**** BEGIN LOGGING AT Thu Sep 08 02:59:56 2005 Sep 08 03:00:01 hrw|work: no argument here Sep 08 03:00:47 alan|home: seven minutes past the hour, every hour (that's PST hours; not ACST hours, which would be 30 minutes different) Sep 08 03:01:15 hrw|work: does bzip2 have an --rsyncable option? Sep 08 03:04:08 (cause we care about the network traffic we use every hour to update the remote hosting site which hosts it, rather than the once-off time for a user to download it) Sep 08 03:06:09 hi Sep 08 03:06:22 hi Sep 08 03:06:29 hi avenger Sep 08 03:07:36 irda.. fail to build Sep 08 03:08:50 rwhitby-away: hmm.. dont know;( Sep 08 03:30:06 hi guys Sep 08 03:30:14 hi Bernardo Sep 08 03:35:04 hi Sep 08 03:35:13 hi bernado Sep 08 03:47:16 RP: Any thoughts on the backlight issues? Anyone know how Opie handles it which allows it to still set the backlight? Sep 08 04:16:55 .... Sep 08 04:17:02 irda-fail :( Sep 08 04:18:57 anyone have any suggestions as to what I might have broken to get a fresh opie-image which doesn't seem to have any fonts? Sep 08 04:19:56 show me /usr/lib/ipkg/status or resulted image Sep 08 04:20:05 you can take it from tmp/rootfs/ Sep 08 04:24:05 having error compiling irda-utils... /tmp/oetmp/staging/arm-linux/include/gconv.h:72: error: parse error before "size_t" Sep 08 04:26:31 gridzero: thx - you found a bug Sep 08 04:27:10 I will fix it in a minute/two probably Sep 08 04:27:44 np... glad to help :) Sep 08 04:29:19 gridzero: can you add "opie-freetype" to OPIE_BASE into pacakges/meta/opie-collections.inc and rebuild opie-image? Sep 08 04:29:40 sure.... Sep 08 04:29:49 thx Sep 08 04:30:03 I have too much changes there to be able to test it Sep 08 04:44:58 gridzero: now I have to pull/merge/update and will push that update Sep 08 04:47:55 hm obexpush needs -Wl,-rpath-link Sep 08 04:54:07 hrw - if you still have a simpad (can't remember) - externe.net/zaurus/simpad-bluetooth : now bluetooth+mmc work Sep 08 04:54:29 along with 128 Mb of RAM, this makes a simpad with super cow power. Sep 08 04:57:04 guylhem1: I never even saw simpad Sep 08 04:57:18 I have a simpad Sep 08 04:59:46 oops sorry hrw Sep 08 05:00:11 woglinde: now running the latest OE image. I'm going to replace the kernel, fix some scripts and hopefully release it soon Sep 08 05:00:18 mmc rocks! Sep 08 05:00:50 guyhelm with the actual hh 2.6 kernel the tft is broken Sep 08 05:03:08 I know. I'm sticking to 2.4 - even if that's not popular :- Sep 08 05:05:37 03hrw 07org.oe.dev * r720e5c15... 10/packages/meta/ (opie-collections.inc opie-image.bb): Sep 08 05:05:37 opie-image: added opie-freetype into OPIE_BASE Sep 08 05:05:37 - was missing due to recent change in opie-freetype and opie-ttf-support depends Sep 08 05:08:10 hrw|work: any reason my stage linux include cause irda-utils fail to build? Sep 08 05:09:24 no Sep 08 05:37:54 guylhem1: if I could get someone to solder me the smc bluetooth module... ever thought of selling "upgrade kits" for the simpad? Sep 08 05:42:21 Bernardo: that would be interesting, but I lack the manufacturing capabilities. I only do reference design and prototypes. Sep 08 05:42:34 But I'm sure I could find a company that could prepare everything. Sep 08 05:43:11 Please ask segor.de - that's where my bluetooth modules come from. I'll contact them. Sep 08 05:44:00 give me your email Sep 08 05:46:10 I'm asking them if they could sell "already made" module, like they do for mp3 Sep 08 05:46:38 that would be nice. Sep 08 05:49:51 did you get my email? Sep 08 05:52:51 send to gmail 67 Sep 08 05:52:54 nope Sep 08 05:55:53 on a private message? Sep 08 05:56:01 ahhh Sep 08 05:56:29 neither Sep 08 06:02:08 ok, will send now Sep 08 06:03:06 shouldn't be a problem, since I just got one of your emails on the mailing list Sep 08 06:03:08 guylhem Sep 08 06:03:11 at gmail.com Sep 08 06:03:11 just read your email, forget it Sep 08 06:04:27 now all I need is a few years free time to start a "kids opie" for my daughter to use the simpad when she grows up a little... :) Sep 08 06:04:44 guyhelm you could contact wolfsoft also Sep 08 06:04:44 lol Bernardo :) Sep 08 06:06:24 woglinde: why? they do 128 M;b of ram - just send them your simpad. Sep 08 06:06:28 use my kernel and it works fine Sep 08 06:09:52 guyhelm if you want sell the mmc bluetooth and on Sep 08 06:10:00 add-on Sep 08 06:10:26 args sorry Sep 08 06:10:34 misunderstand it Sep 08 06:11:53 03mickeyl 07org.oe.dev * r50b29d3f... 10/packages/efl/ (5 files): efl: *) September is the 9th month... *) ship emotion plugins Sep 08 06:15:10 does wolfsoft have a english website? Sep 08 06:15:59 wow, the ram upgrade is expensive... Sep 08 06:18:21 how much? Sep 08 06:19:54 woglinde: wolfsoft only do ram upgrade. I think segor could do mmc/bluetooth Sep 08 06:20:24 I offered them to take my reference design for free (well - I asked another WML since they sold me