**** BEGIN LOGGING AT Sun Apr 01 02:59:57 2007 Apr 01 03:53:48 jacques: EABI BE built. Apr 01 03:57:48 rwhitby, you gonna try it now? Apr 01 03:59:00 yeah, just gonna change ixp4xxbe to nslu2be so it builds an 8MB image for me to flash. Apr 01 04:14:10 jacques: I forgot that Angstrom rootfs is too big for the slug. I'll build MokoSlug BE and try that. Apr 01 04:15:37 ah, d'oh Apr 01 04:15:43 not too big for lude tho Apr 01 04:18:18 no Apr 01 04:18:42 but my lude is currently the mokogateway, so I don't have time to test and then flash back again. Apr 01 04:19:48 my lude is currently in several pieces awaiting ka6sox to tell me about jtag on it Apr 01 04:20:20 but I think it will run from flash in its current state Apr 01 06:27:04 why isent the images beeing bzipped anymore? Apr 01 06:27:17 I am getting a tar file instead Apr 01 06:28:33 not tar.bz2. And also it is beeing created two equal images, one with Angstrom- one without Apr 01 06:29:06 goxboxlive: IMAGE_FSTYPES set correctly? Apr 01 06:29:42 I havent touched local.conf in six months or so Apr 01 06:30:10 goxboxlive: Check that .tar.bz2 is in the list.. IIRC, there's been recent changes to make .tar not imply .bz2 Apr 01 06:30:59 NAbyss: Ok, thx, thats why then, like i said i havent touched in a long time :-) Ill try that out Apr 01 07:22:24 good morning al Apr 01 07:22:26 ll Apr 01 07:22:30 hey koen Apr 01 07:43:46 03koen 07org.oe.dev * r971e3adc... 10/ (5 files in 4 dirs): initscripts: fix checkroot and remove angstrom specific checkroot Apr 01 07:43:51 03koen 07org.oe.dev * r846c3fff... 10/ (1 packages/gpsd/gpsd.inc): gpsd: disable python bindings Apr 01 07:45:53 hi all Apr 01 07:47:40 hey pH5 & gremlin[it] Apr 01 07:47:50 koen: has anyone tried BE eabi at all? Apr 01 07:48:09 (I know you said it was untested - has anyone even got it to boot?) Apr 01 07:48:50 rwhitby: I only know do13 added patches for it, I know it compiles (without insanity) Apr 01 07:48:58 but I don't know of anyone that booted it Apr 01 07:49:02 it compiles with insanity now. Apr 01 07:49:29 I did a mokoslug image, but it didn't boot - haven't had time to debug anything yet. Apr 01 08:01:26 rwhitby: that's sort what I expected, and why I didn't build anything yet Apr 01 08:01:35 no time to debug things like that Apr 01 08:02:02 rwhitby: did you build an EABI/BE kernel? Apr 01 08:02:30 yep, I'm going to perhaps push the build through to fix other simple "missing armeb settings" things like in insane.bbclass Apr 01 08:02:44 looking at perl at the moment Apr 01 08:03:16 03koen 07org.oe.dev * r8510ca2d... 10/ (3 files in 2 dirs): prismstumbler: fix PV Apr 01 08:03:23 03koen 07org.oe.dev * r458c1e5b... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: prefer a prismstumbler that actually builds Apr 01 08:29:24 * koen hugs avahi Apr 01 08:29:39 ssh hx4700.local Just Works(TM) Apr 01 08:39:13 03pH5 07org.oe.dev * rff961d04... 10/ (4 files in 4 dirs): libgsmd: add interpreter-ready.patch for universal Apr 01 08:51:45 pH5: is the hx4700.conf up to date with the kernel changes? Apr 01 08:53:22 koen: yes. the touchscreen drivers are built-in currently, I'd like to make them modular in the future, but in cooperation with tasks/other devices that use the same driver. Apr 01 08:53:50 oh, I didn't test wifi yet. Apr 01 08:55:25 pH5: there's a slight problem with the Xorg R7.2 upgrade path: Xw100 Apr 01 08:55:49 R7.1 xservers seem to dislike tslib/tscalibrate Apr 01 08:56:04 * koen is waiting for krive to recompile Apr 01 08:56:32 did you build Xw100 1.2? I still have 1.1 in the image. Apr 01 08:56:40 1.1 Apr 01 08:56:44 1.2 is building now Apr 01 08:56:48 which has the nice side effect that keylaunch isn't broken. Apr 01 08:56:52 and it doesn't work? Apr 01 08:57:06 I upgraded the kernel modules, reflashed the kernel and now I have a non working touchscreen Apr 01 08:57:20 koen: rm /etc/pointercal Apr 01 08:57:37 the ts-adc-debounce driver returns completely different values than hx4700_ts Apr 01 08:57:42 psokolovsky_: any idea why that is? Apr 01 09:03:21 03rwhitby 07org.oe.dev * r27f7ed93... 10/ (1 packages/perl/perl_5.8.7.bb): perl: Add support for armeb-linux-gnueabi (trivial addition) - now configures and compiles, but still doesn't install. Apr 01 09:15:32 03koen 07org.oe.dev * rbff61468... 10/ (3 files in 2 dirs): xserver-kdrive: re-add w100 support Apr 01 09:17:23 I'd rather have Ximageon, but Xw100 will do for now Apr 01 09:18:32 yeah, where's sirfred if you need him Apr 01 09:19:11 magician and hx4700 kernel modules updated in the feeds Apr 01 09:20:08 morning all Apr 01 09:22:58 pH5: gpe-login.c has #define XTSCAL_PATH "/usr/bin/gpe-xcalibrate.sh", do we still ship that file? Apr 01 09:23:10 mickey|brunswick: guten morgen! Apr 01 09:24:20 koen: no. Apr 01 09:24:24 hey mickey|brunswick Apr 01 09:25:07 Hi! Apr 01 09:25:15 so I really need to dig out the serial cable Apr 01 09:25:19 pH5: of course, different Apr 01 09:36:51 koen: please fix that path, because of it, re-calibration from gpe-login doesn't work (but I bet that's what you see) Apr 01 09:37:35 http://rafb.net/p/UGeBpO13.html Apr 01 09:38:44 yeah, cool Apr 01 09:40:25 ehm Apr 01 09:40:39 why is ts-adc-debounce spamming syslog? Apr 01 09:41:43 03koen 07org.oe.dev * r3b076e1f... 10/ (3 files in 3 dirs): gpe-login: use xtscal instead of a non-existant script to recalibrate Apr 01 09:42:20 * koen wonders where /usr/lib/ts/input.so went Apr 01 09:44:14 psokolovsky_: why? Apr 01 09:44:23 pH5: what? Apr 01 09:44:34 "of course, different" Apr 01 09:45:25 pH5: well, it uses different algo, different params, how could it stay the same? Apr 01 09:47:02 morning all Apr 01 09:47:28 hi RP Apr 01 09:47:38 Hi! Apr 01 09:47:53 psokolovsky_: I wasn't aware that there is a different algo. I thought X/Y readouts are X/Y adc readouts, no matter what. Apr 01 09:48:42 pH5: alas, all those things affect values. I saw /etc/pointercal needing update even for smaller changes. Apr 01 09:50:05 strange Apr 01 09:50:59 hi all Apr 01 09:52:26 pH5: it's not the strangest thing we have around ;-). it's on the same level of strangenes as ts-ads-deb having issues with magician, like adhoc driver - no. Apr 01 09:53:04 psokolovsky_: yes! pretty strange :) Apr 01 09:54:13 we'll solve strangenesses which makes life harder, and just deal with small strangenesses ;-) Apr 01 09:58:48 hey all Apr 01 09:58:52 i can't boot cacko Apr 01 09:59:02 i installed it according to instructions Apr 01 09:59:05 no errors Apr 01 09:59:06 03rpurdie 07bitbake-1.8 * r805 10/ (ChangeLog lib/bb/cache.py): cache.py: Catch truncated cache file errors (from trunk) Apr 01 09:59:08 daurn: see topic Apr 01 09:59:17 but, when it goes to boot Apr 01 09:59:19 daurn: "This is not a distro support channel" Apr 01 09:59:20 it errors on hdd2 Apr 01 09:59:26 koen: it can be Apr 01 09:59:38 where else do you find people that know about zaurus Apr 01 09:59:44 dunno Apr 01 09:59:50 but this is not a distro support channel Apr 01 10:02:35 03rpurdie 07bitbake-1.8 * r806 10/ (5 files in 2 dirs): Add PE (Package Epoch) support from Philipp Zabel (pH5) (from trunk) Apr 01 10:04:06 03rpurdie * r807 10bitbake/lib/bb/taskdata.py: taskdata.py: Remove debug Apr 01 10:04:39 03rpurdie 07bitbake-1.8 * r808 10/ (ChangeLog lib/bb/runqueue.py lib/bb/taskdata.py): Add code to handle inter-task dependencies (from trunk) Apr 01 10:05:42 03rpurdie 07bitbake-1.8 * r809 10/ (ChangeLog lib/bb/parse/parse_py/ConfHandler.py): Allow operations other than assignment on flag variables (from trunk) Apr 01 10:20:04 hi Apr 01 10:21:06 in .bb file, file:// uri where are supposed to point to? Apr 01 10:35:17 diego71: to the files or ${PN} directory in OE Apr 01 10:39:51 i'm still a bit confused by oe name convention Apr 01 10:40:16 what PN does it mean? Apr 01 10:40:46 package name Apr 01 10:42:13 ok, i'm looking the linux directory under packages Apr 01 10:42:25 and i see: Apr 01 10:42:44 triton-kernel_2.6.11.bb triton-kernel-2.6.11 Apr 01 10:42:59 unslung-kernel_2.4.22.l2.3r63.bb unslung-kernel Apr 01 10:43:32 koen: the BE eabi failure to boot could be a different issue - so there is still hope. rebuilding now. Apr 01 10:43:56 so i wonder which is PN in this two example Apr 01 10:44:16 rwhitby: cool Apr 01 10:44:42 rwhitby: are you planning to outsmart debian twice Apr 01 10:44:45 ? Apr 01 10:45:00 or better, in which way i can decide the PN Apr 01 10:45:10 for some reason (it may be one of the nslu2 patches), the cmdline is screwed up in the last four characters. So I've moved console= to the front of the cmdline. I had the same problem with LE EABI. Apr 01 10:45:23 koen: when did I outsmart Debian the first time? Apr 01 10:45:34 rwhitby: by having slugos Apr 01 10:46:05 armeb in debian is going to be a mess Apr 01 10:46:10 nslu2 wasn't even on Debian's radar when we started SlugOS. Apr 01 10:46:12 at least for the people using armeb now Apr 01 10:46:21 what, both of them? Apr 01 10:46:34 since they aren't going to change the arch name like they did for arm -> armel Apr 01 10:46:43 armeb == OABI and EABI Apr 01 10:46:53 armeb was never official, so caveat emptor. Apr 01 10:47:28 good thing I don't buy debian stuff :) Apr 01 10:49:36 if the speedup of arm -> armel that has been reported carries over to armeb->armeb(eabi) for SlugOS, then that should make some people happy. Apr 01 10:50:13 'speedup' Apr 01 10:50:28 if you used softvfp or softfpa you don't get any speedup Apr 01 10:50:57 I wish I had more time in brazil to write a proper reply to that linuxdevices article Apr 01 10:51:39 03ifaistos 07org.oe.dev * r8ee8881d... 10/ (5 files in 4 dirs): Apr 01 10:51:39 packages/perl/perl_5.8.7.bb: A set of patches to make perl usable on powerpc targets. Apr 01 10:51:39 Thanks to Jamie Lenahan for the help Apr 01 11:07:03 Koen how was brazil ? Apr 01 11:07:11 gerwinin: sunny :) Apr 01 11:09:10 hey zecke Apr 01 11:11:23 moin Apr 01 11:11:44 Koen: Saw any nice chicks ? Apr 01 11:11:51 RP: one of the reasons to scope was to md5sum local patches. But generally this is not needed Apr 01 11:12:03 gerwinin: brazil is the land of dead cows and chicks Apr 01 11:12:13 Koen: :) Apr 01 11:13:05 Koen: Eindhoven is het land of nurses and of public boolean sex; while (1) { sex=true; } Apr 01 11:15:00 zecke: That makes sense. Perhaps just make the filename optional? Apr 01 11:16:00 zecke: btw, I've made some backports ready for 1.8.2 Apr 01 11:19:40 RP: PN-PV you mean :) Apr 01 11:20:09 RP: okay, do we want to go through the bugtracker for these backports? Apr 01 11:20:35 RP: What to do with the interactive patch resolver and bitbake? (context switch) Apr 01 11:21:18 Ifaistos: /usr/bin/ld: C.o: Relocations in generic ELF (EM: 20) Apr 01 11:21:19 C.o: could not read symbols: File in wrong format Apr 01 11:21:27 Ifaistos: perl fails to build for efika Apr 01 11:22:01 koen : what's the error ? Apr 01 11:22:23 the one I just pasted Apr 01 11:22:59 hmmm let me see if i forgot to send anything Apr 01 11:23:18 03koen 07org.oe.dev * r83ea8283... 