**** BEGIN LOGGING AT Fri Jun 25 02:59:57 2010 Jun 25 03:23:34 ah? Jun 25 04:01:31 GNUtoo|laptop, there is something interesting in your failing do_rootfs log... search for "check_data_file_clashes" Jun 25 04:09:06 ideally, OE should check for that error message and fail instead of do_rootfs succeeding Jun 25 04:17:32 GNUtoo|laptop, see rootfs_ipk.bbclass:rootfs_ipk_log_check(), it should be trivial to add a check for, e.g. "wants to install file" Jun 25 04:18:26 ...And then i suppose you need to figure out why the Marvell-Licence.txt file already exists. Jun 25 04:38:40 grg: i added "S:2345:respawn:/sbin/getty 115200 ttyS2" entry in /etc/inittab and now it allows the login shell access from the serial console, how do i redirect the runlevel S, 1, 2, ... logs to serial port?? Jun 25 04:39:16 hgs, i still suspect you need to alter the kernel command line Jun 25 04:39:26 hgs, you have console=foo listed twice Jun 25 04:39:58 once for the serial port, once for the framebuffer console Jun 25 04:40:13 those init level messages are probably going to your framebuffer Jun 25 04:40:45 oh ok i will disable my pslash and chk whether i could see init messages Jun 25 04:41:12 GNUtoo|laptop, I've updated the issue with my analysis: http://code.google.com/p/opkg/issues/detail?id=52 Jun 25 04:45:32 hgs, psplash should be able to find your framebuffer without console=tty0 on your kernel command line Jun 25 05:31:32 03Martin Jansa  07org.openembedded.dev * r40447c0205 10openembedded.git/recipes/xorg-lib/xtrans/abstract_socket_fix.patch: Jun 25 05:31:32 xtrans: remove unused abstract_socket_fix.patch Jun 25 05:31:32 Signed-off-by: Martin Jansa Jun 25 05:31:33 03Martin Jansa  07org.openembedded.dev * r5404278ea3 10openembedded.git/recipes/xorg-lib/ (9 files): Jun 25 05:31:33 libx11-native: add provide virtual/libx11-native Jun 25 05:31:41 03Martin Jansa  07org.openembedded.dev * r27ff4654fb 10openembedded.git/conf/ (15 files in 3 dirs): Jun 25 05:31:41 distros: add PREFERRED_PROVIDER_virtual/libx11-native Jun 25 05:31:41 * to be able to use virtual/libx11 in BBCLASSEXTEND = native recipes Jun 25 05:31:41 Signed-off-by: Martin Jansa Jun 25 05:31:43 03Martin Jansa  07org.openembedded.dev * r659f8c0df7 10openembedded.git/recipes/xorg-lib/ (6 files): Jun 25 05:31:43 libxfixes: convert to BBCLASSEXTEND Jun 25 05:31:43 * PR will be bumped later with INC_PR conversion Jun 25 05:31:51 libxdmcp: convert to BBCLASSEXTEND Jun 25 05:31:51 * PR will be bumped later with INC_PR conversion Jun 25 05:32:25 * remove original PROVIDES it's not used anywhere and is broken with BBCLASSEXTEND Jun 25 05:32:25 Signed-off-by: Martin Jansa Jun 25 05:32:25 03Martin Jansa  07org.openembedded.dev * r5467edf73c 10openembedded.git/recipes/xorg-lib/ (6 files): Jun 25 05:32:25 libxkbfile: convert to BBCLASSEXTEND Jun 25 05:32:25 * PR will be bumped later with INC_PR conversion Jun 25 05:33:14 xserver-xorg: add /usr/share/X11/xorg.conf.d to xserver-xorg package Jun 25 05:33:14 * ie for /usr/share/X11/xorg.conf.d/10-evdev.conf Jun 25 05:33:14 Signed-off-by: Martin Jansa Jun 25 05:33:14 03Martin Jansa  07org.openembedded.dev * re03f8bb577 10openembedded.git/recipes/xorg-util/util-macros.inc: Jun 25 05:33:15 util-macros: add BBCLASSEXTEND sdk, because some packages were already depending on it Jun 25 05:58:46 grg: thanks now could see the actual init debug messages redirected to the serial console it is saying "fsck: fsck.auto: No such file or directory" Jun 25 06:00:25 does that mean you need to specify the correct filesystem type in your fstab? Jun 25 06:04:12 grg: shld the chkscript supposed to call fsck.ext3 instead of fsck.auto?? Jun 25 06:04:25 hgs, yeah. Jun 25 06:05:05 it looks like it reads the filesystem type from fstab Jun 25 06:06:08 ok i updated fstab let me verify it Jun 25 06:22:29 grg: i modified fstab etnry as "rootfs /dev/root ext3 defaults 1 1", it again calling fsck.auto. Jun 25 06:23:48 hgs, well, maybe you have a different checkroot script to the one I'm looking at. I suggest you have a read of yours and see why its calling fsck.auto Jun 25 06:28:26 grg: ok i actually changed to "/dev/root / ext3 rw,relatime,errors=continue,data=writeback 1 1" this time called fsck.ext3 appropriately, but it is prompting for user input..here is the logs Jun 25 06:28:43 http://pastebin.com/i3gesu6p