**** BEGIN LOGGING AT Sun May 22 23:59:57 2005 May 23 00:20:11 hehe May 23 00:20:13 mythtv + wireless laptop == fun May 23 00:20:18 cleaning fun May 23 00:25:22 someone must be merging a kernel tree May 23 00:26:01 treke|home: CIA has been down for days, its probably still processing its backlog May 23 00:26:01 heh May 23 00:26:07 well, somehwat down anyway May 23 00:26:17 it looks like someonke is pushing a pile of bk changes May 23 00:26:18 : May 23 00:26:26 oe ones that is May 23 00:26:27 : kergoth * r1.2843.407.1 jonslug/packages/genext2fs/ (3 files): No longer apply the bytes per inode patch to genext2fs cvs, since it was applied upstream, and switch from cvs to the first prerelease of 1.4. May 23 00:26:30 i didnt push that May 23 00:26:52 * kergoth hugs CIA-4 May 23 00:27:04 * CIA-4 hugs kergoth May 23 00:35:03 hehe May 23 00:39:32 luke-jr_: pong May 23 00:39:37 morning all May 23 00:47:52 morning May 23 00:50:17 hi XorA RP May 23 00:50:19 morning XorA May 23 01:11:36 morning May 23 01:11:46 helloing all May 23 01:12:10 hi hrw May 23 01:14:49 ;( build failed May 23 01:15:16 ljp: did not pass opie-image? May 23 01:15:36 not even m4 May 23 01:19:10 hi mickeyl May 23 01:20:53 hey hrw|work May 23 01:20:56 moin mickeyl May 23 01:21:45 ljp: may be paste your error again.. today there is more gurus :) May 23 01:23:36 morning May 23 01:24:21 difrferent error now :) May 23 01:24:22 good morning pb_ May 23 01:24:45 m4-1.4.2.tar.gz seems to be awol May 23 01:25:13 mickeyl: good morning! May 23 01:25:15 hi ljp May 23 01:26:31 hi May 23 01:26:53 got my quickcam working on the c3000 May 23 01:27:01 cute May 23 01:31:27 framerate sucks May 23 01:35:07 If i delete any files from my local OE directory and do a bk pull, will they be restored? May 23 01:35:25 No. May 23 01:35:55 Ah May 23 01:36:13 ~nslookup www.oesources.org May 23 01:36:19 www.oesources.org is 192.216.230.225 May 23 01:36:33 Is it better to delete the entire OE dir and clone a new one then? May 23 01:36:51 Kristoffer: DONT DO THAT May 23 01:36:52 Kristoffer: try "bk -r co -q" May 23 01:36:53 you can do a bk co, right? May 23 01:36:59 ah :) May 23 01:37:13 Kristoffer: sorry hit caps by accident, I havent been able to clone at all recently May 23 01:38:41 If you deleted the files from your working tree, yeah, "bk co" should bring them back. If you deleted them from your local repository, that may not be recoverable: you might have to reclone in that case. May 23 01:39:29 Oki, thx May 23 01:39:41 or, if you mean that you did "bk rm" by mistake, you can use "bk unrm". May 23 01:41:03 I need to study up on the bk commands methink May 23 01:42:12 hmm.. nuclear battery in PDA.. May 23 01:45:47 morning May 23 01:48:47 For the 3 last days, patching stage dies with quilt barfing (cp on a readonly/missing path). May 23 01:49:16 oe problem or bitbake ? May 23 01:49:22 morning May 23 01:50:04 dunno, i don't use quilt May 23 01:50:06 i recommend patcher May 23 01:50:13 echo 'INHERIT += patcher' >> local.conf May 23 01:50:20 I use quilt May 23 01:50:40 proti: almost certainly oe May 23 01:51:01 pb_: OK. Will look on the appropriate channel then (oe ?). May 23 01:51:31 if this is the "permission denied" thing, that's a quilt bug being tickled by an earlier patch failure. May 23 01:51:43 you'll need to clean the offending package and re-build to see the original error. May 23 01:52:02 OK. That is the thing. Except it happened on a brand new build. May 23 01:52:26 Barfed on the gnu-config package patching. May 23 01:52:48 This was the very first time you'd ever run bitbake in that build directory? May 23 01:52:50 That does seem a bit strange. May 23 01:53:29 I guess you'll have to inspect the logs to find out what's going on, in that case. May 23 01:54:01 No. Trying to build the gcc-cross. It went to the glibc package then barfed for the patch. Then did bk pull. And nothing worked anymore. May 23 01:54:05 I'll have a look. May 23 01:54:36 My first inspection in the log didn't show anything (log empty if I remember). May 23 01:54:39 Ah, so it wasn't actually a brand new build? May 23 01:54:47 pb_: same thing I saw. first build May 23 01:55:05 ljp: you had the permission denied on the very first run? May 23 01:55:18 normally you get a patch failure on the first run, and permission denied on all subsequent attempts. May 23 01:55:38 not sure of the first error. tried it a few times May 23 01:55:49 pb_: Here is what I did. rm -rf of the TMP. bitbake gcc-cross. And then it barfed on the gnu-config package. May 23 01:56:09 managed to get up to m4 May 23 02:00:58 proti: what CVSDATE? May 23 02:01:45 hrw|work: No CVSDATE. A fresh copy from the bundled local.conf.sample targeted to openzaurus collie. That's all. May 23 02:03:13 aha May 23 02:03:37 A bitbake question : Why do we have to compute the world dep everytime we build the dep graph ? I removed the code doing it and nothing bad happened (as suspected). I put it in a build_depgraph_world function and called it if world is given as target. May 23 02:08:19 03koen 07 * r1.3370 10openembedded/packages/ (5 files in 4 dirs): May 23 02:08:19 handhelds-sa-2.6_cvs.bb, qte_2.3.10.bb: jornada720 -> jornada7xx May 23 02:08:19 devices, tslib.sh, ts.conf: mvdir jornada720 -> jornada7xx May 23 02:10:00 good morning all May 23 02:10:49 proti: try to set CVSDATE_gnu-config-native = 20050510 May 23 02:11:12 morning koen May 23 02:11:43 robert___: I will. thanks. May 23 02:11:53 koen: morning. May 23 02:12:16 hey proti & mickeyl May 23 02:13:45 proti : or some older CVSDATE . 20050508 .. May 23 02:15:00 * koen points people to http://bugs.openembedded.org/show_bug.cgi?id=2 May 23 02:15:16 recent tree -> recent gnu-config May 23 02:15:26 old tree -> old gnu-config May 23 02:16:05 koen: How old is an "old" tree ? May 23 02:16:29 a week or so May 23 02:17:57 proti: see the bugreport when the cset was pushed May 23 02:17:58 I did a bk pull on friday. So I should get a "new" gnu-config. What kind of new gnu-config ? The head, the 20050510, 20050508 ? May 23 02:18:18 todays gnuconfig should work May 23 02:18:25 koen: bugs.openembedded.org seems down. May 23 02:19:35 20050521 builds ok over here May 23 02:20:34 mickeyl: I sent bero an email that he should contact you or kergoth if he wants commit access May 23 02:20:54 proti: try what koen said May 23 02:25:02 robert___: I'm pulling now. Will give a status in a few minutes. May 23 02:29:02 booohoo.. no m4 May 23 02:29:16 koen: right May 23 02:30:32 ljp: the best opinion about m4 I encountered: "looking at m4 macros makes my eyes bleed" May 23 02:31:29 i dont wanna look at it, just download it :) May 23 02:31:55 :) May 23 02:32:05 :) May 23 02:36:48 re May 23 02:39:07 ~seen zecke May 23 02:39:10 zecke <~ich@83-169-170-199-dynip.superkabel.de> was last seen on IRC in channel #oe, 9h 56m 29s ago, saying: ':}'. May 23 02:39:39 hey hrw May 23 02:39:46 ~seen zecke_ May 23 02:39:48 zecke_ <~ich@83-169-170-199-dynip.superkabel.de> was last seen on IRC in channel #oe, 4d 10h 42m 34s ago, saying: 'cya tomorrow'. May 23 02:40:03 bitbake-qa needs some tweaks May 23 02:40:03 looks like ibot is a bit off May 23 02:41:40 * CoreDump|home flashed an image from this night and hopes the sdcontrol / busybox badness was fixed May 23 02:43:07 koen: do you know how to disable CVS_TARBALL_STASH for bitbake-qa? or SOURCEFORGE_MIRROR? it looks like setting it in testrun.conf has no meaming.. May 23 02:43:46 CVS_TARBALL_STASH = "" in a local.conf May 23 02:44:02 somewhere after bitbake.conf get's parsed May 23 02:45:00 I had only testrun.conf for it May 23 02:45:49 CVS_TARBALL_STASH should only be in local.conf May 23 02:45:55 no idea why qa picks it up May 23 02:47:18 yep May 23 02:50:21 hmm.. now bitbake-qa tries to rm archive after fetch.. May 23 02:53:02 interesting May 23 02:53:29 but only those which cannot be fetched... May 23 02:53:32 it looks like the new "quiet" kernel has some side-effects regarding busybox May 23 02:53:48 so its OK - it remove *.md5 then May 23 02:53:59 kernel-launched scripts are f*cked May 23 02:58:21 If anyone is interested, I've made .bb files for oe to enable Japanese input on GPE. May 23 02:58:50 b0ti|work: cool May 23 02:59:49 ok. noone maintain util-linux so I can clean ithere May 23 03:00:49 I couldn't test it, because my GPE installation is broken, but it should work. bitbake files are here: http://siliconium.net/oe-jpinput/ May 23 03:01:50 dejavu fonts 1.10 released.. May 23 03:02:49 its takes me a while, but ive finally remebered having a c compiler helps oe builds no end May 23 03:02:58 hi all, anybody help out with a couple of newbie OE building questions? May 23 03:03:50 chocolatetpot: sure, fire away May 23 03:04:09 NOTE: fetch opie.core.settings.button_cvs.handhelds.org__20050521.tar.gz - nice uri.. May 23 03:04:23 I'm getting a Python traceback, from the error AttributeError: 'dict' object has no attribute setVar May 23 03:04:42 chocolatetpot: too old bitbake or wrong BBPATH May 23 03:05:26 bitbake was straight from subversion last friday, so perhaps BBPATH then. It was failing on a "do package" of libpng May 23 03:05:34 DIC_SRC = "${WORKDIR}/../${ANTHYNATIVE}/image/${TMPDIR}/staging/${BUILD_SYS}/share/anthy" May 23 03:05:43 * koen is overwelmed May 23 03:05:52 I had successfully built some stuff already - e.g. nano editor May 23 03:06:15 re May 23 03:06:36 wb hrw|work May 23 03:06:51 Parsing not finished. Note to myself : must do something about non-psyco target that are not top notch machines (like 300mhz dual sparc II marchines). May 23 03:07:07 is here anyone w/ a collie using a recent image? Does sdcontrol work for you? May 23 03:08:05 anyone here need a 5000d for spare parts? ;) May 23 03:08:43 Twiun: what did you break? May 23 03:08:53 BBPATH looks ok to me - points to my build and openembedded folders like in the getting started guide May 23 03:09:04 koen: the ram's dead May 23 03:09:16 koen: fails the internal check May 23 03:09:51 could I be missing some required Python modules? May 23 03:12:26 argh.. tmp/staging/i686-linux/bin/ binaries gives me 'permission denied' ;( May 23 03:12:43 hmm bugs.openembedded.org seems to be down May 23 03:12:58 CoreDump|home: treke machines are down May 23 03:13:03 oesources.org too May 23 03:13:15 ah i c May 23 03:13:28 will file the bug later then ;) May 23 03:14:00 ok, 2nd question - has anyone played with the "sharprom-compatible" distro setting