**** BEGIN LOGGING AT Mon May 30 23:59:57 2005 May 31 00:14:52 morning everybody! May 31 00:39:28 | skin.cpp:164: error: `Resource' has not been declared May 31 00:39:29 | skin.cpp:164: error: `findPixmap' undeclared (first use this function) May 31 00:39:29 | skin.cpp:164: error: (Each undeclared identifier is reported only once for each function it appears in.) May 31 00:39:32 | make: *** [qpeobj/skin.o] Error 1 May 31 00:39:35 | FATAL: oe_runmake failed May 31 00:39:37 :( May 31 00:42:04 Anyone an idea on this? Full error is at http://ascent.student.utwente.nl/~ascent/error May 31 00:43:00 morning May 31 00:43:22 howdy May 31 00:49:39 morning all May 31 00:49:57 ascent: thats umbaumba error May 31 00:50:07 a _what_ ? :) May 31 00:50:17 never heard of umbaumba before May 31 00:50:17 NOTE: package opie-mediaplayer2-1.2.0+cvs-20050531-r1: task do_compile: failed May 31 00:50:26 yea May 31 00:50:29 Always add info WHICH package failed and WHEN May 31 00:51:04 well, that's in the error ? May 31 00:51:10 umbaumba is my word for something which is unknown - in that moment because you didnt said which pacakge it was May 31 00:51:18 yes, sorry, I understand now :) May 31 00:51:31 ascent: and that error is upstream one - downgrade CVSDATE to 20050530 and retry May 31 00:52:01 Hm.. Can I do so using bk? May 31 00:52:11 cvs used -r for a date iirc May 31 00:52:47 ascent: add CVSDATE=YYYYMMDD in conf/local.conf May 31 00:53:25 Ah, nice! May 31 00:54:45 Hm.. When I bk pull after that date, it sayus nothing to pull May 31 00:57:09 thats flag for building not for metadata fetching May 31 00:57:27 ah okay :) May 31 00:57:38 I'll run the meta-opie again :) May 31 01:01:15 NOTE: Handling BitBake files: - (1946/2375) [81 %] May 31 01:01:30 That process takes a few minutes, will psyco speed that up, or can't it be sped up ? May 31 01:06:01 currently cant May 31 01:11:05 morning all May 31 01:22:23 hertog: hello local :) May 31 01:23:06 ascent: Ah, Enschede is represented out of proportion here :) May 31 01:23:24 Very good :) May 31 01:24:20 Twente! May 31 01:24:22 * koen hides May 31 01:24:28 :) May 31 01:24:29 brb May 31 01:31:30 morning May 31 01:31:39 hey proti May 31 01:34:45 guten morgen Herr Lauer May 31 01:36:08 Is the bitkeeper sync from oe-devel to openembedded in working order again? May 31 01:36:26 hi mickeyl May 31 01:36:33 mickeyl: can you sync bk? May 31 01:36:39 It usually involves bribing mickeyl May 31 01:38:03 Guten Morgen Koen :) May 31 01:38:14 you want a sync? you get a sync May 31 01:38:15 hold on May 31 01:38:40 now that's what I call service :) May 31 01:39:00 djeez, berlios is still down May 31 01:39:04 let's see if he forgets to turn off the verbosity again.. May 31 01:41:24 SYNC: done May 31 01:41:51 mickeyl: I didn't get what zecke was talking about yesterday. May 31 01:42:07 hrw|work: please remind me to pester drw about findPixmap not in OResource May 31 01:42:15 proti: me neither. i guess it was just late :) May 31 01:42:59 proti: your correction patch raises an interesting issue. somewhere there is either wrong or just ineffective code in bb. Will try to take a look when I have a minute May 31 01:43:16 for now I need to writeup a mail explaining how to use bitbake- i May 31 01:43:27 mickeyl: I suspect zecke put a ignore, because he does not seems to respond when I talk to him. May 31 01:43:48 mickeyl: What kind of issue ? May 31 01:43:48 proti: hah. i'll talk him into removing that May 31 01:44:07 proti: well... an issue in that it looks odd May 31 01:44:42 Be more precise, i can't figure it by myself now. May 31 01:44:53 i can't be more precise May 31 01:44:59 i don't know the code May 31 01:45:06 but your reasoning seemed valid May 31 01:45:21 (iterating over a string as opposed to a list of package names) May 31 01:46:24 Yes, I'm diving in the last part of the bitbake I don't know. I am thinking to auto-evaluate the working set based on the package to build. May 31 01:47:04 Just a closure of the dependencies that bitbake can cache. May 31 01:47:32 mickeyl: Ok for the string, I still don't know why it worked before. May 31 01:48:14 mickeyl: is bitbake -b known to be broken? May 31 01:49:48 koen: no longer in r235 May 31 01:50:24 cool May 31 01:50:36 just as you said that berlios came to life :) May 31 01:52:42 hehe May 31 01:54:34 hrw|work: 20050530 fails too for the mediaplayer, I'm trying a day before now. May 31 01:57:21 mickeyl: how hard would it be to add a "image too big for device" check to OE? May 31 01:57:32 an optional check, of course May 31 01:58:19 Anyone an idea on what date the opie-mediaplayer broke, using the bk://oe-devel.bkbits.net/openembedded/ repos ? May 31 01:58:27 best, use 1.2.0 May 31 01:58:29 koen: look how it is done in linux-openzaurus*.bb May 31 01:58:53 that's a kernel.bb May 31 01:58:55 koen: wouldn't be to hard. May 31 01:59:08 koen: see what i added to the zaurus machine confs May 31 01:59:11 ROOTFS_SIZE May 31 01:59:12 or so May 31 01:59:17 add this for all your machines May 31 01:59:27 and then I'll have a look at adding a programmatic check May 31 01:59:27 I have that May 31 01:59:38 koen: addtask checksize after do_image and then check rOOTFS_SIZE May 31 02:00:00 I was trying to find a way to do "IFSPACELEFTADD = "various apps" May 31 02:02:20 03koen 07 * r1.3453.1.1 10openembedded/conf/machine/nokia770.conf: nokia770.conf: Beginning of a machine.conf for the nokia 770, will need updating after figuring out the bootloader May 31 02:03:19 mickeyl: from the stories I hear, the bootloader and closed modules on the 770 are worse than on zaurus/2.4 May 31 02:06:33 hi people May 31 02:06:54 hey zecke & Arjan May 31 02:07:13 does anybody have the problem with usb ethernet gadget that after some time it just stops working? especially with data flowing? May 31 02:08:08 it seems that from 29 to 30, there we're some major changes :) May 31 02:08:15 it's downloading new old stuff like crazy :) May 31 02:11:47 hi zecke, Arjan May 31 02:12:49 hi hrw|work