on do_stage Mar 04 13:44:01 morning* Mar 04 13:46:24 Eddy: let us have a look. paste your output to a pastebin Mar 04 13:46:51 want me to paste the do_stage log? Mar 04 13:47:10 yep Mar 04 13:47:14 ~pastebin Mar 04 13:47:16 it has been said that pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.com/, or http://pastebin.ca, or http://channels.debian.net/paste Mar 04 13:51:56 pastebin truncates the log file big time Mar 04 13:52:23 can you upload to a webpage? Mar 04 13:52:23 http://pastebin.com/583463 Mar 04 13:52:27 hold on Mar 04 13:57:03 http://cdf-imaging.com/log.do_stage.txt Mar 04 13:57:34 aah yeah Mar 04 13:57:44 an infamous texi error Mar 04 13:57:50 two options Mar 04 13:57:55 a) disable doc generating Mar 04 13:58:05 b) update your sgml,texi infrastructure Mar 04 13:59:36 doc generating, does that mean no help? Mar 04 14:00:04 well, docs are not exactly necessary on embedded devices - most of the time Mar 04 14:00:14 is it just programming docs? Mar 04 14:00:21 like api docs or something? Mar 04 14:00:43 I would assume so since its gnutls huh Mar 04 14:01:07 so how do I disable? Mar 04 14:02:36 also what would I need to keep it enabled? Mar 04 14:03:51 no idea frankly Mar 04 14:04:07 on either option? Mar 04 14:04:08 check ./configure Mar 04 14:04:11 --help Mar 04 14:04:18 perhaps there's something like --disable-docs or so Mar 04 14:04:25 what packages are involved? Mar 04 14:04:36 that I would need to update to keep the docs Mar 04 14:04:38 latex? Mar 04 14:05:42 not sure exactly, latex and some of the sgml tools probably Mar 04 14:05:56 read the Makefile and see what tools it calls Mar 04 14:06:05 hi mickeyl Mar 04 14:06:37 hey chouimat, how are things for you Mar 04 14:07:02 k Mar 04 14:07:09 mickeyl: getting better ... it's a lot of work starting a company Mar 04 14:07:54 mickeyl: and playing with CAN and arm9 too :) Mar 04 14:08:53 i can imagine that, yeah Mar 04 14:08:59 CAN as in CAN-bus? Mar 04 14:10:07 yup Mar 04 14:13:21 ROFL http://www.askmen.com/jokes/2006_mar/mar04.html Mar 04 14:27:05 http://www.askmen.com/jokes/2006_feb/feb08.html Mar 04 14:30:00 that one is a good one too http://www.askmen.com/jokes/2005_dec/dec07.html Mar 04 14:32:35 hehehehe Mar 04 14:43:45 hey mickeyl Mar 04 14:48:14 03koen 07org.oe.dev * r253db2cc... 10/conf/machine/ep93xx.conf: ep93xx: add eraseblock size for the glomation board Mar 04 14:49:39 Could anyone try to help me, please? http://rafb.net/paste/results/KPhv0C75.html When trying to build bitbake gtkmm Mar 04 14:58:54 ArthurMaciel: it looks like you bitbake is too old Mar 04 14:59:22 I have to install a new bitbake or do I have to update the tree (I've updated it already) Mar 04 14:59:31 new bitbake Mar 04 15:00:40 I'll try that. Mar 04 15:16:15 koen: I'm unable to build gcc-cross-initial_4.1-20060217. any idea what I'm missing here? this is arm-linux/libssp/config.log: http://en.pastebin.ca/44446 Mar 04 15:16:30 it doesn't find crt1.o Mar 04 15:18:25 pH5: all -I and -L to the LDFLAGS in the Makefile Mar 04 15:18:49 it's either that, or you built another gcc before 4.1 Mar 04 15:29:14 <_stalker_> hi all Mar 04 15:29:30 <_stalker_> i can't manage to compile hostap-0.4.4-r4 Mar 04 15:29:44 <_stalker_> could anyone tell me how to force bitbake to compile hostap-0.3.9 instead? Mar 04 15:33:14 PREFERRED_VERSION_ = Mar 04 15:33:27 <_stalker_> PREFERRED_VERSION_hostap += "0.3.9" Mar 04 15:33:38 <_stalker_> i tried this and it still wanted to compile version 0.4.4 Mar 04 15:34:01 because it's called hostap-modules, not hostap Mar 04 15:34:05 hi koen Mar 04 15:34:10 <_stalker_> let's try Mar 04 15:34:11 <_stalker_> thx Mar 04 15:39:42 heh\ Mar 04 15:39:53 'avoid divergence, run qtopia on your nokia' Mar 04 15:41:40 :D Mar 04 15:47:31 mickeyl: did you see the tkc apps for the nokia? Mar 04 15:48:00 mickeyl: they rewrote them in *C* Mar 04 15:50:11 funny Mar 04 16:19:38 * pb_ returns! Mar 04 16:20:28 pb_: wb Mar 04 16:34:25 * florian is confused Mar 04 16:34:44 florian: join the club Mar 04 16:39:16 i don't get rid of the broken symlinks in the gcc sdk packages Mar 04 16:40:44 ibot: botmail for ph5: http://wiki.debian.org/ArmEabiPort Mar 04 16:41:09 florian: what do you mean? Mar 04 16:42:03 pb_: during packaging oe complains about some broken links in the package Mar 04 16:42:34 but i guess its ok... they seem to be fixed by library packages... Mar 04 16:43:34 <_stalker_> koen_: hostap-modules-0.3.9 builds fine Mar 04 16:43:49 <_stalker_> but now, it's atmelwlandriver-3.3.5.6 that refuses to build :( Mar 04 16:43:49 florian: which symlinks is it upset about? Mar 04 16:44:07 <_stalker_> it drives me nut Mar 04 16:44:24 just remove the dependency on atmelwlandriver Mar 