and got it to work. It should pick up the arm-linux-gcc-2.95 compiler, but the first test compile it does it seems to be looking for arm-linux-gcc and then failing (not sure if this is the right irc channel for that question tho - aplogies if not) May 23 03:14:16 or even apologies May 23 03:14:30 CoreDump|home: what bug? May 23 03:15:39 /dev/hda5 on /home/hrw/zaurus/oe/build type reiserfs (rw,noexec,nosuid,nodev,user=hrw) May 23 03:15:42 fsck.. May 23 03:15:44 hrw|work: sdcontrol is f*cked in recent images. I believe it to be a busybox thing. Also, "quiet" kernels break scripts launched by the kernel (like a hacked /sbin/init) May 23 03:16:17 when I switched distro to openzaurus-3.5.2 I was able to build which suggested to me that the rest of my config was ok, but maybe not May 23 03:17:37 CoreDump|home: quiet probably have wrong console set May 23 03:18:47 hrw|work: which would it be for the quiet kernels then? /dev/console? /dev/tty0 works for normal kernels May 23 03:19:16 also, the sdcontrol bug also happens when using a non-quiet kernel of an earlier release May 23 03:21:45 ew wtf May 23 03:21:56 awk b0rks May 23 03:22:37 CoreDump|home: set KERNEL_CONSOLE=tty0 in local.conf and test May 23 03:22:43 hrw: any other ideas, or suggestions of where to start to debug? May 23 03:23:08 hrw|work: this helps with /sbin/init, but not w/ sdcontrol May 23 03:24:38 * CoreDump|home builds collie-kernel-58-6-debug to be sure... May 23 03:25:24 chocolatetpot: sharprom thing? hmm.. May 23 03:25:59 chocolatetpot: you have arm-linux-gcc 2.95.3 version in PATH? May 23 03:26:28 hrw: yep, that gets picked up and used when I build for openzaurus May 23 03:27:04 I also have assume provided filed out in local.conf set to arm-linux-gcc-2.95.3 May 23 03:27:23 bit of a poser May 23 03:28:02 hrw|work: busybox's "test" (as in if test blahh; then whatever ;done) breaks in this image when run by sdmgr. It works as expected when run from the shell manually May 23 03:28:52 in the config.log for the first thing it tried to build it was trying to use arm-linux-gcc (without the 2.95) and failing to compile because of that May 23 03:29:02 I have no idea what would cause this, but I do know that 3.5.4 will be useless on collie if this bug isn't found May 23 03:34:30 * CoreDump|home reverts to a working snapshot for the time beeing May 23 03:36:37 May 23 03:36:38 unified build of util-linux May 23 03:36:38 added util-linux 2.12q May 23 03:36:42 May 23 03:37:50 and removed util-linux 2.12n (not fetchable from upstream) May 23 03:38:46 hrw: spotted an =+ rather than += in the sharpromcompatible.conf - is that a syntax error or just something I'm not familiar with? May 23 03:39:27 chocolatetpot: its ok May 23 03:39:41 =+ add on end, += on beginning May 23 03:39:53 or in reverse May 23 03:40:00 ah cool, ok May 23 03:41:26 i'll keep fiddling, if I find anything I'll post back somewhere - either a bug in me, the howto or oe (in order of probability) May 23 03:43:37 cool, due to new super low memory bb I can now OE on colinux at work May 23 03:46:44 XorA: yay :) May 23 03:47:03 ~praise bb team May 23 03:47:04 All hail bb team! May 23 03:47:24 CIA-4: fixed SRC_URI for zsh 4.1.1 (4.2.5 available upstream) May 23 03:48:51 espgs_7.07.1rc1.bb: point SRC_URI to sf.net May 23 03:50:12 ~lart ftp.easysw.com May 23 03:50:12 * ibot brandishes Excalibur! "With this sword, I vanquish thee, ftp.easysw.com!" and lops off ftp.easysw.com's head May 23 03:51:01 03koen 07 * r1.3370.1.1 10openembedded/packages/espgs/espgs_7.07.1rc1.bb: espgs_7.07.1rc1.bb: point SRC_URI to sf.net May 23 04:05:22 CIA-4: fixed SRC_URI for zlib 1.1.4, removed visibility patch from it (as it is for 1.2.2 zlib) May 23 04:12:09 balsa 2.0.17 is not fetchable May 23 04:17:02 heh.. my machine at home lost connection... May 23 04:18:20 hrw|work: that sucks when that happens May 23 04:18:41 hrw|work: normally happens to me when I overload the router, 300 active connections upsets it May 23 04:20:04 I have to prioritize network setup at home machine but lack time/knowledge May 23 04:28:15 hrw|work: you run debian? May 23 04:28:20 yes May 23 04:28:29 hrw|work: apt-get install wondershaper May 23 04:29:03 hrw|work: wondershaper eth0 (downstream-10%) (upstream-10%) quick and easy method May 23 04:29:29 XorA: thx May 23 04:29:43 hrw|work: its not perfect, but damn site better then nothing May 23 04:30:02 fsck.. jffs2 oopses.. May 23 04:30:28 or other.. too fast and too many of them May 23 04:31:41 ooooh, bitbake trunk is loads faster than yesterday's version May 23 04:32:11 koen: and bitbake-qa not quite run with it May 23 04:32:23 koen: I am massively impressed with bb trunk, its running well on a colinux May 23 04:32:49 this should open OE to a new bunch of board office workers May 23 04:32:56 indeed May 23 04:33:33 koen: you wouldnt think english is supposed to be my first language with my spelling :-( May 23 04:34:09 XorA: I know the feeling... ;-) May 23 04:34:09 it isn't my first language, so I hardly notice :) May 23 04:34:24 XorA: aren't you from 'up north'? May 23 04:34:41 koen: I was born in London, just live in Edinburgh these days May 23 04:35:15 * koen starts keyhole May 23 04:35:22 I guess having been brought up on Tyneside, its a wonder I can manage anything resembling English anyway :) May 23 04:35:54 hi, is anybody here? May 23 04:36:25 dmun: hi, no ;-) May 23 04:36:39 dmun: type /who in your irc client May 23 04:37:09 RP: arent you supposed to end every sentence with "like" May 23 04:37:25 XorA: Why aye man ;-) May 23 04:37:56 * XorA imagines a kernel with error messages in geordie May 23 04:38:26 That would be scary :) May 23 04:39:00 sweet, keyhole updated the !USA databases May 23 04:44:20 XorA: so you're somewhere in this picture: http://dominion.kabel.utwente.nl/koen/images/edinburgh.jpg May 23 04:52:27 koen: find yellow dotted box in center of map with edinburgh written beside it, I am midway one block left of the top left corner of that box May 23 04:53:42 koen: and if you follow river into the docks my hose is on left just by first dock you come to May 23 04:54:37 so your office is somewhere at rose or young street May 23 04:55:05 koen: on Princess Street, rose street is to our rear May 23 04:55:21 I read hose as horse, and was wondering for a minute there May 23 04:56:01 koen: http://maps.google.com/maps?q=eh2+4ad&hl=en is me May 23 04:58:14 google maps is pretty nice May 23 04:59:40 for that part of world which they cover May 23 04:59:55 indeed May 23 05:01:19 multimap or streetmap are better maps though from the looks of that one May 23 05:01:30 again, for those areas they cover May 23 05:02:19 I partially take it back - the interface is far better with google May 23 05:05:47 could someone please make a glibc_2.3.4.bb out of the glibc_2.3.3.bb (just copy and rename) May 23 05:05:58 and add that to the OE tree May 23 05:07:18 And anyone feeling up to adding the blackbox package I made before, would be most welcome to May 23 05:17:15 mickeyl: how to get from diag menu (c7x0) to normal? I took battery for some time and its nono, I used extramenu/reset and still in diag... May 23 05:18:56 iirc there is an option named "clear diag flag" May 23 05:19:02 you need to choose that and then reboot May 23 05:19:18 (which sharp has written in japanese) May 23 05:19:22 hehe May 23 05:19:27 of course May 23 05:19:34 thx it was that May 23 05:19:35 no no, the diag is in english May 23 05:19:41 diag menu is engish May 23 05:20:15 ok. hI have nand backup so can go gpe May 23 05:23:12 mickeyl: those big titlebars from 20050520 opie are awful... May 23 05:24:26 can someone ping home.hrw.one.pl? May 23 05:24:37 ~ping home.hrw.one.pl May 23 05:24:38 pong home.hrw.one.pl May 23 05:25:01 ibot: that isn't very helpfull May 23 05:25:27 hrw|work: it isn't responding from here May 23 05:25:55 ok. thx May 23 05:26:27 * koen wonders when people will stop confusing OE and OZ May 23 05:26:46 there shouldn't be a difference between 20050520 qte titlebars and the oz 3.5.3 release May 23 05:26:52 what do you mean? May 23 05:27:58 mickeyl: I think there's a problem with cardmgr, which incorrectly identifies inserted modem card as ttyS0, while it's ttyS3; who should I address this issue to? May 23 05:28:11 mickeyl: titlebar.patch was not added recently? May 23 05:28:31 mickeyl: http://oe-devel.bkbits.net:8080/openembedded/patch@1.3281.1.1 May 23 05:28:36 file://fix-titleheight.patch;patch=1" May 23 05:30:34 hrw|work: just to cvs version. has been added to 1.2.0 May 23 05:31:16 zmiq: no idea offhand. you should check how cardmgr gets this value May 23 05:31:28 oesources.org is down? May 23 05:31:54 chocolatetpot: yes May 23 05:32:11 koen: ok, cheers! May 23 05:32:36 I suspect it will be up again when california wakes up May 23 05:32:58 mickeyl: another issue: has anyone been successful loading ip_tables on oz3.5.3, 2.6.11-20050512 ? May 23 05:34:28 to my knowledge no May 23 05:34:33 i'm not at home, so i can't test right now May 23 05:34:43 what problem do you see ? May 23 05:35:41 lunch time May 23 05:36:39 I see an unknown symbol (-1): no suche file or directory when modprobe ip_tables May 23 05:36:56 s/suche/such May 23 05:40:41 which symbol? May 23 05:41:21 try insmod the offending module then look into logread, it should be printed there May 23 05:42:14 we need Larry Wall May 23 05:42:33 or someone who has been hit py the perl cluebat for a prolonged period of time May 23 05:42:49 s/py/by/ May 23 05:43:21 it's not specified May 23 05:44:12 mickeyl: nf_register_sockopt May 23 05:45:41 ok, now to find out where this symbol is defined and install this module May 23 05:47:42 mickeyl: zmiq net/core/netfilter.c May 23 05:50:18 xora: what module is that then? I have already installed iptable_filter.ko, but when modprobing it I get that ip_tables cannot be loaded error May 23 05:50:31 zmiq: at a guess a module called netfilter May 23 05:51:30 zmiq: although that is wrong looking at my modules list May 23 05:53:07 zmiq: you need to update kernel to that one which have netfilter turned on May 23 05:56:43 zmiq: you need atleast r17 kernel - 20050510 or newer May 23 05:57:00 zmiq: I propose update to r19 kernel from RP site May 23 05:57:51 hrw|work: I've already done so May 23 05:58:07 I'm using 20050512 May 23 05:58:33 from rp or upgrades? May 23 05:58:38 from rp May 23 05:58:48 (that's what you suggested ...) May 23 05:58:57 