Jun 25 06:29:48 hgs, why not specify mmcblk0p1 instead of /dev/root? Jun 25 06:30:02 i mean, /dev/mmcblk0p1 Jun 25 06:30:28 grg: yes iam really confused with those fstab entries, dont know which format to follow Jun 25 06:32:07 the format is generally, filesystem_dev_file_or_whacko_alias /path/for/mount/destination comma,separated,options dump_no pass_no Jun 25 06:32:50 or was it pass_no, then dump_no... i forget Jun 25 06:34:21 grg: finally worked shuuu.... Jun 25 06:36:07 grg: then what about this default entry says "rootfs /dev/root auto defaults 1 1" Jun 25 06:36:20 rootfs seems to be an additional entry Jun 25 06:36:50 that some whacko non unix shit that has crept into linux in recent years Jun 25 06:37:14 i blame udev for those kind of shenanegans Jun 25 06:39:33 grg: oh ok i have other entries "proc, devpts ,usbfs ,tmpfs" follwing similar structure (non unix), should i update them Jun 25 06:39:46 no no Jun 25 06:39:51 those are normal :) Jun 25 06:39:56 :) Jun 25 06:40:12 they don't actually have corresponding /dev/file entries Jun 25 06:43:13 grg: i actually dont have a clue...have lot to learn.. Jun 25 06:43:57 hgs, well, you've walked through this problem in a manner that indicates you are capable of learning. That's a good start. Jun 25 06:44:51 i often get fed up trying to help people when it becomes clear they just want to be spoon fed solutions and cant do any of their own troubleshooting Jun 25 07:10:46 morning Jun 25 07:13:03 morning Jun 25 07:18:20 morning Jun 25 07:54:42 03Martin Jansa  07org.openembedded.dev * r7d9b0799a0 10openembedded.git/recipes/xorg-lib/ (134 files): Jun 25 07:54:42 xorg-lib: use INC_PR Jun 25 07:54:42 Signed-off-by: Martin Jansa Jun 25 07:54:51 03Martin Jansa  07org.openembedded.dev * re50bb7af5b 10openembedded.git/recipes/xorg-lib/ (6 files): Jun 25 07:54:51 libxext: remove PROVIDES_virtclass-sdk Jun 25 07:54:51 Signed-off-by: Martin Jansa Jun 25 07:54:54 03Martin Jansa  07org.openembedded.dev * r2531ea35b2 10openembedded.git/recipes/xorg-lib/ (147 files): Jun 25 07:54:54 xorg-lib: oe-stylize Jun 25 07:54:54 * require in oe-stylize script was moved to top Jun 25 07:54:54 * reminders were removed Jun 25 07:54:55 Signed-off-by: Martin Jansa Jun 25 07:55:05 03Martin Jansa  07org.openembedded.dev * rc34bf3cb88 10openembedded.git/recipes/ (19 files in 14 dirs): Jun 25 07:55:05 recipes: use libxt/libxft/libxdmcp/libxscrnsaver/liblbxutil/libxau/libxext instead of xt/xft/xdmcp/libxss/lbxutil/xau/xext providers Jun 25 07:55:05 * there is only one provider, so use it directly as othere recipes does Jun 25 07:55:05 Signed-off-by: Martin Jansa Jun 25 08:08:33 JaMa: now there is one. in past it was not clean Jun 25 08:09:22 hrw: provider? but OK to clean it now, right? Jun 25 08:09:27 yes Jun 25 08:09:39 before x11 got modular Jun 25 08:09:46 hrw: because most of recipes were using direct depends now.. so better to clean it everywhere Jun 25 08:09:56 yes Jun 25 08:10:09 JaMa: I like your work on x11 stuff Jun 25 08:10:41 good, than I hope you'll ack my mesa cleanup (a bit more controversal because of old versions removal) Jun 25 08:10:52 long time ago I was doing lot of x11 upgrades and it is nice to see that someone continue that Jun 25 08:10:55 but it's not yet pushed - still testing old cruft Jun 25 08:11:11 JaMa: I do not do OE builds so hard for me to ack that Jun 25 08:13:20 ok Jun 25 08:17:54 gm Jun 25 08:20:26 gm Jun 25 08:26:00 03Martin Jansa  07org.openembedded.dev * r4eda57ba4a 10openembedded.git/recipes/libjson/ (libjson/json-c-longlongint.patch libjson_svn.bb): Jun 25 08:26:00 libjson: add svn versions with patch needed for elmdentica Jun 25 08:26:00 Signed-off-by: Martin Jansa Jun 25 08:29:57 hi eFfeM_work Jun 25 08:30:25 good morning Jun 25 11:32:16 Hi All, One doubt plz.. does tmp/ folder needs to be deleted or moved whenever you bitbake different packages.. ex. helloworld and console-image?? Which is more safe? Jun 25 11:35:58 nikunj ???? Jun 25 11:37:32 woglinde : I bitbaked helloworld first.. it had created tmp/ folder.. which has all the associated data. Now If I want to bitbake console-image .. does that tmp/ folder needs to be moved or deleted.?? Jun 25 11:38:04 no why? Jun 25 11:38:14 that would be really stupid Jun 25 11:38:23 we are making some stupid things Jun 25 