**** BEGIN LOGGING AT Mon Mar 13 10:59:57 2006 Mar 13 10:59:57 looks like the problem isn't with my bb or the driver itself, the akita doesn't provide enough power for my wifi dongle Mar 13 11:01:25 need to use powered hub ;( Mar 13 11:01:35 it looks like that Mar 13 11:02:07 with the akita using the charger, it almost works (loads and iwconfig works, but no wifi), using the battery won't work at all Mar 13 11:03:34 should I still post the bb to bugzilla? It should work also with 2.6.x, but, if it needs a powered hub, the driver won't be much help Mar 13 11:04:16 Bernardo: the driver could be useful for slugs and other OE distros Mar 13 11:05:12 XorA: ok, I'll generate the mt diff tonight or tomorrow, and I'll put it in bugzilla Mar 13 11:05:33 I've created bb's both for the cvs version and the "stable" version, using a inc for all common stuff Mar 13 11:26:56 CoreDump|home: re Mar 13 11:27:08 hi, i tried building the .oz354fam083 branch for poodle, but it failed NOTE: package linux-jlime-sh3-2.6.11-r0: task do_compile: failed; buildsystem is opensuse 10 Mar 13 11:27:19 zecke: wb Mar 13 11:27:30 ~zauruskernels Mar 13 11:27:32 rumour has it, zauruskernels is By default Zaurus used 2.4.18 or 2.4.20 which need GCC 2.95.3 to compile - all is described in OE wiki: http://openembedded.org/cgi-bin/moin.cgi/ZaurusKernels Now we work on moving all models to 2.6 kernels: http://openembedded.org/cgi-bin/moin.cgi/ZaurusKernels26 Mar 13 11:28:39 sws: read that wiki page **** BEGIN LOGGING AT Mon Mar 13 13:13:21 2006 Mar 13 13:16:21 | Xcomposite.c:226: error: `XCOMPOSITE_VERSION' undeclared (first use in this function) Mar 13 13:16:29 something to rebuild I think? Mar 13 13:32:57 hrw|work: we need to wait for ph5 Mar 13 13:33:14 hrw|work: maybe remove Xcomposite.h from STAGING_INCDIR as a workaround Mar 13 13:33:58 ok Mar 13 13:43:21 where can I set the path for ccache relative to the build directory? Mar 13 13:44:07 oe is using the directory .ccache in my homedir Mar 13 13:44:21 in local.conf Mar 13 13:44:43 I think there's an example in it on how to do it Mar 13 14:03:43 <_law_> hrw|work, same XCOMPOSITE_VERSION problem here Mar 13 14:04:26 _law_: find tmp/staging/arm-linux/ -name *ompos*|xargs rm;rebuild compositeproto libxcomposite Mar 13 14:04:42 now I have libxpm Mar 13 14:04:56 <_law_> hrw|work, ok thanks i?ll try Mar 13 14:06:45 | /home/hrw/devel/build/0308/tmp/staging/arm-linux/include/X11/Intrinsic.h:174:27: X11/Composite.h: No such file or directory Mar 13 14:12:04 ~seen ph5 Mar 13 14:17:33 hmm, the monotone pulling eat up all my cpu time .... I strongly suggested using git instead of monotone.. Mar 13 14:21:23 morning Mar 13 14:22:04 hi reenoo Mar 13 14:22:43 hey pb__ Mar 13 14:24:11 hi pH5 Rene Mar 13 14:24:50 pH5: libxpm 3.5.4.2 does not build (X11/Composite.h not found). libxpm 3.5.1 builds Mar 13 14:25:42 hi everyone Mar 13 14:26:18 hi ph5 Mar 13 14:28:39 hrw|work: can you show me log? I think that libxpm should not depend on composite in any form. Mar 13 14:29:11 pH5: http://pastebin.com/599719 Mar 13 14:29:46 heh.. looked again - old X11 headers left in staging... Mar 13 14:30:27 mickeyl, koen|sleep: I take it you've disabled my key? Mar 13 14:31:44 huh? Mar 13 14:31:58 enter passphrase for key ID [rw@openembedded.org]: Mar 13 14:31:59 monotone: read from fd 6 (peer ewi546.ewi.utwente.nl) failed, disconnecting Mar 13 14:32:00 i wouldn't do anything like that - at least not without giving notice. Mar 13 14:32:07 enter passphrase for key ID [rw@openembedded.org]: Mar 13 14:32:07 monotone: read from fd 6 (peer monotone.vanille.de) failed, disconnecting Mar 13 14:32:08 let me check the db Mar 13 14:32:14 thanks Mar 13 14:32:15 hrw|work: OE makes leaving behind old header files in staging too easy :) Mar 13 14:32:52 pH5: we need packaging of staging Mar 13 14:33:53 <_law_> hrw|work, same X11/Composite.h problem :-) should i cp manually from libxt-0.1.5-r1/libXt-0.1.5/include/X11/Composite.h ? Mar 13 14:33:56 hrw|work: in this case even versioned DEPENDS would have been good enough, but I agree. Mar 13 14:34:36 _law_: I think that we should rather drop TMPDIR and restart from scratch Mar 13 14:34:38 reenoo: please try again. Mar 13 14:35:11 mickeyl: works, thank you Mar 13 14:35:19 mickeyl: what was the problem? Mar 13 14:35:19 reenoo: i have a slight idea how that came, but I don't support this action. Btw., please try to sort the issues with jnc out. Mar 13 14:35:32 <_law_> hrw|work, this is what i dont wanna hear :-) Mar 13 14:36:17 03rw 07org.oe.dev * rc3ca3e0e... 10/conf/local.conf.sample: local.conf.sample: add hints on ENABLE_BINARY_LOCALE_GENERATION and qemu-native Mar 13 14:36:28 mickeyl: now that I have an email address to that nick I can try to (the list is working as well) Mar 13 14:36:36 reenoo: excellent. Mar 13 14:37:11 reenoo: perhaps we really need to do some kind of challenge-respond via oe@handhelds.org+oe-commits so that we make sure developers are subscribed to both lists. Mar 13 14:37:22 mickeyl: jnc was busy insulting people (zecke) on IRC yesterday Mar 13 14:37:59 i know. that aside, we need to value all contributions and we need to educate folks a bit more diplomatic than just by reverting. Mar 13 14:38:12 agreed Mar 13 14:38:13 jnc == shadows? Mar 13 14:40:28 mickeyl: arguably the education bit should be done before granting push access to avoid this type of crap in the first place. I understand that plain reverting those changesets was a dim plan now though. Mar 13 14:42:58 XorA: so it seems. that was part of the communication problem Mar 13 14:42:59 reenoo: *nod*. usually our policy has been to see what patches people attach to bugzilla, and when we see that they're doing the right things to grant access then. Mar 13 14:43:34 it's good that we actually agree to improve communication anyway. Mar 13 14:43:43 s/improve/try to/ Mar 13 14:43:59 that means also communicating with non-devels or wannabe-devels Mar 13 14:44:15 we will keep getting better :) Mar 13 14:44:57 a list of people (real names + working email (hh.org and hh.org-independent)) with commit access would be a start I guess Mar 13 14:46:04 and we need to fix a couple of misunderstandings. just because it's called .dev and you have push access doesn't mean you can push random untested/broken stuff there Mar 13 14:46:23 maybe one day we will manage kind of forward list on @openembedded.org to make contacting easier Mar 13 14:47:12 sorting out breakage like dropped patches is a major pain because you believe "hey I fixed that last year... wtf?" Mar 13 14:50:09 reenoo: looks like we should improve http://oe.handhelds.org/cgi-bin/moin.cgi/People Mar 13 14:50:55 pH5: on a non-wiki page I should have added Mar 13 14:53:07 reenoo: let first people add themself to People and then someone from proj_oe group on hh.org will move it into static file Mar 13 14:54:24 hrw|work: IMO the registration needs to be part of the "grant push access" procedure Mar 13 14:55:29 but we should not make it so long as NM process in Debian Mar 13 14:56:10 I was thinking few years ago about starting NM to be Debian devel but when I read that it takes years to became one I resigned Mar 13 14:56:33 on average it doesn't take years Mar 13 14:56:42 reenoo: sounds reasonable. also, the people page IMHO should include the oe commit key user names. Mar 13 14:57:32 <_law_> bbl Mar 13 15:03:11 * XorA adds himself to People page Mar 13 15:03:14 hrw|work: it all depends on how talented you are and how quick at responding to mail you and your sponsor are. I know a number of people who've gone through NM fairly quickly. you will notice that I'm not proposing anything like that for OE though. Mar 13 15:03:56 reenoo: I know Mar 13 15:07:13 ~lart moinmoin table syntax Mar 13 15:08:43 hi Mar 13 15:20:23 NOTE: package libxpm-3.5.4.2-r0: task do_build: completed Mar 13 15:21:05 * pH5 is relieved Mar 13 15:21:21 pH5: had to remove libxt headers and rebuild it Mar 13 15:32:50 mickeyl: mail sent Mar 13 15:34:18 france_: morning Mar 13 15:40:03 reenoo: thanks for that. Mar 13 15:44:32 Bernardo: pong (addressed your privmsg, thanks :) Mar 13 15:46:38 * jnc goes "ieeee!" and quickly sets some gmail filters Mar 13 15:49:34 reenoo: nice mail Mar 13 15:55:48 koen|sleep: so file-native is affected by the currently staged file-native? I really shouldn't have to build clean again for this...... Mar 13 16:00:23 anyone played with the alternative backend of qemu ? Mar 13 16:02:56 hrw|work: whom is the bugzilla admin, i would like to update my email address from the serverart one to my gmail address, for consistency Mar 13 16:06:22 jnc: not me Mar 13 16:06:24 jnc: ggilbert Mar 13 16:07:03 thanks pb Mar 13 16:08:06 work time, i will be around later tonight (CST) to follow up on changes to libgcrypt/gnutls Mar 13 16:12:50 hi all Mar 13 16:13:20 hi florian_kc Mar 13 16:13:52 Anyone an idea why bitbake decides to use ggaim_cvs.sourceforge.net__20060204.tar.gz for aim-2.0.0-cvs-mco1-20050904? Mar 13 16:15:00 pH5: hey Mar 13 16:15:10 hey zecke Mar 13 16:15:11 pH5: I don't know if anyone else reported it Mar 13 16:15:31 pH5: XComposite is still broken (hrw,CoreDump), I have breakage in libXfont Mar 13 16:15:39 libxpm? Mar 13 16:16:06 zecke: they are broken mostly by the fact that xorg7 moved some headers Mar 13 16:16:27 I'm not sure if we can do anything about that. Mar 13 16:16:32 pH5: http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=4&logfile=20060312201243.log Mar 13 16:16:52 hrw|work: yes but the local header should get included, not the global one Mar 13 16:17:31 pH5: might be just a missing dependency on freetype2 Mar 13 16:20:17 zecke: hmh, I think you're right.. Mar 13 16:28:48 03pH5 07org.oe.dev * r734ad589... 