| 00:00.58 | *** join/#oe Crofton_|work (n=balister@pool-71-171-22-222.nwrknj.east.verizon.net) |
| 00:04.15 | *** join/#oe BenLauDC (n=benlau@221.125.8.105) |
| 00:11.58 | *** join/#oe rsalveti__ (n=rsalveti@200.184.118.130) |
| 00:32.55 | *** join/#oe Crofton (n=balister@pool-71-171-22-222.nwrknj.east.verizon.net) |
| 00:37.37 | *** join/#oe waite (n=bwaite@c-24-91-81-44.hsd1.ma.comcast.net) |
| 00:48.57 | *** join/#oe mithro (n=tim@unaffiliated/mithro) |
| 01:00.24 | *** join/#oe waite (n=bwaite@c-24-91-81-44.hsd1.ma.comcast.net) |
| 01:03.45 | *** join/#oe hufnus_cicq (n=hufnus_c@69-12-177-67.dsl.static.sonic.net) |
| 01:16.30 | *** join/#oe archae0pteryx1 (n=snewman@207.47.42.130.static.nextweb.net) |
| 01:23.40 | *** join/#oe raster (n=raster@enlightenment/developer/raster) |
| 01:25.18 | *** join/#oe kergoth (n=kergoth@ip98-167-215-86.ph.ph.cox.net) |
| 01:26.35 | *** join/#oe mekius (n=mekius@enlightenment/developer/mekius) |
| 01:33.01 | kergoth | anyone want to poke at this bitbake git repo? I'd like to confirm it seems sane |
| 01:38.27 | *** join/#oe tmartins (n=zero@187.37.57.221) |
| 01:47.20 | *** join/#oe zecke (n=ich@123-192-186-20.dynamic.kbronet.com.tw) |
| 01:53.49 | *** join/#oe aloisiojr (n=aloisio@189.81.35.54) |
| 02:02.57 | *** join/#oe fraxinath (n=quassel@p54AA5A5B.dip.t-dialin.net) |
| 02:08.15 | *** join/#oe brolin (n=brolin@186.81.88.253) |
| 02:20.40 | *** join/#oe gregoiregentil1 (n=gregoire@adsl-71-135-115-239.dsl.pltn13.pacbell.net) |
| 02:33.11 | *** join/#oe waite (n=bwaite@c-24-91-81-44.hsd1.ma.comcast.net) |
| 02:48.16 | gregoiregentil1 | Ping sakoman |
| 02:55.17 | sakoman | gregoiregentil1: pong |
| 02:55.48 | gregoiregentil1 | I'm trying to create a recipe for system-tools-backends |
| 02:55.55 | gregoiregentil1 | but I get a strange po/Makefile error |
| 02:56.01 | gregoiregentil1 | I saw on Google that you hit the same problem |
| 02:56.05 | gregoiregentil1 | did you manage to fix it? |
| 02:56.27 | gregoiregentil1 | My intention is to build gnome-system-tools through liboobs and system-tools-backends |
| 02:56.48 | *** join/#oe raster (n=raster@enlightenment/developer/raster) |
| 02:57.35 | sakoman | gregoiregentil1: yeah, I solved it |
| 02:58.01 | gregoiregentil1 | you are teasing me ;-) |
| 02:58.02 | sakoman | I have about 100 patches to clean up & submit |
| 02:58.17 | sakoman | I've been working on a gnome image :-) |
| 02:58.21 | sakoman | hang on |
| 02:58.52 | gregoiregentil1 | OK. I'm working more on an xfce image and there are some gnome dependencies |
| 02:59.46 | sakoman | http://pastebin.com/m1abd15b1 |
| 02:59.59 | sakoman | Not pretty but it works |
| 03:00.19 | sakoman | what machine are you building for? |
| 03:00.47 | gregoiregentil1 | strange, I removed po from the Makefile.am and it didn't work |
| 03:00.55 | gregoiregentil1 | I'm the guy behind the touchbook... |
| 03:01.03 | gregoiregentil1 | I'm trying to finalize my image |
| 03:01.09 | sakoman | Ah, I want one :-) |
| 03:01.17 | gregoiregentil1 | it would be nice to merge my xfce/gnome efforts with yours |
| 03:01.22 | sakoman | You should try my gnome image on a touchbook |
| 03:01.35 | sakoman | www.sakoman.com |
| 03:01.41 | gregoiregentil1 | I would love to... |
| 03:01.46 | sakoman | in the download section |
| 03:02.15 | sakoman | Still needs a lot of work, but it is getting there |
| 03:02.31 | sakoman | performs better than I expected it to |
| 03:03.10 | gregoiregentil1 | very cool |
| 03:03.43 | gregoiregentil1 | BTW, the download link doesn't work... |
| 03:04.17 | gregoiregentil1 | how could we merge our efforts? |
| 03:04.25 | gregoiregentil1 | I have a bunch of xfce/gnome recipes |
| 03:04.58 | *** join/#oe kergoth (n=kergoth@ip98-167-215-86.ph.ph.cox.net) |
| 03:05.24 | sakoman | you are right - let me give you the direct link, must have cut & pasted wrong |
| 03:06.14 | sakoman | http://www.sakoman.com/feeds/omap3/glibc/images/experimental/ |
| 03:07.06 | sakoman | gregoiregentil1: that's a good question - as I mentioned I have a *huge* number of patches |
| 03:07.18 | gregoiregentil1 | do you have your tree somewhere? |
| 03:07.27 | gregoiregentil1 | BTW, I'm in Mountain View, it would be great to meet one day |
| 03:07.35 | gregoiregentil1 | I guess that you are on the other side of the bay? |
| 03:07.52 | sakoman | no public tree at the moment |
| 03:08.34 | sakoman | but I'll fix that over the next week or so |
| 03:08.38 | *** join/#oe mrmoku|away (n=mrmoku@ppp-93-104-100-166.dynamic.mnet-online.de) |
| 03:08.50 | sakoman | I'm about 4 hours north or you :-) |
| 03:08.54 | sakoman | of |
| 03:09.04 | sakoman | in Shasta county |
| 03:55.12 | *** join/#oe aloisiojr1 (n=aloisio@189.81.59.121) |
| 04:00.20 | *** join/#oe rsalveti (n=rsalveti@189.70.35.140) |
| 04:03.58 | *** join/#oe sabrod (n=sab@mic92-8-82-234-143-184.fbx.proxad.net) |
| 04:08.00 | *** join/#oe aloisiojr2 (n=aloisio@189.81.122.209) |
| 04:17.04 | *** join/#oe kergoth (n=kergoth@ip98-167-215-86.ph.ph.cox.net) |
| 04:24.44 | *** join/#oe aloisiojr (n=aloisio@189.81.91.24) |
| 04:24.52 | *** join/#oe TSCHAKeee (n=TSCHAKee@c-24-128-190-106.hsd1.ma.comcast.net) |
| 04:25.38 | TSCHAKeee | hello, I am having difficulty building xorg-image because madwifi-ng is puking something fierce. I do not need it. Is there a way to prevent madwifi support from being built in for the image that I choose? |
| 04:30.51 | *** join/#oe aloisiojr1 (n=aloisio@189.81.97.147) |
| 05:05.52 | *** join/#oe alecrim__ (n=alecrim@201008202242.user.veloxzone.com.br) |
| 05:19.50 | *** join/#oe Sup3rkiddo (n=sudharsh@unaffiliated/sudharsh) |
| 05:28.23 | *** join/#oe valhalla (n=valhalla@81-174-23-163.dynamic.ngi.it) |
| 05:36.23 | CIA-1 | 03Klaus Kurzmann <mok@fluxnetz.de> 07shr/import * rc46bca4fc1 10openembedded.git/recipes/libgee/libgee_0.1.6.bb: |
| 05:36.23 | CIA-1 | libgee_0.1.6.bb: add missing inherits for vala and autotools_stage |
| 05:36.23 | CIA-1 | Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de> |
| 05:37.07 | *** part/#oe gregoiregentil1 (n=gregoire@adsl-71-135-115-239.dsl.pltn13.pacbell.net) |
| 05:48.57 | CIA-1 | 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * r4007d4d622 10openembedded.git/classes/oestats-client.bbclass: oestats-client.bbclass: support arbitrary URLs (Closes: #5283) |
| 06:04.26 | *** join/#oe pleemans (n=toi@d51A49C45.access.telenet.be) |
| 06:25.04 | *** join/#oe pocek_ (i=pock@87.205.75.129) |
| 06:30.00 | *** join/#oe dth (n=Dieter@p4FDED0F2.dip.t-dialin.net) |
| 06:31.30 | *** join/#oe rkirti (n=oespirit@117.254.30.80) |
| 06:36.20 | *** join/#oe jpeeters (n=jpeeters@193.5-14-84.ripe.coltfrance.com) |
| 06:48.35 | *** join/#oe greentux (n=lemke@BAA0edc.baa.pppool.de) |
| 06:51.04 | *** join/#oe Sup3rkiddo (n=sudharsh@unaffiliated/sudharsh) |
| 07:04.57 | *** join/#oe darion76 (n=darion76@94-248-37-66.dynamic.peoplenet.ua) |
| 07:18.56 | methril|work | good morning |
| 07:19.23 | CIA-1 | 03Stefan Schmidt <stefan@datenfreihafen.org> 07org.openembedded.dev * r4223c584f1 10openembedded.git/conf/checksums.ini: checksums.ini: Change checksums for devmem2.c after upstream changes and review. |
| 07:19.26 | *** join/#oe woglinde (n=henning@p5DDC6F89.dip.t-dialin.net) |
| 07:20.28 | woglinde | hi |
| 07:21.31 | *** join/#oe stefan_schmidt (n=stefan@w1722.wlan.rz.tu-bs.de) |
| 07:23.26 | *** join/#oe Pr0t0N (n=lcintrat@wyplay-fr-4-82.cnt.nerim.net) |
| 07:24.24 | *** join/#oe MWelchUK (n=welchma@65.91.2.71) |
| 07:24.50 | *** join/#oe cyberdeck (n=cyberdec@iss66.vlsi.informatik.tu-darmstadt.de) |
| 07:28.51 | b0fh | morning |
| 07:30.28 | woglinde | hi b0fh |
| 07:31.27 | woglinde | ~ping rp |
| 07:31.28 | ibot | pong rp |
| 07:34.07 | b0fh | hi woglinde |
| 07:56.16 | *** join/#oe rob_w (n=bill@217.237.177.190) |
| 07:57.04 | rob_w | did anybody ever implement the icc from intel as toolchain into oe yet ? |
| 07:57.33 | woglinde | not that I know |
| 07:57.39 | woglinde | did you really need it? |
| 07:57.41 | zecke | rob_w: not heard about it. why icc and not the arm toolchain? |
| 07:57.57 | woglinde | icc for arm? |
| 07:58.02 | woglinde | not that I would know |
| 07:58.07 | woglinde | icc is pure x86 |
| 07:58.24 | zecke | icc for arm, I assumed it exsists due the xscale :} |
| 07:58.27 | zecke | *nevermindà |
| 07:58.45 | woglinde | the performance gain isnt that much |
| 07:59.34 | rob_w | yeah not arm .. i am talking for the atom devices |
| 07:59.41 | rob_w | they say the performance is noticable |
| 07:59.48 | woglinde | hm okay |
| 07:59.58 | woglinde | that might be |
| 08:00.18 | rob_w | but as the icc doenst know many switch of gcc, one will need a wrapper around those switches .. |
| 08:00.33 | *** join/#oe otavio (n=otavio@debian/developer/otavio) |
| 08:00.35 | woglinde | hm I think you could try with external toolchain |
| 08:00.36 | rob_w | intel talks about it in their pdfs but never show such a wrapper |
| 08:01.49 | rob_w | ok so if anyone ever has tested that before .. feel free to point them to me .. i will go ahead and try to use icc nativly now .. when i mangage this i will look into integrating it into oe |
| 08:02.00 | sx | morning |
| 08:02.16 | *** join/#oe pvanhoof (n=pvanhoof@d54C0C0BA.access.telenet.be) |
| 08:03.05 | woglinde | http://macles.blogspot.com/2008/09/intel-cc-compiler-gcc-and-intel-atom.html |
| 08:07.09 | *** join/#oe daq4th (n=darkness@v062.natpool.sourcentral.org) |
| 08:07.25 | *** join/#oe boris_OmegA (n=boris@194.51.141.172) |
| 08:14.43 | sx | is there anyone who can play video on atmel arm system using gstreamer and gtk+ ? |
| 08:16.12 | rob_w | woglinde, nice one |
| 08:17.50 | woglinde | I have no atmel |
| 08:18.02 | woglinde | sx maybee you have to wait for hrw |
