**** BEGIN LOGGING AT Sun Sep 09 02:59:56 2007 Sep 09 04:07:15 mwester: ping Sep 09 06:59:37 03ynezz 07org.oe.dev * re4f5996a... 10/ (5 files in 3 dirs): gcc 4.2: fix uclibc compilation, closes #2959 Sep 09 08:46:19 gm Sep 09 08:51:17 03koen 07org.oe.dev * r15710183... 10/ (1 packages/linux/linux-ezx-2.6.21/a780/defconfig): linux-ezx: enable swap, loo, nfsrootp and wext in a780 defconfig Sep 09 08:51:23 03koen 07org.oe.dev * r2fcb4337... 10/ (1 packages/linux/linux-ezx_2.6.21.bb): linux-ezx: bump PR Sep 09 09:16:58 hya! anyone working with an i.mx31 board? Sep 09 09:18:58 koobla|home: lrg has one at work Sep 09 09:19:17 ah, nice - thx, koen Sep 09 09:19:32 I should add support for some parts of the atlas chip Sep 09 09:21:23 hah! Sep 09 09:21:35 db_rewrite_sql() is the evildoer Sep 09 09:22:12 that inserts a ORDER BY n.nid every where you use ORDER BY n.sticky Sep 09 09:22:27 in other words: the news on the OE site isn't reversed anymore Sep 09 12:00:20 03mickeyl 07org.oe.dev * r0cf7c9f4... 10/ (8 files in 6 dirs): linux-ezx: e6 -> rokre6, e2 -> rokre2 Sep 09 12:00:27 03mickeyl 07org.oe.dev * r3025e5e7... 10/ (1 packages/linux/linux.inc): linux.inc: prepare adding fic-gta01/fic-gta02 support Sep 09 12:01:35 03rwhitby 07org.oe.dev * r1cc47f29... 10/ (4 files in 3 dirs): gsm: Fix default file and init script. GSM_RES was missing, and the tests for null values needed quotes around the value to prevent syntax errors. Sep 09 12:02:01 mickey|away: who is looking after gsmd in OE? Sep 09 12:02:35 rwhitby: mostly pH5 Sep 09 12:03:29 * rwhitby refrains from making any further comments Sep 09 12:05:28 I successfully built and tired openmoko-devel-image for akita. It's working perfectly Sep 09 12:09:53 guys, i can't seem to find SRCREV_${PN} working. anyone used that yet? Sep 09 12:09:59 i.e. Sep 09 12:10:03 SRCREV_oh-puzzles = "20" Sep 09 12:10:16 SRCREV_pn-${PN} is the format AFAIK Sep 09 12:10:28 ooh, interesting Sep 09 12:10:30 missed that Sep 09 12:10:34 thanks, let me try Sep 09 12:10:36 mickey|away: see conf/distro/include/sane-srcrevs.inc Sep 09 12:11:35 koen: are there any grave issues with angstrom-2008.1 currently? Sep 09 12:12:03 heh, indeed. that's it Sep 09 12:12:04 thanks koen Sep 09 12:12:12 koen++ Sep 09 12:12:18 ?karma koen Sep 09 12:12:20 pH5: haven't tested it after the xorg upgrade Sep 09 12:12:52 ok Sep 09 12:12:54 * pH5 switches Sep 09 12:13:36 pH5: I have yet to see any issus with glibc 2.6 and gcc 4.2.1, so things are looking good Sep 09 12:14:58 mickey|away: http://kryogenix.org/code/Jackfield%20and%20Python/pix/guido_3000.jpg Sep 09 12:15:18 ;) Sep 09 12:15:50 http://kryogenix.org/code/Jackfield%20and%20Python/pix/cows.jpg haha Sep 09 12:16:40 zecke: the OE website is fully functional again Sep 09 12:17:08 koen++ and we need a bot that understands that Sep 09 12:17:35 koen: I'm going to buy you a mattress, cover and pillow! Sep 09 12:18:07 mickey|away: zecke and RP were talking about having bitbake output all versions used so distros can lock things down completely Sep 09 12:22:26 who is that guido3000 person? Sep 09 12:22:46 dcordes: guido is the father oh python Sep 09 12:22:51 and python3000 is the new version Sep 09 12:23:02 s/oh/of/ Sep 09 12:23:06 the picture is kind of funny Sep 09 12:23:33 python is funny! Sep 09 12:30:38 :) Sep 09 12:30:41 koen: looks like a good feature to waste time on Sep 09 12:36:32 mickey|away: scrolling is really smooth on the a780, I can't wait to have SRAM support to make it even smoother :) Sep 09 12:36:45 yep Sep 09 12:37:26 I think feedreader2 is the first OSS 'browser' to have kinetic scrolling :) Sep 09 12:37:54 hehe Sep 09 12:41:29 03zecke123 * r974 10bitbake/MANIFEST: Fix the MANIFEST to include