**** BEGIN LOGGING AT Sat Sep 15 02:59:57 2007 Sep 15 03:47:24 I amd fine eith yuo fsrg =d ouink hrt hrrss Sep 15 06:00:58 03khem 07org.oe.dev * rabaaa429... 10/ (3 files in 3 dirs): Delete the fcntl2 patch. Its committed upstream. Sep 15 07:50:39 who looks after the bluetooth recipes in OE? I'd like to change the default device name from "BlueZ (%d)" to something that includes the host name (%h) so you can distinguish multiple OE devices on in the general vicinity ... Sep 15 07:51:18 03koen 07org.oe.dev * r5ee68e0b... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: fix webkit entry, it uses SRCREV_FORMAT, which the generator script doesn't pickup Sep 15 07:51:47 rwhitby: that would be me :) Sep 15 07:51:55 any objection? Sep 15 07:52:01 rwhitby: changing the name would be a very good idea Sep 15 07:52:18 change it to "%h (%d)" perhaps? Sep 15 07:52:22 rwhitby: /etc/bluetooth/hcid.conf ? Sep 15 07:52:26 yep Sep 15 07:52:33 can't you just edit that ? Sep 15 07:52:45 rwhitby: %h would be good Sep 15 07:52:46 I'm talking about the default value for new installatins Sep 15 07:52:53 oh, OK, sorry Sep 15 07:52:58 koen: what about multiple devices on one host? Sep 15 07:53:17 (I presume that's why the %d is there) Sep 15 07:53:22 rwhitby: I also need to look at OE putting in a sane default service type (pda, workstation, phone, etc) Sep 15 07:53:31 rwhitby: I meant %h (%d) is fine :) Sep 15 07:53:37 cool Sep 15 07:54:10 koen: I'm currently working on getting an FSG-3 to do the right thing for Neo's (pand) and Treo's (dund) as a server. Sep 15 07:57:06 I would suspect that anything that OE runs could be just about anything as far service types go, so we should default to the most comprehensive set of service type that is reasonable Sep 15 07:57:15 s/runs/runs on/ Sep 15 08:00:58 rwhitby: the service type (or device type) ends up being the thing GUIs select icons for Sep 15 08:01:24 so you neo will have a phone icon in windows and kde/gnome/etc if you set the service type up correctly Sep 15 08:47:29 did somebody experience problems building libgpg-error_1.4.bb lately? Sep 15 08:48:08 I'm trying to find out if it is my OE or the recipe being broken Sep 15 08:59:42 pH5: could you do me a favour and check if libgpg-error_1.4.bb is building for you? Sep 15 09:00:51 dcordes: I built it yesterday Sep 15 09:01:14 * pH5 bitbake -crebuild -b libgpg-error_1.4.bb Sep 15 09:04:02 NOTE: package libgpg-error-1.4: completed Sep 15 09:06:15 ok I will try with the rebuild command Sep 15 09:18:36 pH5: doing -c rebuild on the package it built just fine. Why didn't it with the runqeue from angstrom-x11-image? Sep 15 09:20:42 dcordes: how would I know, I can't even look at the logs Sep 15 09:21:32 I mean the runqueue from the angstrom image build command does the same thing like calling the build of the package manually, doesn't it? Sep 15 09:23:12 morning all Sep 15 09:23:18 oops there are 400 tasks to go. Don't know what made me think it is finished soon Sep 15 09:23:23 morning RP Sep 15 09:23:36 sorry wrong channel. Sep 15 10:16:46 if I want to use bitbake to compile an autotools package which is already unpacked in a directory ... Sep 15 10:17:00 what should I set the S and SRC_URI vars to ? Sep 15 10:17:41 and is there anything else I should set for inherit autotools to work ? Sep 15 10:19:00 polz: do you want bitbake to "fetch" from that directory, or do you want to build in that directory? Sep 15 10:19:20 if the latter, SRC_URI doesn't matter because you should skip the do_fetch and do_unpack steps Sep 15 10:21:04 pH5: where can I read the instructions on how to do both ? Sep 15 10:21:36 pH5: and what's the usual method ? Fetching from a dir or just building everything where it's supposed to be unpacked ? Sep 15 10:22:50 polz: the usual method is to fetch from a scm or tarball Sep 15 10:23:10 have you read the bitbake and openembedded user manuals? Sep 15 10:23:30 I've read half of the openembedded user manual Sep 15 10:24:09 and I'm looking at www.openembeded.org/wiki/Bbfile Sep 15 10:24:37 but the only SRC_URI mentioned there are CVS and tarball Sep 15 10:29:00 polz: there is no directory tree fetcher, so if you want to build inside ${WORKDIR} and only fetch from the source directory, maybe the easiest thing to do would be to tar that dir and put the tarball in SRC_URI Sep 15 10:31:39 pH5: somehow no package is building anymore. But when I -c rebuild -b *.bb them manually it works fine. Could you take a look at it? http://www.pastebin.ca/698407 Sep 15 10:32:16 after building it manually angstrom-x11-image runqueue is continuing with the nextpackage ending up in the same error Sep 15 10:34:32 pH5: I think I'll put it in SVN. Thank you for the help. Sep 15 10:34:58 polz: good idea. Sep 15 10:36:01 dcordes: no idea. what happened before this behaviour started? Sep 15 10:36:52 perhaps clear your cache or remove the do_install stamps for packages where this happens? Sep 15 10:37:59 pH5: I don't know exactly. I think it was from the beginning I tried to build the angstrom image Sep 15 10:38:40 I built one openmoko image for akita and one for htctornado before Sep 15 10:39:15 maybe RP has an idea. the strange thing is that some of the packages seem to package w/o problem, only the install/*-locale dir is missing. Sep 15 10:40:32 pH5: I didn't see you were proposing to remove the do_install stamps for the package where it happens but it is happening vor every package Sep 15 10:41:10 dcordes: even ones that you didn't build before at all? Sep 15 10:41:14 I rebuild them manually- that works fine- then continue with