**** BEGIN LOGGING AT Thu May 06 02:59:56 2010 May 06 05:30:56 morning everyone May 06 05:35:07 morning May 06 06:02:26 hello JaMa May 06 06:09:28 mmm May 06 06:09:34 hi May 06 06:09:37 QA Issue with staging: fso-specs.pc failed sanity test (tmpdir) in path /home/embedded/oetmp_eeepc701/staging/i686-angstrom-linux/usr/lib/pkgconfig May 06 06:09:42 what does it means May 06 06:09:44 ? May 06 06:09:54 I should look in that pc I bet May 06 06:09:55 but May 06 06:10:02 it didn't do that for arm May 06 06:10:33 but arm was cleaner(packaged staging) May 06 06:12:24 strange strange May 06 06:12:46 fso-specs.pc is not in gnash sources May 06 06:14:24 ah it's a fso thing May 06 06:14:27 I'll delete that May 06 06:16:21 GNUtoo|oeee: I have a patch for that May 06 06:17:26 ah May 06 06:17:44 is it pushed or should I test it? May 06 06:17:52 http://pastebin.com/q3gnbqze May 06 06:17:55 no its not pushed May 06 06:18:05 nor proposed yet May 06 06:18:10 ok May 06 06:18:28 I must have compiled fso by error May 06 06:18:37 because the target is eee701 May 06 06:19:14 or maybe it compiled it as a dep of tangogos or another package May 06 06:21:42 yeah minimal is pulling it in May 06 06:21:53 I guess if you have DISTRO=debug May 06 06:22:01 it pulls lot of fso stuff May 06 06:22:06 due to cli-tool May 06 06:22:11 cli-tools May 06 06:23:10 ok May 06 06:36:52 thanks a lot btw May 06 06:41:07 03Denys Dmytriyenko  07org.openembedded.dev * r7fffc9eaeb 10openembedded.git/recipes/ti/ (matrix-gui/init matrix-gui_svn.bb): (log message trimmed) May 06 06:41:07 matrix-gui: numerous updates and fixes (from Arago) May 06 06:41:07 * Bump the revision May 06 06:41:07 * Fix the license May 06 06:41:08 * Don't set redundant variables May 06 06:41:08 * Display message when calibrating TS May 06 06:41:09 * Install additional binaries/scripts May 06 06:41:11 khem: btw, have you seen that kernel oops log we talked about? May 06 06:41:14 03Denys Dmytriyenko  07org.openembedded.dev * r8a0750d209 10openembedded.git/recipes/ti/am-benchmarks_svn.bb: May 06 06:41:14 am-benchmarks: add ARM Benchmarks dhry/whet-stone, linpack (from Arago) May 06 06:41:14 Signed-off-by: Denys Dmytriyenko May 06 06:41:14 03Denys Dmytriyenko  07org.openembedded.dev * r819cd0db4d 10openembedded.git/recipes/ti/am-sysinfo_svn.bb: May 06 06:41:14 am-sysinfo: add AM SysInfo set, currently mem_util only (from Arago) May 06 06:41:14 Signed-off-by: Denys Dmytriyenko May 06 06:41:15 03Denys Dmytriyenko  07org.openembedded.dev * r49d7eff0e9 10openembedded.git/recipes/ti/ (matrix-tui/init matrix-tui_svn.bb): May 06 06:41:15 matrix-tui: add the Matrix TUI (Text UI) launcher app (from Arago) May 06 06:41:16 Signed-off-by: Denys Dmytriyenko May 06 06:41:27 JaMa: ben a bit budy thesr days May 06 06:42:11 hehe :) May 06 06:42:54 ok, let me know when you have more time and can look at it May 06 06:43:10 * denix0 needs a decypher for this piece: "ben a bit budy thesr days" :) May 06 06:43:53 * JaMa needed 43 sec to decypher it.. May 06 06:44:06 maybe 42, that's it! May 06 06:44:45 although I just found couple typos in my own patch sent to the list... d'oh May 06 06:51:58 denix0: thats how you type when you have 2yr old in one arm :) May 06 06:52:24 heh :) May 06 06:56:21 03Koen Kooi  07org.openembedded.dev * rddceb0bc5e 10openembedded.git/recipes/qt4/qt4-tools-sdk_4.6.2.bb: May 06 06:56:22 qt4-tools-sdk: symlink tools back to their original names to make e.g. eclipse plugins work May 06 06:56:22 Change imported from http://arago-project.org/git/?p=arago.git;a=commitdiff;h=5bbda8657da241ca7c5a48cc497ca1506d6b05b0 May 06 06:56:32 03Koen Kooi  07org.openembedded.dev * r8677716219 10openembedded.git/recipes/mesa/mesa-xlib_7.8.bb: mesa-xlib: add 7.8 May 06 07:03:07 denix0: what is arago project about ? May 06 07:04:22 khem: TI (pico-)distro :) May 06 07:04:58 basically, a playground for stabilizing OE for TI products... May 06 07:06:19 I see May 06 07:06:25 technically, it's just an overlay on top of OE.dev with small fixes and mods May 06 07:06:47 denix0: and which distro do you inherit from ? May 06 07:06:51 most of the stuff ends up in OE anyway May 06 07:06:52 denix0: minimal ? May 06 07:06:59 actually, angstrom May 06 07:07:13 when I started 2 years ago, there was no minimal... May 06 07:07:18 denix0: ok May 06 07:07:28 denix0: you may try now though May 06 07:07:36 I know :) May 06 07:07:57 and it uses eglibc as its main library thats what I am interested in May 06 07:08:02 good morning May 06 07:08:22 hello dth May 06 07:08:53 03Koen Kooi  07org.openembedded.dev * r3359bc5124 10openembedded.git/recipes/libtool/libtool-native_2.2.6b.bb: libtool-native 2.2.6b: convert to new style staging May 06 07:10:40 khem: there are few reasons to keep own mods because of corporate/legal limitations - GPLv3 and patents/IP, export of crypto algos and external CSL toolchain May 06 07:11:28 although, external CSL toolchain is now in OE and works at least in Angstrom... May 06 07:11:30 denix0: I understand May 06 07:11:33 03Koen Kooi  07org.openembedded.dev * rc73f8d610a 10openembedded.git/recipes/unifdef/unifdef-native_2.6.18+git.bb: unifdef-native: convert to new style staging May 06 07:13:48 03Koen Kooi  07org.openembedded.dev * rb1f53557c5 10openembedded.git/recipes/pax-utils/ (4 files): pax-utils: convert to new style staging and BBCLASSEXTEND May 06 07:27:01 hi, How can I hide mouse cursor in Angstrom, any Xorg option is there? May 06 07:27:46 enlightenment is the window manager May 06 07:30:23 Hi1 .... are there any reason why I can't get LD_PRELOAD to work on a glibc target.... apparently LD_PRELOAD is ignore..? May 06 07:30:30 bala_holmes: -nocursor May 06 07:31:01 ANG root@zjama ~ $ Xorg -help 2>&1 | grep cursor May 06 07:31:01 -nocursor disable the cursor May 06 07:43:58 * denix0 wonders what OE devs use as a desktop linux distro... May 06 07:48:52 denix0: you mean build machine ? May 06 07:49:04 yep May 06 07:49:09 I use kubuntu May 06 07:49:17 sometimes suse and fedora May 06 07:49:49 03Khem Raj  07org.openembedded.dev * rc65b917370 10openembedded.git/recipes/gdb/gdb-cross_7.1.bb: May 06 07:49:49 gdb-cross_7.1.bb: Convert to new staging. May 06 07:49:49 Signed-off-by: Khem Raj May 06 07:49:50 03Khem Raj  07org.openembedded.dev * r4134f55dc7 10openembedded.git/recipes/eglibc/ (6 files): May 06 07:49:50 eglibc: Move away from legacy staging. May 06 07:49:50 Signed-off-by: Khem Raj May 06 07:49:51 03Khem Raj  07org.openembedded.dev * rb221714c84 10openembedded.git/recipes/gcc/ (5 files): May 06 07:49:51 gcc-4.5.0: Move away from legacy staging. May 06 07:49:52 * Define EXTRA_OECONF_INITIAL and EXTRA_OECONF_INTERMEDIATE variables May 06 07:49:52 * Remove redundant require statements. May 06 07:49:53 * Cleanup DEPENDS and use NATIVEDEPS. May 06 07:49:53 Signed-off-by: Khem Raj May 06 07:50:16 khem: ah, thanks. I was just going to look at that :) May 06 07:50:32 khem: thanks for unblocking it! May 06 07:50:51 RP: np May 06 07:51:18 I am trying to get a build from scratch using gcc-4.5.0 and uclibc with minimal-uclibc DISTRO May 06 07:52:55 gcc recipes look a lot cleaner now at least :) May 06 07:53:40 RP: http://pastebin.com/NLUHHbzM May 06 07:53:54 is that gettext thing May 06 07:54:26 khem: yes, that one was mentioned on the mailing list - its missing an inherit gettext May 06 07:55:14 http://thread.gmane.org/gmane.comp.handhelds.openembedded/32386 May 06 07:56:47 I want