**** BEGIN LOGGING AT Tue Mar 25 02:59:56 2008 Mar 25 03:17:43 03xora 07org.oe.dev * rbfbb7a4b... 10/ (4 files in 3 dirs): Mar 25 03:17:43 gtkhtml-3.8_3.12.3.bb : update gtkhtml with patch stolen from debian for Mar 25 03:17:43 glib API changes Mar 25 03:31:52 03rwhitby 07org.oe.dev * r089057d0... 10/ (8 files in 5 dirs): linux-ixp4xx: Updated to 2.6.24.4 Mar 25 04:17:04 03rwhitby 07org.oe.dev * r95ddd65a... 10/ (1 classes/rootfs_ipk.bbclass): rootfs_ipk.bbclass: symlink /usr/lib/opkg to /usr/lib/ipkg too Mar 25 05:14:13 !logs Mar 25 05:14:13 Channel logs for #oe are archived at: Mar 25 05:14:14 http://hentges.net/tmp/logs/irc/%23oe Mar 25 05:14:15 Live-logs are available at Mar 25 05:14:16 http://hentges.net/tmp/logs/irc/livelogs/%23oe.livelog Mar 25 05:14:18 See ?? help-logs for usage instructions Mar 25 05:55:13 hi all Mar 25 05:56:11 I'm trying to build nano as a simple test for the openembedded system however it is borking trying to build qemu-native-0.9.1 Mar 25 05:57:07 it seems to be having trouble with make: *** [subdir-ppc-linux-user] Error 2 Mar 25 05:57:26 however I don't have a ppc linux so why is it trying to build that subdir? Mar 25 05:57:38 Is there a config option I have forgotten? Mar 25 06:26:50 03rodrigo.vivi 07org.oe.dev * r6ec6766b... 10/ (8 files in 5 dirs): Mar 25 06:26:50 Adding Usbnet package. Mar 25 06:26:50 Usbnet is a package used by Mamona to setup network through usb. Mar 25 06:26:54 03rodrigo.vivi 07org.oe.dev * r77b913cd... 10/ (6 files in 3 dirs): Mar 25 06:26:54 Adding mamona-sound-n800 package. Mar 25 06:26:54 Used by mamona to configure sound system on nokia N8*0 using dsp. Mar 25 06:26:59 03rodrigo.vivi 07org.oe.dev * rebc2d786... 10/ (1 conf/distro/mamona.conf): Mar 25 06:26:59 Updating Mamona distro conf file. Mar 25 06:26:59 BAD HISTORY: This conf file has changed a lot since last commit. Mar 25 06:27:01 Now on Mamona team will maintain it and commit every changes. Mar 25 06:27:04 03rodrigo.vivi 07org.oe.dev * r6e883229... 10/ (11 files in 3 dirs): Adding tasks and meta packages for Mamona. Mar 25 06:41:28 03rodrigo.vivi 07org.oe.dev * r187b455c... 10/ (8 files in 2 dirs): Mar 25 06:41:28 Adding noemu packages for Mamona. Mar 25 06:41:28 Noemu packages are native packages made to install on a qemu environment as a target package avoiding some unecessary emulations. Mar 25 06:41:28 http://dev.openbossa.org/trac/mamona/wiki/noemu Mar 25 06:59:37 it seemed that bitbake -c rebuild qemu-native fixed it Mar 25 07:00:29 can i get to nano, here's hoping crosses fingers, I like the way that oe downloads the kernel-headers, very good! Mar 25 07:03:08 hi rob_w|laptop Mar 25 07:03:20 hi Mar 25 07:03:31 are you geoffrey ? Mar 25 07:03:41 nope Daryl Mar 25 07:03:46 oh sorry Mar 25 07:04:19 03rodrigo.vivi 07org.oe.dev * r55b6b2c1... 10/ (4 files in 2 dirs): Mar 25 07:04:19 bug fix on noemu packages. Mar 25 07:04:19 These packages was recently moved to packages/mamona but it was missing a change on include paths. Mar 25 07:04:29 s'ok I have been starting to use oe again after a few months, has taken me a while to upgrade everything and get stuff compiling again. Mar 25 07:04:59 i'm looking to get hold of my own copy of dillo and ipatience, I can't find them as pre packaged angstrom anywhere Mar 25 07:05:07 oe is quite bad with me also these days Mar 25 07:05:42 I got some kind of weird ppc error compiling qemu but doing a -c rebuild worked, go figure Mar 25 07:06:50 rob_w|laptop, like the old joke about software engireer fixing a car problem, "everyone get out of the car and then back in again and restart the car" lol Mar 25 07:07:06 hehe Mar 25 07:29:36 someone please merge Mar 25 07:31:29 automerge should merge in a minute Mar 25 07:31:47 * rob_w|laptop waits Mar 25 07:32:00 XorA: How's life in taiwan? Mar 25 07:32:04 hot Mar 25 07:32:18 XorA: uhggg Mar 25 07:32:36 did buy eeePC though Mar 25 07:35:40 XorA: looks kinda fun... Mar 25 07:39:01 * ScaredyCat throws snowblls at XorA Mar 25 07:39:12 snowballs too Mar 25 07:45:33 night all Mar 25 07:49:38 morning Mar 25 07:52:56 yo hrw Mar 25 08:00:09 elo Greame Mar 25 08:01:15 * hrw -> re-reading long thread on oe-private Mar 25 08:07:34 hrw XorA : hi ! Mar 25 08:08:01 yo steliosk Mar 25 08:08:36 * steliosk is happy as today its a holiday here Mar 25 08:09:31 I was happy that Good Friday is free day in UK :D Mar 25 08:10:55 heh Mar 25 08:10:58 * XorA wasnt Mar 25 08:11:27 oh well Easter is after a couple of weeks for us Mar 25 08:11:45 XorA: it is working day in Poland so we were able to finish some stuff which required driving though city Mar 25 08:11:46 today its our natial liberation day Mar 25 08:12:05 steliosk: greek church use dates from ortodox church? Mar 25 08:12:23 hrw : yep. Mar 25 08:12:45 hrw: I didnt get the holiday, I was in .tw Mar 25 08:12:46 your churches looks similiar to ortodox ones Mar 25 08:13:11 XorA: with OM work you get other days Mar 25 08:13:21 hrw : we are orthodox (or lets say most are orthodox 98% or something) Mar 25 08:13:37 heh, I shall take Chinese new year next year :-) Mar 25 08:13:44 heh Mar 25 08:13:55 Greek churches rock Mar 25 08:14:05 * XorA isnt christian, but