11:38:30 woglinde : till now I was deleting tmp/ folder every time I bitbake any package.. Jun 25 11:38:35 but dont expected us this much stupid Jun 25 11:39:55 woglinde : I agree !! Actually I wasnt too sure about whether to keep tmp/ folder or continue with the old one..Afraid about getting errors. Jun 25 11:40:53 woglinde : So ultimately this will again save time in building any package..Right? Jun 25 11:42:15 ????????? Jun 25 11:42:17 *sigh* Jun 25 11:42:35 Nikunj please look into the tmp yourself Jun 25 11:42:39 and find your answer Jun 25 11:46:59 Nikunj, usually you do not need (nor want) to manually delete the tmp directory Jun 25 11:48:14 only for distro bump Jun 25 11:48:16 *g* Jun 25 11:48:25 hi blindvt Jun 25 11:48:44 woglinde, \"folder\" always reminds me on BSE Jun 25 11:49:34 blhm? Jun 25 11:52:54 dunno. folding (proteins) and flapping proteins. Just an association. nm Jun 25 11:58:51 03Martin Jansa  07org.openembedded.dev * r3c8aa3ac19 10openembedded.git/conf/machine/om-gta02.conf: Jun 25 11:58:51 om-gta02: prefer kernel 2.6.32 Jun 25 11:58:51 Signed-off-by: Martin Jansa Jun 25 11:58:52 03Martin Jansa  07org.openembedded.dev * r3322af9410 10openembedded.git/recipes/xorg-xserver/ (2 files in 2 dirs): Jun 25 11:58:52 xserver-xorg-conf: update om-gta02 config for 2.6.32 and newer kernel (evdev node changed for ts) Jun 25 11:58:58 03Martin Jansa  07org.openembedded.dev * r9707d105fa 10openembedded.git/recipes/freesmartphone/fsodeviced_git.bb: Jun 25 11:58:58 fsodeviced: make alsa states for 2.6.31 default Jun 25 11:58:58 Signed-off-by: Martin Jansa Jun 25 11:59:00 03Martin Jansa  07org.openembedded.dev * rc0987e7f50 10openembedded.git/conf/distro/shr.conf: Jun 25 11:59:00 shr: inherit couple of usefull classes: recipe_sanity, insane, devshell and bump DISTRO_PR Jun 25 11:59:00 * gcc major version was changed, better to sync recipes to targets Jun 25 11:59:00 * new QA tasks needs available workdirs Jun 25 11:59:11 03Martin Jansa  07org.openembedded.dev * r733149b626 10openembedded.git/conf/distro/include/preferred-shr-versions.inc: Jun 25 11:59:11 shr: prefer qt-4.7.0-beta1 Jun 25 11:59:11 Signed-off-by: Martin Jansa Jun 25 11:59:15 03Martin Jansa  07org.openembedded.dev * rfe3e3931f5 10openembedded.git/recipes/freesmartphone/ (2 files in 2 dirs): Jun 25 11:59:15 fsogsmd: add patches for 2.6.31+ on om-gta02 Jun 25 11:59:15 Signed-off-by: Martin Jansa Jun 25 11:59:18 03Martin Jansa  07org.openembedded.dev * r168eea64ef 10openembedded.git/recipes/xorg-proto/ (5 files): Jun 25 11:59:18 kbproto: convert to BBCLASSEXTENDS Jun 25 11:59:18 Signed-off-by: Martin Jansa Jun 25 11:59:19 03Martin Jansa  07org.openembedded.dev * rd2bc5b243d 10openembedded.git/recipes/xorg-proto/ (39 files): Jun 25 11:59:19 xorg-proto: remove XORG_PN, it was now used only in DESCRIPTION where BPN is also enough Jun 25 11:59:27 03Martin Jansa  07org.openembedded.dev * rc56b6369c0 10openembedded.git/recipes/xorg-proto/ (24 files): Jun 25 11:59:27 xorg-proto: add nativesdk to BBCLASSEXTEND Jun 25 11:59:27 Signed-off-by: Martin Jansa Jun 25 12:12:54 re Jun 25 12:13:02 hm jama has some freetime Jun 25 12:13:07 03Koen Kooi  07org.openembedded.dev * r721ce80031 10openembedded.git/recipes/binutils/binutils-cross-sdk_cvs.bb: binutils-cross-sdk cvs: fix PV Jun 25 12:23:24 woglinde: not at all (deadline at work ;/).. but I would really like to have it cleaned a bit.. Jun 25 12:25:22 *g* Jun 25 12:28:18 hey guys, anyone know the best way to get Brussels to Munich by train and how long it takes? Jun 25 12:28:42 xora hm Jun 25 12:28:54 xora from brussel to cologne via thalis Jun 25 12:29:04 and from cologone to munich Jun 25 12:29:22 you can use www.bahn.de Jun 25 12:30:34 woglinde: thanks Jun 25 12:30:35 6:50 the fastest Jun 25 12:31:03 in the thalis you can have wlan acess Jun 25 12:31:07 but its not cheap Jun 25 12:31:37 its not for me, friend is planning the journey for herself but got lost on a french language site Jun 25 12:31:45 uh Jun 25 12:32:04 hm for thalis you normaly need reservation Jun 25 12:32:04 figured OE guys tend to scoot around by train a lot :-) Jun 25 12:32:13 yes Jun 25 12:32:28 most of the time cheaper than flight Jun 25 12:32:36 and more comfy Jun 25 12:32:40 planes annoy me Jun 25 12:32:45 hm not really Jun 25 12:32:46 *g* Jun 25 12:33:30 my