**** BEGIN LOGGING AT Sun Feb 12 10:59:58 2006 Feb 12 12:07:24 morning Feb 12 12:41:08 zecke: bug 1540 talks about some opie stuff. To whom should this be assigned? Feb 12 12:41:54 more specifically, it talks about opie-opie-networkapplet and opie-opie-networksettings-wlanplugin Feb 12 12:42:08 pb_: which bugzilla? Feb 12 12:42:14 handhelds.org Feb 12 12:42:25 http://handhelds.org/~bugzilla/show_bug.cgi?id=1540 Feb 12 12:42:27 pb_: I'm tempted to say INVALID Feb 12 12:42:41 pb_: opie-bugs.oszine.de would be the appropriate place Feb 12 12:42:46 just on general principle, or because this bug is actually invalid? Feb 12 12:43:01 pb_: due the usage of the 'wrong' bugtracker Feb 12 12:43:06 okay Feb 12 12:43:23 INVALID is a bit of a harsh resolution for that case. Feb 12 12:43:38 I guess we should figure out how to move bugs between bugzillas. Feb 12 12:43:50 pb_: we use mantis :} Feb 12 12:43:57 doh Feb 12 12:44:17 pb_: I will try to find my handhelds.org bugzilla password and reply Feb 12 12:45:00 okay, cool Feb 12 12:45:12 I added a note to the bug referring to the opie-bugs thing you mentioned Feb 12 12:45:24 pb_: I wonder about the "manual restarting doesn't work" Feb 12 12:45:51 yeah, that's a bit mysterious Feb 12 12:46:33 it does sound like there is some bug there, but I'm not sure exactly what it is Feb 12 12:47:04 I guess I should repair my wifi network so that I can test these things Feb 12 12:47:12 :-} Feb 12 12:47:58 IIRC we are just calling ifup wlan0 Feb 12 13:11:38 did anyone ever manage a berlios project? Feb 12 13:11:47 I can't see where to log in to change the website Feb 12 13:24:47 03pH5 07org.oe.dev * r9d680ed9... 10/ (4 files in 3 dirs): Feb 12 13:24:47 gaim_cvs: drop strftime and desktop-name patches Feb 12 13:24:47 - Drop the configure-strftime.patch again. Upstream decided to always Feb 12 13:24:47 enable the compatibility code when cross compiling. Feb 12 13:24:47 - Delete the unused desktop-name-cvs.patch. Feb 12 13:34:39 * ashak yawns, wish i'd read the SpeedUpMonotne page of the wiki before trying to pull a more up to date copy of org.openembedded.dev Feb 12 13:35:32 192 revision written out of 264, in just over two hours. Feb 12 13:37:09 Bah, no wonder, the cpu frequency scaling on my laptop isn't increasing the cpu frequency. Feb 12 13:42:48 ashak: I have written 452 revs in two hours (on ewi) Feb 12 13:43:09 * ashak blinks Feb 12 13:43:19 ashak: that is on ewi Feb 12 13:43:40 This has sped up not that i've forced the cpu frequency to 1.87GHz Feb 12 13:44:04 s/not/now Feb 12 13:57:57 Hmm, monotone: misuse: tid 438 (packages/linux/linux-openzaurus-2.6.15) clobbered tid 2245 (packages/linux/linux-openzaurus-2.6.15) Feb 12 13:58:19 DUring monotone update Feb 12 13:58:43 Is that anything serious? Feb 12 14:01:22 Hmm, guess so, as when I try to run bitbake is complains: ERROR: Unable to open conf/bitbake.conf Feb 12 14:02:03 Hmm, that appears to be totally missing, in fact teh whole org.openembedded.dev/conf dir is missing Feb 12 14:03:47 time for a new checkout huh? Feb 12 14:05:15 meh, we should _really_ agree on a single server to push to. Feb 12 14:05:27 merging sucks Feb 12 14:08:44 Looks like it. Can't see what's happened though Feb 12 14:08:52 Or why it's happened Feb 12 14:09:18 Unless the error I get during monotone update is what's cuasing it. Feb 12 14:10:03 mickeyl: RP ping Feb 12 14:14:53 morning Feb 12 14:18:58 * ashak sighs: monotone: fatal: std::logic_error: smap.hh:94: invariant 'I(lagged->first != leader->first)' violated Feb 12 14:25:48 zecke: pong Feb 12 14:43:26 hi guys Feb 12 14:44:25 Morning mickeyl. Feb 12 14:46:15 hi mickeyl Feb 12 14:48:49 mickeyl: good morning! Feb 12 14:49:07 03coredump 07org.oe.dev * rcdebeb67... 10/classes/efl.bbclass: efl.bbclass: Fix efl/*.bb ipk generation. Feb 12 14:49:12 03coredump 07org.oe.dev * ra6f5fab8... 10/packages/altboot/ (9 files in 5 dirs): Feb 12 14:49:12 altboot Feb 12 14:49:12 - Do not show option for USB boot on devices w/o USB host Feb 12 14:49:12 - Disable "Boot original INIT" Feb 12 14:49:12 - Do not drop into emergency shell on die() but restart altboot with menu (doesn't work for every plugin, yet) Feb 12 14:49:15 - Some fixes / enhancements for tar.gz installer Feb 12 14:49:17 - Remove "orce" error message when cmdline "-force" is used (should have seen this coming heh) Feb 12 14:50:13 hi twiun Feb 12 14:50:17 hey pb_ Feb 12 14:50:34 so... what's this free beer that renoo mentioned? Feb 12 14:51:18 ah, he was talking about gcj and/or gij on arm Feb 12 14:57:25 got my interest :) Feb 12 15:00:46 hi mickeyl Feb 12 15:00:56 hey RP Feb 12 15:01:31 pb_: Do you fancy reviewing a solution to the debian.bbclass RDEPENDS issue? Feb 12 15:01:53 sure Feb 12 15:04:18 pb_: http://www.rpsys.net/openzaurus/temp/oe_debianfix-r0.patch Feb 12 15:04:45 pb_: There are a few issues it raises that might need dicussion... Feb 12 15:06:14 Re-reading that patch, my functions names look totally wrong. The code itself stands though :) Feb 12 15:06:26 okay, I'll take a look Feb 12 15:13:10 I'm not sure I understand why this is so specific to debian.bbclass. Can you not do all this with generic code in package.bbclass? Feb 12 15:15:24 debian.bbclass is the thing that alters the package naming. The functions could be put in generic code but we'd only want them used when debian.bbclass was active Feb 12 15:15:49 I would have thought you'd want them active whenever PKG is set, irrespective of what is setting it. Feb 12 15:16:26 (for packages like gtk, for example) Feb 12 15:16:46 Yes, I'd agree with that Feb 12 15:16:49 There is also a nasty problem here. Technically, I should add BUILD_ALL_DEPS = "1" to debian.bbclass Feb 12 15:17:23 If that's not added and a RDEPEND hasn't been built, it can put the wrong thing into the package dependency line Feb 12 15:17:39 indeed, that should be added Feb 12 15:19:22 An alternative I wondered about might be to add the debian_pmaps diretctory into the OE metadata itself... Feb 12 15:26:31 Hmm: checking target system type... Invalid configuration `INVALID-linux': machine `INVALID' not recognized Feb 12 15:26:47 ...? That doesn't sound like a very attractive option. Feb 12 15:27:34 To create a 'machine' is it not a simple case of creating a machine.conf for the machine and having a MACHINE="machine" line in local.conf? Feb 12 15:27:42 ashak: sounds like your TARGET_ARCH is... invalid Feb 12 15:28:00 pb_: The point being that debian.bbclass isn't like to change the mappings it makes and if it does, we probably need to know about it Feb 12 15:30:09 pb_: That's what I assumed, but surely TARGET_ARCH comes from the machine.conf files, which has TARGET_ARCH = "arm" in it. bitbake outputs TARGET_ARCH = "arm" before it starts building. Feb 12 15:30:18 That much is certainly true, and having a way to verify that the package naming is correct would be nice. But I think that would be an addition, not an alternative. Feb 12 15:31:15 ashak: yes, indeed it should. I guess you will have to inspect the run.do_configure output to locate the bug. Feb 12 15:38:30 dammit, I was getting on really well with all of this, until today. Now i've spent 3/4 of the day arsing about trying to get it to do something again. God knows what happened to it all :( Feb 12 15:41:49 pb_: So having it check a directory in staging, then a directory within the metadata as a fallback would be acceptable? Feb 12 15:46:21 ah, touching conf/local.conf allowed it to happily find the machine i'd added. Feb 12 15:49:02 * zecke steals part of ebuild.sh Feb 12 15:50:11 hmm, ok, more problems. Tried buildimg opie-image and it's failed with a number of Cannot satisfy the following dependencies for The ones it's failing on are task-opie-bluetooth, task-opie-irda, task-opie-games-extra, task-opie-extra-apps, task-opie-extra-styles. Each of those are missing anywhere from 2-6 dependencies. Shouldn't it just try to build them, seeing as they're dependencies? Feb 12 16:02:47 If I try to build one of the dependancies manually it seems to build it (and its dependancies) ok, so why doesn't opie-image target cause it to try and build that package? Feb 12 16:35:45 03eFfeM 07org.oe.dev * rc1c29580... 