**** BEGIN LOGGING AT Fri May 27 02:59:58 2011 May 27 06:27:32 hi all May 27 07:18:36 good morning May 27 08:45:13 morning all May 27 08:48:17 there is not much activity there! May 27 08:53:02 bluelightning: argh..koen is supporting spitz in the old layer... May 27 08:53:52 I've the new kernel configs almost done. This evening we'll have 2.6.39 May 27 09:05:29 ant_work: cool May 27 09:06:07 morning, bluelightning May 27 09:07:27 hi anarsoul May 27 09:24:38 anarsoul: do you have a defconfig for h1940 I could use? May 27 09:25:14 bluelightning: sure, I can send it to you by email May 27 09:25:28 anarsoul: thanks May 27 09:27:52 bluelightning: done, it's for early 2.6.39 RCs, but should work with 2.6.39 May 27 09:29:43 btw smthg changed in BT config, some thing cannot be modular anymore (L2CAP, ...) May 27 09:29:59 I was migrating from 2.6.38, though May 27 09:30:48 bluelightning: would you add *all* usb modules for BT and networking ? May 27 09:30:55 usb and CF May 27 09:31:30 hi ppl May 27 09:31:41 hey Jay7 May 27 09:31:54 bluelightning: do you support hx4700? May 27 09:32:22 Jay7: to a degree... I have one here and I have been trying to get it working to a satisfactory level lately May 27 09:32:32 still struggling with tslib May 27 09:33:27 hi Jay7 May 27 09:33:51 bluelightning: nice, alexxy said me that hx4700 should work ok with latest kernels May 27 09:33:56 anarsoul: hi May 27 09:34:10 Jay7: it did look like most of the support pieces should be there May 27 09:34:30 my concern is I'll have the same problems with other devices still stuck on 2.6.21-hh May 27 09:34:57 hence why I'm still trying to get the touchscreen to work with the old kernel May 27 09:35:12 bluelightning: I see koen moved A. to tslib_git May 27 09:35:44 ant_work: yes, that definitely doesn't work with older devices; however the odd thing is tslib-1.0 doesn't anymore either :/ May 27 09:36:03 issue are linux-libc-headers May 27 09:36:09 iirc May 27 09:36:24 not directly kernel May 27 09:36:37 bluelightning: maybe it'll be better to get newer kernel for hx4700? May 27 09:36:52 anarsoul: that's the end goal yes May 27 09:37:24 I do have to support some of these older ones for the moment though also May 27 09:37:32 see, for Zaurus we still have big issues with keyboard...but those will be probably solved if we bother Marex... but only using smthg like linux-next May 27 09:37:46 ant_work: ok, interesting I'll check linux-libc-headers May 27 09:37:47 kernel devs have no time for older 2.6.2 May 27 09:37:59 naturally not May 27 09:38:10 it looks like the definition changed after 2.6.37 headers May 27 09:39:22 quoting JaMa May 27 09:39:28 looks like this problem May 27 09:39:30 https://github.com/kergoth/tslib/commit/412d99d8b92c12545f939972146a38c5074f3dcb May 27 09:39:32 or in OE May 27 09:39:33 http://git.openembedded.net/cgit.cgi/openembedded/commit/?id=390e8cd69ee11bcc8a10ea3d887a386eab2579cb May 27 09:40:43 ant_work: thanks May 27 09:41:04 hm...you didn't read zaurus-devel, isn't? May 27 09:42:57 ant_work: yes I'm on that list May 27 09:43:43 see, I have ASSUME_PROVIDED teh linux headers and I still had 2.6.36 May 27 09:43:55 issue raised after I updated to 2.6.37 May 27 09:45:19 my xinit doesnot read xinitrc ..... and by default it just open xterm.... why is it so.... can anyone help me? May 27 10:24:46 bluelightning: do you own ipaq rx3xxx? May 27 10:25:02 anarsoul: no, that's not one of the ones I have May 27 10:25:10 though I could get one easily from ebay May 27 10:32:07 looks like there's no jtag info for rx3715 :( May 27 10:36:19 anarsoul: hope you didn't kill yours... ? May 27 10:36:38 bluelightning: nope, but I plan to port u-boot to it. However, I did not get it yet :) May 27 10:36:50 ah right :) May 27 10:47:29 anarsoul: kexecboot! ;) May 27 10:47:54 Jay7: goal is to boot linux without winmo/haret May 27 10:48:14 I got it on