10/ (3 files in 3 dirs): gpe: adjust some SVN PVs Apr 01 11:26:32 zecke: It seemed pointless for me to file backport requests, the do them myself. Any other reuqests can go through the bugtracker Apr 01 11:27:36 RP: right, what ever suits you more Apr 01 11:29:24 koen : what gcc are you using ? Apr 01 11:29:46 4.1.1 Apr 01 11:30:39 zecke: What do you mean about the intereactive patch resolver? You mean my proposal on the oe list? Apr 01 11:34:30 RP: aeh hmm *hide*, I need to read that Apr 01 11:52:59 koen: BE EABI ixp4xx kernel booted, but problem executing /sbin/init Apr 01 11:53:43 koen: http://pastebin.ca/419510 (haven't started debugging it yet) Apr 01 11:53:57 (ignore the /linuxrc problem - it's unrelated) Apr 01 11:55:55 rwhitby: so we know that the compiler works Apr 01 11:56:11 I wonder what triggers the invalid elf headers Apr 01 11:57:55 koen: also see the cmdline further up that pastebin. the last 3-4 characters are truncated from "noirqdebug". This happens on both LE and BE EABI, but the exact same kernel patches have a correct cmdline on OABI LE and BE. Apr 01 11:58:20 dunno whether that's related or not. Apr 01 11:58:57 weird Apr 01 11:59:08 it is always the last 4 chars or after a fixed number? Apr 01 12:00:15 dunno - when console was the last thing on the line it was enough to take enough off of 115200n8 to kill the console (at first I thought it wasn't booting at all, but then it would magically appear when syslog started_ Apr 01 12:01:12 so it's been the last three characters in the two cases I've seen. Apr 01 12:01:19 (both cases are completely reproducable) Apr 01 12:13:10 RP: could you turn of LZO compression in tosa images, since those are still at 2.6.17? Apr 01 12:14:19 koen: that error line at the end of http://pastebin.ca/419510 looks really weird - all those mixed up characters in the libc pathname .. Apr 01 12:14:37 (that's what made me think of the cmdline problem I'd seen before) Apr 01 12:15:01 rwhitby: it looks me that the kernel panic is messing up the message Apr 01 12:15:30 no, the kernel panic is neatly spliced in the middle of "libraries". Apr 01 12:15:45 it's the stuff printed after "libraries:" which is weird. Apr 01 12:16:14 RP: and http://www.nopcode.org/0xFFFF/ Apr 01 12:18:20 koen: it seems that the last N characters are being truncated from each string which is printed. "/sbin/init" is truncated by one character, and the parts of the pathname of libc.so.6 are each truncated. Apr 01 12:20:07 yet the kernel messages are all printed without truncation Apr 01 12:20:25 (including the panic message in the middle of the /sbin/init error Apr 01 12:23:32 koen : perl build fine here Apr 01 12:23:57 koen : the difference is i have not synced with OE for a about a week Apr 01 12:24:43 perl built fine yesterday, when I built an angstrom-efika image Apr 01 12:25:01 did you rebuild perl-native ? Apr 01 12:25:24 no, since that one is unchanged, right? Apr 01 12:26:20 koen : i cleared both perl and perl-native Apr 01 12:26:49 koen : the thing is perl was building for powerpc but was not working Apr 01 12:27:26 * koen rebuilds perl-native Apr 01 12:31:23 03rwhitby 07org.oe.dev * rd85c16a3... 10/ (1 packages/tasks/task-slugos.bb): task-slugos: Add altboot Apr 01 12:33:11 koen : here is my build -> http://www.pastebin.ca/419538 Apr 01 12:38:01 Ifaistos: ok, rebuild of perl-native fixed it, thanks Apr 01 12:44:34 koen : i think the PR of perl-native should be bumped also Apr 01 12:45:56 koen : BTW i checked perl the last few days and it works fine. I managed to get webmin and a few others apps running without a problem on both 405 and 603 Apr 01 12:47:23 03koen 07org.oe.dev * r54d07772... 10/ (1 packages/perl/perl-native_5.8.7.bb): perl-native: bump PR Apr 01 12:49:38 lunch time.. Apr 01 12:49:42 brb Apr 01 13:12:52 is perl-native's bashism fixed? Apr 01 13:21:45 03freyther 07org.oe.dev * rdd5fc12a... 10/ (1 classes/base.bbclass): Apr 01 13:21:45 classes/base.bbclass: [ftp://ftp.trolltech.com/qt/source/qtopia-core-4.3.0beta.tar.gz] is good enough Apr 01 13:21:45 src_uri as section/groups are unique enough as well. Most of the time Apr 01 13:21:45 (almost always) scoping PN (or PN+PV) is not necessary and in this case Apr 01 13:21:45 the SRC_URI is good enough. Richard thank you for saying the obvious Apr 01 13:33:28 koen: That flasher is interesting :) Apr 01 13:34:15 I thought so :) Apr 01 13:35:07 koen: tosa shouldn't be using the lzo compression? Apr 01 13:35:27 RP: it's at 2.6.17 and the lzo patches don't apply Apr 01 13:35:53 koen: Ah, I broke the kernel build? Apr 01 13:36:12 no, you (or someone else) broke runtime Apr 01 13:36:31 if you flash 2.6.17 + OE built image you get a kernel panic Apr 01 13:36:48 JFFS2 compression type 0x07 not available Apr 01 13:36:49 Error: jffs2_decompress returned -5 Apr 01 13:36:49 kernel panic - not syncing: No init found. Try passing init = option to kernel Apr 01 13:37:04 (according to a fearless angstrom tester) Apr 01 13:38:10 koen: LZO compresison is only enabled in zaurus-clamshell-2.6.conf though? Apr 01 13:40:45 you're right Apr 01 13:40:45 ah, the default command is overridden for tosa... Apr 01 13:40:47 hmmm Apr 01 13:41:35 I did this: http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2007-April/000145.html Apr 01 13:44:34 koen: I've pushed a fix. zaurus-2.6.conf. We should make the jffs2 command defaults more sane so they don't get overridden as much... Apr 01 13:44:34 03rpurdie 07org.oe.dev * r7d07d609... 