**** BEGIN LOGGING AT Thu Mar 29 02:59:56 2007 Mar 29 03:08:46 ibot, seen RP Mar 29 03:09:07 ibot, status Mar 29 03:09:25 rp is currently on #gpe (6h 34m 28s) #oe (6h 34m 28s) #openmoko (6h 34m 28s) #handhelds.org (6h 34m 28s). Has said a total of 2 messages. Is idling for 5h 33m 30s, last said: 'hi CoreDump|home '. Mar 29 03:09:34 Since Tue Mar 27 02:14:46 2007, there have been 106 modifications, 520 questions, 0 dunnos, 0 morons and 456 commands. I have been awake for 2d 54m 39s this session, and currently reference 112741 factoids. I'm using about 19232 kB of memory. With 0 active forks. Process time user/system 8696.81/664.01 child 0.04/0.01 Mar 29 03:10:25 ibot, reluctant botsnack Mar 29 03:10:26 thanks, jacques Mar 29 03:26:07 what reciepe is gdk-pixbuf-csource? Mar 29 03:26:25 gqview needs that as a dep Mar 29 03:29:58 Zero_Chaos: it's part of .. hang on Mar 29 03:30:04 Zero_Chaos: I just used it myself for psplash Mar 29 03:30:54 Zero_Chaos: it should be built as part of gtk+-dev Mar 29 03:31:02 which is a dynamic package that is part of gtk+ Mar 29 03:47:41 jselect: thanks, the deps for gqview should be updated, because it fails when it cannot find gdk-pixbuf-csource Mar 29 03:56:36 jselect: are you sure? it still barfs Mar 29 04:08:56 jselect: Try it on your system. If you build gtk+-dev dynamic package and install it Mar 29 04:09:13 it has in it's install dir the gdk-pixbuf-csource binary Mar 29 04:09:27 but it might be a wrapper for another binary Mar 29 06:28:30 $ arm-linux-gcc -v Mar 29 06:28:31 ... Mar 29 06:28:40 --enable-multilib Mar 29 06:28:45 ... Mar 29 06:28:49 --disable-multilib Mar 29 06:28:50 ... Mar 29 06:29:01 gcc version 4.1.1 Mar 29 06:29:25 Someone could not decide if multilib is useful or not :-) Hmm, what *IS* multilib? Mar 29 06:32:31 iirc it's for mixed 64bit/32bit environments Mar 29 06:32:35 good morning all Mar 29 06:35:05 koen: good morning 2u2 Mar 29 07:07:40 morning all Mar 29 07:16:25 morning all Mar 29 07:36:46 hi all Mar 29 08:17:41 * koen|lab adds "Tanner EDA" to the list of "crap tools that should die" Mar 29 08:19:06 hehe, that sounds like a german tool Mar 29 08:22:48 looks like a win 3.11 tool Mar 29 08:23:14 no auto-refresh, no scrollbars, requires 3 mouse buttons, etc Mar 29 08:23:55 and the 'graphical representation' of an inverter doesn't show the 'blocks', but "include ../foo/bar/mod.md" Mar 29 08:31:55 morning Mar 29 08:33:22 hey XorE Mar 29 08:37:05 koen|lab: yo yo yo, 2 days to release :-D Mar 29 08:37:13 ;) Mar 29 08:37:22 6 days before the project deadline Mar 29 08:37:36 gah, Angstrom is more important :-) Mar 29 08:37:39 and our sigmadelta ADC only has 2 NMOS' so far Mar 29 08:38:09 XorA: does wolfson have a design for a lownoise 5MHz SD-ADC? Mar 29 08:38:33 koen|lab: no idea, I havent been here long enough to absorb product info Mar 29 08:38:50 whats the SD part stand for? Mar 29 08:38:59 sigma-delta Mar 29 08:39:02 its been 12 years since I last did analogues Mar 29 08:40:03 the reverse of http://en.wikipedia.org/wiki/Delta-sigma_modulation Mar 29 08:40:44 I love signal processing, but the crap tools created when neanderthals were around make it less fun Mar 29 08:42:14 XorA: but before a release we need hrw|gone's internal lib code in OE psokolovsky_'s keylaunch changes and a working gpe-conf Mar 29 08:42:45 and once people verify an image boots on their device we can move those images to the RC directory Mar 29 08:42:58 koen|lab: gpe-conf that doesnt turn off screen is easy Mar 29 08:43:17 should we add the patch to OE or will florian make a release? Mar 29 08:43:24 what are the keylaunch changes? Mar 29 08:43:32 patch Mar 29 08:43:36 koen|lab: Signal processing was the main part of my degree, but since I havent used it since passing, I lose all knowledge Mar 29 08:43:43 and it seems keylaunch isn't in angstrom-x11-image Mar 29 08:43:44 is there any way to tell OE to look to an other location for extra shared libs? I am using an external toolchain and want OE to be able to pull in it's shared libs as needed? Mar 29 08:44:02 tkp: look at the sharp-compat distro Mar 29 08:44:11 XorA: thanks Mar 29 08:45:22 * XorA runs his feed generator to see if he can get bluetooth working Mar 29 08:45:27 tkp: the information in the manual is not helping, then please help to improve it Mar 29 08:46:08 zecke: the manual suggests creating a package which installs them manually Mar 29 08:46:29 XorA: the regular angstrom feeds aren't working? Mar 29 08:46:39 I did that, but something is wrong with is, as the -dbg package isn't working correctly Mar 29 08:46:44 pH5, psokolovsky: what is the best kernel version to run on hx4700? Mar 29 08:46:54 koen|lab: they do, but I also like to keep my own feed for when things like mb-panel-2 fsckup my machine Mar 29 08:46:55 ie. gdb can't find the files in .debug for all the libs I installed manually Mar 29 08:47:06 XorA: aha Mar 29 08:47:25 XorA: what is it about that distribution that gets OE to pull in the libs from the prebuilt toolchain tho? Mar 29 08:47:30 tkp: make sure to give back and update the documentation! Mar 29 08:47:36 koen|lab: and one day I shall sell the how to generate one feed from many script :-) Mar 29 08:47:38 koen|lab: latest, 2.6.20-hh5 ;-) (need to add it though) Mar 29 08:47:49 psokolovsky: did you test with OE configuration? Mar 29 08:47:55 tkp: what do you mean by "pull in" use it to link against it? or to package it properly? Mar 29 08:47:57 tkp: no idea, but it uses the toolchain from sharp to do the build Mar 29 08:48:12 zecke: to package it properly Mar 29 08:48:17 we have to do that and/or update defconfig before it can be used as default kernel. Mar 29 08:48:27 ie. to ensure the shared libs from the toolchain get installed on the target Mar 29 08:48:39 ah, yep ;-) Mar 29 08:49:08 I didn't test wifi/bt with 2.6.20-x yet. I guess it'll work, but I'm not sure. Mar 29 08:49:13 koen|lab: in this regard, there's a proposal to move to defconfigman-maintained defconfig for hx4700, like for h3900 & h4000 Mar 29 08:49:30 psokolovsky: did you have a look at my latest mail about defconfig? Mar 29 08:49:45 before proposing a move we should fix it to support everything (wifi, cpufreq, etc) Mar 29 08:50:02 tkp: well you will need to create bb files that puts the libraries into packages Mar 29 08:50:08 pH5: yep, now that adc class is in, highest topic. will reply today (and add magician of course) Mar 29 08:50:16 zecke: right - I created aa package like this (http://rafb.net/p/gmLfUj86.html), but it's not working properly... Mar 29 08:50:17 psokolovsky: great Mar 29 08:50:17 tkp: and oe_libinstall them to make sure RDEPENDS are found properly Mar 29 08:50:43 zecke: is there an example package that does this somewhere? Mar 29 08:51:04 tkp: not that I know Mar 29 08:51:35 tkp: well... you install to /lib which is not part of FILES by default Mar 29 08:51:36 I got it (semi) working by making it more like this: http://rafb.net/p/G1YdZm27.html ... but it's the -dbg package that isn't working correctly Mar 29 08:52:50 tkp: see, your live would be more easy if you would read messages printed by OE Mar 29 08:53:12 tkp: it will tell you which files were not packaged, etc. and concrete errors would be appreciated as well Mar 29 08:53:27 pH5: fyi, I'm adding -hh4/hh5 now Mar 29 08:53:29 tkp: otherwise we need to classify these things as human error which has a lot of different solutions Mar 29 08:54:12 psokolovsky: ok. I'll try/test -hh5 with OE defconfig for magician/hx4700 later today Mar 29 08:54:25 psokolovsky: do you have a patch for the keylaunch issue already? Mar 29 08:54:41 zecke: OE doesn't print messages over here Mar 29 08:54:48 zecke: with bitbake trunk Mar 29 08:55:15 koen|lab: don't use unstable versions. We should increase the sleep... Mar 29 08:55:26 sirfred's imageon patches break compilation of Xsdl :( Mar 29 08:55:28 pH5: nope, but its confirmed to be the case, and OKed by koen to commit. Mar 29 08:55:34 tkp: So you have created packages (ipk,tar) and they contain files? Mar 29 08:56:05 tkp: why do you think you will need -dbg packages? What is the error you get? Mar 29 08:56:06 zecke: yes Mar 29 08:56:40 zecke: I'm trying to debug something on the target, and not having these symbols in there isn't helping Mar 29 08:57:39 tkp: again. What is the error? Do you have debugging symbols in your prebuilt stuff? Mar 29 08:57:52 tkp: what packages are created, what do they contain?... Mar 29 08:58:02 zecke: example: http://rafb.net/p/AqAcTK50.html Mar 29 08:58:34 don't use helix, use GStreamer it comes will the full sourcecode :) Mar 29 08:58:43 zecke: not an option! Mar 29 08:58:49 tkp: but again, either listen or