**** BEGIN LOGGING AT Wed Sep 06 02:59:56 2006 Sep 06 06:15:10 good morning all Sep 06 06:20:33 koen: Hi! ;-) Sep 06 06:28:55 hey koen Sep 06 06:59:21 good morning Sep 06 07:27:14 good morning Sep 06 07:31:45 kergoth: did you make any progress on the conformantdeps stuff? Sep 06 07:37:55 morening Sep 06 07:40:49 hey XorA Sep 06 07:57:59 morning Sep 06 07:58:56 koen: do you have something on ewi which req php4 and will not work under php5? Sep 06 07:59:57 giel: I'm not familiar with the S3 stack Sep 06 07:59:59 hi zecke Sep 06 08:09:31 hrw|work: I have no idea Sep 06 08:10:21 koen: other words: which services works on ewi under apache (other then f-b and directory listings) Sep 06 08:13:10 hrw|work: nothing mission-critical Sep 06 08:14:16 koen: f-b guy asked about possibility to upgrade to php5 Sep 06 08:14:52 hrw|work: on ewi php5 wouldn't be a problem, but on ltg it could be Sep 06 08:14:55 morning all Sep 06 08:14:58 ah Sep 06 08:15:06 hey ade|desk Sep 06 08:18:22 koen: drupal is php4? Sep 06 08:19:44 drupal core and most modules work with php5 Sep 06 08:19:59 I haven't checked the modules ltg and oe.org use for php5 safety Sep 06 08:26:12 g'morning Sep 06 08:28:23 hey bigd0g Sep 06 08:38:37 zecke: i'm not too familiar with it atm, but we're going to use it, so it'll be fun to try and break it with stuff which is standard compliant but 'weird' ;) Sep 06 08:48:26 giel: the other two interesting bits will be inflating the FDT Instance Sep 06 08:48:31 giel: and not sending EXT_FTI headers Sep 06 08:58:50 zecke: are you releasing it as open source? Sep 06 09:01:07 i've got a question - i'm in the processing of flashing my dev board with my OE build Sep 06 09:01:23 do i just drop the rootfs that OE built on there? or did it also build a kernel and put it somewhere? Sep 06 09:01:46 giel: I'm not sure, currently it carries MIT license headers Sep 06 09:04:03 bigd0g: refer to devboard manual Sep 06 09:04:17 bigd0g: OE can generate kernel and rootfs but flashing method depend on devboard Sep 06 09:04:48 where does OE drop the kernel it builds? and what's the command? Sep 06 09:04:58 i thought it would've been built with 'bootstrap-image' Sep 06 09:05:20 i'm following the manual to flash, so no worries there :) Sep 06 09:05:25 giel: it depends if I may release it Sep 06 09:05:50 bigd0g: tmp/deploy/images/ Sep 06 09:06:04 zecke: are you doing this for BMCO or for someone else? Sep 06 09:06:21 uni (FU Berlin) Sep 06 09:06:22 hrw|work: it only has bootstrap images for the fs in there, so i guess i didn't build it. is there a standard command for the kernel? Sep 06 09:06:36 giel: I'm just sitting in the office space of the BMCO Sep 06 09:06:42 (to build it by itself, i mean) Sep 06 09:07:19 but I'm quite confident that I am allowed to release it Sep 06 09:07:34 zecke: that would be really cool Sep 06 09:08:50 bigd0g: bitbake virtual/kernel Sep 06 09:08:56 hrw|work: thanks Sep 06 09:10:31 hrw|work: it built successfully, but there's nothing added to the tmp/deploy/images folder Sep 06 09:10:37 any ideas? Sep 06 09:11:08 bigd0g: which kernel? Sep 06 09:11:45 hrw|work: it should be building the modifed 2.6.17 build by Crofton Sep 06 09:16:06 bigd0g: which is? Sep 06 09:16:32 hrw|work: linux-omap1 Sep 06 09:16:43 hrw|work: are you looking for the name of the bb file ? Sep 06 09:16:49 bigd0g: machine = omap5912osk? Sep 06 09:16:54 hrw|work: indeed Sep 06 09:17:22 bigd0g: ${DEPLOY_D Sep 06 09:17:23 IR_IMAGE}/${KERNEL_IMAGETYPE}-${MACHINE}-${DATETIME}.bin Sep 06 09:17:31 bigd0g: ${DEPLOY_DIR_IMAGE}/${KERNEL_IMAGETYPE}-${MACHINE}-${DATETIME}.bin Sep 06 09:18:06 bigd0g: which is tmp/deploy/images/vmlinux-omap5912osk-*.bin in your case probably Sep 06 09:18:24 what would be a cheap <200 Sep 06 09:18:42 mikearthur: ? Sep 06 09:18:51 what would be a cheap <£200 device that OE builds for that could do with some work, preferably with either a keyboard or wifi Sep 06 09:18:56 hrw|work: ok, thanks. it's definitely not there. i'll 'build linux-omap1' again, i guess Sep 06 09:21:02 mikearthur: wrt54? Sep 06 09:21:05 mikearthur: could you use the metric system? e.g. a NSLU2 with USB keyboard+USB wifi could be below lb200 Sep 06 09:21:53 hrw|work: I was thinking more PDA-like Sep 06 09:22:18 zecke: are you being funny or do you want in euros? Sep 06 09:22:50 mikearthur: hehe, I'm a moron. I forgot that lb means pounds :) Sep 06 09:24:31 :) Sep 06 09:25:35 I just want to keep hacking on oe Sep 06 09:25:47 but once I leave here I'll need my own kit :) Sep 06 09:26:23 mikearthur: nslu is nice target Sep 06 09:28:17 OE can do builds for those 100 GBP Ipaqs cant it? Sep 06 09:29:02 this is quite odd... when i do a 'bitbake virtual/kernel' it reports the build is successful, but i see: Sep 06 09:29:02 Attempted builds: 0 Sep 06 09:30:01 http://rafb.net/paste/results/bDPR0e15.html Sep 06 09:32:03 that means it already build a provider for virtual/kernel Sep 06 09:32:57 can i force it to build the kernel again? it's not listed in my deploy/images dir Sep 06 09:33:07 bitbake -s | grep kernel Sep 06 09:33:21 that should show what OE thinks it has to build for a kernel Sep 06 09:33:22 ok, i'll give that a go. thanks, koen Sep 06 09:41:43 is that "bitbake -s | grep kernel" supposed to take a long time and generate no output? Sep 06 09:41:51 never done this before, just want to make sure nothing's hung Sep 06 09:42:28 can anyone tell me how I can begin to trace this problem down? I've looked and searched, but cannot find what happend or where: http://rafb.net/paste/results/FF9AKy98.html Sep 06 09:43:24 It's complete nonsense to me Sep 06 09:43:40 T0mW: show us that recipe Sep 06 09:43:45 ok Sep 06 09:45:36 This is the recipe: http://rafb.net/paste/results/3iNwvN74.html Sep 06 09:45:52 lots of patches to this one, heh Sep 06 09:46:13 hrw|work: results of bitbake -s | grep kernel :: http://rafb.net/paste/results/nSTO7T62.html Sep 06 09:46:31 hmmm Sep 06 09:46:41 i could have sworn it listed virtuals as well Sep 06 09:48:13 AFAICT, it has something to do with finding the modules within the kernel build dir, it appears to be stripping out maybe modinfo? Sep 06 09:49:41 T0mW: no idea now. sorry Sep 06 09:50:12 hrw|work: how can I tell what classfile was involved ? Sep 06 09:50:32 bigd0g: bitbake -i then: parse;getvar PREFERRED_PROVIDER_virtual/kernel Sep 06 09:50:45 bigd0g: or 'rebuild virtual/kernel' inside of bitbake shell Sep 06 09:51:00 T0mW: it inherit kernel.bbclass Sep 06 09:51:09 hrw|work: ok, i'll give that a try Sep 06 09:51:13 hrw|work: I think kergoth added 'rebuild' to base.bbclass yesterday Sep 06 09:51:28 http://www.openembedded.org/viewmtn/revision.psp?id=331c796786f820575978cca1b06ff1676cbbfb45 Sep 06 09:52:24 bitbake does produce a bootable zImage though Sep 06 09:52:37 koen: running latest (42be14012b6790df46c62295f3a02286acecbd4f), so i think rebuild is in there Sep 06 09:52:59 'bitbake rebuild virtual/kernel' however, doesn't work Sep 06 09:53:06 -c rebuild Sep 06 09:53:13 ahh Sep 06 09:53:34 "bitbake -i" gets to a console, too Sep 06 09:54:28 very good. this looks like it might work. thanks, hrw|work Sep 06 09:55:04 koen: there is no such task as rebuild rather... Sep 06 09:55:22 bitbake -i;rebuild something == bitbake -cclean something;bitbake something Sep 06 09:56:00 +addtask rebuild Sep 06 09:56:01 +do_rebuild[dirs] = "${TOPDIR}" Sep 06 09:56:01 +do_rebuild[nostamp] = "1" Sep 06 09:56:01 +do_rebuild[bbdepcmd] = "" Sep 06 09:56:01 +python base_do_rebuild() { Sep 06 09:56:02 + """rebuild a package""" Sep 06 09:56:03 + bb.build.exec_task('do_clean', d) Sep 06 09:56:05 + bb.build.exec_task('do_' + bb.data.getVar('BB_DEFAULT_TASK', d, 1), d) Sep 06 09:56:10 no such task? Sep 06 09:56:17 ah Sep 06 09:56:26 rebuild is definitely working :) Sep 06 09:58:37 morning all Sep 06 09:58:45 hi Liam Sep 06 09:58:59 hey lrg Sep 06 09:59:34 hey Liam Sep 06 10:06:15 koen: argh, I've forgot the package name from commit line yesterday, I don't think there is anyway to change it later. Sep 06 10:06:35 no sweat Sep 06 10:06:57 we don't expect people to get it right the first try :) Sep 06 10:07:13 koen: i should rtfm more often ;) Sep 06 10:18:08 read the fine manual Sep 06 10:18:29 clearly ;) Sep 06 10:20:12 03lrg 07org.oe.dev * r0cc40044... 10/ (3 files in 2 dirs): added alsa support to generic device table Sep 06 10:20:49 ph34r the lrg Sep 06 10:23:02 Crofton: you have a second? Sep 06 10:23:18 sure Sep 06 10:23:35 i'm the process of updating everything on my board Sep 06 10:23:40 yesterday i got u-boot updated Sep 06 10:23:45 cool Sep 06 10:23:49 today i'm trying to get the kernel and rootfs updated Sep 06 10:23:54 ok Sep 06 10:24:07 i'm starting with the kernel... which i think is zImage-omap5912osk-20060906100754.bin in my deploy/images dir Sep 06 10:24:11 is that correct? Sep 06 10:24:15 yeah Sep 06 10:24:20 ok, i copied it over Sep 06 10:24:26 cp.b'd it to the right spot Sep 06 10:24:32 but... when i try to boot... it says: Sep 06 10:24:50 ## Booting image at 00100000 ... Bad Magic Number Sep 06 10:24:51 drum roll Sep 06 10:24:54 have you seen that before? Sep 06 10:25:04 hmmm Sep 06 10:25:04 sorry ... i was cutting and pasting :P Sep 06 10:25:09 :) Sep 06 10:25:24 I love the dramatic pauses on irc :) Sep 06 10:25:30 as do i Sep 06 10:25:39 hmm Sep 06 10:25:51 can you try the imls commands? Sep 06 10:26:01 such as? Sep 06 10:26:03 the ones that do the image checks Sep 06 10:26:10 imls I think Sep 06 10:26:20 do help at the u-boot prompt Sep 06 10:26:31 I am still at home, no osk handy Sep 06 10:26:42 'imls' lists all the images Sep 06 10:26:51 when i execute it, it returns nothing (since i wiped the board) :) Sep 06 10:27:30 try iminfo? Sep 06 10:27:40 you have the kernel in flash at 0x100000? Sep 06 10:28:03 kernel is in memory at 0x10000000, not flash Sep 06 10:28:09 checksum passed Sep 06 10:28:09 hmmm Sep 06 10:28:32 can you check the version of u-boot? Sep 06 10:28:39 latest... 