10/packages/pvrusb2-mci/pvrusb2-mci_20060209.bb: pvrusb2-mci: added release 20060209 Feb 12 17:14:35 good pax-utils are packaged Feb 12 17:14:53 phew Feb 12 17:15:30 pb_: I'm ripping ebuild.sh to use scanelf to check for some things Feb 12 17:18:06 hmm how does one clear "ie: empty" a tty where no login prompt is running? Feb 12 17:18:26 CoreDump|home: I don't understand that Feb 12 17:18:38 on my Z I can chvt2 Feb 12 17:18:46 but there's no login on that tty Feb 12 17:19:10 when i echo "blahh" > /dev/tty2 "blahh" is displayed on that otherwise empty tty Feb 12 17:19:26 how do I empty the display of tty2? Feb 12 17:21:19 there's an escape code for that Feb 12 17:21:28 bear with me i don't know it offhand though Feb 12 17:22:05 * CoreDump|home googles Feb 12 17:23:42 hi all Feb 12 17:24:43 zecke: there is no 1.3.3 in bitbake svn tags, only a svn://svn.berlios.de/bitbake/branches/bitbake-1.3.3 - I suppose that's not a normal situation :) Feb 12 17:25:13 zecke: should I consider this URL to be 1.3.3 really ? Feb 12 17:25:20 lol Feb 12 17:26:29 hmm the only escape code I can find are for a sh shell. There's no shell running on that tty=) Feb 12 17:27:09 Good evening all Feb 12 17:28:51 hi uv1 Feb 12 17:29:49 btw, has developement on bitbake v2 started, or is it still just in planning stage ? Feb 12 17:30:10 Is it possible to show all DEPENDS / RDEPENDS (recursively) which a package uses ? Feb 12 17:30:11 yann: blanning Feb 12 17:30:15 yann: planning even Feb 12 17:30:56 \033[2J Feb 12 17:31:05 worth a try Feb 12 17:31:43 mickeyl: that works perfectly, thanks! Feb 12 17:31:47 np Feb 12 17:31:58 zecke: is the content of branches/bitbake-1.3.3 really the 1.3.3 release ? Feb 12 17:32:01 VT100/ANSI rocks :) Feb 12 17:32:20 yann: yes, at least this is what I haved packaged and uploaded Feb 12 17:32:29 ok, thx Feb 12 17:32:30 yann: I just recognized that the ./setup.py bdist is fucked up Feb 12 17:33:55 FYI: "\033c" is even better as it moves the "cursor" back to the first line Feb 12 17:35:19 What is linux-jlime? Feb 12 17:36:36 iirc a dedicated distro for HPs Jornada Feb 12 17:36:51 doesn't seem to be pretty maintained though Feb 12 17:37:20 Hmm, why is bitbake trying to build that? Feb 12 17:38:52 probably because that some dependencies are not met for what you want to build Feb 12 17:39:03 use bitbake -nv and have a look at the output Feb 12 17:39:07 Right, that might make sense Feb 12 17:39:16 Ok, i'll try thanks Feb 12 17:57:24 anyone here got the link(s) to OZ3.5.4? Feb 12 17:57:28 I' dlike to test Feb 12 18:01:50 looks like linux-jlime-sh3 is the first in a list of multiple providers for virtual/kernel, which makes total sense :) Feb 12 18:05:54 sounds like you didn't specify which provider you want, then Feb 12 18:06:10 check if your MACHINE=foo is pointing to a valid conf file Feb 12 18:06:17 I haven't, as the kernel for this device requires a patched omap kernel. Feb 12 18:06:17 in which the preference is stated Feb 12 18:06:25 ah Feb 12 18:06:36 you may say ASSUME_PROVIDED += virtual/kernel then Feb 12 18:06:39 SO I only have a partial machine config file atm :) Feb 12 18:06:45 Ah right Feb 12 18:06:53 some packages may need a kernel though Feb 12 18:06:59 see how far you can get Feb 12 18:08:51 THanks, i'll try, fialing that i'll try to work out how I can get it to create and patch the correct kernel for me, i'm sure it's nothing too difficult. Feb 12 18:17:54 Ah, look slike it was required by pcmcia-cs, which is required by bluez-utils-nodbus. Feb 12 18:21:28 which is an RDEPEND for task-opie-bluetooth Feb 12 18:23:14 So is there a simple way to disable certain parts of opie, or the requrement for pcmcia_cs, as i'd be fairly certain that this device doesn't have pcmcia Feb 12 18:32:01 * france is back (gone 12:00:19) Feb 12 18:36:49 hi Feb 12 18:37:41 hi hrw Feb 12 18:37:43 mickeyl|dinner: ping Feb 12 18:38:06 ashak, does your device use any other "plugable" media or devices? Feb 12 18:38:29 i could be wrong but i think they use pcmcia_cs Feb 12 18:38:43 cf, sd, etc Feb 12 18:39:03 emte: cf does as its basically pcmcia. sd doesn't Feb 12 18:39:12 there we go :) Feb 12 18:40:21 whats new in .dev land? Feb 12 18:40:49 hrw: I just pushed fixed for the debian.bbclass renaming? :) Feb 12 18:40:52 i put macchanger on the buglist for someone to add :) Feb 12 18:41:20 ~lart kopete again Feb 12 18:41:20 * ibot urinates on kopete again Feb 12 18:41:31 nice ibot Feb 12 18:42:26 hrw: Finished all the wedding plans now? :) Feb 12 18:42:56 RP: has to find cameraman, dance course and few others Feb 12 18:43:13 03rpurdie 07org.oe.dev * rd9ef02ad... 10/packages/libmatchbox/libmatchbox_svn.bb: libmatchbox_svn: Correct PV field Feb 12 18:43:19 03rpurdie 07org.oe.dev * r68843a39... 10/packages/gcc/gcc-cross_3.3.4.bb: gcc-cross 3.3.4: Use generic includes instead of repeating code (the code is equivalent but the .inc files contain fixes) Feb 12 18:43:23 03rpurdie 07org.oe.dev * r0e49f763... 10/packages/galago/galago-daemon_0.3.4.bb: galago-deamon 0.3.4: Package a missing file Feb 12 18:43:29 03rpurdie 07org.oe.dev * rd15a485e... 10/packages/eds/eds-dbus_svn.bb: eds-dbus: Correct RDEPENDS and PV Feb 12 18:43:34 03rpurdie 07org.oe.dev * ra84be889... 10/conf/distro/ (2 files): preferred_gpe_versions 2.6/2.7: Correct some strings to match PV's that changed and to reflect issues that were fixed in bitbake Feb 12 18:43:34 hrw: The planning seems to take some doing from what I've seen :-/ Feb 12 18:43:39 03rpurdie 07org.oe.dev * r078227ef... 10/packages/fontconfig/fontconfig_2.2.95.bb: fontconfig: Fix handling of fontconfig-utils properly, once and for all Feb 12 18:43:45 03rpurdie 07org.oe.dev * r574359d9... 10/classes/debian.bbclass: Add DEBIAN_NOAUTONAME_ variable to disable debian.bbclass renaming for specific packages. Feb 12 18:43:49 03rpurdie 07org.oe.dev * r4b9cff7a... 10/classes/ (debian.bbclass package.bbclass package_ipk.bbclass): (log message trimmed) Feb 12 18:43:49 Fix package name renaming for classes like debian.bbclass: Feb 12 18:43:49 package.bbclass now tracks any package renaming performed using PKG_*. Feb 12 18:43:49 When a package is built, the runtime variables are translated using Feb 12 18:43:50 this cached information which is stored in staging. Feb 12 18:43:52 From now on, only use package names are valid in runtime variables like Feb 12 18:43:54 RDEPENDS, RRECOMMENDS and RSUGGESTS. eg. fontconfig-dev is correct, Feb 12 18:44:03 hrw, the photographer is the worst out of all the planning Feb 12 18:44:47 emte: I got it Feb 12 18:45:08 a photographer? Feb 12 18:45:33 emte: photographer. have to get cameraman==movie man Feb 12 18:45:37 ah Feb 12 18:45:48 then the hard stuff is done Feb 12 18:45:51 :) Feb 12 18:46:18 aside from paying for it Feb 12 18:46:22 :( Feb 12 18:46:26 emte: thats my common mistake - kamera in Polish is 'movie camera', photo camera is other name Feb 12 18:47:28 ah, yeah living north america one doesnt always think ... must be the air Feb 12 18:48:42 emte: dont know - I rather avoid Us Feb 12 18:50:59 bit father north, where its all about beer, curling, hockey, and eating doughnuts Feb 12 18:51:19 and maple syrup if your on the east coast Feb 12 18:52:06 * emte curses at eet Feb 12 18:52:12 emte: Nope, just sd. it's a /smart/phone. Feb 12 18:53:21 RP: you know - first do all things which need timing, then all the rest Feb 12 18:58:42 someone can build python-pylinda from .dev? Feb 12 18:59:33 hrw: pong Feb 12 18:59:41 ups Feb 12 18:59:43 mickey: python-pylinda fails to patch Feb 12 19:00:05 | 1 out of 2 hunks FAILED -- rejects in file setup.py Feb 12 19:00:05 | Patch honor-datadir.patch does not apply (enforce with -f) Feb 12 19:00:06 nasty Feb 12 19:00:50 mickeyl: what do you think about starting feed build? I will switch all pxa machines to 'arm' in .oz354fam083 and we can start world build (it takes some time) Feb 12 19:01:33 hrw: sounds good to me. should I do the world build?= Feb 12 19:02:01 mickeyl: I do not have machine for it - dont want to do it on ewi as this machine has other usage too Feb 12 19:02:24 no problem. my machine will likely take 2 days or so for it though Feb 12 19:02:56 hrw, can you add macchanger first? Feb 12 19:03:05 mickeyl: machine=collie;bitbake opie-image gpe-image;bitbake -k world? Feb 12 19:03:13 emte: to .dev and .oz354? Feb 12 19:03:13 exactly that Feb 12 19:03:13 03mickeyl 07org.oe.dev * re1bf3143... 