a PySide recipe in OE, but don't have the skills to make it myself. Do I have any other options? Ask nicely (beg and plead) on IRC/Mailing List? May 06 07:57:48 tasslehoff: heh, May 06 07:58:55 03Richard Purdie  07org.openembedded.dev * r1b6f617f5f 10openembedded.git/ (3 files in 3 dirs): May 06 07:58:55 classes/recipes: Catch populate_staging references that were missed in the first pass May 06 07:58:55 Signed-off-by: Richard Purdie May 06 07:59:15 RP http://pastebin.com/bfj4144w May 06 08:00:08 tasslehoff: have you tried to make one on your own? May 06 08:01:10 khem: maybe you forget to bump INC_PR in eglibc staging commit? May 06 08:01:37 khem: I'm rebuilding from scratch.. so I can clean it manually now again.. but maybe someone is not May 06 08:01:50 khem: libtool's staging function doesn't make the directory first May 06 08:01:55 JaMa, I think one needs to build from scratch :) May 06 08:02:08 khem: ok, fair enough :) May 06 08:02:19 RP: koen committed a change just few mins back May 06 08:02:26 may be he forgot May 06 08:02:35 RP, can you please give me a hint whats wrong? http://pastebin.org/204338 May 06 08:02:37 Heinervdm: No. I have tried to make PyQt/SIP compile in dev without luck, and that's where the "don't have the skills" came from :) May 06 08:02:54 khem: I wanted to ask before rebuilding (to save another rebuild in case someone bump it in the end :)) May 06 08:03:22 khem: He converted it but needs an install -d ${D}${bindir} May 06 08:03:28 no, I think with RP's branch merge you have to rebuild everything May 06 08:03:39 I do.. but I started too soon May 06 08:03:50 but, perhaps PySide proves easier. May 06 08:03:53 so I have virtual/kernel and 1/4 image already built May 06 08:04:12 tasslehoff: i think pySlide should be easier, just take one of the ofter python-py* recipes (one of the easy one with just few lines) and replace everything May 06 08:04:16 khem: it shouldn't require a rebuild May 06 08:04:31 khem: although testing one would probably be wise May 06 08:04:50 tasslehoff: it seems to depend only on python-pygame May 06 08:05:06 dth: That looks odd, not sure there is enough context there to comment... May 06 08:05:22 dth: which bitbake version is it? May 06 08:06:09 RP, http://download.berlios.de/bitbake/bitbake-1.8.18.tar.gz May 06 08:06:51 RP: all staging is renamed May 06 08:07:05 03Richard Purdie  07org.openembedded.dev * r0c07ef24bf 10openembedded.git/recipes/libtool/libtool-native_2.2.6b.bb: May 06 08:07:05 libtool-native: Add missing directory creation May 06 08:07:05 Signed-off-by: Richard Purdie May 06 08:07:11 khem: Yes, but we do compat symlinks :) May 06 08:07:20 oh May 06 08:07:49 Heinervdm: PySide, not PySlide :) May 06 08:08:17 Heinervdm: Qt's own Python bindings May 06 08:08:35 oh :) May 06 08:08:47 gn May 06 08:11:31 RP: http://pastebin.com/TuVfrYVk does not seem to fix it May 06 08:12:47 khem: I just pushed a fix, try that May 06 08:13:00 RP: oh i should have put it atop May 06 08:13:06 ok will pull May 06 08:15:50 RP: next one http://pastebin.com/jx0G4d7e :0 May 06 08:17:24 this time I will push :) May 06 08:17:46 khem: ok, its just an ordering problem May 06 08:18:38 03Khem Raj  07org.openembedded.dev * r6be330e09a 10openembedded.git/recipes/libtool/libtool-native_2.2.6b.bb: May 06 08:18:38 libtool-native_2.2.6b.bb: Install remaining missing directories. May 06 08:18:38 Signed-off-by: Khem Raj May 06 08:19:55 03Khem Raj  07org.openembedded.dev * r2efe9d2d01 10openembedded.git/recipes/libiconv/ (libiconv.inc libiconv_1.11.bb libiconv_1.12.bb): May 06 08:19:55 libiconv: inherit gettext May 06 08:19:55 Signed-off-by: Khem Raj May 06 08:23:40 RP: inheriting gettext does not fix the problem for libiconv May 06 08:30:32 strange, same user, libltdl dir has 755, cp: cannot create regular file `/OE/tmpdir-dev-shr/sysroots/x86_64-linux/usr/share/libtool/libltdl/Makefile.am': Permission denied May 06 08:31:30 but build finish ok.. just the error is strange, maybe it's because I'm looking at that dir after finish May 06 08:31:46 JaMa: do a pull May 06 08:31:49 its fixed May 06 08:32:00 pulled 5sec ago May 06 08:32:32 2efe9d2d019f9c0ef9587ebbc21ff22f5de12f58 May 06 08:36:49 03Koen Kooi  07org.openembedded.dev * r71be81bcc9 10openembedded.git/recipes/stage-manager/stagemanager-native_0.0.1.bb: stagemanager-native: convert to new style staging May 06 08:44:49 khem: is gettext being added to the dependencies? May 06 08:45:37 now bitbake stops without complaining May 06 08:45:45 it thinks it has built the target May 06 08:45:46 khem: Looking at the code in that recipe, its plain to see why :/ May 06 08:46:04 DEPENDS is overridden May 06 08:46:27 I fixed it locally lets see May 06 08:46:31 khem: Trouble is gettext is needed since those recipes use gettext autoconf recipes May 06 08:46:45 yeah May 06 08:46:48 so the gettext.m4 file needs to be present to reautoconf May 06 08:48:15 now koen broke stage-manager May 06 08:53:34 khem: broke how? May 06 08:53:50 I dont know its giving me build errors May 06 08:54:09 "UnboundLocalError: local variable 'bb' referenced before assignment" when use bitbake to compile .bb file May 06 08:54:10 | NOTE: make -e MAKEFLAGS= install May 06 08:54:11 | make: *** No rule to make target `install'. Stop. May 06 08:54:20 khem: let me merge that from poky May 06 08:56:11 03Richard Purdie  07org.openembedded.dev * r6bd8937350 10openembedded.git/recipes/stage-manager/stagemanager-native_0.0.1.bb: May 06 08:56:11 stagemanager-native: Sync with Poky to fix builds May 06 08:56:11 Signed-off-by: Richard Purdie May 06 08:58:18 I'm having a go at PySide, starting with apiextractor. I get "Qt qmake not found". I have tried inheriting qmake and depending on qmake-native. Any tips? May 06 08:58:50 khem: try that... May 06 08:58:55 tryig May 06 08:59:17 RP: that works May 06 08:59:51 alright time to sleep May 06 09:02:16 hm, did a fresh pull of head, now getting: UnboundLocalError: local variable 'bb' referenced before assignment May 06 09:02:39 anyone any idea? I already made sure that I installed a clean bitbake 1.8.18 May 06 09:03:10 eFfeM_work: Which distro/machine? May 06 09:03:28 eFfeM_work: There have been a few reports but I don't know why :/ May 06 09:03:30 RP, angstrom/calamari (== powerpc) May 06 09:03:31 and why not bitbake 1.10? May 06 09:04:07 RP: strange error I get when I change DEPENDS = to DEPENDS += in libiconv_1.12.bb bitbake parsed all recipes and then exited happily May 06 09:04:08 because I still had 1.8.18 in my dir (that was the one I was using before I went on vacation; today just returned @ work) May 06 09:04:35 RP: I am usin git master of bitbake see if you can reproduce this error May 06 09:05:31 khem: What should it have done? Tried to build something? May 06 09:05:32 full error @ http://www.pastie.org/948093 May 06 09:07:28 eFfeM_work: ah, I think I understand May 06 09:07:48 RP, great May 06 09:07:58 RP: yes May 06 09:08:06 JaMa: berlios does not yet have 1.10 for d/l May 06 09:08:09 it should have gone on to build console-image May 06 09:08:19 or any target I specify May 06 09:08:24 it just bailed out May 06 09:08:38 unifdef is borked too May 06 09:08:40 man May 06 09:09:20 03Richard Purdie  07org.openembedded.dev * r779458cafb 10openembedded.git/classes/sanity.bbclass: May 06 09:09:20 sanity.bbclass: Fix to work with old versions of bitbake for now May 06 09:09:20 Signed-off-by: Richard Purdie May 06 