a.. yep.. May 23 06:00:00 hrw|work: i gotta go now; be back later; I was just wondering if I was the first one testing this; otherwise, if someone would have it working I'd reinstall my whole stuff May 23 06:00:42 zmiq: probably first May 23 06:01:29 hrw|work: i also think that there's much kernels and modules, and ip_table, ip_ipt, ip-iptables stuff in rp and oz, and that makes it a bit easy to get something wrong ... May 23 06:02:31 We really need to sort out the module dependencies... May 23 06:02:34 hrw|work: who should I get in conatct with to get that solved? May 23 06:05:01 zmiq: That depends which problem you're talking about and how you're proposing its solved... May 23 06:05:48 fsck.. lost g_ether module from filesystem... and card writer is at home ;( May 23 06:06:01 Setting the kernel module dependencies correctly would solve the usability issue - that's an oe problem May 23 06:06:04 hrw|work: as long as you have g-storage May 23 06:06:19 XorA: which I dont have now May 23 06:08:25 zmiq: hrw|work I have iptables running with -r20 May 23 06:08:41 xora: good to hear! May 23 06:09:34 gotta go now. xora, could you please post in oeforums links to the packages you used? thanks, May 23 06:09:55 zmiq: ipkg install *ip* May 23 06:10:14 nothing more specific? May 23 06:10:28 zmiq: no, just all the ipt and ip netfilter modules May 23 06:10:35 ok May 23 06:13:12 * XorA thinks this cpu is lieing when it claims to be a 1.7Ghz P4 May 23 06:15:08 ;) May 23 06:15:19 maybe it's only exagerating May 23 06:15:27 ... by a factor of 10 May 23 06:15:39 Twiun: I swear its slower than my 1Ghz thunderbird May 23 06:16:01 XorA: Is it not a celeron? May 23 06:16:18 XorA: slower p4 are slow May 23 06:16:46 RP: I begin to suspect it is and Compaq and lieing ********'s May 23 06:19:16 in other news, did you guys see the teeny new RiscOS machine? May 23 06:19:26 very tempted to by one as a portable machine May 23 06:19:27 yup May 23 06:19:31 500 gbp May 23 06:19:41 have you a url? May 23 06:20:13 http://www.drobe.co.uk/riscos/artifact1350.html May 23 06:20:38 and http://www.advantagesix.co.uk/products.html May 23 06:20:48 that is asking for an OE based distrib May 23 06:22:02 Hmmm. http://www.rpsys.net/myc3/#myc3 May 23 06:22:07 now how to hide 500 GBP from my mrs May 23 06:23:20 RP: sweet May 23 06:24:33 RP: home built? May 23 06:24:48 XorA: yes. All by me :) May 23 06:25:18 RP: nice, where did the SBC come from? May 23 06:25:56 XorA: There are various people selling them. I can't remember which company I got it from in the end May 23 06:26:05 hey, any breakage produced with the latest bitbake? May 23 06:26:24 The PSU in that is the special feature... May 23 06:26:33 zecke: I've not tried yet :) May 23 06:26:40 zecke: it runs great over here, but hrw said it doens't play well with qa May 23 06:26:43 RP: thats the sort of shit I should do :-) I could probably justify buying parts to the missus May 23 06:26:45 zecke: bitbake-qa not fetch with latest May 23 06:27:04 XorA: The case is the tricky bit May 23 06:27:44 hrw|work: do you've a wrong BBPATH or PYTHONPATH? May 23 06:27:54 or wait May 23 06:29:05 zecke: bittest goes but fails to fetch. when changed PYTHONPATH to older bitbake (~r205) it works May 23 06:29:54 hrw|work: I need to test it tonight then May 23 06:30:13 RP: luckilly the Mrs has an artistic flair and is good at fiddly stuff May 23 06:30:33 zecke: ok May 23 06:30:44 hrw|work: I changed virtually nothing that should play evil games with fetching sourcecode May 23 06:30:54 zecke: my home machine refuse to connect so I cannot tell more May 23 06:31:48 hrw|work: so you've network connectivity problems and wonder why bitbake fails fetching source ;) May 23 06:32:02 zecke: ha ha ha May 23 06:32:25 zecke: bitbake fetched in same time when bittest does not May 23 06:32:37 just kidding your reports are always excellent May 23 06:32:41 I know May 23 06:32:46 :) May 23 06:32:58 zecke: I have your new bb happilly generating an image on a 256meg colinux VM May 23 06:33:26 nice May 23 06:34:06 zecke: I'm OO'ing bitbake tonite, please refrain from checkins until I'm completed May 23 06:34:18 s/I'm/this task has been/ May 23 06:34:32 mickeyl: screw you I break bitbake whenever I want :} May 23 06:34:38 hah May 23 06:34:39 * zecke starts to behave again May 23 06:34:56 mickeyl: zecke Z's at dawn :-) May 23 06:35:38 hehe yeah May 23 06:35:52 I'm really pleased to see bitbake moving forward like this :) May 23 06:36:38 mickeyl: once your OO stuff is in, should we consider doing a 1.2.2 release? May 23 06:37:07 RP: I plan to use your bitbake data instance with bitbake-ng in the OE recipe box May 23 06:37:14 zecke: I'd vote more for 1.3, but yes, a release would be good May 23 06:37:22 zecke: I think the memory improvements justify 1.3 ;-) May 23 06:37:26 mickeyl: changes are minor May 23 06:37:48 mickeyl: on my 32 CPU SGI machine the difference in ram usage is unnoticable May 23 06:38:19 zecke: A x10 reduction in memory usage is "minor"? :) May 23 06:38:48 RP: from 0.002% to 0.00129% ;) May 23 06:38:49 zecke: poser :D May 23 06:39:06 zecke: How much memory has it got? May 23 06:39:28 mickeyl: actually I not even have an account there, but I was jealous when woglinde ran a dmesg May 23 06:39:31 RP: I heard it has an infinite tape. May 23 06:39:54 RP: right infinite tape and 'drum-memory' (Trommelspeicher) May 23 06:40:31 right :) May 23 06:41:51 pb_: infite is always good? May 23 06:41:57 infinite even May 23 06:42:05 zecke: infinite++ is better May 23 06:42:20 which order of inifinity? May 23 06:42:31 koen: aleph 3 May 23 06:44:38 zecke: concerning tapes, I think so May 23 06:44:51 obviously there are other situations where infinite would not be so good May 23 07:07:59 http://www.oesf.org/forums/index.php?showtopic=12764 - for opie users May 23 07:08:01 ops May 23 07:09:07 hrw|work: next oz with ttf instead of qpf? May 23 07:09:31 hrw|work: going to make ttf .ipk for them? May 23 07:10:08 koen: who knows... but I'll vote for dejavu instead of vera May 23 07:10:23 hrw|work: it's all the same to me May 23 07:10:26 XorA: probably yes - first have to check ttf fonts policy in OE May 23 07:10:32 I'm not i18n enough to run into troubles May 23 07:10:38 koen: you dont use accents etc May 23 07:10:55 heh.. qpf archives takes 13M May 23 07:11:15 hrw|work: we only have vera I beleive :-( May 23 07:11:31 XorA: we == !opie? May 23 07:11:50 yep.. only vera.. May 23 07:11:52 bitbake even works without any font May 23 07:12:18 zecke: makes guessing the error kinda difficult though May 23 07:12:45 everything is there $\epsilon$ May 23 07:12:46 XorA: zecke has morse code output on numlock led May 23 07:13:08 hrw|work: dont joke, Ive done similar on embedded boards before May 23 07:13:16 ;)) May 23 07:13:49 XorA: forgot that in OE that joke is not as good as with normal^wother computer related people May 23 07:21:38 hi again all. Trying build packages using OE for sharprom-compatible. Got further, still stuck - any ideas, pointers appreciated... May 23 07:21:48 Hmm, I am having a senior moment May 23 07:21:56 consider the following line May 23 07:22:04 oe_libinstall lib/libxerces ${D}${libdir} May 23 07:22:04 + oe_libinstall lib/libxerces-c ${D}{libdir} May 23 07:22:04 oe_libinstall lib/libxerces-depdom ${D}{libdir} May 23 07:22:22 I think {libdir} should be ${libdir} May 23 07:22:30 I edited sharprom-compatible.conf to comment out all the ASSUME_PROVIDED lines other than the one for arm-linux-gcc-2.35 - this got me further and got packages created May 23 07:22:34 but I think behavior says otherwise? May 23 07:22:42 Crofton: yes, it should be, May 23 07:22:54 shoudl be $libdir ? May 23 07:23:11 packages in the right format etc. However, everything depends on a version of libc which isn't in my ROM. Hence stuck. May 23 07:23:11 ${libdir} May 23 07:24:35 it's after libc version GLIBC_2.3 / 2.3.4 which I assume is newer than the Sharp original rom one May 23 07:25:01 gints|wrk: ping May 23 07:25:34 I am assuming there is some way of telling OE that I have an older libc? May 23 07:25:57 chocolatetpot: you need to assume_provided libc and all other glibc stuff and copy its headers/libs into tmp/staging/arm-linux/ probably May 23 07:26:59 I guess that is that headers and libs matching the sharp ROM? May 23 07:28:01 yep May 23 07:28:10 hrw|work: pong May 23 07:29:55 hrw|work: ah, it's all slowly dawning on me now ;) So I can pull the libs from the device itself, is there an official place to get the headers or is it JFGI time? May 23 07:33:05 gints|wrk: hunkyfonts are vera derived? May 23 07:33:27 hrw|work: do not know realy... May 23 07:33:28 chocolatetpot: get it from sharp toolchain? May 23 07:35:11 hrw|work: may be you can get some info from project sf.net page... May 23 07:35:25 gints|wrk: I'm thinking of adding some ttf fonts into OE but hunkyfonts are miracle to me.. everywhere it is listed that it is vera derived, but author license it under LGPL 2.1 May 23 07:35:49 http://sourceforge.net/projects/hunkyfonts/ May 23 07:36:29 I know May 23 07:36:43 heh May 23 07:36:48 I have to read vera license May 23 07:37:00 you do know that your document is under gpl if you use gpl'ed fonts in it? May 23 07:37:09 s/use/embed/ May 23 07:37:19 thanks to our friends at the fsf May 23 07:37:24 heh .. :) May 23 07:38:06 hrw|work: cheers, I'm there. May 23 07:38:46 oki ... goin home now... May 23 07:39:39 koen: that depends May 23 07:40:17 yeah, you have to embed them in the doc to hint the 'linking' clause May 23 07:40:26 s/hint/hit/ May 23 07:41:14 koen: though FSF made noises of fixing that, as that is sure to make companies avoid *nix utils forever May 23 07:43:49 hmmm May 23 07:44:15 bero hinted that mainting a .bb for cvs and svn exludes us from the free bk license too May 23 07:44:27 excludes* May 23 07:44:38 koen: not that that matters anymore :-) May 23 07:44:42 yeah May 23 07:45:10 any news on the monotone/svk front? May 23 07:46:38 * treke|home doesnt buy that May 23 07:46:59 'morning treke|home May 23 07:47:46 if you follow that line of reasoning, even using svn is going to prohibit you from ever using bk/free May 23 07:49:02 or software developed in cvs/svn May 23 07:50:55 heh - its one thing to prevent others from getting an advantage by building on your software, its quite another thing to completely stifle all mention of competitors May 23 08:01:17 hi offroadgeek