**** BEGIN LOGGING AT Mon Feb 04 02:59:57 2008 Feb 04 03:20:36 VoodooZ__: what machine/arch? Feb 04 03:23:58 VoodooZ__: builds fine here for poodle (arm) Feb 04 03:35:08 hvontres|home: it's for slugosbe on ixp4xx.. I noticed that 2.34 does build ok but 2.36 doesn't. Feb 04 03:35:48 the reason I want 2.36 is that 2.34 has a problem with the gps.h include when using the libgps in code.. Feb 04 03:36:00 VoodooZ__: NOTE: package gpsd-2.36: completed Feb 04 03:36:16 THere's a patch in the latest rev of revs.. Feb 04 03:36:18 odd Feb 04 03:36:50 Mine gave me some error saying i386 is an unknown arch... and I'm cross-compiling using OE.. Feb 04 03:37:04 I'll try other things I guess.. thanks. Feb 04 03:42:28 VoodooZ__: maybe a problem with the config files for slugos.. Feb 04 03:47:22 yeah, I noticed that this package is not directly accessible in my slugos dev env.. I have to use bb -b on the .bb file Feb 04 03:47:51 and that the default .bb file builds all the dbus and python fluff which I don't need for a headless robot controller. Feb 04 03:48:26 I'll be using it to parse the gps data for my gps nav bot. Feb 04 03:49:14 hvontres|home: wow! Now I don't understand! It builds fine... Maybe I had some remnants in the tmp folder.. odd Feb 04 03:58:22 does that error ring a bell? gps.h:264: error: 'PATH_MAX' undeclared here (not in a function) Feb 04 03:59:29 I couldn't find much aside from a potential cross-compiler issue (lots of those around these parts! :) Feb 04 04:10:01 FYI, looks like this might be close: http://gcc.gnu.org/ml/gcc-patches/2005-07/msg00751.html Feb 04 04:13:38 VoodooZ__: nope. Feb 04 04:14:57 actually. adding #include did the trick as per http://www.busybox.net/lists/busybox/2001-November/005206.html Feb 04 04:15:43 * VoodooZ__ hugs google Feb 04 04:24:30 hvontres|home: thanks for your help.. Feb 04 04:25:03 VoodooZ__: No problem :) Feb 04 04:25:33 * hvontres|home didn't really do much.... just rebuilt the package here Feb 04 04:27:38 still.. some of those package might just be busted so your confirmation was appreciated. Feb 04 04:29:20 so now my slug equipped bot can figure out bearing and distance to next coordinate.. just need to merge in my existing behaviour module.. Feb 04 04:29:36 night.. Feb 04 04:31:48 VoodooZ__: good luck. And show us some pictures of the bot :) Feb 04 04:32:08 CoreDump: morning Feb 04 04:32:22 morning Feb 04 04:49:32 hvontres|home: I will got a progress page at http://robotics.no-ip.org.. I'll update with pictures as I progress.. Feb 04 04:49:35 night Feb 04 08:13:09 good morning Feb 04 08:15:00 morning Feb 04 08:20:06 hrw: I'm curious where the postinst scripts are stored in a rootfs Feb 04 08:23:24 xjqian: /usr/lib/ipkg/info/ Feb 04 08:27:12 hrw: thanks. btw RRECOMMENDS kernel-module-loop on altboot doesn't give me the desired effect Feb 04 08:27:25 it should Feb 04 08:28:40 hrw: I think RRECOMMENDS does not pack the recommended package into the image. I tested and it didn't. Feb 04 08:30:04 but it should :( Feb 04 08:33:05 one question is how to debug this, the other is if it should, how is RRECOMMENDS different than RDEPENDS case Feb 04 08:33:52 xjqian: RRECOMMENDS == 'install package if available, do not make problems when user wants to remove it' Feb 04 08:34:20 RDEPENDS = 'really needed for app - will not work at all without it. will not install without it' Feb 04 08:39:28 hrw: my mis-understanding then, so if RRECOMMENDED package is available in the feed, it will be installed also? Feb 04 08:57:28 yes Feb 04 08:59:16 florian, hrw : good morning Feb 04 08:59:55 hi steliosk Feb 04 09:01:20 * florian watches apile of hardware waiting for him Feb 04 09:01:32 hrw : we got an intel DL201GLY2 board to test. Mini ITX format, Celeron 1.2Ghz, passive cooling Feb 04 09:02:09 hrw : and the best thing is price.... less than 70Euro including vat :) Feb 04 09:02:26 hrw: thanks for the explanation. I've always misunderstood RRECOMENDS from the behavior of RSUGGESTS. now i'm clear. Feb 04 09:06:17 steliosk: nice Feb 04 09:07:01 steliosk: normal ATX powersupply required? 2.5" or 3.5" hdd connector? Feb 04 09:07:58 hrw : yeah normal power supply, 3.5" connectors and DLGY2 has also SATA ;) Feb 04 09:08:13 DLGY is only IDE Feb 04 09:08:36 interesting Feb 04 09:08:51 The only "problem" is that the boards are using SYS chipset.... Feb 04 09:09:23 * steliosk wonders why Intel would use a SYS chipset on their boards Feb 04 09:10:11 steliosk: with next flat I would probably look at miniitx boards to make some kind of mythtv box Feb 04 09:10:56 hrw : http://www.intel.com/products/motherboard/D201GLY2/index.htm Feb 04 09:12:10 steliosk: nice Feb 04 09:12:18 large heatsink but nice Feb 04 09:12:55 hrw : yes the heatsink is big. does not fit in 1U boxes Feb 04 09:13:10 can be replaced with smaller Feb 04 09:13:21 hrw : but i think a "stock" 1U would be fine Feb 04 09:57:31 bonjour Feb 04 10:02:31 03koen 07org.oe.dev * r161f933d... 10/ (1 packages/linux/linux.inc): linux.inc: sed out LOCALVERSION, it breaks too many things in kernel.bbclass and depmod Feb 04 10:03:47 Are there known problems with ipkg in angstrom 2008? I can't seem to convince it to install any packages I've built, apparently due to it wanting 'arm' but the packages being 'armv5te' Feb 04 10:04:51 broonie: which device? Feb 04 10:04:59 mainstone Feb 04 10:05:34 It's possible there's something I need to change in my build configuration for the update to 2008... Feb 04 10:07:03 mainstone should want armv5te Feb 04 10:33:24 broonie: