**** BEGIN LOGGING AT Thu Apr 26 02:59:57 2007 Apr 26 05:14:43 CoreDump|home: Thanks for setting up svn for me. I got your message right at the end of the day :) Apr 26 06:19:26 morning all Apr 26 06:24:31 OE will have 5% market share of embedded distros according to the linuxdevices survey :) Apr 26 06:26:33 Ifaistos: whooo.... We WILL take over the world...:) Apr 26 06:40:56 hvontres|home : take over the world one dev board the time :) Apr 26 06:47:42 fun with google maps: http://maps.google.com/maps?f=q&hl=en&q=new+york+to+berlin&layer=&sll=45.39845,-35.859375&sspn=42.449264,81.5625&ie=UTF8&z=3&om=1 Apr 26 06:50:31 point 24 is the fun part.... Apr 26 06:51:34 Ifaistos: but no routes across the pacific yet..:( Apr 26 06:54:41 * Ifaistos goes to make some fresh coffee while he ponders what will happen after OE takes over the world Apr 26 06:55:29 Ifaistos: Robotic coffee makers on wheels :) Apr 26 06:58:52 hvontres|home : so we will need OE hard realtime :) Apr 26 07:00:12 * Ifaistos adds OE hard realtime on its todo list, for things left to do after the world is taken over Apr 26 07:00:40 let's do some paying work now :) Apr 26 07:12:32 03lenehan 07org.oe.dev * rce89c98f... 10/ (4 files in 3 dirs): Apr 26 07:12:32 perl 5.8.8: Move configuration files for libnet and cpan to /etc/perl to Apr 26 07:12:32 match the debian configuration. It moves these configuration files to /etc/ Apr 26 07:12:32 to allow for read only /usr installations. Apr 26 07:12:49 time for bed :) Apr 26 07:14:33 Good morning! Apr 26 07:56:33 morning Apr 26 07:59:40 morning Apr 26 08:11:08 03lenehan 07org.oe.dev * rb72cc605... 10/ (4 files in 3 dirs): (log message trimmed) Apr 26 08:11:08 perl 5.8.8: Minor updates to the packaging: Apr 26 08:11:08 * don't install packlists - we don't need them and it saves space (debian Apr 26 08:11:08 patch) Apr 26 08:11:08 * don't split POSIX module over perl and perl-arch dirs (debian patch) Apr 26 08:11:08 * add a -doc package and put all the random documentation and examples from Apr 26 08:11:10 the perl tree in it. Apr 26 08:12:22 hello Apr 26 08:14:10 * v8jlene declares perl 5.8.8 ready for use Apr 26 08:14:37 sorted Apr 26 08:20:36 building netbase in angstrom fails for me (and thus angstrom-console-image). netbase is looking for non-existant files in the OE tree: http://rafb.net/p/eF5uUR73.html What is the solution? Remove the entries for if-pre-up.d the netbase bb file or copy the files to OE? In that case, where do I get them? Apr 26 08:23:42 remove them and write to ML about it Apr 26 08:23:51 probably someone forgot to add files Apr 26 08:24:04 brb Apr 26 08:26:41 Laibsch: Looks fine here. Do you the empty if-pre-up.d directory in packages/netbase/netbase? Apr 26 08:27:04 you "have" the Apr 26 08:36:30 * zecke wonders why my scanelf fails to find these RPATHs Apr 26 08:37:01 RP: BitBake wise we are in a bad situation :} Apr 26 08:37:39 OE MetaData doesn't really work with 1.6 Apr 26 08:37:45 BitBake 1.8 has issues :} Apr 26 08:38:50 v8jlene: http://rafb.net/p/j9epK273.html has the output of "find packages/netbase" "mtn ls missing" showed the file you were talking about was missing. I will retry compilation (and check if there are other files missing for whatever reason). Apr 26 08:39:12 * Laibsch does mtn revert --missing packages/netbase Apr 26 08:40:10 ~lart ATI Apr 26 08:40:10 * ibot whips out his power stapler and staples ATI's foot to the floor Apr 26 08:49:10 netbase unpacks fine now. Thanks, guys. Apr 26 08:59:57 how to determine which version was the package that i set in RDEPENDS? Apr 26 09:04:54 how to determine the package version which package is set in RDEPENDS? Apr 26 09:13:18 03pfalcon 07org.oe.dev * r558f128d... 10/ (3 files in 3 dirs): (log message trimmed) Apr 26 09:13:18 busybox 1.2.1: angstrom: Make syslogd log to /var/log/messages. Apr 26 09:13:18 * With rotation and size/rotation count limit, which coupled Apr 26 09:13:18 with /var/log being a tmpfs, makes it not worse than canonical Apr 26 09:13:18 circular buffer logging, but much more friendly for people not Apr 26 09:13:19 familiar with busybox. Apr 26 09:13:20 * Closes #2135. Apr 26 09:44:35 ~change 226 GBP to EUR Apr 26 09:44:42 226.00 Pound (GBP) makes 332.063 Euro (EUR) (from http://www.xe.com/) Apr 26 09:47:24 koen: on a spendign spree? Apr 26 09:50:23 I'm returning a batch of unused lenses today Apr 26 09:50:33 then someone posted about http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=300103306550 :) Apr 26 10:02:03 koen: Im assuming thats a good camera :-) Apr 26 10:02:11 koen: as I know nothing about cameras Apr 26 10:03:00 XorA: it's retro Apr 26 10:03:09 koen: ah so the zx81 of cameras :-) Apr 26 10:03:11 and built by german using steel and concrete ;) Apr 26 10:03:22 ah the BBC Master of cameras :-D Apr 26 10:03:54 :) Apr 26 10:04:22 someone remember where in /sys/ I can check does Neo gets 100 or 500mA? Apr 26 10:04:52 * XorA still looks for an affordable zx81/zx80 Apr 26 10:05:50 hrw: /sys/devices/platform/s3c2410-i2c/i2c-0/0-0008 Apr 26 10:05:58 koen@neo1973:/sys/devices/platform/s3c2410-i2c/i2c-0/0-0008$ cat chgmode Apr 26 10:05:58 fast_cccv Apr 26 10:06:07 iirc fast_cccv = 500mA Apr 26 10:06:16 idle == full, not charging? Apr 26 10:06:28 I guess so Apr 26 10:10:54 I need to check does my USB2.0 hub with integrated card-reader is burnt inside (no cardreader visible) or other crap (still have warranty) Apr 26 10:35:42 zecke_: What's the problem with 1.6? Apr 26 10:36:40 Also, which 1.8 problems do we have? Apr 26 10:37:04 !oebug 2133 Apr 26 10:37:05 * * Bug 2133, Status: NEW, Created: 2007-04-23 01:53 Apr 26 10:37:06 * * koen(AT)dominion.kabel.utwente.nl: bitbake 1.8 r816 breaks SRC_DATE Apr 26 10:37:07 * * http://bugs.openembedded.org/show_bug.cgi?id=2133 Apr 26 10:38:07 koen: Any idea why its doing that? Apr 26 10:38:19 "because of bugs" Apr 26 10:38:24 I really have no idea Apr 26 10:38:25 koen: Which SRCDATE was correct and what did the config files say? Apr 26 10:38:50 sane-srcdates has 20051101 for libfakekey Apr 26 10:39:15 it looks like OE/bitbake only applies SRCDATE to ${PN}, but not ${PN}-dbg Apr 26 10:39:38 but it does add a dep for libfakekey- to all apps Apr 26 10:41:41 koen: I made a SRCDATE change to base.bbclass recently when I made those "backwards compatibility" changes. I'm going to bet it was that Apr 26 10:41:52 03lenehan 07org.oe.dev * raa1ed7ac... 10/ (1 packages/linux/linux-titan-sh4/titan-flash.patch): Apr 26 10:41:52 linux-titan-sh4: Update flash driver so that it'll build with recent kernels: Apr 26 10:41:52 * Remove remains of devfs support. Apr 26 10:41:52 * Implement HDIO_GETGEO ioctl in place of getgeo block device operation. Apr 26 10:41:56 03lenehan 07org.oe.dev * red96c8df... 10/ (5 files in 3 dirs): Apr 26 10:41:56 linux-titan-sh4: Add 2.6.21 kernel. This moves away from the old linux-sh Apr 26 10:41:56 cvs and uses the main linux kernel directly. Apr 26 10:42:00 03lenehan 07org.oe.dev * r07175ec3... 