don't! Mar 29 08:58:57 I have the full source code for Helix Mar 29 08:59:13 tkp: Does your prebuilt toolchain contain debugging information? Mar 29 08:59:15 zecke: it's not question of not listening... I need to use Helix Mar 29 08:59:24 zecke: yes Mar 29 08:59:51 tkp: at a guess item=-1077847592 is bad :-) Mar 29 09:00:27 tkp: well, then what packages are getting created? (e.g. pastebin the OE output) Mar 29 09:01:04 tkp: what is your _concrete_ issue with your -dbg packages? Is it not getting created? Is it missing a file? Mar 29 09:01:22 http://rafb.net/p/JNnGJo66.html Mar 29 09:01:50 heh Mar 29 09:01:53 that clearly doesn't work.... I'll paste the output from my fudged version, which semi works Mar 29 09:01:55 "Installing libesd0 blanks the screen on bootup" Mar 29 09:02:26 tkp: You don't listen. If you fail you present a problem how do you expect to get any help? Mar 29 09:02:30 /ignore tkp Mar 29 09:02:37 koen|lab: s/libesd0/gpe-conf/ Mar 29 09:02:47 zecke: what? Mar 29 09:02:56 what am I not listening to? Mar 29 09:03:50 tkp: Q:"What packages are getting created and what do they contain?" ll in a random dir is not an answer! Mar 29 09:04:36 tkp: specially if I ask for output of OE during the creation of the packages. So you don't want to listen, you fail to read, you fail to give an answer, you failed the turing test. Mar 29 09:04:43 stc-shlibs & stc-shlibs-dbg Mar 29 09:04:46 http://rafb.net/p/wm7Ivl35.html Mar 29 09:05:47 tkp: poor boy, you really think you answered that question? Mar 29 09:05:59 tkp: ll in a random dir is not output of OE Mar 29 09:06:12 zecke: hold on please Mar 29 09:06:18 tkp: and what is the packaged content? Mar 29 09:06:32 morning all Mar 29 09:06:47 morn RP Mar 29 09:06:52 yo RP Mar 29 09:06:58 03pfalcon 07org.oe.dev * r143c8e4d... 10/ (3 files in 2 dirs): (log message trimmed) Mar 29 09:06:58 linux-handhelds-2.6: Drop 2.6.20-hh2. Add 2.6.20-hh4. Mar 29 09:06:58 * Changes since 2.6.20-hh3: Mar 29 09:06:58 1. htcathena port added. Mar 29 09:06:58 2. Generic Device GPIO API (aka gpiodev) committed, based on virtual device Mar 29 09:06:58 methods pattern. Mar 29 09:07:00 3. rs232_serial - generic driver for GPIO-controlled RS232 transceiver on Mar 29 09:07:04 03pfalcon 07org.oe.dev * r7ebe9518... 10/ (3 files in 2 dirs): (log message trimmed) Mar 29 09:07:06 linux-handhelds-2.6: Drop 2.6.20-hh1. Add 2.6.20-hh5. Mar 29 09:07:08 * Changes since 2.6.20-hh4: Mar 29 09:07:10 1. ts-adc-debounce: Use spec-originated formula for pressure calculation, Mar 29 09:07:12 still needs tweaking for real world; implement pressure cut off properly, Mar 29 09:07:14 and handle jitter in more sustainable manner: now samples with too much Mar 29 09:07:16 jitter are ignored, not cause TS sampling to abort. Still needs further Mar 29 09:07:18 03pfalcon 07org.oe.dev * r546defc1... 10/ (1 packages/linux/linux-handhelds-2.6_cvs.bb): linux-handhelds-2.6 cvs: HEAD is 2.6.20-hh5 now. Mar 29 09:07:35 zecke: the packaged content of each of the two packages that got created (stc-shlibs & stc-shlibs-dbg) is as per my last paste Mar 29 09:07:50 * koen|lab thought in-kernel processing was considered evil Mar 29 09:08:22 is it possible to compile in locale binaries with qemu-native on x86_64? Mar 29 09:08:44 03pH5 07org.oe.dev * r23f2457b... 10/ (1 packages/xorg-xserver/xserver-kdrive-common.inc): xserver-kdrive-common.inc: shuffle w100/imageon support a bit Mar 29 09:08:53 03pH5 07org.oe.dev * r589b03f4... 10/ (4 files in 3 dirs): Mar 29 09:08:53 xserver-kdrive-common: the tslib is patch not needed anymore in 1.2, Mar 29 09:08:53 the other two have to be modified Mar 29 09:09:00 03pH5 07org.oe.dev * r0e52f21a... 10/ (1 packages/xorg-xserver/xserver-kdrive_X11R7.1-1.1.0.bb): xserver-kdrive: change X11R7.1-1.1.0 to use xserver-kdrive-common.inc Mar 29 09:09:09 03pH5 07org.oe.dev * r533dbe85... 10/ (1 packages/xorg-xserver/xserver-kdrive_1.2.0.bb): xserver-kdrive: change 1.2.0 to use xserver-kdrive-common.inc Mar 29 09:09:17 03pH5 07org.oe.dev * re0bcd233... 10/ (4 files in 3 dirs): xserver-kdrive: change git version to use xserver-kdrive-common.inc Mar 29 09:09:23 03pH5 07org.oe.dev * r53fc298d... 10/ (4 files in 3 dirs): xserver-kdrive: remove xcalibrate-1.2 patch, xcalibrate is upstream after all Mar 29 09:09:27 03pH5 07org.oe.dev * r1e8faa47... 10/ (1 packages/angstrom/task-angstrom-x11.bb): task-angstrom-x11: add keylaunch Mar 29 09:10:09 zecke: here is the output from OE: http://rafb.net/p/s7bxhc80.html Mar 29 09:11:12 I had to modify the bb to only try and install 3 libs - to keep the output small so it would fit in the paste bin (http://rafb.net/p/uB8GQZ12.html) Mar 29 09:11:48 tkp: sorry, the time I'm willing to sacrifce is over. next time try to present your problem in a clean way so that people can actually help you Mar 29 09:12:11 tkp: don't run with -DDD and it helps to paste the output of the tasks in question... Mar 29 09:12:27 zecke: that is just one -D Mar 29 09:13:18 and it creates the following files in the two packages that get created: http://rafb.net/p/pqsZlZ94.html Mar 29 09:13:26 tkp: boy, you seem to expect something from -dbg. So where is reality not matching your expectations and why do you think it is a packaging issue? Mar 29 09:14:53 zecke: when I run things through gdb, it finds the debug files from the .debug directories for all libs, except those installed by my shlibs-dbg package Mar 29 09:15:24 zecke: I'm sorry I presented this problem in a bit of a muddled fashion Mar 29 09:20:26 tkp: but the files are getting installed? So it is unlikely to be a packaging bug Mar 29 09:20:42 tkp: so use objdump and findout where .debug is getting searched Mar 29 09:25:32 what is up with monotone.openembedded.org? I want to push sirfred's gpe-conf fix. Mar 29 09:26:27 pH5: try pushing to ewi or wolfson Mar 29 09:27:18 koen|lab: Seems like it's up again. I'll remember that for the future. Mar 29 09:27:58 03pH5 07org.oe.dev * r7fa457f0... 10/ (4 files in 3 dirs): gpe-conf: fix for http://bugs.linuxtogo.org/show_bug.cgi?id=42 by Manuel Teira Mar 29 09:28:04 zecke: did you see deepfish already? Mar 29 09:28:33 nope Mar 29 09:28:59 koen|lab: a browser? Mar 29 09:31:23 zecke: yes Mar 29 09:31:26 hey mickeyl Mar 29 09:31:31 cheers Mar 29 09:34:41 morning all Mar 29 09:35:42 hey do13_ Mar 29 09:36:03 hi koen|lab Mar 29 09:40:03 re Mar 29 09:41:20 aaaargh Mar 29 09:41:40 selecting text and copying to the clipboard crashes the t-spice app Mar 29 09:43:12 hi mickeyl , do13_, hrw Mar 29 09:43:37 hey Richard, Marcin, mickeyl Mar 29 10:00:56 pH5: the imageon patch changes the way kdrive handles memory, so that might be breaking Xsdl Mar 29 10:02:28 koen|lab: yes. I'll remove imageon from the 1.2 build for now and let sirfred handle that. Mar 29 10:03:07 maybe create xserver-kdrive-imageon recipe? Mar 29 10:03:46 zecke: how can I printk mesg? definition is here http://www.gelato.unsw.edu.au/lxr/source/include/linux/pm.h#L143. my pasebin output is here http://pastebin.ca/414821 generated with this code changes(line566) http://pastebin.ca/414824 Mar 29 10:03:49 hrw: I'd prefer if sirfred could update his patch not to break Xsdl and submit it upstream eventually. Mar 29 10:04:05 like http://www.openembedded.org/repo/org.openembedded.dev/packages/xorg-xserver/xserver-kdrive-imageon_X11R7.1-1.1.0.bb ? Mar 29 10:04:13 koen|lab: ;D Mar 29 10:05:04 hrm Mar 29 10:05:15 CELF is routing me through chicago Mar 29 10:05:50 so instead of taking 9 hours it will take 16 hours... Mar 29 10:06:11 * koen|lab goes to find some books to read Mar 29 10:09:23 * djay-il suggests koen|lab "the tipping point" Mar 29 10:10:38 mr_nice: print msg.event as %d? Mar 29 10:11:06 djay-il: from gladwell? Mar 29 10:11:32 zecke: thx I will try that Mar 29 10:12:45 "If the VBS (bulk-source-voltage) of the NMOS transistor is not zero then this transistor is Mar 29 10:12:45 (probably) placed upside-down in your schematic (source at the top and drain at the bottom). This Mar 29 10:12:45 has no effect on the operation of the transistor (it is a symmetrical device), but the results of the OP Mar 29 10:12:46 analysis are wrong. To correct this, select the NMOS transistor in the schematic of your inverter and Mar 29 10:12:46 push the Flip vert. button" Mar 29 10:12:51 who designs crap like this? Mar 29 10:14:34 03pH5 07org.oe.dev * r8e23de0b... 10/ (12 files in 4 dirs): xserver-kdrive: clean up different patch versions, drop imageon from 1.2 for now Mar 29 10:20:47 03pH5 07org.oe.dev * rb55671f4... 