10/packages/xlibs/libxfont_X11R7.0-1.0.0.bb: libxfont_X11R7.0-1.0.0: add freetype to DEPENDS Mar 13 16:30:20 ~lart gtk/atk libtool Mar 13 16:33:55 reenoo: morning Mar 13 16:34:01 pb_: ping Mar 13 16:34:39 :-( Mar 13 16:34:48 * JustinP misses ibot Mar 13 16:35:53 pH5: I will start a compile on ewi now Mar 13 16:36:19 pH5: you can watch the log on the tinderbox Mar 13 16:37:07 france: hello Mar 13 16:37:28 cu Mar 13 16:39:34 zecke: great, how do I find out the new timestamp / log file name? Mar 13 16:39:42 * pH5 doesn't know a thing about tinderbox Mar 13 16:43:51 pH5: ewi546.ewi.utwente.nl/tinderbox -> Click -> View/Show Mar 13 16:43:57 pH5: and look for the zecke column Mar 13 16:44:14 it is now parsing Mar 13 16:45:50 http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=4&logfile=20060313175122.log Mar 13 16:45:55 follow that link Mar 13 16:46:06 (well it is a clobbered build - Opie already built freetype) Mar 13 16:49:10 zecke: got it, thanks Mar 13 16:50:21 pH5: oh and it failed Mar 13 16:51:06 zecke: you are the sheriff? eleet Mar 13 16:51:30 pb__: could you hook me up with a Use Case Diagram of a Sheriff - so I can judge my role Mar 13 16:51:50 pb__: well in the absence of a build irc bot, I will play the role Mar 13 16:52:50 cya later, pals Mar 13 16:53:08 zecke: later Mar 13 17:04:38 hi ibot! Mar 13 17:07:05 heh Mar 13 17:07:08 * pb__ stabs ibot Mar 13 17:07:15 ~emulate stewie Mar 13 17:07:17 It's like an orgy in my mouth! Mar 13 17:07:49 03pH5 07org.oe.dev * r08d3b9db... 10/packages/xserver/ (xserver-kdrive_20060312.bb xserver-kdrive_cvs.bb): xserver-kdrive_{20060312,cvs}: add videoproto to DEPENDS, a tinderbox build failed with a missing Xv.h Mar 13 17:07:59 Who owns ibot? Mar 13 17:08:13 tim riker Mar 13 17:08:39 pb__: I figured that much out on my own... what's his nick? Mar 13 17:08:44 timriker Mar 13 17:08:56 pb__: ahh... Mar 13 17:34:35 hi all Mar 13 17:35:41 hi koen Mar 13 17:38:31 hey NAiL Mar 13 17:45:37 Building with OE seems to be witchcraft =) Mar 13 17:48:05 Cockroach-: undocumented wicthcraft Mar 13 17:49:21 indeed Mar 13 17:50:08 for some reason I got things to compile at work, but can't seem to reproduce it at hoe Mar 13 17:50:11 *home Mar 13 17:50:14 ...hoe :) Mar 13 17:51:12 glibc-initial_2.3.2+cvs20040726 doesn't want to compile Mar 13 17:51:24 for x86 architecture Mar 13 18:08:16 Cockroach-: what kind of error do you get? Mar 13 18:08:49 koen: I guess it does work, though....I still don't really understand why i have to rebuild, though.... Mar 13 18:09:15 JustinP: becoasue file-native was broken before the fix Mar 13 18:09:21 because* Mar 13 18:09:35 hmmm....but the build should still *work* though...shouldn't it? Mar 13 18:09:47 unless it's using the staged file....but it shouldn't be...right? Mar 13 18:09:54 the build works, as long as you don't have any broken version in staging Mar 13 18:10:37 well, yea, it seemed to work this time..... Mar 13 18:10:49 oh well I guess.....rebuild glibc and gcc yet again.... Mar 13 18:15:41 FYI: glibc 2.4 has been released Mar 13 18:16:10 so when linux 2.6.16 gets releeased all eabi support is available from release version Mar 13 18:40:46 re Mar 13 18:41:16 pH5: hey, do you have time to continue the X.org7 building? Mar 13 18:41:40 zecke: bring it on Mar 13 18:42:11 pH5: should Xv be build? should kdrive depend on it? Mar 13 18:42:21 pH5: do you want to see a config.log? Mar 13 18:42:57 zecke: I already added videoproto, which should have Xv.h, to the kdrive depends, although it doesn't seem to look for it in configure.ac Mar 13 18:52:56 I don't think kdrive has any use for Xv. Mar 13 18:53:16 afaik, that's pretty specific to xfree86/xorg Mar 13 18:54:17 pH5: how to proceed? Mar 13 18:54:41 pb_: unbelievable X has no decent support for Display Postscript Mar 13 18:55:27 zecke: yeah, that is the killer feature that drives me to use qte Mar 13 18:55:57 pb_: and what about the NeWS compability of QtE? Mar 13 18:56:16 zecke: it is like oxygen to me Mar 13 18:56:30 without NeWS compatibility I am nothing Mar 13 18:56:55 Gosling just rocks :} Mar 13 18:56:56 We could either just try another build and see if it builds (did for me, but as we see that doesn't mean a thing) or put --disable-xv into EXTRA_OECONF and -- try another build and see if it works. Mar 13 18:57:54 pH5: I will try again Mar 13 18:57:58 ah damn, still no batteries for my stupid mouse Mar 13 18:58:04 * pb_ throws it out the window again Mar 13 18:58:37 pb_: write it on the todo list right next to "Fix the Window of my car" Mar 13 18:58:40 clearly my policy of benign neglect is not yielding the dividends I hoped Mar 13 18:58:46 pb_: Opie has a great Todolist you could try Mar 13 18:58:58 zecke: funnily enough I have booked an appointment for my car to have a new window fitted Mar 13 18:59:01 on the 27th of this month Mar 13 18:59:25 a whole new roof, actually, not even just a window Mar 13 18:59:42 your car must be pretty broken Mar 13 18:59:52 yeah, it's quite a wreck Mar 13 18:59:55 so you should try on the 26th to get rid of the cats... Mar 13 19:00:11 right, it would be a nightmare to fit the new window and then find that they were on the inside Mar 13 19:00:33 pH5: which kdrivers versions should be buildable? Mar 13 19:00:36 I think they mostly come under cover of darkness though, so this may be a challenge Mar 13 19:01:47 zecke: 20060312 and cvs. 20050207 too, if you follow the comments in the .bb file (change DEPENDS, uncomment a patch) Mar 13 19:02:42 I also bought a new gps system, which I figure should make me pretty much invincible. Neither inclement weather, cats, nor complicated routes will hold any fear for me now. Mar 13 19:02:43 * mickeyl wanders home Mar 13 19:04:20 pH5: failed Mar 13 19:04:29 now trying extra_oeconf Mar 13 19:05:58 zecke: It built videoproto and STAGING_INCDIR/X11/extensions/Xv.h doesn't exist? Mar 13 19:06:47 I have Xv there Mar 13 19:07:03 still it can't include it Mar 13 19:07:12 hmm I'm looking at the include dirs now Mar 13 19:07:20 which source file is it that wants Xv.h? Mar 13 19:07:32 Xext/xvmain.c Mar 13 19:07:41 see http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=4&logfile=20060313200822.log Mar 13 19:08:18 ? Mar 13 19:08:25 sorry Mar 13 19:08:34 I just saw the 'p' Mar 13 19:09:48 ~herring zecke Mar 13 19:09:52 * ibot whacks zecke on the side of the head with a large red herring named alfred Mar 13 19:09:55 ls tmp/staging/arm-linux/include/X11/extensions/Xv.h tmp/staging/arm-linux/include/X11/extensions/Xv.h Mar 13 19:10:11 and that dir is in the include path Mar 13 19:10:45 pH5: I have a videoproto 2.2.2 package Mar 13 19:11:23 zecke: do you understand this? Mar 13 19:11:37 oh, that is strange Mar 13 19:11:42 As I see it, Xv.h is in place, the build command has the correct include path and it still fails? Mar 13 19:11:42 pH5: no Mar 13 19:11:52 pb_: that is X ;) Mar 13 19:12:05 pH5: I'm trying strace now Mar 13 19:12:19 .. Mar 13 19:12:29 Guest210: quite so Mar 13 19:12:35 zecke: it seems Mar 13 19:12:53 pb_: any better option than strace? Mar 13 19:13:15 pH5: am I building the right version of xserver? Mar 13 19:13:16 none that I can think of Mar 13 19:13:17 is there anyone who know about TI's DSC25? Mar 13 19:13:27 that Xv.h isn't a broken symlink or anything like that, presumably Mar 13 19:13:45 pb_: I tried to open it, it worked (C) 1991 DEC Mar 13 19:14:01 that might be a trick of vim, I should try (x)emacs Mar 13 19:14:25 good plan, xemacs should have more of an affinity with this code Mar 13 19:14:32 strace does sound like your best option though Mar 13 19:15:47 zecke: it's the right version - I just noticed that I accidentally copy&pasted the PR from another .bb file.. Mar 13 19:16:11 pH5: also that PACKAGE_ARCH_poodle = "poodle" looks odd Mar 13 19:17:53 probably no worse than anything else that mentions poodle Mar 13 19:18:44 pH5: http://oe.pastebin.com/600215 Mar 13 19:18:53 only trace of Xv.h and it looks good Mar 13 19:19:18 strace -f -o strace.log make Mar 13 19:19:39 hmm arm-linux-gcc does not look for the header Mar 13 19:19:46 hmm could that be a ccache bug? Mar 13 19:22:18 weird Mar 13 19:22:24 now saveproto is failing Mar 13 19:22:47 s/failing/missing Mar 13 19:25:12 gnn, shouldn't they check for this stuff in the configure script.. Mar 13 19:25:27 X folks... ;) Mar 13 19:25:58 pH5: now it stopped failing on Xv.h it decided to fail on saverproto Mar 13 19:26:28 so let's add scrnsaverproto to kdrive DEPENDS Mar 13 19:28:13 pH5: will you add it and push? or should I add it locally? Mar 13 19:28:36 pH5, zecke: I'm probably missing something, but have you guys looked into Gentoo's ebuilds for Xorg7 modular? AFAIK they