**** BEGIN LOGGING AT Wed Aug 12 02:59:57 2009 Aug 12 03:00:52 Hi! Aug 12 03:01:42 I get 'arch/arm/mach-pxa/hx2750.c:446: error: "pxa_init_irq" undeclared here (not in a function)' when compiling linux-rp for hx2000 Aug 12 03:01:48 what can I do about it? Aug 12 03:02:09 J13R: change the .config? Aug 12 03:02:24 J13R: hx2750 should DEPEND/SELECT some pxa interrupt handling... Aug 12 03:05:08 zecke: how do I change the config? won't it be overridden by unpacking next time I launch the compile? Aug 12 03:09:44 J13R: right, you can do something like bitbake -cmenuconfig linux-rp, enable whatever is needed for pxa_init_irq, and copy the config back to the "origin" Aug 12 03:10:09 pxa seems to be enabled though Aug 12 03:10:18 J13R: in general the config item for hx2750 should select the PXA dependencies it needs, I have no idea why this is not the case (I'm not too familiar with it) Aug 12 03:10:20 but the function really does not exist ... Aug 12 03:10:30 J13R: okay, so it is just outdated code :} Aug 12 03:11:27 there is a MACHINE_START section with .init_irq = pxa_init_irq Aug 12 03:11:51 but the only functions I can find are pxa_init_irq_low() and high and gpio Aug 12 03:12:21 as well as pxa_init_irq which calls all three Aug 12 03:13:36 J13R: search for pxa27x_init_irq Aug 12 03:13:38 IIRC hx2750 was already PXA270 Aug 12 03:14:19 yeah, I think in MACHINE_START it should be .init_irq = pxa27x_init_irq Aug 12 03:14:34 em-x270 has that Aug 12 03:15:23 zecke: how can I make this code change so it isn't overridden? Aug 12 03:15:28 i agree, care to make a patch and send it to RP? Aug 12 03:16:03 zecke: no problem, will do (once it compiles) Aug 12 03:16:30 J13R: bitbake -cdevshell linux-rp (if in angstrom) Aug 12 03:16:34 J13R: quilt new make-hx2750.patch; quilt edit arch/arm/mach-pxa/hx2750.c; change the code; quilt refresh Aug 12 03:16:53 J13R: then you can copy the patch to oe/recipes/linux/linux-rp/ and change the .bb file to include the patch Aug 12 03:17:11 yes, angstrom Aug 12 03:17:46 J13R: so if you send a a patch to the OE list, it will show up here: http://patchwork.openembedded.org/project/openembedded/list/, which means I will be likely to land it Aug 12 03:17:47 hehe Aug 12 03:17:53 gnome-terminal: command not found Aug 12 03:19:30 J13R: hehe, look at bitbake.conf and search for TERMCMD, you can set TERMCMD, TERMCMDRUN to use xterm, konsole... Aug 12 03:21:13 I do not have x Aug 12 03:21:45 J13R: look at the file ;) Aug 12 03:21:59 J13R: you will have shell or screen Aug 12 03:21:59 yeah Aug 12 03:26:29 it wants to create the directory "/.screen-profiles" Aug 12 03:26:35 rrr Aug 12 03:27:18 you don't have $HOME set? then use shell :) Aug 12 03:28:06 I don't think shell belongs to the group of options, it has no "xxxrun" Aug 12 03:28:38 i have $HOME set Aug 12 03:28:52 right Aug 12 03:30:09 J13R: you can always go to the WORKDIR and invoke the run.do_compile script by hand Aug 12 03:33:14 * J13R tries something evil Aug 12 03:52:10 "See the note in gcc/gcc_3.4.0.oe" Aug 12 03:52:12 hmmm Aug 12 03:52:45 must have been an important note, to have so many references to it but not exist Aug 12 03:55:11 grg: where do you see this? Aug 12 03:55:21 grg: note that the extension .oe is quite old too Aug 12 03:55:24 mostly in the glibc recipes Aug 12 03:55:27 one has to wonder how this