**** BEGIN LOGGING AT Mon Jul 25 02:59:57 2011 Jul 25 07:58:26 morning all Jul 25 07:59:30 bluelightning: morning and good night! Jul 25 07:59:46 sgw: hi, sleep well :) Jul 25 08:00:16 yup 8:00am go/nogo in the morning, just tucking everything in for the night :-) Jul 25 08:00:36 * sgw -> Zzzz Jul 25 08:34:19 morning all Jul 25 08:35:28 hi RP__ Jul 25 10:29:56 * bluelightning -> office Jul 25 13:06:04 RP__, ping. around ? Jul 25 13:06:26 zeddii: pong Jul 25 13:07:15 I'll send a patch shortly to force 2.6.37 as the kernel version. I was so used to the recipe names 'saving' us from an auto update, that the force to 3.0 was missed (well, that the fact I was using it on purpose for testing). Jul 25 13:08:38 I presume that qemu.inc will fixup core, but I'll have to touch all the meta-yocto and intel ones explicitly. unless you know of a more global way to do it ? Jul 25 13:14:53 zeddii: The only other way would be a default setting in defaultdistro Jul 25 13:15:04 zeddii: then machines could upgrade as required Jul 25 13:17:38 aha. I'll look that way as well. my first effort seems to still be building 3.0 at the moment anyway :) Jul 25 13:21:36 RP__: ./include/tune-xscale.inc: TUNE_PKGARCH = "${@['armv5teb', 'armv5te'][bb.data.getVar('SITEINFO_ENDIANESS', d, 1) == 'le']}"; seems causing some problems.. there is tmp/sysroots/x86_64-linux/usr/bin/armv5teb-oe-linux-gnueabi.gcc-cross-initial/arm-oe-linux-gnueabi-gcc Jul 25 13:21:51 JaMa: I know :( Jul 25 13:22:01 RP__: but eglibc doesn't think it's big endian and fails to find arm-oe-linux-gnueabi-gcc Jul 25 13:22:18 JaMa: Since this whole area is going to get an overhaul with the tune changes shortly I'm working on that rather than trying it patch it up Jul 25 13:22:36 ok, I'll wait a bit more for test then Jul 25 13:22:39 thanks Jul 25 13:41:53 RP__: any comments on siteinfo series? Jul 25 13:42:04 Tartarus: no, looks reasonable Jul 25 13:42:08 k, thanks Jul 25 14:06:10 RP__: ping Jul 25 14:06:21 yuke: pong Jul 25 14:07:08 RP__: i notice you have create several new multilib branch, so which branch should i rebase to? Jul 25 14:07:21 yuke: rpurdie/ml3 Jul 25 14:07:38 RP__: get it Jul 25 14:08:01 yuke: Its changing as I complete code, that branch doesn't work well for ARM atm Jul 25 14:08:18 yuke: When I've finished the ARM code I'll get back to one branch again Jul 25 14:08:40 RP__: i see. BTW, i have new several patches in kyu3/ml, you may want to pick. i will send you one email Jul 25 14:08:59 yuke: ok, thanks Jul 25 14:09:10 yuke: I've started pulling some pieces into master Jul 25 14:10:27 RP__: do you mean pulling the infrastructure change? Jul 25 14:10:51 yuke: I mean I've started to merge the tune file changes Jul 25 14:11:26 RP__: oh, yes, i just see that. that is great Jul 25 14:26:25 RP__ (or anyone). stupid question time. is there any reason why a PREFERRED_VERSION for virtual kernel wouldn't work in a .inc file ? I'm having zero luck convincing bitbake that 2.6.37 should be the default, not 3.0. Jul 25 14:27:45 zeddii: is it being oerwritten somewhere? Should work from a .inc.... Jul 25 14:28:04 zeddii: I assume you mean a .conf file .inc, right? Not a recipe Jul 25 14:28:31 yup. conf .inc, I tried in: meta/conf/machine/include/qemu.inc Jul 25 14:28:44 I'll dump my env and look for overrides. Jul 25 14:30:17 zeddii: does bitbake -e | grep PREFERRED_VERSION show it? Jul 25 14:32:23 it does. I'm using the syntax that you gave me a while ago to force my local.conf to build 2.6.34 vs 2.6.37 .. so that might be the issue. Jul 25 14:39:02 I have to say that the version handling and variable naming makes me feel stupid very quickly. Jul 25 14:39:09 * zeddii thinks he