have it all in portage and working... Mar 13 19:29:17 JustinP: I'm just playing sheriff :) Mar 13 19:29:29 JustinP: I think it is a matter of fixing the DEPENDS Mar 13 19:29:49 JustinP: can you tell me where to find a gentoo ebuild for xserver-kdrive? Mar 13 19:29:57 sure....my point was that the ebuilds probably have all of the depends already set up Mar 13 19:30:15 oh...well....not that one I guess...I figured you were working ont he rest of it as well Mar 13 19:30:19 * JustinP goes back to work Mar 13 19:30:31 JustinP: hehe Mar 13 19:32:08 zecke: I guess it's faster if you add it locally and retry Mar 13 19:32:34 zecke: but let me grep "X11/" Xext/*.c first, I suspect there are more missing dependencies Mar 13 19:32:47 at your order Mar 13 19:33:26 * pH5 feels like the bad gunman ordering the sherrif around Mar 13 19:33:51 hehe Mar 13 19:33:54 meanwhile the sheriff pets his pussy :} Mar 13 19:33:55 /usr/portage/x11-base/kdrive Mar 13 19:34:22 probably not xorg7-ified, though Mar 13 19:34:52 NAiL: do you have a description on how to open the nslu? Mar 13 19:35:10 zecke: in the wiki Mar 13 19:35:19 koen: I have not found one :} Mar 13 19:35:25 Why is it I go away for 5 minutes and my server breaks... :-/ Mar 13 19:35:32 hey RP_ Mar 13 19:35:35 RP_: murphy Mar 13 19:35:38 RP_: or evil gnomes Mar 13 19:35:50 RP_: don't use Windows Vista Beta Mar 13 19:36:02 zecke: http://www.nslu2-linux.org/wiki/HowTo/OpenTheCase Mar 13 19:36:05 zecke: I never planned to ;-) Mar 13 19:36:33 then again...the kdrive ebuild does have x-modular in its inherits... Mar 13 19:36:39 * JustinP shrugs Mar 13 19:37:05 JustinP: it was worth a try Mar 13 19:37:17 koen: way too many steps, is there a flash movie available? Mar 13 19:38:09 heh Mar 13 19:38:35 you can do as I did, read it after you have opened the case Mar 13 19:38:56 ~lart OSX pl2303 drivers Mar 13 19:38:57 * ibot slaps OSX pl2303 drivers around with a large trout Mar 13 19:39:04 I will clean my desk first Mar 13 19:41:08 pH5: bero will put his eabi toolchain .bbs in bugzilla soon(TM) Mar 13 19:49:27 koen: glad to hear that! Mar 13 19:49:44 zecke: It's not very hard to do. Just gently squeeze the nslu where the silver part meets the black. Press the black part gently, and it should be somewhat easy to wiggle apart. Mar 13 19:49:58 (ie, press the perforated part) Mar 13 19:50:44 zecke: can you try to just add scrnsaverproto to kdrive dependencies and rebuild? I checked the other *proto packages, and they should be either pulled by libx11 or unneeded with the current configuration Mar 13 19:51:35 has anyone expiremented with the --as-needed linker flags yet? Mar 13 19:55:17 koen: no, what do you expect from this? Mar 13 19:56:35 pH5: less spurious deps Mar 13 19:57:33 koen: that doesn't sound like a bad thing Mar 13 19:58:19 pH5: it will only link to stuff it needs, not to stuff you tell it to link to Mar 13 20:04:32 pH5: will try Mar 13 20:04:41 hmm I don't find my switch :} Mar 13 20:15:30 pH5: built Mar 13 20:15:54 how cruel, my build just finished parsing Mar 13 20:26:06 03pH5 07org.oe.dev * r6916a0e4... 10/packages/xserver/ (xserver-kdrive_20060312.bb xserver-kdrive_cvs.bb): xserver-kdrive_{20060312,cvs}: add scrnsaverproto to DEPENDS Mar 13 20:42:22 I don't recognize any of the names on the CELF confrence list as having anything to do with OE - did nobody submit a proposal? :( Mar 13 20:44:27 CosmicPenguin: looks like lrg is on the list, though not with anything OE related Mar 13 20:45:40 personally, I... er, forgot all about it Mar 13 20:46:01 heh Mar 13 20:46:16 ~herring pb_ Mar 13 20:46:20 * ibot whacks pb_ on the side of the head with a large red herring named alfred Mar 13 20:48:35 koen|away: man dutch universities are poor, they can not even afford reasonable size of storage Mar 13 20:52:49 mithro: thanks for your postcard Mar 13 21:00:07 zecke: correct, we can't even afford a proper EE lab Mar 13 21:00:23 zecke: so they put a 3MW fm transmitter on the roof Mar 13 21:02:48 I'm freeing space on ewi again Mar 13 21:03:00 Does anyone in here have access to, or could make for me, an image of the OZ 3.5.4 for an SL-C1000 / SL-C3100? Mar 13 21:06:31 Jason_Foss: welcome, images are curently being build. They might have SL-C1000 and SL-C3100 Mar 13 21:06:57 Jason_Foss: I'm not too sure though, the best option is to wait for the OZ3.5.4 builds and see if the SL-C3100 is supported Mar 13 21:08:06 Yeah. Mar 13 21:08:56 I'd be perfectly happy with any kind of 3.5.4 images, though...even the RC ones. Mar 13 21:09:20 Jason_Foss: official builds are not too far away Mar 13 21:10:05 Jason_Foss: oh, see the front page of openzaurus.org Mar 13 21:10:20 Yeah, that's why I'm antsy. Mar 13 21:10:43 Jason_Foss: I don't want to interfer with the OpenZaurus guys here Mar 13 21:11:55 I'm just presently heading into "crunch time" testing an NSA project that we're presenting in less than a month... Mar 13 21:12:43 Jason_Foss: understandable, I don't know what these 2.4.20 issues are Mar 13 21:13:01 Jason_Foss: it might be wise to either wait for RP to turn up, drop a mail on the openzaurus mailing list Mar 13 21:13:11 Jason_Foss: or come at european daytime and talk with hrw Mar 13 21:13:49 And presently, the only way I've managed to get GPE on my Zaurus is to hack the pdaXrom loader to put Familiar on it. Obviously, not the "best" option ;-) Mar 13 21:14:33 hehe Mar 13 21:14:58 Jason_Foss: you could build your own image (might not be possible due your time constraints) Mar 13 21:16:01 I *could*, but, I don't have OE setup...nor do I have any clue as to how I'd go about doing that (I could figure it out, but...that wouldn't be a quick process) Mar 13 21:16:03 Jason_Foss: 3.5.3 has gpe images Mar 13 21:16:20 3.5.3 doesn't support the SL-C3100 Mar 13 21:16:50 Nor does it have GPE images for the SL-C1000 Mar 13 21:17:13 OPIE isn't X-based, so my GUI won't run on it. Mar 13 21:18:06 zecke: can I run Xinerama on the ROAD to make the 2 displays behave as one? Mar 13 21:18:36 koen: hehe, I doubt it Mar 13 21:19:06 koen: the front display is slow and you can only updates row Mar 13 21:19:14 koen: and it is not a color display... Mar 13 21:19:38 koen: but, as the device will be reflashable you are welcome to try it Mar 13 21:20:10 zecke: of course I'll 0v3rlcl0ck it with g3nt00 Mar 13 21:20:27 it will make much faster phonecalls ;) Mar 13 21:20:50 koen: you might consider running gentoo on the wireless modules (an ARM as well) Mar 13 21:21:10 zecke: softMAC? Mar 13 21:22:31 pardon? Mar 13 21:22:34 the wifi module is probably armv7 Mar 13 21:22:43 koen: ah the next ARM Mar 13 21:22:48 conexant softmac wifi module? Mar 13 21:22:53 ehm, arm7 that is Mar 13 21:23:00 stupid arm numbering Mar 13 21:23:01 koen: no conexant Mar 13 21:23:51 Jason_Foss: do you only want a gpe system? Mar 13 21:24:06 Jason_Foss: you could create a gpe feed and install it Mar 13 21:25:54 What do you mean? Mar 13 21:26:42 Jason_Foss: I'm not too sure about your reasoning. If you simply want more modern GPE applications Mar 13 21:27:04 Jason_Foss: you could build them with OE and install them onto your 'bastardized' pdaXrom Mar 13 21:27:15 zecke: que? Mar 13 21:27:25 zecke: how do you install stuff into a ROM? Mar 13 21:27:47 Jason_Foss: but the best option is to either mail OpenZaurus mailing list, or wait for the release master hrw Mar 13 21:27:56 koen: puch it bitwise into the holes? Mar 13 21:27:59 <- is avoiding writing his report Mar 13 21:27:59 punch even Mar 13 21:29:22 Well, the problem is that I need to have some basic rudimentary comtrol over the device itself... Mar 13 21:29:32 Specifically, the ability to dim the screen. Mar 13 21:30:44 Jason_Foss: you can dim the screen with slbl.... Mar 13 21:30:57 slbl? Mar 13 21:31:01 Jason_Foss: your best bet would be to build an image yourself. It's not too bad Mar 13 21:31:06 Jason_Foss: from slutils Mar 13 21:31:17 Jason_Foss: you can also use the utilities in GPE or Opie Mar 13 21:31:50 The utilities in GPE don't work because the version I'm running isn't designed for the hardware I'm running it on. Mar 13 21:32:04 I'd gladly build one for you but my build machine is somewhat slow.... Mar 13 21:32:23 well.....do you have a linux box with a few GB of RAM and some spare time? Mar 13 21:32:27 excuse Mar 13 21:32:31 few GB of HD Mar 13 21:32:41 and 512MB+ RAM Mar 13 21:33:14 Possibly. What I don't have is aaaaany clue how to set up OE. Mar 13 21:33:18 (few being 10GB or so...) Mar 13 21:33:23 Jason_Foss: it's all on the website Mar 13 21:33:29 oe.handhelds.org Mar 13 21:33:55 Yeah...the instructions scared me. Mar 13 21:34:16 http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted Mar 13 21:34:48 Unfortunately, the only computer in this room that has internet access is this one, which is a Windows machine. Mar 13 21:35:01 But if you built an image for me, you'd pretty much be my hero ;-) Mar 13 21:35:06 the use colinux :) Mar 13 21:35:37 You're all speaking Greek to me. *lol* Mar 13 21:35:57 well....I only have a 700Mhz Celeron for building.... Mar 13 21:36:24 Jason_Foss: what machine do you need? Mar 13 21:36:51 SL-C1000 and Sl-C3100 Mar 13 21:36:54 akita or borzoi? Mar 13 21:37:01 why *and*? Mar 