**** BEGIN LOGGING AT Tue Jul 19 02:59:56 2011 Jul 19 05:48:34 03Martin Jansa  07master * r82e76a4d64 10openembedded.git/recipes/navit/navit_svn.bb: Jul 19 05:48:34 navit_svn: return trailing space to SRC_URI Jul 19 05:48:34 * so that existing appends are not broken Jul 19 05:48:34 * it was removed in f8a2cccaca84bd3748e8733ac592ade5c5b8eeec Jul 19 05:48:34 Signed-off-by: Martin Jansa Jul 19 07:54:45 03Joel A Fernandes  07org.openembedded.dev * r14795a3217 10openembedded.git/recipes/beagleboard/beagleboard-test-scripts_git.bb: Jul 19 07:54:45 beagleboard-test-scripts: BeagleBoard userbutton detection for flashing Jul 19 07:54:45 * Build and install a userbutton tool to detect state of the user button on a BeagleBoard from /dev/input/event0 Jul 19 07:54:45 * Install a user.txt and uEnv.txt to /boot, Narcissus will copy from here to the FAT partition of an SD Card Jul 19 07:54:45 Signed-off-by: Joel A Fernandes Jul 19 07:54:45 Signed-off-by: Koen Kooi Jul 19 07:55:18 03chase maupin  07org.openembedded.dev * r2913806fc8 10openembedded.git/recipes/nodejs/ (2 files in 2 dirs): (log message trimmed) Jul 19 07:55:18 nodejs: Add patch to compile with GCC 4.3.3 Jul 19 07:55:18 * Add a patch to nodejs to allow compiling with GCC versions Jul 19 07:55:18 less than or equal to 4.3.3. Jul 19 07:55:18 * Define the function nodejs_get_gcc_ver to allow applying the Jul 19 07:55:19 patch conditionally based on compiler version. Jul 19 07:55:19 * Patch based on post at https://github.com/joyent/node/issues/883 Jul 19 08:41:55 someone also having "4.6.1/ld: cannot find -lm" while building gcc-runtime? Jul 19 08:42:09 just did build from scratch and it's the same Jul 19 08:42:59 also missing "eglibc-2.13-r8+svnr14157/image/usr/bin/localedef and eglibc-2.13-r8+svnr14157/image/usr/lib/gconv" in eglibc-2.13 do_locale_install Jul 19 08:53:43 JaMa: I've seen complaints, yes. Still using 2.12 (Angstrom) Jul 19 08:54:17 2.13 worked before here.. Jul 19 09:01:11 morning all Jul 19 09:01:19 It worked ok for me last time I tried. Let me do a fresh build now. Jul 19 09:01:23 hi bluelightning Jul 19 09:01:30 hi pb_ Jul 19 09:01:58 I think I built from scratch on friday anyway, and rp doesn't appear to have landed any changes since then so I would expect it still to work. but who knows. Jul 19 09:02:12 * pb_ rm -rf tmp* sstate-cache Jul 19 09:02:54 pb_: thanks Jul 19 09:03:59 ok, it's running now, be a few minutes before it gets to gcc-runtime Jul 19 09:21:24 pb_: what about the report of broken thumb code generation? Jul 19 09:22:07 I forgot to verify the binaries.. Jul 19 09:26:41 ant_work: I didn't see that report Jul 19 09:26:51 was it on some list? Jul 19 09:26:56 here on irc Jul 19 09:27:01 ah, I must have missed it Jul 19 09:27:07 can you re-paste the complaint? Jul 19 09:27:10 ensc|w Jul 19 09:28:23 11:42.43 ensc|w why does gcc-configure-cross.inc in oe-core defines do_compile_prepend() but not do_configure_prepend()? Old OE modified configure() and new variant makes build fail when '-mthumb' is required for target arch Jul 19 09:28:44 oh, yeah, I did see that Jul 19 09:28:50 I think he's probably right Jul 19 09:28:53 http://ibot.rikers.org/%23oe/20110718.html.gz Jul 19 09:29:04 around 12 0clock Jul 19 09:29:14 ant_work: real (or other) problem is in zecke-xgcc-cpp.patch Jul 19 09:29:27 hi, tried to avoid pasting Jul 19 09:29:46 XGCC_FLAGS_FOR_TARGET should be set for CPP too Jul 19 09:30:24 so you expect the binaries in rootfs could be b0rked? Jul 19 09:30:54 no; toolchain will not build for thumb-only archs Jul 19 09:31:36 iirc Angstrom was thumb-only (in oe-dev) Jul 19 09:31:40 I am in doubt that anything except ./configure cares about cpp -E Jul 19 09:31:48 :p Jul 19 09:32:20 for armv7 Jul 19 09:32:29 well, seems false alarm then Jul 19 09:33:35 