**** BEGIN LOGGING AT Thu May 12 02:58:19 2005 May 12 02:59:34 DWAIN Distro Without Any Interesting Name May 12 03:01:26 Familiar, kolibri, angstrom are my favourites May 12 03:01:51 and galapagos May 12 03:01:55 what about pdaYrom? May 12 03:02:06 STOP CALLING FLASH ROM May 12 03:02:09 May 12 03:02:50 Good morning May 12 03:02:59 hey Twiun May 12 03:03:01 koen: anyway I guess we can abuse any forum May 12 03:03:34 umbaumba distro May 12 03:04:05 koen: I can vote for it May 12 03:04:11 :) May 12 03:04:40 or the Utopia distro May 12 03:04:50 Utopia is good... May 12 03:04:55 with "that's without a q" as a slogan May 12 03:05:02 Project Utopia to make Linux Userspace fit for PDAs May 12 03:05:21 Secure Utopia for the pointy haired types May 12 03:05:49 cya later May 12 03:07:46 Q^HUtopia Linux May 12 03:15:51 fsck... May 12 03:16:17 RP: 2.6.11 still has usbnet gadget inside or in module? May 12 03:18:20 great... kernel panic May 12 03:18:24 is there a way to let wget detect http://aleron.dl.sourceforge.net/sourceforge/lex/flex-2.5.31.tar.bz2 as a non-existent page? (gives a "please choose another mirror, the one you selected is broken"-page). it screws up the built process for me as some html page gets downloaded, the _fetch task doesn't fail, but populate_staging does then (with an obscure error). May 12 03:19:32 that page exist May 12 03:19:41 so wget is doing the right thing May 12 03:20:23 yes, but the fetch script isn't May 12 03:20:39 is there a way to force wget not to forward "Location:"s? May 12 03:20:45 and/or refreshs May 12 03:20:47 patches are welcome May 12 03:23:05 hrw|work: Its now a module May 12 03:23:52 heh.. hopefully I got bt card May 12 03:23:52 hrw|work: I've just done the sums and tacc_ns is being calclated correctly from the CSD according to that data sheet May 12 03:24:12 RP: o.. I can check it... May 12 03:25:31 tmbinc: chnage your sf.net mirror in bitbake.conf May 12 03:28:57 RP: could you export patch? May 12 03:31:57 hrw|work: Which patch? May 12 03:32:42 oops.. you said that it is already correct.. misunderstood I.. May 12 03:33:37 uv1: Just seen your message - the patch is in openembedded May 12 03:34:18 hrw|work: I've just done this calculation fully and the timeouts are all correct. Your card doesn't respond to requests in the time it says it can May 12 03:35:08 hrw|work: Try changing your patch to add the line writel((timeout + 255 + 1000) / 256, host->base + MMC_RDTO); May 12 03:35:39 increase or decrease the value 1000 until you find the point it works/doesn't work May 12 03:36:02 ok May 12 03:36:05 Once we have that, we can just add a fudge factor in to help slow cards May 12 03:36:23 I don't want to make it bigger than we have to May 12 03:41:20 RP: what is max value? May 12 03:44:07 hrw|work: 0xffff May 12 03:44:12 (in total) May 12 03:44:29 in your case, timeout works out at 30001 May 12 03:51:54 Hi all May 12 03:53:12 RP: I have tested your binary kernel from this morning and am quite happy. But now and then the system wakes up and shows the opie screen without an alarm. It shows the screen for approx 1 sec and then suspends again. May 12 03:58:51 uv1: I'm also seeing something like that here. I'll have to turn debugging back on and have another look at it tonight May 12 03:59:05 uv1: Is your device charging when its doing this? May 12 04:04:16 RP: No charging, was testing auto suspend. Trying with charging now ... May 12 04:11:36 uv1: Is it actually suspending again or does it just blank the screen (you can test by pressing the shift key) May 12 04:12:39 uv1: I need to rewrite all this code as its full of holes as it stands :-/ May 12 04:13:40 RP: Did not happen the last 10 mins while charging. Trying whether screen is blanked or really suspend now May 12 04:14:49 uf.. bnep0 connected May 12 04:18:57 uv1: When the device is misbehaving, I really need to know: a) how it was suspended b) whether it was on ac power c) how charged the battery was d) how long after suspending did it wake up May 12 04:19:06 file-storage gadget works May 12 04:19:42 root@c7x0:~# modprobe g_file_storage file=/dev/mmcblk0 May 12 04:19:56 hrw|work: neat :) May 12 04:19:56 and card is available as usb storage May 12 04:20:01 ah, that answers my questions :) May 12 04:20:44 The latest kernel has a bug. It will wake up after 20minutes after being suspended if on ac power with a fairly full battery May 12 04:20:47 cool, I was going to try that out May 12 04:20:54 What about serial? May 12 04:21:18 but when tried to del file from card whole Z is frozen... maybe its because of file size and my sdcard timeouts May 12 04:22:32 uf.. Z recovered May 12 04:23:27 and file deleted? May 12 04:23:28 hrw|work: Deleting files is slow :-/ May 12 04:23:41 so is copying them :) May 12 04:23:56 argh... May 12 04:24:02 ~lart dt1l_cs May 12 04:24:02 * ibot puts dt1l_cs through a wood chipper May 12 04:24:08 Just be thankful it works at all ;-) May 12 04:24:18 Of course May 12 04:24:29 Seriously, if anyone can find out where the speed issue is, I'd be very insterested May 12 04:25:22 RP: a) AutoSuspend by time (light&Power), b) no AC power, c) nearly full battery. d) Approx 1-2 mins after suspend. Will seriously test the suspend menu option again now with the same conditions, than the autosleep... May 12 04:25:32 beware dt1l_cs based bt cards (like socket rev E,F) May 12 04:27:28 uv1: That makes no sense. Perhaps an alarm hadn't been cleared... :-/ May 12 04:27:53 files which I tried to delete are now input/output errors.. (maybe I should umount card before) May 12 04:28:31 hrw|work: you should umount before using it as a storage device, yes May 12 04:28:43 hrw|work: what's up with the BT kernel module? May 12 04:29:04 lardman|busy: lack of suspend/resume May 12 04:29:09 hrw|work: but if you hadn't umounted it, it shouldn't have worked at all May 12 04:29:14 RP: I just want to have the saem condition. Adjust an alarm, suspend, wait 10 mins and watch for waking up / going to sleep again. May 12 04:30:46 RP: Have another issue right now. The last wake up stopped on the text console (no opie screen any more). The last messages are "corgi_mci_setpower: on\nAllocated rca : 0002\nRestarting tasks... done". Can I do something here without rebooting ? May 12 04:31:32 RP: I think the device is frozen :-( May 12 04:32:21 Try holding down Fn and pressing the right arrow May 12 04:32:33 RP: Nothing May 12 04:32:51 Are the SysRq keys working May 12 04:32:54 ? May 12 04:33:06 RP: SysRq keys ? May 12 04:33:35 Hold down the right hand japanese key, then press Home. Release the jap key continuing to hold Home and press S May 12 04:33:58 It should say "Emergency Sync" May 12 04:34:40 If it says that, userspace had