**** BEGIN LOGGING AT Tue Oct 16 02:59:57 2007 Oct 16 04:17:06 stefan_schmidt: So I found the problem Oct 16 04:17:11 Someone closed SMTP to my server. Oct 16 04:17:16 Thanks for tipping me off. Oct 16 04:18:29 ggilbert: For your private email address problem? Oct 16 04:18:46 ggilbert: Thanks goes to johnlee Oct 16 04:19:25 stefan_schmidt: For the you not getting email from bugzilla Oct 16 04:19:35 ggilbert: Ah, cool Oct 16 04:19:50 the openmoko tries mailing me back, but port 25 is apparently firewalled off Oct 16 04:19:54 ggilbert: But that also fixes that the email admin address on the OE site bounces? Oct 16 04:20:07 ggilbert: ah, spam checks Oct 16 04:20:29 I'll email the network admin to get that fixed. Oct 16 04:20:40 ggilbert: Means there could have been other problem too. Anyway, thanks for fixing. Oct 16 04:20:41 It explains why my spam flow has gone down. Oct 16 04:20:48 It isn't fixed yet :) Oct 16 04:20:53 ggilbert: Just ping me when it's done and I'll try again Oct 16 04:20:54 I just noticed that a second ago Oct 16 04:21:05 ggilbert: k Oct 16 06:26:31 morning Oct 16 06:31:47 mmph... I guess technically it's morning Oct 16 06:34:05 summatusmentis: go to bed then :-D Oct 16 06:37:09 XorA : Morning ! Oct 16 06:37:20 yo steliosk Oct 16 06:39:04 I see you are an early bird today :) Oct 16 06:39:41 lot's of "worms" crawling around i assume ? :) Oct 16 06:41:32 03xora 07org.oe.dev * r91c6cc22... 10/ (3 files in 2 dirs): packages/usbpath/usbpath_svn.bb : DEPENDS on libusb and native does as well. Oct 16 06:42:00 steliosk: just awake and I dont work in an office Oct 16 07:27:36 stefan_schmidt: ok, mail is starting to come through. Give it a little while because I have about a months worth the old mail pouring in Oct 16 07:27:39 but you should be set Oct 16 07:27:59 ggilbert: cool, thanks. Oct 16 07:28:20 ggilbert: I'll try in some hours. Need to finish some coding first. Oct 16 07:28:25 sure thing Oct 16 07:28:51 It's bed time out here, so I'll see if it worked for you in the morning Oct 16 07:29:24 ggilbert: I'll just leave you a message on irc. You read your private msg backlog? Oct 16 07:29:32 yeah I do Oct 16 07:30:05 ggilbert: ok Oct 16 07:30:23 ggilbert: night Oct 16 07:30:29 Night Oct 16 07:35:18 good morning all Oct 16 07:35:31 yo koen Oct 16 07:42:57 "I have been meaning to write a replacement for the OO splashscreen that starts abiword to allow the poor OO users to get some work done before lunch ;)"" Oct 16 07:43:18 Does OE put any information into the rootfs about the build date or name of the build or somthing Oct 16 07:44:39 Hmm.. /etc/issue and /etc/version maybe.. Oct 16 07:46:13 /etc/angstrom-version ? Oct 16 07:47:12 Is that set by bitbake or is it some distro thing? Oct 16 07:54:28 bitbake knows nothing about rootfs'es Oct 16 07:56:14 03daniel 07org.oe.dev * r3fbecafa... 10/ (3 files in 3 dirs): alsa-state: Add state file gsmspeakerout (fic-gta01) Oct 16 08:06:36 hmm, what's the deal with INHIBIT_AUTO_STAGE in autotools.bbclass? As far as I can see recipes need to opt-in to autotools staging by writing something along the lines of do_stage () { autotools_stage_all } so why needs there to be an opt-out way with INHIBIT_AUTO_STAGE = "1"? Oct 16 08:07:10 (I'm asking because I want to implement something similar for mono and am currently pondering whether it should be opt-in or opt-out) Oct 16 08:09:18 Hello? Oct 16 08:12:52 I'm having some problems with the angstrom distro on our new development board... The distro builds succesfully, but when i'm booting, the last line i see at my console is Algorithmics/MIPS FPU Emulator v1.5. After that, nothing happens. When i remove the /sbin/init symlink, i get a kernel panic, obviously. So, it starts sysvinit, but it does nothing! Inittabs and other initstuff is OK. Oct 16 08:12:52 We have used the same configuration, in combination with a pre-build uclib toolchain. That worked perfectly. But we want to use glibc, so it is something with the toolchain, i pressume. Oct 16 08:12:52 I asked this in the Angstrom channel, but i didn't get a reaction..... Oct 16 08:25:34 hrw : what was the name of the tablet pc you had OE running on ? Oct 16 08:26:53 Reneger : You can build OE using your existing toolchain to check Oct 16 08:27:32 Well, it worken friday, we changed it to the standard OE glibc toolchain yesterday, and it didnt start at once. Oct 16 08:28:06 It starts sometimes, after a couple of resets, we get a segmentation fault, and it starts the init. Oct 16 08:28:48 steliosk: I think it was a conf/machine/progear.conf Oct 16 08:29:46 hmmm