ant_work: problem appears as Jul 19 09:33:36 $ echo | arm-linux-uclibceabi-gcc -mcpu=cortex-m3 -E - Jul 19 09:33:36 :1:0: error: target CPU does not support ARM mode Jul 19 09:33:55 heh Jul 19 09:34:13 you have to specify -mthumb explicitly Jul 19 09:34:30 yeah, that's always seemed a little bit bogus to me Jul 19 09:34:32 I've seen this caused funny flames about linux-boot Jul 19 09:34:47 I'm not quite sure why gcc requires you to give this extra "please make it work" option Jul 19 09:35:19 you'd think that -mthumb should just be the default for cortex-m and -mno-thumb should be an error. Jul 19 09:35:46 yes; I would expect it too Jul 19 09:35:54 but, anyway, I agree, it should be using the same flags for cpp as it does for gcc. Jul 19 09:36:53 there are a few other possible cases where configure does "echo "#ifndef __something__\n#error xx\n#endif" | gcc -E and gets the wrong results if the flags are incorrect Jul 19 09:41:05 pb_: I'll try again clean rebuild with added P_V_eglibc-locale.. http://git.openembedded.org/cgit.cgi/openembedded-core-contrib/commit/?h=shr&id=88620d7609ebc998eb68cc6266f9917ec4d8f0c6 Jul 19 09:43:40 NOTE: package gcc-runtime-4.6.0+svnr175150-r9: task do_compile: Succeeded Jul 19 09:43:42 seems ok Jul 19 09:53:02 micro-base-image completed successfully Jul 19 10:01:45 pb_: so, how micro is micro these days? (I've not looked at it but I presume from the name the aim is very small size) Jul 19 10:53:41 bluelightning: right now, a micro-image-build for micro-uclibc on i586 is about 2.3MB before compressing. Jul 19 10:53:58 it used to be a bit smaller than that in oe.dev, we seem to have picked up some bloat in the move to oe-core. Jul 19 10:54:06 that's still pretty small :) Jul 19 10:54:20 presumably there are some optimisations that can be done Jul 19 10:54:45 as for the aim, yeah, basically small size and suitability for actual embedded use (rather than a general-purpose distro that happens to run on "embedded" hardware) Jul 19 10:54:54 yeah Jul 19 10:55:04 for example I just noticed that libgcc_s.so.1 doesn't appear to be getting stripped for some reason Jul 19 10:55:19 and we seem to have 200k of garbage in /etc Jul 19 10:55:50 presumably that 2.3mb is without package management? Jul 19 10:55:55 yeah Jul 19 10:56:12 it does include dropbear as a concession to interactive debugging, but no package management Jul 19 10:56:44 total of 361 files and directories according to "find . | wc -l" Jul 19 10:57:35 the same thing with eglibc is about 1MB bigger, iirc Jul 19 10:58:11 (more or less a direct and obvious consequence of glibc itself being that much bigger than uclibc) Jul 19 11:08:27 pb_: failed the same again.. I will try to debug why, thanks for test Jul 19 11:12:55 pb_: yes about the bloat: I see that even in the initramfs images Jul 19 11:13:12 in my case just empty dirs Jul 19 11:15:08 yeah. some empty dirs, some random files that we don't want. Jul 19 11:15:28 I also seem to be getting a populated static /dev again even though we are meant to be using devtmpfs nowadays Jul 19 11:15:40 not quite sure what's gone wrong there Jul 19 11:15:52 heh, 'defaults' Jul 19 11:16:34 image.bbclass Jul 19 11:16:38 if devtables == None: Jul 19 11:16:39 oh, and sysvinit is still installing a bunch of useless-ish binaries (eg wall). I think it always did that though. Jul 19 11:16:40 devtables = 'files/device_table-minimal.txt' Jul 19 11:19:27 ah right, yeah Jul 19 11:21:47 pb_ in fact I have a device-tables_micro.txt just for booting the initramfs Jul 19 11:22:05 afaik even with devtmpfs we need /dev/console Jul 19 11:23:01 really? I think if you set devtmpfs to auto-mounting then you shouldn't need anything in /dev on the image Jul 19 11:23:18 auto-mounting doesn't work for initramfs Jul 19 11:23:21 ah Jul 19 