ever compiled Aug 12 03:55:37 uclibc and eglibc too Aug 12 03:55:52 J13R: hx2750? well, it is not too mainstream hardware anymore :) Aug 12 03:57:04 grg: hehe :} Aug 12 03:57:31 zecke: really? iPAQs? Aug 12 03:58:13 J13R: from what I see... yes Aug 12 04:07:29 zecke, http://www.lemote.com/english/index.html have cheap mips hardware Aug 12 04:07:57 i'm not familiar with their products, but the loongson chips seem to be gaining momentum recently Aug 12 04:08:46 grg, are you aware of any vendors for those boards/products? Aug 12 04:09:08 sorry, no Aug 12 04:09:39 darn Aug 12 04:09:42 i work on an ingenic jz4730 based device Aug 12 04:09:50 but we're not selling the devices yet Aug 12 04:09:55 i found a dual mips atx mobo, but got no response from the vendor Aug 12 04:09:58 oh neat Aug 12 04:10:12 you can get dev boards from Ingenic i suppose. not sure how much they are Aug 12 04:10:55 probably $$ Aug 12 04:11:42 :-| Aug 12 04:12:39 zecke: so I edited in the work folder, how do I make a git diff? Aug 12 04:16:55 m4t, this mob sell the lemote notebook http://www.tekmote.nl/ Aug 12 04:16:55 J13R: I proposed to use quilt :) Aug 12 04:17:05 J13R: what you are operating on is not a git repository Aug 12 04:17:12 zecke: I didn't get cdevshell to run Aug 12 04:17:17 05:16 < zecke> J13R: quilt new make-hx2750.patch; quilt edit arch/arm/mach-pxa/hx2750.c; change the code; quilt refresh Aug 12 04:17:26 this is unrelated to devshell :) Aug 12 04:17:31 ah ok Aug 12 04:17:37 never used quilt before Aug 12 04:17:39 will try Aug 12 04:17:52 J13R: you will have to use quilt edit before you have done any changes :} Aug 12 04:18:49 I copied away my linux folder, will rebuild and do it then Aug 12 04:19:28 J13R: cool Aug 12 04:19:32 http://s153.photobucket.com/albums/s223/dsobodash/Lemote%20Yeeloong/DSC_0048.jpg Aug 12 04:20:29 grg: ah thanks, I forgot about them Aug 12 04:21:26 grg: stupid question, MIPS64 supports all mip32 instructions too? Aug 12 04:21:32 yeah Aug 12 04:21:52 grg: so like with amd64 I can run a mips32 userspace? Aug 12 04:21:59 yup Aug 12 04:22:35 grg: thanks a lot Aug 12 04:23:48 zecke, let me know if you find a cheap supplier :) Aug 12 04:26:06 there is also another loongson netbook called 'Gdium' Aug 12 04:26:25 grg: well, I'm in Taiwan... my supplier probably won't help you Aug 12 04:29:38 for the USAians http://cgi.ebay.com/NEW-Gdium-Netbook-10-Linux-GDNBL10USK006_W0QQitemZ350233519382QQcmdZViewItemQQptZAU_comp_laptop?hash=item518b8b6516&_trksid=p3286.c0.m14 Aug 12 04:30:35 grg, im sure you've heard of 'onda'? Aug 12 04:30:51 m4t, nope Aug 12 04:31:07 oh, they use ingenic chips in their media players Aug 12 04:33:19 http://www.compware.com/_ATX_MIPS.html is the one i was trying to find more info about Aug 12 04:33:27 doesnt seem to really exist outside of that page Aug 12 04:37:53 interesting looking board. i wonder what people buy that for Aug 12 04:38:25 aahh ... I messed up a little Aug 12 04:38:36 zecke: in which directory is the patch now, and how do I reset Aug 12 04:38:45 (after quilt refresh) Aug 12 04:39:19 ah the .pc folder Aug 12 04:40:33 J13R: no patches are in patches/ Aug 12 04:40:43 J13R: the .pc folder is internal Aug 12 04:50:09 03Graham Gower 07org.openembedded.dev * rdd72daef5f 10openembedded.git/recipes/dbus/ (5 files): (log message trimmed) Aug 12 04:50:09 