**** BEGIN LOGGING AT Fri Aug 12 02:59:59 2011 Aug 12 05:00:52 khem: around? Aug 12 06:57:15 gm Aug 12 06:57:33 did someone saw ericben latetly? Aug 12 06:58:39 In the depency explorer I see that a certain package only depends upon itself. What does that mean? What can I do to ensure that that package is not included? Aug 12 06:59:14 what excatly you mean Aug 12 06:59:19 give an example please Aug 12 07:00:36 if I open the dependency explorer for an image I made, I see that the package mesa only reverse depends upon itself. I want to remove mesa from my image. How can I do this as I don't kow why it is added? Aug 12 07:01:15 use bitbake -g and find out Aug 12 07:03:05 that's what I did and it only told me that it reverse depends on itself...not on any other package in my image Aug 12 07:03:29 mesa is providing virtuals Aug 12 07:04:35 mesa-common.inc:PROVIDES = "virtual/libgl" Aug 12 07:04:58 so look which package needs gl Aug 12 07:05:48 ok, thanks Aug 12 07:10:14 another thing: I've made a recipy that should overwrite a certain shared library that is already provided by another package. But when the package is added to the image, I get an error saying that that file is already provided and the image build then fails. Aug 12 07:10:31 yes Aug 12 07:10:54 thats the correct behavouir Aug 12 07:11:19 I only see one way to solve it Aug 12 07:11:33 working with virtuals Aug 12 07:11:39 make your own layer Aug 12 07:11:46 put both recipe there Aug 12 07:11:53 let them provide virtuals Aug 12 07:12:08 set in local.conf the prefferred one Aug 12 07:12:27 ok, I will try that. Thanks! Aug 12 08:10:13 morning Aug 12 08:10:26 hiho xora Aug 12 08:42:54 does libjpeg provide do conversions in raw format? Aug 12 09:16:37 bluelightning: any news about layer priority? Will be fixed? Removed? Will stay? Aug 12 09:19:11 ant_work: I don't know... the only conclusion we seem to have reached is the "old version in higher priority layer wins by default" behaviour is by design Aug 12 09:25:26 I'd favor an implicit priority given by parsing the string from left to right Aug 12 09:25:44 although bitbake -s is clearly broken in that regard Aug 12 09:25:49 and iirc pb_ seems to prefer too Aug 12 09:26:31 I'm not sure but I think if you remove all priority settings you should get that Aug 12 09:27:08 personally that's a little too implicit for my liking as default behaviour, but I'm willing to be outvoted Aug 12 09:28:26 I'm re-reading the msgs, it looks like pb_ and Khem would get rid of in-layer-priority Aug 12 09:29:33 and kergoth too Aug 12 09:30:12 bluelightning: maybe you should ping and rebump the subject Aug 12 09:30:32 ant_work: yes, I will do that (once I test the priority stuff) Aug 12 09:30:34 atm I have a recipe in 3 layers :/ Aug 12 09:30:54 well the first question in that situation should always be "why?" Aug 12 09:31:01 heh Aug 12 09:31:20 kexec-tools for the curious Aug 12 09:31:21 if there's a good reason, fine Aug 12 09:31:49 different packaging, more grained Aug 12 09:32:08 any reason you couldn't push that to one of the other layers? Aug 12 09:33:28 bluelightning: more scary example is the 'linux' recipe Aug 12 09:36:35 is the recipe in each layer expected to limit its scope adding COMPATIBLE_MACHINE? It would seem so.. Aug 12 09:37:23 then we have different linux.inc and that's another story :/ Aug 12 09:40:09 yeah :( Aug 12 09:40:34 judging by recent meeting minutes there has been talk about plain kernel recipes in oe-core, don't know if that's going ahead or not Aug 12 09:42:59 hi all Aug 12 09:45:58 