bitbake angstrom-x11-image and the before errornous package is left out and it continues with the next one Sep 15 10:50:59 pH5: I think the runqueu is only building packages I din't build before Sep 15 11:02:10 RP: do you have any idea what might be the reason for the problem described? Sep 15 11:51:41 http://ilounge.com/index.php/news/comments/ipod-classic-tests-reveal-audio-problems/ Sep 15 11:51:56 wrong channel sorry Sep 15 12:00:45 dcordes, pH5: I'll bet you went over commit where the install stamps had to be removed due to the pkgmaps changes Sep 15 12:01:22 This issue was mentioned in the commit logs and the mailing list Sep 15 12:02:46 RP: thanks, I didn't notice that issue. Sep 15 12:03:16 RP: pH5 how do I get rid of it? Sep 15 12:03:18 XorA|gone, I can see why that would annoy you Sep 15 12:03:25 RP: while you're here, I was trying to have tasks remove other tasks' stamps if they remove their results (think package removing install stamps, rm_work removing all the others) Sep 15 12:03:30 RP: see http://en.pastebin.ca/698446, http://en.pastebin.ca/698447 Sep 15 12:04:11 but that didn't work because stamp_is_current always returns false if any stamp in the dependency chain is missing Sep 15 12:04:25 any idea where to go from there? Sep 15 12:05:35 I'd like to stamp_is_current return true if the stamp exists and there is no newer stamp in the dependency chain. missing dependency stamps should be ignored. Sep 15 12:05:45 pH5: what about checkign for "newer timestamp" instead of "missing || newer timestamp"? Sep 15 12:05:52 heh Sep 15 12:05:54 * koen too slow Sep 15 12:06:34 I'm trying to build an autotools package using bitbake... Sep 15 12:07:05 I've managed to get it unpacked, but now it's reporting a bunch of possibly undefined macro: errors Sep 15 12:07:31 it can't seem to find AM_PATH_GLIB and AM_PATH_GTK Sep 15 12:07:41 what should I check ? Sep 15 12:08:51 DEPENDS = "gtk+" ? Sep 15 12:09:33 tried that, to no avail. Sep 15 12:09:42 koen, RP: I was wondering if changing its behaviour for build.py might have negative effects on other users (runqueue and cooker use it, too) Sep 15 12:11:23 the error is preceded by "warning: macro `AM_PATH_GLIB` not found in library Sep 15 12:11:49 and it's also complaining about some underquoted definitions Sep 15 12:30:32 RP: http://projects.linuxtogo.org/pipermail/openembedded-devel/2007-September/002935.html is this the mention on the mailing list? Sep 15 12:33:48 pH5: is this spyrna? http://scap.linuxtogo.org/files/43091acec6b18755afc111cb45857f5c.png Sep 15 12:35:07 koen: i wish :) Sep 15 12:35:25 no, there has been no stable locking yet on individual satellites. Sep 15 12:36:03 hmm Sep 15 12:36:05 gtk 2.12.0 Sep 15 12:36:17 angstrom 2008 material :) Sep 15 12:37:37 erm... still noone had problems with gpe-mini-browser? Sep 15 12:48:43 building it? Sep 15 12:50:02 yup Sep 15 12:52:24 Darth_Wader: ERROR: function do_configure failed Sep 15 12:52:36 but it could also be my rather broken setup Sep 15 12:52:53 mine fails at do_configure too Sep 15 12:53:05 http://pastebin.ca/698497 Sep 15 13:04:27 can i make a patch dependent on the SVN rev being at least X ? Sep 15 13:21:47 likewise: ;minrev=133213421 Sep 15 13:22:03 and related: ;maxrev=34343214342 Sep 15 13:22:11 in SRC_URI Sep 15 13:28:40 koen: thanks Sep 15 13:55:16 morning Sep 15 13:56:22 03koen 07org.oe.dev * red2dc1d9... 10/ (3 files in 3 dirs): h2200 defconfig: CONFIG_TOUCHSCREEN_ADC=y Sep 15 13:56:27 03koen 07org.oe.dev * r8071f5c5... 10/ (1 packages/gnome/gconf-dbus_svn.bb): gconf-dbus: DEPENDS += virtual/libintl Sep 15 13:56:32 03pH5 07org.oe.dev * rbb4bc38e... 10/ (1 packages/glibc/glibc-package.bbclass): Sep 15 13:56:32 glibc-package.bbclass: fix do_prep_locale_tree Sep 15 13:56:32 * there are both libgcc_s.so and libgcc_s.so.${someversion} in Sep 15 13:56:32 ${CROSS_DIR}/${TARGET_SYS}/lib Sep 15 14:07:01 anyone out there today? I am running angstrom x11 image. it *should* have kdrive, right? well, when I run the mouse around on a solid color background, the contents of the framebuffer change. Sep 15 14:07:11 when I use a filesystem built by buildroot, it works fine Sep 15 14:07:16 so, I know the kernel and driver are ok. Sep 15 14:07:38 and the directory in build/tmp/work says xserver-kdrive. Sep 15 14:07:49 just wondering what other setup I may need to do with respect to the hardware Sep 15 14:08:17 i was looking for the host.def file but couldn't find it Sep 15 14:08:22 you need to add support to xserver-common and xserver-kdrive-common Sep 15 14:12:37 koen: is there a way to check in the file system to see if it was included? Sep 15 14:12:46 I was thinking it was, but maybe not Sep 15 14:13:06 check for /etc/X11/Xserver Sep 15 14:13:10 is that in the depends of the matchine? Sep 15 14:13:13 ok, let me check that Sep 15 14:13:52 definately don't have that in there. Sep 15 14:15:09 koen: ok, so do common and kdrive-common go in the image file? Sep 15 14:15:27 only one of them Sep 15 14:15:34 both provide /etc/X11/Xserver Sep 15 14:15:54 ok, so I would just want xserver-kdrive-common Sep 15 14:16:02 so, just put that in the RDEPENDS then Sep 15 14:18:09 if you're using angstrom* you'd want xserver-common Sep 15 14:20:32 koen: hmmm. ok. isn't xserver vs kdrive much larger and slower or is that a bad misconception. we only have a arm 9 Sep 15 14:21:07 xserver-(kdrive-)common are just a few text files Sep 15 14:22:06 so I guess I am wondering why I would want xserver-common vs xserver-kdrive-common when it looks like the directory is xserver-kdrive-*** Sep 15 14:22:15 I am fine doing that, just isn't clear why Sep 15 14:22:31 XSERVER ?