locked up or some driver is blocking. If that doesn't work, its truely locked up May 12 04:34:41 RP: Nothing. The only thing that changes is the Charge LED comes back, if reconnected to power. May 12 04:35:10 and you have text on the screen? May 12 04:35:18 RP: yes May 12 04:35:41 If the charge light changes, the kernel is still running May 12 04:35:58 SysRq should also work :-/ May 12 04:36:36 You're going to need to reboot anyway, but try the above but press B instead of S. This should reboot the device May 12 04:36:52 If that works, it sounds like the framebuffer has locked up... May 12 04:38:31 RP: "SysReq + B" switched the display off and I was able to Switch the device ON with ON/OFF button. Beforehand ON/OFF was pressed several times, Thats possibly the reason, why reboot did not work ? May 12 04:38:35 NOTE: [13:38] package linux-openzaurus-2.6.11-r18: task do_package: started May 12 04:38:35 sh: arm-linux-depmod: command not found May 12 04:38:39 strange.. May 12 04:39:34 uv1: just to be clear, the device rebooted? May 12 04:40:34 RP: Not from "SysReq + B" but I was able to switch the device ON by Button. After that it rebooted. May 12 04:41:04 uv1: That's normal for a sysrq + B May 12 04:42:06 That suggests that the consoles had stalled. Very strange :-/ May 12 04:44:01 RP: So "SysRq +B" is in fact halt, instead of reboot ? Just tried if again, same behaviour. May 12 04:44:50 uv1: It does reboot the device into the low level bootloader. The bootloader waits for a keypress of the power button May 12 04:45:04 The device isn't halted as such May 12 04:46:36 RP: OK then, Now performing the "Suspend Menu" test. May 12 04:48:04 I've put a new kernel up at http://www.rpsys.net/openzaurus/packages/ May 12 04:48:22 This should fix the issues I was seeing and maybe some others, I really don't know... May 12 04:49:02 flashing 0x8000 timeout kernel May 12 04:51:15 hmm.. way to small.. May 12 04:51:27 MMC kernel timeout = 300000 May 12 04:51:48 I forgot that it is divided by 256 May 12 04:54:13 RP: If suspended from menu, the effect does not appear obviously. May 12 04:54:26 RP: printk support hexadecimal? May 12 04:54:32 RP: A new kernel again ? Or the one from this morning ? May 12 04:55:52 uv1: Its a new kernel May 12 04:56:19 uv1: It may be that opie's suspend from the menu is doing something funny :-/ May 12 04:56:42 RP: I just observed another issue. If display light is off (light and power) and system goes to suspend, the light is switched before it relly suspends. Will try you new kernel now ... May 12 04:56:42 hrw|work: Yes, it does %x iirc May 12 04:56:54 hrw|work: Just like printf May 12 04:57:10 thx May 12 04:57:55 uv1: You need to talk to the opie people about the light - that's not the kernel May 12 04:59:25 ok. timeout set to 0x800000 - rebuild, reflash May 12 04:59:29 RP: OK will ask the opie people. BTW do I have to reflash the device to get the kernel working, or would a scp of the image be OK ? May 12 05:00:04 You can just reflash the kernel May 12 05:00:15 scp won't work as its stored in a special area of the flash May 12 05:00:26 You don't need to reflash the image itself though - just the kernel May 12 05:00:53 RP: So I assume, if initrd.bin is not there, only the kernel is relfashed ? May 12 05:00:59 uv1: yes May 12 05:01:05 uv1: correct May 12 05:04:49 iirc there was an option in kernel to run kernel directly from flashtype storage if it is mapped in address space May 12 05:07:09 Testing new kernel from RP May 12 05:08:06 fsck. again trapped in chicken<>egg sdcard trap... ;( May 12 05:09:58 someone can explain why etc/unit_conversion.dat is not with install for opie-calculator package? May 12 05:10:11 is it bug? May 12 05:28:33 good afternoon folks May 12 05:29:22 hi chris144 May 12 05:29:29 gints|wrk: yes, if the calculator needs it,but its not there: BUG May 12 05:29:43 hi hrw|work May 12 05:31:39 folks, is anybody aware of a problem with the bytes_per_inode patch in genext2fs-native? see http://pastebin.ca/11580. thanks! May 12 05:32:24 probably patch is not needed now May 12 05:33:09 hrw|work you think is is that simple? ok, i will try it tonight May 12 05:33:37 re May 12 05:35:32 RP: temp wake up is still there ;-( May 12 05:39:26 hi b0ti|work May 12 05:48:40 RP: Sorry, it was not temp wake up, but the formerly mentioned LightOnBeforeSuspend (opie). May 12 05:50:50 ~seen kergoth May 12 05:50:52 kergoth is currently on #devlounge #uclibc #nslu2-linux #oe #elinux #handhelds.org #ol #tacobeam. Has said a total of 200 messages. Is idling for 6h 22m 11s May 12 05:54:54 uv1: That's ok, at least we know what it is now. I didn't realise opie did that :) May 12 05:57:25 hi! anyone from Europe alive? May 12 05:59:28 yeq May 12 05:59:32 yep May 12 06:00:36 mithro: yup May 12 06:00:42 mithro: yes May 12 06:01:27 actually it is Munich, if you have handheld questions, i am not the right guy, general oe I might be able to help with May 12 06:05:13 mithro: Frankfurt is alive May 12 06:05:31 hi schurig May 12 06:06:31 chris144|work: Ei guude wie! May 12 06:07:43 schurig: cool May 12 06:07:48 schurig: I think the proper reply could be a bavarian "servus" :-) May 12 06:08:08 schurig: you wouldn't happen to know if websites (in english) which show fairs from Frankfurt to Egpyt? May 12 06:08:32 anybody else know good websites for looking up flights in europe to egpyt? May 12 06:09:02 mithro: www.lufthansa.com May 12 06:09:55 mithro: I think on http://www.travelocity.com/ you will find all flights all airlines from Frankfurt to egypt... May 12 06:10:44 i'm trying to find the cheapest way to get from europe to egpyt May 12 06:11:15 mithro: swim :-) May 12 06:12:13 :P May 12 06:13:07 only be aware from sharks May 12 06:16:47 it seems cheapest to fly from athens May 12 06:17:23 mithro: the book exhibition and the internationional car exhibution might be big enougth for some egyptian pages somewhere May 12 06:17:51 mithro: the german name for the car exhibition is IAA May 12 06:18:17 mithro: but of course they don't have a egyptian or otherwise arabic web page May 12 06:18:19 http://www.iaa.de/iaa-interim/index.html May 12 06:28:29 koen, i forgot i said id do the efl changes tonight ... they will have to wait untill after my exam tomorrow May 12 06:28:36 which is today May 12 06:29:27 emte: ok May 12 06:30:17 altho i could have filled em with latch and flipflop info :P May 12 06:33:09 mornning May 12 06:50:36 hi chouimat May 12 06:55:25 re May 12 06:55:42 anyone tried to build this with oe already? May 12 06:55:43 http://www.denx.de/twiki/bin/view/Know/MiniFOHome May 12 07:36:31 Is there anyway to tell bk to show you changsets related to a file or directory? May 12 07:38:40 bk sccslog? May 12 