hi pb_ Aug 12 09:46:16 pb_: any more thoughts on priority (see above discussion) Aug 12 09:46:18 ? Aug 12 09:48:14 hi pb Aug 12 09:48:57 morning woglinde Aug 12 09:49:39 hi ka6sox Aug 12 09:52:41 I should sleep...setting up Opal tonight. Aug 12 09:58:02 yo mickey|office Aug 12 09:58:35 & bp_ & woglinde & ka6sox & ant_work & bluelightning & everyone else :-) Aug 12 09:59:38 yo effem Aug 12 09:59:51 ka6sox yeah go take some sleep Aug 12 10:00:40 bluelightning: not immediately. as you say, the current behaviour is by design and I still think it is probably the right one. Aug 12 10:01:12 in most cases, it seems to me that having recipes duplicated across multiple community layers is probably a mistake and just should not happen in general. Aug 12 10:03:43 pb_: what if you don't have control on the outer layers? Aug 12 10:03:51 how can you devel ? Aug 12 10:04:02 I'm not sure I understand what you mean by "outer layers". Aug 12 10:04:25 in the specific case, the pile of layers composed by the Angstrom-scripts Aug 12 10:04:34 hi eFfeM_work, how are things? Aug 12 10:04:40 on teh top of oe-core and meta-oe Aug 12 10:04:57 can you give me a concrete example of a situation which is causing you a problem? Aug 12 10:05:19 mickey|office: fine, busy, had a bumpy real-life ride, but things are moving back to normal Aug 12 10:05:40 haven't decided yet if I'll go to ELC this year Aug 12 10:06:34 pb_: having klibc in meta-nslu2 and meta-zaurus. I committed it in meta-oe as third. Aug 12 10:06:37 jo boi Aug 12 10:06:39 i trying to use libjpeg, does this accept raw bayer and convert it into jpeg? Aug 12 10:06:40 eFfeM_work: i see. i can't make it unfortunately (baby duties), so i'll be absent to a GA for the 2nd year :/ Aug 12 10:06:59 mickeyl oh Aug 12 10:07:01 pity, but baby duties do have prio (been there myself) Aug 12 10:07:15 hope you can make it to fosdem Aug 12 10:07:40 yeah, i'll definitely try that Aug 12 10:07:44 pb_: it is clear to me that once recipe reaches meta-oe or oe-core teh 'custom' layer need to be purged Aug 12 10:07:51 (quite likely I'll go there as brussels is only an hour drive for me) Aug 12 10:07:58 ya, that's handy Aug 12 10:08:10 it's good to reach from FRA as well Aug 12 10:08:13 (that is an hour on weekends, probably 2 during the rush hour) Aug 12 10:08:16 there's a direct train which is pretty fast Aug 12 10:08:22 ah nice Aug 12 10:08:57 pb_: worst episode happens when linux_3.1 from meta-ti tries to build obviously lacking a defconfig Aug 12 10:08:58 perhaps we could even go alltogether… (wife and doughter)… need to think about that Aug 12 10:09:54 well there is the spouses city walk but maybe not too convenient with baby Aug 12 10:10:03 mickeyl pb did this to Aug 12 10:10:05 o Aug 12 10:10:49 yeah, I think I will try to go to fosdem again next year Aug 12 10:10:53 I didn't go this year Aug 12 10:11:21 last year I did indeed take my wife and (then) baby daughter. they didn't go on the organised walk, just spent the day amusing themselves in brussels. Aug 12 10:12:20 I guess emma must have been about 10 months old at that point, probably not vastly different to how old mini-mickey will be by next year :-) Aug 12 10:13:08 pb_: so I would expect the layer specified as last one in bblayers.conf always wins (so I build my kernel). Now, if you say to do this it should even mask recipes with higher PV from other layers, I can live with that Aug 12 10:15:29 and to the fresh fathers: childrens are compatible with flights after 3 years. Before they pay less but you risk to shame yourself on