a defective one and I can't find the bill...) Sep 08 06:24:44 hrw|work: 189 euros Sep 08 06:27:56 too much Sep 08 06:28:01 more then simpad is worth Sep 08 06:36:21 cool. got glibc up to packaging on osx. :) Sep 08 06:36:59 hrw|work: my thoughts exactly Sep 08 06:37:48 fyi gonna submit a couple of quick glibc patches to allow building on case-insensitive case-preserving filesystems (hfs[+], ntfs). just moves from using "oS" and "os" to using "on" and "os". Sep 08 06:46:53 yeah 128 mb is to much Sep 08 06:48:09 args 189 euros Sep 08 06:48:22 damnit whats wrong with me Sep 08 06:49:01 lol Sep 08 07:13:17 whee, the tailor failed again Sep 08 07:13:18 hehe Sep 08 07:13:46 kergoth double cross fingers Sep 08 07:13:51 kergoth: hi Sep 08 07:13:53 whats the monotone command to view the log of a given revision? Sep 08 07:13:56 * chouimat is listening to "The Vice of Killing" by Sodom on Eine Nacht in Bangkok [amaroK] Sep 08 07:14:33 kergoth: mt log revision Sep 08 07:15:12 kergoth: in tailor you use 'monotone automate' or parse output of normal commands? Sep 08 07:16:37 it uses monotone automate Sep 08 07:16:47 i didnt write the monotone support, i'm just fixing bugs :) Sep 08 07:17:11 ah.. Sep 08 07:17:24 i was goign to, but someonje got to it first Sep 08 07:39:05 * RP sees 11 emails from rmk in his inbox :-/ Sep 08 07:39:37 rejected, rejected, rejected? Sep 08 07:39:38 03mickeyl 07org.oe.dev * rfbd660b2... 10/packages/efl++/ (5 files): efl++: bump to v0.0.5 Sep 08 07:40:09 hrw|work: near enough. He's not happy about my patches sent to akpm... Sep 08 07:40:42 akpm is arm kernel? Sep 08 07:41:36 hrw|work: I got fed up with rmk ignoring my emails... Sep 08 07:41:57 and I had other reasons, which I'd explained and have now reexplained to rmk Sep 08 07:46:27 heh.. Sep 08 07:52:16 ~apkm Sep 08 07:52:26 ibot: ? Sep 08 07:54:53 arm patch tracker suxx Sep 08 07:59:40 hello everyone! Has anyone done a .bb for the 'newt' package? Sep 08 08:02:59 * chouimat is away: food Sep 08 08:05:31 some people should be shoot - like programmer which wrote website which I have to fix Sep 08 08:05:47 heh Sep 08 08:07:09 he liked to fetch the same data from database 3 times in a row Sep 08 08:07:51 ow Sep 08 08:08:06 and lot of magic numbers Sep 08 08:08:10 if ($row['dzial_function'] == 4) { Sep 08 08:08:20 wtf is 4? :) Sep 08 08:15:17 hrw|work: akpm = Andrew Morton, maintainer of the -mm tree Sep 08 08:15:25 aha Sep 08 08:15:44 RP: I'm not familiar with all those linux acronyms Sep 08 08:16:28 hrw|work: You are now ;-) Sep 08 08:17:30 hey guys - is there a tutorial or howto on how to create a .bb file? I need to get the 'newt' library added to my image, but having trouble with the basics of 'configuring' the tarball using bitbake Sep 08 08:21:45 uh... e2fsprogs sets sbindir as /sbin. Is there a reason for this? Sep 08 08:23:00 methinks sbindir shouldn't be set in a .bb Sep 08 08:29:22 NAiL: we want fsck.ext2 and friends in /sbin, not /usr/sbin Sep 08 08:30:56 kergoth: base_sbindir? Sep 08 08:31:36 Not all dists will have writable /sbin, will they? :) Sep 08 08:31:37 yeah, it should be setting sbindir to that.. which is still unpleasant, but at least it isnt hardcoded :P Sep 08 08:31:42 * kergoth nods Sep 08 08:31:56 ok, I'll fix that Sep 08 08:32:01 i created those variables with the intent of fixing all the hardcoded paths, but there are a _lot_ of them Sep 08 08:32:15 copy all the *dir variables from bitbake.conf into local.conf Sep 08 08:32:17 change them all Sep 08 08:32:31 i.e. libdir=/libdir, sbindir=/sbindir, etc Sep 08 08:32:33 do a build Sep 08 08:32:37 watch how few packages actually obeyed it. Sep 08 08:32:40 its depressing Sep 08 08:32:53 Its getting better Sep 08 08:32:56 thats one of hte big issues surrounding making openslug packages work on unslung, actually Sep 08 08:33:11 CosmicPenguin: tailor failed again. new error this time. at least i'm not bored.. Sep 08 08:33:16 heh Sep 08 08:33:16 nice variety of failures Sep 08 08:33:18 Trying to make e2fsprogs make mkfs.ext[23] and fsck.ext[23] into symlinks... shave off ~150k IIRC :) Sep 08 08:33:23 * CosmicPenguin is happy to let kergoth work out the bugs... :) Sep 08 08:33:32 NAiL: they should be hardlinks right now, afaik Sep 08 08:33:40 in which case you wouldnt save anything Sep 08 08:33:55 do an ls -l on them and look at hte link count to check Sep 08 08:33:59 hmm.. apparently not, since removing them in flash saves me some space Sep 08 08:34:01 * kergoth prefers symlinks in any case Sep 08 08:34:57 yeah Sep 08 08:35:03 on my desktop, they're hard links. just a bug, i guess Sep 08 08:35:05 ~/code/linux-2.4.24$ ls -l /sbin/fsck.ext2 Sep 08 08:35:05 -rwxr-xr-x 3 root root 134028 2005-03-17 13:09 /sbin/fsck.ext2 Sep 08 08:35:17 odd Sep 08 08:35:58 Does jffs2 handle hardlinks properly? Sep 08 08:36:05 good question Sep 08 08:38:05 well.. it does if I