enjoy... current mainstone linux support is not as good as it used to be. Feb 04 10:33:51 florian: All I care about is ethernet and serial. Feb 04 10:33:55 They seem to work fine. Feb 04 10:34:15 remind me, anyone know a cheap place to get CF ethernet card? Feb 04 10:35:50 * florian wonders why an usb touchscreen uses /dev/input/mouse0 Feb 04 10:36:04 broonie: okay, these work Feb 04 10:38:30 XorA: did you try usb to ethernet? Feb 04 10:38:49 dcordes: yes, I have one Feb 04 10:39:09 XorA: does it work with zaurus usb host? Feb 04 10:39:25 dcordes: yes through a hub, pulls too much power Feb 04 10:39:47 ethernet cfwould be interesting then Feb 04 10:48:23 Hrm. It looks like the conf file I'd been given had a spurious architecture=arm specified in it. Feb 04 10:51:55 ~blame lrg Feb 04 10:51:58 * ibot blames lrg (and Canada) for all the evil in the world Feb 04 10:52:26 Well, we'll have to see if rebuilding the image has the desired effect. Feb 04 11:02:23 03xjqian 07org.oe.dev * r05ed4540... 10/ (6 files in 2 dirs): Feb 04 11:02:23 altboot: add kernel-module-loop as RRECOMMENDS, bump PR. Feb 04 11:02:23 * currently only tosa and collie set CONFIG_BLK_DEV_LOOP=m, other devices built it into kernel Feb 04 11:07:09 i get the error: gconvert.c:48:2: #error GNU libiconv not in use but included iconv.h is from libiconv while compiling glib-2.0 Feb 04 11:07:28 ashu: somehow you built libiconv Feb 04 11:07:39 i have given PREFERRED_PROVIDER_virtual/libiconv="libiconv" in my local.conf file Feb 04 11:07:55 ashu: why? Feb 04 11:08:38 ashu: basically you have buggered your build Feb 04 11:10:08 hrw: you battery totally failed? Feb 04 11:20:52 Xora: am i not supposed to build libiconv? Feb 04 11:21:06 ashu: no Feb 04 11:21:11 ashu: its in glibc already Feb 04 11:21:22 i am using an external libc Feb 04 11:21:43 ashu: check for iconv.h in your external SDK then Feb 04 11:22:11 i have a pre-built libc, i am working with an external toolchain Feb 04 11:22:39 ashu: you need the headers from that glibc otherwise your going to lose Feb 04 11:23:20 XorA: in 760? no - works more or less Feb 04 11:23:34 hrw: no, n810 Feb 04 11:23:38 hrw: your blog post Feb 04 11:24:39 XorA: battery was working nicely before Feb 04 11:25:10 hrw: wierd Feb 04 11:25:20 hrw: charging seems to work here, hope it continues to Feb 04 11:25:25 XorA: looks like it common bug in n800/n810 ;( Feb 04 11:27:31 03tmbinc 07org.oe.dreambox * r6e0ba7d3... 10/ (1 packages/enigma2/enigma2-plugins.bb): enigma2-plugins: parse control files from source to get RDEPENDS, DESCRIPTION Feb 04 11:31:11 Xora: my external toolchain has an header file iconv.h and i have given a TARGET_CPPFLAGS_append = " -I${PRE_BUILT}/include " in local.conf file Feb 04 11:31:41 i dont know why it is not picking it from over there and every time it tries to built libiconv Feb 04 11:31:42 ashu: so ASSUME_PROVIDED += "virtual/libiconv" Feb 04 11:31:53 yes i have given that as well Feb 04 11:32:17 ASSUME_PROVIDED += "libiconv" as well just in case Feb 04 11:44:03 huhu Feb 04 11:45:21 my local.conf file has both the ASSUMED_PROVIDED and somehow libiconv is getting built. glibc fails, see http://pastebin.ca/891127 Feb 04 11:48:15 * XorA is stuck, havent tried to use external toolchain to this level Feb 04 11:49:23 I thought Crofton maintained working omap toolchains Feb 04 11:50:41 Xora: is there anything wrong in my local.conf file? Feb 04 11:51:04 ashu: it looks good to me, but Ive never tried to do what you are doing so I might be overlooking something Feb 04 12:05:49 03koen 07org.oe.dev * r038835a3... 10/ (6 files in 3 dirs): binutils* 2.17: fix avr32 Feb 04 12:10:23 Xora: got glib compiled somehow, i think i have screwed up my build somewhere Feb 04 12:10:24 there was a header file libiconv in my staging directory probably from a previous libiconv compilation Feb 04 12:12:40 stop blaming me ,,,, Feb 04 12:16:20 i think bitbake -c clean libiconv does not remove the files it creates in the staging directories Feb 04 12:16:48 Crofton: i didnt blame you. youre my mentor here, how can i blame you? Feb 04 12:16:54 ashu blames his boss only Feb 04 12:17:08 heh Feb 04 12:19:09 so we can blame ashu and all blame goes to his boss, excellent Feb 04 12:20:00 i want to fix this external toolchain thing, i think the documentation is very poor in OE for interfacing external toolchain Feb 04 12:20:13 so when do we switch to git? Feb 04 12:21:09 zecke: yesterday Feb 04 12:21:55 someone will need to prepare a git for idiots guide .... Feb 04 12:23:19 yes i would need that very much Feb 04 12:23:20 Crofton: I want copy! Feb 04 12:23:31 ashu, I tihink most people use external toolchains to create packages for existing systems originally developed with something other than oe Feb 04 12:23:36 Crofton: git gui Feb 04 12:23:42 urg Feb 04 12:23:44 no gui Feb 04 12:23:49 git: 'gui' is not a git-command. See 'git --help'. Feb 04 12:23:56 I just need to figure out how to us it effectively Feb 04 12:24:04 Crofton: the i in gui is for idiot Feb 04 12:24:05 rofl Feb 04 12:24:09 hrw: then install it Feb 04 12:25:44 doing it now Feb 04 12:26:06 git gui is gitk.. tk ;( Feb 04 12:26:14 I wonder how ugly Tk is now.. Feb 04 12:26:23 hrw: Pretty ugly.. Feb 04 12:26:55 * NAbyss has to put up with a horrible set of Tk scripts at work built atop clearcase.. it's nasty. Feb 04 12:26:58 NAbyss: looks like motif from 80s Feb 04 12:27:04 hrw: Essentially Feb 04 12:27:14 Nothing's moved forward Feb 04 12:28:37 hrw: :) Feb 04 12:29:52 qgit looks better Feb 04 