the plane... Aug 12 10:16:13 ant_work: heh, I think it is probably just the period between about 1 year and 3 years that you have a problem. when emma was really little we took her to the USA and she was fine on the plane, but I certainly wouldn't want to do that now. Aug 12 10:17:16 regarding your klibc thing, yes, I think you need to delete it from your "local" layers if you commit the recipe to an upstream repo and want to start using that one. Aug 12 10:19:10 pb_: (fosdem) awesome, would love to spend a dinner with you again Aug 12 10:19:39 yeah, that would be great Aug 12 10:20:06 pb_: mini-mickey, heh, correct, she's now 2 months Aug 12 10:20:48 oe family conversion? Aug 12 10:21:02 with mickeyl hrw pb and me? Aug 12 10:21:03 lol Aug 12 10:21:34 yeah, a diaper party ;-) Aug 12 10:21:40 na Aug 12 10:21:44 mine is now 5 Aug 12 10:22:45 well mine are 22 and not too interested in computing, so I don't think they'll join (although maybe they would like the beer party) Aug 12 10:23:19 ah, hopefully no diapers involved for you either then Aug 12 10:23:38 eFfeM_work you made something wrong with the computers Aug 12 10:23:51 mine is intressted Aug 12 10:24:11 mine were at 5, but mostly when it came to games etc Aug 12 10:24:29 games is the beginning Aug 12 10:25:34 * mickey|office would be more disappointed if she wasn't interested in music rather than if she wasn't into computers Aug 12 10:25:57 wait for the barbie time Aug 12 10:26:01 and pinky pin Aug 12 10:26:02 k Aug 12 10:26:15 i love barbie Aug 12 10:26:17 :D Aug 12 10:26:29 no diapers for me too Aug 12 10:26:29 (my wife doesn't though...) Aug 12 10:26:32 hrw: :D Aug 12 10:26:38 I mean: for my Mira ;) Aug 12 10:26:41 hehe, clear Aug 12 10:26:47 you're not old enough yet to have a need for diapers Aug 12 10:26:50 but I will probably have to skip fosdem 2012 Aug 12 10:26:53 gm hrw Aug 12 10:27:16 Linaro Connect Q1.12 will be in SanFrancisco week before ELC so right after fosdem Aug 12 10:27:49 and I do not want to fly BRU -> SF during european night Aug 12 10:30:42 speaking about children and computers... https://lh3.googleusercontent.com/-NNoUOjl-nSA/TkQE0P_GQDI/AAAAAAAAAW8/I5zc1P352BI/h301/11.08.2011%2B-%2B1 Aug 12 10:31:27 this is computer which Mira built yesterday for me (with my wife's help) Aug 12 10:31:59 hehe Aug 12 10:32:49 hrw *g* Aug 12 10:33:07 mine plays now mario on the game boy color Aug 12 10:33:20 and sometimes some games on the lego website Aug 12 10:33:31 Mira likes any painting apps on my phone or tablet Aug 12 10:46:42 hrw: excellent work Aug 12 10:46:47 (from mira) Aug 12 12:14:34 hi ant_work Aug 12 12:14:54 I wondered if you could pass the atags like that from /proc/atags: Aug 12 12:15:04 /proc/atags -> kexeced kenrel Aug 12 12:27:37 basically I've some issues booting the SHR or maemo kenrels, and I guess it's atags related Aug 12 12:27:54 for instance one minute delay between kexec -e and the knerel booting Aug 12 16:51:44 03Joshua Lock  07master * rc828ba571c 10bitbake.git/lib/bb/ui/crumbs/hobeventhandler.py: Aug 12 16:51:44 bb/ui/crumbs/hobeventhandler: adapt to reset -> reparse change Aug 12 16:51:44 The API has changed from just a reparse call to a reset call followed by Aug 12 16:51:44 a reparse call. Aug 12 16:51:44 Signed-off-by: Joshua Lock Aug 12 16:51:51 03Joshua Lock  07master * rb3ad7acebf 10bitbake.git/lib/bb/ui/crumbs/hobprefs.py: Aug 12 16:51:51 lib/bb/ui/crumbs/hobprefs: fix erroneous save/reparse Aug 12 16:51:51 We need to unset the reload_required variable once we've triggered a reload Aug 12 16:51:51 