13 21:37:04 you have both? Mar 13 21:37:08 Yes sir. Mar 13 21:37:23 well...only kernel should be different between them I think...and maybe fstab Mar 13 21:37:47 * JustinP starts a 3.5.4 gpe-image for akita Mar 13 21:37:50 As well as two Ipaqs...one of which I fight with on a daily basis. Mar 13 21:37:56 heh Mar 13 21:38:01 well, I'm a Zaurus person Mar 13 21:38:06 I ahve a spitz myself Mar 13 21:38:10 Damned thing randomly decided it wasn't going to mount SD cards anymore. Mar 13 21:38:16 s/ahve/have/ Mar 13 21:38:16 JustinP: 3.5.4 doesn't support cxxx IIRC Mar 13 21:38:27 "doesn't support" meaning that we won't release an image Mar 13 21:38:31 it *works* Mar 13 21:38:39 it's just not stable and has quirks Mar 13 21:38:45 at least last I tried Mar 13 21:38:56 should be "works", not *works* Mar 13 21:39:05 2.6 kernel will be much better Mar 13 21:39:34 JustinP: just pointing out that the image you're building isn't "3.5.4" Mar 13 21:39:42 Jason_Foss: if I do build this you need to know that it's unofficial and not supported Mar 13 21:39:45 true Mar 13 21:39:56 That's fine with me. Mar 13 21:39:58 where can i set the evrsion.... Mar 13 21:40:13 It's more official and unsupported than what I have now, I can assure you ;-) Mar 13 21:40:15 JustinP: use the oz354fam083 branch :) Mar 13 21:40:23 of course I am! Mar 13 21:40:34 for cxxx I'd use .dev Mar 13 21:40:42 otherwise you'd end up with 2.4-crappix Mar 13 21:40:45 ack Mar 13 21:40:48 ok, then Mar 13 21:41:00 * JustinP will use dev....2.6 kernel *is* better Mar 13 21:41:56 a pool of maggots is better than 2.4-crappix Mar 13 21:42:01 images for the c1000, c3100, c3000 will not ship with OZ 3.5.4 Mar 13 21:42:10 They will ship with 3.5.4.1 Mar 13 21:42:11 we know Mar 13 21:42:29 Did someone spell that out for Jason_Foss? Mar 13 21:42:34 Yeah. Mar 13 21:42:58 I don't need it to make my PDA do lots of stuff. I just need it to run a GUI and let me dim the screen. Mar 13 21:43:00 RP_: .fi has internet? Mar 13 21:43:01 ok, ignore me then :-/ Mar 13 21:43:12 koen, it would appear so, yes :) Mar 13 21:43:24 It has lots of snow as well Mar 13 21:43:37 nice Mar 13 21:43:46 I heard it has lots of darkness as well Mar 13 21:44:07 So does North East England so I'm use to that :) Mar 13 21:44:34 The NE UK actually feels colder than here due to the windchill... Mar 13 21:45:05 I just want to know who's wise idea it was to put a hotel next to a hydroelectric powerstation... Mar 13 21:45:21 heh **** ENDING LOGGING AT Mon Mar 13 21:45:31 2006 **** BEGIN LOGGING AT Mon Mar 13 21:46:11 2006 Mar 13 21:46:32 RP_: so it's just like home, except you have no clean clothes and can't sleep? Mar 13 21:46:57 koen, That sums it up, yes :) Mar 13 21:48:04 I will also be doing some 2.6 kernel builds from .dev for spitz soon (as in perhaps over the weekend). These will run on the C3100 (borzoi) Mar 13 21:48:22 Yeah. Mar 13 21:48:27 I appriciate it. Mar 13 21:48:55 Post to the one of the OZ mailing lists so I know where to mention any images I create Mar 13 21:49:22 3.5.4.1 will be branch based but having .dev tested won't do us any harm... Mar 13 21:50:16 RP_: spitz 2.6 == borzoi 2.6? Mar 13 21:50:37 JustinP: Effectively, yes. Mar 13 21:51:18 So, someone mentioned a way I can dim the screen on the Zaurus? Mar 13 21:51:33 Put it in a drawer Mar 13 21:51:34 RP_: I have a draft release policy document for Angstrom which makes it a lot easier to 'inject' new devices and kernels into a release Mar 13 21:51:38 ummm....forget about it... ;-) Mar 13 21:52:00 Jason_Foss: you *can* use the slutils, but you'd have to build it Mar 13 21:52:00 Jason_Foss, 2.4 or 2.6 kernel? Mar 13 21:52:06 2.4 Mar 13 21:52:32 JustinP is right, forget it. Start with a 2.6 kernel. You won't regret it Mar 13 21:52:43 koen, That sounds very useful Mar 13 21:53:20 koen, Does this also cover things like image compatibility between 2.4 and 2.6 kernels? Mar 13 21:53:21 RP_: after mickeyl is done reviewing it we'll post it somewhere public Mar 13 21:53:32 RP_: no Mar 13 21:53:34 The Familiar version I'm using on my iPaq came with a program called bl that does it for me. Mar 13 21:53:42 RP_: as eabi is still 2.6 only Mar 13 21:54:22 koen, good point. This does looks like a good oppertunity to give all the 2.4 based devices a good push :) Mar 13 21:55:42 Justin, about how long does it take to build an image? Mar 13 21:56:24 Jason_Foss: I'm starting from nothing built but glibc....will take a while. Some hours. Mar 13 21:56:45 how can I check endianness ? Mar 13 21:56:56 RP_: glibc 2.4 is out and bero has a .bb for it :) Mar 13 21:57:04 or does anyone know big or little for an sl-6000 Mar 13 21:57:23 koen, Yay :). Has he tried it on a device yet? Mar 13 21:57:31 RP_: not yet Mar 13 21:57:37 REdOG: LE Mar 13 21:58:10 REdOG, run file against a binary and it will tell you Mar 13 21:58:23 but its LE as koen says Mar 13 21:59:03 heh, my version doesn't even have file Mar 13 21:59:06 thanks Mar 13 21:59:39 you can run it as EB as well Mar 13 21:59:47 the cpu doesn't care Mar 13 22:01:39 huh? Mar 13 22:02:31 the cpu doesn't have a fixed endiannes Mar 13 22:02:46 the bootloader and kernel just used it as little endian by default Mar 13 22:02:49 oh, well how does it know the diff Mar 13 22:03:06 but nothing is stopping you from running it in big endian mode Mar 13 22:04:17 except that I do not know how to switch or check its modes Mar 13 22:04:25 :-) Mar 13 22:05:21 I haven't seen any report of someone running it as big endian, so LE would be a very good guess Mar 13 22:05:40 hehe, ok ill stick with that thanks again Mar 13 22:21:34 Justin, you still here? Mar 13 22:25:15 Jason_Foss: yup Mar 13 22:25:21 Jason_Foss: up to the kernel Mar 13 22:28:20 ljp: wb Mar 13 22:29:44 'night all Mar 13 22:32:41 dang quickcam module p.o.s Mar 13 22:34:27 ah sheize Mar 13 22:34:37 drivers/net/wireless/hostap/hostap_ap.c:1604: error: stray '`' in program Mar 13 22:44:10 all: Who is playing with IE and the Cirrus Logic EP-93xx Development Platform? Mar 13 22:44:19 OE, rather :-) Mar 13 22:44:20 hi Mar 13 22:44:25 hi Mar 13 22:46:49 That sounds like an accusation Mar 13 22:49:17 IE ??? Mar 13 22:49:44 likewise me rarelly ... but maybe on #ep93xx u can find more info Mar 13 22:50:14 any news on the "NOTE: package libxcomposite-X11R7.0-0.2.2.2: failed Mar 13 22:50:14 " front? Mar 13 22:50:35 OE, not IE guys. Oh what a typo :-) Mar 13 22:50:48 gremlin[it]: thanks Mar 13 22:52:29 likewise just to know which board with ep93xx u have ? Mar 13 22:55:03 CoreDump|home: hmm Mar 13 22:55:11 CoreDump|home: ph5 is away again Mar 13 22:55:16 =) Mar 13 22:55:32 CoreDump|home: remove your XComposite.h from staging Mar 13 22:55:41 zecke: will try Mar 13 22:57:04 mhentges@oe-akita-head:..Embedded/tmp.dev/staging >find . -name "*XComposite.h*" Mar 13 22:57:04 mhentges@oe-akita-head:..Embedded/tmp.dev/staging > Mar 13 22:57:27 hmm how do I tell "find" to ignore CaSe? Mar 13 22:58:11 -iname Mar 13 22:58:32 excellent, thanks! Mar 13 23:00:12 NOTE: package libxcomposite-X11R7.0-0.2.2.2: completed Mar 13 23:00:20 yay =) thanks zecke! Mar 13 23:00:23 ~curse screen Mar 13 23:00:25 May the fleas of a thousand camels infest your most sensitive regions, screen ! Mar 13 23:00:44 * CoreDump|home hugs screen Mar 13 23:01:23 I love screen, but it keeps locking up and I lose my session Mar 13 23:01:33 last time it was because I accidentally pressed Ctrl-S Mar 13 23:01:39 JustinP: I hate it when that happens Mar 13 23:01:48 Ctrl-Q fixes Ctrl-S Mar 13 23:02:09 and it often kills the screen I'm on when I'm on a bad connection (aka cell phone) Mar 13 23:02:13 ah... Mar 13 23:02:15 * JustinP makes a note Mar 13 23:03:00 if it weren't for these random freezes, screen would be my #1 program heh Mar 13 23:15:50 03yuecelm 07org.oe.dev * r831d6459... 10/packages/neon/ (neon_0.26.0.bb neon_0.24.7.bb): Mar 13 23:15:52 neon-0.26.0: enabled https support Mar 13 23:15:54 - add gnutls to DEPENDS Mar 13 23:15:56 - set DEFAULT_PREFERENCES = "-1" Mar 13 23:16:00 CTRL-A CTRL-D, then exit your connection :-) Mar 13 23:16:37 how can I find out the dependencies of a package (librsvg) in openembedded? Mar 13 23:17:38 likewise: no....doesn't help Mar 13 23:18:00 likewise: bitbake -n libsrvg Mar 13 23:18:18 * JustinP sighs Mar 13 23:18:22 I think this machine is dying Mar 13 23:18:46 I was getting random errors building my kernel yesterday which went away if I retried....then it would hit another one Mar 13 23:18:56 eventually finished.... Mar 13 23:19:26 Jason_Foss: yes, I am still building ;-) Mar 13 23:25:23 03freyther 07org.oe.dev * r284d4243... 10/classes/tinderclient.bbclass: Mar 13 23:25:25 classes/tinderclient.bbclass: Mar 13 23:25:27 Use the MultipleProvider and NoProvider Event to create Mar 13 23:25:29 reports to the TinderBox. Mar 13 23:25:31 The message format can be improved now we at least know Mar 13 23:25:33 if a build failed due a missing provider Mar 13 23:37:07 is diet-x11 compiling? Mar 13 23:37:32 I have seen we have libs from the latest xorg now, but doesn't compile anymore Mar 13 23:38:02 guillermo_: it is fixed Mar 13 23:38:06 03freyther 07org.oe.dev * r82466044... 