**** BEGIN LOGGING AT Tue Jun 28 02:59:58 2011 Jun 28 07:21:54 has anyone sucessfully used touchscreen for mini6410 board ? .....plz suggest Jun 28 07:22:25 the driver support or any info. Jun 28 07:36:29 good morning Jun 28 07:40:13 hello Jun 28 07:40:33 can anyone help me how to change screen resolution with fbdev? Jun 28 08:25:41 Does eGalax work for mini6410 board Jun 28 10:43:22 hi ant_work, florian, all Jun 28 10:43:33 hi pb_ Jun 28 11:21:28 03Robert Yang  07master * r41bef02bef 10bitbake.git/lib/bb/ (build.py cache.py runqueue.py): (log message trimmed) Jun 28 11:21:28 bitbake: Add task specific stamp file support Jun 28 11:21:28 This patch, based on proof of concept code from Richard adds code to Jun 28 11:21:28 bitbake to allow individual tasks to optionally specify their stamp Jun 28 11:21:28 file using the stamp-base flag. This takes the same form as the STAMP Jun 28 11:21:29 variable but can be specified on a per task basis. Jun 28 11:21:30 Code is also added to runqueue to ensure that if two tasks share the Jun 28 12:03:34 for 7 inch LCD connected to the friendlyARM board ,does any one know how is the touch screen connected to the LCD is it through USB or serial port internally Jun 28 12:56:45 guys, do you have a recipe for gsoap >= 2.8.0 ? Jun 28 12:57:00 hello :-) Jun 28 12:57:11 he he, hello otavio :-) Jun 28 12:57:21 ottavio: hey :-) you made me thing if I had written my nick wrong hehe Jun 28 12:57:31 yeah! Jun 28 12:57:34 hehe Jun 28 12:57:37 crap Jun 28 12:57:38 hehe Jun 28 13:03:50 o_O Jun 28 13:24:00 Is this channel the appropriate place to inquire about build errors that we believe are related to bitbake? Jun 28 13:40:25 Is someone seeing this? Jun 28 13:40:27 ERROR: Unable to parse /home/otavio/hacking/el/meta-openembedded/meta-oe/recipes-core/systemd/systemd_git.bb: Exited with "1" Jun 28 13:40:36 just pulled meta-oe Jun 28 13:40:41 and oe-core Jun 28 13:47:28 Got it fixed Jun 28 13:47:39 will prepare the patches for pushing. Jun 28 13:58:31 otavio, I see you are an active developer Jun 28 13:58:40 could you help me with a recipe? Jun 28 13:58:53 ottavio: dunno; what is your doubt? Jun 28 13:59:06 I am trying to write a recipe for gsoap 2.8.3 Jun 28 13:59:15 but compilation fails poorly Jun 28 13:59:27 and I'm not able to understand why Jun 28 14:00:25 ottavio: without a log; I can't help hehe Jun 28 14:01:10 ok, I'll upload to pastebin Jun 28 14:01:12 one sec Jun 28 14:01:17 that I produce it Jun 28 14:01:44 ottavio: paste the recipe too Jun 28 14:04:05 so, otavio, give a look at the recipe http://pastebin.com/HXd1AS72 Jun 28 14:04:09 it consists of three files Jun 28 14:04:36 I basically took the recipe for the previous version of gsoap and tried to use the last version Jun 28 14:05:25 no I'm getting the log Jun 28 14:09:02 * otavio waiting for the build log Jun 28 14:09:32 Is someone here available to help me with an SDK issue? Jun 28 14:09:57 it seems my built SDK lacks GCC and like Jun 28 14:11:24 oh, crazy, otavio, I cleaned up the recipe for posting to pastebin, I thought I changed nothing, but this time it worked..... Jun 28 14:13:27 hehehe Jun 28 14:14:11 you have write access to the OE repository? maybe you want to try it and upload it, because the version is OE is old Jun 28 14:26:21 lol 2 ot[t]avios Jun 28 14:26:30 hi Jun 28 14:26:37 GNUtoo: heeheh Jun 28 14:26:47 GNUtoo: one is good; imagine two of us :-D Jun 28 14:26:49 heheehhe Jun 28 14:26:50 lol Jun 28 14:27:19 * otavio confess; it is strange to read myself hehehe Jun 28 14:27:30 otavio, you live in italy? me too Jun 28 14:27:41 GNUtoo: I don't; Brazil Jun 28 14:27:55 at maybe it's ottavio that does Jun 28 14:28:14 ottavio: do you live in Italy? Jun 28 14:28:19 yes Jun 28 14:28:21 I'm italian Jun 28 14:28:25 ok Jun 28 14:28:36 GNUtoo: btw, did you get oe-core to build an sdk lately? Jun 28 14:28:37 otavio, I tought you were ottavio sorry