the two new files Sep 09 12:42:42 mickey|away: can you apply a patch to OM SVN for me please? Sep 09 12:43:00 mickey|away: http://bugzilla.openmoko.org/cgi-bin/bugzilla/attachment.cgi?id=281&action=view Sep 09 12:44:22 whoever wrote that code that this patch fixes *cannot* have tested it *at* *all* before committing. Sep 09 12:45:07 testing, what's that? Sep 09 12:45:27 "Testing is for users" - Linux Torvalds Sep 09 12:45:32 Linus even Sep 09 12:46:03 * rwhitby lies in wait for mickey|away to show himself as not idle ... Sep 09 12:46:21 hehe.. yeah.. we need some kind of automatic unit testing. Sep 09 12:46:44 no you need 'basic testing' first Sep 09 12:47:05 koen that also. Sep 09 12:47:07 rwhitby: i don't want to work at all today. i just did some small things ;) i'll commit your stuff anyways. Sep 09 12:47:48 mickey|away: many thanks - it saves me doing it in OE and then having it break tomorrow when the real patch is applied. Sep 09 12:49:27 mickey|away: I can't take credit for the patch - cesarb reported it yesterday and attached the patch Sep 09 12:49:39 * koen hints at ;maxrev=foo in SRC_URI Sep 09 12:50:02 koen: ah, yes - I'd forgotten about maxrev. Sep 09 12:53:48 mickey|away: actually, the patch allows dialing when netreg is denied (for emergencies) and also fixes the problem which didn't allow dialling normally. Sep 09 12:55:01 mickey|away: thx Sep 09 13:04:23 zecke are you there? Sep 09 13:04:32 yes Sep 09 13:04:48 want to help to distract my studies? ;) Sep 09 13:04:54 wonderful.. does webkit-gtk build for you atm? Sep 09 13:05:20 it fails on linking GdkLauncher here. rest is fine Sep 09 13:05:31 roh: http://build.webkit.org/, shows it fails Sep 09 13:06:05 oha Sep 09 13:06:09 interresting Sep 09 13:06:29 WebCore::callOnMainThread(void (*)())? Sep 09 13:06:49 apple guys seem to work on a multithreaded rendering engine... Sep 09 13:08:23 nice Sep 09 13:08:40 * koen should try a safari nightly build Sep 09 13:09:21 03mickeyl 07org.oe.dev * r6129c0fa... 10/ (1 packages/e17/exhibit_cvs.bb): exhibit cvs: add dependency on epsilon Sep 09 13:09:26 koen: http://build.webkit.org/changes/2863 so far only icon lookup :} Sep 09 13:09:27 03mickeyl 07org.oe.dev * ra9378a67... 10/ (4 files in 3 dirs): Sep 09 13:09:27 openmoko distro configurations: Sep 09 13:09:27 - remove openmoko.inc which is no longer in use Sep 09 13:09:27 - include preferred-openmoko-versions.inc Sep 09 13:09:27 - set preferred SRCREV for linux-openmoko Sep 09 13:09:32 03mickeyl 07org.oe.dev * rc762de4b... 10/ (5 files in 2 dirs): linux-openmoko: indivate SRCREV in kernel binary version, add dedicated recipe for development kernel (floating SRCREV) Sep 09 13:09:38 03koen 07org.oe.dev * re5f88ce4... 10/ (1 conf/machine/rokre6.conf): rokr6: add machine stub Sep 09 13:09:40 koen: JavaScriptCore should be thread safe already Sep 09 13:09:44 03koen 07org.oe.dev * r541a0c31... 10/ (1 packages/linux/linux-ezx_2.6.21.bb): linux-ezx: expand COMPATIBLE_MACHINE to rokre2 and rokre6 Sep 09 13:11:47 mickey|away: with the gsmd restart in /etc/matchbox/session, I can now make and recieve calls on the neo again. Exactly where I was a week ago ;-( Sep 09 13:12:13 (oh, and the removal of console=ttySAC0 from uboot bootargs_base variable Sep 09 13:16:44 roh: thanks for telling me, I will add a stub for now Sep 09 13:18:29 zecke thanks. i think we should dup the webkit-gtk_svn.bb and pull it to a working rev. which you can bump everytime we verified it works with a _svn_head version Sep 09 13:19:10 sure Sep 09 13:19:24 I think OM should do that with *every* _svn bb Sep 09 13:19:40 (OM, not OE) Sep 09 13:20:00 * koen hints at sane-srcrevs.inc Sep 09 13:20:17 * rwhitby passes the hint to roh and mickey|away