11:23:22 must be done by hand Jul 19 11:23:28 okay, that makes sense then Jul 19 11:23:46 in that case, yeah, probably do need /dev/console in the initramfs Jul 19 11:23:49 http://lxr.free-electrons.com/source/scripts/gen_initramfs_list.sh Jul 19 11:23:57 seems the only node Jul 19 11:25:27 btw, I tend to forget you can 'stack' those initramfs's Jul 19 13:06:10 03Martin Jansa  07master * r5ca8109054 10openembedded.git/ (2 files in 2 dirs): Jul 19 13:06:11 linux-nokia900-meego: update defconfig to build at least kernel-module-tun Jul 19 13:06:11 * some people like me are using openvpn with tun/tap Jul 19 13:06:11 * it was removed in 6a65a05de26f7875ea0cc0d03f09c14c4d151ad2 Jul 19 13:06:11 Signed-off-by: Martin Jansa Jul 19 14:10:13 does anyone have an "idiots guide" to adding a custom user to the 2011.03-maintenance branch? Jul 19 15:29:02 Good afternoon/morning/evening all. Jul 19 15:29:16 ~ugt Jul 19 15:29:17 well, ugt is Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html Jul 19 15:29:22 morning :) Jul 19 15:29:45 :) Jul 19 15:30:44 I have a vendor driver package which wants to see the kernel source tree in order to compile. How would I do that in my recipe for this package? It has out-of-tree drivers. Jul 19 15:31:20 It would be nice if it could be a disposable kernel tree, in case I don't trust this vendor package to not mess it up. Jul 19 15:38:53 is STAGING_KERNEL_DIR what I hope it is? Jul 19 15:40:52 yeah it is Jul 19 15:41:39 bluelightning: so, I was inspired to remove a bit of bloat from uclibc and busybox. micro-base-image now down to 1.9MB after manually stripping libgcc Jul 19 15:41:59 not quite sure what's causing the unstripped library, I suspect db99a65b3e93dfacc27ea821c788f15b5de3a497 is probably to blame but need to confirm that Jul 19 15:42:04 pb_: cool :) Jul 19 15:43:03 pb_: are you using linux-yocto for the kernel or linux.inc based? any customisations? Jul 19 15:43:18 I'm just about to try using linux-yocto Jul 19 15:43:56 I had previously been put off that because it seemed to be a bit scary and yoctoish but it seems worth a try. Jul 19 15:44:18 pb_: it is quite different, definitely Jul 19 15:44:25 pb_: apparently darren hart (dvhart) is working on an ongoing project to remove kernel bloat, based on linux-yocto Jul 19 15:44:50 yeah. I am still a bit unclear what exactly the goals of linux-yocto are, and what exactly I should expect to be different from the upstream kernel Jul 19 15:45:21 I did send a note to the list last week asking about it, and I got a fairly lengthy essay from Bruce Ashfield in response but, unfortunately, he omitted to actually answer the question. Jul 19 15:45:47 pb_: well there's always Darren's ELC talk if you haven't already seen that Jul 19 15:45:53 ah, no, I hadn't seen that Jul 19 15:46:08 I can dig up the link if you like Jul 19 15:46:14 that would be handy, thanks Jul 19 15:46:19 iirc wind river has some sort of different workflow stuff based on it, and assembly of chunks of config files into a single defconfig, or similar, but I'm not sure about actual content Jul 19 15:46:19 heh Jul 19 15:46:32 * pb_ not cool enough to go to elc Jul 19 15:46:51 kergoth: oh right, sounds funky Jul 19 15:47:03 pb_: http://free-electrons.com/blog/elc-2011-videos/ Jul 19 15:47:08 thanks Jul 19 15:47:09 just search for his name Jul 19 15:48:33 doh, what?! crazy bitbake seems to be trying to check gcc out from svn all over again. I'm sure I already have about eight copies of that tarball on hand. Jul 19 15:48:41 must have forgotten to set DL_DIR or something Jul 19 15:48:45 * pb_ stabs self in eye Jul 19 15:49:00 pb_: that sounds painful :( Jul 19 15:50:16 oh, I see, svn rev went backwards when I reverted that patch Jul 19 15:50:26 gotta just wait it out I suppose Jul 19 15:50:55 hi pb Jul 19 