07:40:26 thanks! May 12 07:47:06 is someone interested in testing the final patches? May 12 07:47:59 I can but not now May 12 07:48:09 same here May 12 07:50:16 are they webified? May 12 07:50:20 zecke: I'll happily test them May 12 07:54:45 * zecke wonders where I can put them May 12 07:54:59 zecke: hh.org? May 12 07:56:20 I don't mind putting them on my webserver... May 12 07:57:20 I want to try them, but I'll have to set up a development environment, and I don't have one handy May 12 08:36:20 bye May 12 09:08:20 Bernardo|away: sounds good May 12 09:08:25 Bernardo|away: done :D May 12 09:08:33 03mickeyl 07 * r1.3267 10openembedded/packages/minifo/minifo-modules_0.6.1-pre1.bb: add minifo-modules, a mini fanout overlay file system May 12 09:08:55 hi mickeyl May 12 09:09:23 no having net access is a bitch May 12 09:10:03 no kidding May 12 09:10:17 still no date for having adsl at home May 12 09:12:51 so I can't test that stuff... sounds like a alternative for the rboot scripts. It eevn would allow us to have the /boot partition r/o May 12 09:13:01 I mean / May 12 09:14:41 RP: ping May 12 09:29:17 re May 12 09:30:01 hrw: wb May 12 09:30:18 suspending 'works' now on 2.6/gpe May 12 09:30:32 I only have to ssh in and do 'chvt 1' to get X back May 12 09:30:57 I've got a question on 2.6/husky not resuming May 12 09:32:20 hrmmm Fetch class isn't as fine grained as I'd hoped :p May 12 09:35:27 I'm getting closer to my sd card timeouts May 12 09:36:29 hrw: I'm afraid I'll have to do something like that when we finally have a 2.6 kernel on a cx000 May 12 09:36:51 my 1GB SD card gives lots of timeouts when booting already May 12 09:37:08 Bernardo: I added a new gpsd, could you check if it's ok? May 12 09:37:27 my card reports timeout = 0x493e0 but it is too small for it.. with 0x200000 it works. I will try with 0x80000 now May 12 09:37:49 koen: I'm afraid not... :( I still haven't got adsl switched to my new place, I do hope they'll transfer it before the 19th... May 12 09:38:00 ah, ok May 12 09:38:18 until then, I only have a server connected to the old line, to let me use email and irc when connecting from work May 12 09:39:20 :( May 12 09:39:26 ouch May 12 09:39:27 4,2,Y and flashing May 12 09:39:51 (japanese),cf,y :) May 12 09:40:11 the worst part is that if the "fidelization" period had already finished, I'd be able to have adsl at the new place in 1-2 days. May 12 09:40:38 hmm May 12 09:40:40 huh? May 12 09:40:44 drivers/built-in.o(.init.text+0x9a8): In function `elpp_init': May 12 09:40:45 pxamci.c: undefined reference to `con2fb_map' May 12 09:40:45 make: *** [.tmp_vmlinux1] Fehler 1 May 12 09:41:07 pxamci.c doesn't refer con2fbmap at all, nor does it contain a function names elpp_init May 12 09:41:19 mci is for mmc, right? May 12 09:41:24 yeah May 12 09:41:31 it doesn't relate to the console at all May 12 09:41:33 * mickeyl confused May 12 09:42:31 ok. 0x80000 is too small timeout value May 12 09:42:39 I guess elpp wants an SD card to work :) May 12 09:42:44 heh May 12 09:43:48 and again 42y May 12 09:45:02 why it boots so slow... May 12 09:45:32 btw - I vote for removing banner initscript May 12 09:45:55 no way May 12 09:46:02 ah May 12 09:46:03 wait May 12 09:46:06 banner May 12 09:46:10 not checkversion May 12 09:46:11 right? May 12 09:46:12 :D May 12 09:46:14 yes May 12 09:46:26 hrw: seconded. May 12 09:46:32 yank it May 12 09:46:40 although it's helpful sometimes May 12 09:46:45 best, do it conditionally May 12 09:46:50 yank it for release images May 12 09:47:00 because on debug we sometimes want to know if init starts at all May 12 09:47:37 0x100000 is over twice to 0x493e0 and works May 12 09:47:59 time for 0xC0000 then ;) May 12 09:50:30 binary search May 12 09:50:46 not on start May 12 09:51:08 I had maximum set before and started testing from quite low value May 12 09:52:19 both sd cards which I have here have same timeout value.. HS kingston has it twice smaller but I cannot test it May 12 09:52:39 i have a sandisk and a kingston here May 12 09:52:46 sandisk is 128m, kingston is 512m May 12 09:52:52 both work with the default timeout May 12 09:52:53 sandisk 64m, pretec 256m May 12 09:54:11 hmm.. my sdcontrol use /mnt/ May 12 09:54:37 both works with 0xc0000 but dont with 0x80000 May 12 09:55:11 time for 0xa0000 May 12 09:55:20 re May 12 09:55:24 re zecke May 12 09:55:25 wb May 12 09:57:02 ah, my bt card finally works on my z May 12 09:57:06 bye guys, see you tomorrow May 12 09:57:10 cya May 12 09:57:14 mickeyl: thanks for adding the 8250 patch May 12 09:57:19 np May 12 09:57:21 Bernardo: see you May 12 09:58:55 0xa0000 is too small May 12 09:58:59 hmm i don't get it fbcon.c defines static signed char con2fb_map[MAX_NR_CONSOLES]; and it is in the same built-in.o as the one referring it. now why the hell doesn't that link May 12 09:59:51 mickeyl: ist conf2b_map static? May 12 10:00:08 ya May 12 10:00:38 mickeyl: try EXPORT_SYMBOL(con2fb_map); May 12 10:00:46 ok, will try that May 12 10:00:53 it's referred to via 'extern char con2fb_map[MAX_NR_CONSOLES];' May 12 10:01:01 does that 'signed' matter ? May 12 10:02:14 oh yeah May 12 10:02:25 char = signed char on x86 May 12 10:02:30 and unsigned char on arm May 12 10:02:35 oh May 12 10:02:45 perhaps they changed that since 2.6.11 May 12 10:02:49 * mickeyl checks May 12 10:02:56 yeah forget the EXPORT_SYMBOL, it seems to be only needed for modules... May 12 10:03:57 RP: timeout set to 0xB0000 works for me, 0xA0000 does not. both cards report 0x493e0 May 12 10:04:11 ibot: 0x493e0 * 2 May 12 10:04:31 600000 May 12 10:04:45 0x927c0 May 12 10:05:33 so its over 2.4 times bigger May 12 10:09:09 hrw: That's the value you're setting timeout to? May 12 10:09:30 *sigh* May 12 10:09:34 mickeyl: There is a bug in our arm toolchain which makes those error mesages a bit unhelpful May 12 10:09:39 | drivers/built-in.o(.init.text+0x9a8): In function `elpp_init': May 12 10:09:39 | pxamci.c: undefined reference to `con2fb_map' May 12 10:09:46 adding signed made no difference May 12 10:09:47 RP: i ssee May 12 10:11:11 mickeyl: http://www.arm.linux.org.uk/developer/toolchain/ has a patch which will fix that May 12 10:11:23 I keep meaning to ask one of our gcc gurus about including it May 12 10:11:34 Sorry, binutils... May 12 10:12:14 * zecke gets food May 12 10:12:31 RP: yes May 12 10:13:13 137 timeout = 0xb0000; // hack May 12 10:13:13 138 writel((timeout + 255) / 256, host->base + MMC_RDTO); May 12 10:13:22 in drivers/mmc/pxamci.c May 12 10:14:04 hrw: Ok. That's quite an extended timeout :-/ May 12 10:14:41 RP: much smaller than it was before May 12 10:14:52 00001b9c l O .bss 0000003f con2fb_map May 12 10:14:58 before = with older patch May 12 10:15:06 that