**** BEGIN LOGGING AT Sun Jun 26 23:59:57 2005 Jun 27 00:04:08 anyone here use a zaurus sl-c3100? Jun 27 00:05:09 yes lots do Jun 27 00:05:15 personally . no Jun 27 00:05:30 well maybe not lots 3-5 Jun 27 00:05:46 which is a fair amount concidering the cost of it Jun 27 00:06:38 it is quite expensive Jun 27 00:07:19 I just wanted to as a few questions of some current owners Jun 27 00:09:41 ask someone might awnser, maybe log the channel and check it later Jun 27 00:09:52 both are good methods Jun 27 00:11:16 there are som helpful people in #zaurus now, but thank you for the advice emte Jun 27 00:17:04 np Jun 27 00:26:08 morning all Jun 27 00:26:24 morning Jun 27 00:35:34 hi all Jun 27 00:38:56 hi bluelightning Jun 27 00:39:03 hi RP Jun 27 00:40:45 morning Jun 27 00:41:05 hey hrw|work Jun 27 00:41:17 do you have updated dejavu? Jun 27 00:41:21 in OE Jun 27 00:41:32 gints|wrk: yes 1.11 Jun 27 00:41:46 ... :( 1.12 is coming... Jun 27 00:41:48 gints|wrk: 1.12 is 2005-07-31 scheduled Jun 27 00:42:50 in days... Jun 27 00:44:45 hmmm... ah ok.. it's postponed... Jun 27 00:44:46 it is a month Jun 27 00:46:16 we have dejavu? oops got confused with dejagnu for a second Jun 27 00:56:48 morning hrw Jun 27 00:57:18 hrw|work: Did you try that sysfs battery fix btw? Do you still see the oopes/charging problem? Jun 27 01:01:59 RP: not yet Jun 27 01:02:48 good morning all Jun 27 01:02:55 hi koen Jun 27 01:03:33 morning koen Jun 27 01:04:07 hrw|work: Let me know if any problems remain Jun 27 01:04:16 hey hrw|work & RP Jun 27 01:04:43 RP: ok Jun 27 01:05:13 RP: now I'm trying to get some 3.5.3 upgrades for users - month passed without any Jun 27 01:05:51 hrw|work: Ok, sounds good Jun 27 01:07:19 I'm concious we need to try and make 2.6.12 fit for users now. The main issues would appear to be the CF memory card handling... Jun 27 01:08:07 brb Jun 27 01:08:41 Ä…Ä…Ä… Jun 27 01:08:46 re Jun 27 01:09:02 that was quick :) Jun 27 01:10:09 hi married man Jun 27 01:10:30 ~hail mickeyl for bitbakesh Jun 27 01:10:30 * ibot bows down to mickeyl for bitbakesh and chants, "I'M NOT WORTHY!!" Jun 27 01:10:41 morning Jun 27 01:11:06 bah, I come back to work and find the linux RNDIS gadget driver is busted :-( Jun 27 01:11:06 what's bitbakesh? Jun 27 01:11:09 hi XorA Jun 27 01:11:30 bluelightning: bitbake -i Jun 27 01:11:36 bluelightning: bitbake -i Jun 27 01:12:42 wow, cool :) Jun 27 01:13:03 hi XorA. How it all went well :) Jun 27 01:13:49 s/How/Hope/ Jun 27 01:13:57 * RP needs more caffine Jun 27 01:13:59 RP: went perfect cheers Jun 27 01:14:05 morning all Jun 27 01:14:25 hey ade|desk Jun 27 01:14:30 hi ade|desk Jun 27 01:14:50 RP: RNDIS on 2.6.12-mm1 is hard lockup, I have no serial cable to see an messages :-( Jun 27 01:15:11 hard lockup! ? eek :-( Jun 27 01:15:49 I shall once again blame preempt :) Jun 27 01:15:55 XorA: works for me Jun 27 01:16:04 hrw: RNDIS???? Jun 27 01:16:11 hrw: since when did you use windows Jun 27 01:16:24 agh.. usbnet works Jun 27 01:16:37 hrw: yeah usbnet works, RNDIS is hard lockup Jun 27 01:16:51 even alt-sysreq-b doesnt work Jun 27 01:16:55 ah.. Jun 27 01:17:05 sorry then Jun 27 01:17:42 Debian goes X.Org finally.. Jun 27 01:17:53 XorA: There have been a few nasty kernel changes resulting in a big speed improvement... Jun 27 01:18:42 RP: Ill just have to teach XP, usbnet :-) Jun 27 01:18:57 XorA: Good luck ;-) Jun 27 01:19:19 XorA: If you can pinpoint when it broke, that would probably help... Jun 27 01:20:33 RP: Ill have to make that serial cable finally Jun 27 01:20:35 03pb 07 * r1.3632 10openembedded/packages/alsa/ (4 files in 2 dirs): add handhelds.org alsa driver, courtesy Pierluigi Frullani Jun 27 01:21:35 XorA: argh.. I have to send parts for it to my friend.. Jun 27 01:21:55 hrw: I have the parts, just need 30 mins to solder it Jun 27 01:22:44 ah.. Jun 27 01:23:02 XorA: that fscking tcx3106 too? Jun 27 01:23:03 03pb 07 * r1.3633 10openembedded/classes/kernel.bbclass: Jun 27 01:23:03 also stage headers from drivers/sound, if it exists. Jun 27 01:23:03 patch courtesy Pierluigi Frullani Jun 27 01:23:18 hrw: yes, got one from my old sync cable I borke Jun 27 01:27:15 XorA: I have 3-4 unused ones - just have to package all and send to friend who will make serial cables from it. one for Bernardo, rest for other devs which need Jun 27 01:29:02 XorA: The other way to debug might be to switch to a console when plugging in RNDIS Jun 27 01:29:19 See if the oops shows on the framebuffer Jun 27 01:30:06 RP: it does Jun 27 01:34:49 * XorA needs to relearn debuggin, after 2.5 weeks my brain has forgotten all :-) Jun 27 01:35:10 ~seen zecke Jun 27 01:35:10 zecke <~ich@83-169-170-199-dynip.superkabel.de> was last seen on IRC in channel #handhelds.org, 19d 9h 17m 44s ago, saying: 'good nite'. Jun 27 01:35:20 ~summon zecke Jun 27 01:35:20 apt takes out 20 clean, identical-looking phones, some extra hands, and pretends to be a telemarketer for a large corporation, so he gets delivered a phonelist containing zecke's coordinates Jun 27 01:46:08 hrw: I hope that zecke isn't again ill ... Jun 27 01:46:28 hrw: /me was/is ill (partially broken knee plate) Jun 27 01:48:04 i am compiling gpe-image for osk5912. wanted to know the image works straight away with touch screen. Jun 27 01:48:14 has any one trie gpe-image for osk 5912 Jun 27 01:55:15 hello ? Jun 27 01:56:56 ram123: it should Jun 27 01:57:28 straight away with touchscreen support ? Jun 27 01:57:44 yes Jun 27 01:57:52 fine Jun 27 01:58:12 just wanted to know the development process of openembedded ?. Jun 27 01:58:37 urs is a company which ports kernel and utilities to different processore ? Jun 27 01:58:53 or just a bunch of guys who do that in thier free time ? Jun 27 01:59:10 bunch of guys Jun 27 01:59:29 we do use openembedded in companies Jun 27 01:59:45 and u do all these in ur free time ? Jun 27 02:01:55 mostly, yes Jun 27 02:01:56 03mickeyl 07 * r1.3634 10openembedded/packages/opie-pcmciaapplet/ (opie-pcmciaapplet.inc opie-pcmciaapplet_1.2.1.bb): fix opie-pcmciaapplet preference Jun 27 02:02:27 unfortunately, yes :) Jun 27 02:02:56 * mickeyl still looking for a company that pays him to work on OSS Jun 27 02:03:25 ram123: free time Jun 27 02:03:34 ram123: or free-at-work time Jun 27 02:06:04 great man Jun 27 02:06:12 i dont know what to say. Jun 27 02:06:23 no words to describe the things u do. Jun 27 02:06:32 u people are simple great. Jun 27 02:06:40 it makes life simple for us. Jun 27 02:19:15 koen: were you able to resolve the multiple heads problem on ewi? Jun 27 02:20:01 did pulling to a new db and using that instead help? Jun 27 02:20:20 rwhitby-treo: haven't tried that yet Jun 27 02:21:21 should be quick to do if it's on the same machine ... Jun 27 02:22:10 but no problem or hurry - the db which you get when you pull from ewi only has a single head Jun 27 02:22:39 so it only affects viewmtn