normal Edin to London route, because of all the messing around you have to do, sometimes the train is actually quicker :-) Jun 25 12:33:45 sure Jun 25 12:33:53 03Koen Kooi  07org.openembedded.dev * r371343e2cc 10openembedded.git/recipes/u-boot/u-boot_git.bb: u-boot git: remove ramdisk and user button patches that don't work on revB and revC beagles Jun 25 12:34:51 jama ack for the deleting of old mesa stuff Jun 25 12:36:52 thats 2 acks, Ive just emails mine Jun 25 12:47:32 XorA: thanks :) Jun 25 12:48:04 woglinde: can you ack on email to be correct for everyone? Jun 25 12:50:30 jama done Jun 25 13:04:07 03Martin Jansa  07org.openembedded.dev * rabbc5acc0f 10openembedded.git/recipes/mesa/ (17 files in 3 dirs): Jun 25 13:04:07 mesa: move common definitions to include file, change SRC_URI to freedesktop for releases newer than 7.5 Jun 25 13:04:07 * drop eee901 override (no such machine in OE, dropped from poky Jun 25 13:04:07 http://git.pokylinux.org/cgit.cgi/poky/commit/?id=542eb3b16dc1c6fa7920257219ed5dac0b95c257) Jun 25 13:04:07 Signed-off-by: Martin Jansa Jun 25 13:04:15 preferred-om-2008-versions: remove pinned version of mesa/mesa-dri Jun 25 13:04:16 Signed-off-by: Martin Jansa Jun 25 13:04:16 Acked-by: Henning Heinold Jun 25 13:04:17 Acked-by: Graeme Gregory Jun 25 13:04:17 03Martin Jansa  07org.openembedded.dev * r2f62a96457 10openembedded.git/recipes/mesa/ (6 files in 2 dirs): Jun 25 13:04:18 mesa-dri: remove old 7.0.3 and 6.5.2 relase using mesa-common-old.inc Jun 25 13:04:18 Signed-off-by: Martin Jansa Jun 25 13:04:19 Acked-by: Henning Heinold Jun 25 13:04:19 Acked-by: Graeme Gregory Jun 25 13:04:27 * there was --disable-egl in mesa-dri recipes (there are mesa recipes for this) Jun 25 13:04:27 Signed-off-by: Martin Jansa Jun 25 13:04:28 Acked-by: Henning Heinold Jun 25 13:05:19 Acked-by: Graeme Gregory Jun 25 13:05:20 03Martin Jansa  07org.openembedded.dev * r02fdabdbaf 10openembedded.git/ (11 files in 5 dirs): Jun 25 13:05:20 mesa: move 7.8 release to latest minor 7.8.2 Jun 25 13:05:20 Signed-off-by: Martin Jansa Jun 25 13:05:20 Acked-by: Henning Heinold Jun 25 13:12:40 03Martin Jansa  07org.openembedded.dev * rdb9b2d1745 10openembedded.git/classes/mozilla.bbclass: Jun 25 13:12:40 mozilla.bbclass: use libxt directly Jun 25 13:12:40 Signed-off-by: Martin Jansa Jun 25 13:12:41 03Martin Jansa  07org.openembedded.dev * r46012bde53 10openembedded.git/recipes/vim/gvim_7.2.bb: Jun 25 13:12:41 gvim: use libxt directly Jun 25 13:12:41 Signed-off-by: Martin Jansa Jun 25 13:13:43 03Martin Jansa  07org.openembedded.dev * re1f208e7e1 10openembedded.git/recipes/gcc/gcc-package-sdk.inc: Jun 25 13:13:43 gcc-package-sdk: skip stripping *.py files Jun 25 13:13:43 Signed-off-by: Martin Jansa Jun 25 13:13:43 Signed-off-by: Koen Kooi Jun 25 13:20:48 03Martin Jansa  07org.openembedded.dev * rea59e852d8 10openembedded.git/classes/testlab.bbclass: Jun 25 13:20:49 testlab: from opkg info grep only +- 7 lines around package with status installed of unpacked Jun 25 13:20:49 * silly solution to filter multiple versions shown in installed-packages.txt where one is with Status not-installed and the right one with installed Jun 25 13:20:49 * someone with better idea? Jun 25 13:20:49 Signed-off-by: Martin Jansa Jun 25 13:23:20 03Koen Kooi  07org.openembedded.dev * r2e0f5099da 10openembedded.git/recipes/binutils/binutils-cross-sdk_2.20.1.bb: binutils-cross-sdk: add 2.20.1 Jun 25 13:45:01 03Martin Jansa  07org.openembedded.dev * r69efbdbe8f 10openembedded.git/recipes/xorg-lib/ (18 files): Jun 25 13:45:01 libx11,diet-x11: use INC_PR Jun 25 13:45:01 Signed-off-by: Martin Jansa Jun 25 13:45:13 03Martin Jansa  07org.openembedded.dev * r82dd23467d 10openembedded.git/recipes/xorg-lib/ (40 files in 7 dirs): (log message trimmed) Jun 25 13:45:14 libx11: convert to BBCLASSEXTEND Jun 25 13:45:14 * common patches in new libx11.inc (some needed to be renamed to keep Jun 25 13:45:14 same name across version) Jun 25 13:45:14 * keep one common dir with diet-x11/libx11-trip variants Jun 25 14:13:01 morning Jun 25 14:14:10 hi kergoth Jun 25 14:14:27 hey hrw Jun 25 14:16:37 g'day kergoth Jun 25 14:17:31 fyi, I modified bitbake so it would error out if anyone tries to access a variable that wasn't captured by the signature for the task being executed -- and successfully built