10/ (4 files in 3 dirs): Apr 26 10:42:00 linux-titan-sh4: Remove the older 2.6.17 kernel that was based on the Apr 26 10:42:03 old linux-sh cvs tree. Apr 26 10:42:56 .21? Apr 26 10:43:04 ah, released today Apr 26 10:43:12 koen: Can you try adding the bit marked by "# OBSOLETE in bitbake 1.7.4" in http://www.rpsys.net/openzaurus/temp/oe-depends-fixes-r3.patch back and see if that fixes it? Apr 26 10:45:29 * koen tries Apr 26 10:46:22 It's over 7hrs since the .21 announcement , plenty of time to test a kernel ;) Apr 26 10:46:25 RP: that indeed seems to fix it Apr 26 10:46:33 Linux home 2.6.21 #16 PREEMPT Thu Apr 26 11:29:10 CEST 2007 x86_64 GNU/Linux Apr 26 10:46:48 v8jlene: is perl 5.8.8 ready for use? Apr 26 10:47:00 koen: Ok, lets add that bit of code back and I'll work out why it breaks at some future date ;-) Apr 26 10:47:38 * koen mtn commit Apr 26 10:48:08 koen: Just about. The only outstanding issue is that it doesn't build with uclibc. But apart from that I'm all done on perl itself. Apr 26 10:48:14 argh.. Apr 26 10:48:25 usb 2.0 has limit of cable lenght? Apr 26 10:48:32 koen: I've tested sh4 and i486. It'll obviously need a bit of testing on other targets. Apr 26 10:48:37 hrw: no, USB itself has Apr 26 10:48:40 hrw: 5m iirc Apr 26 10:48:42 sure, don't ask me what it is Apr 26 10:49:53 RP: I don't know. Either it is the usual me-too-broken syndrom or we have grave dependency issues and a stability crisis Apr 26 10:50:08 koen: its ~2m and I got 1.1 speed only Apr 26 10:50:20 RP: at least this my impression from the outside Apr 26 10:50:30 bbl Apr 26 10:50:36 *windoze* Apr 26 10:50:42 zecke_: Unless someone can point me at the problem, I can't fix it... Apr 26 10:50:55 hrw: is it on a desktop? Apr 26 10:51:01 RP: every day we get at least two mails indicating dependency issues Apr 26 10:51:16 RP: pax-utils-native not being built (well this looks like the base.bbclass issue) Apr 26 10:51:32 RP: ipkg-utils-native not present (probably same issue) Apr 26 10:52:13 djay-il: yes Apr 26 10:52:15 revision a17faaae1865241de5e0ea74486fc129ad0655ed should fix all that Apr 26 10:52:17 but admitingly the root-cause is probably us not keeping bitbake 'stable'. I know we change it on OE's request but still :) Apr 26 10:52:39 zecke_: I used to see shasum-native, pax-utils-native and ipkg-native not being built on any I try so I guess I should build an image now I have pulled Apr 26 10:52:48 zecke_: There does seem to be a bitbake 1.6 problem, I just need to try and reproduce it. Last I heard, it occured with bb 1.6.3 but 1.6.6 was untested Apr 26 10:53:04 RP: MACHINE=collie DISTRO=angstrom-2007.1 bitbake r819/1.9 - had to build ipkg-utils-native Apr 26 10:53:21 RP: I hope you don't get me wrong :) Apr 26 10:53:40 hrw: Not good :/ Apr 26 10:53:44 - depends = "${PACKAGE_DEPENDS} fakeroot-native" + depends Apr 26 10:53:44 + depends = "${PACKAGE_DEPENDS} fakeroot-native " + depends Apr 26 10:53:56 Just to make matters worse, I'm away for the next few days without computer access Apr 26 10:54:11 koen: will you be here in ~hour Apr 26 10:54:14 ? Apr 26 10:54:15 RP: take your time, you deserve it :) Apr 26 10:54:19 koen: Can you apply that? Apr 26 10:54:33 RP: applied ~yesterday Apr 26 10:54:41 koen: Ah, good Apr 26 10:54:46 * RP hopes that was at fault Apr 26 10:55:01 RP: Im testing now Apr 26 10:55:11 hrw: no, but I hope to be back in ~2 hours Apr 26 10:55:14 * hrw will restart oabi build from scratch then Apr 26 10:55:32 koen: ok - I have FPUEmu problem with glibc-intermediate for oabi Apr 26 10:59:55 hrw: sometimes I get from the different ports different speeds. Sometimes it is cable wiring. Apr 26 11:00:57 * koen reads http://spooky-possum.org/cgi-bin/pyblosxom.cgi/x86fp.html Apr 26 11:04:02 * hrw watchs distcc monitor Apr 26 11:04:07 hrw: How do you build openembedded-essential? I'd like to take a look at the "source" to apply it so a similar problem I have. Apr 26 11:04:19 apply it TO Apr 26 11:04:23 03koen 07org.oe.dev * rba162bf6... 10/ (1 classes/base.bbclass): base.bbclass: add back a snipped of obsolete code to fix SRCDATES problems, seems to fix #2133 Apr 26 11:06:45 hrw: did you try icecc? Apr 26 11:07:49 no Apr 26 11:09:29 ERROR: Can not check RPATH scanelf not found Apr 26 11:09:34 hrw: it does a pretty good job here - the central scheduler it pretty nice because you do not need to configure the compile machines to use. Apr 26 11:09:39 so pax-utils-native not built then Apr 26 11:10:17 XorA: bitbake -vvv please Apr 26 11:12:36 florian: first I want to get distcc support working here Apr 26 11:15:35 Hi! Does anyone have an idea which package provides tune2fs? It's not in e2fsprogs and e2tools Apr 26 11:18:00 XorA: This is bitbake 1.8 svn? Apr 26 11:19:44 <_law_> koen, new monotone revision? should fix bitbake bug? Apr 26 11:20:18 <_law_> koen, ok i´ll try Apr 26 11:22:04 RP: 1.8.0 Apr 26 11:23:51 zecke|away: of cource when you restart bitbake it goes off doing a whole different set of tasks than it tried last time :-( Apr 26 11:25:42 koen: here yet? Apr 26 11:25:57 /home/hrw/devel/build/oabi/tmp/work/arm-oabi-angstrom-linux/glibc-intermediate-2.5-r5/build-arm-angstrom-linux/libc_pic.os: Apr 26 11:26:01 In function `__monstartup': Apr 26 11:26:03 unwind-pe.c:(.text+0x95c40): undefined reference to `__floatsisf' Apr 26 11:28:14 * florian looks at the KVM console in his browser window... Apr 26 11:28:55 <_law_> XorA, update to 1.8 svn Apr 26 11:29:43 hrw: sorry, I'm leaving just now Apr 26 11:30:04 ok Apr 26 11:31:49 koen: http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=232&logfile=20070426132821.log for future Apr 26 11:32:00 koen: I am scratching this build and go from start Apr 26 11:32:39 hrw: When I was checking out distributed compiling everybody was advising me to use icecc rather than distcc. icecc is working fine for me. Apr 26 11:33:25 I believe there is some step-by-step guide at http://www.openembedded.org/icecc Apr 26 11:33:40 Laibsch: later will check it. Apr 26 11:34:24 Laibsch, hrw: icecc is better suited for *cross*compiling, whereas distcc is not. Apr 26 11:35:21 likewise: now all targets here crosscompile for i586 (from i686 and amd64) Apr 26 11:35:46 hrw: using distcc? Apr 26 11:35:51 yes Apr 26 11:36:46 hrw: what I know as a difference is that with icecc the daemon takes care of distributing the actual cross-compiler (using a tar.bz2 rootfs) where with distcc it's a manual task Apr 26 11:37:09 likewise: here targets will always have same toolchain ;D Apr 26 11:42:55 XorA: 1.8.0 is just plain broken, I'd use the svn branch which i need to make a release from... Apr 26 11:45:01 RP: ok cool, that'll be why so many users are having the problems then Apr 26 11:46:14 RP: let us disable the 1.8.0 release from downloads Apr 26 11:46:16 zecke|away: ok, I agree Apr 26 11:46:19 RP: and there is at least the SRC_DATE issue preventing the release of 1.8.2? Apr 26 11:46:41 zecke|away: No, we've put the workaround back and that should be 'fixed' for now Apr 26 11:46:59 zecke|away: I suspect something in OE is broken for that rather than bitbake itself Apr 26 11:47:55 I've just been waiting to make sure the depdency changes in OE.dev do really fix the problems with bitbake 1.8.x which I'm sure they do. We can release ASAP Apr 26 11:48:13 ~lart windows for being too stupid to say which dll is missing :} Apr 26 11:48:13 * ibot blasts windows to oblivion with a kamehameha wave for being too stupid to say which dll is missing :} Apr 26 11:48:45 RP: if you want to do a release we can wait some more days, otherwise I could do one tonight? Apr 26 11:49:12 zecke|away: I'm tempted to make one now ;-) Apr 26 11:49:52 if you can do that, i would be welcome it Apr 26 11:50:52 1.8.0 is hidden now Apr 26 11:51:51 can you ping me when you are done with the release, so i can update the debian package? Apr 26 11:51:51 morning Apr 26 11:53:25 yay bitbake svn shasum-native is first thing built Apr 26 11:54:12 XorA: I had that before Apr 26 11:54:25 XorA: will it build ipkg-utils-native is a question Apr 26 11:54:32 and it is now building pax-utils-native Apr 26 11:55:08 NOTE: package ipkg-utils-native-1.6+cvs20050404: completed Apr 26 11:55:46 03rpurdie * r817 10/ (5 files in 5 dirs): Release bitbake 1.8.2 Apr 26 11:56:52 Laibsch: openembedded-essential is in contrib/ Apr 26 11:58:53 Thanks Apr 26 12:00:31 I used equivs in the meantime which does what I wanted. Apr 26 12:01:52 Shoragan: released ;-) Apr 26 12:01:58 cool :) Apr 26 12:02:47 pulled, building :-) Apr 26 12:02:59 <_law_> koen, looks better now, but gpephone image requires gconf-dbus (>= 2.16.0+svn20070424) and