| 08:19.52 | *** join/#oe florian_kc (n=fuchs@port-217-146-132-69.static.qsc.de) |
| 08:20.17 | woglinde | jo florian |
| 08:21.03 | sx | hrw use some atmel stuff, right? |
| 08:21.04 | *** join/#oe stefan_schmidt (n=stefan@w1713.wlan.rz.tu-bs.de) |
| 08:25.00 | *** join/#oe pleemans (n=toi@81.164.156.69) |
| 08:30.08 | *** join/#oe woglinde (n=henning@93.220.111.137) |
| 08:33.34 | b0fh | sx: yes, hrw is using an at91sam9263ek. (As do I) |
| 08:34.02 | sx | b0fh, as you do? |
| 08:34.32 | sx | b0fh, have you tried to play video on gtk window using gstreamer's ximagesink? |
| 08:35.14 | b0fh | sx: no, haven't. I'm able to do video using mplayer and fbdev, never tried gstreamer |
| 08:37.15 | sx | b0fh, hmm, I'm writing my own gtk application, the app sometimes must play video on it and sometimes the video stops and comes gtk stuff (buttons, inputbox) |
| 08:37.20 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r11f327a470 10openembedded.git/classes/base.bbclass: |
| 08:37.20 | CIA-1 | base bbclass: try to find the git binaries in PATH instead of BBPATH |
| 08:37.20 | CIA-1 | * Also, how do I get the build banner back? This bug went unnoticed due to it being gone |
| 08:37.37 | *** join/#oe pleemans (n=toi@d51A49C45.access.telenet.be) |
| 08:37.48 | sx | b0fh, but I have problems with gstreamer |
| 08:39.03 | *** join/#oe ZaPPaS (n=moritz@buero2.pi5.physik.uni-stuttgart.de) |
| 08:39.06 | b0fh | sx: sorry, no experience with gstreamer.. |
| 08:40.00 | sx | I'm searching for anyone who success to play video on gtk window using gstreamer on atmel stuff |
| 08:45.20 | CIA-1 | 03Stefan Schmidt <stefan@datenfreihafen.org> 07org.openembedded.dev * r99f9432c08 10openembedded.git/recipes/devmem2/devmem2.bb: devmem2: Bump forgotten PR, thanks zecke. |
| 08:45.27 | b0fh | I'm searching for someone who knows where PAGE_SIZE is declared :-) |
| 08:46.35 | *** join/#oe stefan_schmidt (n=stefan@w1386.wlan.rz.tu-bs.de) |
| 08:48.03 | *** join/#oe hrw (n=hrw@chello089078170228.chello.pl) |
| 08:49.08 | b0fh | hi hrw. |
| 08:49.11 | sx | b0fh, in asm/page.h ? |
| 08:49.40 | b0fh | sx: I know.. and am a bit clueless, when I include it, it still is undefined. |
| 08:50.48 | sx | b0fh, mybe other code undeclares it before your code... |
| 08:51.14 | b0fh | sx: then it must be one of the gcc internals, #include <asm/page.h> is the first line of code..... :-/ |
| 08:51.18 | sx | hi hrw|gone, I have some questions to you |
| 08:52.18 | BET_frogger | b0fh: for userspace code, man sysconf |
| 08:52.37 | b0fh | BET_frogger: thanks. |
| 08:52.46 | sx | b0fh, do you have getpagesize() function? |
| 08:54.01 | b0fh | sx: yes, getpagesize() does work... |
| 08:54.23 | woglinde | PAGE_SIZE one is a common pitfall |
| 08:54.24 | sx | b0fh, it is the same thing like PAGE_SIZE |
| 08:54.39 | woglinde | jepp please use the function |
| 08:54.53 | *** part/#oe b0fh (n=BOFH@ntp.hscg.net) |
| 08:54.58 | *** join/#oe b0fh (n=BOFH@ntp.hscg.net) |
| 08:55.14 | b0fh | woglinde, BET_frogger, sx: thanks, getpagesize() satifies my need. |
| 08:55.54 | sx | where are you hrw ;( |
| 08:56.38 | RP | morning all |
| 08:56.45 | RP | woglinde: pong |
| 09:00.15 | woglinde | jo rp |
| 09:00.47 | RP | woglinde: Whats up? |
| 09:00.56 | woglinde | rp I have a question about RPROVIDES |
| 09:01.03 | woglinde | for all javavm we have |
| 09:01.19 | woglinde | RPROVIDES = "java2-runtime" |
| 09:01.34 | woglinde | unfornatly we can use PREFERRED_PROVIDER on it |
| 09:01.57 | RP | No, PREFERRED_PROVIDER only works for build dependencies |
| 09:02.02 | RP | not runtime ones |
| 09:02.10 | woglinde | and I dont understand really whats going on in lib/bb/provider.py |
| 09:02.34 | RP | R* variables are in the domain of the package manager, not bitbake |
| 09:03.04 | woglinde | so how can we then choose the java-vm which should be used |
| 09:03.52 | woglinde | now we have the problem that only the first one is using |
| 09:04.07 | woglinde | in list of provides |
| 09:04.15 | RP | At what point are you after to choose it? |
| 09:05.13 | RP | Are you after to be able to set a PREFERRED_PROVIDER to select it? |
| 09:05.26 | woglinde | yes |
| 09:06.22 | woglinde | problem is some packages have rdepends on java2-runtime |
| 09:06.32 | woglinde | and so the firs of the list is selected |
| 09:06.33 | *** join/#oe rob_w (n=bill@217.237.177.190) |
| 09:06.50 | woglinde | then bitbake spoils an debug using PREFERRED_PROVIDER |
| 09:07.27 | woglinde | which not works and yesterday I understand wyh examing filterProvidersRunTime |
| 09:07.32 | woglinde | for 2 hours |
| 09:08.38 | RP | woglinde: I'm thinking |
| 09:10.22 | RP | woglinde: This isn't a simple problem |
| 09:10.37 | RP | selecting providers in the runtime space is a painful problem :( |
| 09:11.02 | woglinde | okay |
| 09:11.17 | woglinde | I believe after seeing the filter function |
| 09:11.27 | RP | woglinde: The nearest example I can think of we have in the code is libx11 |
| 09:11.35 | RP | or maybe the xserver selection |
| 09:11.39 | woglinde | hm |
| 09:12.49 | RP | The equivalent of the XSERVER for java would be an RDEPENDS_${PN} =+ "${JAVARUNTIME}" |
| 09:13.18 | RP | but it means you only get one java runtime and rebuilding to change it |
| 09:13.33 | woglinde | rp hm istn it the easies thig to put var = "PREFERRED_PROVIDER_%s = %s" % (item, prefervar) into the filter function? |
| 09:13.56 | RP | woglinde: Absolutely not |
| 09:14.07 | RP | woglinde: This is not a bitbake problem |
| 09:14.14 | woglinde | okay |
| 09:14.15 | woglinde | sorry |
| 09:14.39 | *** join/#oe stefan_schmidt (n=stefan@w1567.wlan.rz.tu-bs.de) |
| 09:15.21 | woglinde | rp in which file exactly is the xserver choosing? |
| 09:15.37 | RP | woglinde: grep for XSERVER |
| 09:16.28 | RP | woglinde: Thinking about this, the RDEPENDS "java-runtime" isn't a problem so much, its just bitbake has real problems trying to work out who provides it :/ |
| 09:16.47 | pb_ | It sounds like it is a bitbake problem to some extent. If bitbake is inferring build dependencies from RDEPENDS and not giving you any way to influence which one it chooses from multiple options, then that seems to be at least a missing feature if not an actual bug. |
| 09:17.17 | RP | pb_: Yes :/ |
| 09:17.20 | pb_ | On the other hand, it sounds like the only real issue is an image construction one and maybe you can just solve that with an explicit (R)DEPENDS in your image file. |
| 09:18.24 | woglinde | hm |
| 09:18.38 | RP | pb_: I think its a bit deeper than that |
| 09:18.41 | woglinde | hm |
| 09:18.50 | pb_ | Obviously that'd mean that typing "bitbake <java-using-package>" on its own in a fresh build directory wouldn't necessarily do what you wanted, but I would be content with "don't do that, then" as a response in that situation. |
| 09:18.53 | woglinde | problem is I want to test the java build only |
| 09:19.00 | woglinde | not to building an image |
| 09:19.05 | woglinde | problem gets deeper |
| 09:19.22 | *** join/#oe marcosmamorim (n=marcos@bd21096b.virtua.com.br) |
| 09:19.23 | RP | woglinde: We need a better solution for you than the XSERVER problem |
| 09:19.29 | woglinde | because the recipes has java-doc package defines which relays on "java2-runtime" |
| 09:20.14 | pb_ | woglinde: in that situation, can you not just build your preferred java2-runtime by hand in advance? |
| 09:20.36 | RP | pb_: Why would building it in advance help? |
| 09:21.01 | woglinde | pb the problem bitbake is choosing for java2-runtime the first providee |
| 09:21.14 | pb_ | RP: does bitbake no longer prefer packages that have already been built when it's trying to select a provider? |
| 09:21.22 | *** join/#oe otavio (n=otavio@debian/developer/otavio) |
| 09:21.28 | woglinde | so when building the vm you want its dragging in the first from the list too |
| 09:21.29 | pb_ | it certainly used to do that. if the code has been deleted then that seems like a bit of a shame. |
| 09:21.30 | RP | pb_: It does not, it was removed as a rather bad idea |
| 09:22.24 | RP | pb_: It meant that you could never tell what "bitbake X" might do without knowning what was in staging |
| 09:22.39 | RP | and that was causing reproducibility issues |
| 09:22.51 | *** join/#oe Spyro (n=ian@benden.mnementh.co.uk) |
| 09:23.15 | pb_ | ah. well, this was exactly the situation that that code was introduced to deal with. it does seem like a bit of a retrograde step to have removed that without providing a replacement. |
| 09:23.48 | RP | pb_: Having to always build A, then B sees like a rather poor way of dealing with this problem |
| 09:24.31 | pb_ | RP: it's a fairly specialised problem, since it only occurs in the case when there is no other way to identify A as the favoured choice. In the "normal" case when you are building a high-level target (i.e. an image) which includes A explicitly, there should be no issue. |
| 09:25.20 | pb_ | RP: conceptually, it is no different to the same issue that you get with regular package management. if I say "apt-get install java-using-package" and I don't already have a jvm installed, apt will pick one for me more-or-less at random. |
| 09:25.29 | RP | pb_: Even if A is included explicitly, I'm not sure that removes the problem? |
| 09:25.32 | pb_ | if I do have a jvm installed already, it will use that one and will not install another random one just because it can. |
| 09:26.08 | pb_ | RP: possibly it doesn't actually fix it at the moment, but I think it would be fairly straightforward to make it so that it did. |
| 09:26.30 | *** join/#oe hemanthariyani (n=a0875145@nat/ti/x-9dc78ea7179c9612) |