12:30:32 one + point for git, gitweb is far supierior to viewmtn Feb 04 12:31:00 XorA: gitk is working well, massive community Feb 04 12:31:10 zecke: I like gitk Feb 04 12:31:26 zecke: for web stuff, gitweb works really well for examining repo Feb 04 12:31:48 every paranoid guy hast to remember the hash before going to bed and verifying it the next morning :) Feb 04 12:32:12 and verifying that his system is working, or hash the files themselves :) Feb 04 12:32:28 zecke: thats what postits are for :-) Feb 04 12:32:41 there is even ssh auth for git these days Feb 04 12:33:31 XorA: my only mistake is sometimes to use a --hard for git-reset Feb 04 12:48:50 hi, people! Feb 04 12:52:14 03koen 07org.oe.dev * r2762975d... 10/ (1 conf/machine/atngw100.conf packages/linux/linux.inc): linux.inc: allow u-boot entrypoint to differ from loadadress and make atngw100 use it Feb 04 13:03:42 ibot: test Feb 04 13:03:43 Oh, no! There's a test and I haven't studied! Feb 04 13:13:50 * polyonymous_ prefers cgit to gitweb Feb 04 13:16:02 polyonymous_: it doesnt really matter, just the fact the web tools are better Feb 04 13:16:45 XorA, all tools are better. Like zecke mentioned before - massive community. Feb 04 13:19:52 i found this talk on git by linus: http://www.youtube.com/watch?v=4XpnKHJAok8 Feb 04 13:20:32 ashu, if that's the talk I'm thinking of it's more of a show rather than tech talk :) Good show, but show. Feb 04 13:21:13 There's also one by Randall which is more of a "how to use it" sort of thing. Feb 04 13:21:39 Perhaps. Personally, I'm more interested in shows, because I know how to use it :) Feb 04 13:21:53 But I doubt git lacks resources on how to use it. Feb 04 13:23:33 i dont think so Feb 04 13:24:13 it it were the case then it wouldnt be popular, the community is massive, lots of ppl are using it Feb 04 13:25:38 ashu, I think it's the other way around - it did lack resources, but as it grew more popular it gained quite a few. Feb 04 13:25:47 but the result is the same :) Feb 04 13:34:29 ~praise dpkg-deb Feb 04 13:34:37 All hail dpkg-deb! Feb 04 13:35:24 the only reason I know to switch to git is so I need to remember less SCM command sets Feb 04 13:35:37 I really do need to improve git skillz Feb 04 13:35:50 for working with u-boot and kernel Feb 04 13:38:07 * mwester doesn't know git, but he knows a great deal of other SCM tools and anything has be better than the abomination known as monotone. Feb 04 13:38:48 zecke: did we have a workaround for the empty directories problem on git? Feb 04 13:39:02 monotone: proof that the consumption of sufficient computing resources is enough to fool people into thinking the tool really works. ;) Feb 04 13:39:40 XorA: touching an empty file :} Feb 04 13:39:58 has anyone used uclinux with OE? is it possible? and what the hell is this uclinux anyway, just a different kernel for no-mmu devices? Feb 04 13:40:17 zecke: cool Feb 04 13:40:33 .gitignore is traditional. Feb 04 13:40:42 XorA: http://repo.or.cz/w/org.openembedded.dev.git?a=blob;f=.mtn2git_empty;h=e69de29bb2d1d6434b8b29ae775ad8c2e48c5391;hb=HEAD Feb 04 13:41:03 broonie: that is stupid because .gitignore has a different semantic ;) Feb 04 13:41:08 zecke: and question 2 was can your mtn2git run both ways? Feb 04 13:41:16 XorA: not yet Feb 04 13:41:25 zecke: I meant in theory Feb 04 13:41:51 in theory it's possible, as long as the commits back are commited with the same key. Feb 04 13:42:01 or the "gateway" has all the private keys ;) Feb 04 13:42:41 XorA: one theoretical issue. mtn wants two merge only two parents with git you can have up to 16 or 32 (or even more) Feb 04 13:42:52 zecke: ah right Feb 04 13:42:55 first two =-> to Feb 04 13:43:12 yeah actually Ive seen mtn go from 4->2 for merges Feb 04 13:43:32 the way to avoid it is not to perform 16->1 merges :) Feb 04 13:43:58 XorA: mtn could merge infinite number of revs, but with their current automate interface they have limited themselves to two Feb 04 13:44:21 two is a pretty good approximation for infinity Feb 04 13:44:43 zecke: anything != 0 is used in maths isnt it? Feb 04 13:44:52 The tutor who tooted the toot tried to tutor two tooters to toot. Said the two to their tutor: is it harder to toot or to tutor two tooters to toot. Feb 04 13:46:50 anyway I think the two options are hg, git. I don't like hg (even if it is mostly written in python) because I remember to type hg recover on the server... Feb 04 13:48:35 which leave us with git Feb 04 13:49:18 btw, hrw, `git gui` is not `gitk` ;-) Feb 04 13:49:32 But I prefer gitk, even though in general I hate tk interfaces. Feb 04 13:49:56 polyonymous_: indeed Feb 04 13:50:02 but both looks like crap Feb 04 13:50:06 qgit is nicer Feb 04 13:50:31 well, I do think gitk looks alright. Its UI elements are ugly, but it hardly has any. Feb 04 13:50:54 besides, I don't use gui much, just to examine history graph. Feb 04 13:52:48 the mtn2git script has one shortcoming, there is a misplaced openembedded directory in the root (I miss some kind of root renaming... but I blame the mtn history) Feb 04 13:53:30 zecke, what does it mean, anyway? Feb 04 13:54:24 polyonymous_: /packages,classes,conf is our directory layout Feb 04 13:54:31 zecke, yes? Feb 04 13:54:42 polyonymous_: now we have another /openembedded/packages/pmvr directory :) Feb 04 13:54:46 ah Feb 04 13:54:59 zecke, well, my mtn2git converter doesn't do it :) Feb 04 13:55:03 But it's unfinished :( Feb 04 13:55:19 I still have to look at it Feb 04 13:55:21 And it's in c++. Feb 04 13:55:27 