h1940 and rx1950 :) May 27 10:48:37 anarsoul: is it easy to flash? May 27 10:48:45 or there is some special HW needed? May 27 10:49:07 Jay7: boot to linux, flash with nanderase/nandwrite May 27 10:49:16 ah, nice May 27 10:49:18 no hw required except device itself May 27 10:49:42 or boot u-boot with haret and do flashing via u-boot May 27 10:50:29 what's wrong with 1st stage bl? May 27 10:50:34 dcordes is working on something similar.. May 27 10:50:43 ah, HTC Dream iirc May 27 10:50:55 ant_work: ? May 27 10:51:03 ant_work: they are booting winmobile :) May 27 10:51:28 Jay7: nope, 1st stage on h1940 and rx1950 boots winmo bootloader May 27 10:51:46 (or u-boot if it's flashed instead of winmo bootloader) May 27 10:51:49 and this winmo fails booting linux? May 27 10:52:18 or is just hardcoded to read from specific flasj address? May 27 10:52:37 ant_work: winmo (winmobile) bootloader can't boot linux May 27 10:52:43 I see May 27 10:52:57 still haret May 27 10:53:14 ant_work: I prefer u-boot :) May 27 10:53:39 and then it's possible use internal flash for rootfs :P May 27 11:04:28 sure, once linux boots you can have rootfs everywhere May 27 11:07:53 Jay7: pong May 27 11:08:05 Jay7: htcleo aka htc hd2 May 27 11:08:14 ah, so :) May 27 11:09:14 Jay7: we have two direct bootloaders now for htcleo and other devices with similar chipsets May 27 11:09:30 Jay7: magldr (closed source project) and lk_bootloader supplied by qualcomm May 27 11:09:55 I fear this won't help much for h1940 rx1950 as it's a totally different plat May 27 11:12:59 anarsoul: see http://htc-linux.org/wiki/index.php?title=LK_Bootloader it is for the qsd8* and msm7* platforms.. it would be nice to boot linux directly on h1940 etc. the htc-linux wiki is supposed to replace hh.org for the htc stuff so feel free to thow anything in there. there's not much on the older platforms present yet (unfortunately) May 27 11:13:46 anarsoul: at least they are not present in the device list at http://htc-linux.org/wiki/index.php?title=Devices May 27 11:19:54 msg RP__ hm, good question. I think GPL 2 was basically the default, but I will check. May 27 11:20:06 heh, missing slash May 27 11:21:03 heh :) May 27 11:21:14 hiho May 27 11:22:59 mm, yeah, no license in the tarball, but the .bb file in oe says "GPL" May 27 11:23:13 I would guess either GPL2 or GPL2+ was probably the original intent May 27 11:23:19 pb_: hence the question :) May 27 11:23:41 pb_: I'd also guess GPLv2 was intended but it isn't entirely clear :/ May 27 11:24:08 yeah, that seems to be what we had in most of the rest of gpe May 27 11:24:09 hi woglinde May 27 11:24:20 see for example the source headers from gpe-calendar, which have the standard GPL2+ boilerplate May 27 11:24:37 * Copyright (C) 2001, 2002, 2003, 2006 Philip Blundell May 27 11:24:42 heh, ten years ago May 27 11:24:59 * RP__ noticed the session scripts seemed to be from 2003... May 27 11:25:14 I am fairly relaxed about making a new release of the session scripts with either GPL2+ or MIT licensing stated explicitly if you want. May 27 11:25:37 pb_: I think even if we put a license file into svn, it would clarify things enough May 27 11:25:43 *g* May 27 11:26:18 okay May 27 11:26:27 I will see if I can trick florian into doing that, otherwise I will sort it out myself later on May 27 11:26:37 pb_: thanks! May 27 11:29:42 hi likewise May 27 11:30:02 woglinde: hi! May 27 11:36:13 RP__: hey, I have created a BSP for some board for poky and wonder if sending 'feedback' to the poky list is appropriate? May 27 11:36:39 hi zecke May 27 12:04:26 zecke: I think so :) May 27 12:20:07 dcordes: h1940 already able to boot linux directly via u-boot, here's my fork of openmoko's u-boot May 27 12:20:12 https://github.com/anarsoul/ipaq-s3c24xx-u-boot May 27 12:20:33 and I have one h1940 with u-boot flashed into nand May 27 12:22:14 same for rx1950 (u-boot