**** BEGIN LOGGING AT Mon Dec 01 02:59:58 2008 Dec 01 05:35:04 * * OE Bug 4893 has been created by  Dec 01 05:35:06 * * gcc-cross-sdk-4.2.4-autobuild Dec 01 05:35:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4893 Dec 01 07:16:44 03Robert Schuster  07org.openembedded.dev * r105b568bcd 10openembedded.git/packages/classpath/ (classpath-0.97.2/gtk-fix.patch classpath_0.97.2.bb): classpath 0.97.2: Updated gtk-fix patch. Dec 01 08:14:41 hi, all! Is it me, or gcc-cross build failed recently? Dec 01 08:30:36 03Koen Kooi  07org.openembedded.dev * r5fd146277d 10openembedded.git/packages/gcc/ (gcc-4.2.3/gfortran.patch gcc-4.2.4/gfortran.patch): gcc 4.2.4, 4.2.4: fix bug #4893 Dec 01 08:33:41 03Koen Kooi  07org.openembedded.dev * rfafec65303 10openembedded.git/packages/opencv/opencv_cvs.bb: opencv cvs: ARM_INSTRUCTION_SET = "arm" Dec 01 08:33:41 03Koen Kooi  07org.openembedded.dev * r542577e49d 10openembedded.git/packages/octave/octave.inc: octave: ARM_INSTRUCTION_SET = "arm" Dec 01 08:41:16 03Koen Kooi  07org.openembedded.dev * r007a03b34f 10openembedded.git/packages/linux/ (4 files in 2 dirs): Dec 01 08:41:16 linux 2.6.28rc6: fix versioning. Dec 01 08:41:16 * as has been pointed out *numerous* times before: Dec 01 08:41:16 PV = "2.6.28rc6" sorts higher than PV = "2.6.28" thus breaking upgrades Dec 01 08:51:53 * raster grumbles @ oe Dec 01 08:51:57 so... what broke Dec 01 08:53:10 why does oe refuse to put 1 of the packages listesinto the image... why... why... Dec 01 08:53:15 it puts the rest in,.. damnit! Dec 01 08:58:28 03Koen Kooi  07org.openembedded.dev * r7063202238 10openembedded.git/packages/linux/linux_2.6.28-rc6.bb: linux 2.6.28rc6: kernel_version is a reserved var, change it to kernel_release Dec 01 09:07:45 03Koen Kooi  07org.openembedded.dev * r7b4484d039 10openembedded.git/packages/images/illume-image.bb: Dec 01 09:07:45 illume image: start cleaning Dec 01 09:07:45 * use IMAGE_LINGUAS to get linguas into the image instead of hardcoding "glibc-binary-locale*" which breaks things when not using qemu Dec 01 09:07:45 * remove postprocess step that can be done with a package Dec 01 09:07:46 MUCH more cleanup is needed, but this should stop the worst from getting copy/pasted by people Dec 01 09:09:42 where is that koen! Dec 01 09:09:44 ... Dec 01 09:10:36 he doesn't use irc Dec 01 09:10:42 damnit! Dec 01 09:26:00 good morning Dec 01 09:27:45 pb_: what do you think about koen closing bugs with ARM_INSTRUCTION_SET = "arm" Dec 01 09:27:49 ? Dec 01 09:28:55 are we moving towards thumb mode for all arm archs? Dec 01 09:29:13 (supporting it) Dec 01 09:34:07 I suppose often the nedded patch is smthg like this: Dec 01 09:34:09 -#ifdef __arm__ Dec 01 09:34:16 +#if defined(__arm__) && !defined(__thumb__) Dec 01 09:40:25 bulls*it, I checked a couple of patches...it really depends case-by-case :-/ Dec 01 09:51:25 ant_work: I would prefer the packages to be fixed "properly" rather than just setting ARM_INSTRUCTION_SET = "arm", but koen's patches don't preclude making a better fix later. Dec 01 09:51:41 as you say, the right answer is often different from one package to the next. Dec 01 09:52:47 ideally yes, we should move towards supporting thumb mode for all packages, not least because support for ARM mode is liable to go away in newer cpus at some point. Dec 01 10:01:01 I see. Do you expect slowdowns? Any benchmark? Dec 01 10:01:56 'cause the image size is much smaller using uclibc than enabling thumb in glibc. Dec 01 10:02:12 Dunno about ram savings Dec 01 10:28:24 03Koen Kooi  07org.openembedded.dev * r3669669b73 10openembedded.git/contrib/angstrom/sort.sh: angstrom feed sorter: don't run script if there are no packages to sort Dec 01 10:41:55 sure, using thumb with glibc is never going to give you the same kind of space savings as uclibc. Dec 01 10:44:32 pb__: in how far is ARM mode going to disappear? like arm cpus and only thumb running on them or do I misread it somehow? Dec 01 10:44:42 about this savings, I found approx 10% for thumb and 30% for uclibc images (console-image even 50%) Dec 01 10:44:51 against glibc/arm Dec 01 10:46:08 he, all nice and well, ...but unfortunately the uclibcgnueabi x11 image fails starting X...sigh Dec 01 10:49:12 BlindMan: well, this is just speculation; I don't know exactly what is planned. But ARM mode has already been deleted in Cortex-M3 for example and I would not be surprised if the same happened to other processors in the future. Dec 01 10:49:56 pb__: what's the artional behind it? using thumb mode is sufficient for the apps anyway? Dec 01 10:50:29 BlindMan: right, exactly. I think the theory is that thumb-2 gives you the best of both worlds so there is no longer any benefit in using arm mode. Dec 01 10:50:56 pb__: how limited