they are cool as hell buildings Mar 25 08:14:56 church cool as hell building... contradiction in terms ? :) like military intelligence Mar 25 08:15:12 steliosk: the ones I was in take art seriously Mar 25 08:15:26 steliosk: I care little for the religious bit though Mar 25 08:15:53 steliosk: ok Mar 25 08:16:22 steliosk: I am used to 'ortodox == prawosławie (russian church)' Mar 25 08:16:25 XorA : yes. usually ti takes a lot of years to finish a church Mar 25 08:17:14 hrw : probably because its closer to Poland Mar 25 08:18:31 steliosk: East part of Poland has lot of ortodox Mar 25 08:19:52 time to do some Dremel work :) Mar 25 08:20:14 brb Mar 25 08:43:11 blast I'm stuck again in my quest to build nano configure: error: GMP 4.1 and MPFR 2.2.1 or newer versions required by fortran Mar 25 08:43:30 it get's stuck on buildng gcc-cross Mar 25 08:49:07 03jeremy_laine 07org.oe.dev * r5d85339a... 10/ (5 files in 3 dirs): opkg: fix md5 hashing on big-endian machines Mar 25 08:50:04 * * OE Bug 4118 has been created by xjqian(AT)gmail.com Mar 25 08:50:06 * * udev-118 failed to create device nodes on Tosa Mar 25 08:50:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4118 Mar 25 08:59:28 configure: error: GMP 4.1 and MPFR 2.2.1 or newer versions required by fortran Mar 25 09:15:43 03koen 07org.oe.dev * rabbe1af3... 10/ (1 packages/mamona/gcc-noemu_4.1.1.bb): gcc-noemu: attempt to fix requires Mar 25 09:19:57 bonjour Mar 25 09:36:05 * * OE Bug 4119 has been created by dv(AT)vollmann.ch Mar 25 09:36:07 * * Missing patch for binutils 2.16.1 Mar 25 09:36:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4119 Mar 25 09:53:04 * * OE Bug 4059 has been RESOLVED (FIXED) by xjqian(AT)gmail.com Mar 25 09:53:06 * * db-native_4.3.29 fetch failure Mar 25 09:53:08 * * http://bugs.openembedded.org/show_bug.cgi?id=4059 Mar 25 09:53:34 03xjqian 07org.oe.angstrom-2007.12-stable * r98f112eb... 10/ (8 files in 2 dirs): (log message trimmed) Mar 25 09:53:34 applied changes from 69b3066d0bd612829be7ae8c9e78a3501e5cdb3c Mar 25 09:53:34 through a6d2d41d484ef7827260add2c94d270077592a01 Mar 25 09:53:34 applied changes from f460277636d213d6a3257745ceaac42b26814877 Mar 25 09:53:34 through 578f199bc6b531a09543f38044f98199ef5f49ad Mar 25 09:53:37 applied changes from e60db6ca0d9d6a0b3b02507a6068c2d803af4d6c Mar 25 09:53:39 through 848c6e3f16104e7b87568edf06d0d9cc79b5e860 Mar 25 09:53:41 03xjqian 07org.oe.angstrom-2007.12-stable * r1aec21ed... 10/ (1 BACKPORTS.txt): BACKPORT: db fixes Mar 25 10:05:04 * * OE Bug 4120 has been created by  Mar 25 10:05:06 * * inetutils-1.4.2-autobuild Mar 25 10:05:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4120 Mar 25 10:13:05 * * OE Bug 4121 has been created by  Mar 25 10:13:06 * * inetutils lacks u-a entries Mar 25 10:13:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4121 Mar 25 10:26:21 03koen 07org.oe.dev * r557257ed... 10/ (1 packages/inetutils/inetutils_1.5.bb): inetutils: update to 1.5 Mar 25 11:06:04 * * OE Bug 4122 has been created by  Mar 25 11:06:06 * * e-wm-0.16.999.041-autobuild Mar 25 11:06:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4122 Mar 25 11:34:48 huhu Mar 25 11:37:33 RP, which of your patches are still needed for usb gadget support for the 2.6.25 kernel and what has made it mainstream? Mar 25 11:45:33 I am working on building/creating an OE distribution for our next gen HDTV hardware. Would this be the correct resource channel? Mar 25 11:46:04 yes Mar 25 11:46:16 excellent. Glad to meet you all :) Mar 25 11:47:13 crweb: just curious, anything already out? asking because I was using Micronas Decypher stuff but as you know they closed down :> Mar 25 11:47:38 svolpe_gerrath: none of them have made mainline Mar 25 11:47:43 The development hardware based on TI Davinci is due out in the next 2 months Mar 25 11:48:21 We have a very small team also working on porting our current generation hardware over to OE (Neuros OSD 1.0) Mar 25 11:48:39 and by small team i mean, I tinker in my free time. Mar 25 11:48:49 I have a DAvinci board around, just did not have the time to try it yet :P Mar 25 11:49:11 I think I built an OE image for it but I had to reflash uboot and just did not get to actually doing that Mar 25 11:49:14 afk, meeting Mar 25 11:49:16 However, our Davinci hardware will be completely OE based Mar 25 11:50:01 Jin^eLD: very interesting. I'm just now starting to look at OE, so hopefully we might beable to get you some real use for your davinci board ;) Mar 25 11:57:20 svolpe_gerrath: have you been following this UDC thread: http://lists.arm.linux.org.uk/lurker/thread/20080317.034947.a1c57ba4.en.html#20080317.034947.a1c57ba4 Mar 25 12:03:51 cbrake, no, thanks for the link Mar 25 12:09:20 hi any collie user here? Mar 25 12:12:37 woof Mar 25 12:13:39 re Mar 25 12:14:09 crweb: well, I curently do not have a project that involves the board so I was mainly looking at it in my spare time Mar 25 12:14:45 Jin^eLD: obviously if you have a build already you are much further than me... any tips? Mar 25 12:14:46 the uboot on my board was just too old, I think that was the only reason why I could not get it going Mar 25 12:15:33 crweb: I think