May 31 02:12:54 mail sent May 31 02:13:08 koen: ugh. i don't like to hear that May 31 02:13:55 i hope they ship it soon though May 31 02:13:57 i want one :) May 31 02:14:41 hi guys May 31 02:14:54 mickeyl: my bitbake failed overnight - similar problem May 31 02:15:16 mickeyl: it tried to do tslib from CVS, failed to fetch/unpack it and died. May 31 02:15:28 why did it fail? May 31 02:15:40 eh, just dug around in my 'old obsolete hardware chest' guess what I found :) a Philips Nini and Velo :) May 31 02:15:44 running bitbake -b openembedded/packages/tslib/tslib_cvs.bb allowed it to start building but it failed to compile May 31 02:16:04 how does it fail to compile? May 31 02:16:41 * koen has a feeling OE support for the hx4700 will improve soon May 31 02:16:44 Traceback (most recent call last): May 31 02:16:46 File "/home/ian/projects/openembedded/stuff/bitbake/bin/bitbake", line 767, in ? May 31 02:16:46 cooker.cook( args[1:] ) May 31 02:16:46 File "/home/ian/projects/openembedded/stuff/bitbake/bin/bitbake", line 631, in cook May 31 02:16:51 followed by more similar... May 31 02:17:26 ah well May 31 02:17:32 you need to post the error May 31 02:17:40 otherwise i have no chance to help May 31 02:18:09 s/error/complete error/ May 31 02:18:15 mickeyl: Im not sure where the error is in all the output :) May 31 02:18:22 hang on May 31 02:18:43 use pastebin May 31 02:18:46 post it completely May 31 02:19:25 http://pastebin.ca/13011 May 31 02:19:32 thats from bitbake -b .... May 31 02:20:02 eek May 31 02:20:12 that's a nasty one May 31 02:20:20 let me try to reproduce that May 31 02:20:47 :) May 31 02:21:02 (thats with the bitbake exec you sent me last night btw May 31 02:21:04 ) May 31 02:21:31 NOTE: package tslib-0.0cvs20050528-r18: task do_package: completed May 31 02:21:43 bummer, it works here :/ May 31 02:21:50 which python do you use? May 31 02:21:53 version May 31 02:21:54 http://pastebin.ca/13012 May 31 02:22:06 thats the earlier failure of bitbake task-bootstrap May 31 02:22:06 Maybe a bitbake --version to check if core match the bitbake ? May 31 02:22:09 hang on May 31 02:22:30 bash-2.05b$ bitbake --version May 31 02:22:31 BitBake Build Tool Core version 1.3.0, bitbake version 1.3.1 May 31 02:22:34 no i mean python May 31 02:23:14 try bitbake -c clean tslib; bitbake -b May 31 02:23:18 python 2.4 May 31 02:23:29 like i said, if it can't download a source file it may confuse the stamps May 31 02:23:43 look out for a .md5 file in DL_DIR where the corresponding source file is missing May 31 02:23:47 remove that .md5 file May 31 02:24:19 bash-2.05b$ ls Download/bbsources/tslib_pubcvs.arm.linux.org.uk__200505 May 31 02:24:21 tslib_pubcvs.arm.linux.org.uk__20050517.tar.gz May 31 02:24:21 tslib_pubcvs.arm.linux.org.uk__20050518.tar.gz May 31 02:24:21 tslib_pubcvs.arm.linux.org.uk__20050531.tar.gz May 31 02:24:42 looks like 0530 is not there May 31 02:24:43 same error after a clean May 31 02:24:47 so bb is correct in that May 31 02:25:14 yes, it appears that if you run bitbake task-bootstrap it fails to fetch it May 31 02:25:16 which md5 files are there ? May 31 02:25:28 if you run -b /path/to/tslib it works ok May 31 02:25:42 mickeyl: cvs checkouts does not give md5 files May 31 02:25:47 for tslib there is no md5 May 31 02:26:02 ah May 31 02:26:49 well it's hard for me to debug when i can't reproduce it May 31 02:26:58 something goes horribly wrong at your end May 31 02:27:24 mickeyl: are you around later today? May 31 02:27:29 you didn't rename your oe directory or use many softlinks? May 31 02:27:34 yeah, most likely May 31 02:27:39 I will see if I can get you a ssh login here May 31 02:27:44 k May 31 02:27:53 can you email me your pubkey ? May 31 02:27:58 sure. which address? May 31 02:28:04 spyro@f2s.com May 31 02:30:13 sent May 31 02:31:42 mickeyl: Nice feature. I think I'll try it soon. May 31 02:31:51 hmm.. switched to utf8 here... and it gives me more problems May 31 02:31:59 proti: :) May 31 02:32:08 16 bit languages must die :) May 31 02:32:25 Spyro: I use 8bit lang May 31 02:33:09 Spyro: 1 billion in china. Don't know if indian language is 16bit. But you'll have to kill a good share of earth'population. May 31 02:33:30 yay! genocide here we come ;-) :-) May 31 02:33:59 Spyro: Do you realize we (8bit languages) could be a minority ? :) May 31 02:34:04 heh that miliard/billion thing... here bilion is 10^12 not 10^9 May 31 02:34:21 proti: but we have too many 8bit langs May 31 02:34:41 at least you dont have to design horrible parsers for 8 bit languages May 31 02:35:20 besides the world pop is 6000 million, of which ~ 1-2 000 million are 16 bitters :) May 31 02:35:35 hrw|work: I propose we have 4bit language, it'll be enough and you can put 2 letters per byte. Yeah ! May 31 02:35:53 proti: youre sick :) May 31 02:36:00 be back in a bit, TTYL :) May 31 02:37:38 for me it will be enough is gvim will recognize iso8859-2 files as non-utf8 ones automatically May 31 02:39:03 hrw|work: I remember 2 chars per byte tricks was used in games to compress the text. May 31 02:39:43 proti: I remember awesome games made in text modes on 8bit Atari - it was hard to believe that this is text mode May 31 02:40:18 hrw|work: One name to mention ? May 31 02:40:21 grm. Also mediaplayer for 29th doesn't compile. What was the last successfull compilation date of that package? Anyone ? May 31 02:40:46 proti: Robbo, Lasermania, Fred, Misja (Mission Shark) May 31 02:40:52 ascent: 0520 May 31 02:41:03 ascent: I didnt build newer May 31 02:41:09 Ok. Let's try that, thanks. May 31 02:41:13 ascent: s/build/tried May 31 02:41:15 ascent: is there any particular reason why you don't build 1.2.0 ? May 31 02:41:27 there haven't been features since May 31 02:41:35 and this version is known to build May 31 02:41:45 mickeyl: in the normal checkout, bk://openembedded.bkbits.net/openembedded, there's a file 404'ing which doesn't exist anymore May 31 02:42:01 I got the tip to try devel from bk://oe-devel.bkbits.net/openembedded/ May 31 02:42:18 And that solved the problem with the non-existing file, but gave me an error with this mediaplayer May 31 02:42:19 ~lart me or gvim for very poor utf8 handling May 31 02:42:21 * ibot drops a baby grand on me or gvim for very poor utf8 handling May 31 02:42:39 mickeyl: tximage-0.2.tar.gz is the one 404'ing May 31 02:42:49 ascent: not anymore May 31 02:42:52 And that file is no where to be googled :) May 31 02:42:57 oh, got fixed this morning ? May 31 02:43:12 yesterday May 31 02:43:22 I bailed out on me last night. May 31 02:43:38 ChangeSet@1.3448, 2005-05-30 13:52:37+02:00, hrw@marcinj.local May 31 02:43:38 tximage, knights: source are now mirrored on openzaurus.org - closed #31 May 31 02:43:41 But, I'll try a pull again soon, right after the 0520 May 31 02:43:43 ascent: it just build like a charm here May 31 02:43:56 s/build/fetched May 31 02:44:06 Okay. Well, then I have to see to get my openembedded back to the regular instead of the oe-devel :) May 31 02:44:18 Should I just delete the whole openembedded dir and clone it ? May 31 02:44:27 or is there a nicer way? May 31 02:45:01 ascent: bk parent May 31 02:45:18 ascent: do not clone or we will punish you May 31 02:46:10 eeks :) Why is that? :) May 31 02:46:13 ascent: both knights and tximage are fetchable and buildable now May 31 02:46:18 cloning locks repo May 31 02:46:26 oh :) May 31 02:46:37 hm, so parent it is. Let's look up the syntax May 31 02:47:12 (or is just running 'bk parent' in my openembedded dir enough?) May 31 02:47:16 In unified format there is this line : '@@ -235,6 +234,65 @@' What means the -235,6 and the +234,65 ? May 31 02:48:06 remove 6 lines starting from 235 and add 65 starting from 234? May 31 02:48:12 ya May 31 02:48:54 substitute a 6 line long block starting in line 235 with a 65 line long block starting in line 234 May 31 02:49:57 Hm, just bk parent shows the normal (non-devel) now. but when I bitbake meta-opie, it builds loads of cvs things. How should I revert that ? May 31 02:50:20 do bitbake gpe-image May 31 02:50:31 ascent: you dont understand how OE works maybe.. May 31 02:50:48 mickeyl: In this hunk, there is only + lines, no -. May 31 02:50:57 hrw|work: not fully yet :) and the wiki gave me the basics, but not enough for this I think May 31 02:51:10 proti: yeah May 31 02:51:18 substitute a 6 line long block starting in line 235 with a 65 line long block starting in line 234 May 31 02:51:26 no need for '-' May 31 02:51:33 because essentially it gets only longer May 31 02:52:32 Ok. But how can patch do a '-r' if the removed lines are not there ? May 31 02:52:46 heh.. which editor other then gvim will fast handle 9M sql dumps? May 31 02:52:55 hrw|work: grep. May 31 02:53:03 proti: I meant editor May 31 02:53:08 sed May 31 02:53:20 proti: why removed ? May 31 02:53:29 proti: a 6 line long block May 31 02:53:34 means 2 * 3 lines of context May 31 02:53:37 nothing to be removed there May 31 02:53:43 see man diff for more details May 31 02:53:55 hrw|work: Since older versions of MS-Word used to make such big files, I guess it can handle it ;) May 31 02:53:57 * Hertog ducks May 31 02:54:02 or rather info diff May 31 02:54:15 mickeyl: or even 'pinfo diff' May 31 02:54:35 mickeyl: I did. info diff == man diff. But I think I get it, I have to take the context line in the count. May 31 02:58:47 hrw|work: What do you want to do in your dumps ? May 31 02:59:42 proti: I have to go through it and get some parts to update other base May 31 03:00:10 proti: in other words - I have to get Audi A8 related info from test server and put on audi.pl database May 31 03:00:23 hrw|work: Ever considered a throw-away perl script to do it ? May 31 03:01:15 proti: its not so simple usually ;( May 31 03:02:19 Then if it a long work, go for xemacs. It is really good at it. May 31 03:02:41 proti: after years with vim going to emacs? strange May 31 03:03:54 I would probably move to kate faster May 31 03:03:55 hrw|work: I use both. I prefer emacs for long work with multiple files and vim for mono edition and quick correction (no more than 2 files). May 31 03:04:10 I used gvim with >500 files loaded May 31 03:05:50 I'm not in a contest, I'm just saying the way I use my tools. I find shortcuts to navigate between files difficult and dangerous. I killed numerous buffers by doing ^w ^o instead of ^w ^p. May 31 03:06:57 sure May 31 03:07:08 So I gave up for multiple files. May 31 03:09:21 In xemacs, I especially like the possibility to put files side-by-side, and having multiple view of the same file by splitting the window. May 31 03:11:00 Plus, compilation in buffer, patch on steroid with ediff, and so on. May 31 03:12:20 last time when I started xemacs I have a problem with finding editor options there.. May 31 03:12:39 games, tools etc was easier to find then editor options May 31 03:14:01 env is too rich in xemacs, I agree with you. May 31 03:14:03 hrw|work: what do you want to edit for? May 31 03:14:15 hrw|work: playing games in your editor is much more fun May 31 03:14:25 XorA: ;) May 31 03:14:41 XorA: SVG tetris was enough May 31 03:24:04 unicode switching is good at home where I can have all in utf8 encoding. here (at work) I usually work with iso8859-2 encoded files (webpages) May 31 03:29:08 hmm.. what should GPE display after the New User screen? I managed to run it on my h1940. now everything works, but gpe just stops after I press OK on the New User screen. I get this black screen with the X cursor May 31 03:29:32 Wrong channel - try #gpe May 31 03:30:14 Arjan: http://handhelds.org/~gpe/gallery/view_photo.php?set_albumName=themes-and-art&id=land_23789 May 31 03:30:43 koen: thanks :) May 31 04:22:43 morning May 31 04:23:24 hi BigAl May 31 04:49:38 ~lart kspread authors for cvs 'export' May 31 04:49:38 * ibot executes killall -HUP kspread authors for cvs 'export' May 31 05:11:13 hey May 31 05:11:46 hey zecke_new May 31 05:12:56 re zecke May 31 05:16:07 wb mickeyl May 31 05:16:11 hi May 31 05:18:13 * zecke continues