damn symbol is there May 12 10:15:07 hrw: Its a definite improvement May 12 10:15:44 hrw: Now the only question that remains is why it takes as long... May 12 10:15:57 hrw: I'll add a patch to incease it May 12 10:16:14 RP: no idea.. both my cards have that problem and both has the same timeout May 12 10:16:40 hrw: It sounds like there may be some CSD data that decodes differently somewhere May 12 10:16:51 maybe May 12 10:16:58 wanna CSD dump? May 12 10:17:09 I have it from when we last discussed it May 12 10:17:20 I manually worked it out and it seems to work out ok May 12 10:17:22 ok May 12 10:17:28 The only difference would be if the spec had changed May 12 10:18:08 mickeyl: You might need to add something like extern ... con2fb_map(...); to the elpp patch May 12 10:18:21 that's already in May 12 10:18:28 mickeyl: Ah :-/ May 12 10:18:50 that's odd: May 12 10:18:50 mickey@r2d2:/local/pkg/oe/c7x0/tmp/work/linux-openzaurus-2.6.12-rc3-mm2-r6/linux-2.6.11$ /local/pkg/oe/c7x0/tmp/cross/bin/arm-linux-nm drivers/video/console/built-in.o |grep con2fb May 12 10:18:50 00001b9c b con2fb_map May 12 10:18:50 U con2fb_map May 12 10:18:50 00001bdb b con2fb_map_boot May 12 10:18:51 00000fec t set_con2fb_map May 12 10:18:59 U is undefined isn't it? May 12 10:19:23 hmm wtf. May 12 10:19:57 I think U is undefined, yes May 12 10:21:28 aaah May 12 10:21:32 here we are: May 12 10:21:41 --- linux-openzaurus-2.6.11-r18/linux-2.6.11/drivers/video/console/fbcon.c 2005-05-12 19:00:30.674948704 +0200 May 12 10:21:41 +++ linux-openzaurus-2.6.12-rc3-mm2-r6/linux-2.6.11/drivers/video/console/fbcon.c 2005-05-12 19:08:48.989565824 +0200 May 12 10:21:41 @@ -110,15 +110,15 @@ May 12 10:21:41 }; May 12 10:21:41 struct display fb_display[MAX_NR_CONSOLES]; May 12 10:21:43 -signed char con2fb_map[MAX_NR_CONSOLES]; May 12 10:21:45 mickeyl: Error: "@" is not a valid command. May 12 10:21:45 -signed char con2fb_map_boot[MAX_NR_CONSOLES]; May 12 10:21:47 +static signed char con2fb_map[MAX_NR_CONSOLES]; May 12 10:21:49 +static signed char con2fb_map_boot[MAX_NR_CONSOLES]; May 12 10:21:51 shut up kbot May 12 10:22:02 they made it static May 12 10:22:11 that affects visibility ? May 12 10:22:15 <--- C newbie May 12 10:22:17 Yes May 12 10:22:20 bummer May 12 10:22:32 should i take that patch back or use EXPORT_SYMBOL ? May 12 10:22:34 ANything static is only seen in the local file May 12 10:22:40 You do that to speed the compile May 12 10:22:45 Just remove the static May 12 10:22:50 ok, thanks May 12 10:22:57 ~lart premature optimizations May 12 10:22:57 * ibot takes a large goose feather pillow and swings it wildly in premature optimizations's direction, hitting premature optimizations and sending premature optimizations flying into the closet May 12 10:23:20 Its fair game if nothing in the kernel uses is :-/ May 12 10:23:57 hrw: I think I'll all a mutliply by 4 to the mmc code May 12 10:24:04 That should solve your problems May 12 10:25:45 RP: ugly hack which should work? :) May 12 10:27:12 hrw: Yes :) May 12 10:27:40 hrw: Your card doesn't respond in the time is says it can so I think its faulty hardware however if we can make it work... May 12 10:28:32 hrw: looks like you built opie-wellenreiter-r1 after sharp-compat. it now needs sharp-compat :/ May 12 10:28:42 RP: and lets see what will lkml say May 12 10:28:47 mickeyl: really? fsck May 12 10:28:51 hehe May 12 10:29:48 Depends: libqpe1 (>= 1.2.0), libopiecore2-1 (>= 1.2.0), libopieui2-1 (>= 1.2.0), libopienet2-1 (>= 1.2.0), libqte2 (>= 2. May 12 10:29:51 3.10), libgcc1 (>= 3.4.3), libc6 (>= 2.3.2+cvs20040726) May 12 10:29:58 get off mickeyl ;) May 12 10:30:16 *shrug* May 12 10:30:22 i judged from the post in the mailing list May 12 10:31:21 oz-'so-called-users'? May 12 10:31:46 mickeyl, hrw: You didn't know that? :p May 12 10:32:00 sharp-hardware iq drop May 12 10:32:00 ya May 12 10:32:07 right May 12 10:32:10 i tend to forget that May 12 10:32:10 :D May 12 10:32:18 it fucks up the dependency generation May 12 10:32:27 sure we know May 12 10:32:35 but i was surprised that hrw would be fooled by that May 12 10:32:39 turns out he wasn't May 12 10:32:43 instead one user smoked crack or so May 12 10:32:44 treke: sharp-compat-libs fucked some changes ago May 12 10:32:44 *shrug* May 12 10:32:56 just one? May 12 10:32:59 that's an improvement May 12 10:33:06 :D May 12 10:33:10 ChangeSet@1.2930.65.1, 2005-03-14 12:26:37+01:00, hrw@marcinj.local May 12 10:33:10 allow packages to not have shlibs info generated - useful with packages like sharp-compat-libs May 12 10:33:32 NOTE: package linux-openzaurus-2.6.12-rc3-mm2-r6: task do_package: completed May 12 10:33:35 w/ elpp May 12 10:33:37 very good, thanks RP May 12 10:33:47 mickeyl: flash it May 12 10:33:53 mickeyl: np. pleased you found the problem :) May 12 10:34:05 lets see how that kernel behaves... May 12 10:34:24 mickeyl: and then look why hostap gives WE0 May 12 10:34:56 ya, will do that May 12 10:35:03 i'll build a complete image now May 12 10:35:07 * mickeyl bitbake opie-image May 12 10:35:17 03mickeyl 07 * r1.3268 10openembedded/packages/linux/ (2 files in 2 dirs): unbreak elpp patch for 2.6.12-rc3 May 12 10:35:19 mickeyl: Hinge bechaviour under opie is the only known broken thing May 12 10:35:33 we better don't touch that until the interface issues have been sorted out May 12 10:35:42 mickeyl: agreed May 12 10:35:54 btw. May 12 10:35:56 mickeyl: I'm waiting for 2.6.12 to ship, then I'll force the issue May 12 10:35:58 May 12 10:36:02 decide on a LED interface please May 12 10:36:04 May 12 10:36:09 ;) May 12 10:36:12 Ah, yes :) May 12 10:36:32 I recommend to use Johns for the time being, but its your call of course May 12 10:36:34 That might be a nice job for tomorrow :) May 12 10:36:45 I could do with something relaxing :) May 12 10:36:49 heh May 12 10:36:55 how's the weather at your place? May 12 10:37:08 I want to hack John's patch a bit May 12 10:37:15 Its rather nice here for a change May 12 10:37:44 then better go out and make something nice May 12 10:38:06 It was out over the weekend and found hail stones and sleet :-/ May 12 10:38:10 I was... May 12 10:38:19 argh.. 10KB/s write speed on sd... killlllllllll May 12 10:38:34 hrw: ouch :-( May 12 10:38:45 eek. nasty. May 12 10:38:53 hrw: Maybe there is something wrong with your clocks for that card... May 12 10:39:04 That would explain the timeouts May 12 10:39:05 * mickeyl hoping for his new guitar to arrive hopefully tomorrow May 12 10:39:08 RP: who knows.. May 12 10:39:10 that would save the long weekend May 12 10:39:56 ok. time to clean house a bit - my younger btoher arrives tomorrow with his girlfriend May 12 10:40:10 I should have learnt to play an instrument when I