Sep 09 13:20:33 rwhitby yes and no. i think its important to keep it building. but not only on a static snapshot.. then we would not recieve enough testing. Sep 09 13:21:08 roh: on the contrary, you would get very good testing of the single version which is built, and you would be able to choose what gets tested next. Sep 09 13:21:23 roh: I'm trying to make bitmake emit all the information to create a sane-srcrevs.inc, then you can easily do snapshots... Sep 09 13:21:36 at the moment, you're not testing anything well, cause *everything* is moving and people are putting their neo's back in the case until it settles down. Sep 09 13:21:43 rwhitby i think we should fix all external repo dependant stuff. meaning everything in our svn should be building anyways. if it doesnt its our responibility to fix it. but we cannot and should not depend on others to drive the same policy Sep 09 13:22:08 roh: right - I'm only talking about non-OM _svn external dependencies. Sep 09 13:22:36 although I think that the people committig to SVN in OM should be taught to bump the PR in OE after they have tested something. Sep 09 13:22:37 rwhitby must not only be svn.. we also pull git for uboot which breaks from time to time... Sep 09 13:22:46 or cvs is in there also.. i think Sep 09 13:22:47 right - any moving target. Sep 09 13:23:17 (bump the PR and change the REV to show the new "unit tested" version) Sep 09 13:23:29 (for OM source code) Sep 09 13:24:02 that would encourage some actual responsibility for not breaking the build. if you changed the _svn REV, then you're saying you tested it. Sep 09 13:24:38 if you haven't tested it, then you *shouldn't* be releasing it onto the masses Sep 09 13:24:43 alternatively do more releases! Sep 09 13:25:18 (where testing is something simple like booting and making a call, not exhaustive or formal testing) Sep 09 13:25:25 * rwhitby goes to bed Sep 09 13:25:37 n8 rwhitby Sep 09 13:27:22 rwhitby: that's why I was arguing to put the kernel patches in OE, since svn breaks on everyday that has ends in 'ay' Sep 09 13:27:47 (but now we have the -dev kernel and a pinned SRCREV, so that's a decent compromize) Sep 09 13:29:13 what is a sane way to add a script to /etc/apmd/suspend.d/ ? Sep 09 13:29:58 *near Sep 09 13:31:34 roh: it should compile now again Sep 09 13:33:44 *rebuild* Sep 09 13:37:44 good morning! Sep 09 13:39:46 hey sakoman Sep 09 13:43:43 been scanning boot messages to make sure nothing funky is happening and discovered some date command strangeness Sep 09 13:44:36 when save-rtc writes the time stamp it ends up as: %2m%2d%2H%2M00%4Y Sep 09 13:45:47 So it seems my busybox date doesn't like the format string very much! Sep 09 13:47:26 could you file a bug on that so RP and mickeyl can look at it when they get back to OE work? Sep 09 13:48:12 all is well if I use %m%d%H%M%Y Sep 09 13:48:20 sure, no problem Sep 09 13:49:14 sakoman: what's your opinion on softfocus filters on TLRs? Sep 09 14:00:34 03koen 07org.oe.dev * r47331541... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007: include preferred-openmoko-versions Sep 09 14:06:20 hello has anyone had any success compliling kernel for a palm zire 31 ? Sep 09 14:25:59 morning Sep 09 14:27:54 hey chouimat|ibook Sep 09 14:57:58 i am having trouble building qemu. the build dies in ops_sse.h when compiling op.c Sep 09 14:58:20 can anyone help? Sep 09 15:01:02 please pastebin the exact error message Sep 09 15:01:05 ~pastebin Sep 09 15:01:05 hmm... pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Sep 09 15:06:51 koen heres me error http://dpaste.com/19057/ Sep 09 15:07:50 koen: never used a commercial soft focus filter Sep 09 15:08:03 rich__: no idea on that, sorry Sep 09 