there is only gconf-dbus 2.16.0+svn20070412 available Apr 26 12:06:15 RP: You might want to update the year for the copyright in the bitbake script Apr 26 12:07:06 Laibsch: Which files or you mean all of them? Apr 26 12:09:33 Next step - blacklist 1.8.0 in OE.dev... Apr 26 12:10:22 what was the problem with 1.8.0? Apr 26 12:10:46 what does this NOTE mean? NOTE: gcc-cross-4.1.2-file://100-uclibc-conf.patch has no section, not checking URI Apr 26 12:14:59 checksum checking? Apr 26 12:16:15 Shoragan: It had some minor gremlins which recent changes in OE.dev have exposed a lot more Apr 26 12:27:13 RP: I meant bitbake/bin/bitbake (I did not look at the others). I am not an expert but as far as I understand you risk your work becoming public domain if you don't claim copyright for every year. Apr 26 12:29:16 Laibsch: that sounds like a American view, in UK at least the (c) date should represent the last date you worked on file, you copyright of course expiring in 50 years from that date Apr 26 12:29:38 http://www.nctexasbirds.com/images/hot_news.jpg Apr 26 12:29:39 Laibsch: I kinda guess RP is not bothered about what happens when he is 70 :-D Apr 26 12:42:21 XorA: the question is - will RP still use his MG then or buy other car? Apr 26 12:44:13 hrw: by then it will be a ball of gaffa/duck/tank tape + engine Apr 26 12:45:00 my bitbake 1.9/r816 does not built ipkg-utils-native or pax-utils-native ;( Apr 26 12:45:28 hrw: OE is up-to-date? Apr 26 12:46:02 87b49518bcc5300a395daeb3510b7801f333eda9 hrw@openembedded.org 2007-04-26T11:01:15 Apr 26 12:46:04 hrw: luckilly wey can buy new MG from chinese :-D Apr 26 12:48:27 hrw: bitbake -vvv does it at least turn up verbatim? Apr 26 12:48:45 hrw: what about the stable branch? Apr 26 12:49:16 checking 1.9 - moment Apr 26 12:51:01 03hrw 07org.oe.dev * rfc756d25... 10/ (4 files in 3 dirs): telepathy-gabble: drop old versions Apr 26 12:54:30 zecke|away: no sign of those tools in -vvv Apr 26 12:57:56 this is bad Apr 26 12:58:05 hrw: try to use 1.8 Apr 26 12:58:49 ok Apr 26 12:59:36 testing 1.8.2 Apr 26 12:59:42 hrw: Look at the output from bitbake -g to tell for certain whats going on for 1.8/1.9 Apr 26 13:00:35 o.. right Apr 26 13:00:47 Is bitbake using the sane srcdates file again? Apr 26 13:00:48 Laibsch: It should be updated next time we touch the file, thats currently policy anyway Apr 26 13:00:54 (afaik) Apr 26 13:02:28 depends.dot:"pax-utils-native" -> "shasum-native" Apr 26 13:02:28 task-depends.dot:"linux-rp.do_package" -> "pax-utils-native.do_populate_staging" Apr 26 13:02:31 1.8.2 Apr 26 13:03:10 switched to 1.9svn nw.. Apr 26 13:05:38 hrw: That looks right Apr 26 13:06:06 no sign of pax in *dot with 1.9 Apr 26 13:06:26 hrw: ok, sounds like 1.9 has a bug then :/ Apr 26 13:06:31 right Apr 26 13:06:49 1.8.2 is more important atm Apr 26 13:07:24 RP: 1.8.2 looks like its working sweet now Apr 26 13:07:41 RP: never got this far without manually building stuff with 1.8.0 Apr 26 13:08:56 I thought people knew 1.8.0 has issues :/ Apr 26 13:09:29 This is the problem when the developer runs svn HEAD :/ Apr 26 13:09:43 RP: not until you said this morning Apr 26 13:10:08 RP: I guess its info that slipped into cracks Apr 26 13:10:37 XorA: I guess so. I'm sure problems were discussed on list, the implications weren't clear though :/ Apr 26 13:11:04 XorA: The problems in 1.8.0 were why some of the RFCs appeared and I made the base.bbclass changes Apr 26 13:13:22 RP: was this the inter task dependencies? Apr 26 13:13:40 XorA: yes Apr 26 13:13:54 RP: ah right, I just didnt connect the two in my head Apr 26 13:14:20 'task-depends.dot:"linux-rp.do_package" -> "pax-utils-native.do_populate_staging"' is inter-task Apr 26 13:14:28 * XorA demotes himself to luser Apr 26 13:14:53 http://blog.haerwu.biz/2007/04/26/how-to-pronounce-haerwu/ Apr 26 13:16:07 hrw: That doesn't help much, I think I'll have to stick to "Marcin" ;-) Apr 26 13:17:06 RP: I know - but this was too nice comment to miss it ;D Apr 26 13:17:40 why will it help English speakers when the phonetics are in wierd ass characters :-) Apr 26 13:18:29 XorA: ipa is used everywhere Apr 26 13:18:48 hrw: Ive never seen it before :-) Apr 26 13:19:20 me guess on pronouncing it would be haar wooo Apr 26 13:22:32 ;D Apr 26 13:23:41 hrw: pretty neat -- first time I've run into IPA. Guess I can say I learned something today :-) Apr 26 13:26:40 ;D Apr 26 13:26:54 interesting bitbake behavior: if my 2.6.16 kernel build fails in a compulab build, bitbake loops forever trying to build u-boot, which is not a dependency in the build. Apr 26 13:33:14 ok. angstrom-oabi fails for me Apr 26 13:33:29 ~summon koen Apr 26 13:33:31 apt takes out 20 clean, identical-looking phones, some extra hands, and pretends to be a telemarketer for a large corporation, so he gets delivered a phonelist containing koen's coordinates Apr 26 13:34:44 likewise: actually my insane.bbclass should catch these rpaths Apr 26 13:35:01 likewise: they are not only ugly but might even be a security issue :) Apr 26 13:35:14 zecke|away: security issue? in what sense? Apr 26 13:35:25 likewise: Imagine you would build in /tmp Apr 26 13:35:48 likewise: and you could create the directory and place an alternative libc here Apr 26 13:35:55 zecke|away: understood Apr 26 13:36:04 likewise: for a mobile system the threat is minor but still :) Apr 26 13:36:50 zecke|away: I just finished a fresh angstrom build for machine guinness, let me check if insane was used and if the libraries hold host rpaths... Apr 26 13:37:07 likewise: my code could be incorrect :) Apr 26 13:37:41 zecke|away: insane should check for strange paths, but which class should mangle them in the first place? Apr 26 13:38:10 likewise: the question is who inserts them Apr 26 13:38:22 likewise: when building only -rpath,/usr is sane! Apr 26 13:39:49 zecke|away: insane does not modify RPATH's, but only checks them, right? Apr 26 13:40:54 right Apr 26 13:42:28 zecke|away: libfontconfig was wrong here, but I had disabled inheritance of insane (I don't know why anymore..). Now re-enabling and rebuilding fontcofig Apr 26 13:43:59 hi Thomas Apr 26 13:44:18 hi everybody Apr 26 13:46:43 hrw: I thought about busybox on collie. One had to port the SD-Driver to busybox to use busybox to boot from SD right? (But on the other side one could just give the right commandline to the kernel.) Apr 26 13:47:11 hrw: In every case one would have to port some drivers to busybox. Apr 26 13:48:48 u-boot you mean? Apr 26 13:50:23 RP: NOTE: package angstrom-console-image-1.0: completed Apr 26 13:50:31 hrw: NOTE: package angstrom-console-image-1.0: completed for collie with OABI Apr 26 13:50:33 XorA: :) Apr 26 13:50:47 lucky bastard Apr 26 13:51:03 hrw: I guess its a distro problem youhave? Apr 26 13:52:27 | unwind-pe.c:(.text+0x1283c): undefined reference to `__subdf3' Apr 26 13:52:38 who knows is it distro problem Apr 26 13:54:22 hrw: yes of course ;) Apr 26 13:55:30 are there problems with angstorm images on collie? Apr 26 13:58:44 thesing: I have problem to build one Apr 26 14:00:58 hrw: works fine on debian testing amd64 for me. Apr 26 14:01:30 debian sid amd64 here Apr 26 14:01:50 I will try later on maluch (debian sid x86) Apr 26 14:03:02 with etch released, sid is more unstable than usual Apr 26 14:04:07 hrw: debian etch here Apr 26 14:05:24 thesing: with over 7 years of experience with debian I am ok with sid Apr 26 14:05:55 hrw: I would guess gcc-4.X.X-cvs is the fault then :-) Apr 26 14:06:16 thesing: home goes collie kernel? Apr 26 14:06:47 s/home/how/ Apr 26 14:08:13 suspend works fine now. The integration in sharpsl_pm is not complete yet. The charging needs some love too. Apr 26 14:09:21 thesing: wow, you are flying Apr 26 14:09:29 But its nearly usable now. I already build an