Jun 27 02:23:19 the alternative is just to wait till the 1st July I guess. Jun 27 02:24:08 * XorA must remeber to learn monotone for July 1st Jun 27 02:39:02 good morning pb_ Jun 27 02:39:56 mickeyl: good morning! Jun 27 02:40:49 drat, my kernel has gone wrong, gotta reboot. Jun 27 02:40:51 * pb_ stabs linux Jun 27 02:42:39 how do u convert a zImage into a uImage (U-Boot) ? Jun 27 02:44:14 it should have made an uImage out of it Jun 27 02:44:34 but take a look at packages/linux/linux-omap1_2.6.12_rc2.bb Jun 27 02:45:04 brb Jun 27 02:45:21 the do_deployy_omap5912osk() function handles the uImage Jun 27 02:45:24 when i made gpe-image i get an zImage in /boot ? Jun 27 02:45:54 not sure on that Jun 27 02:46:33 ±êæ³ Jun 27 02:46:38 uf. Jun 27 02:46:41 re Jun 27 02:47:04 and bye Jun 27 02:47:39 ok i dont have a uImage in ext2.gz file. how do i make it manually ? Jun 27 02:48:34 see the do_deployy_omap5912osk() in packages/linux/linux-omap1_2.6.12_rc2.bb Jun 27 02:49:53 that's better Jun 27 03:00:53 XorA: is Monotone the new SCM for OE ? Jun 27 03:01:26 schurig: read the mailinglist Jun 27 03:01:35 schurig: people seem to be going that way, I have a small personal repository of bb's in it just to learn Jun 27 03:01:50 schurig: http://thread.gmane.org/gmane.comp.handhelds.openembedded/4853 Jun 27 03:01:59 not suscribed at home (I've a broken knee plate and don't work currently) Jun 27 03:02:22 schurig: ouch, hope that heals soon Jun 27 03:06:51 * schurig has no real opposition to Monotone after mickeyl checked it's speed (it improved lately) Jun 27 03:15:06 03pb 07 * r1.3635 10openembedded/conf/machine/ipaq-common.conf: prefer xserver-kdrive on all ipaqs Jun 27 03:16:59 koen: so any problems from your end with regard to nslu2-linux's role in the new monotone setup? Do you forsee any problems in the branch structure, or the procedures we're going to follow to push stuff back to org.openembedded.dev ? Jun 27 03:17:15 03koen 07 * r1.3624.2.1 10openembedded/ (3 files in 3 dirs): gpe-contacts_0.39.bb: fix icon problem, courtesy Florian Boor Jun 27 03:17:24 rwhitby-away: your proposal sounds good Jun 27 03:20:55 just booted gpe-image for osk 5912 Jun 27 03:21:13 i am notgetting x server on the display ? Jun 27 03:21:25 how do i test it ? Jun 27 03:21:43 the startup messages are at http://pastebin.com/304582 Jun 27 03:28:03 hello Jun 27 03:28:05 what should i do ? after i get the shell prompt. Jun 27 03:28:08 ram123: I see no real errors in that Jun 27 03:28:09 does the gpe start by itself after the boot sequence is over or how is it ? Jun 27 03:28:37 it starts itself Jun 27 03:29:33 i dont even see the X mark on the touch screen. Jun 27 03:30:01 what tests should i carry out to analyse the problem ? Jun 27 03:30:07 it should start xtscal first Jun 27 03:30:25 shall i run it manually ? Jun 27 03:31:02 you can try that Jun 27 03:31:09 make sure X is running Jun 27 03:31:24 1 sec Jun 27 03:32:47 i cant find xserver Jun 27 03:33:22 it should be in /usr/bin/Xfbdev Jun 27 03:37:26 get the following error when i run Xfbdev http://pastebin.com/304583 Jun 27 03:38:41 can you symlink your framebuffer to /dev/fb0? Jun 27 03:41:36 <_chronic> pb_: you were kind enough to name yourself in libxml2_2.6.10.bb. I'm trying to get an updated version of this working (libxml2_2.6.19), currently (arm) compile-time borking with "undefined reference to `inithtmlDefaultSAXHandler'". I've updated the header variable settings & looked at the configure options, but can't quite get past this. any thoughts/comments would be appreciated. Jun 27 03:42:40 i am using osk 5912. did not get u ? Jun 27 03:44:29 ram123: it needs a framebuffer and can't seem to find it Jun 27 03:47:57 _chronic: it sounds like you may need to modify the configure options a bit, to either enable the feature that provides that function, or disable the feature that requires it. This is a fairly common problem with libxml: it looks like the upstream maintainers only test it in the default configuration. Jun 27 03:51:26 did not get u which frame buffer u are taling about ?. i also use /dev/fb0 only Jun 27 03:51:51 is not the image built with frame buffer drivers ?. Jun 27 03:52:01 what should i do to make it work ?. Jun 27 03:53:53 <_chronic> pb_: thanks for the tip. will keep looking into config options. Jun 27 03:54:18 ram123: try doing 'cat /proc/fb' Jun 27 03:57:47 i dont have a /proc/fb Jun 27 03:57:49 there is no such file named /proc/fb ? Jun 27 04:00:08 ram123: try replacing the defconfig in packages/linux/linux-omap1-2.6.12-rc2/omap5912osk with a defconfig that has a framebuffer enabled Jun 27 04:01:37 how do i see if the module is there in the filesystem ?. Jun 27 04:01:54 check /lib/modules Jun 27 04:01:59 i am using a lcd and a touch screen Jun 27 04:02:50 i see a fb.ko Jun 27 04:03:04 also an omapfb.ko Jun 27 04:03:22 try doing 'modprobe omapfb' Jun 27 04:03:31 1 sec Jun 27 04:06:47 please see http://pastebin.com/304597 Jun 27 04:06:51 the X mark i get if i try to move it it goes and site at the top left corner ; unable to move it down ? Jun 27 04:09:45 is that the calibration screen? Jun 27 04:10:28 i got a grey screen with an X in the middle ? Jun 27 04:11:05 when i use a pen on the x server it moves with the pen and sits at the top left corner unable to move it down ? Jun 27 04:11:34 try running xtscal Jun 27 04:17:38 koen : getting the error Jun 27 04:17:40 root@omap5912osk:~# xtscal Jun 27 04:17:42 Couldn't open display Jun 27 04:17:44 root@omap5912osk:~# Jun 27 04:19:23 ram123: export DISPLAY=:0 Jun 27 04:19:28 and make sure X is running Jun 27 04:22:32 i ran Xfbdev in background and the X moves but not across the entire screen ?. Jun 27 04:22:48 probably a calibrartion problem ? Jun 27 04:23:03 fine how to make the Gpe to work now ? Jun 27 04:23:03 I think so Jun 27 04:23:27 shall i put modprobe in a Conf file ?. Jun 27 04:23:32 ram123: try doing (after killing X) /etc/init.d/gpe-dm start Jun 27 04:23:45 03XorA 07 * r1.3637 10openembedded/packages/mplayer/ (7 files in 2 dirs): mplayer-atty: version bump to 1.1.5 Jun 27 04:23:45 ok Jun 27 04:29:58 koen: i got enter user password and so on and a keyboard at the bottom Jun 27 04:30:00 is that all ? Jun 27 04:30:04 but due to calibration problem i cannot enter the names. i am trying .... Jun 27 04:30:08 to calibrate properly. Jun 27 04:30:28 can you hold the stylus on the screen? Jun 27 04:30:45 yes Jun 27 04:31:11 i will try to calibrate again. is there any standard way ? Jun 27 04:31:50 running xtscal Jun 27 04:32:02 what does it do ? Jun 27 04:32:20 running xtscal and moving the stylus ? Jun 27 04:32:23 thats all ? Jun 27 04:35:31 xtscal will write the calibaration info into /etc/pointercal Jun 27 04:41:38 actually i am unable to enter letters in username Jun 27 04:41:42 even if i hold on it. i am facing problem it always points to CAPLOCKS key ? Jun 27 04:47:06 when i run Xfbdev after stopping gpe-dm i get segmentation violation: Jun 27 04:47:08 root@omap5912osk:~# Xfbdev Jun 27 04:47:10 _XSERVTransSocketOpenCOTSServer: Unable to open socket for inet6 Jun 27 04:47:12 _XSERVTransOpen: transport open failed for inet6/omap5912osk:0 Jun 27 04:47:14 _XSERVTransMakeAllCOTSServerListeners: failed to open listener for inet6 Jun 27 04:47:16 Segmentation fault Jun 27 04:47:20 when i run Xfbdev after stooping gpe-dm i get segmentation fault Jun 27 04:47:22 root@omap5912osk:~# Xfbdev Jun 27 04:47:24 _XSERVTransSocketOpenCOTSServer: Unable to open socket for inet6 Jun 27 04:47:26 _XSERVTransOpen: transport open failed for inet6/omap5912osk:0 Jun 27 04:47:28 _XSERVTransMakeAllCOTSServerListeners: failed to open listener for inet6 Jun 27 04:47:30 Segmentation fault Jun 27 04:51:28 koen: are u there ? Jun 27 04:54:24 03pb 07 * r1.3635.1.1 10openembedded/packages/linux-libc-headers/linux-libc-headers_2.6.11.1.bb: package and ship linux-libc-headers from here, since glibc doesn't actually include them in libc6-dev Jun 27 05:04:13 unable to run Xfbdev; it is giving segmentation violation ?. Jun 27 05:04:17 i tried it even after a reboot ? Jun 27 05:04:21 first time it worked ?. but now it gives segmentation violation. Jun 27 05:07:19 03koen 07 * r1.3637.1.1 10openembedded/packages/gpe-contacts/ (gpe-contacts_0.40.bb gpe-contacts-maemo_0.40.bb): Add gpe-contacts 0.40 Jun 27 05:07:37 hi all Jun 27 05:08:57 hi florian_kc Jun 27 05:20:56 hi zecke Jun 27 05:21:28 hey Jun 27 05:21:42 cant run my xserver. it is giving segmentation fault. what could be wrong ? Jun 27 05:22:36 sounds like a bug Jun 27 05:25:32 ran it for the first time. it worked. rebooted the system and ran it again. it gives segmentation fault. please see http://pastebin.com/304613 Jun 27 05:26:19 does it need /etc/pointercal ?. it is empty ? Jun 27 05:32:29 my gpe-image build seems to be stalled on : NOTE: Update cvs://anoncvs:anoncvs@cvs.handhelds.org/cvs;module=linux/kernel26;tag=K2-6-12-hh0 Jun 27 05:33:07 oh, of course as soon as I said that it magically started downloading Jun 27 05:33:19 after a several minute delay Jun 27 05:33:21 heh Jun 27 05:33:28 yeah, cvs does that sometimes Jun 27 05:33:53 yeah true, waiting for locks or something Jun 27 05:34:37 pb_, do you know why a supposedly 266MHz ixp420 would dhrystone like a 133MHz ARM ? Jun 27 05:38:58 dunno Jun 27 05:39:04 what does lmbench think the frequency is? Jun 27 05:42:07 hmm I don't recall lmbench being able to tell Jun 27 05:42:39 but it only gets 133.12 bogomips which seems about 1/2 what it should be getting Jun 27 05:44:18 re Jun 27 05:44:41 alcomats suxx Jun 27 05:47:51 jacques: if bogo in bogomips: results_meaning_less() Jun 27 05:48:08 jacques: but yes 'mips' normally just corelate with the MHZ number of the CPU Jun 27 05:48:29 zecke: depends on architecture Jun 27 05:49:00 zecke: on some architectures bogomips ~= mhz on other bogomips = 2xmhz Jun 27 05:49:09 ssh: connect to host oe-devel.bkbits.net port 22: Connection timed out Jun 27 05:49:13 ~lart bk Jun 27 05:49:14 * ibot shoves a crumpet down bk's throat, happy now?! Huh? Want some JAM with that? Jun 27 05:49:45 hrw|work: 2x is just a constant Jun 27 05:49:58 jacques: bogomips don't always correlate very well with "real" mips. it's certainly not impossible to get a factor of 2 difference there, depending on the kernel. Jun 27 05:50:11 jacques: have another go with lmbench, see what you can get out of it. It's usually pretty good at figuring out the clock rate. Jun 27 05:51:57 someone with a 266MHz ixp425 devel board with same kernel gets 266 bogomips Jun 27 05:52:29 okay then something is 'wrong' Jun 27 05:53:12 ah, that does seem suspicious Jun 27 05:53:26 hm... I will reboot husky Jun 27 05:53:52 we can't find any mention of a way to run a ixp420 at 133MHz, so we're kinda stumped Jun 27 05:55:22 we even have a wiki page about it: http://www.nslu2-linux.org/wiki/OpenSlug/TheSlugSpeedProblem Jun 27 06:06:19 jacques: funky Jun 27 06:08:25 pb_, that does seem to be the general concensus Jun 27 06:10:52 jacques: maybe you could use the PMU to gather more information about what's happening. Jun 27 06:11:13 it can count stalls due to core starvation, for example. Jun 27 06:11:59 hi jamey Jun 27 06:12:06 hi pb_ Jun 27 06:51:10 Good afternoon. Jun 27 06:53:36 gpe works on my osk board Jun 27 06:54:05 the filesystem size is 22MB when i mount it on host using -o loop option. Jun 27 06:54:26 i want to reduce the size of this. has anyone worked on that before ?. Jun 27 06:54:38 not sure which libs and executables to remove ? Jun 27 06:54:53 is there any straight forward way ? Jun 27 06:58:24 ram123, congrats! Jun 27 06:58:51 look at the gpe-image.bb file Jun 27 06:59:03 ram123, most should already have been stripped by BB i think ... Jun 27 06:59:24 check the meta and the classes, if they havent you can do it manually Jun 27 06:59:29 i am really happy today. Jun 27 06:59:32 emte, I am pretty sure he means not include programs in the image Jun 27 06:59:46 ram123: probably you need to reduce task-bootstrap or meta-gpe Jun 27 07:00:05 meta-gpe is a small gpe image? Jun 27 07:00:12 thats good, thats eaier Jun 27 07:00:19 easier* Jun 27 07:00:31 hey!! raster Jun 27 07:00:38 been looking for you all weekend .. Jun 27 07:00:45 it was really easy compared to downloading each lib by hand, compiling it, downloading its dependents and so on. Jun 27 07:01:01 what is meta-gpe ? Jun 27 07:01:04 explain to me again how X, X, X, X relates to color .... CMYK? Jun 27 07:01:16 or RGBA Jun 27 07:01:23 whats A? Jun 27 07:01:29 Alpha channel Jun 27 07:01:33 antialiasing? Jun 27 07:01:35 hmm Jun 27 07:01:45 emte: emmie! Jun 27 07:01:55 hey Jun 27 07:02:11 wandering through the smart object stuff Jun 27 07:02:31 what size is meta-gpe ? Jun 27 07:02:51 when mounted onto nfs ? Jun 27 07:02:55 ram123, for the h3600 series GPE builds ~ 14-16M Jun 27 07:03:04 full image Jun 27 07:03:07 mine is osk ? Jun 27 07:03:14 omap 5912 TI Jun 27 07:03:41 depends on device specs if its okay or not Jun 27 07:04:15 i dont knoe the size comprssion for NFS vs CramFS or UnionFS Jun 27 07:04:18 know* Jun 27 07:04:25 maybe hrw|work does Jun 27 07:04:50 NFS and UFS do compression?? Jun 27 07:04:59 NFS doesnt Jun 27 07:05:03 i want a jffs2 image Jun 27 07:05:21 how can i convert the .ext2 file into jffs2 format ? Jun 27 07:05:44 and how much will be the size of the jffs2 filesystem ? any ideas ? Jun 27 07:05:53 ram123: see local.conf and add 'jffs2' to IMAMGE_FSTYPE Jun 27 07:05:58 raster, reading the evas.h i create the touch simular to the way the mouse has been done? Jun 27 07:06:10 given it. but got an .ext2.gz file Jun 27 07:06:16 IMAGE_FSTYPES iirc Jun 27 07:06:25 "Houstion we have (touch) splash down" Jun 27 07:06:32 houston* Jun 27 07:06:37 :P Jun 27 07:07:26 bit