10/packages/python/ (python-2.4.2-manifest.inc python_2.4.2.bb): python 2.4.2: fix dependencies for python-re and python-stringold Feb 12 19:03:20 to both?? Feb 12 19:03:41 i can do that, but for what would we need .dev feed? Feb 12 19:04:00 mickeyl: no .dev feed Feb 12 19:04:06 hrw, i'd think .oz354 its usefull for wireless G cards that dont set a MAC on thier own Feb 12 19:04:15 mickeyl: .oz354fam083 is the only one Feb 12 19:04:25 emte: ok. we have agere support in .oz354fam083 Feb 12 19:04:35 mickeyl: can you login into ewi? Feb 12 19:04:44 hrw: i think i have the pass, but i never tried Feb 12 19:05:13 mickeyl: ah. Feb 12 19:05:20 hrw, yeah i am using the jan29 snap of familiar right now, using wavelan for my WCF45G Feb 12 19:06:18 mickeyl: http://ewi546.ewi.utwente.nl/tmp/hrw/oz3.5.4-test-images/test5/._feed/ - I want to use this structure for 3.5.4 Feb 12 19:06:37 WCF54G* Feb 12 19:07:01 mickeyl: Is there any reason python 2.3.4 doesn't provide python-core btw? Feb 12 19:07:22 You can do stuff like DEPEND_machine, for machine specific dependancies can't you? Does that mean you can also do SRC_URI_machine for machine specific patches or something? Feb 12 19:07:45 ashak: SRC_URI_append_machine = " something " Feb 12 19:07:58 hrw: That's it, thanks Feb 12 19:08:17 hrw, http://bugs.treke.net/show_bug.cgi?id=673 for the bb Feb 12 19:08:27 SRC_URI_machine will also work but might mean you have to duplicate data as it will overwrite SRC_URI Feb 12 19:08:40 ~lart ipkg-make-index for beeing dog slow Feb 12 19:08:41 * ibot beats the living hamstercrap out of ipkg-make-index for beeing dog slow Feb 12 19:10:09 03freyther 07org.oe.dev * r18e05897... 10/ (3 files in 2 dirs): Feb 12 19:10:09 conf/bitbake.conf: Feb 12 19:10:09 - Add GENTOO_MIRROR Feb 12 19:10:09 packages/pax-utils: Feb 12 19:10:09 - Add version 0.1.9 of the pax utilities including Feb 12 19:10:09 scanelf. I will use it in a dedicated QA bbclass Feb 12 19:10:13 to check for common security issues. Feb 12 19:11:49 emte: its nasty recipe.. DEPENDS="" etc.. Feb 12 19:11:54 emte: SRCDATE=PV? Feb 12 19:12:43 emte: http://pastebin.com/551518 is nicer Feb 12 19:12:46 i cant remember where that came from ... Feb 12 19:12:54 its staic anyway Feb 12 19:12:57 NOTE: package macchanger-1.5.0-r0: task do_build: completed Feb 12 19:12:58 RP: there shouldn't be a package named python-2.3.4 at all, since it's just the package name. Feb 12 19:13:03 s/package/bbfile/ Feb 12 19:13:03 mickeyl meant: RP: there shouldn't be a bbfile named python-2.3.4 at all, since it's just the package name. Feb 12 19:13:26 heh Feb 12 19:13:29 he got it wrong Feb 12 19:13:30 ~shoot ibot Feb 12 19:13:32 * ibot shoots ibot in the head with a frozen turkey cannon! Feb 12 19:14:26 he got it right, you got your regex wrong :P Feb 12 19:14:33 emte: http://pastebin.com/551525 updated Feb 12 19:14:56 hrw, yeah that is cleaner, since oe is down its a bit hard to find a reference for whats required in the new metas Feb 12 19:15:39 hmm Feb 12 19:15:55 we no longer need to define ${S} ? Feb 12 19:16:06 mickeyl: python_2.4.2.bb PROVIDES = "python python-core" whereas python_2.3.4.bb doesn't. I wasn't referring to RPROVIDES? Feb 12 19:16:08 S=WORKDIR/PN-PV Feb 12 19:16:36 S = "${WORKDIR}/macchanger-${PV}" isnt that what that is? Feb 12 19:16:49 emte: its not needed in .bb Feb 12 19:16:54 ah Feb 12 19:17:19 emte: its like writing PV=1.5.0 in file_1.5.0.bb Feb 12 19:17:39 emte: working but not needed repeating Feb 12 19:17:45 true Feb 12 19:18:44 pushed Feb 12 19:19:45 # Your system Evas Benchmark: Feb 12 19:19:46 # Feb 12 19:19:46 # EVAS BENCH: 0.367 Feb 12 19:19:49 RP: interesting. i don't see why either of those should PROVIDE python-core Feb 12 19:20:05 thanks, now i can stop compiling it outside and droping it in to use it :) Feb 12 19:20:11 up from 0.348 w/ kernel 2.4 Feb 12 19:20:35 mickeyl: http://bugs.treke.net/show_bug.cgi?id=678 is blocker for release Feb 12 19:20:49 mickeyl: and you are the only one who can resolve it... Feb 12 19:21:58 uh oh... ok, i'll try to handle that. Feb 12 19:23:37 mickeyl: I suspect the meta slugos-native package is doing something evil here as I think it has it in a DEPENDS indriectly Feb 12 19:24:25 03hrw 07org.oe.oz354fam083 * r2f66f394... 10/conf/machine/ (poodle.conf tosa.conf zaurus-clamshell.conf): zaurus machines: switch to tune-strongarm instead of xscale Feb 12 19:24:30 03hrw 07org.oe.oz354fam083 * r3682a18a... 10/packages/macchanger/macchanger_1.5.0.bb: added macchanger 1.5.0 - close #673 Feb 12 19:24:34 03hrw 07org.oe.dev * re4f666d6... 10/packages/macchanger/macchanger_1.5.0.bb: added macchanger 1.5.0 - close #673 Feb 12 19:25:09 hmm i must have missed the discussion as for why we are switching back to arm for the images? Feb 12 19:25:16 for the pxa devices, that is Feb 12 19:26:16 mickeyl: with multimachine config it is easier to build - no need to build all tools twice Feb 12 19:27:27 hrw: which tools? Feb 12 19:27:28 03hrw 07org.oe.oz354fam083 * r8516ef52... 10/packages/lsof/lsof_4.75.bb: lsof 4.75: fixed SRC_URI - close #664 Feb 12 19:27:45 mickeyl: if we want to get any 'pxa' speed then we should also consider tune-pxa270 for cxx00 Feb 12 19:27:56 * Philippe is back (gone 45:11:40) Feb 12 19:28:11 RP: glibc, whole image content Feb 12 19:28:18 hmm Feb 12 19:28:45 hrw: armv5te is a big speed boost, particualry for glibc. tune-pxa270 is much less of an optimisation Feb 12 19:29:05 ok. then I'll revert it Feb 12 19:29:10 yeah, please do that Feb 12 19:29:18 the benefits outweigh imo Feb 12 19:29:31 anyone seen sirfred lately? Feb 12 19:29:43 mickeyl: He is around and still working on aticore Feb 12 19:29:45 i had hoped to get the w100 accel in 3.5.4 by default Feb 12 19:29:50 ah good to know Feb 12 19:30:18 He and another person are making good progress with it. Sadly, I've not had much time to help recently :-/ Feb 12 19:31:05 RP: where does the speed boost come from in armv5te? Feb 12 19:31:34 pb_: Faster memcpy operations (and friends) Feb 12 19:31:59 Which armv5te operations help with memcpy? Feb 12 19:32:17 pb_: preload Feb 12 19:32:31 prefetch even Feb 12 19:32:43 or whatever its called :) Feb 12 19:32:48 pld? You ought to be able to do that in armv4 binaries with no ill effects. Feb 12 19:33:13 pb_: Yes. I'm not sure glibc does by default though Feb 12 19:33:36 ah. well, in that case, it sounds like it would be a better option to patch glibc so that it does, rather than build everything as armv5te. Feb 12 19:34:36 In an ideal world, yes Feb 12 19:36:54 hello ! Feb 12 19:37:11 Sunday and the OE crue is working hard :) Feb 12 19:37:20 not that it matters all that much for openzaurus I guess, but these kinds of changes seem to have an unfortunate habit of sneaking into familiar as well. We definitely don't want armv5te there. Feb 12 19:38:34 pb_: The OZ policy/agreement was that the world builds wouldn't be armv5te but having the images themselves optimised was acceptable/desireable. Feb 12 19:40:02 I don't see the issue for familiar as long as only the feeds themselves are compiled with a machine of the lowest denominator type... Feb 12 19:42:18 mickeyl: I'm going to fix that meta slugos package and remove that PROVIDES python-core line if that's ok with you? Feb 12 19:42:45 ok. I'll revert it for zauruses Feb 12 19:43:27 RP: familiar's policy is that the images and the feeds should match. It isn't acceptable for us to have one build in the images and a differently-compiled one in the feeds. Feb 12 19:44:55 pb_: You don't need a differently compiled one. ipkg is hopefully intelligent enough to know not to install armv5te binaries on a v4 machine Feb 12 19:45:23 So as long as both a v4 and a v5 ipk are in the feed, everything should work... Feb 12 19:45:32 yes, it is, but we don't want to have to double up the packages in the feeds either unless it brings a real benefit. Feb 12 19:45:49 having both architectures in the feeds causes (a) Packages bloat; (b) user confusion; and (c) version skew. Feb 12 19:46:25 in the case of glibc, if we can get the best of all worlds with a single (v4) binary, that's what we should do Feb 12 19:47:00 If that's familar's policy, so be it. Feb 12 19:48:24 * RP -> afk Feb 12 19:48:56 indeed it is. Feb 12 19:51:38 03hrw 07org.oe.oz354fam083 * r1b0ee7db... 10/conf/machine/ (poodle.conf tosa.conf zaurus-clamshell.conf): disapproval of revision '2f66f3942fd222b5e7e77b3dee1f05009eedf27b' Feb 12 19:53:21 ~curse monotone for failing when meld cannot run Feb 12 19:53:24 May you be reincarnated as a Windows XP administrator, monotone for failing when meld cannot run ! Feb 12 19:56:04 mickeyl: what do you think about that feed structure? Feb 12 20:05:06 RP: Do you have any idea why OE selects gcc-corss-kernel-3.3.4 although i declare it to use 3.4.4 ? PREFFERED_VERSION_gcc-cross-kernel = "3.4.4-3.4.4" Feb 12 20:06:34 RP:i am trying to compile the madwifi modules and they require to be compiled by the same compiler version as the kernel... Feb 12 20:07:27 03hrw 07org.oe.oz354fam083 * r0d43130f... 10/packages/module-init-tools/module-init-tools_3.2.2.bb: Feb 12 20:07:27 module-init-tools: change depmod removal - close #671 Feb 12 20:07:27 - "update-alternative --remove depmod" moved to m-i-t-depmod prerm Feb 12 20:07:27 - removed handling of depmod from m-i-t as it is handled in m-i-t-depmod Feb 12 20:07:27 already and user can have only depmod package installed Feb 12 20:07:31 03hrw 07org.oe.dev * r6b5e81c6... 