gpe-image but somehow I can't boot from SD. I haven't figured out what broke this. Apr 26 14:10:30 I get messages like: failed to sync: attempting to kill init. Apr 26 14:10:45 thesing: add rootfs_delay=5 Apr 26 14:10:48 to CMDLINE Apr 26 14:11:10 5 seconds should be more then enough for kernel to recognize card Apr 26 14:12:09 I alread have rootdelay=20 in CMDLINE. Apr 26 14:12:35 I had this working a few weeks ago.. Apr 26 14:15:19 collie + browser + rtm8000 would rock Apr 26 14:15:33 Ill have to see if I can work out why rtm8000 doesnt work on 2.6 Apr 26 14:16:15 whats a rtm8000? Apr 26 14:17:27 ah the audiovox gprs card. Apr 26 14:18:07 * thesing has to buy some supplies for a canoe trip this weekend. Apr 26 14:21:41 recreating all packages takes time... Apr 26 14:22:47 XorA: and both my machines have same gcc version ;D Apr 26 14:24:59 gcc (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) Apr 26 14:25:03 yes Apr 26 14:25:13 gcc (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) Apr 26 14:25:18 amd64 Apr 26 14:25:31 gcc (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) Apr 26 14:25:33 x86 Apr 26 14:26:16 bbiab Apr 26 14:36:28 zecke|away: #2140 Apr 26 14:36:46 zecke|away: NOTE: Checking Package: fontconfig Apr 26 14:36:48 __builtin__:27: DeprecationWarning: raising a string exception is deprecated Apr 26 14:45:12 the rebuild from scratch seems to work Apr 26 14:49:35 koen: not here - but I will retry wit 1.8.2 bitbake Apr 26 14:50:06 * hrw|afk -> food Apr 26 14:51:39 * likewise <- food? anyone? Apr 26 14:53:48 likewise: I'm an idiot Apr 26 14:54:01 koen: ??? Apr 26 14:54:09 if I had a nickel for every time I was an idiot Apr 26 14:54:11 likewise: I was in eindhoven two days ago (being lost), and forgot to pickup the server again Apr 26 14:54:40 koen: ah, well, the server is at gerwinin and I haven't spoken with him for a long time. Apr 26 14:54:49 Apparently I have a SFF SDR board waiting for me at school Apr 26 15:00:27 koen: we need to finally get rid of few files from conf/distro/ Apr 26 15:00:40 yes Apr 26 15:00:58 familiar* openzaurus* for sure Apr 26 15:02:13 openomap can go Apr 26 15:03:27 likewise: could you check my 'path' in #2140? Apr 26 15:03:31 patch* Apr 26 15:04:19 likewise: zecke will say that patch is wrong (and it is), but it should error out more often Apr 26 15:05:32 koen: I was just looking into it, I see insane being invoked but it does not write a log as it should. Apr 26 15:06:13 koen: your patch merely removes the work/ suffix? Apr 26 15:06:35 yes Apr 26 15:07:03 koen: I do not expect anything from it then, but I will apply. Apr 26 15:07:03 since your rpath points to staging (tmp/staging) instead of wordir (tmp/work) Apr 26 15:07:17 koen: ah! Apr 26 15:07:29 koen: you are not an idiot after all ;-) Apr 26 15:07:35 insane should now errors out on anything containing tmp/ Apr 26 15:07:40 error* Apr 26 15:08:00 see http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-April/002006.html Apr 26 15:08:33 bah, thunderbird mangled my patch Apr 26 15:09:44 koen: got it applied Apr 26 15:10:40 koen: am I the only one having full paths? Apr 26 15:11:03 probably not Apr 26 15:21:26 koen: what is needed to get MediaTomb into the feeds? the .bb file in OE was updated to 0.9 version long time ago, but Angstrom and nslu2 linux still have the prerelease packages Apr 26 15:22:02 basically ask someone to build it :) Apr 26 15:22:08 * koen builds mediatomb Apr 26 15:22:11 http://video.google.com/videoplay?docid=-5503582578132361295 Apr 26 15:22:23 * koen reads http://blogs.gnome.org/view/xclaesse/2007/04/26/0 Apr 26 15:26:12 koen: thanks :) Apr 26 15:27:32 * XorA really needs to get a pda with 800x600 and a decent amount of RAM Apr 26 15:30:17 XorA: netbook pro? Apr 26 15:30:29 koen: they are serieous cash :-( Apr 26 15:30:43 and have sucky graphics Apr 26 15:31:22 and all this talk of UMPC and noone seems to have actually made one yet Apr 26 15:31:26 XorA: umpc? Apr 26 15:31:32 Ultra Mobile Apr 26 15:51:17 In trying to bitbake bootstrap-image it's failing on step 976 of 978 task do_rootfs complaining that ipkg-cl 'command not found'. Apr 26 15:51:29 What door do I look behind to try and figure this out? Apr 26 15:52:46 vervain: update OE and use bitbake 1.8.2 Apr 26 15:57:51 hrw: how sure are you the files on the CELF mirror are correct? Apr 26 15:57:57 * koen just ran into a checksum problem Apr 26 15:59:45 koen: not in 100% Apr 26 16:00:07 tell which file and I will check it locally Apr 26 16:00:12 hrw: could you have a look at http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-April/002007.html ? Apr 26 16:00:17 NOTE: package glibc-intermediate-2.5-r5: task do_compile: failed Apr 26 16:00:21 buu Apr 26 16:00:25 mickeyl: good morning! Apr 26 16:00:46 oh, wait Apr 26 16:00:52 hiho Apr 26 16:00:56 Dr. mickeyl, good morning! Apr 26 16:01:01 ;) Apr 26 16:01:03 * mickeyl bows Apr 26 16:01:04 ok Apr 26 16:01:07 I have a toy Apr 26 16:01:08 863bb6462f4ce535399a7c6276ae6776 js-1.5.tar.gz Apr 26 16:01:08 ef72a154320c9949f8de0732e8dbfa186b83feb0a5e890d90807757de9998585 js-1.5.tar.gz Apr 26 16:01:24 mickeyl: what's up doc? Apr 26 16:01:32 http://www.lyrtech.com/DSP-development/dsp_fpga/sffsdrdevelopmentplatform.php Apr 26 16:01:44 except what I received doesn't look like the picture :) Apr 26 16:01:46 hrw: that's what I have, but checksums.ini disagrees :( Apr 26 16:02:07 hrw: still suspecting inguinal hernia, but otherwise good Apr 26 16:02:10 * hrw is writing long post which will probably get lot of 'popularity' Apr 26 16:02:26 mickeyl: ouch! Apr 26 16:04:47 *nod* Apr 26 16:05:01 hi all Apr 26 16:05:09 had that 15 years ago and since then I was always careful not to carry too heavy stuff Apr 26 16:05:26 unfortunately carried some heavy boxes lately :/ Apr 26 16:05:27 mickeyl: did you get that hernia as a "souveneer" from your move? Apr 26 16:05:29 and since then it hurts Apr 26 16:05:36 I'm afraid so, ya Apr 26 16:05:45 one doctor said he can't see any gaps Apr 26 16:05:52 but i don't believe him, 'cause i know how it feels Apr 26 16:05:57 * hvontres|poodle hasn't moved in almost 10 years :) Apr 26 16:05:58 going to visit a specialist next week Apr 26 16:06:13 "I don't see anything" means "pay me to do more expensive tests" Apr 26 16:07:07 mrdata: hi Apr 26 16:07:17 thesing: hi Apr 26 16:07:20 03koen 07org.oe.dev * r3feda625... 10/ (1 classes/gconf.bbclass): gconf.bbclass: depend on gconf-native Apr 26 16:07:24 03koen 07org.oe.dev * r72018d16... 10/ (1 packages/matchbox-wm/matchbox-wm_1.2.bb): matchbox-wm: depend on gconf-native Apr 26 16:07:43 Crofton: Why? did your new toy arrive as a "kit" ? Apr 26 16:08:09 well, it arrived with a CDROM labeled "Integrity" Apr 26 16:08:36 and they redesigned it since they took the original marketing pictures :) Apr 26 16:09:04 Crofton: so, out of curiosity, how bad is "For details about pricing and the platform, contact Lyrtech Apr 26 16:09:17 last I heard "bad" Apr 26 16:09:25 I need to get some updated information Apr 26 16:09:41 Crofton: so you didn't need to fill out the Requisition form then? Apr 26 16:09:45 hopefully, knocking the board count down helped Apr 26 16:09:46 no Apr 26 16:09:53 www.radiochallenge.org Apr 26 16:11:00 Crofton: wow, TWO teams from VT... That's pretty cool.:) Apr 26 16:11:16 yeah Apr 26 16:11:29 there was some fussing that we are both working on the same problem Apr 26 16:11:44 but those guys do not actually speak to us :) Apr 26 16:12:17 Crofton: Just make sure they don't send in a "mole" :) Apr 26 16:12:23 yeah Apr 26 16:13:01 Crofton: good luck :) And may the Schwarz be with you :) Apr 26 16:13:11 http://www.lyrtech.com/DSP-development/dsp_fpga/sffsdrevaluationmodule.php Apr 26 16:13:16 hmm, this is what they sent Apr 26 16:14:29 Crofton: yet still it falls