confused where ecore and evas boundaries are ... pretty tightly integrated Jun 27 07:09:32 ram, no one has built files for gpe on OSK yet Jun 27 07:10:04 Crofton|laptop: pssst, he just did ;) Jun 27 07:10:39 koen, let me restate, no one has made a gpe file specifically for the OSK so the image fits in the flash :) Jun 27 07:11:08 how much flash does it have? Jun 27 07:11:10 * Crofton|laptop just built gpe for the also, but haven't had a chnace to fool with Jun 27 07:11:11 16mb? Jun 27 07:11:13 32 M Jun 27 07:11:28 more than enough Jun 27 07:11:39 emte: touch? eh? Jun 27 07:12:05 raster, yeah, pretty much a key/mouse combo without buttons i expect Jun 27 07:12:31 looks like it shouldnt be hard Jun 27 07:12:45 tho this is the time i wish i had a touchscreen monitor ... Jun 27 07:13:06 Here is a df from my OSK http://pastebin.ca/15978 Jun 27 07:13:23 anyway i have to catch my bus to go to my high reliability soldering course Jun 27 07:13:28 bbl Jun 27 07:13:41 i'll get you portal access tonight raster Jun 27 07:14:27 hjmm Jun 27 07:14:27 ok Jun 27 07:14:35 not sure what u are talking about with "touch" Jun 27 07:14:36 heheh Jun 27 07:16:06 errr - touchscreen? Jun 27 07:16:46 * Twiun strokes his erm, device. Jun 27 07:16:49 * Twiun hides Jun 27 07:19:40 Twiun: heh Jun 27 07:37:14 ~lart bitmover Jun 27 07:37:14 * ibot frags bitmover with his BFG9000 Jun 27 07:38:58 ssh: connect to host oe-devel.bkbits.net port 22: Connection timed out Jun 27 07:39:01 argh Jun 27 07:39:08 ah... so i'm not the only one :-) Jun 27 07:41:11 it is 1st of july ;) Jun 27 07:41:27 heh Jun 27 07:41:43 Oh, my god, I slept for 4 days? Jun 27 07:41:59 So did I... weird Jun 27 07:42:17 zecke: I'm not so drunk to miss 4 days Jun 27 07:42:57 maybe June only has 26 days in Larry's crazy universe. Jun 27 07:43:17 pb_: change from gregorian to larrian time? Jun 27 07:45:01 hm... interesting. Collie APM hangs on suspend looping with "set_power_state: what=1 state=4". Where that combination means ALL DEVICES, BUSY Jun 27 07:46:39 * XorA wonders how to keep a monotone key safe Jun 27 07:47:52 * koen suggest burning it to cd Jun 27 07:48:04 koen: I dont find cd-r's very reliable Jun 27 07:50:25 XorA: just put a million copies of it on the disc Jun 27 07:50:39 koen: yeah I suppose Jun 27 07:50:49 koen: tempted to post it to usenet :-) Jun 27 07:50:57 XorA: craft it by hand, 1111111111 Jun 27 07:51:48 as a side note: we'll use @openembedded.org IDs for the people who have bk access at the moment Jun 27 07:52:18 * CosmicPenguin hugs koen Jun 27 07:55:29 mickey|meeting will have the honour of coordinating the first part of the move Jun 27 07:57:15 koen: so OE is moving from BK to what? (sorry, missed the other half of the story :P) Jun 27 07:57:39 marcan_: monotone Jun 27 07:58:01 ok Jun 27 08:09:29 what TS controler does the Collie have? Jun 27 08:09:40 I know it is compatible with the Philips UCB1200 Jun 27 08:10:42 * CosmicPenguin pukes Jun 27 08:12:30 morning kergoth Jun 27 08:12:37 hey Jun 27 08:12:54 'morning kergoth Jun 27 08:15:26 the collie APM code is truly horrid Jun 27 08:15:37 makes me want to write my own instead of fixing it. It might be easier Jun 27 08:15:57 marcan_: collie 2.4 is scary Jun 27 08:16:06 marcan_: go to 2.6er kernel directly Jun 27 08:16:11 I wont touch it with long stick even if payed for it Jun 27 08:16:23 zecke: 2.6/collie is useless without sd driver Jun 27 08:16:26 yeah, 2.6 is the way to go Jun 27 08:16:32 but no sd, no nothing Jun 27 08:18:11 marcan_: so fix that Jun 27 08:18:19 marcan_: cant be harder than fixing apm Jun 27 08:18:35 SD is pretty simple Jun 27 08:18:52 Then why hasn't anyone done it yet? :) Jun 27 08:18:56 I know nothing about SD Jun 27 08:19:00 CosmicPenguin: I know, I wrote the SD layer we use in 2.6 :) Jun 27 08:19:11 Oh, so that was you? Jun 27 08:19:33 well, I wrote it and a couple of others cleaned it up but yeah, basically. Jun 27 08:19:34 kergoth: could you do some magic with the topic (bugzilla, monotone, etc)? Jun 27 08:20:33 Spyro: I would happily use it, if my damn controller could handle a CMD51 Jun 27 08:21:31 CosmicPenguin: isnt that a basic necessity for SD ? Jun 27 08:21:53 Spyro: heh - yeah, you would think Jun 27 08:22:26 We've got this strange memory mapped controller that handles most of hte bitbanging foo for us Jun 27 08:22:28 :) Jun 27 08:22:49 is it memory card only ? Jun 27 08:22:58 both Jun 27 08:23:01 actually, SDIO too Jun 27 08:23:22 surely for SDIO it has to be able to do just raw commands then? Jun 27 08:23:46 Spyro: on collie SD/MMC is handled by locomo (sharp only chip without any docs) Jun 27 08:23:48 ~locomo Jun 27 08:23:48 i heard locomo is a completely undocumented chip in the Sharp Zaurus Jun 27 08:24:11 * CosmicPenguin calls the hardware guy again to complain Jun 27 08:24:20 Spyro: and only kergoth has kind of 'docs' for it - scan of japanese datasheet (under NDA) Jun 27 08:25:32 what's the point in not documenting a chip that linux has to support? Jun 27 08:25:39 makes me wonder if they just like to make our lives difficult Jun 27 08:26:01 marcan_: collie sd driver is binary only - totally closed source Jun 27 08:26:06 ~lart sharp as usual Jun 27 08:26:06 * ibot does a little 'renice 20 -u sharp as usual' Jun 27 08:26:25 yeah, but all the other stuff - keyboard, display, etc Jun 27 08:26:27 that is LoCoMo too Jun 27 08:26:36 they could at least document *that* Jun 27 08:26:40 and, while I doubt they are setting out to make your life deliberately difficult, they almost certainly don't care in the slightest about making your life easy. Jun 27 08:26:43 and the SD too, now that the spec is open supposedly Jun 27 08:27:20 marcan_: now it is but it was not when collie was released Jun 27 08:27:29 so, make it available now Jun 27 08:27:37 marcan_: linux-2.4.xx-sharp is so awful.... Jun 27 08:27:45 yeah Jun 27 08:27:56 anyway, no use ranting about this here Jun 27 08:28:20 It's sad... nowadays hardware is full of black holes with no docs Jun 27 08:28:28 all pxa based zaurus machines will sooner or later get usable 2.6 kernel but not collie Jun 27 08:28:37 what prevents us from 'dlopening' the driver and treat it the way we treat Windows COFF binaries (NDIS wrapper) Jun 27 08:28:50 or a-300 (but that model is totally unsupported by us) Jun 27 08:31:36 zecke: in principle, nothing. Jun 27 08:32:27 zecke: want to do it? Jun 27 08:32:30 zecke: but, in practice, the fact that the kernel has no internal ABI stability makes it almost impossible to do that kind of thing. Jun 27 08:33:04 hrw|work: maybe next year ;) Jun 27 08:39:51 <_ggilbert> marcan_: it's important to keep in mind that they do have documentation for the chips, just not publically available and not in english :) Jun 27 08:41:05 well, since they made the chips, they better have the docs :P Jun 27 08:41:10 anyway, they could make it availab Jun 27 08:41:12 le Jun 27 08:42:02 marcan_: in industry its never as simple as that :-( Jun 27 08:42:11 yeah, that is the problem Jun 27 08:42:28 heh. 