10/ (1 conf/machine/include/zaurus-2.6.conf): zaurus-2.6.conf: Disable lzo compression by default in the same way as the default jffs2 commandline, fixing problems with old kernels on tosa Apr 01 13:57:38 did anyone use newlocale and strftime_l here? Apr 01 15:15:26 zecke: Did you have a chance to read that mail about interactive tasks? Apr 01 15:26:19 RP: do you remember the subject? Apr 01 15:26:46 zecke: "Interactive task handling", sent on Wednesday Apr 01 15:31:05 RP: sounds sane Apr 01 15:31:53 zecke: I'm open to better suggestions. I was surprised we don't need to touch bitbake :) Apr 01 15:31:56 RP: I dodn't know that TERMRCFILE exists, but if it is honored Apr 01 15:32:20 RP: Better solution: Implement a dynamic automatic patch resolver :) Apr 01 15:32:40 zecke: Feel free to add PATCH_RESOLVER = "auto" :} Apr 01 15:34:10 RP: TERMCMDRUN for konsole and gnome-terminal are potentially harmful as they can return before you would expect it Apr 01 15:34:38 zecke: For gnome-terminal, that is why we have --disable-factory Apr 01 15:35:03 My konsole voodoo might be weaker Apr 01 15:35:16 oh and obviusly everytime I see a hardcoded 'bash' alarm bells ring Apr 01 15:35:30 RP: well, kconsole is quite safe... (it does not exists AFAIK) Apr 01 15:35:44 :} Apr 01 15:36:08 zecke: The "hardcoded" bash is trivial to change ;-) Apr 01 15:37:17 hi Apr 01 15:37:25 RP: yes, but not only by changing the conf file Apr 01 15:37:31 can anybody tell me if openzaurus-devel@lists.sourceforge.net is still active? Apr 01 15:38:05 RP: +f.write("#" + bb.data.getVar('TERMCMDRUN', d, 1)) what is the '#' doing? Isn't it a comment? Apr 01 15:38:09 zecke: are you volunteering to update bitbake in OE? Apr 01 15:38:22 splitnode Apr 01 15:38:25 zecke: Showing what was in that variable for debugging ;-) Apr 01 15:38:38 hehe Apr 01 15:38:40 self.test() Apr 01 15:39:04 koen: once 1.8.2 is out I can do that Apr 01 15:39:48 When do we want to release 1.8.2? How much testing does the 1.8 branch need? Apr 01 15:40:13 *food* :) Apr 01 15:40:21 bbl in an hour Apr 01 15:40:36 as of testing building OpenSlug and Angstrom Images would be enough for me Apr 01 15:40:49 03koen 07org.oe.dev * r3e65d641... 10/ (1 packages/monotone/monotone-6_0.34.bb): monotone: add 0.34 Apr 01 15:41:46 what about bumping trunk to 1.9.1? Apr 01 15:43:47 koen: Do we need to atm? Apr 01 15:43:56 RP: does package_deb.bbclass need to sort .debs like package_ipk.bbclass does? Apr 01 15:44:21 RP: it would be nice to be able to distinguish between trunk 'versions' via the OE startup summary Apr 01 15:46:21 koen: Which bit of package_ipk do you mean? Apr 01 15:47:09 koen: An svn revision would probably be more useful... Apr 01 15:47:18 deploy/ipk//foo.ipk Apr 01 15:47:45 The debian packaging did thing long before the ipk one did ;-) Apr 01 15:47:51 s/thing/this/ Apr 01 15:48:31 I couldn't find it in package_deb.bbclass Apr 01 15:49:15 koen: Look at the definiton of pkgoutdir ;-) Apr 01 15:49:27 Why sort afterwards when you can output there directly ;-) Apr 01 15:49:57 I wonder why vi didn't match that for 'ARCH' Apr 01 15:50:16 RP: ipk.bbclass puts them there directly as well Apr 01 15:51:54 zecke: pdaXrom is awesome, they have tslib 1.3 Apr 01 15:57:57 03koen 07org.oe.dev * ra8568a54... 10/ (1 packages/pulseaudio/pulseaudio_0.9.5.bb): pulse: remove rprovides, since OE changes them to libesd0, thus breaking esd using apps Apr 01 16:15:11 bleh Apr 01 16:15:16 stupid shlib renaming Apr 01 16:17:52 very stupid ;-E Apr 01 16:22:18 RP: http://www.oesf.org/forums/index.php?showtopic=23629&pid=157742&st=0&#entry157742 Apr 01 16:25:22 koen: Interesting. I should look at pulling some of these into OE Apr 01 16:25:51 2.6 on collie looks a lot more usable now Apr 01 16:25:57 ~spell usable Apr 01 16:25:59 'usable' may be spelled correctly Apr 01 16:26:16 koen: Although I suspect the driver is slow :/ Apr 01 16:28:57 koen: yes? Apr 01 16:29:13 HU-Berlin... Apr 01 16:32:07 RP, is there a recommended userspace API to leds class? Apr 01 16:32:41 or is it just write to the brightness files in /sys ? Apr 01 16:32:46 03koen 07org.oe.dev * rd212e3b5... 10/ (1 packages/gpe-conf/gpe-conf_0.2.5.bb): gpe-conf: bump PR Apr 01 16:39:57 jacques: None exists, yet... Apr 01 16:40:12 hi rp Apr 01 16:40:17 hi woglinde Apr 01 16:42:49 RP, so for now writing to brightness files is as good as any? Apr 01 16:45:43 RP: no hal support yet? Apr 01 17:11:25 hi Apr 01 17:11:37 hey Bernardo Apr 01 17:19:20 I know that today is April Fools day (April 1st), but, a pull this morning of the OE.mtn database shows massive changes to the org.openembedded.dev/classes. Apr 01 17:19:34 jacques: yes Apr 01 17:19:37 bitbake chokes at not finding "shasum" Apr 01 17:19:38 koen: Sadly not Apr 01 17:19:52 T0mW: chokes? Apr 01 17:20:03 T0mW, I had that problem last night Apr 01 17:20:12 T0mW: like printing a warning, or like backtrace? Apr 01 17:20:14 sh: shasum: command not found Apr 01 17:20:14 NOTE: The SHA256 Sums do not match. Wanted: 'c0810fb3ddb6cb73c9ff045965e542af6e3eaa7f2995b3037181766d26d5e6e7' Got: '' Apr 01 17:20:14 NOTE: Task failed: Checksum of 'ftp://ftp.astron.com/pub/file/file-4.20.tar.gz' failed Apr 01 17:20:23 T0mW, it's a perl package - Digest-SHA Apr 01 17:20:38 T0mW: which `shasum` Apr 01 17:20:43 why oe doesn't just call sha256sum directly I dunno Apr 01 17:20:45 which shasum even Apr 01 17:20:55 jacques: because OSX doesn't have that :) Apr 01 17:21:00 zecke: none found Apr 01 17:21:01 zecke: update requiredsoftware? Apr 01 17:21:21 I have sha1sum Apr 01 17:21:40 T0mW: jacques : Either read the mailinglist or blogs to get informed Apr 01 17:22:54 zecke: I was just running a final test on the build this morning, moved the source tree out of the way, then went for a fresh load before I packaged this stuff up. I wanted to announce the availability of this OE project for Zipit. Apr 01 17:23:29 zecke: it was just weird that it worked last night, and then this morning I had breakages. Apr 01 17:24:09 T0mW: Check the latest checkins, some new functionality was added Apr 01 17:24:33 T0mW: Check commits, check mailinglists, check blogs the universe is moving on Apr 01 17:25:18 ok, I'm dealing with it. I just didn't know if anyone else knew that things broke. thanks Apr 01 17:25:30 T0mW: we didn't it Apr 01 17:25:46 I didn't. jacques was aware but did not say anything Apr 01 17:26:04 I said quite abit last night Apr 01 17:26:17 everything I said above, including which perl package was needed Apr 01 17:27:15 zecke: shasum should perhaps be added to sanity.conf? Apr 01 17:27:30 jacques: thank you, urpmi perl-Digest-SHA fixed the prob Apr 01 17:27:34 03freyther 07org.oe.dev * rafa30356... 10/ (1 classes/sanity.bbclass): classes/sanity.bblcass: Require shasum since base.bbclass is calling it Apr 01 17:27:46 RP: okay? Apr 01 17:27:58 zecke: Yes, thanks :) Apr 01 17:28:13 RP: I'm updating the wiki now Apr 01 17:28:15 T0mW, np :-) Apr 01 17:28:17 I know about md5sum, just didn't know (or think of) the sha sum Apr 01 17:28:35 next utility required (even if it is required by bitbake already) Apr 01 17:32:00 koen: Required Software was updated as well Apr 01 17:32:07 pH5_, psokolovsky__ : http://projects.linuxtogo.org/plugins/scmsvn/viewcvs.php?rev=9087&root=gpe&view=rev Apr 01 17:32:09 zecke: thanks Apr 01 17:32:30 hi Apr 01 17:33:48 zecke: did you add libdigest-sha-perl for ubuntu users? Apr 01 17:33:49 hrw|gone: do you have access to the source mirror? Apr 01 17:34:08 Bernardo: no, I should scroll down and check that page Apr 01 17:34:17 Bernardo: I hit packages.ubuntu.com for that stuff Apr 01 17:35:06 I just found out I needed that when one of the packages tried to check sha1 sum and found I didn't have shasum installed by default Apr 01 17:36:29 Bernardo: hehe, you are from yesterday :) Apr 01 17:36:36 Bernardo: just before you entered we had this question Apr 01 17:37:06 lol Apr 01 17:37:26 I just managed to get my system up again and startend build oe first thing... :) Apr 01 17:37:32 started Apr 01 17:38:37 had a psu burn down, which in turn burned a hd. The main problem is that it didn't burn at once, started failing, and I didn't notice the emails from mdadm until it was too late and I had lost data... Apr 01 17:40:16 should have opened kvirc 10 minutes earlier... :) Apr 01 17:40:43 Bernardo: upgrade to konversation or use screen+irssi Apr 01 17:41:25 konversation instead of kvirc? Might be a good idea Apr 01 17:41:41 :) Apr 01 17:41:53 03koen 07org.oe.dev * r6bd9066c... 10/ (7 files in 3 dirs): gpe-dm: add 0.50 and clean up files Apr 01 17:42:46 now if I can get lirc-modules to build, I'll be happy Apr 01 17:47:41 03koen 07org.oe.dev * rf3c1220c... 10/ (1 packages/gpe-edit/gpe-edit_0.40.bb): gpe-edit: add 0.40 Apr 01 17:52:09 oehm, RP, zecke, anyone? What would be the recommended way to keep data (cached 'svn info' results, for example) between tasks in a single bitbake run. I just found that I get a fork() for each task, thereby deleting my variables Apr 01 17:52:30 03koen 07org.oe.dev * rbf8b0ba8... 10/ (1 packages/gpe-dm/files packages/gpe-dm/gpe-dm_0.50.bb): gpe-dm: really add 0.50 Apr 01 17:52:34 03koen 07org.oe.dev * ra2a88493... 10/ (3 files in 3 dirs): libmimedir: switch to svn, remove SRCDATE from sane-srcdates.inc Apr 01 17:53:17 can I make a simpad development update bugreport with multiple files for packages/linux/linux/simpad and one propper patch for linux_2.6.20.bb or do I need to split it? there is also a name change for one file. Apr 01 17:53:42 mr_nice: one bug is fine Apr 01 17:53:50 zecke: thx Apr 01 17:53:58 mr_nice: make sure to not have garbage in your patches (it is time to use quilt/stgit) :) Apr 01 17:54:24 hehe, I will double check this time ;) Apr 01 17:56:50 zecke: mtn 0.34 has a 'mtn clone' command, which should make a mtn fetcher a bit easier Apr 01 18:05:25 03freyther 07org.oe.dev * r9328774a... 10/ (1 classes/sanity.bbclass): classes/sanity.bbclass: md5sum is required as well Apr 01 18:05:31 03freyther 07org.oe.dev * rcda80ec8... 10/ (1 conf/checksums.ini): conf/checksums.ini: Remove the PN/PV from the section and only use the SRC_URI Apr 01 18:05:36 03freyther 07org.oe.dev * r2fdc45b3... 10/ (6 files in 3 dirs): Apr 01 18:05:36 packages/libmrss: Apply my patches to libmrss (will be send upstream) Apr 01 18:05:36 -Make sure pubDate is always set for ATOM feeds from planets Apr 01 18:05:36 -Bump the library version to 0.17 Apr 01 18:05:36 -Format Atom dates RFC822 compliant Apr 01 18:05:42 03freyther 07org.oe.dev * r4ee39786... 10/ (3 files in 3 dirs): packages/{libmrss|libnxml}: Make koen happy and replace '//' with '/' Apr 01 18:05:47 03freyther 07org.oe.dev * rfad527d2... 