| 09:27.38 | pb_ | and I think that would give you the most useful "do what I mean" semantics for builds of top level targets. it certainly doesn't seem desirable to require people to spell out all their preferred providers in configuration, since that would lead to a weird split of the desired packages between two different places. |
| 09:27.41 | *** join/#oe SDuensin (n=Scott@12-226-44-250.client.mchsi.com) |
| 09:28.21 | RP | pb_: I'd suggest its not straightforward. Lets say imageA includes javaA, imageB includes javaB. What should happen if I "bitbake imageA;bitbake imageB"? "bitbake imageA imageB"? |
| 09:28.45 | woglinde | hm |
| 09:28.57 | pb_ | RP: you'd get both being built. I don't think there's any issue there. |
| 09:28.59 | woglinde | bitbake should yell and only build one |
| 09:29.20 | pb_ | there's no harm in building two jvms, and each image would get the one that it wants. |
| 09:29.28 | RP | pb_: The package manager will have two RPROVIDES "java-runtime" |
| 09:29.36 | pb_ | RP: sure, but that isn't a problem |
| 09:29.48 | pb_ | again, this is the same issue that package managers face every day. |
| 09:29.51 | RP | pb_: It is as the moment as the result isn't determinsitic |
| 09:30.15 | woglinde | seems no easy solution for today |
| 09:30.28 | pb_ | if opkg is handling that wrong, that's an opkg bug. I could well believe that opkg is indeed broken but that's a separate issue to what bitbake should do. |
| 09:30.46 | RP | pb_: Which should end up in the image? |
| 09:30.50 | pb_ | as I say, apt gets it right (and ipkg has done at some point in the past, even if opkg doesn't today). it certainly isn't an intractable problem. |
| 09:30.54 | pb_ | RP: the one that the image depends on. |
| 09:30.55 | RP | pb_: How should opkg choose? |
| 09:31.28 | pb_ | RP: imageA has an explicit dependency on jvmA, right? so, when opkg is constructing imageA, it will be given "opkg install ... jvmA ... java-using-package". |
| 09:31.32 | RP | pb_: hmm. I still think we're heading for nasty corner cases |
| 09:31.51 | woglinde | besides for images |
| 09:31.58 | pb_ | even given opkg's limited intelligence, it should be able to figure out that jvmA satisfies java-using-packages's need for java-runtime and hence there is no need to find another. |
| 09:32.10 | pb_ | as I say, this is completely standard stuff for virtual packages. |
| 09:32.13 | woglinde | I only want for now bitbake java-vm1 without dragging the second in |
| 09:32.21 | RP | pb_: Having to guarantee there is always an explict dependency there will cause pain |
| 09:32.34 | *** join/#oe ao2 (n=u@2001:1418:117:0:0:0:0:1) |
| 09:32.48 | florian | good morning |
| 09:32.56 | pb_ | RP: how so? nothing too awful will happen if you don't have an explicit dependency, it's just that you won't be able to select which jvm you get (which is the situation today). |
| 09:33.05 | woglinde | yo florian |
| 09:33.06 | pb_ | adding an explicit dependency seems like a completely natural way of making that selection. |
| 09:33.08 | pb_ | florian: good morning |
| 09:33.14 | RP | pb_: You will get a random jvm in your image |
| 09:33.36 | RP | pb_: I'm afraid I object strongly to random output from OE |
| 09:33.36 | pb_ | RP: right, but plenty of users don't care which jvm they get. the whole point of jvms is that, in theory, they are interchangeable. |
| 09:34.11 | pb_ | if you do care which jvm you get, you add a dependency on that one. |
| 09:34.22 | RP | pb_: Isolating this to jvms only misses the point somewhat |
| 09:34.39 | RP | and the point is your build is no longer deterministic |
| 09:34.52 | pb_ | if you care strongly about having all your dependencies nailed down, you could easily enough add a diagnostic mode to opkg to print an error in that situation rather than picking one at random. |
| 09:35.06 | pb_ | RP: it is deterministic, it is just (perhaps) not predictable. |
| 09:36.19 | pb_ | RP: it doesn't apply only to jvms, it applies to more or less all true-virtual packages. |
| 09:36.21 | RP | pb_: deterministic is the wrong word :/ |
| 09:36.37 | RP | pb_: You can have multiple possible outputs from the same input |
| 09:36.57 | *** join/#oe jpeeters (n=jpeeters@193.5-14-84.ripe.coltfrance.com) |
| 09:37.07 | RP | Anyhow, I suggest we go back to woglinde's problem specifically |
| 09:37.40 | pb_ | I think woglinde's specific problem is exactly that one, unless I have understood it wrong. |
| 09:38.15 | pb_ | it's just a worse, and more pathological case of the same thing |
| 09:39.05 | RP | well ok, lets just and get back to a more practical level of discussion |
| 09:39.05 | pb_ | (in that, in his situation, the very package he has asked bitbake to build is a provider of java-runtime and yet bitbake wants to build another one) |
| 09:39.05 | woglinde | pb again I want build java-vm1 without sneaking the first javavm from the list as providees in to satisfy RPROVIDES java2-runtime |
| 09:39.09 | pb_ | woglinde: right |
| 09:39.40 | RP | There should be some way to fool bitbake into doing that with our existing controls |
| 09:39.41 | woglinde | s/as/of |
| 09:39.51 | *** join/#oe ant_work (n=andrea@host214-85-static.34-85-b.business.telecomitalia.it) |
| 09:39.56 | woglinde | jo ant |
| 09:40.14 | RP | woglinde: What happens if you raise the preference of java-vm1 ? |
| 09:40.15 | pb_ | so, I still think it should be fairly straightforward to fix that in bitbake: it is the same issue as the explicit-depends-in-image thing I mentioned above. |
| 09:40.26 | ant_work | morning |
| 09:40.36 | pb_ | I can't think offhand of any way other way to fix that, if bitbake is not respecting the normal preference controls for runtime-inferred dependencies. |
| 09:40.48 | woglinde | rp hm let me test |
| 09:40.59 | ant_work | woglinde: of course you read http://wiki.openembedded.net/index.php/Java |
| 09:41.12 | woglinde | ant the problem is deeper |
| 09:41.36 | ant_work | yes, I'm following |
| 09:44.15 | RP | pb_: How exactly do you propose to fix that in bitbake? |
| 09:45.45 | woglinde | rp problem remains |
| 09:45.56 | RP | woglinde: :( |
| 09:46.11 | woglinde | here is the pastebin |
| 09:46.13 | woglinde | http://pastebin.ca/1503008 |
| 09:47.19 | pb_ | RP: in taskdata.py, where it's considering which provider to use out of the ones that are available, have it inspect the list of packages that have already been selected for building and, if one of those is an eligible provider, use it. |
| 09:47.54 | RP | pb_: taskdata doesn't know which providers have already been selected |
| 09:47.54 | pb_ | I'm not sure offhand whether you could actually do it in a single pass like that or whether it would require a post-process step, but it doesn't seem conceptually all that complicated. |
| 09:48.10 | pb_ | RP: there's no reason it couldn't find out though, right? |
| 09:48.19 | RP | pb_: taskdata lists all possible options and selections is made in runqueue |
| 09:48.50 | RP | woglinde: and which one do you want to select? |
| 09:48.53 | pb_ | ah right, yes. I guess prepare_runqueue() would be the place then. |
| 09:49.05 | woglinde | rp openjdk-6-jre |
| 09:49.17 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r122df42e88 10openembedded.git/recipes/gnome/gnome-bluetooth_git.bb: gnome-bluetooth git: bump SRCREV |
| 09:49.18 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r57b4d3a2ee 10openembedded.git/ (2 files in 2 dirs): libomxil: add 0.9.1 |
| 09:49.18 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * rcc50d445b1 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev |
| 09:49.45 | RP | pb_: How do we know which providers need this special treatment? |
| 09:49.58 | pb_ | RP: there's nothing special, you would do it for all of them. |
| 09:50.15 | pb_ | any time you are faced with multiple providers for the same virtual. |
| 09:50.22 | CIA-1 | 03Klaus Kurzmann <mok@fluxnetz.de> 07shr/import * r7fe7e96160 10openembedded.git/recipes/intone/intone_svn.bb: |
| 09:50.22 | CIA-1 | intone: package all icons / set correct PV |
| 09:50.22 | CIA-1 | Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de> |
| 09:50.34 | pb_ | obviously, if there is only one eligible provider, there's no point in doing this analysis: it wouldn't hurt, but it would be a waste of time since it can't change the outcome. |
| 09:50.50 | RP | woglinde: Have you tried something like PREFERRED_PROVIDER_openjdk-6-jre = openjdk-6-jre ? |
| 09:51.09 | RP | pb_: "all of them" being all possible runtime providers? |
| 09:51.35 | pb_ | RP: all possible runtime providers for the runtime virtual that you are considering, yes. |
| 09:51.59 | RP | pb_: How do you know its a runtime virtual? |
| 09:52.27 | woglinde | rp??? |
| 09:52.32 | woglinde | this is selected as default |
| 09:52.39 | *** join/#oe ArteK_ (n=Artur@81.15.241.96) |
| 09:52.47 | b0fh | hrw: Hmm.. at91.com phpbb seems fscked..: You have been permanently banned from this board.A ban has been issued on your username. |
| 09:52.58 | *** join/#oe mcgeagh (n=McGeagh@xbmc/staff/mcgeagh) |
| 09:53.10 | woglinde | hm no |
| 09:53.15 | woglinde | let me try |
| 09:53.17 | RP | woglinde: What do you mean? I know that line is stange but have you tried it? |
| 09:53.28 | woglinde | its look strange |
| 09:53.33 | woglinde | but I will test as you said |
| 09:53.36 | pb_ | RP: I'm not sure I understand the question. it's a runtime virtual if it came from RDEPENDS, obviously, but I don't think that can be what you were asking. |
| 09:54.41 | pb_ | in any case, the runtime-ness isn't actually significant, you'd want to do this for both types of dependencies I think. |
| 09:54.41 | RP | pb_: I was trying to work out if you meant to treat some subset of he dependencies differently or we apply this treatment accross the whole set of runtime data |