dbus: fix dbus-daemon segfault on mipsel with recent toolchain Aug 12 04:50:09 dbus-daemon segfaults on mipsel when built with -fPIE/-pie, as outlined in Aug 12 04:50:09 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=532821 Aug 12 04:50:10 Removes PIE compilation from the dbus-daemon Makefile as a workaround. Aug 12 04:50:12 Signed-off-by: Graham Gower Aug 12 04:50:14 Acked-by: Khem Raj Aug 12 04:50:39 grg: what is more common, mips or mipsel? Aug 12 04:50:54 zecke, dunno Aug 12 04:51:11 should I send a normal mail for the patch or does some git command do everything? Aug 12 04:51:57 zecke, probably mipsel Aug 12 04:51:59 J13R: you can do a git commit and then a git format-patch Aug 12 04:55:54 oh and OE only has mipsel support? Aug 12 04:57:56 zecke, i think it has both Aug 12 04:58:25 i only have mipsel devices Aug 12 04:59:37 grg: git grep for "mips" only turns up lines with TARGET_ARCH mipsel :} Aug 12 05:00:17 blubb blubb Aug 12 05:00:33 * grg shrugs Aug 12 05:06:37 zecke, m4t another cheap mipsel device is the wrt54g linksys modems/routers Aug 12 05:09:21 yea i have one Aug 12 05:09:50 ive got serial port headers and a max232 circuit as well, i just have been waiting on hooking it all together Aug 12 05:09:56 as it is my primary link to the inet Aug 12 05:10:22 nice Aug 12 05:10:22 * m4t curious to know where his problems with 2.6 are Aug 12 05:10:39 i got bootloader output on an older linksys too Aug 12 05:10:57 i can't play with mine as it is used as the modem for all my housemates Aug 12 05:11:06 yea. its my uplink Aug 12 05:11:34 its a nicer model too, been honing my soldering skills Aug 12 05:11:38 it seems to crawl to a halt when everyone torrents too much.... can't handle the packets :) Aug 12 05:11:41 got to find a time when my hands are steady Aug 12 05:12:05 grg hrm, mine maxes at about 600kb/s, i think it is the wifi limit though, not the hw Aug 12 05:12:16 its the wrtsl54gs, w/usb Aug 12 05:12:26 grg: right, but generally it doesn't have plenty of ram... I search for more like a desktop system (doesn't need to have a screen though) Aug 12 05:12:26 266/32mb/8mb Aug 12 05:12:58 hmm.. mine does 800kb/s ok Aug 12 05:13:44 i think it might peak around that Aug 12 05:14:05 its a wifi uplink though, not ethernet to a cable/dsl Aug 12 05:14:22 we had a nasty heat wave earlier this year, 5 days over 40C. The linksys didn't like that at all :) kept hanging Aug 12 05:14:30 what model is it? Aug 12 05:15:04 wrt54g, not sure of the revision. maybe C... its at least 4 years old Aug 12 05:15:24 cool. theyre nice cheap mips boards Aug 12 05:15:52 the routerstations have caught my eye, gigabit switch, 128mb of ram, and 680-800mhz broadcom chips, 3 mini-pci Aug 12 05:16:01 on the new model, i dont know if its shipping yet Aug 12 05:16:06 ~80$ Aug 12 05:16:30 very nice Aug 12 05:17:08 i was looking for mini-pci sata and video options for a while, planning on building a full featured computer out of one Aug 12 05:22:56 http://thread.gmane.org/gmane.comp.handhelds.openembedded/25674 Aug 12 05:23:44 oh well **** ENDING LOGGING AT Wed Aug 12 05:26:18 2009 **** BEGIN LOGGING AT Wed Aug 12 05:28:14 2009 **** ENDING LOGGING AT Wed Aug 12 05:32:06 2009 **** BEGIN LOGGING AT Wed Aug 12 05:34:16 2009 Aug 12 06:02:44 hrw|gone: ping? Aug 12 06:02:49 J13R: thanks Aug 12 06:25:53 morning