create them manually, atleast Sep 08 08:38:31 03koen 07org.oe.dev * r4c3dd1fc... 10/packages/gaim/gaim_2.0.0-cvs-mco1.bb: Sep 08 08:38:31 gaim 2.0: update bb to fetch patches out of hh.org cvs via cvsweb, courtesy Mads Olesen. Sep 08 08:38:31 closes bug #310 Sep 08 08:38:43 Somebody should make sure genext2fs handles them too Sep 08 08:38:55 and mkcramfs - jffs2 isn't the only filesystem used by bitbakers Sep 08 08:39:26 no, but it's the only fs I use ;) Sep 08 08:39:35 a common sickness Sep 08 08:39:52 i use ext2 all the time for my uml and stuff Sep 08 08:39:57 and tar.bz2 for nfs Sep 08 08:40:26 I would settle for x86 to get 1/8th the attention of Opie/GPE Sep 08 08:40:48 if this tailor ever finishes, i might actually start contributing to oe again :P Sep 08 08:40:53 But no worries, thats human nature Sep 08 08:40:56 you hack on what you care about Sep 08 08:41:38 indeed Sep 08 08:45:25 kergoth: tailor? Sep 08 08:46:34 ~tailor Sep 08 08:46:52 hm Sep 08 08:46:54 ~noibot? Sep 08 08:47:16 ibot: Tailor is a tool to migrate changesets between ArX, Bazaar, Bazaar-NG, CVS, Codeville, Darcs, Git, Mercurial, Monotone, Subversion and Tla repositories. See http://www.darcs.net/DarcsWiki/Tailor Sep 08 08:47:18 kergoth: okay Sep 08 08:47:48 using it to produce an svn repo out of the oe monotone repo, for testing out svk Sep 08 08:52:35 kergoth: neat. quick skim of it.. it should be pretty good at converting a cvs repository to a subversion repository? cvs2svn seems kinda sketchy watching it go Sep 08 08:53:31 a bit of overkill though it looks like Sep 08 08:54:14 yeah it works well for that. i was testing it on tslib's repo Sep 08 08:54:32 retains all history, even sets the svn author attribute if you configure your svn repo to allow it Sep 08 08:54:37 i just pointed my coworker who's investigating cvs->svn migration here Sep 08 08:54:47 neat Sep 08 08:55:31 whats cool is that it can be used not just for migration, but to maintain a mirror. we could use it to make an anonymous svn mirror of oe available for those who hate monotone, for example Sep 08 08:55:35 doesn't look like it does tag/branch preservation Sep 08 08:55:57 nod. thatd be nice Sep 08 08:56:38 that's the main thing we want to get right here Sep 08 08:56:38 itd be hard for it to find the merge points between branches in the cvs repo, since cvs doesnt track such things.. you'd have to tell it the naming scheme of your tags, assuming you tag your merge points on each branch.. Sep 08 08:56:51 yeah Sep 08 08:57:35 course svn doesnt track merge points either, at least not directly. svk does though ;) Sep 08 08:58:17 svk does through log entries right? Sep 08 08:58:31 cu Sep 08 08:59:01 gb2: attributes Sep 08 08:59:06 ah Sep 08 09:02:55 03nail 07org.oe.dev * r2ed62cd8... 10/packages/e2fsprogs/e2fsprogs_1.38.bb: Sep 08 09:02:55 e2fsprogs: Fix the use of sbindir (changed to base_sbindir) Sep 08 09:02:55 -mke2fs: Change mkfs.ext[23] to symlinks Sep 08 09:02:55 -e2fsck: Chanke fsck.ext[23] to symlinks Sep 08 09:03:04 grr... Sep 08 09:03:14 monotone status says it lists the files changed, but I can't get that to work Sep 08 09:03:22 * kergoth considers whatd need changing in tailor to add the ability to retain information about changeset flow among branches/tags Sep 08 09:04:13 gb2: retaining tags/branches alone would be easy.. its just a matter of converting i.e. a cvs tag to an svn cp .. and a bunch of linear changeset histories. gets quite a bit more complex if you want to try to retain the actual flow among the branches by tracking/storing merge points Sep 08 09:06:15 s/cvs tag/cvs rtag/ Sep 08 09:06:25 svn has nothing like cvs tag, other than attributes Sep 08 09:09:40 Good evening. Sep 08 09:09:47 kergoth: yeah some of the tagging stuff doesn't translate great to svn Sep 08 09:10:27 kergoth: one thing our engineers commonly like to do is look at what tags a particular file has, and which tags correspond to particular revisions in the log Sep 08 09:10:33 which svn doesn't do Sep 08 09:12:04 it's been interesting sofar going through the process of figuring out how to redo the layout of our whole repository and how workflow will change Sep 08 09:13:16 03nail 07org.oe.dev * rca142ce9... 