strange.... If it was a toolchain issue, then the produced code would fail to start everytime Oct 16 08:30:38 Hmyeah, or it is the sysvinit in combination with the toolchain... Oct 16 08:32:49 Reneger : Copy the init that works to the OE build fs and try to see if it works Oct 16 08:34:19 I think the init is the same, with the prebuild or the glibc toolchain, but i give it a try. Oct 16 08:41:30 've copied the while /etc/init* from the working version to the current version. No effect :(\ Oct 16 08:42:06 while = whole. Typo ;) Oct 16 08:43:03 morning broonie Oct 16 08:43:52 Morning. Oct 16 08:47:14 So, i really don't know where to find it... I'm working on this together with a colleague, and we both are out of ideas now... Oct 16 08:55:24 morning XorA Oct 16 09:05:21 hey lrg Oct 16 09:13:07 morning Oct 16 09:13:15 steliosk: FrontPath Progear Oct 16 09:17:15 hrw : Thanks ! Oct 16 09:32:04 * * OE Bug 3159 has been created by lynn.lin(AT)avocent.com Oct 16 09:32:06 * * iputils package set ${PN} package empty Oct 16 09:32:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3159 Oct 16 09:32:46 steliosk: want to play with it? Oct 16 09:33:45 No one has any clue where the problem is? :| Oct 16 09:37:25 Reneger : You use busybox for init ? Oct 16 09:38:13 no, sysvinit Oct 16 09:38:48 and both sysvinit show the same problem ? Oct 16 09:40:12 changing from prebuilt-uclibc toolchain to oe-glibc toolchain is a big step Oct 16 09:41:02 the sysvinit on the image, builded with the prebuild is working perfect. With the glibc it won't start. When i remove the /sbin/init symlink, i get a kernel panic, obviously. So it starts, but it doesnt start correctly, i pressume Oct 16 09:52:40 bbl. Lunchtime! Oct 16 09:56:18 03koen 07org.oe.dev * rb363c343... 10/ (1 packages/linux/linux.inc): linux.inc: try harder to find a kernel image in arch//* Oct 16 09:56:25 03koen 07org.oe.dev * rc76a936d... 10/ (6 files in 5 dirs): linux-bfin: add kernel from blackfin svn, currently support adzs-bf548-ezlite Oct 16 09:56:32 03koen 07org.oe.dev * rce2d9b99... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: add blackfin stuff Oct 16 09:56:48 03koen 07org.oe.dev * rd4019fa9... 10/ (1 packages/uclibc/elf2flt_svn.bb): elf2flt: update to match cross dir updates Oct 16 09:56:55 03koen 07org.oe.dev * r5c1eabef... 10/ (4 files in 2 dirs): blackfin machine: use linux-bfin Oct 16 10:10:03 good morning Oct 16 10:10:13 yo florian Oct 16 10:14:12 hi, i got a question regarding MACHINE=wrt54oe Oct 16 10:14:35 -oe ... does this actually build for task-base? Oct 16 10:15:08 well it first fails on uclibc 0.9.29 (i use 0.9.28 now) and now it tells me Unable to determine version for kernel headers Oct 16 10:24:47 nik0n : I think its broken, no one has build wrt54oe for a long time Oct 16 10:32:24 ok Oct 16 10:32:33 gonne have a look at it Oct 16 10:33:55 nik0n : The message you get means that the linux-header version you define is "younger" than what it expects to find Oct 16 10:34:09 for the uclibc version it builds Oct 16 10:36:02 nik0n: wrt54 support in OE is not finished. someone added it years ago and moved to other targets Oct 16 10:36:44 nik0n: using MACHINE=wrt54 DISTRO=wrt54oe you can build openwrt compatible binaries (atleast openwrt 0.9 can use it) Oct 16 10:38:18 Back. Oct 16 10:39:59 koen: consider my next commit and ship it with Angstrom :-D Oct 16 10:44:01 03xora 07org.oe.dev * r0f6db70a... 10/ (3 files in 3 dirs): Oct 16 10:44:01 packages/openssh/openssh_4.6p1.bb : package sftp-server seperately from Oct 16 10:44:01 sftp and make it not depend on anything from openssl. Oct 16 10:44:01 This will allow images to ship it as default without pulling in OpenSSL Oct 16 10:44:23 XorA: cute Oct 16 10:44:40 koen: simple little hack :-D Oct 16 10:48:40 XorA: what do you think about bumping openmoko-panel-gsm to 3185? Oct 16 10:49:23 koen: is it working? Oct 16 10:49:30 koen: Im currently in middle of something Oct 16 10:49:44 haven't tried it yet Oct 16 10:49:49 * koen fires of a build to test Oct 16 10:50:01 * hrw do not touch openmoko panels Oct 16 10:50:15 they were mostly mockups when last tried Oct 16 10:50:25 openmoko-panel-gsm - 0.1.0+svnr3185-r0 on my GTA02 Oct 16 10:50:36 hrw: ? Oct 16 10:50:37 but no SIM in there :-) Oct 16 10:50:49 hrw: But if there go fixes in... Oct 16 10:51:14 hrw: IIRC thos had a look at the disappearing panel problem today Oct 16 10:51:21 cool Oct 16 10:51:27 hrw: Not sure about the status thought Oct 16 10:51:43 it is one of my bugs (with me as reporter I mean) Oct 16 10:51:51 stefan_schmidt: http://scap.linuxtogo.org/ Oct 16 10:51:54 hrw: I know that we have many non-functional applets, but we still should take care that fixes go in. Oct 16 10:52:01 stefan_schmidt: you fixed qvga on the lcm? Oct 16 10:52:10 koen: mostly. :) Oct 16 10:52:33 koen: Still some hardcoded stuff. I'm still working on. Oct 16 10:52:42 stefan_schmidt: I know - I do testing of apps and bump them from time to time in OE Oct 16 10:53:33 * koen notices gsmd svn is still as obsolete as last week Oct 16 10:53:41 hrw: ok, I just saw your * hrw do not touch openmoko panels Oct 16 10:56:05 Henryk: a long time ago the idea was to make autotools stuff do staging by default, but there was some opposition, so that's why it's opt-in Oct 16 10:56:18 koen: ah, ok Oct 16 10:57:19 is it possible to run a specific task (like rm_work, devshell) from bitbake -i? Oct 16 11:01:03 So.... Uhm. Where are the bottlenecks if i go from a precompiled uclibc toolchain to the standard OE glibc 2.5 toolchain? Oct 16 11:05:36 * koen hints openmoko folks at the existence of ;maxrev for patches to svn versions Oct 16 11:06:03 maxrev_ Oct 16 11:06:04 ? Oct 16 11:06:41 if a patch has been applied upstream, you don't remove it from SRC_URI, but add ;maxrev= to SRC_URI Oct 16 11:06:55 Shoragan: minrev/maxrev, mindate/maxdate - way of telling 'this patch applies on svnrev minrev, was applied upstream on maxrev' Oct 16 11:07:31 koen, hrw: regarding om-panel-gsm I think it's pretty stable. Been playing around with it for some time now. Oct 16 11:07:32 I just looked thru the commit list and the last 5 people that removed patches were openmoko people :) Oct 16 11:07:37 and then it 'disappears' when the svnrev is highere? Oct 16 11:08:05 Shoragan: yes, and you get a 'patch .patch' is outdated Oct 16 11:08:06 alphaone: applet maybe is stable because it is very simple. but it use mokogsmd which has to use gsmd... Oct 16 11:08:08 message Oct 16 11:08:16 koen: Ah, I knew that once. But it seems I forgot :-( Oct 16 11:08:33 hrw: mickey has a patch that stops mokogsmd from killing gsmd Oct 16 11:08:35 alphaone: and gsmd crashes so often that mokogsmdconnection restarts it eac time when it cannot connect to gsmd Oct 16 11:08:48 alphaone, we all use the line :) Oct 16 11:08:49 hrw: "Now gsmd only gets killed once on the neo!" Oct 16 11:09:23 This guy is going to gitmo Oct 16 11:09:24 http://www.mercurynews.com/ci_7162255?source=most_viewed&nclick_check=1 Oct 16 11:09:27 koen: instead of 'now gsmd does not need to be killed to be running' ;( Oct 16 11:09:41 right Oct 16 11:09:54 Crofton|home: need login Oct 16 11:09:55 koen, srcrev currently only supports one repository? Oct 16 11:10:03 hmmm Oct 16 11:10:06 hrw: Patches for that feature are welcome ;-) Oct 16 11:10:10 * hrw wonder when gsmd will became stable Oct 16 11:10:11 maybe I have a cookie Oct 16 11:10:19 Shoragan: it support more repositories with SRCREV_FORMAT Oct 16 11:10:24 alphaone: patches or rewrite? Oct 16 11:10:30 Shoragan: but RPs change to bitbake.conf broke it Oct 16 11:10:40 Shoragan: see the webkit-gtk recipe Oct 16 11:10:41 in summary aerial-reconnaissance enthusiast crashes small RC plane on national lab property :) Oct 16 11:10:57 koen, and how does that interact with maxrev? Oct 16 11:11:15 hrw: The question basically is if patches will ever be able to make gsmd stable Oct 16 11:11:40 Shoragan: I have no idea, that's a question for RP :) Oct 16 11:11:49 ah, http://www.bugmenot.com/view/mercurynews.com Oct 16 11:11:53 ah Oct 16 11:13:13 http://diydrones.com/profiles/blog/show?id=705844:BlogPost:5904 Oct 16 11:15:54 Shoragan: It reacts badly with maxrev :( Oct 16 11:16:03 :( Oct 16 11:16:08 koen: How did I break SRCREV_FORMAT? :/ Oct 16 11:16:25 SRCREV=1 in bitbake.conf Oct 16 11:16:47 we probably need a more flexible format for specifying patches and sources Oct 16 11:16:49 after that your can't lock down a SRCREV_FORMAT package Oct 16 11:17:19 RP: btw, SRC_REV=1 clashes with the no-distro goal as well Oct 16 11:18:26 koen: What are you proposing then? :) Oct 16 11:18:41 what I've been saying for years Oct 16 11:18:46 koen: not really no recipe should be pulling from floating svn for no distro Oct 16 11:18:47 create snapshot recipes Oct 16 11:19:00 koen: ahah snao Oct 16 11:19:03 snap Oct 16 11:19:28 koen: We shouldn't need them... Oct 16 11:19:34 so put SRCREV= inside the recipe itself or create a foo_1726.bb next to foo_svn.bb Oct 16 11:19:51 koen: A total waste of space... Oct 16 11:20:39 right now only angstrom-derived distros work in OE, because they include sane-srcrevs.inc, every other distro tries to fetch rev 1 Oct 16 11:21:14 koen: So we should be including one of the files by default and using ?