15:51:04 not sure if you're using them in micro, but when building angstrom at home I've found the angstrom mirrors are so darn slow it's easier to just turn them off Jul 19 15:51:29 in particular the gcc svn snapshots were downloading agonizingly slowly Jul 19 15:51:37 bluelightning narcissus eats all the power Jul 19 15:51:39 hi woglinde Jul 19 15:51:52 or you got the us mirror Jul 19 15:51:54 woglinde: yeah, it's definitely overloaded AFAICT Jul 19 15:52:04 ah, here we go, I found a copy of the tarball on another machine Jul 19 15:52:09 back in business now Jul 19 15:52:18 I don't think I'm using the angstrom mirrors. Jul 19 15:52:45 what host do they run on, discovery? Jul 19 15:53:16 the virtualbox mirrors are overloaded too Jul 19 15:53:18 *g* Jul 19 15:55:16 fwiw, discovery has load 0.22, doesn't seem to be a problem there. Jul 19 15:55:26 Hi, I am stuck... I cannot get a clean build on head of master branch at the moment. My problem is I have a working kernel & fs but cannot figure out where in the tree it was built from. At boot it says "Angstrom 2010.7-test-20101125 mini2440 ttySAC0" any idea where it gets this text from/how I can find this in the git repos history? Jul 19 15:55:31 I think there might be an apache configuration issuette of some kind though, it does seem to be going a bit slow Jul 19 15:55:49 guyvdb_: sounds like a question for #angstrom Jul 19 15:56:26 oops lost connection Jul 19 15:57:04 guyvdb_ its build from the kernelsources Jul 19 15:58:35 I am trying to track down the point (approximate) in the openembedded git repos where I built it... i.e. a commit uuid Jul 19 15:59:12 ? Jul 19 15:59:19 could I not find the text Angstrom 2010.7-test-20101125 in the repos somewhere? Jul 19 15:59:22 the kernel as a recipe Jul 19 16:01:06 er, does that text really come from the kernel? Jul 19 16:01:10 basically the kernel and filesystem that I was building on head of the repos was working great although I had not updated much so it was pretty old. I have now done a git pull and my board hangs after populating udev... i do not know if it is the kernel of the fs Jul 19 16:01:15 it sounds more like /etc/issue to me, that's not a kernel thing Jul 19 16:01:21 but, anyway, this is still a question for #angstrom Jul 19 16:01:28 no the text is at the login prompt... Jul 19 16:01:45 guyvdb_: and in /etc/angstrom-version? Jul 19 16:02:31 ahh ok got a revision number there Jul 19 16:02:34 guyvdb_ its udev Jul 19 16:02:40 which hangs Jul 19 16:02:46 gm ka6sox Jul 19 16:03:02 morning woglinde Jul 19 16:03:07 ok let me reboot with new kernel and pastie the console... one min Jul 19 16:03:53 guyvbd try http://narcissus.angstrom-distribution.org/ might be faster for you Jul 19 16:06:05 < woglinde> guyvbd try http://narcissus.angstrom-distribution.org/ might be faster for you Jul 19 16:06:08 ups Jul 19 16:06:11 -> Jul 19 16:06:13 *g* Jul 19 16:06:19 ok so when I build from my current git repos my machine hangs with hwclock: can't open '/dev/misc/rtc': No such file or directory Jul 19 16:07:03 no its udev Jul 19 16:07:05 woglinde I have been building rootfs/kernel a number of times before... this is only occuring since my last pull of the repos Jul 19 16:07:09 believe me Jul 19 16:07:18 yes Jul 19 16:07:31 so were in udev? Should I look at udev rules? Jul 19 16:07:36 you didnt update for half an year Jul 19 16:07:43 yup Jul 19 16:07:45 oe-cache stuff for udev Jul 19 16:08:50 woglinde can you give a suggestion for me to move forward? what should i investigate? Jul 19 16:09:31 woglinde here is full output of console on boot http://pastie.org/2237858 Jul 19 16:09:46 I know the console output Jul 19 16:09:53 as I said Jul 19 16:10:03 try first narcissus to build you an image Jul 19 16:10:11 than look which version are installed Jul 19 