04 16:44:36 <_stalker_> how can I? Mar 04 16:44:42 only some glibc .so links are wrong in the sdk, but that's the same in the old one. Mar 04 16:44:51 with vi or emacs Mar 04 16:44:58 <_stalker_> yeah, cheers Mar 04 16:45:24 <_stalker_> I don't even know which package needs atmelwlandriver Mar 04 16:45:41 pb_: libggc_s, libstdc++ and libg2c Mar 04 16:46:02 /usr/local/arm/oe/arm-linux/lib/libg2c.so.0.0.0 e.g. Mar 04 16:46:05 <_stalker_> koen|away: could you be a bit more explicit about removing a dep? Mar 04 17:02:19 do we have any 480x640 ipaqs? Mar 04 17:03:07 yes, the hx4700 is 480x640 Mar 04 17:03:14 ok, thanks Mar 04 17:05:33 hmm Mar 04 17:05:53 which machine tag is the hx4700 using? Mar 04 17:06:27 ah, ipaq-pxa270 Mar 04 17:07:16 are there any other pxa270 ipaqs or just the hx4700? Mar 04 17:07:39 i just added some 480x680 pictures for opie on tosa Mar 04 17:07:46 not sure. I think some of the 2000 series might be pxa270 as well. Mar 04 17:08:03 hmm, that's a bit unfortunate. we should have a concrete hx4700 machine tag imo Mar 04 17:08:21 at least for some opie GUI things Mar 04 17:09:06 otherwise it's using the upscaled pixtures for the 240x320 devices Mar 04 17:09:19 ah. "our" policy (in the sense of familiar) was always to avoid a proliferation of separate MACHINE types where it wasn't absolutely necessary. Mar 04 17:09:28 can you not ship both sets of pictures and choose between them at run time? Mar 04 17:09:56 we could, we decided not to do it because of size restrictions Mar 04 17:10:01 actually we have them in four versions Mar 04 17:10:13 e.g.: Mar 04 17:10:13 # Wallpaper and welcome splash Mar 04 17:10:13 PIXMAP_SIZE = "" Mar 04 17:10:13 PIXMAP_SIZE_c7x0 = "-640x480" Mar 04 17:10:13 PIXMAP_SIZE_spitz = "-640x480" Mar 04 17:10:13 PIXMAP_SIZE_akita = "-640x480" Mar 04 17:10:15 PIXMAP_SIZE_borzoi = "-640x480" Mar 04 17:10:17 PIXMAP_SIZE_tosa = "-480x640" Mar 04 17:10:19 PIXMAP_SIZE_simpad = "-800x600" Mar 04 17:10:27 that's for us not be able to use svg Mar 04 17:10:28 heh Mar 04 17:10:33 heh Mar 04 17:10:44 what is it you don't like about svg? Mar 04 17:11:05 i like svg, i just don't think qte can do anything with it Mar 04 17:11:28 ah right Mar 04 17:11:40 that would be a problem, I guess Mar 04 17:11:50 yeah. something to consider for opie 2.0 ;) Mar 04 17:13:00 :-) Mar 04 17:13:16 bbiab, going to do some pruning in the garden before darkness falls Mar 04 17:14:09 good plan. btw., when you have a chance, you should show us a picture of your house and garden to make us envy you Mar 04 17:14:10 :) Mar 04 17:18:27 :-) Mar 04 17:20:11 * florian just found out that feeding babies and using computers at the same time is really difficult Mar 04 17:22:56 heh - doesn't come as a surprise to you or does it? :D Mar 04 17:23:29 not really :-) Mar 04 17:37:25 03koen 07org.oe.dev * rbee378ef... 10/packages/glibmm/ (glibmm_2.4.5.bb glibmm_2.6.1.bb): glibmm: delete old versions Mar 04 17:41:06 03koen 07org.oe.dev * rbee378ef... 10/packages/glibmm/ (glibmm_2.4.5.bb glibmm_2.6.1.bb): glibmm: delete old versions Mar 04 17:47:15 florian: heh Mar 04 17:50:25 hey Mar 04 17:52:41 clearly zecke is a man on a tight schedule, no time to hang around Mar 04 17:56:27 ah, welcome back Mar 04 18:00:08 Hey experts -- I am in seach of a hint: I need to find an elegant way to tell a Makefile if the kernel it's building a module for is a 2.6 or 2.4 kernel --- is the best approach to do something based on "DISTRO" or is there a better place to look for a hint about the kernel used by the target/distro/machine/whatever? Mar 04 18:00:20 c/seach/search/ Mar 04 18:01:10 If it makes a difference, this is for the zd1211.bb Mar 04 18:07:02 mwester: i recommend checking the hostap and orinoco buildsystems, they're doing this. Mar 04 18:07:18 there are some tests to get a reasonably sane heuristic Mar 04 18:07:56 Thanks! Examples are good. :) Mar 04 18:28:20 evening Mar 04 18:28:26 hey Mar 04 18:28:39 hi zecke Mar 04 18:31:40 hi all Mar 04 18:31:47 hey koen Mar 04 18:39:16 hey zecke Mar 04 18:44:56 hi koen Mar 04 18:48:31 hey reenoo Mar 04 18:48:40 hi reenoo Mar 04 18:48:51 hey pb_ Mar 04 18:48:52 hi koen Mar 04 18:49:04 hey pb_ Mar 04 18:51:41 Hey how do you REbuild a kernel image Mar 04 18:51:45 ? Mar 04 18:52:43 I mean what do I tell bitbake Mar 04 18:53:49 clean and then build Mar 04 18:55:24 can i do bitbake -c clean build some_package ? (clean && build in 1 shot) Mar 04 18:55:36 so bitbake -c clean virtual/kernel Mar 04 18:56:00 then bitbake virtual/kernel? Mar 04 18:59:04 ya Mar 04 18:59:13 bbnew () Mar 04 18:59:13 { Mar 04 18:59:13 bitbake -c clean -b $1 && bitbake $BBD -b $1 Mar 04 18:59:13 } Mar 04 18:59:34 why adding stuff into the