Aug 12 06:47:30 zecke: pong for a moment Aug 12 06:49:44 hrw|gone: in patchwork there are ~10 of your patches for the stable branch, were they merged? Aug 12 06:51:14 nope Aug 12 06:51:38 I have to ping stable maintainers Aug 12 06:51:55 or even drop them and create newer ones Aug 12 06:53:06 * hrw|gone -> off Aug 12 06:57:47 any clues as to why the glibc 2.9 recipes use glibc-ports 2.7 ? Aug 12 06:58:50 glibc-ports 2.9 works for me on mipsel, while 2.7 does not. Of course, I can't test all the arm specific hacks... Aug 12 06:59:19 i'm sure most of those hacks aren't necessary for glibc-ports-2.9 Aug 12 06:59:58 grg: no idea+ Aug 12 07:00:56 grg: are all your posted mips patched merged? did i miss one? Aug 12 07:01:15 i'll check Aug 12 07:02:35 pb__: awake yet? Aug 12 07:03:21 zecke, yep it looks like you got them all. Cheers Aug 12 08:32:39 03Thomas Zimmermann 07org.openembedded.dev * rc80831a276 10openembedded.git/ (4 files in 2 dirs): Aug 12 08:32:39 python: add missing dependencies Aug 12 08:32:39 * Added mimetools.py and rfc822.py to python-netclient Aug 12 08:32:39 * Added python-json package Aug 12 08:32:39 * Let python-email depend on python-netclient, because email/utils.py depends on base64.py Aug 12 08:32:40 NOTE: Fixes http://trac.shr-project.org/trac/ticket/549 Aug 12 08:45:22 03Dirk Opfer 07org.openembedded.dev * r5814ae4cde 10openembedded.git/recipes/gcc/ (179 files in 6 dirs): Aug 12 08:45:22 gcc: add new gcc version 4.3.4 Aug 12 08:45:22 Tested on a armv5te platform using thumb instruction set Aug 12 08:45:23 03Dirk Opfer 07org.openembedded.dev * r9e0d36bab8 10openembedded.git/recipes/fuse/ (4 files in 3 dirs): fuse: update to 2.7.4 Aug 12 08:45:24 03Dirk Opfer 07org.openembedded.dev * rb5b13fcd2a 10openembedded.git/recipes/owfs/owfs_2.7p23.bb: owfs: update to 2.7p24 Aug 12 08:45:27 03Dirk Opfer 07org.openembedded.dev * r110422e26f 10openembedded.git/recipes/linknx/ (5 files in 2 dirs): Aug 12 08:45:30 linknx: Fix crash on startup Aug 12 08:45:32 - restructure recipe and add .inc Aug 12 08:45:34 - Add patch to fix a crash on startup during static initialization of a map Aug 12 08:45:36 - Add startup script Aug 12 08:45:38 - Enable curl and smtp Aug 12 08:45:40 - Add recipe for cvs version Aug 12 08:53:59 how does the pstaging work? I can't find it in the docs Aug 12 08:54:41 there are packages generated for staging files? can I use it to do partial builds backed by previous staging packages? how? ;) Aug 12 09:09:04 morning all Aug 12 09:10:56 morning RP Aug 12 09:11:01 RP: ho Aug 12 09:11:05 lrg: and hey! Aug 12 09:11:33 zecke, hey, btw did you reply ? I moved email servers over the weekend and lost 1 days mail Aug 12 09:12:21 lrg: I did Aug 12 09:13:18 zecke: would you be able to forward your reply to me. thanks. my fault for screwing up mail transfer. Aug 12 09:14:11 done Aug 12 09:14:19 zecke: thanks :) Aug 12 09:16:24 03John Willis 07org.openembedded.dev * rc6dd2a039e 10openembedded.git/recipes/rdesktop/rdesktop_1.6.0.bb: Aug 12 09:16:24 rdesktop-1.6.0: Fix build issue Aug 12 09:16:24 Remove duplicate audio-2008.patch patch line from recipe (fixes build issue Aug 12 09:16:24 introduced with 013c006d11a59269b9d3adc6cbb6fb8091144965). Aug 12 09:16:24 Signed-off-by: David-John Willis Aug 12 09:16:27 Signed-off-by: Holger Freyther Aug 12 09:24:15 03Michael 'Mickey' Lauer 