10/ (4 files in 4 dirs): pckages/{libmrss|libnxml}: Remove the md5sum and add md5sum/shasum to the config file Apr 01 18:11:35 Henryk: good question. What kind of data, how often and when will it be accessed? Apr 01 18:12:08 * koen would just do 'svn info' everytime Apr 01 18:16:29 zecke: i have added code to the fetcher classes so that they can retrieve the remote source revision (plus generate an identifier suitable for integration in PV). than I have code in base.bbclass in an after_parse style (though I'm open to better ideas) that calls into the fetcher code and sets variables of the style SRCREV_foo (for foo in svn1, git1, and similar) to be included in PV as PV = "1.2.0+svn${SRCREVsvn1}+git${SRCREVgit1}". Apr 01 18:16:34 can anyone open nslu2-linux.org page Apr 01 18:17:29 due to the new way bitbake works my after_parse code is called once for each task and in turns calls the fetcher revision code once for each task. that code already does caching (in urldata) but due to the same fork() that caching doesn't work Apr 01 18:18:29 calling out to svn info, git ls-remote, etc. once per task is not acceptable (high load, unecessary waiting times) and also not correct (we wouldn't want to allow the source revision to change during a bitbake run). Apr 01 18:19:29 how would it change? Apr 01 18:19:44 you aren't doing a 'svn up' or 'git pull' in between Apr 01 18:19:55 koen: you do remove svn info :) Apr 01 18:20:13 koen: remote svn info. that's the very heart of the functionality i'm implementing Apr 01 18:23:23 zecke: so I need some way to store data through more than a task but not through more than one bitbake run. e.g. I need what I would get by using urldata, where it not for the fork Apr 01 18:24:08 Henryk: well this after parse (anonfuncs) are called during parse and the file gets reloaded/reparsed before executing Apr 01 18:24:25 Henryk: you can either use the global cache and cache dates there? Apr 01 18:24:40 global cache? Apr 01 18:24:58 cache.py Apr 01 18:25:35 we store some (expanded) keys inside a cache to speedup bitbake startup Apr 01 18:27:02 ah Apr 01 18:28:12 http://www.sigmund.nl/strips_eng/set9/7.gif Apr 01 18:28:16 Henryk: I think this would be an obvious space to store small tokens, I think we currently don't expose a way to store extra tokens though Apr 01 18:32:22 which version of shasum do I need for oe? Apr 01 18:32:52 in openembedded, how can i revert to last well known config? will a svn pull of specific branch work ? or do i need to play with mtn Apr 01 18:35:31 I have shasum v. 1.1 which has no -a option and so I am getting an error http://pastebin.ca/419892. Is oe using a patched version of shasum? Apr 01 18:44:10 03freyther 07org.oe.dev * r2f8a1c80... 10/ (1 packages/libmrss/files/fix_atom_date_locale.patch): packages/libmrss: Remove the printf from the patch Apr 01 18:45:07 koen: if you have time can you test the new gtk-webcore and gpe-mini-browser svn? Apr 01 18:45:27 As my h3850 is still not in working order Apr 01 18:55:31 zecke: ok, just to let you know, everything built ok after adding the shasum Apr 01 18:56:12 T0mW: if you are bored feel free to add shasum's :) Apr 01 18:56:27 lol Apr 01 18:56:53 too much code and not enough time. :-( Apr 01 19:02:50 hi T0mW zecke Apr 01 19:03:16 chouimat: good morning Apr 01 19:05:56 ok I found it out oe wants Digest-SHA-5.44 and not shasum Apr 01 19:24:10 anyone tried ext4? Apr 01 19:26:43 hello all Apr 01 19:27:37 likewise: likewise Apr 01 19:27:57 hi likewise Apr 01 19:37:43 [21:35] daniels: need to add a 'no, you cannot optimise fbCompositeGeneral, really, i'm not kidding, just make sure it never gets called' bullet point for our performance team Apr 01 19:37:45 heh Apr 01 19:38:24 oh Apr 01 19:38:31 sorry Apr 01 20:05:24 does any know how to set up a failover ip if busybox's dhcp client is not able to get one ? Apr 01 20:07:47 Ifaistos: install angstrom :) Apr 01 20:08:00 that will run avahi-autoip Apr 01 20:10:45 i have a client that his dhcp server does not assign an ip to a device we have installed, and i want to make sure that at least it will get something we can reach wehn tis happens Apr 01 20:11:29 koen : does the busybox dhcp client has such an option ? Apr 01 20:11:37 no idea Apr 01 20:11:47 I use dhcp or ipv6 Apr 01 20:20:31 for those who want to read on the subject of auto ip configuration -> http://www.science.uva.nl/research/air/wiki/ZeroconfSoftware Apr 01 20:20:32 Ifaistos: look at the udhcpc.d scripts that come with avahi-autoipd (in the avahi/files folder) Apr 01 20:24:57 zecke: it seems llvm-gcc4 will behave just like gcc4 if you want a cross-compiler Apr 01 20:25:51 03koen 07org.oe.dev * r1d81bf38... 10/ (1 packages/openmoko-base/openmoko-session_svn.bb): openmoko-session: depend on mb-panel-2 and mb-wm instead of 'matchbox' Apr 01 20:26:25 koen: compiling was? the gcc4 frontend is ARCH independant (mostly) Apr 01 20:27:42 the people on the mailing list say 'just' and 'only' a lot, but it seems hopefull Apr 01 20:28:54 hehe Apr 01 20:29:03 koen: so angstrom will use llvm built images? Apr 01 20:29:13 maybe Apr 01 20:29:20 I do want to have llvm support in OE Apr 01 20:29:37 koen: are you member of the ADC? Apr 01 20:30:07 I subscribed once to get X11 Apr 01 20:30:34 http://lists.cs.uiuc.edu/pipermail/llvmdev/2007-April/008417.html Apr 01 20:31:24 --target=arm-iphone-darwin. Apr 01 20:31:27 rofl Apr 01 20:32:02 even if he says it is fictional :) Apr 01 20:37:07 * koen heads to bed Apr 01 20:37:09 'night all Apr 01 20:49:53 koen|away: llvm only uses the GCC frontends Apr 01 20:50:08 koen|away: this means being able to parse C, C++, Java into the internal tree representation Apr 01 20:50:20 koen|away: the patches we have are for code generation on a later stage Apr 01 21:02:57 03freyther 07org.oe.dev * r4d6a07a1... 