**** BEGIN LOGGING AT Fri Sep 09 02:59:56 2005 Sep 09 03:10:29 it is so funny that ... other may have use the gicon already why only irda get problem Sep 09 03:11:37 RP: any possible way to by pass the irda make? Sep 09 03:14:11 RP: I try both 3.4.4 and 3.4.3 .. same error :( Sep 09 03:16:11 03koen 07org.oe.dev * r7b9c0ad7... 10/packages/gpe-calendar/gpe-calendar_cvs.bb: packages/gpe-calendar/gpe-calendar_cvs.bb: add (disabled) cvs version of gpe-calendar Sep 09 03:17:16 hi dirk! Sep 09 03:17:23 Hi Richard Sep 09 03:17:28 hi dirk Sep 09 03:17:40 hi hrw|work Sep 09 03:17:45 * koen welcomes the tosa kernelgod Sep 09 03:17:50 AvengerMoJo: I really don't have the time to look into that problem at the moment I'm afraid. I'm trying to do too much at once as it is Sep 09 03:18:18 :) sorry RP Sep 09 03:19:30 hi koen Sep 09 03:19:47 do13: I pushed an updated sharpsl detection patch - I included tosa in it as it saves maintaining separately Sep 09 03:20:01 (I also added akita, borzoi and spitz) Sep 09 03:20:05 AvengerMoJo: did you start a build again from scratch using openzaurus-3.5.4 Sep 09 03:20:10 hi Sep 09 03:20:27 hm montone server on vanille is dead? Sep 09 03:20:41 RP: Is it included in 2.6.13-mm1-rc4? Sep 09 03:20:52 XorA: yes I did Sep 09 03:21:02 woglinde: if so, try one of the backup servers Sep 09 03:21:11 do13: We're up to -rc4 already?! Sep 09 03:21:22 XorA: why would the size_t causing the error Sep 09 03:21:28 do13: Its in mainline now Sep 09 03:21:44 RP: Ups typo: 2.6.13-mm1-r4. Ok Sep 09 03:22:16 do13: Ah, right - yes, the patch I submitted it Sep 09 03:22:21 s/it/is/ Sep 09 03:23:19 koen hm where is backupserver? Sep 09 03:23:21 do13: I needed to move the cxx00 series ahead of tosa in the pxa makefile/kconfig which meant I had to regenerate one of your patches Sep 09 03:23:43 Ok. I recognized this and updated my patches. Sep 09 03:23:59 woglinde: see gettingstarted Sep 09 03:24:31 monotone.nslu2-linux.org, ewi546.ewi.utwente.nl, dominion.kabel.utwente.nl and may be even more Sep 09 03:24:40 AvengerMoJo: the only time Ive seen something like that was when I accidently had wrong gcc/glibc versions together Sep 09 03:24:41 RP: I will test my image after work, if it works we can update the bb file. Sep 09 03:24:45 koen will my system affect the result ? Sep 09 03:24:54 hm should I update the phrasebook Sep 09 03:25:00 AvengerMoJo: or when using ccache you dont ccache -c between changing gcc/glibc versions Sep 09 03:25:03 XorA: will the system gcc affice the build ? Sep 09 03:25:12 do13: There are mmc changes going in which mean you can loose the timer in tosa.c as well. I have a large batch of updates sitting out of tree at the moment whilst I try to react to the mainline feedback I have Sep 09 03:25:20 XorA: what do you mean? Sep 09 03:25:31 ~hail koen for sorucepkg.bbclass Sep 09 03:25:32 * ibot bows down to koen for sorucepkg.bbclass and chants, "I'M NOT WORTHY!!" Sep 09 03:25:55 affact Sep 09 03:26:04 koen: I have two changes for it Sep 09 03:26:26 koen: 1. EXCLUDE_FROM ?= ".pc autom4te.cache" to not have configure tests in diff Sep 09 03:26:44 koen: 2. tar cvzf ${DEPLOY_DIR_SRC}/${P}.orig.tar.gz --exclude-from temp/exclude-from-file $src_tree to get it working here Sep 09 03:28:09 hrw|work: reenoo did all the work (mostly rewriting everything) Sep 09 03:28:11 koen: http://pastebin.ca/22476 in other words Sep 09 03:28:37 put the patch in bugzilla Sep 09 03:28:38 XorA: when should I do the ccache clean? Sep 09 03:28:40 sure Sep 09 03:31:17 RP: rmk send nice comments? Sep 09 03:31:46 do13: I had about 20 emails from him yesterday ;-) Sep 09 03:32:05 koen: http://bugs.openembedded.org/show_bug.cgi?id=313 Sep 09 03:34:27 RP: Yep. I read some comments on lkml: "As a general note, these patches could do with being closer to the Sep 09 03:34:27 coding style, such that things like" Sep 09 03:35:33 do13: linux-arm-kernel had a few more and then we had some discussions privately about the mmc stuff and things Sep 09 03:36:06 Ahhh. Sep 09 03:36:18 AvengerMoJo: I do it when I rm -rf tmp Sep 09 03:36:47 AvengerMoJo: but thats only because Ive seen similar wierdness and I have a FAST machine Sep 09 03:42:15 RP: Any progress on the backlight bug? Sep 09 03:43:20 lardman|work: I haven't looked yet and won't be until I finish sorting the cxx00 patches for mainline Sep 09 03:44:13 lardman|work: If you want to look into it, please feel free. It would be interesting to know if it stops inside corgi functions or if it stops in the backlight class itself Sep 09 03:44:23 you could check by adding printks Sep 09 03:44:39 lardman|work: You patch problem was most unfortunate... Sep 09 03:52:06 XorA: so .. you want me to remove the temp again? Sep 09 03:52:14 XorA: is that what you are saying? Sep 09 03:53:02 XorA: because I did that before .... and error stay .. exact error ... the same as me cleaning and updating the local.conf also Sep 09 03:55:29 rm tmp and update the tree Sep 09 03:58:28 AvengerMoJo: And clear ccache Sep 09 04:00:21 RP: I would look at it but I can't flash anymore :( Sep 09 04:00:54 lardman|work: Why not? Sep 09 04:01:55 morning folks Sep 09 04:02:18 FYI: I have just submitted a patch to enlightenment-devel that adds tslib support to ecore :D Sep 09 04:02:19 RP: Just doesn't work - I just get a black screen and no flash occurs (see: http://www.oesf.org/forums/index.php?showtopic=14894&view=getnewpost) Sep 09 04:02:42 bbiam Sep 09 04:03:03 ...... Sep 09 04:03:06 lardman|bbiam: I think I can explain this Sep 09 04:03:43 hi mickeyl Sep 09 04:03:51 .....RP taht is what I have miss to do ... Sep 09 04:04:15 morining mickeyl Sep 09 04:04:35 cache........I get two problem two because of cache related Sep 09 04:04:44 do13: It seems the pxa i2c driver has hit mainline. Amazing... Sep 09 04:06:54 RP: Yep. I got his mail Sep 09 04:07:00 morning mickeyl Sep 09 04:08:00 RP: .... same problem after ccache -c Sep 09 04:08:16 should I do it again with removing the temp? Sep 09 04:08:39 AvengerMoJo: Do you mean temp or tmp? Sep 09 04:09:04 lardman|bbiam: See my reply on the forum - I suspect you've corrupted NAND Sep 09 04:09:13 right Sep 09 04:10:35 RP: .... do you think that will help? Sep 09 04:11:50 AvengerMoJo: Do I think what will help. Did you remove tmp or temp? Sep 09 04:12:21 RP: do you think I clean up the tmp will help? :) Sep 09 04:12:46 If you haven't remove tmp before, yes. Sep 09 04:13:08 That is what people have suggested you do. Removing temp will make no difference Sep 09 04:13:12 temp != tmp Sep 09 04:13:26 hey mickeyl Sep 09 04:13:34 right :) I mean the tmp for the oebuild Sep 09 04:13:59 * AvengerMoJo is confused .. Sep 09 04:14:25 If you alter important variables like DISTRO, you need to delete tmp and start again Sep 09 04:14:31 and clear ccache Sep 09 04:17:42 I will try one more time :) Sep 09 04:17:48 * AvengerMoJo cross his finger Sep 09 04:18:35 and update the OE tree ofcourse Sep 09 04:19:00 koen will do :) in the script I wrote already Sep 09 04:19:15 * AvengerMoJo now adding ccache -c Sep 09 04:19:52 my internet connection and machine are both very slow so :( Sep 09 04:19:55 :) Sep 09 04:21:04 ccache is using the system ccache right ? :) Sep 09 04:21:10 mickeyl: pb_ : bonos nochos Sep 09 04:21:11 AvengerMoJo: I think thats -C capital actually Sep 09 04:21:13 AvengerMoJo: right Sep 09 04:21:27 AvengerMoJo: but there is kergoths ccache.bbclass to do otherwise Sep 09 04:22:00 zecke: how to tell which one it is using? Sep 09 04:22:40 AvengerMoJo: you use the system one ;) Sep 09 04:22:42 .... XorA.... may be you are right about that Sep 09 04:22:57 XorA: if you are right ... I will kill myself for that Sep 09 04:23:39 XorA: what is the different between -c and -C ? Sep 09 04:23:47 AvengerMoJo: one is cleanup, other is clear Sep 09 04:24:34 ....... Sep 09 04:24:55 * AvengerMoJo need a dictionary for that Sep 09 04:25:19 still confused :) Sep 09 04:26:13 man page help a little Sep 09 04:26:15 hehe Sep 09 04:26:15 :) Sep 09 04:26:47 if because of -c and -C lost me half a day of work ... ... I really need to shoot myself Sep 09 04:28:14 RP: Thanks. I thought the size of the kernels which could be flashed was limited...? Sep 09 04:29:08 lardman|work: OE warn if kernel is too big Sep 09 04:29:11 lardman|work: flasher does not Sep 09 04:29:40 hrw|work: Oh, where is this warning (I've not seen anything obvious)? Sep 09 04:29:46 once i accidantly mv opie-image.bin zimage.bin Sep 09 04:29:50 when i flashed that... boom Sep 09 04:30:01 hail zecke Sep 09 04:30:04 all the nand overwritten .:) Sep 09 04:30:09 hi pb_ Sep 09 04:30:09 hail mickeyl Sep 09 04:30:15 *headache* Sep 09 04:30:20 I'm glad it can be fixed - I was pulling my hair out! Sep 09 04:30:23 lardman|work: linux-openzaurus_*.bb: # Check the kernel is below the 1272*1024 byte limit for the c7x0 Sep 09 04:30:39 lardman|work: die "This kernel is too big for the c7x0 and will destroy your machine if you flash it!!!" Sep 09 04:30:53 hrw|work: Hmm, I've not had it fail on me Sep 09 04:31:08 lardman|work: check it then Sep 09 04:31:26 lardman|work: you can also change updater.sh to check for it Sep 09 04:31:55 hrw|work: Sounds like a plan Sep 09 04:32:23 plan... Sep 09 04:32:26 ~seen coredump* Sep 09 04:32:32 i haven't seen 'coredump*', hrw|work Sep 09 04:32:33 CoreDump|home: ping Sep 09 04:32:37 hi pb Sep 09 04:32:53 In the output of 'ls -rtl', the numbers are bytes? Not MS kb*100 or anything like that? Sep 09 04:33:15 du -h Sep 09 04:33:57 That's not accurate enough Sep 09 04:34:03 It gives it in Mb Sep 09 04:34:25 My largest kernel is 1281272 which is less than the 1272*1024=1302528 Sep 09 04:34:27 du -k Sep 09 04:35:10 Could it be that my kernel partition has bad blocks, so it can't store the whole lot? Sep 09 04:35:11 lardman|work: if [ "`du -b zImage.bin` lt '1302528'" ];then echo 'ops - too big';else echo 'ok';fi Sep 09 04:35:21 lardman|work: s/lt/gt Sep 09 04:36:07 Have you changed it? Where's the updater .bb file? Sep 09 04:36:18 (So I can change it if you've not) Sep 09 04:36:29 lardman|work: I did not change Sep 09 04:37:13 lardman|work: It'd be worth adding it to updater.sh Sep 09 04:37:28 Its in zaurus-updater in OE (I think) Sep 09 04:37:35 RP: where do I find the plain updater? Sep 09 04:37:45 ok, slow typing on my part, soz Sep 09 04:38:13 lardman|work: The kernel .bb should probably be altered so it makes that check directly after it does the compile Sep 09 04:38:25 In any case this won't change things for me - do I put this down to a reduced-size partition, or just a random corruption? Sep 09 04:38:39 I think it was a do_staging append at the moment which means if you use -c compile , you can get bitten Sep 09 04:39:03 RP: I was doing bitbake -b Sep 09 04:39:05 lardman|work: Are you sure you didn't flash a larger kernel when you were testing? Sep 09 04:39:20 RP: does the hx2750 have the same restriction? Sep 09 04:39:25 koen: no Sep 09 04:39:30 RP: Nope, just the ones which I produced and which were placed into tmp/deploy/images Sep 09 04:40:04 RP: in that case updater.sh would be the place to put that check Sep 09 04:40:20 RP: and it would be a good safety check too Sep 09 04:40:26 koen: We didn't have OE generating nice updater.sh when I added it Sep 09 04:40:52 koen: and I still would like the sanity check in the .bb as not everyone will change their updater.sh immediately Sep 09 04:40:55 03koen 07org.oe.dev * rdbf2ceb8... 10/packages/gpe-conf/ (3 files in 2 dirs): packages/gpe-conf/gpe-conf_cvs.bb: add (disabled) cvs version of gpe-conf and add Makefile.* to work around the CVSBUILD 'feature' Sep 09 04:42:04 RP: the kernel .bb is the wrong place for that, unless you use machine overrides Sep 09 04:42:18 RP: and since when do we support stubborn people? Sep 09 04:43:19 people 1337 enough to flash random kernel should be able to figure out how to replace a updater.sh Sep 09 04:43:49 koen: It does use machine overrides quite heavily anyway Sep 09 04:45:13 koen: It will help distro builders - lets them know in advance their kernel won't work Sep 09 04:45:48 with a warning in the middle of a build? Sep 09 04:46:25 koen: with a die Sep 09 04:46:32 ehm Sep 09 04:47:17 so my entire build fails because the kernel builds correctly, but is slightly too big? Sep 09 04:47:33 that's going to annoy me when I want to build a new rootfs Sep 09 04:47:34 It is an incorrect kernel as it won't work on the target device Sep 09 04:47:39 and *not* falsh the kernel Sep 09 04:48:09 Add an option to oe/bitbake to make you see the error then. If it was just a warning, you'd not see it Sep 09 04:48:37 maybe block of oenotes? Sep 09 04:48:54 just add it to updater.sh Sep 09 04:48:55 I'm not going to be held responsible for damageing users devices, especially when we can guard agaist it. Sep 09 04:49:13 koen: I will but the check stays in the .bb as well Sep 09 04:49:21 ok Sep 09 04:49:53 since I can build a oversized kernel and distribute that Sep 09 04:50:07 s/I/zaurus user/ Sep 09 04:50:38 Yes, it was always planned to have checks on both Sep 09 04:50:59 when I added that check, OE didn't generate updater.sh so there was nowhere to add the check to Sep 09 04:51:38 we all love encrypted scripts..... Sep 09 04:51:43 ~lart sharp Sep 09 04:51:44 * ibot puts on some milking gloves. "All right, now, sharp, this won't hurt a bit...." Sep 09 04:56:49 Right, heading home for lunch (to my serial cable), I'll let you know how it goes Sep 09 04:57:52 http://www.fooishbar.org/blog/tech/tomLord-2005-09-09-13-10.html Sep 09 04:57:59 a funny note on distributed scms Sep 09 05:12:21 argh.... Sep 09 05:35:44 03yuecelm 07org.oe.dev * r77d702f8... 