| 09:54.51 | pb_ | RP: no, across the whole set. |
| 09:54.52 | RP | right |
| 09:55.27 | woglinde | rp I was sure I saw in the -DDD output PREFERRED_PROVIDER_openjdk-6-jre is choosing openjdk-6-jre |
| 09:55.30 | woglinde | but it wasnt |
| 09:55.33 | woglinde | so I test now |
| 09:56.54 | woglinde | hm yes worked |
| 09:57.16 | woglinde | args no |
| 09:57.58 | *** join/#oe pleemans (n=toi@d51A49C45.access.telenet.be) |
| 09:57.59 | RP | pb_: At the moment we have the build dependency taskdata and the runtime dependency taskdata. Both are currently sorted by providers.py which is a good fit for the build data but not the runtime data |
| 09:58.05 | woglinde | hm it worked |
| 09:58.12 | woglinde | bitbake DEBUG confused me |
| 09:58.30 | woglinde | and I understand why it works |
| 09:58.47 | RP | woglinde: it at least gives you something to work with |
| 09:58.51 | woglinde | yes |
| 09:58.54 | *** join/#oe Sleep_Walker (n=Sleep@193.179.96.131) |
| 09:58.58 | *** join/#oe marcosmamorim (n=marcos@bd21096b.virtua.com.br) |
| 09:59.05 | woglinde | didnt came to mind to set it |
| 09:59.10 | RP | I don't think pb_'s suggested changes are going to happen overnight :/ |
| 09:59.16 | woglinde | because I thought its a setting by default |
| 09:59.47 | RP | woglinde: If one exists and the others do not bitbake uses it as a hint to choose that one |
| 09:59.59 | pb_ | yeah, probably not. I will add them to my list of planned bitbake changes but there is other stuff in front of them. :-} |
| 10:00.08 | woglinde | thanks anyway |
| 10:00.19 | woglinde | seems my brain was somehow blocked |
| 10:00.26 | woglinde | to see the solution myself |
| 10:00.38 | RP | woglinde: Setting that isn't obvious... |
| 10:00.52 | RP | Its a lot more subtle than I'd like... |
| 10:02.09 | ant_work | woglinde: that's why I hinted at OE Java docs..see Information on specific libraries |
| 10:02.25 | ant_work | .... So bitbake will not chose the wrong one by accident (which would otherwise pull in all kinds of unwanted dependencies). |
| 10:02.38 | *** join/#oe marex (n=marex@thor.hackndev.com) |
| 10:02.53 | woglinde | ant has nothing really to do with the problem |
| 10:03.13 | woglinde | ant you can test it yourself the problem the following way |
| 10:03.20 | woglinde | make to recpes foo-a.bb |
| 10:03.25 | woglinde | foo-b.bb |
| 10:03.43 | woglinde | put in both RPROVIDES = moo |
| 10:04.04 | woglinde | make a third recpie with RDEPENDS = moo |
| 10:04.08 | woglinde | bitbake foo-b |
| 10:04.21 | woglinde | and you will see that foo-a is dragged in for moo |
| 10:04.34 | woglinde | lunch now |
| 10:05.15 | *** join/#oe vhallac (n=vedat@85.103.144.59) |
| 10:05.18 | ant_work | ok, thats RDEPEND, I know, still, as outsider, I've seen that when / if you have different solutions for a preferred provider, you just don't know which one would be selected (e.g. it was for xerver-kdrive, bluez, virtual/kernel) |
| 10:06.49 | vhallac | Hi. I am trying to get libsmbclient to my target, which works fine. But it also pulls in the full samba package, which I don't want |
| 10:07.26 | vhallac | Is there an easy way to get samba out of libsmbclient RDEPENDS? |
| 10:07.40 | pb_ | mickeyl: good morning! |
| 10:07.47 | mickeyl | morning pb_ |
| 10:08.19 | Viltapi | any suggestions for the following: I'm using Apex bootloader for a board OE doesn't support (LPC3131). If I build Apex manually using a arm-linux- toolchain it comes out clean and boots, but if I build it with OE using arm-angstrom-linux-gnueabi- toolchain, it gets built without any error or warnings but does not boot |
| 10:08.49 | Viltapi | manually built Apex's size is around 65kb, with OE it's 86kb |
| 10:09.32 | Viltapi | same sources, same patches, same config, but maybe I need some specific compiler options? |
| 10:10.00 | *** join/#oe markos_ (n=markos@athedsl-425182.home.otenet.gr) |
| 10:10.26 | pb_ | Viltapi: I guess you need to compare the two binaries to see what's different. |
| 10:11.07 | vhallac | I am thinking of adding a samba-server to the end of PACKAGES in samba recipe, and move all the server related files in that package. Is this how I should approach the issue? |
| 10:11.32 | pb_ | vhallac: it sounds like it's just a bug for libsmbclient to depend on samba. I can't think of any good reason why it should be doing that. |
| 10:12.11 | vhallac | I can't find where that dependence is added. I assumed a sub-package in a recipe depends on the main package |
| 10:13.16 | Viltapi | pb_: any suggestion how/with what tool? Diff only states that they differ |
| 10:13.26 | vhallac | package.bbclass seems to have some code that adds dependencies, but I didn't do any nosing around to see if the assumption is correct. |
| 10:13.29 | pb_ | Viltapi: I think you just have to inspect them by hand. |
| 10:13.41 | pb_ | i.e. disassemble both and eyeballs the resulting output |
| 10:13.52 | vhallac | The piece of code that says "if extra_depends == None:". I am on stable.2009 branch, btw. |
| 10:14.27 | pb_ | vhallac: that's in do_split_packages(), which I wouldn't have expected to be used in this case. |
| 10:15.00 | pb_ | if samba really is using do_split_packages() to make libsmbclient then that'd be a bit weird, but you could fix it by setting extra_depends to be empty. |
| 10:15.23 | vhallac | isn't it what handles the case for declarations like: PACKAGES =+ "libsmbclient libsmbclient-dev cifs cifs-doc swat" |
| 10:15.30 | pb_ | no |
| 10:15.47 | vhallac | Ah, my bad, then. |
| 10:16.23 | *** join/#oe pirho (i=pirho@gateway/gpg-tor/key-0x2CEEC9CB) |
| 10:17.24 | *** join/#oe hemanthariyani (n=a0875145@nat/ti/x-3677f08470131ff8) |
| 10:18.25 | vhallac | I am back to square one. Something pulls in samba, and I don't know what. |
| 10:19.32 | vhallac | Viltapi: Did you try manually compiling with the compiler oe generates? |
| 10:19.42 | vhallac | Just to rule out a cmpiler issue |
| 10:21.07 | *** join/#oe markos_ (n=markos@athedsl-425182.home.otenet.gr) |
| 10:21.09 | *** join/#oe a0875145_ (n=a0875145@nat/ti/x-f9900b030ae3ed50) |
| 10:25.05 | b0fh | ~ping hrw |
| 10:25.06 | ibot | pong hrw |
| 10:28.33 | *** join/#oe rkirti (n=oespirit@117.254.30.241) |
| 10:37.04 | Viltapi | vhallac: no, i'll have to try that next |
| 10:41.29 | hrw | morning |
| 10:41.56 | rkirti | good morning hrw |
| 10:42.21 | hrw | someone knows is there a php 5.3.0 (or newer) in Ubuntu? |
| 10:46.16 | valhalla | hrw: not in 8.04LTS (5.2.4) |
| 10:46.42 | *** join/#oe rkirti (n=oespirit@117.254.30.214) |
| 10:48.12 | *** part/#oe ao2 (n=u@2001:1418:117:0:0:0:0:1) |
| 10:48.29 | *** join/#oe mithro (n=tim@unaffiliated/mithro) |
| 10:49.08 | methril|work | hrw: in 9.04 5.2.6.dfsg.1-3ubuntu4.1 |
| 10:49.25 | hrw | thx |
| 10:49.30 | *** join/#oe waite (n=bwaite@c-24-91-81-44.hsd1.ma.comcast.net) |
| 10:50.24 | *** join/#oe Wiedi (n=wiedi@newton-air.w.fruky.net) |
| 10:50.30 | *** join/#oe mnabil (n=mnabil@41.234.68.8) |
| 10:50.44 | *** join/#oe stefan_schmidt (n=stefan@b142.apm.etc.tu-bs.de) |
| 10:52.52 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r755f936dbb 10openembedded.git/conf/distro/angstrom-2008.1.conf: angstrom: change blacklist messages a bit |
| 10:55.38 | b0fh | hi hrw! Can you login to the forum on at91, or is it just me who's blacklisted? |
| 10:55.57 | hrw | I can |
| 10:56.24 | b0fh | hrw: I noticed a sql server is shutting down message and now I'm banned :-( |
| 11:03.53 | woglinde | re |
| 11:07.47 | *** join/#oe rsalveti_ (n=rsalveti@189.70.66.165) |
| 11:07.53 | *** join/#oe birunko (n=birunko@189.70.87.19) |
| 11:14.49 | hrw | ipaq-sleep must die |
| 11:15.38 | florian | :-) |
| 11:15.40 | woglinde | *g* |
| 11:15.46 | woglinde | isnt ipaq dead anyway? |
| 11:16.22 | hrw | woglinde: we still install that stuff in image |
| 11:16.27 | florian | hrw: I don't really agree - if you don't have sleep support working disable apm emulation. |
| 11:16.55 | hrw | cat: can't open '/proc/apm': No such file or directory |
| 11:17.24 | florian | hrw: apm -s ? |
| 11:17.39 | *** join/#oe mnabil_ (n=mnabil@41.234.71.21) |
| 11:17.59 | hrw | florian: echo 'standby'>/sys/power/state is better |
| 11:18.12 | hrw | root@at91sam9263ek:~# apm -s |
| 11:18.13 | hrw | No APM support in kernel |
| 11:18.18 | florian | good |
| 11:18.53 | *** join/#oe jpeeters (n=jpeeters@193.5-14-84.ripe.coltfrance.com) |
| 11:21.52 | b0fh | hrw: I totally agree.. ipaq-sleep = pure evil! |
| 11:23.13 | hrw | b0fh: I knew few more evils |
| 11:23.27 | hrw | but not quite at91 related |
| 11:23.39 | b0fh | hrw: :-) |
| 11:32.38 | *** join/#oe gremlin[it] (n=gremlin@85.32.103.209) |
| 11:43.48 | *** join/#oe mrc3 (n=ddiaz@61.82.185.3) |
| 11:46.20 | *** join/#oe dos11 (n=dos@afiu120.neoplus.adsl.tpnet.pl) |
| 11:55.58 | *** join/#oe GNUtoo (n=gnutoo@host54-85-dynamic.51-79-r.retail.telecomitalia.it) |
| 11:56.02 | *** join/#oe mickey|cafe (n=M@p54B025BA.dip0.t-ipconnect.de) |
| 12:06.02 | hrw | bbl |
| 12:07.29 | *** join/#oe remur_030 (n=karsten@2001:6f8:1c31:0:221:5dff:feb3:8706) |
| 12:08.30 | mckoan | good afternoon |
| 12:10.59 | remur_030 | good day, my build for ntp fails as md5 sums mismatch, problem is according to ntp download location oe got the md5 right, just the dowload is always different |
| 12:11.12 | remur_030 | erm always the same, but always a wrong md5sum |
| 12:11.13 | *** join/#oe mario-goulart (n=user@201-40-162-47.cable.viacabocom.com.br) |
| 12:12.54 | remur_030 | I could obviously replace the md5 sum in the receipe, but I am unsure about the security implications there =/ |
| 12:13.37 | *** join/#oe emachado (n=edjunior@201.82.64.173) |