I just built angstrom for davinci, there is a machine conf for it already so you can simply select angstrom/davinci and try a build Mar 25 12:16:02 would it be better to use angstrom? or build our own distro? Mar 25 12:16:10 we have a lot of our own custom apps Mar 25 12:16:29 well, angstrom is a maintained distro, so even if you roll your own I think it's a good idea to start off it Mar 25 12:16:38 that's what I usually do Mar 25 12:16:55 sec, phone Mar 25 12:19:34 mmm.. no qtopia4-core support :( looks like i might have to create bitbake recipe Mar 25 12:20:33 crweb: talk with Zecke Mar 25 12:21:05 ScaredyCat, you are a collie user? Mar 25 12:25:17 no, sorry bad joke Mar 25 12:29:42 re Mar 25 12:30:04 crweb: you can have your custom apps and keep them separate from the apps that are already in OE Mar 25 12:30:19 you can have a nice setup where youseparate your own development from what's already available Mar 25 12:31:11 very interesting. I'm trying to read the docs, but very much information overload ;) Mar 25 12:32:18 RP: moin Mar 25 12:33:23 RP: copying from usr/lib/evas/enginges/fb/i686-gnu-linux/module.so to usr/lib/evas/loaders/jpeg/i686-gnu-linux/module.so, usr/lib/evas/loaders/png/i686-gnu-linux is certainly not good Mar 25 12:33:49 if you do a md5sum tmp/staging/i686-linux/usr/lib/evas/*/*/*/module.so, it is the same md5sum :) Mar 25 12:33:55 for _every_ module Mar 25 12:34:24 zecke: right, that doesn't make sense. I didn't see that :/ Mar 25 12:34:49 does anybody know the current status of kernel 2.6 on the collie? Mar 25 12:35:13 RP: and the function changed, from just looking at /usr/lib to start to look for every .la file Mar 25 12:35:42 RP: so before it didn't pick up the modules at all, after the change it does Mar 25 12:35:55 zecke: That was intentional Mar 25 12:36:06 RP: right, we want to fix the .la files Mar 25 12:36:27 RP: but we should copy the lib from the dir it is in, this is what -C is for Mar 25 12:37:06 zecke: No, this isn't how that function has ever worked Mar 25 12:37:30 zecke: If it can find it in ${S} and can find the .lai file, it uses that Mar 25 12:37:45 zecke: I don't claim this is right or wrong, just the way its always worked Mar 25 12:39:04 RP: well, see above, you changed how this function always worked :) Mar 25 12:39:31 RP: it didn't care for subdirs before, so ${S} was enough, now it cares for subdirs, now things break miserable Mar 25 12:39:56 zecke: It worked for the examples I looked at with subdirs (gst) Mar 25 12:40:34 RP: the gst modules don't have the same filenames ;) Mar 25 12:41:08 why are not all packages created as ipk `s anymore ? i am missing the whole libs of xserver-kdrive Mar 25 12:41:53 rob_w|laptop: You probably want -c buildall Mar 25 12:42:09 is that new ? Mar 25 12:42:10 RP: I still wonder how oe_libinstall finds module.so at all, when did it start to search for files? the -C is so wrong it should barf out on not finding any module.so Mar 25 12:42:21 rob_w|laptop: yes, ~1 year Mar 25 12:42:28 hmm Mar 25 12:42:52 so i do bitbake xserver-kdrive -c buildall ?? Mar 25 12:45:22 rob_w|laptop: That will package all the dependencies too, yes Mar 25 12:45:30 rob_w|laptop: or set BB_DEFAULT_TASK = "buildall" Mar 25 12:46:03 zecke: The thing is it worked here... Mar 25 12:46:06 in lcoal.conf .. thanks will try that Mar 25 12:46:24 RP: what did work here? staging? do the md5sum on the dirs above, they are all the same Mar 25 12:46:45 RP: if you compile edje-native and then execute edje_cc your machine will halt :) Mar 25 12:46:48 *not funny* Mar 25 12:47:00 zecke: Which files exactly? Mar 25 12:47:06 hey zecke and RP Mar 25 12:47:17 hi mithro Mar 25 12:47:19 md5sum tmp/staging/i686-linux/usr/lib/evas/*/*/*/module.so Mar 25 12:48:00 zecke: ah, so all the files have the same checksum? Mar 25 12:48:03 did any of the Embedded stuff get into GSoC? Mar 25 12:48:30 RP: yes, you pick one module.so and copy it to every dir that contains a .la :) Mar 25 12:48:44 mithro: Openmoko is there Mar 25 12:48:44 zecke: ah is this problem with oe_libinstall causing the edje_cc problem? Mar 25 12:49:14 XorA: segfault on recursion? :) Mar 25 12:49:28 zecke: yes, 3G of ram and die Mar 25 12:49:32 zecke: your working for Openmoko right? Mar 25 12:49:48 mithro: I'm a freelancer :) Mar 25 12:50:03 was going to ask how GSoC was going Mar 25 12:50:10 zecke: ok, the patch you mention did not change the behaviour of oe_libinstall, it changed the behaviour of autotools_stage_all Mar 25 12:50:10 zecke: freelancer working for openmoko? Mar 25 12:50:24 zecke: The autotools function was enhanced to work with subdirectories Mar 25 12:50:48 zecke: it looks like oe_libinstall can't cope when we have libraries with the same name as you summarised though Mar 25 12:50:50 RP: right, the way we use oe_lininstall Mar 25 12:51:20 RP: if we find a .la in a subdir, the .so is supposed to be in the same dir, use -C to go there, copy (to itself) to fix the .la file there Mar 25 12:51:46 XorA: rebuild evas-native after a mtn pull and update :) Mar 25 12:53:04 zecke: not on an eeePC I wont Mar 25 12:53:27 how goes OE? :) Mar 25 12:53:44 You'll not run into any trouble with it using 3G! Mar 25 12:53:48 zecke: Did you commit that patch? Mar 25 12:54:38 mithro: Its still moving forward, packaged staging is nearly a reality :) Mar 25 12:54:53 RP: oh - that sounds spiffy Mar 25 12:55:13 RP: to OM, I think it is sane, it is fixing a serve bug Mar 25 12:55:23 zecke: oe_libinstall does dir=$dir`(cd $dir;find . -name "$dotlai") | sed "s/^\.