had the chance at school... May 12 10:40:15 have fun hrw :) May 12 10:40:23 or even my older younger brother as I have two brothers May 12 10:40:26 I'm musically illiterate :) May 12 10:40:45 RP: It's never too late to learn an instrument May 12 10:40:51 hrw: don't remind me about cleaning. I really need to do some of that :-/ May 12 10:41:00 i should learn something more. i can play ukulele, mandoline, guitar, and piano. May 12 10:41:01 I can play music.. but I need music player for this ;) May 12 10:41:03 heh May 12 10:41:20 mickeyl: I'm now of the opinion I'll stick to what I can do well :) May 12 10:41:29 hehe fair enough May 12 10:41:46 i wish i knew kernel stuff like you do May 12 10:42:13 I am going to spam the oe list with some patches for omap osk board support May 12 10:42:22 then again, it would increase my potential places to hack and i dunno if that's a good thing :)) already doing to much May 12 10:42:26 Crofton: very good May 12 10:42:32 they aren't perfect, but they will help get some things sorted out May 12 10:42:37 mickeyl: On the other hand, your general programming knowledge is much greater than mine ;-) May 12 10:42:59 and some packages we need for our work, mainly omniORB, omniEvents, and XERCES-c May 12 10:43:15 oh good. i gave up on omniorb after struggling with it for some hours May 12 10:43:22 cool May 12 10:43:36 i'll add omniorbPy then May 12 10:43:39 I was motivated, the other ORB I tried really pissed me off May 12 10:43:45 that would be very cool May 12 10:44:21 I'm not 100% certain they are exactly correct yet, but some people have been asking me for this stuff May 12 10:44:22 i hacked on MICO when i was doing my diploma. it has been developed in the research group I was part of May 12 10:44:38 I've heard of that. Not sure why we didn;t use it May 12 10:44:55 Some guys have delusions of running stuff on windows :) May 12 10:44:55 omniorb is the best multipurpose ORB out there - good choice May 12 10:45:03 that's good to hear May 12 10:45:37 The omniORB guy says to use the "old style" make for cross, but I hacked the configure way to work May 12 10:46:10 nice. perhaps we can send some of these patches upstream May 12 10:46:16 mickeyl: In some ways, having a narrower skill set does help focus your effort. I'm trying my best not to get hooked into the non kernel stuff :) May 12 10:46:34 collie booting.. May 12 10:46:45 hrw: What were you using to measure the SD speed? May 12 10:46:57 RP: that's very true May 12 10:46:58 RP: copied two zbedic files using mc May 12 10:47:21 hrw: aha, another mc user :) May 12 10:47:28 I like mc... May 12 10:47:32 * zecke uses mc May 12 10:47:55 I hate mc too often but it does not have alternative... May 12 10:47:56 hmm May 12 10:48:04 i'm not impressed by the magneto review May 12 10:48:06 I was brought up on PCs using a program called Xtree which is very similar in idea May 12 10:48:08 looks ugly as ever May 12 10:48:34 I just wish mc has some of xtree's features... May 12 10:48:48 I wish mc has better ftpfs... May 12 10:49:22 using collie after c7x0 is... May 12 10:49:30 at least I can browse ipk with mc May 12 10:52:42 mickeyl: magneto is more than you see May 12 10:52:57 mickeyl: did you recently fixup the E stuff? ewl_*.bb refers to a fix-configure.patch that didn't seem to make it into bk May 12 10:53:00 mickeyl: ask yourself why microsoft rescheduled its release and you see no gui changes? May 12 10:53:10 mickeyl: they did something with the codebase in that time May 12 10:53:35 jh_: yeah, i fixed everything E. hold on May 12 10:53:39 zecke: right. May 12 10:54:40 i guess what the patch was and it works fine after that :) May 12 10:54:49 /guess/guessed/ May 12 10:54:59 03mickeyl 07 * r1.3269 10openembedded/packages/ewl/ewl/fix-configure.patch: add forgotten patch for ewl. bk sfiles -x rules *cough* May 12 10:55:00 hmm.. too old zbedic on collie May 12 10:55:14 thanks :) May 12 10:55:29 jh_: yeah, that odd AC_C_ACCESSOR. didn't know what that was for so I just removed it heh May 12 10:56:15 cool. still looking forward to Epicenter ;) May 12 10:56:40 ups May 12 10:56:41 right May 12 10:56:47 after I received my Ph.D. though May 12 10:56:56 so... not before end of 2006 May 12 10:57:12 mickeyl: hurry up May 12 10:57:16 heh. cool, maybe raster will be done by then May 12 10:57:26 zecke: i will. May 12 10:57:39 zecke: there's a lot of research fund money around for us :D May 12 10:57:51 hopefully still then May 12 10:57:59 mickeyl: hurry, hurry, hurry May 12 10:58:22 no do it carefully and right(tm) May 12 10:59:43 jh_: epicenter? May 12 11:00:16 the codename for a yet to be started palmtop environment based on the enlightenment foundation libraries May 12 11:00:56 hmm, didn't see that in e-devel yet May 12 11:01:14 ~lart opie networksettings for bad broadcast May 12 11:01:15 * ibot takes a big bite out of opie networksettings's jugular vein for bad broadcast May 12 11:01:36 koen: yeah, because yet to be started by me. May 12 11:01:38 heh May 12 11:01:51 the day E is written in a modern language I'll join May 12 11:02:05 EFL++ May 12 11:02:17 i released the preview code to two people interested in working on it May 12 11:02:31 zecke: what's so modern about c++? May 12 11:02:43 koen: did I say C++? May 12 11:03:01 no, but the little voice in my head said it May 12 11:03:31 you can use ruby or c# if you want May 12 11:03:37 eeek May 12 11:03:52 he said "modern" May 12 11:03:55 koen: modern := { l \elem Language | supports using modern object orientated concepts easily } May 12 11:03:55 not "new" May 12 11:04:06 ada? May 12 11:04:11 koen: and yes c++ allows that but it has a lot of pitfalls May 12 11:04:29 koen: so everything not using gobject and got sane classes could be named modern... May 12 11:04:43 gobject? May 12 11:05:14 koen: look at how much more productive people are using C# and Gtk# May 12 11:05:17 koen: instead of C May 12 11:05:35 zecke: I know, even I have c# patches in gnome-cvs May 12 11:06:49 RP: do you've time for a quick test? May 12 11:07:06 did anyone have a look at the armv5te vs armv5e remarks on the mailing list? May 12 11:07:10 :)) May 12 11:07:24 zaurus said that my dlink is socket connect... May 12 11:08:46 koen: I would even call Java modern and good May 12 11:08:54 brrrr May 12 11:09:02 modern yes, good ? now May 12 11:09:06 no, even May 12 11:09:22 mickeyl: not for GUI, but ok for developing and beginners May 12 11:09:27 hmm May 12 11:09:34 mickeyl: huge classpath, not too many pitfalls May 12 11:09:46 we had a hard time teaching beginners how to grok May 12 11:09:50 hmm.. its not zaurus - its opie May 12 11:09:51 * koen looks at the "visual programming in Java" books on his shelf May 12 11:09:57 public