**** BEGIN LOGGING AT Tue Dec 04 02:59:56 2007 Dec 04 03:24:53 Anyone familiar with udev? I'm trying to figure out how to adjust the automounting features and understand the blacklist file Dec 04 07:18:15 is there a way to redownload a source package? Dec 04 07:19:00 rm the old one + its md5 Dec 04 07:19:39 thanks CoreDump Dec 04 07:19:48 np Dec 04 07:21:52 except... doesn't seem to be redownloading the file, just the md5 Dec 04 07:23:35 summatusmentis: rm tmp/stamps/...? Dec 04 07:25:25 Hm, one of my housemates is looking for a device to act as a mini router.. well, provide ppp over bluetooth at least.. a device that can run as a usb host + bluetooth with OE.. can anyone recommend any hardware? Dec 04 07:30:33 NAbyss: linksys NSLU2 fits your profile Dec 04 07:31:02 <80$ IIRC Dec 04 07:32:41 CoreDump: Cool, thanks Dec 04 07:33:01 np Dec 04 08:15:53 * CM pokes NineX ;) Dec 04 08:45:53 gm Dec 04 08:48:33 morning likewise Dec 04 09:03:40 mickey|bbiab: pong Dec 04 09:03:51 mickey|bbiab: Have you an example of something that broke? Dec 04 09:04:40 I'ld like to unbreak something: | The external toolchain could not be found in /usr/local/angstrom/powerpc! Dec 04 09:06:04 likewise: Using angstrom? Dec 04 09:06:07 yo people Dec 04 09:06:16 RP: yes. Dec 04 09:12:11 likewise: Which bitbake version are you using? Dec 04 09:13:57 RP: 1.8.8 rev 936 Dec 04 09:14:39 morning all Dec 04 09:14:50 everybody uses angstrom \o/ Dec 04 09:15:27 Perhaps bitbake trunk wasn't the best way to test this :/ Dec 04 09:15:59 night rebuild for c7x0 was successful Dec 04 09:16:18 but always nasty NOTEs Dec 04 09:16:20 http://www.pastebin.org/10378 Dec 04 09:16:28 Ah, 1.8.9 will work better for this than 1.8.8 :/ Dec 04 09:16:53 built with 1.8.9 Dec 04 09:17:55 RP: So, 1.8.9 is now required? Dec 04 09:18:12 RP: which is an unstable release series, right? Dec 04 09:18:36 likewise: It won't give the error you're getting as the provider mechanism was made more sane Dec 04 09:18:53 likewise: Its the svn head of the stable branch Dec 04 09:19:14 * RP should release it as 1.8.10... Dec 04 09:20:11 I can add something to angstrom that should fix this for 1.8.8 too... Dec 04 09:20:45 likewise, XorA: The other way to solve this is: PREFERRED_PROVIDER_linux-libc-headers = "linux-libc-headers" Dec 04 09:20:45 PREFERRED_PROVIDER_virtual/${TARGET_PREFIX}binutils = "binutils-cross" Dec 04 09:20:51 in local.conf Dec 04 09:22:00 RP: thanks. It's hard to track BitBake/OE interdependencies for me. I have been looking at this in OE metadata alone... Dec 04 09:23:06 likewise: We set the minimum version in OE which is currently 1.8.8 so it should work, its just bad luck it currently doesn't quite :( Dec 04 09:24:08 * RP will push a change to angstrom to address this Dec 04 09:24:09 RP: ok, well, thanks for caring about this. It's good to see progress on the meta / sdk stuff, I'll take the bleeding for that. You sharpen the edge. Dec 04 09:28:48 ERROR: Multiple files due to be built which all provide virtual/xserver Dec 04 09:28:55 wasn't this solved? Dec 04 09:30:05 leaving imageon and removing older w100 stuff? Dec 04 09:31:14 good morning all Dec 04 09:32:56 morning koen Dec 04 09:33:21 koen: Can I add a couple of PREFERRED_PROVIDERs to angstrom to stop external-toolchain breaking things for bitbake 1.8.8 users? Dec 04 09:33:46 sure Dec 04 09:34:19 * cdm is having a hell of time getting OE to build against a pre-built toolchain. Dec 04 09:34:27 but I don't have anything intelligent to ask yet! :) Dec 04 09:34:45 cdm: You looked at the external-toolchain.bb file? Dec 04 09:35:29 good morgen koen Dec 04 09:35:44 RP -- sec Dec 04 09:35:57 likewise: half negen is veeeel te vroeg voor college ;) Dec 04 09:36:01 hello cdm Dec 04 09:36:20 koen: zelfs voor werk, zit er een beetje doorheen hier (ook al was het half tien) Dec 04 09:36:51 RP -- hurm, is that somewhat new? Dec 04 09:36:52 likewise: mtn up should fix your problem Dec 04 09:37:07 RP: thank a bunch! Dec 04 09:37:10 +s Dec 04 09:37:18 cdm: yes, new as of a couple of days ago Dec 04 09:37:39 * RP wonders what he did to distutils :( Dec 04 09:37:55 ahhh Dec 04 09:37:56 thanks Dec 04 09:38:38 cdm: Can I ask which toolchain you're using? Dec 04 09:38:39 where abouts does it live in the tree? Dec 04 09:38:50 RP: sure -- the latest code sourcery drop Dec 04 09:39:05 03rpurdie 07org.oe.dev * r8c692ec4... 