**** BEGIN LOGGING AT Sat Dec 29 02:59:56 2007 Dec 29 03:16:12 Evening folks Dec 29 03:16:50 mornin' Dec 29 03:17:07 Mornin :) Dec 29 07:17:21 03rwhitby 07org.oe.dev * rfd451af1... 10/ (1 packages/images/unslung-image.bb): unslung-image: Override nslu2_pack_image to not use Apex. Dec 29 07:17:26 03rwhitby 07org.oe.dev * r82edd0ec... 10/ (3 files in 3 dirs): unslung: Bumped version to 6.10-beta, and bumped PR on rootfs to rebuild with new version number Dec 29 07:17:31 03rwhitby 07org.oe.dev * ref09a721... 10/ (1 packages/git/files/git-gui-install-mode-arg-spaces.patch): git-native: Add missing patch Dec 29 07:34:22 morning all Dec 29 07:48:21 hi all again Dec 29 07:49:36 how i can force to use a specific version of gcc to build the kernel? there's a machine that already do it? Dec 29 09:39:18 ggilbert: were you doing some work on OE bugzilla earlier? if so, mail sending is borked in case you didn't know already Dec 29 09:39:30 It's not exactly borked Dec 29 09:39:43 It's more that postfix isn't even on :p Dec 29 09:39:44 well I just got "There was an error sending mail from 'bugzilla-daemon' to 'roman.moravcik@gmail.com':error when closing pipe to /usr/lib/sendmail:" Dec 29 09:39:55 ah, right, that would do it :) Dec 29 09:40:03 np, just letting you know Dec 29 09:40:27 I'm actually sitting here staring at main.cf trying to decide what I'm doing Dec 29 09:41:08 on the brightside, I think diff viewing is now working in bugzilla Dec 29 09:41:38 ggilbert_: seems to be, yeah Dec 29 09:41:43 works pretty well actually Dec 29 09:47:07 hi all, i need to compile the kernel with a specific version of gcc, how can i do it? there is a machine that already need this? Dec 29 10:07:33 morning all Dec 29 10:08:40 gremlin[it]: Look at mnci-ramses_2.4.21-rmk2-pxa1.bb Dec 29 10:09:01 thank RP ! Dec 29 12:26:52 wow, this feud on angstrom-distro-devel is really getting out of hand ... Dec 29 12:28:29 I stopped following it after my last mail. Discussion with the two of them is like talking to a wall Dec 29 12:28:40 somehow I've managed to miss this... Dec 29 12:28:46 hi zecke Dec 29 12:29:12 hey richard Dec 29 12:29:15 RP: how are you? Dec 29 12:29:21 I consider unsubscribing from angstrom-* Dec 29 12:30:06 zecke: ok thanks, you? Dec 29 12:30:47 RP: I sit at the 24c3 so I'm fine :) Dec 29 12:31:00 CoreDump: Which thread are we talking about? Dec 29 12:31:25 zecke: sounds like fun :) Dec 29 12:31:28 OE & Angstrom reputation (was: Bugzilla is not USENET) Dec 29 12:34:52 * rwhitby is glad he isn't involved in that thread ... Dec 29 12:36:48 I can't decide if I want to read the context for this or not... Dec 29 12:37:14 RP: it started at http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2007-December/001179.html, and continued in a new thread at http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2007-December/001264.html Dec 29 12:45:47 basically it is about a user comlaining about the reply he recieved after asking a valid question. Ends in the user getting himself banned due to him being an idiot (but his original point is value fwiw) Dec 29 12:50:24 * XorA|gone wakes up Dec 29 12:50:33 with lots of christmas spirit in the posts between :-) Dec 29 12:51:22 well at least the flame war is on a mailing list not bugzilla, which was my original complaint Dec 29 12:52:22 XorA|gone: got a URL to the bug? Dec 29 12:53:16 rwhitby: I honestly cant remeber it was sometime the day before I sent original email when a bug about international keyboards exploded into a discussion on who was forced out of angstrom by who Dec 29 12:54:28 RP: what is mallum's email address? mallum@o-hand.com? Dec 29 12:56:54 rwhitby: http://bugs.openembedded.org/show_bug.cgi?id=1908 Dec 29 12:57:52 thx Dec 29 12:57:53 rwhitby: but there are a few others Dec 29 14:09:08 hi everybody Dec 29 14:16:05 has anyone booted base-image on anything? Dec 29 14:19:14 thesing: hi Dec 29 14:28:44 hmmm, looks like 450 usd to get a plane ticket for FOSDEM Dec 29 14:37:17 mhhh for me is around 200EUR ... Dec 29 14:38:00 someone have