eek Feb 04 13:55:31 Sorry :) Feb 04 13:55:36 polyonymous_: what is wrong with mine? Feb 04 13:55:52 Wrong? Nothing, I think. Feb 04 13:55:58 lol Feb 04 13:56:18 Actually, I wrote mine when yours wasn't quite working. Feb 04 13:57:17 polyonymous_: what would be cool is to write a SQL statement to convert to git :) Feb 04 13:57:20 zecke, another thing that is "wrong" is that I wanted mine to do two-way operations like git-svn, because I wasn't hoping for OE to switch SCM :) Feb 04 13:57:32 zecke, but git has no sql backend :) Feb 04 13:57:50 polyonymous_: you can then dump the objects Feb 04 13:58:01 polyonymous_: well, I have an empty git2mtn Feb 04 13:58:06 of course I can, so I'd have to do import/export first :) Feb 04 13:58:13 polyonymous_: and why do you start with mtn -> git, when you want git -> mtn? Feb 04 13:58:45 I wanted two-way, not git->mtn. And the important part was to preserve mtn revs information. Feb 04 13:59:50 polyonymous_: the info is preserved, not in the commit msg though Feb 04 14:00:00 ah, where? Feb 04 14:00:14 I have stuff like Monotone-rev-id: 0b957fab68a4cf71d4ced4e51b427c749003ad25 Feb 04 14:00:36 So that the database can be rebuilt. Feb 04 14:00:39 polyonymous_: in two files, mtn => mark, mark => git, git => mark Feb 04 14:00:52 Ah, that too, but once you lose it - you lose it. Feb 04 14:01:09 polyonymous_: you can rebuild it as well Feb 04 14:01:15 How? Feb 04 14:01:38 polyonymous_: the previous rev's can't change, so you can rebuild by height (almost) :) Feb 04 14:02:07 And, btw, if you ever convert OE to git to never go back, I'd recommend that you add this to comments just for history :) Feb 04 14:02:42 I wouldn't rely on this kind of rebuild. Besides, it would take about as much time as initial import. Feb 04 14:03:53 hello i have cp: cannot stat `/oe/packages/linux/defconfig': No such file or directory and defconfig is in /oe/packages/linux/linux-pma430_2.4.19-rmk7-omap1 and i have: nstall -m 0644 ${WORKDIR}/defconfig ${S}/.config and file://defconfig in SRC_URI...what should i do? Feb 04 14:04:35 s/nstall/install/ Feb 04 14:04:46 GNUtoo: create a defconfig for your custom kernel? Feb 04 14:05:12 zecke, i have created it and it's in /oe/packages/linux/linux-pma430_2.4.19-rmk7-omap1/ Feb 04 14:05:36 GNUtoo: it must be pma430-2.4.19 and not _ Feb 04 14:05:53 zecke, ok thanks a lot Feb 04 14:07:19 Hrmpf. Changing the target architecture doesn't appear to have convinced ipkg it wants to install armv5te packages. Feb 04 14:08:52 broonie: bitbake.conf:94? don't know what you want to do though Feb 04 14:09:08 zecke: EPARSE? Feb 04 14:11:09 Ah, I see. TARGET_ARCH is armv5te in my local.conf Feb 04 14:13:05 Bah, humbing. Feb 04 14:13:11 Humbug even. Feb 04 14:14:27 It's also being set in mtn. Though that is setting PACKAGE_EXTRA_ARCHS which bitbake.conf suggests ought to be accepted too. Feb 04 14:14:47 gm Feb 04 14:15:47 hrw: born? Feb 04 14:25:00 Where is ipkg supposed to get the list of architectures it will install from? Feb 04 14:25:24 broonie: the variable in line 94 of bitbake.conf? Feb 04 14:26:24 zecke: Which is in turn populated from machine.conf? Feb 04 14:26:35 ??? Feb 04 14:26:53 sorry, conf/MACHINE/MACHINE.conf Feb 04 14:26:55 broonie: you can set variables at many places, not limitied to machine.cinf Feb 04 14:27:26 Oh, wonderful. Feb 04 14:27:46 broonie: I don't understand your motivation so my answer is quite generic Feb 04 14:28:26 My problem is that I've built mainstone images where ipkg appears to only want to install all,noarch,arm Feb 04 14:29:02 And not armv5te which is what all the package files produced in the build are. Feb 04 14:29:34 I did have TARGET_ARCH=arm in my local.conf but that's now TARGET_ARCH=armv5te with no visible effect. Feb 04 14:29:53 PACKAGE_EXTRA_ARCHS = "armv4 armv4t armv5e armv5te" Feb 04 14:29:57 in mainstone.conf Feb 04 14:31:08 XorA: It has "armv5te" in there in mtn... Feb 04 14:31:41 broonie: so it has Feb 04 14:32:04 broonie: did you rebuild from scrath after fixing TARGET_ARCH? Feb 04 14:32:08 morning Feb 04 14:32:10 Yes. Feb 04 14:32:33 I blew away everything in my build tree except mainstone Feb 04 14:32:38 broonie: although sounds like you have too much in local.conf if you set TARGET_ARCH Feb 04 14:34:10 i had a problem in the ipk directory when i had too many armXXX directories, i was getting an ipk failure error saying it is a old ipk format or something Feb 04 14:34:33 broonie: are you getting a similar error? Feb 04 14:35:11 No, different. Feb 04 14:38:36 hello, my custom kernel doesn't compile because include/linux/autoconf.h is not found(make: *** No rule to make target `include/linux/autoconf.h', needed by `include/config/MARKER'. Stop.) it's a 2.4 kernel(that may be why there is no such file) Feb 04 14:39:58 03koen 07org.oe.dev * r4c232b32... 