1.1.4 Sep 06 10:28:53 U-Boot 1.1.4 (Aug 29 2006 - 17:05:48) Sep 06 10:29:04 ok Sep 06 10:29:40 I am thinking it may not boot from address 0x1000000 Sep 06 10:31:15 if you look in the bb file at the line with mkimage in it, there are some addresses specified Sep 06 10:31:15 hrm... that would be odd Sep 06 10:31:36 the newbie guide says to copy it to 0x1000000 in flash Sep 06 10:31:47 ok, i'll check the bb file Sep 06 10:31:58 hmmm Sep 06 10:32:05 but maybe the addresses are ram Sep 06 10:32:18 I always put my kernel at 0x100000 in flash Sep 06 10:32:43 I may be confused about addresses ..... Sep 06 10:32:53 no, you're right ... 0x100000 ... i added an extra 0 Sep 06 10:33:10 i copied the comand from the newbie guide at the console, tho Sep 06 10:33:18 it decompressess kernel to ram Sep 06 10:33:18 cp.b 0x10000000 0x100000 $(filesize) ... so that's right Sep 06 10:35:42 Crofton: mkimage -A arm -O linux -T kernel -C gzip -a 0x10c08000 -e 0x10c08000 -n "OE" -d Sep 06 10:35:51 that line? Sep 06 10:36:15 yeah Sep 06 10:36:30 thse may define the addresses to unpack the kernel to though Sep 06 10:36:44 based on the number of digits Sep 06 10:36:55 hrm Sep 06 10:37:00 anyway your kernel is in flash at 0x100000 Sep 06 10:37:05 correct? Sep 06 10:37:26 the copy said it was successful, but 'imls' doesn't show it Sep 06 10:37:33 can you type version at the u-boot prompt? Sep 06 10:37:39 try iminfo Sep 06 10:38:10 'lminfo' works ... checksum passed ... but "## Checking Image at 10000000 ..." Sep 06 10:38:20 is that checking that location in flash or memory ? Sep 06 10:39:16 0x10,000,000 is RAM Sep 06 10:41:20 * bigd0g ponders Sep 06 10:41:48 does "cp.b 0x10000000 0x100000 $(filesize)" appear correct? copy from 0x10,000,000 (ram) to 0x100,000 (flash) Sep 06 10:42:08 yeah Sep 06 10:42:17 I always use the hex size Sep 06 10:42:29 did you erase the flash first? Sep 06 10:42:41 it was completely erased except for uboot, yeah Sep 06 10:43:00 i tried the copy again... it says successful... 'imls' still shows nothing Sep 06 10:43:15 try iminfo Sep 06 10:43:28 there are two commands that do slightly different things Sep 06 10:43:40 iminfo reports information for 0x10,000,000 (ram) still Sep 06 10:43:47 also what is the result of typing version at the u-boot pormpt Sep 06 10:43:49 hmmm Sep 06 10:43:52 brb Sep 06 10:44:08 re: u-boot ... same as before ... U-Boot 1.1.4 (Aug 29 2006 - 17:05:48) Sep 06 10:46:27 hrm. Sep 06 10:46:28 progress Sep 06 10:47:08 ## Booting image at 00100000 ... Verifying Checksum ... Bad Data CRC Sep 06 10:48:19 checksum is ok in memory, but not after copy ... hrm Sep 06 10:48:28 (in flash, i mean) Sep 06 10:49:43 hmm Sep 06 10:49:49 what did you change? Sep 06 10:50:09 i just did another "erase 1:8-15" and re-did the copy ... then tried bootm 0x10,000 again Sep 06 10:50:28 erase 1:8-16 and repeat Sep 06 10:50:41 flash Sep 06 10:50:50 ok, trying again Sep 06 10:51:25 back to "bad magic number" Sep 06 10:51:44 ## Booting image at 00010000 ... Bad Magic Number Sep 06 10:52:11 i used the same copy command in all tries. just copying and pasting Sep 06 10:52:12 cp.b 0x10000000 0x100000 $(filesize) Sep 06 10:52:34 rather than use filesize Sep 06 10:52:55 I use the hex value for number of bytes transfered Sep 06 10:53:07 it is shown when the tftpboot finished Sep 06 10:53:13 not sure about kermit Sep 06 10:53:27 yeah... but i'm using kermit... it doesn't show it :( Sep 06 10:53:37 i used the same var yesterday when doing the u-boot upgrade Sep 06 10:53:48 it simpy uses the filesize of the last file transferred Sep 06 10:53:52 s/simpy/simply Sep 06 10:53:59 when you display it, does it make sense? Sep 06 10:54:29 11E68B is var value ... let me convert that Sep 06 10:55:27 imls or iminfo should give you the kernel ifo Sep 06 10:56:17 11E68B(h) = 1173131(d) = ~1.1 Mbytes ... looks right to me Sep 06 10:57:26 yeah, the number is right... 1173067 Bytes = 1.1 MB... acounting for my conversion error Sep 06 10:57:40 (latter info is from lminfo) Sep 06 10:57:46 er... 'imls' Sep 06 10:58:11 bother Sep 06 10:58:22 I was trying to log into my OSK over ssh Sep 06 10:58:37 imls and iminfo data match ... imls is from 0x10,000 and iminfo reports 0x10,000,000 Sep 06 10:58:39 but ctrl a is used by ssh also .... Sep 06 10:58:47 haha... doh ;) Sep 06 10:58:50 and you get bad magic number Sep 06 10:59:13 yep. bad magic number Sep 06 10:59:19 funny error, if you ask me :) Sep 06 10:59:38 wait Sep 06 10:59:44 victory ! Sep 06 10:59:58 I think bad magic number is when u-boot does not recognize the number in thekernel image Sep 06 11:00:15 i erased it, re-copied it... and now it works Sep 06 11:00:18 * bigd0g scratches his head Sep 06 11:00:19 ok Sep 06 11:00:30 erase 1:8-16 ,,,,, Sep 06 11:00:51 bootm 0x100000 works now ... should i try again? Sep 06 11:00:56 sometimes your fingers need training Sep 06 11:01:05 quit while you are ahead Sep 06 11:01:11 yeah Sep 06 11:01:17 perhaps something was slightly off. idk. Sep 06 11:01:39 yeah Sep 06 11:02:18 now the kernel is booting. since it's stuck at sending dhcp requests, i imagine i need to have it hooked by ethernet to me? Sep 06 11:02:30 i guess i've taken it as far as i can via serial Sep 06 11:02:31 or turn off network Sep 06 11:02:35 yeah Sep 06 11:02:38 get a router Sep 06 11:03:00 one of the cheesy linksys ones ... Sep 06 11:03:47 yeah, i have a bunch. they're 3,500 miles away. i'll have to see if there's one lying around here Sep 06 11:04:04 so annoying when you don't have any of your actual stuff handy Sep 06 11:04:08 heh Sep 06 11:04:11 yeah Sep 06 11:04:33 RP: I will enable iptables subsystem in 2.6.16 kernels Sep 06 11:04:35 moving the file system image over serial will likely be annoying Sep 06 11:05:20 here's a random question: how do i get it to auto boot on reset ? Sep 06 11:05:36 right now it just sits at the osk prompt, even with the kernel installed Sep 06 11:06:23 do you get a countdown? Sep 06 11:06:27 nope Sep 06 11:06:35 there is some u-boot var Sep 06 11:06:41 *** Warning - bad CRC, using default environment Sep 06 11:06:42 delay boot or something Sep 06 11:06:49 not sure if that is bad or not Sep 06 11:06:59 that's on a fresh reset Sep 06 11:07:00 I think you need to set up the image address Sep 06 11:07:06 RP, do13, mikearthur: did you ever tried to use CONFIG_USB_SUSPEND option? Sep 06 11:07:08 do a printenv and pastebin Sep 06 11:07:14 ok Sep 06 11:07:43 RP, do13, mikearthur: did you ever tried to use CONFIG_USB_SUSPEND option? this allow to suspend/resume individual usb peripherials. can be usefull on tosa to shutdown internal wifi Sep 06 11:08:08 Crofton: http://rafb.net/paste/results/GkZRxh40.html Sep 06 11:08:56 there are some more that need setting Sep 06 11:09:06 there is one that sets the image address Sep 06 11:09:21 I can pastebin mine in a couple of hours ..... Sep 06 11:09:24 wait Sep 06 11:10:24 http://rafb.net/paste/results/2cR5QU26.html Sep 06 11:10:28 I had saved a copy Sep 06 11:10:33 cool Sep 06 11:10:56 that should solve the booting problem Sep 06 11:11:27 yeah, that looks more like the env on the old uboot Sep 06 11:11:33 probably should've saved it initially Sep 06 11:11:37 hindsight is 20/20 Sep 06 11:14:23 does uboot have a setenv_from_pastebin option? Sep 06 11:21:48 koen: while working on sanitize.py, I found some others typos Sep 06 11:21:59 packages/sn9c102/sn9c102_1.32.bb and packages/tcp-wrappers/tcp-wrappers_7.6.bb need s/DESRIPTION/DESCRIPTION/ Sep 06 11:22:10 packages/xorg-app/xprop_cvs.bb needs s/MAINTER/MAINTAINER/ Sep 06 11:22:20 and packages/otpkeygen/otpkeygen_1.3.0.bb needs s/MAINTANER/MAINTAINER/ Sep 06 11:26:56 Crofton: i've got some weird string " e[[3~add[ " in my env vars any way to remove it? Sep 06 11:27:34 i tried "setenv e[[3~add[" since "setenv " is supposed to remove a var Sep 06 11:27:38 but... it's still there Sep 06 11:27:49 morning all Sep 06 11:27:58 hrw|work: I've never tried USB suspend Sep 06 11:28:04 Crofton: n/m ... i just reset without a save Sep 06 11:28:48 JustinP: I might have a solution to the alsa issues as there have been a lot of changes there and I need to update to a new asoc. The SD issues aren't good though. Were the block i/o errors from SD? Sep 06 11:30:17 03koen 07org.oe.dev * r6be0c561... 10/ (5 files in 5 dirs): various: correct typos spotted by Cyril Romain Sep 06 11:30:51 RP: under 2.4.18 sharp used own usbctl tool to disable/enable usb wifi on tosa. USB suspend maybe work same way - have to check it later Sep 06 11:31:16 ~lart people editing generated autofoo for doing a release Sep 06 11:31:16 * ibot pulls out a ClueBat (tm) and thwaps people editing generated autofoo for doing a release Sep 06 11:32:22 hrw|work: It sounds like a good idea Sep 06 11:32:39 RP: too bad that I forgot to take card reader here... Sep 06 11:34:12 hey RP Sep 06 11:34:20 hi Liam Sep 06 11:35:56 hay Sep 06 11:37:02 hi niv_one_three Sep 06 11:37:15 may I use any compact flash with my zaurus in OE (question mark) Sep 06 11:37:16 hrw|work: looks like you can turn off a usb port just by writing to sysfs Sep 06 11:37:26 XorA: s/port/device Sep 06 11:37:56 * hrw|work want kernel config builder in OE... Sep 06 11:38:08 2. does anyone have a udev rule to load usbnet handey (question mark) Sep 06 11:38:08 hrw|work: well you could suspend the host driver which removes all ports :-) Sep 06 11:38:17 hrw|work: We should add a target that runs make menuconfig :) Sep 06 11:38:27 do_configure_prepend() { combine 1 2 3 4 5 into .config} Sep 06 11:38:28 In fact, I might well do that... Sep 06 11:38:55 RP: out 2.6.16 defconfigs differ too much in !machine related parts... Sep 06 11:39:06 XorA: on tosa I do it by rmmod Sep 06 11:39:12 XorA: on each suspend Sep 06 11:39:19 hrw|work: I can see advantages and disadvantages in doing that for the