10/ (3 files in 2 dirs): xserver-kdrive: fix patch directory name for 1.2 Mar 29 10:23:14 "console=ttyS0,115200 console=tty1" in CMDLINE will put kernel output on both or only first one? Mar 29 10:23:38 both Mar 29 10:25:18 koen|lab: yes :-) Mar 29 10:28:05 thx Mar 29 10:30:07 djay-il: thanks, I'll add it to the list Mar 29 10:30:26 mickey|hri: can python have oneline if? Mar 29 10:30:46 "if something then do this else do that" way Mar 29 10:30:56 koen: sure Mar 29 10:31:45 btw, I'm having failed fetch, how do I test the links? Mar 29 10:36:14 koen: I might be wrong, but pcre-4.4.bb cannot be built, and libpcre-4.4 as well, as they don't exist on URI anymore Mar 29 10:37:09 sorry, libpcre-4.4.bb Mar 29 10:53:13 does oe_libinstall only work on libtooled libs? Mar 29 11:09:12 http://pastebin.ca/414870 - for review Mar 29 11:13:12 how do I add my own package? Mar 29 11:13:17 into dev? Mar 29 11:17:45 submit to bugtracker Mar 29 11:17:59 but I should create it first, right? Mar 29 11:18:25 is it possible to have multiple package versions? If so, which one got chosen for build? Mar 29 11:20:11 gi all Mar 29 11:20:14 eh hi Mar 29 11:24:32 koen: the gcc-cross 4.1.2 made modprobe (from modutils) segfault, but I did not dig further. It works with gcc-cross-4.1.1 Mar 29 11:37:33 RMK uses gcc 4.4 Mar 29 11:37:59 03pH5 07org.oe.dev * r0530cbc2... 10/ (2 files in 2 dirs): xserver-kdrive-1.2: fix enable-tslib.patch Mar 29 11:38:17 hi Mar 29 11:38:46 exit Mar 29 11:40:13 pH5: is it wise to try kdrive 1.2 now? Mar 29 11:40:47 hi, all! Mar 29 11:40:54 !oebug 2060 Mar 29 11:40:56 * * Bug 2060, Status: NEW, Created: 2007-03-29 04:46 Mar 29 11:40:56 * * slapinid(AT)gmail.com: Angstrom build failure (using 4.1.2 toolchain) Mar 29 11:40:57 koen|lunch: yes. Xfbdev works on the magician Mar 29 11:40:57 * * http://bugs.openembedded.org/show_bug.cgi?id=2060 Mar 29 11:47:59 pH5: libxcalibrate doesn't have a release yet? Mar 29 11:50:27 is OE.dev supposed to build using 4.1.2 gcc-cross, or not? Mar 29 11:50:38 psokolovsky_, Mar 29 11:51:19 koen|lunch: I don't know of any. Mar 29 11:53:01 ok Mar 29 11:54:40 bug 2060 seems to be libtool problem, am I right? Mar 29 11:55:21 slapin_nb: looks that way Mar 29 11:55:27 slapin_nb: try rebuilding libtool cross Mar 29 11:56:08 koen|lab: I'm attaching some of our configuration patches to the freedesktop.org bugzilla. Mar 29 11:56:20 pH5: thanks Mar 29 11:56:44 although I suspect kdrive bugzilla is just a fancy /dev/null Mar 29 11:58:35 03xora 07org.oe.dev * r30dd54ac... 10/ (5 files in 3 dirs): Mar 29 11:58:35 packages/apt/apt_0.6.46.2.bb Mar 29 11:58:35 packages/apt/apt.inc Mar 29 11:58:35 packages/apt/apt-package.inc : build and package shared libs at all times. Mar 29 11:58:50 koen|lab: yeah, then it's just for my own conscience ;) Mar 29 11:59:46 koen, it is build from scratch. So I doubt that will help... I'm trying it though. Strange it breaks that late though... Mar 29 12:06:47 03pH5 07org.oe.dev * r8bddcec8... 10/ (1 packages/xorg-xserver/xserver-kdrive/devfs.patch): xserver-kdrive: drop devfs.patch, nobody uses it anymore. Mar 29 12:06:51 03pH5 07org.oe.dev * r8c94a3c2... 10/ (2 files in 2 dirs): xserver-kdrive-1.2: simplify enable-tslib.patch Mar 29 12:07:15 aargh Mar 29 12:07:17 XorA, pH5: ipkg-make-index doesn't know about epochs.... Mar 29 12:07:24 koen|lab: heh heh Mar 29 12:07:32 koen|lab: btw mtn pull and build apt Mar 29 12:08:25 anybody remembers where to get fosdem materials? Mar 29 12:08:34 slapin_nb: fosdem.org Mar 29 12:08:44 or .fosdem.org Mar 29 12:09:03 morning Mar 29 12:10:28 viviji1, vivijim: apt builds now Mar 29 12:11:36 koen|lab: so that proves the debian EABI effort is stalled :-) Mar 29 12:16:27 how does ipkg-make-index decide to move packages to the morgue? Mar 29 12:16:46 if (packages.packages.has_key(pkg_key)): Mar 29 12:16:46 old_filename = packages.packages[pkg_key].filename Mar 29 12:17:03 that seems to the case it's looking at, but I have no idea what it does Mar 29 12:17:14 (and why it doesn't handle epochs) Mar 29 12:17:24 koen|lab: all I know is +cvs is guaranteed to end up morgued Mar 29 12:18:57 * florian seems to trigger all nasty bugs Mar 29 12:19:32 hi koen, I've noticed, thanks! :) Mar 29 12:19:38 florian: you built sdk with oe? Mar 29 12:19:39 configure tries to run a test that is cached in the site files already... Mar 29 12:19:46 hrw: i try Mar 29 12:19:52 florian: me too Mar 29 12:20:10 you have two weeks left ;) Mar 29 12:20:19 hrw: it looks like all the stuff is in bad condition... Mar 29 12:21:10 hrw: meta-sdk needs to be updated, it builds blsolete packages... not speaking about the compilers. Mar 29 12:21:14 koen|lab: on elc you will present working recipes? Mar 29 12:21:21 koen|lab: there is epoch support in ipkg.py, perhaps it doesn't work. Mar 29 12:21:27 florian: I first want to build gcc-cross-sdk working Mar 29 12:21:45 koen we decided to use oe to build mamona... First of all we are trying to code a stable distro conf files and then we will ask for repository write permission and start to really contribute to oe... Mar 29 12:21:56 viviji1: cool! Mar 29 12:22:03 koen|lab: the important part in -make-index is that packages.add_package(pkg) either adds the package (if it is newer) and returns 0, or doesn't add it at all and returns 1. Mar 29 12:22:18 ah Mar 29 12:22:19 http://rafb.net/p/jPbfQw83.html Mar 29 12:22:33 I can't make heads or tails of that snippet Mar 29 12:22:41 its all python to me Mar 29 12:23:18 hrw: i found out that it is quite likely that one of the patches breaks it - without all the patches it builds (but of course it is not what you want). Mar 29 12:23:31 http://rafb.net/p/cX8T7R74.html <- add_package function Mar 29 12:24:21 hrw: I try to define a useful toolchain to become imported into Scratchbox - that's more idiot-proof ;) Mar 29 12:24:23 http://rafb.net/p/Kdw1qW81.html <- compare version function Mar 29 12:26:18 now that &%§$ startup-notification breaks... lf_cv_sane_realloc is in the site files but it still tries to run the test Mar 29 12:26:36 http://rafb.net/p/nv3dBS73.html <- parse_version seems to handle epochs Mar 29 12:26:58 koen|lab: I'd try put a debug print in compare_version. "%s vs %s, %s wins". Mar 29 12:27:17 s/put/to put/ Mar 29 12:27:31 pH5: fyi, I have fix for i-m-i braindeadness regarding version compare Mar 29 12:36:42 pH5, psokolovsky_, XorA: http://rafb.net/p/GNVmXE66.html always falls through to "we don't need no stinking epochs" Mar 29 12:38:07 koen|lab: so the version parsing is broken. is .epoch always 0? Mar 29 12:38:18 koen|lab: i.e. you suspect that it mis-parses epoch'ed version string? because "math" is right apparently ;-) Mar 29 12:40:01 !oebug 2061 Mar 29 12:40:02 * * Bug 2061, Status: NEW, Created: 2007-03-29 05:42 Mar 29 12:40:03 * * pmiscml(AT)gmail.com: Fix broken version comparison in ipkg-make-index Mar 29 12:40:04 * * http://bugs.openembedded.org/show_bug.cgi?id=2061 Mar 29 12:40:42 koen|lab, pH5, RP: Can you gentlemen review ^ and tell me if it's ok as is, or would need RFC on the list? Mar 29 12:41:12 hmmm Mar 29 12:41:14 print "We don't need no stinking epochs. self: %s, ref: %s " % self.epoch, ref.epoch Mar 29 12:41:20 that isn't valid python apparently Mar 29 12:41:40 koen: use braces after % Mar 29 12:41:47 ( ... ) Mar 29 12:42:25 psokolovsky_: there's debug left in your patch Mar 29 12:43:01 koen|lab: yeah, so people can try that it is (now) works as expected. can be removed for commit, of course Mar 29 12:43:08 psokolovsky_, koen|lab btw, I always wanted to ask for ref what an epoch is regarding to versions? is it debian-specific thing? Mar 29 12:43:29 slapin_nb: kind of super-version Mar 29 12:43:45 also, an integer only Mar 29 12:44:27 http://rafb.net/p/USDVZ641.html <- epochs are always zero Mar 29 12:45:37 heh Mar 29 12:45:49 ackage: xserver-kdrive Mar 29 12:45:49 Version: 1.2.0-r2 Mar 29 12:46:02 IOW, OE doesn't store the epoch in the resulting package Mar 29 12:46:14 lol Mar 29 12:46:44 fields.append(["Version: %s-%s\n", ['PV', 'PR']]) Mar 29 12:46:59 koen: ouch. Mar 29 12:47:12 I guess that needs some form of if(PE>0){'PE'} Mar 29 12:47:15 so I didn't commit my OE changes, RP only updated bitbake. Mar 29 12:47:22 let me see... Mar 29 12:47:29 psokolovsky_, so it overrides normal version? like most significent part of it? Mar 29 12:48:51 slapin_nb: consider it is extra component preprended on the left of normal version. but it's purpose is to allow big changes of versioning schemes (like we have now) Mar 29 12:52:14 03pH5 07org.oe.dev * r4887947d... 10/ (1 classes/package_ipk.bbclass): package_ipk.bbclass: add epoch support Mar 29 12:52:18 03pH5 07org.oe.dev * r7ac1e35a... 10/ (1 conf/bitbake.conf): bitbake.conf: add epoch support Mar 29 12:52:23 koen: sorry about that Mar 29 12:52:46 psokolovsky_, ah, I see, thanks! Mar 29 12:53:01 nice Mar 29 12:54:04 do you Mar 29 12:54:07 oops Mar 29 12:54:30 I have a problem fetching from anoncvs.gnome.org...anyone can help? Mar 29 12:54:42 djay-il: GNOME switched to svn Mar 29 12:54:57 yea, and how would I solve it? Mar 29 12:55:01 djay-il: update OE and make check if this issue was already fixed Mar 29 12:55:11 djay-il: if not fix the SRC_URI pointing to anoncvs Mar 29 12:55:23 djay-il: and don't forget to send a patch/file a bugreport Mar 29 12:55:26 OE is superuperupdated Mar 29 12:56:10 hm, ok Mar 29 12:57:03 bug 1921 Mar 29 12:57:10 zecke_: thanks Mar 29 12:58:00 is it right that package_deb.bbclass sets Architecture to TARGET_ARCH and not PACKAGE_ARCH? Mar 29 12:59:07 03pH5 07org.oe.dev * r16aa4528... 