task-boot, task-base, task-proper-tools, etc with it :D Jun 25 14:17:43 so i guess the functionality for the varref stuff is there Jun 25 14:17:58 now can focus on performance and improving the code, maintenance wise Jun 25 14:18:16 excellent Jun 25 14:19:11 then need to figure out how to fully integrate it -- i think i'll need to add a setVar notification mechanism to dirty cached expanded variables Jun 25 14:19:20 then perhaps see about making the parser directly produce this tree Jun 25 14:19:43 and of course, to use this for stamps would require that we make tasks stay within WORKDIR, and start tracking WORKDIR contents Jun 25 14:19:54 * kergoth makes a list Jun 25 14:21:15 it's too bad we can't safely cache pythonsnippets, since we can't assume they have no side effects Jun 25 14:21:48 functional programming ftw Jun 25 14:23:39 heh, the problem is we have snippets doing all sorts of things that can't function properly without some side effects -- at a minimum, I/O Jun 25 14:23:51 could add the ability to explicitly specify referenced files Jun 25 14:23:55 in addition to referenced variables Jun 25 14:24:00 then utilize that in the caching Jun 25 14:24:20 * kergoth thinks this is one giant nested can of worms Jun 25 14:26:20 so Jun 25 14:26:35 i'm trying to gather as much information as possible on the history of OE, more than what i happen to recall Jun 25 14:26:53 about the original discussions for the design when we started the project Jun 25 14:27:06 we no longer seem to have the original wiki, sadly Jun 25 14:27:23 but, timriker's bot, ibot/purl/jbot has logs going back to 2002 for #openzaurus, #elinux, #openembedded, #oe Jun 25 14:27:26 so that seems promising Jun 25 14:28:12 i can remember quite a bit about why we didn't go with the various alternatives, but i figure getting exact info is better than trusting my memory Jun 25 14:28:59 the most annoying thing is trying to remember which projects were even around back then, vs ones that showed up later :) Jun 25 14:29:16 "well, i know what issues exist with FOO, but was that even an option then?" Jun 25 14:29:19 * kergoth rolls eyes Jun 25 14:32:42 that's promising, i told timriker that schurig had added some more detail about the ideas "we discussed the other week" on the oe wiki, this was 20030218 at 21:39.29 Jun 25 14:37:16 haha, tim fishslapped me for creating #openembedded on 20030123. course, at the time it wasn't a project of its own really, it didn't have much of an identity Jun 25 14:37:19 but still :) Jun 25 14:38:52 hah Jun 25 14:39:38 he worried a bit about fracturing, i think, even then he wanted #elinux / elinux.org as the central embedded linux information repository, but I don't think he realized we were more about the metadata than the documentation Jun 25 14:41:20 okay, the design discussions had to occur after 20030123 - around the time we created #openembedded and got the domains, and near / shortly after 20030218, when i told tim about updates to the wiki pages Jun 25 14:41:25 getting there Jun 25 14:41:27 :) Jun 25 14:42:48 ah ha! Jun 25 14:42:51 http://web.archive.org/web/20030513111311/www.openembedded.org/phpwiki/index.php/BuildSystem Jun 25 14:42:54 thank you wayback machine Jun 25 14:42:55 <3 Jun 25 14:43:13 that includes logs of the original discussions between tim and i that got me mulling over the thing Jun 25 14:43:23 woot. Jun 25 14:44:20 hahaha Jun 25 14:44:23 schurig's ebuild page Jun 25 14:44:25 "What's not so good Jun 25 14:44:25 With it's over 3000 packages, portage is sometimes slow when it comes to dependency calculations. On my 1GHz box it takes 1-3 seconds. They are right now thinking about a Berkely DB where they store this info. However, the dependency tracking works very well." Jun 25 14:44:29 1-3 seconds, i wish.. Jun 25 14:44:30 ;) Jun 25 14:44:45 heh, puny 3000 packages Jun 25 14:44:51 hehe, indeed Jun 25 14:53:45 http://web.archive.org/web/20030606074302/openembedded.org/phpwiki/index.php/OeTODO - heh - oeconfig .. we never did get such a thing Jun 25 14:54:18 haha Jun 25 14:54:19 "enhance task dependency checking.. force a redownload/rebuild/etc if SRC_URI changes from what it was previously, rebuild if the build steps change, etc. Essentially store a CRC of the variable which applies to a given task, and force a clean and rerun