'grep "RDEPENDS =" */*.bb | wc -l' gives a faintly depressing result. Jun 27 08:42:41 <_ggilbert> marcan_: plus Sharp USA simply didnt even have them :p Jun 27 08:42:44 marcan_: Ive seen many cases where someone designed chips/software and the docs never got written, when they quit the design is lost Jun 27 08:42:48 03XorA 07 * r1.3637.2.1 10openembedded/packages/mozilla/ (4 files in 2 dirs): firefox: version bump to 1.0.4 Jun 27 08:43:43 heh. Funny, I always have to write at least a basic doc of what I do (especially hardware), else I get messed up writing the software :) Jun 27 08:43:44 and I know from my own experience of Japanese corps they will not give anything to subsidaries Jun 27 08:45:38 marcan_: it is sharp - they dont document Jun 27 08:47:50 ok. I go home Jun 27 08:47:52 cu Jun 27 08:52:08 03pb 07 * r1.3638.1.2 10openembedded/packages/glibc/glibc-cvs-2.3.5/ (dyn-ldconfig-20041128.patch ldsocache-varrun.patch): forgotten files Jun 27 08:52:17 03pb 07 * r1.3638.1.1 10openembedded/packages/glibc/ (4 files in 2 dirs): glibc updates Jun 27 09:02:50 hi Jun 27 09:07:07 Hello. Jun 27 09:07:34 Do you know some tool like valgrind that could be compiled for the c7x0 ? Jun 27 09:08:29 SirFred: imho valgrind is x86 only Jun 27 09:08:47 florian_kc: Yes, I think so. That's why I said 'like valgrind'. ;-) Jun 27 09:09:56 SirFred: maybe memprof... but why you don't run that software on a pc? Jun 27 09:10:25 florian_kc: Because the target runs on the c7x0. I'm looking for a failure. Jun 27 09:10:49 SirFred: you don't have its source? Jun 27 09:10:58 florian_kc: Yes, I have it. Jun 27 09:11:54 SirFred: so compile it for x86, all these tools are performance killers. Jun 27 09:12:12 florian_kc: Yes, I know. But the software is too low level to test on x86. Jun 27 09:12:36 SirFred: uh nasty Jun 27 09:12:45 And anyway, it shares memory with other processes, so ,perhaps it would not be a good idea. Jun 27 09:12:53 florian_kc: Thanks. Jun 27 09:20:47 ok, I *think* I know what the heck is going on with suspend on collies Jun 27 09:21:32 the suspend signal sends an APM SUSPEND event, which sends SIGSTOP to all processes. Then, it queues the event and guess what, waits for all APM users to respond. Of course, apmd is now stopped, it will never respond. Jun 27 09:21:36 marcan_: we all know: ugly embeddix code full of nasty bugs ;) Jun 27 09:22:20 koen: yeah, that is a pretty good generalization Jun 27 09:22:23 I just wanted the details :P Jun 27 09:27:57 03pb 07 * r1.3642 10openembedded/classes/linux_modules.bbclass: respect PARALLEL_INSTALL_MODULES Jun 27 09:30:14 What a surprice. Since when something like printf( "%s", 0 ) prints (null) instead incurring in a good SIGSEGV? Jun 27 09:30:22 s/surprice/surprise Jun 27 09:30:59 SirFred: I've seen that on cash registers too, where the display prints "(null)" in some modes Jun 27 09:31:01 03pb 07 * r1.3643 10openembedded/conf/machine/ (7 files): pass extra argument to linux_module_packages() Jun 27 09:31:06 i'm guessing it's the same thing :) Jun 27 09:31:14 gb2: It seems to be libc. Jun 27 09:31:18 yeah Jun 27 09:31:19 it is Jun 27 09:31:23 it's convenient Jun 27 09:32:10 Well, I wanted a core. Jun 27 09:32:15 :) Jun 27 09:33:36 abort() Jun 27 09:33:48 *(int *)0 = 0; Jun 27 09:33:52 gb2: Yes, I made it. Jun 27 09:34:19 03pb 07 * r1.3644 10openembedded/conf/machine/thinclient-common.conf: forgot one Jun 27 09:37:51 * mr_claus is away: "sport" Jun 27 09:37:52 03pb 07 * r1.3645 10openembedded/packages/ (libosip/libosip_0.9.7.bb linphone/linphone-hh_0.12.2.hh1.bb): correct SRC_URI Jun 27 09:41:41 03pb 07 * r1.3646 10openembedded/conf/machine/h3900.conf: update kernel26 module list for h3900 Jun 27 09:43:51 re Jun 27 09:47:04 03pb 07 * r1.3648 10openembedded/packages/kbdd/kbdd_cvs.bb: fix DEPENDS Jun 27 09:47:09 03pb 07 * r1.3647 10openembedded/packages/meta/task-mythfront.bb: mythfront updates Jun 27 09:53:45 03pb 07 * r1.3649 10openembedded/packages/meta/task-bootstrap.bb: bump PR Jun 27 09:59:06 03pb 07 * r1.3650 10openembedded/classes/linux_modules.bbclass: avoid "no such file" when building native packages Jun 27 10:13:55 <[g2]> kergoth, ping Jun 27 10:32:18 * chouimat is away: food Jun 27 10:33:59 hey zecke Jun 27 10:34:40 <[g2]> koen, I've got a question for you, kergoth, mickey|meeting and other devs Jun 27 10:35:05 <[g2]> I've just about got the native-development environ all setup for openslug Jun 27 10:35:22 <[g2]> an important issue is the kernel include headers Jun 27 10:36:04 <[g2]> I was thinking that an ipkg should be created from the kernel builds that has this available Jun 27 10:36:34 [g2]: don't the linux-libc-header do about the same? Jun 27 10:36:35 <[g2]> I think it's the asm-arm, asm-generic and linux directories Jun 27 10:36:50 [g2]: -rw-rw-r-- 1 pb pb 1129408 Jun 27 12:36 tmp/deploy/ipk/linux-libc-headers-dev_2.6.11.1-r1_arm.ipk Jun 27 10:36:53 there it is! Jun 27 10:36:56 heh Jun 27 10:37:41 <[g2]> pb_ are those the sanitized headers ? Jun 27 10:37:53 yeah Jun 27 10:37:54 gb2: yes Jun 27 10:39:01 <[g2]> so how different will that be from the actual tree ? Jun 27 10:39:37 <[g2]> meaning the kernel tree Jun 27 10:40:00 * [g2] should probably just diff Jun 27 10:40:04 not very Jun 27 10:40:28 but yeah, you probably should Jun 27 10:40:43 <[g2]> is that what you guys are using in building your native development environs ? Jun 27 10:41:20 yeah, though I haven't actually tested it yet Jun 27 10:41:45 <[g2]> well I've built perl and python and the kernel Jun 27 10:41:50 cool Jun 27 10:41:58 <[g2]> all the perl tests pass and all but 3 of the python tests pass Jun 27 10:42:19 <[g2]> also apache and php5 built and run Jun 27 10:42:49 <[g2]> I'll have to take a look at what's in there Jun 27 10:42:58 I think I fixed the issue with collie suspend hangs Jun 27 10:43:04 <[g2]> THX guys Jun 27 10:43:23 but I still can't believe how the heck whoever programmed this expected apmd to respond after shutting down all processes and sending the SUSPEND signal to everyone Jun 27 10:43:36 * pb_ go home now Jun 27 10:43:37 later all Jun 27 10:43:43 marcan_: sharp magic Jun 27 10:43:56 yeah Jun 27 10:45:30 it was like send_event() and later queue_event() Jun 27 10:45:43 So, it shuts down everything and then queues it up for apmd to respond Jun 27 10:45:44 how fun Jun 27 12:10:46 03pb 07 * r1.3653 10openembedded/ (packages/meta/task-bootstrap.bb conf/machine/h3900.conf): more h3900 modules Jun 27 12:14:06 argh Jun 27 12:14:31 now I fixed APM (I hope), but usbnet dies on the third resume and takes the whole kernel with it Jun 27 12:20:02 ~usbdmonitor Jun 27 12:20:15 ibot ? Jun 27 12:20:39 usb has been flakey on that retarded embedix kernel since day 1 Jun 27 12:20:49 