defconfigs... Sep 06 11:40:10 RP: me too Sep 06 11:41:31 hrk|work; did you ever came across CF cards that didnt work in OE Sep 06 11:41:40 niv_one_three: s/OE/OZ? Sep 06 11:42:18 hrw; I got a simpad and a new Jordana Sep 06 11:42:32 btw jordana 720 at 75$ Sep 06 11:42:33 florian_kc: can you look into .oz354x to check why gconf-dbus is not building? Sep 06 11:42:35 at buy.com Sep 06 11:42:42 ~change 75 usd to pln Sep 06 11:42:51 75.00 United States Dollar (USD) makes 231.662 Zloty (PLN) (from http://www.xe.com/) Sep 06 11:43:06 nice price Sep 06 11:43:35 so, should all CF work. even x40 and x60. Sep 06 11:43:42 niv_one_three: yes Sep 06 11:43:44 (question mark) Sep 06 11:43:50 10x Sep 06 11:44:11 niv_one_three: any cf memory card will work. all will work slow Sep 06 11:44:21 2.does any one here uses a udev rule to load usbnet (question mark) Sep 06 11:44:51 niv_one_three: echo 'g_ether'>/etc/modutils/g_ether;update-modules Sep 06 11:45:00 niv_one_three: if it is on device Sep 06 11:45:14 lrg: ASoC 0.11.2 seems to have the same problem on spitz :-( Sep 06 11:45:21 niv_one_three: on host I have usbnet in kernel Sep 06 11:45:29 RP: :( Sep 06 11:45:46 hrw; i mean on the pc side Sep 06 11:46:03 niv_one_three: on host (pc side) I have usbnet in kernel Sep 06 11:46:23 RP: can you send me the debug again, it shouldn't be the bit size atleast Sep 06 11:46:59 lrg: ok Sep 06 11:47:25 hrw|work: ok Sep 06 11:47:47 thx florian_kc Sep 06 11:49:51 hrw; i mean have the iptables commands run when the simpad is connected to the usb Sep 06 11:52:51 niv_one_three: add them into usb0 definition on pc side Sep 06 11:54:26 hrw|work: btw, what new features can we expect in f-b? Sep 06 11:54:55 morning Sep 06 11:55:08 lrg: sent Sep 06 11:55:45 koen: searching after extra fields, better package view, better view in small resolutions Sep 06 11:56:55 ok Sep 06 11:57:04 * koen looks at Packages.filelist Sep 06 11:59:00 gnr Sep 06 11:59:22 I see no sane way to parse that in one go without using idiotic amounts of memory Sep 06 11:59:51 food time Sep 06 11:59:53 koen: feed it into SQL database Sep 06 12:00:27 yeah Sep 06 12:00:44 it also lacks version info for the packages Sep 06 12:01:23 XorA: i wonder what would be a sane table layout for that Sep 06 12:02:41 koen sane?? there is none :) Sep 06 12:11:05 * bigd0g goes looking for a router Sep 06 12:23:01 koen: It does look like ipkg-make-index needs altered to add a version column in that Sep 06 12:26:59 Crofton: another silly question: what's the default login/pass ? :) Sep 06 12:27:06 root Sep 06 12:27:10 return? Sep 06 12:27:12 oh, no pass Sep 06 12:27:13 haha Sep 06 12:27:21 that's... secure ;) Sep 06 12:27:34 well Sep 06 12:27:47 better than having a board you can't log in to Sep 06 12:27:57 indeed. it appears to have worked Sep 06 12:28:01 ssh forces you to at least change from the default Sep 06 12:28:03 cool Sep 06 12:29:44 hi folks Sep 06 12:33:10 koen: ipkg-sqlite would be interesting Sep 06 12:35:58 florian_kc: I suspect wrong pkgconfig search path Sep 06 12:36:39 hrw|work: the error was a missing MODINFO / MODULE_DESCRIPTION macros in a kernel module source Sep 06 12:36:51 T0mW: popular in 2.4... Sep 06 12:37:00 gak! Sep 06 12:37:16 2.4 was a bit of a free-for-all Sep 06 12:50:54 hrw|work: it would indeed be very interesting Sep 06 12:55:03 folks, anybody knows when svn.berlios.de will be back up? Sep 06 12:55:35 nope Sep 06 12:55:40 hi koen Sep 06 12:55:43 hey chris144 Sep 06 12:56:21 it seems berlios does "we're down for maintenance at" . date(today) . "for 10 till 24" Sep 06 12:56:27 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Sep 06 12:57:41 ok Sep 06 13:04:04 oh ya , when will ikpgfind be back (qestion mark) Sep 06 13:06:27 try asking in a channel that cares Sep 06 13:06:32 e.g. handhelds.org Sep 06 13:06:53 we have http://ewi546.ewi.utwente.nl/tmp/hrw/fb/ and http://www.angstrom-distribution.org/repo/ Sep 06 13:09:35 ipkgfind is now in rewritting phase - porting to php5 etc. Sep 06 13:09:56 I had an offer of doing that but lacked time Sep 06 13:12:06 so, when will it be back/ Sep 06 13:12:29 why not leave it on, while rewriting/ Sep 06 13:13:39 koen: now add a query page to the tinderbox? to see all builds of PN, PV, PR Sep 06 13:20:06 http://pastebin.ca/162361 <- nice security issue Sep 06 13:20:21 koen; sorry Sep 06 13:20:28 to many screens Sep 06 13:21:06 ok , being a newbie to linux on simapd - should i try angstorm (question mark) Sep 06 13:21:18 i am good i in gentoo linux Sep 06 13:21:19 .. Sep 06 13:22:40 wow http://www.angstrom-distribution.org/repo is sweet Sep 06 13:31:07 nobody like me because i didnt red koen response in time ;-) Sep 06 13:31:13 sorry ;-( Sep 06 13:33:01 hrw|work: the new tosa patches should be done today Sep 06 13:36:36 mikearthur: nice Sep 06 13:37:04 mikearthur: did you read Liam mail about creating wm97xx bus? Sep 06 13:37:18 hrw|work: thats been done Sep 06 13:37:23 aha Sep 06 13:39:23 hrw|work: the ipkgfind situation is funny, since I did ~90% of the work needed to get it working with php5 last year Sep 06 13:39:55 I even sent the patches upstream Sep 06 13:49:32 c;) Sep 06 13:57:15 ah, it seems berlios is back up Sep 06 14:00:31 hrw|work: it builds here :-} Sep 06 14:02:53 florian_kc: pkgconfig changes? Sep 06 14:03:10 florian_kc: uninstall pkgconfig from host enviroment Sep 06 14:04:16 hrw|work: patches done, sent to liam Sep 06 14:04:49 mikearthur: thx Sep 06 14:05:02 mikearthur: you have tosa with you? Sep 06 14:05:17 hrw|work: you mean the physical device? Sep 06 14:05:23 mikearthur: can you build kernel with USB_SUSPEND enabled and test can it be used to switch on/off internal wifi? Sep 06 14:05:27 mikearthur: yes Sep 06 14:05:58 how would I test that? Sep 06 14:06:37 mikearthur: help from that kernel option explain that Sep 06 14:06:40 ok Sep 06 14:06:50 mikearthur: http://pricejapan.com/front/e_good_info.php?code=31&category=5 Sep 06 14:06:59 less than 200GBP Sep 06 14:07:09 mikearthur: Device drivers -> USB support -> USB selective suspend/resume and wakeup Sep 06 14:07:36 do I need any recent patches? haven't updated my tree in a while Sep 06 14:07:53 mikearthur: no - its mainline thing Sep 06 14:08:44 I'll give it a go Sep 06 14:08:48 thx Sep 06 14:09:12 hrw|work: a cool result of the new patches are you can modprobe/rmmod tosa_ts.ko and see the interference off/on :D Sep 06 14:09:49 mikearthur: I think that tosa_ts will end in kernel rather Sep 06 14:10:01 yeh, it probably will Sep 06 14:10:04 but Sep 06 14:10:05 :D Sep 06 14:10:13 its cool to be able to see before/after Sep 06 14:10:26 hrw|work: i moved the pkg-config binary away before building Sep 06 14:10:37 s/the/the host Sep 06 14:10:56 koen: what is that? spitz or something? Sep 06 14:11:00 mikearthur: akita Sep 06 14:11:25 mikearthur: spitz is akita+microdrive (in short words) Sep 06 14:11:26 the akita is slightly over 200gbp iirc Sep 06 14:11:41 florian_kc: hm. Sep 06 14:12:00 ~change 30200 jpy to gbp Sep 06 14:12:37 3200 is 286GBP Sep 06 14:12:41 includng shppng Sep 06 14:13:03 * koen wonders how XorA knows that so well Sep 06 14:13:13 koen: page is open, Im buying one Sep 06 14:13:21 :p Sep 06 14:13:23 florian_kc: then I do not know what is fscked.. I cant build it on work machine or on ewi Sep 06 14:13:52 XorA: don't you already have a similar thing? Sep 06 14:13:58 XorA: so you now will join club of 'multiple zaurus users'? Sep 06 14:14:19 hrw|work: finally I will be in the elite :-) Sep 06 14:14:33 mikearthur: I need one for ANgstrom, and one to use day to day Sep 06 14:14:41 XorA: ok Sep 06 14:14:46 XorA: currently I have 4 Z at home Sep 06 14:15:06 mickey|sports: still want poodle? Sep 06 14:15:39 hrw|work: your just a 4 armed wierdo :-) Sep 06 14:16:12 http://cgi.ebay.co.uk/SHARP-ZAURUS-SL-C3100-LINUX-PDA-CF-CARD-WiFi-more_W0QQitemZ180024892213QQihZ008QQcategoryZ38331QQssPageNameZWDVWQQrdZ1QQcmdZViewItem Sep 06 14:16:18 Hello. Sep 06 14:17:31 hi sirfred Sep 06 14:19:09 So, the C3100 has host USB capabilities? What other zauruses are able of that? Sep 06 14:19:22 c6000, c1000, c3x00 Sep 06 14:19:33 http://www.kdedevelopers.org/node/2320 Sep 06 14:20:01 koen: Thx Sep 06 14:20:23 chouimat: :) Sep 06 14:22:52 ;))) Sep 06 14:23:09 hrw|work: kernel compiled and running, have a shell, what now? Sep 06 14:23:29 mikearthur: lsusb show you 2 usb devices? Sep 06 14:23:47 shows none :) Sep 06 14:24:15 mikearthur: mount -t usbfs none /proc/bus/usb;lsusb Sep 06 14:26:04 * cyrilRomain hadded the Format Guidelines into the sanitize.py script Sep 06 14:26:21 mikearthur: then cd /sys/bus/usb/ and experiment with power/state of usb wifi device Sep 06 14:26:45 cyrilRomain: where is latest version? Sep 06 14:27:07 hrw|work: I'll pastebin it Sep 06 14:27:40 ok Sep 06 14:28:26 sanitize.py: http://rafb.net/paste/results/qvMxlB34.html Sep 06 14:28:51 so , must I build angstrom using bitbake , or are there common bin available/ Sep 06 14:29:23 I think that as I read more I will ask less FAQ Sep 06 14:29:43 niv_one_three: angstrom doesn't work on strongarm (yet), try using openzaurus-unstable Sep 06 14:29:51 ~lart gcc from breaking armv4 Sep 06 14:29:56 s/from/for/ Sep 06 14:30:08 It seems that oprofile-0.9.1+cvs20060214 sources are not available neither at ewi nor oesources.org. And sf cvs seems not to work. Sep 06 14:30:23 koen; oh I see Sep 06 14:30:35 cyrilRomain: #6 "spaces or tabs on empty lines" -> \s in perl regurals Sep 06 14:31:02 oh , ya I almost forgot this; where to seek drivers for atheros(question mark) Sep 06 14:31:15 niv_one_three: atheros == madwifi? Sep 06 14:31:39 they are called madfi, I got a Nec pc card I want to use in my simpad Sep 06 14:31:48 madwifi Sep 06 14:32:00 this is y I asked 4 ipkgfind Sep 06 14:32:17 niv_one_three: you have CardBus card, not pcmcia Sep 06 14:32:37 oh, I will google the difference Sep 06 14:32:50 niv_one_three: cardbus = 32bit, pcmcia = 16bit. Sep 06 14:32:56 so, can it work (question mark) Sep 06 14:32:57 16 extra wires Sep 06 14:33:02 niv_one_three: cardbus slot accept pcmcia cards but not viceversa Sep 06 14:33:17 cool Sep 06 14:33:21 hrw|work: no /proc/bus/usb :D Sep 06 14:33:30 mikearthur: modprobe ohci-hcd Sep 06 14:33:34 where to seek the ipkg for madwifi Sep 06 14:34:17 hrw|work: sorted now Sep 06 14:34:26 niv_one_three: this card cant be used in simpad Sep 06 14:35:08 hrw; sweet how(question mark) Sep 06 14:36:57 hey mallum Sep 06 14:37:02 hey koen Sep 06 14:37:28 koen: did you ever have much luck actually running iwmmx built images ? Sep 06 14:38:17 hi mallum Sep 06 14:38:27 hey hrw|work Sep 06 14:38:40 hrw|work: yes, matching \s* is what isspace() do but the isspace() handle the case were the line is empty (without space nor tab) differently. \s* will match empty line, not isspace() Sep 06 14:38:50 aha cant Sep 06 14:38:54 mallum: yes Sep 06 14:38:55 cyrilRomain: ah. did not know Sep 06 14:39:48 koen: X, gdk-pixbuf etc not all crashing left right and center ? Sep 06 14:40:19 mallum: not if you login as root Sep 06 14:40:41 * CosmicPenguin ducks and tries to build python for the target Sep 06 14:41:06 koen: ? Sep 06 14:41:39 mallum: X stuff breaks when run as a user, but not when run as root over here Sep 06 14:42:01 * mrz80 is back (gone 01:45:35) Sep 06 14:42:02 XorA, pH5 and I theorize shm is broken in kdrive Sep 06 14:43:00 koen: well kdrive has never run not as root, it actually does a uid check on startup Sep 06 14:43:36 I can do Xfbdev -ac & as root Sep 06 14:43:40 * XorA had wondered if /dev/shm was needed Sep 06 14:44:04 * mallum wonders how Xephyr runs Sep 06 14:44:59 but I think the bug snt actuay in kdrive, it just happens to trigger it Sep 06 14:45:08 s/snt/isnt/ Sep 06 14:46:04 koen: tomas here is close to having the fbmmx foo in kdrive ported to iwmmx Sep 06 14:46:17 a, way cool Sep 06 14:46:22 mallum: very cool Sep 06 14:46:38 I guess we can bother him next month at OEDEM Sep 06 14:46:49 koen: but it seems a reliable toolchain cannot be created to test :( Sep 06 14:46:52 Xw100 + mmx stuff would fly on my hx Sep 06 14:46:58 mallum: correct Sep 06 14:47:12 mallum: either it's that csl crap, or gnu crap Sep 06 14:47:43 cyrilRomain: do you know where in debian I can find odict python module? Sep 06 14:47:49 ImportError: No module named odict Sep 06 14:47:56 koen: so you dont have one either ? Sep 06 14:48:15 I've changed cvs://anonymous@cvs.sourceforge.net/cvsroot/oprofile;module=oprofile with cvs://anonymous@oprofile.cvs.sourceforge.net/cvsroot/oprofile;module=oprofile in the oprofile_cvs.bb and now it's checking out correctly, at least. Sep 06 14:48:25 mallum: the glibc 2.4/gcc4.1/binutils 2.17 is working pretty good Sep 06 14:48:36 sirfred: .dev or .oz354x? Sep 06 14:48:53 hrw|work: .dev I guess Sep 06 14:48:54 mallum: but we are seeing the shm issues, which we can't seen to pinpoint Sep 06 14:49:37 hrw|work: I'm pulling from org.openembedded.dev, if that's what you're asking. Sep 06 14:49:59 RP: mmcblk0: error 1 sending read/write command Sep 06 14:50:12 hrw|work: http://www.voidspace.org.uk/python/odict.html Sep 06 14:50:22 RP: end_request: I/O error, dev mmcblk0, sector0 Sep 06 14:50:22 hrw|work: I don't think you can find it in debian Sep 06 14:50:27 mallum: I suspect nptl is wonky on arm Sep 06 14:50:55 RP: sector is 0, 2, 249, or 251 (7 or so of each) Sep 06 14:51:13 koen: hmmm Sep 06 14:51:30 Hi! Sep 06 14:51:35 RP: also, the SD card seems to be fine when I plug it into my PC, so it doesn't seem to be corrupting it, just failing to read.... Sep 06 14:51:36 mallum: the csl stuff doesn't suffer from that, since they hacked a shimlib into glibc 2.3 Sep 06 14:51:41 How do I register for OE wiki? Sep 06 14:52:00 koen: I think this is what tomas last attempted the csl foo Sep 06 14:52:21 cyrilRomain: got it in rest2web Sep 06 14:52:31 hrw|work: ah Sep 06 14:52:35 psokolovsky: ATM you can't Sep 06 14:52:45 psokolovsky: it's an item for discussion at OEDEM Sep 06 14:52:47 * cyrilRomain looks into Gentoo for a similar package Sep 06 14:53:04 koen: ok, will fix typos anonymously Sep 06 14:53:16 mallum: tomas was using glibc 2.4 Sep 06 14:53:42 koen: just in case, it lacking a changelog entry for edit sucks too Sep 06 14:53:51 JustinP: I'm afraid I have no idea what's wrong with the mmc code, unless its some timeout issue :-( Sep 06 14:53:55 cyrilRomain: can you change it to not generate empty lines between first lines? SECTION/HOMEPAGE/etc Sep 06 14:54:03 psokolovsky: the whole wiki is a bit sucky Sep 06 14:54:18 hrw|work: yes. Sep 06 14:54:27 RP: :-(. I only have one type of SD card so I can't test to see if it's an imcompatibility introduced with my card(s). Sep 06 14:54:36 RP: have you tested SD at all on spitz? Sep 06 14:54:44 mallum: angstrom is sticking with glibc 2.4 and nptl, since we'll have to use that eventually Sep 06 14:54:53 hrw|work: remove line 351 ;) Sep 06 14:55:02 mallum: I just wished that less people would resort to using CSL stuff Sep 06 14:55:15 koen: whats CSL? Sep 06 14:55:38 cyrilRomain: thats not solution Sep 06 14:55:39 mikearthur: Code Sourcery Labs, the dudes that work on toolchains paid by ARM Sep 06 14:55:48 k Sep 06 14:55:56 cyrilRomain: PR = "r5"\n\nSRC_URI should stay Sep 06 14:56:00 * JustinP wishes there was an easy way to install both sets or kernel modules... Sep 06 14:56:01 mikearthur: but they also make money with supporting their own patched-to-hell toolchains Sep 06 14:56:05 heh Sep 06 14:56:08 hrw|work: ah ok Sep 06 14:56:19 koen: don't submit the patches upstream? Sep 06 14:56:20 mikearthur: which is why upstream gcc and glibc are missing various bits Sep 06 14:56:25 cyrilRomain: same with }\n\ndo_install() {\n Sep 06 14:56:33 mikearthur: the gcc release dudes are from CSL Sep 06 14:56:48 hrw|work: this is still possible, but can you please put those rules to the wiki ? So that I'll update the script Sep 06 14:56:52 ....modules.tgz.... Sep 06 14:56:55 ok Sep 06 14:58:00 RP: Did you have any PXA UDC/g_ether issues with 2.6.17? Sep 06 14:58:45 psokolovsky: http://article.gmane.org/gmane.linux.ports.arm.kernel/27280 Sep 06 14:59:09 koen: drupal wiki does not allow to describe changeset? Sep 06 14:59:54 hrw|work: the wiki module doesn't do that Sep 06 14:59:55 cyrilRomain: first changeset: added RCONFLICTS, SRCDATE, removed WORKDIR Sep 06 15:00:07 koen: note exactly that. 2.6.17 appears to have problems not present in 2.6.16. I wonder, what's RP solution. Sep 06 15:00:21 JustinP: My two SD cards and an MMC card are fine Sep 06 15:00:31 psokolovsky: Which pxa? Sep 06 15:00:32 hrw|work: ok Sep 06 15:00:51 RP: I have issues on 255 ;-) Sep 06 15:01:11 RP: hmmm...odd..... I have 2 SanDisk Ultra II SD+USB cards. both worked fine in 2.6.16, both fail in 2.6.17 Sep 06 15:01:35 psokolovsky: I don't use the udc much and probably haven't tested it under 2.6.17 Sep 06 15:02:36 JustinP: do you have serial cable? Sep 06 15:02:49 JustinP: Do you see that error straight away or only after using the cards for a while? Sep 06 15:03:34 RP: gotcha. because pH5 confirmed it, wonder if for hx4700 (pxa270). so, there're messed __init/__exit stuff in g_ether. fixed in git, don't see the fix yet in 2.6.17.* Sep 06 15:04:08 psokolovsky: pxa27x is a totally different drive Sep 06 15:04:09 r Sep 06 15:05:13 hrw|work: I have one but it is only read-capable (last I tried it) Sep 06 15:05:36 psokolovsky:are you having problems with g_ether or g_serial? Sep 06 15:06:37 psokolovsky: I have been using g_ether on my poodle for a while now. In fact, I am connected through g_ether right now.:) (on a pxa250) Sep 06 15:06:42 RP: the cards don't work at all. those messages come up.....just after an ALSA failure during boot.... Sep 06 15:07:04 hvontres|poodle: Yes, g_ether on 2.6.17, with ipaq h4000. I have a solution though, just wonder if someone cares to push it into 2.6.17 mainline ;-) Sep 06 15:07:12 RP: ejecting/reinserting the card shows nothing in dmesg Sep 06 15:07:44 psokolovsky: What problems were/are you seeing? Sep 06 15:07:58 JustinP: I'm afraid you might have to dig into the changes between 2.6.16 and 2.6.17 - I don't know what the problem is and it all works for me :-( Sep 06 15:08:29 hvontres|poodle: are you using 2.6.17? Sep 06 15:08:32 JustinP: ALSA issues are being worked on Sep 06 15:08:42 RP: I know, I trust you ;-) Sep 06 15:08:47 JustinP: You have a cxx00, right? Sep 06 15:09:01 RP: I just thought it was interesting that the mmc failures happen after the asoc/alsa failure Sep 06 15:09:06 RP: C3000, yes Sep 06 15:09:21 I just tested MMC on a c3000... Sep 06 15:09:32 psokolovsky: I have searched for you tomorrow morning :( Sep 06 15:09:32 psokolovsky: I'm sorry for not having comitted your patch yet Sep 06 15:09:39 RP: I'll look but I doubt I'm going to find anything..... Sep 06 15:09:44 JustinP: You could try disabling ALSA incase it did something weird with the device I guess Sep 06 15:09:45 psokolovsky: Yes. I am currently using a build from July 24th. I might have to try out the newer build from OZ 3.5.4.2-rc2 Sep 06 15:10:18 psokolovsky: If I don't do it until tomorrow noon I owe you more than one drink Sep 06 15:10:19 JustinP: A binary search is guaranteed to give results eventually and just takes time input Sep 06 15:10:30 RP: well, the ALSA failure just before the mmc errors is a failure to prepare an IOCTL IIRC Sep 06 15:10:41 binary search? Sep 06 15:10:47 zecke: Hi, np. But it looks ok? Would it be committed to trunk or 1.6 branch too? Sep 06 15:11:13 JustinP: compile kernel with MMC_DEBUG, connect cable and save log from 'insert card, dd if=/dev/mmcblk0 of=/dev/null bs=1024' Sep 06 15:11:39 hvontres|poodle: Hm, well, that's strange. The breakage I see is that it doesn't work at all for static build, for dynamic, you got to be connected