of that task if the CRC changes." Jun 25 14:54:31 it only took us 7 years, and we still aren't there yet Jun 25 14:54:53 doh Jun 25 14:54:58 bad work by us Jun 25 14:55:04 :) Jun 25 14:55:22 well, to be fair, we did have more important things to do Jun 25 14:55:25 like making sure shit actually built Jun 25 14:55:27 but still Jun 25 14:57:07 oh, that reminds me Jun 25 14:57:15 the varref stuff completely ignores overrides Jun 25 14:57:28 need a wrapper class to associate a conditional to a value Jun 25 14:57:30 most likely Jun 25 14:57:38 then the conditional value could sit anywhere in the tree Jun 25 14:57:51 hmm, how to store the condition Jun 25 15:10:55 * kergoth tries to make something coherent out of this scattered information Jun 25 15:11:44 holy cow koen, according to ohloh he has nearly double the commits of the next person below him for OE Jun 25 15:15:08 kergoth: you forgot that Koen was a man who wrote OE in one day? Jun 25 15:15:15 aka first monotone commit Jun 25 15:15:24 * kergoth chuckles Jun 25 15:16:07 the first oe commits were in cvs or something, but I don't think the OE repo has the full history Jun 25 15:16:15 (i.e. pre-oe-split) Jun 25 15:16:29 kergoth: cvs? before bitkeeper even? Jun 25 15:17:12 kergoth: now we have history since bitkeeper but BK commits are disaster due to a way how bitmover exported it Jun 25 15:17:49 yeah, before bk Jun 25 15:18:18 check out the bitbake repo, i spent a lot of time on that one, and i think it still may be msising a few things Jun 25 15:18:54 http://git.openembedded.org/cgit.cgi/bitbake/commit/?id=af645b03b585a9c92dc111188756b8f6badfec5f Jun 25 15:19:21 i think that one was cvs, i worked to convert all the authors to current info and removed extra metadata in the commit logs Jun 25 15:19:24 hi kergoth Jun 25 15:19:43 2003-02-21 04:03:45 sounds about right, this may be everything Jun 25 15:19:56 hey woglinde Jun 25 15:21:44 that first parser wasn't very pythonic python ;) can tell i was a noob Jun 25 15:22:39 0a866478cc2ae23a4d60ea407b0d6d89146f3016 is when holger schurig started pulling over portage bits Jun 25 15:22:51 *g* oe history Jun 25 15:22:55 :) Jun 25 15:23:20 hi hrw btw. Jun 25 15:24:10 * kergoth ponders what to do with all this info Jun 25 15:24:50 maybe a basic timeline showing certain big events in the project lifetime Jun 25 15:24:57 supplemented by additional information about design decisions, etc Jun 25 15:25:36 kergoth: nice idea Jun 25 15:25:50 someone was asking me about why we did things the way we did Jun 25 15:26:00 i'm hoping this sort of information will help that Jun 25 15:27:44 hmmm Jun 25 15:28:13 * kergoth wonders where he put the dirs where he was working on fixing up the bitbake repository, since there were oe bits there too, could graft them onto the main repo Jun 25 15:31:45 hmm, anyone know if any of the OE presentations covered much about our history, other than just very brief onceovers? Jun 25 15:31:52 alternatively, any that cover the high level concepts well? Jun 25 15:32:06 seems like most are just quick introductions + some recipe or implementation details Jun 25 15:32:11 * kergoth downloads some Jun 25 15:40:54 03Martin Jansa  07org.openembedded.dev * rd3649ee3a5 10openembedded.git/recipes/xorg-xserver/ (xserver-xorg_1.8.0.bb xserver-xorg_1.8.1.bb): Jun 25 15:40:54 xserver-xorg: remove D_P -1 from 1.8 versions, because mesa is now new enough Jun 25 15:40:54 Signed-off-by: Martin Jansa Jun 25 15:44:23 kergoth: didn't someone render codeswarm video? would be interesting to comment design decisions in video in form of subtitles or something :) Jun 25 15:44:32 i was thinking about that too :) Jun 25 15:45:11 could show a movement along a timeline as the codeswarm proceeds, with a subtitle showing the event in the timeframe Jun 25 15:46:46 * JaMa would like to see SRCREVs in recipes and checksums in recipes commits :) Jun 25 15:47:42 JaMa: I used gource Jun 25 15:48:17 http://www.youtube.com/watch?v=06RdXKHwV6U Jun 25 15:48:50 ah, nice Jun 25 15:48:59 * JaMa saves link for later, corporate fw here blocks youtube Jun 25 15:49:03 greg gilbert :( Jun 25 15:49:24 RIP Jun 25 15:49:42 yep Jun 25 15:50:16 good that his family handled exports of data from his farm Jun 25 15:50:29 did they? that's impressive Jun 25 15:50:47 i should put my passwords