10/packages/module-init-tools/module-init-tools_3.2.2.bb: Feb 12 20:07:32 module-init-tools: change depmod removal - close #671 Feb 12 20:07:34 - "update-alternative --remove depmod" moved to m-i-t-depmod prerm Feb 12 20:07:36 - removed handling of depmod from m-i-t as it is handled in m-i-t-depmod Feb 12 20:07:38 already and user can have only depmod package installed Feb 12 20:28:29 Ifaistos: you have misspelt "preferred" Feb 12 20:28:52 also, "3.4.4-3.4.4" looks erroneous; it should probably just be "3.4.4" Feb 12 20:29:25 also, some of the kernels will force a particular version, irrespective of your preference. look in the .bb file and see if it defines any KERNEL_CCSUFFIX. Feb 12 20:31:32 pb_: It was and since it did not work i tried alternatives. It looks like the x86/epia kernels are using KERNEL_CCSUFFIX = "-3.3.4" Feb 12 20:32:43 pb_ : or to be exact the epia 2.6.11 i used to make my 2.6.12.bb Feb 12 20:35:09 pb_ : Is the order of presidence (which can mody which) like this local.conf -> distro -> machine -> meta -> .bb Feb 12 20:36:56 mickeyl: ping? Feb 12 20:39:14 03freyther 07org.oe.dev * rffe7fe8b... 10/classes/insane.bbclass: Feb 12 20:39:14 org.openembedded.dev/classes/insane.bbclass: Feb 12 20:39:14 -Add the skeleton for the QA bbclass Feb 12 20:48:11 JustinP: have fun explaining him how todo that hehe Feb 12 20:48:41 CoreDump|home: I did it yesterday with SmrtJustin Feb 12 20:48:48 CoreDump|home: or was that Friday....? Feb 12 20:48:59 whenever it was, I remeber it. Feb 12 20:49:32 But I somehow lack the motivation to explain it the hundredth time...so I say nothing heh Feb 12 20:50:51 maybe 3.5.4 should include .tar.gz images. Would make things a heck of a lot simpler for n00bs Feb 12 20:51:05 zecke: I (finally) replied to your email. sorry about the delay. Feb 12 20:51:06 CoreDump|home: 3.5.4-test5 already contain them Feb 12 20:51:15 really? nice =) Feb 12 20:51:16 pb_: don't worry Feb 12 20:52:08 ~praise tarballs/ Feb 12 20:52:10 All hail tarballs/! Feb 12 20:52:27 Guess I need to work on kexec then =) Feb 12 20:52:53 CoreDump|home: for 3.5.4.1? Feb 12 20:53:19 that, too. But I really don't want to flash a2.4 kernel anymore Feb 12 20:55:33 zecke: there are other strange things in svn, like svn://svn.berlios.de/bitbake/tags/bitbake-1.3.1/trunk/bitbake Feb 12 20:55:45 psst Feb 12 20:57:41 good nite Feb 12 20:58:04 CoreDump|home: yes, it should include tgzs Feb 12 20:58:05 n8 zecke Feb 12 20:58:55 damn it.....libSM thing again Feb 12 20:59:06 * JustinP goes for food Feb 12 21:10:33 03coredump 07org.oe.dev * rcf224fd4... 10/packages/gpe-bootsplash/ (files/tty2-message.patch gpe-bootsplash_1.15.bb): Feb 12 21:10:33 gpe-bootsplash: Feb 12 21:10:33 - Display a message informing the user that he can chvt 1 to see the boot-messages Feb 12 21:10:33 - Do a chvt 1 if gpe-bootsplash errors out Feb 12 21:10:39 03coredump 07org.oe.dev * rcb780a1c... 10/packages/gpe-bootsplash/ (4 files in 2 dirs): gpe-bootsplash: Add an opie-theme Feb 12 21:16:27 ~lart x86 coders Feb 12 21:16:28 * ibot steals x86 coders's mojo Feb 12 21:53:13 evening Feb 12 21:57:53 cu Feb 12 22:14:26 * RP returns Feb 12 22:48:49 * Philippe is away: visual contact - melancholic dreams Feb 12 23:00:20 * JustinP pushes a fix for xmonobut... Feb 12 23:00:47 Hi! Feb 12 23:01:00 Anyone with glibc in OE experience around? Feb 12 23:01:24 I'm trying to create a .bb file for glibc-2.3.6 Feb 12 23:04:23 03justinp 07org.oe.dev * ra32a4791... 10/packages/xmonobut/xmonobut_0.4.1.bb: xmonobut_0.4.1: add missing DEPENDS on libmatchbox Feb 12 23:05:35 pb_: are you still around? Feb 12 23:14:00 RP: yes Feb 12 23:18:13 pb_: I'm trying to patch up gcc-cross-sdk a bit. Can you give me a clue what the sdk-libstdc++-includes.patch does? Feb 12 23:19:27 Anyone has a clue where this sort of an error comes from? /usr/bin/install: cannot remove `/usr/lib/libc.a': Permission denied Feb 12 23:19:44 Why doesn't proper DESTDIR make it to make-s command line? Feb 12 23:20:40 I don't remember exactly. I guess you should try without the patch and see what happens. Feb 12 23:20:56 | LANGUAGE=C LC_ALL=C; export LANGUAGE LC_ALL; make -r PARALLELMFLAGS="" CVSOPTS="" -C /home/OpenEmbedded/build/tmp/work/glibc-2.3.6-r1/glibc-2.3.6 objdir=`pwd` install Feb 12 23:21:05 no destdir :( Feb 12 23:21:19 RP: I saw your commits, looks like you may have fixed my lib* auto-munging RDEPEND problem (I haven't tested though) Feb 12 23:21:42 pb_: What would be the difference between a "normal" cross gcc and a one for the sdk, apart from their installed path? Feb 12 23:22:04 JustinP: Yes, its highly likely those patches will help. I did notice you were having problems Feb 12 23:22:12 RP: none, in theory Feb 12 23:22:50 pb_: Ok, that makes me feel slightly better about what I'm doing to the sdk .bb file, thanks :) Feb 12 23:23:11 uhmm.. is there any pastebin where anybody can consolate my frustration ? Feb 12 23:23:17 *could Feb 12 23:23:22 okay, very good Feb 12 23:23:40 good night all Feb 12 23:23:49 Night, pb. Feb 12 23:23:55 'night pb_ Feb 12 23:23:55 'night pb_ Feb 12 23:25:43 night pb_ Feb 12 23:26:03 night all Feb 12 23:26:16 RP: can I stop the auto-lib* stuff for one package in a bb? Feb 12 23:26:31 I'd much rather move my -utils PACKAGEs back into the main bbs... Feb 12 23:26:57 JustinP: Yes - see the DEBIAN_NOAUTONAME_xxx I detailed in the changelog Feb 12 23:27:02 ok Feb 12 23:27:06 JustinP: or see fontconfig for an example Feb 12 23:27:23 I thought I saw a commit like that, just wanted to make sure Feb 12 23:27:32 * JustinP will try it as soon as this build finishes Feb 12 23:27:41 RP: thanks Feb 12 23:45:13 hello Feb 12 23:45:39 if anybody is feeling like helping out, i just followed what GettingStarted indicates Feb 12 23:45:42 http://rafb.net/paste/results/Y0Heh262.html Feb 12 23:47:10 how did you set DISTRO? Feb 12 23:47:31 in conf/distros/familiar.conf Feb 12 23:47:45 what? Feb 12 23:47:49 in your local.conf Feb 12 23:47:50 no, wait, in build/conf Feb 12 23:47:53 yeah that one Feb 12 23:47:59 build/local.conf Feb 12 23:48:02 and what did you set it to? Feb 12 23:48:07 and that's still not right Feb 12 23:48:16 DISTRO=familiar Feb 12 23:48:17 it's conf/local.conf in your build dir Feb 12 23:48:27 that's wrong Feb 12 23:48:28 yeah build/conf/local.conf Feb 12 23:48:39 you have to use a version Feb 12 23:48:52 familiar-unstable for the .dev branch Feb 12 23:49:02 ah! Feb 12 23:49:09 i'll try that Feb 12 23:49:57 i don't have to clear the cache, right ? Feb 12 23:50:28 nope Feb 12 23:50:46 it's fetching again... it'll take like 30 mins Feb 12 23:53:53 hi is there a semi stable that will compile Feb 12 23:54:16 bootstrap-image for the h2200 Feb 12 23:54:51 I just want an env to work on the buil for Feb 12 23:55:02 hx2xxx Feb 12 23:55:32 but want to check my setup worksfirst Feb 12 23:56:13 bleh! Anybody knows what on earth is wrong when bitbake -b file and bitbake task-bootstrap give different results on the same freaking package? Feb 12 23:56:56 When I do bitbake task-bootstrap, glibc build dies like this: | install: cannot stat `/home/OpenEmbedded/build/tmp/work/glibc-2.3.6-r1/libc/sunrpc/rpcsvc/bootparam_prot.h Feb 12 23:58:24 When I do bitbake -b packages/glibc/glibc_2.3.6.bb it dies like this: /usr/bin/install -c -m 644 /home/OpenEmbedded/build/tmp/work/glibc-2.3.6-r1/build-i586-linux/libc.a /usr/lib/libc.a Feb 12 23:58:24 /usr/bin/install: cannot remove `/usr/lib/libc.a': Permission denied Feb 12 23:58:39 Any ideas would be very welcome. Feb 12 23:58:52 