are the memory controllers in the new chipsets ? Dec 01 10:51:04 or at least, not enough benefit to justify the extra silicon area that is required to support two different instruction sets. Dec 01 10:51:28 ant_work: which new chipsets do you mean? Dec 01 10:51:35 new cortex Dec 01 10:51:43 pb__: kind of funny development, actually ... what about register/memory addressing then? all indirect then? Dec 01 10:52:06 iirc one point was thumb is well suited for 16bit bus Dec 01 10:52:06 BlindMan: er, arm has always been fully indirected. even arm mode has never had an address literal mode. Dec 01 10:52:20 ant_work: cortex is quite a wide ranging family, you can't really generalise about it. Dec 01 10:52:34 pb__: ah! prolly too much mips asm floating around in my head then ;) Dec 01 10:52:39 ok, I'm comparing to pxa255 :-) Dec 01 10:53:01 ant_work: cortex-m1 is very different to cortex-a9 for example. Dec 01 10:55:50 I guess meanwhile the minimum is a 32bit bus. I see Cortex-A8 has 64bits External Memory Bus width Dec 01 10:55:53 well, regarding cortex-m3 - is it me, or this CPU is so dramatically different from normal ARM, that it can't exit thumb state? Dec 01 10:56:43 and this curious stack implementation in ARMv7 - is it only in m3, or in a8/9 too? Dec 01 10:59:12 slapin: right, that's what we were just talking about. cortex-m3 has no arm decoder, it can only execute thumb(-2) code. Dec 01 11:00:26 pb__: do you know of better resources regarding this? discussing that a bit more thoroughly? Dec 01 11:01:17 BlindMan: only the datasheets. they're available on the arm.com site. Dec 01 11:01:30 pb__: ah ok, thanks Dec 01 11:01:39 as I understand, on v7 arms, ARM instruction set runs better than thumb, if you don't care for bigger code Dec 01 11:03:26 slapin: I think that's probably true, but I suspect it is more of a compiler issue than an architectural one. Dec 01 11:03:29 and killing ARM instruction set seems pointless to me, because it is much cleaner than thumb/thumb2 and I don't want to see ARM as messy as x86. ARM instruction set seems to me as one of ARM strong points. Dec 01 11:05:02 yes, arm is perhaps a more elegant instruction set than thumb-2. but I suspect at some point the question will become "how much do you want to pay to retain this extra instruction set for purely aesthetic reasons?". Dec 01 11:05:49 given a sufficiently good compiler I think thumb-2 ought to work out just as fast as arm but with better code density, so there would be no technical reason to prefer arm state anymore. Dec 01 11:06:38 are the saving that significant? Dec 01 11:06:42 right now I suspect that gcc probably is not a sufficiently good compiler. I don't know whether realview is or not. Dec 01 11:06:43 well, so we'll get another x86 here, and as there is one already, there might be no place for ARM CPUs on market in future. Dec 01 11:07:26 slapin: hey, intel wants to eradicate arm anyway ;) Dec 01 11:07:46 slapin: do you really think that the majority of cpu purchasing decisions are driven by the "neatness" of the instruction set? Dec 01 11:07:48 slapin: (or anandtech, it's a bit blurry ;) Dec 01 11:08:33 keeping a redundant set of decoders around in the silicon will add cost, add power, and reduce speed. those three things are far more likely to influence purchasing decisions. Dec 01 11:08:43 pb__: more what is "certified" which mostly doesn't need to have technical reasons but just some words attached to some piece of software Dec 01 11:08:49 I think that it really depends. at least for me, it worked perfectly till now. Dec 01 11:09:29 anyway, as I say, this is all just speculation. maybe cortex-m3 is just a one-off. Dec 01 11:09:45 we will have to wait and see what happens in the future. Dec 01 11:10:02 I hope there will be some armv5tish on opencores some day then. Dec 01 11:10:24 pb__: it would be just consistent Dec 01 11:10:34 what architecture does opencores support nowadays? Dec 01 11:10:45 There is a reasonable amount of hand optimised ARM assembler out there in binary form. Dec 01 11:10:52 armv4t to armv5t is not much of a leap, certainly if you don't care about the edsp bits. Dec 01 11:11:48 opencores has only OpenRISC, lotso of 8bit CPUs, some MIPS-likes. Dec 01 11:48:04 * * OE Bug 4738 has been UNCONFIRMED by Dec 01 11:48:06 * * python-numpy-1.1.1-autobuild Dec 01 11:48:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4738 Dec 01 11:51:59 thebohemian: ping Dec 01 11:59:58 hi zecke Dec 01 12:03:11 moin Dec 01 12:25:57 gm Dec 01 12:26:58 hi all Dec 01 12:28:38 hi florian Dec 01 12:36:31 zecke: pong Dec 01 12:37:42 thebohemian: I assume mirko pinged you as well? I assume you wanted to talk about the devroom Dec 01 12:38:04 zecke: he pinged me Dec 01 13:00:02 03Koen