09:09:20 eFfeM_work: try pulling May 06 09:09:25 RP: it there something missing in 1.10 bitbake branch to pack it as 1.10.0? It works great here May 06 09:09:34 RP: I think unifdef-native needs porting from poky too May 06 09:09:42 dth: this probably fixes your problem too May 06 09:10:02 khem: yes May 06 09:11:46 khem: fix pushed May 06 09:11:56 03Richard Purdie  07org.openembedded.dev * r5469b3efd2 10openembedded.git/recipes/unifdef/unifdef-native_2.6.18+git.bb: May 06 09:11:56 unifdef-naitve: Fix from Poky May 06 09:11:56 Signed-off-by: Richard Purdie May 06 09:12:22 and I found a bug in poky in unifdef :) May 06 09:12:41 RP pulled, also had to install chrpath, now bb is parsing so guess it is ok May 06 09:12:42 I fixed it May 06 09:13:59 RP, after pull and installing CHPATH it seems to run, thank you for yor service! May 06 09:15:37 03Khem Raj  07org.openembedded.dev * r9966b81a7a 10openembedded.git/recipes/unifdef/unifdef-native_2.6.18+git.bb: May 06 09:15:37 unifdef-native_2.6.18+git.bb: Fix new style staging. May 06 09:15:37 Signed-off-by: Khem Raj May 06 09:16:27 good morning (ugt) May 06 09:18:40 khem: We both pushed fixes :/ May 06 09:19:07 RP: I just merged your into mine May 06 09:20:11 RP: us the directory install is required if we have NATIVE_INSTALL_WORKS = "1" May 06 09:20:15 s/us/is May 06 09:23:54 khem: ah, right, yes :) May 06 09:24:54 RP: check the bitbake error if you can reproduce it May 06 09:25:02 actually I can see it all the time May 06 09:25:22 khem: Whats the step to reproduce? May 06 09:25:23 even when I comment out the DEPENDS line in libiconv_1.12.bb May 06 09:25:43 on latest master of both oe and bitbake May 06 09:26:01 comment out DEPENDS in libiconv_1.12.bb May 06 09:26:08 then do bitbake libiconv May 06 09:26:33 hi, I'm currently building for a x86 machine, using bitbake 1.8.18 / angstrom / linux 2.6.33 / custom image based on base-image + mplayer grub compat-wireless : are there specifics things to test regarding this night's merges ? May 06 09:26:39 hopefully someother recipe is not providing libiconv for you May 06 09:30:21 RP things are building for me too, tnx, one small issue left: python-native do_populate_sysroot (at least I think it is that one) gives several pages of messages about new RPATH May 06 09:30:43 eFfeM_work: some new code is a little too verbose. Don't worry about it May 06 09:31:39 RP, i don't worry thought it was maybe unknown, so wanted to report it May 06 09:32:43 btw what happened to the TSC? still didn't see any reply/decision on the glib proposal (actually didn't see anything from TSC on the mailing list in the last few weeks) May 06 09:34:08 eFfeM_work: there is a meeting today. May 06 09:34:53 khem: it just builds libiconv here May 06 09:36:15 RP, ah ok, is membership changed? I recall something about an election in april, but didn't see any back on that either May 06 09:37:25 eFfeM_work: It was all discussed on the list... May 06 09:37:40 eFfeM_work: whats the subject line of the glib proposal? May 06 09:38:28 eFfeM_work: the removal of old versions? May 06 09:44:13 RP, yes removal of the old versions May 06 09:44:39 RP and search did not give me an article on the new tsc; guess I overlooked it May 06 09:44:56 eFfeM_work: are you a member of the e.V.? May 06 09:45:01 nope May 06 09:45:02 it was on the e.V. members list May 06 09:45:33 was told i could get on after the election (someone suggested a membership drive) May 06 09:46:13 eFfeM_work: Yes, that sounds like a good idea May 06 09:46:14 e.V members list is not publicly accessible I think, as TSC is for oe, i had expected a note on the oe ml May 06 09:46:19 XorA: ping ;-) May 06 09:46:50 eFfeM_work: Yes, we should pass that message on... May 06 09:48:24 hm dev head fails to compile for ppc: May 06 09:48:29 /home/frans/workspace/GTD100/C/Linux/openembedded/tmp/cross/ppce500v2/lib/gcc/powerpc-angstrom-linux-gnuspe/4.3.3/include-fixed/limits.h:122:61: error: limits.h: No such file or directory May 06 09:48:43 RP: here May 06 09:48:44 in configure May 06 09:48:59 afk for 1.5 hr or so May 06 09:51:03 I met a problem building utils-linux-ng-native : it seems to be gone after adding gettext-native to DEPENDS_virtclass-native in util-linux-ng.inc May 06 09:51:44 XorA: We were just talking about membership elections for the e.V. May 06 09:52:07 XorA: We have plans to run some next, right? May 06 09:52:46 RP: Im away next week, but I plan to start a membership drive after I get back May 06 09:53:15 RP: in meantime anyone can start a vote May 06 09:54:13 XorA: Thats fine, I'm not putting pressure on, just checking it was planned. I' hadn't remembered, you obviously have so we're fine :) May 06 09:54:26 so if there is a group of people wanting in, someone who is a member collect the names and send out a Vote Proposal, Ill handle the rest May 06 10:02:10 03Steffen Sledz  07org.openembedded.dev * r31af13b2a1 10openembedded.git/recipes/linux-libc-headers/ (2 files in 2 dirs): May 06 10:02:10 linux-libc-headers: recipe for 2.6.24 added May 06 10:02:10 Signed-off-by: Steffen Sledz May 06 10:20:33 XorA: I will ask people to come forward and collate the names May 06 10:27:13 here is the error met on utils-linux-ng-native : http://pastebin.org/204521 May 06 10:27:44 I running with BB_NUMBER_THREADS = "4" and it seems gettext-native is not yet built May 06 10:34:36 RP: hi, we have our TSC meeting tonight? May 06 10:34:52 zecke: yes. Did you get the email with the room details in? May 06 10:35:12 ericben: is gettext-native being added to the DEPENDS? May 06 10:35:20 RP: I think so. May 06 10:35:39 no May 06 10:35:44 zecke: ok, great May 06 10:36:36 shouldn't that be in DEPENDS_virtclass-native ? May 06 10:36:45 RP: shouldn't that be in DEPENDS_virtclass-native ? May 06 10:36:49 ericben: The inherit gettext should add it May 06 10:37:14 RP: then there is a strange thing May 06 10:38:11 RP : I've cleaned tmp and starting gain, I'll check the dependencies and let you know May 06 10:38:40 ericben: the way to check would be "bitbake util-linux-ng-native -e | grep ^DEPENDS May 06 10:38:52 and looking locally, its not listed :/ May 06 10:39:27 RP: OK, that's why adding gettext-native to DEPENDS_virtclass-native in util-linux-ng.inc fixed the thing for my second try May 06 10:44:00 ericben: its the virtclass-native override thats doing it May 06 10:44:06 I guess we need to be explicit there May 06 10:44:57 RP: DEPENDS_virtclass-native="autoconf-native automake-native libtool-native gnu-config-native coreutils-native zlib-native ncurses-native lzo-native" May 06 10:45:01 RP: DEPENDS_virtclass-nativesdk="autoconf-native automake-native libtool-native gnu-config-native coreutils-native" May 06 10:45:19 RP: DEPENDS_GETTEXT="gettext-native" May 06 10:46:06 03Richard Purdie  07org.openembedded.dev * rf17f015c5e 10openembedded.git/recipes/util-linux-ng/util-linux-ng.inc: May 06 10:46:06 util-linux-ng: Add gettext-native dependency explicitly due to override usage May 06 10:46:06 Signed-off-by: Richard Purdie May 06 10:46:13 ericben: try that :) May 06 10:46:23 RP: OK I understand the overide. Ipull and relaunch the build, I'll let you know May 06 10:48:07 hi eric May 06 10:48:56 as u answered about QWS_MOUSE_PROTO environment parameter, there was no such device /dev/input/touchscreen0 when i did a simple ls /dev May 06 10:49:01 am i missing something here May 06 10:49:48 is there anything extra i