10/ (3 files in 2 dirs): angstrom: Define some extra PREFERRED_PROVIDERs to ensure external-toolchain isn't touched Dec 04 09:39:07 cdm: Poky has a .bb file for an older csl toolchain too Dec 04 09:39:14 cdm: packages/meta Dec 04 09:39:35 yeah, I saw the old one. Dec 04 09:41:05 that's a great reference! thanks Dec 04 09:43:57 likewise: which kernel are you using on your efika nowadays? Dec 04 09:44:12 One thing I am unclear about, who generates the shlibs/ files? Dec 04 09:44:26 RP:sorry, my wireless needed to be kicked. Dec 04 09:45:03 cdm: Usually package.bbclass although with an external toolchain, the external toolchain .bb needs to Dec 04 09:46:53 RP: ahhh -- is that how you actually make the ipk's depend on the toolchain runtime? Dec 04 09:49:33 time for bed Dec 04 09:50:53 koen: morning Dec 04 09:50:58 koen: may I ask you about something really puzzles me since long time? Dec 04 09:51:03 sure Dec 04 09:51:20 koen: it's about make -j4 Dec 04 09:51:32 koen: I'd use -j3 Dec 04 09:51:47 koen: on parallel Dec 04 09:52:08 koen: it's proved faster Dec 04 09:52:26 I use the 2*(Ncpu) rule of thumb Dec 04 09:52:47 koen: Nr cpu + 1 ... was so for the last 20 years... Dec 04 09:53:18 on my build machine -j4 is faster than -j3 for most builds Dec 04 09:53:25 koen: but today perhaps has changed.... Dec 04 09:53:39 thanks to fast scsi disks :) Dec 04 09:54:08 koen: I use SCSI and 2 cpu since 1990 Dec 04 09:54:17 koen: 2x P133 Dec 04 09:54:28 koen: :-) Dec 04 09:55:09 * CoreDump just ordered a shiny quad box ;) Dec 04 09:55:16 koen: I'll check the build time on core2 dual with both flags and let you know Dec 04 09:56:45 koen: BTW how is developing on Tiger? Dec 04 09:57:24 koen: would you suggest it as *easy* testbed? Dec 04 09:59:24 koen: I am using the -CFS patched kernel for my EFIKA, let me check. Dec 04 09:59:56 koen: linux-efika_2.6.20.20.bb Dec 04 10:00:25 koen: should bring it to .23 or .24 soon Dec 04 10:00:26 ant|work: I have leopard, but I am not a developer Dec 04 10:00:48 likewise: I have 2.6.20.20-cfs as well Dec 04 10:00:55 likewise: but I think I messed up OF Dec 04 10:01:17 koen: good to know Leopard works Dec 04 10:01:32 koen: the OF non-volatile stuff? Dec 04 10:01:52 likewise: dunno, it tries to execute something bogus as bootscript Dec 04 10:02:10 and the only thing I ever typed into OF was 'boot efika-kernel' Dec 04 10:02:20 maybe some garbage on serial or something Dec 04 10:03:32 koen: does it reproduce (same garbage every time)? Dec 04 10:04:05 it looks like something echos the serial output back to OF Dec 04 10:04:13 and OF output isn't valid OF input Dec 04 10:05:17 koen: eek. could be cable, shortcut, or your terminal program echo'ing back. Dec 04 10:05:50 RP: what phase is that normally done? Dec 04 10:06:00 the generation of the shlib files. Dec 04 10:08:25 do_stage or do_install? Dec 04 10:10:42 cdm: As I said earlier, do_package Dec 04 10:10:55 cdm: See package.bbclass Dec 04 10:12:30 so I see in there where it looks in the files to generate dependancies Dec 04 10:13:04 but I don't see where it generates those files or their contents -- but I must be daft. Dec 04 10:14:23 n/m Dec 04 10:14:24 see it now Dec 04 10:15:05 I have a question with angstrom distro. The kernel in use is a 2.6.22. But during boot I get an error "WARNING: Couldn't open directory /lib/modules/2.6.22: No such file or directory" and in the /lib/modules is only an 2.6.20 directory. Dec 04 10:19:34 mickey|bbiab: I just bitbake'd python-pycario and twisted and both built fine :/ Dec 04 10:26:10 mickey|bbiab: and python-pyopenssl. You'd best tell me what fails ;-) Dec 04 10:36:14 RP: thanks -- I figured out how to coax package.bbclass to find the bits it was missing before. Dec 04 10:50:17 good monring Dec 04 10:53:14 hi! Dec 04 10:54:45 hi mickeyl Dec 04 10:56:18 mickeyl : hi ! Dec 04 10:57:03 mickeyl : any chance to get the pdf's from you or should i proceed and send the rest of invoices ? Dec 04 11:01:33 * mickeyl fights against the scanner atm. Dec 04 11:01:40 could be there in a couple of minutes Dec 04 11:01:48 now where does these scans go... Dec 04 11:02:25 skoroneos at digital-opsis dot com Dec 04 11:03:40 * koen proposes that mickeyl should get a mac Dec 04 11:07:22 ok, i think