10/packages/avetanabt/ (avetanabt_svn.bb avetanabt_cvs.bb): avetanabt_svn: rename to avetanabt_cvs because sources are in a cvs repository Sep 09 05:43:50 FUCK! Sep 09 05:44:14 Why if I try to build OZ-3.5.4/uclibc it ALWAYS try to rebuild ALL? Sep 09 05:44:54 or does 'bitbake' is incompatible with 'bitbake -i'? Sep 09 05:45:25 both should use the same stamps Sep 09 05:47:58 ok, guys, i'll leave for a clerical weekend. Sep 09 05:48:04 koen: please have a look at the server, will you? Sep 09 05:48:10 should it good word Sep 09 05:48:18 'should' is good word Sep 09 05:48:31 mickeyl: sure Sep 09 05:48:37 thanks. I'll be back sunday evening Sep 09 05:48:45 have fun Sep 09 05:48:52 thanks Sep 09 05:50:31 trying another attempt Sep 09 05:52:50 fsck fsck fsck Sep 09 05:53:17 ~lart bitbake Sep 09 05:53:17 * ibot whacks bitbake upside the head Sep 09 05:55:11 ~lart bootloader for changeging the cmdline Sep 09 05:55:11 * ibot brandishes Excalibur! "With this sword, I vanquish thee, bootloader!" and lops off bootloader's head for changeging the cmdline Sep 09 06:06:07 ~lart bootloaderguys for not mentioning "set" in the help command Sep 09 06:06:07 * ibot teaches bootloaderguys the basics, including how to RTM for not mentioning "set" in the help command Sep 09 06:06:21 hihi Sep 09 06:10:22 heh.. you have control over bootloader atleast Sep 09 06:17:16 RP: Worked, thanks Sep 09 06:24:46 And cpufreq is now running after my little misunderstanding with patch ;) - I have /sys/devices/system/cpu/cpu0/cpufreq/ along with it's files. Time for some testing me thinks. Sep 09 06:25:14 s/it's/its - aargh, don't tell anyone I typed that Sep 09 06:25:48 lardman|work: cool Sep 09 06:26:24 koen: Should I produce a patch to the defconfig? Sep 09 06:26:56 three more days until I can pick up OEs tosa Sep 09 06:29:39 Good Morning Sep 09 06:30:27 Anyway, cpufreq.patch and defconfig-c7x0.patch here if anyone wants to have a go themselves: http://people.bath.ac.uk/enpsgp/Zaurus/cpufreq/ Sep 09 06:31:18 Seems I was a bit eager to get alot of apps loaded in my openembedded image, and I'm receiving blockset errors on the creation of the image. My type is 'ext2' instead of 'jffs'. I'm assuming the best course of action would be to go back to a 'jffs' image? Sep 09 06:31:46 or are there some tricks for the 'ext2' type image? Sep 09 06:34:10 lardman|work: works? Sep 09 06:39:08 morning Sep 09 06:51:49 yesterday someone in here pointed out, that bitbake is able to create initrds. read some docs and tried a few things, but couldn't find a way to create initrds via bitbake. Sep 09 06:51:56 any hints? Sep 09 06:53:30 set the image type to ext2 and gzip it after wards Sep 09 06:53:51 bitbake can create any rootfilesystems. what you put in them is up to you. take a -image.bb and tweak it so that the initrd contains what you want it to contain. Sep 09 06:54:30 koen - only word of caution, don't make the 'ext2' image have too many packages. :-( That's the problem I have right now, and it won't build the image Sep 09 06:54:48 didn't know that the initrd is just a zipped image Sep 09 06:54:48 jffs type works in building the image though Sep 09 06:55:11 dwildes: you need to set the ext2 image size larger then. Sep 09 06:55:38 kergoth: would love to! where do I specify? :) Sep 09 06:55:39 Henry78: IMAGE_FSTYPES = "ext2.gz" Sep 09 06:56:05 dwildes: grep EXT2 conf/*/*.conf :P Sep 09 06:56:37 from local.conf: "Valid are jffs2, tar, cramfs" Sep 09 06:56:39 ? Sep 09 06:56:56 Henry78: that info is out of date Sep 09 06:57:09 kay. i'll try Sep 09 06:57:16 kergoth: thanks - but I already done that in the conf directory - nothing there except: distro/openslug.conf:OPENSLUG_EXT2_PROGS Sep 09 06:57:31 thanks Sep 09 06:57:47 dwildes: hmm, i thought at least one distro set it. just look in bitbake.conf then. thats where the defaults are specified, and the image command that uses it. Sep 09 07:00:58 Found it: IMAGE_ROOTFS_SIZE_ext2 = "65536". Thanks kergoth :) Sep 09 07:03:15 dwildes: so i haven't to grep, where to find the rootfs_size_ext2? Sep 09 07:04:18 np Sep 09 07:04:29 hi kergoth Sep 09 07:04:41 hey Sep 09 07:05:17 henry78: it was in the bitbake.conf Sep 09 07:05:30 thanks Sep 09 07:05:52 henry78: org.openembedded.dev/conf/bitbake.conf for me - well it is now ;-) Sep 09 07:06:01 np - thanks to kergoth for the tip Sep 09 07:06:37 hey kergoth Sep 09 07:06:39 hey pb_ Sep 09 07:06:43 is anybody of you booting from an cf-card? Sep 09 07:06:50 pb_: any chance you would come to LCA 2006? Sep 09 07:06:56 Henry78: i am Sep 09 07:07:04 Henry78: x86-based Sep 09 07:08:11 i'm on an arm, but that shoudn't matter too much. Sep 09 07:08:32 Henry78: i boot from SD card Sep 09 07:08:34 are you booting directly from your cf? Sep 09 07:08:41 Henry78: I've been doing on a pxa270 with a CF microdrive Sep 09 07:09:08 The kernel is flash based though Sep 09 07:09:11 yes - directly from CF internally as the OS, got a usb keychain to save configs & files I want to keep Sep 09 07:09:17 RP: thats pretty the thing i want to do. Sep 09 07:09:53 but on my h2200 i have to use haret - at least this was the last info i could find (pretty outdated i think) Sep 09 07:09:55 Henry78: I pushed some patches recently to help wit this Sep 09 07:10:29 RP: already in CVS? Sep 09 07:11:14 RP: any chance you could get to LCA 2006? Sep 09 07:28:30 Henry78: it's not in hh.o CVS yet Sep 09 07:29:56 i anyways have to get it flying before i start to patch anything. Sep 09 07:30:11 kergoth: ping Sep 09 07:31:31 btw Henry78, since you have a h2200, you might like to know that I just got the touchscreen to return linear results Sep 09 07:32:14 fantastic! Sep 09 07:32:38 yes, nothing's stopping us from using it fully, except a few bugs. I think we're pretty much feature-complete now Sep 09 07:32:48 mreimer: pong Sep 09 07:32:50 mreimer: congrats on the ts Sep 09 07:32:54 kergoth: thanks Sep 09 07:33:10 do you know why the wiki h2200 page is so out-of-date? Sep 09 07:33:27 Henry78: I'm too busy working on it :-) Sep 09 07:33:33 *lol* Sep 09 07:33:58 if i manage to finish a working image, i'll wirte something about it Sep 09 07:33:59 kergoth: quick question: I'm noticing that tslib coordinates can go negative sometimes. would it be ok to pin them to min or max on the given axis? Sep 09 07:34:14 Henry78: ok. check the archives too for lots of discussion Sep 09 07:34:24 Henry78: I do try to keep the to-do list up-to-date Sep 09 07:34:55 mreimer: hmm, perhaps a new 'limit' module for that? Sep 09 07:35:13 sure Sep 09 07:35:27 keeps it configurable Sep 09 07:35:29 mreimer: saw that. but only had a quick look at the archives Sep 09 07:35:30 * kergoth nods Sep 09 07:35:58 mreimer: already anxious what i might find Sep 09 07:36:14 Henry78: pretty much everything works, except for a few bugs. Sep 09 07:36:31 Henry78: the only thing I haven't done yet is to read the backup battery level Sep 09 07:39:01 mreimer: if that's the last thing not working, i'm sure you get it Sep 09 07:39:33 Henry78: I think that will be easy, but I'll probably leave that for after we get some bugs fixed. Sep 09 07:40:43 don't hesitate to do work on the bugs first, battery level isn't important, i think Sep 09 07:41:09 Henry78: right, especially the backup battery. We can already monitor the main battery level. Sep 09 07:41:56 i tried out one of the precompiled packages, and noticed it gladly Sep 09 08:35:00 hi dirk Sep 09 08:37:24 hi hrw|work Sep 09 08:37:57 hmmm Sep 09 08:38:30 03hrw 07org.oe.dev * r96432990... 10/packages/initscripts/ (3 files in 3 dirs): Sep 09 08:38:30 initscripts: added keymaps definitions for collie and poodle Sep 09 08:38:30 - they are extracted from collie_keymap.patch by RP request Sep 09 08:39:56 hrw|work: thanks. We can nuke that from the kernel now :) Sep 09 08:40:25 RP: but thats your part ;) Sep 09 08:40:43 hrw|work: I'll do it with the next commit. It will be a big one :-/ Sep 09 08:42:23 ok Sep 09 08:42:31 RP: mm1-r5? Sep 09 08:42:46 I have mm1-r4.1 now (cpufreq by lardman) Sep 09 08:43:10 I need to sort out the spitz mainline patches above all else. cpufreq can wait Sep 09 08:43:53 hmmmm Sep 09 08:44:00 need to add pptp stuff to oe.. Sep 09 08:44:16 kergoth: and mppe? Sep 09 08:44:20 * kergoth nods Sep 09 08:44:21 root@c7x0:/sys/devices/system/cpu/cpu0/cpufreq# echo ondemand >scaling_governor Sep 09 08:44:32 i need it for vpn stuff Sep 09 08:44:37 kergoth: it is there Sep 09 08:44:45 and my husky now report 100bogomips instead of 400 Sep 09 08:44:45 kergoth: pptp worked but is in nonworking Sep 09 08:45:02 cat /dev/urandom >/dev/null make it 400 again, break make it 100 Sep 09 08:45:03 kergoth: pppd can do mppe, and before it could I had a patch for ARM ;) Sep 09 08:45:08 cya later Sep 09 08:46:02 ah Sep 09 08:47:04 hrw|work, lardman|work: very cool Sep 09 08:47:04 hrw|work: Any signs of problems with it? Sep 09 08:47:27 particularly CF related? Sep 09 08:47:30 or SD? Sep 09 08:48:15 RP: just started to look Sep 09 08:50:46 by bt card was able to configure, do 'hcitool scan' Sep 09 08:51:29 cf memory card also without problem Sep 09 08:51:42 even if you change speed once its configured? Sep 09 08:51:57 http://www.samba.org/ppp/features.html Sep 09 08:52:04 I set it to 'ondemand' so its changed by kernel Sep 09 08:52:19 100-400-100-200-400-100-300-100 etc Sep 09 08:52:51 http://palmtelinux.sourceforge.net/ mentions OE Sep 09 08:52:53 root@c7x0:~# cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies Sep 09 08:52:53 99532 199065 298598 398131 Sep 09 08:52:57 hrw|work: ok. We just need to be sure the peripherals can cope with the speed changes Sep 09 08:53:31 and the kernel finally compiles again.... Sep 09 08:53:58 RP: at home I can look with hostap based card Sep 09 08:54:11 RP: aha - usbnet works all the time Sep 09 08:54:39 also after resume Sep 09 08:55:10 and its good to have autorotate working again in opie ;) Sep 09 08:55:38 hrw|work: Yes. It should stay working now as well. That is all in mainline Sep 09 08:56:47 koen: and it run Familiar Linux v0.8.3 Sep 09 08:56:48 ;) Sep 09 08:57:44 arf... my gpe imae build failed... Sep 09 08:58:35 http://pastebin.ca/22501 Sep 09 08:58:43 any advice ? Sep 09 09:00:41 alan|home: try 'rebuild module-init-tools-3.2-pre7' in bitbake shell Sep 09 09:00:53 I dont know why 3.1 are choosen... Sep 09 09:00:54 ok thanks hrw Sep 09 09:01:17 anyway I have to push 3.2-pre9 which I have in local repo Sep 09 09:01:25 but now time to home Sep 09 09:01:49 cu all Sep 09 09:01:59 have a nice weekend Sep 09 09:02:08 thanks. bye hrw Sep 09 09:04:38 RP: ...for the third time .. it fail in the same place with the same error :( Sep 09 09:05:28 anyone still here? :) Sep 09 09:05:48 AvengerMoJo: still irda ? Sep 09 09:05:57 right :( Sep 09 09:06:04 AvengerMoJo: Still here but still busy :-( Sep 09 09:06:07 is your tree up to date ? Sep 09 09:06:13 alan|home: yes Sep 09 09:06:38 ccache -C Sep 09 09:06:38 cd /opt/oe/database Sep 09 09:06:38 monotone --db=/opt/oe/database/OE.db pull monotone.vanille.de org.openembedded.dev Sep 09 09:06:38 cd /opt/oe/database/org.openembedded.dev Sep 09 09:06:38 monotone update Sep 09 09:06:38 cd /opt/oe/build Sep 09 09:06:40 bitbake opie-image Sep 09 09:10:13 RP: that's ok I understand :) I'm just wondering ... what may be the cause of it ..... .... blindlessly doing this ... really killing me :( Sep 09 09:14:05 crosscompile gcc 3.4.4 is correct right? Sep 09 09:18:35 AvengerMoJo: I use 3.4.3 but others happily use 3.4.4. The above looks ok to me... Sep 09 09:22:57 RP: it look to me is the include file problem ..... I build with normal make in my system ... Sep 09 09:22:58 RP: Are you running opie? Sep 09 09:23:01 RP: it pass Sep 09 09:23:30 RP: I take out the orginal include path ... and find out missing irda.h Sep 09 09:23:51 RP: there are multiple irda.h in the staging ... which one should be used? Sep 09 09:25:02 RP: it pass :) Sep 09 09:25:19 RP: using the irda package path... seem to be the configuration problem then Sep 09 09:26:23 arm-linux-gcc -march=armv5te -mtune=xscale -I/opt/oe/build/oetmp/staging/arm-linux/include -I../include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -c irattach.c <--- this is the correct compile line Sep 09 09:26:48 arm-linux-gcc -march=armv5te -mtune=xscale -I/opt/oe/build/oetmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -I/usr/include -I../include/ -c irattach.c <---- this is the bitbike build line Sep 09 09:27:18 so ... /usr/include really mess it up Sep 09 09:27:34 RP: ... can I fix this? :) Sep 09 09:29:42 AvengerMoJo: Yes, that is a problem - it was pulling in files from your system instead of from staging. The question is where did like /usr/include come from? Sep 09 09:30:12 RP: that's the first feeling ... and question remain .... I don't know where it is coming from :) Sep 09 09:30:30 RP: could you tell me how to trace it ... I don't know the build system good enought to tell Sep 09 09:30:36 Have a look in the run_* files in the temp directory Sep 09 09:30:55 RP: ok :) thank ... I think I can fix this now hehe :P Sep 09 09:31:04 See if the configure/compile ones show anything obvious Sep 09 09:31:44 export includedir="/usr/include" Sep 09 09:32:47 it is clearly using export includedir="/usr/include" Sep 09 09:33:39 in oe_runconf() Sep 09 09:33:46 it is just like normal configure Sep 09 09:33:58 doesn't look like it is using the stage at all Sep 09 09:34:50 where does the run_* file generated from? Sep 09 09:35:27 bitbake generates them. Sep 09 09:35:34 probably from one of the class files Sep 09 09:35:55 in openembedded/classes/*.bbclass Sep 09 09:36:55 ok I will look into that also Sep 09 09:39:46 RP: you said bl was working for Opie users? Sep 09 09:40:01 seem to be the oe_runconf () in the autotools.bbclass is not right Sep 09 09:40:48 RP: where is the "includedir" assigned from? Sep 09 09:43:21 lardman|work: It works here. Does it not work for you? Sep 09 09:43:41 AvengerMoJo: Offhand, no idea I'm afraid Sep 09 09:43:52 RP: cross.bbclass:prefix = "${CROSS_DIR}" Sep 09 09:44:08 I trace all the var to CROSS_DIR Sep 09 09:44:43 where is CROSS_DIR should be define? ... local.conf? Sep 09 09:44:59 bitbake.conf? Sep 09 09:45:24 let me see Sep 09 09:45:26 RP: I use GPE and it doesn't - I was wondering how the open method in libopie2 was different from cat'ing a value into the file Sep 09 09:46:33 RP: CROSS_DIR = "${TMPDIR}/cross" this will be correct then ... Sep 09 09:46:50 RP: how come it end up getting /usr/include .... Sep 09 09:49:40 RP: this doesn't look right to you right Sep 09 09:49:42 http://pastebin.ca/22505 Sep 09 09:51:29 that should come from autotools.bbclass this file right Sep 09 09:52:19 then cross.bbclass:includedir = "${exec_prefix}/include" Sep 09 09:57:50 cross.bbclass:prefix = "${CROSS_DIR}" Sep 09 09:57:50 cross.bbclass:exec_prefix = "${prefix}" Sep 09 09:58:19 ....now I'm back to square one :) why it is become /usr/include? Sep 09 09:58:34 or where does it come from? :) Sep 09 09:59:36 AvengerMoJo: what becomes /usr/include? Sep 09 09:59:55 I'm off, night all Sep 09 10:00:03 zecke: I my irda build Sep 09 10:00:30 make: Entering directory `/opt/oe/build/oetmp/work/irda-utils-0.9.16-r2/irda-utils-0.9.16/irattach' Sep 09 10:00:31 ccache arm-linux-gcc -march=armv5te -mtune=xscale -I/opt/oe/build/oetmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 ******-I/usr/include******** -I../include/ -c irattach.c Sep 09 10:00:41 the star is where cause my error Sep 09 10:01:02 and I don't know where is it coming from Sep 09 10:01:32 AvengerMoJo: is that in the Makefile? Sep 09 10:01:44 AvengerMoJo: it is likely not to come from cross.bbclass Sep 09 10:02:11 zecke: right I have just come to that ... the bbclass file is good Sep 09 10:03:16 yay free extra shot. 4 shot mocha = happy kergoth Sep 09 10:03:28 zecke: it is also in the Makefile Sep 09 10:03:30 must be starbucks Sep 09 10:03:43 nope Sep 09 10:03:43 zecke: should I patch the Make file? Sep 09 10:03:48 they cant actually make singles with their new machines Sep 09 10:03:54 ~lart Rasmus Sep 09 10:03:55 * ibot cats /dev/urandom into Rasmus's ear Sep 09 10:03:57 ~lart Wez Sep 09 10:03:58 * ibot hooks into a hydrant and hoses Wez down Sep 09 10:04:00 ~lart Pierre Sep 09 10:04:01 * ibot cuts off Pierre's head with a halberd that could have been a little bit sharper Sep 09 10:04:17 JustinP, more people asking about e-image? Sep 09 10:04:25 bwahaha! Sep 09 10:04:35 you obviously don't know the PHP world Sep 09 10:04:43 no, it's unrelated to this channel, actually Sep 09 10:04:50 no, i generally avoid php unless i have to write it Sep 09 10:04:52 I just wanted to express myself *somewhere* Sep 09 10:04:58 hehe Sep 09 10:05:03 ~botsnack Sep 09 10:05:04 kergoth: :) Sep 09 10:06:10 AvengerMoJo: so if it is in the Makefile, it will be used Sep 09 10:06:21 bah time to run downtown then come back and finish my pcb Sep 09 10:06:27 AvengerMoJo: now find out why it is there Sep 09 10:06:43 zecke: no configurare has been apply Sep 09 10:06:51 configure is empty Sep 09 10:06:52 i am curious ... is