**** BEGIN LOGGING AT Tue Sep 06 02:59:57 2005 Sep 06 03:38:52 man Sep 06 03:38:57 reading seems to be a lost art Sep 06 03:40:38 ... Sep 06 03:40:55 koen|slug: what kind of device you are running oe on? Sep 06 03:53:24 AvengerMoJo: dual opteron Sep 06 03:53:55 koen|slug: why read when you can make arse of yourself Sep 06 03:57:59 "six young germans have beaten the record for showering last weekend" Sep 06 03:58:22 now we know why mickeyl was showering that long :) Sep 06 03:58:26 koen|slug: one of them could be mickeyl but he is not so young.. Sep 06 03:58:40 koen|slug: perhaps he trained them Sep 06 03:59:14 I have to fix fuses at home.. my router goes mad since yesterday so I cannot connect to home machine ;( Sep 06 03:59:53 hrw|work: /home on my server got click of death yesterday :-) Sep 06 04:10:50 ~lart nail for commit logs Sep 06 04:10:51 * ibot_ does a little 'dpkg -P nail' action for commit logs Sep 06 04:53:48 When the kernel is compiled, should defconfig-c7x0 be copied over .config in the kernel source dir? Sep 06 04:55:53 it is done in do_configure_prepend iirc Sep 06 04:56:46 okay, thanks Sep 06 04:59:43 I've just sent preliminary spitz+borizoi patches to LKML/akpm :) Sep 06 05:00:00 cool Sep 06 05:00:16 rmk was only cc'd which I expect I'll pay for but there was no other way to do it... Sep 06 05:01:07 I'm probably going to push the battery/charging code soon so if anyone has any bugs, now would be a good time to let me know about them Sep 06 05:01:24 I also need to finish removing sharpisms as a matter of urgency now :) Sep 06 05:11:42 hmm.. opie say that I have 0% battery Sep 06 05:12:41 battery_percentage 0 battery_voltage 0 Sep 06 05:12:53 I have to flash 2.6.13-mm1 Sep 06 05:15:43 hrw|work: might well be the same thing I was seeing Sep 06 05:16:11 hrw|work: in fact I imagine it must be the same Sep 06 05:17:32 .... Sep 06 05:18:16 hrw|work: what kind of system are you testing your 2.6.13 on Sep 06 05:18:23 AvengerMoJo: c760 Sep 06 05:18:37 AvengerMoJo: and ofcourse I use 2.6.13 on my desktops Sep 06 05:22:01 oh Sep 06 05:24:40 hrw|work: If you want to use it in the meantime just plug in the ac power for a moment Sep 06 05:27:25 hrw|work: Yes, its a known issue - upgrading to 2.6.13-mm1 should fix it Sep 06 05:27:59 what's the status on the udev troubles? Sep 06 05:32:13 ok Sep 06 05:32:21 lardman|work: I will flash 2.6.13 soon Sep 06 05:33:21 I've just flashed 2.6.13-mm1 with cpufreq enabled (at least I thought it should have been), but no /sys/devices/system/cpu/cpu0/cpufreq ... Sep 06 05:34:16 * koen|uni imagines lardman|work adding a lot of printk's Sep 06 05:34:57 I'm hopefully I just forgot to patch something - the cpufreq dir in the source is full of .o files, so I got something right at least Sep 06 05:43:31 koen|uni: udev troubles? Sep 06 05:43:42 lunch time.. Sep 06 05:43:53 Someone needs to spend some time and write some nice udev scripts... Sep 06 05:54:24 bbl Sep 06 06:09:37 re Sep 06 06:35:30 RP: 2.6.13-mm1 does not build for collie Sep 06 06:54:20 hrw|work: That doesn't surprise me given the last JL patches were against 2.6.11/// Sep 06 06:54:40 yes Sep 06 06:54:53 hrw|work: If there was any interest in collie 2.6, I'd try and update it so it built... Sep 06 06:55:09 I also suspect tosa is broken at the moment... Sep 06 06:55:38 any not your in other words Sep 06 06:56:58 hrw|work: c7x0 and cxx00 have me slightly overloaded at the moment ;-) Sep 06 06:58:58 ~lart kbabel for crashing on opie translations Sep 06 06:58:58 * ibot_ explains, ever so gently, that if kbabel doesn't give the channel more information, they can't help for crashing on opie translations Sep 06 06:59:39 hrw|work: I