| 12:16.51 | remur_030 | http://ftp.riken.jp/pub/net/ntp/ntp4/ntp-4.2/ has the proper version with fitting md5sum |
| 12:17.10 | *** join/#oe mickey|cafe (n=M@p54B02369.dip0.t-ipconnect.de) |
| 12:19.39 | *** join/#oe vo5 (n=vo@bd21096b.virtua.com.br) |
| 12:45.27 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r517bcd18c8 10openembedded.git/recipes/dhcp/dhcp_3.0.2.bb: dhcp 3.0.2: fix QA error |
| 12:53.56 | *** join/#oe Laibsch (n=Laibsch@p5B3B2669.dip.t-dialin.net) |
| 13:02.05 | *** join/#oe waite (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 13:07.14 | mckoan | gremlin[it]: welcome back, it was a long time! :-D |
| 13:08.26 | gremlin[it] | mckoan, yes here on IRC are some weeks i didn't login ... |
| 13:11.14 | mckoan | gremlin[it]: too busy with the new home? |
| 13:12.10 | gremlin[it] | mckoan, with work and with home yes ... |
| 13:14.20 | mickey|cafe | florian: ping |
| 13:37.11 | *** join/#oe waite1 (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 13:39.19 | *** join/#oe Laibsch (n=Laibsch@p5B3B2669.dip.t-dialin.net) [NETSPLIT VICTIM] |
| 13:39.19 | *** join/#oe shoragan (n=shoragan@debian/developer/shoragan) [NETSPLIT VICTIM] |
| 13:39.37 | *** join/#oe waite (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 13:40.45 | *** join/#oe shoragan (n=shoragan@debian/developer/shoragan) |
| 13:42.22 | *** join/#oe ptitjes_ (n=didier@AVelizy-154-1-30-98.w82-124.abo.wanadoo.fr) |
| 13:43.08 | chouimat | morning |
| 13:43.26 | mickey|cafe | morning chouimat |
| 13:46.18 | *** join/#oe waite (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 13:49.25 | *** join/#oe aloisiojr (n=aloisio@200.184.118.130) |
| 14:00.00 | *** join/#oe aloisiojr1 (n=aloisio@200.184.118.130) |
| 14:01.09 | *** join/#oe MWelchUK_ (n=welchma@65.91.2.71) |
| 14:03.30 | *** join/#oe mickey|train (n=M@80.187.208.186) |
| 14:04.35 | *** join/#oe xxiao (i=c058a823@gateway/web/freenode/x-20115dea7d8eddcc) |
| 14:07.26 | *** join/#oe e-ffi (n=cybercom@dslb-088-068-162-126.pools.arcor-ip.net) |
| 14:08.19 | *** part/#oe DuckFault (n=DuckFaul@rrcs-71-43-244-114.se.biz.rr.com) |
| 14:11.13 | mickey|cafe_ | yo hrw |
| 14:11.17 | mickey|cafe_ | florian: ping |
| 14:11.25 | florian | mickey|cafe_: pong |
| 14:11.28 | mickey|cafe_ | ah |
| 14:11.46 | hrw | mickey|cafe_: elo |
| 14:11.51 | hrw | mickey|cafe_: Latte? |
| 14:11.53 | mickey|cafe_ | florian: GA invitations via E-Mail are valid. Where do we have a list of e-mail addresses for the founding members? |
| 14:11.57 | mickey|cafe_ | hrw: yeah :) |
| 14:12.21 | hrw | mickey|cafe_: going to ELC Europe? |
| 14:12.25 | gremlin[it] | hi mickey|cafe_ hrw florian |
| 14:12.39 | florian | mickey|cafe_: I have attaced adresses to the members list in git. All founding members are marked as well. |
| 14:13.12 | mickey|cafe_ | florian: ok, great. which repo again? (sorry, my memory) |
| 14:13.44 | mickey|cafe_ | hrw: good question... early november, right? |
| 14:13.55 | hrw | mickey|cafe_: 15-16 October |
| 14:14.05 | mickey|cafe_ | oh |
| 14:14.15 | gremlin[it] | I cannot be present at OE GA, please prepare the "proxy document" so i can sign it and send |
| 14:14.21 | florian | mickey|cafe_: the e.v. one with all the documents. if you don't have it at hand i can send you the file |
| 14:14.21 | hrw | mickey|cafe_: or even 14th if you count some workshops (I will be part of one) |
| 14:14.39 | mickey|cafe_ | florian: yes, please. i need to write that down somewhere |
| 14:14.48 | mickey|cafe_ | i will send the official invitation then |
| 14:14.50 | florian | mickey|cafe_: do we have a location? |
| 14:14.51 | mickey|cafe_ | gremlin[it]: will do |
| 14:15.02 | florian | mickey|cafe_: oh... good guy :-) |
| 14:15.12 | mickey|cafe_ | florian: Robert wanted to organize something in tarent |
| 14:15.39 | florian | is a little bit stressed... usually the summer is relaxed but this year is different |
| 14:18.59 | mickey|cafe_ | no problem |
| 14:19.06 | mickey|cafe_ | just mail me the doc and i'll take care |
| 14:21.20 | mickey|cafe_ | got it, thans |
| 14:22.41 | *** join/#oe pcgeil (n=steffen@p5B17DDC5.dip.t-dialin.net) |
| 14:22.48 | mickey|cafe_ | so, ich marschier mal heim |
| 14:22.53 | mickey|cafe_ | wanders home |
| 14:25.49 | *** join/#oe woglinde (i=woglinde@g225073001.adsl.alicedsl.de) |
| 14:29.34 | *** join/#oe florian (n=fuchs@port-217-146-132-69.static.qsc.de) |
| 14:29.50 | florian | re |
| 14:30.06 | woglinde | jo florian |
| 14:32.25 | rkirti | When one uses SRC_URI_append_foo , what is foo supposed to be ? is it restricted to ${MACHINE} ? |
| 14:35.50 | *** join/#oe rsalveti (n=rsalveti@200.184.118.130) |
| 14:36.01 | florian | rkirti: you can do things like SRC_URI_append_beagleboard |
| 14:39.02 | *** join/#oe flo_lap (n=fuchs@port-217-146-132-69.static.qsc.de) |
| 14:46.37 | pb_ | rkirti: see ${OVERRIDES} |
| 14:47.18 | *** join/#oe pirho (i=pirho@gateway/gpg-tor/key-0x2CEEC9CB) |
| 14:52.31 | *** join/#oe waite1 (n=bwaite@206.83.81.178) |
| 14:52.57 | *** join/#oe kergoth (n=kergoth@65.200.49.156) |
| 14:57.37 | *** join/#oe waite1 (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 15:09.16 | *** join/#oe ctusar (n=ctusar@router2.videon-central.net) |
| 15:21.11 | rkirti | florian, pb_ : thanks |
| 15:22.46 | sx | is there gst-plugin-sdlvideosink in OE angstrom ? |
| 15:23.08 | b0fh | sx: still nog progress with gstreamer? |
| 15:23.19 | sx | :) |
| 15:23.48 | *** join/#oe noglitch (n=Miranda@mail.atmel.fr) |
| 15:23.58 | b0fh | sx: I lack time (and proper gstreamer skills).. |
| 15:25.28 | sx | b0fh, I've tried it with mplayer. I firstly call the mplayer with system(""); and after X min stop it using killall but my gtk app doesn't start very fast |
| 15:25.44 | sx | I don't like the delay |
| 15:28.58 | b0fh | sx: you're trying something like this http://tinyurl.com/lbjgqs ? |
| 15:30.28 | sx | b0fh, you are the best :D |
| 15:30.40 | b0fh | xs: hartly :-P |
| 15:31.44 | b0fh | xs: that page was last updated in januari 2008, don't get your hopes up to high. Good luck! |
| 15:34.54 | *** join/#oe noglitch (n=Miranda@mail.atmel.fr) |
| 15:42.09 | sx | b0fh, thanks, I'll try this. |
| 15:43.38 | *** part/#oe ArteK (n=Artur@81.15.241.96) |
| 15:49.04 | *** join/#oe waite (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 15:55.20 | *** join/#oe Sup3rkiddo (n=sudharsh@unaffiliated/sudharsh) |
| 15:56.12 | *** join/#oe GNUtoo (n=gnutoo@host54-85-dynamic.51-79-r.retail.telecomitalia.it) |
| 16:01.13 | CIA-1 | 03Henning Heinold <heinold@inf.fu-berlin.de> 07org.openembedded.dev * r4e31a26bb6 10openembedded.git/conf/checksums.ini: checksum.ini: adding some java related stuff |
| 16:01.14 | CIA-1 | 03Henning Heinold <heinold@inf.fu-berlin.de> 07org.openembedded.dev * rf3f3a1378a 10openembedded.git/ (conf/checksums.ini recipes/ecj/libecj-bootstrap_3.5.bb): libecj-bootstrap: update to version 3.5 |
| 16:01.16 | CIA-1 | 03Henning Heinold <heinold@inf.fu-berlin.de> 07org.openembedded.dev * rfe804aca3e 10openembedded.git/recipes/classpath/classpath-tools-native.bb: |
| 16:01.16 | CIA-1 | classpath-tools-native: native package should inherit native |
| 16:01.17 | CIA-1 | * workaround for do_stage |
| 16:01.19 | CIA-1 | * not bump PR because package doesnt produce something |
| 16:01.21 | CIA-1 | 03Henning Heinold <heinold@inf.fu-berlin.de> 07org.openembedded.dev * r54c0d124fc 10openembedded.git/recipes/guile/guile-native_1.8.2.bb: |
| 16:01.24 | CIA-1 | guile-native: make sure we don't have -Werror as compiler flag |
| 16:01.26 | CIA-1 | * only for version 1.8.2 |
| 16:01.28 | CIA-1 | * bump PR |
| 16:01.30 | CIA-1 | 03Henning Heinold <heinold@inf.fu-berlin.de> 07org.openembedded.dev * r32a67c42a4 10openembedded.git/conf/distro/chinook-compat.conf: |
| 16:01.33 | CIA-1 | chinook-compat.conf: set preferred version for guile-native |
| 16:01.35 | CIA-1 | * avoids using of newer autotools |
| 16:06.35 | *** join/#oe pleemans (n=toi@d54C2AAB7.access.telenet.be) |
| 16:25.06 | *** join/#oe mnabil_ (n=mnabil@82.201.217.239) |
| 16:43.23 | kergoth | zecke: ping |
| 16:43.58 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r998d60e2d5 10openembedded.git/recipes/binutils/ (binutils-2.18/parse-neon-vmov.diff binutils_2.18.bb): |
| 16:43.58 | CIA-1 | binutils 2.18: don't emit symbols for neon registers |
| 16:43.58 | CIA-1 | * this fixes building in NEON code into binaries that do runtime detection but are compiled for non-NEON platforms (e.g. pixman) |
| 16:45.20 | zecke | kergoth: sorry, need to sleep first |
| 16:45.27 | kergoth | hehe, np |
| 16:45.35 | zecke | kergoth: or, whats up? |
| 16:46.05 | kergoth | 1) going to go ahead and change some of the ast nodes to accept arguments at construction rather than a regex match object, 2) please look at the bitbake git repo whenever you're awake and have the time ;) |
| 16:48.00 | zecke | kergoth: I did 2nd) a bit... I have no hard feeling |
| 16:48.13 | *** join/#oe pleemans (n=toi@d54C2AAB7.access.telenet.be) |
| 16:48.24 | kergoth | k, unless someone yells, i think i'll ask whomever maintains git.openembedded.org to set up the repo |
| 16:48.52 | zecke | kergoth: 1st) I wonder what to do for the actual parsing... I still want to have a grammar |
| 16:49.07 | kergoth | zecke: i wrote a replacement parser using pyparsing that expresses the grammar as pyparsing parse expression objects |
| 16:49.11 | kergoth | but its like twice as slow as regex |
| 16:49.13 | kergoth | it made me sad |
| 16:49.14 | kergoth | hehe |
| 16:49.24 | zecke | kergoth: 1st) I think something better would be adding profile of anon python funcs and reporting, I think the most speedup can be reached by fixing base.bbclass and autotools.bbclass to compute less and do the depdency injection differently |