it normal -O2 is used? Sep 09 10:07:02 so ... Makefile must be from the source Sep 09 10:07:38 seems a bit odd to try and optimise for an arm/pxa chip Sep 09 10:07:58 or i've just never noticed Sep 09 10:08:04 zecke: SYS_INCLUDES = -I/usr/include Sep 09 10:08:09 zecke: from the source Sep 09 10:08:16 zecke: in the make file Sep 09 10:08:38 AvengerMoJo: so patch it ;) Sep 09 10:08:49 and file a bug report at bugs.openembedded.org Sep 09 10:09:23 zecke: ok ... I need to make sure it is working first Sep 09 10:09:36 zecke: so I add the patch in the bb file first right? Sep 09 10:09:46 :-) Sep 09 10:09:51 hi foka Sep 09 10:09:52 :P Sep 09 10:10:01 AvengerMoJo: it is up to you, but I would do it the following way Sep 09 10:10:08 AvengerMoJo: quilt new fixincludes.patch Sep 09 10:10:16 foka: I think I finally figure out the problem Sep 09 10:10:17 AvengerMoJo: quilt edit the 'thefile' Sep 09 10:10:20 AvengerMoJo: change it Sep 09 10:10:31 AvengerMoJo: ../temp/run.do_configure Sep 09 10:10:39 AvengerMoJo: ../temp/run.do_compile Sep 09 10:10:43 AvengerMoJo: and if it compiles Sep 09 10:10:47 AvengerMoJo: quilt refresh Sep 09 10:10:59 Good for you! Found a bug in the OE configure scripts? :-) Sep 09 10:11:00 AvengerMoJo: and then you could add it to the bb file and then do a bugreport Sep 09 10:11:17 zecke: thanks alot Sep 09 10:13:28 foka: it is in the source Sep 09 10:14:42 I see. Sep 09 10:18:41 "It is not that I "think" this, it is that this is how it is. If *you* Sep 09 10:18:43 had ever built any images form OE you would understand." Sep 09 10:18:52 what is Justin Patrins nick? Sep 09 10:19:34 zecke: JustinP Sep 09 10:19:35 justinp Sep 09 10:20:37 now he is on ignore... idiot Sep 09 10:20:53 good plan Sep 09 10:21:52 * pb_ go home now Sep 09 10:21:54 later all Sep 09 10:22:54 zecke: done Sep 09 10:23:24 is the patch automagically in the bb file now? Sep 09 10:24:25 I guess not :) Sep 09 10:26:19 zecke: I just add this fixincludes.patch file into the bb file right? Sep 09 10:29:50 sorry, general liinux question... i want to export some environment variables from a script (ie. $BBPATH etc) so i can just run ./env-oe . how can i d o this? Sep 09 10:30:42 you cant _run_ a script and have its variables go into your current shell Sep 09 10:30:55 because running it forks a new process. variables cant go into the parent process, only children Sep 09 10:30:57 JustinTodd: . script Sep 09 10:31:01 JustinTodd: source script Sep 09 10:31:06 you want to do a . ./oe-env or source Sep 09 10:31:10 yeah Sep 09 10:31:12 oh ok Sep 09 10:31:13 both should work, both are bashism? Sep 09 10:31:16 that runs the contents in the current shell Sep 09 10:31:20 rather than spawning a new one Sep 09 10:31:36 zecke: dunno if they're susv3, but '.' is supported by ash, 'source' is not. Sep 09 10:31:44 s/ash/busybox ash/ Sep 09 10:31:49 so the former would be more portable Sep 09 10:31:56 we had a discussion on #maemo about it Sep 09 10:32:00 right because #!/bin/sh spawns a new shell and the enviroment variables pertain only to that shell... even if you export them. Sep 09 10:32:13 the conclusion was '.' is better because it saves precious flash :) Sep 09 10:32:18 thx Sep 09 10:32:27 right Sep 09 10:32:29 hehe Sep 09 10:33:32 * kergoth hugs ccache Sep 09 10:33:42 * AvengerMoJo pass irda now!!! Sep 09 10:33:48 grats Sep 09 10:38:34 how long does bugs server create an account email? Sep 09 10:39:39 AvengerMoJo: it should work almost immediately Sep 09 10:40:05 zecke: that's what I through... Sep 09 10:40:47 AvengerMoJo: maybe check with your spam filter Sep 09 10:41:20 hum Sep 09 10:46:22 "Mythbusters: Cooling a six-pack" Sep 09 10:46:27 * koen turns on the tv Sep 09 10:46:59 koen: I want that in Germany as well... Sep 09 10:47:11 you don't have discovery channel? Sep 09 10:47:46 koen: I've in pay-tv Sep 09 10:48:09 * koen wants DVB-T Sep 09 10:48:20 "This HOWTO is intended to be very much a Bazaar style development. If it were to be any more open, bits would fall out." Sep 09 10:49:01 kergoth: http://www.fooishbar.org/blog/tech/tomLord-2005-09-09-13-10.html Sep 09 10:50:07 heh Sep 09 10:50:08 the best part: "It's sort of ironic that the lead developer in a distributed versioning system was able to have such a chokehold on development, but it was true." Sep 09 10:50:17 hehe Sep 09 10:52:53 zecke: I think my email is filtering the bugs server Sep 09 10:53:42 and geoinformatics people present? Sep 09 10:53:43 koen: haha Fluff will not use Qt :} Sep 09 10:53:54 if so: FIX YOUR AUTOFOO Sep 09 10:53:59 zecke: no? Sep 09 10:54:05 evil efl tech? Sep 09 10:54:20 koen: it (Qt) forbids developing of a commercial alternative to Qtopia Sep 09 10:54:29 heh Sep 09 10:54:32 how open Sep 09 10:55:40 zecke: did you see gtk-qt? Sep 09 10:55:57 it's a gtk theme engine which uses qt as rendering backend Sep 09 10:56:01 * CosmicPenguin pukes Sep 09 10:56:33 CosmicPenguin: is thunar in OE yet? Sep 09 10:57:07 No Sep 09 10:57:10 never heard of it before Sep 09 10:57:22 I thought you were the xfce man Sep 09 10:57:47 http://thunar.xfce.org/wiki/ Sep 09 10:58:07 I haven't been keeping up on Xfce advancements lately Sep 09 10:58:26 But if its better then xffm, then I'm in Sep 09 10:59:08 that and it'll order your pizza too Sep 09 10:59:45 koen: well that means Gtk-- or reviving of Inti Sep 09 11:05:06 koen: any chance you could get to LCA? Sep 09 11:12:01 mithro: unlikely, but I'd like to keep dreaming :) Sep 09 11:15:01 how much would it cost for you to get there? Sep 09 11:15:14 zecke bug report submited Sep 09 11:15:20 where to send the patch Sep 09 11:15:43 mithro: sydney? Sep 09 11:15:55 nope in New Zeeland Sep 09 11:16:32 lca is in june? Sep 09 11:16:45 * koen opens the linux.conf.au website Sep 09 11:17:09 linux.conf.au is Australia's national Linux conference. linux.conf.au 2006 will be held in Dunedin, New Zealand from the 23rd - 28th January 2006 at the University of Otago. Sep 09 11:17:23 AvengerMoJo: append it to the bug report Sep 09 11:17:35 zecke: ok Sep 09 11:19:21 mickey|weekend: what do we do? let us be forced to swallow Qt4? Sep 09 11:19:28 mickey|weekend: use a different toolkit? Sep 09 11:19:45 zecke: mickey|weekend will be back on sunday Sep 09 11:21:02 mithro: about E1400 according to expedia.com, but I think some better searching can shave off ~E250 Sep 09 11:21:35 hmmm Sep 09 11:22:15 so about 1200 euro? Sep 09 11:23:09 kergoth: how much would it cost you? Sep 09 11:23:17 * kergoth checks Sep 09 11:23:36 zecke: done Sep 09 11:24:08 ow. Sep 09 11:24:54 mithro: that is straight to dunedin Sep 09 11:25:01 * koen checks the fare to wellington Sep 09 11:25:12 2,460.87 USD for the flight alone, from here Sep 09 11:26:07 ah, E800 to wellington Sep 09 11:26:22 i looked at doing that this past year, it was US$1400 for the flight Sep 09 11:26:23 ~lart gdal authors Sep 09 11:26:23 * ibot keeps mailing gdal authors free America Online CDs until he drowns Sep 09 11:26:51 Try going round the world, it's often cheaper. Sep 09 11:26:58 and i'd have had to bring my gf too, so that's $2800 Sep 09 11:27:05 if I pass the stagingdir to gdal's configure it automatically prepends /usr/lib/ .... Sep 09 11:27:11 hmm, yeah, gf.. Sep 09 11:27:13 * kergoth ponders Sep 09 11:27:28 03jbowler 07org.oe.dev * r9dc7ab94... 10/packages/mt-daapd/ (4 files in 2 dirs): Sep 09 11:27:29 mt-daapd: fix 0.2.1.1 to work with 'iTunes5' Sep 09 11:27:29 The fix applies a patch copied from www.mt-daapd.org, the patch is copied Sep 09 11:27:29 because the version at the web site has path names (well, no path names) Sep 09 11:27:29 which don't work with OE. This commit also adds a .bb for mt-daapd_0.2.2, Sep 09 11:27:29 but since that hasn't been released it doesn't work (yet) and is DPed out. Sep 09 11:27:42 yeah, can you really get away with a vacation to .au alone? :) Sep 09 11:27:49 indeed Sep 09 11:27:52 doubtful :) Sep 09 11:28:09 uh oh Sep 09 11:28:15 * koen hides this window from his gf Sep 09 11:28:17 I figure we'll go in a year or two, and take some extra time and go to new zealand Sep 09 11:28:18 lol Sep 09 11:29:10 * mithro ponders Sep 09 11:29:20 * kergoth ponders too Sep 09 11:30:58 i'm just pondering if it's worth going on a fundraising drive to get some of you to LCA :P Sep 09 11:33:12 koen: I still hope it is not true... Sep 09 11:33:28 zecke: the no competition clause? Sep 09 11:33:30 mithro: CELF? Sep 09 11:33:41 koen: right, I think the Qtopia license has it Sep 09 11:33:44 CELF? Sep 09 11:33:55 USD2101-2521 NY->SeaTac->honolulu->bangkok->kuala lumpur->sydney->bangkok(sic)->armman->casablanca->ny Sep 09 11:33:58 consumer electronic linux forum Sep 09 11:34:04 http://www.qantas.com.au/regions/dyn/nl/specials/ActionFaresNZ?lk=promonl Sep 09 11:34:29 jbowler: wow thats a way to go Sep 09 11:34:37 we could rent a boat and do a round trip Sep 09 11:34:47 from europe to Amercia to New Zealand Sep 09 11:34:52 Australia is annoying - those guys need to shift it north a few thousand miles Sep 09 11:35:20 jbowler: it's actually in New Zealand next year :P Sep 09 11:36:17 Ah, then stay down there a year and bum around. Sep 09 11:36:47 koen: my ibook rocked ;) Sep 09 11:37:03 mithro: http://tree.celinuxforum.org/CelfPubWiki/OpenEmbedded Sep 09 11:37:21 zecke: that's expected behaviour :) Sep 09 11:38:01 haha, inotify crashed my kernel Sep 09 11:38:12 koen: now my gal al... owns it... Sep 09 11:38:16 he he - my name is mentioned there :P Sep 09 11:39:53 zecke: D'oh Sep 09 11:42:25 koen: I hope to have another one in a week or so... Sep 09 11:43:33 hmm, what ipkgs do I need to install to run mb-desktop-xine? I've installed mb-desktop-xine already, but can't see how to run it Sep 09 11:43:54 maybe i should hit them up :P Sep 09 11:44:30 mreimer: see the mb-xine doc Sep 09 11:44:34 it's a plugin Sep 09 11:44:54 or module (whatever mb calls it) Sep 09 11:48:17 any suggestions on how to make it do something? Sep 09 11:51:55 mreimer: there's a doc in the source tree which describes what to do Sep 09 11:52:01 reading it Sep 09 11:52:08 I can't remember how to do it right now Sep 09 11:52:10 re Sep 09 11:52:16 np, now I'm on the trail Sep 09 12:02:43 bummer, matchbox-desktop hangs when I tap on an mp3 Sep 09 12:12:12 http://kergoth.com/files/routerideas.lua Sep 09 12:18:20 kergoth: what's that for? Sep 09 12:19:10 just toying around with a possible firewalling/routing configuration tool that lets you define your policies independent from the actual implementation (i.e. iptables commands or direct manipulation of the tables via netlink) Sep 09 12:21:20 ah Sep 09 12:21:31 interesting Sep 09 12:21:36 I've never seen lua before Sep 09 12:21:52 yet another scripting language I don't understand Sep 09 12:21:54 it'd be nice if there was a very forgiving language for users to write in Sep 09 12:22:22 lua is dead simple, often used as the logic for games, since even the UI folks with no programming experience can use it Sep 09 12:22:44 yeah, it's got lots of {}'s though Sep 09 12:22:46 lua is also very small Sep 09 12:23:07 {} is a table, the main lua data structure Sep 09 12:23:12 used for everything from arrays to classes Sep 09 12:23:23 it looked like that Sep 09 12:23:49 a.B is identical to a["B"] :) Sep 09 12:23:53 functions are first class objects Sep 09 12:23:58 I've learned that the more idiot proof you can make a system, the better though. Sep 09 12:24:01 nod Sep 09 12:24:53 I maintained our demo for a year or two, I learned all the sorts of stupid things marketing people do when they're capable of clicking on things. Sep 09 12:25:01 what about idiot proofing an idiot trap? Sep 09 12:26:59 the most frightening are the people with some engineering background, but not enough. so they think they can go in and fix a problem. Sep 09 12:28:13 we can't ? Sep 09 12:29:09 "sales engineer" sort of people Sep 09 12:29:23 who go off to sell things, and take a demo with them Sep 09 12:29:33 demo goes on the fritz, and they go in and start editing things Sep 09 12:29:45 hehe Sep 09 12:29:48 or go to radio shack and buy a soldering iron Sep 09 12:30:11 cya tomorrow Sep 09 12:30:16 that one really scared me, demo hardware was fortunately fine. Sep 09 12:30:28 well, optimally the device will have web and menu interfaces, but i like having the full flexibility of lua for policy definitions for the advanced user, rather than representing every possible complex setup via those interfaces Sep 09 12:30:42 plus itd be cool to be able to emit bsd pf or cisco from the same policies Sep 09 12:30:46 * kergoth ponders Sep 09 12:30:59 sounds interesting Sep 09 12:37:05 hey guys Sep 09 12:37:17 re Sep 09 12:37:35 let's say I have support for a new board.....what would be the best way to go about getting it merged in? Sep 09 12:44:53 MSpin: submit the files to bugzilla Sep 09 12:45:12 koen|slug: k Sep 09 12:54:12 there, #316 :) Sep 09 12:58:02 gb2: what about something like http://kergoth.com/files/routerideas2.lua , for a less confusing syntax for the user? Sep 09 12:58:15 just as valid lua as the other, just no {} Sep 09 13:02:57 kergoth: less confusing, less sexy Sep 09 13:03:02 hehe Sep 09 13:03:52 the PF.manip:new()'s are unnecessary, the initialization of the rule could initialize those as well Sep 09 13:03:58 make it even easier Sep 09 13:04:25 I'm going to sleep now ;) Sep 09 13:04:30 night Sep 09 13:06:42 kergoth: somewhat clearer. Sep 09 13:08:32 could kill the call to add_rule() and add a field in the rule that specifies which chain, in which table, its in, to make it a bit more declarative Sep 09 13:08:47 rule:new() would queue it up for addition to the chain implicitly Sep 09 13:08:58 * kergoth thinks Sep 09 13:09:21 it really feels like you're giving the user far more than enough rope though Sep 09 13:10:24 could go with a purely declarative non-lua configuration syntax Sep 09 13:10:30 would limit the flexibility, but shorten that rope Sep 09 13:10:35 yeah Sep 09 13:10:50 cisco-ish style stuff i think is designed that way Sep 09 13:11:16 (i really hate IOS access-list stuff though) Sep 09 13:11:52 i just want a user to be able to say "look, _this_ is what i want. do what you need to do to make it happen." Sep 09 13:12:17 yeah Sep 09 13:14:06 is there anything other than lex for writing parsers? Sep 09 13:14:11 something sane perhaps Sep 09 13:15:19 good question. 