Kooi  07org.openembedded.dev * r800de10f9a 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Dec 01 13:00:02 opkg(-nogpg): bump SRCREV to 4838 Dec 01 13:00:02 * runtime tested on beagleboard and omap3evm Dec 01 14:50:51 do anybody knows anything about opie2? Dec 01 14:51:27 I'd like to contact developers or something Dec 01 14:52:05 didnt it die? Dec 01 14:52:22 ljp seems to be here though, ask him Dec 01 14:54:56 03Koen Kooi  07org.openembedded.dev * rdaca2dc14f 10openembedded.git/contrib/angstrom/build-feeds.sh: angstrom feed builder: add armv6-novfp Dec 01 14:57:56 morning Dec 01 14:58:24 morning hrw Dec 01 14:59:01 ljp, hi! what is status of Opie II project? is it live? Dec 01 15:02:50 03Koen Kooi  07org.openembedded.dev * r4d54b7f1ed 10openembedded.git/packages/images/beagleboard-demo-image.bb: beagleboard demo image: include more linguas Dec 01 15:07:27 slapin: why not check opie II website or their subversion repo? Dec 01 15:09:55 hrw, hi! well, I see that last commit is from very long ago, but some people tell me that project is live somewhere, so I ask and hope somebody knows. Dec 01 15:10:20 hrw, just asking won't hurt, I think Dec 01 15:32:08 03Koen Kooi  07org.openembedded.dev * ra6f9994982 10openembedded.git/packages/fontconfig/ (fontconfig.inc fontconfig_2.6.0.bb): fontconfig: bump PR for cache change Dec 01 15:36:48 morning Dec 01 15:36:55 hi Chris Dec 01 15:37:36 03Stefan Schmidt  07org.openembedded.dev * rac5d71cc7b 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Dec 01 15:37:39 03Stefan Schmidt  07org.openembedded.dev * r4f41031526 10openembedded.git/packages/openmoko2/openmoko-calculator2_svn.bb: openmoko-calculator2: Add license as band-aid to get it build Dec 01 15:38:27 sleeeepy Dec 01 15:38:38 boo for going back to work after holiday Dec 01 15:44:04 muhdays? Dec 01 16:01:19 Jay7: hi Dec 01 16:01:26 ant_work: rere Dec 01 16:02:07 ant_work: what do you thinking about kdrive -> xorg move? :) Dec 01 16:02:53 he..what can I say..x11 image built yesterday won't start X properly... Dec 01 16:03:01 huhu Dec 01 16:03:32 ant_work: it is kdrive fail? Dec 01 16:03:42 Jay7: was uclibcgnueabi, it is rebuilding now glibc Dec 01 16:03:46 or we have xorg already? Dec 01 16:03:55 ah Dec 01 16:04:27 I'll have to see what's failing...it doesn't come to gpe-login Dec 01 16:04:53 Jay7: the console-image is 50% smaller with uclibc... Dec 01 16:05:01 wow Dec 01 16:06:08 ant_work: can uclibc and glibc be "mixed"? some binaries (console) with uclibc, some (x11) - with glibc? Dec 01 16:06:32 hm.. but this should increase memory usage.. Dec 01 16:13:41 Jay7: do you think we could give name to the items in the 'bootlist' of kexecboot (labels)? Dec 01 16:14:37 ant_work: I think we should split gui from kexecboot.. Dec 01 16:14:47 like Paul says Dec 01 16:15:02 oh, I misunderstood then Dec 01 16:15:43 I thought he was planning to extend kexecboot with uniboot features, in C Dec 01 16:16:03 then we can just call gui from "backend" with prefilled list of devices/partitions Dec 01 16:16:32 and get return code on exit Dec 01 16:16:54 hm.. looks like xdialog/cdialog Dec 01 16:17:48 so you would allow text console too (telnet) ? Dec 01 16:18:15 when we will support network then yes Dec 01 16:18:18 it is possible Dec 01 16:18:40 I don't know...kexecboot is only a remedy for broken bootloaders...I don't see any use let say on routers/headless machines Dec 01 16:18:49 just use telnetd + cdialog as frontend :) Dec 01 16:19:19 Jay7: all in 1.2 Mib? Dec 01 16:19:35 we have no space on zauruses Dec 01 16:19:42 I have report this in mail :) Dec 01 16:19:55 he...for the other there is uniboot Dec 01 16:20:52 then unifying uniboot and kexecboot is meaningless Dec 01 16:22:20 just let Paul use GUI from kexecboot in uniboot :) Dec 01 16:26:53 Omegamoon: hello Dec 01 16:27:35 Jay7: Omegamoon was trying xorg with debian rootfs iirc Dec 01 16:28:02 Omegamoon: have you any success/story? :) Dec 01 16:29:25 03Dmitry Baryshkov  07org.openembedded.dev * r1eaea9d442 10openembedded.git/ (2 files in 2 dirs): Dec 01 16:29:25 sharprom-compatible: update to reflect current toolchain classes Dec 01 16:29:25 Current toolchain classes want to also have virtual/arm-linux-libc-initial Dec 01 16:29:25 So, add it to sharprom-toolchain-native PROVIDES and to PREFERRED_PROVIDER of Dec 01 16:29:27 sharprom-compatible. Dec 01 16:29:29 Signed-off-by: Dmitry Baryshkov Dec 01 16:29:31 03Dmitry Baryshkov  07org.openembedded.dev * r8045219455 10openembedded.git/packages/zaurus-updater/ (zaurus-updater.bb zaurus-updater/updater.sh): Dec 01 16:29:34 zaurus-updater: add support for home partition reflashing Dec 01 16:29:36 Sometimes one needs to clean/reflash the home partition. Dec 01 16:29:38 Add support for