10/conf/bitbake.conf: Mar 13 23:38:08 conf/bitbake.conf: Mar 13 23:38:10 guillermo_: ph5 fixed it Mar 13 23:38:10 cpio calls makeinfo which is part of texinfo on my Mar 13 23:38:12 Ubuntu installation. This is why I add texinfo-native Mar 13 23:38:14 as ASSUMED_PROVIDED package to bitbake.conf. Mar 13 23:38:26 03freyther 07org.oe.dev * r200e22ce... 10/packages/cpio/cpio_2.5.bb: Mar 13 23:38:28 packages/cpio/cpio_2.5.bb: Mar 13 23:38:30 Add texinfo-native to the DEPENDS of this Mar 13 23:38:32 package as makeinfo is used in do_install Mar 13 23:39:40 great thx Mar 13 23:55:44 Anyone alive? Mar 13 23:56:10 Jason_Foss: just ask Mar 13 23:57:49 Just wondering if anyone was alive ;-) Mar 13 23:58:04 I just got back from a hardware run. Mar 13 23:58:34 Trying to get a wireless card that'll work with these stupid Zauruses... Mar 13 23:59:09 you should have asked before you bought....or checked ut the mailing list archives for every other time this was asked Mar 13 23:59:19 or checked Linux compatibiity lists.... Mar 13 23:59:29 LinkSys WCF12 is known to work Mar 13 23:59:39 Ambicom is also known to work Mar 14 00:00:57 That's actually the card I have in my hand...the WCF12 Mar 14 00:01:57 When we bought the Zauruses, we bought cards that worked w ith the default OS...but the cards aren't supposed in non-Qtopia based OSes as of yet. Mar 14 00:03:58 ummm...Qtopia isn't an OS Mar 14 00:04:03 which card? Mar 14 00:04:17 it's all Linux Mar 14 00:04:23 Planex GW-CF11X Mar 14 00:05:05 did it come with binary drivers or soemthing? Mar 14 00:05:33 It came with windows drivers. Mar 14 00:06:04 ummm...Zaurii aren't windows.... Mar 14 00:06:12 Yes, I know. Mar 14 00:06:16 so it just worked in the Sharp ROM? Mar 14 00:06:21 g or b? Mar 14 00:06:29 It's a B. Mar 14 00:06:46 And the card is recongised and works under OPIE or Qtopia. Mar 14 00:08:03 then it works in Linux.... Mar 14 00:09:42 it doesn't matter what UI you're running Mar 14 00:09:56 likely you just don't have the right modules in your hacked OZ rom... Mar 14 00:10:15 It's a Familiar rom, actually. Mar 14 00:14:15 (not that it makes a difference...but...) Mar 14 00:19:33 Don't we have an OE elevator pitch somewhere on the twiki? Mar 14 00:19:38 and why am I so stupid as to not be able to find it Mar 14 00:20:00 ? Mar 14 01:22:03 great....X11R7 is now the default in dev..... Mar 14 01:22:08 * JustinP waits for the inevitable Mar 14 01:25:33 crash? Mar 14 01:34:48 03yuecelm 07org.oe.dev * rbc5b4df3... 10/packages/neon/neon_0.26.0.bb: neon-0.26.0: fixed typo, thanks to Justin Patrin Mar 14 01:55:31 JustinP, sounds liek more fun Mar 14 01:55:34 like* Mar 14 01:56:10 lets see what the message i didnt mean to send got for response ... Mar 14 01:58:27 hmm just reenoo Mar 14 02:02:54 he didnt understand my logic comment but thats okay Mar 14 02:23:33 Justin, you there? Mar 14 02:27:38 Jason_Foss: yes Mar 14 02:27:50 still building Mar 14 02:32:06 Cool. Mar 14 02:32:20 Sorry if I'm being annoying. Mar 14 02:33:48 What can you tell me about installing ACPI? Mar 14 03:22:02 anyone else getting this? Mar 14 03:22:05 build/tmp/work/arm-linux/glibc-2.3.5+cvs20050627-r3/build-arm-linux/csu/divdi3.os Mar 14 03:22:05 | /tmp/ccLLYwX5.s: Assembler messages: Mar 14 03:22:05 | /tmp/ccLLYwX5.s:573: Error: symbol definition loop encountered at `__divdi3_internal' Mar 14 03:55:43 yuck Mar 14 03:56:09 someone broke glibc :( Mar 14 03:56:23 wonder if uclibc will behave Mar 14 04:00:19 oi jnc Mar 14 04:00:24 'allo Mar 14 04:00:36 i see your stuff stayed this time Mar 14 04:00:44 it's not staying for long Mar 14 04:00:47 i have to fix it Mar 14 04:00:50 no? Mar 14 04:01:06 you broke it already? Mar 14 04:01:22 i cant even compile that far Mar 14 04:01:31 someone broke glibc Mar 14 04:01:41 hah, no i know it's broken. there's other things in the distro that need fixing. i want to get everything fixed up by the time the other bugs are resolved Mar 14 04:01:46 indeed Mar 14 04:02:02 i'm going to be helping out where needed Mar 14 04:02:22 first, would like to change bugzilla email ID so that it's consistent with my mailinglist addy Mar 14 04:02:43 tahts under profile settings no? Mar 14 04:03:20 it's the login name, i need the bugtracker admin to change it Mar 14 04:03:25 http://bugs.openembedded.org/userprefs.cgi?tab=email Mar 14 04:03:49 oh wait ... Mar 14 04:03:51 hmm Mar 14 04:05:23 which admin would you get to change it? Mar 14 04:06:30 cmon uclibc, give me some love Mar 14 04:08:30 emte: i asked once, someone said ggregory i think i remember that different Mar 14 04:08:38 ggsomething Mar 14 04:08:49 ah Mar 14 04:08:56 treke Mar 14 04:09:24 i'm going to make a post to the ML about it, since i don't have an email addy Mar 14 04:11:42 Anyone in here know how to dim the screen on a Zaurus? Mar 14 04:11:48 Programmatically. Mar 14 04:12:10 jnc, try treke@handhelds.org Mar 14 04:12:43 Jason_Foss: it should be covered by a /proc or /sys export Mar 14 04:13:04 Jason_Foss: see the 'slutils' package 'slbl' program Mar 14 04:14:15 emte: thanks Mar 14 04:14:25 Will that give me some code that will do it for me? Because the Linux I'm using on my Zauri is pretty bastardized. Mar 14 04:14:27 Jason_Foss: take a look at slutils for 2.4 kernel Mar 14 04:14:41 Jason_Foss: my build is failing....someone broke gnome-themes Mar 14 04:15:53 How do I find such a thing? I'm not getting any good results from my search... Mar 14 04:16:50 * emte mentions that Jason_Foss is for some unkown reason running familiar on his zaurus Mar 14 04:17:12 i am actually supprised it runs for the most part Mar 14 04:17:13 Jason_Foss: heh...that's because the source is in OE Mar 14 04:17:25 packages/slutils/slutils-0.1.0/slbl Mar 14 04:17:57 <--- doesn't have OE Mar 14 04:18:31 just grab a snapshot and check out the dev branch Mar 14 04:18:56 Jason_Foss: What device? Mar 14 04:19:02 okr http://oe.pastebin.com/601060 Mar 14 04:19:48 I'm trying to get all of the following equipment to work: 2 iPaqs, 2 SL-C1000s, and 1 SL-C3100 Mar 14 04:20:03 The iPaqs have a program called BL that handles the backlights. Mar 14 04:23:34 Which apparently looks for /dev/touchscreen/0 Mar 14 04:23:43 Jason_Foss, your going to have to install OE and somehow merge the OZ device code into familiar Mar 14 04:24:27 ugh....darn X11R7... Mar 14 04:24:32 broke the build Mar 14 04:25:04 JustinP, i wish i could get that far Mar 14 04:25:22 and no default_pref to turn it off and no preferred-versions for older x..... Mar 14 04:25:42 set virtual Mar 14 04:25:47 emte: glibc isn't broken....at least not AFAIK... Mar 14 04:25:56 i cant compile anything now Mar 14 04:25:58 I just updated and built.. Mar 14 04:26:01 wait... Mar 14 04:26:08 I didn't rebuild glibc today...yesterday Mar 14 04:26:37 i updated ~ 1hour ago Mar 14 04:26:46 tryoing to fix the problem Mar 14 04:27:18 http://oe.pastebin.com/601071 Mar 14 04:27:35 no glibc commits in a couple of days... Mar 14 04:27:59 Hm, I built from scratch on 2006-03-11 Mar 14 04:28:14 strange Mar 14 04:28:19 Does anyone know where the touchscreen is under dev? Mar 14 04:28:33 Jason_Foss: look at that pastebin link I posted for you Mar 14 04:28:34 http://oe.pastebin.com/601073 Mar 14 04:28:41 Jason_Foss: http://oe.pastebin.com/601060 Mar 14 04:29:02 second one is the logfil;e Mar 14 04:29:57 Jason_Foss: may i ask what the project is for ? Mar 14 04:30:11 i myself have a C-3000 and C31000 Mar 14 04:30:21 err... s/31000/3100/ Mar 14 04:30:43 trying to get the bugs worked out mostly, then sell them off and keep one as a toy Mar 14 04:32:20 Well... Mar 14 04:32:25 Without being too long winded... Mar 14 04:32:50 The project deals with distributed-database transactions through a mobile ad-hoc network. Mar 14 04:32:53 * JustinP starts an oz354 build.... Mar 14 04:33:27 There are a series of laptops with a modified version of mySQL5 on them... Mar 14 04:33:43 And then the PDAs run client software that communicates with the servers. Mar 14 04:34:00 maybe i should start trying to fit the h36 stuff in OZ since taht seems to be the only useable distro now Mar 14 04:38:49 is anyone besides myself even trying to build familiar anymore? Mar 14 04:40:03 emte: I'm just building OZ here Mar 14 04:40:46 yeah i am not sure anyone else is since koen left Mar 14 04:41:15 which means things arnt getting reported or fixed Mar 14 04:43:46 * Jason_Foss codes happily. Mar 14 04:43:52 night Mar 14 04:44:41 night Mar 14 04:44:44 emte: What's the difference between the distros, anyways? Mar 14 04:45:08 mainy specific device support, licensing, and screen size Mar 14 04:45:15 mainly* Mar 14 04:45:40 licensing? Mar 14 04:46:26 yes, thats what the whole onenssl thing for the last week or two has been in regards to Mar 14 04:46:32 openssl* Mar 14 04:47:00 openssl cant be distributed with familiar because of its licensing Mar 14 04:47:32 so wpa_suplicant was crippled so that openssl wasnt used Mar 14 04:48:15 there are lots of packages with the same issue Mar 14 04:53:43 hey mithro Mar 14 04:53:43 less licensing issues than legal issues stemming from licensing incompatibilities Mar 14 04:54:20 well, bitbake has been taught to adhear to distro licensing Mar 14 04:58:14 hey emte Mar 14 04:59:36 maybe i need to go back to mar 1 so that things