into the "if you need to ask, you can't afford it" price range :) Apr 26 16:14:45 yep Apr 26 16:14:57 the marketing dude is out of town Apr 26 16:15:02 I'll ask when he is back Apr 26 16:15:41 let me look up the FPGA Apr 26 16:18:34 OF python Q: If I have a list of strings, how can I check if a string B occurs as a substring in that list? Apr 26 16:18:45 s/OF/OT Apr 26 16:18:51 hvontres|poodle, the FPGA is about a $500 part Apr 26 16:19:13 that probably sets the "gross" pricing level Apr 26 16:19:34 hvontres|poodle: was that a spaceballs reference? Apr 26 16:19:52 koen|away: ask hrw for the source of js-1.5.tar.gz and diff Apr 26 16:20:17 zecke: hrw got the same checksum as I have Apr 26 16:20:20 NOTE: The MD5Sums did not match. Wanted: 'ea291c89e68832d570d9d5e007218bd6' and Got: '2d59fd4e5f2756fed3717b93ec0e5df6' Apr 26 16:20:20 NOTE: Task failed: Checksum of 'http://xorg.freedesktop.org/releases/individual/xserver/xorg-server-1.2.0.tar.bz2' failed Apr 26 16:20:23 and another one Apr 26 16:21:05 weird :) Apr 26 16:21:20 koen: what if the file is only partly downloaded? Apr 26 16:21:38 is ea291c89e68832d570d9d5e007218bd6 at least duplicated in the file? Apr 26 16:22:09 ~quote armhy Apr 26 16:22:12 ARMHY: 8.33 (8.39/8.16), 4/26/2007 12:06pm, Opened 8.16, Volume 955,749, Change +0.41 Apr 26 16:22:23 * Crofton hearts his portfolio Apr 26 16:25:45 thesing: did your patch for sa1100-generic.c work for collie? Apr 26 16:26:43 mrdata: yes it now suspends and resumes fine. Apr 26 16:28:09 thesing: okay, then i must use the wrong way and will try it again Apr 26 16:30:25 mrdata: maybe there is an additional problem on simpad, collie uses sharp scoop as pcmcia chip. Apr 26 16:30:43 koen: Of course :) Apr 26 16:30:54 * hvontres|poodle loves mel brooks movies :) Apr 26 16:32:37 likewise: for line in list: Apr 26 16:32:40 so far only 3 problems with a rebuild from scratch Apr 26 16:33:12 hvontres|poodle: nope, that's the buggy solution, it does not check for *sub*strings Apr 26 16:33:35 hvontres|poodle: I think it must be if any((substring in a for a in listofstrings)): Apr 26 16:34:06 hvontres|poodle: thanks for helping though, I'm noob on python. Apr 26 16:34:39 likewise: that was only the first line. try this: http://pastebin.ca/459646 Apr 26 16:35:43 * hvontres|poodle is trying to become a pythonista...:) Apr 26 16:36:44 koen: have you seen any strange looking warnings when building 2.6.20 kernels ? Apr 26 16:37:08 hvontres|poodle: thanks, trying your approach (mine hid the 'line' variable in the test scope) Apr 26 16:37:46 koen: I keep getting one about a dir not being there and the name of that dir had the output of uname -r on my build machine in it Apr 26 16:37:58 cu Apr 26 16:38:34 ~lart lyrtech for using a different serial cable than the OSK Apr 26 16:38:35 * ibot smacks lyrtech up side the head with a clue-by-4 for using a different serial cable than the OSK Apr 26 16:38:48 insane.bbclass is about to break havoc on all your builds - MUUUHAAAAAHAAAA (with kind thanks to zecke) Apr 26 16:39:14 likewise: oh oh, what did I help you do? Apr 26 16:40:14 bitch about bad rpath's? Apr 26 16:40:32 hvontres|poodle: haven't seen that yet Apr 26 16:41:10 koen: ahh, great. Apr 26 16:41:28 Good evening Apr 26 16:41:38 hey sirfred Apr 26 16:41:43 ~lart kernel guys for adding obscure CONFIG_UNAME_RELEASE var without documentation Apr 26 16:41:43 * ibot flings poo at kernel guys for adding obscure CONFIG_UNAME_RELEASE var without documentation Apr 26 16:42:02 Hello koen Apr 26 16:42:31 hey HopsNBarley Apr 26 16:42:43 hi koen! Apr 26 16:42:48 HopsNBarley: hello Apr 26 16:43:07 hi likewise! Apr 26 16:43:51 koen: you might want to remove insane.bbclass as it will now fail on our shared libs... Apr 26 16:44:02 koen: 'remove', as in 'not use'. Apr 26 16:44:04 03likewise 07org.oe.dev * rf3f908d8... 10/ (1 packages/uclibc/uclibc.inc): uclibc.inc: also mangle KERNEL_HEADERS not only KERNEL_SOURCE. Apr 26 16:44:11 03likewise 07org.oe.dev * r038bc21b... 10/ (1 classes/insane.bbclass): insane.bbclass: Fix check for bad RPATHs. This is gonna break builds. Needs fixes in the shared library handling. Apr 26 16:44:52 likewise: let's fix the bugs instead :) Apr 26 16:45:07 likewise: the fix is only a single chrpath command away, right? Apr 26 16:45:28 koen: yes, but we do not build the chrpath package yet, it needs a .bb Apr 26 16:45:55 koen: And I did not yet find the proper place where to invoke it. Apr 26 16:46:12 thesing: patch from you works like a charme, i have made a extra mistake while changing sa1100-generic.c Apr 26 16:46:18 koen: and better still, we might not need chrpath, but should be able to specify the RPATH rather... Apr 26 16:46:58 thesing: simpad do suspend and also resume, the wrong pointer is lost, it is now NULL and thats okay ;-)) Apr 26 16:46:58 likewise: lets leave it enabled for a while to see which packages break and then evaluate it again Apr 26 16:47:07 mr-data: great. One bug less in the kernel. Apr 26 16:47:26 likewise: could you send a mail to the list about that? Apr 26 16:47:27 thesing: yes, next i must get the touchscreen working Apr 26 16:48:16 thesing: and also the usage of syslogd Apr 26 16:48:32 likewise: and your change still only checks for WORKDIR, not for TMPDIR and/or staging Apr 26 16:49:06 koen: ah yes, I reverted before doing the final fix, just to make sure all my debugging stuff was removed. I forgot to add your approach... Will do shortly Apr 26 16:50:45 thesing: and i forgot, many thanks for your quick-response and right solution for that problem Apr 26 16:57:26 03likewise 07org.oe.dev * r0a4ab83c... 10/ (1 classes/insane.bbclass): insane.bbclass: Broaden check for RPATHs by removing work/ path component. Apr 26 16:59:04 mrdata: no problem. Apr 26 17:12:33 hrw|gone: could you summarize the thread at http://www.mail-archive.com/pld-devel-pl@lists.pld-linux.org/msg14773.html ? Apr 26 17:20:48 koen: > >-#if defined(MNT_FORCE) && !defined(__sparc__) && !defined(__arm__) Apr 26 17:20:49 > >+#if defined(MNT_FORCE) Apr 26 17:21:26 'this looks like old workaround for old headers on sparc and arm. in glibc 2.3.6 have all what is needed' Apr 26 17:21:35 thats all Apr 26 17:22:06 koen: any chance you could add kexec-tools to the angstrom feed? Apr 26 17:22:26 aren't they there already? Apr 26 17:22:47 koen: if they are, the feedbrowser is missing them Apr 26 17:23:24 Thanks for the pointer to the new bitbake release - I now have a OE bootstrap image. Apr 26 17:24:36 hvontres|poodle: I'll add them when my build from scratch completes Apr 26 17:25:08 koen: np. Also, do you think it would be a good idea to include them with the Console image? Apr 26 17:25:16 not really Apr 26 17:25:38 only a small portion of the machines supports kexec / has it enabled Apr 26 17:26:21 koen: ok. Apr 26 17:26:53 hvontres|poodle: feel free to add an angstrom-rescue-image :) Apr 26 17:27:32 koen: hehe...you are reading my mind again, aren't you :) Apr 26 17:27:45 yes Apr 26 17:28:52 koen: As soon as I can figure out a clean way to deal with that @#$%^ing built in CMDLINE... Apr 26 17:28:52 bugs.oe.org seems to be down again.... Apr 26 17:29:04 nm, nope, back now Apr 26 17:29:09 going in and out today... Apr 26 17:29:31 * hvontres|poodle does not grok kexec at all Apr 26 17:29:38 I just managed a pull Apr 26 17:39:48 I am running a build for the OSK, should I expect problems from the RPATH change? Apr 26 17:40:45 koen: ping Apr 26 17:40:59 shreyas: pong Apr 26 17:43:33 koen: i created a initscripts directory containing scripts i needed to be copied over to the root_fs. I added if a dist=amida_dist to the initscripts_1.0.bb, now when i run bitbake olai-image it just hung, i digged around a bit and looked like ipkg-cl hung here in gdb traces Apr 26 17:43:58 #2 0xa7fb19fa in default_ipkg_response_callback ( Apr 26 17:43:58 question=0x8069c58 " Configuration file '/etc/inittab'\n ==> File on system created by you or by a script.