hints about an hotel, such last year we was about all all in the same place Dec 29 14:38:48 gremlin[it]: the Ibis Hotel is just in front of the Train station Dec 29 14:39:59 I suspect we will stay at the same place as last year Dec 29 14:41:01 astrid I think Dec 29 14:42:22 yes wan't bad ... but i suspect this year the Astrid could be quite full... Dec 29 14:42:36 why? Dec 29 14:42:58 wan't -> wasn't ... but this year the breakfat isn't included :( Dec 29 14:43:13 i try to book and sometime it give me all full Dec 29 14:43:36 via expedia ok, via direct hotel site all booked ... :( ... really don't know Dec 29 14:58:16 anyone around to help a newb? Dec 29 14:58:29 just ask and see if you get a response Dec 29 14:59:45 what's the best wy to determine/understand why packages are being skipped? Dec 29 15:00:49 fun_in_oe: skipping in which sense? not created? not build? not installed in a image? Dec 29 15:01:38 well when i run bitbake it tells me parsing finished, 4167 cached, 1 parsed, 197 skipped Dec 29 15:02:41 fun_in_oe: ah, technical answer: When parsing these packages raise an SkipException Dec 29 15:02:47 gremlin[it], sounds like we need a native speaker to contact the hotel .. Dec 29 15:03:07 fun_in_oe: why do they do it? Because some of these packages don't make sense for your TARGET_ARCH, MACHINE, DISTRO... Dec 29 15:03:17 matchbox-applet-startup-monitor seems to have a broken .desktop file... Dec 29 15:03:36 mb-applet-startup-monitor.desktop: error: value "Panel;Utility;MB;" for key "Categories" in group "Desktop Entry" contains an unregistered value "MB"; values extending the format should start with "X-" Dec 29 15:04:23 okay so, maybe my direction is off, so these skipped files may not be the issue Dec 29 15:04:41 in short, i'm trying to add bluetooth to the image Dec 29 15:04:50 should i change it as suggested? Dec 29 15:05:31 Shoragan: hey, where are you guys? Dec 29 15:05:34 i've tried to insert via the old way by adding bluez-utils to local.conf Dec 29 15:06:31 and dependancies don't look to be installed Dec 29 15:06:41 zecke, nemo's lounge Dec 29 15:06:58 Shoragan: do you have a free power socket over there? and a chair? Dec 29 15:07:39 chair yes, power should be possible Dec 29 15:08:00 fun_in_oe: which image is lacking bluez-utils? do you have bluetooth as machine feature? Dec 29 15:08:13 okay, then I'm running, hopefully my battery will last Dec 29 15:08:57 good luck Dec 29 15:09:34 where would i be defining it as a machine feature Dec 29 15:09:37 ? Dec 29 15:11:22 fun_in_oe: what is your machine (MACHINE="yourmachine" in your local.conf) ? Dec 29 15:13:13 fun_in_oe: bluetooth is set as MACHINE_FEATURES for machine that has bluetooth (see in conf/machine/.conf) Dec 29 15:14:35 off of the build directory? Dec 29 15:14:42 yes Dec 29 15:15:27 fun_in_oe: what is MACHINE set to in your build/conf/local.conf ? Dec 29 15:16:05 well, here's the squirley part, in conf, i only have local.conf and some other auto conf files Dec 29 15:16:11 no machine dir Dec 29 15:16:20 by 'yes' I meant the machine configuration files are in org.openembedded.dev/conf/machine/ Dec 29 15:16:34 oh Dec 29 15:16:49 as well, there is no machine variable set in local.conf Dec 29 15:16:56 Crofton|home, i think also an email to the Hotel could help Dec 29 15:17:50 fun_in_oe: hm in your build/conf/ directory, can you 'grep MACHINE *' ? Dec 29 15:18:29 fun_in_oe: MACHINE should be set in one of your file in build/conf/ Dec 29 15:19:40 i find it in auto.conf Dec 29 15:19:43 sound right? Dec 29 15:21:39 fun_in_oe: why not, but to what is it set ? Dec 29 15:22:06 tsi-gatewayexa Dec 29 15:22:37 and there is no corresponding .conf under org.openembedded.dev/conf/machine/ Dec 29 15:22:51 i'm completely stumped Dec 29 15:23:27 i've been thrown into an unknown environment tasked with adding bluetooth to the image Dec 29 15:24:05 after countless hours of reading, re-reading and attempting builds, i don't know where to look Dec 29 15:24:11 fun_in_oe: doh! Dec 29 15:24:24 fun_in_oe: do you know what image you