in a fire safe or something, just in case Jun 25 15:50:49 heh Jun 25 15:51:04 right, and write a will :-} Jun 25 15:51:09 that too Jun 25 15:51:10 kergoth: did not we had bugtracker at his machine? Jun 25 15:51:28 * kergoth doesn't remember Jun 25 15:51:35 i know he ran the source mirror Jun 25 15:51:41 yeah, I think a whole pile of oe bits were hosted on treke.net at one point Jun 25 15:52:01 this was before melo, obviously, and I think it pre-dated ltg as well Jun 25 15:52:30 bugs.treke.net Jun 25 15:56:37 kergoth: LTG is 2006, Greg died in Feb 2008 Jun 25 15:57:02 http://linuxtogo.org/here-we-are http://article.gmane.org/gmane.comp.handhelds.openembedded/17660/ Jun 25 15:57:52 ugh Jun 25 15:58:18 heart attack at 26 due to diabetes that he didn't know about? Jun 25 15:58:22 god, what a shame Jun 25 15:58:31 * kergoth didn't realize the cause Jun 25 16:03:17 hi mrmoku Jun 25 16:04:46 woglinde: hi :) Jun 25 16:09:02 wonder how much work itd be to track down when various oe based projects / commercial releases of oe based products happened Jun 25 16:10:24 have a nice weekend Jun 25 16:10:30 bye hrw Jun 25 16:11:34 kergoth: http://www.openzaurus.org/wordpress/category/releases/ will give you OZ track Jun 25 16:12:20 cool, thanks Jun 25 16:24:00 can I instruct bitbake to run rm_work as soon as possible instead of as one of the tasks executed last? Jun 25 16:24:29 I mean while building and image or so (clearly it should be last for particular recipe) Jun 25 16:24:57 * JaMa has rather small /tmp Jun 25 17:15:49 JaMa: hmmm how small is your tmp Jun 25 17:20:26 khem: 12G Jun 25 17:20:59 khem: but sometimes there is another build already running (some other OE distro or portage workdir is also there) Jun 25 17:22:32 khem: I'm building image as sequence of smaller tasks (which cleans before finish) but would be nicer to just run rm_work as soon as workdir is not needed anymore by following tasks Jun 25 17:23:13 khem: btw I'm rebuilding gcc-4.5 images without optspace disable (seeing your fixes before), do you consider it safe now? Jun 25 17:25:48 JaMa, right, but exactly when can you be certain there are no tasks that do anything with workdir anymore? Jun 25 17:26:01 its not just do install. do deploy copies files out of there too, etc Jun 25 17:26:49 now, i could see being a bit smarter than wiping workdir. after install and deploy have been run, wipe ${S}/${B}, perhaps Jun 25 17:26:51 * kergoth shrugs Jun 25 17:29:01 kergoth: honestly I have no idea how the order of tasks is decided (between different recipes)... I only know how is task order defined with addtask for one recipe Jun 25 17:29:54 there are many flags that control cross-recipe task deps, in addition to DEPENDS/RDEPENDS based Jun 25 17:30:15 see deptask, depends, recrdeptask, rdeptask, etc Jun 25 17:30:30 may want to look at the bitbake -g output Jun 25 17:31:52 if it's really hard to be a bit smarter then I'll stay with sequence of targets and let bitbake experts like you handle it if they find it doable and usefull :) Jun 25 17:33:29 we'll get there. in the future we're looking to better track what the file/dir input/output is for individual tasks Jun 25 17:33:37 with that information, we could really know when we're done with a given dir Jun 25 17:56:08 JaMa: yes optspace should work ok now Jun 25 18:14:56 will is be the Swiss or spaniards .. lets see Jun 25 18:15:09 * khem plans to watch at the local St. Johnn Jun 25 18:16:15 hi khem Jun 25 18:16:34 khem I am really lucky that italy is out Jun 25 18:16:50 woglinde: heh Jun 25 18:17:26 woglinde: spain and chile will be interesting Jun 25 18:17:27 spain will follow today Jun 25 18:17:41 woglinde: I dont think so Jun 25 18:17:44 03Steve Sakoman  07org.openembedded.dev * rdd0c9cb1e6 10openembedded.git/recipes/xorg-xserver/ (xorg-xserver-common.inc xserver-xorg_1.7.4.bb): Jun 25 18:17:44 xserver-xorg: remove dependency on xorg-minimal-fonts Jun 25 18:17:44 this is no longer necessary, and infact leads to a conflict with font-alias Jun 25 18:17:44 Signed-off-by: Koen Kooi Jun 25 18:18:05 woglinde: it could be that spain and swiss loose and Honduras goes thru :) Jun 25 18:18:25 but I guess spain will play a boring draw Jun 25 18:18:37 *g* Jun 25 18:18:52 the brazil - portugese game was shit Jun 25 18:20:32 yeah Jun 25 18:21:42 