'lemon' seems a bit nicer than bison for that piece, but i dunno about lexers Sep 09 13:15:40 kergoth: lemon rocks, bison is yeah 30 years behind Sep 09 13:15:56 kergoth: Marc Singer used lemon on his Bitbake C parser Sep 09 13:16:00 yeah Sep 09 13:16:33 kergoth: flex is good as a lexer Sep 09 13:17:02 http://re2c.org/ looks nice Sep 09 13:17:26 "re2c is on the order of 2-3 times faster Sep 09 13:17:26 than a flex based scanner, and its input model is much more Sep 09 13:17:26 flexible." Sep 09 13:18:13 zecke: re2c leaves more in your hands, so you can do things like use mmap() .. Sep 09 13:18:17 easily Sep 09 13:18:21 so it seems, anyway Sep 09 13:18:58 kergoth: ah that is possible with flex as well Sep 09 13:19:05 what's mmap have to do with lexing? Sep 09 13:19:26 gb2: reading from the file Sep 09 13:19:36 mechanism for getting input, rather than read() Sep 09 13:19:46 gb2: you lex any input Sep 09 13:20:01 ah, flex is limited to read()'ing? Sep 09 13:20:13 i thought it was, but i havent messed with it in ages Sep 09 13:20:54 YYLEX_INPUT can be used to even lex from buffers or mmaped files (not zero copy though) Sep 09 13:20:59 I wish there was a reliable way to really discard data so it gets freed rather than the kernel keeping stuff around in the page cache Sep 09 13:24:03 kergoth: http://cvs.arm.linux.org.uk/cgi/viewcvs.cgi/tslib/configure.ac.diff?r1=1.2&r2=1.3 Sep 09 13:24:17 kergoth: could you change 'dejitter' to something else ;) Sep 09 13:24:28 hehe Sep 09 13:24:31 forgot about that Sep 09 13:24:38 * kergoth emails himself so he wont forget Sep 09 13:25:02 that module is obsolete now anyways Sep 09 13:25:33 hm? Sep 09 13:25:45 oh, the h2200 Sep 09 13:25:58 where is it being linearized then? Sep 09 13:26:26 on the asic Sep 09 13:26:33 ah Sep 09 13:26:36 kergoth: mreimer just obsoleted it Sep 09 13:26:37 mreimer found the right registers Sep 09 13:26:41 cool Sep 09 13:27:02 well then, we should yank it from cvs and tag the damn 1.0 tslib release already Sep 09 13:27:08 * kergoth keeps forgetting to do that :P Sep 09 13:27:11 damn I'm too lazy into looking why gnutls.org does not serve files... Sep 09 13:31:24 zecke, because there is nothing there? Sep 09 13:33:00 and ftp.gnutls.org is at max users Sep 09 13:34:34 M$ is better, I never get MAX users there ;) Sep 09 13:34:40 emte: good explanation Sep 09 13:34:49 lol Sep 09 13:35:46 well you did say you were feeling a bit lazy on the url entry front Sep 09 13:36:20 i just needed 4 charachters and a right mouse-click Sep 09 13:36:24 emte: I could download it from somewhere else... Sep 09 13:36:36 nah, what fun is that? Sep 09 13:36:38 emte: but I'm still tired from last month... Sep 09 13:37:04 actually ... did gaim just release a new version? Sep 09 13:37:16 that might explain the max user thing Sep 09 13:37:34 emte: nope Sep 09 13:37:43 man tomorrow is simpad and oe hacking... Sep 09 13:37:56 zecke: was that siemens thingie dvb-h or dvb-t? Sep 09 13:38:02 tell you what? ... i'll send you to the beach instead Sep 09 13:38:03 dvb-h Sep 09 13:38:27 not sure what authority i have but ... Sep 09 13:39:12 koen: terratec cynergy t² is an awesome device Sep 09 13:39:17 i shoudl go buy a lotto ticket ... i need a decent eda Sep 09 13:39:30 koen: actually I know the guys who designed it Sep 09 13:39:33 koen: www.qanu.de Sep 09 13:40:16 oooh Sep 09 13:40:21 that looks cool Sep 09 13:40:49 koen: you can even tune the dvb-h channels and it works with dvbnet Sep 09 13:40:57 nice Sep 09 13:41:00 koen: and it works with my ibook Sep 09 13:41:11 koen: it worked with the ibook i had and lost to al... Sep 09 13:41:33 *g* Sep 09 13:42:05 woglinde: you know her ;) Sep 09 13:42:08 zecle you wanted to go to bed Sep 09 13:42:24 woglinde: I'm waiting for my breezy update to finish... Sep 09 13:43:13 root@h3900:/usr/bin# /etc/init.d/ipaq-sleep stop Sep 09 13:43:13 Stopping sleep daemon: ipaq-sleep/sbin/start-stop-daemon: unrecognized option `--oknodo' Sep 09 13:43:39 hmm, that old problem again Sep 09 13:44:09 busybox? Sep 09 13:44:19 3yes Sep 09 13:44:22 -3 Sep 09 13:44:41 this is 1.00 Sep 09 13:45:10 I don't think any busybox version knows about oknodo Sep 09 13:47:31 mreimer: congrats again ;) Sep 09 13:47:37 zecke: how much does that dvb-h thingy cost? Sep 09 13:47:48 thanks zecke :-) Sep 09 13:48:01 koen: siemens one? Sep 09 13:48:08 crap Sep 09 13:48:10 dvb-t Sep 09 13:48:13 the qanu one Sep 09 13:48:23 koen: from terratec ~80 euro Sep 09 13:48:48 * kergoth ponders Sep 09 13:50:05 now I only have to wait 3 years or move to 'a large part of the country' Sep 09 13:50:23 for 'a large part of the country' read 'Amsterdam' Sep 09 13:57:52 * kergoth sends the zecke/justinp thread to /dev/null Sep 09 13:59:01 ~cluebat freyther Sep 09 13:59:03 * ibot pulls out a ClueBat (tm) and thwaps freyther. Sep 09 13:59:42 I mailed him in private trying to explain but apparently he can't read. Sep 09 13:59:43 you both know how bitbake works, so stop filling my mailbox with worthless drivel. Sep 09 13:59:49 no, he doesn't Sep 09 14:00:01 and I wanted to make sure that the OP understood as well Sep 09 14:00:18 perhaps you're unfamiliar with the concept of a communications breakdown. Sep 09 14:00:25 and anyone else who happened to be listening Sep 09 14:00:32 which is clearly what was happening there Sep 09 14:00:36 I explained in detail Sep 09 14:00:44 and he even thinks my patch does nothing when it clearly does Sep 09 14:00:50 that isnt what he said. Sep 09 14:01:27 well that's sure what it sounds like Sep 09 14:01:39 I read his comment as "your patch does the same thing as a floating CVSDATE" Sep 09 14:02:03 no, he thought you were claiming that the existing behavior was other than what it was, and that you then wrote a patch that changed it to what you had been claiming it was all along Sep 09 14:02:06 communications breakdown. Sep 09 14:02:11 he thought you were claiming something that you werent Sep 09 14:02:16 now stop filling my mailbox with crap Sep 09 14:02:29 * kergoth goes back to doing networky things Sep 09 14:02:45 kergoth: q: removing the tarball should make a floating cvsdate update DL_DIR/cvs/foo, right? Sep 09 14:02:56 bb should not* Sep 09 14:03:25 koen: floating cvsdate doesnt do a cvs up, it does a cvs up -D with the date, at least last time i touched the code. Sep 09 14:03:36 to ensure that the tarballs are always reproducable Sep 09 14:03:46 that's what I thought Sep 09 14:03:52 reproducability is a major concern in a build infrastructure like ours Sep 09 14:04:33 which is why the 'removing the tarballs makes it fetch again' was wrong in that thread Sep 09 14:04:59 thats correct, it would do an up, but not change anything, just retar the existing directory Sep 09 14:05:27 does a CVSDATE of today pull current HEAD or the code from 00:00 of today? Sep 09 14:05:45 it isnt current head. it does a cvs up -D [today's date] Sep 09 14:05:51 so its up to cvs what time it pulls from Sep 09 14:05:52 likely 00:00 Sep 09 14:05:54 I understand that Sep 09 14:05:59 but you'd want to read the docs on cvs to find that out Sep 09 14:06:06 well, this is why I was asking, I wasn't sure what CVS would do Sep 09 14:06:24 "man cvs" doesn't say AFAICS Sep 09 14:07:11 someone proved it kept at midnight a year ago Sep 09 14:07:42 midnight as in 00:00 or as in 24:00? Sep 09 14:07:45 I assume 00:00 Sep 09 14:08:22 00:00 Sep 09 14:12:32 and I suppose the patch was a bit flawed as it wasn't specifying a -D and previous updates *could* have Sep 09 14:12:46 so the patch has been updated to use "-D now" instead of nothing Sep 09 14:13:05 i read "Use the most recent revision no later than DATE_SPEC." from the cvs info page as meaning that 20050101 = 20050101 00:00, but yeah, its unclear. Sep 09 14:14:51 I had previously assumed that it would continue to update until the end of the day but I suppose that would defeat the purpose of -D Sep 09 14:15:18 indeed Sep 09 14:15:30 The specification is "sticky" when you use it to make a private Sep 09 14:15:30 copy of a source file; that is, when you get a working file using Sep 09 14:15:30 `-D', CVS records the date you specified, so that further updates Sep 09 14:15:30 in the same directory will use the same date (for more information Sep 09 14:15:30 on sticky tags/dates, *note Sticky tags::). Sep 09 14:15:32 heh Sep 09 14:15:35 anyway, this feature isn't meant for production systems. It's supposed to be a handy fix for people using OE while developing an application Sep 09 14:15:42 nod Sep 09 14:16:14 * koen always sets cvsdates to tomorrow to test Sep 09 14:16:16 just as floating CVSDATEs *shouldn't* be used in OE....unfortunately we still have some (many) Sep 09 14:16:22 though i'd argue that the same functionality could be implemented via a new do_fetch/do_unpack Sep 09 14:16:29 opie is one big cvsdate for one Sep 09 14:16:36 if you're developing, you probably have a checkout handy that you're working in Sep 09 14:16:45 in which case you want oe to build in there, and not fetch at all Sep 09 14:16:49 or unpack, for that matter Sep 09 14:16:52 S = /path/to/tree Sep 09 14:17:04 along with a no-op unpack/fetch, yeah Sep 09 14:18:01 then you're probably want to nostamp just about everything as well Sep 09 14:18:04 so it always rebuilds it Sep 09 14:18:50 note that its possible to put your .bb for the app _in_ the app's source tree as well, and have bitbake build it right there, against those sources. then you'd manipulate that .bb to have the special "live development" bits like the modified S Sep 09 14:18:59 heh Sep 09 14:19:07 * kergoth hugs bitbake/oe Sep 09 14:20:10 * ldc needs more ram to speed up the compilation of his oe-buildroot :P Sep 09 14:33:30 * JustinP smacks himself Sep 09 14:33:39 I suppose this is really much simpler...... Sep 09 14:34:08 that and I did somehting incorrect...::sigh:: Sep 09 14:34:30 hehe Sep 09 14:34:36 * kergoth kicks expansys.. ship my phone damnit Sep 09 14:34:48 call them :) Sep 09 14:35:28 the phone isnt in stock yet, waiting on it to arrive. they have a live feed of arrivals at their warehouse and it isnt on the list yet :P Sep 09 14:35:34 * kergoth just needs to keep waiting, he expects Sep 09 14:35:38 grrr.. lack of a proper dma engine sucks Sep 09 14:35:45 heh Sep 09 14:35:50 what are you working on, gb2? Sep 09 14:36:00 right now trying to keep a fifo from overflowing :) Sep 09 14:36:15 hmm, i havent been on any irc networks other than freenode since i was in texas. should fix that Sep 09 14:36:20 heh, fun Sep 09 14:36:36 the dma engine on this card can't do scatter gathering, and i'm forced to do DMA in chunks of 128k, which is kinda annoying with 4k pages Sep 09 14:37:15 i'm wondering how many 128k chunks i can actually allocate Sep 09 14:38:05 * koen looks at http://www.giantshoulderinc.com/ Sep 09 14:39:17 so after I build some packages how do I make them available to the Z. Do you mount a NFS and direct my ipkg.conf to my NFS mount point? Sep 09 14:39:19 hm, i'm still on the dns extensions working group mailing list Sep 09 14:39:22 * kergoth unsubscribes Sep 09 14:39:45 jjbenham: topic, 3rd item Sep 09 14:39:56 I -really- need to get my home server migrated Sep 09 14:40:00 i've been so lazy about that Sep 09 14:40:12 and I should do something for my birthday tomorrow Sep 09 14:40:15 migrated? Sep 09 14:40:21 heh, happy birthday Sep 09 14:40:22 how old? Sep 09 14:40:29 gb2: congrats Sep 09 14:40:39 i'll be 26 Sep 09 14:40:48 i bought a new box for it in like march or so Sep 09 14:40:51 it's been sitting there Sep 09 14:40:56 it's even had a disk die Sep 09 14:41:02 so i bought replacement disks Sep 09 14:41:03 ah Sep 09 14:41:09 but i haven't migrated the old servers over Sep 09 14:41:27 I've been annoyed at my email being slow to read, the new server should help that. Sep 09 14:41:38 * koen goes to sleep and will check the loft specs next week Sep 09 14:42:21 ok, this should be the final version of the patch....basically 2 lines changed Sep 09 14:44:10 it's supposed to be chilly tonight, and a high tomorrow of 70, i think i'll go for a nice walk outside for my birthday and maybe read a book. Sep 09 14:45:05 * kergoth listens to the new tatu on his iriver Sep 09 14:45:30 i'm really liking this cd. somewhat more refined than their last one Sep 09 14:46:01 new tatu? Sep 09 14:46:04 which iriver? Sep 09 14:46:24 yeah, tatu has a new cd coming out in a couple weeks Sep 09 14:46:26 h300 Sep 09 14:46:48 I have the h120, and i replaced the disk in it Sep 09 14:46:57 i keep failing to carry it around Sep 09 14:47:09 hehe Sep 09 14:47:15 i use mine at work constantly, but tahts about it Sep 09 14:47:23 just picked up a http://www.expansys-usa.com/product.asp?code=114076 to use with it Sep 09 14:47:24 i should just leave it and the charger at work Sep 09 14:47:46 I have shure e3c headphones, they're great, except one of my ears doesn't seem to tolerate them :( Sep 09 14:47:55 :( Sep 09 14:48:30 * JustinP uses his Spitz to play MP3s Sep 09 14:48:51 what's a good media player for handhelds? Sep 09 14:49:10 what are the features you're looking for? Sep 09 14:49:20 I use mpd Sep 09 14:49:24 (on my spitz) Sep 09 14:49:35 and I usually use ncmpc for a frontend (ncurses based) Sep 09 14:49:45 03rpurdie 07org.oe.dev * rb2fa0df3... 10/packages/linux/linux-openzaurus_2.6.13-mm1.bb: Sep 09 14:49:46 linux-oz-2.6: Rehash spitz patches after advice from mainline. Add delay to Sep 09 14:49:46 mmc_detect_change() and update corgi and spitz to use this. hx2750 and tosa need Sep 09 14:49:46 updating. More patches pushed. hrw moved the collie/poodle keymap into Sep 09 14:49:46 initscripts so the collie.keymap patch is no longer needed. Sep 09 14:49:48 http://www.gamestop.com/product.asp?product%5Fid=880784 Sep 09 14:49:49 there's also glurp and gmpc (if you like X frontends) Sep 09 14:49:53 kergoth: i was considering getting this Sep 09 14:50:12 and I also use mpc for command-line scripts (and control via my CE-RH2 remote with keylaunch) Sep 09 14:50:19 JustinP: not too fancy. just want to play mp3s from sd/cf. Sep 09 14:50:21 it's flash, and has an SD/MMC slot for expansion. i think it'd probably be a bit saner for my use than the h120 Sep 09 14:50:30 mreimer: in that case mpd should be great Sep 09 14:50:37 mreimer: what Z do you have? Sep 09 14:50:46 JustinP: no Z, h2200 Sep 09 14:50:57 mreimer: oh ;-) heh, sorry, I just assume Sep 09 14:50:58 gb2: oooh, sd/mmc slot.. thats nice Sep 09 14:51:06 wonder how big it is, physically Sep 09 14:51:44 kergoth: http://www.alienware.com/intro_pages/alienware_ceiv.aspx Sep 09 14:51:48 Is sleek and lightweight, measuring 3.1 x 2 x 0.7 inches and weighing in at just 1.3 ounces. Sep 09 14:52:10 mreimer: well...with no keyboard ncmpc would be hard to use ;-). But gmpc and glurp should work well as long as you run GPE and not Opie. Sep 09 14:52:44 considering the small amount of storage, thats a bit bigger than i'd have expected, but not too bad Sep 09 14:52:48 mreimer: although I don't know about, say, mapping keys to music functions. As I said I use the audio remote for my Zaurus and map some shell scripts to it. Sep 09 14:52:51 sure looks sharp Sep 09 14:52:51 yeah. Sep 09 14:52:58 JustinP: yeah, no kb here, but I am running GPE. Of course, I would like to run e ;-) Sep 09 14:53:05 it's got most of the features in the one a friend and I have been designing Sep 09 14:53:47 mreimer: then mpd should work pretty well. E *should* run....but I have no ipaq and i"m not sure if anyone has tested it yet ;-) Sep 09 14:54:34 JustinP: I'm tryin'! but somehow I think a ~500K e-image rootfs is quite big enough. (don't know how that happened) Sep 09 14:54:59 JustinP: thanks for the recommendations Sep 09 14:55:04 mreimer: ? 