if you can patch _that_, you're our hero :) Jun 27 12:21:03 (you are anyway for removing that right-button thing) Jun 27 12:27:24 mickey|meeting: Hi, Mickey. Jun 27 12:28:02 evening Jun 27 12:28:21 mickey|meeting: I've just found a new bug in QLinuxFbScreen. Jun 27 12:28:34 congrats :) Jun 27 12:29:13 mickey|meeting: Thanks.It's related with the offscreen memory management. Under certain conditions, a substraction overflows and everything goes to hell. Jun 27 12:29:55 * chouimat is back. Jun 27 12:30:33 mickey|meeting: after seeing what APM is like, I'm afraid of what usbnet will look like Jun 27 12:31:02 anyway suspend/resume works now, but only as long as you don't touch usbnet or so it seems Jun 27 12:31:03 SirFred: nasty. that makes 3 patches pending. the sync() thing, the overflow thing and the freak-out-on-suspend one :) Jun 27 12:31:07 marcan_: excellent Jun 27 12:31:46 mickey|meeting: I send a bug report about the sync() thing, everything I got is an automatic response, by now. Jun 27 12:32:01 SirFred: you have a commercial developer license? Jun 27 12:32:09 mickey|meeting: What's that? Jun 27 12:32:11 mickey|meeting: :) Jun 27 12:32:13 heh Jun 27 12:32:18 then don't expect anything from qt-bugs Jun 27 12:32:23 we went through that before Jun 27 12:32:30 a 'don't ignore me' license? Jun 27 12:32:31 mickey|meeting: Well, so, lets break the binary compatibility. Jun 27 12:32:37 SirFred: *nod* Jun 27 12:33:43 OK. That last overflow bug was the main reason for rotation not working fine. Jun 27 12:34:18 BTW, does anyone know what that flakey screen on tosa is about? Jun 27 12:34:21 the yellow/brown lines Jun 27 12:34:26 marcan_: no one knows :/ Jun 27 12:34:30 Anyway, I've found some problems rotating while the pixmap cache is at it's limits. For example, with konqueror-embedded running. Jun 27 12:34:32 looks really strange Jun 27 12:34:35 like a hardware glitch Jun 27 12:34:56 marcan_: yeah. i dunno how far dirk is with fb on 2.6 Jun 27 12:35:02 we should ask him if he's sees the same things Jun 27 12:35:12 yeah Jun 27 12:35:28 fixing it in 2.4 is pointless Jun 27 12:35:31 mickey|meeting: One question about the patches. I suppose they got applied in the same order than specified in SRC_URI, don't they? Jun 27 12:35:33 *nod* Jun 27 12:35:36 SirFred: yes Jun 27 12:36:00 mickey|meeting: who do I send the APM patch to? :) Jun 27 12:36:10 mickey|meeting: OK. I'll try to prepare that two patches. Jun 27 12:36:12 marcan_: mickeyl@handhelds.org Jun 27 12:36:40 SirFred: very good Jun 27 12:36:50 mickey|meeting: I've correctly configured my emacs tabstops. You'll be proud. ;-) Jun 27 12:36:56 hehe Jun 27 12:38:38 oh, mickey|meeting, can you test the APM stuff with Opie? My collie is... not too well suited for it :P (broken touchscreen) Jun 27 12:38:58 marcan_: sure. i'll launch a complete image build with the latest stuff right after applying the patch Jun 27 12:39:06 cool :) Jun 27 12:39:31 I've tested apm -s and suspending with cancel. Both seem to go OK Jun 27 12:39:40 (except for the usbnet thing) Jun 27 12:40:05 mickey|meeting: you are still on meeting? Jun 27 12:42:17 hmm. not sure Jun 27 12:42:20 headache Jun 27 12:42:52 marcan_: could you describe in non-kernel-programmer words what you did to fix it? the patch is interestingly small Jun 27 12:43:14 the kernel wasn't hanging up really, but it was deadlocking with apmd Jun 27 12:43:30 it shuts down everything and then attempts to wait for apmd to acknowledge the suspend Jun 27 12:43:34 which is of course pointless Jun 27 12:43:43 uh oh Jun 27 12:43:47 so I switched everything around and now it shuts everything down after querying apmd Jun 27 12:43:47 ~lart lineo Jun 27 12:43:47 * ibot nabs the moon and broadsides lineo with the sea of tranquility Jun 27 12:44:51 It all became apparent after I riddled everything with debugs and connected the serial console, hacked the code to ignore the non-incoming-acknowledge after a while, and saw how apmd actually re-suspended the system after wakeup (that was the late ack :) ) Jun 27 12:46:18 The sooner we get away from 2.4, the better... Jun 27 12:46:24 no kidding Jun 27 12:47:06 so the reason for that problem to appear in later reason is different apm userland timing Jun 27 12:47:26 s/reason/releases/ Jun 27 12:47:35 03mickeyl 07 * r1.3637.3.1 10openembedded/packages/linux/ (2 files in 2 dirs): Apply patch from Hector Martin that fixes collie suspend/resume deadlock w/ apmd. (Yippee :) Jun 27 12:47:48 * mickeyl build opie-image Jun 27 12:47:59 so I have to rebuild all collie kernels for upgrades again.. Jun 27 12:48:06 let me test a bit first Jun 27 12:48:25 yep Jun 27 12:48:57 apmd was probably sending the signal fast enough, or maybe a different kernel config caused different suspend behavior Jun 27 12:49:20 hmm Jun 27 12:49:23 thinking about that Jun 27 12:49:36 timing varies on the amount of resume scripts in apm.d Jun 27 12:49:43 this is what we changed in releases Jun 27 12:49:43 true Jun 27 12:49:53 could have been that Jun 27 12:50:07 * mickeyl shakes head Jun 27 12:50:14 or could have been the alignment between the moon and the sun and a vudu 12412 km from here Jun 27 12:50:15 ~lart lineo just because Jun 27 12:50:15 * ibot offers lineo just because some herring Jun 27 12:50:17 who knows Jun 27 12:50:42 the code is REALLY wicked Jun 27 12:50:53 lots of magic constants and strange and/or pointless behavior Jun 27 12:51:15 like, one of those dudes set up a wait queue for retreiving the status of the suspend, after the suspend Jun 27 12:51:21 marcan_: There normally is a reason. It might not be a good reason but there normally is one :-/ Jun 27 12:51:23 which is pointless because the suspend routine never fails Jun 27 12:52:10 marcan_: It should be able to and that's one of the problems with 2.6. It was never completed Jun 27 12:52:52 yeah Jun 27 12:53:50 ibot: pester do13 and bigal to do more work on 2.6 for tosa/spitz Jun 27 12:53:50 do13 and bigal to do more work on 2.6 for tosa/spitz: Are we there yet? .. Are we there yet? .. Are we there yet? see http://beverlys.net/LJ/BuggingYou.swf Jun 27 12:54:11 heh Jun 27 12:54:18 that shockwave file is pretty nice Jun 27 12:54:27 I'll look at that next thing, when I have some time Jun 27 12:55:10 my collie has just about served its purpose, and once apm works I'm not touching it much more (it's being used for a hardware project since it has no touchscreen) Jun 27 12:55:15 but my tosa needs some work Jun 27 12:55:27 coolness :) Jun 27 12:55:40 i wish i got one of those CF jacket expansion while they were on sale :/ Jun 27 12:56:25 I wish I got my tosa when they were cheaper :/ Jun 27 12:57:52 RP: 2.6.x is not yet stable, 3 months work for somebody... Jun 27 12:58:34 * mr_claus is back (gone 03:20:51) Jun 27 12:59:35 france_: 2.6.x is perfectly stable on the c7x0 Jun 27 13:00:03 RP: stress test it and