07org.openembedded.dev * ree14422b23 10openembedded.git/ (3 files in 2 dirs): python: merge some improvements done in various oe.dev-boycott branches back into oe.dev Aug 12 09:25:57 mickeyl: ping Aug 12 09:26:14 hi zecke Aug 12 09:26:18 zecke: pong Aug 12 09:26:22 morning pb__ Aug 12 09:26:26 yo mickeyl Aug 12 09:26:42 mickeyl: have you seen the python patch earlier today? there was a similiar patch in the dreambox branch.. do you prefer any of the two Aug 12 09:27:18 no, didn't spot it Aug 12 09:27:18 pb__: hey! http://patchwork.openembedded.org/patch/651/ { Aug 12 09:27:18 .refuse = 0, Aug 12 09:27:18 .close_side = 1, Aug 12 09:27:18 .close_cause = 5, Aug 12 09:27:18 }, Aug 12 09:27:24 *sigh* Aug 12 09:27:35 have the commit ref offhand? Aug 12 09:27:55 pb__: http://patchwork.openembedded.org/patch/691/ and http://patchwork.openembedded.org/patch/651/ list your name, do you know the status of these patches? Aug 12 09:28:47 mickeyl: c80831a276 in .dev and 177228f80c113c15d3cf437264bc56f712d9e92c in dreambox Aug 12 09:30:41 zecke: 691 is not applied, I think it was waiting for a core h4x0r to approve it. Aug 12 09:31:06 pb__: do you want to land it? Aug 12 09:31:13 651 is still slightly incomplete. I should get back to that later. Aug 12 09:31:38 zecke: I think it (691) would be a nice thing to have but I don't feel all that strongly about it. Aug 12 09:35:46 good morning everyone Aug 12 09:47:06 03Michael 'Mickey' Lauer 07org.openembedded.dev * rc9bff79ca0 10openembedded.git/recipes/python/python-phoneutils_git.bb: python-phoneutils: new recipe; python bindings for libphoneutils Aug 12 10:01:16 hi Aug 12 10:01:20 mmm Aug 12 10:01:35 when will shr-import be merged in org.openembedded.dev? Aug 12 10:02:15 work on that is underway Aug 12 10:02:22 ok thanks a lot Aug 12 10:02:23 (see progress mails) Aug 12 10:02:27 ok Aug 12 10:02:49 dbus/*, python/*, freesmartphone/*, vala/* are done Aug 12 10:02:57 I've a missing dep on intltool-0.40.3-r2 Aug 12 10:03:00 missing are the SHR apps and configs Aug 12 10:03:06 I had SHR branch.... Aug 12 10:03:12 s/had/was Aug 12 10:03:19 where should I push the change? Aug 12 10:03:51 btw for htc-dream support I'm waiting for #htc-linux guys to push openmoko-msm-2.6.29 in linuxtogo git Aug 12 10:03:57 s/push/import Aug 12 10:08:34 mickeyl, ^^^ Aug 12 10:19:27 good morning Aug 12 10:22:45 I bet I have to push in org.openembedded.dev but try it before Aug 12 10:23:00 I will have to rebuild everything tough Aug 12 10:43:25 what could be causing "-gnueabi" being added to our pstage packages? Aug 12 10:48:26 Stecchino: it is not the only thing added right? Aug 12 10:49:25 zecke: example: staging-linux-bc750-bc750-angstrom-linux-gnueabi_2.6.28-r2_i686-linux.ipk Aug 12 10:50:05 i realize know it's the elf class (or something similar). But for other targets we use OE for it's not being added Aug 12 10:51:06 Stecchino: guessing here, the layout of the ipk/ directory is per architecture Aug 12 10:51:39 Stecchino: the pstage directory is quite flat. to support building multiple machines in one tree we put the target triplet in there Aug 12 10:52:05 Stecchino: e.g. my header files for a arm build might be different for mips Aug 12 10:53:11 zecke: build/tmp/deploy has armv5te, x86_64, all but also bc750 (a target) Aug 12 10:54:53 Stecchino: okay let me try again