Jun 28 14:28:46 otavio, I didn't got it to build an image Jun 28 14:28:50 I tried many times Jun 28 14:28:54 GNUtoo: no; ottavio is ottavio and I am otavio Jun 28 14:28:56 hehehe Jun 28 14:28:58 CRAP Jun 28 14:29:00 every time there were a new breakage Jun 28 14:29:00 hehe Jun 28 14:29:02 AAAAHHHHHHHHH!!!! Jun 28 14:29:06 :-) Jun 28 14:29:20 lol Jun 28 14:29:47 GNUtoo: i got it building but the tarball lacks toolchain Jun 28 14:29:51 GNUtoo: gcc and like Jun 28 14:30:45 03Richard Purdie  07master * r9d43e32798 10bitbake.git/lib/bb/data_smart.py: Jun 28 14:30:45 bitbake/data_smart: Optimise the data store iterator Jun 28 14:30:45 Since we're going to creat the seen set() anyway, we might as well use Jun 28 14:30:45 it directly. If we don't do this, we see thousands of function calls Jun 28 14:30:45 with associated overhead on profiles. Jun 28 14:30:46 Signed-off-by: Richard Purdie Jun 28 14:30:50 03Richard Purdie  07master * r2b5548c591 10bitbake.git/lib/bb/data_smart.py: (log message trimmed) Jun 28 14:30:50 bitbake/data_smart: Don't export deleted/empty entries in the list of keys Jun 28 14:30:50 If you d.delVar(), you expect the variable to be gone. Even empty Jun 28 14:30:50 variables continue to exist in the datastore and are still user visible Jun 28 14:30:50 I am looking forward to someone to help me to understand how does it is suppose to work on oe-core so I can fix mine Jun 28 14:30:51 unfortunately. The COW siutation means you can't just remove it Jun 28 14:30:51 since it might unmask a variable from an inner copy. Jun 28 14:30:52 This patch therefore stops empty variables from appearing in key lists Jun 28 14:30:52 03Richard Purdie  07master * r8d3c899e0a 10bitbake.git/lib/bb/parse/ast.py: (log message trimmed) Jun 28 14:30:53 bitbake/ast: Add optional argument for BBCLASSEXTEND Jun 28 14:30:53 Add an optional argument to BBCLASSEXTEND entries which gets passed to Jun 28 14:30:54 the extention class as BBEXTENDVARIANT. Also add BBEXTENDCURR whic Jun 28 14:31:21 is set to the current extension class name. Jun 28 14:31:22 This mode functions slightly differently to the previous BBCLASSEXTEND Jun 28 14:31:22 code in that PN is not changed. Jun 28 14:31:22 03Richard Purdie  07master * rd924ff9ede 10bitbake.git/lib/bb/data_smart.py: (log message trimmed) Jun 28 14:31:22 bitbake/data_smart: Don't track overrides in deleted variable names Jun 28 14:31:22 When we delete a variable we no longer expect it to override other Jun 28 14:31:23 variables. Jun 28 14:31:23 To do this we remove it from the list of active overrides at deletion Jun 28 14:31:24 time. It turns out we already had to do this at override expansion time Jun 28 14:31:24 so this cleans up the code to be more consistent as an added bonus. Jun 28 14:31:25 03Richard Purdie  07master * r664b85742d 10bitbake.git/lib/bb/parse/ast.py: Jun 28 14:31:25 bitbake/ast: Call expandkeys after the RecipePreFinalise event Jun 28 14:32:03 sgw: are you used to oe-core nativesdk ? Jun 28 14:32:14 sgw: I am lost on it and looking for help Jun 28 14:32:15 heh Jun 28 14:32:21 This means the event handler can change variables such as PN and those Jun 28 14:32:22 changes will be reflected in the updated variable key names. Jun 28 14:32:22 Signed-off-by: Richard Purdie Jun 28 14:32:22 (3 lines omitted) Jun 28 14:34:19 otavio: seems I may have come in the middle of a question, I am heading out in 5 minutes, could you send me an email with the issue and detail.s Jun 28 14:34:45 sgw: basically I made one sdk and it lacks gcc and like Jun 28 14:34:52 sgw: probably I am doing something stupid Jun 28 14:35:06 sgw: should be simple to fix but I got no good example to look at Jun 28 14:37:35 otavio: I will be back online in about 35 minutes, but in meetings until 10:00, might be able to help after that. Jun 28 14:37:50 sgw: ok; mind to ping me then? Jun 28 14:38:03 10:00am my time which is 2:30 