= in that file Oct 16 11:21:34 that's the wrong way around Oct 16 11:22:36 We should be using ?= in that file and including it by default then? :) Oct 16 11:23:03 heh Oct 16 11:23:04 no Oct 16 11:23:15 we shouldn't be using sane-* *at all* Oct 16 11:23:49 and certainly not in a way that mandates is (SRCREV=1 in bitbake.conf) Oct 16 11:23:51 BTW linux.inc and kernel.bbclass seem to class, if you have KERNEL_IMAGETYPE="uImage" kernel build seems to generate a zImage, but kernel.bbclass tries to install a uImage Oct 16 11:24:31 XorA: That suggests the uboot image creation tool is missing Oct 16 11:24:50 koen: So you're suggesting we remove SRCREV=1 from bitbake.conf? Oct 16 11:25:22 RP: revert it to get_srcrev(d) thingy as a short term solution Oct 16 11:25:50 and move this discussion to the ml for medium and long term solutions Oct 16 11:26:00 koen: and what will angstrom default to? Oct 16 11:26:15 RP: there is a install_prepend in linux.inc that goes about linking stuff to fake it up Oct 16 11:26:21 I have no idea what angstrom will default to Oct 16 11:26:35 koen: I think this needs to be discussed on the mailing list Oct 16 11:26:48 RP: doesnt look clean to me Oct 16 11:27:59 XorA: It sounds messy :/ Oct 16 11:28:53 * XorA is hungry, can look again later Oct 16 11:40:03 XorA: the update the panel-gsm stops matchbox-panel2 from crashing over here :) Oct 16 11:42:50 Uhr. Sorry i quitted... Had some network problems or whatshowever. Oct 16 11:45:14 XorA: sorry, spoke too soon Oct 16 11:45:55 03koen 07org.oe.dev * rcdb9c74e... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: update linux-bfin Oct 16 11:46:01 03koen 07org.oe.dev * r03afc682... 10/ (1 packages/linux/linux-bfin/adzs-bf548-ezlite/defconfig): linux-bfin: re-enable touchscreen since r3731 fixed it Oct 16 11:46:09 koen: no applets again? Oct 16 11:46:26 guess Oct 16 11:46:36 gsmd got restarted and mb=panel2 collapsed Oct 16 11:48:14 But euh, sorry to nag, anyone had an idea for my problem? Oct 16 11:52:03 koen: ok, so normal bug Oct 16 11:52:19 Reneger: you will get bigger rootfs Oct 16 11:53:35 I'm using a 128 MB flash disk. So that's not a problem Oct 16 11:57:47 Reneger : Use your existing toolchain and build a glibc OE based fs to test Oct 16 11:58:17 Reneger : The compare the 2 sysvinit binaries and check for differences Oct 16 11:59:08 i have done some mipsel builds or a routerboard 5xx but that was several months back Oct 16 11:59:30 Mipsel aint a very common CPU... Oct 16 11:59:43 and months in OE timeframe is a very long time :) Oct 16 11:59:57 hehehe Oct 16 12:00:22 is you machine mips or mipsel ? Oct 16 12:00:48 mipsel Oct 16 12:00:52 afaik Oct 16 12:01:04 Crofton|home: how's the chumby? Oct 16 12:01:05 It's a secure processor. Oct 16 12:01:12 boring Oct 16 12:01:16 hmm wait i think i have an fs build for the routerboard maybe you can test that Oct 16 12:01:26 hopefully the next sw update makes it more exciting Oct 16 12:01:32 :) Oct 16 12:01:33 it really needs to stream web radio Oct 16 12:01:34 steliosk: comparing binaries between uclibc and glibc doesn't help Oct 16 12:01:44 so it can be a clock radio Oct 16 12:02:09 current release uses 2.6.19 I think Oct 16 12:02:34 If I completely tire off it, I replace sw with angstrom :) Oct 16 12:03:44 * likewise_ prepares for toolchain PV bumpage. Oct 16 12:03:58 koen : Well if debug info is there you can see which function fcks Oct 16 12:08:18 Reneger : Try this fs -> http://www.digital-opsis.com/owmnr/images/test/rb500/owmnr-standard-rb500_200604261021.rootfs.tar.bz2 Oct 16 12:08:54 thnx! I'll keep u informed! Time for coffee and a smoke now ;) Oct 16 12:09:31 anyone know the diference betwen microSD and microSD trans flash? Oct 16 12:09:38 Reneger : Its a wifi router image -> http://owmnr.digital-opsis.com for details Oct 16 12:09:47 Crofton|home: only name Oct 16 12:09:49 besides $10 (7 euro) Oct 16 12:09:57 Crofton|home: its same card format/size Oct 16 12:10:03 good Oct 16 12:10:13 Crofton|home: transflash is old name Oct 16 12:10:15 transflash is the old name Oct 16 12:10:16 * steliosk thinks that Crofton has become very good converting $->E Oct 16 12:10:19 they had both and hte TransFlash was on sale Oct 16 12:10:28 stelios, google :) Oct 16 12:10:56 you would think they would put the old name on sale .... Oct 16 12:13:25 "On Friday, Mickey mentioned to me that libmokogsmd would probably not be Oct 16 12:13:26 used in the final software stack. Instead, he would prefer a