Aug 12 10:54:56 seems like TARGET_OS is changed (by way of patching angstrom-glibc.inc) to not have "-gnueabi" Aug 12 10:55:05 so that is the cause Aug 12 10:55:09 don't ask me why though Aug 12 10:56:29 zecke: Since i'm fairly unfamiliar with OE you might want to be a little bit more pedantic with me, sorry :) Aug 12 10:57:12 i also know of only one ipk directory it's in build/tmp/deploy ;) Aug 12 10:57:37 correction build/tmp/deploy/glibc Aug 12 10:58:13 Stecchino: what is your question? Why do we add angstrom-linux-gnueabi to the pstage package name? What gnueabi is supposed to mean? Aug 12 10:59:11 the question is already solved for me kind of. I just asked here to get some insight of where to look. Aug 12 10:59:32 oka Aug 12 10:59:33 y Aug 12 10:59:35 Let me guess at answers to your questions: gnueabi is an executable format Aug 12 10:59:59 and adding it to the packagename is needed because it's incompatible with certain kernels Aug 12 11:00:28 hmmm Aug 12 11:00:39 I remember having to configure th kernel differently recently becaus init was being killed. Finally know why :) Aug 12 11:01:11 zecke: feel free to laugh at me if I'm wrong, just let me know the correct answers when you do :) Aug 12 11:01:28 * Stecchino could google but is lazy :) Aug 12 11:01:30 Stecchino: the debian wiki has a good introduction to EABI Aug 12 11:02:06 Stecchino: ABI defines struct sizes, padding (to a degree) but also calling convetion (including syscalls) Aug 12 11:02:36 time to search food Aug 12 11:02:39 ah, kernel syscall ABI. I see. Aug 12 11:02:43 zecke: bon appetit Aug 12 11:18:33 03Marcin Juszkiewicz 07org.openembedded.dev * r0fae80c38c 10openembedded.git/recipes/tasks/task-self-hosting.bb: task-self-hosting: do not use Monotone, OE switched to GIT Aug 12 11:18:35 03Marcin Juszkiewicz 07org.openembedded.dev * rb5f769c7e0 10openembedded.git/recipes/tasks/task-self-hosting.bb: task-self-hosting: use task-native-sdk, req python Aug 12 11:20:46 zecke: yes? Aug 12 11:35:25 morning Aug 12 11:35:28 mickey|lunch: ping Aug 12 12:04:01 zecke: I dropped most of my stable/2009 patches from patchwork Aug 12 12:06:39 hi Aug 12 12:07:10 is there way to have 2 kernel images built for a machine? My situation is, I want to build a kernel for target and also an extra kernel for qemu so I can test the image. Aug 12 12:07:52 I know I can build a separate kernel with bitbake but I'd like it to be part of the image so modules for both get installed in the rootfs etc. Aug 12 12:27:25 mickey|lunch: check http://diff.pastebin.ca/1526420 please. It fixes some bugs in Python packaging. Aug 12 12:28:05 mickey|lunch: 1. python-devel is named python-dev as it should be. As bonus it gets dependencies on all *-dev pacakges required Aug 12 12:28:17 mickey|lunch: 2. python-modules package do not depend on python-dev anymore Aug 12 12:28:46 mickey|lunch: 3. python-man package was renamed to python-doc as ${PN}-doc is what OE uses for manpages Aug 12 12:28:57 mickey|lunch: look, test, ack, push Aug 12 12:59:02 File "/usr/lib/python2.6/commands.py", line 55, in getstatusoutput Aug 12 12:59:02 pipe = os.popen('{ ' + cmd + '; } 2>&1', 'r') Aug 12 12:59:02 OSError: [Errno 38] Function not implemented Aug 12 12:59:14 thats bitbake trunk on sheevaplug/angstrom Aug 12 13:11:01 03Klaus Kurzmann 07shr/import * r44e4f4387a 