before modprobing it, or it won't work. Sep 06 15:11:49 psokolovsky: good question Sep 06 15:11:53 JustinP: Pick a git revision half way between 2.6.16 and 2.6.17, see if it works. google git bisect Sep 06 15:12:07 psokolovsky: I think, trunk first and then file another report for 'backport' Sep 06 15:12:29 JustinP: You repeat until you find the guilty change. Should take a maximum of 20 or so interations Sep 06 15:12:40 zecke: I'm using it with 1.6.1 for some time now, so it appears to be pretty stable. You see the code, anyway ;-) Sep 06 15:12:43 zecke: ok Sep 06 15:12:49 berlios is down again... Sep 06 15:12:58 RP: Hello. About oprofile. Sep 06 15:13:17 RP: Do I need to copy the zImage used as kernel to the zaurus fs? Sep 06 15:13:31 sirfred: no Sep 06 15:13:32 sirfred: No, the vmlinux file Sep 06 15:13:49 (uncompressed) Sep 06 15:13:52 RP: The one found in the kernel work directory? Sep 06 15:14:00 sirfred: yes Sep 06 15:14:27 RP: I see. I suppose I could try that. :-) (I see what you mean now by binary search, I just didn't realize you meant git revs) Sep 06 15:14:29 RP: I've seen there's a daemon or something so. Sep 06 15:14:51 oprofiled, do I need to just run that indicating the vmlinux file? Sep 06 15:15:08 hrw|work: I can possibly try that at home, but dmesg shows nothing right now in insert, only boot Sep 06 15:15:14 sirfred: No, use opcontrol --vmlinux=/somepath/vmlinux Sep 06 15:15:21 hrw|work: perhaps it would show mor eon boot, though Sep 06 15:15:28 RP: OK. Let's see. Sep 06 15:15:28 sirfred: Basically just use opcontrol and opreport Sep 06 15:15:43 RP: Thanks. Sep 06 15:16:24 http://www.openembedded.org/styleguide - with example recipe Sep 06 15:16:49 hrw|work: ok thx Sep 06 15:17:05 hrw|work: comments outside "" or {} bloc will not be easy to handle Sep 06 15:17:07 JustinP: Boot without inserting to capture the log later Sep 06 15:17:18 (inserting the card that is) Sep 06 15:17:25 psokolovsky: Hmm, maybe its a 255 specific thing. On my pxa250 it has been working since I tried the first 2.6.17 kernel back in june Sep 06 15:17:47 cyrilRomain: I try to avoid STH = "STHGG" # comments Sep 06 15:18:02 hvontres|poodle: ok, so maybe after all there're patches in OZ and I missed them. Sep 06 15:18:15 hrw|work: this is not the problem Sep 06 15:18:28 hrw|work: the problem is when the script parse the file, it sort lines Sep 06 15:18:49 and sorting comments accordingly is not easy unless there are within a "" or {} bloc Sep 06 15:18:57 cyrilRomain: ;( Sep 06 15:18:59 anyway, I'll try to find a solution for that Sep 06 15:19:30 anyone in here use KDE? Sep 06 15:19:35 hi everyone Sep 06 15:19:52 mikearthur: I used to Sep 06 15:20:01 tomas_f: hi Sep 06 15:20:05 mikearthur: I do Sep 06 15:20:05 mikearthur: I use it on my machine at home Sep 06 15:20:09 I'm just looking for anyone still using it to try and confirm some bugs for me :) Sep 06 15:20:21 tomas_f: Frydrych? Sep 06 15:20:24 https://bugs.kde.org/show_bug.cgi?id=126715, https://bugs.kde.org/show_bug.cgi?id=131869, https://bugs.kde.org/show_bug.cgi?id=126414 Sep 06 15:20:25 RP: ok. will try to get to it when I get back home (on the way to work right now) Sep 06 15:20:32 try and confirm them Sep 06 15:20:34 hrw|work: yep Sep 06 15:21:21 JustinP: ok, I wish I had better news :-/ Sep 06 15:21:27 RP: I'll try to do the bisecting today as well Sep 06 15:21:30 thanks :) Sep 06 15:21:36 hi tomas :) Sep 06 15:21:46 RP: work on ALSA is good news ;-) And the git bisecting should hopefully show what the problem is Sep 06 15:21:50 tomas_f: I remember one girl with that surname when I was at school years ago Sep 06 15:22:40 koen, I am trying to build an image for Akita using gcc4.1 -mtune iwmmx, but various apps segfault at boot time Sep 06 15:23:38 koen, I hear you have the 4.1+iwmmx toolchain working? Sep 06 15:24:12 RP: Should opcontrol terminates inmediatly ? Sep 06 15:24:18 RP: Or should it block the terminal? Sep 06 15:24:32 terminate is Ok I think Sep 06 15:25:07 sirfred: It should terminate and leave the daemon logging/running (is that's what you asked it to do) Sep 06 15:25:12 RP: There's a bad line in the opcontrol script at line 1517, btw. Sep 06 15:25:17 RP: It has just 'n'. Sep 06 15:25:39 sirfred: Someone needs to tell upstream... :) Sep 06 15:25:42 RP: Well, perhaps that's the problem, I didn't tell i to do nothing. Just runned opcontrol --vmlinux=path/to/vmlinux Sep 06 15:25:44 RP: :) Sep 06 15:26:15 sirfred: and did it terminate? Sep 06 15:26:20 RP: Yes. Sep 06 15:26:23 that's fine then Sep 06 15:26:34 RP: But there's no oprofiled daemon running... Sep 06 15:26:54 no, it would just know when vmlinux is now :) Sep 06 15:27:11 You need to start it. Have a look at the oprofile site, or opcontrol's help Sep 06 15:27:23 I can't remember the syntax off the top of my head Sep 06 15:27:39 hrw|work: after the 'PR' variables, the script now inserts empty lines between variables/routines Sep 06 15:27:52 RP: OK. Thanks. Sep 06 15:28:19 hrw|work: please feel free to add Format and/or Style Guidelines, especially from the list I sent to the mailing list Sep 06 15:28:39 cyrilRomain: sure - when/if find time Sep 06 15:28:49 sirfred: opcontrol --start to start, opcontrol --shutdown to finish, opreport tosee a profile Sep 06 15:28:57 hrw|work: :] Sep 06 15:29:11 sirfred: You can enable callgraphing for more detailed logs Sep 06 15:29:15 RP: Great! Sep 06 15:43:14 tomas_f: which version of busybox do you have? Sep 06 15:43:28 tomas_f: <1.2.x segfaults like crazy on EABI Sep 06 15:43:40 let me check Sep 06 15:44:45 1.01 Sep 06 15:45:07 yep, that will segfault like crazy Sep 06 15:45:19 koen: why does busybox care for ABI? Sep 06 15:45:19 tomas_f: install coreutils, or upgrade to busybox 1.2.x Sep 06 15:45:28 zecke: no idea, but it does Sep 06 15:47:07 koen, Ok, will try Sep 06 15:48:30 RP: I'm not able to put it to work. I'm starting with --start, then stopping with --shutdown (here opcontrol says 'Daemon stuck shutting down; killing !') but opreport always say me the same "opreport error: /var/lib/oprofile/samples/current/{kern}/g_ether/{dep}/{kern}/g_ether/CPU_CYCLES.100000.0.all.all.all: Invalid argument". Any idea? Sep 06 15:50:09 sirfred: I've never seen the daemon fail to unload :-( Sep 06 15:50:24 lrg: bad news - same probably with 0.11.3 :-( Sep 06 15:50:29 problem Sep 06 15:50:38 RP: Under 2.6.17 ? Sep 06 15:51:19 sirfred: I can't remeber which kernel versions I've used. I think 2.6.17 is ok though Sep 06 15:51:23 tomas_f: if you want to cross ref some versions, this is what I have installed on my hx: http://www.openembedded.org/~koen/installed-versions.txt Sep 06 15:51:33 thx Sep 06 15:52:23 RP: The only related line in the kernel messages is: "oprofile: using arm/xscale1" Sep 06 15:52:29 cu Sep 06 15:52:35 I suppose it's dumped by the module while loading Sep 06 15:52:38 * lrg is puzzled Sep 06 15:53:11 RP: can you send me another dump when you get a minute. Sep 06 15:53:58 RP: Anyway, lsmod tells that oprofile has one user, and no oprofiled daemon is running. I'm going to restart. Sep 06 15:56:52 lrg: sent Sep 06 15:57:00 RP: thanks Sep 06 15:57:07 sirfred: ok, strange :-/ Sep 06 15:59:33 RP: R7 is still 0xe (which is really strange), should be 0x2 Sep 06 15:59:56 RP: looks like something is happening with the params Sep 06 16:00:19 lrg: strange... Sep 06 16:00:38 * lrg investigates further.....has he screwed up the patch somehow...... Sep 06 16:02:13 lrg: The patch on the site is the one I applied and it did say ASoC 0.11.3 as I keep wondering if I did install it correctly... Sep 06 16:05:03 RP: No luck. Same results. :-( Sep 06 16:05:40 a+ Sep 06 16:05:44 bbl Sep 06 16:16:40 RP: it looks installed correctly, the new debug is visible Sep 06 16:28:16 later Sep 06 16:34:22 that's all for now. bb tomorrow. thanks again, Crofton Sep 06 16:35:22 Are you aware of any file name length limitation in jffs2 ? Sep 06 16:36:38 Or path depth? Sep 06 16:39:53 bbl Sep 06 16:55:22 re Sep 06 16:58:32 hrw: I managed to handle comments in the sanitize script :) Sep 06 16:59:31 hrw: the rule is 'bloc of comments above variables/routines are kept, and bloc of comments with an empty line below are skiped' Sep 06 16:59:58 hrw: comments inside routines/variables are kept as is Sep 06 17:00:45 s/skiped/skipped/ Sep 06 17:00:49 PR = "r1"\n#comment\n#comment\n#comment\ndo_install() is ok but }\n\n#comment\n#comment\n\nEOF is not? Sep 06 17:01:11 hrw: no Sep 06 17:01:19 hum... Sep 06 17:02:06 well, maybe can I let comments at the end of the file, if when }\n\n#comment\n#comment\nEOF Sep 06 17:02:18 http://pastebin.ca/162561 is ok for me Sep 06 17:02:22 s/if/i.e./ Sep 06 17:02:39 quite ok - }\n# -> }\n\n# Sep 06 17:03:06 hrw: it is not ok for me :-( Sep 06 17:03:28 hrw: because to which bloc is the first comment bloc attached ? Sep 06 17:03:40 hrw: it is related to the PR ? Sep 06 17:03:57 cyrilRomain: yep its wrong... PR = "r1"\n\n#comment is ok Sep 06 17:04:10 http://pastebin.ca/162563 Sep 06 17:04:45 hrw: the problem if that the sanitize script sort all the variables and routines Sep 06 17:05:01 I know ; Sep 06 17:05:02 ( Sep 06 17:05:23 hrw: so I don't know where to put such bloc execpt at the end or beginning of the file :( Sep 06 17:06:05 RP: can you bzip2 patches bigger then 50-100K? Sep 06 17:10:36 hrw: I could, it just tends to irritate me when I try to do anything with them Sep 06 17:10:56 RP: ok Sep 06 17:12:57 hrw: so I'll put such bloc of comments at the beginning of the sanitized file Sep 06 17:13:31 hrw: it is anyway better than skipping them Sep 06 17:13:40 yep Sep 06 17:13:59 cyrilRomain: can you put them \n separated? Sep 06 17:14:22 hrw: i.e. as they were in the original file ? Sep 06 17:14:27 cyrilRomain: can we put sanitize.py in the contrib/ dir of org.oe.dev? Sep 06 17:14:29 yes Sep 06 17:14:37 hrw: ok Sep 06 17:15:18 koen: ok, np (especially since this last version should not remove any file content as previous versions did) Sep 06 17:15:40 * cyrilRomain pastebin it Sep 06 17:16:09 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Sep 06 17:17:12 sanitize.py v0.3: http://rafb.net/paste/results/20b2iE11.html Sep 06 17:17:43 koen: I think it is time now to really test it Sep 06 17:20:31 from my tests, the rule that is mostly not respected is 'don't use tab (use spaces instead)' Sep 06 17:20:58 cyrilRomain: do we have that rule? Sep 06 17:21:01 the script expands the tabs to 8 characters Sep 06 17:21:11 zecke: yep, in the Format Guidelines Sep 06 17:21:16 zecke: mickeyl spoke his preference on the ml Sep 06 17:21:33 cyrilRomain: you get bitten by tabs. vs. spaces with _append and python code :) Sep 06 17:21:45 I need to subscribe to the lists... Sep 06 17:21:58 indeed :) Sep 06 17:22:27 someone feels like adding me to the mailman directly? Sep 06 17:23:21 03Cyril 07org.oe.dev * r5c42f9ce... 10/ (1 contrib/sanitize.py): sanitize.py: add a script that does the heavy lifting for making recipes conform to the styleguide in the wiki Sep 06 17:24:53 I need better desk.. and cleaning this Sep 06 17:27:05 zecke: done Sep 06 17:27:18 any feedbacks (especially bugs) welcome ! Sep 06 17:28:45 cyrilRomain|bbl: http://dominion.kabel.utwente.nl/koen/cms/index.php/v/Koen/Macro/Beestjes/IMG_4937_001.jpg.html?q=gallery <- bug Sep 06 17:31:42 koen: lol ;) Sep 06 17:32:38 koen: Have few mins? Would like to talk about machine addition/support policy. Sep 06 17:32:56 koen: is it a photo you take with one of the lens from your buy list ? Sep 06 17:33:09 sure, but I think that's best discusse done the ml Sep 06 17:33:09 03hrw 07org.oe.oz354x * r69f0a166... 10/ (3 files in 3 dirs): linux-openzaurus 2.6.16: backport of USB MaxPower limit information Sep 06 17:33:30 cyrilRomain|bbl: no, this with the lens I bought a few years ago :) Sep 06 17:33:41 koen: Just don't want to show my ignorance that much ;-) would like to get initial guidance on that. Sep 06 17:33:47 koen: nice photo anyway Sep 06 17:34:00 * cyrilRomain|bbl just use wheel to see the lens specs Sep 06 17:34:14 koen; and the best way to start is probably to ask why hx4700 is called ipaq-pxa270 in OE? Sep 06 17:34:44 * cyrilRomain|bbl tends to prefer real bugs than computer ones ... Sep 06 17:35:26 psokolovsky: it supports the hx2xxx as well Sep 06 17:35:37 psokolovsky: but mostly 'historical raisins' Sep 06 17:36:58 kown: ??? how it could. Those are pretty different devices. Not speaking about kernel, they have different screens, and that affects userland, no? Sep 06 17:37:12 hi, is it normal that mtn starts spittting chars to me: Sep 06 17:37:12 mtn: <<><<<<<<<<><<<><<<><<<<><><>><>>><>>>rc>rcrccrcrcrccrcrcrccrcrcrcc Sep 06 17:37:18 it didn't do this before .... Sep 06 17:37:22 psokolovsky: thats how familiar guys added them iirc Sep 06 17:37:42 eFfeM: messed up TERM var? Sep 06 17:38:07 eFfeM: I get that when I do "ssh foo.com 'mtn pull bla'" Sep 06 17:38:17 koen: hx4700 vs hx2xxx descrepancy is exactly why I ask. So, far, it seems that angstrom goes w/hx4700, and hx2xxx only in OZ. But they would obviously class if used in one distro. Sep 06 17:38:18 koen, doubt it, I also get mtn: ticks: >="bytes in"/1024, <="bytes out"/1024, c="certs in"/3, r="revs in"/1 Sep 06 17:38:45 eFfeM: yep, that's exactly what I get over ssh Sep 06 17:38:51 koen, hm, I did make update from the nslu2 master makefile, so maybe it is in there. Sep 06 17:38:57 this is natively not over ssh Sep 06 17:39:43 koen, brainwave: normally I do make update, this time I did make update 2>&1 | tee LOG Sep 06 17:39:54 that is probably the difference Sep 06 17:40:27 psokolovsky: the arrangement is far from perfect, I'd like to have 1 machine.conf per distinctive machine in OE, which can use includes for shared stuff Sep 06 17:40:36 psokolovsky: that's why I added the h5xxx.conf Sep 06 17:41:02 multimachine images are still a laudable goal, though Sep 06 17:41:05 koen: perfect. So, we can call ipaq-pxa270.conf in its current form a bug? Sep 06 17:41:15 psokolovsky: yes Sep 06 17:41:35 koen: Nice, I'll send messages discussing that to the list. Sep 06 17:42:29 we can make that a point to discuss at OEDEM if it becomes controversive Sep 06 17:42:51 someone know what logic is in power/state? Sep 06 17:42:53 koen: Then, what policy we have for adding new devices. You know, there's an unbound number of PocketePC and Palm devices, many of them already having ports on the level that it's time to support building userland for them. Sep 06 17:42:58 which values == suspend etc Sep 06 17:43:17 hrw: echo "mem" >... to suspend Sep 06 17:43:21 psokolovsky: I would prefer MACHINE for each model Sep 06 17:43:33 psokolovsky: I want to suspend one usb device not whole machine Sep 06 17:44:15 hrw: for FB, it's 4 for suspend, 0 wakeup, not sure about other device classes ;-( Sep 06 17:44:20 MACHINE=zire71, tungstent1, tungstent2, tungstent3, zire72 Sep 06 17:44:24 psokolovsky: most of those are probably rebadged HTC devices Sep 06 17:44:40 FB uses vesa powermodes Sep 06 17:44:42 koen, hrw: So, ok with having ~20 more machine.conf's? ;-) Sep 06 17:45:04 psokolovsky: if they can be grouped that I prefer grouped - like we have with c7x0 or spitz Sep 06 17:45:28 which are exactly the same machines exact for flash sizes Sep 06 17:45:52 psokolovsky: storage is cheap, and we have multimachine.bbclass now Sep 06 17:46:23 hrw: They can be groupped. But on different levels. For each distinct mchaine hardware, we'd have specific kernel, as it's too long a path for multimachine kernels. Sep 06 17:46:57 koen: hehe, which email did you use? Sep 06 17:47:01 hrw, koen: but regarding userspace, it's indeed waste and not scalable to support indvidual images. Sep 06 17:47:07 psokolovsky: misc kernel == misc machine for me Sep 06 17:47:57 hrw: what you mean by "misc"? Sep 06 17:49:16 psokolovsky: linux-openzaurus is ok for: c7x0 (5 models), collie, poodle, tosa, akita, spitz (3 models), hx2750 - so it generate 7 kernels for 7 OE-machines (13 hardware models) Sep 06 17:49:21 psokolovsky: I had a lot of stupid bugs in familiar because people insisted on multimachine images Sep 06 17:50:22 psokolovsky: if linux-palm will be for 10 hardware models but 6 kernels would be enough for whole set then 6 OE-machines need to be created (if they do not have to be different in other packages) Sep 06 17:50:44 koen: but it was in Familiar ;-) I guess it worth second try based on previous experience. Anyway, it's "extra" aim, after there will be indeed many machines to support. Sep 06 17:51:21 psokolovsky: I think it's easier to 'merge' machines later on that to split them Sep 06 17:51:41 hrw: Sure, I understand that. But so far, situation is not that bright: even for close machines, there may be wild varrying kernels ;-( Sep 06 17:52:38 I'd like to have them seperate in OE, but keep all info in e.g. the xserver-common package Sep 06 17:52:53 koen, hrw: But again, for starters, you're ok with adding separate conf for each machine with own hardware set, right? Sep 06 17:52:58 yes Sep 06 17:53:00 yes Sep 06 17:53:15 hrw, kown: Wonderful, thanks. Sep 06 17:53:22 thats easiest way Sep 06 17:54:35 So, I'm planning to do some proactive measures on all tiny hh.org and hackndev.com ports, to bring there attention that they could, and should, submit patches upstream. Now I know that there's backingf or that ;-) Sep 06 17:54:50 for that Sep 06 17:55:08 psokolovsky: those could need some poking indeed Sep 06 17:55:17 yeah Sep 06 17:55:24 I tried to do that for the hackndev people Sep 06 17:55:41 but they want to re-invent the wheel for everything Sep 06 17:56:14 Alex' eagerness to learn by doing can be a bit annoying Sep 06 17:56:51 and "OMG g3nt00 on pd4"" people should be cluebatted Sep 06 17:57:14 Yes Sep 06 17:57:25 koen: not all as far as I saw from their forums. Anyway, there're just to many closed realms and great deal of feudalism there. people just lack information ;-) Sep 06 17:57:32 get it working for, OMGLOL it later ;) Sep 06 17:57:48 ~lart lack of docs on linux internals Sep 06 17:57:49 * ibot steals lack of docs on linux internals's mojo Sep 06 17:57:52 s/for/first/ Sep 06 17:58:25 linux/Documentation/power/ is only about swsusp... Sep 06 17:58:57 psokolovsky: making a machine+kernel doesn't take too long, so you could lead by example Sep 06 17:59:41 psokolovsky: write use case - some text about how you added h4000 into OE and why Sep 06 17:59:46 we lack such texts Sep 06 17:59:54 koen: yeah, I'm going too ;-). it would be too wasteful to went all the path I've done only to support h4000 ;-) Sep 06 18:00:01 ok ;-) Sep 06 18:00:30 org.oe.documentation is open for everyone Sep 06 18:08:23 Alien3 Sep 06 18:11:25 psokolovsky: The problem with hx2xxx vs. hx4700 is I added the hx2xxx support to the oz kernel tree and they have different kernels. Different kernels really needs different machine.conf files. I see no reason why a hx2xxx image shouldn't run on a hx4700 though Sep 06 18:12:01 RP: git bisect looks nice ;-) Sep 06 18:12:12 JustinP: It can be handy :) Sep 06 18:13:11 RP: unfortunately I don't need to use it :-o Sep 06 18:13:22 RP: Glad you agree that they should have separate machine.conf's. As for image, you mean userland, right? Sure, it will work. Maybe not exactly perfect, but sure will. Sep 06 18:14:00 RP: moving away /lib/modules/2.6.17/kernel/sound and rebooting causes my SD to work Sep 06 18:14:17 RP: so it's something to do with loading alsa/snd that borks my SD Sep 06 18:14:22 JustinP: eek :-/ Sep 06 18:14:42 JustinP: Can you tell me which asoc patches that kernel had applied please? Sep 06 18:14:51 RP: surely Sep 06 18:15:44 RP: linux-openzaurus-2.6.17-r22 Sep 06 18:15:58 RP: I see in patches/: asoc-v0.11pre12.patch asoc_makefile-r0.patch