hours from now. No problem I will do that. Jun 28 14:38:14 * sgw -> breakfast Jun 28 14:38:21 sgw: thanks in advance Jun 28 15:48:41 damnit Jun 28 15:48:48 i keep running into annoying limitations in GitPython Jun 28 15:49:35 turns out its implementation of the config file parser uses RawConfigParser, and doesn't enhance it to allow multiple entries with the same name Jun 28 15:49:50 git, of course, allows multiple entries with the same name, e.g. for fetching multiple sets of refs Jun 28 15:58:05 gah Jun 28 15:58:16 and it chokes and dies if it encounters a tag ref that isn't in refs/tags/ Jun 28 15:58:39 kergoth: nice robust exception handling! Jun 28 15:59:25 it's less an exception problem than just an implicit assumption in the code. it knows the correct ref, it should use it, but instead it reconstructs it -- ValueError: Reference at 'refs/tags/766da90/tags/1.0' does not exist Jun 28 15:59:39 the real path is refs/remotes/766da90/tags/1.0, in this case Jun 28 16:00:31 kergoth: ah. Jun 28 16:02:21 * kergoth mutters Jun 28 16:02:23 looks like the server for "file" just went away Jun 28 16:02:27 this is like the third problem i've had with this thing Jun 28 16:03:08 and yet, its interface is far superior to that of pygit2 Jun 28 16:03:12 maybe is hould give dulwich another shot Jun 28 16:05:29 argh, i don't think i can work around it this time Jun 28 16:05:32 * kergoth sighs Jun 28 16:09:39 * kergoth tries dulwich Jun 28 16:15:06 and dulwich blows up even sooner Jun 28 16:15:07 lovely Jun 28 16:26:13 dulwich is broken, gitpython has multiple bugs blocking me, and pygit2's interface sucks. Jun 28 16:26:16 * kergoth grumbles Jun 28 16:26:24 guess i can try pygit2 anyway, or try to fix gitpython.. Jun 28 16:27:22 maybe i'll just do my own python git wrapper for now, that'd probably be the quickest solution, if not the prettiest Jun 28 16:31:42 kergoth: it seems like a work duplication Jun 28 16:31:56 * kergoth really thinks his git cache stuff would be a better way to implement bb.fetch2.git, but has to be able to get it implemented in python to do it cleanly.. Jun 28 16:32:16 03Tom Rini  07master * r82c21e61d1 10openembedded.git/classes/cpan.bbclass: Jun 28 16:32:16 cpan.bbclass: Make cpan_do_configure undo the perl.real unwrapping Jun 28 16:32:16 The cpan build system unwrapps our perl -> perl.real shell script Jun 28 16:32:16 and generates Makefile's that call perl.real directly and thus fail Jun 28 16:32:16 to be relocable. So at the end, regex them back to calling perl. Jun 28 16:32:17 Signed-off-by: Tom Rini Jun 28 16:32:27 03Tom Rini  07master * rd8f645c96f 10openembedded.git/recipes/file/ (file-5.05/reloc.patch file_5.05.bb): Jun 28 16:32:27 file: Fix relocation patch Jun 28 16:32:27 We weren't calling the correct binreloc function (and we needed to add Jun 28 16:32:27 /misc to the path we got) and we need to pass in ENABLE_BINRELOC to have Jun 28 16:32:27 the changes take effect. Finally, add in a new patch header. Jun 28 16:32:27 Signed-off-by: Tom Rini Jun 28 16:32:52 kergoth: while you're on it, mind to look why AUTOREV seems to not work with GIT? Jun 28 16:35:24 Is anyone testing relocated sstate builds in oe-core atm? Jun 28 16:35:54 Unless perl 5.13 or 5.14 (whichever oe-core has) has fixed userelocableinc I bet it needs that same kind of fix instead of what we do atm :( Jun 28 16:39:30 Tartarus: https://github.com/kergoth/reloc-tests - haven't touched it in a long time, but that's basically my relocation testing script adjusted for oe-core, in a repository similar to my more current oe-core-template Jun 28 16:51:22 What does it mean for bitbake to "stage" something? Jun 28 16:51:43 Is that the part where packages and compiled and installed into /sysroot ? Jun 28 16:51:50 *are Jun 28 16:52:14 joelagnel: installed into sysroot, yes Jun 28 16:52:58 in the old scheme there was a special do_stage for this; now this