send me the resultes???? Jun 27 13:00:24 france_: stress test in what way? Jun 27 13:00:27 it survided my linux tag stress test Jun 27 13:00:31 survived even Jun 27 13:00:41 mickeyl: very nice Jun 27 13:00:46 i don't use it as a web server though ;) Jun 27 13:00:57 RP: there are several standard sets (freshmeat.net) Jun 27 13:01:07 RP: my favorite is stress-kernel Jun 27 13:01:39 RP: 2.6.x does not survive well on my x86 box, I would be shocked it if it did well on my ARM box Jun 27 13:02:21 RP: then again most folks probably do not use their pda hard enough to notice. Jun 27 13:02:54 03pb 07 * r1.3653.1.1 10openembedded/packages/linux/handhelds-pxa-2.6_2.6.12-hh1.bb: update handhelds-pxa-2.6 to 2.6.12-hh1 Jun 27 13:03:18 france_: Are these tests in oe? Jun 27 13:03:33 france_: 2.6.x does not survive on x86? That's strange. Jun 27 13:04:06 RP: nope Jun 27 13:04:08 hy all Jun 27 13:04:26 marcan_: not really since Linus changed the development model at the last OLS Jun 27 13:04:40 marcan_: 2.6.x is not considered stable. Jun 27 13:05:01 marcan_: he felt that since the distro's test and make it stable, that 2.6.x did not need to be a stable kernel. Jun 27 13:05:25 marcan_: RH, SuSe, Mandrake, etc... put many man months into a kernel before puttting in their distros Jun 27 13:05:38 marcan_: somebody needs to take the time and fix the hh.org kernel. Jun 27 13:05:51 must be mine is stable because I run the Gentoo flavour Jun 27 13:06:05 marcan_: or you do not stress test your system, Jun 27 13:06:24 define stress test Jun 27 13:06:54 marcan_: run all the systems, memory, hard drives, dma, videos, serial, parrell, pci, usb, cpu, fpu, etc.... Jun 27 13:07:01 marcan_: about 40 processes in all............. Jun 27 13:07:10 france_: I don't use the hh.org tree for a reason. Too little quality control IMO :-( Jun 27 13:07:31 RP: your welcome to conribute. Jun 27 13:07:52 RP: but I am talking about the regular 2.6.x kernel, not the hh kernel Jun 27 13:07:58 RP: they both have the same issues. Jun 27 13:08:18 france_: I've tried. I've enough problems with standard kernels without the extra ones in hh.org Jun 27 13:08:24 heh Jun 27 13:11:38 Ciao all Jun 27 13:12:15 hi pigi Jun 27 13:12:42 hi pb_ I have reopened #120 to add the fixes for sa arch . Jun 27 13:12:48 hey Pigi Jun 27 13:13:09 From now on we have alsa in oe ( and familiar then ) Jun 27 13:13:11 hi koen Jun 27 13:13:29 Pigi: Alsa hero!! Jun 27 13:13:33 ~alsa Jun 27 13:13:33 from memory, alsa is the Advanced Linux Sound Architecture, or at http://www.alsa-project.org/ Jun 27 13:13:46 ~praise alsa Jun 27 13:13:46 All hail alsa! Jun 27 13:13:51 Pigi: okay, very good Jun 27 13:13:51 heh Jun 27 13:21:14 I dunno. I regularly stress most of the systems on my 2.6 box, including gigabit ether, HDD, DVD drive, CPU, video, all at the same time. it seems pretty solid here... Jun 27 13:21:25 bash-2.05b$ uptime Jun 27 13:21:33 21:21:16 up 63 days, 8:11, 64 users, load average: 0.08, 0.40, 0.36 Jun 27 13:22:08 even my buggy pvr is doing >100 days uptime Jun 27 13:22:15 ~hail ivtv drivers Jun 27 13:22:15 * ibot bows down to ivtv drivers and chants, "I'M NOT WORTHY!!" Jun 27 13:22:39 Spyro: lots of users, but the load average needs to be much higher, then run the box out of dram/swap, etc.... Jun 27 13:22:45 I just have to rmmod and insmod them every week or so Jun 27 13:23:13 france: thats the *current* load avg. Im not *doing* anything Jun 27 13:23:14 koen: I thought you were on holiday? Jun 27 13:23:18 21:23:02 up 56 days, 23:25, 3 users, load average: 1.41, 1.53, 1.29 Jun 27 13:23:33 That's also 2.6. Not many users but overworked :) Jun 27 13:23:55 and it has been known to run out of swap... Jun 27 13:23:58 lightly overworked, try a load average of 40 plus Jun 27 13:23:58 its kinda hard to run this box outa RAM/swap now since its got 2G RAM and 0B swap, but it used to be forever on the edge of overload back when it was 256MB/500MBswap on NFS Jun 27 13:24:38 france_: It regularly sees a load of 10 or so... Jun 27 13:24:55 Spyro: my alpha has 16G on each processor and I can do it. Jun 27 13:25:06 france_: starting thursday Jun 27 13:25:12 Spyro: it takes some time, but it is doable. Jun 27 13:25:14 france: sure, on some artificial test or a massive compile Jun 27 13:25:21 koen: :-D enjoy.... Jun 27 13:25:28 Spyro: Jun 27 13:25:32 hi france Jun 27 13:25:44 pb_: hello! Jun 27 13:26:14 PARALLEL_MAKE="-j" on bitbakeng :) Jun 27 13:26:19 pb_: jamey setup that mailing list for you, but I think he forgot to tell anybody. Jun 27 13:26:48 koen: I can't wait :) Jun 27 13:27:25 03pb 07 * r1.3656 10openembedded/classes/update-rc.d.bbclass: avoid crash when PACKAGES is empty Jun 27 13:28:02 heyho Jun 27 13:28:25 who're the guys working on bitbake_qa / tinderbox / web-interface? :) Jun 27 13:29:02 and who can help me with my plans for the integration of a package configuration system? ;) Jun 27 13:30:10 bb_qa and tinderbox are zecke Jun 27 13:30:33 france: heh Jun 27 13:30:48 france: but unless my 2G is faulty (it tests OK) then when my box was loaded at 256M it should stress t around aas much as maxing out my 2G ould Jun 27 13:31:11 zecke: you there? ;) Jun 27 13:31:21 no Jun 27 13:31:58 Spyro: there are some types of h/w that get funky over 1G then 4G depending upon the memory mode of the CPU/MMU Jun 27 13:32:33 Spyro: lots of things break on my alpha when I put more than 4G of dram in it (per cpu) the first time. Jun 27 13:33:26 TheCount: what kind of config system? Jun 27 13:34:38 TheCount: I've been looking at something like packages.debian.org Jun 27 13:34:46 zecke: damn. I would have given you a user otherwise Jun 27 13:35:20 koen: a web-based release-management system - the package configuration part (which is currently missing) needs some integration which dies into the build/configure process Jun 27 13:35:37 koen: I'd like to do that this week, if y'all are willing to make this possible, it will happen. Jun 27 13:39:38 mickey_away: away :((( Jun 27 13:40:40 TheCount: do you have an example of such systems? I'm affraid I'm not really understanding what you're aiming at Jun 27 13:43:25 koen: think of it as a frontend for bitbake as buildsystem which allows you to see a summary of the builds of a project. a project consists of packages which are maintained by developers and have options (i.e. kernel version, autoconf options, etc) Jun 27 13:48:11 a kind of config-browser Jun 27 13:55:37 koen, before you leave for vacation, what will happen to oe after 1 of july ? Jun 27 13:56:09 Pigi: http://thread.gmane.org/gmane.comp.handhelds.openembedded/4853 Jun 27 13:56:10 We're all leaving to work on ant Jun 27 13:57:13 I have read that thread ( well almost ) but on the oe wiki there aren't instruction, or at least I haven't seen Jun 27 13:57:32 CosmicPenguin: not ant!!!!! Jun 27 13:58:17 koen: no? We already signed you up - we