10/ (1 classes/package_deb.bbclass): package_deb.bbclass: add epoch support Mar 29 13:02:55 03pH5 07org.oe.dev * rc3806da7... 10/ (1 packages/xorg-lib/libx11_1.1.1.bb): libx11-1.1.1: makekeys.diff is applied upstream Mar 29 13:03:05 * koen|lab for bb in `grep -rn "^PE =" . | awk -F: '{print $1}' | awk -F/ '{print $3}' | awk -F_ '{print $1}'` ; do bitbake -c rebuild $bb ; done Mar 29 13:08:05 pH5: could you send a mail to the list saying we now have proper epoch support? Mar 29 13:09:52 pH5: xorg-font need mkfontdir mkfontscale in rdepends Mar 29 13:10:34 * hrw goes to push some updates Mar 29 13:10:47 hrw: can you fix it? Mar 29 13:10:56 * pH5 tries to find his email client Mar 29 13:11:19 anybody could explain to me how could I get list of .bbs executed for some particular target? Mar 29 13:11:25 I have partial fix Mar 29 13:11:29 slapin_nb: bitbake -nv? Mar 29 13:12:09 -v is not needed. -n is enough Mar 29 13:14:41 03hrw 07org.oe.dev * r0d2e69a0... 10/ (1 classes/package.bbclass): Mar 29 13:14:41 package.bbclass: added support for private libraries (used only in package) Mar 29 13:14:41 - if package contain libraries which are not used outside then add PRIVATE_LIBS Mar 29 13:14:41 variable with names of them to not generate shlibs for them. Mar 29 13:14:45 03hrw 07org.oe.dev * rd480f9ca... 10/ (5 files in 3 dirs): Mar 29 13:14:45 dates/openmoko-dates: mark libgtkdatesview.so as PRIVATE, make them conflict Mar 29 13:14:46 - both packages provide libgtkdatesview.so.0 which is not used outside of dates Mar 29 13:14:48 - both packages provide binaries of same names Mar 29 13:14:52 03hrw 07org.oe.dev * rfa0b91e7... 10/ (1 conf/distro/include/preferred-xorg-versions-X11R7.1.inc): X11R7.1: prefer libxfont-native 1.2.0 as we use freetype 2.3.x Mar 29 13:14:56 03hrw 07org.oe.dev * r40299631... 10/ (5 files in 2 dirs): acpid: use update-rc.d class Mar 29 13:15:01 03hrw 07org.oe.dev * r06419316... 10/ (1 packages/aumix/aumix_2.8.bb): aumix: added ncurses to dependencies Mar 29 13:15:09 03hrw 07org.oe.dev * r410b421e... 10/ (1 packages/efl/edje-native_0.5.0.037.bb): edje-native: fix staging Mar 29 13:15:16 03hrw 07org.oe.dev * r626f2c6d... 10/ (1 packages/gpe-mini-browser/gpe-mini-browser_svn.bb): gpe-mini-browser: do not comment code but remove it. we have SCM Mar 29 13:15:25 03hrw 07org.oe.dev * rddfdd404... 10/ (1 packages/gstreamer/gst-plugins-good_0.10.5.bb): gst-plugins-good: depend on gst-plugins-base (from Poky) Mar 29 13:15:29 I just found that running with -b doesn't runs out of memory on my VPS Mar 29 13:15:31 03hrw 07org.oe.dev * r3d41fdda... 10/ (6 files in 2 dirs): hostap-daemon: do not depend on hostap-modules as they are 2.4 only now Mar 29 13:15:37 03Marek 07org.oe.dev * r9cb8a3c2... 10/ (1 packages/picocom/picocom_1.4.bb): picocom: fix building - close #1972 Mar 29 13:15:42 while just bitbake target does. Mar 29 13:15:43 hrw: what depends on libxfont-native? Mar 29 13:15:43 03hrw 07org.oe.dev * r791cff22... 10/ (1 packages/starling/starling_0.1.bb): starling: fixed dependencies - close #2055 Mar 29 13:15:47 r0d2e69a0... changeset will be reverted Mar 29 13:15:50 03hrw 07org.oe.dev * r33a1439e... 10/ (1 packages/uucp/uucp_1.07.bb): uucp: package /usr/spool/uucp/ (empty) for cu Mar 29 13:15:53 pH5: mkfontdir-native Mar 29 13:16:14 pH5: some stuff which is needed by font-misc-misc Mar 29 13:16:17 ah. so I should provide libfont-native for 1.2.7, too. Mar 29 13:16:25 thx Mar 29 13:18:26 hrw: mkfontdir-native needs libxfont-native added to its DEPENDS, then. Mar 29 13:18:57 maybe Mar 29 13:19:10 ~lart python for lack of string.isEmpty() Mar 29 13:19:11 * ibot slaps python upside and over the head with one freakishly huge killer whale named hugh for lack of string.isEmpty() Mar 29 13:20:00 String objects... tsts Mar 29 13:28:29 03hrw 07org.oe.dev * r1dce7867... 10/ (1 classes/package.bbclass): disapproval of revision '0d2e69a0248ac99b9f990dfb473d63c9653f0f7c' Mar 29 13:28:34 03hrw 07org.oe.dev * r0374910c... 10/ (1 classes/package.bbclass): package.bbclass: fix build problem related to private libs Mar 29 13:32:49 03pH5 07org.oe.dev * rde029d6b... 10/ (1 conf/machine/magician.conf): magician.conf: add ohci-hcd kernel module, don't build jffs2 images for now Mar 29 13:35:14 strings as objects - such a waste of resources! hrw: use perl instead! ;) Mar 29 13:35:33 pH5: MACHINE_FEATURES += usbhost? Mar 29 13:35:44 oh. Mar 29 13:35:49 can I commit http://en.pastebin.ca/415004? Mar 29 13:36:19 pH5: we have an scm, so no need to comment :) Mar 29 13:36:19 what about the phone MACHINE_FEATURE - is there a way to pull gsmd into an angstrom-x11-image already? Mar 29 13:36:38 pH5: add it to taskbase or task-angstrom? Mar 29 13:36:46 eh :) Mar 29 13:37:23 we can assume that those who build their own image have a current u-boot loaded, right? Mar 29 13:37:49 yes Mar 29 13:38:04 if you want to flash an image you should be smart enough to look up how to flash it Mar 29 13:39:48 03pH5 07org.oe.dev * rca921d74... 10/ (1 conf/machine/fic-gta01.conf): fic-gta01.conf: remove obsolete padding Mar 29 13:44:24 03pH5 07org.oe.dev * r65a1a083... 10/ (1 conf/machine/magician.conf): Mar 29 13:44:24 magician.conf: add usbhost MACHINE_FEATURE, drop kernel modules that are Mar 29 13:44:24 taken care of by task-base Mar 29 13:54:34 03pH5 07org.oe.dev * r52476389... 10/ (1 packages/gsm/files/gsmd): gsmd: remove unneeded stty call from init script Mar 29 13:54:38 03pH5 07org.oe.dev * rf727073e... 10/ (1 packages/linux/linux-handhelds-2.6/hx4700/defconfig): linux-handhelds-2.6: update hx4700 defconfig for 2.6.20 Mar 29 13:54:42 03pH5 07org.oe.dev * ra6b118bb... 10/ (1 packages/linux/linux-handhelds-2.6/magician/defconfig): linux-handhelds-2.6: update magician defconfig for 2.6.20 Mar 29 13:54:47 03pH5 07org.oe.dev * r78fcd28c... 10/ (3 files in 2 dirs): puzzles: add libxt dependency Mar 29 13:57:42 hello everyone Mar 29 13:58:13 hi vlo Mar 29 13:58:22 03pH5 07org.oe.dev * re1bd1634... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom-2007.1.conf: add avahi-autoipd Mar 29 14:00:01 I have cross-built a C++ program I have written for our robot, and when I run it on the OE system, I get this lib information version "warning": http://pastebin.ca/415028 Mar 29 14:00:14 someone knows how to fix it ? Mar 29 14:01:06 I have bever seen this :) Mar 29 14:02:28 the waiting for connection is ok, it's in my program, written by me ;) (just in case) Mar 29 14:03:10 vlo: I would say, ignore it :) Mar 29 14:03:25 vlo: we might have just stripped this information from the binary Mar 29 14:03:54 hrw: wow, you're still using UUCP? (I remember using it in ancient times between Linux 2.0 and SCO whatever) Mar 29 14:05:27 schurig: neo1973 came with it. I switched to picocom Mar 29 14:06:57 zecke: OK ... I could ignore it, the program seems to work correctly, but I just wanted to ask in case it was something well known ... (I guess I'm not the first one writing C++ app with STL on OE systems) Mar 29 14:09:18 mmh, 1mb left on my mem, need to clean some apps to do a successful ipkg upgrade Mar 29 14:10:41 vlo: e.g. inhibit stripping of binaries and check if the error goes away Mar 29 14:13:23 zecke: euh, what do you mean by this ? Mar 29 14:20:05 I don't see this on my c++ stuff Mar 29 14:20:22 vlo, I think there is some variable you can set to prevent stripping occuring Mar 29 14:22:25 vlo: I assume that arm-linux-strip removes something from libstdc++ what it would like to have Mar 29 14:22:50 vlo: we have a flag IIRC it is speleed INIHIBT_STRIP which inihibits the call of arm-linux-strip Mar 29 14:27:43 :D I ar u suftwear unjunear, hoo nedes speel chekkers? Mar 29 14:28:30 vlo: I once wrote an STL programm for OE, with gcc-4.1.1, glibc2.3.5+cvsSOMETHING and did not see your linker message Mar 29 14:28:48 vlo: I could link without problem against libstdc++.so Mar 29 14:29:02 vlo: what happens when you "file libstdc++.so" ? Mar 29 14:30:09 vlo: I get this: Mar 29 14:30:12 $ file tmp/cross/arm-linux/lib/libstdc++.so.6.0.8 Mar 29 14:30:12 tmp/cross/arm-linux/lib/libstdc++.so.6.0.8: ELF 32-bit LSB shared object, ARM, version 1, not stripped Mar 29 14:34:30 hi lrg Mar 29 14:40:24 schurig: 'cu' is part of uucp Mar 29 14:40:43 schurig: 'cu' supports colour terminals and resizing while minicom doesn't Mar 29 14:40:48 hey lrg Mar 29 14:41:59 koen|lab: tried picocom? Mar 29 14:42:12 enough for dumb usage of gsm modem ;D Mar 29 14:42:36 hrw: no yet Mar 29 14:42:40 not* Mar 29 14:46:18 morning all Mar 29 14:46:22 do anybody knows some tool for python memory profiling? Mar 29 15:06:55 * koen|lab ponders about the consequences of 60 times oversampling a 5MHz signal Mar 29 15:09:13 well, you use a lot of extra cpu time .... Mar 29 15:10:34 Crofton|laptop: Sigma-delta ADC Mar 29 15:10:39 so no CPU :) Mar 29 15:11:07 koen|lab: 300Mhz signal then ? :-) Mar 29 15:11:18 but not very many bits, right? Mar 29 15:11:43 who can i set a french keyboard in console with familiar ? Mar 29 15:11:46 djay-il: 600Mhz (nyquist) Mar 29 15:12:05 Crofton: 60dB SINAD, so ~10 bits Mar 29 15:12:13 koen|lab: ? why the doubling..? Mar 29 15:12:45 djay-il: you normally sample at 2x the frequency you want Mar 29 15:12:51 ok Mar 29 15:13:04 ha, 60x10 nice Mar 29 15:14:02 actually, electronics in lab doesn't work around me...so it's no surprise Mar 29 15:17:50 Crofton|laptop: the team with the lowest noise && lowest power ADC designs wins some beer :) Mar 29 15:18:09 hmmm, you have a better program than ours Mar 29 15:18:43 we get grades as well :) Mar 29 15:18:46 this is some form of class project? Mar 29 15:19:41 course project Mar 29 15:19:56 the ESP course has lab assigments, an exam and 3 projects Mar 29 15:20:03 and beer Mar 29 15:20:06 :) Mar 29 15:20:11 only for the SD project Mar 29 15:24:00 RP,hrw : you might want to merge back the working recipe for apt into poky Mar 29 15:25:22 koen|lab: thx Mar 29 15:26:37 hrw: and add an EXTRA_OECONF_append_poky Mar 29 15:31:53 cu Mar 29 15:55:26 * koen reads http://viewmtn.angrygoats.net/revision/file/577c6b34e5a05584265da13eed21531c2c2a4121/NEWS Mar 29 16:14:37 Keysym not found, for key definition '????Pressed XF86PowerDown' Mar 29 16:15:17 keylaunch fails with that. I remember I've seen it before. is this a problem with my libx11? Mar 29 16:15:35 linux-libc-headers iirc Mar 29 16:17:12 XStringToKeysym("XF86PowerDown") returns 0. But why? XStringToKeysym("F12") works. Mar 29 16:26:18 I would like to put a file both in STAGING_DATADIR and the base package. It looks like do_stage() is working with files that are in the install/${PN}-dev/usr directory. Can I make do_stage() install files from install/${PN} instead? Or can I make a file appear in both places from do_install()? Mar 29 16:28:21 apaulsen: do_stage operates on ${S}, not on ${D} Mar 29 16:29:12 good evening to all Mar 29 16:29:36 hey gremlin[it] Mar 29 16:29:44 hi koen ! Mar 29 16:29:59 apt working, PE working, xorg 7.2, gsmd fixes Mar 29 16:30:13 seems today was a productive day for OE engineers Mar 29 16:32:23 koen: still, it doesn't look like we'll make 2007.1 :) Mar 29 16:32:40 2007.1 was 2 months ago :) Mar 29 16:33:59 koen: I started reading that 2007.Q1 in february :) Mar 29 16:34:22 koen@h2200:~$ uptime Mar 29 16:34:23 16:34:55 up 43 days, 19:01, 1 user, load average: 1.00, 1.00, 1.00 Mar 29 16:34:30 angstrom seems to be pretty stable Mar 29 16:40:39 pH5: what would you want in an angstrom release that isn't present now? Mar 29 16:42:48 X11R7.3 Mar 29 16:42:57 ot: is there a way to compile/emulate openmoko on x86 desktop to programm an applikation for it? Mar 29 16:44:04 pH5: btw, no more diet-x11? Mar 29 16:44:19 mr_nice: MACHINE=native ; bitbake openmoko-devel-image Mar 29 16:44:32 koen: thx Mar 29 16:44:49 koen: I will try that Mar 29 16:45:44 koen: no time for that yet. did you have time to work on your idea of more than two libx11 flavours? Mar 29 16:46:00 not yet Mar 29 16:46:05 I forgot about ut Mar 29 16:46:08 actually, I just notice that even basic libx11-1.1.1 doesn't build correctly Mar 29 16:46:11 thanks for reminding me Mar 29 16:46:29 * koen looks at libx11_1.1.1-r0_armv5te.ipk Mar 29 16:46:29 up to install everything is fine, but package generation breaks. I wonder if PE is confusing shlibs Mar 29 16:46:35 see Mar 29 16:47:20 pH5: http://rafb.net/p/nU1Y3577.html Mar 29 16:48:04 hm.. Mar 29 16:50:18 btw, we need http://www.youtube.com/watch?v=0MUOn_nJmRA for people with usbhost and mice Mar 29 16:50:56 strange. last time my libx11 was empty. but still, shouldn't it be renamed to libx11-6 or something like that? Mar 29 16:51:09 pH5: bitbake -i? Mar 29 16:51:39 bitbake -i generates broken packages in 1.8 and trunk Mar 29 16:51:58 it should be renamed, but it isn't 'cause libx11-xcb is there as well Mar 29 16:52:36 aah, thanks. how can I fix that? Mar 29 16:53:00 (no, not bitbake -i. but I can't reproduce it now. maybe it's just a thinko) Mar 29 16:53:13 PACKAGE =+ libx11-xcb, FILES_libx11-xcb = "" Mar 29 16:53:17 * pH5 imagines 1st person shooters on hx4700 with touchscreen + mouse Mar 29 16:54:23 two handed interaction with gimp or blender would be nice, but not necessarily for handheld devices. Mar 29 16:55:57 koen|away: thanks, that was what I needed Mar 29 17:06:18 hi koen, I'm starting mamona conf files from scratch and I could not figure out how to compile to arm eabi instead of pure arm... I'm trying to read the angstrom files trying to figure it out but without success... can you help me to figure it out? Mar 29 17:06:44 vivijim: look for TARGET_OS Mar 29 17:09:16 koen: thanks for that PACKAGE hint Mar 29 17:10:55 pH5: http://rafb.net/p/3990oY50.html Mar 29 17:10:58 like that? Mar 29 17:11:17 | configure: error: XCB requires xsltproc. Mar 29 17:11:18 koen uhm got it... but I can not understand that expression from angstrom... is the simple corect way to use linux-gnueabi or linux-gnueabi-arm ? Mar 29 17:11:24 NOTE: package libxcb-0.9.93: failed Mar 29 17:11:29 bitbake 1.6 Mar 29 17:11:49 vivijim: linux-gnueabi Mar 29 17:12:01 koen thanks Mar 29 17:12:13 psokolovsky_: seems you need xsltproc-native Mar 29 17:12:18 bbl Mar 29 17:15:00 i.e. it's not PE leak? apparently no Mar 29 17:15:13 and ercipe has DEPEND on it, but commented. weird. Mar 29 17:16:24 aha, so there's no xsltproc-native. cool. Mar 29 17:16:42 vivijim: what is mamona? Mar 29 17:17:00 vivijim: SDK/distro for the nokia tablets Mar 29 17:17:11 ehm Mar 29 17:17:16 s/vivijim/florian/ Mar 29 17:17:23 florian: http://blog.vivi.eng.br/ Mar 29 17:17:41 but there's libxslt-native. weird. Mar 29 17:17:45 florian: sorry: http://labs.vivi.eng.br/blog/?p=10 Mar 29 17:19:42 vivijim: oh, that sounds pretty good :-) Mar 29 17:19:48 florian: we already had a linux from scratch running over qemu to build all deb packages, but now we are moving starting to use OE as our build system to increase or community interaction and avoid monkey work! ;) Mar 29 17:20:21 s/or/our/ Mar 29 17:20:52 vivijim: that's a good idea... OE is something quite a lot projects are involved in Mar 29 17:24:23 yes, we believe so! Mar 29 17:33:02 does anyone know which package provides adds dbm support ? Mar 29 17:33:10 pH5: could you commit the ./xserver-kdrive/optional-xkb.patch that applies to kdrive 1.2.0? Mar 29 17:33:13 gdbm does not work for perl Mar 29 17:33:48 koen florian or someone else: the arm eabi is softfloat by default, but n800 is hardfloat... so how can I say that into my distro? TARGET_FPU_n800="hard" and TARGET_FPU_nokia770="soft" ? it will work? (yes I'm starting to code a conf file to n800 too) Mar 29 17:34:02 vivijim: yes Mar 29 17:34:39 Good evening all Mar 29 17:34:44 hey sirfred Mar 29 17:35:02 koen: Next week I'm borrowing a hx4700 Mar 29 17:35:08 sirfred: cool Mar 29 17:35:15 koen: As I have a pair of holy days, I expect to make some progress. Mar 29 17:35:25 sirfred: ph5 should have instruction on how to boot linux from sd instead of flashing it Mar 29 17:35:39 koen: That should be nice, I don't want to modify my friend's pda Mar 29 17:35:59 So, I don't need to touch the internal flash, do I ? Mar 29 17:36:02 vivijim: interesting... why dies the n800 default to hardfloat? Mar 29 17:36:08 s/dies/does/ Mar 29 17:38:45 florian: I don't have idea.. all I know is that n800 use harfloat vfp2 ... koen: just "hard" or it needs some more specific like "vfp2" ? Mar 29 17:40:09 It seems there're still some problem with gpe backlight management Mar 29 17:40:43 sirfred: oh Mar 29 17:40:46 vivijim: 'hard', eabi defaults to vfp Mar 29 17:41:00 vivijim: ok :-) Mar 29 17:41:09 florian: Oh, you're the florian that accepted my patch, aren't you? :-) Mar 