10/ (3 files in 3 dirs): libmad: update avr32 patch Feb 04 14:41:22 The kernel uses autoconf???? Feb 04 14:42:40 mwester, no but make *config generate autoconf.h Feb 04 14:43:17 Ah, ok - so it's not the kernel but the config stuff that would be broken then. Feb 04 14:43:31 oh why oh why is kdrive 1.4 so broken :-( Feb 04 14:45:16 * broonie tries *again* with a minimal local.conf Feb 04 14:45:44 GNUtoo, you did at least a 'make oldconfig' ??? Feb 04 14:45:52 no Feb 04 14:46:03 i'll do it Feb 04 14:46:27 but i copied the .config to defconfig Feb 04 14:48:22 gremlin[it], is copying the .config to defconfig ok or should i also do a oldconfig? Feb 04 14:51:30 we'll see if it compile with the make oldconfig and the defconfig Feb 04 14:55:02 thanks a lot it worked Feb 04 15:00:48 bye Feb 04 15:01:09 bye, hrw|gone Feb 04 15:03:33 03xora 07org.oe.dev * rfe3ce31a... 10/ (1 conf/distro/angstrom-2008.1.conf): angstrom-2008.1.conf : update gtk+ and glib Feb 04 15:03:47 Crofton: if your offer to commit patches for the omap osk still stands, have you seen bug 3731? Feb 04 15:41:20 * Crofton curses oe dns ... Feb 04 15:42:28 apaulsen, this works with angstrom from .dev with gpe-image? Feb 04 15:42:54 Crofton: yes Feb 04 15:43:04 ok, I'll push in a bit Feb 04 15:43:11 thanks Feb 04 15:43:23 need to get he cat of me Feb 04 15:43:29 the cat off me Feb 04 15:43:46 he likes to help me type and he reads what you guys say Feb 04 15:49:19 * mwester wonders what the cat will do with that information; one can never trust a cat, you know. Feb 04 15:50:12 mwester: you know the difference between a dog and a cat? Feb 04 15:51:16 * mwester likes to think he does... but sometimes his dog makes him wonder Feb 04 15:51:47 chouimat|work: I think you have a cat, no? :) Feb 04 15:52:09 mwester the dog when call him it come right away. the cat take the message and will do whatever he wants Feb 04 15:52:13 mwester yes Feb 04 15:52:21 apaulsen, what directory is that file in? Feb 04 15:52:26 hehe - that sums it up well. Feb 04 15:52:30 in the future, can you use mtn diff Feb 04 15:53:50 Crofton: yes, sorry. I was in a hurry and forgot :( . I'll put up an mtn diff in a minute. Feb 04 15:55:06 Hi. Can't understand... Familiar and familiar-like distributives is soft-float point based, because xscale have't float-point coprocessor. But in disassemble code "libc.so.6" I found "fmls, stfs" and etc. instructions (float-point instructions). More then, I can't compile my test application on arm-linux-gcc with soft-float: "libc.so.6 uses hardware FP, whereas hello.o uses software FP". That wrong? Feb 04 15:55:10 thanks Feb 04 15:55:20 I am having trouble finding the file to patch Feb 04 15:55:26 brb Feb 04 16:05:24 Crofton: I added the mtn diff. This should probably be pushed upstream, but I'm not sure how to go about that. Feb 04 16:13:37 apaulsen, this seems different than the previous patch Feb 04 16:13:41 hmm Feb 04 16:13:48 add the ti-osk patch to the bb file Feb 04 16:14:03 and the actual patch applies to the XSserver source? Feb 04 16:14:24 yes, is there a better way? Feb 04 16:14:30 sorry for seeming like an idiot, I do not look at thies area of code often Feb 04 16:14:33 no Feb 04 16:14:42 I am just making sure I understand Feb 04 16:15:10 Xserver is just a shell script, this patch just inserts two lines into it Feb 04 16:15:30 yeah Feb 04 16:16:52 ok applied Feb 04 16:17:11 you will likely want to revert the change on your copy, then pull again after I push Feb 04 16:17:33 Crofton: sounds good, thanks Feb 04 16:22:58 Crofton : has this anything to do with touch ; i have run into an issue of running Xserver with touch support, X does not recognise touch Feb 04 16:23:44 probably, you are working with the OSK also? Feb 04 16:24:37 ashu: Is your board recognised by /etc/X11/Xserver? Feb 04 16:25:15 ashu: If it's not in the big case statement there try adding a match for it specifying -mouse tslib Feb 04 16:26:44 yes sir Feb 04 16:28:05 * * OE Bug 3741 has been created by dp(AT)xora.org.uk Feb 04 16:28:07 * * kdrive 1.4 broken for tslib when rotated Feb 04 16:28:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3741 Feb 04 16:28:19 broonie: i have just an SecurityPolity file in /etc/X11/xserver Feb 04 16:28:45 what do you mean by "Is your board recognised?" can you elaborate? Feb 04 16:29:05 * * OE Bug 3731 has been RESOLVED (FIXED) by philip(AT)balister.org Feb 04 16:29:06 ashu: In your root filesystem? Feb 04 16:29:07 * * omap OSK needs an entry in Xserver Feb 04 16:29:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3731 Feb 04 16:29:44 has anyone had success building Angstrom-2008.1 for a x86 target recently? Feb 04 16:30:04 glibc-intermediate_2.6.1 is failing for my for the x86 machine Feb 04 16:30:47 hello, bootstrap-image failled: http://pastebin.com/m46ad0be6 what does things like "task-distro-base: unsatisfied recommendation for avahi-daemon" means? Feb 04 16:31:08 yeah in my rootfs: ls -l /etc/X11/xserver/ Feb 04 16:31:18 -r--r--r-- 1 root root 2952 Feb 4 2008 SecurityPolicy Feb 04 16:31:33 xserver != Xserver Feb 04 16:32:05 apaulsen, pushed you should for the automerger to merge the heads though Feb 04 16:32:10 bbl, lunchtime Feb 04 16:33:43 ashu: /etc/X11/Xserver is a script that detects your platform and sets your xserver arguments Feb 04 16:34:42 i dont have that script Feb 04 16:41:19 ashu: is xserver-common installed? Feb 04 16:47:15 03andrew.paulsen 07org.oe.dev * r3a89dabf... 10/ (3 files in 3 dirs): xerver-common : Add support for TI OSK board with Mistral LCD. From bug #3731. Feb 04 16:54:06 anyone? Feb 04 17:01:12 hi, all! Is there anybody who have knowledge of at91sam9260 CPU? Feb 04 17:04:29 slapin: some, yes Feb 04 17:05:51 apaulsen, could you please explain SDRAM initialization? 