flashing home.bin/HOME.BIN files onto "home" mtd Dec 01 16:29:40 device. Dec 01 16:29:42 Signed-off-by: Dmitry Baryshkov Dec 01 16:29:44 03Dmitry Baryshkov  07org.openembedded.dev * r1823b120ab 10openembedded.git/ (3 files in 3 dirs): Dec 01 16:29:47 liblinebreak: Unicode line breaking library Dec 01 16:29:49 Signed-off-by: Dmitry Baryshkov Dec 01 16:29:51 03Dmitry Baryshkov  07org.openembedded.dev * r7561550296 10openembedded.git/packages/libopie/ (libopie2/libopie2-tosa.patch libopie2_1.2.4.bb): Dec 01 16:29:58 libopie2: add patch to keep up with changed tosa kernel interface Dec 01 16:30:00 Some kernel interfaces have changed on tosa. This patch fixes the backlight setting. Dec 01 16:30:02 More to come later. Dec 01 16:30:04 Signed-off-by: Dmitry Baryshkov Dec 01 16:32:25 * Jay7 wants office suite for opie :) Dec 01 16:32:36 psokolovsky: hello! Dec 01 16:32:48 Hi! Dec 01 16:33:52 hrw, RP, zecke|uni: Hi folks, did you receive my key for git access? Thanks! Dec 01 16:34:14 psokolovsky: Mickeyl and Koen are proper persons Dec 01 16:35:48 hrw: does Koen hang on irc nowadays? Dec 01 16:35:57 psokolovsky: about uniboot vs kexecboot Dec 01 16:36:09 Jay7: yep Dec 01 16:36:21 psokolovsky: is uniboot now doing all you wants? Dec 01 16:36:27 except GUI :) Dec 01 16:37:07 Jay7: I have trivial fix to commit to make kexec'ing work again for it Dec 01 16:37:43 Jay7: I also found that other stuff it relies on (e.g. busybox) has some issues in .dev, I kinda fixed that too. Dec 01 16:37:49 Jay7: so, mostly, yes Dec 01 16:37:55 psokolowsky: pls check rootdelay / rootwait Dec 01 16:38:24 ant_work: it works for me ;-) Dec 01 16:38:32 psokolovsky: just kexecboot is created as "a remedy for broken bootloaders" (c) ant_work Dec 01 16:38:58 I mean rootwait should be the preferred way Dec 01 16:39:06 bringing this on wide arena should take more time Dec 01 16:39:36 Jay7: well, I guess, we found a ground on mailing list: both things continue to be used by respective parties ;-). Dec 01 16:39:51 :) Dec 01 16:40:03 yep, something like :) Dec 01 16:40:03 ant_work: is there a way to implement it in userspace? I guess yes, but it's still not a big priority. Dec 01 16:40:59 rootwait in userspace? Dec 01 16:41:25 sorry, I miss the point too. Implement what in userspace? Dec 01 16:42:15 ant_work: you talk and rootwait the preferred way, right? ;-) so, yes it's about rootwait Dec 01 16:42:49 Ah, yes. Copy and paste Dec 01 16:42:50 rootdelay=n (wait for n seconds then try to mount root) Dec 01 16:42:56 rootwait (wait until root is available for mount, as long as it takes) Dec 01 16:43:01 While the delay is certainly usable, the automatic rootwait is much nicer. Dec 01 16:43:11 psokolovsky: while [ \! -d $ROOTDEVICE ]; do sleep(1); done Dec 01 16:43:14 my point is reducing boottime Dec 01 16:43:40 ant_work: you tried to implement and debug it from userspace? if so, snippet would be appreciated. or kexeboot does it? Dec 01 16:43:49 Jay7: heh, not so simple ;-) Dec 01 16:44:11 kexecboot uses rootwait, it's kernel feature Dec 01 16:44:15 it magically worked Dec 01 16:44:25 s/rootdelay=x/rootwait/ Dec 01 16:45:01 Jay7: weird. rootdelay for sure doesn't work with initramfs, neight *wait should. Dec 01 16:45:31 ah..interesting..thesing had it working before our patch Dec 01 16:46:42 ant_work: there reason it shouldn't work must be simple: kernel with initramfs doesn't know what root is! it's initramfs' job to select one and prepare for booting. Dec 01 16:47:01 ah... Dec 01 16:47:11 anyway, thanks for suggestion. should be well implementable in userspace, I'll add that to the queue... Dec 01 16:47:30 we have enough time to device to settle Dec 01 16:47:40 because of gui :) Dec 01 16:49:02 possibly ;-) Dec 01 16:49:14 when kexecboot can't see device it doesn't put it in list Dec 01 16:49:21 this is other reason :) Dec 01 16:50:29 the problem is that ROOTDEVICE is merely an mknod entry ;-). somebody should create it out of something. that something for block devices is apparently /proc/partitions, so first that settles, them /dev/* should be created. that's rootwait impl. I can imagine right now. Dec 01 16:50:33 so all devices should be inserted and detected by kernel before it appears Dec 01 16:51:18 Jay7: by kernel drivers, sure Dec 01 16:51:34 psokolovsky: did you mean nfs/loop devices? Dec 01 16:51:52 this kind of devices can have such problems.. Dec 01 16:51:53 Jay7: nope, only real block ones Dec 01 16:53:47 Jay7: do you think we can get rid of '/bin/sleep 3' in the initscript? I know we'll eliminate the need of a shell probably... Dec 01 16:54:01 but in the meanwhile... Dec 01 16:54:18 ant_work: we should but can't say when :) Dec 01 16:54:47 what was that delay for? Dec 01 16:55:06 don't know.. thesing