d-bus based Oct 16 12:13:26 interface to GSMD" Oct 16 12:13:33 and there was much rejoicing! Oct 16 12:13:35 yay! Oct 16 12:14:23 heh Oct 16 12:14:26 PhoneKit Oct 16 12:14:56 koen: where it is? Oct 16 12:15:06 om.framework Oct 16 12:15:27 that one I do not read Oct 16 12:18:03 hrw: http://dominion.kabel.utwente.nl/koen/cms/unbreaking-openmoko Oct 16 12:20:02 steliosk: Thnx! Oct 16 12:20:43 koen: hehe Oct 16 12:21:21 florian: maybe phoneserver could hook into PhoneKit Oct 16 12:21:44 to get lips -> dbus -> Oct 16 12:22:41 koen: true, in fact it would be better if PhoneKit just exposed the lips api Oct 16 12:23:04 * * OE Bug 3160 has been created by autobuild(AT)openembedded.org Oct 16 12:23:06 * * matchbox-panel-2-0.1+svnr1626-r8-do_configure Oct 16 12:23:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3160 Oct 16 12:23:39 hmm... which change did I miss if gconf-dbus always tries to build rev. 1 from svn? Oct 16 12:24:14 * hrw finally booted 2.6.23 on c7x0 Oct 16 12:33:54 ~lart 2.6.23 anyway Oct 16 12:33:54 * ibot puts 2.6.23 anyway through a wood chipper Oct 16 12:34:08 it forced me to remove pcmcia from kernel.. Oct 16 12:35:30 I know an unemlpoyed tech writer Oct 16 12:35:51 I'd like to ask her to look at some of our documentation Oct 16 12:36:00 What would be a good starting point? Oct 16 12:38:20 we only have one canonical place for docs: org.openembedded.documentation Oct 16 12:38:29 hi Oct 16 12:38:43 first problem, she is a windoze luser :) Oct 16 12:39:10 I am trying to suck her in to doing some work while she finds another job Oct 16 12:39:28 explaining that contributing to open source makes you more valuable Oct 16 12:40:00 however explaining mtn to this sort of person is , shall we say challenging :) Oct 16 12:40:15 I'll check out that branch and see what I can do Oct 16 12:40:27 Crofton: http://www.openembedded.org/~koen/usermanual.pdf Oct 16 12:40:36 Crofton: the result of 'make pdf' Oct 16 12:40:38 hmm Oct 16 12:40:42 I'll email that Oct 16 12:41:26 she was doing contract work for meat packing plant ... Oct 16 12:45:12 steliosk: The archive you sended works partially. It starts, i see the init: entry, only it crashes at the syslogd, because the /proc isnt mounted correctly, but thats another problem. Oct 16 12:47:58 Hopefully she can do some editing for us Oct 16 12:48:33 crofton she should learn programming Oct 16 12:48:45 we should learn to write :) Oct 16 12:49:02 I'm not sure which would be easier :) Oct 16 12:51:59 steliosk: The only negative point of this image, it is using uclibc, if i'm correct.... :| Oct 16 12:52:25 interesting, the KDE-ev has "Patrons" Oct 16 12:53:14 €10k per year, right> Oct 16 12:54:21 I only saw the concept in a report :) Oct 16 12:56:21 I am assuming we will read the KDE-ev documents :) Oct 16 13:03:27 ow to gain 30GB of disk space? Oct 16 13:03:33 remove few builddirs Oct 16 13:04:25 hrw *g* Oct 16 13:13:21 woglinde: I am dropping one of my oldest builddir now - hope to get lot of space from it Oct 16 13:18:51 i mickeyl Oct 16 13:19:32 That was a long dinner :) Oct 16 13:19:35 hi mickeyl Oct 16 13:19:54 hey mickeyl Oct 16 13:20:05 hi guys Oct 16 13:20:17 RP: indeed. i tend to sleep long after dinnfer ;) Oct 16 13:23:41 koen: angstrom-2008.1 includes angstrom-2007.1 which includes sane-srcrevs which sets all necessary svn revisions, right? Oct 16 13:24:42 florian: which sets most svn revisions, yes Oct 16 13:27:48 koen: ok... this does not seem to happen here, I bet its a typo somewhere :-/ Oct 16 13:35:34 hmm Oct 16 13:35:40 meta-toolchain still doesn't love me Oct 16 13:36:19 meta-toolchain does not like me either :-( Oct 16 13:36:41 03xora 07org.oe.dev * r699a123d... 10/ (1 packages/openmoko2/openmoko-appmanager2_svn.bb): packages/openmoko2/openmoko-appmanager2_svn.bb : fix packaging Oct 16 13:36:46 03xora 07org.oe.dev * r9f4762fe... 10/ (1 packages/tasks/task-openmoko.bb): Oct 16 13:36:46 packages/tasks/task-openmoko.bb : add appmanager2 and browser2 to images Oct 16 13:36:46 for testing Oct 16 13:36:53 03koen 07org.oe.dev * r0cf8f2c5... 