I've out of ideas myself and the default glibc doesn't compile for x86 :( Feb 12 23:59:28 I mean glibc_2.3.5+cvs20050627 Feb 12 23:59:31 anttix: did you setup your env right? BBPATH and such? Feb 12 23:59:33 anttix: I belive the epia machine might work for x86 Feb 12 23:59:48 what glibc does it use? Feb 13 00:00:05 glibc_2.3.5+cvs20050627 will die with an error in signal.S Feb 13 00:00:07 Harvy: the .dev branch is failry stab;e, but the oz354fam083 branch is for the new release, so it's more stable Feb 13 00:00:43 anttix: Not if it applies extra patches or changes the settings which I believe it does Feb 13 00:00:54 as long as I can get a compile for Feb 13 00:00:55 I have: export BBPATH=$thisdir/build:$thisdir/org.openembedded.dev Feb 13 00:01:13 I'll check that out, thanx for the hint. Any ideas when newer glibc will be in dev branch? Feb 13 00:01:29 something that I think is good ie h2200 Feb 13 00:01:47 RP: FYI i am making progress on compiling with gcc4 initial cross Feb 13 00:01:55 Oh, and linuxthreads will not compile on x86 unless You set the --without-tls option to configure. Feb 13 00:02:13 then I can start working om the hx2xxx Feb 13 00:02:25 And when using NPTL glibc-intermediate needs to be updated, because it suffers from that problem too, and it suffers from a lack of GCC 4.0 support. Feb 13 00:02:27 shadows: ok, good. I have had a bootstrap image out of gcc4 before... Feb 13 00:02:40 RP: bootstrap works great, thank you for that Feb 13 00:03:10 RP: i'm now up to the gtk-webcore stuff. i've submitted bugs and fixes for expat and libsvg Feb 13 00:03:17 anttix: We need someone to give the glibc .bb files some TLC Feb 13 00:03:41 http://bugs.openembedded.org/show_bug.cgi?id=672 and http://bugs.openembedded.org/show_bug.cgi?id=675 respectively Feb 13 00:05:24 shadows: Looks good. We should get them committed Feb 13 00:05:32 JustinP: thanks for the pointer is there a Feb 13 00:05:43 hmm, strange, I get Host bugs.openembedded.org not found: 3(NXDOMAIN) Feb 13 00:05:54 anttix: read the topic Feb 13 00:06:00 i'm not sure who to ask to commit those, so if you don't mind i'll leave that alone for now Feb 13 00:06:01 site with the branches Feb 13 00:06:12 Oops, I already saw that, yes :P Feb 13 00:06:12 they need to be tested Feb 13 00:06:28 shadows: I can commit them as new .bb files. Feb 13 00:08:27 shadows: Acutally could you test and see if we can change the do_stage function to simply autotools_stage_all() for libsvg? Feb 13 00:09:46 RP, what do You mean by give glibc .bb files some TLC? What is TLC? Feb 13 00:10:05 anttix: Tender Loving Care. We lack a glibc maintainer at the moment Feb 13 00:10:21 We lack lots of maintainers :-( Feb 13 00:10:38 Oh! Nice acronym :) Feb 13 00:10:39 RP: sure... it will be a few min Feb 13 00:14:47 I'm in the middle of creating a .bb file for glibc_2.3.6, however I don't have an ARM machine to test it on :( My main interest is in x86. Feb 13 00:16:44 Now if I only could get this freaking thing to build properly :( I'll try delete-ing the tmp directory, maybe it's cluttered with some silly stuff from my experiments that confuses the bitbake. Feb 13 00:18:09 that's interesting, your target is x86 :) Feb 13 00:19:07 anttix: An x86 glibc would be happily added to OE. Someone else might make it work on arm Feb 13 00:24:03 RP: autotools_stage_all() works fine for libsvg_0.1.4.bb Feb 13 00:24:26 bug updated to reflect the change (with patch) Feb 13 00:25:13 Ok, I'll send the .bb-s to list when I'm ready? Or You prefer another way of submitting them? Feb 13 00:25:18 shadows: thanks Feb 13 00:25:27 anttix: submit to bugzilla Feb 13 00:25:58 anttix: there's some trouble with the DNS, ordinarily the bugtracker is http://bugs.openembedded.org/ Feb 13 00:26:02 Hmm, I create a new bug against glibc? Feb 13 00:26:09 * anttix nods Feb 13 00:26:12 you can reach the bugtracker using http://bugs.treke.net Feb 13 00:26:12 anttix: yes Feb 13 00:26:58 or do what i do, and edit your hosts file to add an entry for bugs.openembedded.org :) Feb 13 00:29:58 ok tried oz354fam083 branch and just gotgnoff: CSMan (Nick collision from services.) Feb 13 00:29:58 CSMan_ is now known as CSMan Feb 13 00:29:59 Signoff: tmbinc (Read error: 104 (Connection reset by peer)) Feb 13 00:29:59 Now if I only could get this freaking thing to build properly :( I'll try delete-ing the tmp directory, maybe it's cluttered with some s Feb 13 00:30:24 try again Feb 13 00:30:26 gnoff: CSMan (Nick collision from services.) Feb 13 00:30:26 CSMan_ is now known as CSMan Feb 13 00:30:26 Signoff: tmbinc (Read error: 104 (Connection reset by peer)) Feb 13 00:30:27 Now if I only could get this freaking thing to build properly :( I'll try delete-ing the tmp directory, maybe it's cluttered with some s Feb 13 00:32:02 Harvy: are you okay? Feb 13 00:32:14 * shadows blinks Feb 13 00:32:22 ok tried oz354fam083 branch and just got ERROR: Nothing provides runtime dependency python Feb 13 00:32:22 NOTE: no buildable providers for ipkg-utils-native Feb 13 00:32:22 ERROR: dependency ipkg-utils-native (for task-bootstrap) not satisfied Feb 13 00:32:23 NOTE: no buildable providers for task-bootstrap Feb 13 00:33:02 lost in clip sorry for the noise Feb 13 00:33:53 Harvy: did you install all of the prereqs for bb and OE? Feb 13 00:34:25 * RP suspects that somehow .dev and .oz354fam083 have got confused Feb 13 00:34:42 the ozfam branch should not be trying to resolve runtime depends Feb 13 00:35:40 I think so, .dev got further at least i got a gcc-cross.. Feb 13 00:37:32 * poli didn't know it would take so long to get gcc-cross. Feb 13 00:40:31 hrm.... gtk-webcore metadata needs some love Feb 13 00:40:51 host env is amd64 if that helps Feb 13 00:44:30 Harvy: talk to shadows Feb 13 00:44:43 Harvy: amd64 won't fully build much ATM Feb 13 00:46:39 at last, do you have anybody who is Feb 13 00:46:59 heh heh, amd64 Feb 13 00:47:07 running anything like me for a host Feb 13 00:47:19 Harvy: gcc3 on amd64 build host for crosscompile is fubar, that much i know Feb 13 00:47:48 i've got an athlon64 3000xp+ box here Feb 13 00:48:29 yess i spent a year working on x86_32 to 64 Feb 13 00:50:15 I will just nuke a partition and put a 32 linux on if that will Feb 13 00:50:18 help Feb 13 00:50:35 my preference is slack Feb 13 00:50:36 it depends on how you want to help Feb 13 00:50:49 i am focused on making amd64 a viable build platform Feb 13 00:51:00 it's cheap, and lightning quick Feb 13 00:51:24 tel me about it Feb 13 00:51:41 i have a aspire 1520 Feb 13 00:52:03 just a bit warm tho Feb 13 00:52:33 and overheats on XP64 installs Feb 13 01:01:23 shaddows Feb 13 01:01:31 'night all Feb 13 01:01:40 where are you based Feb 13 01:02:20 if amd64 can e made a good build platform Feb 13 01:02:45 it would help Feb 13 01:04:03 ttfn will log Feb 13 01:09:15 03rpurdie 07org.oe.dev * r5cf1d188... 10/packages/expat/ (expat-2.0.0/autotools.patch expat_2.0.0.bb): expat: Add 2.0.0 thanks to Eric Shattow (shadows) (#672) Feb 13 01:09:21 03rpurdie 07org.oe.dev * rd48a9722... 10/packages/cairo/ (libsvg-0.1.4/gcc4_and_expat.patch libsvg_0.1.4.bb): libsvg: Add 0.1.4 thanks to Eric Shattow (shadows) (#675) Feb 13 01:09:25 03rpurdie 07org.oe.dev * rb8bb8279... 10/packages/dpkg/ (dpkg-1.13.11/autofoo.patch dpkg_1.13.11.bb): dpkg: Add 1.13.11 Feb 13 01:12:00 RP: sleep well Feb 13 01:12:58 RP: something tells me your recent changes killed something... Feb 13 01:13:14 RP: lots of "cannot satisfy deps" which didn't happen yesterday.... Feb 13 01:17:46 yay, bugs closed. Feb 13 01:17:47 RP++; Feb 13 01:29:24 JustinP: i changed the DISTRO line to "familiar-unstable" in build/conf/local.conf but i still get the same error Feb 13 01:41:52 CSMan, JustinP was incorrect, you DO need to clear the cache Feb 13 01:44:13 Harvy: i'm located in the Chicago, IL USA area Feb 13 01:44:46 eric_shattow.timezone = CST; Feb 13 01:44:48 :) Feb 13 01:54:52 emte: yeah, i did, i noticed by the timestamp Feb 13 01:56:02 emte: do you know if the url where the files are pulled from are correct? Feb 13 01:56:04 CVS_TARBALL_STASH = "http://www.oesources.org/source/current/" Feb 13 01:56:06 that one Feb 13 01:56:42 put it in your browser and find out if its working Feb 13 01:58:37 k Feb 13 02:03:49 CSMan: yes, that's right. Most things aren't on there Feb 13 02:05:55 I need build tools to make a gtk app Feb 13 02:14:51 CSMan: what I meant is that most of the sources aren't on oesources Feb 13 02:15:11 CSMan: we definately have stuff in OE to build GTK apps (such as GPE) Feb 13 02:15:42 ok Feb 13 02:16:23 JustinP: i did change the line in DISTRO , so I'm supposed to get familiar-unstable displayed at the beginning of the build, right ? Feb 13 02:17:16 I'm not sure, I don't think so , actually Feb 13 02:17:22 but it needs to be that way Feb 13 02:20:40 no it wont Feb 13 02:20:47 it only shows familiar Feb 13 02:42:55 03justinp 07org.oe.dev * r224679d0... 