may know Dec 01 16:55:25 can you try remove this? Dec 01 16:55:37 I'll do :D Dec 01 16:55:50 is it successfull? :) Dec 01 16:56:02 give me 2-3hours ;-) Dec 01 16:57:01 np :) Dec 01 17:07:30 psokolowsky: another question. Is it right you wrote a tool to manage the defconfigs? Legend? Dec 01 17:07:56 actually it is like total anarchy.... Dec 01 17:40:32 bbl Dec 01 17:56:55 Jay7: hi Dec 01 17:57:35 Omegamoon: have you any success with xorg? Dec 01 17:58:09 Jay7: success with xorg? Dec 01 17:58:29 on zauruses Dec 01 17:59:07 Jay7: It works, but it's slow Dec 01 17:59:09 ant says that you have tried xorg :) Dec 01 17:59:29 Omegamoon: more slower than kdrive? Dec 01 18:02:08 yes, I tried to see whether or not Ubuntu Netbook Remix worked, which needs opengl Dec 01 18:02:31 hm.. Dec 01 18:03:06 and what about rotating? can screen be rotated without restarting X? with xrandr e.g. Dec 01 18:03:19 xorg has opengl support, so I used it. Dec 01 18:03:34 rotation works Dec 01 18:03:55 ok, thanks for info :) Dec 01 18:30:08 mwester: was your ssh key added to the OE repository? Dec 01 18:30:36 Yes. :) Dec 01 18:30:41 (thanks!) Dec 01 18:48:36 03Holger Hans Peter Freyther  07org.openembedded.dev * r42c5793dce 10openembedded.git/ (2 files in 2 dirs): Dec 01 18:48:36 tichy: This project moved from svn to git, catch up Dec 01 18:48:36 - Move the SRC_URI to get, increase the PV after Dec 01 18:48:36 the change Dec 01 18:48:36 - Build the latest git version. Dec 01 20:12:40 zecke|uni: ping Dec 01 20:16:38 zecke|uni: nevermind, found the issue Dec 01 20:36:04 question, how do I install a fresh copy of bitbake?' Dec 01 20:36:24 ? Dec 01 20:36:28 there is only setup.py, and little doc. Dec 01 20:36:43 you dont have to install it at all, you can use it where its at Dec 01 20:36:57 if you want to install it, there are packages floating around for various distros, or you can use setup.py Dec 01 20:37:18 setup.py is standard python distutils. run python ./setup.py, it shows usage info Dec 01 20:38:39 kergoth: g'day Dec 01 20:39:33 It has this. Dec 01 20:39:36 [wweng@localhost bitbake-1.8.10]$ ./setup.py Dec 01 20:39:38 usage: setup.py [global_opts] cmd1 [cmd1_opts] [cmd2 [cmd2_opts] ...] Dec 01 20:39:40 or: setup.py --help [cmd1 cmd2 ...] Dec 01 20:39:41 or: setup.py --help-commands Dec 01 20:39:43 or: setup.py cmd --help Dec 01 20:39:48 yes. Dec 01 20:39:49 it does Dec 01 20:39:56 so run python ./setup.py --help-commands Dec 01 20:39:59 like it tells you to Dec 01 20:40:02 and see what commands you can use Dec 01 20:40:15 you probably want 'install' or so Dec 01 20:40:25 hey pb_ Dec 01 20:41:40 hmm. Dec 01 20:41:49 How do I specify prefix? Dec 01 20:42:01 python ./setup.py install --help Dec 01 20:42:03 like it also told you Dec 01 20:42:13 amazing how helpful reading the --help info is Dec 01 20:44:41 * Jay7 is using 'not installed' version of bitbake Dec 01 20:44:57 hi Dec 01 20:46:12 Hello. I have an error during openmoko-image, webkit-gtk-0.1+svnr36540-r7, do_compile. How can i fix that error (Failed to read names from file: ./WebCore/html/HTMLTagNames.in at WebCore/dom/make_names.pl line 190. )? Dec 01 21:01:32 or maybe it is unbuildable at all? Dec 01 21:05:48 loo Dec 01 21:11:58 03Chris Larson  07org.openembedded.dev * rd4709e08fe 10openembedded.git/classes/src_distribute.bbclass: src_distribute.bbclass: Unbork do_distsrcall, fixes the issue seen by Rob Tow. Dec 01 21:20:29 kalemas: you have issues building webkit-gtk? Dec 01 21:20:32 03Holger Hans Peter Freyther  07org.openembedded.dev * rae6902838c 10openembedded.git/ (5 files in 4 dirs): Dec 01 21:20:32 bind: Add bind version 9.3.6 Dec 01 21:20:32 - Split the bind-utils into a new package Dec 01 21:20:32 - epoll fails because it can not run the test. Do not use Dec 01 21:20:32 epoll for bind... even if unlikely a resulting binary might Dec 01 21:20:34 run on linux 2.4 and then things don't work... Dec 01 21:20:42 03Holger Hans Peter Freyther  07org.openembedded.dev * r31120dc4d2 10openembedded.git/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: Add bind-utils to the feed Dec 01 21:25:08 zecke -> < kalemas> Hello. I have an error during openmoko-image, webkit-gtk-0.1+svnr36540-r7, do_compile. How can i fix that error (Failed to read names from file: ./WebCore/html/HTMLTagNames.in at WebCore/dom/make_names.pl line 190. )? Dec 01 21:25:52 nice name Dec 01 21:26:07 hey ReaperOfSouls :) Dec 01 21:26:25 that would make a good name for a new GarbageCollecting patent/algorithmn Dec 01 21:26:25 Whats going on... Dec 01 21:27:39 you should hang out in #montavista more :) Dec 01 21:28:22 Yeah, my dircproxy client died and I was focused on something else and never popped back on. Dec 01 21:29:00 ah Dec 01 21:36:05 hi ant Dec 01 21:36:26 