29 17:41:28 sirfred: yes :-) Mar 29 17:41:36 florian: Nice to meet you Mar 29 17:41:42 vivijim: 'require conf/machine/include/tune-arm1136jf-s.conf' Mar 29 17:41:58 in n800.conf Mar 29 17:42:00 sirfred: are there more issues? Mar 29 17:42:01 It seems that when the backlight goes off , it never goes on again, but I'm not sure, I've just experimented it Mar 29 17:42:13 oops Mar 29 17:42:16 florian: I executed gpe-conf task_backlight and says: Mar 29 17:42:17 on 0 Mar 29 17:42:18 vivijim: you can use mx31ads.conf for an example (also 1136 core with vfp) Mar 29 17:42:30 It seems an odd value to be on Mar 29 17:42:32 I need to try this with an up to date image here as well... Mar 29 17:43:01 florian: Perhaps it's just a missconfiguration here Mar 29 17:43:45 A question about gstreamer. Should it inherit gconf, is that needed? Mar 29 17:53:14 hi all Mar 29 17:53:29 mr_nice: hi Mar 29 18:04:01 hi, does open embedded or open moko have a Google SoC proposal regarding Java (phoneME advanced) support? Mar 29 18:08:10 morning Mar 29 18:08:19 vivijim: Poky has n800 config and kernel Mar 29 18:08:28 Hello hrw Mar 29 18:08:45 vivijim: Poky has n800 config and kernel Mar 29 18:09:10 I do not plan to integrate it in OE (lack of hardware to test and dislike of maemo) Mar 29 18:09:31 hi greentux Mar 29 18:09:58 morning all Mar 29 18:10:04 hey cyrilRomain Mar 29 18:10:05 hi hrw, and you have more or less sparetime in your new job? Mar 29 18:10:15 hi koen Mar 29 18:10:20 koen: did you tried dates/openmoko-dates with new package.bbclass? Mar 29 18:10:24 re Mar 29 18:10:25 hi cyrilRomain Mar 29 18:10:36 hrw: yes, but I haven't checked depends yet Mar 29 18:10:43 hi hrw :) Mar 29 18:10:49 greentux: usually less Mar 29 18:10:59 hrw: but can you share it with us? Mar 29 18:11:03 koen: rm tmp/staging/*/shlibs/*dates* first Mar 29 18:11:17 vivijim: Poky SVN is open for anonymous access Mar 29 18:11:22 hrw: thats the freelancer problem Mar 29 18:11:46 # glibc: Mar 29 18:11:46 PREFERRED_PROVIDER_virtual/libiconv ?= "glibc" Mar 29 18:11:46 PREFERRED_PROVIDER_virtual/libintl ?= "glibc" Mar 29 18:11:46 PREFERRED_PROVIDER_virtual/libc ?= "glibc" Mar 29 18:11:46 #NPTL stuff Mar 29 18:11:47 PREFERRED_PROVIDER_virtual/${TARGET_PREFIX}libc-for-gcc = "glibc-intermediate" Mar 29 18:11:49 PREFERRED_VERSION_glibc ?= "2.5" Mar 29 18:11:51 PREFERRED_VERSION_glibc-intermediate ?= "2.5" Mar 29 18:11:53 PREFERRED_VERSION_glibc-initial ?= "2.5" Mar 29 18:11:58 ~pastebin Mar 29 18:12:02 koen, sirfred: are there news on imageon on hx4700? Mar 29 18:12:03 vivi|note: rafb.net/paste Mar 29 18:12:04 extra, extra, read all about it, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Mar 29 18:12:09 pH5: not yet Mar 29 18:12:27 pH5: Next week I'm borrowing one from a friend Mar 29 18:12:32 greentux: but I like that job ;) Mar 29 18:12:52 hrw thats the most important Mar 29 18:13:07 sirfred: ah, good :) Mar 29 18:13:19 pH5: An easy and *fast* procedure to have linux running on a SD ? Mar 29 18:13:32 pH5: Without need to modify anything in the machine, is that possible? Mar 29 18:13:33 sirfred: haret? Mar 29 18:13:49 haret. Mar 29 18:13:55 koen sory about this paste, what I want to know is that is this sufficient to make it build using glibc-2.5... because my build is using glib2.3 and broken :( Mar 29 18:13:57 sirfred: let your friend do full backup first Mar 29 18:14:01 hrw: :) Mar 29 18:14:15 vivi|note: copy the angstrom configs :) Mar 29 18:14:22 03florian 07org.oe.dev * re537e433... 10/ (1 packages/libgpewidget/libgpewidget_0.115.bb): libgpewidget: Add 0.115 Mar 29 18:14:28 03hrw 07org.oe.dev * ra5a1bb0d... 10/ (3 files in 3 dirs): sapwood: fixed packaging, recreated patch Mar 29 18:14:30 hrw: I just expect not to return him a beautiful brick Mar 29 18:14:34 03pH5 07org.oe.dev * r39d93a99... 10/ (1 packages/gimp/gimp_2.3.10.bb): gimp: fix debug packaging Mar 29 18:14:36 vivijim: not copy - include them Mar 29 18:14:38 03pH5 07org.oe.dev * ra169800c... 10/ (4 files in 3 dirs): Mar 29 18:14:38 libgsmd: add TI/HTC GSM chipset support and ldisc switching support for magician Mar 29 18:14:38 - this is a temporary measure and will be reverted once machine plugin support is in place Mar 29 18:14:42 03pH5 07org.oe.dev * rdc90a1a9... 10/ (3 files in 3 dirs): magician: add N_TIHTC to handhelds-linux-2.6 defconfig and machine.conf Mar 29 18:14:49 03pH5 07org.oe.dev * r7bdad482... 10/ (1 conf/machine/magician.conf): magician.conf: autoload n_tihtc.ko Mar 29 18:14:56 03pH5 07org.oe.dev * r9bc9e7c2... 10/ (1 conf/machine/hx4700.conf): hx4700.conf: add voltage switching, bluetooth modules Mar 29 18:15:05 03pH5 07org.oe.dev * r4c365b71... 10/ (3 files in 3 dirs): Mar 29 18:15:05 libgsmd: gsmd/magician: add patch to wait for the interpreter Mar 29 18:15:05 to be ready before sending Mar 29 18:15:13 03pH5 07org.oe.dev * rc017d627... 10/ (1 packages/xorg-lib/libx11_1.1.1.bb): libx11-1.1.1: fix packaging Mar 29 18:15:23 03koen 07org.oe.dev * r07c919d1... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: prefer gcc 4.1.2 as native compiler (cross versions are still 4.1.1) Mar 29 18:15:31 03hrw 07org.oe.dev * rb4020cab... 10/ (5 files in 2 dirs): php: dropped 4.3.6/4.3.10 Mar 29 18:15:43 cia awake finally Mar 29 18:15:45 go cia Mar 29 18:16:05 hrw: sorry, where should I include? Mar 29 18:16:38 heh. I'm third active dev on cia.vc, that's a first. Mar 29 18:16:41 vivijim: if you use own distro then you can copy from angstrom ones or include angstrom one in your distro config Mar 29 18:16:42 03pH5 07org.oe.dev * rbb10d8a3... 10/ (2 files in 2 dirs): xserver-kdrive-1.2.0: missing optional-xkb.patch Mar 29 18:16:44 I see that haret is a bootloader that runs from windows CE. Mar 29 18:16:46 03koen 07org.oe.dev * r37479993... 10/ (2 files in 2 dirs): kdrive 1.2: rediff optional-xkb.patch Mar 29 18:16:53 heh Mar 29 18:17:02 that's probably a sign I should think more before committing. Mar 29 18:17:02 sirfred: exactly Mar 29 18:17:07 ~haret Mar 29 18:17:11 methinks haret is HAndheld Reverse Engineering Toolkit, available at anoncvs@cvs.handhelds.org:/cvs/haret (CVS sources) or projectpage: http://www.handhelds.org/moin/moin.cgi/HaRET Mar 29 18:17:15 That will meet my needs. Mar 29 18:17:21 mrdata: hi Mar 29 18:17:33 ~You are wise Mar 29 18:17:35 sirfred: what are you talking about? Mar 29 18:17:39 nothing... Mar 29 18:17:44 sirfred: this and card which will have rootfs on it Mar 29 18:18:04 mrdata: I have tryed to find the bug but I didn't found out much Mar 29 18:18:22 mr_nice: hi, not so bad Mar 29 18:18:23 hrw: nice Mar 29 18:18:53 sirfred: best way is sd card with two partitions: p1/vfat with kernel and haret, p2/ext2 with rootfs Mar 29 18:19:05 mr_nice: the bug would not be trivial, so we need some time for it Mar 29 18:19:09 * sirfred takes notes Mar 29 18:19:28 sirfred: so pocketpc will use p1 and allow to run haret (which will read kernel from card) and kernel need to have root=/dev/mmcblk0p2 in cmdline Mar 29 18:19:31 hi Lorn Mar 29 18:19:42 hrw: Seems reasonable Mar 29 18:20:11 sirfred: the funniest thing is that I never owned pocketpc device, never had such one for longer then 20 minutes in hands and never ever saw haret working ;D Mar 29 18:20:11 hrw: I've seen you closed my bug about starling. Have you been able to make it run on a zaurus? Mar 29 18:20:15 sirfred: that's the way I always do it. and if there is wm2005 on the hx4700, you don't even have to backup the data. Mar 29 18:20:22 mrdata: here is my latest pastebin http://pastebin.ca/414854. I also have messed up the adress of r3 now. this is a result of adding {} and a printk inside the brackets Mar 29 18:20:28 hrw: ok, I can include the angstrom-glibc.inc but I believe that I'm missing something.. I believe that even including I'll get the glibc2.3 instead of 2.5 Mar 29 18:20:31 sirfred: I built it for neo before but it does not run Mar 29 18:20:55 hrw: I'm getting a segfault in play_list_set_sink, but perhaps it's a gstreamer issue. Mar 29 18:21:02 vivijim: angstrom use glibc 2.5 on arm Mar 29 18:21:05 I'm getting very ugly warnings. Mar 29 18:21:17 sirfred: starling is from gstreamer 0.8 times Mar 29 18:21:19 pH5: ok. Mar 29 18:21:21 hrw: oh Mar 29 18:21:33 hrw: starling svn should be 0.10 ready Mar 29 18:21:38 neal rewrote it :) Mar 29 18:21:46 koen: but OE one is not ;D Mar 29 18:21:48 koen: I'm using starling_svn.bb I've just created. Mar 29 18:21:54 someone need to update it then Mar 29 18:21:57 koen: And I'm getting the same sigsegv. Mar 29 18:22:00 03koen 07org.oe.dev * r915e7bca... 10/ (2 files in 2 dirs): disapproval of revision '3747999334cad4bb22159396976a802d18c50703' Mar 29 18:22:06 hrw: I can open a bug