10/conf/distro/openslug.conf: Sep 08 09:13:16 openslug.conf: Remove reiserfs support from default image. Sep 08 09:13:16 Saves ~1mb in flash. If anyone wants it, they can ipkg install it. Sep 08 09:13:20 kergoth: best solution i've thought of for that tag case i've thought of involves walking the tree Sep 08 09:14:09 nod. svn cp all the files with a given tag to the corresponding subdir within a dir in tags/? Sep 08 09:14:34 kergoth: that's the standard way of converting yeah Sep 08 09:16:04 well, adding conversion of rtags to tailor should be painless. adding conversion of tags slightly less so. adding conversion of actual merge point metadata much less so ;) Sep 08 09:16:12 its just python Sep 08 09:18:54 morning kergoth Sep 08 09:19:04 hey Sep 08 09:21:53 kergoth: what do you mean by rtags? Sep 08 09:21:56 vs. tags Sep 08 09:22:19 cvs doesn't make a distinction. a tag is just applying the tag to the revision which is in your working dir Sep 08 09:22:32 ahh right Sep 08 09:22:46 the distinction is only in the user's interface, not how the info is stored Sep 08 09:22:47 gotcha Sep 08 09:22:53 * kergoth misunderstood how it worked Sep 08 09:23:04 the thing i was thinking of though, cvs log foo.c will show you all the tags on that file. but you can't get the same sort of thing out of svn Sep 08 09:23:35 yeah, itd have to traverse your tags/ directory Sep 08 09:23:41 so you'd have to teach it about your repo layout Sep 08 09:23:54 yep. it would be easier if svn allowed you to follow a copy forward Sep 08 09:24:14 you'd track back, and follow all copies forward. as it is now, you have to look at a lot more. Sep 08 09:24:57 actually, i dont think svn even tracks back, at least not via metadata. you'd have to check the top revision of the svn repo to see where it branched off. no merge point metadata, remember Sep 08 09:25:00 I was wondering if you could build up a tree representation in a database, and have a tool query that, and make it capable of being updated from new revisions applied to the repo Sep 08 09:25:27 you can look at the head of all your tags, and walk backwards and see where it was copied from Sep 08 09:25:33 thats essentially how tailor works now. it has an internal changeset store that it uses, and each type of scm has code to convert data to/from that Sep 08 09:25:42 not a full tree though, just a changeset stack Sep 08 09:25:51 * kergoth gets a soda Sep 08 09:26:01 we're doing a poland springs order :) Sep 08 09:26:14 we discovered they're really cheap if your order directly Sep 08 09:26:24 NOTE: package e-image-1.0: completed Sep 08 09:26:42 now what? I have a bulid/tmp/rootfs that only has ipkg stuff and an Xfbdev in it Sep 08 09:33:19 the rootfs is only 524288 bytes long; JustinP, I didn't know you were able to condense e-image down that small ;-) Sep 08 09:33:47 hehe Sep 08 09:33:48 impressive! Sep 08 09:34:06 recursive gzip, maybe? ;-) Sep 08 09:35:54 hi ppl Sep 08 09:36:01 just a short question: Sep 08 09:36:24 are there linux pdas/handhelds out there for which flashplayer6 exists?? Sep 08 09:36:28 hey mreimer Sep 08 09:36:34 hey koen|slug Sep 08 09:36:37 mreimer: removing 20MB of fonts helps :) Sep 08 09:36:46 wow, yeah Sep 08 09:37:28 koen|slug: what do you think about renaming h3900.conf to ipaq-pxa.conf? That would seem to reflect its purpose better Sep 08 09:38:12 I proposed that a while ago Sep 08 09:38:28 maybe it's time to propose it again Sep 08 09:38:47 I second your proposal Sep 08 09:38:58 the proposal is now open for discussion :-) Sep 08 09:39:01 sumone knows? Sep 08 09:39:13 actually ipaq-pxa would be confusing as well Sep 08 09:39:39 because of the pxa270 version? Sep 08 09:39:42 as we use it for asus (and probably axim) harwdare too Sep 08 09:39:51 ah Sep 08 09:39:58 so how about pxa.conf? Sep 08 09:40:15 the pxa270 is necessary because we can't build a mixed pxa255 and pxa270 kernel Sep 08 09:40:17 (yet) Sep 08 09:40:43 a hint where to find out would be helpful, too .. google, macromedia & linux sites didn't help much yet :/ Sep 08 09:40:44 we support pxa250/255/26x with h3900.conf, right? Sep 08 09:40:44 handheld-pxa would be my suggestion Sep 08 09:41:12 and rename ipaq-pxa270 to handheld-pxa270? Sep 08 09:41:26 phaidros_: I think then nokia 770 supports flash, but I'm not sure Sep 08 09:41:36 and merge h1910/h1940/h3600 to handheld-pxa? Sep 08 09:41:59 h3600 is sa Sep 08 09:42:05 oh yeah Sep 08 09:42:07 and one of the 1900s is samsung Sep 08 09:42:26 the 1910 config was necessary because of the different jffs2 params Sep 08 09:42:31 doh, that's right Sep 08 09:46:20 conf/distro/familiar-0.8.3.conf has a typo: PREFERRED_VERSION_hostap-modules ?= "0.3.9 Sep 08 09:46:26 notice the lack of a closing quote Sep 08 09:46:26 which is why I want something like a'subclass' Sep 08 09:47:03 koen|slug: you mean so we could subclass, say, h3900 and h2200 from handheld-pxa? Sep 08 09:47:50 yes Sep 08 09:47:56 I like that idea Sep 08 09:48:01 and have less useless modules lying around Sep 08 09:48:07 yes please! Sep 08 09:48:35 implementing it isn't straightforward Sep 08 09:48:49 and it's a touchy topic Sep 08 09:48:55 03koen 07org.oe.dev * r9fe75450... 