are supposed to build ? Dec 29 15:24:42 pretty much all i've been given is the bitbake recipe for the image Dec 29 15:25:37 well, there's a .bb task created for the image, when i build it, bluetooth nor it's dependancies are rolled into it Dec 29 15:26:33 anyone seen a call for papers for the embedded room at FOSDEM? Dec 29 15:26:41 fun_in_oe: you may just need to add 'task-base-bluetooth bluez-utils' to the IMAGE_INSTALL of your .bb file Dec 29 15:26:58 fun_in_oe: but I'm afraid i'll not be able to help you much :-/ Dec 29 15:28:20 something i said? Dec 29 15:28:32 thanks btw! Dec 29 15:29:38 fun_in_oe, more likely lack of time or expertise :) Dec 29 15:36:37 fun_in_oe: no, more likely lack of time actually :) Dec 29 15:37:00 okay so, adding those to the .bb produced errors Dec 29 15:37:26 my oe environment is not current Dec 29 15:38:08 the error is: Cannot satisfy the following dependancies kernel-module-bluetooth Dec 29 15:38:50 so, how to add the bluetooth module? Dec 29 15:42:26 i've edited the defconfig file to read CONFIG_BT=y Dec 29 15:45:42 fun_in_oe: s/time/expertise/ even :( Dec 29 15:45:50 * cyrilRomain looks at kernel-module-bluetooth Dec 29 15:47:26 um ... i'm having a very senior moment Dec 29 15:47:40 that SHOULD BE CONFIG_BT=m Dec 29 15:47:43 no? Dec 29 15:47:48 srrry about caps Dec 29 15:50:06 fun_in_oe: whatever I think Dec 29 15:50:34 no, that fails too Dec 29 15:51:01 i'm having a tough time finding kernel-module-bluetooth under packages Dec 29 15:51:13 maybe i'm looking in the wrong place? Dec 29 15:51:22 fun_in_oe: yes, we just have to understand why the dependency is not met Dec 29 15:51:35 fun_in_oe: no there is no kernel-module-bluetooth .bb file Dec 29 15:51:51 ahhhh okay, so i'm less crazy than i thought Dec 29 15:51:53 fun_in_oe: because kernel module are handled differently Dec 29 15:52:53 is there anywhere you can point me for more information on how they are handled? i've been hard pressed to find such info but, that's not to say i haven't missed it in my frustration Dec 29 15:54:52 * cyrilRomain forwards the question to core developers Dec 29 15:57:40 fun_in_oe: with CONFIG_BT=y in the defconfig, and only bluez-utils (not kernel-module-bluetooth) does it fails too ? Dec 29 15:57:43 heh, cute. It sure looks like they used a N770 for the LCD in this device: http://netronome.com/web/guest/products/open_appliances Dec 29 16:00:11 omg, i hope they know what they are getting themselves into with those 770s Dec 29 16:00:27 i never had much luck w/them always flakey hardware issues Dec 29 16:06:27 fun_in_oe: well I think with CONFIG_BT=y in your defconfig, adding bluez-utils in the IMAGE_INSTALL of your image .bb is a good start Dec 29 16:07:58 fun_in_oe: but I don't know the repercution of not having bluetooth in MACHINE_FEATURES, so probably you may try to add MACHINE_FEATURES += "bluetooth" in your local.conf Dec 29 16:09:20 whois fun_in_oe Dec 29 16:09:27 done and building now Dec 29 16:09:37 * cyrilRomain crossing fingers Dec 29 16:09:44 thank you so much cyrilRomain ! Dec 29 16:10:17 more than anything, i thank you to help me regain some sanity Dec 29 16:10:34 being thrust into an environment w/o notes != fun Dec 29 16:10:57 fun_in_oe: I can understand that Dec 29 16:12:44 still failing w/kernel-module-bluetooth dependancy Dec 29 16:13:09 fun_in_oe: without kernel-module-bluetooth in the IMAGE_INSTALL of your image .bb ? Dec 29 16:15:24 yeah, only thing under image_install is task-base-bluetooth (which probably calls kernel-module-bluetooth) and bluez-utils Dec 29 16:15:42 :( Dec 29 16:16:16 is it possible there was no module available for my kernel version? Dec 29 16:16:20 2.6.12? Dec 29 16:17:44 fun_in_oe: nop, CONFIG_BT is already there in kernel 2.6.7 Dec 29 16:17:58 riight Dec 29 16:18:32 isn't there somewhere in the tree is ought to be able to manually build the kernel-module-bluetooth? Dec 29 16:19:27 I think is is worth taking a look at kernel.bbclass and module.bbclass in org.openembedded.dev/classes/ directory Dec 29 16:23:56 how