May 23 08:01:30 CosmicPenguin: could you update gftp to 2.0.18 final? May 23 08:02:20 hi hrw|work May 23 08:13:14 cu May 23 08:34:52 koen - I've got a couple of changes to portmap which I would like to check in. The global one is to remove the --oknodo from start-stop-daemon on 'stop', 'cause busybox does not support it. May 23 08:36:33 The other one is INITSCRIPT_PARAMS_openslug to move the rc[06].d/S??portmap to K??portmap so that it happens before networking goes down and so that it works on openslug (portmap is in rc[1-5].d/S??portmap too, so /etc/init.d/rc does not run it in rc6.d) May 23 08:41:13 sounds good to me. May 23 08:41:19 i don't know why we're using portmap at all May 23 08:41:36 i don't see a problem in mounting nfs with -nolock May 23 08:41:40 *shrug* May 23 08:41:58 Well, openslug is a file server, so locking is pretty much a requirement ;-) May 23 08:42:11 that's right May 23 08:42:17 I don't see why any other system is putting portmap in though. May 23 08:42:19 makes sense for your target platform May 23 08:43:04 Ok, my change fixes openslug so that it shuts down portmap correctly, other platforms may still have a problem, but they should probably not be including it... May 23 08:43:38 ya. we will reconsider that May 23 08:43:54 we're close to revamping the bootstrap task anyway May 23 08:44:05 towards a bit more granularity May 23 08:44:54 re May 23 08:45:00 wb May 23 08:45:41 Ah, I've been encountering problems with some things not happening - e.g. the rc6.d sequence leaves some stuff mounted on openslug and it won't reboot. May 23 08:45:55 mickeyl: what do you think about dropping qpf-bitstream-vera? qt/e 2.3.10 smoothsans is the same May 23 08:46:10 hrw: fine with me May 23 08:46:14 ok May 23 08:46:28 I plan to create ttf-fonts/ dir and place some recipes there May 23 08:46:33 cool May 23 08:46:44 ~praise hrw May 23 08:46:46 All hail hrw! May 23 08:46:50 someone can update gsl to 1.6 and try to build it? May 23 08:47:52 not now May 23 08:47:54 later tonite May 23 08:48:01 but first I have to get sound working in wine May 23 08:48:03 i'm doing a full opie-image with my OO'ed bitbake to test May 23 08:48:22 long time since last time I did OO... May 23 08:48:42 php and OO is a no-no for me - too simple things in sites to go OO May 23 08:48:50 right May 23 08:50:33 uah, i should unsubscribe from qt4-preview-feedback May 23 08:52:48 zecke: bittest works here with bitbake r205 (latest) May 23 08:59:24 hmm May 23 08:59:28 gsl 1.6 no worky May 23 08:59:39 | In file included from fp.c:36: May 23 08:59:39 | fp-gnuc99.c: In function `gsl_ieee_set_mode': May 23 08:59:39 | fp-gnuc99.c:81: error: `FE_TONEAREST' undeclared (first use in this function) May 23 08:59:39 | fp-gnuc99.c:81: error: (Each undeclared identifier is reported only once May 23 08:59:39 | fp-gnuc99.c:81: error: for each function it appears in.) May 23 09:00:47 ok. same as here May 23 09:00:52 mickeyl: sorry about the konq/e bug report in mantis... I don't know what I was thinking ;) May 23 09:00:56 thats why I not pushed 1.6 May 23 09:01:02 offroadgeek: nm :) May 23 09:01:29 offroadgeek_: we have to draw a clear line between our stuff and the external programs or else we'll get swamped with reports for things no one wants to look at ;) May 23 09:02:01 offroadgeek: did you get pocketop working? May 23 09:11:57 i have the feeling some people are using oe@handhelds.org as a write-only-list May 23 09:12:37 at least I seem to seldomly get answers when asking about issues with patches appearing there May 23 09:12:38 heh May 23 09:17:11 heh May 23 09:22:27 heh.. supermemo under wine cannot open sound device ;( May 23 09:27:54 anyone know what happened to this tech? Fujitsu's lab site doesn't have info http://www.mobilemag.com/content/100/102/C2393/ May 23 09:28:32 CIA-4: fixed SRC_URI for linux-jlime May 23 09:28:42 ~shoot jlime maintainer May 23 09:28:44 * ibot shoots jlime maintainer in the head with a cork gun! May 23 09:32:04 would I be right in assuming today is monday? May 23 09:32:50 all day May 23 09:32:56 ~monday May 23 09:32:57 i guess monday is when everything breaks for no apparent reason, creating so many problems that it takes you until friday to get back to normal. at which point one more thing breaks that takes you the whole weekend to fix May 23 09:33:00 until tomorrow May 23 09:39:07 Hmm.. I'd think that with a 'bitbake e2fsprogs-native', I'd get a native mke2fs --- but that doesn't work.. May 23 09:39:49 stupid fc3 ext2 extensions.. May 23 09:39:52 heh May 23 09:40:02 fc3 is going to be a proprietary Linux May 23 09:40:07 s/fc3/fcX/ May 23 09:40:21 too many specials all over May 23 09:40:22 fcX? May 23 09:40:24 it's sad May 23 09:40:27 "going to be"? heh May 23 09:40:32 fedora core May 23 09:40:39 ok, already is May 23 09:40:50 it seems. May 23 09:41:36 aren't they going to be a bit more open again? May 23 09:41:46 read something abt that.. May 23 09:42:08 188M of stuf to recompile wine.. May 23 09:42:24 RP: reping May 23 09:42:29 oesources.org down? May 23 09:43:38 Hmmpf, no wonder treke.net doesn't work either.. ;) May 23 09:43:42 There is no staging build/tmp/staging/x86_64-linux/sbin at all.. Is sbin stuff skipped perhaps, for native? May 23 09:44:38 luke-jr_: hello May 23 09:44:47 RP: Hey May 23 09:44:49 * kergoth yawns and stretches May 23 09:46:48 hh kergoth May 23 09:47:15 ~tosa May 23 09:47:16 well, tosa is the Sharp SL-6000 PDA, or a dog (see http://www.dogbreedinfo.com/tosa.htm) May 23 09:47:46 Good evening. May 23 09:48:53 RP: Still have problems in waking up from suspend now and then. In this case my CL860 stay on the text screen and does not react any more (instead of SysReq + Boot). May 23 09:53:16 uv1: gpe or opie? May 23 09:54:00 03mickeyl * r206 10bitbake/bin/bitbake: (log message trimmed) May 23 09:54:00 Slight OO refactoring of BitBake command line utility in order to May 23 09:54:01 prepare for the BitBake shell which will need to have instance based May 23 09:54:03 variables as opposed to module global ones. May 23 09:54:05 Sorry, this diff is larger than it should be, but I had to reorder May 23 09:54:07 methods and change indenting all over the place. What I basically did May 23 09:54:09 is: May 23 09:54:34 CIA-4: congratulations. you have a talent for trimming at the interesting places May 23 09:54:34 heh May 23 09:54:55 RP: opie May 23 09:55:02 ;) May 23 09:55:30 we have too many kernels in OE:) bitbake-qa is fetching kernels for over hour here.. May 23 09:55:54 someone maintain ipod kernel? May 23 09:56:00 uv1: I've also seen that but it was there before I started fiddling with the power code. I have no idea what the problem is :-( May 23 09:56:37 anyone have a recent mirror of Dirk's patch site? May 23 09:56:42 uv1: 2.6.11 or 2.6.12-...? May 23 09:56:43 hrw: is the qa thing checking (amongst things) the availability of sources of packages? May 23 09:57:08 ok, this time I broke bitbake. please throw some testcases at it May 23 09:57:11 RP: 2.6.11. I think it is related due to timing. It looks like if opie does not come up quick enough, the system gets stuck. May 23 09:57:14 mickeyl: how much is missing for the shell? May 23 09:57:16 (opie-image seems to work) May 23 09:57:38 Hertog: source availability, patches applyability May 23 09:57:38 * zecke thinks it is time to read some serious python code May 23 09:57:40 zecke: a weekend of work May 23 09:57:59 mickeyl: bbsh? May 23 09:58:00 zecke: some command execution logic and perhaps some more fine granular split of the bitbake methods May 23 09:58:03 hrw: ya May 23 09:58:17 hrw: ah, ran into a few 'by hand', May 23 09:58:29 Hertog: send patches to bugzilla then May 23 09:58:30 uv1: or maybe its too quick. I don't think I see this under 2.6.12... May 23 09:58:37 hrw: sorry, I was afk... no I'm not trying to get pocketop working (at least on my 1000) because of the ill placement of the ir port May 23 09:58:59 ah.. remember.. May 23 09:59:04 mickeyl: ah cool we now could easily, add -p for profiling the cooker May 23 09:59:16 zecke: exactly May 23 09:59:43 mickeyl: also I plan to improve caching May 23 09:59:58 i want to finish that shell thing because it will improve the way I work. i type way too much and always the same things. bbnew , it breaks, file, parse , etc. May 23 09:59:59 pb_: can I find out that something changed in a directory by looking at stat? May 23 10:00:04 pb_: even for subdirectories? May 23 10:00:10 RP: Does 2.6.12 work with the rest meanwhile (resume, suspend, network )? May 23 10:00:27 hrw: was planning on that, but since it where a couple of them, I was wondering about a more 'structured aproach' (mv-ing the sources dir and -c fetch) but it seems that is being taken care of now? May 23 10:00:49 zecke: I don't think so May 23 10:00:51 Hertog: you can still help May 23 10:00:57 hrw: will do May 23 10:01:08 zecke,pb: time to write a python inotify wrapper :D May 23 10:01:18 beep May 23 10:01:23 mickeyl: for your shell yes May 23 10:01:33 mickeyl: I want to cache the dep tree May 23 10:01:36 Hertog: look into last changesets - many apps are old - maybe you can update them etc May 23 10:01:36 anyone know if there is a biometrics system that work in *nix? May 23 10:01:53 mickeyl: good plan :-) May 23 10:02:16 do13: Some of the more recent files seem to have disappeared from your website May 23 10:02:27 zecke: did you see "11:03 < proti>"? May 23 10:02:46 hrw: nope May 23 10:02:50 emte: http://dominion.kabel.utwente.nl/koen/pda/?page=software&sub=gpe-fsi&PHPSESSID=b229174185e4ebfdeed94766be5bf230 May 23 10:02:51 zecke: "23 11:03 < proti> A bitbake question : Why do we have to compute the world dep everytime we build the dep graph ? I removed the code doing it and nothing bad happened (as suspected). I put it in a build_depgraph_world function and called it if world is given as target." May 23 10:02:55 he is on /ignore May 23 10:03:05 zecke: ah ;) May 23 10:03:39 Hertog, i guess i should have specifically said a usb-biometric ... May 23 10:03:43 hrw: idiot, I've explained it three times May 23 10:03:50 ;) May 23 10:03:51 ok May 23 10:03:54 hrw: and then he got onto the /ignore list May 23 10:04:17 hrw: see we now build the world tree as well and it is even faster than any previous bitbake version May 23 10:04:21 that one as well doesnt do any auth May 23 10:04:38 aha May 23 10:04:38 heh.. linux-2.6.12-rc2 as whole archive.. May 23 10:05:25 emte: duh the 'todo' lettering was so big, my brain didn't parse them May 23 10:07:06 hrw: actually we did worse than the worldgraph May 23 10:07:19 hrw: iterating over each file three times to get the PackageNames May 23 10:07:56 Hertog, there is one progect at FM but i cant find any status on it ... May 23 10:08:38 http://fvs.sourceforge.net/index.html May 23 10:11:14 actually it looks like it may work ... May 23 10:11:43 koen, probably already knows about the library May 23 10:13:53 re May 23 10:17:38 RP: ok, i have no problems with your 0.15rc1 May 23 10:17:45 RP: I'll compile one myself now May 23 10:18:36 mickeyl: ok, I'm pleased it works for us both :) May 23 10:18:54 ya, works fine. no BAP entry and great speed May 23 10:20:19 mickeyl: excellent. I suspect its the makefile causing problems for 0.15rc1. 