| 16:49.24 | kergoth | still, was fun to play with |
| 16:49.35 | kergoth | yeah, that's a good idea |
| 16:50.52 | pb_ | morning kergoth |
| 16:51.57 | *** join/#oe ArteK (n=Artur@81.15.241.96) |
| 16:52.38 | zecke | nite |
| 16:56.52 | kergoth | hey pb_ |
| 16:59.39 | zecke | nite |
| 16:59.59 | kergoth | night zecke |
| 17:00.01 | kergoth | sleep well |
| 17:00.17 | zecke | kergoth: kergoth oh, so you agree that the change makes sense and we should continue from there? |
| 17:00.46 | kergoth | the parser rework in general? absolutely, it's nicely done |
| 17:02.10 | zecke | kergoth: cool, do 1) or my 1b)... and let us put the stuff on git.openembedded.org (or gitorious...). we need to ask RP to have a close look at the repository and let him enough time |
| 17:02.26 | kergoth | yeah, sounds good to me |
| 17:02.31 | remur_030 | how can I override the kernelsetting from the distro.conf file? |
| 17:02.58 | remur_030 | angstrom builds 2.6.21 with hh patches, I want to try 2.6.29 though |
| 17:03.39 | remur_030 | I modified the angstrom distrofile PREFERRED_VERSION_linux-handhelds-2.6 ?= "2.6.21-hh20" to say 2.6.29 here, this build n2.6.16 with hh patches though =( |
| 17:07.48 | *** join/#oe archae0pteryx (n=snewman@207.47.42.130.static.nextweb.net) |
| 17:08.47 | kergoth | remur_030: assuming a 2.6.29 hh recipe exists, you'd just set the preferred version, yes. if there isnt such a recipe however, that wont do much |
| 17:09.05 | kergoth | mickeyl: was it you that's running git.openembedded.org right now? |
| 17:09.38 | mickeyl | no longer, no |
| 17:09.45 | remur_030 | kergoth: linux_2.6.29.bb is in the recipes/linux directory, I assume that should do? |
| 17:09.47 | mickeyl | cbrake is now in charge of it |
| 17:10.04 | kergoth | remur_030: that's linux, not linux-handhelds, so setting the preferred version for linux-handhelds isn't going to do it |
| 17:10.14 | remur_030 | Oh no there is no hh recipe, I want to try with 2.6.29 |
| 17:10.19 | kergoth | remur_030: you'll need to set your preferred provider for virtual/kernel to Linux as well as setting the version to 2.6.29 |
| 17:10.26 | kergoth | cbrake: boo |
| 17:10.44 | remur_030 | argh, overread that it's for linux-handhelds |
| 17:11.39 | remur_030 | I assume this is done in build/conf/local.conf eh? |
| 17:12.42 | kergoth | local.conf is generally where you'd set such things, yes |
| 17:12.51 | kergoth | unless you're maintaining your own distro or something :) |
| 17:13.24 | remur_030 | nope I am starting slow =) |
| 17:13.36 | kergoth | good plan :) |
| 17:18.04 | CIA-1 | 03Stanislav Brabec <utx@penguin.cz> 07org.openembedded.dev * r335fde990d 10openembedded.git/ (4 files in 3 dirs): devhelp: New package, version 0.23. |
| 17:18.25 | remur_030 | kergoth: I don't really find documentation about how to use the providers |
| 17:18.48 | kergoth | PREFERRED_PROVIDER_<provider> = |
| 17:18.56 | kergoth | PREFERRED_PROVIDER_virtual/kernel = "linux" in your case, most likely |
| 17:19.13 | kergoth | every recipe provides functionality, PROVIDES lists the DEPENDS that it will be able to satisfy |
| 17:19.22 | kergoth | for most recipes, that's just package name, package name - package version, etc |
| 17:19.30 | kergoth | but some, like the kernel, add others. virtual/kernel, here |
| 17:19.52 | cbrake | kergoth: hello |
| 17:19.54 | kergoth | PREFERRED_PROVIDER is the way you tell bitbake which of the multiple recipes that satisfy that dependency you want it to actually build |
| 17:20.08 | kergoth | cbrake: can you create a bitbake git repo on git.openembedded.org, for me? |
| 17:20.29 | cbrake | kergoth: sure, just a minute |
| 17:20.31 | remur_030 | kergoth: ah I see, thanks |
| 17:20.33 | kergoth | cool, thanks |
| 17:20.39 | kergoth | remur_030: np |
| 17:20.46 | kergoth | bitbake -e linux|grep \^PROVIDES= |
| 17:20.55 | kergoth | bitbake -e|grep \^PREFERRED_PROVIDER_ |
| 17:27.56 | cbrake | kergoth: ok bitbake repo set up -- let me know if anything does not work |
| 17:30.12 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r734fd70cf4 10openembedded.git/recipes/dvsdk/ (2 files in 2 dirs): gstreamer-ti: add patch to hardcode combo location so users don't have to start apps from $datadir/ti/ all the time |
| 17:30.23 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * rd0e0bfef85 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev |
| 17:31.34 | *** join/#oe bluelightning (n=blueligh@pdpc/supporter/active/bluelightning) |
| 17:31.44 | kergoth | cbrake: anything special we need to do to get CIA notifications for the bitbake repo? |
| 17:33.48 | rkirti | what package exactly provides libintl.h ? Its there in my staging dir, but mysteriously missing in the image :-( |
| 17:35.03 | mario-goulart | rkirti: from libc? |
| 17:35.11 | cbrake | kergoth: hack some horrible scripts on the server |
| 17:35.14 | *** join/#oe jackpot (n=jackpot@216.58.130.98) |
| 17:35.24 | kergoth | arg, damnit, i need to fix up the tags |
| 17:35.26 | kergoth | grumbles |
| 17:35.34 | cbrake | kergoth: cgit seems broke for some mysterious reason |
| 17:35.38 | cbrake | kergoth: on the bitbake repo |
| 17:35.39 | kergoth | weird |
| 17:36.10 | kergoth | do me a favor and wipe the bitbake repo and re-create it, please. i don't want people using what i just pushed, cause i need to re-filter-branch the tags :) |
| 17:36.13 | rkirti | mario-goulart: not sure if its from libc. Looks like it is generated by passing certain options to gettext package. But even without gettext, I think I had it in my staging dir |
| 17:36.18 | kergoth | shakes head at himself |
| 17:36.35 | rkirti | mario-goulart: if it were libc, I cant explain why its missing in my image |
| 17:36.38 | jackpot | cbrake: have you looked at the mono git patches yet? |
| 17:37.27 | mario-goulart | rkirti: do you have libc headers on your image? Or just the libraries? |
| 17:38.01 | kergoth | cbrake: nevermind, i got it, don't bother wiping it (if you have, its no big deal, i can re-push, but dont need it cleared, got the tags fixed) |
| 17:38.10 | kergoth | forgot a git fetch --tags when i pulled over from a previous git repo :) |
| 17:38.57 | rkirti | mario-goulart: I am using a tweaked gpe-image |
| 17:39.50 | *** join/#oe xxiao (i=c058a823@gateway/web/freenode/x-3793d4f2db357ed1) |
| 17:39.50 | cbrake | kergoth: ok, wiped |
| 17:40.09 | cbrake | jackpot: I've looked at them, but I'm afraid that is about it |
| 17:40.24 | jackpot | cbrake: ok |
| 17:40.35 | cbrake | jackpot: let me get them applied and going on a test build ... |
| 17:40.46 | mario-goulart | rkirti: just a guess, but maybe you need to make your image depend on a -dev package from libc. |
| 17:40.47 | jackpot | cbrake: ok |
| 17:41.15 | kergoth | cbrake: that cgit error is weird. |
| 17:41.37 | cbrake | jackpot: it will probably be later today until I get my build updated |
| 17:41.42 | *** join/#oe mnabil_ (n=mnabil@82.201.217.239) |
| 17:41.50 | jackpot | cbrake: alright |
| 17:41.54 | rkirti | mario-goulart: it is a glibc header indeed. thanks. |
| 17:41.59 | woglinde | ~seen khem |
| 17:42.03 | ibot | khem <n=khem@adsl-71-146-11-93.dsl.pltn13.sbcglobal.net> was last seen on IRC in channel #oe, 18h 4m 30s ago, saying: 'hmm'. |
| 17:43.10 | mario-goulart | rkirti: you're welcome. |
| 17:44.11 | CIA-1 | 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * rfdbbb6d779 10openembedded.git/recipes/xorg-lib/pixman_git.bb: pixman git: only enable neon on armv7a platforms |
| 17:44.15 | kergoth | cbrake: could the cgit error be a permissions problem? did you create the repo with init --shared? *shrug* |
| 17:44.20 | woglinde | hm what was the trick to get old behavior for e2fsprog-libs back? |
| 17:44.30 | woglinde | need it for chinook-compar |
| 17:44.31 | woglinde | t |
| 17:44.40 | woglinde | which relays on older autotools |
| 17:44.40 | cbrake | kergoth: I created it with gitosis |
| 17:44.45 | kergoth | ah |
| 17:44.58 | kergoth | even more weird then, why would cgit think its not a valid repo.... hmm |
| 17:45.24 | cbrake | kergoth: yeah, ohh well, more things for the todo list ... |
| 17:45.29 | kergoth | hehe |
| 17:45.54 | kergoth | well at least its there now, i think this is as good as the repo is going to get, think i'll go toss a DONT_USE_THIS_USE_GIT in the svn repo or something |
| 17:46.13 | cbrake | kergoth: great, glad its being moved to git! |
| 17:46.45 | cbrake | kergoth: do you have a list of committers you want added? |
| 17:48.16 | kergoth | zecke, RP, hrw at a minimum. not sure who else messes with bitbake, but i guess any core type person :) |
| 17:53.37 | kergoth | erk, i hate it when i can't remember where i put some changes |
| 17:54.05 | kergoth | somewhere in these 34 or so build areas is a version of this collection with a change to autotools to patch it to remove the crosscompile fallback from AC_RUN_IFELSE |
| 17:58.18 | woglinde | ~lart util-linux-ng guys for not adding libtool macros in configure.ac |
| 17:58.18 | ibot | takes a rusty axe and swings it violently, taking util-linux-ng guys's head off for not adding libtool macros in configure.ac |
| 17:58.43 | woglinde | why always me |
| 17:58.47 | woglinde | *sigh* |
| 17:58.57 | woglinde | I should quit software stuff |
| 17:59.02 | woglinde | so many broken thinks |
| 17:59.03 | kergoth | what do you mean? all they have to do is call AC_PROG_LIBTOOL |
| 17:59.08 | rkirti | :-) autotools deals it blows again |
| 17:59.15 | woglinde | kergoth right |
| 17:59.22 | woglinde | and thats what the forget |
| 17:59.26 | kergoth | ugh |
| 17:59.46 | woglinde | and by the why AC_PROG_LIBTOOL shouldnt be used for newer configure.ac |