//;s/\/$dotlai\$//;q" to find the .lai Mar 25 12:55:38 zecke: We need to point into ${S}, not ${STAGE_TEMP} Mar 25 12:55:41 RP: I argue it is correct, I still need to build a whole image Mar 25 12:56:21 RP: this will not make any difference, it will still find a random module.so and copy it everywhere Mar 25 12:56:30 RP: fs's tend to be stable in the way they list a directory Mar 25 12:56:48 zecke: Its rather subtle change in behaviour since it makes libtool use ${STAGE_TEMP} instead of ${S} and it therefore won't find the .lai files Mar 25 12:58:09 zecke: I agree what is there is broken if we have multiple modules with the same name but your patch changes some behaviour in a subtle way you don't acknowledge ;-) Mar 25 12:58:10 RP: but then it won't find the module.so at all? Mar 25 12:58:54 RP: http://pastebin.com/m2c3f953d - any idea how is it possible, that ep93xx-gpio-interrupt-debounce.diff is applied twice? Or how to debug how it's so? Mar 25 12:59:05 RP: I just woke up, but I think copying the .so from where the .la is found is sane :) Mar 25 12:59:38 zecke: That isn't what oe_libinstall has ever done though Mar 25 13:00:52 RP: no, but we don't change oe_libinstall and the -C in oe_libinstall is there to go close to the lib Mar 25 13:01:13 RP: so what my change does is making it more clear where to find the .so Mar 25 13:01:35 be it .libs/ (which doesn't exist after install....) Mar 25 13:02:04 RP: taking the step back, we don't want to call oe_libinstall at all, we want to fix the .la files ;) Mar 25 13:02:38 RP: we know that make install will put the files in the right directories, we should honor that and just fix the la files Mar 25 13:03:28 zecke: Its not as simple as that Mar 25 13:03:39 zecke: Do you understand what oe_libinstall does? Mar 25 13:04:31 RP: at some point i did Mar 25 13:04:58 zecke: I actually agree with what you're saying, we probably shouldn't be using it. The fact is we did and do though. Mar 25 13:06:02 my point is that function does very different things if you pass it ${STAGE_TEMP} instead of ${S} Mar 25 13:06:39 RP: to bring is to the same point. We talk about autotools_stage_all how a recent change triggered a side effect when the .la(i) name is not unique Mar 25 13:07:08 right Mar 25 13:08:19 and let me get breakfast :) Mar 25 13:08:20 zecke: Do you agree oe_libinstall changes behaviour depending on whether you pass it somewhere in ${S} or somewhere in ${STAGE_TEMP} ? Mar 25 13:08:40 RP: Yes because we don't find .lai files below STAGE_TEMP Mar 25 13:08:55 RP: we will not find the .so/.la as well :) Mar 25 13:09:25 not quite. But at least we agree it will change behaviour :) Mar 25 13:10:42 kergoth: Are you awake by any chance? Mar 25 13:11:39 RP: but why do we call oe_libinstall from autotools_stage_all? My assumption is we want to fix .la files after having installed them :) Mar 25 13:12:54 zecke: Honestly, I'm not sure why we call it from there. I suspect its a historical relic but I'm nervous about changing things I don't understand Mar 25 14:09:58 03pfalcon-EbJVxxdVoWb2StjH6ZfwpA 07org.oe.dev * ra97cd6be... 10/ (1 BACKPORTS.txt classes/patch.bbclass): (log message trimmed) Mar 25 14:09:58 patch bbclass: backport Mar 25 14:09:58 Revision: a99dc9b34551aa3889d7057538bf2e37b0471af1 Mar 25 14:09:58 Ancestor: 33c44c7c7ce5fce5da6c869be191126d5fca6f96 Mar 25 14:09:58 Author: pfalcon-EbJVxxdVoWb2StjH6ZfwpA@public.gmane.org Mar 25 14:10:00 Date: 2008-01-28T00:11:15 Mar 25 14:10:02 Branch: org.openembedded.dev Mar 25 14:20:14 RP: I'm back but not for long :) Mar 25 15:09:43 is there an option in opkg.conf to avoid sig download ? Mar 25 15:14:04 * * OE Bug 4123 has been created by gwossum(AT)acm.org Mar 25 15:14:06 * *  support for base-image Mar 25 15:14:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4123 Mar 25 16:42:14 hi all Mar 25 16:42:20 Could not initialize SDL - exiting Mar 25 16:42:31 is someone using qemu 0.9.1 ? Mar 25 16:42:37 I am getting above error Mar 25 16:42:40 0.9.0 works ok Mar 25 17:00:09 * ScaredyCat blames florian Mar 25 17:04:57 ScaredyCat: ! Mar 25 17:05:18 it always works... Mar 25 17:05:35 monotone.oe stops working, I blame you and it works again Mar 25 17:05:53 heh Mar 25 17:32:55 /quit narf Mar 25 17:33:06 *sigh* Mar 25 17:34:14 03woglinde2 07org.oe.dev * read8fe4e... 10/ (3 files in 3 dirs): Mar 25 17:34:14 linux-2.6.24: new rtc-sa1100 fixes for the simpad Mar 25 17:34:14 * new rtc-sa1100 fix Mar 25 17:34:14 * bump pr Mar 25 17:34:14 * fixes bug 4094 Mar 25 17:35:31 :) Mar 25 17:43:04 * * OE Bug 4084 has been RESOLVED (FIXED) by heinold(AT)inf.fu-berlin.de Mar 25 17:43:06 * * Simpad - Angstrom 2007.12 R10 hangs with soft lockup - hwclock Mar 25 17:43:08 * * http://bugs.openembedded.org/show_bug.cgi?id=4084 Mar 25 17:48:05 FYI: still