500K? Sep 09 14:55:30 -rw-r--r-- 1 mreimer mreimer 401013 Sep 8 11:20 e-image-h3900-20050908165014.rootfs.tar.bz2 Sep 09 14:55:36 ? Sep 09 14:55:39 strange.... Sep 09 14:55:40 I thought, wow, JustinP's *good* Sep 09 14:55:45 hehe Sep 09 14:55:56 the build completed without error, but obviously stuff is missing Sep 09 14:56:01 likely it doesn't get everything because it's for familiar.... Sep 09 14:56:08 I'm building for OZ Sep 09 14:56:12 about all that's in there is Xfbdev and ipkg Sep 09 14:56:19 my bz2 for e-image-core is around 17M Sep 09 14:56:27 nice Sep 09 14:56:50 h2200 can boot from sd/mmc, so size doesn't really matter to us Sep 09 14:56:53 are you using the latest ipkg? the one with broken Depends: handling? Sep 09 14:57:02 lemme check Sep 09 14:57:09 try cleaning your ipkg stuff and rebuilding it Sep 09 14:57:38 it should use the old version now...I got bit by that myself when it was first checked in Sep 09 14:57:45 and it won't downgrade automatically... Sep 09 14:57:46 I have 0.99.153-r0 Sep 09 14:57:55 ok Sep 09 14:58:01 yep Sep 09 14:58:03 153 is broken Sep 09 14:58:11 doesn't follow Depends Sep 09 14:58:22 minor oversight Sep 09 14:58:25 heh Sep 09 14:58:33 clean it out and make sure you get 152 when you rebuild Sep 09 14:58:38 *should* only take cleaning ipkg Sep 09 14:58:43 not the entire tmpdir Sep 09 14:58:55 I'll "rebuild ipkg-native" in bitbake -i Sep 09 14:59:01 ah, heh Sep 09 14:59:03 also ipkg Sep 09 14:59:13 you don't want your installed ipkg to be broken Sep 09 14:59:14 yeah, seeing how I'm on a piii 450, I'm not in a hurry to delete tmp Sep 09 14:59:19 ok Sep 09 14:59:20 yeah Sep 09 14:59:31 I don't support you've used "reparse"? ;-) Sep 09 15:00:22 no, what's that do? Sep 09 15:00:49 lets you reparse a single provider rather than everything when in the shell Sep 09 15:01:05 it's a patch that I submitted last week, just got in a few days ago Sep 09 15:01:32 so if you edit your bb files you can just reparse the ones you edited instead of waiting for a "parse" Sep 09 15:01:32 *nice* Sep 09 15:01:40 it was surprisingly simple Sep 09 15:02:01 now if bb could be taught to use that when doing a big parse... Sep 09 15:02:13 shouldn't be much more than a matter of checking .bb timestamps, right? Sep 09 15:02:27 well....this is targeted, you have to specify what you wat to reparse Sep 09 15:02:39 ghoing through and checking the timestamps on 2500 files still takes a while Sep 09 15:02:43 ah Sep 09 15:02:51 does it update dependencies? Sep 09 15:02:59 for .bb files, yeah. .conf/.bbclass changes would still result in a reparse of the entire world Sep 09 15:03:00 heh Sep 09 15:03:16 that's the next thing that should be added, though...a reparse of everything whichh just checks timestamps Sep 09 15:03:16 * kergoth ponders picking bitbake-ng back up Sep 09 15:03:44 true, that would need to follow reverse-imports/includes to be correct Sep 09 15:03:48 It should take a fraction of a second to read the timestamps of 2500 files - its not a if it needs to read the files themselves :-/ Sep 09 15:03:55 ok, then Sep 09 15:04:14 it takes a while now because it's re-reading the parsed (cached) files Sep 09 15:04:26 (thunder; should I unplug?) Sep 09 15:05:00 it shouldn't be too hard to fix this to just re-stat, but the import/include situation would need to be handled (perhaps the info is already in there...I don't know) Sep 09 15:05:01 I survied our last one. A friend lost a hub and his ADSL modem :-/ Sep 09 15:05:40 :-( Sep 09 15:06:19 JustinP: remember that things like := are evaluated immediately, which means the context (.bbclass/.conf) directly affects those, not at expansion time, but immediately. Sep 09 15:06:32 JustinP: which makes it quite difficult to try to track back without a substantial rework Sep 09 15:06:43 afaik. Sep 09 15:07:04 kergoth: Yep, it's a can of worms. Sep 09 15:07:14 kergoth: are inter-file dependencies stored anywhere? Sep 09 15:07:40 theres an __inherit_cache or something you can check to see what was inherited, but nothing for include, iirc Sep 09 15:08:00 building ipkg-native: | cp: cannot stat `/opt/oe/org.openembedded.dev/packages/ipkg/ipkg-0.99.152/terse.patch': No such file or directory Sep 09 15:08:00 | Failed to import patch terse.patch Sep 09 15:08:08 and those are all immediate operations.. we've got a declarative format but with imperative commands throughout. its a mess. Sep 09 15:08:18 heh Sep 09 15:08:30 ah well. gets the job done Sep 09 15:08:41 RP: what's the word on the spitz 2.6 kernel? Is sound working? I also would want my CE-RH2 remote to work...(yes, I know, add it myself....I tried a bit but I'm fairly well lost. I'll try again sometime soon.) Sep 09 15:08:59 kergoth: heh...yeah, it's confusing to combine the 2 Sep 09 15:09:29 kergoth: is there anything except include/inherit that would need to be checked for file-based dependencies? Sep 09 15:09:50 JustinP: No, I haven't touched sound. I dare say the changelog might mention it when I add it... Sep 09 15:10:06 JustinP: iirc, no, but those are enough to make life very difficult. Sep 09 15:10:13 but i'm going from memory here, and mine can rarely be trusted ;) Sep 09 15:10:21 As for your remote, I don't have one and won't be writing a driver for something I can't test. I don't do virtual programming Sep 09 15:10:22 i've got the issues documented somewhere in the -ng design docs Sep 09 15:10:27 somewhere.. Sep 09 15:10:28 hmm, both ipkg and ipkg-native fail to build Sep 09 15:11:21 RP: ah....sorry, I had thought it was working for some reason. I'll do my best to get the remote working myself, I understand not wokring on things you don't have. Sep 09 15:11:28 mreimer: when did you last update OE? Sep 09 15:11:35 about five minutes ago Sep 09 15:11:42 ok, maybe 20 by now Sep 09 15:11:48 i'll update again Sep 09 15:11:58 JustinP: For now, my priority is to get the basics into mainline as a foundation. Don't underestimate how much time that takes ;-) Sep 09 15:12:05 mreimer: that shouldn't matter then Sep 09 15:12:24 mreimer: hmmm...I see a terse.patch in my ipkg/files directory Sep 09 15:12:34 | cp: cannot stat `/opt/oe/org.openembedded.dev/packages/ipkg/ipkg-0.99.152/terse.patch': No such file or directory Sep 09 15:12:47 it's not looking for it in files... Sep 09 15:12:50 RP: oh, I know that upstream pushing can take a while. Especially for something as important as the kernel Sep 09 15:13:13 * mreimer runs monotone diff Sep 09 15:13:48 i don't have an ipkg-0.99.152 directory Sep 09 15:13:57 neither do I Sep 09 15:14:05 it's building, though.... Sep 09 15:14:43 mreimer: do you have files/terse.patch? Sep 09 15:14:55 nope, but I'm mt reverting ipkg/files/terse.patch Sep 09 15:14:59 now I do Sep 09 15:15:08 huh, how'd I lose that? Sep 09 15:15:46 ah, much better Sep 09 15:19:39 does glurp work on little screens? Sep 09 15:26:03 not sure Sep 09 15:28:14 I only have a spit Sep 09 15:28:15 z Sep 09 15:28:21 can you test for me? ^_^ Sep 09 15:28:39 gmpc *might* be better.... Sep 09 15:28:56 ok, I'll bitbake 'em and find out Sep 09 15:29:07 :-) Sep 09 15:29:35 you don't *need* mpd on your ipaq to test the frontends. You can always connect to another box running mpd as well Sep 09 15:29:44 rebuild cvs versions of tons of stuff, blah blah blah Sep 09 15:29:50 good idea Sep 09 15:30:03 can the frontends just play files from the local filesystem? Sep 09 15:30:07 no Sep 09 15:30:10 mpd is the player Sep 09 15:30:16 and it plays from the machine it's running on Sep 09 15:30:31 you also have to update its library when you change the files Sep 09 15:30:53 so pop out a card, pop in another, "update" and it has the new files in its DB Sep 09 15:31:35 i see Sep 09 15:32:34 it's really meant for a large music collection, which is why it uses its own DB rather than querying the filesystem Sep 09 15:32:42 makes sense Sep 09 15:32:58 nice and quick, though :-) Sep 09 15:33:05 and I like that it's scriptable through mpc Sep 09 15:33:08 seems like it wouldn't take long to whip up a simple player with pygtk or something Sep 09 15:33:16 I'm sure it wouldn't Sep 09 15:33:29 do you use yours over wifi? that would be handy Sep 09 15:33:32 you could use the code in mpc or ncmpc as a base Sep 09 15:33:51 I don't now, but it could be made to....via a mounted samba/NFS share or something Sep 09 15:34:35 I used to have my mp3s served locally via apache to my h2200 over bluetooth--got my music across the room wirelessly :-) Sep 09 15:34:49 I just used pocket ie to browse the m3u's Sep 09 15:35:25 You could also use bashmp as a base, which just uses netcat, sed, and grep to control mpd Sep 09 15:35:34 * JustinP has never had a bluetooth enabled device Sep 09 15:35:36 get about configuration problem Sep 09 15:35:44 NOTE: Running /opt/oe/build/oetmp/work/bluez-utils-nodbus-2.20-r6/bluez-utils-2.20/configure --build=i686-linux --host=arm-linux --target=arm-linux --prefix=/usr --exec_prefix=/usr --bindir=/usr/bin --sbindir=/usr/sbin --libexecdir=/usr/libexec --datadir=/usr/share --sysconfdir=/etc --s Sep 09 15:35:56 JustinP: :-) might be tricky with no keyboard Sep 09 15:36:10 this time my be autoconf failing? Sep 09 15:38:03 in my compile log Sep 09 15:38:07 ccache arm-linux-gcc -march=armv5te -mtune=xscale -DHAVE_CONFIG_H -I. -I/opt/oe/build/oetmp/work/bluez-utils-nodbus-2.20-r6/bluez-utils-2.20/tools -I.. -I/opt/oe/build/oetmp/work/bluez-utils-nodbus-2.20-r6/bluez-utils-2.20/common -I/opt/oe/build/oetmp/staging/arm-linux/include -I/usr/include -I/usr/include -I/opt/oe/build/oetmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -MT hcitool.o -MD -MP -MF " Sep 09 15:38:33 mreimer: I meant that you should be able to look at its scripts and fairly easily understand the network commands you'd need for your own frontend Sep 09 15:38:40 ah Sep 09 15:38:51 bbimage is running btw, looking good Sep 09 15:38:57 ccache arm-linux-gcc -march=armv5te -mtune=xscale -DHAVE_CONFIG_H -I. -I/opt/oe/build/oetmp/work/bluez-utils-nodbus-2.20-r6/bluez-utils-2.20/tools -I.. -I/opt/oe/build/oetmp/work/bluez-utils-nodbus-2.20-r6/bluez-utils-2.20/common -I/opt/oe/build/oetmp/staging/arm-linux/include ************-I/usr/include -I/usr/include****************** -I/opt/oe/build/oetmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers - Sep 09 15:39:06 AvengerMoJo: please use a pastebin Sep 09 15:39:09 ~pastebin Sep 09 15:39:13 hmm... pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca/ Sep 09 15:39:20 JustinP: ops :) Sep 09 15:39:25 JustinP: ok Sep 09 15:39:50 I was thinking it is just one line hehe :P one long line hehe Sep 09 15:40:04 AvengerMoJo: what's the problem? As long as the staging include is before the system one you should be ok Sep 09 15:40:35 JustinP: it messup the compile : ( Sep 09 15:41:34 03jbowler 07org.oe.dev * rc66761d9... 10/packages/gcc/ (3 files in 2 dirs): Sep 09 15:41:34 gcc(libgcc_s.so.1): fix the arm interwork _call_via_rX out of line ops Sep 09 15:41:34 The interworking symbols were being built but not exported from libgcc Sep 09 15:41:34 (i.e. they were not in the version file), added a specific version file Sep 09 15:41:34 for them and updated t-linux. Forced the cross build to always generate Sep 09 15:41:34 ARM instructions (the thumb version does not currently work), bumped Sep 09 15:41:36 PRs. Sep 09 15:42:06 JustinP: I get the same problem earlier with irda-utils Sep 09 15:42:26 JustinP: if I take out the -I/usr/include it compile just fine Sep 09 15:42:44 JustinP: should that even in there anyway :( ? Sep 09 15:42:54 AvengerMoJo: no, it shouldn't be in there Sep 09 15:43:08 JustinP: NOTE: package e-image-1.0: completed Sep 09 15:43:21 AvengerMoJo: if you can patch the autoconf stuff to fix that it would be nice ;-) Sep 09 15:43:27 mreimer: :-) Sep 09 15:43:31 -rw-r--r-- 1 mreimer mreimer 30026978 Sep 9 15:37 e-image-h3900-20050908165014.rootfs.tar.bz2 Sep 09 15:43:35 that's more like it Sep 09 15:43:51 yes Sep 09 15:44:20 mine is 30898141 Sep 09 15:44:24 JustinP: btw, to get e-wm to build, I had to mess with a couple of .edc files; seems like the #ifdef stuff in .edc files isn't working. (I'm hosting on debian unstable) Sep 09 15:44:30 so that's probably right ;-) Sep 09 15:44:37 JustinP: I'm going to ..... :) Sep 09 15:44:37 heh Sep 09 15:44:55 mreimer: do you have patches? specific build errors? Sep 09 15:45:20 the ol' error about RAW COMP LOSSY etc. Sep 09 15:45:26 mreimer: the #ifdef stuff is processed with cpp.... Sep 09 15:45:45 03jbowler 07org.oe.dev * r7ee529a6... 10/packages/busybox/busybox_1.01.bb: busybox: Enable busybox_1.01 on nslu2 Sep 09 15:45:50 I don't have a real fix, just a hack--I just pulled the appropriate defines out of the ifdef at the top Sep 09 15:45:59 hmmmm.... Sep 09 15:46:06 well if it doesn't work you'll know why ;-) Sep 09 15:47:00 I wish I remembered how I fixed that particular problem....perhaps we should compare cpp versions or something Sep 09 15:48:14 JustinP: take your pick: cpp cpp-2.95 cpp-3.3 cpp-3.4 cpp-4.0 cppw Sep 09 15:48:30 but shouldn't the cpp in staging be used? Sep 09 15:48:51 not sure Sep 09 15:48:57 edje *may* be using host cpp Sep 09 15:49:11 in fact it probably is, I didn't check that at all Sep 09 15:49:12 so any instructions on booting from this e-image? I'm going to boot from nfs Sep 09 15:49:25 just boot and entrance should come right up Sep 09 15:49:33 of course there's no keyboard... Sep 09 15:49:41 uh....yeah..... Sep 09 15:49:44 I have serial console and usb Sep 09 15:49:45 I forgot about that :-| Sep 09 15:49:49 :-) Sep 09 15:49:50 oh, ok Sep 09 15:49:56 you'll need a keyboard on the device Sep 09 15:49:59 just to let you know ;-) Sep 09 15:50:03 my h2200 has a Dell multimedia keyboard :-) Sep 09 15:50:07 ok, then Sep 09 15:50:10 03jbowler 07org.oe.dev * r5bf3b4ef... 10/ (5 files in 5 dirs): Sep 09 15:50:10 uclibc, alsa, monotone-5, util-linux: enable thumb on ucslugc, disable it on these packages Sep 09 15:50:10 This commit causes thumb build for everything in ucslugc-packages except Sep 09 15:50:10 these packages and libgcc (from gcc). libgcc and uclibc still need some Sep 09 15:50:10 work, the other packages reveal compiler bugs (monotone), source bugs Sep 09 15:50:10 (alsa) or missing configuration (thumb arch support in util-linux) and so Sep 09 15:50:12 are disabled for the moment. Sep 09 15:50:25 jbowler: what's "thumb"? Sep 09 15:50:34 mreimer: cpp --version Sep 09 15:50:39 mreimer: cpp (GCC) 3.4.4 (Gentoo 3.4.4, ssp-3.4.4-1.0, pie-8.7.8) Sep 09 15:50:54 I'm not jbowler, but "thumb" refers to 16-bit ARM instructions (rather than 32-bit) Sep 09 15:50:57 JustinP: the 16 bit ARM instruction set Sep 09 15:51:19 JustinP: cpp (GCC) 4.0.2 20050821 (prerelease) (Debian 4.0.1-6) Sep 09 15:51:24 "thumb" is to "ARM" as ... thumb is to arm :P Sep 09 15:51:36 :-) Sep 09 15:51:37 Unless someone can tell me otherwise I suspect this is the first full running non-test thumb linux system Sep 09 15:51:40 JustinP: but should the temp/run.configure file ...... using /usr/include ? Sep 09 15:52:21 thumb is part of ARM and arm (as in __arm__) is also part of ARM Sep 09 15:52:25 AvengerMoJo: huh? Sep 09 15:52:28 thumb-only? Sep 09 15:52:32 jbowler: and you've lost me ;-) Sep 09 15:52:49 smaller instructions = tighter code (maybe) Sep 09 15:52:50 mreimer: well that could be it....it seems more and more people are using GCC 4....and it's breaking all sorts of stuff Sep 09 15:53:03 betcha Sep 09 15:53:11 The processor is an ARM, it has two instruction sets - 32 bit or 16 bit (encodings of the same set actually), __thumb__ is set for 16 bit, __arm__ is set for 32 bit in GCC - this is confusing... Sep 09 15:53:11 JustinP: I check the cross bbclass file earlier ... it should correct ... it should have the configure package using /usr/include ... Sep 09 15:53:28 JustinP: I'm still tracking down where the /usr/include coming from Sep 09 15:53:34 AvengerMoJo: well...yes, it *should* be. Likely something in the *package* is causing the problem Sep 09 15:53:41 AvengerMoJo: probably has something