10/conf/distro/familiar-0.8.3.conf: conf/distro/familiar-0.8.3.conf: add missing ", thanks to Matt Reimer for spotting it Sep 08 09:48:58 why? Sep 08 09:49:00 thanks koen Sep 08 09:49:13 control vs incompatible rootfses Sep 08 09:49:59 by "incompatible" do you mean that not having one rootfs to rule them all? Sep 08 09:50:06 yes Sep 08 10:27:10 good evening! Sep 08 10:28:18 i'm at the point to assemble my first initrd. but i couldn't fine any detailed documentation. does anybody know such? Sep 08 10:30:31 Henry78: the *-image.bb should give a decent starting point. Sep 08 10:30:51 * Disconnect watches glibc package (possibly successfully!! :) ..) it's 1000 or so locale packages. Sep 08 10:32:12 thats funny. in an hour I have to go meet with mainframe guys to discuss linux on the mainframe. from zaurus to mainframe in seconds.. wonder if I'll sprain something. Sep 08 10:32:55 it's hard to be as flexible as linux. Sep 08 10:33:05 thanks for the tipp, i'll have a look. Sep 08 10:34:58 yay! osx-oe builds past glibc now. Sep 08 10:35:07 * Disconnect doubts it builds -much- past glibc, but every step is progress... Sep 08 10:36:31 can't find much in the *-image.bb which helps me to start building my first(!) initrd Sep 08 10:39:15 Henry78: pick one you like, copy it over, change the dependencies.. Sep 08 10:39:35 anyone want to test my build-on-osx patch against a non-osx target? (in theory it'll work fine.. in theory..) Sep 08 10:41:11 ..if so its linked from the BuildOnOSX wiki page Sep 08 10:41:50 Disconnect: change the dependencies? how? Sep 08 10:45:34 take xfce-image.bb .. its got a nice simple (and short) dependency setup.. determine what you need on your initrd and go from there. Sep 08 10:45:59 then just copy over the packages? Sep 08 10:46:00 if you haven't already, read the (somewhat sparse) bb docs in the wiki Sep 08 10:46:12 no, then just bitbake the image. Sep 08 10:46:19 i'll have a look at the docs Sep 08 10:46:20 eg 'bitbake test-image' Sep 08 10:47:00 oho. bitbake is able to do images! that's nice. so i'll read some documentation now Sep 08 10:47:31 err: wanted to say: "bitbake is able to creat initrds" Sep 08 11:05:25 hi Sep 08 11:07:12 * chouimat is back. Sep 08 11:12:09 kergoth? Sep 08 11:12:14 03jbowler 07org.oe.dev * rd69e4956... 10/packages/uclibc/ (4 files in 2 dirs): Sep 08 11:12:14 uclibc: put ARM thumb patches from 0.9.27 into 0.9.28 Sep 08 11:12:14 Some of the thumb patches had already been pushed into the uclibc release, Sep 08 11:12:14 this commit adds corrected/updated versions of most of the ones which had Sep 08 11:12:14 not been accepted. The controversial 'change the ldso symbol resolver' Sep 08 11:12:15 patch is not in this set - it will be re-worked to match the changes being Sep 08 11:12:17 made in binutils for the same problem. Sep 08 11:15:40 lamikr: yes? Sep 08 11:21:12 Are you maintaining linux-libc-headers? I think there is small omap error, but I do not know how it would be fixed best. Sep 08 11:23:16 I think linux-libc-headers-2.6.11.1.bb should have option to copy arch omap files to arch dir in a similar way than it is done if arch is arm. (cp -a include/asm/arch-omap include/asm/arch) The problem is just how to make the if correctly as omap processors are also arm ... Sep 08 11:24:22 I have actually tested that and it has not caused any problems while building or executing gpe-image in h6340. Sep 08 11:25:17 userspace applications should not be depending on asm/ headers matching runtime Sep 08 11:25:28 nor should they expect linux/ headers to match the runtime kernel Sep 08 11:28:39 Ok, it just looked weird as there were also these tests like: if test "$ARCH" = "arm"; then cp -a include/asm/arch-omap include/asm/arch. But if you think this is ok, then ok. Sep 08 11:29:34 Sorry... Copy pasteing wrong... It has Sep 08 11:30:14 if test "$ARCH" = "arm"; then cp -a include/asm/arch-ebsa285 include/asm/arch. Why this? Sep 08 11:32:12 hey mithro Sep 08 11:34:54 what bb file contains the wxwigets bitbake configuration script? greping "wx" doesnt find anything Sep 08 11:36:07 i guess i could just get vlc-gpe because im quite sure it uses wxwidgets Sep 08 11:40:09 RP: 2.6.13-mm1 backlight seems to be totally broken, takes 30 secs to write a number to /sys/class/backlight/corgi-bl/brightness and then just turns off backlight no matter what number is sent Sep 08 11:58:02 hey koen|slug Sep 08 12:00:40 RP: actually make that is never finishes the write :-( Sep 08 12:02:37 backlight control is fluff anyway Sep 08 12:02:50 koen: I like to see my screen :-( Sep 08 12:04:01 * koen would like to have gprs on his PDA Sep 08 12:04:04 ~stab ppp Sep 08 12:04:07 * ibot runs at ppp with an origami Swiss Army knife, and inflicts a nasty paper cut. Sep 08 12:04:14 wazzup with ppp? Sep 08 12:04:36 it gets a termreq after a succesfull PAP login Sep 08 12:04:41 hey Pigi Sep 08 12:04:50 Ciao all Sep 08 12:04:53 hey koen Sep 08 12:04:54 while giving the same output in ppp.log as my powerbook Sep 08 12:05:02 koen: :-( I use it on Z Sep 08 12:46:10 03koen 07org.oe.dev * rcf261b9a... 10/packages/galago/ (libgalago-0.3.3/no-check.patch libgalago_0.3.3.bb): libgalago: add libgalago to OE Sep 08 12:52:54 03koen 07org.oe.dev * rf4f18562... 