bisecting the E-native breakage Mar 25 17:48:19 5861933efac1910d02cd49fed73a291599ca8779 (2008-03-19 14:06:21) ok Mar 25 17:48:23 0b604857bbf871639fdb43ee8380222e8ef64bb7 (2008-03-19 23:25:13) broken Mar 25 17:48:31 so i guess i will soon have the rev to blame Mar 25 17:48:47 mickeyl: ejde-native failing? Mar 25 17:49:03 yes, edje_cc miscompiling so that it SIGSEGVs Mar 25 17:49:08 i guess it's your fault Mar 25 17:49:11 but i need more data :D Mar 25 17:49:18 mickeyl: zecke found the problem and yes, I was guilty Mar 25 17:49:22 ooh Mar 25 17:49:26 why does no one tell that to me Mar 25 17:49:28 *sigh* Mar 25 17:49:33 mickeyl: http://repo.or.cz/w/org.openembedded.dev.git?a=commit;h=b374d082c249cbf528013e1d86b3f462e7433fcb was at fault Mar 25 17:49:48 only after it's eaten all the ram in the universe Mar 25 17:49:52 mickeyl: We're still discussing the right way to fix it ;-) Mar 25 17:50:02 it's still broken though Mar 25 17:50:20 RP: ok, cool. Mar 25 17:50:27 RP: please keep me posted. Mar 25 17:50:33 the damn thing is... Mar 25 17:50:36 i had heaps of changes Mar 25 17:50:39 during my brazil visit Mar 25 17:50:46 and so of course i thought one of these broke it Mar 25 17:51:04 to make it worse, E upstream did some breakage as well Mar 25 17:51:23 mickeyl: The question is whether autotools_stage_all should be using oe_libinstall Mar 25 17:52:00 mickeyl: zecke says not and I'd tend to agree but not using that is a major change in the way we do things Mar 25 17:53:29 ok, i don't have enough insight into that to comment atm. Mar 25 17:53:47 i will revert your change locally and try to fix the upstream breakage Mar 25 17:53:51 so that we deal with less variables Mar 25 17:53:53 *phew* ;) Mar 25 17:54:13 mickeyl: Its that change combined with a number of others recently. Mar 25 17:54:52 all that makes me think we need two+1 branch Mar 25 17:54:59 or, two+n Mar 25 17:55:35 native packages weren't using autotools_stage_all, koen proposed making some tweaks which I agreed with not realising everything would 'explode'. Nearly everything works again now thanks to some quick fixes except for this issue... Mar 25 17:56:01 Either change wouldn't do this on their own but together it breaks Mar 25 17:56:05 hehe Mar 25 17:56:06 right Mar 25 17:56:11 sometimes these things happen Mar 25 17:56:17 did you happen to read my pkgconfig thingy? Mar 25 17:56:53 a while ago? Mar 25 17:56:58 three days or so Mar 25 17:56:58 or something recent? Mar 25 17:57:02 perhaps four Mar 25 17:57:18 the deal is, i need to rely on autotools pkgconfig staging Mar 25 17:57:26 because pkgconfig.bbclass unconditionally stages _all_ .pc files Mar 25 17:57:30 I don't think I've seen that, was it to the oe list? Mar 25 17:57:50 whereas some software (E for instance) relies on the presence of .pc files for configuration Mar 25 17:58:00 no, i didn't sent it to the list Mar 25 17:58:03 just a comment in a commit Mar 25 17:58:05 let me dig it out Mar 25 17:59:23 http://lists.linuxtogo.org/pipermail/openembedded-commits/2008-March/013669.html Mar 25 18:00:29 mickeyl: In a way I'd say that is as it should be and should be the default for autotools Mar 25 18:00:43 mickeyl: Perhaps autotools should set a var which disables the line in pkgconfig? Mar 25 18:01:03 RP: yes, i'd rather see it solved that yay Mar 25 18:01:10 s/yay/way/ Mar 25 18:01:28 I'm about to be called away for food but basically I agree and its a good candidate for a cleanup Mar 25 18:01:34 back shortly Mar 25 18:01:34 ok, cool Mar 25 18:01:38 bon appetit Mar 25 18:01:48 and don't worry, we'll get everything back on track Mar 25 18:01:55 i actually have a good feeling atm. Mar 25 18:07:45 hi! when i boot my new kernel with Samsung NAND i get `bad eraseblock error` for each block written using Atmel SAM-BA. It seems that its connected with ECC checking, but i switch it off! Mar 25 19:05:25 woglinde_: wb Mar 25 19:05:44 jo Mar 25 19:13:41 cool, a tool forces me to decide in which branch I will add a file to not provoke a NCC :) Mar 25 19:49:04 re flo Mar 25 19:49:56 re Mar 25 19:50:57 morning all Mar 25 19:55:59 gm Mar 25 19:58:29 hi Fique, hi cyrilRomain Mar 25 20:00:11 flo_lap, gtodo's new item screen doesn't fit unless fullscreen on maemo Mar 25 20:00:18 s/gtodo/gpetodo Mar 25 20:00:40 also, I've been adding items and they don't appear on the summary screen Mar 25 20:00:41 and Mar 25 20:00:47 I have no option to turn that screen off Mar 25 20:01:15 also, the status combobox wastes quite a lot of space Mar 25 20:01:30 and the details text-box has no borders, making it impossible to see it Mar 25 20:02:33 pvanhoof: the latter is an open bug against the maemo theme... its open since a time before maemo was opened to the public Mar 25 20:02:50 ok Mar 25 20:03:46 The + button doesn't work for my fingers, the OK button of the new item window dows Mar 25 20:03:57 so if that + button would be bigger, I could more easily create a new item Mar 25 20:04:04 same for the comboboxes Mar 25 20:04:13 although they somewhat work Mar 25 20:04:24 * flo_lap needs more time Mar 25 20:04:26 but for example the due checkbox, again, doesn't work for fingers Mar 25 20:04:41 same for