15:08:09 sakoman: stockings? Sep 09 15:08:28 ok thanks anyway Sep 09 15:08:38 no, but I have tried petroleum jelly on clear filter Sep 09 15:09:00 did that give the wanted effect? Sep 09 15:09:10 smear around outer margins for small effect, increase width for more effect Sep 09 15:09:19 yes, it's kind of cool Sep 09 15:09:38 Can't see it in action ahead of time on a tlr though :-) Sep 09 15:09:58 so it will be a surprise after you develop Sep 09 15:10:05 I just won an ebay action for rollei softfocus filters so I can put one on the viewing lens as well :) Sep 09 15:10:12 works great for slr or view camera Sep 09 15:10:22 Ah, perfect Sep 09 15:10:26 yes Sep 09 15:10:37 Curious how they will work for you Sep 09 15:10:41 let me know Sep 09 15:10:51 and a rolleiparkeil for correcting parallax when using closeup filters Sep 09 15:11:04 (germans and product names.....) Sep 09 15:11:14 I sense the beginning of an obsession! Sep 09 15:11:22 :D Sep 09 15:11:44 (if it only would stop raining in my free time) Sep 09 15:12:26 I thought you spent all your free time here :-) Sep 09 15:13:06 Hasn't rained here in at least 4 months! Sep 09 15:21:37 you're in SoCal? Sep 09 15:23:44 koen: far northern CA, near Mt Shasta & Mt Lassen Sep 09 15:24:11 ah right, cause it *never* rains in socal ;) Sep 09 15:24:31 so the song says! Sep 09 15:24:48 I used to work in a supermarket that had that song on their muzak cd Sep 09 15:25:01 so 8 times each saterday... Sep 09 15:25:15 arghh! Sep 09 15:25:39 hehe! Reality is so different -- but it doesn't sound as nice in a song to speak of mudslides or wildfires ;-) Sep 09 15:26:15 just checked my buildroot image and busybox date doesn't like numerics in the format there either Sep 09 15:26:42 busybox is something the cat would drag home Sep 09 15:26:54 perhaps it is a libc/uclibc thing Sep 09 15:26:56 it's funny for someone, but you end up with the mess Sep 09 15:27:16 oh! Sep 09 15:27:25 I remember something int he busybox defconfig about that Sep 09 15:27:46 koen: why did alberthijn fire you? Sep 09 15:28:03 zecke: I quit because I moved :) Sep 09 15:28:23 koen: My cousin said he quit as well... Sep 09 15:28:39 which flavour of vla should I bring? Sep 09 15:28:52 classic (vanilla) Sep 09 15:29:08 * koen makes a note of that Sep 09 15:29:10 alternatively bring some cassis sirup Sep 09 15:29:26 for on the vla or drink? Sep 09 15:29:34 drink Sep 09 15:29:58 koen: is this it? # CONFIG_FEATURE_DATE_ISOFMT is not set Sep 09 15:30:51 03koen 07org.oe.dev * r63dfdf9c... 10/ (5 files in 5 dirs): linux-ezx: update a1200, e680, rokre2 and rokre6 defconfigs to match a780 defconfig. Device owners: please review Sep 09 15:30:53 sakoman: that looks like the right one Sep 09 15:32:15 worth a try, though the universal fix just might be to change the save-rtc format to +%m%d%H%M%Y since that will do the right thing whether the option is selected or not Sep 09 15:32:33 that would work as well Sep 09 15:32:36 (I think) Sep 09 15:32:55 If the docs for date are correct it will :-) Sep 09 15:33:14 that would IMO be the preferred solution Sep 09 15:33:20 agreed Sep 09 15:33:46 I'll file the bug and let the experts decide :-) Sep 09 15:34:15 In the meantime I'll just override set-rtc.sh Sep 09 15:35:05 format change works on my settup -- don't know about a libc build though Sep 09 15:35:53 is there any kernel in OE yet that is usable with htcwizard/other omap850 htcs? Sep 09 15:36:48 dcordes: maybe the htc tornado one Sep 09 15:37:20 I built one angstrom image with the htctornado machtype Sep 09 15:37:29 but that oopses on my htcprophet Sep 09 15:37:38 prophet is also omap850 device Sep 09 15:39:48 I succesfully tried a kernel made for wizard on the prophet but