playing with a umts modem sample May 31 05:18:32 afternoon May 31 05:19:05 hi Rene May 31 05:19:35 hey Marcin May 31 05:35:19 zecke: ping May 31 05:36:20 ibot: ~lart zecke for putting ignore on proti and forgetting about it. May 31 05:36:38 ~lart zecke for putting ignore on proti and forgetting about it. May 31 05:36:38 * ibot whips zecke with a wet and grimy noodle just because for putting ignore on proti and forgetting about it. May 31 05:40:41 proti: buuh May 31 05:40:58 proti: you're lucky, irssi ignores you only on my home machine May 31 05:40:58 zecke: How do you do ? May 31 05:41:11 I had a nice visit at the dentist today May 31 05:41:20 zecke: The one you pinged me yesterday ? May 31 05:42:23 proti: Three things to summarise. Good to kill code duplication, bad to lose functionality, bad to emulate a Source Control Management (commenting stuff out instead removing it) May 31 05:42:57 proti: I think we should return the tuple (latest, preferred...) May 31 05:43:34 quadruple even May 31 05:43:44 mickeyl: n-tuple May 31 05:44:01 yes, n is always correct May 31 05:49:34 zecke: 1) lastest is used only for printing. Not used elsewhere May 31 05:50:47 2) As I said in my mail, I comment out the printf to remember it (i.e. decide wether removing the latest or modifying the function to return the latest too. May 31 05:52:41 3) What the code need the selected, not the history about what did what, how and when. I propose to add a function if noone can live without the latest version. May 31 05:54:08 1) it is used. removing functionality is not an option May 31 05:54:19 2) we've a SCM for that May 31 05:54:46 3) adding two functions does not kill code duplication? May 31 05:55:59 zecke: Not used. In most case, latest == preferred. In case of lastest <> preferred, bitbake just print it. May 31 05:56:18 If I read the code correctly. May 31 05:58:34 2) Sorry, I meant the patch was not final. I wanted you to have a look and give me feedback to what you prefered : 2 strings return value or 4 strings return value. May 31 05:59:39 proti: if by "just print it" you're referring to showVersions(), this is important functionality and you should not just arbitrarily decide to remove it. May 31 06:01:03 pb_: Correct. That's why I wanted to discuss about it. To me showversion is something like debug string (i.e. not in the common path). May 31 06:02:24 It's not in the common path, but it's user-visible functionality (i.e. it's for people maintaining metadata, not for people maintaining bitbake itself). May 31 06:03:21 proti: you annoy me. Either it is used then leave it, or it is not used then remove it and don't talk too much about it May 31 06:03:52 proti: see pb_ latest answer May 31 06:07:06 Ok, I wasn't aware of this feature. May 31 06:09:18 hi guys May 31 06:09:23 any updates to bb ? May 31 06:10:31 * proti will prepare another patch. May 31 06:12:15 Spyro: no, mickeyl refused to commit os.system('sudo rm -rf /') May 31 06:17:59 zecke: tch, mickeyl is no fun May 31 06:18:37 hi reenoo May 31 06:19:50 hey pb_ May 31 06:25:57 if we are to factor a common findBestProvider out, then just return 4 values May 31 06:26:03 one code path discards two of 'em May 31 06:26:08 the other code path uses all May 31 06:26:13 voila, problem solved May 31 06:26:21 right May 31 06:27:09 the additional overhead of computing and then discarding the extra pair should be negligible. May 31 06:29:52 * france is away: Away May 31 06:29:53 sure May 31 06:30:22 pb_: hi, whats the time in the UK? May 31 06:41:11 france: hello! it's 14:40 over here. May 31 06:44:29 france: hey May 31 06:44:31 mickeyl: ack May 31 06:45:18 pb_: thanks! as I run out the door. May 31 06:51:06 proti: could you do that? May 31 06:55:48 mickeyl: What ? The BestProvider thing ? May 31 06:56:20 Yes, I'm doing it at the moment. May 31 06:56:44 zecke: LOL :) May 31 06:58:31 hey Spyro May 31 06:58:38 you migrated to this channel now :P May 31 07:05:24 proti: cool. appreciating that. May 31 07:13:24 ~seen [cc]smart May 31 07:13:32 [cc]smart <~smart@gw.ptr-62-65-149-158.customer.ch.netstream.com> was last seen on IRC in channel #asterisk, 2d 23h 16m 43s ago, saying: 'not while doing.... while trying to initiate'. May 31 07:15:16 time to buy wrt54 May 31 07:15:18 hi May 31 07:19:14 uh oh, more zeckes May 31 07:22:52 wahhaaaa May 31 07:23:18 stop fork()-ing May 31 07:23:36 koen: I've clone() and vfork() May 31 07:24:01 stop those too :) May 31 07:24:57 SELinux or Solaris10? May 31 07:25:33 crumbs, talk about a rock and a hard place May 31 07:30:09 Spyro: pong May 31 07:34:10 ~change 419 pln to eur May 31 07:34:15 419.00 Zloty (PLN) makes 101.432 Euro (EUR) (from http://www.xe.com/) May 31 07:35:08 ~poodle May 31 07:35:11 methinks poodle is sharp sl-b500/5600, or a dog May 31 07:36:58 hrw|work: what do you want to buy? :) May 31 07:39:25 mickeyl: wrt54gs May 31 07:41:48 mickeyl: so finally my wifi card will get some use May 31 07:42:04 thx goes to [cc]smart May 31 07:58:19 This is what I feared. Something is wrong in pkg_pn. May 31 07:58:48 Here is an extract with pdb May 31 07:58:51 p pkg_pn['btxml'] May 31 07:59:07 ['/home/seb/dev/zaurus/openembedded/packages/btxml/btxml.bb'] May 31 07:59:18 (Pdb) p pkg_pn['gpe-clock'] May 31 07:59:18 [['/home/seb/dev/zaurus/openembedded/packages/gpe-clock/gpe-clock_0.14.bb', '/home/seb/dev/zaurus/openembedded/packages/gpe-clock/gpe-clock_0.20.bb']] May 31 07:59:54 Hum, I will hunt down this one. May 31 07:59:56 is that new? May 31 08:00:46 how do you teach OE that you have a 64bit kernel, but a 32bit userspace? May 31 08:01:57 Well when only one files provides the pn, files is a string, else, it is an array. May 31 08:02:59 pkg_pn[pn] is either a string or an array. Not good when one does for f in pkg_pn[pn] (either f is a string or a character of the string). May 31 08:04:03 proti: IIRC it should always be one array May 31 08:04:39 any amd64 users out here? May 31 08:05:05 koen: you could abuse the compiler optimisation flags