works from nand), but I keep this one with default bootloader May 27 12:27:43 but I wrote no documentation on reflashing due to lack of time May 27 12:29:31 hey. do bblayers work in bitbake 1.12? May 27 12:34:15 yes May 27 13:03:20 clear May 27 13:03:31 oops, wrong window :) May 27 13:03:52 reset May 27 13:08:07 florian: good job it wasn't the login screen you thought you were typing into ;) May 27 13:08:20 *g* May 27 13:19:33 ouch...base-files is the reign of anarchy... May 27 13:20:02 yeah May 27 13:20:03 investigating about lack of /tmp in tmpfs May 27 13:20:50 and fwiw /media/ram could go from default fstab May 27 13:21:34 and usbfs is deprecated :) May 27 13:21:39 yes yes May 27 13:21:43 clean it up May 27 13:41:08 ok, will need 2 acks May 27 13:41:44 (even though each machines deploys own fstab :p) May 27 13:42:06 or almost May 27 13:42:31 btw remove simpad from the bad guys ;) May 27 14:31:51 bbl May 27 14:39:32 Hi, I deleted my u-boot bin image from tmp/deploy/images/xxx/ and cannot figure out how to just rebuild/package uboot. bitbake u-boot does not seem to do it. May 27 14:40:21 bitbake -c clean u-boot && bitbake u-boot May 27 14:40:33 woglinde: thx May 27 14:42:22 heh, maybe we need a --rebuild the way oe-lite has :) May 27 14:42:49 yeah, that would be handy. particularly in the presence of sstate. May 27 14:43:03 * kergoth nods May 27 14:44:16 if you look at oe-lite, it has a rebuild, a rebuildall, and a version that rebuilds something + its deps excluding native/cross. course, we couldn't go that far without hardcoding knowledge of the type of recipe we're dealing with, but the notion of 1) rebuilding vs running a single task and 2) running any arbitrary requested operation against a recipe and its deps vs just the recipe (thereby avoiding the foo/fooall pattern) would be nice May 27 14:45:06 right May 27 14:56:08 where is the device table "sucked in" ? I see in openembedded/files there are various device tables. Are these required from the distro? May 27 14:59:00 Generally it comes from initscripts.bb. May 27 15:03:19 grep is our friend ;) May 27 15:03:41 is console-image one of the most basic bootable recipies or is there something more trivial yet bootable? May 27 15:03:53 heh, apparently the git urls to git.debian.org changed, need to fix a couple recipes May 27 15:04:16 micro-base-image is fairly minimal, I think it has less stuff in than console-image May 27 15:04:44 is it glibc or uclibc? May 27 15:04:54 you can build it for either, that's a DISTRO choice May 27 15:05:08 ahh ok.. checking it out now thx May 27 15:09:10 https://simple-note.appspot.com/publish/nhc70W - initial thoughts after first pass review of the oe-lite codebase, parser-rewrite branch May 27 15:12:44 interesting May 27 15:15:27 is oe-lite a new initiative? May 27 15:16:47 not actually all that new anymore but yes, it's a sort of oe/bitbake splinter initiative May 27 15:17:35 it does some things better than oe, including a few long-standing wishlist items, but it isn't entirely compatible and it has some limitations of its own May 27 15:19:36 Esben Haabendal is the oe-lite guy. we had a few discussions on the mailing list about merging bits of oe-lite technology into oe, but he doesn't seem to have any interest in anything other than a wholesale adoption of oe-lite's approach. May 27 15:20:11 which is fine, and I can sort of see his point of view, but I think it's too much of a big bang for oe to swallow in one go and we don't seem to have gotten anywhere with it. May 27 15:21:54 I think its a case of differing priorities. we clearly care about compatibility much more than they do, which isn't too surprising May 27 15:22:12 right May 27 15:24:14 is it also bitbake based or does it have its own "bitbake" implementation? May 27 15:24:16 I'd certainly like to take their PLY based parsing, revert the changes to the file