told them you had mad Java skilz... :D Jun 27 13:58:25 03pb 07 * r1.3657 10openembedded/packages/dbus/ (4 files in 2 dirs): first cut at dbus 0.34 Jun 27 13:59:11 pb_: could you lower the default pref for that dbus? Jun 27 13:59:19 to what? Jun 27 13:59:52 oops Jun 27 13:59:54 spoke to soon Jun 27 13:59:56 sorry Jun 27 14:01:04 koen: more or less ... rather less. Jun 27 14:01:09 *I* usually forget stuff like that Jun 27 14:01:34 koen: a kind of overview with the possibility to change thinks, manage customers, manage developers, watch the buildsystem, and decide about releases Jun 27 14:01:57 RP: where can I find kernel patches for hx2750? Jun 27 14:01:57 koen: would/should work well in a business environment, as that was what we designed it for in the last months Jun 27 14:03:54 jamey: they should be here http://projects.o-hand.com/hx2750/ Jun 27 14:04:19 jamey: but someone hasn't mirrored them there yet ;) Jun 27 14:04:51 that is where I looked for them Jun 27 14:05:38 jamey: they will be hidden on rpsys.net somewhere then :/ Jun 27 14:06:06 or in OE Jun 27 14:06:28 check the linux-openzaurus(!) .bb Jun 27 14:17:21 jamey: http://www.rpsys.net/openzaurus/patches/ipaq/ are the main ones Jun 27 14:17:48 jamey: Some of the ones in the patches directory might be needed to apply cleanly Jun 27 14:19:00 jamey: ignore the w1 ones Jun 27 14:19:08 * CosmicPenguin hugs Spyro Jun 27 14:20:18 03koen 07 * r1.3640 10openembedded/ (2 files in 2 dirs): Add 480x640 support to kmode.patch so the hx4700 works with kdrive Jun 27 14:26:22 CosmicPenguin: a little intimate arent we? Jun 27 14:37:14 You betcha Jun 27 14:38:20 jamey: I've also added them to o-hand.com... Jun 27 14:39:19 * koen nominates updating gnome in OE as his summer waste of time project Jun 27 14:40:31 03zecke123 * r268 10bitbake/bin/bitbake: Jun 27 14:40:32 Patch by Yann Dirson to show a more verbose error log Jun 27 14:40:32 when not beeing able to parse a file Jun 27 14:42:19 If I have Makefile what I should patch to include SDL.h from staging dir.. how to do that? Jun 27 14:42:44 ${STAGING_INC_DIR} should point to that Jun 27 14:43:12 so I should create patch dinamicaly Jun 27 14:43:20 or? Jun 27 14:44:24 I have makefile with CFLAG= -g -O2 -I/usr/include/SDL -D_REENTRAT -DPLAY_MOD -DUNIX Jun 27 14:44:52 gints|wrk: use 'sdl-config' to get the include directory Jun 27 14:45:08 gints|wrk: but I'm no expert on gnu make Jun 27 14:46:36 zecke: problem is that source contains ready Makefile and I need to change it... Jun 27 14:47:18 so I should overide do_configure with changing Makefile? Jun 27 14:51:17 gints|home: use sdl-config to get CFLAGS + LFLAGS Jun 27 14:51:25 gints|home: now the question is how to call the binary Jun 27 14:52:35 `${STAGING_BIN_DIR}/sdl-config` ? Jun 27 14:55:01 zecke: with sdl-config I can get this path... Jun 27 14:55:01 zecke: but what then.. Jun 27 14:55:26 gints|home: patch the Makefile Jun 27 14:56:01 zecke: main question is when I should do patching.. Jun 27 14:56:10 03jcrouse 07 * r1.3643 10openembedded/packages/acpid/ (acpid_1.0.3.bb acpid-1.0.3/acpid_1.0.3.bb): Jun 27 14:56:11 acpid_1.0.3.bb: Jun 27 14:56:11 Fix for gcc 4 Jun 27 14:56:12 do_patch Jun 27 14:57:35 'night all Jun 27 14:57:51 03jcrouse 07 * r1.3644 10openembedded/ (2 files in 2 dirs): Jun 27 14:57:51 gcc40.patch: Jun 27 14:57:51 Add the patch Jun 27 14:57:51 .del-acpid_1.0.3.bb~90bc4a32d33fdf1a: Jun 27 14:57:51 I don't know what I was thinking Jun 27 14:57:56 night koen Jun 27 14:58:08 * CosmicPenguin needs a swift kick in the head Jun 27 14:58:13 heh Jun 27 14:59:04 nite all Jun 27 17:07:54 hi Jun 27 17:07:55 hola Jun 27 17:17:46 i have a prblem, i can't build any package i always have the same error ERROR: /home/arrase/cross-gcc/ipaq/openembedded/packages/ipkg-utils/ipkg-utils-native_1.6cvs.bb depends on itself (eventually) Jun 27 17:18:20 i can't find a fix for this can you help me? Jun 27 17:41:38 i have a prblem, i can't build any package i always have the same error ERROR: /home/arrase/cross-gcc/ipaq/openembedded/packages/ipkg-utils/ipkg-utils-native_1.6cvs.bb depends on itself (eventually) Jun 27 17:41:46 i can't find a fix for this can you help me? Jun 27 18:32:11 i have a prblem, i can't build any package i always have the same error ERROR: /home/arrase/cross-gcc/ipaq/openembedded/packages/ipkg-utils/ipkg-utils-native_1.6cvs.bb depends on itself (eventually) Jun 27 18:32:13 i can't find a fix for this can you help me? Jun 27 19:46:48 hi,all **** BEGIN LOGGING AT Mon Jun 27 20:57:45 2005 Jun 27 23:12:29 hey there he is Jun 27 23:13:24 is the portal info ... i hate to say it but i have questions in the forum for you when you have time :( Jun 27 23:13:47 but i also have some concept images done Jun 27 23:13:54 maybe thats a consolation ... Jun 27 23:14:02 uploading em in a few Jun 27 23:14:51 bah Jun 27 23:14:56 wont be around for long Jun 27 23:15:02 so faer i have gotten NOTHING done today Jun 27 23:15:24 np Jun 27 23:15:42 its casual, maybe you will find time next week Jun 27 23:16:16 ok Jun 27 23:16:18 top-left Jun 27 23:16:23 is where all co-ords start Jun 27 23:16:32 pixesl are orders row by row, topp left to top right Jun 27 23:16:36 and line after line until the bottom Jun 27 23:16:47 as per the usual way any linear frameubffer is done Jun 27 23:16:59 alpha is the upper 8 bites of each 32bit value Jun 27 23:17:08 0xff == solid Jun 27 23:17:13 0x00 = transparent Jun 27 23:17:24 0x80 = partially transparent Jun 27 23:17:36 right now pixels are in non-pre-multilied space Jun 27 23:17:46 i may change to pre-multipied alpha at some time (maybe soon) Jun 27 23:18:09 k Jun 27 23:18:18 i think i understand most of that Jun 27 23:22:20 there Jun 27 23:22:24 posted the reply Jun 27 23:22:28 its there to refer to Jun 27 23:27:34 thnks Jun 27 23:27:46 hmm i just remembered why i hate pngs Jun 27 23:28:17 anyway the concept images are in News->e17biohazard Jun 27 23:30:00 need to put that white bg back in them ... Jun 27 23:34:54 just looking Jun 27 23:34:59 looks ok :) Jun 27 23:35:14 i always thought it was time to change the device ui Jun 27 23:35:22 to not be task oriented but object oriented Jun 27 23:35:27 i dont launch app x Jun 27 23:35:37 i just so "this message from mr X." Jun 27 23:35:41 and it opens so u can read it Jun 27 23:35:48 or i select a "new message" object Jun 27 23:39:10 but i do agree witht he menu and 1 app fulls creen thing Jun 27 23:41:27 yeah the three defailts will prob be filemanager, messaging or some type and media i'd expect Jun 27 23:41:42 or whatever you'd liek to set em as Jun 27 23:42:12 hmm why didnt 1 and 2 get whit bgs ... Jun 27 23:42:28 anyway ... like the logo? Jun 27 23:43:21 not bad :) Jun 27 23:45:36 aha fixed 1 and 2 Jun 27 23:46:09 hmm one is missing ... Jun 27 23:48:50 there the media one is back ... **** ENDING LOGGING AT Mon Jun 27 23:59:56 2005