command when you can easily do things via shell? Mar 04 18:59:47 isn't that the unix philosophy? one tool = one job Mar 04 19:00:00 indeed Mar 04 19:00:08 bitbake -i knows about "rebuild" though Mar 04 19:00:30 hmmm should of tried that Mar 04 19:00:45 before I got an answer I just manually cleared out the stamps anyways Mar 04 19:00:54 for the kernel Mar 04 19:00:55 fwiw, the unix philosophy doesn't dictate that CLIs have to be cryptic Mar 04 19:02:00 or, in this case slow (if you omit the -b ...) Mar 04 19:02:22 indeed. that's why i made bitbake -i Mar 04 19:04:53 Did you make bitbake mickey1? Mar 04 19:05:25 or just patched it? Mar 04 19:05:32 bitbake is a collaborative effort Mar 04 19:05:38 i worked on different parts Mar 04 19:05:44 interactive mode, caching, ui Mar 04 19:05:57 pythonizig Mar 04 19:06:00 pythonizing, even Mar 04 19:06:00 beer drinking Mar 04 19:06:09 heh, yeah that as well Mar 04 19:06:45 sweet Mar 04 19:06:53 You a big gentoo person? Mar 04 19:07:36 no, not at all. my code is roughly 25% of bitbake anyway. the major contributors were Chris Larson, Holger Schurig, Phil Blundell Mar 04 19:07:51 Chris and Holger looked at gentoo Mar 04 19:07:58 when starting bitbake Mar 04 19:08:35 Why didn't they just use Gentoo? Mar 04 19:08:41 iirc, portage doesn't (yet) handle cross compiling. Mar 04 19:08:49 right Mar 04 19:08:55 ah Mar 04 19:09:05 it can build crosstoolchains apparently Mar 04 19:09:16 Yea I know that Mar 04 19:09:19 mingw Mar 04 19:09:25 besides, you can't just use a desktop distribution on embedded systems. there are people who think this is the way to go (*cough*), but most of us #oe folks don't Mar 04 19:09:29 koen: that's rather trivial to do Mar 04 19:09:56 reenoo: yeah, but I switched to debian before I tried to do that for my sparc Mar 04 19:10:39 Why is ipkg used and not dpkg? Mar 04 19:10:45 in this project Mar 04 19:11:08 ipkg just more lightweight or something? Mar 04 19:11:25 oe itself doesn't really "use" either ipkg or dpkg Mar 04 19:11:41 familiar, for example, uses ipkg because dpkg is missing a pile of functionality and apt is too heavyweight Mar 04 19:12:04 while we are on the topic, what host (x86) command can i use to "dump" a package - files names, dependencies, etc? Mar 04 19:12:09 openzaurus has it too Mar 04 19:12:18 indeed it does Mar 04 19:12:36 HopsNBarley: what do you mean by "a package" in that context? Mar 04 19:12:43 if you're talking about a .ipk file, you want dpkg-deb Mar 04 19:12:48 oh, sorry, an .ipk file Mar 04 19:13:17 e.g. dpkg-deb -c to see the contents, or dpkg-deb -I to see the control file Mar 04 19:13:33 thanks! hold on while i find it... (-; Mar 04 19:13:38 ~lart gimpOSX Mar 04 19:13:38 * ibot gives gimpOSX an extra strength ACME sleeping pill, sending gimpOSX to sleep for 150 years, and awakening to seven strange dwarfs and a large apple Mar 04 19:14:15 will libusb run on oe? Mar 04 19:14:28 on Zaurus SL C1000 Mar 04 19:14:32 it has USB host Mar 04 19:14:36 no, but you can use it in the image distro OE creates Mar 04 19:14:50 i.e. gphoto2 works on my nslu2 Mar 04 19:15:05 thats exactly the program I'm concerned about Mar 04 19:15:07 great! Mar 04 19:15:28 so what does it have to be statically linked or something? Mar 04 19:16:30 i can emerge dpkg v1.10.28, should that be okay? Mar 04 19:17:22 yeah, any version should be fine Mar 04 19:18:07 * pb_ shoots all online retailers in the uk Mar 04 19:18:17 trying to buy a new cd player for my car is a nightmare Mar 04 19:18:47 koen: What do you mean no? Isn't using it in the image distro = running on oe? Mar 04 19:19:16 Eddy: that would be "running on openzaurus", or whatever your image distro actually is. Mar 04 19:19:43 but openzaurus is openembedded right? Mar 04 19:19:52 no Mar 04 19:20:10 running on oe itself would be tricky, unless your cpu can execute .bb files directly. I'm not sure that even mickeyl is in that position. Mar 04 19:20:22 hehe Mar 04 19:20:38 So oe is just the environment that you create the images with? Mar 04 19:20:44 yes Mar 04 19:21:07 But you guys host openzaurus and familiar repository? Mar 04 19:21:34 right? Mar 04 19:22:52 pb_: thanks! works great! Mar 04 19:25:47 cool Mar 04 19:25:56 * pb_ heads off to tackle a large pile of ironing Mar 04 19:26:10 back... later Mar 04 19:26:21 later pb_ Mar 04 19:27:18 So are you guys openzaurus devs and openembedded devs? Mar 04 19:28:37 I'm just trying to oe for the very first time and my question may be somewhat stupid, but why do I consistently fail to bake any *-image? Mar 04 19:29:33 non-image targets seem to work... Mar 04 19:30:32 bugs Mar 04 19:30:59 either with your conf or the .bb files Mar 04 19:31:04 or your environment Mar 