do i configure bitbake to include modules from a different kernel package in the image? Dec 29 16:26:43 okay so, to clarify ... Dec 29 16:27:51 so long as i have CONFIG_BT set to y or m in defconf, bitbake ought to automagically call/create the kernel-module-bluetooth ? Dec 29 16:28:19 getting this module built is out of my hands? Dec 29 16:28:52 what i'm trying to ascertain is whether this may just be a matter of updating the kernel/oe environment Dec 29 16:28:58 fun_in_oe, that is what I expect Dec 29 16:29:11 look in the work directory to see what modules where built Dec 29 16:29:55 fun_in_oe: hmm can you please pastebin your image's .bb ? Or if you don't want, take a look at the packages/images/mokogateway-image.bb Dec 29 16:30:31 hope it help Dec 29 16:31:04 where should i see them in the work dir? Dec 29 16:34:36 qork/machine name / linux I think Dec 29 16:34:46 where the kernel source gets unpacked and built Dec 29 16:39:10 okay, i see a bunch of modules built Dec 29 16:41:01 so they should all have corresponding ipks, right? Dec 29 16:45:10 i would think so ... $find ./ -name kernel-module*.ipk Dec 29 16:45:11 ? Dec 29 16:47:26 look in tmp/deploy..... Dec 29 16:47:37 there should be an ipk directory Dec 29 16:48:04 ipks are in the tmp/deploy//ipk/ Dec 29 16:48:19 yep Dec 29 16:48:23 ipks are in the tmp/deploy//ipk// even Dec 29 16:48:53 but so, now how to force an ipk build from modules? Dec 29 16:48:56 fun_in_oe: btw when you change the defconfig you have to 'bitbake -c rebuild theKernel' (I'm not sure bitbake takes defconfig changes into account after a kernel is already built) Dec 29 16:49:14 I am pretty sure it will pachake up all modules that come from kernel build Dec 29 16:49:55 when you say "theKernel" what do you mean? Dec 29 16:50:11 the name of the bb file Dec 29 16:50:14 i've been doing $bitbake -crebuild Dec 29 16:50:37 that won't work Dec 29 16:50:39 fun_in_oe: no, just the kernel Dec 29 16:50:48 ahhhhhh! ic Dec 29 16:50:52 THANK YOU Dec 29 16:50:54 -c rebuild only rebuilds the package name you give it Dec 29 16:51:17 alternatively, you can increment PR in the bb file after changing defconfig Dec 29 16:51:32 after the kernel build successfully, we can 'bitbake yourImage' again to generate again the rootfs Dec 29 16:52:01 okay so, how to determine whatis "theKernel"? Dec 29 16:52:37 jilles: exact, except that incrementing PR requires more space on disk :} Dec 29 16:53:05 fun_in_oe: in the tmp/work// directory Dec 29 16:53:39 fun_in_oe: there is the kernel directory Dec 29 16:54:30 fun_in_oe: probably linux-something or xanadux-something Dec 29 16:56:59 okay, under work//linux2.6.12 i have a defconfig file that does not have CONFIG_BT enabled Dec 29 16:58:12 fun_in_oe: then set CONFIG_BT=y to the right defconfig located in the packages/linux/linux2.6.12/ Dec 29 16:59:20 * cyrilRomain think you might avoid calling bitbake by changing directly the defconfig you found, but this is more error prone Dec 29 17:03:50 * fun_in_oe crosses fingers, building kernel Dec 29 17:07:56 w00t! Dec 29 17:08:07 looks lie the module built Dec 29 17:08:17 ok Dec 29 17:09:30 you can check that by looking into the work//linux2.6.12/image/ Dec 29 17:10:16 then 'bitbake ' Dec 29 17:10:27 now i should be able to $bitbake right? Dec 29 17:10:31 right Dec 29 17:10:57 except I fear the dependency issue will be still there .. Dec 29 17:11:00 * fun_in_oe gets giddy as he sees dependancy packages being pulled Dec 29 17:11:11 ? Dec 29 17:11:47 well, blueprobe was pulled/built Dec 29 17:12:35 fun_in_oe: yes, and bluez-utils will be correctly included in your image Dec 29 17:13:05 * cyrilRomain is just wondering what task-base-bluetooth actually do Dec 29 17:15:23 okay, completed w/o errors (that i can see) Dec 29 17:15:35 fun_in_oe: the image built ? Dec 29 17:15:55 looks like it Dec 29 17:15:58 :) Dec 29 17:16:20 any quick way to check to see what all is included in the image? Dec 29 17:16:38 in tmp/rootfs directory Dec 29 17:16:41 i think i had been $ar x ing the ipk Dec 29 17:18:03 ipks can still be installed latter with ipkg (at the contition your image includes ipkg of course) Dec 29 17:18:24 saaweet Dec 29 17:18:36 bbiab, time to fetch and push the image Dec 29 17:18:46 thanks SOOOO much cyril! Dec 29 17:18:53 fun_in_oe: you welcome Dec 29 17:45:12 03jluebbe 07org.oe.dev * r35be16b6... 