wish we knew what had happened to JL. I suspect he's switched Unis and we don't have a valid address for him now Sep 06 06:59:42 morning kergoth Sep 06 06:59:59 probably Sep 06 07:00:28 hey Sep 06 07:00:32 hi kergoth Sep 06 07:00:48 hrw|work: Pavel Machek has been doing a good job of getting bits of collie into mainline Sep 06 07:03:03 I noticed it Sep 06 07:03:47 its funny that upgrading kernel module require usage of -force-downgrade option Sep 06 07:03:50 hi kergoth Sep 06 07:06:46 hrw|work: 2.6.13-rc5-mm1 > 2.6.13-mm1 in ipkg's eyes no doubt... Sep 06 07:08:21 indeed, since "r" is (famously) later in the alphabet than "m". Sep 06 07:08:59 Good evening. Sep 06 07:10:07 RP: rc6-mm2 -> mm1 here ;) Sep 06 07:13:04 hi sirfred Sep 06 07:13:10 sirfred: hello Sep 06 07:13:11 hrw|work: ;-) Sep 06 07:13:17 RP, hrw|work : hello Sep 06 07:13:33 RP: I've read that the corgi w100fb is going to be merged in the 2.6 kernel. Congrats. Sep 06 07:13:58 sirfred: and SD support and corgi keyboard/touchscreen stuff etc. :) Sep 06 07:14:17 sirfred: and maybe even spitz support if I can get away with it Sep 06 07:14:40 RP: Great. Sep 06 07:16:09 someone know where I can find info what for are other choices in sharpmenu (which is used to flash)? Sep 06 07:16:21 hrw|work: concics.net Sep 06 07:16:30 conics.net even Sep 06 07:16:47 FSCK... rootfs got corrupted ;( forgot about collie initrd on card ;(((((((( Sep 06 07:17:23 buuu Sep 06 07:21:46 I broke flash process so now kernel starts with looots of jffs2 errors.. Sep 06 07:22:18 so today I will test my brand new opie-image variant on c7x0 Sep 06 07:24:00 RP: do you own the SD stuff now, then? Sep 06 07:25:40 CosmicPenguin: own it? no, just been helping it along where I can... Sep 06 07:26:44 Pierre Ossman took Ian Molton's patches and pushed them before I got around to it so I've just added support to them. I also added a few pxa specific bits... Sep 06 07:27:00 and corgi specific bits, obviously :) Sep 06 07:27:08 someone needs to write 4 bit support for the pxa270 now Sep 06 07:30:05 Ok, just checking Sep 06 07:30:11 I have a few minor changes that might be useful too Sep 06 07:30:20 Did you guys already fix the problem with 1.10 cards? Sep 06 07:31:03 I'm not sure I even know of the problem... Sep 06 07:31:26 Better question - who should I send patches to? Sep 06 07:31:37 Patches against what? Sep 06 07:31:43 and do they touch SD? Sep 06 07:31:48 latest hh.o, I guess Sep 06 07:31:55 * CosmicPenguin pulls to make sure he's up to date Sep 06 07:32:32 hh.org patches go to the kernel-discuss. As SD is going into mainline, someone needs to keep mainline upto date now though Sep 06 07:33:20 If they're raw MMC related, they go to rmk, SD patches to akpm, probably via Pierre. Or I can deal wih them... Sep 06 07:34:15 heh - yea, the problem was fixed Sep 06 07:34:30 RP: how different are MMC and SD when it comes to kernel? (other than write lock switch in SD) - does they got any differ at speed? and does mmc+ are supported? Sep 06 07:34:51 CosmicPenguin: Can you look at 2.6.13-mm1 and check its also been allied there please? Sep 06 07:35:25 hrw|work: No difference in speed really. Just some different commands and different card init procedure Sep 06 07:35:33 I don't know much about mmc+ Sep 06 07:35:35 thx RP Sep 06 07:35:58 RP: I'll try - I'll bet it has Sep 06 07:58:12 RP: (since you were talking about SD ;) has anything been done towards SDIO support yet? Sep 06 07:59:44 gridzero: give someone SDIO card with documentation and you will see ;) Sep 06 08:00:05 Bah! Documentation is for wimps! :) Sep 06 08:02:16 gridzero: theres only one portion of the SDIO spec that is not freely available Sep 06 