10openembedded.git/recipes/freesmartphone/fso-abyss_git.bb: Aug 12 13:11:01 fso-abyss: switch to SRCPV and adjust PV Aug 12 13:11:01 Signed-off-by: Klaus Kurzmann Aug 12 13:36:51 strange it is Aug 12 13:45:25 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r87215cba66 10openembedded.git/conf/machine/at91sam9260df.conf: Add support for at91sam9260ek booting from internal dataflash Aug 12 13:45:25 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r5579ce1b54 10openembedded.git/conf/machine/at91sam9g20nf.conf: Add support for at91sam9g20ek booting from NAND flash Aug 12 13:45:26 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r9ba96853e2 10openembedded.git/conf/machine/ (6 files): Change erronous use of LINUX26_CONFIG to LINUX26_DEFCONFIG Aug 12 13:45:26 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r008eaaef25 10openembedded.git/conf/machine/at91sam9260dfc.conf: Add support for at91sam9260ek booting from dataflashcard Aug 12 13:45:29 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r2c542e776f 10openembedded.git/conf/machine/at91sam9g20dfc.conf: Add support for at91sam9g20ek booting from dataflashcard Aug 12 13:45:32 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r735c680a16 10openembedded.git/conf/machine/at91sam9260nf.conf: Add support for at91sam9260ek booting from NAND flash Aug 12 13:45:37 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * rc593b76475 10openembedded.git/conf/machine/include/AT91.inc: Remove u-boot-utils from root fs Aug 12 13:45:40 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r0a0e90dfb3 10openembedded.git/recipes/linux/linux.inc: Fix syntax errors when doing make _defconfig in kernel build Aug 12 13:45:43 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r30a1560d4b 10openembedded.git/recipes/u-boot/u-boot_2009.08-rc2.bb: Desperate attempts to make u-boot populate the deploy/../images directory Aug 12 13:45:46 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r072cdc2d20 10openembedded.git/conf/machine/at91sam9g20df.conf: Add support for at91sam9g20ek booting from internal dataflash Aug 12 13:56:17 Hi! Aug 12 13:57:46 I have a problem with pulseaudio: the do_install fails with a libtool error: 'error: cannot install `libcli.la' to a directory not ending in /usr/lib/pulse-0.9.15/modules/' Aug 12 13:57:56 can someone help? Aug 12 14:24:58 fix makefile to honor DESTDIR Aug 12 14:28:48 hrw: I'll try Aug 12 14:28:49 :-) Aug 12 14:57:15 Will bitbake fall back to downloading a tarball from a mirror if an svn server for a bb file were to be temporarily or permanently unavailable? Aug 12 14:57:37 Laibsch: no Aug 12 14:57:54 OK Aug 12 14:58:32 I'm in the process of setting up a complete source mirror for OE (I got really fed up with those 404s all the time) Aug 12 14:59:00 Laibsch: where? Aug 12 14:59:09 one will be on melo Aug 12 14:59:23 I'm thinking of round-robin for community mirrors Aug 12 14:59:24 with speed limit set? Aug 12 14:59:34 Let's see if it becomes a problem first Aug 12 14:59:54 I'm currently in the process of collecting the sources Aug 12 15:00:10 They aren't being served, yet. Aug 12 15:00:20 hrw: you expect bandwidth problems? Aug 12 15:00:20 the new LTG machine might be able to handle something like this too. Aug 12 15:00:25 cool Aug 12 15:00:38 I was hoping for some redundancy Aug 12 15:00:51 yep, that would be