= "xserver-kdrive-fbdev" Sep 15 14:22:40 that's the line in angstrom-x11-image Sep 15 14:22:44 so maybe that is why i am confused Sep 15 14:22:46 :) Sep 15 14:24:14 require xserver-kdrive-common.inc <--that line is that the top of the kdrive bb file Sep 15 14:33:15 03koen 07org.oe.dev * r3a47f20d... 10/ (4 files in 4 dirs): pointercal: add file for a780 Sep 15 14:42:47 so, I see the files there koen, but still, when I move the mouse around on the screen, I see some of the colors changing under the mouse. Sep 15 14:43:02 like it is interfacing incorrectly with the hardware and writing back to the wrong location. Sep 15 14:43:08 almost like an endian problem, but only for the mouse! Sep 15 14:43:22 and it doesn't happen with the tiny-x with an old buildroot. Sep 15 15:01:07 03koen 07org.oe.dev * r4131d7b7... 10/ (2 files in 2 dirs): Sep 15 15:01:07 kdrive-1.3-18bpp.patch: add, but not activate a patch to have 18bpp support in kdrive 1.3 Sep 15 15:01:07 * people with 18bpp hardware (motorola a1200, motorola rokre6, gumstix lcd) should test this and activate it if it works Sep 15 15:03:51 koen: is there a place in kdrive where it will take some preset hw descriptions vs just requesting that info from the framebuffer? Sep 15 15:04:56 christopher: yes, in the files I pointed you to an hour ago Sep 15 15:05:31 koen: sorry, I thought those looked like scripts and if I just ran Xfbdev, they would not be involved Sep 15 15:22:16 koen: ok, I see the Xsession file where it spits out the resolution and bpp. I specified that to match the hardware but no luck there. Are there any other files that you know of that could make a difference of how the 16 bit words are read and written back to the frame buffer outside of the kernel? Sep 15 15:22:55 not that I know off Sep 15 15:25:15 koen: drats. somehow this is interacting different than the other kdrive from buildroot. somehow, when the mouse moves, it must be doing a read modify write and it must get an address off by 1, or swap the bytes. I've done captures and the bytes change from 00F800F8 to F8F800F8, like it read out the F8, but wrote it back to the wrong location Sep 15 15:27:22 does anyone out there know where the mouse tracking is done? perhaps I can look there to see if anything is using some hardware defines/vars ? Sep 15 15:36:11 03koen 07org.oe.dev * r6fefb143... 10/ (3 files in 3 dirs): linux-ezx: add 18bpp support to pxafb and enable it for a1200, rokre6 owners will have to make a small change to arch/arm/mach-pxa/ezx-e6.c Sep 15 15:40:49 RP: do you know that bitbake trunk looks for classes in conf/distro/include/classes 12 times before searching machine/include/classes, build/classes and finally classes? http://en.pastebin.ca/698614 Sep 15 15:42:02 is there any chance that the ezx people will try to push pxafb upstream? Sep 15 16:16:55 can we look at mkaing linux.inc depend on uboot-utils and not the gta uboot? Sep 15 16:18:43 pH5: a very small change Sep 15 16:19:06 Crofton|home: we can, but I know the gta one is working an based on a recent uboot Sep 15 16:19:28 but we only use it for mkiamge Sep 15 16:19:45 I think the gta one falied to fetch last night and stopped my build Sep 15 16:19:48 I am annoyed :) Sep 15 16:20:35 and eventually, I would hope all teh gta stuff moves into uboot Sep 15 16:20:46 and the gta specific uboots go away Sep 15 16:21:11 I don't think it can fail to fetch Sep 15 16:21:21 since it fetches tarballs from the angstrom servers Sep 15 16:21:37 home, something croaked the build and I restarted it Sep 15 16:21:50 The console it failed on is at school Sep 15 16:22:04 I check for actual failure in a bit Sep 15 16:22:28 still, I think there are lots of reasons to use uboot-utils ... Sep 15 16:22:54 argh Sep 15 16:23:23 uboot-utils was horrendously outdated and broken at the time I made linux.inc Sep 15 16:23:25 http://rafb.net/p/KoEXKf60.html Sep 15 16:23:30 I fixed it Sep 15 16:23:34 so I still have a strong aversion against it Sep 15 16:23:45 heh Sep 15 16:23:51 I know it wasn't building etc Sep 15 16:23:56 had weirs patches Sep 15 16:24:01 take another look Sep 15 16:24:25 mtd-utils-native-1.0.0+git-r4 died on me also Sep 15 16:24:56 arg rap Sep 15 16:25:04 forget that last pasteb Sep 15 16:25:09 wrong fargin machine Sep 15 16:25:51 * Crofton|home is confused by which machine he is working on Sep 15 16:27:09 in spite of me being a complete idiot, I still think we should have linux.inc use uboot-utils for mkiamge .... Sep 15 16:27:51 hhi Sep 15 16:28:11 RDEPENDS_kernel-base = "" Sep 15 16:28:14 Crofton|home: mkimage? as long as it generate working kernels I do not care where from it came Sep 15 16:28:21 should remove uImage from teh jffs2 image right? Sep 15 16:28:49 Crofton|home: depends on method used to load kernel Sep 15 16:29:15 you can load kernel from raw flash partition, from jffs2 flash partition... Sep 15 16:29:30 on gumstix, kernel load from flash partition Sep 15 16:29:42 but it is also appearing in image Sep 15 16:29:52 I thought we went through this two weeks ago Sep 15 16:30:04 Crofton|home: raw separate flas partition? then RDEPENDS_kernel-base = "" should be ok Sep 15 16:30:25 except it is not working for me, but I swear it worked once Sep 15 16:32:09 I did not tested it Sep 15 16:35:20 I am going to nuke tmp and check on the machien I thought I had it going on Sep 15 16:35:25 hi all Sep 15 16:35:35 steliosk: hi man Sep 15 16:35:42 steliosk: how weather in greece? Sep 15 16:35:48 hrw : hot Sep 15 16:36:11 good Sep 15 16:36:19 when are you going? Sep 15 16:36:25 