is handled as part of normal installation Jun 28 16:53:06 (in recipes that is) Jun 28 16:53:20 ah Jun 28 16:53:36 i.e. if you see do_stage it's legacy, ideally you shouldn't copy it for new recipes Jun 28 16:54:45 bluelightning, also how do we know if dependencies are statically compiled? Do we look at DEPENDS? (or is this upto the build system) Jun 28 16:55:01 like statically compiled libraries Jun 28 16:55:30 joelagnel: I'd have to defer to someone else on static linking; I've not done any within OE Jun 28 16:55:57 bluelightning, oh ok. thanks for your answer though Jun 28 16:56:22 bluelightning, mind if i just tell you what it is I'm doing? Jun 28 16:57:06 joelagnel: not sure but if you mean how does it know how to determine runtime dependencies when something might have been statically linked in instead - this should be handled; the system checks shared library dependencies and statically linked executables shouldn't have any I wouldn't think Jun 28 16:57:24 joelagnel: sure, fire away Jun 28 16:58:04 BlindMan, its like I am building the "gstreamer-ti" recipe, but now I changed some code in "ti-dmai" which AFAIK is statically linked into gstreamer Jun 28 16:58:29 I cleaned and rebuilt ti-dmai but gstreamer still seems to use the "old" ti-dmai before my modifications Jun 28 16:58:57 Is this even possible, because I did do a clean on ti-dmai before rebuilding both Jun 28 16:59:13 *sorry, I was supposed to address you as bluelightning ;) apologies. Jun 28 16:59:31 no worries :) Jun 28 16:59:57 joelagnel: you rebuild gstreamer-ti afterwards right? Jun 28 17:00:12 bluelightning, yeah I cleaned and rebuilt it Jun 28 17:00:32 it sounds like what you're describing shouldn't happen... a rebuild should have staged the new libs and then gstreamer-ti should have picked them up Jun 28 17:00:48 bluelightning, are static libraries staged as well? Jun 28 17:01:24 probably they are, nvm :) Jun 28 17:01:43 Also, does a -c clean remove what was staged earlier? Jun 28 17:01:57 packaged staging should ensure that I think Jun 28 17:02:11 in any case a rebuild will overwrite those files Jun 28 17:02:24 ah ok, yeah makes sense, hmm Jun 28 17:02:55 I guess the first thing to check would be to look at what ti-dmai has put in the sysroot Jun 28 17:03:27 bluelightning, logs would tell me? maybe the run.do_ scripts? Jun 28 17:04:05 or maybe a `find` in sysroot :) Jun 28 17:04:45 log.do_install would be the first port of call (there's no do_stage for ti-dmai it seems) Jun 28 17:20:19 * joelagnel wants lightning back :(( :) Jun 28 18:11:44 jo crofton Jun 28 18:11:47 italia Jun 28 18:16:47 yeah Jun 28 18:16:49 hiking Jun 28 18:17:03 trying to dl email Jun 28 19:07:51 anyone know if there's a larger png of the angstrom logo then the 98x64 one in the kexec/files/angstrom dir? Jun 28 19:07:59 angstrom/compass Jun 28 19:09:15 jconnolly ask koen Jun 28 19:11:02 thanks woglinde Jun 28 19:16:36 jconnolly: I think cbrake should also have Jun 28 19:17:13 thanks khem. emailed koen Jun 28 19:17:24 contrib/artwork also has .svg Jun 28 19:17:38 err thats for oe Jun 28 19:17:42 not for angstrom Jun 28 19:18:10 khem right Jun 28 19:18:23 that's why I pointed to koen Jun 28 19:18:44 I think actually the graphic desginer we use at buglabs did the oe logo ;D Jun 28 19:18:53 yes Jun 28 19:19:01 our small upstreaming contribution Jun 28 19:19:03 jconnolly: yes I remember Jun 28 19:19:33 jconnolly: btw. contrib/angstrom has some stuff too Jun 28 19:19:49 jconnolly: it has angstrom.png and .svg too Jun 28 19:19:58 oh excellent Jun 28 19:20:01 I'll look Jun 28 19:20:35 k Jun 28 19:20:53 that's perfect khem thanks a lot Jun 28 19:21:01 np Jun 28 20:24:51 * otavio still missed on nativesdk problem Jun 28 21:58:55 heh AVM vs. Cybits interesting legal battle Jun 29 01:55:54 hello Jun 29 02:45:18 otavio, hi Jun 29 02:45:34 joelagnel: hi **** ENDING LOGGING AT Wed Jun 29 02:59:56 2011