10/ (3 files in 3 dirs): packages/libmrss: Add a memory leak patch and bump the PR Apr 01 21:16:20 zecke: [20:14] pH5: what's wrong with sha1sum ? Apr 01 21:16:30 zecke: I don't know. any reason not to use that? Apr 01 21:16:59 which sha1sum Apr 01 21:17:00 no sha1sum in /space/dports/bin /space/dports/sbin /Library/Frameworks/Python.framework/Versions/Current/bin /bin /sbin /usr/bin /usr/local/bin /usr/sbin Apr 01 21:17:25 hm. it is in debian's coreutils Apr 01 21:19:10 pH5: I did this on OSX and it had shasum installed Apr 01 21:19:18 pH5: and my feisty seemed to have it installed as well Apr 01 21:19:35 debian doesn't even have a package for it. Apr 01 21:19:45 pH5: Ubuntu has :) Apr 01 21:19:48 and to make matters more interesting, shasum and sha1sum don't produce the same result Apr 01 21:20:01 pH5: debian has a package Apr 01 21:20:04 really? Apr 01 21:20:12 pH5: sha1, I think this generates sha1 sums Apr 01 21:20:21 pH5: where shasum -a256 generates sha256 sums? Apr 01 21:21:05 pH5: I installed libdigest-sha-perl which seems to have something that works Apr 01 21:22:20 and there seems to be something similar for sha1sum : libdigest-sha1-perl Apr 01 21:22:29 But it is not obvious at all Apr 01 21:23:27 shasum from libdigest-sha1-perl and sha1sum from coreutils produce the same result Apr 01 21:23:48 pH5: so we might want to call sha256sum? Apr 01 21:26:23 pH5: philippe : I'm happy to use a tool more commonly installed. The question is, is sha256sum more commonly installed? Apr 01 21:26:31 zecke: that'd be better. sha256sum is part of coreutils. Apr 01 21:26:49 libdigest-sha-perl is really non-obvious Apr 01 21:26:52 hi Apr 01 21:27:27 pH5: feel free to change base.bbclass and the call to os.popen("shasum -a256 -p " + localpath) Apr 01 21:27:52 pH5: e.g. my coreutils on OSX do not have a sha256sum :} Apr 01 21:28:32 hi, all! anybody have an idea about status of page_to_bus on ARM? when it is last exieted if ever? any links on its history? Apr 01 21:28:36 unnh, and sha256sum from coreutils doesn't support -p Apr 01 21:28:51 pH5: -b for binary Apr 01 21:29:02 ah, ok. Apr 01 21:29:05 -B even Apr 01 21:29:15 just skip that Apr 01 21:29:54 the sha256sum manpage says -b binary and -t text mode (default). Apr 01 21:30:15 --help doesn't list that :) Apr 01 21:30:27 ~lart slightly different implementations Apr 01 21:30:28 * ibot changes slightly different implementations's permissions to 0777 and tells the world Apr 01 21:33:28 pH5: make sure to update sanity.bbclass and the RequiredSoftware entry in the wiki Apr 01 21:34:28 so just switch to ("sha256sum " + localpath) and hope they produce the same result on all platforms? Apr 01 21:34:38 yes Apr 01 21:34:43 ok, will do. Apr 01 21:35:10 pH5: or pass -b Apr 01 21:35:14 people will complain :) Apr 01 21:35:34 does osx sha256sum support -b? Apr 01 21:35:42 OSX has no sha256sum :) Apr 01 21:35:46 aha. Apr 01 21:35:51 and even macports don't install anything like this Apr 01 21:36:03 I will alias sha256sum to openssl locally :) Apr 01 21:36:06 RP, ping Apr 01 21:36:07 :) Apr 01 21:36:14 as OSX is not the primary platform for OE yet Apr 01 21:38:16 nite all Apr 01 21:44:46 03pH5 07org.oe.dev * r15e084fa... 10/ (1 classes/base.bbclass classes/sanity.bbclass): base.bbclass, sanity.bbclass: use sha256sum instead of shasum -a256 Apr 01 21:48:44 The next question is how to populate the checksums.ini Apr 01 21:49:02 a) Trust the source we currently use and use the md5sum and sha256sum of already downloaded files Apr 01 21:49:12 b) go to upstream and check their sums Apr 01 21:49:28 c) only at new sums Apr 01 21:51:58 a) practical, but not completely secure. b) secure, but not all upstream sources have sums. c) practical near term, but doesn't really solve any overall problem. JMO, that and USD 5 will get you a fancy cup of coffee. Apr 01 21:53:18 mwester: right a) practical and not secure but fits our current trust model Apr 01 21:53:47 Trial balloon: b) where-ever possible, a) where not, and provide an definite date or event by which all sources must be summed. Apr 01 21:53:48 a should be enough, any forgery would eventually be discovered upstream Apr 01 21:53:50 slapin_nb: pong Apr 01 21:55:05 RP, hi! Sorry for disturbance, do you remember what is page_to_bus and how to get rid of it to make old driver to compile with new kernel? Apr 01 21:55:32 RP, arch is arm Apr 01 21:55:36 15e084fa8118e1a49559b4193f626d0c5ad7cce0 pH5@openembedded.org 2007-04-01T21:42:48 Apr 01 21:55:36 mtn: 80791bbfbbfd682c0e16ebb739bfb7041451a1c1 freyther@openembedded.org 2007-04-01T21:01:12 Apr 01 21:56:03 we have two heads again... Apr 01 21:56:16 philippe: sorry, I'll merge Apr 01 21:56:17 CoreDump|home: how so? For projects that are no longer very active, or where OE is behind some releases, how would a forgery be discovered? Apr 01 21:56:39 I agree that a) is better than nothing, though. Apr 01 21:56:53 slapin_nb: virt_to_bus(page_address(page)) ? Apr 01 21:57:11 RP, is it correct to replace it by page_to_phys? Apr 01 21:57:18 pH5: Thanks! Apr 01 21:57:19 mwester: Once _upstream_ detects the breach, they will update the sources and the checksum will no longer match. If upstream does not detect the breach we are f*cked in any case =) Apr 01 21:58:11 Anyway I just realized why I could not merge myself... Apr 01 21:58:15 stupid me Apr 01 21:58:38 slapin_nb: Possibly is the best answer I can give Apr 01 21:59:23 RP: you mean virt_to_bus(page_address(page))? Apr 