0.15rc2 doesn't work though May 23 10:23:32 time for food - back shortly May 23 10:24:53 bon appetit May 23 10:25:16 smacznego RP May 23 10:25:42 or "smach nego" as someone said here few days ago :) May 23 10:35:06 who maintain matchbox* packages? 0.8.x are not fetchable May 23 10:42:15 * mickeyl looks over to the gpe crowd May 23 10:42:57 mickeyl: zaurus-clamshell-2.6.conf is good place to add usb gdaget kernel modules? May 23 10:43:15 which one would that be ? May 23 10:43:19 s/one/ones/ May 23 10:43:35 I added all May 23 10:43:50 basically g-ether and g-storage May 23 10:44:01 ah so they're modular now? May 23 10:44:15 yes, zaurus-clamshell-2.6 is the right place then. please insert them at the proper position May 23 10:44:16 yes May 23 10:44:44 does usb gadgets imply usb-host functionality? May 23 10:44:49 no May 23 10:44:54 usb gadget is usb client API May 23 10:45:13 usb host is usb host :) May 23 10:45:18 ah, ok, was wondering which Z was having usbhost ;) May 23 10:45:26 spitz,akita,tosa have May 23 10:45:56 ~akita May 23 10:45:57 methinks akita is the Sharp SL-C1000, or a dog May 23 10:46:13 k May 23 10:49:51 03RP 07 * r1.3374.1.1 10openembedded/packages/linux/linux-openzaurus_2.6.12-rc4-mm2.bb: linux-oz-2.6: tosa updates from Dirk May 23 10:50:01 * RP is gone again May 23 11:14:41 haha debian WIP fsck is fscked again May 23 11:19:52 it anyone (preferably in Europe) able to ship to Poland relatively easily? May 23 11:20:10 is anyone... sorry May 23 11:20:24 I think any EU member qualifies? May 23 11:20:37 sure May 23 11:21:29 zecke: mind a private chat? May 23 11:22:43 pb_: you have a chance to read that bitbake-ng email? May 23 11:23:28 offroadgeek: if it is a ibook I can't gurantee to pass it to poland May 23 11:25:08 zecke: I wish it was! May 23 11:25:40 offroadgeek: query me May 23 11:39:45 03zecke123 * r207 10bitbake/TODO: May 23 11:39:45 bitbake/TODO: May 23 11:39:45 -Not much is done yet but at least fake that we've done May 23 11:39:45 something May 23 11:40:46 03zecke123 * r208 10bitbake/lib/bb/__init__.py: May 23 11:40:46 bitbake/lib/bb module: May 23 11:40:46 Bump the version number to be different from the May 23 11:40:46 last version May 23 11:41:40 03zecke123 * r209 10bitbake/MANIFEST: May 23 11:41:40 bitbake/MANIFEST: May 23 11:41:40 -Include data_smart.py in the MANIFEST May 23 12:12:04 RP: can you tell me where the irq type should be set up for a CF card driver? I was told on linux-usb-devel that it should be done outside the pcmica driver. May 23 12:31:04 koen: I bet we will use monotone, it has lua in it May 23 12:32:46 if it can perform, anyway May 23 12:32:53 heh May 23 12:35:17 hi kergoth May 23 12:35:21 hey May 23 12:35:24 finally heading out to mn today May 23 12:35:38 man, was finishing up cleaning my closet May 23 12:35:45 found a small uhaul box half full of books May 23 12:35:50 :\ May 23 12:35:53 hehe May 23 12:35:56 no way thats fitting in the car May 23 12:36:14 kergoth: my father is coming next weekend and I have to clean the place ... May 23 12:36:23 heh, fun fun May 23 12:36:49 not really May 23 12:42:50 kergoth: have a safe drive May 23 12:47:28 May 23 12:49:11 kergoth: toi toi toi May 23 13:01:32 I can't seem to get started on building. My problem seems to be in patching, it can't patch quilt. The error, it seems that the patch that I have brough in with the system (gentoo, and there's been nothing extra) seems awfully picky about the name of the patch directory, it gets unhappy about there being anything but an exact match. Anyone have any guesses here? May 23 13:03:08 quilt searches _up_ fromk the directory its run in all the way to /, searching for the top of the source tree (since you might be calling it from a subdir of the tree) May 23 13:03:22 it recognizes the top of the source tree by the existance of a directory called "patches" May 23 13:03:33 patcher does the very same thing, but is less likely to cause problems since it looks for ".patches" May 23 13:03:50 just rename the rogue patches dir, or INHERIT += "patcher" in your local.conf May 23 13:06:48 kergoth, the problem I'm hitting is in the rebuilding of quilt (which I suppose could use quilt to do) May 23 13:07:06 the building of quilt doesnt use quilt to apply patches, just 'patch' May 23 13:07:28 but again, inheritnig patcher would work around the problem, since you wouldnt need to build quilt ;) May 23 13:07:32 * kergoth shrugs May 23 13:07:36 kergoth: we should add a mkdir -p patches; somewhere before doing quilt upgrade May 23 13:07:38 I don't think it is using quilt, I think the prpblem is in 'patch' May 23 13:07:48 didnt i just say that? May 23 13:07:50 i could have sworn i did May 23 13:08:18 heh May 23 13:08:18 mmm.... cake May 23 13:08:26 kergoth: no didn't show up in the log May 23 13:08:40 were you folks atlking about this before I got here? I may have stepped into an open hole accidentally ... May 23 13:08:59 chuckr: sorry to have confused you May 23 13:09:34 i'ts my problem to avoid confusion, I'm very sorry to mess things up, but I *do* have a problem May 23 13:09:41 first of all, we cant do a damn thing to diagnose the problem with your patch May 23 13:09:45 especially not without exact errors May 23 13:09:51 would you like us to pull a guess out of our collective ass? May 23 13:09:53 we certainly good May 23 13:10:01 second, once again, you can work around it May 23 13:10:05 I'm very new at this, please, what should I do? May 23 13:10:11 .. May 23 13:10:22 WORK AROUND IT, OR GIVE US EXACT FUCKING ERROR MESSAGES May 23 13:10:42 The scary thing is - this guy is going to be driving across the country in a few hours... :) May 23 13:11:00 well jesus, how many times do we need to say things before they sink in May 23 13:11:18 I'll come bak later (I have the exact error messasges, but certain people get mad when stuff like that is just posted) May 23 13:12:46 huh? May 23 13:13:06 as long as you dont flood the channel, which is just basic irc ettiquitte, you'll be fine May 23 13:13:14 CosmicPenguin: stay inside ;) May 23 13:14:38 * offroadgeek wonders if kergoth ever experiences road rage May 23 13:14:49 amen May 23 13:15:31 yes, idiots piss me off everywhere, on the road and off. May 23 13:16:25 chuckr: its very simple - if you have a problem, then report it with the exact error messages, and we'll try to help May 23 13:16:32 * offroadgeek laughs May 23 13:16:40 chuckr: yeah, we might tell you "thats been fixed, pk pull and try again" May 23 13:17:20 I have a bad reaction to geting yelled at, sorry, the problem, I'm not sure how much I can paste here, is there someplace else I could paste maybe 20 lines? May 23 13:17:26 But if you don't tell us the error - then we have to guess, and we're just not very good at that - especially since it seems to work on everybody elses machine May 23 13:17:31 ~pastebin May 23 13:17:32 i guess pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca May 23 13:17:47 ok, gvei me a moment May 23 13:18:53 kergoth: You don't turn AC/DC on when driving? Don't you? May 23 13:20:43 hehe May 23 13:20:46 * offroadgeek wonders if he should send kergoth an Enya CD for the drive May 23 13:21:14 ok, it's at http://pastebin.ca/12454 May 23 13:21:34 kergoth: how long is the tip back home? May 23 13:21:48 I don';t really know yet what parts are needed and what parts are fluff May 23 13:21:58 zecke: AC/DC is pretty harmless in my experience. dnb is evil. May 23 13:22:02 chouimat|tv: iirc its about 18 hours or so striaght through May 23 13:22:23 dnb? May 23 13:22:29 drum and base May 23 13:22:31 ok some people are simply clueless when they ask for features May 23 13:22:51 it looks like (to me) that somehown the wrong amount of level are stripped off in the -p argument to patch, but I don't know why this is May 23 13:23:00 chuckr: its time to check assumptions. the patch is failing to apply, but is it failing due to the patch, or due to what its trying to patch? make sure it was able to unpack the quilt sources. tmp/work/quilt-whatever/quilt-0.32/ May 23 13:23:43 betcha cvs freaked out on him May 23 13:24:23 I'm not precisely sure, but I might add in this little bit of info, I went looking for quilt myself, and it seems that there are several different projects named quilt, and they have numbers that are closely alike May 23 13:24:38 kergoth: makes you lose any sense of speed (wording?) in my experience. May 23 13:24:59 I dono't know what version quilt is being built here, but the patch itself is named 32, it quilt #32 the sourve? May 23 13:25:18 32 is in the patch name May 23 13:25:44 I don't know how to check what version quilt is being built May 23 13:26:22 ERROR: quilt- failed on do_patch May 23 13:26:27 chuckr: NOTE: package quilt-native-0.39: started May 23 13:26:28 something like that should be in the log May 23 13:26:38 That should be a dead giveaway May 23 13:26:50 where is the log left? May 23 13:26:59 you mean stdout? May 23 13:27:19 chuckr: the line I posted is from *your* pastebin May 23 13:27:27 ok May 23 13:28:25 chuckr: hm? its applying install.patch. May 23 13:28:54 ls tmp/work/quilt-native-0.39-r0/quilt-0.32/ May 23 13:28:58 the problem I have, I see what seems to be a fetch of a version of quilt that I can't find at all on the net, it's named quilt_savannah.nongnu.org_VER_0_39_.tar.gz, that's a weird name for a package May 23 13:29:07 thats a pull from cvs. May 23 13:29:14 ok May 23 13:29:26 I was trying to find it in ftp sites May 23 13:29:38 well? are the sources there? May 23 13:29:42 ls tmp/work/quilt-native-0.39-r0/quilt-0.32/ May 23 13:29:50 moment