16:10:24 ok Jul 19 16:12:07 woglinde should i build it with udev as the /dev manager? On Narcissus Jul 19 16:12:42 why not Jul 19 16:12:48 depends on your needs Jul 19 16:12:58 I can't tell you your needs Jul 19 16:19:05 thinking of libgcc_s, does anybody remember what the construct is that requires it to be installed? Jul 19 16:19:12 something to do with thread cancellation, if I remember right Jul 19 16:19:24 woglinde: ok the rootfs from Narcissus booted fine with the kernel I built ... now to figure out what I am doing wrong Jul 19 16:20:29 pb sorry no Jul 19 16:20:30 ok that is built from a diffenet branch to me... let me check that out... thx for the help woglinde Jul 19 16:20:34 khem might know Jul 19 16:20:44 guyvdb_ sure nor prob Jul 19 16:20:57 yeah, I guess I can look back in the old logs or something Jul 19 16:21:04 it'd be nice to get rid of it, even stripped it's 200kB or so Jul 19 16:21:39 pb_, do you know of any reason that melo would have ntop running as a daemon? Jul 19 16:21:42 pb uhm 200kb is nothing Jul 19 16:21:49 we have 32 gig mmc now Jul 19 16:21:50 ka6sox: nope. Jul 19 16:22:03 woglinde: well, 200kB is more than 10% of my current image size Jul 19 16:22:05 pb_, thanks Jul 19 16:22:12 pb ah Jul 19 16:22:16 pb .so or .a? Jul 19 16:22:20 .so Jul 19 16:22:32 hm you could look at the symbols Jul 19 16:22:37 readelf *g* Jul 19 16:22:39 and, sure, you can buy a 32GB MMC but it still costs a finite amount of money Jul 19 16:22:53 (and most of the cheap ones are garbage) Jul 19 16:23:19 pb hm my transced 8 gig class10 for 11 euros are okay Jul 19 16:23:25 woglinde: well, I know what symbols are in it, that's not really the point Jul 19 16:23:40 woglinde: all the transcend ones we have tested have been useless Jul 19 16:24:01 they seem to just start randomly losing data after a while Jul 19 16:25:06 hm my not Jul 19 16:25:15 and buffered them a lot Jul 19 16:25:25 with swap sometimes and compiling Jul 19 16:25:57 oh well, I guess you have been lucky Jul 19 16:26:23 yeah maybee Jul 19 16:26:23 with almost every one we have tried, if you put a big file (eg a movie or something) on the card and then read it a few thousand times, after a while the data you get back becomes corrupt. Jul 19 16:26:35 but sure afterwars a I ordered them Jul 19 16:26:44 a looked up some reports about them Jul 19 16:26:59 hm oh Jul 19 16:27:04 didnt test with movie Jul 19 16:27:23 * pb_ go home now Jul 19 16:27:24 later all Jul 19 16:28:14 till later pb Jul 19 16:28:46 One man's trivial space saving is another's 2 cents / pence / eurocents off the BOM Jul 19 16:29:23 cya pb_ Jul 19 16:31:29 woglinde just to confirm the current repo is : git://git.openembedded.org/openembedded ? Jul 19 16:36:48 hmm If i do a git branch -l I only have a master branch from git://git.openembedded.org/openembedded but the build Narcissus did for me was from branch 2011.03-maintenance what am I missing? Do I have a clone of an old repo? Jul 19 16:38:14 branch Jul 19 16:38:21 type git branch -r Jul 19 16:38:27 you will see all remote branches Jul 19 16:39:09 guyvdb_: read Pro Git Jul 19 16:39:25 k .. gonna have a google Jul 19 16:39:40 anyone know how to disable a touchpad in X? I keep bumping the damn thing and selecting things i don't want to Jul 19 16:39:49 at the very least i want to disable tap-to-click Jul 19 16:39:53 hrm Jul 19 16:41:28 hm input section in xorg.conf? Jul 19 16:41:42 but that isnt well documented Jul 19 16:53:12 hmm, anyone play with tmux or byobu? Jul 19 17:10:37 kergoth: I have, why do you ask? Jul 19 17:11:10 just curious about people's thoughts. debating diving into tmux to replace screen, but byobu looks quite convenient Jul 19 17:12:23 I've used both and the only real difference that I've noticed is that ability to divide the screen