build again Mar 14 05:05:39 pb_: what's your bugtracker OE email? Mar 14 05:05:54 want to assign some glibc bugs and patches to you Mar 14 05:07:52 pb@reciva.com is my guess Mar 14 05:08:17 * jnc tries it Mar 14 05:08:35 worked, thanks Mar 14 05:08:49 i just searched for glibc bugs Mar 14 05:08:52 :) Mar 14 05:09:29 ah, good one Mar 14 05:09:44 i'm taking it upon myself to assign bugs :( Mar 14 05:11:58 usually people pickup bugs Mar 14 05:12:08 unless they are already working on it Mar 14 05:12:13 * jnc winces Mar 14 05:12:20 reenoo re-opened bug 702 Mar 14 05:13:28 looks liek he sort of appologized? Mar 14 05:14:03 he was right, you know, i'm just kind of amazed that he keeps making me do more and more unnecessary work Mar 14 05:14:21 lol Mar 14 05:14:54 like, it doesn't end with a public statement about how he made me do more work unnecessarily in the first case, he keeps doing it Mar 14 05:15:01 you will be ready to consult for micky and appl soon Mar 14 05:15:32 ugh Mar 14 05:15:49 why is glibc failing :( Mar 14 05:17:10 my IRC name is 'Eric Shattow' same as the bug report from 2 weeks ago Mar 14 05:17:42 ? Mar 14 05:17:43 why is gnome-themes failing? Mar 14 05:17:49 eh. Mar 14 05:17:50 * emte wonders where that came from Mar 14 05:17:53 it's just funny Mar 14 05:18:23 JustinP, because i willed everything to break until glibc works for me :) Mar 14 05:18:34 i'm not all that mysterious. i always use my name when appropriate Mar 14 05:18:44 emte: :) Mar 14 05:19:12 i'm going to do a full rebuild tonight i think Mar 14 05:19:25 i was hoping to test out the glibc 2.4 stuff Mar 14 05:19:31 build familiar Mar 14 05:19:33 :) Mar 14 05:19:38 then you can be in my boat Mar 14 05:19:43 hey, i said i was going to Mar 14 05:19:52 and i'll know atleast one other person builds it Mar 14 05:20:12 oh i thought you ment OZ Mar 14 05:20:14 post your local.conf somewhere again, i'll fire it up Mar 14 05:20:30 its not a conf problem Mar 14 05:20:36 i can build OZ Mar 14 05:20:49 but taht doesnt work too well on an ipaq Mar 14 05:24:24 i'm talking about you posting your conf, so that i may attempt to build it here Mar 14 05:25:48 hmm Mar 14 05:25:54 i broke pastebin Mar 14 05:27:49 * emte stuffs more hamsters in the interweb Mar 14 05:30:22 eentaarr weeeb Mar 14 05:34:08 lol Mar 14 05:34:22 i am sure i broke pastebin Mar 14 05:34:35 time for raftb i guess Mar 14 05:34:41 rafb Mar 14 05:35:38 emte: what do you think mickey's bugtracker id is Mar 14 05:35:38 http://rafb.net/paste/results/BibGr034.html Mar 14 05:36:15 no idea ... i am not sure a bug has ever been assigned to him Mar 14 05:36:38 * jnc uses a deep and promising voice IT IS GOING TO HAP PEN Mar 14 05:36:53 he's listed as a maintainer Mar 14 05:37:18 http://bugs.treke.net/show_bug.cgi?id=306 Mar 14 05:38:13 i think this is the most i've ever searched BT Mar 14 05:38:20 :P Mar 14 05:38:23 wahahaha Mar 14 05:38:27 thanks Mar 14 05:38:58 Jey Justin, you still up? Mar 14 05:40:27 Jason_Foss: yeh Mar 14 05:44:19 Cool. That link you gave me = huge help. Mar 14 05:44:38 :-) Mar 14 05:44:55 can give you the full source if you want (with includes ;-) Mar 14 05:47:00 I'm actually recoding it so that it follows the same prescripts as the BL program on the iPaq Mar 14 05:47:13 That way, the client code doesn't need to give a crap what kind of PDA it's on. Mar 14 05:48:45 grrrr Mar 14 05:48:52 why the hell doesn't /etc/passwd exist?? Mar 14 05:49:15 it was diasbled? Mar 14 05:49:19 disabled* Mar 14 05:49:24 ::shrug:: Mar 14 05:51:15 JustinP: some systems lack an /etc/passwd Mar 14 05:51:31 i'm pretty sure that its absence would be a bug though, on an embedded system Mar 14 05:52:00 i belive its part of the no-root password Mar 14 05:52:16 Jason_Foss: AFAIK, the BL program has a zaurus patch floating around somewhere Mar 14 05:52:26 patch (re-)submissions welcomed Mar 14 05:52:43 jnc, he is using in/on familiar Mar 14 05:53:20 so i am not sure how valid it would be Mar 14 05:55:33 hmm Mar 14 05:55:40 looking for koen's bugtracker id Mar 14 05:58:53 oooh. got it now Mar 14 06:00:52 emte: there's a zap-root-password post-image Mar 14 06:01:02 emte: it's failing because /etc/passwd is missing... Mar 14 06:02:19 JustinP, ah Mar 14 06:02:34 taht sounds suspiciously like an old bug Mar 14 06:04:53 current oz354.....for akita.... Mar 14 06:05:41 i am sure this was a past issue Mar 14 06:08:52 JustinP, [Oe], gpe-image fiales Mar 14 06:08:58 fails Mar 14 06:09:01 thread Mar 14 06:10:41 From: Michael Kropfberger Mar 14 06:10:41 To: oe@handhelds.org Mar 14 06:10:41 Date: Thu, 10 Mar 2005 16:48:22 +0100 Mar 14 06:10:44 if taht helps Mar 14 06:13:32 * emte chuckles at jnc Mar 14 06:30:32 hmm Mar 14 06:30:37 it's okay Mar 14 06:30:39 :) Mar 14 06:30:42 i think i got a few down Mar 14 06:30:55 i'm now hacking on gcc 4.1/glibc 2.4 things Mar 14 06:31:18 i read your addon to my ... previous rant Mar 14 06:31:22 under the pretense that breakage in OE is going to persist until EABI toolchain is integrated Mar 14 06:33:00 Hey Justin, if you can't get the image to build...feel free to give up. I think my bastardized version will be OK now that I have that code you gave me. Thanks for all your help today. Mar 14 06:34:43 the lucky coincidence that qemu manages to build at all, is due to an abuse of gcc assumptions that are and never were gauranteed Mar 14 06:35:04 s/are/are not/ Mar 14 06:37:42 * emte shrugges Mar 14 06:38:14 i half wonder if qemu is being shoved in because it already workis in scratchbox Mar 14 06:38:47 it's being shoved in to fix a problem that is a result of assuming everyone wants a gabazillion locales generated Mar 14 06:43:25 i still only 1/2 understand taht reasoning Mar 14 07:13:32 03jnc 07org.oe.dev * rbb3ad83d... 10/packages/gcc/ (6 files in 2 dirs): Mar 14 07:13:34 Pushing gcc 4.1.0 after a short review. Mar 14 07:13:36 Not sure if it builds, leaving preference at "-1". See OE bug #378. Mar 14 07:13:38 Thanks to Bernhard Rosenkraenzer for the submission. See OE bug #773. Mar 14 07:25:55 03jnc 07org.oe.dev * r6a070585... 10/packages/glibc/ (16 files in 3 dirs): Mar 14 07:25:57 Pushing glibc 2.4 per OE bug ticket #741. Thanks to Bernhard Rosenkraenzer Mar 14 07:25:59 for the patch. Not including CVS modifications, reference bug report if Mar 14 07:26:01 they are needed, or add them yourself. Mar 14 07:26:03 This is not tested, setting preference to "-1". Mar 14 07:26:33 anyone interested in EABI should be hacking on those builds now. Mar 14 07:26:37 'nite Mar 14 07:26:50 <_law_> hi Mar 14 07:26:52 <_law_> ERROR: Cannot satisfy the following dependencies for gpe-task-base: Mar 14 07:26:53 <_law_> libxpm4 (>= 3.5.4.2) Mar 14 07:27:27 <_law_> whiche package provides libxpm4 ? Mar 14 07:28:04 xpm maybe ? Mar 14 07:28:59 night jnc Mar 14 07:30:02 <_law_> org.openembedded.dev/packages/xpm/ is empty Mar 14 07:30:25 <_law_> libxpm_3.5.1.bb libxpm_3.5.4.2.bb libxpm_cvs.bb Mar 14 07:30:41 <_law_> found in xlibs dir Mar 14 07:32:31 <_law_> but i cant find libxpm4 Mar 14 07:33:00 we have to get rid of jnc ASAP Mar 14 07:33:55 he has little to no understanding of toolchains, oe and his comments can only hurt Mar 14 07:34:00 the project Mar 14 07:35:19 hi all! Mar 14 07:35:58 hi Mar 14 07:36:00 * emte shrugges Mar 14 07:37:25 i think comments are open game as for anything beyond that it requires a project admin to grant repo access Mar 14 07:38:20 emte: right, I wonder how he got access :} Mar 14 07:38:46 before i forget ... do you know what broke glibc? Mar 14 07:39:15 adding and setting that flag suppresses the qemu problem but glibc is now hosed Mar 14 07:39:52 i am not seeing anything about it on the buglist or mailing list Mar 14 07:41:59 <_guillermo_> I am getting a XCOMPOSITE_VERSION' undeclared (first use in this function) error with libxcomposite Mar 14 07:43:15 <_guillermo_> yesterday I was told this was fixed. Is there an trick to make it work? Mar 14 07:43:29 <_law_> _guillermo_, find tmp/staging/arm-linux/ -name *ompos*|xargs rm -r; Mar 14 07:43:41 <_guillermo_> _law_: thx Mar 14 07:44:01 <_law_> _guillermo, i have got some success with this Mar 14 07:45:06 emte: He is that kind of guy who thinks "Let me bring my conflict resolution skills to OE", well the only conflict is between him and the rest of the team :} Mar 14 07:45:12 _guillermo: remove the Xcomposite.h header Mar 14 07:45:29 _guillermo: ph5's attempt to change the include order failed, we need to improve it Mar 14 07:45:47 <_guillermo_> has anyone buit a whi Mar 14 07:45:57 _guillermo: http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=4&logfile=20060314001553.log Mar 14 07:45:59 <_guillermo_> whole image with gcc 4.1? Mar 14 07:46:26 _guillermo: no, gcc3.4.4 Mar 14 07:46:50 i dont even think gcc-4.0.x will build a whole image ytet Mar 14 07:46:53 emte: I have built glibc 2.3.5 for OpenSLUG and OpneZaurus yesterday - without qemu Mar 14 07:47:08 zecke, yeah it seems familiar is hosed Mar 14 07:47:28 | /tmp/ccOmnmqR.s: Assembler messages: Mar 14 07:47:28 | /tmp/ccOmnmqR.s:573: Error: symbol definition loop encountered at `__divdi3_internal' Mar 14 07:48:26 i am starting to wonder if anyone is taking care of any new issues with