format, and drop it into bitbake and see how it goes May 27 15:24:35 guyvdb_: it's based on bitbake, but they've run with it, rewriting parts on their own May 27 15:24:37 ok there is my answer May 27 15:24:38 :) May 27 15:26:37 it's certainly interesting. at the very least worth monitoring. it can act as a sort of testbed for prototyping possible future bitbake changes. they can blaze a trail, we can follow more gradually, if we choose to, or go a different directione n tirely May 27 15:26:53 and the recipies... does it use oe recipies or are they not compatiable formats with each other? May 27 15:27:12 i think the format is the same rihgtn ow, but the prototype parser rewrite is moving toward a slightly different format May 27 15:27:26 e.g. anonymous python is going away fo rthem in favor of 'hooks', which lets them exert control over evaluation order May 27 15:27:33 e.g. this snippet runs before or after that one May 27 15:27:52 I see May 27 15:27:56 which is interesting, considering some of the issues we've run into in the evaluation order area May 27 15:28:07 they've also started moving toward types in the parser May 27 15:28:13 strings must be quoted May 27 15:28:18 non-strings can be of particular type May 27 15:28:25 e.g. True/False, or an integer May 27 15:28:38 more like a regular language, in that way. i wonder if they're going to allow variable references that way May 27 15:28:41 could be interesting May 27 15:28:44 e.g. FOO = BAR + "baz" May 27 15:29:07 Interesting... my python is not that good... more of a ruby guy.. but it makes sense May 27 15:30:35 personally, though, I'd want to add support for a new, entirely different format, at some point, rather than trying to hack this one into shape. but who knows when that'd happen :) May 27 15:31:08 the recipies are really a DSL ... have you looked much at rake or rspec which are ruby DSLs ... a format in that vein could be cool May 27 15:32:33 well, part of hte reason we went directly parsed the way we are is that we wanted it to be easy to manipulate the format later on. my only concern with going with an internal dsl like that would be, if we start taking advantage of, say, ruby language capabilities, it would be difficult to change formats at any point in the future May 27 15:32:48 of course, you could more tightly control what capabilities you use, and avoid the issue May 27 15:33:34 have any of you guys looked at the Nix package manager? May 27 15:33:38 really interesting May 27 15:35:20 Also creating an external DSL with something like ANTLR could be a way May 27 15:46:25 guyvdb_: well, that's just one of many ways of creating a parser. we have a handcrafted one, oe-lite has a PLY-based one May 27 15:46:36 i did a pyparsing one as well May 27 15:46:47 how do i modify my .config kernel config file within the oe workflow? May 27 15:47:13 bitbake -c menuconfig virtual/kernel; bitbake -c compile -f virtual/kernel May 27 15:47:17 I see May 27 15:47:23 assuming the change is temporary, as that only changes it within the source tree May 27 15:47:43 to make it permanent you'd have to copy the new .config out of ${S} to the defconfig in the metadata repository, or a local layer May 27 15:47:47 yea understand... that will work for now thx May 27 15:47:50 np May 27 15:48:54 if i have defconfig in a local layer, it will override any other defconfig? May 27 15:48:57 ooh, hey, my angstrom oe-core build actually succeeded. thats my first oe-core-based build to successfully complete in a long time May 27 15:49:15 heh May 27 15:49:16 very good May 27 15:49:16 guyvdb_: if the local layer is earlier in the bbpath than the others May 27 15:49:28 pb_: how's micro doing? May 27 15:49:51 yeah, doing fairly well. still working on oe-core-izing a few bits of functionality from oe that it relied on May 27 15:50:01 if I add task with addtask and without "before" parameter