1) is it possible using pllB as MCK source at 48 MHz? and 2) how could I write mode register? Feb 04 17:06:34 apaulsen, also, is DBGU supposed to work with PLLA at 200Mhz and MCK=PLLA/2? Feb 04 17:07:04 apaulsen, I mean serial port of DBGU Feb 04 17:07:49 slapin: I just used the atmel dev board and the vanilla kernel. there weren't any extra patches or anything Feb 04 17:12:06 apaulsen, I use custom board with 2xmicron 128Mbit chips with 32bit data bus, and I'm unable to setup SDRAM properly, I either get C33C... patterns of FFFFFFFF patterns when read after some time, and proper values when read immediately, what could be a prooblem? Feb 04 17:14:18 * broonie is still having trouble installing any packages on angstrom-2008.1 Feb 04 17:14:33 apaulsen, all timings are per datasheet, I think only about PLL configurations and mode register settings, because DBGU serial on my board for some unknown reason works only when MCK is on PLLB with divider /2 and with default 48Mhz setting. Feb 04 17:14:45 slapin: have you looked at the signals on a scope? I would ask around at the at91 forums. Feb 04 17:15:42 broonie: strange. I upgraded my test board to 2008.1 few days ago. Feb 04 17:16:01 apaulsen, I have no equipment enough to measure near-100Mhz values, and found no info on at91 forum, thanks, I will try to ask. Feb 04 17:16:30 Oh, FFS. As well as all the feeds there's /etc/ipkg/arch.conf. Feb 04 17:16:36 Which I've been deleting along with the feeds. Feb 04 17:16:50 broonie: oops Feb 04 17:17:19 * broonie notes that he did go looking for a configuration file for architectures initially but couldn't find TFM for ipkg configuration. Feb 04 17:20:32 Thanks :) Feb 04 17:27:36 03xora 07org.oe.dev * r7a8c7842... 10/ (6 files in 4 dirs): Feb 04 17:27:36 libtorrent_0.12.0.bb, rtorrent_0.8.0.bb : a small working bittorrent client Feb 04 17:27:36 ideal for embedded systems. Has advanced features such as PEX and DHT Feb 04 17:33:23 * mwester applauds XorA's commit - that'll make some happy SlugOS users. Feb 04 17:33:40 mwester: testing it and it works well on arm Feb 04 17:34:03 * XorA needs to purchase slug now I thinks Feb 04 17:34:19 would anyone be interested in/opposed to a change to ipaq-sleep to make it monitor %cpu instead of load average? Also, I'm thinking of making some changes to make it work smarter with cpufreq. Feb 04 17:34:28 It seems to be the best of the bunch on the NSLU2, but it's only been available in the optware feeds up to now; now it'll be native. Feb 04 17:35:45 well I think I got the compile options right, Im sure someone will moan if I havent Feb 04 17:36:53 pity slug is lacking in ram :-( Feb 04 17:38:40 Yeah, but they're cheap. :) Feb 04 18:02:18 !seen psokolovsky Feb 04 18:02:19 psokolovsky (n=psokolov@nat-users.is.ua) was last seen quitting from #oe 23 days, 1 hour, 41 minutes ago stating (Read error: 110 (Connection timed out)). Feb 04 18:02:23 geez Feb 04 18:03:15 03mwester 07org.oe.dev * r3828ccf5... 10/ (5 files in 3 dirs): SlugOS: updated boot scripts to support nfs rootfs with the new network driver. Feb 04 18:14:28 bye all! Feb 04 18:18:11 anyone? Feb 04 18:37:59 yeah GNU Feb 04 18:45:21 erk, Ive never even heard of bakefile before :-( Feb 04 18:46:57 re Feb 04 18:55:31 GNUtoo: are you still asking about this? http://pastebin.com/m46ad0be6 Feb 04 18:57:04 XorA: oh no... not another trademark dispute.... Feb 04 18:57:18 XorA: bitbake bakefile.... Feb 04 19:03:26 XorA: oh... Feb 04 19:04:05 florian: oh?? Feb 04 19:04:41 XorA: oh trademark dispute... Feb 04 19:04:54 what? Feb 04 19:05:26 nothing, there is a program called bakefile, and hvontres|work made a joke :-) Feb 04 19:15:49 XorA: phew :-) Feb 04 19:35:20 apaulsen, yes Feb 04 19:36:05 XorA, wew Feb 04 19:38:10 just built X windows from OE for omap, my X does not recognise touch, how do i check if X has touch enabled? Feb 04 19:38:22 GNUtoo: did you do anything strange like move your ipkg directory? Feb 04 19:38:41 all i can see is touch is producing interrupts Feb 04 19:39:38 ashu: did you find out if you have a /etc/X11/Xserver? Feb 04 19:40:49 apaulsen, no just bitbake bootstrap-image Feb 04 19:42:07 hi everybody Feb 04 19:42:26 GNUtoo: it looks like it can't find your ipkg directory for some reason. I really don't know why this would happen...did you change your machine type? Feb 04 19:43:38 apaulsen, no i made a fresh tmp and my machine is a custom machine that didn't change Feb 04 19:43:40 ashu: can you run "awk 'BEGIN { FS=": " } /Hardware/ { print $2 } ' hvontres|work: did you have success with the initramfs? Feb 04 19:46:18 GNUtoo: that log doesn't show the build output from all those packages...were there any errors? Can you check if the .ipks were built and placed in tmp/deploy/something? Feb 04 19:47:08 i have an omap2430 board Feb 04 19:47:31 i have not compiled Xomap i have got Xfbdev instead Feb 04 19:47:37 ashu, send me one :) Feb 04 19:48:26 i shall try :-) Feb 04 19:48:43 ooohhh free hardwarew!!!! :-) Feb 04 19:48:58 ashu, do you have a link? Feb 04 19:50:54 i have a soft and a hard link Feb 04 19:51:46 ashu: no wireless link? Feb 04 19:52:14 sorry, i know only soft and hard links, wireless links are out of my reach Feb 04 19:52:47 Crofton: what are you looking for? Feb 04 19:53:31 what I really want is an omap board that is setup for software defined radio work Feb 04 19:53:39 ashu: :-) Feb 04 19:53:48 Crofton: you could get an N800 Feb 04 19:54:01 omap2420 and it has a radio Feb 04 19:54:14 I like to make my own radio :) Feb 04 19:54:23 can it do usb high speed? Feb 04 19:54:38 an omap with high speed usb could talk to a usrp Feb 04 19:54:50 think so... I am not sure Feb 04 19:54:52 of course, usrp2 has gig-e interface Feb 04 19:55:01 so I no longer worry about that Feb 04 19:55:18 philippe, any word on the embedded room schedule? Feb 04 19:55:53 Crofton: sent it to the Fosdem organisation Feb 04 19:56:13 great Feb 04 19:56:38 I can post it somewhere if you want. There still might be some changes as I am waiting for confirmations Feb 04 19:57:18 well, I am going regardless :) Feb 04 19:57:26 so no need to post Feb 04 19:57:43 Crofton: ok :-) Feb 04 19:58:37 if I had not made up my mind, I would want to peek :) Feb 04 20:03:04 03philippe 07org.oe.dev * r865eaa99... 