they're now on my usb stick Dec 04 11:07:25 * mickeyl agrees with koen Dec 04 11:08:14 koen: without Leoptard I assume :-) Dec 04 11:09:07 hmm hey this scanner rocks Dec 04 11:09:10 it can scan to usb Dec 04 11:09:15 without a PC Dec 04 11:09:17 and it works (tm) Dec 04 11:09:32 RP: pleaes try anything distutils that compiles external modules, i.e. python-pyxml or python-pysoappy Dec 04 11:10:35 RP: there's a small flaw in packaged-staging2.bbclass: do_deploy gets skipped, so you can't build e.g zaurus or neo images Dec 04 11:11:10 steliosk: sent! :D Dec 04 11:11:28 koen: no-one uses those machine anyway :-) Dec 04 11:11:30 koen: Yes, I mentioned it in the notes. Whilst it gets skipped, if it didn't it would have to build the kernel to have a kernel to deploy Dec 04 11:12:19 mickeyl: ok, so I didn't break all distutils ;-) Dec 04 11:12:28 RP: it just highlights the big flaw in the current do_deploy methods Dec 04 11:12:42 koen: Its a flaw? Dec 04 11:12:54 or is the flaw in packaged-staging? Dec 04 11:13:19 We could put deploy/images under staging control... Dec 04 11:13:36 most do_deploy methods go around install and staging for no apparent reason Dec 04 11:13:51 mickeyl: NOTE: package python-pyxml-0.8.4-ml0: task do_build: completed Dec 04 11:15:16 koen: That isn't really true... Dec 04 11:15:32 koen: stage and install never touch deploy/images Dec 04 11:17:08 RP: we could just have do_deploy get its stuff either from a generated packages (e.g. zImage from kernel-image.{ipk,deb}) or from staging Dec 04 11:17:09 mickeyl: NOTE: package python-soappy-0.11.6-r0: task do_build: completed Dec 04 11:17:14 argh, multiple heads. Dec 04 11:17:27 yay, OE automerger! Dec 04 11:17:38 XorA and rpurdie Dec 04 11:17:46 koen: That is another option Dec 04 11:18:38 hello Dec 04 11:18:49 in essence do_deploy is only there to get something into deploy so users can find it more easily Dec 04 11:18:51 what happened to the SRC_TARBALL_STASH proposal? Dec 04 11:19:57 mickeyl : got it thanks ! Dec 04 11:20:29 RP: koen: I like the idea of do deploy operating on packages Dec 04 11:21:24 XorA: it would solve the 'dfu-util gets only into one deploydir' problem Dec 04 11:21:45 koen: and the random file called uImage in kernel-image but no idea what file it is Dec 04 11:21:55 right Dec 04 11:23:11 the autobuilder used to run into trouble because I cleaned out deploy from time to time Dec 04 11:23:11 03xora 07org.oe.dev * r5f67fa16... 10/ (1 conf/distro/include/angstrom-2007-preferred-versions.inc): Dec 04 11:23:11 angstrom-2007-preferred-versions.inc : increase the Angstrom version of Dec 04 11:23:11 kismet to 2007-10-01 as it has a number of wanted improvements as detailed Dec 04 11:23:11 in #3387 no objections were raised on mailing list. Dec 04 11:23:50 steliosk: sorry again. sometimes i suck Dec 04 11:24:10 RP: *sigh* Dec 04 11:24:17 RP: org.oe.dev unchanged? Dec 04 11:25:22 mickeyl: This is a clean angstrom build with no changes against oe.dev apart from the external toolchain stuff Dec 04 11:25:41 and that is disabled in this build... Dec 04 11:26:52 hmm Dec 04 11:26:57 this is very strange Dec 04 11:27:08 03jeremy_laine 07org.oe.dev * r3e94a66e... 10/ (4 files in 2 dirs): sofia-sip: split glib-specific parts to ${PN}-glib Dec 04 11:27:43 mickeyl: In theory we haven't hit the rough patch with the staging layout changes yet since I'd just been preparing the land, I haven't flicked the layout change switch yet Dec 04 11:28:08 RP: ok, sorry for blaming you then. i'll find out what's wrong here Dec 04 11:28:12 It is possible I replaced something with something that wasn't equivalent though... Dec 04 11:28:40 mickeyl: Can you pastebin what you're seeing? Dec 04 11:29:24 RP: need to do some more researches first. will get back to you Dec 04 11:29:27 thanks Dec 04 11:29:38 mickeyl: ok Dec 04 11:29:54 * XorA watches monotone pipe shrink to 1byte/s Dec 04 11:30:33 XorA: wait till it reaches 0.3 bits/s Dec 04 11:30:54 the mtn server is down to ~15 concurrent connections after rwhitby pointed the makefile to the openmoko server Dec 04 11:31:12 (from 80-100 concurrent connections) Dec 04 11:31:48 koen: now imagine if you had to teach all those people how to change their sync machine manually in their setup ... Dec 04 11:32:14 I wonder what the 'distributed' in monotone