May 23 13:30:08 or in tmp/work/quilt-native-0.39-r0/quilt/, or similar May 23 13:30:15 heh - I was going to say... :) May 23 13:30:19 :) May 23 13:31:41 I'm looking at /usr/tmp/work/quilt-native-0.39-r0/temp/, it's all full of patch logs May 23 13:32:09 yes, that isnt what i said May 23 13:32:10 second please May 23 13:32:23 is anything else in tmp/quilt-native-0.39-r0/? May 23 13:32:32 please I need a moment! May 23 13:32:35 er s/tmp/tmp\/work/ May 23 13:32:35 heh May 23 13:34:11 there are two sorts of files, run.do_patch files and log.do_patch, and each maybe 10 (I've been screwing with it a bit) May 23 13:34:22 yes, there are May 23 13:34:25 once again, that isnt what i asked May 23 13:34:43 I'm not arguing with you Kergoth! May 23 13:35:07 i'm going to go be productive. i'll come back in 20 minutes, when perhaps you'll have answered the question May 23 13:35:13 chuckr: go back one directory now May 23 13:35:15 cd .. May 23 13:35:22 yeah? May 23 13:35:33 whats in that directory May 23 13:35:51 I'm trying really, but between trying to read here and go back there, I miss something sometimes, I;'m really sorry if I do that May 23 13:36:11 you mean the parent of that dir with all the log files? May 23 13:36:15 yes May 23 13:36:20 moment May 23 13:37:45 b-june:root:/home/chuckr>ls -l /usr/tmp/work/quilt-native-0.39-r0/ May 23 13:37:46 total 1 May 23 13:37:48 drwxrwxr-x 2 root root 48 May 23 18:35 quilt May 23 13:37:49 drwxrwxr-x 10 root root 632 May 23 18:35 quilt-0.39 May 23 13:37:51 drwxrwxr-x 2 root root 768 May 23 21:02 temp May 23 13:37:52 b May 23 13:38:08 is that what you wanted? May 23 13:38:29 There you go May 23 13:38:59 what does it show you that I'm blithely missing? May 23 13:39:27 that its building 39? May 23 13:39:36 Which of the two quilt directories has anything in it? May 23 13:39:43 oh May 23 13:40:00 quilt is empty May 23 13:40:17 the other is Quite full May 23 13:40:46 I want to go look at those log files May 23 13:42:39 chuckr: the OE repo shows the S dir to be "${WORKDIR}/quilt", so I'm not exactly sure where the quilt-0.39 directory came from May 23 13:42:59 But thats your problem - there isn't any source in tmp/quilt-native-0.39-r0/quilt/, so there is nothing to patch May 23 13:43:40 Make sure that quilt_savannah.nongnu.org_VER_0_39_.tar.gz exists in your source directory, and make sure that its a good file, and not zero length May 23 13:43:45 OK, the S dir is what exactly? May 23 13:44:17 The source directory - thats where the patches will be applied May 23 13:46:47 anyone any idea why PREFERRED_PROVIDER_virtual/kernel_ doesn't work? May 23 13:48:00 i'm trying to substitute it with the alternative PREFERRED_PROVIDERS syntax now but it would be nice to know why it doesn't work May 23 13:54:54 where does that file quilt_savannah.nongnu.org_VER_0_39_.tar.gz come from? Is it a direct cvs pull? May 23 13:57:26 hmm zecke did we change something in syntax ? May 23 13:57:38 PREFERRED_PROVIDERS_append = "foo:bar" in a .conf no longer works May 23 13:57:40 whereas May 23 13:57:46 PREFERRED_PROVIDERS += "foo:bar" still works May 23 13:57:59 however, we need the append syntax for the overrides May 23 13:58:35 chuckr: yeah, more or less. we save our copies to use cvs update subsequently May 23 13:58:57 chuckr: if i were you i'd remove my entire build tree, echo "INHERIT += patcher" >> conf/local.conf, rebuild stuff and be happy May 23 14:00:21 mickeyl: do we've a test case for that at all? May 23 14:00:36 yeah May 23 14:00:40 zaurus-clamshell-2.4.conf May 23 14:00:42 ah no May 23 14:00:46 that one uses the old form May 23 14:00:50 well, from the hints in the files I read, I thought I would get my sources vi oesources, will that work for this? May 23 14:01:08 probably not May 23 14:02:19 zecke: May 23 14:02:26 TESTVAR_append = "foo" May 23 14:02:26 TESTVAR += "bar" May 23 14:02:30 in your local.conf May 23 14:02:34 bbread|grep TESTVAR May 23 14:02:36 the thing is, the name that os is asking for, quilt_savannah.nongnu.org_VER_0_39_.tar.gz, is a really weird name, that doesn't exist in anyone's ftp directories May 23 14:02:38 foo is not there :( May 23 14:02:52 brb May 23 14:02:57 chuckr: read again what kergoth said May 23 14:03:01 chuckr: you know cvs? May 23 14:03:02 chuckr: so what? oe creates that tarball after checking it out of cvs. May 23 14:03:06 chuckr: you know what tar is? May 23 14:03:12 sure May 23 14:03:23 chuckr: you know you can get any data out of cvs? May 23 14:03:27 ok May 23 14:03:34 chuckr: you know that cvs puts load on servers? May 23 14:03:45 I didn't realize that's what he was talking about May 23 14:04:01 chuckr: do you know that it is smart if 1000 people get a cvs snapshot for a date, one generates a snapshot? May 23 14:04:18 chuckr: then ask, don't assume May 23 14:04:27 what with 5 different conversaations going on, I guess I am mis-guessing whats for me and what isn't May 23 14:04:56 ok then , where does oe store that file? May 23 14:04:56 treke-: what was that python based 'tinderbox'? May 23 14:05:08 buildbot May 23 14:05:11 chuckr: whereevery you say it will be stored (DL_DIR) May 23 14:05:12 chuckr: May 23 14:05:13 the problem I have, I see what seems to be a fetch of a version of quilt that I can't find at all on the net, it's named quilt_savannah.nongnu.org_VER_0_39_.tar.gz, that's a weird name for a package May 23 14:05:13 thats a pull from cvs. May 23 14:05:13 <-- woglinde has quit ("zapp") May 23 14:05:13 ok May 23 14:05:19 dont say 'ok' if you dont know. May 23 14:05:56 don't know what? tar and cvs?? I know both well enough May 23 14:06:11 i told you it was from cvs. what part of that is hard to understand? May 23 14:06:28 why does it take people a half an hour to slowly but surely, step by step, hammer this into your skull, when i've told you already? May 23 14:06:44 you cant be this dense. you wouldnt survive outside. May 23 14:07:37 kergoth`bbl: is it larger than a bread box? *insane* May 23 14:08:13 hehe May 23 14:10:52 why is USB so ugly under Linux? May 23 14:10:56 * zecke dumps tinderbox and stops trying to use perl May 23 14:11:03 s/USB/everything/g May 23 14:11:34 or maybe not May 23 14:15:02 hmm May 23 14:16:47 mickeyl: does that bug need a fast solution? May 23 14:16:56 no May 23 14:17:01 it's an oddity May 23 14:17:08 somehow also behaved to bbread as it seems May 23 14:17:15 doesn't interfere with normal operation May 23 14:17:19 but lets keep it in mind May 23 14:17:50 03mickeyl 07 * r1.3372 10openembedded/conf/machine/zaurus-clamshell-2.4.conf: zaurus-clamshell-2.4.conf: use old syntax for adding stuff to PREFERRED_PROVIDERS. PREFERRED_PROVIDER_virtual/kernel_ doesn't seem to work May 23 14:18:00 03mickeyl 07 * r1.3371 10openembedded/packages/orinoco/ (3 files in 2 dirs): add note for orinoco 0.15rc1, disable pci modules for now on 0.15rc2 (breaking w/ 2.6.12) May 23 14:18:04 03mickeyl 07 * r1.3370 10openembedded/packages/opie-pcmciaapplet/opie-pcmciaapplet.inc: fix opie-pcmciaapplet APPNAME May 23 14:18:25 s/behaved/related/ May 23 14:18:34 where's my sense of the english language gone May 23 14:18:36 ~lart me May 23 14:18:36 * ibot stamps mickeyl on the forehead with the official Troll marker May 23 14:19:12 mickeyl: hehe same applies to my skills.. it must the nerve gas the us troops lost while flying from Rammstein towards Iraq May 23 14:19:24 hehe May 23 14:19:44 ~lart me for forgetting words... May 23 14:19:44 * ibot beats zecke severely about the head and shoulders with a rubber chicken for forgetting words... May 23 14:19:59 mickeyl: when I do PREFERRED_PROVIDERS_append = "abc:def" May 23 14:20:17 mickeyl: it recognizes the _append as a flag May 23 14:20:55 try my small testcase in local.conf with bbread May 23 14:21:00 there _append doesn't work at all May 23 14:21:03 very odd May 23 14:21:13 or i am missing something May 23 14:21:32 mickeyl: and after I call update_data abc:def is there May 23 14:22:45 'night all May 23 14:23:01 mickeyl: so an update_data() is missing May 23 14:23:22 i see May 23 14:23:26 that would be a bbread problem right? May 23 14:27:43 mickeyl: try adding bb.data.update_data(d) before emitting the env vars May 23 14:28:06 mickeyl: I hate perl May 23 14:28:20 yahoo May 23 14:28:22 ~praise zecke May 23 14:28:24 All hail zecke! May 23 14:28:26 that's it May 23 14:28:52 i like bbread, because of its quickness May 23 14:32:54 from looking at bin/bitbake PREF_PROV_virual/kernel = "foo" should work May 23 14:32:59 dunno about spitz May 23 14:33:17 it does work, however not when adding an override May 23 14:33:24 * RP is back for a bit May 23 14:33:27 PREF_PROV_virtual/kernel_spitz just does nothing May 23 14:33:36 but we have the old syntax, so it's ok May 23 14:33:40 just something for the records May 23 14:33:44 RP: i'm confused May 23 14:34:09 mickeyl: About the virt/kern provider override? May 23 14:34:16 RP: nah. spectrum_cs. my 0.15rc1 works perfectly with 2.6.12-rc4. May 23 14:34:52 RP: i will do a test with a 2.4 embedix asap. maybe it was just there where i've seen all the problems May 23 14:34:56 mickeyl: yes PREF_PROV_virtual/kernel_spitz is the key May 23 14:35:10 RP: i remember though that we just got OOPSes with earlier 2.6 May 23 14:35:10 mickeyl: That does seem strange... May 23 14:35:25 mickeyl: I don't see how I should know _spitz is an override May 23 14:35:29 mickeyl: We do see firmware problems with 0.15rc2 May 23 14:35:47 mickeyl: Was that with or without the makefile fix? May 23 14:35:48 zecke: i guess we can ignore this special case, but we need to keep the other syntax around May 23 14:36:00 RP: without any fix. stock 0.15rc1 in oe May 23 14:36:11 mickeyl: _append, _prepend_, are speical 'override' key words May 23 14:36:16 RP: yeah, confirmed. we get a hermes timeout on 0.15rc2 before it attempts to load the fw May 23 14:36:29 mickeyl: When you do that you're building against your native kernel you know ;-) May 23 14:36:49 Maybe you had a different native kernel then... May 23 14:36:53 RP: of course May 23 14:36:55 that could be it May 23 14:37:02 RP: will you apply the fix? May 23 14:37:07 zecke: will you apply the bbread fix or should i? May 23 14:37:16 CosmicPenguin: what does 'scalar' mean for perl? May 23 14:37:26 mickeyl: I'm deep into perl ATM please commit yourself May 23 14:37:29 mickeyl: Yes, I was just waiting to see how you got on. I'll fix it and make 0.15rc1 the default May 23 14:37:31 