04 19:31:09 Well, likely... Mar 04 19:31:19 But where to dig? Mar 04 19:31:31 What package did it stop on? Mar 04 19:31:41 zecke: Is this what you meant? http://dominion.kabel.utwente.nl/koen/OE/people/ Mar 04 19:32:12 Well, it only has to do with -image-s. The suspicious output is: ipkg: Cannot create directory `/usr/lib/ipkg': Permission denied Mar 04 19:32:30 polyonymous: that one is harmless Mar 04 19:32:41 polyonymous: search for 'ERROR' in the logs Mar 04 19:32:52 Well and then it comes with sheer unsatisfaction Mar 04 19:33:01 ERROR: Cannot satisfy the following dependencies for task-bootstrap: Mar 04 19:33:11 And, I believe, the list contains everything. Mar 04 19:33:49 that means OE can't find the stuff in deploy/ Mar 04 19:34:09 which usually means that it either wasn't built or was removed Mar 04 19:34:10 lemme check these packages... Mar 04 19:34:28 Well, it was built and I haven't removed it, but I'll check the specific files now Mar 04 19:35:03 anyway, Thank you, I'll see to it now and, I'm afraid, will come back Mar 04 19:36:23 And yes, while 'handling bb files' it complains about unparsed line in pax utils native, I believe. Mar 04 19:36:59 (the line is require pax-utils-${Psomething} or something like that Mar 04 19:37:35 Hey just out of curiosity does anybody here know if the Zaurus C1000 can use a USB To Go cable instead of buying that zhost cable, and still provide a powered USB port? Mar 04 19:38:29 you mean using usb otg instead of zhost? Mar 04 19:39:19 pb_: glibc 2.3.6 lacks the eabi bits, right? Mar 04 19:39:39 pb_: the debian eabi wiki is a bit unclear on that Mar 04 19:40:00 yeah Mar 04 19:40:08 hell, yes, the packages aren't built. Aren't they supposed to be built automatically as a dependencies? Mar 04 19:40:23 polyonymous: they are supposed to Mar 04 19:40:34 polyonymous: I suspect you are using an old version of bitbake Mar 04 19:40:37 koen: any ideas what am I doing wrong? Mar 04 19:40:43 polyonymous: try upgrading to the latest svn Mar 04 19:40:49 Well, I thought so... I tried svn version and it doesn't work at all. Mar 04 19:40:58 koen: I'm missing a picture: yours ;-) Mar 04 19:41:03 * koen looks at zecke for making a new bitbake release Mar 04 19:41:13 I have 1.3.2, I believe Mar 04 19:41:22 pleemans: yeah, I was too lazy to dig out a decent picture :) Mar 04 19:41:32 The trunk version fails Mar 04 19:41:38 polyonymous: you need 1.3.3.x iirc Mar 04 19:41:56 koen: Thanks for putting the others from fosdem up though. Mar 04 19:41:59 Are there any unmentioned dependencies for latest bb? Mar 04 19:42:32 TypeError: pkgdata() takes exactly 2 arguments (3 given) Mar 04 19:43:01 File "./bin/bitbake", line 988, in collect_bbfiles Mar 04 19:43:31 doesn't look like missing deps on the other hand... Mar 04 19:45:17 Would you recommend any other tag or maybe specific revision of trunk? Mar 04 19:46:44 r384 works for me Mar 04 19:47:12 Hmm... this is head revision... doesn't work for me... Mar 04 19:47:59 oops. It does. Mar 04 19:48:08 Had to remove leftovers of 1.3.2 Mar 04 19:48:32 time to watch taggart Mar 04 19:48:37 Thank you. Mar 04 19:49:52 Trying to build an envirnoment I can run on the machine that has cacko to figure out why exactly I can't employ it as my main OS at the moment, and, hopefully, do something to make it better ;-) Mar 04 20:03:03 phew, ironing mostly under control again Mar 04 20:03:05 koen|away: yes, it does Mar 04 20:03:35 ah cool, saves me from cvs hell Mar 04 20:05:26 and off course my home bitbake machine crashed when I'm away Mar 04 20:07:48 er, that's to say, "yes it does lack them" Mar 04 20:08:13 ah, hmm Mar 04 20:08:25 I think 2.4 will have the eabi bits in though, if you don't mind waiting a bit Mar 04 20:09:23 pb_: Lennert and I will continue to try to get one of the cvs snapshots to compile Mar 04 20:09:33 Lennert via crosstool, me via OE Mar 04 20:09:44 righto Mar 04 20:10:11 I suspect the HEAD is slightly bit-rotten at the moment, so it probably needs patching somewhat. Mar 04 20:11:42 I'm still trying to get it to buld in OE with the old ABI Mar 04 20:13:05 yeah, good plan Mar 04 20:13:34 most of the code is shared; you should even be able to turn on nptl in the old abi if you want. Mar 04 20:14:19 that's kind of mandatory Mar 04 20:14:42 you get a lot of 'the sky will fall down' type of messages if you don't use ntpl Mar 04 20:15:36 true enough, though you don't need to take them too seriously. Mar 04 20:15:57 people with old (i.e. 2.4) kernels will still need a non-__thread build, which means linuxthreads. Mar 04 20:17:15 I'll ignore that for now. armv5/2.6 seems like the easiest thing to start with Mar 04 20:18:00 okay Mar 04 20:18:09 people with 2.4 