stands for :-) Dec 04 11:32:31 rwhitby: one line people can paste into shell from an email Dec 04 11:32:49 likewise: all nslu2-linux people sync to a distributed version of org.oe.dev Dec 04 11:33:09 koen: and you think that more than 50% would get that right? Dec 04 11:33:26 I have no idea Dec 04 11:33:49 (am I being elitist enough to join the OEDEM club?) ;-) Dec 04 11:34:11 rwhitby: no, you're too helpful Dec 04 11:34:30 rwhitby: 50% is probably an overestimate ;-) Dec 04 11:34:34 mickeyl : np i understand. You'll buy beers next time we meet to cover all the sweat i spend on this :) Dec 04 11:34:35 It has always puzzled me that 'mtn up' is too hard, but deleting stamps isn't for users Dec 04 11:34:43 rwhitby: (I'm out of my league here, I'm a mere contributor :-) ) Dec 04 11:35:29 even more so, because I haven't come across a document that tells (how) to delete stamps Dec 04 11:35:41 * RP never did reply to the makefile thread and probably should have done :/ Dec 04 11:37:40 * rwhitby goes to bed before a makefile discussion breaks out here. Dec 04 11:38:58 rwhitby: It won't come from me, I don't have the time :( Dec 04 11:39:45 rwhitby: Just do understand I'm not against the makefile and think it is needed and does a lot of good. I just wish it worked a little differently... Dec 04 11:40:37 bitmake Makefile update x11-image :-) Dec 04 11:41:33 hehe Dec 04 11:44:15 steliosk: (beer) deal! Dec 04 11:44:19 arghs, was disconnected Dec 04 11:44:21 here my sermon Dec 04 11:44:29 oh no Dec 04 11:44:31 it's gone Dec 04 11:44:33 *sigh* Dec 04 11:45:54 RP: i _think_ i got it Dec 04 11:47:33 mickeyl, after the driectfb changes, I also built and ran directfb-examples on the OSK Dec 04 11:48:09 Crofton|home: all worky? Dec 04 11:49:09 looks like it Dec 04 11:49:26 we had some questions on the background when run on the gumstix Dec 04 11:49:34 we had the same background on teh OSK Dec 04 12:01:47 hi Dec 04 12:01:54 ewi has network outage again Dec 04 12:02:31 it has? Dec 04 12:02:37 I just logged into it Dec 04 12:02:45 and linuxtogo still has the ssh tunnel up Dec 04 12:03:00 so maybe it is somewhere between me and ewi Dec 04 12:14:03 koen: FYI the random file called uImage in kernel-image is the U-BOOT kernel Dec 04 12:15:05 * otavio is having trouble to get a gethostbyaddr to work inside of the image. It always return null ... any idea? Dec 04 12:15:28 ant|work: not when you use vendor patches Dec 04 12:15:39 then it's a file called uImage that might or might not boot Dec 04 12:15:59 which is why OE generates its own uImage instead of trusting the kernel makefiles Dec 04 12:18:38 koen: ...sorry..deformation...I was reading about u-boot Dec 04 12:21:56 koen, we should save the URL to rmk's comments on the uImage subject :) Dec 04 12:22:38 koen, are you also the website guy? Dec 04 12:23:20 koen:Crofton: http://www.seicento.selfip.org/zaurus/issues/using_u-boot_with_other_roms.html Dec 04 12:24:10 koen:Crofton: shorter manual http://www.embedian.com/index.php?main_page=uboot_manual Dec 04 12:24:41 koen:Crofton: I have the serial cable and will do som tests Dec 04 12:37:13 and re Dec 04 12:37:20 I will be offline today Dec 04 12:37:30 so bye all Dec 04 12:37:51 hm. Dec 04 12:44:19 hrw , alter ego ? Dec 04 13:06:56 * Genesis discovers parted mklabel msdos -_- Dec 04 13:19:49 who feels able to fix a kernel module and being paid for that ? Dec 04 13:19:55 * nud is looking for company names Dec 04 13:21:21 bec systems (aka cbrake) or o-hand (aka RP) come to mind Dec 04 13:23:05 koen: that's the one I knew too, but I wondered if there were more of them Dec 04 13:23:09 thanks anyway Dec 04 13:23:58 a friend of mine is kernel devel and freelance Dec 04 13:24:03 mathdesc@scourge.fr Dec 04 13:24:22 feel free to mail him of my part Dec 04 13:25:08 http://scourge.fr/mathdesc/index.php?lang=en&content= Dec 04 13:25:48 ( company website under construction ) Dec 04 13:34:16 mickey|breakfast, you are keeping east coast hours ... Dec 04 13:45:53 guys, 2.6.24 has some SDIO stuff in it, right? any pointers to the project site or additional info? Dec 04 13:47:17 Jin^eLD: iirc pierre closed his wiki Dec 04 13:48:02 closed? why that? Dec 04 13:48:16 it had lots of obsolete info Dec 04 13:48:34 see http://kernelnewbies.org/Linux_2_6_24#head-b11faf7df86672a665b8f4cce03f8c42b712f475 