so that we don't cause a reload each time the preferences dialog is shown and Aug 12 16:51:52 dismissed, regardless of whether anything has changed. Aug 12 16:51:52 Signed-off-by: Joshua Lock Aug 12 16:51:53 03Joshua Lock  07master * r2843645755 10bitbake.git/lib/bb/ui/crumbs/configurator.py: Aug 12 16:51:54 bb/ui/crumbs/configurator: introduce writeConfFile method for all writes Aug 12 16:51:54 Configuration files are written in several places, this refactors the code Aug 12 16:51:55 to use a common method. Aug 12 16:51:56 Signed-off-by: Joshua Lock Aug 12 16:51:56 03Joshua Lock  07master * ra98f698fe9 10bitbake.git/lib/bb/ (command.py cooker.py): (log message trimmed) Aug 12 16:51:57 bb/command|cooker: refactor the reparseFiles logic Aug 12 16:51:57 Turn the reparseFiles logic into a command to reset the cooker's state Aug 12 16:51:58 machine and a noop which triggers a cache rebuild. Aug 12 16:51:58 The resetCooker command resets the cookers state machine such that a cache Aug 12 16:51:59 update will be triggered by any async command which requires the cache. Aug 12 16:51:59 The reparseFiles command remains as a noop async command that has the Aug 12 16:55:19 http://www.npr.org/blogs/pictureshow/2011/08/11/139548633/german-exhibit-reveals-secret-spy-photos?ft=1&f=97635953&sc=tw&utm_source=twitterfeed&utm_medium=twitter Aug 12 16:55:42 photo 4 of 6 must be the guy infiltrating the CCC :) Aug 12 17:09:19 crofton Aug 12 17:09:22 *g* Aug 12 18:59:02 I am trying different kernel from 2.6.35 to 3.0 and I am having problems with my rootfs. Aug 12 19:00:03 I just changed from 3.0 to 2.6.35 and cleaned my images. Aug 12 19:00:33 davidlt: you are moving from 2.5.35 to 3.0? Aug 12 19:00:34 I build a new image (rootfs), but it has /lib/modules/3.0.0/ in rootfs Aug 12 19:00:42 m Aug 12 19:00:52 msm: nope, from 3.0 to 2.6.35 Aug 12 19:01:38 But is still somehow mixes with 3.0.0, some files at least Aug 12 19:02:11 Running clean build would probably solve this, but my build machine is not fast enough for me. Aug 12 19:03:02 It looks like 2.6.35 ipks are marked as old, maybe that's the case. Aug 12 19:04:44 how did you override 3.0? Aug 12 19:05:12 are you manually build the 2.6.35 recipe or did you set PREFERRED_PROVIDER back to 2.6.35 Aug 12 19:06:14 (a) added DEFAULT_PREFERENCE = "-1" to all other images (b) defined preferred version of linux-omap3 and set MACHINE_KERNEL_VERSION in local.conf Aug 12 19:07:47 The kernel is build, but rootfs is bit mixed. My wireless driver sits in 3.0.0 modules and is not loaded by 2.6.35 Aug 12 19:14:03 how do I see bb.debug messages? Aug 12 19:14:11 can you see them after a build? or only by doing -D Aug 12 19:17:21 davidlt: is there a provider for linux-libs-headers? Aug 12 19:17:25 or perhaps a separate one for modules as well? Aug 12 19:20:16 msm: there is for linux-libc-headers set to .32 Hmm… My OLDEST_KERNEL ?= "2.6.39" Aug 12 19:20:55 does your kernel recipe build modules for you? Aug 12 19:23:16 All modules should be built-in into the kernel Aug 12 19:24:05 lol Aug 12 19:24:25 we had this discussin 20 years ago Aug 12 19:26:26 woglinde: in this channel? Aug 12 19:26:33 no Aug 12 19:27:08 btw. linus and tannenbaum Aug 12 19:30:07 morning Aug 12 19:31:08 msm: richard's recent bitbake logging changes will ensure that debug messages from tasks make it into the task logs regardless of what debug level you set for the bitbake UI Aug 12 19:31:25 see the mailing lists Aug 12 19:33:44 re ka6sox Aug 12 21:48:47 khem: ping **** ENDING LOGGING AT Sat Aug 13 02:59:57 2011