\n ==> File also in package provided by package maintainer.\n What would you like to do about it ? Yo"...) at libipkg.c:439 Apr 26 17:44:36 since there seems to a busy box inittab somewhere in the build tree too, whats the best way to work around such a thing? Apr 26 17:48:12 Anobody know why bitbake gpe-image is triggering totem videoplayer build? Apr 26 17:48:23 s/Anobody/Anybody Apr 26 17:48:54 I've also seen something about gdb that annoyed me, but, since it didn't failed... Apr 26 17:50:53 sirfred: looks like ./task-gpe.bb:123: totem" Apr 26 17:51:17 so the task-gpe which is used by gpe-image builds totem Apr 26 17:51:51 shreyas: But it's under RDEPENDS_gpe-task-desktopapps Apr 26 17:52:45 sirfred: but I think bb 1.8 turns RDEPENDS into DEPENDS at build time Apr 26 17:53:12 hvontres|poodle: But the rdepends are for gpe-task-desktopapps. And I think it's not used for gpe-image. Apr 26 17:53:38 At least, that's what I found looking at gpe-image.bb Apr 26 17:53:56 No traces of gpe-task-desktopapps Apr 26 17:57:51 sirfred: hmmm, good point. Unless task-gpe wants to build everything in PACKAGES..... Apr 26 17:58:35 hvontres|poodle: Sure? Apr 26 17:59:06 That should be changed lately. I've build openzaurus gpe images and never got those packages included. Apr 26 18:00:45 sirfred: was that out of the .dev tree or .oz354x? Apr 26 18:00:52 hvontres|poodle: .dev Apr 26 18:01:07 sirfred: which version of bitbake? Apr 26 18:01:23 hvontres|poodle: 1.8.2 Apr 26 18:01:31 hvontres|poodle: But I also suffered the problem with 1.8.0 Apr 26 18:01:39 sirfred: strange indeed Apr 26 18:07:57 * JustinP growls Apr 26 18:22:07 JustinP, I am starting builds for angstrom for the efika and the OSK Apr 26 18:22:25 so I can see if the RPATH stuff effects those Apr 26 18:22:36 Crofton__: ummm....ok, then. (I know nothing about efika or OSK) Apr 26 18:22:43 machines I have Apr 26 18:22:44 :) Apr 26 18:22:45 ah, I see Apr 26 18:22:56 good luck, then Apr 26 18:23:13 I'm not updating ATM....I don't want OE to just fubar on me Apr 26 18:23:37 hopefully it isn't too hard to fix if it occurs Apr 26 18:23:47 * JustinP wouldn't know Apr 26 18:23:47 what do you build for? Apr 26 18:23:50 spitz Apr 26 18:24:00 ~spitz Apr 26 18:24:00 and other Zaurii if people need me to Apr 26 18:24:11 (C3000) Apr 26 18:24:12 rumour has it, spitz is the Sharp SL-C3000, or a dog Apr 26 18:24:26 still no CE-RH1 owners around, eh? Apr 26 18:24:30 I am not real familiar with the zaurii stuff Apr 26 18:24:33 ~osk Apr 26 18:24:39 well, osk is the OMAP5912 Starter Kit, a $295 development board with 32/32, serial, ethernet, usb host/client, CF, jtag, and expansion connectors available from http://www.spectrumdigital.com/cgi/catalog.cgi?show_product=701875 or http://tree.celinuxforum.org/pubwiki/moin.cgi/OSK, or http://www.ti.com/omap5912osk Apr 26 18:24:56 well, the zaurus is a dead line now :-(( Apr 26 18:26:41 shreyas: don't use your own distro, use DISTRO=angstrom-2007.1 Apr 26 18:28:13 Bitbake says that "task-gpe.do_package_write" -> "totem.do_package" Apr 26 18:28:51 whee, finally got an image to build again Apr 26 18:29:08 You, lucky man Apr 26 18:30:27 In the depends.dot I can see that gpe-task-desktopapps depends on totem. Apr 26 18:31:37 the first bad rpath is in ... Apr 26 18:31:40 *drumroll* Apr 26 18:31:45 :) Apr 26 18:31:45 libgcc Apr 26 18:33:50 Humm, perhaps this is the line producing the funny totem effect: Apr 26 18:33:53 DEPENDS = "${MACHINE_TASK_PROVIDER} task-gpe" Apr 26 18:33:57 in gpe-image.bb Apr 26 18:34:10 sirfred: are you building for angstrom? Apr 26 18:34:16 JustinP: Yes. Apr 26 18:34:51 sirfred: then why not build angstrom-x11-image? Apr 26 18:34:55 If I understand correctly, task-gpe is the whole .bb with all the gpe-task-* Apr 26 18:35:06 JustinP: Is angstrom-x11-image providing a gpe environment? Apr 26 18:35:34 AFAIK the "gpe" image is pre-angstrom. angstrom-x11-image seems to be the angstrom-blessed X/gpe image Apr 26 18:35:52 JustinP: Well, I expected it was a bare x image, looking at the name. Apr 26 18:35:53 X11, not X/gpe Apr 26 18:36:07 s/x/X of course Apr 26 18:36:39 it aims to give you a X11 env where you can alter settings and install more stuff using the gui Apr 26 18:36:54 koen: What gui? Apr 26 18:37:11 currently matchbox-panel, matchbox-desktop and matchbox-wm Apr 26 18:37:21 Oh, so they're included. Apr 26 18:37:22 coupled with gpe-dm, gpe-login and gpe-conf Apr 26 18:37:53 So, that's what I'm looking for. matchbox-*-2 or 1.x versions? Apr 26 18:38:02 1.x Apr 26 18:38:08 2.x is not ready for users Apr 26 18:38:14 (even if openmoko uses it) Apr 26 18:38:44 koen: I was trying a poky snapshot, and I really liked what I saw. Of course it's not stable enough. Apr 26 18:38:56 I like it as well Apr 26 18:39:04 but right now it's a showcase Apr 26 18:39:22 At least something with the right size for the 640x480 zaurii Apr 26 18:39:34 shouldn't all the images be put in packages/images? This way one would find them more easy. Apr 26 18:41:09 thesing: in theory, yes Apr 26 18:42:54 I didn't know that angstom uses different images until now... Apr 26 18:44:32 So someone should put all tasks under packages/tasks and all images under packages/images. One could have subdirectories for angstrom, openmoko etc there. Apr 26 18:45:27 thesing: subdirs wouldn't work due to the default bb conf option... Apr 26 18:47:25 I suppose thats why every distro gets its own directory in packages for now. Apr 26 18:47:35 like|shop: the next commit is for you Apr 26 18:49:59 I see that angstrom-x11-image is trying to build eds-dbus. Is this for the o-hand PIM apps ? Apr 26 18:50:04 Are they included? Apr 26 18:54:24 03koen 07org.oe.dev * re0e204be... 10/ (1 packages/chrpath packages/chrpath/chrpath-native_0.13.bb): chrpath: add native version so we can munge bogus rpath entries Apr 26 18:57:01 I see RPATH failure in libgcc and libstdc++ Apr 26 18:58:13 03koen 07org.oe.dev * r4363d4f2... 10/ (1 classes/insane.bbclass): insane.bbclass: depend on chrpath-native Apr 26 18:58:34 koen, is this going to fix libgcc? Apr 26 18:58:50 Crofton: no, but it's laying the foundations for it Apr 26 18:58:57 thanks Apr 26 18:59:13 I are officialy broke :) Apr 26 19:02:38 koen, ping me when I have a chance getting stuff to build again and I will restart builds Apr 26 19:05:22 03crofton 07org.oe.dev * r7859b104... 10/ (1 conf/distro/openomap.conf): openomap.conf : Angstrom runs fine on OSK so no further need for custom distro. Apr 26 19:21:59 koen: re? Apr 26 19:22:13 yes Apr 26 19:22:19 I got a present for you Apr 26 19:23:07 koen: tnx! was already d'loading and playing with chrpath when I got back from |shop, |food and |tv :-) Apr 26 19:23:08 likewise: http://ewi546.ewi.utwente.nl/tmp/chrpath.patch Apr 26 19:23:32 koen: ok tnx mtn: adding packages/chrpath/chrpath-native_0.13.bb Apr 26 19:26:05 Now, bitbake angstrom-x11-image is trying to fetch abiword. Is this expected? Apr 26 19:26:06 http://blog.haerwu.biz/2007/04/26/openzaurus-time-is-over-long-life-angstrom/ Apr 26 19:26:29 koen: maybe we should even delete the rpath unless we exactly know the location on target, because /lib might be /usr/lib as well. Apr 26 19:26:37 but I'll run a full build with this in place Apr 26 19:26:47 hrw: s/life/live/ Apr 26 19:27:06 ops Apr 26 19:27:16 likewise: I'm an rpath and python n00b, as you might have guessed from that patch Apr 26 19:27:36 url wrong, title ok ;D Apr 26 19:28:14 fixed both Apr 26 19:28:20 hrw: could you make the default text a bit darker, it's a bit hard to read on my laptop Apr 26 19:28:35 (I'm tired, so that could be it as well) Apr 26 19:28:53 koen: my site? Apr 