kernels will also need a non-eabi glibc or patch their kernels won't they? Mar 04 20:18:28 yeah, that's also true Mar 04 20:18:45 I think koen was saying that you get the warning messages when trying to select linuxthreads even if you are using the old abi, though Mar 04 20:18:46 and armv4 can't use eabi due to the thumb stuff at the moment Mar 04 20:19:35 pb_: yah, some people are trying to push nptl... Mar 04 20:20:09 didn't linuxthreads got dropped from glibc cvs as well? Mar 04 20:20:44 no, it's in ports Mar 04 20:20:59 aha Mar 04 20:21:14 the sky hasn't fallen down, butg is hiding in ports/ Mar 04 20:21:21 s/tg/t/ Mar 04 20:21:41 oh, right, you were trying to build without either nptl or linuxthreads? that will indeed definitely not work. Mar 04 20:21:48 you need one or the other thread library. Mar 04 20:22:21 you can get it through configure with --with-iknowitbreaks or something like that Mar 04 20:23:23 but I'm trying with ntpl Mar 04 20:24:43 pb_: patching every single app doing the usual #ifdef __ARM__\n do_some_byteswapping is also going to be fun. Mar 04 20:25:00 reenoo: right Mar 04 20:26:44 reminds me of the fact that I still need to figure out how to serialize doubles in a language independent way (read: the gui is written in java) Mar 04 20:26:58 ~lart sun Mar 04 20:26:58 * ibot cats /dev/urandom into sun's ear Mar 04 20:41:12 hmm... does bitbake keep distfiles? Mar 04 20:45:28 ahh, it does ;-) Mar 04 21:05:16 03koen 07org.oe.dev * rbee378ef... 10/packages/glibmm/ (glibmm_2.4.5.bb glibmm_2.6.1.bb): glibmm: delete old versions Mar 04 21:05:21 03koen 07org.oe.dev * r7d24213f... 10/conf/machine/ep93xx.conf: ep93xx: add USE_VT = 0 Mar 04 21:05:25 03mickeyl 07org.oe.dev * r1db6857c... 10/packages/opie-taskbar/ (opie-taskbar.inc opie-taskbar_1.2.1.bb opie-taskbar_cvs.bb): opie-taskbar: enable dedicated pictures for tosa Mar 04 21:10:46 pb_: regarding the localedef issue, how would we go about packaging pre-generated files separately? localedef appears to generate a single archive file. Mar 04 21:11:32 * RP wonders if anyone will patch up 2.4.18... Mar 04 21:12:18 pb_: for nptl I need to buid gcc-cross-initial (without nptl) -> glibc-initial (without nptl) -> gcc-cross (with nptl) -> glibc (with nptl), right? Mar 04 21:17:22 reenoo: I think there's a flag to make it build ye olde separate archives Mar 04 21:18:02 koen: something like that, though gcc doesn't really care whether you have nptl or not Mar 04 21:18:22 pb_: yeah, there is. didn't know the old format generated separate archives. thanks Mar 04 21:18:56 so | configure: error: compiler support for __thread is required is a bit of a red herring Mar 04 21:19:17 well, that means your gcc doesn't understand __thread. Mar 04 21:19:28 which probably means it wasn't configured for eabi, or something. Mar 04 21:20:30 heh Mar 04 21:20:31 http://sources.redhat.com/ml/crossgcc/2004-08/msg00138.html Mar 04 21:22:12 so with or without the nptl dir configure bombs out Mar 04 21:22:23 (everything with old abi) Mar 04 21:22:40 hmm.. is there any particular reason for libgcc_s.so.1 not to be installed to staging? Mar 04 21:23:07 I think it goes into cross/, and that ought to be sufficient. Mar 04 21:23:21 not that installing it into staging would hurt, but it shouldn't be required there Mar 04 21:24:13 scary, googling my error turns up mails from rday Mar 04 21:24:43 pb_: qemu doesn't seem to grok multiple prefixes to search for libraries Mar 04 21:25:04 oh, right, yeah, that sucks Mar 04 21:25:21 I remember when I was playing with localedef under qemu I had a little script thing to make a special semi-chroot for it to run in. Mar 04 21:26:09 heh Mar 04 21:26:46 reenoo: I've seen it complete localegen using qemu systememulation on RP's laptop at fosdem Mar 04 21:27:26 which is kinda useless since qemu lacks a block driver, so you can't get it to save the result :( Mar 04 21:27:42 koen: network device support works though Mar 04 21:28:01 RP: sweet Mar 04 21:28:21 reenoo: (which is in http://www.freedesktop.org/~pb/glibc-qemu.patch if you're interested) Mar 04 21:28:27 koen: well. the plan was to use user space emulation, which should be sufficient. Mar 04 21:29:22 reenoo: Its just nice we can do complete system emulation to - it will boot all the way into X Mar 04 21:30:00 RP: the colours are still wrong? Mar 04 21:30:09 koen: yes. Its qemu at fault Mar 04 21:35:13 RP: sure. it's just a bit pointless from within OE. Mar 04 21:36:52 reenoo: Not really. It gives us development without hardware potential which will has a lot of appear to certain users Mar 04 21:37:20 reenoo: well and you can use it on systems not providing userspace emulation Mar 04 21:37:32 hey zecke Mar 