vertically in tmux. Jul 19 17:12:28 * kergoth nods Jul 19 17:12:34 screen can too, but only if its patched, which sucks a bit Jul 19 17:12:46 I didn't know that. Jul 19 17:14:43 I use byobu Jul 19 17:14:47 http://thechangelog.com/post/3164295957/teamocil-ruby-yaml-tmux-layouts is rather cute, too, in the tmux area Jul 19 17:14:53 ka6sox: ah, cool Jul 19 17:15:24 makes things consistent. Jul 19 17:21:35 both are nice....I just have been using screen for like 20+yrs. Jul 19 17:21:45 so got used to it Jul 19 17:22:00 yeah, I know what you mean Jul 19 17:22:14 bbiab...heading to office Jul 19 17:27:10 Tartarus: do you feel like backporting f6433e75a047a2d38ab60efea8a7ab7cc3dcd822 from oe.dev to oe-core by any chance? Jul 19 17:28:45 It needs to but it's not high on my list if you wanna do it for me :) Jul 19 17:29:05 righto Jul 19 17:29:22 I am standing by for an impending visit from my mother in law right now, but I will do it tomorrow if I get a moment Jul 19 17:32:28 thanks Jul 19 17:32:42 * Tartarus really wants to get his current tangents done so he can get back to siteinfo Jul 19 17:33:19 BTW working on the multilib and tuning stuff.. Jul 19 17:33:32 there are a few siteinfo tables scattered.. it would REALLY be nice if that stuff could move to the tuning files.. Jul 19 17:33:47 'er.. not siteinfo, I mean architecture info, related to the siteinfo Jul 19 17:33:58 like the get_siteinfo_list... Jul 19 17:34:04 yeah, urk Jul 19 17:34:13 need to get RP to finish reviewing the main patch I sent there Jul 19 17:34:22 Since that will make life much easier I hope for multilib Jul 19 17:34:34 I head back on vacation (ComicCon) tomorrow.. Jul 19 17:34:47 BTW I have an oe-core branch now for the multilib stuff.. Jul 19 17:36:16 fray: Ah, I was in SD last week. Jul 19 17:37:10 * fray works on pushing out his oe-core/multilib stuff.. Jul 19 17:57:09 kergoth: is there any advantage of using files/ folder instead of ${PN} folder to hold the recipe patches Jul 19 18:02:14 msm: Can you go down the hall, I'm assuming anyhow, and remind galak he should be hanging out in here now that he's posting oe-core patches? ;) Jul 19 18:02:35 khem: only if the file needs to be available to mjultiple recipes with different names, which is unlikely Jul 19 18:03:08 khem: for the case where there's one recipe in the recipes dir, I always preferred 'files' as it bugs me to repeat the recipe name like 42 times in the path, but it seems like the preference is for ${BPN} nowadays Jul 19 18:03:39 Tartarus: haha, yea. he tends to avoid irc because too many people bug him… I can serve as his liaison. Jul 19 18:04:11 heh Jul 19 18:04:57 Don't have anything too big to bug him about, just wanna make sure that with that series we're still normally ending up as linux-powerpc-gnuspe and that openssl is just odd Jul 19 18:05:09 since i'd swear we don't usually have linux-gnueabi-arm Jul 19 18:12:35 Tartarus: i mentioned your request to gim Jul 19 18:12:39 him* Jul 19 18:12:57 k Jul 19 18:12:59 thanks Jul 19 19:50:03 perhaps a silly question is there a way to tell python I no longer care about a variable so it can clean up the memory behind it? Jul 19 19:52:54 fray: python has a gc Jul 19 19:52:57 fray: 'del' - http://docs.python.org/reference/simple_stmts.html#del Jul 19 19:53:40 khem, I know.. but I wanted to set the life span of a variable myself.. to the inside of a for ... in ..: statement Jul 19 19:53:50 for scope Jul 19 19:54:05 I believe if I don't use del, then the variable will hang around until the end of the function.. since scoping appears to be on the function side.. Jul 19 19:54:55 yes Jul 19 19:55:06 its either func level or global Jul 19 19:55:30 although you can use nested function Jul 19 19:55:38 ya.. I'm making a copy of 'd'.. so I want it to be even less