| 18:00.36 | woglinde | LT_INIT it is now called |
| 18:00.55 | kergoth | ah |
| 18:01.00 | kergoth | good to know |
| 18:02.12 | woglinde | *g* I have a bookmark folder with autoconf automake libtool m4 gettext so I can use open all in tabs option |
| 18:02.37 | kergoth | hehe, thats a good idea |
| 18:03.03 | woglinde | so now back to stinky util-linux-ng |
| 18:03.17 | woglinde | hm I sure whats going wrong |
| 18:03.17 | kergoth | anyone seen opkg leaving around /tmp/opkg-intercept-* dirs after a build from oe? its odd, i looked at the opkg code, and it shouldn't be happening unless the thing crashes, since it cleans up after itself |
| 18:03.26 | woglinde | the called autostuf ages ago |
| 18:03.36 | woglinde | and then shipping the copy libtool script |
| 18:04.00 | woglinde | and then somehow removed the libool macro |
| 18:04.32 | woglinde | and nobody cared because the always used configure alone |
| 18:05.01 | kergoth | smooth |
| 18:05.29 | woglinde | but I am puzzled why they was so clever to use the m4-dirmacro which 80% dont specify |
| 18:05.33 | woglinde | were |
| 18:08.34 | woglinde | hm okay |
| 18:08.42 | woglinde | they are using LT_INIT |
| 18:08.57 | woglinde | so I was unfair |
| 18:13.09 | *** join/#oe waite1 (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 18:13.54 | *** join/#oe stephank (n=traveler@2002:52c5:cf78:7:21c:c4ff:fece:ea94) |
| 18:21.17 | remur_030 | hrm setting PREFERRED_PROVIDER_virtual/kernel= in build/conf/local.conf doesn't get imported it seems? it still only uses the one from the angstrom descriptiob |
| 18:23.15 | kergoth | is angstrom not defining it with ?= |
| 18:23.31 | kergoth | the distro .conf is loaded after local.conf, so if it doesn't use ?=, which is 'set only if unset', it will override you |
| 18:23.51 | kergoth | you can use OVERRIDES to override the distro from local.conf, however, because override specific variables are processed at the very end of the parsing |
| 18:23.59 | kergoth | PREFERRED_PROVIDER_virtual/kernel_local = "linux" |
| 18:24.07 | kergoth | bitbake -e | grep \^OVERRIDES= |
| 18:25.45 | *** join/#oe archae0pteryx1 (n=snewman@207.47.42.130.static.nextweb.net) |
| 18:25.52 | woglinde | hm kergoth the problem remains with older libtool stuff |
| 18:26.26 | kergoth | hmm, i wonder if we could use the python modules for looking at syntax trees of python code to analyze the inline python snippets to be able to safely cache their results |
| 18:26.29 | woglinde | seems somthing in libtool commandline parsing goes wrong |
| 18:26.46 | kergoth | whats the problem you're seeing? |
| 18:28.02 | remur_030 | kergoth: you are amazing, the ?= bet was right, angstrom sets PREFERRED_PROVIDER_virtual/kernel with = |
| 18:28.07 | woglinde | ../../../libtool: line 764: X--tag=CC: command not found |
| 18:28.26 | kergoth | ugh |
| 18:28.32 | woglinde | but strangly some other libtool commands above are working |
| 18:28.33 | kergoth | I've seen that, and i don't remember how i got around it |
| 18:28.36 | kergoth | :( |
| 18:28.58 | woglinde | I am forced to using older autotools stuff because of chinook-compat |
| 18:31.19 | woglinde | hm okay has to do with libtool 1.5 and 2.x mix |
| 18:31.58 | rkirti | notices libc.a not being part of her image |
| 18:32.02 | rkirti | is that usual ? |
| 18:32.25 | kergoth | I'm not sure why you'd want a .a in your image, unless you're going to be doing development on the target |
| 18:32.32 | kergoth | either way, add libc6-dev to your image if you want it |
| 18:32.44 | rkirti | kergoth: libc.so is missing too |
| 18:32.51 | kergoth | same. |
| 18:33.25 | kergoth | you don't need libc.so unless you're developing |
| 18:33.33 | kergoth | you have /lib/libc.so.6, which is what things end up linked against |
| 18:34.12 | kergoth | generally, libwhatever.so is a symlink to the real soname of the library, to make it easier to link against. |
| 18:34.24 | kergoth | libc is slightly more complex since it provides /usr/lib/libc.so as a linker script, not a symlink, but its the same idea |
| 18:34.40 | rkirti | kergoth: I see. thanks |
| 18:34.42 | rkirti | kergoth: the only reason I am tracing is because I want localization support, which in this case translates to calls to the dgettext function. |
| 18:34.44 | kergoth | np |
| 18:34.47 | woglinde | hm ah gotcha |
| 18:35.02 | woglinde | m4 dir is full of the newer libtool macros |
| 18:35.55 | *** join/#oe waite (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 18:37.10 | woglinde | hm |
| 18:37.33 | woglinde | how do I do_configure prepend only for chinook_compat distro |
| 18:38.13 | Tartarus | do_configure_prepend_chinook_compat doesn't work due to too many _s ? |
| 18:38.32 | woglinde | hm |
| 18:38.34 | woglinde | dont know |
| 18:38.40 | woglinde | but dont believe |
| 18:38.43 | woglinde | let me try |
| 18:39.35 | kergoth | chinook_compat isn't in OVERRIDES, pn-chinook_compat is |
| 18:39.41 | kergoth | bitbake -e | grep \^OVERRIDES= |
| 18:40.00 | *** join/#oe Gnutoo (n=gnutoo@host54-85-dynamic.51-79-r.retail.telecomitalia.it) |
| 18:40.14 | kergoth | do_configure_prepend_pn-chinook_compat () should work |
| 18:40.16 | woglinde | aehm |
| 18:40.28 | Tartarus | er, distro is chinook-compat isn't it? |
| 18:40.31 | Tartarus | or did it get renamed? |
| 18:40.34 | woglinde | whats the pn doing? |
| 18:40.42 | woglinde | tartarus right |
| 18:40.58 | kergoth | oh, right, I'm a moron, misread :) |
| 18:41.10 | Tartarus | so: do_configure_prepend_chinook-compat should work |
| 18:41.15 | Tartarus | right? :) |
| 18:45.36 | woglinde | hm looking fine it seemd |
| 18:47.46 | woglinde | strike it worked out |
| 18:48.33 | *** join/#oe TheJim01 (n=rjh@74-93-123-129-pennsylvania.hfc.comcastbusiness.net) |
| 18:48.49 | TheJim01 | Hello |
| 18:50.13 | woglinde | hm what the hell libsm needs e2fsprogs-lib? |
| 18:51.53 | remur_030 | kergoth: the prefered kernel gets chosen in the machine.conf, not the angstrom.conf |
| 18:52.06 | remur_030 | just in case another one asks in the future =) |
| 18:52.16 | kergoth | same answer, though :) |
| 18:52.20 | remur_030 | yeah |
| 18:52.33 | kergoth | both machine and distro are loaded after local.conf, because local.conf is where MACHINE/DISTRO are usually defined |
| 18:52.40 | kergoth | which annoys me, but we can't fix that anytime soon :) |
| 18:53.49 | remur_030 | I read somewhere there is a way to define a chain, guess one could easily set up a pre and post .conf file |
| 18:54.54 | *** join/#oe jconnolly (n=jconnoll@firebug.buglabs.net) |
| 18:56.31 | *** join/#oe oneshel (n=jim@c-98-216-198-0.hsd1.nh.comcast.net) |
| 18:58.20 | TheJim01 | I know this isn't a distro channel, but is anyone familiar with poky? |
| 18:58.45 | *** join/#oe pcgeil (n=steffen@p5B17DDC5.dip.t-dialin.net) |
| 18:59.01 | *** join/#oe timtimred (n=meh@79-75-141-152.dynamic.dsl.as9105.com) |
| 19:00.12 | *** join/#oe grma (n=gruberm@chello212186029093.tirol.surfer.at) |
| 19:12.12 | *** join/#oe florian (n=fuchs@f054216153.adsl.alicedsl.de) |
| 19:12.29 | florian | re |
| 19:17.19 | kergoth | TheJim01: I'm sure plenty of folks have poked around at it, not sure if any experts with it are around, though. just ask the question and wait and see :) |
| 19:20.23 | kergoth | whee, i think adding limitations to inline python snippets will be quite easy with the ast module in 2.6, once we switch to it |
| 19:20.32 | kergoth | the NodeVisitor is quite easy to use |
| 19:20.56 | kergoth | just tested a visitor that freaks if i try to look up any names other than bb.which or bb.data.getVar :) |
| 19:22.25 | kergoth | should poke at compiler.ast, too |
| 19:23.07 | kergoth | wait, hmm |
| 19:23.20 | *** join/#oe waite (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 19:29.36 | *** join/#oe user2 (n=3MX@80.48.165.42) |
| 19:30.54 | CIA-1 | 03Stefan Schmidt <stefan@datenfreihafen.org> 07org.openembedded.dev * radd719a9d4 10openembedded.git/conf/checksums.ini: |
| 19:30.54 | CIA-1 | checksums.ini: Sort newly added java checksums. |
| 19:30.54 | CIA-1 | This helps to avoid fuzz in other changes to this file. |
| 19:33.08 | *** join/#oe Spyro (n=ian@5aca0349.bb.sky.com) |
| 19:33.17 | TheJim01 | I'm not sure if it's an issue with poky, or an issue with OpenEmbedded. When I try to bitbake their example, I get an error that oe_runmake failed in gmp-native-4.2.1. It seemed to be building fine before that. Any idea what could be causing that? |
| 20:01.59 | kergoth | you'll need to be more specific |
| 20:02.09 | kergoth | when a compile fails, put the do_compile log on a pastebin |
| 20:03.34 | *** join/#oe waite1 (n=bwaite@206.83.81.178.ptr.us.xo.net) |
| 20:07.57 | kergoth | http://gist.github.com/152235 .. nifty |
| 20:10.23 | *** part/#oe ArteK (n=Artur@81.15.241.96) |
| 20:10.39 | TheJim01 | Sorry, I should have thought of that earlier. http://pastebin.com/m629b5035 |
| 20:13.04 | woglinde | hrms libsm needs libuuid |
| 20:13.11 | woglinde | thats where the mess comes from |
| 20:23.27 | *** join/#oe mwester (n=mwester@nslu2-linux/mwester) |
| 20:28.59 | remur_030 | how do I properly add a kernelconfig for machine=magician to linux_2.6.29.bb? I created recipes/linux/linux/magician and added the defconfig there, as it seems to autoload these for all the other machines |
| 20:29.22 | woglinde | nope |
| 20:29.22 | remur_030 | but that was unsuccessful |
| 20:29.25 | woglinde | you have to do it |
| 20:29.44 | woglinde | in recipes/linux-versionbla/ |
| 20:29.47 | woglinde | ups |
| 20:29.59 | woglinde | recipes/linux/linux-versionbla/ |
| 20:30.32 | woglinde | where linux-versionbla is either plain vanillla or something strange like -rp_herr_trunk_git-4556 |