May 31 08:05:24 zecke_: gcc ICEs all over the place May 31 08:05:30 RICH_AND_WEALTHY_BOY = "yes" could be introduced for that May 31 08:05:42 the cross compiler? May 31 08:05:50 hold on May 31 08:06:17 gcc-cross-initial May 31 08:06:25 /home/koen/OE/build/tmp/familiar/work/x86_64-linux/gcc-cross-initial-3.4.3+csl-arm-20050416-r4/ May 31 08:06:37 note the x86_64 bit in the path May 31 08:07:15 I upgraded the kernel from k7 to k8 yesterday May 31 08:07:16 koen: look at bitbake.conf on how HOST_SYS is defined May 31 08:07:26 zecke_: RICH_AND_WEALTH_UNI May 31 08:07:40 right that can happen as well May 31 08:07:57 koen: I am amd64 user May 31 08:08:10 debian? May 31 08:08:14 koen: ubuntu May 31 08:08:19 same difference May 31 08:08:27 any special trick for OE? May 31 08:08:32 koen: and yes, its OE is fscked in 64bit mode May 31 08:08:47 koen: gcc doesnt work very well, libtool is screwed May 31 08:09:13 do I need a 32 bit kernel, or can I use a 64bit kernel? May 31 08:09:15 koen: I have a 32bit chroot for compiling in as I get fedup on instability in 64bit mode May 31 08:09:46 koen: a chroot solves all the problems May 31 08:09:52 all my userspace is 32 bit too, so that shoudln't matter, right? May 31 08:10:05 * koen clueless on 64bit issues May 31 08:10:27 koen: I have a 64bit kernel with a 32bit userspace in a chroot and suffer no problems May 31 08:10:44 koen: when I use 64bit kernel and 64bit userspace Im lucky to get stuff to compile May 31 08:15:06 anyway to force OE to go 32bit? May 31 08:15:40 koen: if your userspace is 32bit then your OE will be, although you might need to run linux32 before bitbake to change the uname May 31 08:16:45 let's try that May 31 08:18:21 that's working like a charm May 31 08:18:24 XorA: thanks May 31 08:19:15 zecke_: you need to run 'linux32 bitbake' on ewi546 till I find a better solution May 31 08:21:22 koen: linux32 bash :-) May 31 08:21:36 linux32 screen :) May 31 08:21:45 koen: that as well May 31 08:23:54 03koen 07 * r1.3457 10openembedded/conf/machine/nokia770.conf: nokia770.conf: remove unneeded wifi and pcmcia drivers May 31 08:24:55 ~lart nokia for closed source wlan drivers May 31 08:24:56 * ibot accelerates a free AOL cd to 50,000 rpm and lets nokia feel it for closed source wlan drivers May 31 08:29:12 >.> May 31 08:29:51 koen: do you know which module they've? May 31 08:30:07 koen: Philips? Samsung? some other Asian? May 31 08:30:37 koen: has anyone forced open a 770 yet and looked inside for info ? May 31 08:30:46 zecke_: no idea yet May 31 08:30:58 ade|desk: I know people with devs boards May 31 08:31:10 ade|desk: those lack the fancy casing May 31 08:31:20 its a start i guess :) May 31 08:31:45 Anyone have any idea why I wouldn't be able to mount mtdblock3 on my Z? May 31 08:31:48 'morning kergoth May 31 08:32:49 it does have one huge 123.5mb jffs2 partition May 31 08:46:51 hi mickeyl May 31 08:48:01 cu May 31 10:05:17 Can someone point me to some documentation on the bitbake.conf Im getting an error that it cant be parsed. May 31 10:05:53 update bitbake and/or check for typos May 31 10:09:56 03koen 07 * r1.3458 10openembedded/packages/rxvt-unicode/ (rxvt-unicode-5.4/xwc.patch rxvt-unicode_5.4.bb): Add new, but not too new rxvt-unicode May 31 10:12:05 tpdd: I had the same problem yesterday, did you just get bitbake installed? May 31 10:16:10 yes i just did. May 31 10:29:16 is that guy I was just talking to still here? May 31 10:29:20 I forget who you are May 31 10:31:21 19:19 < tpdd> Zero_Chaos: did you find a fix? May 31 10:31:28 that one? May 31 10:31:33 koen: thanks May 31 10:31:41 tpdd: yes I did....you still here? May 31 10:32:59 tpdd: that fix was in the local.conf file, I forget exactly where, but it may have been BBFILES, gimme a second, I'll see if I can mess mine up again :-) May 31 10:33:09 yup still here May 31 10:34:10 tpdd: well, I can't break mine like that again, can you post your local.conf somewhere? May 31 10:39:51 tpdd: n/m, I figured it out, try this...export BBPATH=/home/zaurus/src/build:/home/zaurus/src/openembedded May 31 10:40:11 tpdd: add that to your .profile as well, obviously make any needed adjustments May 31 10:50:14 ah crap, a 200G HDD just bit the dust May 31 10:50:15 CoreDump|home: doh! May 31 10:50:23 :( May 31 10:50:26 CoreDump|home: ouch, sorry man May 31 10:51:15 suckage May 31 10:51:37 i am not able to build mysql :-( mysql-native_4.1.10a not found :-( May 31 10:53:18 CoreDump|home: not in warranty? May 31 10:53:26 are 200 GB HDs that old? May 31 10:53:43 luke-jr_: I'm sure its the data loss that is most painful May 31 10:53:56 ~lart western digital May 31 10:53:56 mickey|sports: are you here? May 31 10:53:56 WD stink May 31 10:53:56 * CoreDump|home totally agrees May 31 10:53:56 should have bough a seagate May 31 10:53:58 * ibot pushes the wall down onto western digital whilst whistling innocently May 31 10:54:03 WD rules =p May 31 10:54:15 an newer version wont work :-( May 31 10:54:24 WD has 3 year warranty =p May 31 10:54:46 Zero_Chaos: Zero_Chaos the HDD was part of a RAID, no data-loss :) May 31 10:55:12 CoreDump|home: woohoo! At least that is good, go return that bitch and buy a seagate :-) May 31 10:55:32 I wish I had kept the bill May 31 10:55:33 heh May 31 10:57:08 hmm 200GB SATA ~100? May 31 10:57:09 Zero_Chaos: added that to my path and I have the same issue. echo $BBPATH shows the right path. May 31 10:57:38 tpdd: did you adjust my BBPATH to your configuration, or did you just use mine? May 31 10:57:43 * luke-jr_ is sticking with PATA until SATA has more testing May 31 10:58:07 CoreDump|home: go buy another WD 200GB, then return the broken one and say it didn't work and you don't want it anymore :_) May 31 10:58:08 hmm May 31 10:58:12 tpdd: I had the same error, and it turned out it was a typo in my config May 31 10:58:18 where to get jffs2 support in kernel config? May 31 10:58:22 heh May 31 10:58:45 CoreDump|home: WD will send the replacement before you send the bad HD ;) May 31 10:59:09 