**** BEGIN LOGGING AT Thu May 30 03:00:00 2013
May 30 04:41:17 <Net147> how would I make it so a recipe is never automatically added as a runtime dependency for another recipe?
May 30 04:59:08 <Net147> I guess I would use EXCLUDE_FROM_SHLIBS to make sure other packages don't automatically depend on a package
May 30 06:46:47 <Net147> I get xterm: Cannot allocate color "rgb:5c/5c/ff" when opening xterm on target device. any ideas?
May 30 07:52:48 <mckoan> good morning
May 30 08:11:13 <bluelightning> morning all
May 30 08:28:31 <Net147> for PACKAGECONFIG, if I want to add xcms to it do I just put PACKAGECONFIG += "xcms" in distro conf file?
May 30 08:30:08 <rburton> no, because you need to specify what packageconfig you want to add to
May 30 08:30:42 <rburton> use PACKAGECONFIG_pn-[recipename]
May 30 08:30:43 <Net147> I want to enable xcms in libx11 recipe
May 30 08:31:10 <rburton> not sure that you can do += with overrides like that
May 30 08:31:30 <rburton> so you'll probably need PACKAGECONFIG_append_pn-[recipe] = " xcms"
May 30 08:31:52 <Net147> literally or do I need to replace "recipe" with something?
May 30 08:33:54 <rburton> you need the put the recipe name there
May 30 08:34:10 <rburton> oh hang on, xcms in libx11?
May 30 08:34:12 <rburton> really?
May 30 08:34:15 <rburton> you know that, well, doesn't work.
May 30 08:34:54 <Net147> how so?
May 30 08:36:21 <rburton> well, it's barely used. what did you want it for?
May 30 08:36:50 <Net147> I am trying to work out why xsetroot -solid rgb:5c/5c/ff or xterm -bg rgb:5c/5c/ff doesn't work in my image
May 30 08:37:25 <Net147> xsetroot giving unknown color and xterm giving unable to allocate color
May 30 08:38:09 <Net147> it seems to fail on the XParseColor API call which is part of xlib
May 30 08:38:20 <Net147> and I am thinking, maybe it has been disabled somehow
May 30 08:38:28 <rburton> thats not xcms
May 30 08:38:36 <rburton> maybe you're missing rgb.txt
May 30 08:38:48 <Net147> okay. how would I get it installed?
May 30 08:41:41 <Net147> looks like I have /usr/share/X11/rgb.txt
May 30 08:41:47 <Net147> I will try adding it to xorg conf
May 30 08:42:23 <rburton> it should find it automatically
May 30 08:43:52 <Net147> I added rgbpath to xorg conf, it still doesn't help
May 30 08:44:25 <rburton> i thought you could only pass #hex values like html to X
May 30 08:44:42 <Net147> 5c/5c/ff is hex values
May 30 08:45:08 <Net147> if I run xsetroot on a normal Linux distribution and point DISPLAY to the target device, it accepts the rgb value fine
May 30 08:45:38 <rburton> try #5c5cff
May 30 08:46:06 <Net147> rburton: that works
May 30 08:46:27 <Net147> rburton: but how come rgb doesn't work?
May 30 08:47:17 <Net147> maybe it needs some locale support to know what "rgb:" is?
May 30 08:47:36 <rburton> ah, that's handled by xcms
May 30 08:47:47 <rburton> just use #5c5cff instead of enabling xcms just for that
May 30 08:48:27 <rburton> its a non-trivial lump of code, enable it if you want, but this is the first time i've seen someone want it :)
May 30 08:48:34 <Net147> it's causing ls --color to not work in xterm
May 30 08:48:44 <Net147> the directories are not colored purple-blue
May 30 08:48:52 <Net147> they are set to the foreground color instead
May 30 08:50:07 <Net147> so I would like my directory listings to be properly colored
May 30 08:50:09 <Net147> =)
May 30 08:50:26 <rburton> luckily you can use packageconfig to turn it on :)
May 30 08:51:16 <Net147> if I have something_append = "a", and then something_append = "b", what will something contain?
May 30 08:52:04 <rburton> ab
May 30 08:52:21 <rburton> note that you need a leading space for something like packageconfig
May 30 08:52:30 <rburton> or ba, depending on parse order
May 30 08:52:40 <Net147> so can I technically do PACKAGECONFIG_append = " xcms" ?
May 30 08:53:19 <rburton> that will probably add xcms to every packageconfig statement ever
May 30 08:53:24 <rburton> which is why i said to use _append_pn-libx11
May 30 08:53:42 <bluelightning> agreed, no probably about it, it definitely will :)
May 30 08:54:07 <Net147> PACKAGECONFIG_append_pn-libx11 = " xcms" in my distro config then?