abend Dec 01 21:37:31 ant___: now with 3 underscores! :) Dec 01 21:37:43 ant___: how is your build? Dec 01 21:37:57 he..I'm gaining degrees..pb docet Dec 01 21:38:17 I'm rebuilding kexecboot without 3secs delay Dec 01 21:38:22 almost done Dec 01 21:38:37 did you test it? Dec 01 21:38:49 recompiling right now...1 min Dec 01 21:39:00 ah.. almost :) Dec 01 21:39:32 I want run build of x11-image tonight Dec 01 21:39:42 then I'll try to debug the issue with x-11 image / uclibc Dec 01 21:39:57 just to check 'buildability' Dec 01 21:40:17 :-) Dec 01 21:40:28 seems fine..rebuilt from scratch yesterday Dec 01 21:41:03 by the way upgrade akita stats on oestats :) Dec 01 21:41:45 he he Dec 01 21:42:10 btw, is now u-boot kernel image building working? Dec 01 21:43:03 I hope so...let me see Dec 01 21:43:37 I have KERNEL_IMAGETYPE and CUSTOM_ROOTFS_SIZE_akita in my local.conf Dec 01 21:43:50 Is it enough? Dec 01 21:44:48 well, I suppose yes. add your patch if you want to automate things Dec 01 21:45:06 flashing Dec 01 21:45:23 I can flash by hand :) Dec 01 21:46:22 I just want to upgrade my old -stable 'distro' on my akita.. Dec 01 21:46:30 sorry, I'm flashing...I'm too lazy to write a nandlogical wrapper Dec 01 21:46:52 ah :) Dec 01 21:47:50 Jay7: I was refering to zaurus-2.6.inc_kernel_imgtype.diff Dec 01 21:47:57 of you Dec 01 21:48:17 hm.. my old patch? Dec 01 21:48:33 It should be already in .dev Dec 01 21:48:47 is it? Dec 01 21:49:13 let me test kexecboot...right now... Dec 01 21:49:18 rebooting Dec 01 21:49:34 mount succesful Dec 01 21:49:54 yes, yes, 3 secs gained :-) Dec 01 21:49:56 hm.. Dec 01 21:50:06 no, patch is not in tree Dec 01 21:50:59 so.. build should break on creating fs image.. Dec 01 21:51:49 well, it does not... Dec 01 21:52:02 I don't have that patch atm Dec 01 21:52:20 did you use u-boot atm? Dec 01 21:52:27 wasn't for u-boot? Dec 01 21:52:38 :-), no, since months Dec 01 21:52:51 with zImage here is no problems :) Dec 01 21:53:13 andrea@mizar /oe/build $ bitbake u-boot-git Dec 01 21:53:13 NOTE: Handling BitBake files: | (6260/6260) [100 %] Dec 01 21:53:14 but with uImage build breaks because of zImage absence Dec 01 21:53:23 *should break Dec 01 21:53:27 I'll tell you in a while Dec 01 21:54:33 ah, sh@t, I need to set UBOOT_MACHINE Dec 01 21:54:48 for c7x0 -> corgi Dec 01 21:55:11 hm.. shoud I set something here? Dec 01 21:55:24 no, akita == akita, you lucky b* Dec 01 21:55:30 good :) Dec 01 21:59:04 Jay7: look at attachments of bug 3738..old memories (UBOOT_ENTRYPOINT, ...) Dec 01 22:01:28 zaurus-2.6.inc in this bug is obsoleted by something but is not in final diff :) Dec 01 22:01:47 ah.. no Dec 01 22:01:52 here is other changes Dec 01 22:02:19 btw NOTE: package u-boot-git: completed Dec 01 22:10:41 woglinde: ping Dec 01 22:11:08 ant___: I've found my patch in bugzilla Dec 01 22:11:13 !oebug 4544 Dec 01 22:11:14 * * Bug 4544, Status: NEW, Created: 2008-09-08 01:12 Dec 01 22:11:15 * * andrea.adami(AT)gmail.com: unlock uImage for Zaurus-clamshells Dec 01 22:11:16 * * http://bugs.openembedded.net/show_bug.cgi?id=4544 Dec 01 22:11:20 good Dec 01 22:11:39 somebody should commit it oneday... Dec 01 22:11:41 can somebody say why it is not in tree? :) Dec 01 22:11:47 ^^ Dec 01 22:12:16 how to find this 'some body'? :) Dec 01 22:13:00 Jay7: ok, now I retried the x11-image compiled in uclibcgnueabi. It is looping at Please wait: booting... Dec 01 22:13:13 hm.. Dec 01 22:13:31 what stage is it? Dec 01 22:13:33 I have the x cross-cursor on the screen and can move it touching Dec 01 22:13:35 kernel booting? Dec 01 22:13:38 ah Dec 01 22:13:56 then kdrive is up and working Dec 01 22:14:06 I saw a suspect log... warning invalid date 2%%x&%% Dec 01 22:14:14 before Dec 01 22:14:33 something wrong with some of xsession scripts or with gpe-dm Dec 01 22:15:11 no, with xinit scripts Dec 01 22:15:19 xsession starts after login Dec 01 22:15:20 he..I can even suspend/resume the loop :-) Dec 01 22:15:44 can you ssh to Z? Dec 01 22:16:18 I got to console Dec 01 22:16:32 ctrl+alt+backtab several times Dec 01 22:17:33 nothing in dmesg Dec 01 22:18:08 try to export DISPLAY=":0.0" and run rxvt e.g. Dec 01 22:18:14 ah.. /etc/init.d/gpe-dm start Dec 01 22:18:17 that is Dec 01 22:19:18 is it started? Dec 01 22:19:30 gpe-dm Dec 01 22:19:55 it loops Dec 01 22:20:14 hm.. Dec 01 22:20:15 bot *only* in ANGSTROM_MODE="uclibc" Dec 01 22:20:35 a.k.a. uclibcgnueabi Dec 01 22:21:00 glibc version is ok Dec 01 22:21:12 (same metadata checkout) Dec 01 22:21:31 please try to reproduce on akita Dec 01 22:21:38 oh.. Dec 01 22:21:46 it can take time :) Dec 01 22:21:49 c7x0 is a different beast (w100fb) Dec 01 22:22:17 try to move /etc/X11/Xinit.d/99gpe-login somewhere Dec 01 22:22:35 I'm pretty sure