10/ (3 files in 3 dirs): linux-openmoko-devel: add a recipe for the current patch queue Dec 29 17:45:12 03jluebbe 07org.oe.dev * r35be16b6... 10/ (3 files in 3 dirs): linux-openmoko-devel: add a recipe for the current patch queue Dec 29 17:45:39 03jluebbe 07org.oe.dev * r35be16b6... 10/ (3 files in 3 dirs): linux-openmoko-devel: add a recipe for the current patch queue Dec 29 17:45:39 03jluebbe 07org.oe.dev * r35be16b6... 10/ (3 files in 3 dirs): linux-openmoko-devel: add a recipe for the current patch queue Dec 29 17:46:32 03jluebbe 07org.oe.dev * r35be16b6... 10/ (3 files in 3 dirs): linux-openmoko-devel: add a recipe for the current patch queue Dec 29 17:48:07 03jluebbe 07org.oe.dev * r35be16b6... 10/ (3 files in 3 dirs): linux-openmoko-devel: add a recipe for the current patch queue Dec 29 18:03:25 ~kick CIA-35 Dec 29 18:03:25 * ibot kicks CIA-35 Dec 29 18:03:26 ow Dec 29 18:03:38 That'll teach it to spam the channel. :P Dec 29 18:04:03 :D Dec 29 18:15:47 ~seen RP Dec 29 18:15:48 rp is currently on #maemo #handhelds #oe #openmoko #openzaurus. Has said a total of 100 messages. Is idling for 5h 39m, last said: 'I can't decide if I want to read the context for this or not...'. Dec 29 18:25:15 cool! looks as though bluetooth is in the kernel except, none of the bluetooth packages seem to have been installed Dec 29 18:25:46 thanks again cyrilRomain ~ Dec 29 18:25:48 ! Dec 29 18:36:38 fun_in_oe: try adding task-base-bluetooth in your image .bb Dec 29 18:37:15 unbelievable, now there's no eth0 interface whereas before there was Dec 29 18:37:38 * fun_in_oe is about to tell this outfit to take their fubar build env and shove it Dec 29 18:37:50 :( Dec 29 18:38:12 i thought the WHOLE point behind using bitbake was to streamline building Dec 29 18:38:18 fun_in_oe: can you pastebin your image .bb file please ? Dec 29 18:38:38 seems as though whoever set up this env had NO clue about bitabake/oe Dec 29 18:38:44 (not that i have much more) Dec 29 18:38:58 fun_in_oe, what build env are you talking about? Dec 29 18:39:12 the one i've been forced to use Dec 29 18:39:21 OE? Dec 29 18:39:48 well, oe as layed out by someone Dec 29 18:39:55 :) Dec 29 18:40:06 if you find something better, let us know Dec 29 18:40:13 not knocking OE Dec 29 18:40:26 but I think if you invest some time now, you will be reqrded ... Dec 29 18:40:52 good, that is what I wanted to clarify, it is easy to misunderstand people in irc Dec 29 18:40:56 knocking the implementation of it by someone else who didn't take ANY notes Dec 29 18:41:13 especially given not everyone here is a native english speaker Dec 29 18:41:45 how in the hell does an ether interface go away after adding bluetooth to tthe kernel?!?! Dec 29 18:42:02 cyrilRomain: i'd love to but, it's on another machine Dec 29 18:42:50 back to square one ... now to figure out how to add ethernet back Dec 29 18:42:51 fun_in_oe, any strange boot messages? Dec 29 18:42:55 * fun_in_oe sighs Dec 29 18:43:28 sadly, OE does not do everything for us :) Dec 29 18:43:43 that's a downer :( Dec 29 18:44:01 no but, i CAN do some of the stuff i'm getting hung up on Dec 29 18:44:09 provided someone configs it right Dec 29 18:44:25 if i'm not mistaken Dec 29 18:44:53 yeah, it is just a matter of figuring out why the interface disappeared Dec 29 18:45:31 * cyrilRomain hopes the kernel config was not previously changed and built by hand from the work directory .. Dec 29 18:45:36 there is a complaint upon boot Dec 29 18:46:25 hvontres|home: hi Dec 29 18:49:07 Configuring network interfaces ... cat: /etc/tsi//eth*mac: No such file or directory Dec 29 18:49:32 fun_in_oe: hmm except the CONFIG_BT=y changes, you only added 'bluez-utils' to the existing IMAGE_INSTALL right ? Dec 29 