static void main() May 12 11:09:59 instead of May 12 11:10:01 print "hallo" May 12 11:10:09 mickeyl: yeah you're right about that May 12 11:10:16 mickeyl: www.bluej.org is a good work around May 12 11:10:22 mickeyl: and my Java Console as well May 12 11:10:34 mickeyl: '10 PRINT "hallo"' May 12 11:10:41 20 GOTO 10 May 12 11:10:53 20 RUN May 12 11:11:12 6510 asm was fun May 12 11:11:14 tigrux-enchilada: get me some nachos May 12 11:11:22 mickeyl: z80 asm is more fun May 12 11:11:38 (as far as asm counts as fun) May 12 11:11:47 yeah May 12 11:11:48 6510? you ... C= user May 12 11:11:56 pic16 asm is fun May 12 11:12:04 lol why doesn't gnome-session start gnome panel... May 12 11:12:12 68K was like a high level programming language compared to z80 and 65xx May 12 11:13:23 koen: OT: would you call the dutch healthcare system good? May 12 11:13:29 bother patch got stopped for moderation May 12 11:14:02 Crofton: are you subscribed? if not, do that and send it again, please May 12 11:14:17 zecke: I've been outside the public healthcare system most of my life, and it work relatively good May 12 11:14:19 I am, it is 55K and message size is 40K May 12 11:14:54 I think Kergoth is mod May 12 11:14:58 Crofton: send a link to it then May 12 11:15:00 zecke: moving to .nl? May 12 11:15:15 yeah, I need to turn on httpd on this machine May 12 11:15:32 thttpd rule May 12 11:17:50 cu May 12 11:21:32 RP: http://page.mi.fu-berlin.de/~freyther/bitbake_patches.tar.bz2 May 12 11:21:55 RP: I will publish that soon, it would be nice if you could test May 12 11:26:21 zecke: sorry, was busy coding :) May 12 11:26:26 zecke: I'll give that a test May 12 11:26:33 RP: it is a quilt directory May 12 11:26:41 do quilt push until you're done May 12 11:30:23 zecke: That's quite a neat way to do it :) May 12 11:31:01 RP: saves the xx in front of the file May 12 11:37:44 zecke_: I've set it away... :) May 12 11:39:41 25%, memory usage looks stable :) May 12 11:47:03 OK links to patches on the way May 12 11:51:05 Crofton: what patches? May 12 11:51:42 for OSK support May 12 11:51:49 omap 5912 osk May 12 11:54:05 zecke: Its started building and is looking good... May 12 11:54:27 zecke: I'll just leave it running and see if it has any problems... May 12 11:56:19 good May 12 11:56:32 I'll reply to your mail on bitbake devel now May 12 11:56:41 yay for Crofton May 12 11:57:43 Heh, we'll see. I'm sure I have a few more bits that are broke yet to find May 12 11:58:08 But I needed to clean up my local stuff before I completely forgot what I was doing .... May 12 11:58:51 I've burned myself a couple of times editting files without telling bk .... May 12 12:06:57 Crofton: Permission denied May 12 12:08:12 hmmm May 12 12:08:51 crap May 12 12:09:49 s/www/omap-dev/ May 12 12:24:06 evening May 12 12:24:59 hey pb_ May 12 12:25:05 'night May 12 12:25:08 hey SirFred May 12 12:26:37 03pb 07 * r1.3271 10openembedded/packages/mythtv/mythtv_0.16.bb: correct "--enablee" typo in mythtv_0.16 May 12 12:27:27 03pb 07 * r1.3272 10openembedded/packages/mythtv/mythtv_0.17.bb: update mythtv to 0.17 May 12 12:29:52 03pb 07 * r1.3273 10openembedded/packages/xserver/ (xserver-xorg/imake-installed.patch xserver-xorg_6.8.99.5.bb): update xserver-xorg to 6.8.99.5 May 12 12:30:54 03pb 07 * r1.3274 10openembedded/packages/dri/drm-kernel_cvs.bb: use linux-core directory instead of linux May 12 12:31:13 pb_: dvb season? May 12 12:32:03 yah May 12 12:32:29 my mythtv frontend box crashed, and I decided to make a new image for it May 12 12:32:55 03pb 07 * r1.3275 10openembedded/packages/xinerama/xinerama_1.0.2+cvs20050505.bb: new xinerama snapshot May 12 12:56:57 re May 12 12:57:15 hi zecke May 12 14:04:45 evening May 12 14:04:50 hi May 12 14:05:06 moin mickeyl May 12 14:08:10 03koen 07 * r1.3267.1.1 10openembedded/packages/kanatest/ (kanatest_0.3.6.bb kanatest.inc files/kanatest.desktop): Add kanatest_0.3.6.bb, kanatest.inc, kanatest.desktop, courtesy Robert Marcano May 12 14:31:38 zecke: bitbake died - AttributeError: 'NoneType' object has no attribute 'endswith' May 12 14:35:04 RP: anymore output? which package? May 12 14:37:04 zecke: I'd lost the scrollback - I'm rerrunning it May 12 14:38:42 zecke: usbutils-0.70 May 12 14:39:16 RP: not too interesting ;) May 12 14:39:30 I'm know sending the files anyway :} May 12 14:39:38 * zecke is currently writing the mail May 12 14:39:59 endswith... May 12 14:40:05 probably something expecting a string, but got a None May 12 14:40:48 mickeyl: That's what it looked like... May 12 14:41:24 hmm only base.bbclass uses endswith May 12 14:47:42 zecke: http://www.rpsys.net/openzaurus/temp/rp1 May 12 14:50:10 Of course looking at that log, its using modules from /usr/oe/bitbake :-/ May 12 14:50:34 it should be using modules from bitbake3 which is the hacked version... May 12 14:52:21 I see it myself here May 12 14:53:47 I suspect an empty string became a "None" somehow... May 12 14:55:22 03mickeyl 07 * r1.3269.1.1 10openembedded/ (4 files in 2 dirs): remove opie-mobilemsg, it has been removed upstream also May 12 14:59:36 g'night May 12 14:59:43 'night mickey|zzZZzz May 12 15:18:39 RP: sent May 12 15:18:48 now I'll look at usbutils May 12 15:21:25 zecke: sweet May 12 15:21:27 zecke: good job May 12 15:27:23 zecke: The email sums up the position nicely May 12 15:28:20 CosmicPenguin: Well I finally found the time to actually finish cleaning this code up May 12 15:28:48 Now we really need kergoth to have a look at the patches... **** ENDING LOGGING AT Thu May 12 15:28:53 2005 **** BEGIN LOGGING AT Thu May 12 15:30:10 2005 May 12 15:30:11 zecke: is the code in the mainline? May 12 15:30:27 well not yet May 12 15:30:49 RP: weird it is someone calling the fetcher with an empty dict... May 12 15:33:02 g'day treke May 12 15:36:39 zecke: very weird... May 12 15:39:45 RP: well May 12 15:39:57 fetcher.py uses 'd' May 12 15:40:13 it is either a global 'd' when running a task May 12 15:40:18 or bb.data.init() May 12 15:40:30 so majorily f*cked up May 12 15:41:40 At least we know where the problem is... May 12 15:42:29 that is a guess May 12 15:42:45 and I don't know how it works for the other tasks May 12 15:43:11 wow hotshot2calltree finished May 12 15:47:46 CosmicPenguin: just need to write expire code now :p May 12 15:50:11 treke: sweet May 12 15:50:37 I need to give zecke's code a quick check too May 12 15:50:42 in fact - no time like the present May 12 15:53:43 RP: it is the fetcher May 12 15:55:34 zecke: what is the fetcher? May 12 15:56:49 pb_: fetcher.py May 12 15:56:55 lib/bb/fetcher.py May 12 15:57:10 oh, right May 12 15:57:16 yes, that is indeed the fetcher May 12 15:57:17 heh May 12 15:57:19 def foo(d = bb.data.init()): May 12 15:57:37 