need to install.... May 06 10:50:51 03Koen Kooi  07org.openembedded.dev * r69f21ab99b 10openembedded.git/recipes/unifdef/unifdef-native_2.6.18+git.bb: unifdef-native: one NATIVE_INSTALL_WORKS is enough May 06 10:53:03 kindly tell May 06 10:58:42 hi all, i bitbaked my-own-cutomized-image recently in which i have just the necessary libraries to run a qt embedded program. The app is running fine on qemuarm machine apart from the fact that it is not behaving like a touch screen. The moyse pointer is not moving. I was suggested to set QWS_MOUSE_PROTO environment parameter to /dev/input/touchscreen0 but i do not have such a device node in my image.. May 06 10:59:59 and even when i set the variable the problem still persists. Is this has anything to do with the fact that the image is derived from base-image.bb and only is a bare minimum May 06 11:00:02 ?? May 06 11:00:09 kindly help May 06 11:02:02 03Richard Purdie  07org.openembedded.dev * r8cf219c597 10openembedded.git/classes/native.bbclass: May 06 11:02:03 native.bbclass: Clear PACKAGES variable when using BBCLASSEXTEND (from poky) May 06 11:02:03 Signed-off-by: Richard Purdie May 06 11:02:23 Spyzer: I think this means nobody has an answer for you offhand I'm afraid May 06 11:13:27 RP: any extra time-slice to look at the deprecated staging of klibc? May 06 11:16:02 :( May 06 11:18:01 ant_work: Didn't I give you some examples the last time we discussed this? May 06 11:18:18 I somehow failed to implement those :/ May 06 11:18:40 I ended up with missing klcc... May 06 11:19:06 ant_work: Firstly, why can't the do_install just do a oe_runmake install? May 06 11:20:12 do we know STAGING_DIR_TARGET ? May 06 11:20:29 ant_work: do_install installs into ${D} May 06 11:21:11 that's the point, last time klcc landed somewhere but not where expected May 06 11:21:20 ant_work: Thats fine, we can move it May 06 11:21:38 "oe_runmake install", then tweak as needed May 06 11:21:57 Whats currently in there looks nasty May 06 11:28:33 RP: is not trivial, that's why I'm bothering you again... May 06 11:29:14 ant_work: trouble is I have no experience with klibc May 06 11:29:41 klibc has a good momentum right now upstream May 06 11:30:14 all just works (more or less) May 06 11:31:27 the only issue is they broke compatibility with pre 2.6.28 kernels after klibc 1.5.15 May 06 11:34:30 i just saw the recipe namely qpe-inputhelper which helps opie and qtopia for USB devices, can this help me to enable touchscreen(wacom-tablet) in my image?? May 06 11:38:27 Spyzer: if you lack /dev/input/touchscreen0 something is wrong with your device manager (or the kernel driver is not compiled/loaded) May 06 11:39:12 anyone an idea where limits.h should come from? trying to build eglibc with gcc 4.3.3 but configure says it is missing May 06 11:39:27 well i also bitbaked a opie-image but there the mouse works fine and kernel is same May 06 11:39:45 found some in sysroots, but guess they are not in the path May 06 11:39:50 is the issue distro or GUI related May 06 11:39:53 ?? May 06 11:39:55 probably using /dev/input/eventx May 06 11:40:30 the problem is there is no such directory as input in /dev although i do have /dev/eventx should i try those instead May 06 11:40:33 ?? May 06 11:40:43 yep May 06 11:40:53 check dmesg May 06 11:41:20 eFfeM_work: known issue with certain libc and gcc versions interacting badly May 06 11:41:47 RP, feared that, used to work :-( May 06 11:42:09 is there a known workaround ? May 06 11:42:35 eFfeM_work: some gcc recipes have or had hacks May 06 11:42:47 eFfeM_work: check the scm history? May 06 11:42:59 RP, ok will dig into it May 06 11:43:49 may u please tell me how can i utilize the information from dmesg May 06 11:44:14 it is showing something about the wacom-tablet i passed as a parameter to qemu May 06 11:44:15 it should list the devices using event interface May 06 11:44:33 anyway, look at this similar thing I found http://www.groomlakelabs.com/?q=linux_input_usb_touchscreen_howto May 06 11:45:51 well no, instead it is saying input: Wacom PenPartner as /devices/pci0000:00/0000:00:0c.0/usb1/1-3/1-3:1.0/input/input1 May 06 11:46:03 but i don't even have a /dev/input directory May 06 11:46:19 and dmesg is listing something like an input directory May 06 11:46:46 hi all, i'm trying to build a qt widget with bitbake, but it fails with: [...].../arm-angstrom-linux-gnueabi/bin/ld: cannot find -lQtOpenGLE May 06 11:47:10 Spyzer: OPIE uses udev to populate the device tree May 06 11:47:55 I mean opie-image May 06 11:48:19 ok the url u posted, helped me tell that event1 is the handler May 06 11:48:35 lets see if setting the QWS_MOUSE_PROTO variable will now work May 06 11:48:49 btw can i use udev too May 06 11:50:26 what is lQtOpenGLE good for? May 06 11:51:49 thanks a lot it started working May 06 11:51:51 yepeee May 06 11:51:56 :) May 06 11:55:10 you see, I had one issue where Symlink /dev/input/touchscreen0 was pointing to event1 instead of 2 May 06 11:55:14 :D May 06 12:01:14 good morning May 06 12:02:37 03Steffen Sledz  07org.openembedded.dev * r9da5a071dc 10openembedded.git/recipes/linux-libc-headers/linux-libc-headers_2.6.24.bb: May 06 12:02:38 linux-libc-headers_2.6.24: legacy staging removed May 06 12:02:38 Signed-off-by: Steffen Sledz May 06 12:03:17 in the makefile, I commented "-lQtOpenGLE" out and now it compiles. Does someone know what this flag is exactly good for? May 06 12:08:19 bleikert: -l is for linking libraries May 06 12:09:14 bleikert: in your case it wants to link to the QtOpenGL module for Embedded (E) but it does not exist... I think it is only built in some circumstances (like beagleboard with binary blob GL impl). May 06 12:11:18 zecke: thanks, i know what -l is good for ;) however, i was a little bit confused why QtOpenGLE was pulled in May 06 12:11:47 zecke: i tracked it down, and it was in the Makefile from the beginning, wrong configuration in QtCreator i think ;) May 06 12:13:00 zecke: i wrote a recipe and bitbake pulls the source from git and alters the Makefile, i first thought the flag was set by oe :) May 06 12:13:10 zecke: anyway, thanks for your help! May 06 12:17:41 RP: now glibc_2.9 fails at do_package May 06 12:18:43 RP : error is : http://pastebin.org/204626 May 06 12:28:12 I found an old post about PyQt. http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-August/002803.html. Was this indeed the end-of-life for PyQt? May 06 12:43:21 khem: I have 2 issues with http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=4134f55dc7d99516d750b24a60e5cec400ae5ef3 (eglibc: move away from legacy staging) May 06 12:43:32 first is, shouldn't there be a PR bump in this May 06 12:43:46 2nd and more serious is that it does not seem to stage limits.h any more May 06 12:44:14 (this used to be done in eglibc-initial) May 06 12:48:23 ericben: does bitbake pax-utils-native resolve it? May 06 12:48:31 RP: no already tried it May 06 12:48:44 RP: in fact scanelf is not installed anywhere under tmp/ May 06 12:49:40 RP: it's only found in tmp/work/i686-linux/pax-utils-native-0.1.19-r0/ May 06 12:50:07 RP: but if I understand well should'nt I find it into sysroot or something like this ? May 06 12:50:25 ericben: you should, yes May 06 12:51:01 RP: OK, do you want any log which could help ? May 06 12:52:56 ericben: Its ok, I can see the problem. Its broken in both OE and Poky May 06 12:52:58 :/ May 06 12:52:59 RP: also