manipulating the items in the treeview Mar 25 20:04:49 make them bigger, like the rows of Modest Mar 25 20:04:58 people do use fingers :) Mar 25 20:05:17 I'd rather have two columns than super small rows Mar 25 20:05:23 pvanhoof: right, the whole thing was designed for much more low resolution displays. Mar 25 20:05:41 pvanhoof: some more ways for structuring items would be useful too Mar 25 20:05:48 okay, that's okay, just make the models generic and quickly create a new ui Mar 25 20:06:09 that's exactly why you do a MVC Mar 25 20:07:31 flo_lap, and it would be nice if the device would take the items from my phone's meetings and calendar Mar 25 20:07:43 and put them there if the phone's bluetooth is connected Mar 25 20:08:08 no need for me to sync device -> phone, just show the items on my phone Mar 25 20:08:19 and if the phone is gone, no need to keep showing them Mar 25 20:08:20 :) Mar 25 20:08:38 just readonly would be sufficient already Mar 25 20:09:14 pvanhoof: right, i'm tempted to start a meta-project for maemo and phone integration but it would be hard to get that sponsored and free time is too limited Mar 25 20:09:25 flo_lap, but it feels to much like wince app .. Mar 25 20:09:32 don't try to cram that much things on one screen Mar 25 20:09:35 just make a wizard-like Mar 25 20:09:48 new item can be four steps, that's fine Mar 25 20:09:57 just make it easy to go to the next step Mar 25 20:10:08 wince developers always get that wrong too Mar 25 20:10:42 pvanhoof: good point... i just ran into the same with the multisync-gui Mar 25 20:10:43 *sigh* bluez-utils forces in the end the build of pyhton-native Mar 25 20:11:01 flo_lap, really, three or four buttons is too much already Mar 25 20:11:06 one screen, one action Mar 25 20:11:17 maybe three related actions Mar 25 20:11:31 but selecting the category, that's a different screen Mar 25 20:11:39 and why we need gst-plgugins-esound Mar 25 20:11:47 cancel, that's just the [X] button Mar 25 20:11:47 pvanhoof: it was easy to make it work in maemo, not much more complicated to look correct but its quite some effort to turn it into a useful application Mar 25 20:11:55 and you suddenly have 40 x 500 pixels more Mar 25 20:12:18 pvanhoof: yes sure... Mar 25 20:12:26 that's room to make the comboboxes larger, finger-compatible Mar 25 20:13:17 yep Mar 25 20:13:54 did i mention that cmake sucks? Mar 25 20:13:57 ;) Mar 25 20:14:04 it can't suck more than autotools Mar 25 20:14:12 ah Mar 25 20:14:19 and make that scrollbar bigger Mar 25 20:14:31 I have to use my nail .. Mar 25 20:15:03 and allow me to see which item I will open, by selecting it Mar 25 20:15:08 in the end all suckz more then autotools Mar 25 20:15:09 and only open at double click, not single click Mar 25 20:15:22 actually, just mimic modest's treeview Mar 25 20:15:48 you can put the priority and the status at the second line, and the summary in bold on the first line Mar 25 20:16:01 flo_lap, you want the code how to do that kind of CellRenderers? Mar 25 20:16:22 https://svn.tinymail.org/svn/tmut/trunk/src/modest-widgets/ Mar 25 20:16:42 https://svn.tinymail.org/svn/tmut/trunk/src/modest-widgets/modest-hbox-cell-renderer.h and https://svn.tinymail.org/svn/tmut/trunk/src/modest-widgets/modest-vbox-cell-renderer.h Mar 25 20:17:03 djcb gave me permission to use them in TMut, although that license doesn't require that permission Mar 25 20:17:16 if you want I ask him whether it's okay for u to use them in gpetodo too Mar 25 20:17:31 the really bad experience with cmake was that it is the same bad mixture of strange magic stuff you can't guess Mar 25 20:18:54 flo_lap: but there are less magic in cmake than in autotools, don't you think ? Mar 25 20:20:17 * cyrilRomain is going to switch to omake for personal project, although I'm quite happy with cmake Mar 25 20:21:15 pvanhoof: doesn't look that complicated... Mar 25 20:21:56 omake ? Mar 25 20:22:04 flo_lap, no it's easy Mar 25 20:22:21 chouimat|work: http://omake.metaprl.org/index.html Mar 25 20:22:46 chouimat|work: it is not well known but has good properties Mar 25 20:22:47 flo_lap, https://svn.tinymail.org/svn/tmut/trunk/src/tmut-folder-view.c at tmut_folder_view_instance_init Mar 25 20:22:54 also look at header_view_compact_header_cell_data Mar 25 20:23:05 I already extracted that code out of modest for you :) Mar 25 20:23:15 cyrilRomain: interesting ... Mar 25 20:23:24 you can build any kind of cell in your treeview that way. You have a hbox and a vbox to crap things into Mar 25 20:23:55 cyrilRomain: well... the advantage of cmake is that it is faster. not sure which involves more magic - but with autotools is easier to find out which options you have (as a user) Mar 25 20:24:56 flo_lap, header_view_compact_header_cell_data is the function that happens whenever a render is needed Mar 25 20:24:57 flo_lap: ccmake show you all the available option, or I'm missing what you mean by 'option' Mar 25 20:25:30 so it gets all the widgets inside its hbox and vbox .. and it renders the item into them Mar 25 20:27:20 flo_lap, given that a pocketpc has a very high screen, it would look kinda neat on that too Mar 25 20:27:32 you can