10/packages/efl/edje-utils_0.5.0.023.bb: edje-utils: fix DEPENDS Feb 13 02:50:09 I'm trying to add a .bb for enca (http://trific.ath.cx/software/enca/) and it's configure script wants to compile a couple test programs Feb 13 02:50:33 if I just get rid of the tests it seems to use some default that causes it not to compile Feb 13 02:51:30 johnX: generally the best thing to do in that case is to use AC_CACHE_CHECK or similar so that autoconf stores the result of that test in its configure cache. then obtain the correct results for arm and put them in the site/arm-linux (or whatever arch/os you're on) Feb 13 02:51:49 ok, I'll look at that Feb 13 02:51:59 OE sets autoconf's CONFIG_SITE variable. essentially handing it prepopulated test results for the target in question so it doesnt have to run the test Feb 13 02:52:39 ah, great Feb 13 02:52:51 kergoth: I'll have to remember that one Feb 13 02:53:30 also note that some tests can just be not run when crosscompiling. generally sanity checks and consistency checks, the rare case where the result of the test is never used for anything but making sure things are sane Feb 13 03:10:01 i don't know what to do ... i keep getting something like this : Feb 13 03:10:03 http://rafb.net/paste/results/Y0Heh262.html Feb 13 03:10:11 i won't be able to build anything this way =( Feb 13 03:12:36 I don't know if I know enough to help...but I could take a look at your local.conf if you paste it up there Feb 13 03:13:04 ok man Feb 13 03:15:00 http://rafb.net/paste/results/TgJjGj66.html Feb 13 03:15:55 Hi. anyone here ever work with embedded WIFI? I'm having a hell of a time finding off the shelf components that don't require cardbus or USB Feb 13 03:16:15 what do you mean Feb 13 03:16:22 like minipci? Feb 13 03:16:27 you want to enable wifi using another component Feb 13 03:16:32 ? Feb 13 03:17:02 most devices attach themselves to the USB, PCI, or CardBus/PCMCIA busses Feb 13 03:17:07 yes I have a microcontroller sending data which I'd like to make wireless using wifi, in the absolute smallest space possible Feb 13 03:17:17 oh Feb 13 03:17:26 sounds experimental Feb 13 03:17:40 so you're looking for a serial -> wifi Feb 13 03:17:43 IOW, a pre-programmed chip with a bunch of GPIOs and an antenna are all I think I need, theoretically Feb 13 03:17:48 yes Feb 13 03:18:06 i know there are some units which do serial -> ethernet Feb 13 03:18:10 doesn't the ipaq come with a serial connector Feb 13 03:18:18 the manufacturers had wifi offerings, i did not look at them Feb 13 03:18:26 i have one that does parallel->ethen =P Feb 13 03:18:49 CSMan: what i'm referring to is a full internet services implementation Feb 13 03:18:59 Iron_Monk: you've looked at the gumstix projects? Feb 13 03:19:06 not yet... Feb 13 03:19:07 ok Feb 13 03:20:53 if you're not up to integrating a wireless controller and running a uClinux kernel with wifi support on it, you could add a 2nd computer to the design Feb 13 03:21:16 err Feb 13 03:21:22 i'm not sure about uClinux and wifi Feb 13 03:21:24 that was a guess Feb 13 03:22:02 Iron_Monk: why not integrate the wireless chipsets onto your design? Feb 13 03:22:05 this device really only needs to send UDP packets Feb 13 03:22:18 shadows - that's what I'd like to do. Feb 13 03:22:22 hmm Feb 13 03:22:31 i don't know nearly anything about that Feb 13 03:23:14 I swear, WIFI chips must be proprietary ASICS loaded with the interface in firmware. I simply cannot find a "widget" WIFI µc! Feb 13 03:23:17 does wifi have to be integrated, can the client connect your device using an AP bridge? Feb 13 03:23:49 or just implement miniPCI Feb 13 03:23:58 and use miniPCI wifi adapters Feb 13 03:24:11 Isn't that cardbus? PCMCIA? Feb 13 03:24:15 no Feb 13 03:24:18 it's miniPCI Feb 13 03:24:21 a PCI bus Feb 13 03:24:47 ok I'll look into that... I have some other requirements I'll check for Feb 13 03:24:51 laptop computers, commodity router designs, most embedded wifi implementations for sale use these Feb 13 03:25:22 if you need an example you could look at the WRAP reference design, or the MESH CUBE 3G (meshcube.org? i think) Feb 13 03:27:22 johnX: did you anything unusual or incorrect in my local.conf Feb 13 03:27:35 or anybody, please Feb 13 03:27:36 http://rafb.net/paste/results/TgJjGj66.html Feb 13 03:27:40 CSMan: you're missing a verb Feb 13 03:27:40 CSMan: Nothing looks wrong to me... Feb 13 03:27:48 *notice Feb 13 03:27:49 lol Feb 13 03:28:04 yeah, some words don't make it all the way to my fingers Feb 13 03:28:18 let's hope they make it to the config files Feb 13 03:28:28 true Feb 13 03:28:51 i can't even get a silly package to build Feb 13 03:28:52 CSMan: but like I said I'm not sure if I'd spot it even if something was wrong Feb 13 03:29:14 the one thing I can thing of is to double a triple check your paths Feb 13 03:29:22 yeah, i'm doing everything that says in the GettingStarted page Feb 13 03:31:33 Hrm, miniPCI is only slightly smaller than cardbus Feb 13 03:31:41 CSMan: Could I see /tmp/storage/tmp/work/i386-freebsd5/quilt-native-0.42-r0/temp/log.do_patchcmd.87628 Feb 13 03:34:19 ok Feb 13 03:35:24 http://rafb.net/paste/results/L46aQu68.html Feb 13 03:35:36 that's all is there Feb 13 03:36:24 CSMan: I'm at a loss Feb 13 03:36:38 yeah, I as well Feb 13 03:36:42 CSMan: Which is not to say that someone smarter than me couldn't figure it out... Feb 13 03:36:53 i have no idea... Feb 13 03:36:57 CSMan: start with putting all your build time exports and things into a file Feb 13 03:36:59 but i think it has to do with quilt Feb 13 03:37:01 i call mine "env-vars" Feb 13 03:37:08 and i use bash command source env-vars Feb 13 03:37:11 before building Feb 13 03:37:18 this way it is in plain sight Feb 13 03:37:26 you mean BBPATH and PATH ? Feb 13 03:37:29 yes Feb 13 03:37:41 in 1 script invoking the package to be built ? Feb 13 03:37:54 http://oz.pastebin.com/552151 Feb 13 03:37:56 liek BBPATH=tadada PATH=tadada bitbake nano in the same script Feb 13 03:38:04 heh, oz Feb 13 03:38:25 the zaurus has way more support than the ipaq Feb 13 03:38:33 but i chose ipaq because it's fast Feb 13 03:44:22 neither vanille.de or ewi seems to be working... Feb 13 03:49:24 CoreDump|afk: grrrrr Feb 13 03:49:32 CoreDump|afk: my changes to efl.bbclass got reverted Feb 13 03:51:29 ok I found an 802.11b WIFI family of chips, all I need now (I think) is an RF module? Feb 13 03:52:54 btw, is it possible to play around RFs of a wifi device ? Feb 13 03:53:16 like i don't know, snoop on cellular phones ? Feb 13 03:54:45 CSMan: Nope. Feb 13 03:55:16 Unless you've got a software-defined radio.. it's impossible to do that kind of thing. Feb 13 03:56:55 NA|Zzz: so a wifi device has it's frequency level hardwired Feb 13 03:57:44 not really Feb 13 03:57:51 it operates by spec Feb 13 03:58:06 which means it utilizes spec freqs Feb 13 03:58:20 and that usually means a range Feb 13 03:58:44 and that range is definitely out of what its standard specifies, right ? Feb 13 03:58:51 *NOT out Feb 13 04:01:13 CSMan: Wifi devices are engineered to operate at 2.4 or 5.8GHz, depending on spec.. and the modulation/demodulation's (usually) done in hardware.. means, unlses you've got data coded in the corerct modulation method at those specific freqeuncy bands... you can't do much with it. Feb 13 04:01:39 GSM mobile phones operate at 850/900/1800/1900MHz, which is way off the 2400 or 5800 for Wifi. Feb 13 04:02:18 Add to that that GSM phones are using encrpytion between themselves and the tower (and, of course, the legal implications of listnening in)... it's highly unlikely, and seriously unwise. Feb 13 04:02:50 heh...back in the bad old days of analog celluar listening in was possible Feb 13 04:03:28 Of course, I'm not qualified to giev legal advice etc etc.. but, yeah, it's certainly illegal in many jurisdictions, and (IMHO) unethical. Feb 13 04:03:42 johnX: Yeah, a long time (here at least) since teh AMPS mobile networks have operated Feb 13 04:03:53 Damn I can't type properly with this lag Feb 13 04:07:24 NA|Zzz: interesting points you have Feb 13 04:08:02 uhmm... cordless phones operate at 2.4+ Ghz =P Feb 13 04:08:16 and that is pretty analog, eh? Feb 13 04:08:17 I thought you were referring to cellular telephones Feb 13 04:08:24 NA|Zzz: yeah, i was Feb 13 04:08:31 but then i thought about frequencies Feb 13 04:09:22 shadows: i'm bitbakin the #!