work/i686-linux/pax-utils-native-0.1.19-r0/sysroots is empty May 06 12:53:06 RP: OK May 06 12:54:20 ericben: that is the problem May 06 13:01:07 ericben: Try the commit I just pushed May 06 13:01:12 03Richard Purdie  07org.openembedded.dev * r58ddfb17c7 10openembedded.git/recipes/pax-utils/ (pax-utils_0.1.13.bb pax-utils_0.1.19.bb): May 06 13:01:12 pax-utils-native: Make sure prefix variable is correct during do_install May 06 13:01:12 Signed-off-by: Richard Purdie May 06 13:04:33 03Stefan Schmidt  07org.openembedded.dev * r6324db5d8b 10openembedded.git/recipes/ibrdtn/ (5 files): ibrdtn: Add development versions with DP = -1 May 06 13:04:34 03Stefan Schmidt  07org.openembedded.dev * reba935085e 10openembedded.git/recipes/ibrdtn/ibrdtn_0.1.2.bb: May 06 13:04:34 ibrdtn: Add latest stable version. May 06 13:04:34 This version does no longer need the gcc4.4 patch as it was applied upstream. May 06 13:08:26 RP: is there a pw on our TSC chat room? May 06 13:08:47 RP: problem fixed, thanks May 06 13:09:35 zecke: no, your jabber account should be allowed in May 06 13:09:39 RP: please ignore. :) May 06 13:09:48 RP: it opened the chat window on a different desktop May 06 13:20:30 kergoth: ping? May 06 13:21:16 anyone knows if GNUmake has a rule that is executed at exit? May 06 14:03:00 I still have a problem with linux's sysroots : .config 's link to config-2.6.33 is missing May 06 14:04:22 kernel-config link exists which is strange as both are created in kernel.bbclass May 06 14:54:38 03Koen Kooi  07org.openembedded.dev * r8025a2ace5 10openembedded.git/recipes/gstreamer/gst-plugins.inc: gst-plugins: build some more plugins (e.g. mms, aacparse, faad) May 06 14:57:51 pb_: You've remembered the TSC meeting? May 06 14:58:05 mickey|sports, mickey|sports: TSC meeting? May 06 14:58:12 mickey|office: ^^^ May 06 14:58:58 kergoth: you too ^^^ :) May 06 14:59:26 RP: yes, I think I am already in the room May 06 14:59:40 at least, it says "connected" May 06 14:59:46 pb_: we cant see you May 06 14:59:51 oh, hm May 06 15:00:05 let me try again, maybe empathy is in some bad state May 06 15:00:32 pb_: 4 people in room now May 06 15:01:23 empathy xmpp implenetation is a bit ropey to be honest May 06 15:01:34 pb_: oh, are you happy with empathy? Where you in need for an OTR plugin? May 06 15:01:39 in fact XMPP is a huge embarressment for FOSS software :-) May 06 15:02:01 zecke: this is the first time I have tried to use it May 06 15:02:11 and, based on this experience, I guess I am not all that happy :-} May 06 15:02:13 * RP gave up on empathy May 06 15:02:19 RP: what are you using? May 06 15:02:26 pb_: pidgin May 06 15:02:43 okay, I'll try that one May 06 15:02:52 or, could we just have the meeting in irc instead? we are all here, right? May 06 15:03:30 First item on the agenda is to formalise a meeting location for the new tsc May 06 15:03:48 We want to hold it here? May 06 15:04:44 well, maybe not "here" as in this very channel, but it seems like irc would be a bit less of a nuisance than xmpp if we are all using it already. still, if I am the only one who lacks a working xmpp client then I guess I should just sort myslf out :-} May 06 15:04:44 Lets use a different channel, I'd prefer not to flood here with our discussion May 06 15:04:58 pidgin installed now, just need to set it up May 06 15:05:11 We're missing mickey|office mickey|sports :/ May 06 15:06:30 aww, already 17 May 06 15:06:37 i'll be right there May 06 15:07:40 * kergoth doesn't really have an opinion on the location, as long as it happens and isn't horribly difficult May 06 15:07:58 hm, still no luck with this. I do at least get an error now, it says "407 Registration required". May 06 15:08:16 maybe I am doing it wrong May 06 15:08:24 pb_: Is it connection to jabber at all? May 06 15:08:43 RP: I can see you in my buddy list, so I guess so May 06 15:09:34 darn May 06 15:09:41 i have lost the room address AGAIN May 06 15:09:51 no fricking backup here May 06 15:09:56 tsc-meeting@conference.jabber.xora.org.uk May 06 15:09:58 right May 06 15:10:00 thanks May 06 15:11:19 pb_: try again, your added without typo now May 06 15:13:23 Hmm May 06 15:13:33 ppc stuff seems broken somewhere, digging more to see what/where May 06 15:14:44 Tartarus: hmm :/ May 06 15:15:12 eglibc May 06 15:15:21 * Tartarus waits for do_compile to fail again May 06 15:15:36 Tartarus: There are rumours that Khem's commits this morning weren't 100% successful :/ May 06 15:16:00 RP: youve been listening to election stuff too much today :-) May 06 15:16:16 RP: Did a pull right after your merge and started my loop May 06 15:16:57 Tartarus: odd then as that would be before that commit. We did push some fixes this morning... May 06 15:17:01 XorA: heh :) May 06 15:22:43 New problem - indirectly-related to the device_table commit from the other day, but really an image generation issue: May 06 15:23:08 RP: gnutls blew up here May 06 15:24:08 do_rootfs is stuck; opkg-cl is (according to the log file) prompting for permission to overwrite a file that's already in the image (the device_table). Am I missing a flag in the recipe that can make opkg NOT prompt (it'll never get an answer, so it certainly shouldn't prompt, perhaps fail instead...) May 06 15:26:28 (can't test RP's new branch merge until I can get a baseline image built. :( ) May 06 15:26:55 I can request that we "un-commit" the device_table change, i suppose. May 06 15:33:19 mwester: old opkg? iirc the fix for hanging prompts should be in it, as far as i know May 06 15:33:20 hmm May 06 15:43:33 * kergoth pokes pb_ May 06 15:48:59 kergoth: mm? May 06 15:49:19 nothing now, was when you disconnected from jabber :) May 06 16:09:23 ah, heh May 06 16:10:03 now its kergoth's turn May 06 16:13:13 RP: lzma for arm is almost upstream (in Russel's tree). We need to fix a couple of thing to accomodate lzma kernels/initramfs May 06 16:13:43 ant_work: cool. btw, we finally fixed the packaged-staging architecture fields in poky May 06 16:13:56 I'll fire a build later on May 06 16:14:44 great work, thx May 06 16:18:43 hi, I did: May 06 16:19:29 bitbake boost with preferred version 1.41 May 06 16:19:36 then I did bitbake -c clean boost May 06 16:19:44 then bitbake boost with preferred version 1.40 May 06 16:19:49 and it didn't build the ipks May 06 16:19:57 of the 1.40 version May 06 16:20:25 is there a way to make it build the ipks...I'ld like to test gnash 0.8.7 on my eee701 before commiting May 06 16:20:31 s/commiting/pushing May 06 16:21:03 Tartarus: Sorry, tsc meeting distracted me - blew up how? May 06 16:25:23 any idea why I get this parsing errors http://pastebin.org/204884 ? May 06 16:32:39 i'm trying to build using an external toolchain May 06 16:33:08 but its failing looking for a nonexistant directory: http://pastebin.org/204893 May 06 16:33:17 followed the intrustructions in the manual May 06 16:43:42 bbl May 06 16:51:22 hey, I'm trying to generate an image from angstrom-distribution.org's narcissus page.. but the opie packages fail May 06 16:53:04 there any way to get these packages to work? May 06 17:18:11 03Koen Kooi  07org.openembedded.dev * re80334081b 10openembedded.git/recipes/angstrom/angstrom-uboot-scripts/ (dsplink-512MB.cmd dsplink-bootchart.cmd sh.cmd): angstrom-uboot-scripts: add some more May 06 17:26:59 Hi, all May 06 17:27:05 hi May 06 17:27:20 is there a person who maintains wxWidgets package? May 06 17:27:52 I don't know