waste more space in height than in width on those devices, rigth? Mar 25 20:27:37 s/waste/use Mar 25 20:27:56 so in stead of a column, you just make an extra row per item Mar 25 20:28:13 and you get finger-compatibility for free Mar 25 20:28:19 pvanhoof: yes right Mar 25 20:28:35 that's exactly why modest was done this way Mar 25 20:29:30 on a mobile, people don't do 50 tasks and need complex list screens with 5 columns and tree nodes Mar 25 20:29:37 they want to focus on a few tasks Mar 25 20:29:52 so if you can show 5 or 6 items per screen, that's fine Mar 25 20:29:58 just show them very well Mar 25 20:30:59 and it kinda helps you on the performance and memory consumption too, if you do that right :D Mar 25 20:31:12 winwinwin Mar 25 20:38:16 cyrilRomain: yes, but it s a little bit less obvious that the help option of a configurre script. Mar 25 20:39:21 pvanhoof: you know why i really would love to work on these things the whole day? Mar 25 20:39:40 :) Mar 25 20:39:45 same here, same here Mar 25 20:47:47 flo_lap: indeed Mar 25 20:54:17 cyrilRomain: Do you know offhand how I add a custom section to a cmake generated Makefile? Mar 25 20:54:47 flo_lap: a new target ? Mar 25 20:55:14 cyrilRomain: yes... a packaging target Mar 25 20:55:40 flo_lap: for packaging you might prefer using CPack Mar 25 20:55:54 which is provided with cmake Mar 25 20:56:17 flo_lap: otherwise you still have ADD_CUSTOM_TARGET Mar 25 20:57:15 cyberdeck: ah cool.. let me see Mar 25 21:00:17 cyrilRomain: ok, looks like debian packages are a task for ADD_CUSTOM_TARGET :) Mar 25 21:00:31 flo_lap: :) Mar 25 21:02:48 flo_lap: http://www.cmake.org/Wiki/CMakeUserUseDebian might interest you Mar 25 21:05:59 cyrilRomain|bbl: indeed - thanks! Mar 25 21:07:17 03woglinde2 07org.oe.dev * rb7c384ae... 10/ (4 files in 3 dirs): Mar 25 21:07:17 pop-1.13: fix uclibc builds Mar 25 21:07:17 * add iconv.patch so libpopt.la finds all symbols even for libicionv Mar 25 21:07:17 * bump pr Mar 25 21:30:15 could someone confirm that gst-plguins do not need esound as dependency anymore? Mar 25 21:35:04 woglinde_: I just got mq200 to show a picture with u-boot on my simpad now only pcmcia is missing :) Mar 25 21:35:40 mr_nice cool Mar 25 21:41:05 mr_nice: cool! Mar 25 21:44:02 woglinde_: tobi is only woglinde3_1/2? Mar 25 21:44:20 slims wrote a gpl driver for redboot which I was able to port to u-boot Mar 25 23:29:43 03koen 07org.oe.angstrom-2007.12-stable * r963f8fa0... 10/ (5 files in 2 dirs): native bbclass: use autotools_stage_all where appropriate Mar 26 00:16:30 Do i just place my custom machine config in my build/conf/machines/dm320.conf or do i put it in the org.openembedded.dev/ area? Mar 26 00:17:02 I'm not understanding how to merge my custom stuff, and the OE configs i guess.. Mar 26 00:35:23 I believe you should put it in org.openembedded.dev/conf/machine/ Mar 26 00:41:18 loudawg: how would I go about distributing that to my users though? Would i have to commit to OE's tree? Mar 26 00:42:35 I'm attempting to build a distribution for my employers new hardware release Mar 26 00:48:27 crweb: I think cbrake has done a bunch of stuff with "Overlays". He has some good writeups here: http://bec-systems.com/web/ Mar 26 00:48:48 crweb: you would put it under org.openembedded.dev, and then when it is ready you would attach it to an entry in the OE bug tracker, and perhaps send an email to the mailing list announcing it. Mar 26 00:49:06 (assuming you want it to become part of the standard OE metadata) Mar 26 00:49:29 Once you have a history of good patches, you can then apply for direct MTN commit access. Mar 26 00:49:45 crweb: what's a dm320? Mar 26 00:50:25 TI DM320 (TMS320C54x) Neuros OSD 1.0, I'm using existing hardware to base my research for our new TI Davinci product Mar 26 00:54:26 crweb: I'm very new to OE, and am still trying to learn all of this myself. Basically, if you want it to be part of the OE tree, you're gonna have to create a bug and attach the config Mar 26 00:54:34 oh, oops, it's already been answered :-P Mar 26 00:55:11 well, I just want to make a distribution our team can use Mar 26 00:55:26 being in the official tree isn't really a goal, but would be cool Mar 26 00:57:29 I suppose you can create your own repository or something, and just change your local.conf to use it, which you can distribute. Just an idea Mar 26 01:00:02 I've been struggling for like 2 weeks now just to learn enough about OE to build nothing more than a console-image for my dell axim x51v. LoL Mar 26 01:00:38 you're lucky, my axim barely has kernel support :> Mar 26 01:00:45 It looks like nobody is really working on it so I'm gonna see what I can do to breathe some life back into the project Mar 26 01:00:49 which axim do you have? Mar 26 01:00:52 x3i Mar 26 01:01:07 supposedly the support is enough to boot a console on the serial port Mar 26 01:01:18 but I'm aiming for other hardware at the moment Mar 26 01:01:23 ah yeah I think I have one of those laying around somewhere too, but I never bothered playing with it much since it's older Mar 26 01:01:33 yeah Mar 26 01:02:13 yeah I'm just trying to find something useful for my x51v. I