then this task is called only on user request, right? May 27 15:50:02 I can build working images now but they have a bunch of unwanted cruft in May 27 15:50:09 update-rc.d, update-passwd, stuff like that May 27 15:51:10 hii May 27 15:51:35 i am trying to compile newer version of icewm in openembedded May 27 15:51:52 I need it because savedefconfig is not available in older kernels and user calls it manually with -c savedefconfig anyway https://gitorious.org/shr/shr-experimental/commit/ac864e16230e3d6a92d465d4729da1b0a5a5e474 May 27 15:52:25 JaMa: yes. both the after and before clauses are optional. they're just mechanisms to add dependencies May 27 15:53:00 ok, I'll push it then May 27 15:57:05 03Martin Jansa  07master * r03578a4aa3 10openembedded.git/recipes/linux/ (4 files in 3 dirs): May 27 15:57:05 linux-openmoko-2.6.3[79]: fix nand with badblocks May 27 15:57:05 Signed-off-by: Martin Jansa May 27 15:57:16 03Benjamin Deering  07master * r4c746a073c 10openembedded.git/recipes/xorg-xserver/xserver-xorg-1.10.1.901/randr-support.patch: May 27 15:57:16 xserver-xorg-1.10.1.901: fix screen rotation May 27 15:57:16 * This should fix screen rotation for 2 and 3 orientation May 27 15:57:16 * this is still broken in 1.9.4 May 27 15:57:17 Signed-off-by: Martin Jansa May 27 15:57:18 03Martin Jansa  07master * r480be229f1 10openembedded.git/recipes/xorg-xserver/ (4 files in 2 dirs): May 27 15:57:18 xserver-xorg: upgrade from 1.10.1.901 to 1.10.1.902 May 27 15:57:18 Signed-off-by: Martin Jansa May 27 15:57:19 03Martin Jansa  07master * r8408a25f08 10openembedded.git/recipes/linux/linux.inc: May 27 15:57:19 linux: add new kernel task savedefconfig May 27 15:57:20 * like Andrea did for linux-kexecboot.inc in 3e4c19efcdac75f7c3af575e9be1f9ef75c4efa9 May 27 15:57:20 * because there is no "before build" param in addtask it's not problem May 27 15:57:20 when kernel doesn't support savedefconfig target yet and user has to May 27 15:57:21 call it manually with -c savedefconfig May 27 15:57:21 Signed-off-by: Martin Jansa May 27 15:57:46 load 13 and it's friday.. :/ May 27 16:05:37 there is some compilation issues with icewm package May 27 16:05:48 /usr/bin/ld: note: 'lround@@GLIBC_2.1' is defined in DSO /lib/libm.so.6 so try adding it to the linker command line May 27 16:06:16 trying to build the newer version of icewm May 27 16:25:31 03Lianhao Lu  07master * r4d0e79e559 10bitbake.git/ (4 files in 2 dirs): May 27 16:25:31 Add PR service deamon to bitbake May 27 16:25:31 Added the initial implementation of the server side PR service. May 27 16:25:31 Signed-off-by: Lianhao Lu May 27 16:25:31 Signed-off-by: Richard Purdie May 27 16:25:36 03Richard Purdie  07master * rece05edd7c 10bitbake.git/lib/bb/cooker.py: (log message trimmed) May 27 16:25:36 cooker.py: Don't show spurious warnings for collections of .bbappend files May 27 16:25:36 Seeing warnings like: May 27 16:25:36 WARNING: No bb files matched BBFILE_PATTERN_yocto '^/xxx/meta-yocto/' May 27 16:25:37 are not encouraging to users and we shouldn't show these if we found May 27 16:25:37 .bbappend files (but no .bb files). This change stops these warnings May 27 16:25:38 from appearing. May 27 16:25:38 03Yu Ke  07master * re912122706 10bitbake.git/lib/bb/fetch2/git.py: May 27 16:25:39 fetch2/git: change default protocol from rsync to git May 27 16:25:40 Signed-off-by: Yu Ke May 27 16:25:40 Signed-off-by: Richard Purdie May 27 16:25:41 03Yu Ke  07master * rd424ecd751 10bitbake.git/lib/bb/fetch2/git.py: May 27 16:25:41 fetch2/git: add document for git fetcher supported options May 27 16:25:42 Signed-off-by: Yu Ke May 27 16:25:42 Signed-off-by: Richard Purdie May 27 16:25:43 03Yu Ke  07master * rbd51659f5e 10bitbake.git/lib/bb/fetch2/git.py: (log message trimmed) May 27 16:26:02 03Yu Ke  07master * rc0eb89054a 10bitbake.git/lib/bb/fetch2/git.py: (log message trimmed) May 27 