10/ (1 packages/gpe-mini-browser/gpe-mini-browser2_svn.bb): gpe-mini-browser2: add build recipe for gpe-mini-browser2, the new GPE browser based on WebKitGtk Feb 04 20:03:38 i have done some work on sdio wireless on omap5912 Feb 04 20:03:38 that was ages ago Feb 04 20:03:47 GNUtoo: I think the problem may be in your machine.conf Do you have armv5te as a PACKAGE_EXTRA_ARCHS? Feb 04 20:04:10 okay, there is a board called RDK board manufactured by mistral, which can be attached to the OSK 5912 board Feb 04 20:04:10 The RDK board has a SDIO interface - you can plug in a sdio wireless card and work Feb 04 20:04:24 apaulsen: i ran the script /etc/X11/Xserver - it invokes "Xfbdev -br -pn", is there anything i need to check? Feb 04 20:05:30 ashu: can you run just that one line I sent? it's the guts of module_id() Feb 04 20:10:38 when i run it i get OMAP 2430 SDP Feb 04 20:11:17 i am able to recieve the touch screen inputs in /dev/input/event1 when i try to cat it - i get some values, i see the interrupts Feb 04 20:11:48 i have made a symbolic link between /dev/input/event1 and /dev/input/touchscreen0 Feb 04 20:14:29 ashu: try editing the /etc/X11/Xserver file and make a copy of the "Generic OMAP1510/1610/1710" case, but change the string to "OMAP 2430 SDP" Feb 04 20:15:33 and use event1 instead of event0 Feb 04 20:19:05 apaulsen, no i have only that: http://pastebin.com/m5c0a957 but i'll search for it in my distro and it's requires Feb 04 20:20:43 i get the error: Couldn't find pointer driver /dev/input/event1 Feb 04 20:20:50 apaulsen, no don't have it Feb 04 20:21:29 GNUtoo: which ipkg directory is busybox in? Feb 04 20:22:44 ashu: I didn't think that would work, but it was worth a try. So you have no /dev/input/touchscreen0 or ts0? Feb 04 20:22:44 apaulsen, /oe/tmp/deploy/glibc/ipk/arm-oabi Feb 04 20:23:08 no i dont have it Feb 04 20:23:36 i created a link between /dev/input/touchscreen0 and /dev/input/event1 and it didnt work Feb 04 20:24:03 when i do a cat /dev/input/event1 - i get some thing when i touch the touchscreen Feb 04 20:25:04 I think a driver interprets the event1 output and generates the touchscreen0 output Feb 04 20:26:13 ashu: what does your platform have for a touchscreen device? Feb 04 20:27:40 ADS784X touchscreen Feb 04 20:27:43 apaulsen, according to http://focus.ti.com/general/docs/wtbu/wtbuproductcontent.tsp?contentId=4672&navigationId=12609&templateId=6123 it has one Feb 04 20:32:53 apaulsen: see http://pastebin.ca/891650 Feb 04 20:34:00 GNUtoo: that arm-oabi seems to be the problem - your rootfs script isn't looking there. I'm not exactly sure why or how to fix it. It looks like angstrom adds that directory to the package config only for certain machines. You might try editing the Angstrom config line that deals with ABI. I'm just guessing so if anyone actually knows, please correct me. Feb 04 20:36:26 apaulsen, before that i had a problem with ABI: i modified that : ARM_ABI ?= "${@['','oabi'][bb.data.getVar('MACHINE',d) in ['collie','h3600', 'h3800', 'simpad', 'htcwallaby','pma430']]}" and added the pma430 but it didn't do nothing...i had to add TARGET_OS = "linux" after the require line for angstrom Feb 04 20:37:18 GNUtoo: are you sure you want oabi? Feb 04 20:37:30 apaulsen, stuck to 2.4 kernel Feb 04 20:41:12 GNUtoo: it looks like that should have done the trick. Did you also add a FEED_ARCH_pma430 and PACKAGE_EXTRA_ARCHS_pma430 to conf/distro/include/angstrom.inc? Feb 04 20:41:29 apaulsen, mabe i've found the cause...it's in include/angstrom.inc Feb 04 20:41:50 no i didn't thanks i'll do that... Feb 04 20:42:11 ashu: what does your etc/ts.conf look like? Feb 04 20:43:30 apaulsen, stupid question is arm v4t and arm v4l the same thing? Feb 04 20:43:55 GNUtoo: I doubt it, but I don't know the difference Feb 04 20:44:07 apaulsen: http://pastebin.ca/891662 Feb 04 20:45:59 ashu: try running these before starting X: Feb 04 20:46:02 TSLIB_TSDEVICE=/dev/input/event1 Feb 04 20:46:07 export TSLIB_TSDEVICE Feb 04 20:47:30 apaulsen: i have done it - export TSLIB_PLUGINDIR=/usr/lib/ts Feb 04 20:47:56 export TSLIB_TSDEVICE=/dev/input/event1 ; export TSLIB_CONFFILE=/etc/ts.conf Feb 04 20:51:03 ashu: I think I have exhausted my knowledge of the touchscreen now :( It sounds like tslib has a problem with event1 for some reason Feb 04 20:52:03 i am exhaused as well my friend Feb 04 20:52:23 did you add those exports and the Xserver entry before you made your link from event1 to touchscreen0? Maybe tslib can't make a touchscreen0 since it already exists Feb 04 20:53:46 i think so Feb 04 20:54:19 mabe it's the same thing....i have Processor : Arm925Tsid(wb) rev 2 (v4l) but wikipedia says arm9 are v4t Feb 04 20:54:40 i will do all over again, i have done everything i could Feb 04 20:55:28 i was asking in the case someone else knew...and thanks a lot for your help Feb 04 21:17:31 apaulsen, after making the change should i rebuild evrything from scratch deleting tmp? Feb 04 21:18:01 GNUtoo: I don't think that would be required, but if all else fails... Feb 04 21:31:47 does anyone know if there's an option for busybox to have a separate command history for each shell or is 1.7.2 just broken this way? Feb 04 21:38:09 thesing: no not yet :( I might try later this week Feb 04 21:44:42 apaulsen: i dont get a touchscreen node in /dev Feb 04 21:44:57 i really dont have a clue what could be wrong Feb 04 21:45:30 my touchscreen driver works, is there any configuration file i need to update for X to recognise touch Feb 04 21:48:23 ashu: just that Xsession line as far as I know Feb 04 21:50:08 Xserver, I mean Feb 04 21:50:30 have you tried a "-mouse tslib" instead of the event1? Feb 04 21:56:11 Xsession?? Feb 04 21:56:11 sorry mate, i am bugging you too much, its just that i cant make it work, i appreciate youre help and time Feb 04 21:57:53 ashu: I meant /etc/X11/Xserver Feb 04 21:58:48 why conf/machine/include/tune-arm9tdmi.inc does not have -march=armv4 inside TARGET_CC_ARCH but only has -mcpu=arm9tdmi -mtune=arm9tdmi Feb 04 21:58:49 ? Feb 04 22:00:00 it works :-) Feb 04 22:00:29 thanks buddy, thanks a ton Feb 04 22:01:26 awesome, do you know what your screen's DPI is? Feb 04 22:02:40 ashu: Feb 04 22:03:08 i dont know Feb 04 22:03:39 can we give the size as well i have a custom board which is half VGA Feb 04 22:04:37 ashu: yes, I think there are examples in the current Xserver script Feb 04 22:05:30 ashu: once you get it figured out, you can probably get Crofton to add your lines to the Xserver patch he pushed for me earlier today Feb 04 22:07:27 okay, i will Feb 04 22:07:57 ashu, put the patch on the bugtracker Feb 04 22:08:05 use mtn diff to create a patch Feb 04 22:09:31 morning all Feb 04 22:10:02 my packages have too many customised files. however i will try checking out a new rep and doing a diff Feb 04 22:11:36 mtn diff packages/xserver-common shoudl do the trick Feb 04 22:12:50 whenever i invoke matchbox-desktop i get too many of these Errors "converting from UTF-8 to STRING: Conversion from character set 'UTF-8' to 'ISO-8859-1' is not supported" Feb 04 22:13:51 okay i will Feb 04 22:39:03 has anyone worked here on gnash? Feb 04 22:40:37 ashu, Genesis has Feb 04 22:41:22 wanted to know if it supports alsa Feb 04 22:42:14 is there any light weight flv player other than gnash? Feb 04 22:43:19 ashu, gnash is the best one, mabe konqueror-embedded has compatibility with older flash(see compilations options) Feb 04 22:44:26 okay, i have already started a build. i need to add only gnash in one of image/XX.bb files isnt it?. it will pull whatever is needed Feb 04 22:44:29 isnt it? Feb 04 22:44:44 ashu, http://www.konqueror.org/embedded/#HowdoIrunFlashpluginswithKonquerorEmbedded Feb 04 22:45:15 ashu, you have gnash 0.8.1 bb from gnash repository... Feb 04 22:46:03 i have gnash_0.8.0.bb Feb 04 22:46:05 ashu, you also need a browser...you could use firefox without a lot of options like Genesis did Feb 04 22:46:27 i just want a flash player, i dont need a browser Feb 04 22:46:56 i think it should provide some sort of gui interface Feb 04 22:47:50 ashu, http://cvs.savannah.gnu.org/viewvc/gnash/gnash_0.8.1.bb?revision=1.1&root=gnash&view=markup Feb 04 22:47:50 there is also gxine and libxine that can play, i tried to compile it outside OE and the voice is too choppy and it finally breaks down Feb 04 22:48:13 ashu, what about vlc? Feb 04 22:48:41 ashu, (but it plays only flash videos) Feb 04 22:49:08 vlc would do as well, i dont know how much size it needs Feb 04 22:50:50 ashu, ouch you'll need to update it because the most recent vlc in org.openembedded.dev is 0.8.4 and the current is 0.8.6d Feb 04 22:51:26 thanks, i will Feb 04 23:04:09 apaulsen, ping Feb 04 23:07:00 tftpboot is haning loading gpe-image onto the osk Feb 04 23:08:30 i am facing the same problem Feb 04 23:08:33 Crofton: yes, it runs out of ram. Feb 04 23:08:47 I split my image into two chunks and load them separately Feb 04 23:08:50 grrrr Feb 04 23:09:04 stupid computers Feb 04 23:09:25 do you have notes on how to do this .... Feb 04 23:09:51 I have some scripts Feb 04 23:10:22 Crofton: i am able to download the image but it hangs i am not able to get the shell prompt at all Feb 04 23:10:49 i think the new busybox has too many scripts and i have to wait and wait till i get the busybox prompt Feb 04 23:11:23 I have 32m of ram Feb 04 23:11:43 and can't load a 24M image .... Feb 04 23:12:04 can't load a 21M image without splitting it either Feb 04 23:12:24 I bet uboot loads somewhere dumb :) Feb 04 23:20:28 why does one get user.warn gconfd (root-315): errors? Feb 04 23:20:28 i get plenty of them Feb 04 23:21:11 Crofton: this is the script we use to split the rootfs and it also generates an autoscript for uboot http://www.pastebin.ca/891864 Feb 04 23:22:54 thanks Feb 04 23:22:59 Let me give this a try Feb 04 23:25:04 I'll need to look at this tomorrow Feb 04 23:25:41 Crofton: you will need mkimage that comes with uboot Feb 04 23:26:12 yeah, I might try and skip that bit :) Feb 04 23:26:57 I need to figure out what the script does Feb 04 23:29:49 it really works great, but don't ask me to explain how it works Feb 04 23:30:34 ok :) Feb 04 23:30:41 I guess I just need to try Feb 04 23:30:59 does gnash really need qt-embedded? the depends dont say much about it - i think the link from where qt-embedded is downloaded is broken Feb 04 23:32:05 Crofton: my coworker wrote that script so I can get answers to your questions Feb 04 23:32:15 ok Feb 04 23:32:29 tomorrow I will try to go through it Feb 04 23:32:39 I really should see where u-boot loads Feb 04 23:32:45 that may be the real problem .... Feb 04 23:55:20 /quit Feb 05 00:42:25 * florian shoots mtn one more time **** ENDING LOGGING AT Tue Feb 05 02:59:57 2008