asoc_platform_dev_fix-r0.patch asoc_poodle_morehack-r0.patch Sep 06 18:16:36 JustinP: thanks. dmesg just said something about an invalid ioctl? Sep 06 18:16:55 That version is worryingly recent :-/ Sep 06 18:17:26 I can confirm 0.11.3 doesn't break SD although it doesn't give sound... Sep 06 18:18:22 RP: ALSA sound/core/oss/pcm_oss.c:574: snd_pcm_oss_prepare: SNDRV_PCM_IOCLT_PREPARE failed Sep 06 18:18:26 RP: x2 Sep 06 18:20:39 JustinP: Thanks. That sounds like memory corruption which concerns me a bit. I'll run it past Liam as something odd it going on in ASoC :-/ Sep 06 18:21:49 JustinP: The latest asoc is asoc-v0.11.3.patch in the usual place - you could try that and see if it upsets SD for you... Sep 06 18:26:20 RP: ok, I'll try it Sep 06 18:30:46 RP: asoc_platform_dev_fix-r0.patch fails 1 out of 1 hunk FAILED -- rejects in file sound/soc/soc-core.c Sep 06 18:31:14 JustinP: Just drop that patch Sep 06 18:31:54 * RP -> back 30mins Sep 06 18:31:59 RP: ok, now asoc_poodle_morehack-r0.patch fails. I suppose I can drop that as it's poodle? Sep 06 18:32:39 * JustinP drops it and patching finishes Sep 06 18:32:58 * mrz80 is back (gone 01:16:50) Sep 06 18:37:05 hi Sep 06 18:37:24 hey woglinde Sep 06 18:37:38 i want more time Sep 06 18:43:21 * chouimat is away: shower Sep 06 18:43:54 CoreDump|afk: [1] Normal Boot, would be nice if it would honor alternate kernel choice Sep 06 18:44:38 CoreDump|afk: also, possibly, might be nice if there was an option for "Kexec chosen kernel now". A menu option after that one perhaps.... (i.e., choose a kernel, then choose the "kexec now" option) Sep 06 18:56:47 * chouimat is away: Gone away for now. Sep 06 18:56:54 * chouimat is back. Sep 06 19:10:34 03eFfeM 07org.oe.dev * r337cfe2d... 10/ (4 files in 3 dirs): puvrusb2-mci: added 20090903 snapshot; made bb file style guide compliant Sep 06 19:30:16 CoreDump|afk: ah yes, another one. "init X menu" or something similar. To have altboot not use the saved option (go to menu) on reboot. Sep 06 19:31:56 heh Sep 06 19:32:10 I need to get a new vi plugin that recognizes require Sep 06 19:37:39 when you use udev, you you need to mount the usbdevfs? Sep 06 19:37:57 RP: newest asoc, no sound (as advertised), no mmc errors on boot, when I try to mount I get the mmlblk0 errors in dmesg, same as before Sep 06 19:38:11 RP: so it's still happening...I'll try moving away the sound modules again Sep 06 19:40:08 re Sep 06 19:42:06 http://www.openembedded.org/contact needs to be updated with the new mailing list. Sorry to ask for someone to do that (I cannot edit this page). Infos here: http://www.openembedded.org/openembedded-mailing-lists-move Sep 06 19:45:44 cyrilRomain: fixed Sep 06 19:46:32 RP: moving away sound/ makes SD work again Sep 06 19:46:57 RP: so something in the sound module is causing my SD to fail....if doing DEBUG_MMC and serial would help let me know Sep 06 19:49:00 * chouimat is away: Pub :) Sep 06 19:50:30 koen: thanks :) Sep 06 19:58:57 that was tons of fun....building xserver-xorg took all of my RAM and my machine almost swapped itself to death Sep 06 19:59:40 sounds like my gentoo machine 5 years ago Sep 06 20:00:07 when I woke up everyday someone either bumped glibc or X Sep 06 20:00:15 and that hurts on a 110MHz sparc5 Sep 06 20:00:33 X -> 18 hours, glibc -> 15 hours Sep 06 20:01:17 ouch Sep 06 20:01:38 gentoo users are .... .... for me Sep 06 20:01:44 koen: well doing that on my Shark has a locked up system on compiling glibc Sep 06 20:02:22 zecke: my sparc has 256MB ram Sep 06 20:02:37 and it *had* a fast 9GB scsi drive Sep 06 20:02:50 anyone knows a good C++ lib (smaller than stl) providing file access and a hashmap and red-black-tree (map) Sep 06 20:02:51 I do not see sense of using source driven distro which you update each day Sep 06 20:03:11 zecke: maybe see if uclibc++ provides that? Sep 06 20:03:12 * kergoth shrugs Sep 06 20:03:21 hrw|tv: debian built without hw multiply Sep 06 20:03:32 hrw|tv: you can update whenever you want :-P Sep 06 20:03:32 hrw|tv: I ended up building my own .deb for openssl Sep 06 20:03:37 so, oesources.org expires on friday Sep 06 20:03:41 anyone want it? :p Sep 06 20:03:49 hmmm Sep 06 20:03:50 ggilbert_: domain? Sep 06 20:03:55 koen: ubuntu has OpenSSL update Sep 06 20:03:56 yes Sep 06 20:04:02 koen: has LTG another wiki? Sep 06 20:04:13 zecke: how do you mean? Sep 06 20:04:48 koen: from an article at heise.de the alioth break-in sounded like a GForge exploit (Pmwiki) Sep 06 20:04:53 JustinP: c00l gent00 d00ds has 'while true;do emerge update;emerge buildall;done' in /etc/rcS.d/S01gent00 Sep 06 20:05:06 zecke: iirc the gfroge wiki is disabled Sep 06 20:05:51 hmmm Sep 06 20:05:54 hrw|tv: that was a while ago, you don't update more than once a day? Sep 06 20:06:03 my sparc still has a patch for sane I need to send upstream Sep 06 20:06:34 zecke: debian archive is updated daily Sep 06 20:06:46 &$*@&$*@ pdiffs Sep 06 20:07:17 koen: I love them too Sep 06 20:07:23 back to alien3 Sep 06 20:07:51 * koen wonders why one would call a show 'numb3rs' Sep 06 20:08:20 Sounds like a no :p Sep 06 20:08:48 ggilbert_: I think sources.oe.org would be a bit clearer Sep 06 20:09:00 Ok. Letting it expire Sep 06 20:09:01 but that's only my opinion Sep 06 20:10:35 ggilbert_: send a note to mickeyl & kergoth Sep 06 20:11:20 they are trying to attract a 7eet audience ... Sep 06 20:15:37 03mickeyl 07org.oe.dev * r2af79ceb... 10/ (1 packages/xorg-proto/calibrateproto_git.bb): calibrateproto: s/DATE/SRCDATE/ Sep 06 20:18:24 hrw|tv: there are l4m3rZ in every project/distro. Gentoo is still a good distro. Sep 06 20:19:48 My Zaurus SL-c3200 has come to sweden now :D expecting it deliverd either tommorow or friday Sep 06 20:23:45 JustinP: I know Sep 06 20:25:57 is it bad if base-files contains dangling symlinks? Sep 06 20:26:38 Crofton: dangling symlike are pretty relative Sep 06 20:27:04 Crofton: they could point to /bin/foo, which wouldn't exist on your host, but would exist on your rootfs Sep 06 20:27:11 ok Sep 06 20:27:48 that will go away once we switch to chroot based building Sep 06 20:29:31 good :) it is a confusing messgae Sep 06 20:29:45 cu all Sep 06 20:30:01 ah.. one more thing.. Sep 06 20:31:30 to push Sep 06 20:32:20 CIA-9: ping Sep 06 20:32:21 cu Sep 06 20:33:44 03hrw 07org.oe.oz354x * r6be20709... 10/ (1 conf/distro/openzaurus.conf): openzaurus.conf: oesources.org are history - domain expired Sep 06 20:36:36 ggilbert_: thanks for having created oesources.org in the first place Sep 06 20:36:47 JustinP: Thanks for the debugging - I'll mention it to Liam Sep 06 20:37:26 * CosmicPenguin salutes oesources.org Sep 06 20:37:37 * Kristoffer hails oesources.org Sep 06 20:38:49 ~priase ggilbert Sep 06 20:39:40 err, is oesources.org gone now? Sep 06 20:40:24 Kristoffer: it's still there for another two days Sep 06 20:40:32 unless I get the urge to renew the domain again :) Sep 06 20:40:51 but seeing as it's been mostly abandoned for a year now, I've not been feeling the urge to do so Sep 06 20:40:58 ggilbert: I assume the content will then be offline? Sep 06 20:41:09 Yeah Sep 06 20:41:28 If you guys want me to keep it up in the current state, I don't have a big problem renewing it for another year Sep 06 20:41:39 I just noticed the renewal notice today :) Sep 06 20:42:07 ggilbert: The content is mainly of benefit to OZ iirc? Sep 06 20:42:30 It wasn't OZ specific. It's just not been updated in ages Sep 06 20:42:43 We need some decent scripts to run a mirror server... Sep 06 20:42:45 I think it had familiar stuff in it Sep 06 20:43:15 familiar kept source archives I think so they should be covered... Sep 06 20:43:18 If someone wrote the scripts, I'd even be glad to set them up. I just don't have the drive to do the coding for something that I haven't used in so long Sep 06 20:43:45 * RP adds mirror scripts to the oedem agenda Sep 06 20:46:11 ggilbert: oesources.org was already on the agenda :) Sep 06 20:46:44 hehe good :p Sep 06 20:46:57 It would be something nice for people to have access to Sep 06 20:47:23 It would be really useful to have Sep 06 20:47:28 yeah Sep 06 20:47:37 ggilbert_: it would be nice if we could point sources.oe.org to your hd :) Sep 06 20:47:43 So what is oedm? Sep 06 20:47:48 koen, go for it Sep 06 20:47:51 ggilbert_: and zecke wrote bittest with oesources in mind Sep 06 20:47:53 ggilbert_: a small meeting in berlin Sep 06 20:48:03 My problem has always been a lack of a machine with the bandwidth. If I had access to one, I'd proably have set something up by now... Sep 06 20:48:03 koen: did I? Sep 06 20:48:06 I don't mind continuing to supply space and bandwidh Sep 06 20:48:08 zecke hm I think it isnt small now Sep 06 20:48:10 http://www.openembedded.org/wiki/OEDEM Sep 06 20:48:29 and I am nervous Sep 06 20:48:40 woglinde: it is small enough I could cover for the dope Sep 06 20:48:51 woglinde: why? not trusting our room allocation strategy? Sep 06 20:49:00 zecke yes Sep 06 20:49:08 I already ordered a batch of leffe that I can share with mickeyl :) Sep 06 20:49:11 woglinde: create a backup plan Sep 06 20:49:27 I am not good at backup Sep 06 20:49:34 woglinde: mail our dean Sep 06 20:49:38 I can only mange the first Sep 06 20:49:41 plan Sep 06 20:49:51 but that with sucess Sep 06 20:50:45 zecke hm maybee we could spent one evening at mexico city Sep 06 20:50:51 RP: thanks. If I can test anything else let me know. Sep 06 20:51:24 JustinP: Will do, thanks Sep 06 20:51:46 woglinde: I really hope steve rents this mitte office Sep 06 20:51:55 woglinde, zecke: Is the size of the meeting getting too big? :-/ Sep 06 20:52:15 woglinde: when step in front of the door and we have enough bars to get lost Sep 06 20:52:33 RP: no, woglinde is not trusting my 'strategy' for the room allocation Sep 06 20:52:56 woglinde: another backup is using the meeting room at ROAD on the weekend, it has no i-net