16:26:02 fetch2/git: enhance the unpack by using "git checkout" May 27 16:26:02 current git fetcher unpack method only checkout index and working tree, May 27 16:26:02 but did not did not update the git branch in ref/heads, so user may not May 27 16:26:03 get right info in ${S} by using git. May 27 16:26:03 this patch enhance the unpack by using git checkout to fix this issue. May 27 16:26:04 Fix bug [YOCTO #1089] May 27 16:26:04 (36 lines omitted) May 27 16:38:59 who do I talk to about getting a login for the oe wiki? May 27 16:39:21 need to update OpenEmbedded_Developers May 27 16:49:22 ? May 27 16:49:31 dvhart you can make an account yourself May 27 16:49:44 but proably you cannt edit all sites May 27 16:49:46 *g* May 27 16:49:58 woglinde, hrm ok, didn't find a create account with the login page May 27 16:50:01 woglinde, looking again May 27 16:51:04 http://www.openembedded.org/index.php?title=Special:UserLogin&returnto=Main_Page May 27 16:51:53 woglinde, that provides a username/password/and remember form with log in and email new password May 27 16:51:57 woglinde, no create account May 27 16:52:11 and entering one in that doesn't exist, just says "there is no user by the name of..." May 27 16:52:18 am I missing something? May 27 16:52:41 oh sorry May 27 16:52:42 hm May 27 16:52:51 ;) May 27 16:53:34 dvhart: last I checked the OE wiki account creation stuff was broken, possibly deliberately due to spammers May 27 16:53:41 you'd have to check with ka6sox-away May 27 16:53:48 okydoky then May 27 16:54:03 yes May 27 16:54:08 somewhat annoying, I would also like to know what the long time solution to this is May 27 16:54:23 s/time/term May 27 16:54:30 I have a similar problem with rt.wiki.kernel.org, had clean up some spam late last night May 27 16:54:32 irritating May 27 16:54:37 so I can sympathize May 27 16:55:11 afk May 27 17:05:46 bluelightning: yes its deliberately disabled to create accounts automatically in OE wiki May 27 17:06:02 and captcha is introduced for editing May 27 17:06:07 blame spammers May 27 17:06:21 * khem tried ff 5beta May 27 17:06:40 pb_: ping May 27 17:17:22 03Darren Hart  07master * r10f3ca52dc 10bitbake.git/lib/bb/fetch2/git.py: May 27 17:17:22 fetch2/git: ensure network check log matches actual command May 27 17:17:22 The git command string logged via check_network_access() does not match May 27 17:17:22 the actual command executed in a few places. Ensure that it does. May 27 17:17:22 Signed-off-by: Darren Hart May 27 17:17:23 Signed-off-by: Richard Purdie May 27 17:17:32 03Darren Hart  07master * r9a28f7744e 10bitbake.git/lib/bb/fetch2/git.py: May 27 17:17:32 fetch2/git: use logging.debug() and clarify messages May 27 17:17:32 Replace a call to print() with logging.debug() and flesh out the May 27 17:17:32 message to clarify the state being reported. May 27 17:17:33 Signed-off-by: Darren Hart May 27 17:17:33 Signed-off-by: Richard Purdie May 27 17:18:48 you know what would be cool May 27 17:18:58 if PV was an equinox P2 omni version May 27 17:19:16 entirely self describing, you can fit any versioning scheme into it, rather than being locked into the debian-like versioning mechanism May 27 17:19:32 course, we'd still need to translate into the binary package manager's scheme, but.. May 27 17:39:35 RP__: I totally confused myself with uclibc patches and uclibc being in two layers bit me in the ... I have sent a correction patch May 27 17:39:42 please consider May 27 17:40:13 khem? koen delete uclibc from meta-oe May 27 17:40:59 yes its gone now May 27 17:43:00 khem what about the two patches I commited to old master? May 27 17:45:54 03Tom Rini  07master * r2549167478 10openembedded.git/recipes/sysvinit/ (2 files in 2 dirs): May 27 17:45:54 sysvinit: Upstream patch for newer libcs May 27 17:45:54 ENOIOCTLCMD isn't defined in newer libcs. The fix here comes from Ubuntu May 27 