good to have... Aug 12 15:01:04 Laibsch: I can upload some sources Aug 12 15:01:09 cool Aug 12 15:01:13 my DL_DIR has 29GB of data Aug 12 15:01:17 nice Aug 12 15:01:21 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * re68d0a4e53 10openembedded.git/recipes/linux/linux_2.6.30.2.bb: Split download of linux-2.6.30.2 into linux-2.6.30.tar.bz2 and patch-2.6.30.2.bz2 Aug 12 15:01:21 03Ulf Samuelsson 07ulf/linux-2.6.30.2 * r187fbec73b 10openembedded.git/conf/checksums.ini: Add checksum for path-2.6.30.2.bz2 Aug 12 15:01:26 we could use the statistics of these mirrors to identify problems :) Aug 12 15:01:27 would have to write python script for it first Aug 12 15:01:29 hrw: would you be able to act as a mirror? Aug 12 15:01:34 Laibsch: no Aug 12 15:01:37 ok Aug 12 15:01:48 Laibsch: moving IP and just 1Mbps upload Aug 12 15:01:52 florian: yes, I was thinking about that, too Aug 12 15:02:08 too many hits on the mirror -> look at bb file update Aug 12 15:02:15 But in the end, it wouldn't be necessary anymore Aug 12 15:02:27 In the past I used to update the SRC_URI Aug 12 15:02:35 But that can quickly become a full-time jobs Aug 12 15:02:37 jobs Aug 12 15:02:38 job Aug 12 15:03:21 hrw: I'm thinking about partial mirrors. I have less upstream, yet, I hope I can help mirroring. Aug 12 15:03:49 My current idea is to have a) complete mirror(s) -> melo and possibly others and b) partial mirrors Aug 12 15:04:12 Laibsch: doesn't partial mirroring defeat the purpose of eliminating 404s? Aug 12 15:04:37 I want to change base.bbclass so that at the end b is hit first (round-robin, community-supplied mirrors) and if the source isn't there, go to a) Aug 12 15:04:43 kg4ysn: no Aug 12 15:04:43 bb in some Aug 12 15:04:51 * hrw -> food Aug 12 15:05:09 kg4ysn: if a b type mirror gives you a 404, bitbake will continue to go to a Aug 12 15:05:14 the full mirror Aug 12 15:05:50 Laibsch: that makes sense, but you're still incurring a 404. is your desire to just eliminate the possibility of a source going away entirely? Aug 12 15:06:01 only when all options give 404 will bitbake error out Aug 12 15:06:15 yes, bitbake doesn't care about an intermittent 404 Aug 12 15:06:28 as long as it can get the source from somewhere. Aug 12 15:06:38 the user won't even really notice most of the time. Aug 12 15:06:47 Laibsch: how much storage space / traffic per month are you anticipating? Aug 12 15:07:01 I thought that 10G may be enough for current .dev Aug 12 15:07:09 I'm already at 18G Aug 12 15:07:15 and the fetching hasn't ended Aug 12 15:07:30 But that includes stuff I will eliminate from mirroring Aug 12 15:07:47 like some avi files, or svn/git/cvs Aug 12 15:08:07 Laibsch: any idea on traffic load? Aug 12 15:08:09 I currently don't really know how much HD it will consume Aug 12 15:08:17 no idea on traffic, but Aug 12 15:08:32 first of all, if it becomes too much, we can mirror and spread the load Aug 12 15:08:48 second, the source mirror should only be hit when upstream source has vanished Aug 12 15:09:09 IOW, if bandwidth becomes a problem, we can always update the SRC_URI and lessen the load Aug 12 15:09:24 I think in the end, it should be manageable Aug 12 15:09:47 $ du -sh /home/rolf/oe/sources/svn/ Aug 12 15:09:47 6.9G /home/rolf/oe/sources/svn/ Aug 12 15:09:57 That's already 7G I won't be mirroring Aug 12 15:10:00