26 19:29:12 can do Apr 26 19:29:12 yes Apr 26 19:33:03 now better? Apr 26 19:33:41 yes! Apr 26 19:33:42 thanks Apr 26 19:34:11 koen: that makes two of us. Apr 26 19:34:30 I just got a strange warning among the many bitbake messages I am used to Apr 26 19:34:34 :4: RuntimeWarning: tmpnam is a potential security risk to your program Apr 26 19:35:03 This was apparently while packaging linux-rp-2.6.20 Apr 26 19:35:14 Should I be afraid ;-) Apr 26 19:35:17 ? Apr 26 19:35:41 Laibsch: only if you have a multi-user computer Apr 26 19:37:02 Linux is always multi-user Apr 26 19:37:03 Laibsch: by any chance, did you get a WARNING from bitbake about a missing dir right before that? Apr 26 19:37:14 hvontres|poodle: no Apr 26 19:37:28 Laibsch: as in: "multiple people accessing your computer" Apr 26 19:38:28 koen: Yes, I understood that correctly. Was interrupted and could not reply fast enough Apr 26 19:38:38 Well, this computer acts as a webserver Apr 26 19:38:49 But I am the only one with real access to it. Apr 26 19:38:58 Laibsch: iirc tmpnam create files in /tmp that are a security risk Apr 26 19:39:07 OK Apr 26 19:39:25 It was just that the format of the message was so out of the ordinary Apr 26 19:39:26 I forget if it's because of permissions or because the names aren't really random Apr 26 19:41:46 likewise: could you reply to my message with your suggestions? Apr 26 19:42:02 koen: on the list? Apr 26 19:45:37 koen: It's the non-randomness of the file names. The function just increments a counter...so you could guess the next one Apr 26 19:45:51 likewise: yes Apr 26 19:45:54 chouimat: morning :) Apr 26 19:45:56 koen: will do Apr 26 19:46:04 hvontres|poodle: ah, right, thanks Apr 26 19:46:17 hvontres|poodle: bah Apr 26 19:47:28 koen: you're welcome :) Apr 26 19:48:02 likewise: beware: my patch mixes spaces and tabs :( Apr 26 19:48:31 koen: will you re-commit? Apr 26 19:51:45 NOTE: Fixing RPATH for you in /data/build/koen/OE/build/tmp/angstrom/work/armv5te-angstrom-linux-gnueabi/gcc-cross-4.1.2-r2/install/libstdc++/usr/lib/libstdc++.so.6.0.8 :) Apr 26 19:53:00 cu Apr 26 19:53:24 koen: cool. I forgot to remove --dry-run on your patch, so my build was... dry-running as well. sigh. need coffee. Apr 26 19:54:40 koen@bitbake:/data/build/koen/OE/build/tmp/angstrom/deploy/glibc/ipk/armv5te$ chrpath -l lib/libgcc_s.so.1 Apr 26 19:54:41 lib/libgcc_s.so.1: no rpath or runpath tag found. Apr 26 19:55:11 so it seems it deletes rpath instead of putting in /ib Apr 26 19:55:14 hvontres|poodle: http://www.nctexasbirds.com/images/hot_news.jpg Apr 26 19:56:30 koen: hmm, I will do some manual runs. I don't think it is bad for rpath to be empty (as it was wrong in the first place anyway) Apr 26 19:56:51 koen: the optimum though, if it's present *and* correct :-) Apr 26 19:58:15 indeed Apr 26 19:59:35 likewise: I'll be busy with getting our bar ready for http://batavierenrace.nl/ this weekend, so don't be affraid to commit that patch is no one has serious problems with it Apr 26 20:00:19 koen: it will probably be considered a hack by some, but from the target's perspective, we did an excellent job :-) Apr 26 20:00:38 koen: and yes, I'll be eager to learn where in OE this *should* go. Apr 26 20:01:08 likewise: the rpath should be checked again after the mungin IMO Apr 26 20:01:21 but as I said, I'm an rpath n00b Apr 26 20:01:32 koen: good point. Apr 26 20:01:43 koen: and where is that bar gonna be *exactly* ? Apr 26 20:01:46 :-) Apr 26 20:02:12 "here" Apr 26 20:02:26 at the usual place on the university Apr 26 20:02:40 but we'll also have a mobile one at the restart Apr 26 20:03:16 some people take that etappe just to be able to drink beer immediately after going to the checkpoint :) Apr 26 20:07:13 you have to run, you can't ride a bike? Apr 26 20:08:00 * likewise opens a beer Apr 26 20:10:08 likewise: good idea Apr 26 20:10:29 chouimat: the implementation is ever better :-) Apr 26 20:10:39 Crofton: there's a runner and a biker who has all the red bull and cookies Apr 26 20:10:52 likewise: will in about 45 minutes ... Apr 26 20:11:30 likewise: when I will go to the pub ... and since I only had 2 hours of sleep last night tonight will be cheap ;) Apr 26 20:12:08 * Crofton needs to figure out how to unbrick the board I have, before I brick it .... Apr 26 20:12:58 koen: chrpath must probably be a 'cross' tool, as it complains about the lib not being x86_64. However, my first chrpath-cross .bb (only added 'inherit cross') executable fails the same. Apr 26 20:13:31 Crofton: I can tell you have experience Apr 26 20:13:32 * koen feels a 'qemu-arm chrpath' coming Apr 26 20:13:42 argh, no Apr 26 20:13:46 hope not Apr 26 20:14:06 likewise: I don't hope so either Apr 26 20:14:18 likewise: but chrpath -l works Apr 26 20:15:44 likewise, davinci board with Integrity installed Apr 26 20:15:59 Need to figure out if I can get uboot going on it Apr 26 20:17:23 koen: on what host/target combo did that work? Apr 26 20:23:04 a780/x86-32 Apr 26 20:24:19 chouimat: well, just look at where this gem came from :) Apr 26 20:25:14 hvontres|poodle: hehe Apr 26 20:31:25 http://www.slate.com/id/2164828/nav/tap1/ Apr 26 20:31:38 the second half is interesting Apr 26 20:36:46 Crofton: Thinking about one thing mentioned, it's a bit to be expected that the 'rest of the world' keeps an eye on US news more than vice-versa, as the US is quite big compared to other independent nations. Although we now have a European Union, I don't think the US sees .eu countries as equal, whereas somehow vice-versa, we do. Apr 26 20:42:28 The interesting thing is the range of nationalities affected Apr 26 20:43:06 Crofton: as in: all? Apr 26 20:43:52 !2140 Apr 26 20:45:40 i can feel both the EU and pacific rim breathing down the back of my neck ;P Apr 26 20:46:57 !oebug 2140 Apr 26 20:47:44 * * Bug 2140, Status: NEW, Created: 2007-04-25 14:48 Apr 26 20:47:45 * * likewise(AT)gmx.net: RPATH of shared libraries contains host path, causing overhead Apr 26 20:47:46 * * http://bugs.openembedded.org/show_bug.cgi?id=2140 Apr 26 20:49:10 ~fishslap Apr 26 20:49:14 ~fishslap zecke Apr 26 20:49:16 * ibot slaps zecke up side the head with a wet fish. Apr 26 20:49:16 *tired* Apr 26 20:49:55 zecke: go to sleep Apr 26 20:50:57 zecke: we got your RPATH code working. Now all builds with insane class break because we have illegal RPATH's :-) Apr 26 20:53:59 woglinde: hi Apr 26 20:54:19 hi Apr 26 20:54:45 likewise: well, I want to see what I did wrong Apr 26 20:54:55 likewise: but bugs.oe.org doesn't respond Apr 26 20:55:36 likewise: IIRC I had a hard time to verify the code as most libs were sane back when I wrote it Apr 26 20:55:47 likewise: they were sane because reenoo spend time on fixing them Apr 26 20:56:11 zecke: You did "if bad_path in txt:" where bad_path was a string, txt was a list. That only works for perfect matches. Apr 26 20:57:15 zecke: this is where we'are at now: some experimental chrpath stuff included: http://rafb.net/p/QvV43q92.html Apr 26 20:58:15 zecke: btw, all libraries are now insane again, starting with glibc... Apr 26 20:58:54 zecke: so I really wonder about reenoo's work (who is reenoo btw, do not recall an IRC nick like that?) Apr 26 20:59:20 likewise: ask koen, or better don't :) Apr 26 21:03:54 later Apr 26 21:05:01 later Apr 26 21:05:07 don't get drunk Apr 26 21:05:19 What is the purpose, then? Apr 26 21:05:25 ;-) Apr 26 21:08:07 mickeyl: hello, good Dr Lauer :) Do you remember any sort of dependency analysis tools for python back when you had free time? I am starting to think about improving the way our python feed works. Apr 26 21:14:25 RP: there is a new mail by Sean Kelly Apr 26 21:14:35 RP: it is about mixing QtopiaPlatform and Gtkmm,Matchbox Apr 26 21:17:45 Is there a way to create a dependency tree from bitbake? I