10/ (1 conf/distro/include/sane-srcrevs.inc): sanesrcrevs: bump some openmoko stuff, tested on a780 and gta01 Oct 16 13:37:03 ok, so the first thing is it bails out on accessing pkgconfig Oct 16 13:37:06 | /home/pkg/oe/spitz/tmp/work/i686-armv4t-sdk-angstrom-linux-gnueabi/meta-toolchain-openmoko-1.0-r0/temp/run.do_populate_sdk.16182: line 453: cd: /home/pkg/oe/spitz/tmp/work/i686-armv4t-sdk-angstrom-linux-gnueabi/meta-toolchain-openmoko-1.0-r0/sdk/image//usr/local/openmoko/arm/arm-angstrom-linux-gnueabi/lib/pkgconfig: No such file or directory Oct 16 13:37:18 i'll add a safety check for that Oct 16 13:38:07 mickeyl whats meta-toolchain for? Oct 16 13:38:36 to generate an external toolchain that you can use with or without OE Oct 16 13:39:08 So I can get a grad-student to create a program on a gumstix to use xmmx Oct 16 13:39:15 without having to explain OE :) Oct 16 13:39:25 er wireless-mmx Oct 16 13:40:11 Crofton: iwmmxt? Oct 16 13:40:16 yeah Oct 16 13:40:28 it is "interesting" from a signal processing point of view Oct 16 13:42:39 mickeyl are okay Oct 16 13:43:13 NOTE: package meta-toolchain-openmoko-1.0: completed Oct 16 13:43:18 ok, that's better for a start Oct 16 13:43:18 ???? Oct 16 13:43:21 * mickeyl pushes the fix Oct 16 13:43:21 * florian is cinfused Oct 16 13:43:26 hehe Oct 16 13:43:55 woglinde: read my blog :) Oct 16 13:44:03 the section about why external toolchains are good Oct 16 13:44:16 mickeyl I did Oct 16 13:44:17 and why stopping some people from using OE is good as well :D Oct 16 13:44:20 koen: So if i have things like SRCREV_pn-gconf-dbus="606" in my environment the setup is correct, right? Oct 16 13:44:38 florian: looks like it Oct 16 13:44:43 so, did anyone send the 'let's remove *sdk*' proposal yet? Oct 16 13:44:57 Could anyone point me at a recipie for getting an OE built image booting on a mainstone board? Oct 16 13:45:33 Anyone with a spare mainstone around for testing? ;) Oct 16 13:45:53 broonie: MACHINE="mainstone" DISTRO="angstrom-2007.1" iirc Oct 16 13:46:29 * florian checks if his bitbake is up to date Oct 16 13:47:23 hrw: Yeah, I'm using angstrom-2008.1 but I'd expect that to work also... Oct 16 13:48:05 AFAICT the main problem I've got is that the kernel images I'm getting are noticably bigger than blob wants to let me use. Oct 16 13:48:39 So long as I'm not barking up completely the wrong tree... Oct 16 13:49:40 mickeyl: I don't think so... Oct 16 13:50:36 koen: looks like it does not work with latest bitbake 1.8 Oct 16 13:50:45 broonie: are your kernel images about 4MB? Oct 16 13:51:00 florian: I'm running 1.8 head, so it works Oct 16 13:51:47 cbrake: zImage comes out at 1.2M, Image at 2.5M Oct 16 13:52:37 mickeyl: any chance that your fix will make meta-toolchain work for other distros than openmoko? Oct 16 13:52:39 broonie: ok, don't know then. There was a binutils problem that makes images 4MB Oct 16 13:52:48 koen: ok, same here... then it starts to become intersting :-) Oct 16 13:52:54 broonie: liam is supposed to maintain that :-D Oct 16 13:53:23 * cbrake has a mainstone, but have not used it in a long time ... Oct 16 13:53:26 Esben: depends on what your problems are. Oct 16 13:54:36 http://blog.haerwu.biz/2007/10/16/floppies/ Oct 16 13:55:24 XorA: Heh. Oct 16 13:56:07 * florian still has some floppy drives :-) Oct 16 13:56:16 * woglinde too Oct 16 13:56:24 florian: I have one in drawer but no idea does it work Oct 16 13:56:32 * broonie suspects this is a bitrot issue. Oct 16 13:56:39 but the computer is now virtulized Oct 16 13:56:50 broonie: ask Liam about his key, then laugh manaically Oct 16 13:57:46 He already mentioned that. Oct 16 13:58:26 wasn't I supposed to create a 'liam-is-a-git-and-lost-his-key@openembedded.org' for him? Oct 16 13:59:11 koen: liam-says-england-rules@openmoko.org would be funnier Oct 16 14:00:00 aliases are cheap ;) Oct 16 14:00:17 * koen writes some more ugly php code for feed-browser Oct 16 14:00:31 liam-I-lost-key-again-`date +%s`@openembedded.org? Oct 16 14:00:38 mickeyl: I don't remember exactly what was broken here, I''ll try again . Oct 16 14:00:39 koen: you ugly! Oct 16 14:00:43 lol Oct 16 14:01:00 Esben: right, do that Oct 16 14:02:04 * koen prepares a funeral pyre for libmokogsmd Oct 16 14:02:58 03mickeyl 07org.oe.dev * r91b1890c... 10/ (7 files in 3 dirs): xserver-kdrive-common: install xsplash ppms for branding Oct 16 14:02:59 morning Oct 16 14:03:04 03mickeyl 07org.oe.dev * re2ab58c9... 