17:45:54 which in turn comes from upstream. May 27 17:45:54 Signed-off-by: Tom Rini May 27 17:52:09 woglinde: which ones May 27 17:53:28 RTLD no May 27 17:53:36 and PATH $ORIGIN May 27 18:04:12 RP__: posted my impressions May 27 19:10:29 kergoth: around ? May 27 19:31:35 khem: somewhat May 27 19:31:37 whats up May 27 19:50:53 nitink: hi May 27 19:57:20 pb_: hi May 27 19:57:40 pb_:your patch for gcc 4.6.0 armv7 issue is working May 27 19:58:12 pb_: And also it is a worthy candidate for upstream gcc May 27 20:02:31 nitink: check ML, khem replied that it was already discussed upstream May 27 20:06:35 JaMa|Off:thanks , seeing the discussion on the ML now May 27 20:33:00 yeah, it sounds like it is not quite the right patch for upstream May 27 20:33:39 I think it is fine as a workaround in oe though for the time being. it won't generate quite the optimal code, but it is still going to be better than any of the alternatives: -O1 or -march=armv5 will give worse code, and an ice is clearly no good at all May 27 20:38:52 hi, May 27 20:38:56 how do udev compat work? May 27 20:39:21 I need udev 165 maximum for the bug device(bug.conf, a machine) May 27 20:39:39 ? May 27 20:39:46 why? May 27 20:42:04 because of the kernel version: 2.6.27 May 27 20:42:08 it blocks boot May 27 20:42:12 basically udev runs May 27 20:42:19 and then the device nodes aren't created May 27 20:42:31 which block the serial console at the end of the boot May 27 20:42:53 basically : May 27 20:43:26 udev 165(which memleaks as you say) : boot successfull with 2.6.27 May 27 20:43:47 udev 168: boot fails with 2.6.27 and succeed with mainline kernel May 27 20:44:03 mainline has only serial and microsd patch is currently beeing reviewed May 27 20:45:19 hm May 27 20:45:27 looks port the kernel May 27 20:45:30 at least to .32 May 27 20:47:10 okay, my personal oe-core setup template now has a convenience makefile like slugos & co May 27 20:47:11 hehe May 27 20:49:29 woglinde, that's long May 27 20:49:44 I don't want to port the kernel to 32 May 27 20:49:58 there is mainline but it has nearly nothing as supported May 27 20:50:05 there is an udev compat thing May 27 20:50:18 but I've no idea on how to add that for my machine May 27 20:50:26 kergoth: heh, very good May 27 20:50:29 and I need udev, not mdev May 27 20:50:39 since there are rules for that machine May 27 20:53:15 I was skeptical, but i can see the desire to avoid screwing with your current shell's environment. of course, its easier if you don't need a bitbake copy in your project area -- most probably don't, but i do May 27 20:53:16 heh May 27 20:53:55 yeah, I keep ping-ponging between having a local bitbake and using an installed one, according to how out of date the one in debian is at any given moment :-} May 27 20:54:00 hehe May 27 20:54:12 pb_: I'd like to see it on pypi May 27 20:54:18 pip install --user bitbake would be lovely May 27 20:56:11 it's rather annoying that you can't easily source the setup script with an oe-core setup and then source one from another, due to the way the wrapper script runs the main bitbake. ends up calling itself May 27 20:56:56 yeah May 27 20:57:30 way way too easy to source it twice, forgetting which shells had it loaded already. guess i could make sure they all mangle the PS1 May 27 20:57:30 heh May 27 20:58:08 * JaMa|Off is mangling PS1 also with MACHINE switch :) May 27 20:58:23 :) May 27 20:58:36 right, shame you can't represent skull and crossbones in ascii May 27 20:58:38 and DISTRO :) May 27 20:58:54 * incandescant has skull and cross bones when his git tree is dirty May 27 20:59:12 joshual@vorpal:~/Projects/Yocto/bitbake[xmlrpc☠] May 27 20:59:34 heh, that's cute May 27 20:59:35 cute May 27 21:00:29 it's OK on the SSD but in larger git trees on a spinny disk the PS1 script can take some time May 27 21:07:30 anarsoul: awesome! I will fill that into