with my .bb Mar 29 18:22:13 Doing it in a minute Mar 29 18:22:22 ok, I will add it then tomorrow probably Mar 29 18:22:29 sirfred: does svn version works? Mar 29 18:22:39 hrw: No, same errors as I've said Mar 29 18:22:41 mrdata: I cannot find out hte right definitions for struct resource and struct device which are part of the struct platform_device. I only fund something linke struct device; without any information about the proper types. so I was not able to add printks for that values Mar 29 18:23:00 hrw: But my gstreamer install, well, postinstall is trying to use gconftool-2 for registering something. Mar 29 18:23:10 I don't know if its state on my zaurus is good Mar 29 18:23:30 hrw: I was wondering if inherit gconf is needed for gstreamer plugins. Mar 29 18:23:42 sirfred: there is gst-plugin-gconf Mar 29 18:24:02 I do not know who use it Mar 29 18:24:07 hi _law_ Mar 29 18:24:20 hrw: I'm a totally newbie in gstreamer. Mar 29 18:24:32 mrdata: do you know what programm counter means? I found that as a definition for PC Value Mar 29 18:24:37 sirfred: I was doing some work on gstremer lately but it was not x11 based Mar 29 18:24:38 <_law_> hi hrw Mar 29 18:25:09 mr_nice: and in most programming environments PC === programm counter Mar 29 18:25:48 mr_nice: struct device is defined in /include/linux/device.h search for "struct device {" Mar 29 18:26:53 hrw: It is bug 2063 Mar 29 18:27:20 !oebud 2063 Mar 29 18:27:26 !oebug 2063 Mar 29 18:27:27 * * Bug 2063, Status: NEW, Created: 2007-03-29 11:31 Mar 29 18:27:28 * * manuel.teira(AT)telefonica.net: New starling svn package Mar 29 18:27:29 * * http://bugs.openembedded.org/show_bug.cgi?id=2063 Mar 29 18:27:34 mr_nice: for sa1100 i must read the manual before i can it say right, but i know that the programm counter Mar 29 18:28:36 mrdata: thx I found the definition Mar 29 18:28:37 mr_nice: holds the value of the address of the actual order Mar 29 18:29:40 mrdata: ah, so the adress is wrong? Mar 29 18:30:32 mrdata: I will add more prinks and see what happens then. Mar 29 18:33:09 mr_nice: the program counter should be in the range from 0 to 64MB inside the address range of the sa1100 Mar 29 18:34:22 gpe-session-scripts RDEPENDS on matchbox-panel (>= 0.9.2-r12), but angstrom-x11-image do_rootfs fails because it doesn't think matchbox-panel 0.9.3-r2 satisfies that dependency :( Mar 29 18:35:14 pH5: You can remove the version numbers Mar 29 18:35:28 as soon as the scsi controller stops being an ass I can commit it Mar 29 18:35:36 Cpu0 : 0.0%us, 0.0%sy, 0.0%ni, 0.0%id,100.0%wa, 0.0%hi, 0.0%si, 0.0%st Mar 29 18:35:37 Cpu1 : 0.0%us, 0.0%sy, 0.0%ni, 0.0%id,100.0%wa, 0.0%hi, 0.0%si, 0.0%st Mar 29 18:39:32 hi Mar 29 18:39:54 hello all Mar 29 18:40:07 hi Matt Mar 29 18:40:20 mreimer_: Hello Mar 29 18:42:50 Would anyone here be interested in a little shell script that makes psplash fill the progress bar on a timer? It writes out how long a boot-up takes (say, 46 seconds) and then on the subsequent boot it fills the progress bar so that it's full by that time Mar 29 18:43:19 makes it way easier than trying to psplash-write in all the initscripts Mar 29 18:45:53 koen: ok, that helped Mar 29 18:45:57 hi mreimer_ Mar 29 18:47:41 hey mreimer_ Mar 29 18:48:13 mreimer_: I am no oe dev or something but doing that is not a good idea I think. e.g if the system hangs on dhcp request or if it fscks something the bar gets useless then. Mar 29 18:49:01 03koen 07org.oe.dev * r56b25a11... 10/ (1 packages/gpe-session-scripts/gpe-session-scripts_0.67.bb): gpe-sessionscripts: remove PV in depends Mar 29 18:49:07 mr_nice: yes, that's the downside of it Mar 29 18:50:00 FWIW, nokia uses that method on their tablets Mar 29 18:51:23 mr_nice: like zecke yesterday said, i had begin with a fresh vanilla kernel again Mar 29 18:51:41 mr_nice: Unable to handle kernel paging request at virtual address 645f6370, same by me now Mar 29 18:52:32 mr_nice: like zecke yesterday said, i had begin with a fresh vanilla kernel again Mar 29 18:52:48 mr_nice: Unable to handle kernel paging request at virtual address 645f6370, same by me now Mar 29 18:53:26 hi all, what is the package responsable to setup (or not) a valid root password ??? Mar 29 18:53:33 hrw: your INTERNAL thingy works. Thanks! Mar 29 18:53:51 mrdata: yes, and I am was able to destroy that adress by adding a {... printk} to the if statement Mar 29 18:54:03 PREFERRED_PROVIDER_virtual/${TARGET_PREFIX}libc-for-gcc = "glibc-intermediate" Mar 29 18:54:09 what this virtual means? Mar 29 18:54:38 cu all Mar 29 18:54:44 vivi|note: it means that there are more providers for it (e.g. uclibc, glibc, newlib, etc) Mar 29 18:54:51 klibc Mar 29 18:56:16 uhm... because I'm including angstrom-glibc.inc and using this options like angstrom does, but it still insist in using glibc2.3 :( and I don't have more ideas about how to change this... Mar 29 18:56:43 do you have any idea where can I making a mistake? Mar 29 18:57:50 what is the package responsable to setup (or not) a valid root password in building stage, not in runtime startup/config of target ? Mar 29 18:59:19 mrdata: It is hard for me to add all the needed prinks there are a lot of things I don't know which %{s,i,d...} I need. But I will try to get a complete printk output of the struct Mar 29 19:00:56 mrdata: :) http://kernelnewbies.org/FAQ/LinuxKernelDebug101 -> dev_info Mar 29 19:02:29 please no one can help me ??? Angstrom don't work on ipaq h3600 ... i cannot build the debug image cause it's too big ... but normal one have some but and inital root password setup con't run ... so i have a system wgere i cannot login !!! Mar 29 19:02:49 s/some but/some bug/ Mar 29 19:05:34 gremlin[it]: I'm just guessing - try running with kernel param init="/bin/busybox sh" Mar 29 19:05:37 or something like that... Mar 29 19:05:40 mr_nice: the struct expanding is like a game, i need realy a posibility to use a debugger Mar 29 19:06:59 ok let's try mmp, thanks ... Mar 29 19:07:02 mr_nice: then we had a clean view of the result from the code compiling (adresses and so one) Mar 29 19:07:35 mmp ... if build in debug more final image already have a empty (valid) root password ... :) Mar 29 19:08:10 gremlin[it]: but why there would be some password for non-debug image? Mar 29 19:08:39 mrdata: the problem is to find one for arm Mar 29 19:09:15 mmp yes ok .. but if i know the package that do i can force this one only to beuild in debug mode ;) Mar 29 19:10:25 gremlin[it]: hmm, yes, but I'm afraid I can't help you with this much, I don't know angstrom very well.. Mar 29 19:11:22 mrdata: hm kdb seems to be available for arm Mar 29 19:11:54 mrdata: https://gforge.inria.fr/plugins/scmsvn/viewcvs.php/vanilla-kernels/linux-2.6.20/Documentation/kdb/kdb_bp.man?rev=1341&root=kerrighed&view=markup Mar 29 19:11:59 mr_nice: for the right usage of %{s,i,d ...} -> http://home.fhtw-berlin.de/~junghans/cref/FUNCTIONS/format.html Mar 29 19:17:26 mrdata: my problem is that sometimes there are kernel own type definitions. and I don't know if the types are big enough to keep the data. or if they make sense. Mar 29 19:17:28 mr_nice: thx!, struct resource is defined in /include/linux/ioport.h Mar 29 19:17:40 mrdata: thx Mar 29 19:18:05 mrdata: I will read the kdb manual Mar 29 19:18:37 mrdata: found a good start http://www-128.ibm.com/developerworks/linux/library/l-kdbug/ Mar 29 19:18:57 mr_nice: any type definitions should be a part of 8 Mar 29 19:21:42 mr_nice: and than came the processor part (8; 16; 32 bit) organization, plus c-language how big is a byte, word, int, long ... Mar 29 19:22:44 mr_nice: and the operating-system-logic (16; 32 bit) Mar 29 19:23:16 mrdata: thx, I will read that definitions a second Mar 29 19:42:20 03florian 07org.oe.dev * r7c241467... 10/ (1 packages/tasks/task-sdk.bb): task-sdk: Update various bits to current names. Mar 29 19:42:25 03florian 07org.oe.dev * r5a68fd7d... 10/ (1 packages/meta/meta-sdk-sbox.bb): Mar 29 19:42:25 meta-sdk-sbox: Start of a scratchbox based SDK. Mar 29 19:42:25 Currently creates a rootstrap which can be used after importing the crosscompiler. Mar 29 19:42:25 Compiling seems to work but running applications doesn't. Mar 29 19:45:31 no way ... no enought space to write /etc/shadow :( .... grrrrrr Mar 29 19:45:54 ~ibot last blowing libraries Mar 29 19:45:58 ~ibot lart blowing libraries Mar 29 19:46:07 * ibot shoots blowing libraries in the head Mar 29 19:46:53 gremlin[it]: remove some not-so-neccessary packages... Mar 29 19:47:09 or libraries:) Mar 29 19:48:58 i try ... but i don't know exactly what to delete ... i cut off about 1M of /usr/local/share/X11 ... but still no space :( Mar 29 19:51:01 pH5: INPUT_MISC ?