should just go back to bed Jul 25 14:47:01 finally. gak. Jul 25 14:52:23 hey RP__ Jul 25 14:52:25 zeddii: what was the issue? Jul 25 14:52:39 If I post a patch putting ${HOST_CC_ARCH} into CPP, is that alright with you? :) Jul 25 14:53:35 Tartarus: looking at bitbake.conf, yes Jul 25 14:53:45 k, thanks Jul 25 14:54:06 now to double check what the tools folks would call this option and maybe i should be shoving it into TOOLCHAIN_OPTIONS Jul 25 14:55:15 Tartarus: hmm... you're giving me an hint: klcc-cross has same PACKAGE_ARCH as gcc-cross, isn't? Jul 25 14:55:45 being a perl werapper around xgcc Jul 25 14:55:58 hm Jul 25 14:56:14 Tartarus: what do u need to put into TOOLCHAIN_OPTIONS Jul 25 14:56:33 ant_work: yes klcc-cross is most probably like gcc-cross Jul 25 14:56:38 khem: Making CodeSourcery toolchains build Jul 25 14:56:46 and for ia32 it's a little bit of extra fun Jul 25 14:57:00 ant_work: sounds right Jul 25 14:57:08 khem: so hybrid-packaging is not a nonsense (klibc is still machine_arch) Jul 25 14:57:14 Tartarus: ok Jul 25 14:57:44 Tartarus: precompiled Jul 25 14:59:00 khem: hm? Jul 25 14:59:23 Anyhow, apparently I should be shoving -msgxx-glibc into TOOLCHAIN_OPTIONS since that's where we throw in --sysroot Jul 25 15:00:21 of course CSL toolchain is precompiled so you are trying to glue it into oe-core I suppose Jul 25 15:00:50 or compiling it from source using oe ? Jul 25 15:01:54 kergoth is fiddling with bringing oe-core's external-toolchain-csl back to life Jul 25 15:02:00 and then throwing in the stuff we've added Jul 25 15:02:15 * Tartarus is doing product work Jul 25 15:02:40 ok Jul 25 15:07:53 mornin Jul 25 15:16:09 hi fray Jul 25 15:16:25 I'll have a multilib arm patch ready shortly so don't work on that :) Jul 25 15:25:58 Tartarus: Did you/kergoth see the comments on the list about the external toolchain Jul 25 15:26:09 Tartarus: someone posted a list of things they had to tweak to make it work Jul 25 15:26:09 I did, I think he did Jul 25 15:26:30 Tartarus: I replied and the person agreed to come back and help try and fix things Jul 25 15:29:24 Yeah, it'll be a little tricky, for everything to work right Jul 25 15:29:31 I kinda wonder what we want to do about this Jul 25 15:30:12 (Which is to say, for at least many arches we need non-standard flags throw into TARGET_CC_ARCH or TOOLCHAIN_OPTIONS for things to really work) Jul 25 15:30:54 * Tartarus will try and do the consistency fixes that using the IA32 toolchain showed up, eg all the places where we were assuming gcc/cross-gcc could just be invoked unadorned, but that Jul 25 15:30:57 's not the case) Jul 25 15:37:31 hmm what is FEED_ARCH now Jul 25 15:38:27 morning folks Jul 25 15:39:24 morning Jul 25 15:40:28 RP__ hmm so FEED_ARCH is not TUNE_PKGARCH ? Jul 25 15:40:45 * khem needs to change the code which used it Jul 25 15:41:41 khem: it should be the same Jul 25 15:42:08 hi fray Jul 25 15:43:27 RP__: ok I will use it instead of FEED_ARCH now Jul 25 16:25:08 Sigh, TOOLCHAIN_OPTIONS must be ld safe, so I'll put HOST_CC_ARCH into CPP into my next series Jul 25 16:34:38 Good Morning All ! Jul 25 16:41:20 hi nitink Jul 25 16:41:31 hi RP__ Jul 25 16:45:33 RP__ I see rpurdie_ml4 branch Jul 25 16:45:40 shall I switch to it ? Jul 25 16:46:50 nitink: no, stick with ml3 for now Jul 25 16:47:02 ml4 just has arm pieces in it which I'm discussing with people atm Jul 25 16:47:40 RP__: ok Jul 25 17:08:40 * sgw sez: git.yoctoproject.org is moving IP address we are going to mark it read-only Jul 25 17:09:07 * sgw sez: once your DNS catches up you should find the writable git. Jul 25 17:10:45 You can visit git.yoctoproject.org and check the banner to see if