hmmm debug version of uclibc libc.so.0 on mips is 102M Jun 25 18:21:54 something is broken Jun 25 18:22:04 either the linker or the script Jun 25 18:22:39 hm why? Jun 25 18:22:45 debug ius heavy Jun 25 18:23:03 think about git Jun 25 18:23:05 that remains to be seen Jun 25 18:23:07 git-native Jun 25 18:23:52 CFLAGS += -O0 -g3 -DDEBUG Jun 25 18:23:55 ah Jun 25 18:24:00 thats why Jun 25 18:55:30 what was chilean goalkeeper doing Jun 25 19:54:27 03Martin Jansa  07org.openembedded.dev * r7f80b1581d 10openembedded.git/recipes/dvbtools/dvb-apps_1.1.1.bb: Jun 25 19:54:27 dvb-apps: rename evtest to dvb-evtest to workaround conflict with standalone evtest recipe Jun 25 19:54:27 * input-utils were preferred instead of included evtest already in 2005 Jun 25 19:54:27 http://www.linuxtv.org/pipermail/linux-dvb/2005-November/006402.html Jun 25 19:54:27 Signed-off-by: Martin Jansa Jun 25 19:54:31 03Martin Jansa  07org.openembedded.dev * r1036e9e2d9 10openembedded.git/conf/distro/shr.conf: Jun 25 19:54:31 shr: drop unused DISTRO_* definitions Jun 25 19:54:31 Signed-off-by: Martin Jansa Jun 25 19:54:37 03Martin Jansa  07org.openembedded.dev * r02535ca958 10openembedded.git/recipes/shr/initscripts-shr_0.0.1.bb: Jun 25 19:54:37 initscripts-shr: RCONFLICTS is enough and installed version is controlled by IMAGE_INITSCRIPTS Jun 25 19:54:37 Signed-off-by: Martin Jansa Jun 25 20:20:13 ieehks Jun 25 20:20:15 http://bugs.python.org/issue9075 Jun 25 20:20:57 lol hi Jun 25 20:21:20 woglinde, btw you're interested in fastjar, I saw on lwn.net that it got a CVE Jun 25 20:21:40 gnutoo hm maybee I will be soon fastjar release manager Jun 25 20:21:49 ok Jun 25 20:21:53 dalibor is searching someone Jun 25 20:21:58 ah ok Jun 25 20:22:11 but didnt notice the cve Jun 25 20:46:06 03Henning Heinold  07org.openembedded.dev * r06cb3e1e79 10openembedded.git/recipes/fastjar/ (fastjar_0.98.bb files/jartool.patch): Jun 25 20:46:07 fastjar: apply fix for CVE-2010-0831 Jun 25 20:46:07 * bump PR Jun 25 20:48:39 kergoth: I thought that maybe you'd find it interesting, in case you don't know it already: http://www.stanford.edu/~pgbovine/incpy.html Jun 25 20:50:17 mario-goulart, interesting. i think i'd rather use a persistant memoizing decorator and do it explicitly, rather than trusting that that thing knows which functions are side effect free, but I'll have to check it out, thanks for the link Jun 25 20:51:42 kergoth: you're welcome Jun 25 20:57:58 03Martin Jansa  07org.openembedded.dev * r5e3295a867 10openembedded.git/recipes/tangogps/ (4 files): Jun 25 20:57:58 tangogps: add 0.99.4 version and drop older 0.99.x Jun 25 20:57:58 Signed-off-by: Martin Jansa Jun 25 21:18:01 03Koen Kooi  07org.openembedded.dev * r4be4a137d3 10openembedded.git/recipes/orc/ (4 files in 2 dirs): orc: add 0.4.5, rumoured to have a neon backend Jun 25 22:05:39 khem: good news. bug 5435 can be closed. it was one big low level problem. we just need to set the processor in certain mode. some hardware bug workaround Jun 25 22:34:57 03Andrea Adami  07org.openembedded.dev * rc727649644 10openembedded.git/recipes/kexecboot/ (kexecboot.inc kexecboot_git.bb): Jun 25 22:34:58 kexecboot_git: bump to ac70f55e2b49f23963a8f057330eb47836a01c2f Jun 25 22:34:58 * add forgotten initrd/initramfs support, Jun 25 22:34:58 * make package machine-specific Jun 25 22:34:58 * (even if the binary itself is arch specific and could be run as standalone), Jun 26 00:06:12 03Andrea Adami  07org.openembedded.dev * rd0f0af2e64 10openembedded.git/recipes/kexecboot/ (2 files in 2 dirs): Jun 26 00:06:12 linux-kexecboot: add pending patch for c7x0 to 2.6.34 Jun 26 00:06:12 * unbreak corgi mmc/sd detection. Jun 26 00:06:12 Signed-off-by: Eric Miao Jun 26 00:06:22 03Andrea Adami  07org.openembedded.dev * r7598af259a 10openembedded.git/recipes/kexecboot/ (2 files in 2 dirs): Jun 26 00:06:23 linux-kexecboot: add pending patch for c7x0 to 2.6.32 Jun 26 00:06:23 * unbreak corgi mmc/sd detection. Jun 26 00:06:23 Signed-off-by: Eric Miao Jun 26 00:13:37 03Andrea Adami  07org.openembedded.dev * r9e6b31163c 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Jun 26 00:13:37 linux: add pending patch for c7x0 to 2.6.34 Jun 26 00:13:37 * unbreak corgi mmc/sd detection. Jun 26 00:13:37 Signed-off-by: Eric Miao **** ENDING LOGGING AT Sat Jun 26 02:59:57 2010