| 20:30.54 | woglinde | depending on the machine.conf for the device |
| 20:30.57 | remur_030 | uh ok, but he loads linux.inc so I thought it'd use the linux directory as well, or how does one tell it's alix build to use the defconfig in linux/linux/alix? |
| 20:31.32 | remur_030 | woglinde: ok I'll do this, I just don't get what all the defconfs in /linux/linux/$MACHINE are doing there? |
| 20:31.57 | woglinde | please look under recipes/linux/linux-2.6.28/ or recipes/linux/linux-2.6.29/ and you will dig it out |
| 20:33.08 | remur_030 | woglinde: yeah these are quite empty, I take it this is because the defconfigs weren't just moved over? |
| 20:33.31 | woglinde | haeh? |
| 20:33.40 | woglinde | then you dont using oe-dev |
| 20:33.55 | remur_030 | uh no, stable/2009 |
| 20:34.07 | woglinde | sorry I dont use stable |
| 20:34.19 | woglinde | so cant say how its handeld there |
| 20:38.00 | remur_030 | woglinde: oh right, in git they are filled, hrm, still I am somewhat confused about it |
| 20:39.32 | remur_030 | anyways he still doesn't respect the new defconfig in that directory either, shouldn't he automatically rebuild the kernel? |
| 20:39.44 | woglinde | nope |
| 20:40.03 | remur_030 | bitbake virtual/kernel then? |
| 20:40.07 | woglinde | do bitbake -c configure -f linux-image |
| 20:40.48 | remur_030 | nothing provides linux-image? |
| 20:41.01 | woglinde | hm yes then virtual/kernel |
| 20:41.11 | remur_030 | already tried, same |
| 20:41.19 | woglinde | hm |
| 20:41.21 | woglinde | strange |
| 20:41.29 | remur_030 | nvm, type |
| 20:41.41 | remur_030 | *typo, jesus! |
| 20:42.14 | woglinde | *g* |
| 20:52.32 | *** join/#oe ant__ (n=andrea@host47-49-dynamic.5-87-r.retail.telecomitalia.it) |
| 20:56.37 | remur_030 | woglinde: working perfectly (I guess, the built modules look right), thanks |
| 20:58.05 | woglinde | no prob |
| 20:58.22 | woglinde | be lucky you dont hit the problems I hit |
| 20:58.52 | *** join/#oe Martin-B (n=Martin@pool-41-65-198-89.dbd-ipconnect.net) |
| 21:14.55 | *** join/#oe brolin (n=brolin@186.81.88.253) |
| 21:25.10 | *** join/#oe woglinde (i=woglinde@78.52.225.154) |
| 21:28.27 | *** join/#oe putumutukas (n=putukas@132.29.190.90.dyn.estpak.ee) |
| 21:28.49 | putumutukas | Crofton do you happen do be around ? |
| 21:29.12 | Crofton | yeah |
| 21:29.18 | Crofton | but I haven't found an osk :( |
| 21:29.26 | putumutukas | i've been trying to experiment with my omap |
| 21:29.35 | Crofton | any luck? |
| 21:29.48 | *** join/#oe dijenerate (n=dijenera@64.210.44.91) |
| 21:30.00 | putumutukas | and found out that the no space error comes from the failsystem |
| 21:30.17 | putumutukas | 'cause i used a .29 kernel |
| 21:30.33 | putumutukas | and put my old working fs from .20 kernel |
| 21:30.38 | putumutukas | and i run |
| 21:30.48 | putumutukas | although missed the modules and stuff |
| 21:31.23 | Crofton | ok |
| 21:31.34 | putumutukas | do you have any idea what i should try or look for |
| 21:31.44 | Crofton | so maybe try commenting stuff out in fstab? |
| 21:32.11 | putumutukas | how can i do that *-) |
| 21:32.53 | Crofton | you are building the fs using OE? |
| 21:33.05 | putumutukas | yes |
| 21:33.46 | Crofton | look at the basefiles recipe |
| 21:34.10 | Crofton | in the base-files directory andd a directory called omap5912osk |
| 21:34.19 | Crofton | and put your new fstab there |
| 21:35.10 | putumutukas | the fstab is there |
| 21:35.18 | putumutukas | in omap5912osk dir |
| 21:35.32 | putumutukas | # fstab for omap5912osk |
| 21:35.32 | putumutukas | rootfs / auto defaults 1 1 |
| 21:35.32 | putumutukas | proc /proc proc defaults 0 0 |
| 21:35.32 | putumutukas | sysfs /sys sysfs defaults 0 0 |
| 21:35.32 | putumutukas | tmpfs /dev tmpfs defaults 0 0 |
| 21:35.32 | putumutukas | devpts /dev/pts devpts gid=5,mode=620 0 0 |
| 21:35.33 | Crofton | ok, so make changes to that |
| 21:35.34 | putumutukas | tmpfs /dev/shm tmpfs mode=0777 0 0 |
| 21:35.35 | *** join/#oe Spyro (n=ian@5aca0349.bb.sky.com) |
| 21:35.36 | putumutukas | tmpfs /var/volatile tmpfs mode=0755 0 0 |
| 21:35.38 | putumutukas | tmpfs /media/ram tmpfs defaults 0 0 |
| 21:35.51 | Crofton | what is the actual error again? |
| 21:36.07 | putumutukas | udevd-event[2054]: unable to create db file '/dev/.udev/db/\x2fdevices\x2fplatfo |
| 21:36.08 | putumutukas | rm\x2fohci\x2fusb1\x2fusb_endpoint\x2fusbdev1.1_ep00': No space left on device |
| 21:36.27 | putumutukas | the errors come after i enamble the tmpfs in defconfig |
| 21:36.34 | Crofton | hmm |
| 21:36.37 | putumutukas | before that it builds the kernel and fs |
| 21:36.48 | putumutukas | but lacks to use usb and other stuff |
| 21:36.59 | Crofton | tmpfs is ram based |
| 21:37.05 | Crofton | I wonder what is running it out of ram |
| 21:38.19 | putumutukas | in my silly mind it's somehow connected to the last 3 lindes which include the magic word tmpfs :) |
| 21:38.24 | putumutukas | lines |
| 21:40.39 | remur_030 | woglinde: what problems did you hit with the new kernel? |
| 21:43.24 | woglinde | remur its not about kernelstuff |
| 21:45.38 | remur_030 | ok |
| 21:50.06 | *** join/#oe Widget (n=widget@87-194-215-95.bethere.co.uk) |
| 21:50.12 | Widget | 'lo |
| 21:54.21 | woglinde | lol |
| 21:54.28 | Widget | i'm halfway through an OE build at work and have run out of disc space - can I apply a rm_work retroactively across what I've built already? |
| 21:54.44 | woglinde | you can update from an older windows version to windows 7 |
| 21:55.02 | jackpot | barely |
| 21:55.17 | woglinde | widget yes |
| 21:55.34 | Widget | woglinde: good. how? :) |
| 21:55.47 | woglinde | put INHERIT += "rm_work" into your local.conf |
| 21:56.15 | woglinde | so the last step for each recipe is the rm |
| 21:56.30 | Widget | won't that only apply to new packages built from that point on? |
| 21:56.30 | woglinde | only the temp dir with the logs will be left over |
| 21:56.39 | woglinde | right |
| 21:56.40 | Tartarus | Widget, no, it should work |
| 21:56.51 | woglinde | tartarus hm |
| 21:56.57 | Tartarus | Since a new task will have been added, and not yet done |
| 21:57.11 | Widget | the disc is unfortunately full now, and I'm barely a third of the way in, with 7.5G of arm binaries.. |
| 21:57.31 | Tartarus | Widget, did you do a 'bitbake world' ? or is that the usage on 'tmp' ? |
| 21:57.48 | Widget | bitbake omap3-minimal-image |
| 21:58.08 | Tartarus | k |
| 21:58.25 | Tartarus | So anyhow, yes, you should be able to bring in rm_work, and stuff will get deleted |
| 21:58.48 | remur_030 | qt4-embedded is taking ages, is there a way I can track an ongoing compile? |
| 21:58.55 | woglinde | remur yes |
| 21:59.25 | remur_030 | woglinde: I hope you'll be so kind to tell me =) |
| 21:59.28 | woglinde | look at $oetmpdir/work/archyoubildfor/qt-emebdded-version/temp/log.do_compile.pid |
| 21:59.47 | Widget | hmm, my disc is perhaps not large enough at all then, as I enabled it and I've not seen any appreciable drop - might find a spare USB HDD and build on that |
| 22:00.02 | woglinde | remur and yes |
| 22:00.09 | woglinde | qt is one of the longest build |
| 22:03.19 | remur_030 | woglinde: ohhhh it really finished, thanks for the tip with the do_compile |
| 22:03.30 | remur_030 | must have been close to half an hour |
| 22:04.11 | Widget | an unrelated question, can bitbake be told to explain its dependency choices? I asked it to build ppp and it started libX11 |
| 22:04.19 | remur_030 | it said NOTE: nothing to compile even though it was still running |
| 22:04.31 | *** join/#oe woglinde_ (i=woglinde@78.52.236.192) |
| 22:04.53 | *** join/#oe brolin (n=brolin@186.81.88.253) |
| 22:07.20 | mwester | Widget, look at the bitbake options -- you can cause it to emit the dependencies as a "dot" graph file. |
| 22:07.58 | Widget | mwester: perfect, thanks very much |
| 22:11.51 | Widget | i think i shall get some sleep and try these tips at work tomorrow |
| 22:11.53 | Widget | thanks guys |
| 22:12.01 | woglinde | good nite |
| 22:12.02 | *** join/#oe Spyro (n=ian@benden.mnementh.co.uk) |
| 22:12.04 | woglinde | hm too late |
| 22:15.28 | *** join/#oe guillaum1 (n=gl@AMontsouris-153-1-87-171.w90-2.abo.wanadoo.fr) |
| 22:24.33 | remur_030 | why can't one unpack the images using tar xvjf btw? |
| 22:27.35 | woglinde | hm??? |
| 22:29.42 | *** part/#oe jackpot (n=jackpot@216.58.130.98) |
| 22:30.20 | remur_030 | I tried unpacking the Angstrom rootfs tar.bz2, using plain tar xvjf it fails, if i use bzcat and pipe it into tar x it works |
| 22:31.02 | RP | kergoth: cool on the bitbake repo, will take a look asap |
| 22:31.06 | woglinde | dont know |
| 22:31.10 | woglinde | he rp |
| 22:31.22 | RP | hi woglinde |
| 22:31.27 | putumutukas | Crofton seems that if i comment the lines out in fstab, -c clean console-image and then build it again the file size remains the same :s |
| 22:33.48 | putumutukas | and of course the not-enough-space problem as well |
| 22:34.25 | putumutukas | can i somehow get the bitbake to rebuild the base files as well ? |
| 22:35.22 | woglinde | bitbake -c clean basefiles && bitbake basefiles |
| 22:35.45 | putumutukas | thank you woglinde :) |
| 22:57.49 | *** join/#oe putukas (n=putukas@132.29.190.90.dyn.estpak.ee) |
| 23:05.56 | kergoth | RP: mind looking at the commit on the tip of kergoth/cacheinvalidate in the bitbake git repo? feel free to tell me I'm crazy |
| 23:50.32 | *** join/#oe fraxinas (n=quassel@p54AA5B1F.dip.t-dialin.net) |
| 23:53.14 | *** join/#oe JaMa (n=martin@161-24.13.24.78.awnet.cz) |
| 23:55.03 | woglinde | wuha |
| 23:55.30 | woglinde | after fighting 4 days or more, I can build java packages for n8xx again |
| 23:55.39 | woglinde | good nite |