zecke: k May 23 14:37:41 Should I nuke 0.13e-2.6? May 23 14:37:47 RP: let me wait for some experiences with 2.4.18-embedix and 2.4.20-embedix first May 23 14:37:53 RP: no please keep it alive for the time being May 23 14:39:02 mickeyl: Note I'm not saying 0.13e, just the 2.6 version... If 0.15rc1 doesn't work with 2.4, I'd just lock 2.4 users into 0.13e May 23 14:39:22 RP: oh right, of course. yeah go ahead May 23 14:39:28 03mickeyl * r210 10bitbake/bin/bbread: May 23 14:39:28 update data to make _append's etc. appear. patch courtesy BB-Meister May 23 14:39:28 Zecke May 23 14:40:00 zecke: scalar forces the expression to be evalulated in scalr context May 23 14:40:09 wb CoreDump|home May 23 14:40:15 zecke: The only time I've ever seen it used is to get the number of elements in an array May 23 14:40:18 thx mickeyl May 23 14:40:21 what's those horror stories with debug and non-debug kernels ? May 23 14:40:42 ah May 23 14:40:55 debug = kernel console is on May 23 14:41:23 wtf is a scalar context :} May 23 14:41:45 CosmicPenguin: tinderbox uses scalar a lot, but then it look like arrays May 23 14:42:26 wow bero May 23 14:43:40 ~lart RP May 23 14:43:40 :) May 23 14:43:40 * ibot drops a truckload of VAXen on RP May 23 14:43:41 CosmicPenguin: when to use $ and when @? May 23 14:44:02 03mickeyl 07 * r1.3374 10openembedded/conf/machine/zaurus-clamshell-2.4.conf: zaurus-clamshell-2.4: fix _append'ism problem with spaces May 23 14:45:16 zecke: use $ when you want to dererence the array, and @ when you want to deal with the array as a whole May 23 14:45:25 unbelievable how bugzilla changed our bug tracking moral May 23 14:45:30 we actually use the tracker May 23 14:45:35 * mickeyl shakes head May 23 14:45:36 print $array[1] vs sort @array May 23 14:45:37 mickeyl: sorry. Someone told me append always adds a space :-/ May 23 14:45:45 RP: hehe, nevermind :) May 23 14:45:58 RP: += does it, _append doesn't May 23 14:46:21 mickeyl: Ah, right. It should probably match... May 23 14:46:32 mickeyl: now if you could close bugs with #12-DONE in your commit message May 23 14:46:38 yeah. then again, sometimes it can be necessary to have exact control May 23 14:46:49 zecke: oh. ok. May 23 14:46:54 <--- bugzilla newbie May 23 14:46:56 mickeyl: no no May 23 14:47:02 mickeyl: kde's bugzilla got that feature May 23 14:47:07 mickeyl: std. bugzilla not May 23 14:47:09 _append and _prepend have to stay for the delayed operation capability and override syntax May 23 14:47:17 bleh May 23 14:47:31 kergoth`bbl: no intention to break our recipe box ;) May 23 14:47:38 yeah and sometimes it's useful to append something directly without a space May 23 14:47:39 my luggage is full of dishes, pots, pans, and alcohol. May 23 14:47:49 need to go to bed May 23 14:47:51 g'night guys May 23 14:47:53 night May 23 14:47:54 mickeyl: cya May 23 14:47:59 n8 mickeyl May 23 14:48:05 RP: good work with spectrum_cs. i'll keep you posted when i checked it on 2.4.18+2.4.20 May 23 14:48:17 mickeyl: ok, thanks - g'night May 23 14:49:09 zecke: The latest bitbake is much quicker - nice work! :) May 23 14:49:32 zecke: I've been using it with -b all day so I didn't notice until now! May 23 14:50:26 * kergoth`bbl wonders how apartments feel about tenants just leaving shit they dont want laying around :P May 23 14:50:27 RP: I'm asking myself how I can find out if the cache is still valid May 23 14:51:04 zecke: The only thing you can really use is timestamps... May 23 14:51:33 RP: yes 2*n stats May 23 14:51:47 kergoth`bbl: what are they gonna do to you - hunt you down? May 23 14:51:52 zecke: inotify/dnotify... May 23 14:51:58 for a build daemon anyway May 23 14:52:00 heh May 23 14:52:01 kergoth`bbl: yes when I'm running May 23 14:52:02 CosmicPenguin: heh May 23 14:52:05 kergoth`bbl: hehe May 23 14:52:18 CosmicPenguin: they didnt charge me a deposit when i got the place, so they cant jsut not give it back :) May 23 14:52:42 kergoth: worst that happens is you lose a deposit, so you're probably fine May 23 14:52:47 yep May 23 14:52:58 in which case i'm just about ready to go :) May 23 14:53:00 screw em May 23 14:53:04 what are you leaving? May 23 14:53:11 zecke: Can you cheat and use directory modify stamps at all? May 23 14:53:15 *burn that place* May 23 14:53:33 RP: that is the question but my posix knowledge in that regard is limited May 23 14:53:45 RP: actually it won't work because the way we use BBFILES May 23 14:54:10 gb2: heh, there s atno of stuff i dont really need. crapload of boxes from old purchases, just all sorts of kitchen crap, an old flaky 19" CRT in the closet, etc May 23 14:54:29 heh May 23 14:54:29 er s/there s atno/there's a ton/ May 23 14:54:37 are you allowed to throw CRTs out there still? May 23 14:54:56 dunno about texas. probably needs to go to a recycling facility or whatver May 23 14:54:57 it's like $25 to get rid of one here May 23 14:55:00 zecke: It would rule out groups of files. If a file wasn't in BBFILES but was modified, it would just trigger code to examine the directory contents... May 23 14:55:02 nod May 23 14:55:17 work's dumping all our old 21" CRT's at the end of this week **** BEGIN LOGGING AT Mon May 23 15:05:11 2005 May 23 15:05:25 * Hertog wonders what's wrong with KDE on linux *grin* May 23 15:06:04 i thought it was windows at first, until i saw the qemu icons May 23 15:06:06 heh May 23 15:06:07 ew damn, it looked like windows on the first glance. my apologies :) May 23 15:06:11 :) May 23 15:06:29 windows is rarely without a taskbar May 23 15:06:41 or with a kasbar May 23 15:06:43 but you can "hide" it May 23 15:06:50 i usually autohide mine May 23 15:06:53 still leaves a stripe May 23 15:06:58 thats true May 23 15:07:23 one things sucks tho... I find myself searching for my mousepointer regulary May 23 15:07:24 on XP it's completely gone IIRC May 23 15:08:01 heh, are there any first person shooter games that can deal with a multimonitor setup, showing views to your sides on the left and right monitors? May 23 15:08:07 thatd be pretty cool, if it could keep up May 23 15:08:44 gues the games could (nvidia can act as 1 monitor on those) but it's a 6600 AGP + a 5200 PCI, the last one drags it down... badly May 23 15:08:49 * kergoth`bbl nods May 23 15:09:22 that's why next setup needs two pci-x cards, not in sli mode (and am not that big on games anyway) May 23 15:10:54 (but with three of these baby's, who needs a 20'' 1600x1200 screen :)) May 23 15:11:31 Quick question about make: If I call make SOMEVAR=xxx and Makefile contains SOMEVAR = "yyy", what does SOMEVAR get set to? May 23 15:12:42 * zecke looks away May 23 15:13:43 RP: yyy May 23 15:13:50 err May 23 15:13:53 it depends May 23 15:13:59 if SOMEVAR=xxx is in the environment, its yyy May 23 15:14:05 if SOMEVAR=xxx is on the cmdline, its xxx May 23 15:14:07 er May 23 15:14:24 yeah May 23 15:14:24 heh May 23 15:14:30 xxx is in both the environment and commandline May 23 15:14:35 then its xxx May 23 15:14:42 it appears to still become yyy May 23 15:14:43 unless you're doing evil things May 23 15:14:49 oe does some evil things by default. May 23 15:14:57 unless you inherit autotools May 23 15:15:06 see the default EXTRA_OEMAKE May 23 15:15:06 kergoth`bbl: without options to make it'll be yyy May 23 15:15:16 Its using inherit module May 23 15:15:17 reenoo_: ? May 23 15:15:36 reenoo_: thats what i just said. file overrides env, and cmdline overrides file, unless you're modifying its behavior May 23 15:15:42 which we do May 23 15:16:15 make -e MAKEFLAGS= .. that causes env to override file, but only at the top level of make calls. MAKEFLAGS= causes it to not pass that -e down into submakes May 23 15:16:22 kergoth`bbl: RP was asking about "make FOO=bar" not "oe_runmake FOO=bar" May 23 15:16:28 i also interferes, iirc, with passing other vars from the cmdline into submakes May 23 15:16:42 reenoo_: if the .bb is calling 'make', its broken. May 23 15:16:52 one question: is there any way to get oe to be more verbose in the fetch target, which I think is failing, and where does it drop it's cvs pulls? May 23 15:17:09 The make commandline "inherit module" seems to generate is "make SOMEVAR=xxx" May 23 15:17:33 and yet the make file is behaving as if SOMEVAR=yyy which isn't what I want May 23 15:17:42 I could patch the Makefile... May 23 15:17:54 you havent given us enough information to diagnose further. any number of things could cause that May 23 15:18:24 for example, if the makefile does 'override SOMEVAR=yyy', then it always overrides both env and cmdline. May 23 15:19:16 The Makefile does nothing with the variable except set it and use it May 23 15:19:32 (Its orinoco-modules I'm looking at) May 23 15:19:38 zecke: so are you going with tinderbox or buildbot? May 23 15:20:14 again, by default, cmdline overrides file May 23 15:20:19 something is changing that. May 23 15:20:25 i couldnt say what without it in front of me May 23 15:21:17 RP: have you tried doing this make command outside of oe? May 23 15:21:48 and insert a $(warning SOMEVAR is $(SOMEVAR), origin is $(origin SOMEVAR)) in the makefile May 23 15:21:55 on a new empty line, not in a target May 23 15:22:11 kergoth`bbl: I'll give that a try, thanks May 23 15:22:27 origin shows where it came from May 23 15:22:31 file, environment, or commandline May 23 15:22:42 just to check sanity :) May 23 15:26:26 hmm.. there's something about conway's game of life that makes it highly addictive >:) May 23 15:29:23 * reenoo_ stops playing with gpe-life and heads of to bed May 23 15:29:45 one of these days I'll have to rewrite the view and release it.. May 23 15:30:02 anyways. 