Dec 04 13:48:38 thanks Dec 04 13:49:11 it links to the lkml.org thread Dec 04 13:50:39 I'll probablys have to write an sdio driver Dec 04 13:50:44 so looking for info :) Dec 04 13:50:49 let's see... Dec 04 13:52:07 but what he writes in the thread - is it already in 24-rc3 or is it still only in the mmc tree? Dec 04 13:52:22 I think it's in rc3 Dec 04 13:52:33 everything in the kernel newbies is pulled from -rcX Dec 04 13:52:37 AIUI Dec 04 13:52:41 aha.. Dec 04 13:53:20 mickey|breakfast: does pulse really need libtool .24 or is it just that .24 works because it doesn't have the OE rpath patches? Dec 04 13:55:05 hmm libtool rearing its ugly head Dec 04 13:55:24 I added .24 because one of the netfilter packages would not build with .10 Dec 04 13:56:04 * XorA is suprised we dont have libtool, glibtool, klibtool :-) Dec 04 13:56:28 XorA: kde had its own patched copy of libtool in their svn repo Dec 04 13:56:41 before they switched to scons/cmake/make-du-jour Dec 04 13:57:04 one of the reasons why there is no kde in OE Dec 04 13:59:08 elibtool Dec 04 13:59:11 xlibtool Dec 04 13:59:14 :-D Dec 04 13:59:21 uclibtool Dec 04 14:00:17 Damn. With updated patchset tosa can suspend, but doesn't resume. Dec 04 14:00:32 lumag: cool, at least it saves power :-) Dec 04 14:00:40 lumag: BTW did you check out the wolfson repo? Dec 04 14:01:04 yeah. But I looked into it, but didn't applied it yet. Dec 04 14:01:30 lumag: cool, wasnt sure if you managed to see my offline messages Dec 04 14:01:58 hey vivijim Dec 04 14:02:52 XorA: hi! Dec 04 14:02:53 And it doesn't save power: screen goes shining white :) Dec 04 14:03:23 lumag: they sell that for mobile phones as a torch, costs like 5 GBP, quick sell it Dec 04 14:04:19 Hmmm. Good idea :) Dec 04 14:07:43 t- 50 minutes Dec 04 14:08:36 Maybe there is a problem with tosa_pm. It's a mess. Dec 04 14:12:44 Genesis: and what name should I put after "I'm contacting you on behalf of ..." ? :-) Dec 04 14:21:16 koen: that's a good question. i have no idea. what's actually required is libltdl 1.5.24, not libtool per se Dec 04 14:21:43 Crofton|home: your defense is today, right? or was it already? Dec 04 14:22:22 mickeyl: from his previous comment its in 38 minutes Dec 04 14:23:49 aah Dec 04 14:23:54 what a relaxed guy Dec 04 14:24:00 i was nuts the whole day before Dec 04 14:24:31 Crofton|home: always sounds kinda chilled to me :-) Dec 04 14:28:57 yea Dec 04 14:30:41 ok Dec 04 14:33:11 mickeyl: libltdl is part of libtool though. There shouldn't be any real reason we can't use libtool 2.5.24 yet leave libtool-cross and -native as 2.5.10 Dec 04 14:33:21 1.5.10 even Dec 04 14:34:16 yeo Dec 04 14:34:19 that'd be best Dec 04 14:35:06 I've finally had to look at the libtool 'code' and it makes me ill... Dec 04 14:35:12 hehe Dec 04 14:35:14 i can imagine Dec 04 14:35:24 Fixing things kergoth broke about 5 years ago :) Dec 04 14:36:00 :-D Dec 04 14:37:15 nud : Ronan Bignaux Dec 04 14:37:20 To be fair its not as if anyone had used OE for darwin builds before... Dec 04 14:37:34 Genesis: merci Dec 04 14:37:39 de rien ;) Dec 04 14:41:33 does anyone here do remote ipkg package installation? if yes, how? Dec 04 14:42:11 other than ssh into the target system, and manually doing ipkg install Dec 04 14:42:54 you can use serial Dec 04 14:43:03 mbuf: never tried it, but its theoritically possible to use sshfs to mount / on device into a directory on host, then use ipkg on host with the right switches to do the install Dec 04 14:43:19 yeap with fuse Dec 04 14:44:24 minicom + lrzsz works when ethernet is not available Dec 04 14:44:32 XorA, i see Dec 04 14:45:17 morning Dec 04 14:45:39 XorA, but, that will be again manual, right? can it be automated from remote invocation? Dec 04 14:46:06 ishould work on that too Dec 04 14:46:37 mbuf: dont see why not Dec 04 14:46:50 Genesis, "ssh root@192.168.0.202 ipkg install foo.ipk" segfaults for me (target OM+Neo1973) Dec 04 14:46:50 mbuf: but Ive never tried these tricks Dec 04 14:48:59 XorA: ipkg probably needs a specific terminal and console emu to launch diff and friends Dec 04 14:52:50 koen: well its got that my way :-) Dec 04 14:54:52 drat, it seems the akita updater.sh is still not working properly Dec 04 15:42:57 XorA: replied to your email, CC:ed hopefully the appropriate persons and