04 21:37:53 s/appear/appeal/ Mar 04 21:39:05 RP: err, yes. but I don't want to generate and boot a complete image just to run a single command from a function in OE. Mar 04 21:39:50 reenoo: When see from that POV its not much use. When you consider you could run all postinsts though... Mar 04 21:40:34 and just becuase its not desireable for your particular use, it doesn't make it "pointless from within OE" Mar 04 21:43:02 we were talking about running localedef, weren't we? Mar 04 21:43:38 as for postinst scripts, half of them are device specific and need the actual hardware. Mar 04 21:44:45 koen and I discussed that the other day and concluded that these "half" that are device speficic could be written in a manner to cope... Mar 04 21:45:14 sure. just like the way we (ab)use $D now Mar 04 21:45:59 reenoo: No, we could do better than that IMO, especially if we start to use some kind of package handling the machine specifics Mar 04 21:46:10 (like devmand/zaurusd) Mar 04 21:46:22 all emus have one inherent flaw which i think is why reenoo is balking, it is extreamly rare that code on an emu will be have the same as a real device Mar 04 21:46:32 behave* Mar 04 21:47:00 emte: Its actually surprisingly good, right down to images oopsing in the same way as on the real hardware IME ;-) Mar 04 21:47:46 I agree you'd never want to 100% rely on it instead of hardware but if it runs on the emulator, the odds are good on the real hardware Mar 04 21:48:00 RP: dunno. I for one wouldn't want to be forced to use whatever package that would be. Mar 04 21:48:00 RP, my experiance comes from both software and hardware emus for Microchip and Altera chips Mar 04 21:48:38 03freyther 07org.oe.dev * r2718aaed... 10/packages/lirc/lirc_0.8.0.bb: Mar 04 21:48:38 openembedded/packages/licv/lirc_0.8.0.bb: Mar 04 21:48:38 lircexec.init is only installed to the tmp Mar 04 21:48:38 directory if we are building for NSLU2. So Mar 04 21:48:38 only install it when building on NSLu2. Mar 04 21:48:38 Earlier lirc bbfiles are not affected Mar 04 21:49:08 but yeah it is a valid approximation in the absence of real hw Mar 04 21:49:25 reenoo: The current machine specifics handling is a total mess with no structure and is spread around several different projects. Far from ideal... Mar 04 21:50:09 RP: that's a different story. as a familiar developer I may agree. as an OE developer I don't. Mar 04 21:51:24 effectively forcing (well. we don't have proper support for anything else) people to use ipkg is bad enough Mar 04 21:51:50 why do "thread local storage" and "transport layer security" use the same TLA? Mar 04 21:53:02 reenoo: I guess is comes down to a choice of whether machine specific tweaks should be centrally handled of handled adhoc and distributed... Mar 04 21:54:28 RP: no. the point is that people can do whatever they want in their postinst Mar 04 21:54:48 RP: they need a simple way to check when and how the script is run Mar 04 21:55:07 RP: offline, emulation, target Mar 04 21:55:43 I've never disagreed with that... Mar 04 21:56:07 heh Mar 04 21:56:09 nevermind Mar 04 21:58:23 reenoo: I agree you need the above. I also think a lot of our current machine specific postinsts could be better handled centrally by some package which would remove some of the need for the above Mar 04 22:02:38 'night all Mar 04 22:23:38 pb_: that patch looks promising Mar 04 22:23:54 * reenoo plays with it a bit Mar 04 22:26:26 RP: ping Mar 04 22:26:39 03freyther 07org.oe.dev * r478afa57... 10/packages/opie-pimconverter/opie-pimconverter.inc: Mar 04 22:26:39 packages/opie-pimconverter: Mar 04 22:26:39 Fix the RDEPENDS to sqlite-bin instead of Mar 04 22:26:39 libsqlite-bin Mar 04 22:26:47 zecke: pong Mar 04 22:27:23 RP: in the quest of fixing meta-opie I'm now at opie-sshkeys Mar 04 22:27:28 RP: it RDEPENDS on ssh Mar 04 22:27:46 RP: but dropbear is only PROVIDIG ssh and not RPROVIDING it Mar 04 22:28:02 do you have an idea what the best fix could look like? Mar 04 22:30:14 zecke: I guess it also needs to RPROVIDE it. I'm not sure there is much other option... Mar 04 22:30:25 arrgh Mar 04 22:30:41 ~lart whoever bumped PR on glibc for SRC_URI changes Mar 04 22:30:41 * ibot duct-tapes whoever bumped PR on glibc to the floor and drools on him for SRC_URI changes Mar 04 22:31:47 * florian uploads a new oe-sdk Mar 04 22:31:59 That did irritate me a bit too... Mar 04 22:32:32 * RP attempts to get the SA/PXA RTC driver into mainline... Mar 04 22:32:42 RP: Don't worry... i read RP the first eime ;-) Mar 04 22:32:46 eh time Mar 04 22:34:18 reenoo: when changing RPROVIDES, should PR be bumped? Mar 04 22:34:32 zecke: yes Mar 04 22:35:14 but (generally) not when just changing SRC_URI. Mar 04 22:35:25 hi pb Mar 04 22:35:33 hi woglinde Mar 04 22:35:37 hi Mar 04 22:40:49 hi Bernardo Mar 04 22:42:15 03freyther 07org.oe.dev * r679b6743... 