hmm May 31 11:03:09 Zero_Chaos I used mine (/home/trichmon/stuff/build:/home/trichmon/stuff/openembedded) May 31 11:04:41 tpdd: and it still doesn't work? what is your error? and is there a /home/trichmon/stuff/openembedded/conf/bitbake.conf? May 31 11:05:07 error is ERROR: Unable to open conf/bitbake.conf May 31 11:05:27 and i do have the /home/trichmon/stuff/openembedded/conf/bitbake.conf May 31 11:05:30 tpdd: you have the latest bitbake? May 31 11:05:34 r235 from svn May 31 11:05:47 tpdd: that us $explitive up.... I had that same error and I needed BBPATH to fix it.... May 31 11:06:01 Just got it today following the instructions on the EO wiki May 31 11:06:11 oe wiki May 31 11:09:03 tpdd: so you are in /home/trichmon/stuff/build and your path is PATH=/home/zaurus/src/bitbake/bin:/usr/local/arm/2.95.3/bin:$PATH and your BBPATH is PATH=/home/zaurus/src/bitbake/bin:/usr/local/arm/2.95.3/bin:$PATH and it doesn't work? (modify my username to yours but you get what I'm driving at) May 31 11:09:16 oops May 31 11:09:29 I meant BBPATH=/home/zaurus/src/build:/home/zaurus/src/openembedded May 31 11:21:40 luke-jr_: no they bloody well will not. May 31 11:21:50 I "installed" bitbake so I thought i didnt have to add it to the path May 31 11:22:07 luke-jr_: wd will do that only if you let them charge it onto your card first, and they can be complete shits about uncharging it May 31 11:22:24 Spyro: Worked fine for me May 31 11:22:32 luke-jr_: plus they refused to refund me the price of the crap drives I had to keep returning them May 31 11:22:40 tpdd: Well, can't hurt, right? May 31 11:22:56 tpdd: past that, I really can't help you, I just setup yesterday :-) May 31 11:22:58 I had three shitters in a row, from different batches. May 31 11:23:02 took them 8 months May 31 11:23:06 tpdd: BBPATH != PATH May 31 11:23:14 now I have seagates and they never break May 31 11:23:34 mickey|sports: ping May 31 11:23:46 Spyro: Some guy and I bought like 4 HDs from the same batch-- all died at the same time May 31 11:24:03 Spyro: in that case, tho, there was no warranty at all because of some reseller :( May 31 11:24:10 luke-jr_: thats not uncommon but these three were from different batches May 31 11:24:17 one from a different factory! May 31 11:24:37 How long is Seagate's warranty? May 31 11:25:08 luke-jr_: 2 years I think but all my seagates (I've owned about 20-30 over the years) have outlasted the 5 year point May 31 11:25:24 If they're all good, why the shorter warranty length? May 31 11:25:44 luke-jr_: 2 years isnt short May 31 11:25:57 Shorter than 3 May 31 11:26:12 luke-jr_: warranties are a weird thing anyhow. I know of one company that garauntee the motor in their product for 12 years... May 31 11:26:21 I have some HDs from >5 years ago that still work, BTW May 31 11:27:36 If I had to guess, they probably tried figuring out the point where they'd have the fewest failures while not losing sales for poor support May 31 11:27:44 and came up with 2 years May 31 11:30:16 treke-: Yes, but since WD's warranty is 3 years, it would suggest their HDs fail less in 3 years than someone w/ less a warranty May 31 11:35:00 Ok i have the BBPATH setup per the wiki instructions, and I have added bitbake/bin to my path. Still get the same error. Guess i will scrap the install and start over. May 31 11:35:07 luke-jr_: probably May 31 11:35:32 tpdd: I wouldn't do that, wait till mickeyl is back, I'm sure he will know (he fixed mine yesterday) May 31 11:35:33 treke-: what error? May 31 11:35:41 though i sort of ignore it since I like seagate drives May 31 11:35:45 zecke: ? May 31 11:36:00 tpdd: what error? May 31 11:36:13 treke-: *sorry* I need more brezel to wake up May 31 11:36:55 ERROR: Unable to open conf/bitbake.conf May 31 11:37:10 echo $BBPATH May 31 11:37:34 /home/trichmon/stuff/build:/home/trichmon/stuff/openembedded May 31 11:38:00 ls /home/trichmon/stuff/openembedded/conf/bitbake.conf ? May 31 11:38:23 yup May 31 11:38:41 zecke: that is about where I started scratching my head.... May 31 11:38:48 i posted my local.conf at http://www.eou.edu/~trichmon/local.conf May 31 11:39:04 tpdd: bitbake -v -DDD world May 31 11:39:16 tpdd: bitbake.conf is 644 and trichmon:users right? May 31 11:40:09 re May 31 11:40:34 tpdd: the content of local.conf should be not of any matter at that point May 31 11:41:19 bitbake.conf is 644 and owned by trichmon:trichmon May 31 11:41:25 mental note: don't try stupid things on primary box May 31 11:41:27 why am I not surprised jffs2/mtdram/etc doesn't work on x86_64? =p May 31 11:41:31 bitbake -v -DDD worlkd is : May 31 11:41:53 DEBUG: CONF reading /home/trichmon/stuff/openembedded/conf/bitbake.conf May 31 11:42:00 DEBUG: CONF file 'conf/site.conf' not found May 31 11:42:06 DEBUG: CONF file 'conf/auto.conf' not found May 31 11:42:13 DEBUG: CONF including /home/trichmon/stuff/build/conf/local.conf May 31 11:42:20 ERROR: Unable to parse conf/bitbake.conf May 31 11:42:35 weird May 31 11:43:34 zecke: 'echo MACHINE-c7x0 > conf/auto.conf' -> cannot parse local.conf May 31 11:43:43 after fixing the typo it ran ok May 31 11:44:37 tpdd: your config looks about right May 31 11:49:37 tpdd: that is really weird May 31 11:50:12 tpdd: so bbpath looks right, so build/conf/local.conf must be provoking a parser error May 31 11:58:04 03reenoo 07 * r1.3450.1.1 10openembedded/packages/linux/ (2 files in 2 dirs): handhelds-sa: upgrade to -hh40. DEFAULT_PREFERENCE=-1 since not yet tested on a device. May 31 11:58:15 hi May 31 11:58:57 ccsmart: I got cash - wrt54gs ordered May 31 11:59:07 ccsmart: thx again May 31 12:01:26 Hrw|husky: congrats May 31 12:01:52 opieirc over bt now May 31 12:02:42 Hrw|husky: neato May 31 12:03:57 konq/e lack statusbar... May 31 12:04:28 but to read wiki is ok May 31 12:05:13 thanks zecke I will go through it and see if I can spot anything. May 31 12:05:33 Hrw|husky: personally, the fact that kong/e views all files (instead of downloading things like .tar.gz) is far more annoying than lack of