18:50:03 yep, just what we covered here Dec 29 18:50:14 which is what is SO frustrating Dec 29 18:50:58 the only reason i can think of that the eth0 int would go away is so weirdness in the env Dec 29 18:51:02 * fun_in_oe shakes fist Dec 29 18:51:53 fun_in_oe, what device are you working with? Dec 29 18:52:06 tsi-gatewayexa Dec 29 18:52:56 unbelievable!!!! Dec 29 18:53:04 what is that? Dec 29 18:53:08 * cyrilRomain could not find info for that hardware, except http://slightlyorange.net/wiki/index.php5?title=Wearable:Arm7 and http://slightlyorange.net/wiki/index.php5?title=Wearable:GatewayExpress , which relates to OE Dec 29 18:54:25 yeah, those pages are of no help Dec 29 18:54:31 thanks though! Dec 29 18:54:31 yep :) Dec 29 18:55:56 okay, i'm calling it quits for the year Dec 29 18:56:07 talk to us next year :) Dec 29 18:56:12 Happy New Year! Dec 29 18:56:16 this project i got put on is such bullshit and i think i'm going to let the client know it Dec 29 18:56:30 Where are you? Dec 29 18:56:53 geo? US Dec 29 18:56:57 same here Dec 29 18:57:01 Blacksburg, VA Dec 29 18:57:26 k, thanks again everyone!!! Dec 29 18:57:29 be well! Dec 29 18:59:55 hi all Dec 29 19:28:18 Crofton|home: you are the gumstix maintainer? Dec 29 19:33:49 sort off :) Dec 29 19:34:10 gumstix is working on moving to OE and they have some stuff in svn also Dec 29 19:34:42 I'm behind moving stuff from that into .dev due tio last months defense and remodeling efforts Dec 29 19:34:52 er this motnhs defense :) Dec 29 19:45:01 http://jwboyer.livejournal.com/19484.html Dec 29 19:51:11 Crofton|home: ok, just wanted to say I modified the 2.6.23 kernel recipe lately, gumstix is one of the users I am aware of. Dec 29 19:55:46 ok Dec 29 19:56:11 sakoman, works of a copy of .dev Dec 29 19:56:20 and has his own kernel recipe Dec 29 19:56:39 He plans to update his copy of .dev to teh angstrom stable branch at some point Dec 29 20:00:29 RP: hello... I was AFK at the home center :) Dec 29 20:09:18 hi Dec 29 20:10:00 what is the expected behaviour by fetch error e.g. SRC_URI down ? Dec 29 20:10:20 I'm stumbling in Fetch failed: http://www.zlib.net/zlib-1.2.3.tar.bz2 Dec 29 20:10:36 and I provided the sources in my source dir Dec 29 20:11:05 should I create myself the md5? Dec 29 20:11:47 I thought if the sourcefile exists the md5sum must be calculated over there and not on the SRC_URI Dec 29 20:12:44 ant|away: If you provided the sources yourself, you must also provide the MD5. And yes, if you create the md5sum yourself it does rather defeat the purpose. Dec 29 20:12:58 ok, let see Dec 29 20:18:22 oops NOTE: The MD5Sums did not match. Wanted: 'dee233bf288ee795ac96a98cc2e369b6' and Got: '733e25769de02aacaf84b4e230db21c7' Dec 29 20:19:23 uh-oh. You've been trojaned! Dec 29 20:19:40 :p Dec 29 20:20:15 http://prdownloads.sourceforge.net/libpng/zlib-1.2.3.tar.bz2 Dec 29 20:21:22 +mhentges@mhcln04:~/tmp >md5sum zlib-1.2.3.tar.bz2 Dec 29 20:21:22 dee233bf288ee795ac96a98cc2e369b6 zlib-1.2.3.tar.bz2 Dec 29 20:21:53 damaged d/l on your end Dec 29 20:22:03 dee233bf288ee795ac96a98cc2e369b6 is correct on my copies. Dec 29 20:22:36 dee233bf288ee795ac96a98cc2e369b6 here too Dec 29 20:22:41 [arklinux@localhost sources]$ md5sum zlib-1.2.3.tar.bz2 Dec 29 20:22:46 Crofton: I've booted base-image on at least three different ixp4xx OE devices Dec 29 20:22:49 733e25769de02aacaf84b4e230db21c7 zlib-1.2.3.tar.bz2 Dec 29 20:23:04 I search another source... Dec 29 20:23:14 delete zlib-1.2.3.tar.