hrw : preparing to come down ? Sep 15 16:36:27 tuesday early morning Sep 15 16:36:34 steliosk: yes Sep 15 16:36:48 hrw : Crete ? Sep 15 16:36:55 steliosk: Zakhyntos Sep 15 16:37:10 nice Sep 15 16:37:40 steliosk: Ania is pregnant so we prefer to go somewhere where we concentrate on doing nothing instead of forcing itself during days Sep 15 16:38:08 that's called "vecations" :) Sep 15 16:38:44 so you flight directly there or through Athens ? Sep 15 16:38:57 steliosk: directly - charter flight Sep 15 16:38:58 ixp4xx hackers ? how to pull in ethernet driver bb Sep 15 16:39:20 interfaith: ixp4xx OE hackers use 2.6.2x wit opensource driver Sep 15 16:40:05 hrw : you are lucky. Tomorrow we have elections so all the commotion will be over by the time you arrive Sep 15 16:40:11 steliosk: the bad thing is that 'catherine palace' where we will stay lack webpage so no info does it have wifi Sep 15 16:40:18 steliosk: ;) Sep 15 16:40:30 hmm this is my platform however it seems no bb runs to load the ixp425_eth code Sep 15 16:40:56 interfaith: look in packages/ixp4xx? Sep 15 16:40:59 there is an ixp425 bb directory Sep 15 16:41:13 so? Sep 15 16:41:29 however its not obvious where the build process uses this Sep 15 16:41:36 hrw : My 2cnts of advice.... don't take ANY pc/pda/thing that resembles a computer with you, or your wife will make sure you will not forget it for the rest of your life Sep 15 16:41:47 steliosk: ;) Sep 15 16:41:51 ...and probably in afterlife also Sep 15 16:42:27 steliosk: we were thinking about getting 770 - its crappy enough to not want to use it and good enough to browse pages to check what to visit Sep 15 16:42:43 bitbake virtual/kernel does not seem to go near the ixp425 bb files Sep 15 16:42:51 hrw : i think you will find a few internet cafe on the islnad Sep 15 16:42:58 steliosk: last book about Greece islands which I saw had 5 pages about zakhyntos Sep 15 16:43:12 interfaith: virtual/kernel is KERNEL not external modules Sep 15 16:43:14 ah, my problem is RDEPENDS_kernel-base works for Angstrom minimal image Sep 15 16:43:20 but not for my custom image Sep 15 16:43:35 bitbak ? modules ? Sep 15 16:43:50 interfaith: build ixp425 stuff by hand and test it on device by installing it additionally (by own image for start maybe) Sep 15 16:44:07 ok..thx Sep 15 16:44:10 steliosk: how much for 1h of inet cafe is in Greece now? Sep 15 16:45:39 in athens is like 1E per hour, not sure about the islands. maybe a bit more expensive because its a tourist area Sep 15 16:45:58 affordable ;) Sep 15 16:48:50 hrw : btw wheb you get some free time could you put ipkg back in minimal-image Sep 15 16:49:10 steliosk: do it by yourself - I will be offline until oedem probably Sep 15 16:49:51 hrw : ok Sep 15 16:50:11 03koen 07org.oe.dev * re8137fb0... 10/ (4 files in 4 dirs): linux-ezx: fix a1200 and rokre6 defconfig, should fix #3009 Sep 15 16:52:23 * hrw -> out Sep 15 16:53:04 see you in the evening today or after my honeymoon or on oedem ;) Sep 15 16:53:33 is there a 'make modules' cmd or must it be bitbake -b module.bb Sep 15 16:54:03 baking a single bb file will not likely run Sep 15 16:55:26 * steliosk goes to watch Eurobasket semi-final Sep 15 16:56:41 some bitbake cmd to build ixp425_eth.ko ? Sep 15 17:14:22 interfaith: bitbake 'name-of-recipe-file-without_version.bb' -> bitbake ixp425-eth for you Sep 15 17:26:37 03pH5 07org.oe.dev * rd8e2c140... 10/ (1 conf/distro/include/preferred-xorg-versions-X11R7.3.inc): Sep 15 17:26:37 preferred-xorg-versions-X11R7.3: fix versions once and for all Sep 15 17:26:37 * the X11R7.x-a.b.c nomenclature is gone Sep 15 17:59:50 03koen 07org.oe.dev * ra3ee5d88... 10/ (4 files in 4 dirs): linux-ezx: more a1200 and rokre6 tweaks, also remove size check since boot_usb can handle 1MB images quite well nowadays and blob2 can flash like 1.3MB to mtd4 Sep 15 18:25:42 koen: do I have to do anything special to enable SRCREV_FORMAT support? Sep 15 18:26:04 for me webkit-gtk fails trying to download revision "r25582-r25571" from svn Sep 15 18:26:12 (bitbake trunk) Sep 15 18:36:37 any ixp4xx bitbaker ? ixp425_eth bitbaker.. here it errors out on missing intel files ? Sep 15 18:37:09 such as IxEthMii.h there must be another intel package needed to bake it Sep 15 18:49:50 koen, your commit for webkit broke the fetching for me Sep 15 18:50:34 NOTE: Update svn://svn.webkit.org/repository/webkit/trunk/;module=JavaScriptCore;proto=http Sep 15 18:50:36 svn: Syntax error in revision argument '25582-r25571' Sep 15 18:50:50 this is oe rev 5ee68e0ba9fef03037462ab400a3ec03e1333719 Sep 15 18:51:04 oe-commits] org.oe.dev sane-srcrevs: fix webkit entry, it uses SRCREV_FORMAT, which the generator script doesn't pickup Sep 15 19:00:20 03mickeyl 07org.oe.dev * r69932acf... 10/ (1 packages/linux/linux.inc): linux.inc: add sizecheck, add convenient support for root-nfs-over-usb Sep 15 19:00:24 03mickeyl 07org.oe.dev * rf7f91115... 10/ (1 packages/linux/linux-ezx_2.6.21.bb): linux-ezx 2.6.21: fix formatting (please STOP mixing tabs and spaces), catch up with linux.inc improvements Sep 15 19:00:29 03mickeyl 07org.oe.dev * r6b8b18ee... 10/ (5 files in 4 dirs): openmoko-sound-theme-standard: add classic ringtone, load on startup, bump PR Sep 15 19:00:33 03mickeyl 07org.oe.dev * r62bdd985... 10/ (1 packages/linux/linux-ezx_2.6.21.bb): linux-ezx 2.6.21: repair formatting again Sep 15 19:18:11 pH5: hmmm Sep 15 19:20:07 pH5: ah, fetch -f -D finally errored out Sep 15 19:20:51 oy, lart and praise caching :) Sep 15 19:23:40 ok, 'fixed' Sep 15 19:30:42 03koen 07org.oe.dev * r6cfe3b77... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: SRCREV_FORMAT is broken, so work around it by setting a global (and hence bogus for most webkit svn modules) SRCREV Sep 15 19:31:18 good morning Sep 15 19:32:00 I'm having a bit of a problem adding perl to my image Sep 15 19:32:14 distro is angstrom Sep 15 19:32:18 urg Sep 15 19:32:22 what is problem? Sep 15 19:32:31 * Crofton|home knows little about perl though Sep 15 19:32:36 I add perl to my IMAGE_INSTALL Sep 15 19:32:53 I get the 10K perl binary in /usr/bin Sep 15 19:33:09 and a /usr/lib/perl subdirectory created Sep 15 19:33:23 and then just a broken link to libperl Sep 15 19:34:18 did you look at what other perl related ipks are available Sep 15 19:34:33 Yes they all seem to be there in deploy Sep 15 19:34:35 maye broken RDEPENDS? Sep 15 19:34:48 I need to run over to school and check some things Sep 15 19:34:50 bbiab Sep 15 19:34:55 I think broken packaging Sep 15 19:35:03 the broken link shouldn't be there Sep 15 19:35:03 Should I try to explicitly include the lib-perl package? Sep 15 19:35:15 is libperl5_somethingsomething.ipk installed? Sep 15 19:35:55 checking Sep 15 19:36:43 ipkg list_installed | grep libperl Sep 15 19:37:22 still waiting for flash burn :-) Sep 15 19:37:37 Is there a way to tell from the deploy directory? Sep 15 19:37:50 Or rootfs directory? Sep 15 19:39:32 Looking in /usr/lib/ipkg in rootfs I see libperl and perl Sep 15 19:40:29 libperl5.list includes: Sep 15 19:40:36 /usr/share/perl/5.8 Sep 15 19:40:36 /usr/lib/perl/5.8 Sep 15 19:40:36 /usr/lib/libperl.so.5 Sep 15 19:40:36 /usr/lib/libperl.so.5.8.8 Sep 15 19:40:57 So by all rights it should be there in the image Sep 15 19:41:05 indeed Sep 15 19:41:08 You may be right that it is a packaging bug Sep 15 19:41:27 try comparing your ipkg against http://www.angstrom-distribution.org/unstable/feed/armv5te/base/libperl5_5.8.8-r25_armv5te.ipk Sep 15 19:41:42 (no garantee that the angstrom one is correct, just curious about differences) Sep 15 19:42:26 what is the best tool to compare the two? Sep 15 19:42:48 ar x foo.ipk ; tar zxf data.tar.gz Sep 15 19:51:17 hmm, the RDEPENDS_kernel-base trick seems to work for angstrom-minimal-image, but not my custom image Sep 15 19:51:27 03koen 07org.oe.dev * rd683d126... 10/ (1 packages/linux/linux-ezx_2.6.21.bb): remove size check again Sep 15 19:51:35 03koen 07org.oe.dev * r208a6cd5... 10/ (1 packages/linux/linux-ezx_2.6.21.bb): linux-ezx: disable backlight patch for rorkre6 as it only turns of bl at this point Sep 15 19:52:23 Crofton: works for my custom image, leaves a dangling link, but uImage is gone Sep 15 19:52:44 I suspect I did not copy quite the right stuff from minimal image Sep 15 19:52:53 koen: perhaps operator idiocy on the perl thing. stay tuned :-) Sep 15 19:53:05 I think I will require angstrom-minimal-image.bb and use extra packages to add Sep 15 19:54:00 or just include task-boot as part of your IMAGE_INSTALL Sep 15 19:54:13 I think I do already .... Sep 15 19:54:18 need to review this Sep 15 19:54:27 hmm . . . that worked for me Sep 15 20:00:22 thoroughly odd, I have the uIamge on my custom and not in minimal Sep 15 20:26:02 sometimes I feel cursed Sep 15 20:26:29 he smc91x autoload is not happening either Sep 15 20:32:24 Crofton: Bummer! Sep 15 20:32:38 Are the kernel-modules in the image? Sep 15 20:32:43 yeah Sep 15 20:32:47 well the network ones Sep 15 20:33:04 so it comes right up when you do the modprobe Sep 15 20:33:48 Did you change /etc/network/interfaces to add auto eth0 Sep 15 20:33:50 ? Sep 15 20:34:01 no Sep 15 20:34:08 Well, there you go! Sep 15 20:34:29 Do that and it should work Sep 15 20:36:54 hi, all. I have a question about packaging: I'm seeing "ERROR: QA Issue: non debug package contains .debug directory: ..." and I wonder what's the best way of automatically excluding the .debug dirs and their contents from a quite large directory hierarchy? Sep 15 20:37:08 I've been working on getting the erlang package building with the new release, R11B-5 Sep 15 20:37:41 hmm, I just had that problem Sep 15 20:37:48 where did I do it Sep 15 20:37:53 wild cards .... Sep 15 20:38:00 cxxtest? Sep 15 20:38:27 Crofton, I'm hoping the solution is pretty straightforward :-) Sep 15 20:39:11 hm, I don't have cxxtest in my packages directory Sep 15 20:39:19 yeah Sep 15 20:40:25 I was thinking cppunit, but I can find what I am looking for Sep 15 20:40:32 hmmm, the kernel tester Sep 15 20:41:27 maybe I need an explicit manifest X-( Sep 15 20:42:25 ahhh, maybe it's something like: if I put the .debug dirs explicitly into some other package, the -libs package won't redundantly pick them up? Sep 15 20:42:27 I'll try that Sep 15 20:42:46 yeah Sep 15 20:42:50 (I'll try putting them into the existing -dbg package, which explicitly mentions a few .debug dirs already and could do with a few more) Sep 15 20:42:55 there should be examples in some bb files Sep 15 20:42:58 exactly Sep 15 20:43:15 you can use wildcards Sep 15 20:43:24 * tonyg nods Sep 15 20:43:52 so far, I've found bitbake underdocumented and a bit scary, but basically pretty awesome to work with... /me makes a note to submit squeak-x.bb Sep 15 20:44:12 as in: does amazing things; but is impenetrable and a bit voodooish :) Sep 15 20:46:46 now it tries to bring the interface up, but the module is still not loaded Sep 15 20:47:01 is there a way of telling bitbake to start again from a particular phase? (e.g. do_package) Sep 15 20:47:07 while not forcing a total reconf and recompile? Sep 15 20:47:32 tonyg: bitbake -f -cpackage foo ; bitbake foo Sep 15 20:47:42 pH5, thanks Sep 15 20:47:50 warning though, package needs install results but deletes them subsequently Sep 15 20:48:01 so bitbake -f -cinstall foo ; bitbake foo might be a better idea Sep 15 20:48:15 is there a list of valid args to -c anywhere? (I've looked but am either blind, or...) Sep 15 20:48:22 -clisttasks, I think Sep 15 20:48:24 pH5, right, install's probably more what I'm after Sep 15 20:48:36 pH5, heh Sep 15 20:48:46 'cause the tasks can change depending on what classes a recipe inherits Sep 15 20:49:15 doh, should have RTFM more carefully. it's even mentioned in "bitbake -h" output Sep 15 20:56:57 Crofton: do you get any error messages? Sep 15 20:56:58 03crofton 07org.oe.dev * r973c3c04... 