added everything I know sofar Dec 04 15:43:35 koen: cool, I have come late to the party, so I dont know if the ground was already covered Dec 04 15:43:52 I think all of us have :/ Dec 04 15:44:49 * koen still wonders what happened to the original OE server from gerwinin Dec 04 15:45:18 ~seen gerwinin Dec 04 15:45:21 gerwinin was last seen on IRC in channel #oe, 27d 15h 40m 29s ago, saying: 'zecke: good night :) need to work some with the sdcc compiler :('. Dec 04 15:45:33 real life I guess Dec 04 15:45:47 no it can not be done i tell you Dec 04 15:46:07 last I heard was that the server was split into 2 half functioning servers awaiting parts Dec 04 15:46:13 but that was like 5 months ago Dec 04 15:46:48 what is wanted from this server, power or disk ? Dec 04 15:47:51 for mtn we just need something with ~500MB ram and a gig of diskspace Dec 04 15:48:03 for tinderbox we would need more ram and more diskspace Dec 04 15:48:16 for the autobuilder we would need even more ram and even more diskspace Dec 04 15:49:23 so depending on the server we can migrate 1 or more services from ewi546 to there Dec 04 15:49:32 * CoreDump has traffic to spare but his servers CPU is too weak for mtn and the like Dec 04 15:50:27 CoreDump: I suspect we can get hardware, rack space is the difficulty Dec 04 15:50:45 I see Dec 04 15:50:58 hardware is a one off cost at worst Dec 04 15:51:07 rack space needs to be budgeted :-( Dec 04 16:12:52 got to go now Dec 04 16:13:01 bye! Dec 04 16:24:07 food Dec 04 16:40:35 RP: ping Dec 04 16:53:35 yeapah \o/ Dec 04 16:53:39 first nao boot on OE Dec 04 16:53:41 ;D Dec 04 16:54:33 swap(nao,oe) Dec 04 16:58:55 done, I have the signatures Dec 04 17:03:28 Crofton|home: grats Dec 04 17:03:32 morning Dec 04 17:03:55 thanks Dec 04 17:20:29 Crofton|home: congrats! Dec 04 17:32:33 03Laibsch 07org.oe.dev * r124fe75a... 10/ (3 files in 2 dirs): uicmoc4-native: Drop non-compiling 4.2.3 and 4.3.0 (closes bug 2511). Dec 04 17:32:45 03koen 07org.oe.dev * rfc4bb63f... 10/ (3 files in 3 dirs): console-mipv6-image: add skeleton image for ipv6 mobility folks Dec 04 17:32:49 03koen 07org.oe.dev * re7392dcc... 10/ (1 contrib/angstrom/build-release.sh mtn:execute true): build-release.sh: script that tells the angstrom autobuilder what to build Dec 04 17:33:35 Crofton|home: You defended your thesis, what are you going to do next?...:) Dec 04 17:39:40 arm-poky-linux-gnueabi-ld: no machine record defined Dec 04 17:39:52 someone get something like this during kernel build? Dec 04 17:51:09 hi greentux, hi pH5 Dec 04 17:51:18 hej florian Dec 04 17:51:23 hi florian Dec 04 18:12:42 Hi Khem, I think some time ago koen told me to ask you about this: http://en.pastebin.ca/805813. Dec 04 18:12:50 I need this patch to make gcc 4.2.2 compile because otherwise it fails on the psignal declaration error due to the difference in parameter signedness between libiberty and glibc. Any comments? Dec 04 18:16:05 * Khem looks Dec 04 18:18:21 pH5: what target are you compiling for Dec 04 18:19:00 Khem: armv4t and armv5te angstrom-linux-gnueabi Dec 04 18:19:48 pH5: is it for uclibc or glibc Dec 04 18:20:04 oh you said glibc Dec 04 18:20:39 pH5: does it happen when you build toolchain from sratch ? Dec 04 18:23:13 Khem: hm, good question. I think it did. But I'll retry to be sure. Dec 04 18:24:15 pH5: Retry to build from scratch Dec 04 18:24:28 something could be messed up in glibc headers Dec 04 18:37:22 http://www.balister.org/~balister/card.png Dec 04 18:38:24 03mickeyl 07org.oe.dev * rfd202300... 10/ (1 packages/sword/sword_1.5.10.bb): add sword 1.5.10 Dec 04 18:38:32 03mickeyl 07org.oe.dev * ra6bb9a36... 10/ (4 files in 2 dirs): add python-sword 1.5.10, remove older versions that no longer build. closes #2790 Dec 04 18:38:44 03mickeyl 07org.oe.dev * r96d78702... 