if the maintainer file is maintained... May 06 17:28:06 I am trying to build my program by using "g++ ..." and getting a lot of errors May 06 17:28:32 or maybe I need some additional variable exported May 06 17:28:41 ? May 06 17:28:54 but there is a MAINTAINER file in the root dir of oeo May 06 17:29:09 it's actually "arm-angstrom-linux-gnueabi-g++" May 06 17:29:10 IgorK, then you should ask here or on the mailing list May 06 17:29:59 is there a special procedure to compile the program with the OE toolchain? May 06 17:30:28 cause I can build it with the x86 one without any errors... May 06 17:31:05 hej , do I have to prepare for some work after the "big" merge? or can I start a fresh build and watch tv ? ;) May 06 17:37:21 any idea why I get this parsing errors http://pastebin.org/204884 ? May 06 17:38:32 Here is my command and error messages: http://pastebin.com/GBu0K0RB May 06 17:38:35 soltys: Information not available for target 'i386-freebsd' May 06 17:38:52 soltys: I dont know if someone here used freebsd build boxes May 06 17:39:16 ahm.. May 06 17:39:37 IgorK: error: wx/wx.h: No such file or directory May 06 17:39:59 you need that dependency installed May 06 17:42:24 khem: it is installed and I point to it by "-I" May 06 17:43:42 and why do I get all those "termios.h" errors? May 06 17:46:19 IgorK: the above error says that your inc paths are not complete May 06 17:47:01 Tartarus: how did the build go ? May 06 17:47:07 IgorK: do you call this command on your own , or does bitbake generate it ? May 06 17:47:12 uclibc builds are showing interesting problems May 06 17:47:26 now I am doing a eglibc build myself to see if something is broke May 06 17:47:34 as Frans reported May 06 17:48:09 mlip: I think its not bitbake May 06 17:48:47 it definitelly points in the wrong include dirs ? May 06 17:49:36 I didn't have time to investigate the new tmp structure but, -I should point in sysroots or staging May 06 17:50:00 -I/home/arm/OE/temp/work/armv4t-angstrom-linux-gnueabi/wxwidgets-2.8.9-r7/image/usr/include/wx-2.8/wx May 06 17:50:11 is definitely wrong May 06 17:52:00 khem: uclibc+gcc-4.5 or even with older? May 06 17:52:13 mlip: it is not bitbake command May 06 17:52:45 khem: and you can skip eglibc+armv4 test build as it went fine here as reported May 06 17:53:41 IgorK: Depending on when OE has been updated the last time you should point the -I directories somewhere in tmp/staging/... (old) or tmp/sysroots/... May 06 17:54:33 JaMa: cool May 06 17:54:41 JaMa: uclibc+gcc-4.5 May 06 17:54:51 problem is in libiconv actually May 06 17:54:55 which is needed for uclibc May 06 17:55:02 but not for eglibc/glibc May 06 17:55:09 I think uclibc itself has no issues May 06 17:55:58 khem: and no other issues with gcc-4.5? like mine busybox segfaulting all over the image and oops in kernel? May 06 17:56:08 mlip: where do point it to: http://pastebin.com/h7TLAXnu May 06 17:56:26 JaMa: no :) May 06 17:56:31 JaMa: not on qemu May 06 17:56:34 lucky you :) May 06 17:56:41 JaMa: I havent tried osk yet May 06 17:57:16 JaMa: you told me its on both armv4 and armv5 for you right ? May 06 17:57:17 IgorK: I would use ./temp/staging/armv4t-angstrom-linux-gnueabi/usr/include/wx-2.8/wx/wx.h May 06 17:57:43 IgorK: But you _should_ use the OE toolchain, this really makes this kind of things A LOT easier May 06 17:57:57 khem: yes May 06 17:58:25 mlip: IgorK you should build a recipe for your package and then it will generate the run.do_compile script for you May 06 17:58:28 once you bitbake it May 06 17:58:43 i know ;) May 06 17:58:46 khem: for armv5t it was also built with newer glibc than previous image, but symptoms are the same, so I guess the cause is in gcc in both cases May 06 17:58:46 then you could use that script to compile again and again if you have to May 06 17:59:06 JaMa: yes could be that some optimization is borked May 06 17:59:21 JaMa: where is the stack trace of kernel May 06 17:59:27 JaMa: let me have a look May 06 17:59:53 khem: I have 2 identical minimal images images built with different gcc if you want May 06 18:00:01 mmt looking for trace May 06 18:01:05 http://paste.pocoo.org/show/209449/ May 06 18:02:11 * JaMa off for a while.. May 06 18:11:58 khem: I know, I tried, but "bitbake my_program" didn't produce anything... May 06 18:12:32 mlip: Will try it in a coujple of minutes, thx May 06 18:14:22 IgorK: If you have to compile things more than once (and ppl who do build OE usually have to), you should try to investigate the mistakes you did when creating a recipe May 06 18:14:35 and fix these instead of trying to build a program on your own May 06 18:17:23 * khem wishes to have 'dwim' switch on the keyboard May 06 18:18:20 ;) May 06 18:55:37 hi, is patchwork working fine ? I sent a patch and it didn't appear in patchwork. May 06 18:58:34 mlip: it looks like I am compiling using the bitbake recipe. Now I need to fix some stupid errors going from newer release on x86 to older one on arm-angstrom ;-) May 06 18:59:09 thank you 4 sticking with me and helping even if not helping at all ;-) May 06 20:08:31 * mwester kicks opkg in the **** May 06 20:08:40 * kergoth helps May 06 20:48:01 hi May 06 20:54:21 hi woglinde May 06 20:54:28 hi rp May 06 20:54:32 thanks for the merge May 06 20:54:50 woglinde: np :) May 06 20:54:53 RP: for your delight I just finished a qemuarm build minimal/console-image May 06 20:54:59 woglinde: Nice to finally have that sorted :) May 06 20:55:03 khem: great! :) May 06 20:55:18 RP: were you able to reproduce the issue with libiconv ? May 06 20:55:24 khem: no :/ May 06 20:55:31 hm which issuse? May 06 20:55:32 minimal-image for igep board at 719/1133 May 06 20:55:39 I will try it now May 06 20:55:51 khem: it doesn't make any sense. I can only assume some error was causing bitbake to exit but the source of the error was not obvious May 06 20:56:05 perhaps -DDDv would make it more clear? May 06 20:56:13 RP: yeah I have seen issues when you have a package in ASSUME_PROVIDED May 06 20:56:20 and then you try to bitbake the same May 06 20:56:25 bitbake gets confused May 06 20:56:58 I tried -DDD and it exited gracefully showing me no sign of errors why it did so May 06 20:57:17 khem: ASSUME_PROVIDED is a known issue May 06 20:57:25 * RP has that one on his bug list May 06 20:57:28 RP: btw the build was using gcc 4.5.0 and eglibc with new staging changes May 06 20:58:11 khem: cool - how is 4.5.0 looking? May 06 20:58:18 * RP needs to update the gcc in poky... May 06 20:59:23 RP: mixed May 06 20:59:34 JaMa reported crashes in kernel May 06 20:59:40 for armv4 and v5 May 06 20:59:48 I have no issues with it on qemu May 06 21:00:34 RP: the newly built stuff boots fine on qemu arm so double delight May 06 21:02:44 n8, thx for the merge ! May 06 21:03:44 khem: an added bonus :) May 06 21:09:51 RP: now you can sleep with happiness May 06 21:11:09 lol May 06 21:12:24 khem: Its nice to have something work out well :) May 06 21:12:48 and incandescant did help with that May 06 21:13:45 yeah heh May 06 21:13:54 I will know more on my other box which is slow May 06 21:14:01 is running a build for osk May 06 21:14:48 woglinde: the only german co-worked I knew is quitting so I wont have anyone to practice my german skills May 06 21:15:08 khem why he is quitting? May 06 21:15:24 he is not happy and google is paying him more May 06 21:15:26 :) May 06 21:15:31 khem hms May 06 21:15:33 thats why