htc-linux.org wiki May 27 21:07:59 dcordes, hi May 27 21:08:01 but those devices are not htc, right? :) May 27 21:09:29 anarsoul: iirc most of the old ipaq pdas are htc manufactured May 27 21:09:39 GNUtoo: hi May 27 21:11:15 cotulla is still working on the driver, he asked me some questions on alsa, but he hasn't finished it yet May 27 21:11:46 for the htc dream I've added some DPI in oe May 27 21:11:57 wifi still have no PSM May 27 21:12:07 and mplayer has issues with the .mmap alsa driver May 27 21:12:12 but gstreamer work fine May 27 21:16:41 GNUtoo: roger that May 27 21:23:59 03Tom Rini  07master * re28053063e 10openembedded.git/recipes/openssl/openssl.inc: May 27 21:23:59 openssl.inc: Fix PowerPC64 builds May 27 21:23:59 No INC_PR bump, no PowerPC64 configs here yet. May 27 21:23:59 Signed-off-by: Tom Rini May 27 21:34:36 Can someone with bitbake commit rights to push the fixes of server sent to bitbake-devel ml? oe-core is broken without those May 27 22:01:24 otavio: it isn't quite as simple as that :/ May 27 22:06:24 03Joshua Lock  07master * r0ac48f913a 10bitbake.git/lib/bb/ui/uievent.py: (log message trimmed) May 27 22:06:24 uievent: fix queueing of events for xmlrpc before UI has loaded May 27 22:06:24 The change to Queue up events before the UI is spawned broke the xmlrpc May 27 22:06:24 server because the uievent implementation of BBUIEventQueue expects pickled May 27 22:06:24 strings for its queue_event() method. May 27 22:06:24 This is because the RPC exposed event.send() method must accept pickled May 27 22:06:25 strings, but for xmlrpc event.send() is just mapped to queue_event(). Work May 27 22:06:26 03Joshua Lock  07master * r6c412f009e 10bitbake.git/lib/bb/server/xmlrpc.py: May 27 22:06:26 server: add updated/fixed xmlrpc server from Poky May 27 22:06:27 Signed-off-by: Joshua Lock May 27 22:18:57 03Tom Rini  07master * r946ce467a6 10openembedded.git/recipes/boa/boa_0.94.13.bb: May 27 22:18:57 boa: Add bison-native to DEPENDS May 27 22:18:57 Won't build without this, so no PR change. May 27 22:18:57 Signed-off-by: Tom Rini May 27 22:40:29 03Tom Rini  07master * r802ee3a680 10openembedded.git/recipes/boa/boa_0.94.13.bb: May 27 22:40:29 boa: Also flex-native May 27 22:40:29 Signed-off-by: Tom Rini May 27 23:58:44 03Andrea Adami  07org.openembedded.dev * r32e15bf046 10openembedded.git/recipes/linux/linux-kexecboot-2.6.39/ (5 files in 5 dirs): May 27 23:58:44 linux-kexecboot: add remaining 2.6.39 defconfigs for Zaurus devices. May 27 23:58:44 Signed-off-by: Andrea Adami May 27 23:58:48 03Andrea Adami  07org.openembedded.dev * r90c0288366 10openembedded.git/recipes/linux/linux/ (6 files in 6 dirs): May 27 23:58:48 linux: : sync Zaurus _git defconfigs with 2.6.39 ones. May 27 23:58:48 Signed-off-by: Andrea Adami May 27 23:58:56 03Andrea Adami  07org.openembedded.dev * r4e47f09065 10openembedded.git/recipes/linux/linux-kexecboot/ (6 files in 6 dirs): May 27 23:58:56 linux-kexecboot: sync Zaurus _git defconfigs with 2.6.39 ones. May 27 23:58:56 Signed-off-by: Andrea Adami May 27 23:58:57 03Andrea Adami  07org.openembedded.dev * r2eb9da04b8 10openembedded.git/recipes/linux/linux-2.6.39/ (6 files in 6 dirs): May 27 23:58:57 linux: add remaining 2.6.39 defconfigs for Zaurus devices May 27 23:58:57 * minor updates to spitz defconfig May 27 23:58:58 Signed-off-by: Andrea Adami May 28 00:06:56 03Andrea Adami  07org.openembedded.dev * r84867d40f6 10openembedded.git/recipes/linux/linux-kexecboot_2.6.39.bb: May 28 00:06:56 linux-kexecboot: set DEFAULT_PREFERENCE = "1" for 2.6.39 on Zaurus. May 28 00:06:56 Signed-off-by: Andrea Adami May 28 00:07:06 03Andrea Adami  07org.openembedded.dev * rd452edb9ad 10openembedded.git/recipes/linux/linux_2.6.39.bb: May 28 00:07:06 linux: set DEFAULT_PREFERENCE = "1" for 2.6.39 on Zaurus. May 28 00:07:06 Signed-off-by: Andrea Adami **** ENDING LOGGING AT Sat May 28 02:59:57 2011