* from your downloads directory, and rerun the fetch. Dec 29 20:23:31 http://www.zlib.net/zlib-1.2.3.tar.bz2' failed Dec 29 20:23:44 It redirected me... Dec 29 20:23:54 http://easynews.dl.sourceforge.net/sourceforge/libpng/zlib-1.2.3.tar.bz2 Dec 29 20:24:20 dee233bf288ee795ac96a98cc2e369b6 zlib-1.2.3.tar.bz2 <==== md5sum is correct from that source. Dec 29 20:25:09 it failed three times...no redirection Dec 29 20:25:27 doing now 4th try Dec 29 20:25:48 ERROR: TaskFailed event exception, aborting Dec 29 20:26:12 Resolving www.zlib.net... 0.0.0.0 Dec 29 20:26:20 Connecting to www.zlib.net|0.0.0.0|:80... failed: Connection refused. Dec 29 20:26:31 ant_: wget it from mwester source :) Dec 29 20:26:52 yes, I know, but this bug should have been already resolved Dec 29 20:27:02 or I think so... Dec 29 20:27:19 adding new mirrors Dec 29 20:28:50 ant_: but there is a certain number of retry before falling back to another mirror Dec 29 20:29:04 maybe it is 5 (?) Dec 29 20:29:54 rwhitby, what is in their fstab? Dec 29 20:30:54 I wondering what exactly is making busybox mount choke Dec 29 20:31:12 the Angstrom ones just use the default fstab Dec 29 20:32:46 nothing jumps out at me comparing them Dec 29 20:32:47 yep. unchanged from the default. Dec 29 20:32:59 (just looked on the real device) Dec 29 20:33:04 bother :) Dec 29 20:33:43 oh well, back to house wiring ... Dec 29 20:34:06 ant_: yes, in bitbake.conf you can see 'wget -t 5', so it tries 5 times Dec 29 20:35:29 cyrilRomain: no, the build breaked 4 times... Dec 29 20:35:39 cyrilRomain: here Dec 29 20:36:07 cyrilRomain: and I'm connected with the same host, so DNS is ok Dec 29 20:36:17 cyrilRomain: here Dec 29 20:38:43 ... Dec 29 20:39:01 cyrilRomain: don't ever update your working Gentoo... Dec 29 20:40:00 cyrilRomain: no way to have qemu compiled Dec 29 20:40:31 cyrilRomain: I have one last try: I saw in your conf ccache disabled Dec 29 20:40:48 cyrilRomain: FWIW Dec 29 20:41:23 ant_: disabling ccache will not solve the problem Dec 29 20:41:40 cyrilRomain: I guessed that! Dec 29 20:42:08 cyrilRomain: probably linux-headers 2.6.23 Dec 29 20:42:31 cyrilRomain: you use .22 isn't ? Dec 29 20:43:03 ant_: kernel ? headers ? Dec 29 20:43:25 cyrilRomain: both Dec 29 20:43:36 ant_: yes 2.6.22-r2 Dec 29 20:44:27 cyrilRomain: but vapier @Gentoo admitted the problem and masked older gcc3 Dec 29 20:44:41 cyrilRomain: so must be distro-specific Dec 29 20:44:46 :( Dec 29 20:45:45 cyrilRomain: you know how long it'd take to recompile ALL with old headers... Dec 29 20:45:48 ant_: does then vapier gave you an advice to build qemu ? Dec 29 20:46:01 cyrilomain: ...so I installed ArkLinux in another part Dec 29 20:46:16 cyrilRomain: not to me, but read here: Dec 29 20:47:07 http://packages.gentoo.org/package/sys-devel/gcc Dec 29 20:47:33 from today ;-) Dec 29 20:48:05 your arch is hppa ? Dec 29 20:48:22 no, x86 Dec 29 20:48:31 Core2 Duo Dec 29 20:48:36 32 bit Dec 29 20:51:33 ant_: so gcc-3.3.6-r1 is available again Dec 29 20:51:44 OE_REVISION = "...54cafe77..." - and having coffee Dec 29 20:51:57 cyrilRomain http://sources.gentoo.org/viewcvs.py/gentoo-x86/sys-devel/gcc/ChangeLog?view=markup Dec 29 20:52:29 cyrilRomain: BTW I failed with a binary gcc-3.3.6 Dec 29 20:52:49 cyrilRomain: so my hope are not so... Dec 29 20:53:36 ant_: :-/ Dec 29 20:54:36 ant_: and no much luck with ArkLinux ? Dec 29 20:54:41 cyrilRomain: what really disturbs me is that the gcc-3.3.5-ark is ok Dec 29 20:54:50 -( Dec 29 20:55:25 cyrilRomain: ArkLinux is very very quick to install for OE development Dec 29 20:55:45 cyrilRomain: I'm finishing an updated how-to Dec 29 20:55:53 ArchLinux too =) Dec 29 20:56:11 cyrilRomain: never tested :-) Dec 29 20:57:01 cyrilRomain: bero, the father of ArkLinux was/is an OE developer I think Dec 29 20:57:30 cyrilRomain: or at least friend of the cousin of Koen..;-))) Dec 29 20:57:38 :) Dec 29 20:58:04 shame it did'nt build ArkLinux with OE then ;-p Dec 29 20:58:09 s/it/he/ Dec 29 20:58:53 cyrilRomain: NOTE: package qemu-native-0.9.0+cvs20070613: completed Dec 29 20:59:01 cyrilRomain:- ))) Dec 29 20:59:49 ant_: I think he is/was not a real OE dev, otherwise he would never choose rpm for packages ;-p Dec 29 21:00:05 ant_: cool, so you have qemu thanks to OE :) Dec 29 21:00:08 cyrilRomain: there is synaptic also Dec 29 21:00:55 cyrilRomain: oops Kinaptic (KDE) Dec 29 21:01:04 don't know Dec 29 21:01:19 cyrilRomain: it's the apt-get stuff Dec 29 21:01:39 cyrilRomain: emerge was the first love, you 'll never forget it ! Dec 29 21:01:41 a new one ? Dec 29 21:03:07 ant_: I still love emerge and bitbake