connection but plenty of space Sep 06 20:53:24 I'm not worried that we will find a nice and dry place Sep 06 20:54:26 hmm i guess i could get a place with network connectivity for the weekend Sep 06 20:54:45 florian: don't worry :) Sep 06 20:54:45 You have at least two backup plans that I know of which is much more organised than some things I've seen :) Sep 06 20:54:57 heh Sep 06 20:55:05 http://scan.coverity.com/ <- apache-httpd had 6 defects in seven lines Sep 06 20:55:14 hehe Sep 06 20:55:47 RP: I think we all do not have high comfort criterias Sep 06 20:56:58 zecke: I think you'd be right :) Sep 06 20:57:54 okay back to coding Sep 06 21:01:44 I'm going to read a book now Sep 06 21:01:46 later Sep 06 21:02:57 I think lack of connectivity will actually increase productivity Sep 06 21:03:15 and we can always sync our monotone db's via an ad-hoc network ;) Sep 06 21:03:27 sure :) Sep 06 21:03:36 but mtn port is firewalled anyway :) Sep 06 21:04:36 hmm... us that udp? Sep 06 21:04:43 s/us/is Sep 06 21:06:03 zecke: mtn can sync over ssh Sep 06 21:06:05 any bored udev gurus around? Sep 06 21:07:09 koen: what makes you think ssh is not firewalled? Sep 06 21:07:16 later :) Sep 06 21:08:09 * RP could rarely be described as bored :-/ Sep 06 21:08:32 what a well known situation :-) Sep 06 21:09:20 one can be bored an busy at the same time Sep 06 21:09:28 right Sep 06 21:09:45 I am trying put together a fstab file for the omap5912osk Sep 06 21:10:22 after the first boot I get the .udev/ already exists on the static /dev Sep 06 21:10:24 messgae Sep 06 21:11:46 I also see that message Sep 06 21:12:52 doesn't it bother you? Sep 06 21:12:56 :) Sep 06 21:14:15 RP: could be my imagination, but Xw100 doesnt seem to cause static :-) Sep 06 21:15:01 XorA|gone: Its just a bug thats difficult to reproduce Sep 06 21:15:16 RP: nah, with xmms it became very easy Sep 06 21:15:32 XorA|gone: pssst gnome 2.16 is out Sep 06 21:15:38 koen: bastards Sep 06 21:15:56 crap Sep 06 21:16:03 I forgot to look at the .pc thing Sep 06 21:20:10 RP: actually its gone totally even with Xfbdev, now did he fix something in 0.11 final :-) Sep 06 21:32:45 anyone built opie-eye lately? :D Sep 06 21:32:45 hmmm Sep 06 21:33:07 I wonder if the 'Workaround for r7 corruption in Thumb syscalls' thing could fix all the segfaults in EABI :) Sep 06 21:39:08 Err, bitbake 1.7.0 recommended or what? Sep 06 21:40:30 I have 1.4.5... Sep 06 21:40:35 Kristoffer: what. or what ? Sep 06 21:40:47 http://www.openembedded.org/wiki/GettingStarted Sep 06 21:41:04 the recommended version is bitbake-1.6 Sep 06 21:41:07 I didnt mean it like that :D Sep 06 21:41:22 :0 Sep 06 21:41:24 hmm, I'm behind Sep 06 21:45:12 XorA|gone: I worry about justinp's problem :-/ Sep 06 21:46:03 03tmbinc 07org.oe.dreambox * r7ff88b28... 10/ (4 files in 4 dirs): base-files/dm*/fstab: add /tmp tmpfs Sep 06 21:46:05 RP, kernel issues? Sep 06 22:08:51 can I update http://www.openembedded.org/usefultargets with all meta and tasks ? Sep 06 22:10:17 cyrilRomain: are *all* meta tasks really useful? ;-) Sep 06 22:10:34 psokolovsky: maybe not Sep 06 22:11:23 cyrilRomain: did you add any SECTION value validation to your script? Sep 06 22:11:27 ha, should have read more carefully the title 'usefulTargets' Sep 06 22:11:49 * cyrilRomain checks Sep 06 22:12:19 psokolovsky: yes it is Sep 06 22:12:33 psokolovsky: is it discarded ? Sep 06 22:13:20 psokolovsky: but the script does not validate the content of the file Sep 06 22:13:21 cyrilRomain: wow, cool. then, what would be nice to add to wiki is RFC list of sections, from recent maillist discussion or your script. Sep 06 22:13:46 cyrilRomain: nope, actually I haven't tried it yet. Sep 06 22:14:33 psokolovsky: if I have enough time, I'll try to integrate the script with bittest Sep 06 22:15:32 cyrilRomain: Nice. it's good thing even on its own. we all should get used to it. Sep 06 22:15:49 psokolovsky: sanitize.py checks the Format Guidelines, bittest checks variables value content Sep 06 22:16:33 cyrilRomain: cool, there's always something new to learn ;-) Sep 06 22:16:34 psokolovsky: but it is far from a perfect tool because there are weird cases Sep 06 22:16:50 like bloc of comments in the middle of the .bb file Sep 06 22:16:55 :-/ Sep 06 22:17:17 cyrilRomain: I guess, so far it's early to use it as a pre-commit hook ;-) Sep 06 22:17:37 psokolovsky: yes ;) Sep 06 22:18:56 psokolovsky: the commited version does not have many comments, especially required python module :-/ Sep 06 22:19:31 psokolovsky: if you want to test and use it Sep 06 22:19:33 cyrilRomain: btw, is dependency a single py file or the whole package? Sep 06 22:19:47 a single .py file Sep 06 22:20:11 psokolovsky: but it is included in package on Debian and Gentoo Sep 06 22:20:20 * cyrilRomain pastebin a tip Sep 06 22:20:45 cyrilRomain: nice Sep 06 22:21:05 psokolovsky: http://rafb.net/paste/results/8m4iQV85.html Sep 06 22:21:49 cyrilRomain: wow, py2.4 requirement is rough ;-) Sep 06 22:22:00 psokolovsky: well, I'm not sure about that Sep 06 22:22:26 cyrilRomain: Ok, I'll test it on 2.3 Sep 06 22:22:26 psokolovsky: but I used functions that seems new in python documentation Sep 06 22:22:38 psokolovsky: ok, I think it should work Sep 06 22:22:59 psokolovsky: I'm not a python expert so I guess I only used default functions ;) Sep 06 22:23:22 ok Sep 06 22:23:54 cyrilRomain: Anyway, thanks for selecting Python and not other language! ;-) Sep 06 22:25:49 * cyrilRomain checks for the python documentation Sep 06 22:27:20 hum.. I avoided the 'Deprecated string functions' but the documentation does not tell to which version it is deprecated Sep 06 22:27:44 psokolovsky: anyway if you have any problem with the script, feel free to ask me Sep 06 22:28:00 ok Sep 06 22:30:00 psokolovsky: what did you mean by 'RFC list of sections' ? Sep 06 22:30:53 the script sort from the list of known variables names and routines (see http://www.openembedded.org/wiki/StyleGuide) Sep 06 22:31:22 cyrilRomain: Well, I just got an idea you want to add something to wiki which would help bring more order into OE metadata. There was recently a discussion for SECTION var value should be used. I can dig a link for you. Sep 06 22:32:31 cyrilRomain: http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-August/000073.html Sep 06 22:32:38 psokolovsky: ah ok. As I said bittest is more responsible to check the values. Sep 06 22:32:41 psokolovsky: thx Sep 06 22:32:46 ok Sep 06 22:33:46 any idea when chroot build env is planned? Sep 06 22:48:21 anyone here know much about procmail or fetchmail? Sep 06 23:30:18 any bitbake devs here ? Sep 06 23:33:40 as soon as I modify a file from the bitbake branch, bittest says 'Use PYTHONPATH to point to bitbake/lib' Sep 06 23:54:24 cyrilRomain: I've coded a bit of bitbake (have a patch pending) but I don't know what bittest is.... Sep 06 23:59:08 JustinP: well, I think it was tabs vs spaces somewhere Sep 06 23:59:27 JustinP: it works now after rewriting reading from my patch Sep 06 23:59:54 JustinP: bittest is a test tool for bitbake Sep 07 00:00:11 JustinP: bittest can check for patch_checker', 'source_checker', 'doc_checker', 'content_checker', 'depends_checker Sep 07 00:00:56 JustinP: e.g. source_checker checks for SRC_URI availability, content_checker for .bb file content, etc. Sep 07 00:03:57 JustinP: I'm trying to make source_checker not downloading packages (just checking file is there) to speed up checking for SRC_URI Sep 07 00:10:28 cyrilRomain: ah yes ;-) Sep 07 00:10:50 cyrilRomain: so you're editing python? yeah, consistent whitespace is *very* important for it Sep 07 00:11:29 JustinP: I know one have to be very carefull about that Sep 07 00:11:51 JustinP: I think it is my vim settings that are different from my host system and my chrooted environment Sep 07 00:12:08 JustinP: and this messed things up without I could see it Sep 07 00:12:35 well, first sucess on tarball :) Sep 07 00:20:23 sucess on local files now Sep 07 00:22:02 cyrilRomain: sanitize.py has dos lineendings. bother to fix? Sep 07 00:23:18 psokolovsky: i'm not sure how to fix this the proper way Sep 07 00:23:33 * cyrilRomain looks into python doc Sep 07 00:23:38 fromdos it and commit ;-) Sep 07 00:23:52 cyrilRomain: I mean, fiel you committed has Sep 07 00:23:59 file even Sep 07 00:24:19 psokolovsky: ah the script itself :) Sep 07 00:24:41 psokolovsky: no problem at all, I'll fix it in the next version :) Sep 07 00:25:07 * cyrilRomain suspect to have a really bad .vimrc Sep 07 00:25:39 psokolovsky: maybe it is because I 'send' it to koen via pastebin (?) Sep 07 00:25:45 cyrilRomain: so this sooner, if possible ;-) Sep 07 00:25:57 cyrilRomain: easily Sep 07 00:26:03 psokolovsky: I don't have push access Sep 07 00:26:21 psokolovsky: so do it yourself if it urge for you Sep 07 00:26:25 cyrilRomain: Ah, sorry. I'll fix it then. Sep 07 00:26:31 psokolovsky: thanks Sep 07 00:32:56 03pfalcon 07org.oe.dev * rff29c494... 10/ (1 contrib/sanitize.py): sanitize.py: Fix line-endings (dos->unix). Sep 07 00:39:33 03pfalcon 07org.oe.dev * ra83e90d6... 10/ (1 contrib/sanitize.py mtn:execute true): sanitize.py: Make executable. Sep 07 00:41:25 8-O 2:55 am ! Sep 07 00:41:35 'night all Sep 07 01:43:56 hi everybody, what's the difference between "RDEPENDS" and "RRECOMMENDS"? Sep 07 02:04:45 03pfalcon 07org.oe.dev * rca5e13e4... 10/ (1 contrib/sanitize.py): sanitize.py: Drop unimportant Python 2.4 dependency. Sep 07 02:09:19 O_Neil: RDEPENDS is a must, ipkg won't let you install package unless all RDEPENDS are satisfied. RRECOMMENDS just to note a relation to another package. There's even weaker form RSUGGESTS. Sep 07 02:11:35 recommends get installed automatically, but can be removed. suggests are recommended to the user, b ut not automatically installed Sep 07 02:13:56 thanks **** ENDING LOGGING AT Thu Sep 07 02:59:56 2006