10/ (1 conf/machine/omap5912osk.conf): omap5912osk.conf : Prevent installation of uImage in /boot. Sep 15 20:57:02 03crofton 07org.oe.dev * r1d49e257... 10/ (1 conf/machine/gumstix-verdex.conf): gumstix-verdex.conf : Prevent installation of uImage in /boot. Sep 15 20:57:06 03crofton 07org.oe.dev * rec0908fb... 10/ (1 conf/machine/gumstix-connex.conf): gumstix-connex.conf : Prevent installation of uImage in /boot. Sep 15 20:57:21 when I turn auto on Sep 15 20:57:38 it tried to bring up the interface, but the module does not load Sep 15 20:57:42 I am moving some thigns around Sep 15 20:57:55 hi all Sep 15 20:58:05 for gumstix, I am going to install all modules by defautl Sep 15 20:58:12 flo_lap, hello Sep 15 20:58:33 Crofton: that make sense for the average gumstix user Sep 15 20:58:45 yeah Sep 15 20:58:50 oh, what caused stefan -> stefan|hospital ?! Sep 15 20:58:59 especially when you are getting started Sep 15 20:59:09 flo_lap, no idea Sep 15 20:59:52 Crofton: not sure why you are having so much networking difficulty. That part is rock solid for me Sep 15 21:01:01 thanks for your help, guys, looks like I have some candidate packages for erlang. whee! now to test them... Sep 15 21:03:40 I bet there is something foul in my image Sep 15 21:03:46 let me try minimal image ... Sep 15 21:04:27 flo_lap: Snowboarding in a fun park... Sep 15 21:04:31 flo_lap: ... broken right arm. Sep 15 21:05:09 stefan|hospital: argh... nasty! Sep 15 21:05:45 Crofton: minimal perl is working for me now Sep 15 21:05:53 excellent Sep 15 21:05:59 So I'm almost there Sep 15 21:06:03 that is the one you needed right? Sep 15 21:06:04 stefan|hospital: but a hospital with internet connection... that's a little bit of luck at least ;) Sep 15 21:06:13 was using microperl Sep 15 21:06:18 ah Sep 15 21:06:28 but the minimal perl install might be small enough Sep 15 21:06:46 how much flash are you working with? Sep 15 21:06:52 need to figure out why my OE image is so much fatter than buildroot though Sep 15 21:07:01 I see a few obvious things Sep 15 21:07:10 such as? Sep 15 21:07:52 like libstdc++ Sep 15 21:08:06 ah Sep 15 21:08:12 libiconv Sep 15 21:08:20 not sure what is pulling them in! Sep 15 21:08:59 avahi seems like it might be bigger than mdns Sep 15 21:09:01 flo_lap: yeah Sep 15 21:09:03 flo_lap: Not really. That's what mobile internet flats are for. :) Sep 15 21:09:05 flo_lap: But GPRS sucks Sep 15 21:09:08 also it segfaults :-) Sep 15 21:09:46 I work on size after I get everything functional Sep 15 21:09:57 yeah Sep 15 21:10:23 Hopefully I can just trim back the packages that don't really need to be there Sep 15 21:10:52 that should help with sucking extra libs also ... Sep 15 21:11:07 stefan|hospital: oh okay... yes, the latency must be a pain Sep 15 21:11:35 Yup, hopefully it will be easy to get rid of them Sep 15 21:11:38 my wife messed up her ACL on skis in a terrain park a few years ago .... Sep 15 21:12:49 Crofton: my flash is 16MB, but I'm trying to keep enough room free for a year's worth of historical data Sep 15 21:13:16 ah, I was wondering what was driving you Sep 15 21:13:18 flo_lap: yeah,besides that it's pretty stable. But I'm waiting for my UMTS card to arrive anyway. Sep 15 21:13:38 It's those nice history plots! Sep 15 21:14:00 And 1 year is a nice marketing claim Sep 15 21:14:30 As well as being interesting as a user Sep 15 21:15:04 stefan|hospital: i'm not sure if that would be much better... Sep 15 21:15:15 marketing is the root of all evil :) Sep 15 21:16:34 hm, spoke too soon wrt erlang packages... i wonder if there's some way of telling bitbake to automatically ignore .debug/ directories on a per-package basis, without having to construct a full manifest listing only those files I *do* want to include Sep 15 21:16:39 flo_lap: My hope would be that the 3G cell is empty. But in the end the card was ordered before the accident happened. Sep 15 21:17:10 Crofton: Yeah, but I'm the marketing person too. At least the meetings are short :-) Sep 15 21:18:26 one year looks a lot better than 11 months in the borchure Sep 15 21:20:11 stefan|hospital: mobile internet access is nice anyway, even if the performance sucks... it gives a nice feeling of freedom :) Sep 15 21:21:04 Crofton: and the annual plots would look silly with only 11 months filled in! Sep 15 21:21:18 flo_lap: I fully understand you :D Sep 15 21:21:36 stefan|hospital: i can imagine why ;) Sep 15 21:22:53 stefan|hospital: even the "poor man's solution" (http/wap only flat) i use is extremely useful... Sep 15 21:23:11 (insane.bbclass?? hmm) Sep 15 21:28:21 flo_lap: I have one SIM for the UMTS card and one with WAP flatrate for the mobile Sep 15 21:30:17 stefan|hospital: that reminds me that i have a sim for my laptop too :) Sep 15 21:31:22 hmmm, 4M of modules it appears .... Sep 15 22:05:09 03daniel 07org.oe.dev * rbba23ca6... 