10/packages/dropbear/ (dropbear_0.45.bb dropbear_0.46.bb dropbear_0.47.bb): Mar 04 22:42:15 packages/dropbear/dropbear*: Mar 04 22:42:15 Also RPROVIDE ssh and sshd. After this change Mar 04 22:42:15 meta-opie is buildable again. **** BEGIN LOGGING AT Sat Mar 04 22:47:34 2006 Mar 04 22:53:12 phew finally: http://handhelds.org/~florian/sdk/oe-sdk-20060304232054.tar.bz2 Mar 04 22:53:35 Stable tree with new 3.4.4 compiler... Mar 04 22:53:46 all hail florian! Mar 04 22:53:55 excellent Mar 04 22:54:00 ~praise florian Mar 04 22:54:06 not really Mar 04 22:54:12 All hail florian! Mar 04 22:54:13 ~praise RP Mar 04 22:54:15 All hail RP! Mar 04 22:54:23 ~praise RP for fixing up the sdk .bbs Mar 04 22:54:25 All hail RP for fixing up the sdk .bbs! Mar 04 22:55:47 so, it's finally possible to build compatible c++ apps with the sdk again Mar 04 22:56:08 In the end i didn't more than copying that one bb Mar 04 22:56:08 But the result seems to work at least... i just built gpsd and libgpewidget. Mar 04 22:56:59 hrm. wtf is going on with freenode Mar 04 22:59:40 I hope so... Mar 04 22:59:41 We havn't had this for quite sometime now. Mar 04 23:00:03 wow.. whodathunkit. an sdk Mar 04 23:01:10 it has been quite a while Mar 04 23:02:09 what has been quite a while? we've had an sdk for ages. just not a gcc 3.4 one Mar 04 23:02:18 hey how do you include an extra package in gpe-image? Mar 04 23:02:19 Eddy: at image build time or after installation? Mar 04 23:02:19 build time preferably Mar 04 23:02:45 installation time I just copy it over right? Mar 04 23:02:57 the .ipk **** BEGIN LOGGING AT Sat Mar 04 23:04:08 2006 **** BEGIN LOGGING AT Sat Mar 04 23:09:36 2006 Mar 04 23:09:47 the ipkg install does also? Mar 04 23:09:54 Eddy: yes Mar 04 23:09:55 not just the build Mar 04 23:09:58 sweet Mar 04 23:10:03 thanks a lot florian Mar 04 23:10:47 later everyone Mar 04 23:13:20 hi kergoth Mar 04 23:23:41 03freyther 07org.oe.dev * r54075f67... 10/packages/gnuz/gnuz_0.3.bb: Mar 04 23:23:41 packages/gnuz/gnuz_0.3: Mar 04 23:23:41 Update the URL for the gnuz levels to the Mar 04 23:23:41 ewi mirror of hrw. Mar 04 23:30:33 good nite Mar 04 23:32:39 'night zecke Mar 04 23:34:12 'night zecke Mar 04 23:36:08 03freyther 07org.oe.dev * rd9027dec... 10/packages/meta/task-qpe.bb: Mar 04 23:36:08 packages/meta/task-qpe: Mar 04 23:36:08 We have irk targus and irk belkin but no 'irk' Mar 04 23:36:08 package. Fix the depends Mar 04 23:40:50 re Mar 05 00:01:19 feel better now, but heckofalot of kernel-module deps dissatisfaction on bitbake's part ;-) and one libgpg-error0.... Mar 05 00:28:00 'night all Mar 05 00:28:06 'night RP Mar 05 01:49:35 * france is back (gone 02:41:07) Mar 05 01:50:32 * france is away: Away Mar 05 02:18:20 Still can't satisfy tons of task-bootstrap dependencies... Mar 05 03:18:17 guys, does anyone has an idea of how to fix it? http://rafb.net/paste/results/I2oLDj45.html The first part occurred when I ran "bitbake opie-image". Mar 05 03:23:32 Arthur, you should update your bitbake first. Mar 05 03:26:35 Koen told me that. I thought my distro had updated but I was wrong. Sorry for asking before checking. Mar 05 03:32:35 could you tell which is the latest version, please? Mar 05 03:36:57 Arthur: you should download bitbake Mar 05 03:37:01 and run it that way Mar 05 03:37:10 having bitbake in the distro may not be sufficient Mar 05 03:37:38 use at least revision 383 Mar 05 03:53:32 well, I'm running 1.3.3.2 now. Is it enough? Mar 05 03:56:02 Arthur, hmm. you'd better use the svn version . Mar 05 03:57:52 where can I get instruction on how to get the svn version? I don't understand much about version control systems... Mar 05 03:59:24 it's in the wiki Mar 05 03:59:49 22:50 < ibot> well, gettingstarted is Mar 05 03:59:49 http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted Mar 05 04:12:32 thank you. Mar 05 04:12:40 I've updated to revision 384 Mar 05 04:12:49 Let's do the testing "bitbake nano" Mar 05 04:21:31 leoncamel and jnc, the same error after installing bitbake from svn: http://rafb.net/paste/results/q3K7ms57.html Mar 05 04:31:06 does anyone have any idea on that? Mar 05 04:37:19 Arthur: are you sure your BBPATH and python paths and all are set up right? Mar 05 04:37:33 Arthur: perhaps you have 2 bitbakes installed in various places? Mar 05 04:38:31 no. I've removed the one from my distro and the only bitbake is that of /home/arthur/oe/bitbake/bin/ Mar 05 04:38:35 I've already checked. Mar 05 04:43:41 Arthur: perhaps there's an old compiled lib cached somewhere? Mar 05 04:46:28 I doing a