that was obsolete and discontinued. I was just curious if there's something omap850ish except of the htctornado machtype Sep 09 15:41:46 pH5: did you read prophet related stuff by chance? Sep 09 15:44:45 dcordes: no Sep 09 15:52:42 pH5: with your recommendations I managed to start the boot process with the htctornado kernel on prophet Sep 09 15:52:58 unfortunately it oopses Sep 09 16:33:28 koen: let me introduce you :) Sep 09 16:37:38 EXCITING! Sep 09 16:40:40 hi i am trying to create an fxload package but it fails with a missing linux/usb/ch9.h missing (its in linux-libc-dev on debian). What do i have to say in my bb file to get it added? Thanks Sep 09 16:40:58 bye Sep 09 16:44:26 tstone: linux-libc-headers Sep 09 16:45:21 pH5: so i just put a DEPENDS = "linux-libc-headers" into the bb file? (sorry this is my first bitbake file) Sep 09 16:46:34 tstone: yes, if fxload directly references a linux header file, this is the correct way. Sep 09 16:46:54 zecke: MORE EXCITING Sep 09 16:47:00 although I think it should have been staged anyway by a previous glibc build? Sep 09 16:47:06 koen: :) Sep 09 16:47:45 netscape 2.0 gold Sep 09 16:47:49 those were the times Sep 09 16:49:40 pH5: it still doesn't find ch9.h. :-( Any ideas? Sep 09 16:50:15 CFLAGS += " -I${STAGING_INCDIR} " ? Sep 09 16:51:47 koen: nope this doesn't help Sep 09 16:52:29 tstone: then you should get more specific. pastebin the recipe and compile log, for example. Sep 09 16:53:31 pH5: there is no usb directory in build/angstrom/staging/armeb-angstrom-linux-gnueabi/include/linux which seems to be the directory added with the CFLAGS statement Sep 09 16:55:12 tstone: and linux-libc-headers is built and staged? because I have that header file in there. Sep 09 16:56:36 pH5: yep, fxload_0.0.20020411-r1.bb http://pastebin.com/m1ca77cef Sep 09 16:57:19 the compile log error is "ezusb.c:32:29: error: linux/usb/ch9.h: No such file or directory" Sep 09 16:57:58 i patched ezusb.c to include ch9.h instead of linux/usb.h which is what the debiananistas also did... Sep 09 16:59:27 i have an ch9.h header in ./build/angstrom/staging/ixp4xxbe-angstrom-linux-gnueabi/kernel/include/linux/usb/ch9.h Sep 09 17:06:56 morning all Sep 09 17:18:23 hey RP Sep 09 17:21:23 hi all. I have a package (squeak smalltalk) I've managed to get building within OE as part of OpenMoko. How should I start going about getting the code into the distribution? Sep 09 17:21:29 hi RP Sep 09 17:21:50 tonyg: attach the mtn diff to the oe bugtracker Sep 09 17:22:47 koen, thanks. I'll potter around the website Sep 09 17:23:25 it works with "CFLAGS += " -I${STAGING_INCDIR}/../../ixp4xxbe-angstrom-linux-gnueabi/kernel/include/ "" Sep 09 17:23:44 tstone: but the kernel/ include is not the correct one, linux-libc-headers should stage the cleaned up headers into $STAGING_INCDIR Sep 09 17:24:15 it's a bug if it doesn't to that for ixp4xxbe Sep 09 17:29:18 pH5: what should i do next? Sep 09 17:29:30 may this be a problem with my local.conf? Sep 09 17:37:24 tstone: or bitbake -crebuild linux-libc-headers and check the logs (does ch9.h appear in log.do_stage, are there errors?) Sep 09 18:08:44 pH5: it builds without errors, in the log.do_stage there is a row "UNIFDEF include/linux/usb_ch9.h" Sep 09 18:09:38 aha! olde kernel headers version Sep 09 18:09:53 linux/usb_ch9.h was renamed to linux/usb/ch9.h in recent kernels Sep 09 18:10:05 try using the 2.6.20 kernel headers (or newer) Sep 09 18:14:24 quote of the day "to err is human ... but to blame someone else shows management potential" Sep 09 18:33:55 pH5: sorry for the delay, that is linux-libc-headers-2.6.20-r7 Sep 09 18:35:04 really? oh well, if the change was that recent, you can either upgrade to 