= m - is this for UINPUT? Mar 29 19:51:48 psokolovsky_: yes Mar 29 19:51:54 ok Mar 29 19:56:16 03koen 07org.oe.dev * r33ef9583... 10/ (1 packages/meta/meta-sdk-sbox.bb): meta-sdk-sbox: remove hardcodes references, reuse functions instead of copy-pasting them, and we have an SCM, so no need to comment stuff Mar 29 19:56:34 gremlin[it]: weird.. Mar 29 19:56:44 1MB should be enough for such simple operation Mar 29 19:58:00 yes .... seem nothing change if i delete things :( Mar 29 19:58:55 NOTE: preferred version 2.5 of glibc not available NOTE: multiple providers are available (glibc, glibc-intermediate); NOTE: consider defining PREFERRED_PROVIDER_virtual/arm-linux-gnueabi-libc-for-gcc NOTE: package linux-libc-headers-2.6.15.99: started NOTE: package linux-libc-headers-2.6.15.99: completed NOTE: package glibc-2.3.5+cvs20050627: started Mar 29 19:59:26 koen any idea why the build is using glib2.3 insteadof glib2.5 Mar 29 19:59:37 vivijim: try including angstrom-2007.1.conf Mar 29 20:00:04 koen: http://bugs.webkit.org/show_bug.cgi?id=11308 building WebKit is a mess Mar 29 20:00:12 koen: specially if you don't use xcodebuild :} Mar 29 20:00:21 zecke: duh Mar 29 20:00:28 told you a few weeks ago Mar 29 20:01:04 koen: even the cmake stuff is orphaned Mar 29 20:01:49 ok ... but removing a big library (gtk) ... but i can do nothing without it !!!! Mar 29 20:30:07 ok finallu h3600 barelly run Angstrom Mar 29 20:40:32 hello all Mar 29 20:41:32 hi likewise! Mar 29 20:41:50 hi HopsNBarley Mar 29 20:44:11 HopsNBarley: I'll be in CA from 15 to 21 april Mar 29 20:44:18 celf confirmed my flight(s) Mar 29 20:44:54 koen travel-pingeon :) Mar 29 20:47:11 koen, you fly direct to west coast? Mar 29 20:48:50 Crofton|laptop: no, chicago in between Mar 29 20:48:59 can't really complain, though Mar 29 20:49:10 more time to read books :) Mar 29 20:49:26 I'm no where near chicago :) Mar 29 20:51:05 What should my bb look like to make it fetch CVS head on every build? Mar 29 20:54:32 likewise: do_configure_prepend() {cvs up -dP} Mar 29 20:54:38 that's the best I can come up with Mar 29 20:55:13 other ways are fumbling with CVSDATE or setting nostamp on fetch Mar 29 20:56:30 SRCDATE now for cvs? Mar 29 20:57:11 zecke_: that doesn't work as you'd expect (see the openmoko hacks) Mar 29 20:59:43 it works as advertised Mar 29 21:00:19 koen: or to liberally cite Chris: "I concentrated on a couple of usecases, this was not one of them" Mar 29 21:00:25 Any reason for gstreamer plugins being configured with --disable-x ? Mar 29 21:00:44 sirfred: no good reason Mar 29 21:00:49 sirfred: they can be used on directFB only Mar 29 21:01:27 zecke_: What do you mean? Mar 29 21:01:36 zecke_: Is there any reason preventing them to be used on x? Mar 29 21:02:08 sirfred: do you want to compile directFB if you build X? Do you want it the other way around? Mar 29 21:02:41 zecke_: No, but you don't need to install the x related plugins in a DirectFB environment, isn't that enough? Mar 29 21:02:55 sirfred: http://www.directfb.org/images/RIP-X.png :) Mar 29 21:03:11 zecke_: :) Mar 29 21:03:15 zecke_: Who are they? Mar 29 21:03:53 sirfred: Convergence folk (left-to-right) Neo (gimp), ?, Dok, Mitch? (Gtk/Gimp) Mar 29 21:04:27 sirfred: I think tmbinc added gstreamer, he is not using X at all. Forcing him to compile X will ~tripple his build times Mar 29 21:04:42 sirfred: if you want to build X, would you be happy to build QtopiaCore and directFB all the time? Mar 29 21:04:47 zecke_: Perhaps we could have different bb files. Mar 29 21:05:06 zecke_: No, of course. But what about gst-plugins-good-x and gst-plugins-good-directfb packages? Mar 29 21:05:21 sirfred: packages or different recipes? Mar 29 21:05:28 zecke_: Recipes, I mean Mar 29 21:06:10 sirfred: I think this is one of the fundamental issues we have to solve in OE. (automatic feature configuration based on 'flags') Mar 29 21:06:20 *agree* Mar 29 21:06:26 zecke_: That should be nice, yes. Mar 29 21:06:42 some global customization parameters Mar 29 21:06:55 he tmbinc :) Mar 29 21:06:58 tmbinc_: we need to help you merging, if I just would have more *time* :} Mar 29 21:07:02 like a global "debug" flag, which allows packages to automatically build non-debug builds (not just stripping) Mar 29 21:07:15 zecke_: if i just had more time, i'd be happy to merge :) Mar 29 21:07:23 ;) Mar 29 21:08:02 sirfred: create a recipe that provides the additional X sinks Mar 29 21:08:12 So, the feature we need in OE is a time expander. Mar 29 21:08:40 * zecke_ was happy to see Wim added proper MP4 container parser and h264 depayloading to GStreamer Mar 29 21:09:49 hm, wouldn't fixing gstreamer be better? Mar 29 21:09:53 to not rely on X? Mar 29 21:10:07 it would then autodetect if X is there or not. I just added --disable-x because it wouldn't compile otherwise Mar 29 21:10:08 tmbinc_: THere are things like xv sinks that are X specific Mar 29 21:10:18 gstreamer's configure should be smart enough to disable these sinks then Mar 29 21:10:33 It might even be already fixed Mar 29 21:10:40 hi Mar 29 21:10:58 you could then work out a mechanism so that X is only build before if you want X Mar 29 21:11:00 tmbinc_: It disables them, if it doesn't find the dependencies. Mar 29 21:11:10 tmbinc_: well, generally we want to avoid those coincidences. We want to have a tight control on what features get build Mar 29 21:11:18 tmbinc_: If you don't have libxv , then the xv sink is not compiled. Mar 29 21:11:25 sirfred: it didn't, back when i tried. otherwise i wouldn't had added the --disable-x Mar 29 21:11:37 it was a "get this thing compiling, *now*"-type of "fix" Mar 29 21:11:38 tmbinc_: ok Mar 29 21:11:57 tmbinc_: I have compiled them without that flag. Mar 29 21:12:15 maybe it's already fixed? or was only relevant for one of the plugin packages? Mar 29 21:12:16 tmbinc_: But I have libxv and friends in the staging, of course. Mar 29 21:12:21 oh ok Mar 29 21:12:27 well i didn't Mar 29 21:12:41 tmbinc_: I don't know what would happen if I didn't have that staged Mar 29 21:13:51 zecke_: true. but --disable-x is superflous, isn't it? X is either there or not. If it's not there and should be, that's a bug in OE. If it's not there and gstreamer fails to detect that, it's a bug in gstreamer. We should not override the whole configure mechanism just because something *might be* buggy - or do i miss the point? Mar 29 21:14:18 tmbinc_: well. e.g. angstrom could support directFB and X based UIs Mar 29 21:14:39 tmbinc_: and then gstreamer support would depend on the order of the build Mar 29 21:14:42 maybe a post check if the wanted modules were built? but that's part of the image building process, which should depend on the required modules. If something is incorrectly not build, then it will be detected there. Mar 29 21:14:58 zecke_: hm right. then gstreamer should depend on something like "x_or_not_x" Mar 29 21:16:17 tmbinc_: E.g. all possible dependencies should build a graph. Each node is connected with a maybe-edge. Externally these maybe-edges can be defined and then we replace these maybe-edges with true edges/DEPENDS and if nothing more is to do, we are done and can save the configuration :} Mar 29 21:16:31 tmbinc_: as to disable debug completely, there is a global CPPFLAG for that Mar 29 21:17:11 zecke_: i'm not talking about -g or something. gstreamer for example has a --disable-debug (which is non-default) Mar 29 21:21:16 can anyone help me a bit with bootstrap-image? I'm getting the output here: http://pastebin.ca/415812 Mar 29 21:23:01 nail hm seems some quoting is wrong Mar 29 21:23:20 yeah, but where does it come from? Mar 29 21:23:43 bad .bb file or something in your local.conf? Mar 29 21:25:53 why does it bomb out in bootstrap-image? Mar 29 21:26:02 task-base builds just fine Mar 29 21:41:35 koen: do_configure_prepend() {cvs up -dP} would not work if stamps are still made Mar 29 21:53:16 hi Mar 29 21:57:56 Good night Mar 29 22:29:58 koen|away: http://www.satine.org/archives/2007/03/05/the-skia-source-code-dilemma/ Mar 30 00:31:00 evening gents Mar 30 01:35:58 What is the difference between qte and qte-mt? Mar 30 01:36:06 Which one should I prefer? Mar 30 01:36:22 qte-mt-static fails to build horribly (complete bitbake crash) Mar 30 02:02:33 03coredump2 07org.oe.dev * ra9934146... 10/ (3 files in 2 dirs): altboot: Add symlink for /sbin/altboot Mar 30 02:29:20 03Laibsch 07org.oe.dev * re7bda6cb... 10/ (1 packages/btxml/btxml.bb): btxml: clean up bb file (reorder fields). no semantic change, no PR bump. Mar 30 02:29:25 03Laibsch 07org.oe.dev * r338ea31d... 10/ (7 files in 2 dirs): gpe-bluetooth: unify **** ENDING LOGGING AT Fri Mar 30 02:59:56 2007