May 06 21:15:49 hm I am still fighting with the ti stuff May 06 21:15:55 no luck yet May 06 21:15:58 he is from Nurnburg May 06 21:16:01 still linker error May 06 21:16:09 woglinde: hmm really thats bad May 06 21:16:16 dont know if I should try binutils 2.18 May 06 21:16:44 you can try that but I am not hopeful if it will work May 06 21:17:09 hm why? May 06 21:17:32 not much has changed in this area May 06 21:18:16 but how did it worked before? May 06 21:18:57 I dont think I am the only one hitting this problem May 06 21:20:01 ok try it with 2.18? May 06 21:20:56 hm..here console-image failed, but opie-image and x11-image not May 06 21:21:04 nothing in the logs.. May 06 21:21:17 ant strange May 06 21:21:18 well.. smthg May 06 21:21:20 cp: cannot stat `/oe/build/tmp/work/c7x0-angstrom-linux-gnueabi/angstrom-gpe-task-base-1.0-r40/image/*': No such file or directory May 06 21:21:28 cp: cannot stat `/oe/build/tmp/work/c7x0-angstrom-linux-gnueabi/angstrom-x11-base-depends-1.0-r43/image/*': No such file or directory May 06 21:21:37 I lost head of logs :/ May 06 21:21:51 ant thats comes from bad do_install May 06 21:22:33 hm..I have 2 run.base_do_fetch.. May 06 21:22:58 this is probably a commonly asked question. i can't find it on the wiki. when submitting a patch to the ml, do I attach the patch or paste in the message? May 06 21:23:02 for every package May 06 21:23:17 kgilmer: git send-email is best May 06 21:23:27 wow... May 06 21:23:33 thanks khem May 06 21:23:48 i will look into that May 06 21:24:07 kgilmer: otherwise you can follow the directions what kernel devs do May 06 21:24:41 basically patch should be visible so when I download the email as mbox and do patch then it should apply May 06 21:24:51 ok khem thanks May 06 21:24:53 kgilmer I attache the patch May 06 21:25:05 woglinde: depends on MUA May 06 21:25:09 you use May 06 21:25:14 mut May 06 21:25:15 some attach them inline some dont May 06 21:25:16 mutt May 06 21:25:20 * kgilmer does not trust his webmail. May 06 21:25:22 yeah mutt is ok May 06 21:25:38 but if you have those funky MS outlook express stuff it can get painful May 06 21:26:33 kgilmer: Documentation/email-clients.txt file in lnx srcs May 06 21:26:40 explains how to do it May 06 21:26:45 for different mailers May 06 21:26:49 ah nice. thx khem. May 06 21:27:10 but with git based projects git send-email is the best option May 06 21:31:28 exit May 06 21:31:33 :) May 06 21:32:28 nite rp May 06 21:33:04 hm I wonder why console-image needs 4550 taks May 06 21:33:06 woglinde: not quite yet, wrong terminal :) May 06 21:33:33 hm May 06 21:33:35 nite rpOTE: Gettext required but not in DEPENDS for file /devel/arm/oetmp-ang/work/i686-linux/coreutils-native-7.2-r1/coreutils-7.2/configure.ac. May 06 21:33:38 Missing inherit gettext? May 06 21:33:39 args May 06 21:34:48 hm quilt-native needs some tweaking too May 06 21:35:51 hm should converted to BBCLASSEXTENDED too May 06 21:37:40 woglinde: should be recipes in Poky for some of them May 06 21:37:44 ah, btw we should update the oe & yourdistro: at least gentoo requires chrpath on host May 06 21:37:56 woglinde: We added lots of gettext sanity checking May 06 21:38:25 rp yo May 06 21:39:13 RP: is it normal I have two run.base_do_fetch for each package? May 06 21:39:34 ant__: I suspect not May 06 21:39:45 ah, 2 run.staging_helper too May 06 21:40:18 interestingly the former have same time, the latter not May 06 21:41:30 one run one hour after...perhaps the second or the third image built by my script? May 06 21:41:52 now I'll investigate about failing console-image, the first May 06 21:42:36 the second image should find staged packages, isn't? May 06 21:43:54 uh.. 2 run.read_subpackage_metadata too May 06 21:44:12 anybody sees the same? May 06 21:49:20 PARALLEL_MAKE = "-j5" May 06 21:49:20 BB_NUMBER_THREADS = "4" May 06 21:49:22 here May 06 21:52:09 there must have been some race issue, relaunching the script now console-image is built May 06 21:52:24 damnit, why is this not working.. May 06 21:52:52 running into this problem when building ncurses: tic: symbol lookup error: tic: undefined symbol: _nc_check_termtype2 May 06 21:55:55 btw after last commits here the build-time is about 50% as it was yesterday May 06 21:56:41 ant__: slower or faster? May 06 21:56:53 is much faster May 06 21:56:58 ant__: good :) May 06 21:57:24 congrats! May 06 21:57:57 I have to understand why console-image failed May 06 21:58:00 no logs May 06 21:58:20 just one *run.SRC_DISTRIBUTECOMMAND May 06 21:58:30 no other run. May 06 21:59:33 (executable, I mean) May 06 22:00:38 interesting..now is stalled on bitbake package-index May 06 22:00:43 Configuring staging-package-index-c7x0-angstrom-linux-gnueabi. May 06 22:00:43 NOTE: Staging package found, using it for /oe/openembedded/recipes/meta/package-index.bb. May 06 22:00:51 since 3-4 mins.. May 06 22:03:02 he, it seems it is running an empty run.packagestage_scenefunc May 06 22:07:37 opkg-cl waiting input? May 06 22:07:50 well patch submitted. hope it's ok. thx again khem for your help. May 06 22:09:58 hm May 06 22:10:00 stage-manager-ipkg -f /oe/build/tmp/pstage/opkg.conf -o /oe/build/tmp list_installed | awk '{print $1}' > /oe/build/tmp/pstage/installed-list May 06 22:10:58 or like the second one: May 06 22:10:59 < #stage-manager-ipkg -f /oe/build/tmp/pstage/opkg.conf -force-depends -o /oe/build/tmp install /oe/build/tmp/pstage/angstromglibc/staging-package-index-c7x0-angstrom-linux-gnueabi_1.0-r0_i686-linux.ipk May 06 22:11:14 force-depends? May 06 22:12:54 pls try bitbake package-index May 06 22:14:24 mwester: probably I hit your opkg-cl issue May 06 22:16:23 Keyboard Interrupt, closing down... May 06 22:16:24 Waiting for 1 active tasks to finish: May 06 22:16:24 1: package-index-1.0-r0 do_setscene (pid 621) May 06 22:16:53 sorry for the noise but there is smthg wrong somewhere May 06 22:36:36 RP: rebuilding from pstage May 06 22:36:37 NOTE: package stagemanager-native-0.0.1-r12: task do_compile: Succeeded May 06 22:36:38 NOTE: Running task 17 of 5034 (ID: 1135, /oe/openembedded/recipes/stage-manager/stagemanager-native_0.0.1.bb, do_install) May 06 22:36:38 cp: cannot stat `/oe/build/tmp/work/i686-linux/stagemanager-native-0.0.1-r12/sysroot-destdir//oe/build/tmp/*': No such file or directory May 06 22:39:26 and now again..stuck at task do_setscene May 06 22:39:33 NOTE: package m4-native-1.4.14-r0.0: task do_setscene: Started May 06 22:40:17 bitbake 1.10.0 May 06 22:43:18 hmmm May 06 22:50:23 now rebuilding from scratch I've seen a line scrolling.. May 06 22:50:28 missing gettext ? May 06 22:50:31 ... May 06 22:50:41 retry May 06 22:51:42 like woglinde reported, should be coreutils-native May 06 22:54:37 RP: have a chance to look at this code I've been playing with? the main question, I think, is the approach. The use of a tree of components to represent the value, then collapsing it on expand, rather than doing re.sub's. We could add variable reference tracking without doing this, by maintaining separate tables manipulated by the re.sub functions, but I think this is a bit cleaner. thoughts? May 06 23:00:39 kergoth: I haven't had a chance to look in detail yet, sorry :( May 06 23:00:45 no worries May 06 23:00:50 any thoughts on the concept? May 06 23:01:26 can always refactor the code, but if the direction is wrong, it'd need to be scrapped :) May 06 23:01:31 uh oh...NOTE: package bzip2-full-native-1.0.5-r1: task do_configure: Started May 06 23:01:31 grep: /oe/build/tmp/work/i686-linux/bzip2-full-native-1.0.5-r1/bzip2-1.0.5/.pc/bzip2-1.0.5-autoconfiscated.patch/configure.ac: Permission denied May 06 23:01:40 Your current approach sounds reasonable as much as I understand it May 06 23:01:41 * kergoth thinks ant__ gets the weirdest errors May 06 23:02:00 quilt gives strange permissions on its files May 06 23:02:06 i just added proper checking for recursion in expansions.. never did get that added to the original expand() :) May 06 23:02:08 quilt has problems too May 06 23:02:21 NOTE sub.current=none May 06 23:02:26 Why something is trying to grep that in do_configure on the other hand... May 06 23:02:39 kergoth: I'd love nothing more than to be looking at that code :/ May 06 23:03:01 poor RP, you need some spare time, or work needs to let you do more fun stuff :) May 06 23:04:38 RP: definitely quilt and coreutils natives need a check May 06 23:06:06 and somewhere there are too many slashes May 06 23:06:08 NOTE: package file-native-5.04-r2.0: task do_populate_sysroot: Started May 06 23:06:15 sry May 06 23:06:16 NOTE: package file-native-5.04-r2.0: task do_populate_sysroot: Started May 06 23:06:18 kergoth: I need a holiday. One I don't spend trying to fix a house ;-) May 06 23:06:22 NOTE: package file-native-5.04-r2.0: task do_populate_sysroot: Started May 06 23:06:22 hehe May 06 23:07:07 i own a house, but its just a townhouse, all exterior is covered by the HOA May 06 23:07:37 * mwester owns a house, but seems to live in a hotel mostly. May 06 23:07:43 * RP is trying to write a letter to the neighbour before he goes away tomorrow explaining some acts of parliment to him after he suggested I could always read said acts May 06 23:08:08 - /oe/build/tmp/work/i686-linux/file-native-5.04-r2.0/sysroot-destdir///oe/build/tmp/sysroots/i686-linux/usr/lib/libmagic.so.1.0.0: RPATH=/oe/build/tmp/sysroots/i686-linux/usr/lib May 06 23:08:15 three /// May 06 23:08:29 He doesn't agree I have a right to demolish a shared wall. Its going to be fun May 06 23:08:44 ugh, that sounds painful May 06 23:08:52 RP: ouch May 06 23:08:53 * RP is spending 50% of the time away from home atm :/ May 06 23:09:11 I will rebuild it, bigger than it currently is May 06 23:09:33 its a long story... Maybe some other time :) May 06 23:09:43 heh :) May 06 23:11:05 one can choose the wife but not the neighbours :/ May 06 23:11:41 This guy has managed to rub me up the wrong way... May 06 23:11:51 there was a movie with Belushi wrt May 06 23:12:54 RP: what if I tell you a friend of mine has been suited because of the 'sound' of his Harley? May 06 23:13:13 neighbour called police... May 06 23:13:19 ant__: He doesn't know about my offroad bike yet ;-) May 06 23:13:26 hehe May 06 23:13:26 :D May 06 23:13:48 He did remark "you're not going to be revving engines up all the time are you?" May 06 23:14:27 To which I replied, no whilst pondering whether to invite my brother around with his Ducati and its Termis May 06 23:15:10 The offroad bikes ins't that bad really but he'll hate the idea of it May 06 23:16:26 He tried to explain to my builders they were doing the work wrong a few weeks ago May 06 23:17:17 Their reply "if we did it like that it would do X and Y, just like this bodge up job thats been done here" whilst running a hand over his front wall May 06 23:17:33 I even managed a straight face May 06 23:17:44 Anyhowm don't get me started May 06 23:18:29 * kergoth thinks about how to do something vaguely like static single assignment form with bitbake variables May 06 23:21:31 ah May 06 23:21:32 NOTE: package atk-native-1.30.0-r0: task do_populate_sysroot: Started May 06 23:21:32 new rpath '$ORIGIN/../../../../usr/lib:$ORIGIN/../../../../../..' too large; maximum length 50 May 06 23:21:40 need more ? ;) May 06 23:24:23 same for libdbus-glib-1.so.2.1.0: n May 06 23:25:39 curl-native too seems missing gettext May 06 23:26:03 hmm.. thus far I've managed to avoid making the Value objects know the key they're stored under in the metadata.. i wonder if that's best May 06 23:32:23 http://kergoth.pastey.net/136227 .. not the prettiest display of the path with the recursion.. /me thinks May 06 23:32:43 without the values knowing their names, not sure it can get much prettier, unless I teach the values to work back to their unexpanded string forms May 06 23:32:51 then itd at least show ${FOO} -> ${BAZ} May 06 23:34:13 maybe i should show the initial item twice, making it clear there's a loop May 06 23:34:18 ${FOO} -> ${BAZ} -> ${FOO} May 06 23:34:24 Tartarus: I hit the same chrpath lenght limits with natives as you expected (Was Re: [oe] Request for branch merge) May 06 23:34:44 NOTE: package gcc-cross-intermediate-4.3.3-r11.1: task do_populate_sysroot: Started May 06 23:34:44 new rpath '$ORIGIN/../../../../usr/lib:$ORIGIN/../../../../../..' too large; maximum length 50 May 06 23:36:13 kergoth: It looks clear to me and we probably want to stop printing the stack when we know what the error is and provide a message May 06 23:36:23 but then I'm used to bitbake tracebacks :/ May 06 23:36:26 yeah definitely. this isn't integrated into bitbake yet, so :) May 06 23:36:44 RP: where is set that lenght 50? May 06 23:37:36 ant__: its the length of the original string May 06 23:37:56 ok, I feared was Gentoo's chrpath May 06 23:44:36 oh..renderproto fro console-image is wonderful... May 06 23:45:47 We really need to build and use our own chrpath :) May 06 23:46:00 Tartarus: not as easy as it sounds May 06 23:46:06 think about it May 06 23:46:10 RP: Yes May 06 23:46:19 RP: But we need it all the same. May 06 23:46:29 RHEL isn't usable now May 06 23:46:46 Or we could just link $ORIGIN ;) May 06 23:47:03 Tartarus: It will break, badly May 06 23:47:08 What? May 06 23:47:16 Tartarus: We've done this one ;-) May 06 23:47:39 linking $ORIGIN is fine May 06 23:48:25 I'm just reminding you that RHEL isn't an easy OE host now May 06 23:48:49 Tartarus: chrpath isn't available on that platform from anywhere? May 06 23:48:59 fwiw it seems the strings are 53 chars long..what about $OR ;) May 06 23:49:36 RP: It's not in rhn, no. May 06 23:50:05 Tartarus: Sad. I'm sure we can come up with a way to make something work re: a chrpath-native May 06 23:52:37 ok, after having posted a ton of issues I shall go and sleep... May 06 23:52:42 'night guys May 07 01:04:16 woot May 07 01:04:38 got a recursion check at parse time working in addition to the one at expansion time May 07 01:04:39 kergoth.RecursionError: Recursive variable reference for BAZ via 'FOO' -> 'BAR' -> 'BAZ' May 07 01:07:10 sorry to interrupt kergoth, but umm, was wondering if anyone knew if its possible to render output in a not so standard resolution like 880x800 May 07 01:07:49 "render output"? May 07 01:35:19 kergoth: I mean the screen resolution May 07 01:35:28 as opposed to the usual 1024x768 etc May 07 01:38:07 that's not enough information. i'm sure such a thing is possible, *if* all the bits and pieces support it, from the hardware and drivers on up to the UI you're using May 07 01:38:08 * kergoth shrugs May 07 01:40:48 hmm May 07 01:41:02 is the fact that I have a gumstix with an omap3530 help? May 07 01:41:07 does* May 07 01:41:19 omap3530 + oe May 07 01:41:24 with linux kernel .29 May 07 01:45:38 maybe? anything else I can say that might help? **** ENDING LOGGING AT Fri May 07 02:59:56 2010