Dec 29 21:03:24 cyrilRoman: emerge vs. apt : 1-0 Dec 29 21:03:32 ant_: but I'm looking for a more robust package management Dec 29 21:03:59 cyrilRomain: more robust than what? Dec 29 21:04:05 ant_: because they are all based on a destructive upgrade model Dec 29 21:04:26 cyrilRomain: ah, I see, you need feeds Dec 29 21:04:34 ant_: something that does not break your distro and does not hurt brain ) Dec 29 21:04:36 ;) Dec 29 21:05:06 with emerge, upgrade are not atomic, i.e. your system can be temporary in a broken state Dec 29 21:05:25 and sometimes not only temporary :-( Dec 29 21:05:29 cyrilRomain: I never had problems with Gentoo (in 6 years) Dec 29 21:05:50 ant_: I update my system constantly and never had *big* problem Dec 29 21:05:52 cyrilRomain: upgraded from gcc-2.95 to 4.x Dec 29 21:05:59 cyrilRomain Dec 29 21:06:11 cyrilRomain: during the years Dec 29 21:06:23 but sometimes I need to revdeprebuild, e.g. mplayer is broken until the revdeprebuild finish Dec 29 21:06:27 cyrilRomain soon comed the bomb: the new baselayout 2 Dec 29 21:06:32 and this is not acceptable for a prodution system Dec 29 21:06:42 cyrilRomain: hardened ? Dec 29 21:06:47 ant_: no Dec 29 21:07:10 cyrilRomain: you can choose a very old and stable profile Dec 29 21:07:51 ant_: I prefer to have up-to-date version of software Dec 29 21:08:08 cyrilRomain: he he Dec 29 21:08:34 cyrilRomain: if we talk about postfix, samba, php, apache, YES Dec 29 21:08:40 ant_: but with a better package managment, one could have up-to-date system without (even temporary) breakage Dec 29 21:09:32 cyrilRomain: there are nice scripts for partial rebuild Dec 29 21:09:54 ant_: 'partial rebuild' ? Dec 29 21:10:18 cyrilRomain: yes, avoiding emerge -e system Dec 29 21:10:30 cyrilRomain: 2 times Dec 29 21:10:38 ant_: you speak about emwrap.sh for example ? Dec 29 21:10:50 cyrilRomain: yes Dec 29 21:11:10 cyrilomain: and don't miss udept Dec 29 21:11:24 ant_: yes it is honestly a good attempt to make thing better, but still not the solution Dec 29 21:11:50 cyrilRomain: I think Gentoo is the state-of-the-art Dec 29 21:12:09 cyrilRomain: didn't find any better Dec 29 21:12:11 ant_: I think just a part of it :) Dec 29 21:12:33 cyrilRomain: we speak for the PRESENT Dec 29 21:12:36 ant_: I mean as for package managment system Dec 29 21:13:52 ant_: ant_ for the present, there is http://nix.cs.uu.nl/nixos/ which is IMO _one_ damn good attempt to fix all the present package managemnt system Dec 29 21:14:00 cyrilRomain: we are going OT... somebody will *curse us Dec 29 21:14:33 cyrilRomain: interesting BTW Dec 29 21:14:45 ant_: you right Dec 29 21:15:14 cyrilRomain: he he Dec 29 21:15:47 * cyrilRomain know package management is out of topic in OE, but not when ipkg upgrade show its limit, hehe Dec 29 21:16:10 cyrilomain: the topic is TABOO :-) Dec 29 21:17:08 well, back to code :) Dec 29 21:17:26 cyrilRomain: was a pleasure, see you Dec 29 21:43:22 cyrilRomain: no, just wait for Linux to have union mounts, and base the package system on that. Like QNX does. Dec 29 21:50:10 * cyrilRomain loves how OE split packages :) Dec 29 22:40:16 ~lart C for its inherent lack of correctness Dec 29 22:40:17 * ibot beats C severely about the head and shoulders with a rubber chicken for its inherent lack of correctness Dec 29 23:05:14 cyrilRomain: ;-) Dec 29 23:05:28 cyrilRomain: NOTE: preparing tree for binary locale generation Dec 29 23:05:41 cyrilRomain: qemu works ! Dec 29 23:07:42 * qemu galore: ant is now generating ALL locales Dec 29 23:07:44 ant_: ah glad to read that :) Dec 29 23:08:41 cyrilRomain: once finished 'll switch again to Gentoo and retry qemu with the patched gcc-3.3.6 Dec 29 23:09:48 cyrilRomain: released today (the binaries I installed were older !) Dec 29 23:10:15 cyrilRomain: cross the finger.. Dec 29 23:10:25 * cyrilRomain crosses fingers for ant_ Dec 29 23:10:37 cyrilRomain: I'll let you know (tomorrow...) Dec 29 23:10:48 cheers - bonne nuit Dec 29 23:10:57 ant_: thx, bonne nuit :) **** ENDING LOGGING AT Sun Dec 30 02:59:57 2007