know there is alldepends.dot etc. But that is just A depends on B. I want a tree. "bitbake --dry-run --verbose -D -D -D $package|grep -i path" used to do that but recent version of bitbake don't produce this tree like structure. Apr 26 21:21:37 likewise, zecke: http://oe-devel.bkbits.net:8080/openembedded/?PAGE=patch&REV=1.2930.38.74 Apr 26 21:22:18 Images for angstrom are generated in build/tmp/deploy/glibc/images/$machine ? Apr 26 21:22:29 sirfred hehe Apr 26 21:22:31 What that glibc stands for ? Apr 26 21:22:44 glibc, as opposed to uclibc Apr 26 21:22:49 koen: funny, that's the first thing I tried before we started on insane.bbclass... Apr 26 21:22:49 koen: I don't know why the checksums are wrong, you need to compare them with hrw's source mirror Apr 26 21:22:57 koen oh hm Apr 26 21:23:04 koen: and then feel to update, for curiosity send me both tarballs Apr 26 21:23:19 I thought it is a bug in bitbake-1.8 Apr 26 21:23:22 sirfred: set ANGSTROM_MODE="uclibc" in local.conf to build a uclibc image Apr 26 21:23:34 is uclibc building now? Apr 26 21:24:37 I wonder why packages like gnumeric or gdb was generated after bitbake angstrom-x11-image Apr 26 21:24:54 What could be the reason? Apr 26 21:25:01 sirfred florian said becacuse of the tasks Apr 26 21:25:07 sirfred: bitbake 1.recent builds everything listed in the task .bb you depend on Apr 26 21:25:12 and he will like into it Apr 26 21:25:17 look Apr 26 21:25:19 damnit Apr 26 21:25:33 koen, woglinde : Thanks Apr 26 21:25:57 Well, copying the image to the CF. Let's see. Apr 26 21:28:55 I bet the ts is rotated Apr 26 21:31:45 libgcc_so.1 : cannot open shared object file... Apr 26 21:32:21 * likewise mumbles uh oh Apr 26 21:32:32 Unable to login Apr 26 21:33:08 Well, enough for today Apr 26 21:33:31 At least suspending works fast. ;-) Apr 26 21:34:13 Good night Apr 26 21:35:44 zecke: you have some ideas? Apr 26 21:37:44 ljp: more context please :) Apr 26 21:38:21 about mixing QtopiaPlatform and Gtkmm,Matchbox Apr 26 21:39:19 ljp: I would have hinted XQt as well Apr 26 21:39:19 ljp: and it offers rootless mode Apr 26 21:39:35 ljp: so with glib-mainloop integration and XQt you would be able to bridge Apr 26 21:45:37 there may be some reasearch somewhere about an integrations layer between qt/qtopia and gtk Apr 26 21:46:03 ljp: do you need any help with OE? Is your toolchain working? Do you need a hand on Qtopia Core integration? Apr 26 21:47:20 core?! bah! Apr 26 21:47:36 currently reasearching uclibc and qtopia phone... Apr 26 21:47:47 ljp: send patches :) Apr 26 21:47:48 core is for sissies Apr 26 21:47:59 ljp: Well your timezone classes will fail on uclibc for sure Apr 26 21:48:01 i just might Apr 26 21:48:21 ljp: Qtopia and X11, At least QtopiaCore offers embedding of foreign windows Apr 26 21:48:25 great.. thats one thing in qtopia I am respnsible for Apr 26 21:48:36 ljp: so you could theoretically build a XEmbed->QtopiaCore bridge Apr 26 21:51:48 ljp: do you know if your Berlin office has a serial for the GP? Apr 26 21:51:58 ljp: in that case I could borrow it and compare it to mine Apr 26 21:53:52 dont think so, but I can ask.. Apr 26 21:54:43 I fail to see what is wrong with mine... but I have not invested too much *time* Apr 26 21:55:03 ljp: so you have QtopiaCore in OE? Any updates to the 'ugly' version we have in OE? Apr 26 21:55:54 zecke whats wrong with your GP I want test phoneme on it Apr 26 21:56:12 not at this time... qtopia phone keeps its own version of qtopia core, so i sont use the one already there Apr 26 21:56:18 woglinde: 17 binary only modules Apr 26 21:56:28 * ljp grumbles Apr 26 21:56:37 zecke hehe Apr 26 21:56:52 woglinde: but for phoneME this would be okay Apr 26 21:57:20 woglinde: I can give it to you, for some reason I didn#t get it charged today thoguh Apr 26 21:57:53 zecke hm whats with your n770 it gets dusty here Apr 26 21:58:31 woglinde: do you want to return it? Forward it to NoB/ruediger so he can toy with squaek on it Apr 26 21:59:55 ljp: what will Trolltech do (besides this challenge) to get more devs interested in Qtopia? Apr 26 22:00:08 ljp: what will trolltech do with the interested devs? Apr 26 22:01:09 zecke okay I will bring it monday with me Apr 26 22:01:43 we may give away some gp.. depending on criteria Apr 26 22:05:38 ljp: well FIC financed the qemu Neo port AFAIK Apr 26 22:05:56 ljp: and you could create a gp board now easily Apr 26 22:13:16 morning Apr 26 22:17:20 hi xora Apr 26 23:26:38 03likewise 07org.oe.dev * r08836743... 10/ (1 classes/insane.bbclass): insane.bbclass: RPATH check was always true due to scanelf output including the file path. Apr 26 23:26:42 03likewise 07org.oe.dev * r7b9112b3... 10/ (1 classes/insane.bbclass): insane.bbclass: Commented out scanelf format debugging. Apr 27 00:13:57 03likewise 07org.oe.dev * rc29b7de6... 10/ (1 classes/insane.bbclass): insane.bbclass: Re-add -B to scanelf to make it correctly check single RPATH occurences. Apr 27 00:17:52 likewise: almost done? Apr 27 00:41:08 could someone upload qt-x11-free-3.3.7.tar.bz2 for me? the troll ftp is down Apr 27 00:43:14 nm, found a useable mirror Apr 27 00:53:00 03lenehan 07org.oe.dev * r51c79e4e... 10/ (1 packages/perl/perl_5.8.8.bb): perl 5.8.8: Fixes to be able to build under uclibc. Apr 27 01:00:54 CoreDump|home: wb :) Apr 27 01:01:02 hi hvontres|poodle Apr 27 01:01:18 CoreDump|home: I have a dumb svn question... how do I get my key to work??? Apr 27 01:02:23 hmm? well your SSH private key will authenticate your SSH pub key stored on the server. Apr 27 01:02:39 ssh hvontres@hentges.net should work just as well Apr 27 01:03:08 CoreDump|home: ahhh... hehe..I am logged in as 'henry' at home :) Apr 27 01:03:21 ahh Apr 27 01:03:39 replace $HOST w/ $USER@$HOST Apr 27 01:04:04 CoreDump|home: ok.. np Apr 27 01:37:46 I found that the SRC_URI for libpcre is no longer valid... however there is a valid url available... is this something that I should report as a bug / patch? Apr 27 01:47:21 vervain: don't bother Apr 27 01:47:43 03coredump2 07org.oe.dev * r177ed520... 10/ (1 packages/libpcre/libpcre_4.4.bb): libpcre: Fix SRC_URI Apr 27 01:54:43 CoreDump: Whoops - already did 2141 Apr 27 01:55:42 !oebug 2141 Apr 27 01:55:44 * * Bug 2141, Status: RESOLVED (FIXED), Created: 2007-04-26 18:51 Apr 27 01:55:44 * * wcravens(AT)cortex-it.com: libpcre SRC_URI is incorrect Apr 27 01:55:45 * * http://bugs.openembedded.org/show_bug.cgi?id=2141 Apr 27 01:55:46 ;) Apr 27 01:56:19 perfect Apr 27 01:57:43 That version of the library is pretty old... I realize there could be dependancy issues but is it worth looking into using a more recent version? Apr 27 01:58:03 * CoreDump|home dunnos Apr 27 01:59:39 n8 all Apr 27 02:08:45 03lenehan 07org.oe.dev * r39b0c54a... 10/ (10 files in 3 dirs): (log message trimmed) Apr 27 02:08:45 perl 5.8.8/cpan: Fix the installation paths for cpan modules. The installed Apr 27 02:08:45 files for perl modules built using cpan will end up in different places Apr 27 02:08:45 depending on which version of perl they are being built with. Modules that Apr 27 02:08:45 explicitly set various FILES_ values were using the paths that are only Apr 27 02:08:45 valid for the older versions of perl. Calculate and set the correct path in Apr 27 02:08:49 cpan.bbclass and use that in the FILES_ variables so that it'll be correct Apr 27 02:23:57 03lenehan 07org.oe.dev * ra1a53de4... 10/ (1 classes/cpan.bbclass): Apr 27 02:23:57 cpan: Use CCLD for the linking instead of ${TARGET_SYS}-gcc. Without this Apr 27 02:23:57 native packages break because -gcc doesn't exist, just gcc. CCLD Apr 27 02:23:57 gets us the correct CC to be used for linking in both cases, so we use that. Apr 27 02:33:08 hi, what is "PREFERRED_PROVIDER_virtual/xxx"'s usage? **** ENDING LOGGING AT Fri Apr 27 02:59:57 2007