10/ (4 files in 4 dirs): openmoko-sound-theme-standard2: inherit openmoko2 so sources get fetched from the correct location Sep 15 22:05:42 Crofton: yikes 4M! Might want to scale back a bit on what gets built :-) Sep 15 22:05:57 yeah Sep 15 22:06:04 but we can tune that in the kconfig Sep 15 22:06:25 better not Sep 15 22:06:28 I'll worry about that when things get a little stable Sep 15 22:06:34 Yeah, I think Craig turned just about everything on figuring he could just install what he wanted Sep 15 22:06:35 koen, ? Sep 15 22:06:47 only include the modules you want runtime instead of doing "RDEPENDS = kernel-modules" Sep 15 22:06:52 I think the config could still use a review :) Sep 15 22:07:17 koen: that's the approach I am using on my builds Sep 15 22:07:22 otherwise you have reconfigure and recompile the kernel every time you need a module install of installing it from the feeds Sep 15 22:07:26 koen, the headache is, for the general case, we have now way of knowing what modules users want Sep 15 22:07:49 so we include a few and leave the to the user Sep 15 22:08:10 Craig includes all the modules required to support the gumstix hw line Sep 15 22:08:12 that may be the way to go given the quantity :) Sep 15 22:08:23 sakoman: as he should :) Sep 15 22:08:25 If you want anything else you are on your own Sep 15 22:08:32 Of course! Sep 15 22:08:33 grrr, this stinking smc91x won't autoload for me Sep 15 22:08:54 anyway... Sep 15 22:08:58 it's midnight Sep 15 22:09:00 'night all Sep 15 22:09:02 gn Sep 15 22:09:09 gn Sep 15 22:09:32 how does the autloading work? Sep 15 22:10:05 I'm not an expert at that I'm afraid Sep 15 22:12:14 sakoman, do you have easy access to a running gumtix? Sep 15 22:12:20 sure Sep 15 22:12:26 can you see if you have a file /etc/modules? Sep 15 22:12:37 yes Sep 15 22:12:42 what is in it? Sep 15 22:13:12 ### This file is automatically generated by update-modules" Sep 15 22:13:12 # Sep 15 22:13:12 # Please do not edit this file directly. If you want to change or add Sep 15 22:13:12 # anything please take a look at the files in /etc/modutils and read Sep 15 22:13:12 # the manpage for update-modules. Sep 15 22:13:13 # Sep 15 22:13:15 ### update-modules: start processing /etc/modutils/pxamci Sep 15 22:13:19 pxamci Sep 15 22:13:21 ### update-modules: end processing /etc/modutils/pxamci Sep 15 22:13:23 ### update-modules: start processing /etc/modutils/smc91x Sep 15 22:13:25 smc91x Sep 15 22:13:27 ### update-modules: end processing /etc/modutils/smc91x Sep 15 22:13:28 pastebin .... Sep 15 22:13:41 ok Sep 15 22:13:45 Sorry! Sep 15 22:13:45 I lack this file :) Sep 15 22:14:39 Do you have modules.dep in /lib/modules/2.6.21? Sep 15 22:14:40 running update-modules on the gumstix fixed it Sep 15 22:14:48 I think so Sep 15 22:14:52 make sure Sep 15 22:14:52 this .debug thing is breaking me, slowly but surely Sep 15 22:15:02 tonyg, how so? Sep 15 22:15:36 Does anyone know the name of the package that tests the kernel interfaces? Sep 15 22:15:51 ok Sep 15 22:16:08 update-modules is not happening when I build the image Sep 15 22:16:25 I wonder if this is another minimal versus console difference? Sep 15 22:17:47 Don't know -- that problem went away for me. Perhaps when I started using my own image ?? Sep 15 22:18:16 yeah, I need to look at some thing Sep 15 22:18:28 Crofton, if you take a look at the very bottom of erlang_R11B.1.bb, you'll see definitions of FILES_${PN}... I cannot find the magic incantation required to make insane.bbclass stop complaining about some of the .debug directories that end up in the "erlang-libs" package. The packages defined are "erlang", "erlang-libs", "erlang-dbg", and the doc and locales ones. "erlang-libs" looks to be the only custom one, so I'm wondering if there's some hidden magic aroun Sep 15 22:18:28 d. Sep 15 22:19:36 can you pastebin the messages? Sep 15 22:19:42 sure, will do. Sep 15 22:20:24 (... although I've already upgraded to R11B.5, so the fine detail will be slightly different - but R11B.1 fails for the same reasons) Sep 15 22:20:27 * Crofton wonders how acceptable it is to cut and paste large chunks of thesis.tex into conference paper.tex ... Sep 15 22:20:28 oh wait Sep 15 22:20:34 thats fine Sep 15 22:20:38 no it doesn't, it fails for other HIPE-related reasons. damn Sep 15 22:20:47 hhe Sep 15 22:20:49 anyway, I'll pastebin the .5 output Sep 15 22:25:28 sakoman, I think my problem revolves around only using task-boot and not task-base Sep 15 22:25:56 I think you are correct Sep 15 22:26:13 task-base has a lot of hairy looking bits :) Sep 15 22:26:39 That's why I created my own task-base-headless Sep 15 22:27:05 Also to reduce build time & image size :-) Sep 15 22:27:18 heh Sep 15 22:27:34 I think so guys are looking at breaking some of the dependency chains Sep 15 22:28:00 Just too big a jump from task-boot to task-base Sep 15 22:28:12 yeah, I kind of sense that Sep 15 22:28:42 koen's position is that it is "just" a build time cost Sep 15 22:28:46 not install Sep 15 22:28:50 yeah Sep 15 22:29:09 but *damn* that build time is a pain in the ass and a real productivity killer Sep 15 22:29:18 I understand that, but I also have seen traffic where they are making an effort to cut some of the worse build tiem offendors :) Sep 15 22:29:29 Excellent! Sep 15 22:29:54 are you on .dev list? Sep 15 22:30:02 Yes Sep 15 22:30:18 Don't get me wrong, I actually think task-base is a really good idea Sep 15 22:30:19 so you saw Marcin's comments on task-base being too big Sep 15 22:30:26 Yes Sep 15 22:31:21 http://rafb.net/p/jh1EP292.html Sep 15 22:31:28 We really need to start exercising mtn's branches and see if