it has to do with timezone Dec 01 22:22:36 or replace gpe-login inside with rxvt Dec 01 22:23:05 just to check who fails Dec 01 22:23:24 date: invalid date.... Dec 01 22:23:30 never seen before Dec 01 22:23:32 hehe.. Dec 01 22:24:03 you needed guru :) Dec 01 22:24:25 woglindeeee Dec 01 22:24:32 is thumb enabled? Dec 01 22:24:36 no Dec 01 22:25:14 what? Dec 01 22:25:26 sorry to disturb Dec 01 22:25:36 yeah Dec 01 22:25:44 I am lookin anime with joost Dec 01 22:25:54 I suspect the x-11 image could have some uclibc related problems Dec 01 22:26:18 more like binutils problems Dec 01 22:26:22 with date :) Dec 01 22:26:28 I see a log about date: invalid date 2%%xxx Dec 01 22:26:35 or with timezone data Dec 01 22:27:06 ant___: try to remove /etc/localtime if exists Dec 01 22:27:52 iirc the TZ var need to be set Dec 01 22:28:32 yes, he uClibc time functions get timezone information from the TZ environment variable Dec 01 22:28:32 i have no TZ on my akita Dec 01 22:28:37 ah Dec 01 22:28:45 then yes.. Dec 01 22:32:23 http://wiki.openwrt.org/OpenWrtDocs/WhiteRussian/Configuration#head-037616e62986b57c5fc312bc143b16d6f0687d6b Dec 01 22:37:34 hm.. Dec 01 22:37:58 we should provide such data when building uclibc 'world' Dec 01 22:40:30 i can only notice that error occurs during make, and some files pointed in log does not exists Dec 01 22:41:21 Jay7: look at /etc/profile in rootfs... export TZ Dec 01 22:42:08 if [ ! -e /etc/localtime ] Dec 01 22:42:14 :-/ Dec 01 22:45:24 hmm.. settimeofday() failed Dec 01 22:55:14 ant___ that could be a uclibc error Dec 01 22:55:30 or gpe initscripts... Dec 01 22:55:46 glibc-centric Dec 01 22:56:46 Jay7: I'm sending you the size diffs uclibc/glibc eabi per e-mail Dec 01 22:57:03 ant___: interesting :) Dec 01 22:57:11 it's awesome Dec 01 22:57:40 woglinde did most of the dirty work Dec 01 22:57:45 I have found that patch from oebug 4544 does not apply clean Dec 01 23:01:03 Jay7: btw I did recompile kexecboot only in glibc. It's one byte more (#sleep 3) Dec 01 23:01:16 :) Dec 01 23:03:14 03Fraxinas  07org.openembedded.dreambox * r6cc9ad363b 10openembedded.git/packages/initscripts/initscripts-1.0/dm8000/bootup: don't mount cifs shares at bootup time (no network present then) Dec 01 23:03:14 03Fraxinas  07org.openembedded.dreambox * r6acebcbcd6 10openembedded.git/packages/initscripts/initscripts-1.0/netmount.sh: mount all type cifs, smbfs, nfs shares on net up Dec 01 23:03:15 03Fraxinas  07org.openembedded.dreambox * r6c0cc70700 10openembedded.git/packages/dvdauthor/ (dvdauthor_0.6.14.bb files/dvdauthor-flush-progress.patch): flush VOBU progress output (to be able to parse progress for DVDBurn job) Dec 01 23:03:17 03Fraxinas  07org.openembedded.dreambox * rfe99e67054 10openembedded.git/packages/gstreamer/gst-plugin-dvbmediasink.bb: use original dvbmediasink for dm7025. fixes mpeg2 / mpeg1 playback issue on ati. on behalf of Nix_niX Dec 01 23:03:23 03ghost  07org.openembedded.dreambox * re5bdf17967 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.opendreambox.org/git/fraxinas/opendreambox-1.5 into org.openembedded.dreambox Dec 01 23:03:27 03acid-burn  07org.openembedded.dreambox * rb86ca4ec75 10openembedded.git/packages/samba/ (6 files in 3 dirs): Dec 01 23:03:30 add samba-3.0.24 Dec 01 23:03:32 fix nmbd startup Dec 01 23:03:34 cleanup Dec 01 23:03:36 03Fraxinas  07org.openembedded.dreambox * r098d6077c1 10openembedded.git/packages/initscripts/initscripts-opendreambox_1.0.bb: create netmount script in ifup Dec 01 23:03:39 03ghost  07org.openembedded.dreambox * r9a3396b5f5 10openembedded.git/packages/gstreamer/gst-plugin-dvbmediasink.bb: downgrade dvbmediasink for dm7025 to SRCDATE 20080907, update to 20081127 for all other hardware Dec 01 23:03:51 03ghost  07org.openembedded.dreambox * rcf68f00063 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.opendreambox.org/git/acidburn/opendreambox-1.5 into org.openembedded.dreambox Dec 01 23:03:55 03ghost  07org.openembedded.dreambox * rc8b581f920 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.opendreambox.org/git/fraxinas/opendreambox-1.5 into org.openembedded.dreambox Dec 01 23:03:59 03ghost  07org.openembedded.dreambox * rc6bd31065a 10openembedded.git/packages/gstreamer/gst-plugin-dvbmediasink.bb: gst-plugin-dvbmediasink.bb: another SRCDATE update Dec 01 23:04:02 03ghost  07org.openembedded.dreambox * r9de5fec7b5 10openembedded.git/packages/avahi/ (avahi.inc avahi_0.6.21.bb): avahi: remove dependency to libnss-mdns to fix DNS Problems when the local domain is named ".local" Dec 01 23:04:10 03ghost  07org.openembedded.dreambox * r1d13a03834 10openembedded.git/packages/initscripts/initscripts-1.0/ (5 files in 5 dirs): finally dont mount nfs,cifs,smbfs before network is up Dec 01 23:04:13 