2.6.22 or patch fxload to use the old header path. Sep 09 18:40:55 pH5: thanks for your help, i go with patching to usb_ch9.h, only one dumb newbie question: where is the created package file? Is there any? Sep 09 18:42:28 ok, i found the message that these files are installed but not shipped in any package. Sep 09 18:44:11 it seems as if ${sbindir} is not sbin. Sep 09 18:44:55 whoho: fxload_0.0.20020411-r1_armv5teb.ipk Sep 09 18:46:04 anybody interested integrating this with the oe distibution? Sep 09 18:49:06 attach the output of mtn diff to the bugtracker :) Sep 09 18:50:48 koen: i used my own package directory so the output of mtn diff will be pretty boring ;-) Sep 09 18:51:03 i suppose i should copy the stuff to the packages directory in advance? Sep 09 18:59:44 tstone: yup, copy to packages/fxopen and then mtn add and mtn diff the .bb recipe. Sep 09 19:00:35 s/load/open/ Sep 09 19:01:17 hehe, and that the other way around. Sep 09 19:01:25 * pH5 decides brain needs a break Sep 09 19:18:43 hi all Sep 09 19:36:05 http://bugs.openembedded.org/show_bug.cgi?id=2965 Sep 09 19:36:28 thanks again pH5 Sep 09 19:40:22 tstone: np Sep 09 19:44:35 tstone: can you try if this works for you, too: http://en.pastebin.ca/689014 Sep 09 19:46:19 I'm not sure for what linux version the usb headers were shifted, but we only have 2.6.20 and 2.6.22 in oe anyway. Sep 09 19:52:24 tstone: also, I propose to change the recipe (with the help of oe-stylize.py) like this: http://en.pastebin.ca/689029 Sep 09 19:52:26 shall I commit? Sep 09 19:52:48 pH5 Sep 09 19:53:20 pH5: your diff didn't go through (or it has lost formatting on pastebin) i am just creating a new diff, just a moment Sep 09 19:53:53 you can get the raw like this: http://en.pastebin.ca/raw/689029 Sep 09 19:54:22 http://en.pastebin.ca/raw/689014, sorry **** ENDING LOGGING AT Sun Sep 09 19:54:28 2007 **** BEGIN LOGGING AT Sun Sep 09 19:56:12 2007 Sep 09 19:57:52 tstone: can I see the compile log? I could build this with both linux-libc-headers 2.6.20 and 2.6.22 (for armv5te and armv4t) Sep 09 19:59:06 http://en.pastebin.ca/689039 Sep 09 20:00:02 tstone: you're still using the recipe that has the CFLAGS include ./../ixp4xxbe-angstrom-linux-gnueabi/kernel Sep 09 20:00:42 try this: http://en.pastebin.ca/raw/689041 Sep 09 20:01:20 pH5: i am just trieng my own patch Sep 09 20:02:53 dumb question: how to i apply oe_stylize.py Sep 09 20:03:08 my patch works btw Sep 09 20:04:22 just call it with the .bb recipe as a parameter and it'll spit out a version that it thinks fits the oe style best on stdout Sep 09 20:04:55 Good afternoon all Sep 09 20:05:42 remind me. What's the bb recipe flag to indicate a recipe should not be built regardless of version data Sep 09 20:05:44 ? Sep 09 20:06:00 hillct: you mean DEFAULT_PREFERENCE? Sep 09 20:07:01 pH5: perhaps Sep 09 20:07:06 I'll check it out Sep 09 20:07:16 pH5: my two version are: spandsp_0.0.4+20070512.bb spandsp_0.0.4+20070908.bb Sep 09 20:07:33 shouldn't it be pulling the newer one by default? Sep 09 20:08:08 or is 20070908 alphebetically lesser than the other? Sep 09 20:08:35 hillct: yes, unless the older one has a higher DEFAULT_PREFERENCE value, or somewhere a PREFERRED_VERSION_spandsp = "0.0.4+20070512" is set. Sep 09 20:08:48 pH5: no more maintainer field? Sep 09 20:08:49 alphabetically, 20070908 > 20070512 Sep 09 20:09:31 checking Sep 09 20:09:36 tstone: right. there is a MAINTAINERS file in the tree root and author info is in the mtn commits after all. Sep 09 20:11:04 no default preference or preferred version is set as far as i can tell Sep 09 20:11:20 pH5: there is a new attachemnt on http://bugs.openembedded.org/show_bug.cgi?id=2965 Sep 09 20:11:25 with your suggestions