08:02:20 of course, its the most important part Sep 06 08:02:55 But its probably easily reverse engineerable by those not living in countries that discourage such activities Sep 06 08:04:11 Norway? :) Sep 06 08:04:18 heh Sep 06 08:04:18 sure Sep 06 08:04:20 why not Sep 06 08:20:12 seen this? http://www.geekzone.co.nz/content.asp?contentid=5142 Sep 06 08:21:12 hmm, is pppoe in oe? Sep 06 08:21:43 looks like bruno added ppp-dsl, which is convenience scripts to glue pppoe to pppd Sep 06 08:21:47 but i dont see pppoe itself Sep 06 08:23:23 gridzero: I don't know of anyone working on SDIO support. There's a lot of people would like to see it though. I think everyone has been letting SD test the water :) Sep 06 08:24:06 cu all Sep 06 08:24:09 time to go home Sep 06 08:24:10 Some bad news. MTD people are removing the flags that let the large partitions on husky and borzoi work in a way compatibile with 2.4 :-( Sep 06 08:24:34 Or even worse, in a way compatible with the bootloader perhaps... Sep 06 08:29:21 hmmm Sep 06 08:29:23 http://fjortis.info/pub/debian/ifupdown-0.6.7+iproute2-untested-experimental.patch Sep 06 08:29:29 makes ifupdown use iproute2 instead of ifconfig Sep 06 08:29:31 * kergoth likes Sep 06 08:34:01 hmm, busybox iproute2 doesnt support policy routing. cant say i'm surprised, but good to know Sep 06 08:38:45 damnit, why isnt hostfs working in my uml.. Sep 06 08:38:46 hmm Sep 06 08:59:17 03yuecelm 07org.oe.dev * r111dbb37... 10/packages/avetanabt/ (files/avetanalocaldevice.patch avetanabt_svn.bb): avetanabt_cvs: add version.xml and patch\n - add version.xml to have version information about stack at runtime\n - add patch to set property bluetooth.connected.devices.max to 7 Sep 06 09:09:58 <_law_> where can i find mkfs tool mtd-utils? Sep 06 09:10:41 koen mickeyl: what is the status of uclibc 0.9.28? I just tried it (on ucslugc), it works fine... I'd quite like to move to it (I need to to make progress on ucslugc) Sep 06 09:12:08 hey CosmicPenguin, you about? Sep 06 09:13:48 jbowler the slug was mips architecure right? Sep 06 09:14:18 arm Sep 06 09:14:25 intel xscale Sep 06 09:14:37 hm oh okay Sep 06 09:14:49 woglinde: XScale arm, big endian Sep 06 09:14:54 hmm, current version of tailor seems to support monotone as a source Sep 06 09:14:56 * kergoth tests Sep 06 09:15:48 The uClibc.config files need updating for all architectures because (I believe) they changed the supported processor list (just x86 changes), but that's easy to do. Sep 06 09:16:32 oh wacky.. this is armv5teb Sep 06 09:16:38 mithro: morning Sep 06 09:17:31 jbowler did you sucessfull compile opie? Sep 06 09:17:34 i should dig my nslu2 out of my closet Sep 06 09:17:46 hey jbowler Sep 06 09:17:54 i forget, was it you who did ipkg? Sep 06 09:18:03 'did ipkg'? Sep 06 09:18:12 s/did/wrote the original/ Sep 06 09:18:16 I wonder how well this will like armeb Sep 06 09:18:24 pigi is hte current ipkg maintainer Sep 06 09:18:34 woglinde: no, just ucslugc (i.e. MACHINE=NSLU2) - absolutely no problems though so I suspect anything will work. Sep 06 09:18:39 no idea who wrote it originally. it was one of the handhelds.org guys Sep 06 09:18:55 ahh thats right Sep 06 09:19:08 i was after jbowler because of a kernel panic and some problems with upslug2 :P Sep 06 09:19:42 There's an openslug kernel panic still in the slugbug database (and I think I've seen the same thing). Sep 06 09:19:50 mithro: what's the upslug2 problem? Sep 06 09:20:01 jbowler oh okay Sep 06 09:20:34 morning folks Sep 06 09:20:38 he mickeyl Sep 06 09:20:39 woglinde: I simply removed the patches (all of the ones which aren't already fixed in .28 were added by me anyway ;-) Sep 06 09:20:43 jbowler: please feel free to update our uclibc Sep 06 09:20:56 jbowler: sometimes the slug doesn't come back after a flash with upslug2 Sep 06 09:21:21 mickeyl: Ok - I'll post a big warning on oe-handhelds (because, of course, it's not binary compatible - complete rebuild time) Sep 06 09:21:34 jbowler: my kernel panic just disappeared with my last image which kind of annoying Sep 06 09:21:45 RP: for reference, the hinge stuff can't be the reason for a higher CPUload if Opie is idling, it's just one fd more to the huge main loop select(). Sep 06 09:22:06 mithro: "doesn't come back?" - what state are the LEDs in? Sep 06 09:22:21 jbowler hm I will try to build opie with a uclibc-toolchain Sep 06 09:23:44 RP: then again, the Opie launcher is infamous for having too much polling going around anyway... Sep 06 09:23:55 jbowler: solid yellow Sep 06 09:24:25 * mickeyl mutters about that being corrected in project fluffie Sep 06 09:25:35 mithro: sounds like it is stuck inside RedBoot. You can try 'upslug2' (no arguments) and see if it is really in upgrade mode. Sep 06 09:25:54 mickeyl: tailor supports monotone as a source now. tailor is that project for migrating changesets between different source control tools. if this works, we can easily 1) maintain an anonymous svn, if that's something people want, and/or 2) I can show people the advantages of svk, and we can see if people would prefer that to mt. Sep 06 09:26:04 morning mickeyl Sep 06 09:26:13 mithro: if it is inside RedBoot a power cycle is harmless, but hangs like that in RedBoot suggest a hardware problem. Sep 06 09:26:24 kergoth: ok, very good. Sep 06 09:26:35 i flashed the same image with upslug original and no problems Sep 06 09:26:36 mickeyl: ok, something else must be going on then... Sep 06 09:27:58 mithro: so run 'upslug2 --verify -i foo.img' on the system flashed by upslug, if it verifies there is no problem in upslug2 Sep 06 09:29:17 okay, next time it occurs i will do that Sep 06 09:30:24 mithro: I was assuming you still have the image you flashed with upslug, i.e. I'm saying using upslug2 to verify the upslug result. Sep 06 09:30:38 That tests upslug2 Sep 06 09:31:31 upslug 1.17 verify can also be used for the inverse test, as of 1.17 upslug verify accepts an image argument (and does not write it to flash - i.e. if you don't say -u). Sep 06 09:34:53 My concern about the mtd support for spitz/borzoi is wrong a non-event. What the person meant was virtual blocks have been removed from the mtd code entirely which is fine for us. Sep 06 09:35:08 jbowler: well it happens only on like 1 in 10 flashes, the machine reboots and only ever gets as far as the yellow light - reflashing with upslug original and the machine works fine again Sep 06 09:35:24 RP: *phew* good to know Sep 06 09:36:05 I thought he meant they'd removed the option to disable them which would be a different matter... Sep 06 09:36:22 mithro: (is this openslug?) so you get 'rebooting... done' then a solid amber or does the machine beep then give a solid amber? Sep 06 09:36:23 yah, definitly Sep 06 09:36:42 The new mtd code sounds quite good for us really - I hope it hits mainline soon... Sep 06 09:36:50 we should get the summary patch Sep 06 09:36:57 *nod* that'd be pretty useful Sep 06 09:37:15 OT: we should check if the kexec patch against 2.6.12 still applies Sep 06 09:37:19 jbowler: it's ucslugc compiled myself Sep 06 09:37:38 In cpufreq.c there's a line which says: module_param(debug, uint, 0644); Would/should this appear in sysfs, or how else can I set the debug level? Sep 06 09:38:54 You either set it when you modprobe it or yes, it should be in sysfs Sep 06 09:39:02 I think anyway Sep 06 09:39:22 Okay. It