and while executing the functions d is a global variable... May 12 15:57:44 or bb.data.init()... May 12 15:58:30 mm, yeah. I'm not quite sure why that bb.data.init() is there. May 12 15:58:48 iirc, the fetcher is only ever called from inside do_fetch(), and d always exists in that context. May 12 15:59:17 it is called in one other place May 12 15:59:21 I started writing some unit tests for bb.data a while back May 12 15:59:22 its called in a couple places. May 12 15:59:24 maybe I should post those. May 12 15:59:26 anyone who needs localpath() May 12 15:59:38 kergoth: more than we already have May 12 16:00:03 the 'default argument' d is a complete mess May 12 16:00:18 kergoth: does localpath get called outside of task context? May 12 16:00:38 I guess it could do, theoretically. May 12 16:02:12 I've been doing it May 12 16:02:34 ah. May 12 16:02:54 clearly treke.net is the place where theory and practice collide. May 12 16:03:06 now I can't guarantee what I've done is in any way a legal usage of the bb libraries May 12 16:03:11 but the code works right now :p May 12 16:03:24 heh May 12 16:03:39 I didn't notice any documentation saying "don't use this" :p May 12 16:04:06 well, given the state of the bitbake documentation, "it works right now" is probably the best definition of legal usage. May 12 16:24:38 'night all May 12 16:48:58 RP: good nite May 12 17:04:54 Are the suspending issues specific to 2.6.12? May 12 17:24:42 ooops May 12 17:24:54 nothing is handling the case of a partial download :p May 12 17:25:39 so I guess that's next May 12 17:40:24 darn it May 12 17:41:32 oe doesnt seem to be able to tell me that wget died May 12 17:42:33 ibot CP May 12 17:42:34 i heard cp is Copies one or more files to another file or directory. Syntax: cp (file1) (file2) ...(fileN) (destination). Where file1 through fileN are the files to copy, and destination is the destination file or directory. Example: cp ../frog joe copies the file ../frog to the file or directory joe. May 12 17:42:41 ibot CosmicPenguin May 12 17:42:42 well, cosmicpenguin is Jordan Crouse or jordan.crouse@amd.com (work) or jordan@cosmicpenguin.net (home) May 12 19:03:06 kergoth May 12 19:05:47 I want to usr my local arm-linux-gcc. how can i do May 12 19:06:38 I want to use my local arm-linux-gcc. how can i do May 12 19:07:16 oh, dumbness. The filename for the tarball on upstream's web server is "fooblah.bz2". so bitbake's fetch uncompresses the bz2, but doesn't figure out that it should extract the tarball. May 12 19:09:25 keturn: heh, bitbake's fetcher goes by filename.. not much you can do about that other than bitch at upstream, or write a custom fetch routine, or something May 12 19:10:06 kergoth:I want to usr my local arm-linux-gcc. how can i do May 12 19:10:12 kergoth:I want to use my local arm-linux-gcc. how can i do May 12 19:10:19 yes, repeating helps May 12 19:11:46 robert: i heard you the first time. stop repeating or you'll be kicked. May 12 19:12:58 So can I override oe_unpack_file for just this .bb? May 12 19:13:19 I write wrong word ."usr" "use". May 12 19:13:33 ? May 12 19:13:42 you did it wrong and corrected it not once, but twice May 12 19:13:52 i dont need to read it 4 times. May 12 19:13:53 oh .sorry May 12 19:14:27 i dont have my oe setup in front of me, but what you'll want to do is modify ASSUME_PROVIDED so that oe believes your system already has the c library and compiler May 12 19:14:43 soemthing like ASSUME_PROVIDED = "virtual/libc virtual/${TARGET_PREFIX}gcc" May 12 19:14:46 or similar May 12 19:17:15 thanks May 12 19:19:09 kergoth: got a few minutes? May 12 19:20:20 a few. leaving shortly. whats up May 12 19:20:50 http://openembedded.org/cgi-bin/moin.cgi/GettingStarted May 12 19:20:59 I've started to try a do a build May 12 19:21:13 just want to know if the info there is complete enough May 12 19:21:33 yes, thats the definitive place to get to the point where you can do builds May 12 19:21:37 uh ... wasn't me that put the comment about you btw May 12 19:22:01 should really lock that page so only oe devs can modify it May 12 19:22:13 I added a few things re: freebsd there May 12 19:22:19 nod May 12 19:22:22 should break it down May 12 19:22:31 add subpages for different build distros/os's May 12 19:22:34 lock the main one May 12 19:22:45 I changed a few lines that were confusing too May 12 19:22:48 cool May 12 19:23:06 (since I'm new at it, I could see it with new eyes) May 12 19:23:41 Create a dir to download bitbake to and cd to that dir (cd /stuff/bitbake) <= I think this line is wrong btw May 12 19:24:08 so I added the line below it, b/c if I did that, it'd create bitbake/bitbake May 12 19:24:43 I dont know if it is the same with linux May 12 19:25:20 for me, it had to be cd /stuff only May 12 19:25:21 i think the idea behind that was to create a single dir to hold bitbake, oe, and the build dir May 12 19:25:27 but /stuff is probablys ufficient for that May 12 19:25:32 so yeah, that is odd May 12 19:25:47 we really need an approval process for a document like that May 12 19:25:51 it doesnt belong in a wiki anymore May 12 19:25:57 ok May 12 19:26:11 you know, put it in a content management system or something. let people submit changes that get approved by us May 12 19:26:19 new users see from the best perspective, but they also dont understand the system as well May 12 19:26:21 that makes sense May 12 19:26:24 so the changes need checking May 12 19:26:37 I'm at the local.conf stage May 12 19:26:38 but yeah, thanks for helping fix it up May 12 19:26:49 dont quite understand all the options yet, but I'll figure it out May 12 19:27:36 I'll probably add more freebsd stuff as it comes up May 12 19:27:41 thanks for your time May 12 19:30:32 grr. I hate running into limitations of python that require me to jump into C for only like 3 lines of C code. grr. May 12 19:33:57 robert: see the asusoe.conf distro file for an example of how I did an external toolchain. May 12 19:36:39 ok May 12 19:37:18 okay. hacked around that fun by overriding FETCHCOMMAND and providing a localpath parameter to the SRC_URI. May 12 19:37:40 gb2: what 3 lines can't you do in Python? May 12 19:42:14 struct sock_filter s_filter[4] = { ... }; May 12 19:42:14 struct sock_fprog fp = { .len = 4, .filter = &s_filter; }; May 12 19:42:14 setsockopt(s, SOL_SOCKET, SO_ATTACH_FILTER, &fp, sizeof(fp)); May 12 19:46:27 python's socket code has a setsockopt(), which allows you to feed in something created using struct to create arbitrary c-like structs which translates to passing in a pointer to the struct when calling the syscall. May 12 19:46:53 but, attaching a socket filter requires a struct containing a pointer to the filter. and there's no