the transition has completed for you. Jul 25 18:25:56 halstead: can we change the default theme for cgit on git.yoctoproject.org ? Jul 25 18:26:06 halstead: its quite unreadable for diffs Jul 25 18:29:20 zeddii: Should we be setting the PREFERRED_VERSION in poky.conf or default distro? Jul 25 18:30:15 I have no strong opinion. I didn't see a default distro in meta-yocto, so I picked on poky.conf. Jul 25 18:30:40 khem, I'm the system admin so I don't work on the theme of the site but I'll see if I can track down the right person. Jul 25 18:31:00 halstead: please please do Jul 25 18:31:03 * Tartarus agrees with khem Jul 25 18:32:25 Tartarus, khem Can you paste in a URL that you think is a particularly good example of where the theme doesn't work? Jul 25 18:33:13 http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=819f18f8bc000f13b644edc194d2a12b4ea5fecf Jul 25 18:34:40 Thanks Tartarus. Jul 25 18:43:21 halstead: default is good Jul 25 18:43:34 I mean default theme of cgit Jul 25 18:44:29 khem, I'll note that too. Jul 25 18:46:28 halstead: e.g. this here http://git.openembedded.org/cgit.cgi/bitbake/ is more soothing to eyes Jul 25 20:59:45 * RP__ was responsible for that theme fwiw Jul 25 20:59:47 halstead: ^^^ Jul 25 21:00:33 The diffstats are harder to read but I don't seem to have a problem with the diffs themselves. Must be my monitor settings... Jul 25 21:01:19 It was done with a dark theme to match the other Yocto sites. I suspect some tweaks to the specific components that are causing problems to increase contrast would help... Jul 25 21:06:16 RP__: yeah, it's your monitor I guess Jul 25 21:06:24 I find it unreadable for anything non-trivial Jul 25 21:08:07 Tartarus: Anything in particular (red, green, white?) or all of it? Jul 25 21:08:16 RP__, I like the consistency with the main project color palette. I've been told it's also very difficult to see when displayed on an LCD projector. Jul 25 21:08:34 For me the green on gray is the hardest part Jul 25 21:08:48 Then the diffstat Jul 25 21:08:52 I cannot read the diffstat Jul 25 21:09:11 diffstat is marginal for me too Jul 25 21:11:46 RP__: diffs in general are hard to read for me Jul 25 21:12:09 RP__: that may be they should use more lighter color for text Jul 25 21:12:27 but I think using default theme would be much better if you ask me Jul 25 21:41:18 fray: ping Jul 25 21:56:14 fray: talked with HJ, he left comments on the bugzilla Jul 25 22:37:08 khem: FWIW, git.pokylinux.org shows the same thing with the other theme :) Jul 25 22:37:53 heh Jul 25 22:38:08 Well, is that theme really better for you? Jul 25 22:38:24 Tartarus: which theme? Jul 25 22:38:26 It's nice to not be cookie-cutter sometimes, but that's it buying us Jul 25 22:38:36 the yoctoproject one that khem and I can't stand ;) Jul 25 22:42:39 arg, blue on grey is bad Jul 25 22:42:53 well, look slike th e(in)famous termcolor used by mc Jul 25 22:42:54 http://box-look.org/content/preview.php?preview=1&id=135870&file1=135870-1.png&file2=&file3=&name=%22term+color%22+mc+skin&PHPSESSID=9c2bcf9f7e8ddb9ddcdf1e7a4bffa5ca Jul 25 22:43:00 the dark background makes it hard Jul 25 22:43:11 this almost beats it... Jul 25 22:53:54 RP__: it seems yocto colors are sacred for the git.yp.org Jul 25 22:54:57 I dont know if folks who read it find not an issue I am fine. I just was raising my own discomfort it doesnt have to change if everyone is ok with it Jul 25 23:31:39 khem: Lets see if we can improve it retaining the light on dark theme Jul 25 23:32:05 khem: I don't have time right now to look at it though and mention pokylinux.org just as a helpful work around **** ENDING LOGGING AT Tue Jul 26 02:59:56 2011