a statusbar May 31 12:06:37 zero:maybe. I dont use it too often May 31 12:07:03 Hrw|husky: just sharing pet peeves ;-) May 31 12:08:02 zecke: is ther some option to get task switching in opie? May 31 12:08:18 zecke: some qcop maybe? May 31 12:08:54 Hrw|husky: what kind of task switching? (rising an application?) May 31 12:09:29 zecke: from one running to other May 31 12:09:47 like alt-tab on many desktops May 31 12:10:33 Hrw|husky: with QPE/Application/appname show() one can put it in the foreground May 31 12:10:41 Hrw|husky: but we have no 'listing' tasks May 31 12:11:03 ok May 31 12:13:19 just hit the icon on the taskbar May 31 12:13:29 :) May 31 12:14:10 ljp: it need hand.. clamshells are not good to hold in one hand and writing May 31 12:14:59 ljp: and konq/e use taskbar as statusbar often.. May 31 12:15:07 konq/e sux May 31 12:16:50 :))) May 31 12:17:25 so... May 31 12:17:35 is jffs2/mtdram/etc broken? :| May 31 12:17:49 it causes both my systems to hang when I try mount -t jffs2 /dev/mtdblock0 m May 31 12:17:59 Hrw|that's the only task switching in opie for now May 31 12:19:18 ok May 31 12:20:15 03koen 07 * r1.3458.1.1 10openembedded/packages/maemo/osso-gwconnect_0.70.1.bb: Rename: packages/maemo/osso-gwconnect_0.70.2.bb -> packages/maemo/osso-gwconnect_0.70.1.bb. some upstream person removed the sources. ~lart maemo/nokia May 31 12:22:40 I wonder why module-init-tools are built for h3600 May 31 12:23:01 reenoo_: missing PREFERRED PROVIDER? May 31 12:23:54 by default all images have 2.4 and 2.6? May 31 12:24:04 Hrw|husky: no May 31 12:25:21 zecke: why would that be governed by PREFERRED_PROVIDER? May 31 12:25:57 reenoo_: I don't know but I remember seeing message like "consider setting preferred provider for depmod" May 31 12:28:46 Does OE have mkfs.jffs2 anywhere? May 31 12:29:10 luke:mtd-utils May 31 12:30:01 Luke-Jr: yes (staging) May 31 12:30:16 hrw: thanks; any idea if it works on /dev/mtdblockX directly? May 31 12:30:28 zecke: nm. task-bootstrap pulls it in. May 31 12:30:46 luke: nope May 31 12:31:09 And I don't suppose anyone's heard of a fsck.jffs2, right? May 31 12:31:15 reenoo_: iirc module-init-tools is in every 2.4 image May 31 12:31:28 reenoo_: since all images should be able to run 2.4 and 2.6 May 31 12:31:39 reenoo_: at least, that's what pb envisioned May 31 12:31:53 koen: I can use a 2.6 image w/ a 2.4 kernel... May 31 12:32:14 the other way around counts May 31 12:32:32 koen: I guess someone would have to finish the h3600 port for that to become a reality May 31 12:32:33 since ipaqs are pretty much 2.4 only atm May 31 12:32:42 koen: I know; I've only tested one way, though May 31 12:38:33 ~seen offroadgeek May 31 12:38:33 ibot: ping May 31 12:38:41 offroadgeek was last seen on IRC in channel #openzaurus, 5d 3h 57m 55s ago, saying: 'that only lists the feed's ipks (no screenshots or descriptions), but if you want to know about a specific package, just google the name and you'll get all the info you need'. May 31 12:38:42 pong May 31 12:38:55 can I dd mkfs.jffs2 directly to /dev/mtdblock3 or is the header file (used on rootfs.img files) needed? May 31 12:40:02 ibot: ping May 31 12:40:05 pong May 31 12:40:17 uf.. my husky lost bnep0 May 31 12:41:27 strange.. May 31 13:14:44 <[cc]smart> Hrw|husky: good to know it arived. np.... May 31 13:15:25 ccsmart: so few days more and bt card will go to drawer.. May 31 13:16:02 <[cc]smart> bt is nice too, just not for networking :O) May 31 13:16:45 and my cf bt has borken driver which does not survive suspend May 31 14:02:30 mickey|sports: hello? May 31 14:12:41 koen: how was the system called? May 31 14:13:43 the hostname? May 31 14:13:53 ewi546.ewi.utwente.nl May 31 14:13:56 koen: yupp, I forgot to write it down May 31 14:13:57 thanks May 31 14:36:06 does some one have a url for bbc? May 31 14:36:19 as in telly? May 31 14:36:33 bbc.co.uk May 31 14:36:48 Hertog: no as in bitbake c parser May 31 14:37:17 eh...should've figured that out on my own ;) May 31 14:38:08 :) May 31 14:45:07 Hmm, machine is building for hours now (GPE), not complaining, but is there any way to tell it which locales (not) to build? May 31 14:47:27 Hertog: no :} May 31 14:47:44 zecke: I'm noticing a pattern with kitchen sync May 31 14:47:51 sub todos seem to break it May 31 14:54:40 treke: THANK YOU May 31 14:55:19 :) May 31 14:55:28 treke: that is something I've never really used/tried May 31 14:56:00 hehe May 31 14:56:22 it seems like if either side has subtodos, ks just wiped the file :) May 31 14:56:39 treke: could you try to confirm it? May 31 14:56:50 So can you use the updater.sh for OZ 3.5.3 to load a custom rom made with OE or do you have to build a new one? May 31 14:57:00 zecke: sure May 31 14:57:06 tpdd: dunno May 31 14:57:14 treke: I hope to have my kolab account soon May 31 14:57:20 zecke: sure. I'll do some simpler test cases May 31 14:57:24 neat May 31 14:57:42 I'm thinking of setting up a kolab server somewhere May 31 14:58:22 need to see if I can get it sandboxed into receiving mail through the system postfix May 31 14:58:38 I dont want to use their mta and imap server for all my mail :) May 31 14:59:33 treke: I'm sure you can use their stuff a secondary server(s) May 31 14:59:50 I'm sure I can do it somehow May 31 14:59:57 treke: The only downside on Kolab is it comes with its own Linux Distro May 31 15:00:51 the existing all in one design makes my butt itch May 31 15:02:30 apropos bero's comment on bug 50, the arm bits have been removed from the glibc cvs tree. May 31 15:02:34 proti: File "/home/zecke/bitbake/bitbake/bin/bitbake", line 282, in findBestProvider May 31 15:02:37 pv,pr = self.status.pkg_pvpr[f] May 31 15:02:40 KeyError: '/' May 31 15:02:58 proti: looks like file_set is a string? Did you mean that problem earlier today? May 31 15:03:06 ~lart zecke for now having used quilt May 31 15:03:07 * ibot executes killall -KILL zecke for now having used quilt May 31 15:03:38 bugzilla seems to be going over better than rt ever did May 31 