to do with libtool Sep 09 15:53:49 * JustinP has used a few sed hacks to get around those.... Sep 09 15:53:56 AvengerMoJo: see the pcre bb file ;-) Sep 09 15:53:57 mreimer: no, it's actually not thumb only at present because I still have problems with libgcc and uclibc (so they are 32 bit for the moment). Sep 09 15:54:09 what is pcre? Sep 09 15:54:11 jbowler: how much savings do you see? Sep 09 15:54:25 AvengerMoJo: Perl Compatible Regular Expressions Sep 09 15:54:35 oh :) Sep 09 15:54:36 it seems to be averaging out to around 80% of the 32 bit size (for the code segment) Sep 09 15:54:44 that's decent Sep 09 15:55:01 do you have any idea what kind of effect thumb has on performance? Sep 09 15:55:28 Well, it should be %x2 - i.e. 80% size implies 160% of the original time Sep 09 15:55:59 But it won't work out that way because the time to transfer a cache line from SDRAM to cache is significant - i.e. it will be better than that. Sep 09 15:56:20 I don't have a good test case yet, since I still have an arm libc. Sep 09 15:56:36 sounds like a lot of work Sep 09 15:57:30 A lot of patches to gcc and uclibc, that's all - apart from that only three packages failed to build and I haven't found one that doesn't run yet. Sep 09 15:58:07 mreimer: I don't see a cpp ni staging.... Sep 09 15:58:30 JustinP: so you are telling me to over write the do_configure function is that what you are telling me? Sep 09 15:58:49 JustinP: yeah, me neither. Sep 09 15:59:20 AvengerMoJo: no, I was just mentioning that I'd gotten around similar (linking) problems by using a simple sed script on libtool Sep 09 15:59:34 AvengerMoJo: a proper patch to the .in files would be better if you can manage it Sep 09 16:00:10 then again I'm not seeing a gcc either....WTF? Sep 09 16:00:16 JustinP: booting... Sep 09 16:00:23 JustinP: ok ... but I need to figure out where is causing the do_configure getting "usr/include" in the first place :) Sep 09 16:00:27 argh, will probably have to mess with udev, etc. Sep 09 16:00:47 AvengerMoJo: somewhere in the package's configure Sep 09 16:00:50 JustinP: if it is not from bbclass file ... where will be from? Sep 09 16:01:02 AvengerMoJo: it's a package problem, like I said before Sep 09 16:01:16 AvengerMoJo: probably the configure script Sep 09 16:02:22 JustinP: you mean the source's configure script? Sep 09 16:02:26 yep Sep 09 16:03:02 JustinP: but in the run_configure file ... is shows that do_configure running the configure with --prefix=/usr/include Sep 09 16:03:13 ops .. not exactly --prefix Sep 09 16:03:14 :) Sep 09 16:03:15 AvengerMoJo: that's correct Sep 09 16:03:40 AvengerMoJo: it's *supposed* to be configured that way. That (should) have nothing to do with -I Sep 09 16:03:54 JustinP: oh Sep 09 16:04:07 JustinP: now I'm real confused :) Sep 09 16:04:13 AvengerMoJo: --prefix is run-time Sep 09 16:04:21 (well in this situation it is) Sep 09 16:04:27 oh .... Sep 09 16:04:33 wait...--prefix is /usr/include? Sep 09 16:04:42 that doesn't sound right.... Sep 09 16:05:30 on .. --includedir=/usr/include --oldincludedir=/usr/include \ Sep 09 16:05:33 I mean this Sep 09 16:05:42 --prefix=/usr Sep 09 16:05:50 ah Sep 09 16:06:05 well I *think* that's ok....maybe.... Sep 09 16:06:14 check a working package, see if it has similar things Sep 09 16:06:52 JustinP: thanks good idea :) Sep 09 16:08:27 JustinP: you are right ... they are all like this :) Sep 09 16:08:40 AvengerMoJo: as I thought. So it's actually the package's fault. Sep 09 16:08:52 mreimer: I'm preparing a fix for edje-native...hoepfully Sep 09 16:09:06 JustinP: great, I'll test it for you Sep 09 16:11:50 what's the var for the staging "cross" directory.... Sep 09 16:12:50 Alright people, I'm off to Germany - be good while I'm gone... :0 Sep 09 16:12:52 :) even Sep 09 16:12:58 http://www.tshirthell.com/store/product.php?productid=525 Sep 09 16:13:01 have a good trip Sep 09 16:13:13 have fun Sep 09 16:13:20 thx Sep 09 16:14:32 JustinP: does it look like the problem to you BLUEZ_CFLAGS="" test -d "${bluez_prefix}/include" && BLUEZ_CFLAGS="$BLUEZ_CFLAGS -I${bluez_prefix}/include" Sep 09 16:15:36 AvengerMoJo: if ${bluez_prefix} == '/usr' then yes else no Sep 09 16:16:06 well ... :) Sep 09 16:16:35 JustinP: I still trying to figure out there that two /usr/include come from :) hehe Sep 09 16:17:53 how can I get the real target-sys when doign a native build...? Sep 09 16:19:18 JustinP: I have a login screen, though it doesn't fit :-) Sep 09 16:19:26 JustinP: I think I find it AM_CFLAGS = -I/usr/include -I/usr/include Sep 09 16:19:29 mreimer: not surprised. can you login? Sep 09 16:19:34 AvengerMoJo: in what file? Sep 09 16:19:46 Makefile Sep 09 16:19:51 JustinP: I dunno; is there something like xkbd? Sep 09 16:20:12 mreimer: I thought you said you had a USB keyboard Sep 09 16:20:30 yeah, that I use through ssh/serial, but not hooked right up to the h2200 :-( Sep 09 16:21:12 I see Sep 09 16:21:13 well Sep 09 16:21:23 is there a way to bypass entrance? Sep 09 16:21:37 JustinP: I comment the line seem to work Sep 09 16:21:57 you *may* be able to fix it by altering the /entrance/autologin/mode entry in /etc/entrance_config.db, but I've never done that before.... Sep 09 16:22:15 AvengerMoJo: look for the similar line in Makefile.in or Makefile.am Sep 09 16:22:57 JustinP: back to where I told you earlier AM_CFLAGS = @BLUEZ_CFLAGS@ @USB_CFLAGS@ Sep 09 16:23:05 the BLUEZ :) Sep 09 16:23:56 so ... JustinP: I'm not sure where the fix should be done in ... Makefile.in or aclocal.m4 Sep 09 16:24:08 ops Sep 09 16:24:14 I mean acinclude.m4 Sep 09 16:24:20 whichever makes sense.... Sep 09 16:24:29 a patch is preferred over a sed hack ;-) Sep 09 16:24:34 JustinP: if I patch include.m4 ... it will be one file Sep 09 16:25:09 JustinP: but Makefile.in will be couple files because there are sub directory Sep 09 16:25:22 well then...I would think the m4 file would be the right now Sep 09 16:25:34 heh...I used := for the first time.... Sep 09 16:25:47 REAL_TARGET_SYS := ${TARGET_SYS} Sep 09 16:26:08 JustinP: then again I'm not sure the acinclude.m4 ... the BLUEZ really get any use for other things Sep 09 16:29:33 * JustinP knows next to nothing about autoconf Sep 09 16:31:37 JustinP: yeah :) Sep 09 16:31:50 JustinP: fix it :) it build Sep 09 16:31:54 :-) Sep 09 16:32:24 comment test -d "${usb_prefix}/include" && USB_CFLAGS="$USB_CFLAGS -I${usb_prefix}/include" Sep 09 16:32:40 and test -d "${bluez_prefix}/include" && BLUEZ_CFLAGS="$BLUEZ_CFLAGS -I${bluez_prefix}/include" Sep 09 16:35:47 JustinP: that's bee-yoo-tee-ful! Sep 09 16:36:08 JustinP: the trick was killing entranced and its wrapper, starting X, and then just running 'enlightenment' Sep 09 16:36:36 well, yeah, that would also work ;-) Sep 09 16:37:18 JustinP: after I update it and generated the patch ...do I need to go back and update the bb file? Sep 09 16:37:50 AvengerMoJo: well...if you want the patch to be applied you have to put it in the right directory and alter the bbfile, yes Sep 09 16:38:06 although I don't think that commenting the lines out is the best solution... Sep 09 16:38:48 JustinP: hum .... but the main CFLAGS is already good for the build ... Sep 09 16:40:11 ::shrug:: Sep 09 16:40:15 if it works go for it Sep 09 16:40:21 JustinP: hehe :) Sep 09 16:40:23 03justinp 07org.oe.dev * r8db15a99... 10/packages/efl/edje-native_0.5.0.013.bb: edje-native: Use the staged cross-compile binaries to avoid host version mismatches Sep 09 16:40:29 my fixes have been mostly sed hacks so you're still doing better than me ;-) Sep 09 16:40:39 mreimer: can you please try that? Sep 09 16:40:41 oops... Sep 09 16:40:42 one sec Sep 09 16:40:52 JustinP: will do Sep 09 16:41:28 JustinP: in the bb file there isn't patch file .. I should add SRC_URI = file://fixincludes.patch;patch=1 ?? Sep 09 16:41:53 mreimer: I'm pushing a PR bump as well :-/ Sep 09 16:41:58 k Sep 09 16:42:09 AvengerMoJo: yes, the patch goes in the SRC_URI....it needs to be added, though Sep 09 16:43:06 03justinp 07org.oe.dev * rc3b3a0af... 10/packages/efl/edje-native_0.5.0.013.bb: edje-native: bump PR Sep 09 16:44:01 ~lart JustinP for forgetting to bump PR Sep 09 16:44:01 * ibot does a little 'renice 20 -u JustinP' for forgetting to bump PR Sep 09 16:44:10 after I update the database bb file ... it will update the cache bbfile in tmp right? Sep 09 16:44:17 yep Sep 09 16:44:29 JustinP: thanks Sep 09 16:44:32 if you were using "bitbake -i" you could just reparse that one file ;-) Sep 09 16:44:50 cool thanks Sep 09 16:45:20 * AvengerMoJo still adjusting himself with the bitbake build system Sep 09 16:45:21 "reparse package" ^_^ Sep 09 16:45:36 I just added that. It's why I wasn't using the shell before Sep 09 16:45:50 :) Sep 09 16:46:19 JustinP: should I file a bug report for this also Sep 09 16:46:44 AvengerMoJo: yes, if it's a fix for something file a bug report and your patch Sep 09 16:47:17 * JustinP hopes that fix workes Sep 09 16:47:29 JustinP: It work for me :) Sep 09 16:47:37 is reparse listed by "help"? I don't see it. I must need to upgrade bitbake? Sep 09 16:47:43 JustinP: I'm back to the bitbake opie-image now Sep 09 16:47:44 mreimer: all I tested was rebuilding e-wm after an edje rebuild. I didn't test runtime. Sep 09 16:47:53 JustinP: and all the error's gone Sep 09 16:48:34 mreimer: yes, it's listed by "help". It's pretty new. It's the latest commit. Heh...only comitted yesterday in fact. Sep 09 16:48:44 that's my problem Sep 09 16:49:05 I have another patch pending as well for CVSDATE = "now" Sep 09 16:49:17 nice, I could use that when testing kernel builds Sep 09 16:49:18 although that one's really only changes in 2 lines Sep 09 16:49:32 mreimer: do you check your kernel out of CVS? Sep 09 16:49:42 mreimer: or are you a kernel dev? Sep 09 16:49:52 I'm mostly a kernel dev for h2200 Sep 09 16:50:09 so my dev I do in a local tree, but I build handhelds-pxa-2.6 for gpe-image Sep 09 16:50:24 ah Sep 09 16:50:29 :-) then this should help you Sep 09 16:50:39 http://bugs.openembedded.org/show_bug.cgi?id=315 Sep 09 16:52:00 mreimer: if you test that patch please let me know if it works :-) Sep 09 16:52:06 almost done parsing Sep 09 16:52:09 hehe Sep 09 16:52:19 did you update bb for the reparse command? Sep 09 16:52:29 no, not enough CPU (on the phone) Sep 09 16:52:33 ah Sep 09 16:53:41 * AvengerMoJo back on the opie-image road Sep 09 16:53:50 AvengerMoJo: which package was that for? Sep 09 16:53:54 oh yeah bluez Sep 09 16:54:00 right Sep 09 16:54:10 bluez-utils-2.20 :) Sep 09 16:54:19 bluez-utils-nodbus-2.20-r6 Sep 09 16:56:35 JustinP: could you check for me see I'm doing it right ? http://bugs.openembedded.org/show_bug.cgi?id=317 Sep 09 16:58:52 AvengerMoJo: looks ok Sep 09 16:59:02 thanks Sep 09 16:59:26 * AvengerMoJo wish he can file the image today so he can do some hacking ... Sep 09 16:59:35 file the image? Sep 09 16:59:39 finish :) Sep 09 16:59:40 hehe Sep 09 17:00:02 oh.. one more thing ... is the meta-opie stuff do ? Sep 09 17:00:42 JustinP: rebuilt edje-native; now rebuilding e-wm Sep 09 17:00:51 whoa, don't build gcc-cross-initial Sep 09 17:01:56 ? Sep 09 17:02:30 dependency; I don't want to wait for that, so I'll use filebuild Sep 09 17:03:15 * JustinP wonders how you can build without gcc-cross.... Sep 09 17:03:36 I already have it, I just don't want to build the latest greatest version Sep 09 17:03:47 r1 is probably good enough Sep 09 17:04:34 JustinP, if he is building for x86 Sep 09 17:12:19 * AvengerMoJo is wondering why no one else seem to get the problem he get .... Sep 09 17:13:36 JustinP: it built successfully Sep 09 17:15:07 :-) Sep 09 17:15:12 thanks! Sep 09 17:15:14 np Sep 09 17:41:04 ugh Sep 09 17:41:13 i just tried to understand debian's kernel-package Sep 09 17:42:01 good luck Sep 09 17:42:55 I'm thinking I might just start over from scratch. Sep 09 17:43:21 gb2: ...? Sep 09 17:43:32 gb2: did you try to use kpkg? Sep 09 17:43:45 I just need something to package a custom built kernel making "kernel-image" and "kernel-headers" packages to be used on a debian system. Sep 09 17:44:01 with a few customizations in the packaging process Sep 09 17:44:24 gb2, youu trying manual or using KDE's kbuild tool? Sep 09 17:44:35 make-kpkg looks like it's overengineered perl written by 10,000 monkeys on some sort of bad batch of acid Sep 09 17:44:45 hehe Sep 09 17:45:03 that rather sounds like the kernel doesnt it? :P Sep 09 17:45:04 emte: i don't really want even more levels of indirection Sep 09 17:45:05 gb2: basically you just use make-kpkg kernel-image ... then you are done :) Sep 09 17:45:29 gb2, jsut wasnt sure how you were doing it Sep 09 17:45:50 if you update something .. you can add --revision someXXX then you are done with different deb package Sep 09 17:45:58 gb2: shouldn't be that bad Sep 09 17:46:14 avenger: i know how it's supposed to work. the problem is this situation requires a bit more munging of the files being packaged Sep 09 17:46:39 sarge's make-kpkg also can't properly package 2.6 kernels Sep 09 17:46:47 the one in testing can Sep 09 17:47:30 hum ... can you just pin the testing make-kpkg install and use that? Sep 09 17:47:38 I want to modify it to add a kernel-dbg package as well, but make-kpkg is too bloated to be able to really do that without spending a while understanding it. Sep 09 17:48:54 I wonder if doing this whole process will reveal why debian's separated debuginfo is broken Sep 09 17:51:02 hehe Sep 09 17:51:03 possible Sep 09 17:51:43 gb2: try debian channel I think some guys in there really may help you get to what you want Sep 09 17:52:21 heh Sep 09 17:52:29 :P Sep 09 17:53:21 last time i followed a coworker submitting a bug into debian, the maintainers didn't seem to actually understand. Sep 09 17:55:26 this was a valgrind issue, and the maintainers seemed like they had a user's prospective of the problem rather than a toolchain/library internals view of the problem. Sep 09 17:56:21 gb2: hum Sep 09 17:56:55 gb2: sometime there are two kind of maintainer :) one is more hacker ... one is more usability type I guess Sep 09 17:58:31 yeah. Sep 09 17:59:01 i've found package consistency is sorely lacking across debian's packages. there's a LOT of different styles and levels of experience. Sep 09 17:59:57 gb2: hum :) well.. better get some support then none Sep 09 18:00:53 or use another distro :) Sep 09 18:01:33 lol Sep 09 18:01:55 I use fedora personally, but work uses debian Sep 09 18:01:56 gb2: too many ... I prefer using debian and build from source after install Sep 09 18:02:14 then again .. that's just me ... Sep 09 18:02:27 too many? Sep 09 18:02:28 gentoo take too long to install ... Sep 09 18:02:33 too many distribution Sep 09 18:02:57 I do like Gobo linux :) but not enought package and support Sep 09 18:03:45 I think part of the problem debian has is that it's too large, and has rule by comittee, so stuff drags on, policy becomes long and bloated, and QA suffers somewhat from that. Sep 09 18:04:05 ubuntu seems to be making debian quite a bit more usable Sep 09 18:04:16 yeah Sep 09 18:04:55 I've ranted before about debian's libc shipping with broken separated debuginfo, ubuntu's works fine. they're the same package though when i went to look. Sep 09 18:05:10 i didn't go digging much beyond that Sep 09 18:07:37 oe ...using daily opie package ... is that needed? Sep 09 