doesn't appear in sysfs, and it should be compiled in, rather than a module... Sep 06 09:41:11 ah, kernel command line - I don't suppose that's very easy to access on a Z though Sep 06 09:41:34 lardman|work: It should be in /sys/module Sep 06 09:42:53 RP: Okay Sep 06 09:43:20 RP: it's not though, so I must have done something wrong Sep 06 09:46:09 okay, I updated OE and rebuilded from scratch Sep 06 09:46:23 when compiling mplayer the compiler still segfaults Sep 06 09:49:02 I'm not sure if module params appear for compiled in modules. Its an interesting one as they probably should Sep 06 09:49:44 question...what is the difference between the gaim package and the qpe-gaim package? Sep 06 09:49:51 that is..what does the qpe mean? Sep 06 09:49:55 qtopia/opie Sep 06 09:49:56 RP: I have a feeling that I may have to enable something else to get the debugging options to appear Sep 06 09:50:06 qpe = qt palmtop environment, which was the old name for qtopia Sep 06 09:50:18 RP: Would a built-in module fail to load silently? Sep 06 09:50:41 lardman|work: Depends how you load it ;-) Sep 06 09:50:55 RP: Shouldn't it just be loaded by default? Sep 06 09:51:10 Well, its init function should get called automatically, yes Sep 06 09:51:21 RP: I meant fail to load and not complain about it? Sep 06 09:51:57 If its builtin and your init function will get called. What happens then is up to the module Sep 06 09:52:12 They can fail silently if that's the way the module is written... Sep 06 09:52:24 Yep, as koen said, time to add printk's Sep 06 09:53:28 lardman|work: Have you a link to the patch - I'll glance over it if you like Sep 06 09:53:53 RP: http://people.bath.ac.uk/enpsgp/Zaurus/cpufreq/cpufreq.patch Sep 06 09:56:51 lardman|work: cpufreq_register_driver can probably fail, sliently as there's no warning message if it does Sep 06 09:57:19 there's that dprintk which requires that I define something else in defconfig Sep 06 09:57:51 I'll edit the function so that I don;t need to pass a parameter on the kernel command line and see what I get out Sep 06 10:00:04 lardman|work: Probably best Sep 06 10:00:52 RP: would I be better making the whole thing a module, that way I could pass the debug var on the command line? Sep 06 10:01:30 lardman|work: It saves reflashing your kernel every 5 minutes - you can just scp a new module in... Sep 06 10:01:45 RP: yep Sep 06 10:02:31 Does anyone know why cpufreq has never made it into the arm kernel btw? Did rmk block it? Sep 06 10:02:55 heh, wasnt it rmk that started cpufreq back in the day in the first place? Sep 06 10:03:51 kergoth: probably. rmk doesn't seem to like the pxa... Sep 06 10:04:21 * kergoth nods Sep 06 10:05:11 checking 1.1 gigs of source into svn takes a while. Sep 06 10:05:34 RP: Hmm, I fear it can't actually be a module - parts of it can, but not the guts Sep 06 10:06:10 lardman|work: That code you showed me is modular? Sep 06 10:06:37 RP: Is that rhetorical? - I don't know Sep 06 10:06:55 lardman|work: Ah, you need to make the config entry tristate, not bool Sep 06 10:07:33 RP: Is it not like that for a reason though? Sep 06 10:07:49 RP: Some things can't be modules can they? Sep 06 10:08:24 lardman|work: Compile it and see. My gut feeling is it will work although as you say, things can stop it Sep 06 10:08:36 Depends how well written its been... Sep 06 10:08:42 RP: Plus it specifically says in the cpufreq.c file that the debug param must be passed on the kernel command line... Sep 06 10:09:37 I'd hack that top #define to map it to a straight printk for testing Sep 06 10:09:48 RP: yep Sep 06 10:21:07 hmm is there a problem with the alsa driver packaging atm? Sep 06 10:23:17 or i seem to