familiar Mar 14 07:50:18 emte: probably not. As the internal old CRL circle in Cambridge decided to use something else for familiar Mar 14 07:50:30 emte: anyway I'm going to uni now Mar 14 07:50:49 maybe i need to submit a bug to remove fam083 from the oz-fam branch Mar 14 07:50:53 k Mar 14 07:50:58 have fun at class Mar 14 07:52:52 <_guillermo_> emte: familiar and OE have no future, AFAIK Mar 14 07:53:31 actually that has not been decided yet according to france Mar 14 07:54:21 but that does not negate the fact things have suddenly been broken for other distros Mar 14 07:55:01 if it did then woul would all have to run gentoo because the kernel developers do Mar 14 07:55:05 would* Mar 14 07:55:22 er i ment we Mar 14 07:56:06 my issue is taht the oz-fam branch is supposed to be relativly stable Mar 14 07:58:50 <_guillermo_> I hope angstrom is able to replace familiar before the end of the year, but I think it's difficult Mar 14 08:00:45 as far as i know angstroms goals are not to replace familiar at all Mar 14 08:01:11 <_guillermo_> I know, but it could be used like that. Mar 14 08:01:38 the same could be said debian vs rh Mar 14 08:01:43 <_guillermo_> otherwise if familiar dissapears ( or is not supported) there's gonna be a hole in OE Mar 14 08:01:46 not exactly true Mar 14 08:02:05 yes and no Mar 14 08:02:33 to me it seems oe has a problme with device centric development Mar 14 08:02:37 problem* Mar 14 08:02:56 angstrom hopefully helps to change this Mar 14 08:03:49 oz is nice and all, but only if you own a zaurus Mar 14 08:04:38 the same can be said of almost all the distros currently in oe Mar 14 08:05:49 but its double edged Mar 14 08:06:52 people prefer to work on thier own specific devices, unfort this usually means extra or usable features never get moved to other devices Mar 14 08:08:55 damnit Mar 14 08:09:10 which fuckwith thinks he 0wnz my spare time? Mar 14 08:09:22 ah, it's jnc Mar 14 08:09:28 jnc: stop assigning bugs to me Mar 14 08:10:22 emte: angstrom will replace familiar, openzaurus and opensimpad Mar 14 08:10:32 i do wish we had a java programmer here tho Mar 14 08:10:45 emte: since familiar now suffers from extreme NIH symptoms Mar 14 08:12:13 koen, my concern as stated above, is that the new changes to OE have propigated into the semi-stable branch Mar 14 08:12:55 que? Mar 14 08:13:11 org.openembedded.{oz354fam083} Mar 14 08:13:28 and i cant even get glibc to compile anymore Mar 14 08:13:35 that's your problem Mar 14 08:13:39 it works fine over here Mar 14 08:14:03 if i roll monotone back to march 1st it builds fine Mar 14 08:14:55 does taht still make it my issue? Mar 14 08:14:58 that* Mar 14 08:14:59 yes Mar 14 08:15:06 hrm Mar 14 08:15:08 both you and jnc have trouble reading Mar 14 08:15:20 what am i supposed to be reading? Mar 14 08:15:29 the mailinglist and the commit messages Mar 14 08:15:41 i have, and i searched them for my problem Mar 14 08:15:48 there is nothing about it Mar 14 08:18:06 i dont even know how to address the problem i have, this is the first time i've ever got "symbol definition loop " errors Mar 14 08:18:18 <_guillermo_> maybe I'm missing something,but I don't think there's a need to be so rude Mar 14 08:18:24 there haven't been commit to glibc or its dependencies to .oz354fam083 in over a month Mar 14 08:18:33 iknow Mar 14 08:18:40 so something on your build host changed that fucks up the build Mar 14 08:18:43 thats why i dont understand this Mar 14 08:18:46 hence "your problem" Mar 14 08:19:36 * emte gives up and rolls back monotone Mar 14 08:20:29 and if someone dares to say "qemu/glibc need to be opt-in" ...... Mar 14 08:21:40 * koen quotes zecke "I've just explained it and I'm not going to repeat myself" Mar 14 08:22:24 <_guillermo_> *may* I ask about xcomposite? Mar 14 08:22:41 <_guillermo_> removing Xcomposite.h from staging fixed the first error Mar 14 08:23:25 <_guillermo_> but then I get a lot of non declared function errors Mar 14 08:25:24 sounds like pH5 needs to fix some more stuff :) Mar 14 08:26:37 <_guillermo_> btw what benefits does xorg 7.0 bring to oe? Mar 14 08:29:34 katossi: using releases instead of $random cvs version Mar 14 08:30:04 katossi: and we can re-use patches from debian/suse/redhat easier to fix bugs Mar 14 08:31:20 morning Mar 14 08:31:37 koen: great thx Mar 14 08:31:39 katossi: and we're only using a subset of xorg7, not the complete xlibs+xapps+xserver Mar 14 08:31:43 hey hrw|work Mar 14 08:32:02 speaking about .oz354fam083 - I'm thinking about creating .oz354x branch only for OZ so we could ignore all Familiar work Mar 14 08:32:18 sure Mar 14 08:32:57 hrw|work: seriously? Mar 14 08:33:00 last I heard the hh.org people were complaining that .dev didn't build for 0.8.3 Mar 14 08:33:22 I guess that means the relations are completely broken now Mar 14 08:34:02 that gives an indication how clueless they are about OE Mar 14 08:34:20 and how they 'problems' are purely figments of their imagination Mar 14 08:34:29 but I will shut up about that now Mar 14 08:34:38 hey zecke Mar 14 08:34:40 emte: sorry, my fault Mar 14 08:34:50 emte: the oz,fam branch builds familiar Mar 14 08:35:12 emte: we might consider removing it from oe.dev but koen probably talked about it already Mar 14 08:35:14 katossi: yes Mar 14 08:35:44 zecke: from the people with push access only reenoo cares about familiar right now Mar 14 08:36:03 zecke: but: http://handhelds.org/hypermail/familiar-dev/current/0764.html Mar 14 08:37:50 zecke: isn't it cute that someone is managing our free time? Mar 14 08:37:58 morning all Mar 14 08:38:17 it made my day, waking up and finding out someone assigned a gazillion bugs to me Mar 14 08:38:18 hey alan|home Mar 14 08:46:27 morning Mar 14 08:47:43 koen: yeah jnc must die Mar 14 08:47:45 hey XorA Mar 14 08:47:51 zecke: that's a bit harsh Mar 14 08:47:54 koen: I need help, could you review a mail for me Mar 14 08:48:13 sure Mar 14 08:48:30 * XorA prepares his asbestos suit Mar 14 08:49:46 koen: http://pastebin.ca/45612 Mar 14 08:50:08 koen: I remember Wim, who said RPM is all you need, jnc is of the same class Mar 14 08:52:51 zecke: from my dealings with him, he is very focused on "his" problems, and gets tunnel vision Mar 14 08:53:28 zecke: the qemu stuff is already opt-in Mar 14 08:54:22 zecke: http://pastebin.ca/45614 Mar 14 08:56:03 XorA: http://handhelds.org/hypermail/oe/current/7052.html Mar 14 08:56:21 XorA: I hope I didn't wake up the ricers ;) Mar 14 08:57:17 koen: that is iMail not copying '>' into the mail Mar 14 08:57:28 koen: is too harsh (leave alone the spelling and grammar mistakes) Mar 14 08:57:51 koen: bizarre, he has just detailed the standard operation of mem*, who made the ones in glibc so bad? Mar 14 08:58:39 XorA: no idea, but OE had (has?) a arm-memcpy.patch for glibc Mar 14 08:59:10 koen: I have a feeling mplayer-atty does as well Mar 14 08:59:16 sdl, qt/e and x could also be patched Mar 14 09:00:09 the kdrive faster-rotated patch uses a similar theory Mar 14 09:02:36 time to fix libxpm Mar 14 09:04:02 <_law_> hrw|work, you mean libxpm4 (>= 3.5.4.2) problem? Mar 14 09:04:18 koen: then we just need CFLAGS=-O8 and we sorted :-) Mar 14 09:04:31 heh Mar 14 09:04:38 I kinda like -Os Mar 14 09:04:54 koen: is it too harsh, or can I post it? Mar 14 09:05:04 _law_: I want to change packaging Mar 14 09:06:08 mickeyl: Fix EABI integration, now! Mar 14 09:06:44 <_law_> any ideas how i can fix the libxpm4 problem? (isnt available anymore) Mar 14 09:07:31 _law_: w8 for my test/push Mar 14 09:07:42 <_law_> hrw|work, ok Mar 14 09:07:51 zecke: try this one: http://pastebin.ca/45620 Mar 14 09:08:28 Packaged contents of libxpm into /home/hrw/devel/build/0308/tmp/deploy/ipk/libxpm4_3.5.4.2-r1_armv5te.ipk Mar 14 09:08:40 morning Mar 14 09:08:44 hey CoreDump|home Mar 14 09:08:55 hi CoreDump Mar 14 09:09:19 koen: too late :} Mar 14 09:09:54 koen: I don't want to destroy his believe it is broken... Mar 14 09:10:44 zecke: i wonder howmany times we have to say "qemu is already opt-in" before emte and jnc will get the point Mar 14 09:11:21 _law_: pull Mar 14 09:12:47 koen: I think emte is close Mar 14 09:12:58 koen: his patch has some points Mar 14 09:13:30 03hrw 07org.oe.dev * r9a981595... 