'night all May 23 15:31:31 one question: is there any way to get oe to be more verbose in the fetch target, which I think is failing, and where does it drop it's cvs pulls? May 23 15:32:24 chuckr: CVSDIR = "${DL_DIR}/cvs" in local.conf, I think May 23 15:32:51 thank you May 23 15:33:06 (that's where you can tell it where to drop it. as of more verbose... much more verbose than default doesn't seem possible ;) May 23 15:34:53 kergoth`bbl: Thanks for the tip - I've found the problem - make calls itself and doesn't preserve the commandline, hence the variables on the commandline disappear... May 23 15:36:03 perhaps, but my build, in the fetch phase, makes absolutely no comment beyond "do_fetch: completed", for the fetch target, that is May 23 15:36:08 RP: yeah, kinda figured thatd do it. oe doing make -e MAKEFLAGS= has the same sort of behavior. MAKEFLAGS is how things are passed into submakes. so we're intentionally making it only survive into the toplevel one May 23 15:36:19 s/thatd do it/thatd be it/ May 23 15:36:56 RP: just make sure its calling $(MAKE) to recurse rather than 'make', is generally the rule of thumb. and make sure nothing mangles the flag vars like that May 23 15:37:06 whew, time to start cleaning the car and packing it May 23 15:38:59 kergoth`bbl: Its calling $(MAKE). I know the variable (KERNEL_SRC) doesn't survive, I just need to work out why :) May 23 15:39:12 whats the url to that paste site? :) May 23 15:39:19 ~pastebin May 23 15:39:19 ~pastebin May 23 15:39:21 hmm... pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca May 23 15:39:34 thx May 23 15:40:26 I cant seem to clean my kernel package, checkout http://pastebin.ca/12467 for more info May 23 15:41:00 I've made some changes to the kernel bb file but it still uses the old http adress. IT should get resetted when i cleaned the package and tried to rebuild, but it didnt May 23 15:42:46 Kristoffer: There's a syntax error on a .bb file which seemed to upset it. Setting TARGET_FPU="" would be a quick hack. You could also try cleaning the package directly May 23 15:43:05 bitbake -b /somepath/somefile.bb -c clean May 23 15:43:55 shouldnt "bitbake -c clean linux-jlime-sh3; bitbake linux-jlime-sh3" do the trick? May 23 15:44:38 The file that has the dl info for quilt is packages/quilt/quilt_0.39.bb, and the line that specs the info is "SRC_URI = "cvs://anoncvs:@savannah.nongnu.org/cvsroot/quilt;method=ext;rsh=ssh;module=quilt;tag=VER_$ May 23 15:45:11 Kristoffer: That removes the working directory. It doesn't clear bitbake's .bb file cache though May 23 15:45:20 but I don't know how oe is using it, in cvs alone, or maybe ssh is in there also? It's faling, I need to test it, how to make the testing verbose? May 23 15:45:33 You can ignore my comments about FPU_TYPE as the error you have is none fatal May 23 15:45:33 RP, ah, got it thx. May 23 15:45:40 chuckr: its calling the cvs command. the cvs command is using ssh. May 23 15:45:46 and cleaning is working... May 23 15:46:12 chuckr: cvs -d:ext:anoncvs:@savannah.nongnu.org/cvsroot/quilt co -P quilt <-make sure you can run that manually at a commandline May 23 15:56:31 Look at http://pastebin.ca/12469 for a record of testing the download spec of quilt May 23 15:58:24 RP, getting same issues even after cleaning with -b May 23 15:59:09 Kristoffer: Try wiping bitbake's cache - tmp/cache May 23 15:59:36 or touch local.conf or something - I've not idea if that's working atm or not May 23 16:00:19 will do, thx for the tips May 23 16:00:32 kergoth`bbl: As far as I can see this Makefile will never work with options passed as MAKEFLAGS=... - the only way to make it work is to edit it and force a value to something sane from the enviroment May 23 16:00:37 RP: it should find out that the jhl config was changed May 23 16:00:39 is there any possible way to replace the downloaded file with a file fetched via ftp? I want this to move forward May 23 16:01:12 chuckr: Yes, place it in your sources directory and create a dummy file as an md5 stamp May 23 16:01:36 chuckr: set the taballstash? May 23 16:01:47 chuckr: if you can't use CVS you will have many problems May 23 16:04:29 I have 5 different cvs archives on my machines, I have no problem with cvs at all, I use it constantly May 23 16:10:12 RP: i thought you said that .bb wasnt using oe_runmake?> May 23 16:10:32 RP: if its using oe_runmake, you'll want to override EXTRA_OEMAKE to remove -e and MAKEFLAGS, and explicitly set the variables you need on the cmdline May 23 16:13:59 kergoth`bbl: I didn't say that - I said it was using "inherit module" May 23 16:14:59 * RP wonders about telling kergoth`bbl to read what people type but decides against it ;-) May 23 16:17:13 kergoth`bbl: The makefile is rather tortured - it calls off into the kernel's makefile which then returns to the module makefile. I'll add a simple patch to the .bb files which hack set the variable directly from the envrionment and then it'll save any confusion May 23 16:21:47 I copied the cvs test line, exactly as I got it, and got an error, mishandling passwords for the anoncvs (versus pserver) method, and got en error. Tell me, dId I miscopy your instrctions? May 23 16:22:46 RP, now its working May 23 16:28:25 RP: you mean, kinda like how i read: May 23 16:28:26 The make commandline "inherit module" seems to generate is "make SOMEVAR=xxx" May 23 16:28:33 like that? May 23 16:28:45 ;) May 23 16:30:29 gmake has the export command May 23 16:31:23 indeed. make's recursion is one big hack, really. wasnt well thought out, imho May 23 16:32:21 Yeah, and then people try to play games with including files and stuff, and then it just gets uglier from there May 23 16:32:29 ~lart microwindows May 23 16:32:29 * ibot calls microwindows on the phone ... the lights are on but nobody's home May 23 16:37:06 kergoth`bbl: That doesn't say "it doesn't use oe_runmake" ;-) May 23 16:37:49 kergoth`bbl: Although I admit, I did get the commandline wrong :) May 23 16:37:51 that says its calling make, which it isnt. May 23 16:38:00 its calling oe_runmake, which does a bunch of mangling May 23 16:38:10 and then calls make... May 23 16:38:12 if i'd known it was calling oe_runamke, i couldve told you the problem right there. May 23 16:38:17 heh May 23 16:38:56 ah well May 23 16:39:11 I've learnt something from the experience :) May 23 16:39:18 :) May 23 16:39:23 the gnu make online docs are actually quite good May 23 16:39:28 i highly recommend perusing them sometime May 23 16:40:16 I will do. I doubt I'd have been able to find out what I needed from them that quickly though May 23 16:40:35 RP: check this out.. if we call make with -e but without MAKEFLAGS, and set CFLAGS in the environment, right.. then the toplevel make tries to manipulate it.. say by adding something to the cflags, thatll take effect, but when the -e hits the submake, it pulls the one from the env again! May 23 16:40:51 pain in the ass May 23 16:41:05 so we compromise by setting it only at toplevel, and cross our fingers and hope it makes it into the submakes right May 23 16:41:06 btdt May 23 16:41:32 the alternative is to set all the variables on the commandline, which is a pain in the ass in a different way :\ May 23 16:42:14 I can see the problem :-/. Its a similar story with this variable I'm haing issues with... May 23 16:42:29 submakes don't see it... May 23 16:42:32 I have been checking patches, the patches for quilt, the oe ones, are out of date, I couldn't find even any file to apply the wiggle patch for (second file, apprently oe allows multiple patches in one file) May 23 16:43:11 that doesnt have anything to do with your problem. CP already told you the problem is that oe is expecting the sources to be in the 'quilt' directory, but they arent May 23 16:43:27 i'd advise fixing one thing at a time May 23 16:46:39 I've just tried building without a -b on the commandline on this box and got irritated when it started building a cross compiler. The existing cross compiler dates from Feb13th :) May 23 16:47:11 crazy progress May 23 16:48:17 On other boxes I rebuild it reguarly but this one's gone a long time without change... May 23 16:50:01 * Geo_KM is away: Away at the moment May 23 16:50:31 * Geo_KM is back May 23 17:02:37 03RP 07 * r1.3376 10openembedded/ (2 files in 2 dirs): orinoco-modules: PCI patch no longer needed. Someone should have wondered why it was including pci in the first place... May 23 17:02:38 *tired* May 23 17:02:46 03RP 07 * r1.3375 10openembedded/packages/orinoco/ (orinoco-modules_0.15rc1.bb orinoco-modules_0.15rc2.bb): Merge May 23 17:02:50 03RP 07 * r1.3369.1.8 10openembedded/ (6 files in 4 dirs): orinoco-modules: 0.15rc1 works with arm + 2.6 kernels - make default. Remove 0.13e-2.6 as unneeded now. Add makefile fixes. May 23 17:03:30 And on that note, I'll go to bed May 23 17:05:40 how do I dereference something with perl? May 23 17:09:17 zecke: you about? May 23 17:09:57 ooops May 23 17:09:58 shit May 23 17:13:42 treke: hey May 23 17:14:22 you wouldnt happen to have a qt windows license, would you? :p May 23 17:14:44 treke: not directly May 23 17:14:49 academic program though May 23 17:14:59 started writing an app for someone and realized that this qt3-win32 thing isnt actually complete :) May 23 17:15:06 whew May 23 17:15:10 god i'm out of shape May 23 17:16:24 treke: could try porting to Qt4, but you probably knew that... May 23 17:16:27 kergoth`bbl: good luck May 23 17:16:32 luke-jr_: I checked already May 23 17:16:38 code for qt4 doesnt seem to be up May 23 17:16:41 or I'm blind May 23 17:16:56 There's alphas or betas... May 23 17:17:03 just need to fill in the cracks in my luggage and bring them out, then somehow find a way to fit the computer, clothes, and lots of random parts and things from my closet May 23 17:17:05 * kergoth`bbl sighs May 23 17:17:06 for x11 and macosx. Haven't seen the windows code May 23 17:17:14 treke: Qt4/Win will be released when 4.0.0 is releases May 23 17:17:18 yeah May 23 17:17:21 ah May 23 17:17:22 that's what I figured May 23 17:24:07 ljp_: http://www.trolltech.com/newsroom/announcements/00000206.html <- oh oh May 23 17:24:38 ljp_: I don't like that, TT is really changing... May 23 17:27:52 zecke: you see that ben meyer is going to work for trolltech? May 23 17:28:55 treke: I've heard it, but I wonder what his tasks will be and where he will work... May 23 17:30:01 treke: http://www.informit.com/bookstore/product.asp?isbn=0131240722&rl=1 <- just for the title,picture,isbn-# May 23 17:30:30 yeah May 23 17:30:36 found it just checking local stock May 23 17:30:47 treke: do you know if he will be working on Qtopia? May 23 17:31:23 no idea May 23 17:31:31 I'm guessing not since he's going to norway May 23 17:32:38 I'm going to work for Trolltech when hell freezes and they open an office in my town... May 23 17:32:41 whatever comes first May 23 17:32:44 good nit **** BEGIN LOGGING AT Mon May 23 18:02:35 2005 **** BEGIN LOGGING AT Mon May 23 18:18:06 2005 May 23 18:34:46 hi all May 23 18:35:20 proti: is it ok? May 23 20:28:03 :q May 23 20:39:21 :} May 23 21:40:16 * Geo_KM is away: Away at the moment May 23 23:39:05 * Geo_KM is back **** ENDING LOGGING AT Mon May 23 23:59:56 2005