be missing fakeroot ... Sep 06 10:23:41 hmm Sep 06 10:27:58 Hmm, I see nothing in the logs from cpufreq. Can something be compiled, but not actually included in the kernel? Sep 06 10:28:28 thats a module is it not? Sep 06 10:28:38 no, compiled in Sep 06 10:29:02 or were you talking about fakeroot? Sep 06 10:29:14 nah i found it, i didnt have it installed Sep 06 10:30:18 * lardman|work heads off for supper Sep 06 10:37:05 jbowler: who is the expert on ucslugc? Sep 06 10:37:55 mithro: me Sep 06 10:41:16 he he, thought so - but thought i would ask first Sep 06 10:41:32 jbowler: ever managed to get alsa-utils to compile? Sep 06 10:42:49 mithro: everything that consistently compiles is in ucslugc-packages.bb. I known nmap builds, but nothing else. Sep 06 10:43:01 hej Sep 06 10:43:37 hey hrw Sep 06 10:44:20 which console ftp client has resume support in put? Sep 06 10:46:52 yafc... uf Sep 06 10:47:08 it looks like I have to find better hosting Sep 06 10:47:10 ;( Sep 06 10:48:20 he he, is there any easy way to find out if it's a uclibc compile problem? Sep 06 10:49:27 hmm.. I have to read about that gentoo USE flag.. Sep 06 10:50:31 it appears that uclibc sys_siglist is optional? Sep 06 10:50:43 any reason it's not turned on the ucslugc? Sep 06 10:50:43 its basically a space seperated list of words. if a given word is in the list, then a given piece of functionality is enabled. oe does the same thing, just with lots of variables instead of one. Sep 06 10:51:17 "use_wifi use_bt use_irda use_ipsec use_ipv6 is_vga" type Sep 06 10:51:18 ? Sep 06 10:51:50 yeah Sep 06 10:51:57 exactly Sep 06 10:52:00 its really quite limited Sep 06 10:52:02 flat namespace Sep 06 10:52:32 I'm working on opie-image/meta-opie recently and that kind of stuff really starts to be needed here Sep 06 10:52:44 so OE rocks while gentoo sucks? ;) Sep 06 10:52:52 well, oe's method is unorganized Sep 06 10:53:03 while theirs is more clean, but limited Sep 06 10:53:06 so we both suck Sep 06 10:53:33 ahh okay Sep 06 10:53:39 :) Sep 06 10:53:40 have_big_flash and is_vga are needed Sep 06 10:53:53 mithro: but both in other way ;) Sep 06 10:54:09 hrw: consider also things we already have vars for, like devfs, i18n, etc Sep 06 10:54:18 kergoth: yep Sep 06 10:54:25 i'd advise creating something vaguely USE-like, but with a non-flat namespace Sep 06 10:54:31 support a configuration hierarchy Sep 06 10:54:59 just for future flexibility Sep 06 10:55:37 USE = "network/bt network/wifi network/ipsec network/ipv6" so we can just add all with "if USE contain 'network' then add all from it" Sep 06 10:55:43 exactly. Sep 06 10:57:22 bittorrent sucks a bit.. fetching whole discography needs wait until it is done - trying to listen when it is not finished is hard as files have holes inside ;( Sep 06 10:57:26 should be easy to do. a couple quick def'd python functions and you're done Sep 06 10:57:47 yeah, that does suck about it Sep 06 10:58:38 hrw: some clients let you preference the packets at the beginning Sep 06 10:58:47 ooh, i didnt know that Sep 06 10:59:06 mithro: yep - but I have 70% already and dont want to read bittornado manual again Sep 06 10:59:26 kergoth: that "quick def'd python functions" scares me Sep 06 10:59:30 ~hail hh.org Sep 06 10:59:32 haha Sep 06 10:59:33 * ibot_ bows down to hh.org and chants, "I'M NOT WORTHY!!" Sep 06 11:01:16 how to put 700K file fast on my webpage? scp it to hh.org and ftp it from it to my hosting Sep 06 11:01:43 hrw: read the manual? it's simple. Click "Details", right click on a file, choose the priority Sep 06 11:02:09 JustinP: ever used bittornado curses? Sep 06 11:02:46 btdownloadcurses --one_option_for_something 1 --another_option_which_will_do_other_thing 2 torrent_url Sep 06 11:03:02 hehe Sep 06 11:03:28 it makes me tired of those long options Sep 06 11:03:49 no one ever did a bash tab complete for it then Sep 06 11:04:55 def opie_more_apps_on_big_flash_devices_depends(d): Sep 06 11:05:20 thats a pretty one Sep 06 11:05:48 DEPENDS = "meta-opie ${@opie_more_apps_on_big_flash_devices_depends}" Sep 06 11:05:49 too? Sep 06 11:06:23 and I'm thinking or opie-image_more_apps_on_big_flash_devices_depends() Sep 06 11:06:44 def if_in_use(val, d) .. DEPENDS += "${@if_in_use('vga', 'extra stuff', d)}" or something Sep 06 11:07:10 erm, def if_in_use(value_to_check, return_if_true, d) Sep 06 11:07:18 kergoth: ok - first want to get all deps done - then can clear it Sep 06 11:07:38 ${@} syntax is so butt-ugly Sep 06 11:07:46 * kergoth really wants bitbake-ng with a new file format :P Sep 06 11:08:26 hmm.. I have to start adding comments to meta-opie.bb Sep 06 11:09:25 I dont remember why I created task-opie-extra-settings ;) Sep 06 11:09:37 hehe Sep 06 11:10:37 ok. rescructured a bit - now it contain all which can be removed as not so needed Sep 06 11:10:47 opie-language opie-doctab opie-mediummount opie-networksettings-wlanplugin opie-networksettings-pppplugin Sep 06 11:19:05 kergoth: call it bitbake-aargh Sep 06 11:19:28 hehe Sep 06 11:20:01 ~lart Photoshop Sep 06 11:20:02 * ibot_ nukes Photoshop with a single large nuke Sep 06 11:20:14 1GB RAM not enough for a 300K jpeg? I don't think so Sep 06 11:20:27 depends on your layers and masks Sep 06 11:20:48 it's just a fricking jpeg! Sep 06 11:21:05 photoshop has a nasty issue with allocating the image size per layer if your undo steps are too large Sep 06 11:21:08 233K. How can 1GB RAM and >1GB HD space not be enough to open that thing Sep 06 11:21:14 again...it's a joeg Sep 06 11:21:16 jpeg Sep 06 11:21:27 jpegs son't have "undo steps" or layers AFAIK Sep 06 11:21:45 nothing to say its a valid jpeg either Sep 06 11:21:56 maybe its decompression bomb? Sep 06 11:22:07 200Kx200K one color jpeg? Sep 06 11:22:14 no Sep 06 11:22:18 opens fine in anything else Sep 06 11:22:36 including The Gimp, MS Photo Viewer, Firefox....hell, gmail even thumbnailed it fine Sep 06 11:23:15 bah Sep 06 11:23:26 open in MS Photo Editor, re-save, Photoshop opens it Sep 06 11:23:29 there is the posibility of a cache lock ... Sep 06 11:23:40 I'm pretty sure this jpeg was made in photoshop as well... Sep 06 11:23:43 doesnt usually happen often in ps Sep 06 11:24:17 heh... I want python doc made like php one Sep 06 11:24:25 where examples are examples Sep 06 11:24:26 if ps crashes whil worling on a large image its chache locks and undo histories stay in place and get reloaded when its restarted Sep 06 11:24:34 working* Sep 06 11:24:35 nothing like that Sep 06 11:24:43 yeah Sep 06 11:24:57 heck, if I resave in MSPE with 100% quality it's a bigger file but still opens fine Sep 06 11:25:00 stupid PS Sep 06 11:25:02 my only other guess is that it might be the new jpeg format Sep 06 11:25:31 kergoth: bitbake-ng will allow to use other langs then python and lua? maybe I will start to use php inside recipes ;) Sep 06 11:25:48 * koen cheers for php Sep 06 11:26:04 hrw: yeah, not just flexibility wrt to embeddedl anguage, but also flexilbiity wrt different recipe formats in general Sep 06 11:27:06 * koen votes for replacing '-f' with ', damnit' Sep 06 11:27:20 function is_in_use($checkvar){ return eregi($checkvar, $GLOBALS['USE']);} kind Sep 06 11:43:47 * AvengerMoJo setting a new build enivornment to test oe Sep 06 11:44:21 koen what version of gcc you suggest for my crosscompile? Sep 06 11:46:06 anything without 'csl' in it's name Sep 06 11:47:09 koen