we can create a workflow where we have a stable release for people making product Sep 15 22:31:43 and a dev branch where we can address the systemix issues Sep 15 22:31:48 Definitely! Sep 15 22:32:21 lack of a stable branch is one of the biggest issues (for me anyway) Sep 15 22:32:53 Well, got to run. My wife and I are heading out to the Redding beer & wine festival Sep 15 22:33:01 um Sep 15 22:33:05 that sounds good Sep 15 22:33:19 Yeah, beautiful evening for it too Sep 15 22:33:35 Good luck with your modules issue Sep 15 22:33:41 thanks Sep 15 22:33:45 I think I see the answer Sep 15 22:33:55 in task-base? Sep 15 22:34:06 yeah Sep 15 22:34:12 just include task-base :) Sep 15 22:34:37 So there you go :-) Sep 15 22:34:40 yeah Sep 15 22:34:51 Go take a nap while it builds :-) Sep 15 22:35:03 work on paper Sep 15 22:35:08 Later! Sep 15 22:35:51 tonyg, something like Sep 15 22:36:56 FILES..-dbg += " ${libdir}/erlang/lib/*/*/.debug ${libdir}/erlang/lib/*/*/*/.debug " Sep 15 22:36:58 may help Sep 15 22:42:23 Crofton, already tried something like that; I'll try exactly that, and report back Sep 15 22:42:25 thanks Sep 15 22:43:18 basically, when faced with lots of .debug dirs in many directories, use */.debug */*/.debug ..... Sep 15 22:43:34 There was one package I had to fix up like that Sep 15 22:43:52 the thing that's puzzling me is that one of the reported problems is in a dir that should already have been caught by the glob in the FILES-dbg spec Sep 15 22:48:18 damn. that didn't make any difference :-( Sep 15 22:48:24 bother Sep 15 22:48:42 for now, I don't even *care* about the .debug directories... (where do they come from anyway??) could I add something to the .bb to blow them away? Sep 15 22:48:45 I wonder of the {libdir} is not quite the right one Sep 15 22:48:53 oo, maybe that's it Sep 15 23:01:02 tonyg, look up how libdir is defined Sep 15 23:18:09 any ixp4xx bitbaker ? Sep 15 23:43:31 OMG. It was an order-dependent problem, I think. Putting the -dbg package *first* in PACKAGES means it takes ownership of all the .debug dirs ... somehow Sep 15 23:43:41 wow Sep 15 23:44:15 the packages are built, and even look sane Sep 15 23:44:23 *sigh* Sep 15 23:44:28 well, cool. now to test them Sep 15 23:46:11 03mickeyl 07org.oe.dev * rd23a37e3... 10/ (5 files in 3 dirs): linux-openmoko-devel: refactor to use linux.inc Sep 16 00:21:30 03pfalcon 07org.oe.dev * rc7197314... 10/ (1 packages/linux/linux-handhelds-2.6/hx4700/defconfig): Sep 16 00:21:30 linux-handhelds-2.6: Update hx4700 defconfig from defconfigman r395: Sep 16 00:21:30 * Enable PCMCIA Ether, USB Ether, USB Serial drivers. Sep 16 00:21:37 03pfalcon 07org.oe.dev * rd04d77b9... 10/ (24 files in 24 dirs): Sep 16 00:21:37 linux-handhelds-2.6: Enable binfmt misc and ipv6 consistently for all machines. Sep 16 00:21:37 * defconfigman r397 Sep 16 00:21:48 03pfalcon 07org.oe.dev * rbea2ad8e... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs.inc: Add linux-hacknev-2.6 rev. Sep 16 00:21:54 03pfalcon 07org.oe.dev * r06f3043e... 10/ (1 packages/linux/linux-handhelds-2.6/hx4700/defconfig): Sep 16 00:21:54 linux-handhelds-2.6: hx4700: Enabled USB Printer/Modem drivers. Sep 16 00:21:54 * defconfigman r398. Sep 16 00:22:00 03pfalcon 07org.oe.dev * r790d7127... 10/ (24 files in 24 dirs): Sep 16 00:22:00 linux-handhelds-2.6: Enable TUN and, where applicable, PPPoE. Sep 16 00:22:00 * defconfigman r400. Sep 16 00:22:05 03pfalcon 07org.oe.dev * ra91b0a05... 10/ (1 packages/linux/linux-handhelds-2.6/hx4700/defconfig): linux-handhelds-2.6: hx4700: Enable USB WiFi drivers. Sep 16 00:22:09 03pfalcon 07org.oe.dev * rff6c60fd... 10/ (1 packages/linux/linux-handhelds-2.6/hx4700/defconfig): Sep 16 00:22:09 linux-handhelds-2.6: hx4700: Dirty magic to enable AX8817X USB Ether driver. Sep 16 00:22:09 * It manages to depend on CONFIG_NET_ETHERNET even though according to Kconfig Sep 16 00:22:10 it's aux switch symbol only. Sep 16 00:22:12 * defconfigman r404. Sep 16 00:22:14 03pfalcon 07org.oe.dev * r62ef8bc0... 10/ (23 files in 23 dirs): Sep 16 00:22:16 linux-handhelds-2.6: Enable NET_WIRELESS_RTNETLINK. Sep 16 00:22:18 * defconfigman r404. Sep 16 00:22:20 03pfalcon 07org.oe.dev * r3c1efe69... 10/ (1 packages/linux/linux-handhelds-2.6/hx4700/defconfig): Sep 16 00:22:22 linux-handhelds-2.6: Fix Hermes/Atmel PCMCIA WiFi support. Sep 16 00:22:24 * defconfigman r405 support. Sep 16 00:22:28 03pfalcon 07org.oe.dev * r77359aaf... 10/ (1 packages/linux/linux-handhelds-2.6/h2200/defconfig): (log message trimmed) Sep 16 00:22:31 linux-handhelds-2.6: h2200: Update to defconfigman config. Sep 16 00:22:33 * Consistent with the rest of machines. Sep 16 00:22:35 * Diffs with previous manual config: Ethernet options relevant only for Sep 16 00:22:37 PCI/ISA turned off. No ACX USB support - this driver is not in mainline, and Sep 16 00:22:39 version in HH.org is severely hacked for slavemem/PCMCIA support. Better Sep 16 00:22:43 modularization. Sep 16 00:22:45 03pfalcon 07org.oe.dev * reed6759f... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.21-hh16.bb): linux-handhelds-2.6 2.6.21-hh16: Bump PR for defconfigs changes. Sep 16 00:22:48 03pfalcon 07org.oe.dev * r53d56e79... 10/ (4 files in 3 dirs): linux-handhelds-2.6 2.6.21: Drop few older versions. Sep 16 00:22:51 03pfalcon 07org.oe.dev * r8e566bee... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.21-hh9.bb): linux-handhelds-2.6 2.6.21-hh9: Drop old version. Sep 16 01:29:33 03likewise 07org.oe.dev * r8036d6ea... 10/ (3 files in 2 dirs): lighttpd: Moved from 1.4.16 to 1.4.18. **** ENDING LOGGING AT Sun Sep 16 02:59:56 2007