May 30 08:54:23 <rburton> yes
May 30 08:56:38 <Net147> rburton: thanks =)
May 30 12:04:40 <Net147> rburton: you're right. xlib --disable-xcms was the reason why rgb: wasn't being recognized.
May 30 12:05:33 <Net147> rburton: and also why the colors weren't working
May 30 12:08:31 <ndec> hi. i was wondering what the license is on the various *.png file in yocto git. in the ./documentation. i think it's not obvious, as there is no mention of the documentation in LICENSE. and i'd like to know how i can reuse the images in a presentation.
May 30 12:11:03 <rburton> Net147: it might be worth a revisit and enabling xcms by default
May 30 12:11:22 <Net147> rburton: I agree. I spent many hours trying to figure it out
May 30 12:13:47 <Net147> rburton: was digging into xterm and xsetroot source code to trace the problem
May 30 12:28:45 <Net147> rburton: may I leave that in your experienced hands? or should I open a bug report? =)
May 30 12:30:23 <rburton> Net147: open a bug so it doesn't get forgotten
May 30 12:38:43 <Net147> rburton: done - https://bugzilla.yoctoproject.org/show_bug.cgi?id=4576
May 30 12:38:44 <yocti> Bug 4576: normal, Undecided, ---, ross.burton, NEW , libx11: Xcms should be enabled by default
May 30 12:39:26 <rburton> merci
May 30 12:52:14 <Net147> rburton: I would have thought it is  a regression because it was enabled by default previously for normal libx11
May 30 12:53:33 <rburton> which wasn't the default libx11, the trim version was
May 30 12:53:38 <rburton> (iirc)
May 30 12:53:54 <Net147> I see
May 30 12:57:25 <Net147> what about custom distros where people set PREFERRED_PROVIDER_virtual/libx11 = "libx11"?
May 30 12:59:27 <rburton> for them, its been broken for about a year
May 30 12:59:38 <rburton> i'll sort it (or you can send a patch)
May 30 12:59:44 <rburton> i'm away next week so i may not get to it
May 30 12:59:54 <rburton> if you do send a patch, before/after binary sizes would be interesting
May 30 13:00:19 <Net147> well I tested on my desktop Linux OS
May 30 13:03:10 <Net147> --disable-xcms: 12100710 bytes, --enable-xcms: 12112575 bytes
May 30 13:03:48 <rburton> non-trivial but not massive
May 30 13:04:24 <rburton> toggle it to enabled then people can turn it off if the 12k bothers them :)
May 30 13:08:17 <Net147> more actually if I recompile both with same compiler/libs
May 30 13:08:18 <Net147> --disable-xcms: 12100710 bytes, --enable-xcms: 12184174 bytes
May 30 13:08:56 <Net147> 83464 bytes
May 30 13:09:18 <Net147> I guess newer GCC must be adding bloat
May 30 13:09:42 <rburton> is that stripped?
May 30 13:09:51 <Net147> yes
May 30 13:09:59 <rburton> the best way would be to do the change in oe-core and then use buildhistory-diff
May 30 13:16:13 <Net147> perhaps
May 30 13:16:56 <Net147> rburton: I haven't got llvmpipe to work yet. swrast driver fails to load. something about not being able to find an llvm ostream symbol
May 30 13:17:06 <rburton> shame
May 30 13:17:31 <Net147> rburton: I could try a newer LLVM. was using LLVM 2.9 but it's up to 3.2 now.
May 30 13:18:48 <Net147> is there any reason crossscripts is placed after the native sysroot bin instead of before?
May 30 13:18:52 <Net147> in PATH
May 30 13:20:48 <Net147> I have llvm-config native version in native sysroot bin, and I also have an llvm-config script in crossscripts.
May 30 13:21:23 <Net147> so it picks the native one when I am building a recipe for target that depends on llvm
May 30 13:23:46 <Net147> maybe it should be if building native recipe then native sysroot bin comes first, otherwise the crossscripts path comes first?
May 30 14:09:44 <Net147> rburton: isn't Mesa at 9.1.3 now?
May 30 14:10:45 <Net147> rburton: http://lists.freedesktop.org/archives/mesa-announce/2013-May/000051.html
May 30 14:12:33 <rburton> Net147: they need to update their web site more often
May 30 14:12:44 <rburton> damnit
May 30 14:12:53 <Net147> rburton: yeah. their website does lag behind
May 30 14:13:01 <Net147> rburton: I have noticed that in the past too
May 30 14:13:27 <Net147> rburton: I only know that it's at 9.1.3 because of package updates on my desktop Linux distribution
May 30 14:16:30 <Net147> rburton: I noticed in the release announcement they use --with-llvm-shared-libs. maybe I should build LLVM with --enable-shared. perhaps that would solve my problem
May 30 14:39:58 <ivali> Is there an option to make yocto build stripped version of vmlinux?