thanks ... why is that? Sep 06 11:47:38 the csl version in OE is broken Sep 06 11:47:58 * koen has a feeling oesources is not running anymore Sep 06 11:48:11 koen oh hms Sep 06 11:48:41 koen I'm having a xscale 270 .. do you think gcc 3.3.3 mcpu=xscale will be better then generic arm cpu type? Sep 06 11:49:27 tuning for xscale should give you 10% better performance with memcopy according to RP Sep 06 11:49:56 koen I see ... Sep 06 11:50:15 I have a fundamental question. When using OE sometimes it will fetch a package-x.y.z-native. What does "native" signify? does it mean its building it for the cross compilers platform and NOT the target system platform? if so, why would it do this? Sep 06 11:50:28 yes Sep 06 11:50:36 because your build machine's tools arent always suitable. Sep 06 11:52:01 koen this is what I have been using for my cross compiler ...http://kegel.com/crosstool/ <--- what about you? Sep 06 11:52:21 OE :) Sep 06 11:52:25 <_law_> cant find mkfs.ext3 i have installed e2fsprogs already Sep 06 11:52:43 I usually add the OE built crosscompilers to my path Sep 06 11:52:55 or just make a .bb Sep 06 11:53:02 lardman: e2fsprogs-mkfs too? (or how it was named) Sep 06 11:53:34 koen sorry I'm still fairly new to oe... where can I read up on oe way of building toolchains? Sep 06 11:53:48 oe builds itself a toolchain when it needs a toolchain Sep 06 11:53:51 what do you want to know? Sep 06 11:53:56 <_law_> hrw, thats the reason, thanks :-) Sep 06 11:54:26 kergoth: oh :) I didn't notice that Sep 06 12:00:50 my opie-image/meta-opie works ok for c7x0 - time for tests with collie Sep 06 12:01:01 excellent Sep 06 12:01:06 hms csl didnt include the uclibc patches Sep 06 12:01:17 don't use csl Sep 06 12:01:24 *g* Sep 06 12:01:36 if its broken, why is it still there? :P Sep 06 12:01:38 mickeyl: force OZ to not use csl or set def_pref=-1 for it Sep 06 12:01:42 at least the version in OE is known to miscompile stuff Sep 06 12:01:46 kergoth: good question. Sep 06 12:01:50 * mickeyl looks to the toolchain guys Sep 06 12:01:59 opensimpad still uses csl Sep 06 12:02:02 no one looks back heh Sep 06 12:02:06 tailor of monotone->svn failed :( Sep 06 12:02:16 silly bug it looks like though Sep 06 12:02:23 it tried to svn add a file in a subdir before adding the subdir itself :) Sep 06 12:03:11 * koen is puzzled by ppp bugs Sep 06 12:03:20 I get the exact same sequence on laptop and ipaq Sep 06 12:03:33 excpet the ipaq gets a [LCP TermReq id=0x3] Sep 06 12:04:00 * koen prints the ppp manual and sets it on fire Sep 06 12:04:23 hehe Sep 06 12:04:26 ooh, ppp. fun fun Sep 06 12:04:30 * kergoth gets to dive into pppoe soon Sep 06 12:04:41 extra_stuff := "${@bb_check_value('ROOT_FLASH_SIZE', 17, 'task-opie-extra-games task-opie-extra-apps task-opie-extra-styles', d)}" Sep 06 12:05:11 hmm s/>=/> will make it more clear Sep 06 12:05:44 function name is not best anyway Sep 06 12:06:01 check_value? Sep 06 12:06:12 this duplicates base_conditional a bit Sep 06 12:06:32 or do you want >= ?` Sep 06 12:06:58 mickeyl: I will look into base_cond no worry Sep 06 12:07:17 well, the problem is if you want to have >= you end up writing an own function anyway Sep 06 12:07:41 OE is missing a lot of basic stuff Sep 06 12:08:03 *nod*. the question is, is adding such things as functions to base.bbclass the right thing or not Sep 06 12:08:06 versioned depends (samba >= 3.0.0), simple if a > b structures, etc Sep 06 12:08:27 12145920 opie-image-collie-20050906190424.rootfs.jffs2 - nice Sep 06 12:08:37 cool Sep 06 12:09:10 mickeyl: http://home.hrw.one.pl/tmp/ipk-status.txt - can you look is everything in place? Sep 06 12:09:27 uah Sep 06 12:09:29 heh Sep 06 12:09:31