10/ (1 packages/python/python-pygtk_2.10.4.bb): python-pygtk 2.10.4: fix dependencies and packaging. closes #2905 Dec 04 18:46:56 mickey|sports: ping Dec 04 18:47:33 "my_var = bb.data.getVar('VARIABLE', d, 1) or 0" - it gives me 'my_var' as string - how to make integer from it? Dec 04 18:48:37 bye Dec 04 18:51:37 ok. found Dec 04 18:51:43 my_var = int(my_var) Dec 04 18:51:46 hrw: int (x) ? Dec 04 18:51:52 ah there you go Dec 04 19:20:28 bye Dec 04 19:24:49 Crofton|home: Nice to see the word "Polytechnic" in a school name again :) Dec 04 19:25:21 * hvontres|work went to CalPoly, San Luis Obispo, the best small eng school there is :) Dec 04 19:27:04 heh Dec 04 19:27:24 they like to call it Virginia Tech because people have trouble spelling polytechnic Dec 04 19:28:34 Crofton|home: and VP dosn't sound quite so intimidating...:) Dec 04 19:28:50 heh Dec 04 19:30:46 Crofton|home: so now that you have an MS, are you going to avoid real life a bit longer (ie PHD) ? Dec 04 19:31:14 no PHD, hopefully I can avoid reality by consulting Dec 04 19:32:27 Crofton: heh..."Consulting: If you can't be part of the solution, there is plenty of mony to be made by prolonging the problem" -- one of my favorite demotivators Dec 04 20:09:33 re Dec 04 21:07:54 03mickeyl 07org.oe.dev * r0bcc89a2... 10/ (1 packages/python/python-pygtk_2.10.4.bb): python-pygtk 2.10.4 package demo seperately and fix demo dependencies Dec 04 21:28:55 vivijim, ping Dec 04 21:29:30 hi Crofton Dec 04 21:29:45 I replied to your email about directfb Dec 04 21:29:55 do you know of anything that we broke with the commit? Dec 04 21:30:12 wol I'm going to read Dec 04 21:30:33 we did look at packages that listed directfb and it didn;t look like any compiled before the change Dec 04 21:30:36 for me, evas has broken Dec 04 21:30:43 and after we can run the example Dec 04 21:31:13 is that in .dev? Dec 04 21:31:16 uhm... but why can't the exemple work with the standard? Dec 04 21:31:35 the include files are not where -examples expected them Dec 04 21:37:40 Crofton isn't the exemple wrong? for trying the wrong include directory? Dec 04 21:37:43 vivijim, whch bb file is evas in? Dec 04 21:37:45 heh Dec 04 21:37:51 we had to chose one to fix Dec 04 21:38:09 and could not find any examples that suggest directfb was correct Dec 04 21:39:08 Crofton: http://www.openembedded.org/repo/org.openembedded.dev/packages/efl1/evas_cvs.bb Dec 04 21:41:43 Crofton I've noticed that the standard of directfb include dir is the /usr/include/directfb... And another thing that I've noticed is that using /usr/include directly it creates directories with generic names like "core".... For instance this core directory should be just part of directfb include... Dec 04 21:42:19 brb Dec 04 21:44:02 morning everyone Dec 04 21:47:50 vivijim, Dec 04 21:48:04 sadly the examples used the headers from /usr/include Dec 04 21:48:11 not /usr/include/driectfb Dec 04 21:49:00 sakoman, made the change, I suspect becasue it was less intrusive than changing examples Dec 04 21:49:15 I would create a bug and mention the problem on the listserv Dec 04 21:52:09 Crofton: I checked the directfb web site and they used include there as well for the sample code. Dec 04 21:52:11 uhm... do these examples belong to the directfb developers? Dec 04 21:52:14 Crofton ^ Dec 04 21:52:18 Yes, they do Dec 04 21:52:25 ah, the culprit :) Dec 04 21:52:33 I hate it for the same reason you do Dec 04 21:52:39 sakoman, are you interested in evas? Dec 04 21:52:48 Don't know what it is :-) Dec 04 21:52:58 So I don't know! Dec 04 21:53:06 wal it is insane... so they don't have a standard... afff Dec 04 21:53:15 hehehehe Dec 04 21:53:18 yeah Dec 04 21:53:39 Yeah, sometime I need to check to see what gtk+-directfb expects Dec 04 21:53:59 For gumstix that will make the decision Dec 04 21:54:02 Can one of you open a bug for tracking directfb problems? Dec 04 21:54:35 I am guessing this will crop up again Dec 04 21:54:57 I am guesing the directfb in OE community consists of you two .... Dec 04 21:55:02 You mean just a generic bug for directfb? Dec 04 21:55:09 maybe ... Dec 04 21:55:22 not sure if a generis, or make specific bugs Dec 04 21:55:36 maybe just create bugs for specific problems Dec 04 21:55:51 I think once I have a look at gtk+-directfb I'll know what to file the bug against :-) Dec 04 21:56:08 examples has the lowest priority Dec 04 21:56:13 ok, lets create specific bugs for broken directfb stuff Dec 04 21:56:29 Something real that uses it should be the deciding factor Dec 04 21:56:41 I'll try and get a little interested next week Dec 04 21:56:54 need to do thesis revisions this week and get ETD form signed ..... Dec 04 21:59:20 and go over resume :) Dec 04 22:00:38 Oh no! Revisions :-( Dec 04 22:01:06 apparently I have a fondness for comma splices, and one of my committee