find -name bitbake into the whole system to check. Mar 05 04:47:05 don't you have any other shot? Is it surely an outdate package? Mar 05 04:56:38 just my best guess....unless the newest bitbake doesn't work.... Mar 05 04:59:02 heh Mar 05 04:59:17 <_skelley> hi Mar 05 04:59:41 Arthur: strangely bitbake has segfaulted for me for a long time when doing, say, "bitbake nano". But bitbake -i and -b work fine for me Mar 05 04:59:46 Arthur: try bitbake -i Mar 05 05:02:03 <_skelley> How does one get an OE build to use an xscale processor with softfloat? Does OE actually build the cross compiler? Mar 05 05:45:44 hi, folks. please have a look at this : http://rafb.net/paste/results/mOC4yW86.html Mar 05 05:45:47 any comments ? Mar 05 05:55:27 * jnc looks Mar 05 05:56:22 either your bitbake and/or config files are incorrect, I think Mar 05 05:56:49 100%[====================================>] 11,265,381 725.21K/s ETA 00:00 Mar 05 05:56:52 holy crap Mar 05 05:57:02 that was from ftp.kernel.org to my cablemodem connection at home Mar 05 05:57:40 wonder if comcast is hosing the upload speed to accomplish that Mar 05 06:00:47 jnc, could you have a look at http://rafb.net/paste/results/mOC4yW86.html ? Mar 05 06:03:14 leoncamel: i just did Mar 05 06:03:17 and i told you Mar 05 06:03:23 00:47 < jnc> either your bitbake and/or config files are incorrect, I think Mar 05 06:04:58 jnc, hmm, I don't think so. I bitbake boostrap-image successed . Mar 05 06:05:26 jnc, do you know the error locate at line 21~22 ? Mar 05 06:15:36 what branch are you using, revision? Mar 05 06:15:44 where's your config file? Mar 05 06:17:38 jnc, here is local.conf : http://rafb.net/paste/results/kkH23q72.html Mar 05 06:18:45 jnc, here is machine configuration : http://rafb.net/paste/results/KOLVmY89.html Mar 05 06:20:01 jnc, here is distro config : http://rafb.net/paste/results/X2JTor43.html Mar 05 06:28:58 okay Mar 05 06:30:06 hmm, I add PREFERRED_PROVIDERS for machine config, please ignore line 11. Mar 05 06:32:10 jnc, Maybe, a package miss a dependence of "depmod" ? Mar 05 07:04:14 leoncamel: it seems to be working on my system with today's .dev branch, and suggestions per the GettingStarted wiki Mar 05 07:52:40 morning Mar 05 07:53:35 hey Mar 05 08:26:13 so it sounds like familiar has axed OE/bitbake ? Mar 05 08:35:14 ummmm....WTF? Mar 05 08:35:22 what are you talking about? Mar 05 08:37:18 do you mean that they're broken or that familiar is leaving? I haven't heard anything... Mar 05 09:33:10 strange Mar 05 09:33:17 i haven't heard anything about that previously Mar 05 09:53:02 koen is no longer the maintainer of familiar and is unknown whether they'll continue to use OE... Mar 05 10:04:21 yikes Mar 05 10:04:23 RP: hi Mar 05 10:05:11 RP: can i do anything further to nail down what code needs to be added to 2.6.x and make it feature complete versus the sharprom for cxx00? Mar 05 10:06:19 also, i've spotted a CompactFlash format GPRS adapter, if i do end up with it, i may be mailing it around to different devs and get it supported ;) Mar 05 10:12:49 hi Mar 05 10:16:21 hail zecke Mar 05 10:17:33 jnc: We're also most feature complete already. The missing functionality is the pxa27x cpufreq and voltage changing and the USB client RNDIS handling Mar 05 10:17:37 hi zecke Mar 05 10:22:30 okay Mar 05 10:22:50 i understand time is short. i will bugtest anything though Mar 05 10:23:07 03freyther 07org.oe.dev * rd61d5ef3... 10/packages/zshopi/zshopi_0.2.bb: Mar 05 10:23:07 packages/zshopi/zshopi: Mar 05 10:23:08 DEPEND on the native version of opie-lrelease Mar 05 10:28:19 anyone got good experience with fuzzing tools? Mar 05 10:43:04 * zap опять отправляет cs.ozerki.net на перезагрузку, всем сорьки Mar 05 10:57:50 03freyther 07org.oe.dev * r74c13ba7... 10/packages/meta/task-qpe.bb: Mar 05 10:57:52 packages/meta/task-qpe.bb: Mar 05 10:57:54 -qpf-bitstream-vera.bb does not provide itself Mar 05 10:57:56 but -large and -small variants. Apply this knowledge Mar 05 10:57:58 and depend on the -large variant Mar 05 10:58:00 -Sample applies to the mono version of vera Mar 05 10:58:12 03freyther 07org.oe.dev * re30ba676... 10/packages/meta/task-qpe.bb: Mar 05 10:58:12 zecke: running bittest again? Mar 05 10:58:14 packages/meta/task-qpe.bb: Mar 05 10:58:19 There is no qpf-gentium, the closest is a ttf-gentium Mar 05 10:58:39 also: DIE, qpf, DIE Mar 05 10:59:03 koen: no, build a meta-opie-all Mar 05 10:59:20 koen: The, qpf, the? Mar 05 10:59:22 koen: but we have 4gb space now again :} Mar 05 10:59:28 these disks are too small Mar 05 10:59:52 yeah, I should request some budget again Mar 05 10:59:53 koen: is diet-x11 building for you? **** ENDING LOGGING AT Sun Mar 05 10:59:56 2006