10/packages/galago/ (libgalago-0.3.3/no-check.patch files/no-check.patch): galaga no-check.patch: move to files, since it can be shared with galago-daemon Sep 08 12:53:21 may the quilt be with you Sep 08 12:56:04 03jbowler 07org.oe.dev * rb8529852... 10/packages/freeze/files/freeze: Sep 08 12:56:04 freeze: Add support for uclibc_svn build Sep 08 12:56:04 freeze did not have patterns to handle svn, as opposed to cvs, packages. Sep 08 12:56:04 This commit just adds the pattern required to recognise the uclibc_svn Sep 08 12:56:04 package. Sep 08 13:11:24 ~lart me Sep 08 13:11:25 * ibot beats koen severely about the head and shoulders with a rubber chicken Sep 08 13:11:46 the check for /proc/self/maps can be avoided by setting --disable-binreloc Sep 08 13:12:02 * koen tries to remember which packages failed on that Sep 08 13:33:46 03koen 07org.oe.dev * r45599c4d... 10/packages/galago/galago-daemon_0.3.4.bb: galago-daeom: add 0.3.4 to repo Sep 08 13:33:50 03koen 07org.oe.dev * r5d7867f5... 10/packages/galago/libgalago_0.3.3.bb: packages/galago/libgalago_0.3.3.bb: fix staging and inherit pkgconfig Sep 08 13:36:20 XorA|gone: backlight is a known issue Sep 08 13:36:32 Although I don't think its my fault Sep 08 13:47:30 hello everyone! I finally got 'newt' to compile, but now the other program I'm trying to link to '-lnewt' says it can't find it. Sep 08 13:47:50 dwildes: did you stage the libraries? Sep 08 13:48:21 I done a --verbose of 'ld' - and it appears it the paths are looking in: 'tmp/cross/i386-linux/bin/../../i386-linux/', but mine installed to: 'tmp/staging/i386-linux/lib' Sep 08 13:48:37 sounds like that program isnt obeying oe's LDFLAGS. fix that. Sep 08 13:49:25 JustinP: I put these two entries in 'do_stage' 'oe_libinstall -so libnewt ${STAGING_LIBDIR}' and 'install -m 0644 newt.h ${STAGING_INCDIR}' Sep 08 13:49:56 kergoth: thanks - I'll grep for some other 'LDFLAGS' options and see what I find out Sep 08 13:50:15 oe's ldflags includes a -L that points at staging, so it looks there for libraries. Sep 08 13:50:27 dwildes: sounds like LDFLAGS then Sep 08 13:51:07 yeah - cause I can run the 'ld' with the '-L' flag pointing to my staging directory, and all is well Sep 08 13:51:21 cool! I'll get this fixed in a jiffy now - thanks guys Sep 08 13:54:42 RP: irda fail to build ....any idea? Sep 08 14:23:55 is there documentation outlining how to add projects to open embedded? Sep 08 14:24:34 if it exists, its in the wiki :P Sep 08 14:29:55 and there is a very strong body of prior work... :) Sep 08 14:32:02 hehe Sep 08 15:00:01 Husam: I asked from oe the linux-libc-headers maintainer whether it is a bug that it does not copy asm/arch-omap headers but he sayd that no. According to him "userspace applications should not be depending on asm/ headers matching runtime nor should they expect linux/ headers to match the runtime kernel" Sep 08 15:00:35 Sorry, wrong newsgroup :-) Sep 08 15:00:42 hehe Sep 08 15:02:42 He's right, though Sep 08 15:03:09 Kergoth: I and Husam were really not sure with this earlier when we earlier had X crashes and therefore I started copying those omap-headers. Now I can stop that weisless hapit :-) Sep 08 15:03:19 * france is back (gone 16:18:56) Sep 08 15:03:33 ah. those headers shouldnt have anything to do with that Sep 08 15:06:26 Yeah, but it is just sometimes hard to find/understan real info how gcc/toolchain really works and then you make desperate things... Another thing we have wondered that should we force the gcc always to use 32 bit alignment... Anyway that's not todays problem as my image seems to work now ok. Sep 08 15:07:14 whats the nature of the crash? Sep 08 15:07:17 segmentation fault, or..? Sep 08 15:07:30 is it predictable, or random? Sep 08 15:07:32 reproducable, rather Sep 08 15:10:20 Earlier the crash was always predictable, no matter from the day when we make the build. We were always able to calibrate ts and enter name and password in the gpe's login screen. But after logging in the screen went black for a while and then login screen reappeared. (X crashed I think) Sep 08 15:11:33 We resolved the problem by adding command line boot parameter: "mem=64M at 0x10000000" and that helped for some reason. According to OMAP docs the start address is 0x1000000 Sep 08 15:14:47 nite all Sep 08 15:52:56 03florian 07org.oe.dev * r3d296e3e... 10/packages/gpe-login/ (files/busybox-bad-perms.patch gpe-login_0.82.bb): gpe-login: Use patch by Philippe De Swert to work around bugs in busybox related to handling of links. Sep 08 16:13:21 i must say new hw sure helps monotone ... Sep 08 16:13:34 under 30seconds a rev now :) Sep 08 16:16:25 yep Sep 08 16:19:09 $700 worth of hw tho :( Sep 08 16:22:47 * JustinP wishes he had $700 to spend on hardware...:-( Sep 08 16:24:25 not much of a choice on my part, the ide controller on my other mobo finally died