May 30 14:46:27 <ivali> I want to build with no debugging symbols
May 30 14:56:19 <Net147> is bitbake broken? 'tuple' object has no attribute 'quit'
May 30 14:58:30 <Net147> I reverted to 5cb59cc6910d8f3165528c4a71e29e4de897d242. seems to be working.
May 30 14:58:44 <Net147> who broke bitbake?
May 30 14:59:04 <kergoth> use git-bisect
May 30 15:04:11 <Net147> kergoth: 0fc3a1eddfbab5cbf61c028cf8bc0d6b27b6c420
May 30 15:07:18 <Net147> fails when I doing clean build, PRSERVER localhost
May 30 16:11:33 <rburton> anyone seen "undefined reference to yywrap" when trying to configure a package that uses AC_PROG_LEX?
May 30 16:12:08 <zecke> rburton: yacc vs. bison? flex not installed?
May 30 16:13:30 <rburton> its there
May 30 16:14:12 <rburton> generates the C and compiles it, then fails
May 30 16:14:53 <fray> the newer version of flex and bison seem to be more demanding..
May 30 16:15:03 <fray> some of the deprecated items people have been using for years appears to have finally gone
May 30 16:19:03 <rburton> oh, that's not the failure
May 30 16:19:16 <rburton> oh goddamniut
May 30 16:19:19 <rburton> "flex.real"
May 30 16:19:27 <rburton> is very much not "^flex "
May 30 16:20:04 <kergoth> there was a recent mail thread where someone got bit by that sort of version check due to the wrapper use, not sure if it was ever resolved, though
May 30 16:20:45 <fray> ahh.. ya thats a fun one..
May 30 16:21:01 <rburton> it wasn't
May 30 16:21:21 <fray> I don't think you can exec from shell w/o being forced to change arg0..
May 30 16:22:40 <rburton> you can do it with bash
May 30 16:23:02 <fray> I didn't realize you could..
May 30 16:23:04 <rburton> $ cat mwhaha
May 30 16:23:06 <rburton> exec -a mwhahaha ./myname
May 30 16:23:08 <rburton> $ cat myname
May 30 16:23:10 <rburton> echo Hello.  I am $0
May 30 16:23:12 <rburton> $ ./mwhaha
May 30 16:23:14 <rburton> Hello. I am mwhahaha
May 30 16:23:24 <fray> ahh nice
May 30 16:23:36 * rburton will try
May 30 16:23:39 <fray> so that might be a fix for some of the 'real' wrappers.. use bash and the exec -a
May 30 16:32:49 <rburton> hm, the ones that bounce though env lose the -a
May 30 16:35:33 <rburton> (because env does another exec)
May 30 16:36:40 <fray> ahh
May 30 16:36:50 * rburton thinks he has a fix
May 30 16:36:56 <rburton> instead of env, use "eval $@"
May 30 16:38:45 <rburton> $ flex --version
May 30 16:38:47 <rburton> flex 2.5.35
May 30 16:38:49 * rburton cheers
May 30 16:39:16 <fray> I think most of the wrappers are generated using the wrapper class (might be wrong though).. so if there is a better way, that'd be the place to make the initial change
May 30 16:39:25 <rburton> yeah, fixed it in the class
May 30 16:41:40 <rburton> oh, eval won't work
May 30 16:41:49 <rburton> because it won't export them to the subshell
May 30 18:02:31 <mranostay> hi Jefro
May 30 18:03:24 <Jefro> mranostay heya
May 30 18:04:53 * mranostay curses proxies
May 30 22:04:12 <kergoth> hmm, oe-core has a different systemd version than poky does, afaict. oe-core has 204, poky has 197, both master. this doesn't seem quite right
May 30 22:07:14 <fray> that seems odd..  they're normally in sync
May 30 23:20:27 <mranostay> hi fray!
May 30 23:42:20 <Net147> I wonder if Yocto/OE should be using its own built perl for pod2man, text2html, etc. instead of the host perl. perl 5.18 on host is causing documentation build breakages.
May 30 23:43:04 <kergoth> meta              = "(detachedfromecb30a9):ecb30a910b8c2f3d39ab4d3bf5f3eb1898097b76"
May 30 23:43:07 * kergoth scratches head
May 30 23:43:41 <Net147> kergoth: did you just do a git bisect or something?
May 30 23:44:28 <kergoth> no, having floating checkouts is what i intended, but clearly that formatting is wrong and redundant
May 31 00:18:45 <Net147> sgw1: will the slides for the presentation be made available online?
May 31 01:20:31 <sgw1> Net147: yes, the LF folks will be putting the slides up after they get back next week
**** ENDING LOGGING AT Fri May 31 02:59:58 2013