01 22:00:13 slapin_nb: I meant that page_to_phys() might work ok Apr 01 22:00:30 RP, thanks a lot! Apr 01 22:00:36 slapin_nb: Its difficult to say without context really... Apr 01 22:01:00 CoreDump|home: the use case where upstream does not detect is the one I'm considering; that's where a) is better than nothing (will detect future tampering with a source) Apr 01 22:01:11 RP, xscale based board, camera interface driver. Apr 01 22:02:02 mwester: that's as good as it gets IMO. Apr 01 22:02:23 I also wonder how checksumming of CVS / SVN / MTN checkouts is performed Apr 01 22:02:35 I will go for a) and then make checksums mandantory Apr 01 22:03:01 CoreDump|home: for CVS/SVN we need to use svn export and in this case can checksum the tarballs of the stash Apr 01 22:03:25 CoreDump|home: for developing against active trees like OpenMoko we need to be able to disable it Apr 01 22:03:34 * mwester wonders at the practicality of a utility to test the OE checksums database against "sourcename".md5 automatically... Apr 01 22:03:37 zecke: or checksum every single file =) Apr 01 22:03:48 Gack! Apr 01 22:04:14 mwester: I will do this tomorrow, against an untrusted source mirror of my choiche :} Apr 01 22:04:37 CoreDump|home: for git, you can do verify tag but still need to trust that linus is linus Apr 01 22:04:42 * mwester will be happy to help if he can Apr 01 22:05:07 CoreDump|home: for mtn you just do a pull and wait :) Apr 01 22:05:22 heh Apr 01 22:06:34 I'd be happy to provide sha256sums of my source directory Apr 01 22:06:45 03philippe 07org.oe.dev * rdb9dae5e... 10/ (3 files in 3 dirs): Apr 01 22:06:45 packages/gtk-webcore/osb-nrcore_svn Apr 01 22:06:45 * Update svn build information because of major changes Apr 01 22:06:45 * added patch to correct build and silence gtk-webcore console output Apr 01 22:06:45 * packages/gtk-webcore/files/build_silence.patch Apr 01 22:07:33 Could be interesting if we had multiple source directories... Apr 01 22:07:45 indeed Apr 01 22:08:17 philippe: is someone working on the webcore branch again? Apr 01 22:08:21 mickey|zzZZzz: ping?! Apr 01 22:08:50 zecke: yes the original author has synced up with webkit a bit and a number of patches have been going in recently Apr 01 22:09:09 From what I tested today it has improved a lot compared to the latest svn Apr 01 22:09:47 philippe: is that from the s60 branch or trunk of webkit? Apr 01 22:10:44 zecke: http://www.pastebin.ca/420143 Apr 01 22:11:22 * zecke hopes mickeyl will finally judge the GSoC proposal and that I can work on WebKit/Gdk this summer :} Apr 01 22:11:51 zecke: would be nice... I will gladly test it with gpe-mini-browser Apr 01 22:12:11 * zecke crosses his fingers Apr 01 22:12:44 philippe: thanks for the info. current safari is progress as well Apr 01 22:13:38 well I cannot test it for ARM atm unfortunately because I have no platform with decent support anymore... Apr 01 22:14:08 Well the neo (openmoko) should do soon... Apr 01 22:14:29 eek, I might start a build for my shark and see how far I get Apr 01 22:15:17 zecke: sounds great. Let me know how the performance is. Apr 01 22:15:48 koen told me a while ago that the previous svn builds would segfault in his builds Apr 01 22:19:43 zecke: http://hentges.net/tmp/sha256sums.txt ( http://hentges.net/tmp/sha256sums.txt.sig) Apr 01 22:21:14 gotta sign those signatures! Apr 01 22:21:16 :D Apr 01 22:21:25 =) Apr 01 22:21:34 CoreDump|home: that is the easy part, now match these to the origin (SRC_URI) Apr 01 22:21:55 CoreDump|home: and you work at night, so you must be some kind of vampire, I think I do not trust vampire :) Apr 01 22:22:02 =D Apr 01 22:22:03 hehehe! Apr 01 22:22:26 zecke: verification is only done on download or not? Apr 01 22:22:29 that implies parsing .bb files, no? Apr 01 22:22:44 CoreDump|home: yes, after fetching the files Apr 01 22:22:54 * mwester means matching to SRC_URI requires parsing the .bb files Apr 01 22:22:55 CoreDump|home: this makes us require the do_fetch which we already have Apr 01 22:23:11 right Apr 01 22:23:21 mwester: we have a small secret utility (bittest) which allows to write easy tests Apr 01 22:24:18 BTW: when are dispose and finalize called in a GObject life cycle? Apr 01 22:26:22 night all Apr 01 22:26:28 n8 philippe Apr 01 22:31:30 zecke: is the checksumming code already written? If so, which version of BB if required? Apr 01 22:31:52 CoreDump|home: comparing the checksum? Apr 01 22:31:56 yes Apr 01 22:32:05 CoreDump|home: the is code is written and found in base.bbclass Apr 01 22:32:16 CoreDump|home: and take a look at checksums.ini on the format of this file Apr 01 22:52:45 good night :) Apr 01 22:56:11 'night zecke Apr 01 22:56:29 RP: in case this was lost ACK on TERMCMD Apr 01 22:57:05 zecke: ok, thanks :) Apr 01 22:57:15 but remove kconsole Apr 01 22:57:30 k-farta-morgana new console Apr 01 22:57:52 zecke: Right, I'll either test and fix or remove Apr 01 22:58:09 remove Apr 01 23:00:22 zecke: I suspect a --type "bitbake.random number" might do what we need but I will remove if you're happier ;-) Apr 01 23:01:37 -e cmd -T title Apr 01 23:02:07 yes, I'm thinking of trying to solve the wrong problem :/ Apr 01 23:02:35 e.g. if screen + xterm works it is okay to start with Apr 01 23:02:47 later! Apr 01 23:10:58 03rpurdie * r810 10bitbake/ (ChangeLog bin/bitbake lib/bb/cooker.py): cooker.py: Make -I option a generic one to extend ASSUME_PROVIDED Apr 02 02:05:42 hmm **** ENDING LOGGING AT Mon Apr 02 02:59:56 2007