permenantly Sep 08 16:51:41 i wish alsa had a decent manual ... Sep 08 16:55:05 Morning :) Sep 08 16:56:02 hey Sep 08 17:22:03 hum .... Sep 08 17:29:02 anyone get some idea on why /opt/oe/build/oetmp/staging/arm-linux/include/gconv.h staging header will affect irda build? Sep 08 17:38:20 so people whats up? Sep 08 17:42:59 bad :) Sep 08 17:43:19 my build fail on irda ... Sep 08 17:54:13 AvengerMoJo, with out more info its kinda hard to tell Sep 08 17:57:52 but my guess would be that irda might be calling it for charachter support Sep 08 18:42:28 03jbowler 07org.oe.dev * r4b93f0ae... 10/packages/uclibc/uclibc-0.9.28/thumb-resolve.patch: (log message trimmed) Sep 08 18:42:28 uclibc: fix STT_ARM_TFUNC handling in the uclibc ld.so Sep 08 18:42:28 STT_ARM_TFUNC handling was broken in uclibc because it is an ARM specific Sep 08 18:42:28 hack and uclibc code did not have support for it. Added the support in a Sep 08 18:42:28 way which I hope will be acceptable to uclibc and which should remain Sep 08 18:42:29 compatible with forthcoming binutils changes. ARM Thumb code is fully Sep 08 18:42:32 working with this commit with the possible exception of uclibc/gcc itself Sep 08 18:50:08 03jbowler 07org.oe.dev * r67f2ba05... 10/packages/uclibc/ (3 files in 3 dirs): Sep 08 18:50:08 uclibc_svn: fix the SVN version to build the uclibc SVN head Sep 08 18:50:08 The SVN version now copies 0.9.28 patches, the armeb case should work, Sep 08 18:50:08 other architectures almost certainly require the uclibc-cvs/*/uClibc.config Sep 08 18:50:08 copy of the config file to be deleted (if this is done the build will pick Sep 08 18:50:09 up the config file in uclibc-0.9.28). Sep 08 19:27:01 hmm i think my acpi is lying Sep 08 19:27:13 is there a way to adjust it? Sep 08 19:34:21 emte: right ..... I just wonder why gconv.h will fail the compile ... is my kernel hearder version mis-match? Sep 08 19:47:16 what version of openzaurus I should use in my local.conf Sep 08 19:47:57 the last release Sep 08 19:48:21 official release Sep 08 19:49:39 looks lik taht is 3.5.3 Sep 08 19:49:46 nice spelling Sep 08 19:51:10 hum Sep 08 19:51:18 RP: :) you here? Sep 08 19:54:03 you could always read the oz website as well ... Sep 08 19:59:22 03nail 07org.oe.dev * re55a199a... 10/conf/distro/openslug.conf: disapproval of revision ca142ce9ae4b3c4535d5ed6d23aed2559ef91c6b Sep 08 20:02:29 emte: don't really know where to look right now :0 Sep 08 20:02:32 hehe Sep 08 20:02:58 what are you looking for? Sep 08 20:03:03 I'm trying to build from 0 again Sep 08 20:03:12 the setup stuff for oz? Sep 08 20:03:23 emte: why is my error doesn't happen in others :) Sep 08 20:03:36 what error? Sep 08 20:04:27 compiling the irda-utils :) Sep 08 20:04:44 it fail ... but I just don't know why Sep 08 20:04:57 gcc problem or the linker or just the kernel hearder .. Sep 08 20:06:12 which oz version are you trying to build? Sep 08 20:07:25 emte: I just user openzaurus Sep 08 20:07:41 emte: now I'm trying openzaurus 3.5.4 Sep 08 20:08:46 i wouldnt count on 3.5.4 to work Sep 08 20:08:55 there is a reason it isnt a release yet Sep 08 20:10:00 emte: :) then I'm wondering how others get that image build :) Sep 08 20:10:32 have they? Sep 08 20:10:41 full images? Sep 08 20:14:52 emte: that's why I'm asking RP :) hehe Sep 08 20:36:02 hmm ... someone was playing with pcmcia-cs ? Sep 08 20:37:06 | lspnp.c:499: error: invalid lvalue in assignment Sep 08 20:37:10 nice pretty error Sep 08 21:18:48 hum :o Sep 08 23:06:25 l8r All ;) Sep 08 23:46:37 hi all! Sep 08 23:47:02 hi Sep 09 00:12:25 I get the same error :) Sep 09 00:12:55 any one get time to take a look Sep 09 00:14:15 I get the log here :) Sep 09 00:32:55 hi ! Sep 09 00:51:49 03koen 07org.oe.dev * r74c80fe1... 10/packages/avahi/avahi_0.4.bb: packages/avahi/avahi_0.4.bb: add 0.4, disable gtk and Qt Sep 09 00:54:55 morning Sep 09 00:55:00 hey XorA Sep 09 00:55:16 koen|sleep: monotone while sleeping is not advised Sep 09 00:55:27 :) Sep 09 01:00:16 hi koen ! Sep 09 01:00:34 hey alan|home Sep 09 01:01:09 morning all Sep 09 01:01:19 hey RP Sep 09 01:02:40 morning Sep 09 01:02:46 hi hrw|work Sep 09 01:03:06 hey hrw|work Sep 09 01:03:20 hrw|work: OMG Ive just looked at what the marketing guys want our new website to be, Im doubting the guy who wrote it actually knew html its so bad Sep 09 01:03:31 XorA: url? Sep 09 01:04:09 hrw|work: dont die laughing Sep 09 01:05:15 XorA: another dreamweaver user.. Sep 09 01:08:42 hrw|work: abuser Sep 09 01:12:09 morning all Sep 09 01:12:21 sick bastards: http://ipodlinux.org/Video_Player Sep 09 01:12:23 hey ade|desk Sep 09 01:17:14 hi ade|desk Sep 09 01:19:53 i've just finished my new OE installation but i must have done something wrong. i have this error : "[alan@192 build]$ bitbake nano Sep 09 01:19:54 Traceback (most recent call last): Sep 09 01:19:54 File "/usr/local/bin/bitbake", line 