10/ (1 packages/meta/meta-toolchain.bb): meta-toolchain: only touch pkgconfig files if they're actually present Oct 16 14:03:05 hi chouimat Oct 16 14:05:56 * lrg ears are burning Oct 16 14:06:30 broonie: best to use git HEAD for mainstone kernel atm Oct 16 14:15:04 * * OE Bug 3160 has been RESOLVED (FIXED) by Oct 16 14:15:06 * *  matchbox-panel-2-0.1+svnr1626-r8-do_configure Oct 16 14:15:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3160 Oct 16 14:15:33 lrg: I've been trying with the asoc dev branch, shouldn't have diverged too far I'd have thought? Oct 16 14:16:31 broonie: asoc dev branch is never merged with mainline unless you do it manually Oct 16 14:18:08 XorA: That looks to have happened recently enough. Oct 16 14:19:59 Yup, similar sizes. Oct 16 14:22:42 broonie: main branch works on mainstone. I'll merge today if I get a chance Oct 16 14:23:03 koen: that fire may be premature... Oct 16 14:23:25 libmokogsmd is not the problem Oct 16 14:24:03 it isn't the solution either Oct 16 14:24:14 (to use some sofistry) Oct 16 14:24:30 hmm Oct 16 14:24:55 * mickeyl swallows angry replies Oct 16 14:25:07 but with PhoneKit in place, one can swap the gsm implementations on the fly Oct 16 14:25:37 heh, right. can't see you using the QDexmuer Oct 16 14:25:43 or whatever that's been called Oct 16 14:25:52 s/see/wait to see/ Oct 16 14:30:28 * mwester brings several buckets of gasoline to the funeral pyre koen is preparing... Oct 16 14:31:06 *sigh* Oct 16 14:37:22 03mickeyl 07org.oe.dev * r951d4751... 10/ (1 packages/openmoko2/openmoko-common2_svn.bb): openmoko-common2: remove pixmaps that ought to be part of openmoko-Xserver from here Oct 16 14:38:38 heh Oct 16 14:38:40 networking/ifupdown.c:(.text.doit+0x32): undefined reference to `_BUG_fork_is_unavailable_on_nommu' Oct 16 14:40:27 koen uh Oct 16 15:02:02 russf, ping Oct 16 15:02:35 any news after getting u-boot to build? Oct 16 15:33:30 mtn: 8f416f1cd7eff65f0a75eb6db80cd5f7291ad959 xora@openembedded.org 2007-10-16T15:18:13 Oct 16 15:33:30 mtn: 951d47515dbbef9738cdf3e9159a0e58597f3608 mickeyl@openembedded.org 2007-10-16T14:03:00 Oct 16 15:33:30 which is is the main one ? Oct 16 15:36:07 rob wait 5 minutes then automerger did his work Oct 16 15:39:02 bye Oct 16 15:49:22 Crofton|home: Yes. u-boot worked fine. Exactly the code on the pastebin page. Oct 16 15:49:43 I have not tested write to NAND yet, but everything else has worked fine. Oct 16 15:49:49 how about the kernel? Oct 16 15:50:45 I had to leave early this AM, and I'm just about ready to test again. Will give feedback in 30 mins. What's best process to get that patch upstream? Oct 16 15:50:52 ok Oct 16 15:51:02 send it to the uboot list Oct 16 15:51:20 maybe you shouldl send it to me and I can put it in git and make certain the format is ok Oct 16 15:51:26 Wolfgang is realy picky Oct 16 15:51:41 I have a git u-boot checkout already Oct 16 15:52:24 03xora 07org.oe.dev * r7e856ff4... 10/ (3 files in 3 dirs): uboot/files/makefile-no-dirafter.patch : this patch is more general Oct 16 15:52:36 03xora 07org.oe.dev * r8f416f1c... 10/ (2 files in 2 dirs): packages/uboot/uboot-openmoko_1.2.0+gitf34024d4a328e6edd906456da98d2c537155c4f7+svn2943.bb : add patch to stop pulling in files from host by accident Oct 16 15:52:45 03mwester 07org.oe.dev * rafda8aa1... 10/ (1 packages/meta/slugos-packages.bb): slugos-packages: Added curl and cups Oct 16 15:52:57 * Crofton heads for lunch Oct 16 15:52:59 03daniel 07org.oe.dev * rd63819fa... 10/ (1 conf/distro/openmoko.conf): conf/distro/openmoko.conf: Remove month name from DISTRO_VERSION Oct 16 15:53:10 03xora 07org.oe.dev * r1d2da7c2... 10/ (1 packages/uicmoc/uicmoc4-native_4.3.1.bb): Oct 16 15:53:10 packages/uicmoc/uicmoc4-native_4.3.1.bb : fix problem some people are Oct 16 15:53:10 seeing building this. Fix from #747 on bugs.om.org Oct 16 15:53:23 03mickeyl 07org.oe.dev * rd9316b24... 10/ (1 packages/openmoko2/openmoko-session2.bb): openmoko-session: set pulsesink and pulsesrc as gstreamer default output Oct 16 15:53:30 03mickeyl 07org.oe.dev * r085ed827... 10/ (1 packages/openmoko2/openmoko-sound-system2_0.1.0.bb): openmoko-sound-system: depend on gst-plugin-pulse Oct 16 15:53:37 03mickeyl 07org.oe.dev * rae6ce280... 10/ (1 packages/openmoko2/openmoko-mediaplayer2_svn.bb): openmoko-mediaplayer2: add dependency on curl Oct 16 15:53:50 Crofton|home: I'd love to you look at it and check the format in git. And feel free to pass it on. Oct 16 15:53:55 one mo... Oct 16 15:54:54 Crofton|home: ignore the yellow lines in this: http://pastebin.com/m1c7764e5 Oct 16 15:55:36 crofton, these are diff -u diffs. Copy them from the input field below - not the processed one at the top. Oct 16 15:56:11 Crofton|home: or send email address for the files themselves. russf at topia.com Oct 16 16:11:28 Crofton: Couple of messages above tagged Crofton home - you may have missed. Oct 16 16:38:58 how about qemu `s gcc4.x dislikeness .. is it kinda save to go for --disable-gcc-check Oct 16 16:39:32 most likely not. Oct 16 16:40:03 qemu is doing strange things, like splicing up the gcc generated code and very much depends on the structure that 3.x generates Oct 16 16:42:25 crof Oct 16 16:42:34 Shoragan: how's 2.6.23 looking