18:08:07 everyday I build ..it will download the lately daily cvs tarball ... Sep 09 18:08:37 I hate that Sep 09 18:08:42 no, it's not needed Sep 09 18:08:48 it *should* have a fixed CVSDATE Sep 09 18:08:52 * AvengerMoJo will able to provide an opie archive ... Sep 09 18:09:03 if you could make a note of all of the opie packages which get rebuilt every day I'd be grateful Sep 09 18:09:27 JustinP: what do you mean by that Sep 09 18:09:33 JustinP: I don't really get it Sep 09 18:09:43 or if you could add a CVSDATE = "20050909" to them all and give me a monotone diff that would also be good Sep 09 18:10:05 AvengerMoJo: all of those packages are set up to pull from CVS but don't have a CVSDATE in the bb file Sep 09 18:10:08 they should have one Sep 09 18:10:20 I see Sep 09 18:10:58 *actually* they should be renamed to have a CVSDATE in their filename and use CVSDATE = "${PV}" Sep 09 18:11:01 may be I can write a script and add or remove all the opie package assigned by date? Sep 09 18:11:17 oh what is {PV} by the way Sep 09 18:11:22 actually...perhaps I could write a simple script to find such things.... Sep 09 18:11:33 ${PV} is the version in the filename Sep 09 18:11:41 everything after the first _ I believe Sep 09 18:12:07 I see Sep 09 18:13:05 so ... what I can do .. is write a script to add, remove and edit all the opie-bb file with the CVSDATE Sep 09 18:16:19 JustinP: where is the bb file getting the CVSDATE from? Sep 09 18:16:32 monotone itself? Sep 09 18:19:42 AvengerMoJo: this depends.... Sep 09 18:19:51 oh Sep 09 18:20:09 AvengerMoJo: for the ones which change every day a CVSDATE should be added to each of the bb files Sep 09 18:20:39 right Sep 09 18:20:58 JustinP: but what I mean is .. where is the variable value feeding from? Sep 09 18:21:30 JustinP: does monotone get the default value? Sep 09 18:21:40 if there is no CVSDATE set the bitbake sets it to today automatically Sep 09 18:21:49 monotone has nothing to do with building Sep 09 18:21:53 it's only the SCM Sep 09 18:22:32 cvsdate = bb.data.getVar('CVSDATE_%s' % pn, d, 1) Sep 09 18:22:32 if cvsdate != None: Sep 09 18:22:32 bb.data.setVar('CVSDATE', cvsdate, d) Sep 09 18:22:39 in the base.bbclass Sep 09 18:22:49 does it have anything to do with it? Sep 09 18:24:03 bb.data <--- what is this Sep 09 18:25:29 I need more background understand on the build system... it is still too many unkown to me Sep 09 18:26:41 don't worry about the bbclass Sep 09 18:26:44 that's perfectly fine Sep 09 18:26:47 it's meant to be that way Sep 09 18:26:57 we need to add a CVSDATE to the files which don't define one Sep 09 18:28:10 you mean override it will some specific day right Sep 09 18:28:36 what we can do also is a globle include file for all opie Sep 09 18:28:47 then put the CVSDATE in the include file Sep 09 18:29:00 and put the include line in all the opie-cvs bb file Sep 09 18:29:09 that should fix it Sep 09 18:30:55 JustinP: what do you think? Sep 09 18:32:26 JustinP: what is why I'm thinking bbclass file.. if we can update just the bbclass for opie and define the CVSDATE in it ... it should update all the bb file doesn't it? Sep 09 18:34:06 if they all include the same file I supose that will work, yes Sep 09 18:35:25 JustinP: I will try to get you the diff after I get this build hehe :P Sep 09 18:38:39 JustinP: I'm just thinking where should this file located Sep 09 18:39:59 well...if it's all pulled fromt he same place and interconnected it makes some sense to use the same CVSDATE for them all. But if they're disparate packages it makes more sense to put it in each of the files directly. Sep 09 18:42:14 JustinP: I see ... I guess still better write a script to generate the CVSDATE into the file will be better Sep 09 18:42:45 and in that case we actually should do a rename wiht CVSDATE = "${PV}" Sep 09 18:42:55 see e17/e-wm_xxxxx.bb Sep 09 18:45:00 where is PV define? Sep 09 18:45:38 I'm not sure where it's *set* but that doesn't matter as much as the value it gets. As I said before it's the version part of the filename Sep 09 18:45:53 ok Sep 09 19:24:38 ~lart Flash Sep 09 19:24:38 * ibot takes a rusty axe and swings it violently, taking Flash's head off Sep 09 19:39:53 JustinP: .. I see some bb file get a line like PV= "1.2.0+cvs-${CVSDATE}" Sep 09 19:40:15 JustinP: if we define CVSDATE=${PV} .... won't that be a problem Sep 09 19:41:48 yes Sep 09 19:44:39 .... :) Sep 09 19:45:04 JustinP: so .. should i update the CVSDATE="20050906" as a fix date then? Sep 09 19:54:38 AvengerMoJo: sure Sep 09 19:55:14 JustinP: I wrote a script to add the CVSDATE for all the bb file Sep 09 19:55:23 all the cvs.bb Sep 09 19:55:35 I mean .. all the opie*cvs.bb Sep 09 19:55:39 ok Sep 09 19:55:41 does it work Sep 09 19:55:53 the file update work Sep 09 19:56:22 "Lain has no mail" Sep 09 19:56:31 I will look for CVSDATE=* and the delete that line if exist ... then add your own CVSDATE="2005XXXX" line Sep 09 19:56:40 right before the PV line Sep 09 20:00:11 nono Sep 09 20:00:17 don't delete previous CVSDATES Sep 09 20:00:23 only add it to ones that don't have it Sep 09 20:27:59 JustinP: ok Sep 09 20:28:06 do you want the script Sep 09 20:28:40 sure Sep 09 20:33:22 JustinP: http://pastebin.ca/22556 Sep 09 20:36:21 * AvengerMoJo is excited .. getting into image final stage .. Sep 09 20:36:50 take me almost 5 days ...to build this ... Sep 09 20:38:11 JustinP: ... stupid question .. where is the output fs image? Sep 09 20:38:39 tmp/deploy/images Sep 09 20:38:48 * JustinP hacks some Flash Sep 09 20:39:39 thanks Sep 09 20:39:54 JustinP: does the script useful? Sep 09 20:39:59 not sure Sep 09 20:40:05 sorry, I'm working right now Sep 09 20:40:37 ok :) Sep 09 20:40:41 * AvengerMoJo is done!!! Sep 09 20:40:53 * AvengerMoJo is going to test it on his C3000 Sep 09 20:42:13 good luck Sep 09 20:43:32 * AvengerMoJo is crossing his finger Sep 09 20:45:00 AvengerMoJo: What image are you testing? Sep 09 21:04:32 oh Sep 09 21:05:07 .... is the SanDisk being support in kernel now? Sep 09 21:05:24 can't find any good how to ... :( Sep 09 21:14:28 ? Sep 09 21:14:36 do you mean Secure Digital? Sep 09 21:16:20 no :) Sep 09 21:16:35 JustinP: SanDisk ... the small flash disk Sep 09 21:16:56 I get a internal reader in my laptop and never able to get it to work Sep 09 21:17:11 but when I plugin .. i get this cs: memory probe 0x0c0000-0x0fffff: excluding 0xc0000-0xcffff 0xd8000-0xfffff Sep 09 21:17:44 ? Sep 09 21:18:03 ::shrug:: Sep 09 21:52:12 hi all Sep 09 21:52:40 I finally made it here :) Sep 09 21:54:18 :) Sep 09 21:54:25 and you do Sep 09 21:55:43 I am doing update now with monotone but its very slow Sep 09 21:56:08 it is :) Sep 09 21:56:23 did you download the image Sep 09 21:56:51 yes I did :) Sep 09 21:57:11 :) Sep 09 21:57:12 I didn't at first, and waiting tooo long Sep 09 21:57:17 that's should be faster Sep 09 21:57:27 yup :) Sep 09 21:57:31 like forever Sep 09 21:57:53 bitbake task-bootstrap -> will this build a zImage file or fully working oz? Sep 09 21:58:12 that will give you all of the ipks for a bootstrap image Sep 09 21:58:20 bitbake bootstrap-image Sep 09 21:58:31 including hostap? Sep 09 21:58:37 ? Sep 09 21:58:40 look at it Sep 09 21:59:09 should be in packages/meta Sep 09 21:59:20 yea, I see it Sep 09 21:59:30 I need help here. Sep 09 21:59:39 so ask a question Sep 09 21:59:42 lol Sep 09 21:59:43 :) Sep 09 21:59:49 hold on, my typing is slow Sep 09 21:59:57 you should look for DEPENDS and RDEPENDS Sep 09 22:00:04 DEPENDS is what to build for the current file Sep 09 22:00:14 RDEPENDS is what the ipk created will depend on Sep 09 22:00:31 JustinP: more usb card read work ... but I couldn't seem to find any info on line that help me get my onboard cardreader to work at all Sep 09 22:00:40 %s/more/most Sep 09 22:01:10 AvengerMoJo: so you have an SD card reader on your laptop? Are you sure it's a SanDisk card reader? Sep 09 22:01:20 yes Sep 09 22:01:25 positive Sep 09 22:01:31 just making sure Sep 09 22:01:41 :) Sep 09 22:01:43 laptops are generally less supported than desktops Sep 09 22:01:43 I got a linksys WCF12 wifi card, and I am using C3100 with oz 3.5.4pre the one Mickey compiled. Sep 09 22:01:51 but it's better than it used to be Sep 09 22:02:03 hum Sep 09 22:02:12 man... you get C3100 Sep 09 22:02:20 :) Sep 09 22:02:21 I want to patch hostap so it can allow my linksys card to inject packets Sep 09 22:02:42 yea I got one, the first day sharp released it :) Sep 09 22:02:52 abm_y4k: :) Sep 09 22:03:00 abm_y4k: I get a C3000 Sep 09 22:03:14 abm_y4k: but I get it for work really Sep 09 22:03:15 but I am having difficulty with this toy Sep 09 22:03:26 hi,all Sep 09 22:03:37 I am a system admin, I need this toy for work as well :) Sep 09 22:03:46 hehe ok Sep 09 22:03:57 I have a C3000 Sep 09 22:03:59 that's a high end network monitoring tool Sep 09 22:03:59 works very well Sep 09 22:04:01 :) Sep 09 22:04:12 JustinP: :) Sep 09 22:04:15 I m a beginner with Open Embedded? I want to know , which developet mainboard is fine to me ? Sep 09 22:04:21 I've been using an Ambicom CF 802.11b card without problems Sep 09 22:04:30 works with wpa_supplicant as well Sep 09 22:04:31 yes, especially when I have the aircrack working with hostap Sep 09 22:04:59 tengulre: what do you mean by "which mainboard"? Are you looking to create a new machine? Sep 09 22:05:09 I am currently research on wifi security.. but can't use this toy Sep 09 22:05:34 JustinP, I want to select a board and CPU! Sep 09 22:05:35 tengulre: :) ....what are you trying to do Sep 09 22:05:47 so any steps for me to get hostap patch without doing much of works? Sep 09 22:06:06 try atomchip.com Sep 09 22:06:25 that thing rocks ;) Sep 09 22:07:20 * JustinP loves his C3000 Sep 09 22:07:36 abm_y4k: yes, bootstrap includes hostap for the clamshells Sep 09 22:08:18 abm_y4k: see conf/machine/zaurus-clamshell-2.4.conf Sep 09 22:09:00 JustinP: so I just patch the hostap that allow firmware updates, and bitbake bootstrap-image should work? Sep 09 22:10:41 abm_y4k: you'll need to put your patch in the hostap folder and alter the bb file to include the patch Sep 09 22:11:23 oh no.. I have to learn about bb file :( Sep 09 22:12:21 it's not hard Sep 09 22:12:33 and you can ask questions here ;-) Sep 09 22:12:48 some people will help and others won't Sep 09 22:12:57 I *usually* help Sep 09 22:12:58 sure I will :) Sep 09 22:13:26 lol Sep 09 22:13:27 but what I need to do is very simple just uncomment "PRISM2_NON_VOLATILE_DOWNLOAD" in the hostap_config.h Sep 09 22:13:42 do I still need to alter the bb file? Sep 09 22:14:13 you can build that package seperately Sep 09 22:14:30 or you can add a patch into the bb file Sep 09 22:14:38 bitbake hostap ? Sep 09 22:14:48 or you can add a quick sed to do_compile_prepend() Sep 09 22:14:59 ok got it Sep 09 22:14:59 do_compile_prepend() { Sep 09 22:15:10 i'll try both way Sep 09 22:15:23 sed -i s:#blah:blah: ${S}/hostap_config.h Sep 09 22:15:23 } Sep 09 22:16:23 what is ${S}? Sep 09 22:16:29 source dir? Sep 09 22:17:31 oh no, monotone is still updating.... Sep 09 22:17:46 yes Sep 09 22:17:53 thanks Sep 09 22:17:56 when is your snapshot from? Sep 09 22:18:19 http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 Sep 09 22:18:31 I meant when did you download it? Sep 09 22:18:32 the one on handhelds.org Sep 09 22:18:37 today Sep 09 22:18:40 oh Sep 09 22:18:50 how many revisions is monotone working on? Sep 09 22:19:12 rev in or out? Sep 09 22:19:23 monotone: connecting to monotone.vanille.de Sep 09 22:19:23 monotone: finding items to synchronize: Sep 09 22:19:23 monotone: certs | keys | revisions Sep 09 22:19:23 monotone: 7830 | 31 | 2050 Sep 09 22:19:23 monotone: bytes in | bytes out | certs in | revs in | revs written Sep 09 22:19:23 monotone: 426.5k | 118.0k | 650 | 125 | 1 Sep 09 22:19:25 monotone: verifying new revisions (this may take a while) Sep 09 22:19:27 monotone: bytes in | bytes out | certs in | revs in | revs written Sep 09 22:19:30 monotone: 469.5k | 118.0k | 650 | 125 | 91 Sep 09 22:19:46 in Sep 09 22:19:51 please don't flood Sep 09 22:20:04 sorry! is cpu matters ^_~ Sep 09 22:20:41 I got a celeron 468.308MHz only :( Sep 09 22:22:14 anyways, should be done over night. thanks for the help JustinP and AvengerMoJo! Sep 09 22:22:59 ZZzzz now. good night all! Sep 09 22:23:39 :) Sep 09 22:37:15 JustinP: .... what is the filename I need to update to ? Sep 09 22:37:26 JustinP: I mean the image for flashing the rom Sep 09 22:38:04 AvengerMoJo: I don't understand Sep 09 22:38:28 the update.sh.split need to be rename to ... I forget what Sep 09 22:38:58 JustinP: also I won't need initrd.img? Sep 09 22:39:14 .split? Sep 09 22:39:29 you should have update.sh, zImage.bin, and initrd.bin Sep 09 22:39:30 JustinP: updater.sh.spitz Sep 09 22:39:31 :) Sep 09 22:39:36 oh Sep 09 22:39:46 opie-image-spitz-20050909125153.rootfs.img updater.sh.spitz Sep 09 22:39:46 opie-image-spitz-20050909125153.rootfs.tar.bz2 zImage-2.6.13-mm1-spitz-20050909125153.bin Sep 09 22:39:54 this is all I get from deploy Sep 09 22:40:06 I don't see initrd Sep 09 22:41:17 yep, and all 3 are right there Sep 09 22:42:02 rootfs is the initrd? Sep 09 22:42:04 ok Sep 09 22:42:09 thanks Sep 09 22:42:17 that confused me for a while .:) Sep 09 22:42:32 make sur eyou use the .img, not the .bz2 Sep 09 22:42:55 oh Sep 09 22:42:55 ok Sep 09 22:49:30 ... ops Sep 09 22:49:40 JustinP: ERROR flush Write Sep 09 22:54:25 JustinP: any idea? Sep 09 22:55:46 you're suing the right updater.sh? Sep 09 22:55:54 you used the .img for initrd? Sep 09 22:57:27 right Sep 09 22:57:34 no idea Sep 09 22:57:38 I think it has finish updating the kernel Sep 09 22:57:42 was your card formatted right? Sep 09 22:57:50 how? Sep 09 22:58:22 FAT12 Sep 09 22:58:39 .... Sep 09 22:59:00 FAT12 .. Sep 09 22:59:54 one of the FATs at least,.... Sep 09 23:00:19 I think so ... I will try again .. Sep 09 23:00:25 will the kernel being too big Sep 09 23:01:04 -rw-are--are-- 1 alex alex 1292460 9月 10 13:20 zImage-2.6.13-mm1-spitz-20050909125153.bin Sep 09 23:01:10 1.3M Sep 09 23:03:52 did you alter the config at all? Sep 09 23:04:17 * JustinP still uses 2.4 since 2.6 doesn't have sound Sep 09 23:05:24 no :) Sep 09 23:06:04 it seem to write ... for a little bit then it die Sep 09 23:10:00 in the offical... there is one like this .. mv pivotboot-image-spitz-20050428205242.rootfs.img initrd.bin Sep 09 23:10:08 pivotboot-image .. what is that Sep 09 23:10:49 because I think I still missing the initrd .. opie-image is for hdimage1.tgz... Sep 09 23:14:13 should the kernel need the initrd ? :) Sep 09 23:21:11 oh....heh Sep 09 23:21:12 sorry Sep 09 23:21:25 yeah, for the spitz you need an initrd *and* an hdimage Sep 09 23:22:12 you need to convert the opie-image*****.bz2 to a tgz and name it hdimage1.tgz Sep 09 23:22:30 (I've never "flashed" an hdimage) Sep 09 23:22:45 I see Sep 09 23:22:49 yes, you need a pivotboot-image Sep 09 23:22:55 "bitbake pivotboot-image" Sep 09 23:22:59 JustinP: also why does it come with bz2? not tgz? Sep 09 23:23:11 AvengerMoJo: not sure...possibly the conf Sep 09 23:23:14 hi all, back :) Sep 09 23:23:16 easy enough to fix Sep 09 23:23:22 bzip2 -d xxx.bz2 Sep 09 23:23:30 gzip xxx Sep 09 23:24:04 right ... but that is kind of silly to not do it in the bb Sep 09 23:24:08 directly Sep 09 23:24:14 ::shrug:: Sep 09 23:24:18 I'm sure it's configurable