knows what they are .... Dec 04 22:01:58 Crofton: Oh no, a defrocked english major ... Dec 04 22:02:10 heh Dec 04 22:03:24 Crofton: heh..actually, just translate the whole thing in to Germaan :) Dec 04 22:03:38 arg Dec 04 22:03:57 * hvontres|work spent first 10 years in German school and still uses German rules for comas in English.. Dec 04 22:04:46 Ah, now that I know what a comma splice is -- just search and replace every other comma with a period :-) Dec 04 22:04:54 apparently Dec 04 22:05:23 I have a fondness for -- splices :-) Dec 04 22:05:31 I do not like to break up my thoughts with periods ..... Dec 04 22:05:34 Crofton: heh, that should be easy enough for sed :) Dec 04 22:05:56 Well they always need something to criticize Dec 04 22:06:12 If it is style that means you got the content nailed Dec 04 22:06:18 yeah Dec 04 22:06:26 I need to look over the comments Dec 04 22:06:31 that is tomorrows task Dec 04 22:06:43 Not the tsc2003 driver? Dec 04 22:06:50 heh Dec 04 22:07:03 I will try that if it looks like people are getting some where Dec 04 22:07:09 I need to go over my todo list Dec 04 22:07:23 sakoman: I was taking a look to directfb-examples and I believe the best way is to pass an extra flag: CPPFLAGS=-I/usr/local/include/directfb/ Dec 04 22:07:30 what do you think? Dec 04 22:08:17 I like that -- does it work? Dec 04 22:08:45 I'm for anything you two agree on, and having -examples work :) Dec 04 22:08:52 I still think we should see what other major packages like gtk+ expect Dec 04 22:08:58 yeah Dec 04 22:09:30 If everything we care about (except examples) expects include/directfb I will be happy Dec 04 22:09:41 Avoids the pollution Dec 04 22:09:47 yeah Dec 04 22:10:01 I did show db_cpuload today :) Dec 04 22:10:07 On the tsc2003: I think that one issue is that penirq isn't connected on the gumstix hw Dec 04 22:10:30 The driver author thinks we can work around that fairly easily Dec 04 22:11:04 thats good Dec 04 22:11:10 you found the original author? Dec 04 22:11:25 Line 4 of the source code :-) Dec 04 22:12:35 ah Dec 04 22:13:25 alright, I need to go home Dec 04 22:13:31 bbiaw Dec 04 22:47:09 hey all Dec 04 22:47:22 anyone know why doing PR += "foo" would have PR be "r0 foo" ? Dec 04 22:48:59 because thats what += does Dec 04 22:50:10 Bah, that would xplain it then :) Dec 04 22:50:15 Is there a concat w/o space? Dec 04 22:50:58 PR="${PR}foo" Dec 04 22:51:57 NOTE: :variable PR references itself! while evaluating: Dec 04 22:52:02 oh - I guess not Dec 04 22:52:04 ...everything that uses PR :( Dec 04 22:52:08 That ustacouldawork Dec 04 22:52:13 must have changed Dec 04 22:52:15 good thought tho, i think i tried that earlier Dec 04 22:54:08 Tartarus: maybe PR_tmp="${PR}foo";PR=$PR_tmp ? Dec 04 22:55:31 same :( Dec 04 22:57:49 03mickeyl 07org.oe.dev * rab7efb1c... 10/ (1 packages/tasks/task-python-sharprom_20060425.bb): task-python-sharprom: catch up with renamings and unavailable packages. closes #3322 Dec 04 23:11:00 Tartarus; would this work: Dec 04 23:11:11 PRTMP := PR Dec 04 23:11:28 er, PRTMP := ${PR} Dec 04 23:11:43 PR = "${PRTMP}foo" Dec 04 23:11:50 ... Dec 04 23:12:13 or better yet; does this work: PR := "${PR}foo" Dec 04 23:12:14 ? Dec 04 23:13:50 gcl, that's looking like it Dec 04 23:13:53 no immediate errors, thanks Dec 04 23:14:12 * Tartarus thought he tried a :=, but mustnot have Dec 04 23:17:12 mranostay, ping Dec 04 23:17:13 ga Dec 05 00:02:16 03coredump2 07org.oe.dev * rc44db3b9... 10/ (3 files in 3 dirs): zaurus-updater: Try to unb0rk the flashing of Akita kernels. Still needs work Dec 05 00:20:10 03mickeyl 07org.oe.dev * r8f4d485f... 10/ (1 packages/pulseaudio/pulseaudio_0.9.8.bb): pulseaudio 0.9.8 needs dbus now Dec 05 00:20:25 03mickeyl 07org.oe.dev * rdfdb2e5e... 10/ (1 packages/python/python-pygtk_2.10.4.bb): python-pygtk 2.10.4 python-tokenize is included in python-lang Dec 05 00:20:34 03mickeyl 07org.oe.dev * rbf13c465... 10/ (3 files in 3 dirs): pulseaudio 0.9.8 needs patch for corner case when you build with dbus but without HAL Dec 05 00:45:18 Hi Dec 05 00:46:01 I think c7x0.conf need one more entry: PREFERRED_PROVIDER_virtual/xserver = "xserver-kdrive-imageon" Dec 05 00:46:25 this apparently solves the "multiple provider" problem Dec 05 01:26:07 hello **** ENDING LOGGING AT Wed Dec 05 02:59:56 2007