10/packages/xlibs/libxpm_3.5.4.2.bb: libxpm 3.5.4.2: split sxpm/cxpm into separate packages Mar 14 09:16:09 from my experience with Wim, we need to show jnc his borders as early as possible Mar 14 09:16:26 wim? Mar 14 09:16:42 hrw|work: not wimpie (not ns2) Mar 14 09:16:51 who was Wim? Mar 14 09:16:55 hrw|work: the guy who said FLTK and RPM will solve everything Mar 14 09:17:05 ah.. I dont even remember him Mar 14 09:17:07 hrw|work: he was the Opie version of jnc Mar 14 09:17:09 hail zecke Mar 14 09:17:22 good morning all Mar 14 09:17:25 hi pb Mar 14 09:17:30 good morning pb_ Mar 14 09:17:35 pb_: you wouldn't hail me, if you read my badly phrased and written rants Mar 14 09:17:51 zecke: that is why I keep you on /ignore Mar 14 09:18:14 gosh I'm talking with a bot :} Mar 14 09:18:54 better than talking with certain other individuals, I suppose Mar 14 09:26:21 <_law_> hrw|work, looks better now, thanks Mar 14 09:45:51 hi lala Mar 14 09:55:19 now this is intresting Mar 14 10:07:26 I'm a bit lost with the metadata so I cannot find the reason for the next behaviour: I'm trying to build 'glibc-initial_2.3.2+cvs20040726.bb' which fails when the DL_DIR in my local.conf points to a empty directory (OE downloads what it needs). I get a 'cvs/glibc-initial' directory inside my DL_DIR (among some packages of course) but the configure fails. HOWEVER... Mar 14 10:08:39 I've been toying with oe for some time, and before had another dir pointed at DL_DIR. There is a 'cvs/glibc' directory (no -initial) and building 'glibc-initial_2.3.2+cvs20040726.bb' works like a charm Mar 14 10:09:36 Cockroach-: how is that related to DL_DIR? the configure failure Mar 14 10:09:44 Cockroach-: is there any particular error message, or does it just fail in some general kind of way? Mar 14 10:10:50 zecke: well I had no idea why the code would compile at work but not at home Mar 14 10:11:12 zecke: so I decided to start from scratch and force the oe to download everything again Mar 14 10:11:31 Cockroach-: the glibc people moved coded on the server around Mar 14 10:11:55 Cockroach-: this is why we download a cvs co from familiar.hh.org for some versions of glibc Mar 14 10:13:59 pb__: http://pastebin.ca/45637 if intrested Mar 14 10:15:04 zecke: how is that done "by the book"? Surely not by tweaking the .bb files directly? local.conf black magic? Mar 14 10:15:24 Cockroach-: automatically :} Mar 14 10:15:31 that sounds like you have a damaged cvs checkout (i.e. one without the arm port) Mar 14 10:16:01 yeah, what zecke said Mar 14 10:16:26 Well if it would wokr automatically, I wouldn't be here?-) Mar 14 10:16:35 crying Mar 14 10:17:44 but yeah, that's what I get for trying things with the root of all evil, familiar-0.8.3. I just have no choice Mar 14 10:18:36 Cockroach-: familiar 0.8.3 doesn't use glibc 2.3.2 Mar 14 10:19:09 Cockroach-: pastebin your local.conf please Mar 14 10:19:11 pb__: is do_munge still needed with glibc 2.4? pbrook mentioned something about putting in the ports/ subdir was enough Mar 14 10:19:52 zecke: that can be so, but when building meta-sdk for x86 architecture, it does seem to do so. Mar 14 10:20:33 Cockroach-: I think you need to donate your working tarball of glibc Mar 14 10:20:52 pb_: so 2.3.2+cvs seems broken :} Mar 14 10:21:07 hi, koen .. I am compiling the test iwmmxt cross toolchain. I thinks the option "--with-float=softfp" should be "--with-float=soft", right? Mar 14 10:21:32 good morning Mar 14 10:21:57 zecke: http://pastebin.ca/45639 local.conf Mar 14 10:23:01 Cockroach-: MACHINE=x86? Mar 14 10:23:03 leoncamel: no, it should be -mfloat-abi=softfp, the TARGET_FPU="soft" takes care of making it softfloat, the mfloat-abi specifies softvfp Mar 14 10:23:20 The original metadata related question was, that where does it say in glibc-initial_2.3.2+cvs20040726.bb to use 'cvs/glibc' if it exists instead of 'cvs/glibc-initial' Mar 14 10:24:14 zecke: indeed, as I said, I need a way to develope applications on x86, linked to same libraries (on x86 arch) which are available in the arm architecture Mar 14 10:24:26 ... as I said many, many days ago =) Mar 14 10:24:47 Cockroach-: why do you want to build a glibc2.3.2 then? Mar 14 10:24:56 Cockroach-: that is most likely untested Mar 14 10:25:24 zecke: hey, I only want a working build, the 2.3.2+cvs... comes straight from the .bb files Mar 14 10:25:28 I have nothing to do with that Mar 14 10:25:30 =) Mar 14 10:25:50 Cockroach-: well, have fun :) Mar 14 10:25:51 I just type "build meta-sdk" and sob when it fails Mar 14 10:26:20 I have no idea which versions are tested, I thought the whole idea of oe is to keep track of dependencies by itself Mar 14 10:26:46 but probably this is an issue of familiar being unmaintained? Mar 14 10:26:50 Cockroach-: x86 is not a supported architecture for familiar 0.8.3 I guess Mar 14 10:27:16 Cockroach-: the issue is you stick the finger into the ass of familiar and woner it breaks Mar 14 10:27:24 Cockroach-: ideally it shouldn't break, but it does Mar 14 10:27:24 koen, how about I am using "-mfpu=softvfp" instead of "-msoft-abi=softfp" ? Mar 14 10:27:28 but what about the earlier issue, that I got it to work. Mar 14 10:27:44 Cockroach-: closes guess is glibc CVS is corrupted (as pb_ said) Mar 14 10:27:57 koen, I can't find any information for "-msoft-abi=softfp".. Mar 14 10:28:12 Cockroach-: build the md5 sums of both tarballs (working, non working) Mar 14 10:28:17 Cockroach-: and see if they matcg Mar 14 10:28:26 koen, BTW, Is the "-msoft-abi" is part of EABI ? Mar 14 10:29:13 leoncamel: eabi is softfloat by default, but you can mix soft and hardfloat binaries at will Mar 14 10:29:21 pb_: can glibc 2.3.2 be removed? Mar 14 10:29:29 zecke: I've already checked that they don't. the other being 'cvs/glibc' and the other 'cvs/glibc-initial' puzzle me. Mar 14 10:29:57 Cockroach-: okay, the glibc people decided to movel files around physically Mar 14 10:30:04 Cockroach-: donate your tarball somewhere Mar 14 10:31:03 Cockroach-: ah glibc-initial 2.3.2+cvs is the latest :} Mar 14 10:31:35 weird Mar 14 10:32:25 weird indeed Mar 14 10:32:53 Cockroach-: hmm for me no glibc-initial is build Mar 14 10:33:06 Cockroach-: does your bitbake warn you about multiple providers? Mar 14 10:33:08 zecke: you probably don't have nptl Mar 14 10:33:20 koen: I don't have the issue Mar 14 10:34:55 koen, ok. I will compiling the bootstrap-image today with "-msoft-abi=softfp". Tomorrow, the I will test is in our board. Mar 14 10:35:03 where could I change the desktop folders in gpe? hacking the matchbox bb files? Mar 14 10:35:17 Cockroach-: NOTE: multiple providers are available (glibc, glibc-intermediate);NOTE: consider defining PREFERRED_PROVIDER_virtual/arm-linux-libc-for-gcc Mar 14 10:35:21 Cockroach-: do you see the same? Mar 14 10:37:59 zecke: I remember seeing those in some stages in my experiments, but now I can't see them. Of course it's possible that 'rebuild glibc-initial' still takes stuff from some cache which I don't know about. I try to do a clean build and tell the how it goes. Mar 14 10:38:37 Cockroach-: try setting it glibc instead of glibc-initial Mar 14 10:39:22 koen, shall I use the patch you give me, two days ago ? use "--with-float=softfp" ? or modified it into "-mfloat-abi=softfp" ? Mar 14 10:39:39 modify it Mar 14 10:39:47 I didn't read the gcc manpage correctly Mar 14 10:40:16 koen, ok. Testing start... Mar 14 10:41:19 zecke: setting what? I misinterpret far to easily, so please be exact =). Mar 14 10:42:00 Cockroach-: hehe, I try to resist to not argue on that Mar 14 10:42:23 Cockroach-: set a preferred provider for ?-libc-for-gcc Mar 14 10:42:31 Cockroach-: but consult the warning when building Mar 14 10:43:09 koen, where can I get the manual of EABI ? Mar 14 10:43:15 zecke: ok. I just have to wait a bit since "Handling BitBake files" :) Mar 14 10:45:10 hehe Mar 14 10:45:22 Time to get a cup of coffee. Thank god I was able to purchase an old, cheap ass laptop so I can (in theory) extend the time I have for my thesis work by typing in the bed :) Mar 14 10:45:53 this whole oe & sdk thing proved far more complicated than I though. Mar 14 10:45:57 *thought Mar 14 10:46:18 leoncamel: the 'best' thing I have found is http://www.codesourcery.com/gnu_toolchains/arm/faq.html#q_ABI_what_is_abi Mar 14 10:46:52 Cockroach-: please help us with the documentation - probably after your thesis Mar 14 10:47:49 I've consider it to be an valid option after I get the thesis of my back Mar 14 10:49:32 zecke: when you got those errors concerning multiple providers, what providee where you building? Mar 14 10:49:42 glibc-initial? meta-sdk? Mar 14 10:49:52 Cockroach-: e-image Mar 14 10:50:02 and it selected glibc Mar 14 10:50:48 zecke: (what's that?) I deleted the temporary folder and cvs from DL_DIR, started 'build glibc-initial' and got no warnings Mar 14 10:51:17 Cockroach-: well, that is obvious :) Mar 14 10:51:44 Cockroach-: you said it to build glibc-initial, while if you build meta-sdk you will have a dependency like virtual/?-libc-for-gcc Mar 14 10:52:00 Cockroach-: and both glibc and glibc-initial can provide that Mar 14 10:53:02 Hmm. Scary. this might start to make sense Mar 14 10:54:38 Cockroach-: ? is for me arm-linux Mar 14 10:55:35 hi, I wanted to know: do you mostly use uclibc or glibc ? Mar 14 10:55:47 some packages seems broken with the uclibc Mar 14 10:55:59 victorz: familiar and openzaurus are using glibc Mar 14 10:56:00 zecke: I might get it to work with your advices, let's see Mar 14 10:56:17 Cockroach-: let us at least hope it Mar 14 10:56:27 so nobody using uclibc ? Mar 14 10:56:28 Cockroach-: please conserve your working glibc sourceball please Mar 14 10:56:40 victorz: OpenSLUGuc is using uclibc Mar 14 10:57:18 ok. I'm trying to build familiar/openomap with uclibc Mar 14 10:57:23 victorz: but they only test with ~1000 packages Mar 14 10:57:34 do it makes sense if i report patches for uclibc ? Mar 14 10:57:41 victorz: sure! Mar 14 10:57:48 :) Mar 14 10:58:03 ok, i have some... but where must i post them ? Mar 14 10:58:12 i'm lost with all bugtrackers... Mar 14 10:59:51 victorz: bugs.openembedded.org **** ENDING LOGGING AT Tue Mar 14 10:59:57 2006