to avoid huge swarms of memory hanging over a function that may take a while to execute Jul 19 19:55:41 and then scope will limit to the inner function Jul 19 19:55:51 the gc can do it's thing whenever Jul 19 19:56:31 heh you are looking for free() equivalent it seems Jul 19 19:57:03 ya.. del should work Jul 19 19:59:03 fray: actually python has block scope Jul 19 19:59:09 and global scope to be precice Jul 19 19:59:12 precise Jul 19 19:59:27 blocks can be functions Jul 19 20:00:27 so when you do b = list(a + i for i in range(10)) Jul 19 20:00:38 then i is scoped to that block only within a function Jul 19 20:00:48 makes sense Jul 19 20:00:59 you could use something like that Jul 19 20:01:07 although I dont know what you are trying to do Jul 19 20:03:05 some prototype code.. so it'll get heavily reviewed (I assume).. so if del isn't right, someone should catch it Jul 19 20:03:20 hi khem Jul 19 20:05:01 hoi kergoth Jul 19 20:11:50 03Christopher Larson  07master * r807fb0b3bb 10bitbake.git/lib/bb/cache.py: Jul 19 20:11:50 cache: fix remnant broken 'info' reference from recent cache changes Jul 19 20:11:50 Signed-off-by: Christopher Larson Jul 19 20:13:48 what were to reasons to move opkg's statedir from /usr/share into /var? This makes upgrading impossible when /var is on a tmpfs. Jul 19 20:15:49 03Simon Busch  07master * rf9f1458cb9 10openembedded.git/recipes/freesmartphone/libmsmhll_git.bb: Jul 19 20:15:49 libmsmhll: add the hci link layer implementation library fo msmcomm Jul 19 20:15:49 Signed-off-by: Simon Busch Jul 19 20:15:59 03Simon Busch  07master * rc171f15176 10openembedded.git/recipes/freesmartphone/msmcomm.inc: Jul 19 20:15:59 msmcomm: bump SRCREV for latest updates Jul 19 20:15:59 Signed-off-by: Simon Busch Jul 19 20:16:00 03Simon Busch  07master * r394db2469c 10openembedded.git/recipes/linux/linux-palmpre_git.bb: Jul 19 20:16:00 linux-palmpre: bump SRCREV for latest updates Jul 19 20:16:00 Signed-off-by: Simon Busch Jul 19 20:16:00 03Simon Busch  07master * r7fd50fc9cf 10openembedded.git/recipes/freesmartphone/msmcommd_git.bb: Jul 19 20:16:00 msmcommd: add libmsmhll as dependency and bump PR Jul 19 20:16:01 Signed-off-by: Simon Busch Jul 19 20:16:01 03Simon Busch  07master * re561219496 10openembedded.git/ (2 files in 2 dirs): Jul 19 20:16:02 linux-palmpre: update defconfig and bump MACHINE_KERNEL_PR Jul 19 20:16:02 Signed-off-by: Simon Busch Jul 19 20:51:01 Hi woglinde Jul 20 01:50:44 hey galak Jul 20 01:51:03 Did you get a chance to see if with those patches stuff is normally linux-powerpc-gnuspe and just openssl is weird? Jul 20 01:51:44 Tartarus: don't follow Jul 20 01:52:56 w/ your gnuspe changes, do most things eval out to linux-powerpc-gnuspe, ala linux-arm-gnueabi (unless oe-core has gone crazy in this dept..) Jul 20 01:54:34 having to look for linux-gnuspe-powerpc seems very odd Jul 20 01:54:40 but, openssl is an odd configure Jul 20 01:54:50 yeah, it appears linux-powerpc-gnuspe is normal Jul 20 01:55:03 yeah Jul 20 01:55:10 was able to build all of core-image-sato and only failures where openssl & flac Jul 20 01:55:20 flac because it was trying to mix altivec & spe which is not allowed Jul 20 01:55:29 an openssl because of its odd configure Jul 20 01:55:45 k Jul 20 01:58:15 what's the normal process with patches getting picked up an applied? Jul 20 01:59:53 Getting back on track for that, expect sgw to pick it up and put it in a consolidated pull request pretty soon Jul 20 02:00:01 then RP pulls it Jul 20 02:04:57 is there a way to just a source tree for a package after patches applied? Jul 20 02:05:22 I want to generate a new patch for binutils and trying to figure easiest way to make sure I generate one that will apply Jul 20 02:06:26 bitbake -c patch foo **** ENDING LOGGING AT Wed Jul 20 02:59:57 2011