26, in ? Sep 09 01:19:54 import bb Sep 09 01:19:54 ImportError: No module named bb" Any idea ? Sep 09 01:22:49 ah, you installed bitbake, but didn't adjust PYTHONPATH Sep 09 01:23:01 * koen never installs bitbake Sep 09 01:27:09 koen: ok. Why is installing bitbake not the best method ? Sep 09 01:27:31 problems like yours Sep 09 01:28:42 hehe :) Sep 09 01:29:12 thanks koen. i'll run the other way. Sep 09 01:33:33 it works just fine now. thanks again koen :) Sep 09 01:44:12 RP: :) Sep 09 01:44:12 RP: I couldn't seem to pass the irda compile.... Sep 09 01:44:12 RP: can you look at my log and show me some direction/ Sep 09 01:47:09 Hi. How do I overcome those "Link tests are not allowed after GCC_NO_EXECUTABLES" configure errors when compiling gcc-cross-4.0.0 (Akita, OZ 3.5.4)? Sep 09 01:47:53 raphman: not using gcc 4 ;) Sep 09 01:47:57 raphman: or googling Sep 09 01:48:24 * koen is very happy with familiar's locked down gcc Sep 09 01:48:43 koen: that probably was a smart decision Sep 09 01:48:44 and it's iconv defines Sep 09 01:49:02 raphman: gcc 4.0 is most probably untested on every OE distribution Sep 09 01:49:27 didn't choose it explicitly . Is ist selected automatically for OZ 3.5.4? Sep 09 01:49:45 raphman: try setting a preferred provider to some other gcc Sep 09 01:49:51 heh Sep 09 01:50:06 raphman: and just out of interest please pastebin the error message and config.log Sep 09 01:50:07 looks like the default_preference in the .bbs doesn't work well Sep 09 01:50:11 ~pastebin Sep 09 01:50:12 from memory, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca/ Sep 09 01:52:26 koen ... I can't pass irda ... anyone else get the same problem? Sep 09 01:53:06 AvengerMoJo: what is pass irda? Sep 09 01:53:41 AvengerMoJo: Using DISTRO=openzaurus wouldn't work. You need openzaurus-3.5.4 Sep 09 01:54:29 I hate to say it but if you were using DISTRO=openzaurus, you probably need to start again :-( Sep 09 01:54:52 Keep your sources directory though! Sep 09 01:55:35 RP: I clean the temp and build it with 3.5.4 already ... seem to be same problem Sep 09 01:55:55 DISTRO = "openzaurus-3.5.4" Sep 09 01:56:05 KERNEL_VERSION = "2.6" Sep 09 01:56:51 Ok, what was the error - pastebin the log with the error in Sep 09 01:57:20 it will be somewhere like build/tmp/work/irda*/temp Sep 09 01:57:31 * koen pokes at the OZ config files Sep 09 01:58:32 03koen 07org.oe.dev * rd657642f... 10/conf/distro/openzaurus-3.5.4.conf: conf/distro/openzaurus-3.5.4.conf: select glibc as provider for iconv and intl Sep 09 02:02:55 ... Sep 09 02:02:55 so PREFERRED_VERSION_binutils-cross = "2.15.94.0.1" <-- this is my problem? Sep 09 02:02:56 RP: how to send you the log? Sep 09 02:03:22 zecke: http://pastebin.ca/22472 Sep 09 02:04:41 zecke: sorry, How do I find out which virtual class provides the compiler and which packages are available. Sep 09 02:05:09 raphman: you could use the shell Sep 09 02:05:33 raphman: or spy at the familiar-0.8.3.conf and use the content to set the preferred provider Sep 09 02:05:56 raphman: and your config.log please Sep 09 02:09:44 sorry .. what did I miss Sep 09 02:09:44 :) Sep 09 02:09:50 RP: where to send log? Sep 09 02:11:00 :) Sep 09 02:12:51 zecke: sorry: http://pastebin.ca/22473 Sep 09 02:14:22 AvengerMoJo: pastebin Sep 09 02:14:25 ~pastebin Sep 09 02:14:27 rumour has it, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca/ Sep 09 02:15:00 We've just had a report that 2.6 on Borzoi works :) Sep 09 02:16:34 RP: cool Sep 09 02:18:11 AvengerMoJo: i've been saied you needed to modprobe an irda module. pxa_ircp, or something like that... hrw|work : i need your memory ^^ Sep 09 02:19:24 alan|home: pxa_irfcp IIRC Sep 09 02:19:39 alan|home: but when /proc/interrupts shows STUART you already lost Sep 09 02:21:35 arf... does it mean irda file copying doesn't work yet ? Sep 09 02:22:25 alan|home: does it affect the compile? Sep 09 02:22:49 alan|home: look into packages/linux/linux-openzaurus_*.bb Sep 09 02:22:52 alan|home: works fine in 2.6.13-rc5-mm1 Sep 09 02:23:07 alan|home: havent tested 2.6.13-mm1 yet Sep 09 02:23:17 RP: here you go http://pastebin.ca/22474 Sep 09 02:26:24 hum :( Sep 09 02:40:31 AvengerMoJo: I don't know offhand... Maybe someone else here recognises the problem? Sep 09 02:42:12 RP: tried -mm2? Sep 09 02:44:17 hrw|work: Not yet. Its not going to happen either as the patch series are too much of a mess Sep 09 02:44:29 I'm trying to revise patches in -mm Sep 09 02:44:51 ok - thx for info Sep 09 02:46:01 hrw|work: I have a big list of things I need to change and am having nightmares atm. -mm2 would finish me off ;-) Sep 09 02:46:30 RP: ok - I prefer to have you alive Sep 09 02:58:59 :( Sep 09 02:59:08 I'm not sure where to start .. :) Sep 09 02:59:44 gicon.h problem ... in staging ... does it come from kernel-header or glibc? **** ENDING LOGGING AT Fri Sep 09 02:59:56 2005