Sep 09 23:24:25 JustinP: there aren't any source code files for the packages :( Sep 09 23:24:29 I don't use them, as I said...at least not for "flashing" Sep 09 23:24:37 so I must edit the bb file Sep 09 23:24:38 I unzip directly to a partition myself Sep 09 23:24:52 I've left my Sharp ROM info untouched all of this time ;-) Sep 09 23:25:13 JustinP: I see .. so you still get all the good stuff :) Sep 09 23:25:14 hehe Sep 09 23:25:18 abm_y4k: no, of course not, that would make the repository *huge*. And there's no reason to store them in the OE rep Sep 09 23:25:41 abm_y4k: you don't *have* to alter the bb file...you can unpack, then edit the source, then compile, but that's not reproducible Sep 09 23:25:55 AvengerMoJo: I haven't botted in the Sharp ROM for a long, long, long time Sep 09 23:26:00 AvengerMoJo: I should really get rid of it Sep 09 23:26:04 yea I just found out Sep 09 23:26:07 AvengerMoJo: I also created a swap partition for myself Sep 09 23:26:15 me too Sep 09 23:26:22 I see that : Sep 09 23:26:25 last time Sep 09 23:26:31 I did that last time Sep 09 23:26:32 :) Sep 09 23:26:45 AvengerMoJo: and I also use file images for my development so I don't kill my main install Sep 09 23:26:48 but one crazy point in the start up Sep 09 23:26:48 JustinP this is the line after altered yours: Sep 09 23:26:53 do_compile_prepend() { Sep 09 23:26:53 sed -i s:\/\* \#define PRISM2_NON_VOLATILE_DOWNLOAD \*\/:\#define PRISM2_NON_VOLATILE_DOWNLOAD: ${S}/hostap_config.h Sep 09 23:26:53 } Sep 09 23:26:56 if I didn't load the mplayer module it won't load at all Sep 09 23:27:17 do I need \ for special chars? Sep 09 23:27:21 JustinP: you mean you build stuff directly in the C3000 Sep 09 23:27:37 abm_y4k: only for ones that sed sees as special. Probably not / Sep 09 23:27:49 AvengerMoJo: ? mplayer "module"? Sep 09 23:27:50 what about *? Sep 09 23:27:59 and #? Sep 09 23:28:04 JustinP: forget the name .. bbvd Sep 09 23:28:17 AvengerMoJo: I *have* compiled on the Z. I compiled parted in the Sharp ROM and I compiled emacs in OZ (that took forever). Sep 09 23:28:30 abm_y4k: I don't know, "man sed" Sep 09 23:28:56 JustinP: right ... then what do you mean by using the image for development Sep 09 23:29:30 AvengerMoJo: I mean that when I create a new image I test is using a file image instead of using a partition Sep 09 23:29:42 I see Sep 09 23:29:43 get you Sep 09 23:29:53 my C3000 don't get wireless Sep 09 23:29:55 so :( Sep 09 23:30:01 and if you wanted ot run, say, Opie and GPE in seperate environments you could do it that way Sep 09 23:30:04 I get a card but it didn't work Sep 09 23:30:06 AvengerMoJo: using what card? Sep 09 23:30:12 AvengerMoJo: mine has wireless Sep 09 23:30:15 one sec Sep 09 23:30:22 even WPA Sep 09 23:30:33 that's cool Sep 09 23:30:41 my is ET-CWB1000 Sep 09 23:30:46 one more question, I see there is lots of machine conf, how do I compile for borzoi? Sep 09 23:30:49 EagleTec Sep 09 23:30:55 I have seen someone talk about it Sep 09 23:31:03 but I never get the driver Sep 09 23:31:15 AvengerMoJo: never heard of it Sep 09 23:31:23 abm_y4k: you set it in your local.conf Sep 09 23:31:31 oh ic Sep 09 23:31:32 ok Sep 09 23:31:34 thx Sep 09 23:31:42 JustinP: which one you are using Sep 09 23:31:51 AvengerMoJo: Ambicom Sep 09 23:32:02 WL1100C-CF Sep 09 23:32:05 I see Sep 09 23:32:17 works without any problems whatsoever Sep 09 23:32:31 can I put two thing in one line bitbake Sep 09 23:32:45 AvengerMoJo: such as? Sep 09 23:33:09 bitbake pivotboot-image opie-image Sep 09 23:33:42 heh Sep 09 23:33:43 yeah Sep 09 23:34:02 ok Sep 09 23:34:18 thanks Sep 09 23:34:39 ~lart my powerbookf or being slow Sep 09 23:34:39 * ibot pushes the wall down onto my powerbookf or being slow whilst whistling innocently Sep 09 23:35:34 hehe Sep 09 23:35:43 JustinP: what kind of powerbook you get Sep 09 23:35:56 are you using that for building oe also Sep 09 23:36:06 bwahahah! Sep 09 23:36:07 no Sep 09 23:36:20 I'm using it because it's the only thing I have that I can get Flash MX to run on Sep 09 23:36:28 it's a Lombard....400Mhz G3.... Sep 09 23:36:33 Oh :) Sep 09 23:36:40 I try not to use macs but I'm forced to here Sep 09 23:36:42 really .. Sep 09 23:36:46 why Sep 09 23:37:03 new mac is pretty nice I think Sep 09 23:37:17 I like my Linux, thanks Sep 09 23:37:28 and macs are far too expensive Sep 09 23:39:22 understatement Sep 09 23:39:26 I bought this one because I thought my wife wanted it Sep 09 23:39:35 and i doubt they will get cheaper going x86 Sep 09 23:39:56 I see Sep 09 23:40:31 job better hope his server market holds out Sep 09 23:41:25 hehe JustinP flashing done!!!! Sep 09 23:41:32 ^_^ Sep 09 23:41:46 god :) ... finally Sep 09 23:41:55 PCnet/FAST+ 79C972 ... you think thats my gigabit card? Sep 09 23:42:12 its either that or my fiber card Sep 09 23:42:16 JustinP: this is what I set: MACHINE = "borzoi" TARGET_ARCH = "arm" TARGET_OS = "linux" DISTRO = "openzaurus-3.5.4" KERNEL_VERSION = "2.4.20" is this correct? Sep 09 23:44:18 ... unable to initialize console? Sep 09 23:44:56 you don't want/need TARGET_ARCH or TARGET_OS Sep 09 23:45:10 and you don't need a kernel version Sep 09 23:45:23 (the right value for 2.4 would be "2.4" but that's the default) Sep 09 23:45:30 AvengerMoJo: fun Sep 09 23:45:40 :) Sep 09 23:45:52 so far so good Sep 09 23:48:37 alrite Sep 09 23:48:39 thanks Sep 09 23:51:20 JustinP: time to take a break Sep 09 23:51:23 seem nice Sep 09 23:51:26 AvengerMoJo: break well Sep 09 23:51:42 ~lart people for changing implementation when they're supposed to be finishing the app Sep 09 23:51:43 * ibot judo chops people for changing implementation when they're supposed to be finishing the app Sep 09 23:53:27 JustinP, isnt that when you find the implementation is wrong? :P Sep 09 23:54:34 emte: no, I don't think so..... Sep 09 23:54:55 lol Sep 09 23:54:58 anyway I'm dragged back in without any real Flash experience and no insight into the code-base to finish this thing...::sigh:: Sep 09 23:55:30 sorry i cant help you, i only know director's actionscript Sep 09 23:57:33 hum Sep 09 23:57:53 JustinP: I guess 2.4 kernel still better? Sep 09 23:58:20 "better" in what way? more fully featured? yes. more stable? probably not Sep 09 23:59:21 JustinP: really? why is 2.4 less stable? Sep 09 23:59:50 why is the system benchmark doesn't get C3000 :) Sep 09 23:59:58 maybe I can add it Sep 10 00:00:02 eh? Sep 10 00:00:15 JustinP: I see that the sed command you gave me is a bit diff from the linux one Sep 10 00:00:15 JustinP: yours: sed -i s:\/\* \#define PRISM2_NON_VOLATILE_DOWNLOAD \*\//:#define PRISM2_NON_VOLATILE_DOWNLOAD: ${S}/hostap_config.h Sep 10 00:00:15 JustinP: linux: sed -i 's/\/\* #define PRISM2_NON_VOLATILE_DOWNLOAD \*\//#define PRISM2_NON_VOLATILE_DOWNLOAD/' filename Sep 10 00:00:16 the 2.4 kernel si embeddix and inherently unstable and just plain crap Sep 10 00:00:39 abm_y4k: I didn't give you that line Sep 10 00:00:50 abm_y4k: yes, you should use '' when you have spaces and things like * Sep 10 00:01:04 JustinP: the newest 2.4 kernel is 2.4.20? Sep 10 00:01:17 so I can just insert the linux line in and should work rite? Sep 10 00:01:25 linux line in? Sep 10 00:01:47 this one: sed -i 's/\/\* #define PRISM2_NON_VOLATILE_DOWNLOAD \*\//#define PRISM2_NON_VOLATILE_DOWNLOAD/': ${S}/hostap_config.h Sep 10 00:01:48 AvengerMoJo: the embeddix kernel which Sharp released for the Cxx00 devices pretends to be 2.4.20, yes Sep 10 00:01:56 abm_y4k: use what works..... Sep 10 00:02:16 JustinP: if I switch back to 2.4 do I need to clean up everything? Sep 10 00:02:28 oh I though the bb file will be different, so I just want to double check. Sep 10 00:02:34 sorry for bothering Sep 10 00:02:37 you probably need to rebuild with that, yes Sep 10 00:03:01 abm_y4k: depending on where you're putting things it's basically the same as writing ont he command-line Sep 10 00:03:12 abm_y4k: yu don't need the '' ont he command-line in some cases Sep 10 00:04:46 thankx Sep 10 00:07:58 in the hostap package, there are different version, 0.3.7,0.4.4,0.3.9 what should to just build the hostap package? bitbake hostap? Sep 10 00:08:44 yes Sep 10 00:08:55 ok thankx again Sep 10 00:09:16 that will choose the "preferred" or "newest" version depending on vairous thingd Sep 10 00:25:15 what is this mean: ERROR: Edit your conf/local.conf: BBFILES? Sep 10 00:27:42 it means "edit your conf/local.conf" Sep 10 00:28:28 perhaps your BBFILES setting is wrong? Sep 10 00:31:25 oh Sep 10 00:38:43 I have lots of errors: Sep 10 00:38:50 NOTE: Handling BitBake files: \ (0090/2737) [ 3 %]ERROR: binutils.inc:107: unparsed line: 'pkg_postinst_${PN}-symlinks () {' while parsing /usr/src/zaurus/oz/org.openembedded.dev/packages/binutils/binutils-cross_2.16.bb Sep 10 00:39:04 whats wrong? Sep 10 00:39:47 probably something wrong with your configuration Sep 10 00:39:50 or you environment Sep 10 00:39:55 did you set up your BBPATH? Sep 10 00:40:36 you also need to alter your path if you're building a 2.4 Zaurus kernel (after installing the 2.95 toolchain) Sep 10 00:40:40 this is all in the wiki Sep 10 00:41:35 [ztester@fc1 oz]$ echo $PATH Sep 10 00:41:35 /usr/src/zaurus/oz/bitbake:/usr/kerberos/bin:/usr/local/bin:/bin:/usr/bin:/usr/X11R6/bin:/home/ztester/bin Sep 10 00:41:35 [ztester@fc1 oz]$ echo $BBPATH Sep 10 00:41:35 /usr/src/zaurus/oz/build:/usr/src/zaurus/oz/org.openembedded.dev Sep 10 00:41:54 this shold be fine rite? Sep 10 00:43:29 well...I don't see the 2.95 toolchain in your PATH but that shouldn't cause that error Sep 10 00:44:21 oh I found this: ASSUME_PROVIDED = "virtual/arm-linux-gcc-2.95" Sep 10 00:44:33 maybe this is the causes Sep 10 00:44:49 no, that shouldn't cause a problem Sep 10 00:44:50 hm nope, still give me the same error Sep 10 00:44:54 not that error anyway Sep 10 00:45:12 and you need that ASSUME_PROVIDED and that toolchain if you're building a 2.4 embeddix kernel Sep 10 00:45:27 you may also want to set up your PYTHONPATH Sep 10 00:45:32 export PYTHONPATH="/home/papercrane/oe/bitbake/lib:$PYTHONPATH" Sep 10 00:45:48 it could be a problem in your local.conf Sep 10 00:49:06 will this a problem: NOTE: Psyco JIT Compiler (http://psyco.sf.net) not available.? Sep 10 00:50:26 the error above does not always show up, only on some .bb files Sep 10 00:50:57 you don't need Psycho, but it can make it faster Sep 10 00:51:04 if you get those parse errors you have a problem Sep 10 00:51:58 do you think bitbake version problem? Sep 10 00:52:55 same error: unparsed line: 'pkg_postinst_${PN} () Sep 10 00:53:10 something about this pkg_postinst function Sep 10 00:53:12 do you have the newest version from svn? Sep 10 00:53:46 I did this: svn co svn://svn.berlios.de/bitbake/trunk/bitbake Sep 10 00:53:58 ok Sep 10 00:54:10 perhaps you need to set your PYTHONPATH as well? Sep 10 00:54:24 echo PYTHONPATH is empty Sep 10 00:56:18 I did a little tweak to python, I have manually install Python 2.4.1 and my python version was previously 2.2 on FC1, will this cause the problem? Sep 10 00:56:53 I don't know Sep 10 00:57:08 I suggest you set your pythonpath to include yout bitbake directory Sep 10 00:57:21 and it could eb a conflict of python versions as well.... Sep 10 00:57:25 Jeez...FC1?? Sep 10 00:57:54 Fedora core 1 Sep 10 00:58:46 I know what it means Sep 10 00:58:50 aren't they at 4 now? Sep 10 00:58:56 yeah Sep 10 00:59:16 * JustinP doesn't understand people who don't update Sep 10 00:59:30 update is a matter of perspective Sep 10 00:59:56 * JustinP whistles Sep 10 01:00:05 haha yeah, Sep 10 01:00:20 you can still manually compiel and keep your system "updated" Sep 10 01:00:20 oh well, let me backup my data, and install FC4 than Sep 10 01:00:38 do you *really* have to re-install to update FC? Sep 10 01:00:44 * JustinP whistles again Sep 10 01:00:46 no Sep 10 01:00:47 from fc1 yes Sep 10 01:00:57 emte: reallY? Sep 10 01:01:00 there is a problem between fc1 and fc2 Sep 10 01:01:06 arrrr Sep 10 01:01:34 wheeee Sep 10 01:01:44 or atleast it still existed the last time i tried to update my gf's system Sep 10 01:02:13 how about go straight to fc4 from fc1? Sep 10 01:02:17 which i should reinstall one day ... Sep 10 01:03:04 i doubt you can, or that the time is well spent Sep 10 01:03:11 reinstall time :( Sep 10 01:03:18 would be faster to reinstall Sep 10 01:03:24 thanks for the help all :) Sep 10 01:03:29 bibi Sep 10 01:03:45 I'll be back!~~ Sep 10 01:03:50 would take many many hours to update that far Sep 10 01:04:09 and maybe 30min-1 hour to reinstall Sep 10 01:04:11 yea, simply re-install better for me :) Sep 10 01:05:56 anyone ever use MS's UNIX Services? Sep 10 01:07:17 * JustinP uses Gentoo and simply updates every few days....painless Sep 10 01:13:36 yeah i need to update mysystem cept its a big one this time Sep 10 01:13:44 libc and glibc Sep 10 01:14:03 ::shrug:: just start it and go to sleep Sep 10 01:14:12 thena revdep-rebuild Sep 10 01:14:14 updating those in any fashion makes me a bit nervous Sep 10 01:14:28 I have had no issues with the last few on my system Sep 10 01:14:37 * JustinP is off to sleep Sep 10 01:14:38 'night Sep 10 01:14:41 night Sep 10 01:17:35 JustinP: have you heard of this driver spectrum24t-1.03_cf.tar.gz Sep 10 02:16:59 hum ... Sep 10 02:25:05 JustinP: any idea what they are talking about I'm using the exact card http://www.oesf.org/forums/index.php?showtopic=12949&hl=cwb1000 Sep 10 02:32:51 hey guys I'm taking a OE hiatus Sep 10 02:33:00 Justin pissed me off just too much Sep 10 02:34:12 anyone having probs building 3.5.4 opie-mediaplayer1 ?? twice its failded on meta-opie Sep 10 02:35:12 mmm... "NOTE: package module-init-tools-cross-3.1: failed" AND "ERROR: Build of module-init-tools-3.2-pre7 failed"... what did i do wrong ? Sep 10 02:36:33 alan|home: what about checking the logs? Sep 10 02:45:23 zecke: i'm a bit lost with all this. I'm doing my best to understand how everything work, but it's difficult : computering is not my default formation. Here is the first error : http://pastebin.ca/22501 and here is the second : http://pastebin.ca/22564. Do you think you could explain me what i did wrong and let me find the correct thing to do myself ? Sep 10 02:47:30 alan|home: Just to understand my reasoning. I think this is FreeSoftware and we're all here to improve our skills Sep 10 02:47:42 alan|home: from your first log, what could be the error? Sep 10 02:47:51 alan|home: at least on which line is the error? Sep 10 02:48:17 alan|home: the second error looks weird though Sep 10 02:49:33 zecke: i'm i right thinking the main error is "ERROR: TaskFailed event exception, aborting" ? Sep 10 02:50:14 alan|home: let us start with the first log Sep 10 02:50:26 alan|home: as a said the second (TaskFailed) looks weird anyway Sep 10 02:53:09 zecke: i see in the first log "/usr/bin/ld: cannot find -lc". is my /usr/bin/ld version out dated ? i don't even know what ld is used for... Sep 10 02:55:20 alan|home: ld is the linker Sep 10 02:55:30 alan|home: on linking 'libc' is not found Sep 10 02:55:35 whatever libc is ;) Sep 10 02:56:26 alan|home: do you do parallel make? Sep 10 02:57:16 zecke: no. Sep 10 02:58:07 alan|home: what distribution are you using for compiling - you may be missing the native libc static libs Sep 10 02:58:39 mandriva 2005 (newbie distro ; ) ) **** ENDING LOGGING AT Sat Sep 10 02:59:56 2005