on gto01 and gta02? Oct 16 16:42:53 koen, good on 23 Oct 16 16:43:00 err, gta01 Oct 16 16:43:09 i don't have access to a 02 right now Oct 16 16:43:27 so you don;'t know if spi works yet after the latest changes? Oct 16 16:43:44 Crofton: Finally. Confirmed. This is not decomp output tripping over input or any such booboo. It really is a kernel problem. So I'm diving into printascii immediately to see what's going on. Oct 16 16:44:53 ok Oct 16 16:45:04 maybe compare with similar boards Oct 16 16:45:32 I'd also try talking with the vendor to see if they have any patches not in upstream u-boot Oct 16 16:45:54 Crofton: fair comment. Perhaps see what "invalid compressed format (err=2)" shoudl mean Oct 16 16:46:18 yeah, that would be a good start Oct 16 16:46:43 You think it might still be something not being setup in advance for the kernel?? Oct 16 16:47:35 I jsut received the book you recommended, so that will hand hold a bit. Aldo Dan Bovet's book from O-reilly Oct 16 16:49:13 I dont' think it's uboot. If you remember, the old u-boot worked with the pre-kernel-decomp code. And I can't imagine that the bootm would be doing any additional hardware setup jsut for that case. Oct 16 16:49:40 I'll so some checking - talk is cheap ;) Oct 16 16:57:19 koen, laforge did some changes to 2.6.22, which i ported forward, so it might work now :) Oct 16 16:57:38 russf, I remember u-boot could decompress, we also know the kernel decompress is much faster Oct 16 16:57:47 and people really want faster boot times Oct 16 17:00:39 right. But I think I'm basically done with uboot, modulo any bugs that crept in, now I'll focus on the kernel. We also want the option to boot right into kernel, without u-boot. (not my pref, actually) Oct 16 17:49:51 I get vmlinux in my image, which I don't want. Oct 16 17:49:51 I took it out of PACKAGES in my linux .bb ie. PACKAGES = "kernel kernel-base kernel-image kernel-dev" Oct 16 17:49:51 What's the best way ? Oct 16 17:59:48 dfoley: FILES_kernel-image_cm-x270 = "" is one way Oct 16 18:08:34 cbrake: this leaves me with vmlinux and no zImage (not what I needed) Oct 16 18:15:55 question: directFB will run "out of the box" on a /dev/fb0 device, right? so as long as I've got fb0 (and mouse device) it should work? any recommended first program to test it? Oct 16 18:54:04 * * OE Bug 3161 has been created by  Oct 16 18:54:06 * * h5000 lacks usbhost module Oct 16 18:54:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3161 Oct 16 18:54:17 * * OE Bug 3163 has been created by  Oct 16 18:54:19 * * h5xxx doesn't report second battery Oct 16 18:54:21 * * http://bugs.openembedded.org/show_bug.cgi?id=3163 Oct 16 18:55:04 * * OE Bug 3162 has been created by  Oct 16 18:55:06 * * Angstrom h5000 support metabug Oct 16 18:55:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3162 Oct 16 18:55:16 * * OE Bug 3164 has been created by  Oct 16 18:55:18 * * h5000 sleeves don't work Oct 16 18:55:20 * * http://bugs.openembedded.org/show_bug.cgi?id=3164 Oct 16 19:56:02 03mickeyl 07org.oe.dev * r5e37b755... 10/ (2 files in 2 dirs): xserver-kdrive-common_openmoko: fix (cosmetic) merge damage in Xserver script Oct 16 19:56:06 03todd 07org.oe.dev * r282e9d1e... 10/ (11 files in 9 dirs): ghi270: Provide initial support for Grayhill ghi270 platform. Oct 16 20:16:59 Crofton: OK. I'm done for today and at a show tomorrow. I'll track down that kernel message tomorrow night and keep you in the loop. Let me know what you are up to regarding u-boot. Best. --r. Oct 16 20:24:39 mtn: error: multiple update candidates remain after selection Oct 16 20:26:21 choose one with 'mtn update -r', mickey1 todayT19:12:31 or todd todayT19:10:58 ? Oct 16 20:28:22 re Oct 16 20:58:40 stappers: choose one, or wait some time until someone merges them Oct 16 21:03:29 stappers: merged Oct 16 21:03:37 bugger knows why automerger isnt running Oct 16 21:21:48 say what you want about the cia and gitmo, at least they use open source sw Oct 16 21:21:55 http://plone.net/sites/cia-central-intelligence-agency Oct 16 21:22:11 * Crofton|home is going through the CMS selection process at school Oct 16 21:26:27 XorA|gone: your merge was succesfull Oct 16 21:26:52 XorA|gone: your merge was succesfull. TNX Oct 16 23:32:00 just to reconfirm myself: is this ok, style- and functionality-wise? http://pastebin.ca/739220 Oct 16 23:32:54 (Goal: Mono/.Net doesn't have a clean header/object-file separation. therefore grab all .dll's from the packages and put them into staging so that they can be found by subsequent builds) **** ENDING LOGGING AT Wed Oct 17 02:59:56 2007