/usr/local/bin/bash way Feb 13 04:09:28 shadows: let's hope it works Feb 13 04:09:33 o_O Feb 13 04:10:46 There are seome that operate at 2.4GHz, but, again, you get the problem where the wifi cards are not designed to receive analog transmissions.. IIRC, wifi uses PSK and CCK.. both which wouldn't really assist you with receiving analog transmissions from cordless telephones. I'm not sure what DECT phones use, but I'm prertty sure they're not 2.4GHGz Feb 13 04:11:31 o-kay Feb 13 04:11:48 who wants to look at gtk-webcore with me and fix bugs? Feb 13 04:11:56 i see some volunteers..... anyone? Feb 13 04:12:03 * CSMan steps back Feb 13 04:12:05 =P Feb 13 04:12:25 shadows: how much machine power you need for that ? Feb 13 04:12:35 none at all Feb 13 04:12:38 i have 233Mhz of full power =P Feb 13 04:12:51 (no, it's not a PDA) Feb 13 04:12:54 more brainpower than anything Feb 13 04:13:08 what is that package for? Feb 13 04:13:13 http://bugs.mysql.com/bug.php?id=6554 describes the bug i'm running into Feb 13 04:13:37 and the package i'm seeing it with is osb-browser in org.openembedded.dev/packages/gtk-webcore/* Feb 13 04:13:59 trouble begins when you are using gcc4 as the cross initial compiler Feb 13 04:14:32 osb-browser ? is it openSSL ? Feb 13 04:14:38 so for the moment, the problem is defined as static variables defined inside of functions are causing link and build failures when compiled using gcc4 Feb 13 04:15:25 the method to fix this problem is defined as look for instances where static variables are defined inside of functions, and attempt to rework the code to define static variables instead at source file scope Feb 13 04:15:39 static variables defined inside Feb 13 04:15:45 maybe you have to declare them before ? Feb 13 04:16:03 shadows: Ew, sounds like the work I'm doing at present... reworking some seriously dodgy code with 700+ global gvars. Feb 13 04:16:05 the code i've seen assumes gcc3 specific behavior Feb 13 04:16:12 NA|Zzz: yuck ;) Feb 13 04:16:26 iirc , for gcc3 you declare them before using them in the functions Feb 13 04:16:39 yeah that's how you're *supposed* to do it Feb 13 04:16:57 this code is ported from Apple safari webcore Feb 13 04:17:05 shadows: why don't you pull that file and try to compile it alone , or replicate the same situation Feb 13 04:17:10 It's so tempting to wander downstairs and LART the original developers. Feb 13 04:17:17 naturally Apple programmers think they're hot stuff, and do not need to write proper code Feb 13 04:17:45 shadows: it's the same everywhere else, because they believe they are doing the *hot* stuff Feb 13 04:17:57 debian, kde, gnome Feb 13 04:18:14 anyhow, there's bug reports in OE about this Feb 13 04:18:17 less talk, more code :) Feb 13 04:18:29 did you try to isolate the problem Feb 13 04:20:59 CSMan: follow along with the bugreports ;) Feb 13 04:29:07 fixed. Feb 13 04:29:12 * shadows looks happy Feb 13 04:29:31 okay, who wants to take bets on the next package to fail with gcc4 ? Feb 13 04:29:38 >:-) Feb 13 04:30:50 Doesn't appear to affect OPIE, so I'm good ;) Feb 13 04:31:06 OPIE builds with gcc4 cross initial? Feb 13 04:31:16 03justinp 07org.oe.dev * r584bf1a4... 10/classes/efl.bbclass: efl.bbclass: fix changes reverted by a bad merge Feb 13 04:31:47 shadows: I've got gcc4 on the host.. Feb 13 04:32:12 no no, i mean setting the preferred gcc cross and gcc cross initial to be gcc 4.0.2 Feb 13 04:32:20 regardless of the compiler on your host machine Feb 13 04:32:31 that is what i'm doing Feb 13 04:34:54 Ahh Feb 13 04:35:07 Didn't think about the xcompiler.. too early in the day Feb 13 04:38:35 final run/rebuild before i submit this stuff to the bugzilla Feb 13 04:42:19 damn this crap! Feb 13 04:43:48 YES DAMN IT TO HELL! Feb 13 04:44:17 I'm going nuts looking for RF modules! Feb 13 04:44:47 I found the *perfect* device. It has no datasheets and apparently isn't for sale in the US> Feb 13 04:47:34 Iron_Monk: care to elaborate? Feb 13 04:52:57 I'm seeing about ordering a sample Feb 13 04:53:03 RF module. Feb 13 04:53:28 ok, forget this, dependencies in dev are screwed up Feb 13 04:53:39 http://bugs.openembedded.org/show_bug.cgi?id=684 is ready for review Feb 13 04:53:40 shadows: thanks a lot for the expat...it broke a bunch of my builds now... Feb 13 04:53:42 something to which I may hurl a crapload of data at about 9 Mbps and have it reliably spit that out into the 2.4 gHz rance Feb 13 04:53:43 range Feb 13 04:53:49 JustinP: hah, what's wrong? Feb 13 04:53:55 JustinP: i'm willing to help fix that Feb 13 04:54:01 shadows: in that it's libexpat1 and my builds depend on libexpat0.... Feb 13 04:54:17 your builds need to be updated, no big deal. sorry about that Feb 13 04:54:22 I know Feb 13 04:54:26 and I have no idea which ones they are Feb 13 04:54:36 wait a second, doesn't that mean the expat build is broken? Feb 13 04:54:36 * JustinP greps Feb 13 04:54:47 I think it's part of the debian naming thing Feb 13 04:55:05 if your builds depend on libexapt0, then the previous version should be built instead no? Feb 13 04:55:14 i think i screwed it up if that's the case Feb 13 04:56:32 no Feb 13 04:56:40 they were build with expat 1.95 Feb 13 04:56:44 then the new one got built Feb 13 04:56:51 and it replaced the old one in the feed... Feb 13 04:56:56 I guess... Feb 13 04:57:05 hm Feb 13 04:57:17 hmmm...actually, there's nothing in th emorgue... Feb 13 04:57:35 if you run into an example of how this breaks, and what might be done about it... i'd like to jump in and help out with that Feb 13 04:57:59 shadows: just build a gpe image and you'll see it Feb 13 04:58:00 :-/ Feb 13 04:58:15 if i can get that far ;) Feb 13 04:58:25 i just resolved the osb-browser problem Feb 13 04:58:33 dbus-1 Feb 13 04:58:37 the last person who tried gcc4 got to this point, and gave up Feb 13 04:58:42 gconf-dbus Feb 13 04:59:13 gtk+.... Feb 13 04:59:15 geez Feb 13 04:59:49 okay Feb 13 04:59:53 resuming the gpe-image build Feb 13 05:00:53 it looks like trying to get custom wireless is a highly specialized endeavor. Feb 13 05:01:09 grrrr...I have no idea what's going on Feb 13 05:01:20 holy cow bell, bat man. gpe-image is building! it's at do_rootfs Feb 13 05:01:22 yahho! Feb 13 05:01:24 probably have to start form scratch again...of course who knows if that would even help... Feb 13 05:01:46 JustinP: yes, if you started with a tmp that was before the expat commit... and you updated Feb 13 05:01:52 it's time to rm-fr oetmp Feb 13 05:03:05 damn it Feb 13 05:03:13 god damn pain in the ass Feb 13 05:03:29 * JustinP twitches Feb 13 05:03:47 hi all. Feb 13 05:03:59 * shadows launches a box of camisole tea at JustinP's general direction "quick! before it gets ugly" Feb 13 05:04:04 RP: did you try kexec, yet? I tried loading the branch kernel2.4 and all is get is a kernel panic. Do I need to empty the kernels CMDLINE or something? Feb 13 05:04:52 rm -rf tmp && cp -R gcc-cross.tmp tmp Feb 13 05:05:08 could the OE build for MIPS? Feb 13 05:05:18 oh damn...I should have tested my other thing first... Feb 13 05:05:20 oh well Feb 13 05:05:25 * JustinP switches to stable Feb 13 05:06:10 woowowww..... Feb 13 05:06:13 gpe-image has built Feb 13 05:06:15 * shadows dances Feb 13 05:06:29 RP: pinggg---aggge Feb 13 05:06:41 RP: gcc4 cross initial now builds gpe-image Feb 13 05:06:48 * CoreDump|afk congratulates shadows on a job well done Feb 13 05:07:50 * JustinP congratulates shadows as well Feb 13 05:08:32 * CoreDump|afk heads to work Feb 13 05:09:29 now time for other build targets Feb 13 05:09:54 opie-image will be in-tarrr-eest-eeng Feb 13 05:10:22 * JustinP listens to his hard drive crank Feb 13 05:17:34 * Iron_Monk tosses a rock at global specifications Feb 13 05:21:33 people are always talking about making miscellaneous devices wireless. HOW ARE THEY DOING THIS Feb 13 05:22:16 they use chucknorrisfacts and go killing. Feb 13 05:52:18 hey shadows , you made it Feb 13 05:52:28 yeah Feb 13 05:52:33 it wasn't so bad Feb 13 05:52:38 everybody make it, and i can't even get bitbake nano to work =( Feb 13 05:55:24 CSMan: did you set up your BBPATH? Feb 13 05:55:48 yeah, i put everything in a nice script Feb 13 05:56:22 #!