way to do that that i've found. May 12 21:30:37 could someone recompile a prog for me? May 12 21:31:50 depends what it is May 12 21:33:40 i need pppd to do this: http://developerboards.cingular.com/cngddp/board/message?board.id=techsupport&message.id=952 for my ipaq's gprs card.. the sony erricson thinks the server is 127.0.0.2 so i get loopback error when trying to pppd call gprs May 12 21:34:05 has ppd been fixed yet? May 12 21:34:09 i tried making a patch but, im so confused May 12 21:34:12 ... probably May 12 21:34:15 i still get loopback detected May 12 21:36:23 hmm May 12 21:37:02 why is AT&T violating standards May 12 21:37:11 because they are evil May 12 21:38:16 opening 127 that way would seriously cause a major loss of security on your system May 12 21:42:19 did you read the rest of tho posts? May 12 21:42:22 the* May 12 21:43:06 they talk about modifying other files ,,, May 12 21:44:31 yea, i try all the ppp files but I still get loopback detected when i call gprs May 12 21:45:50 okay this is the one to follow May 12 21:45:53 http://www.eskimo.com/~roger/files/doc/howto-gc82.html May 12 21:46:10 even has the patch file for you May 12 21:47:26 he has a better solution as well May 12 21:47:45 you use 10.0.0.1 instead of opening up your loopback May 12 21:48:41 i take it you already know how to make bb build stuff? May 12 21:49:18 use bkvi to add file://attgprs.patch May 12 21:49:21 not reall I know in the bb file I add the patch under SRC_URI as file://pppd.patch but building it for sa ipaq.. not really; MACHINE in conf/local.conf ? May 12 21:49:21 or something May 12 21:49:56 you need to make sure the patch by that name is in the directory May 12 21:50:16 in the same directory as the bb? May 12 21:50:25 and as long as your already setup and have a working build system you should be good May 12 21:50:29 yes May 12 21:50:46 most of the patch dir are the bb name with a - not a _ May 12 21:51:02 I suppose I do, Im using the snapshot. /stuff .. May 12 21:51:22 and you have built something that works? May 12 21:51:30 never :-/ May 12 21:51:53 well task-bootstrap might be a good start to see if your system works May 12 21:54:13 bitbake task-bootstrap? May 12 21:54:36 yup May 12 21:54:48 i get ERROR: Nothing provides task-bootstrap May 12 21:55:57 then you have not set it up properly May 12 21:57:02 i exported BBPATH, setup conf/local.conf what else? May 12 21:57:35 BBFILES May 12 21:58:28 you rock emte May 12 21:58:32 thanks May 12 21:58:37 a lot May 12 21:59:08 handling bitbake files :D May 12 22:16:00 will bitbake make an ipk automatically or do i need to specifiy IMAGE_FSTYPES="ipk" ? May 12 22:16:12 eh? May 12 22:16:18 IMAGE_FSTYPES has nothing to do with ipks May 12 22:16:26 absolutely nothing May 12 22:16:41 ok May 12 22:16:41 IMAGE_FSTYPE is the filesystem type emitted when creating an image. like cramfs, ext2, jffs2, tar May 12 22:16:51 has nothing to do with packaging May 12 22:16:52 ah May 12 22:17:08 to emit ipks, either set DISTRO to a distribution that uses them, or add package_ipk to your INHERIT May 12 22:19:07 hmm May 12 22:19:16 kergoth, tell me whats wrong with my loop May 12 22:19:57 * emte waits for kergoth to say i wrote it May 12 22:20:07 hehe May 12 22:20:22 for (cH = 0; arrayName[cH] != '\n'; cH++) May 12 22:20:23 { May 12 22:20:23 printf("\n Input your [%i] entry: ", cH); May 12 22:20:23 scanf("%lf", &arrayName[cH]); May 12 22:20:23 } May 12 22:20:32 its just not exiting properly May 12 22:21:20 it stops on an extra return but sorta hangs in nomans land May 12 22:31:11 http://www.pdalive.com/showarticle.php?threadid=8532 May 12 22:31:57 ooo i want one May 12 22:33:52 probably be over 1k May 12 22:34:33 looks like a neat keyboard, but might be hard to type May 12 22:35:12 yeah i liked it until it said its also a phone May 12 22:35:26 that alone doubles the price here May 12 22:47:50 kergoth: ping May 12 22:48:31 treke|home: pong May 12 22:49:16 is it actually possible to build two packages that both use the same filename upstream? May 12 22:49:26 as a src_uri May 12 22:50:05 hmm i cant push ... May 12 22:50:12 guess i'll email May 12 22:50:16 treke|home: heh, only if the file really is the same. May 12 22:50:32 hehe May 12 22:50:36 treke|home: only real workaround for that would be to change DL_DIR to include the package name or something, but then you couldnt share the downloaded file between packages. May 12 22:50:39 so .. May 12 22:50:41 tradeoff May 12 22:50:45 kergoth: just noticed that while I was doing the mirror stuff May 12 22:50:53 :p May 12 22:51:24 ah May 12 22:51:41 nothing we've run into yet, so not a big deal May 12 22:51:54 probably just going to store the files under the full url May 12 22:52:26 treke|home: i could see hacking do_fetch to store the md5 of the downloaded file, and its downloaded path, in a single download digest. then include PN in DL_DIR. then look it up in the digest before downloading to see if anyone else already has downloaded it. May 12 22:52:41 that would solve it without having to download stuff multiple times May 12 22:52:51 but that obviously adds a fair bit of additional complexity May 12 22:52:58 yeah May 12 22:53:28 I'm completely bypassing bitbake for files that are already downloaded May 12 22:54:13 which is really the main reason I'm not just using bitbake as written May 12 22:54:22 ah May 12 22:56:55 with the expiring, i think I'm going to have it move expired source offline, but not delete it until it's no longer being referenced in a bb May 12 22:57:16 that way people get an error, but we can always mirror it elsewhere permenantly May 12 23:07:55 expiring? May 12 23:09:05 kergoth: what's the PREFERRED_PROVIDERS lines all about? May 12 23:09:21 read the bitbake manual May 12 23:09:26 ok May 12 23:11:57 note: link doesn't work May 12 23:12:03 http://openembedded.org/cgi-bin/moin.cgi/GettingStarted May 12 23:12:13 We recommend reading the BitBake users manual (you can build the docs in bitbake/doc/manual, or go to http://bitbake.berlios.de/manual/) for background information May 12 23:12:47 looks like berlios is having problems May 12 23:12:57 oh. ok May 12 23:13:43 in a .bb file, oe use SRC_URI to get the package.How to make oe to find a package that is exist? May 12 23:14:04 i dont understand what you're asking May 12 23:14:34 you mean use _local_ sources instead of downloading them? May 12 23:14:42 you can use file:// urls in SRC_URI May 12 23:14:46 I mean I dont want to download a package from net May 12 23:14:54 yes May 12 23:15:40 then use a file:// url in the SRC_URI line May 12 23:16:54 like file:///root/*.tar.bz2 ? May 12 23:17:00 yes May 12 23:17:10 oh .thanks **** ENDING LOGGING AT Thu May 12 23:59:56 2005