Sep 09 20:12:27 tstone: you really need the CFLAGS? Sep 09 20:12:36 dunno Sep 09 20:12:45 will check Sep 09 20:13:49 no it is not needed Sep 09 20:15:22 ok, I'll remove the CFLAGS, s/sbin/${base_sbindir}/ and commit Sep 09 20:16:01 would bb ignore an invalid recipe file and fall back to a previous versinon without any sort of error or warning? Sep 09 20:16:49 pH5: wait you have to patch the maintainers file so i can brag inf front of my collegues :-) Sep 09 20:18:00 tstone: initially the maintainers file was not supposed to support bragging but to help users find a person that can help them when they have problems with a certain recipe Sep 09 20:18:40 ;) <-- better not forget that one Sep 09 20:23:24 aparently, values after a + sign in the filename aren't used to calvulate version, so the only way to indicate a neweer version in the case of the filenames i mentioned was to increment the revision Sep 09 20:23:26 pH5: i just created another attachement for bug 2965 Sep 09 20:24:15 not sure if the first part is correct, but incrementing the revision did the trick Sep 09 20:25:12 tstone: with that oe@iss... email address mtn commits will show up as authored by "oe" Sep 09 20:26:00 pH5: perhaps i should change it into ohh ;-) Sep 09 20:26:33 tstone: that's just fyi :) Sep 09 20:27:46 along te same lines, when a package name changes, and (in this case) the old package was at revision 3, do I need to start the new package at revision 4 in order to get REPLACES to operate properly? or can I start the new package at revision 0? Sep 09 20:28:26 hillct: there is no REPLACES, only RREPLACES Sep 09 20:28:33 yeah Sep 09 20:28:42 and no, you don't need to keep the revision number for that. Sep 09 20:28:51 typo on my part. thanks Sep 09 20:29:01 you can even replace foo_1.5-r3 with bar_0.1-r2 Sep 09 20:29:51 so RREPLACES takes effect for all revisions of foo Sep 09 20:31:19 hillct: it's just a filed in the package control data to tell ipkg that any foo is obsoleted by any bar. Sep 09 20:31:53 yeah. I just wanted to be sure it wasn't rev-specific. thanks Sep 09 20:33:48 hillct: np Sep 09 20:36:01 03oe 07org.oe.dev * r7b0ea327... 10/ (5 files in 4 dirs): fxload: a firmware uploader for ezusb chips Sep 09 20:36:06 03pH5 07org.oe.dev * r7d51500e... 10/ (1 MAINTAINERS): MAINTAINERS: add Tim Sander so he can brag in front of his colleagues Sep 09 20:38:41 :D Sep 09 20:40:46 pH5: hey, that makes bragging more difficult 8-) Sep 09 20:44:23 tstone: :D Sep 09 20:44:27 good night everyone Sep 09 20:44:33 good night Sep 09 20:48:05 is there a preferred method for implemention optional build steps? Sep 09 21:11:48 i have another question: i want to add the usb_snd_audio_x2y driver for a kernel build what else do i have to do besides editiing defconfig? Sep 09 21:42:31 03hrw 07org.oe.dev * r6cbf0719... 10/ (10 files in 2 dirs): Sep 09 21:42:31 linux: changed way of adding uboot-mkimage dependency so #2926 should be fixed Sep 09 21:42:31 Something is wrong wen it comes to overrides and adding into variable - bug Sep 09 21:42:31 #2966 opened. Sep 09 21:51:51 koen: isnt 2966 of your bug you found in gtk+? Sep 09 21:51:59 a varient of Sep 09 22:00:08 Is there a way I can cause a recipe to save a tarball, and use it in preference to performing a certain stap in the recipe, much like bitbake retrieves local tarball cache versions of svn checkouts when possible? Sep 09 22:17:05 03rwhitby 07org.oe.dev * r47ab531b... 10/ (1 packages/tasks/task-openmoko-feed.bb): task-openmoko-feed: Added ipkg-utils, mplayer, synergy Sep 10 00:02:17 guys, floating recipe is one that has several versions of a package available? isn it? Sep 10 00:09:29 hello **** ENDING LOGGING AT Mon Sep 10 02:59:57 2007