/usr/local/bin/bash Feb 13 05:56:22 export BBPATH=/tmp/storage/build/:/tmp/storage/org.openembedded.dev/ Feb 13 05:56:22 export PATH=/tmp/storage/bitbake/bin:$PATH Feb 13 05:56:22 bitbake nano Feb 13 05:56:27 um Feb 13 05:56:31 why a script? Feb 13 05:56:34 that doesn't make sense Feb 13 05:56:40 you would not want it to be in a script Feb 13 05:56:46 someone suggested it Feb 13 05:56:54 in fact if you had read what _I_ suggested... Feb 13 05:56:59 i think it's pretty much the same as entering it Feb 13 05:57:10 no Feb 13 05:57:18 executing a script spawns a new instance of bash Feb 13 05:57:31 yeah, that's why i have those exports Feb 13 05:57:32 sourcing a file runs it as if you just typed it in yourself Feb 13 05:57:36 there's a marked difference. Feb 13 05:58:03 i thought you meant how the paths are handled, not how the system allocates processes =) Feb 13 05:58:07 what i suggested was to do what i do, which is use bash and source a file containing the directives, which i gave an example of Feb 13 05:58:51 ok Feb 13 05:58:57 how do you source a file again ? Feb 13 05:59:57 03justinp 07org.oe.dev * r9458ed3a... 10/conf/distro/preferred-e-versions.inc: preferred-e-versions: add PREFERRED_VERSIONs for gnome-vfs-dbus and gcc Feb 13 05:59:59 source filename Feb 13 06:01:02 ok, i did Feb 13 06:01:39 never used source Feb 13 06:01:49 remove the "bitbake nano" and #! Feb 13 06:02:00 source the setup, then run bitbake Feb 13 06:02:06 bitbake -i is your friend Feb 13 06:02:21 right Feb 13 06:02:29 it's ok using the cache this time Feb 13 06:02:33 bitbake -i was rather helpful in development Feb 13 06:02:34 it'll take 5 minutes Feb 13 06:02:48 * shadows extracts gcc4 cross compiled gpe-image to his Z Feb 13 06:02:53 i just need ONE to build so i can continue building other packages with -i Feb 13 06:03:57 i'm not nearly impressed enough with CSMan's hardware specs to take time to give him a walk-thru Feb 13 06:04:04 * shadows jokes Feb 13 06:04:47 alright, I finally got enca to build, just need to sort out packaging for it...then onto fbreader Feb 13 06:05:32 * JustinP starts an oz354fam083 build Feb 13 06:07:28 heh Feb 13 06:07:47 it'll be like day 1: bitbake this, day 2: bitbake that Feb 13 06:11:52 'night Feb 13 06:12:26 * JustinP has a rev ready for oz354fam083 for the new e stuff...just have to test... Feb 13 06:15:58 91% Feb 13 06:16:45 it seems to work Feb 13 06:17:50 same error. Feb 13 06:17:55 * CSMan has a headache Feb 13 06:18:07 tomorrow will be another day Feb 13 06:18:16 CSMan: mate, do you have a faster build machine? Feb 13 06:19:00 yeah, my ipaq has a 400Mhz processor Feb 13 06:19:09 heh, heh. Feb 13 06:19:13 =P Feb 13 06:19:30 i actually built some packages using 0.8.2 tools and NFS Feb 13 06:19:48 okay, who's awake and willing to commit some patches? i've got bug reports ready Feb 13 06:19:51 but I needed to blow from time to time to don't let the cpu fry Feb 13 06:21:08 CSMan: in what country do you reside? Feb 13 06:21:23 reside, that'd be .ca Feb 13 06:21:42 spare pc, anyone ? Feb 13 06:23:02 good nite. Feb 13 06:23:12 'night Feb 13 06:23:16 better luck tomorrow Feb 13 06:24:15 nope, sqlite is borked Feb 13 06:24:21 lame Feb 13 06:24:32 totally. Feb 13 06:24:42 worst, assumption, evah! Feb 13 06:25:26 it's not like your work will go to waste...I imagine that transitioning to gcc 4 is the long term goal anyways... Feb 13 06:25:42 gpe-confd: /home/jnc/opensource/openzaurus/build/oetmp/work/arm5vte-linux/sqlite-2.8.16-r1/sqlite-2.6.18/src/btree.c:702:sqliteBtreeOpen: Assertion `sizeof(ptr)==sizeof(char*)' failed. Feb 13 06:25:45 Aborted Feb 13 06:26:04 ~lart worst assumption evah Feb 13 06:26:05 * ibot cats /dev/urandom into worst assumption evah's ear Feb 13 06:26:42 hey maybe you guys can help Feb 13 06:26:53 i don't know much about non-x86 platforms and pointer sizes Feb 13 06:28:15 is '-DINTPTR_TYPE int' appropriate for our targets? Feb 13 08:45:59 anyone with commit access awake? Feb 13 08:47:14 http://bugs.openembedded.org/show_bug.cgi?id=540 the sqlite bb patch needs love Feb 13 08:47:25 i've tested it, and works right this time. Feb 13 09:33:36 hey Feb 13 09:33:43 hi Feb 13 09:36:11 shadows: I can't reach bugs.openembedded.org :-P Feb 13 09:36:24 NAiL: add it to your hosts file Feb 13 09:36:29 aha Feb 13 09:36:36 use the IP of bugs.treke.net Feb 13 09:36:57 192.216.230.225 bugs.openembedded.org Feb 13 09:40:15 ok, so the attachment to bug 687 is the one you want committed then? Feb 13 09:40:35 use bugs.treke.net instead Feb 13 09:40:36 morning Feb 13 09:40:45 NAiL: yes, please. i've thoughouly tested it Feb 13 09:41:02 pardon my limp spelling :( Feb 13 09:41:07 * NAiL takes a look Feb 13 09:41:31 and you have a x64 host, I presume? Feb 13 09:41:33 yes Feb 13 09:41:35 (just skimmed the first bug) Feb 13 09:41:52 * NAiL logs into x64 build host ;) Feb 13 09:42:34 the sqlite code is so horrible, i looked for the least-impact way to make this work, and also so we could change it from bb to bb Feb 13 09:45:15 brb Feb 13 09:45:21 also i will be submitting a version bump shortly Feb 13 09:45:32 .17 of sqlite is out Feb 13 09:45:36 aha Feb 13 09:45:44 and it still has the same problem? Feb 13 09:46:40 yeah, that retard logic is well documented in the sqlite sources and seems that they expect you to use a commandline -D'FOO_TYPE=keyword' if crosscompiling Feb 13 09:46:51 *sigh* Feb 13 09:47:32 A comment in the .bb on why the -D is there would be nice though. Just in case someone gets a hold of a 64bit pda :-P Feb 13 09:47:42 the upgraded apm[d] packages (3.2.2-r6), are they only for the >=OZ3.5.4 releases? Feb 13 09:47:42 agreed Feb 13 09:48:00 i tried stuffing a comment in there, but was not able to figure a good place for it Feb 13 09:48:03 morning all Feb 13 09:48:09 IMO having it documented in the bug tracker is enough Feb 13 09:48:22 can search for a closed bug report pretty easily Feb 13 09:49:17 and, if you're clever, that line should raise some alarm bells in ones mind when reading it Feb 13 09:49:40 versus say a patch to the make.mk outputting a value to config.h which did not exist previously. Feb 13 09:49:42 ..since apmd -r6 crashes/TERMs when the Z suspends Feb 13 09:50:06 accumulator: not sure Feb 13 09:51:09 NAiL: if there's a 64-bit target, we could add logic to the bb file Feb 13 09:51:20 wasn't able to get any messages from apmd why it crashed though Feb 13 09:51:37 shadows: good point Feb 13 09:52:30 NAiL: i've successfully built (on amd64 host) the gpe-image target using gcc4 4.0.2 cross and cross initial Feb 13 09:53:04 nice. Far from the targets I'm building though ;) Feb 13 09:53:25 next up is opie-image, and e-image Feb 13 10:38:04 <_law_> hi Feb 13 10:38:20 <_law_> any ideas why my gpe-image doesnt build? http://pastebin.com/552379 Feb 13 10:38:31 <_law_> libgpg-error is available Feb 13 10:39:07 you need to inspect the logfile to find the real error Feb 13 10:39:20 ignore the output from the log_check thing, it is bogus Feb 13 10:40:24 i'd appreciate if someone could look at http://bugs.openembedded.org/show_bug.cgi?id=688 Feb 13 10:40:28 it's ready for review. Feb 13 10:40:37 http://www.hrw.one.pl/2006/02/13/two-years-with-zaurus/ Feb 13 10:40:49 * shadows reads Feb 13 10:40:56 <_law_> http://pastebin.com/552379 is from the logfile Feb 13 10:41:40 not the correct logfile Feb 13 10:41:51 log.do_rootfs.23440 is the file you need to look at Feb 13 10:44:52 <_law_> pb__, http://helmsdeep.homelinux.net/~pat/log.do_rootfs.23440 Feb 13 10:49:23 if someone with commit access can please look at http://bugs.openembedded.org/show_bug.cgi?id=685 also, i've tested that fairly well Feb 13 10:54:22 _law_: well... Feb 13 10:54:23 ERROR: Cannot satisfy the following dependencies for gpe-task-settings: Feb 13 10:54:23 libipkg0 (>= 0.99.155) Feb 13 10:54:30 I'm guessing that this is the error. Feb 13 10:56:00 <_law_> pb_, ok this is the error thanks Feb 13 10:57:40 when i use the dhcp3 client no client script es be installed Feb 13 10:57:51 have i just forgot a package? Feb 13 10:58:00 or is it really missing? **** ENDING LOGGING AT Mon Feb 13 10:59:56 2006