bought it a year or two ago with all these cool ideas of using it as my organizer and syncing it between work and home over bluetooth, etc etc. Didn't work out so well considering I use Linux at both work and home Mar 26 01:02:14 :-P Mar 26 01:02:20 haha Mar 26 01:02:34 sits unused mostly, right? Mar 26 01:02:45 too much of a pain to always boot into windows just to use the PIM stuff hehe Mar 26 01:02:54 yeah Mar 26 01:03:02 yep, sits around collecting dust. Occasionally gets fired up for use as a road navigation device Mar 26 01:03:26 so I figured I'd make a long-term effort of getting linux running on it, at whatever cost Mar 26 01:03:38 it's good experience at the least :) Mar 26 01:04:12 yeah, that's one of my motivations. I program for mostly userspace right now at work and want to learn more aspects like maybe kernel work and/or driver work Mar 26 01:05:01 and it's just fun doing geeky things with toys :-) Mar 26 01:07:31 I've got a bitbake rolling openmoko for a motorola a1200 at the moment Mar 26 01:08:36 oh, sweet. Will it work with your service and all that too? Mar 26 01:08:45 with luck Mar 26 01:09:31 i just looked up the phone. didn't realize it runs Linux by default. Very cool Mar 26 01:09:34 yeah Mar 26 01:10:13 but some people at work and I want to mess around with OE/openmoko/other phone UIs Mar 26 01:10:52 probably also do some real-world use of some stuff we've written for the kernel Mar 26 01:10:53 Can't blame you. I'd be all over that too! Mar 26 01:12:31 Can any gurus here point me in the right direction as to where I need to start looking to fix a build failure of linux-handhelds? It can't find the initramfs or something. I don't have the exact error right now since bitbake is cooking away again after a clean. This is for an Axim x50v Mar 26 01:14:58 microlith: I'll see if I still have my x3 at home later. If so, maybe I can play with that a bit too. I'm willing to tear it apart for any info or whatever too if you or anyone else could use it Mar 26 01:15:28 ah I've already popped it open once Mar 26 01:16:19 oh you know what, my mistake. I looked at the models and the older one I have is actualy an x5. Oops Mar 26 01:16:52 haha Mar 26 01:19:06 how do I interface with i2c bus drivers that i built into the kernel? Mar 26 01:19:36 particularly a driver for a power regulator Mar 26 01:20:18 That would depend on the driver, I think. But usually it would expose the user-controllable items in /sys Mar 26 01:22:01 do I just access them using open() in c ? Mar 26 01:23:48 Normally one would do things like echo "1" >/sys/whatever/whereever/driver/enable_power_for_widget Mar 26 01:23:51 from a shell. Mar 26 01:24:12 The equivalent in C is to open the file, and write "1 Mar 26 01:24:37 " to it. (sorry, that was supposed to be a newline character, not an actual newline after the 1) Mar 26 01:25:21 oh ok Mar 26 01:27:15 got it thanks Mar 26 02:32:32 Crofton: Welcome back :) How's the cave doing? Mar 26 02:32:42 heh Mar 26 02:32:51 not sure, I am on vacation :) Mar 26 02:32:59 In Colorado for some skiing Mar 26 02:33:16 Crofton: Then what the @#$% are you doing in here ? Mar 26 02:33:55 heh Mar 26 02:34:01 deleting email Mar 26 02:34:10 dealing with a lease Mar 26 02:34:42 Fun.... :) Mar 26 02:41:09 I have OE installed on a VM (the actual folder is on a usb hd) and I recently switched from kubuntu in a vm to a full install. my question is can I use the install on the usb drive as it is, or do i need to reinstall it fresh (im hoping to avoid having to rebuild all those packages Mar 26 02:43:48 03rwhitby 07org.oe.dev * r1af32cac... 10/ (4 files in 3 dirs): slugos-init: Changed boot scripts to only load modules if they are not already loaded Mar 26 02:48:46 can I access functions in device drivers from userspace? I am trying to interface with the driver linux/drivers/i2c/chips/tps65010.c . Mar 26 02:49:37 Upgrading kernel on root from 2.6.24.4+svnr1056-r0 to 2.6.24+svnr1039-r1... Mar 26 02:49:44 Does opkg fix that problem? Mar 26 02:52:09 Darkrift2: you should be able to move the whole thing over. you might need to adjust your BBPATH and some of the settings in your local.conf to update the path. Mar 26 02:52:58 actually the path wouldnt need changing since it was a usb drive, i just mount the drive at /stuff like i did last time.. im just wondering if the setup installs anything on the system that i will be missing Mar 26 02:53:42 Darkrift2: not that I know of. Unless you installed bitbake itself. Mar 26 02:53:49 nope :) Mar 26 02:53:56 that saves me a bit of time :)( Mar 26 02:54:33 Darkrift2: Of course the whole thing would probably run faster if you used an internal HD instead :) Mar 26 02:55:04 Darkrift2: I've moved OE trees between different systems several times, it works nicely. The only issue I've encountered has been when I've forgotten to install some packages on the new host. Mar 26 02:56:23 yeah, prob woudl be faster Mar 26 02:56:34 but it is usb2.0 so im hoping its pretty fast Mar 26 02:56:43 im on a laptop and my internal drive isnt that big Mar 26 02:57:03 Darkrift2: Which Version of Kubuntu are you on ? **** ENDING LOGGING AT Wed Mar 26 02:59:56 2008