03ghost  07org.openembedded.dreambox * rcb910cb88a 10openembedded.git/packages/gstreamer/gst-plugin-dvbmediasink.bb: gst-plugin-dvbmediasink.bb: fix typo Dec 01 23:04:18 03ghost  07org.openembedded.dreambox * rf863933bec 10openembedded.git/packages/initscripts/initscripts-opendreambox_1.0.bb: initscripts-opendreambox_1.0.bb: create missing directory Dec 01 23:04:18 Jay7: strange..yesterday the difference was much bigger! Second e-mail Dec 01 23:04:21 03ghost  07org.openembedded.dreambox * rb066c0087a 10openembedded.git/packages/dreambox/dreambox-secondstage.bb: dreambox-secondstage.bb: cleanup, update dm800 secondstage, add possibility to update secondstage via onlineupdate Dec 01 23:04:27 03ghost  07org.openembedded.dreambox * r3fbcae6e8a 10openembedded.git/packages/dreambox/dreambox-dvb-modules.bb: dreambox-dvb-modules.bb: update dm8000 drivers to 20081201 Dec 01 23:04:30 03ghost  07org.openembedded.dreambox * r0d89bda14f 10openembedded.git/packages/dreambox/dreambox-dvb-modules.bb: dreambox-dvb-modules.bb: update dm800/dm8000 drivers, add dependency to dreambox-secondstage also for dm800 Dec 01 23:04:36 03ghost  07org.openembedded.dreambox * r127a3802a1 10openembedded.git/packages/tuxbox/dvbsnoop.bb: dvbsnoop.bb: update to latest version Dec 01 23:04:39 03ghost  07org.openembedded.dreambox * r3035765726 10openembedded.git/packages/dreambox/dreambox-secondstage.bb: dreambox-secondstage.bb: update 8000 2ndstage to 70 (this fixes some mpeg decoder problems) Dec 01 23:04:43 03ghost  07org.openembedded.dreambox * r7effccb955 10openembedded.git/packages/base-files/base-files_3.0.14.bb: base-files: save and restore old resolv.conf on package install Dec 01 23:08:43 woglinde: if you have time look at this... Dec 01 23:08:49 http://imagebin.ca/view/IniidR4.html Dec 01 23:08:54 http://imagebin.ca/view/T9KOZY.html Dec 01 23:10:57 how to revert all files in git working dir to state from branch? Dec 01 23:11:09 something like mtn revert Dec 01 23:11:13 git reset Dec 01 23:11:32 git reset --hard 123455667 Dec 01 23:11:45 ant which one? Dec 01 23:12:03 I see glibc images lost size in one day...strange Dec 01 23:12:27 uclibc console-image was about 50% Dec 01 23:17:23 now is only 10% smaller (jffs2) Dec 01 23:18:26 Jay7: perhaps I misunderstood your question about 'revert'. What do you wanna do? Dec 01 23:19:02 ant___: no, git reset --hard do what I want :) Dec 01 23:19:19 I have applied patch but don't commit it Dec 01 23:19:39 then I create new patch with git diff Dec 01 23:20:05 then I wants to reset workdir to state before patch :) Dec 01 23:20:16 Jay7: http://git.or.cz/gitwiki/GitCheatSheet Dec 01 23:21:54 btw I prefer to have my local branch Dec 01 23:25:34 ant___: I have updated oebug 4544 with new patch Dec 01 23:25:57 10x for cheat sheet :) Dec 01 23:26:08 I definitely should print this out :) Dec 01 23:26:43 is a link from OE Wiki !100x Dec 01 23:27:04 haha Dec 01 23:27:08 joost rockz Dec 01 23:27:10 http://www.joost.com/059011j/t/Fist-of-Bruce-lee#id=059011j Dec 01 23:29:12 ant___: there was issues with authomated mkimage for u-boot Dec 01 23:29:46 was this resolved? Dec 01 23:30:17 well, I remember I booted an auto-created uImage Dec 01 23:30:34 remember this is not compressed Dec 01 23:31:04 you can add a flag and squeeze more kb Dec 01 23:31:29 I had standard kernel + kexecboot using u-boot Dec 01 23:31:43 (zipped uImage) Dec 01 23:31:48 there was some issues with compression too :) Dec 01 23:32:04 it just works (tm) Dec 01 23:32:26 http://bugs.openembedded.net/show_bug.cgi?id=3151 Dec 01 23:35:23 Jay7: andrea@mizar ~/devel/u-boot-dev $ ./mkimage -A arm -O linux -T kernel -a 0xA0008000 -e 0xA0008000 -n "Linux-kexecboot-2.6.26-r2" -d ./zImage-kexecboot-2.6.26-r2-c7x0.bin.gz -C gzip Dec 01 23:35:23 Image Name: Linux-kexecboot-2.6.26-r2 Dec 01 23:35:23 Created: Sun Aug 17 00:07:44 2008 Dec 01 23:35:23 Image Type: ARM Linux Kernel Image (gzip compressed) Dec 01 23:35:23 Data Size: 1394944 Bytes = 1362.25 kB = 1.33 MB Dec 01 23:35:24 Load Address: 0xA0008000 Dec 01 23:35:26 Entry Point: 0xA0008000 Dec 01 23:35:30 andrea@mizar ~/devel/u-boot-dev $ Dec 01 23:36:32 ok Dec 01 23:36:39 Jay7: kexecboot.bin.gz Dec 01 23:36:39 I have started build of x11-image Dec 01 23:36:54 and gzip again Dec 01 23:37:16 dual gzipped? Dec 01 23:37:34 yes, 4kb saved! there was no ext3 support, though Dec 01 23:37:58 compressing the decompressor Dec 01 23:37:59 hm.. Dec 01 23:38:18 just try Dec 01 23:38:51 I needs some masterplan :) Dec 01 23:39:02 what I should check and try :) Dec 01 23:39:12 1)wife Dec 01 23:39:19 or you're dead ! Dec 01 23:39:43 hehe :) Dec 01 23:39:57 good night Dec 01 23:40:07 yep.. there is time :) Dec 01 23:40:20 nite pplz **** ENDING LOGGING AT Tue Dec 02 02:59:57 2008