15:03:44 yeah May 31 15:03:59 I should sign up for a bugzilla account so I can actually post comments in the bugs. May 31 15:04:06 hehe May 31 15:04:17 every few days I do a run to see if anyone needs update permissions May 31 15:04:52 I suspose I should fix the theme so that it doesnt require a 1280x1024 window May 31 15:04:58 heh May 31 15:05:19 treke: will it be easy to send mails automatically to other email addresses as well? May 31 15:05:29 zecke: ? May 31 15:05:30 treke: simple as creating an account? May 31 15:05:43 treke: I would like to get each and every bugzilla mail to some address May 31 15:06:05 yeah, you can register to get copies of all the bug mails May 31 15:06:48 proti: your patch does not work May 31 15:08:06 * zecke continues with his tinderbox master plan May 31 15:08:13 very good May 31 15:08:18 I've now have a special base.bbclass May 31 15:08:22 bitbake opie-image May 31 15:08:37 and it mails the building with all logs as tinderbox report(s) May 31 15:08:42 cute May 31 15:09:29 now I need some glue to update openembedded, clean caches + (sources) and then start bitbake xyz May 31 15:12:00 * RP sleeps May 31 15:12:01 proti: + pkg_pn[pn] = [] May 31 15:12:02 + for p in p_list: May 31 15:12:02 + pkg_pn[pn] = priorities[p] + pkg_pn[pn] May 31 15:12:23 hmm.. kernel modules appear to have $PV in the name.. how is that expected to work? (do_rootfs blows up for obvious reasons) May 31 15:12:39 RP: *whispers* It's the tooth feh, how does the bbc url look like? May 31 15:12:58 RP: is it green? May 31 15:13:26 proti: first you wipe pkg_pn globally away (mickeyl am I right?) and then May 31 15:13:45 for each p in p_lost you set pkg_pn[pn] instead of appending? May 31 15:14:07 ah you missed the extra [] May 31 15:14:09 zecke: I'm afraid you'll have to explain... May 31 15:14:29 RP: could you give me the bitbake-c parser url? I would like to look at the grammar May 31 15:14:41 ~bbp May 31 15:14:43 i guess bbp is A preliminary .bb parser written in C using 'lemon' for the parser. See http://projects.buici.com/oe/. May 31 15:16:38 RP: thanks and sleep well May 31 15:18:33 zecke: I expect I will :). g'night! May 31 15:22:11 RP: ping May 31 15:27:15 if I want to make my own version of opie-image, how do I make it so bk pull doesn't overwrite it? May 31 15:29:33 Zero_Chaos: setup a seperate directory outside of openembedded and copy opie-image to it May 31 15:29:49 Zero_Chaos: then rename opie-image to opie-image-zc or something like that May 31 15:30:03 Zero_Chaos: and update your BBFILES May 31 15:30:07 that too May 31 15:30:12 ^^ May 31 15:30:36 zecke: can you tell me how to properly set my BBILES, just a : and add another path? May 31 15:31:04 I think it's a space... May 31 15:31:15 mine is BBFILES := "${OEROOT}/openembedded/packages/*/*.bb ${OEROOT}/ljr.pkgs/*.bb" May 31 15:31:16 Zero_Chaos: a space May 31 15:31:21 Luke-Jr: thanks May 31 15:31:25 zecke: thanks May 31 15:32:43 now how do I set bitbake to not pad the initrd for collie and what is a realistic max size? May 31 15:33:07 Zero_Chaos: change the IMAGE_CMD_jffs2 in your conf/local.conf May 31 15:34:18 zecke: I only see IMAGE_FSTYPES = "jffs2 tar" May 31 15:36:25 Zero_Chaos: check the bitbake.conf May 31 15:36:33 Zero_Chaos: you can then just redefine the cmd May 31 15:36:58 zecke: thanks May 31 15:38:39 this line doesn't talk about padding: May 31 15:38:40 IMAGE_CMD_jffs2 = "mkfs.jffs2 --root=${IMAGE_ROOTFS} --faketime \ --output=${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs.jffs2 \ ${EXTRA_IMAGECMD}" May 31 15:38:54 but this one does:EXTRA_IMAGECMD_jffs2 = "--pad --little-endian --eraseblock=0x40000" May 31 15:39:04 should I just remove --pad? May 31 15:39:36 Zero_Chaos: you could May 31 15:39:50 just place the variable you want to redefine in conf/local.conf May 31 15:40:10 zecke: do you know if it will work? I know the image won't load onto the Z that way, but would it properly build so I can see true file size? May 31 15:41:04 Zero_Chaos: it will build May 31 15:41:41 okay, so how big can it be before it will fail? I remember 1 meg less than the padded size, or something like that, right? May 31 15:47:20 Zero_Chaos: I really don't know May 31 15:50:25 can anyone tell me where things like ${OPIE_PIM} are defined? May 31 15:51:11 Zero_Chaos: in the *.inc file? May 31 15:51:14 or meta-opie? May 31 15:51:49 hmmm, I will have to poke around for that May 31 15:55:30 zecke: FYI it was in the .inc file. Thanks for your help May 31 15:57:33 03zecke123 * r236 10bitbake/bin/bitbake: (log message trimmed) May 31 15:57:34 bitbake/bin/bitbake: May 31 15:57:34 Create a common method to get the preferred version May 31 15:57:34 given a package name. May 31 15:57:34 This patch was done by proti (seb BEI frankengul.org) May 31 15:57:34 and made work by myself. The list in list logic does not May 31 15:57:36 seem 'obvious' and I need to consider how to kill it. May 31 15:58:20 ~lart FreeBSD installer May 31 15:58:20 * ibot lowers FreeBSD installer's priority May 31 16:19:38 haha burncd made FreeBSD crash May 31 16:20:03 it feels like Linux years ago May 31 16:20:25 but on the other hand it feels mature as well May 31 16:26:07 good nite May 31 16:27:35 damn I want a Nokia770 badly, I would even port aliens to GTK+ for it... May 31 16:29:16 hehehe May 31 16:29:22 how much are they? May 31 16:30:30 99Euro May 31 16:30:55 treke: http://www.heise.de/newsticker/meldung/60111 <- 100 units to the price of 99Euro each May 31 16:31:04 treke: the money will be donated to the GNOME Foundation May 31 16:31:17 500 instad of 100 May 31 16:31:21 ah May 31 16:32:40 hehe. any special requirements? My german isn't so hot still May 31 16:33:00 zecke: uname -a May 31 16:33:11 zecke: maybe your system or the port is out of date May 31 16:33:29 and yes, the 770 looks yummy May 31 16:33:50 treke: the whole thing is called "Developer Device Program", the revenue will be donated to the GNOME Foundation, 500 units in total... May 31 16:34:05 treke: and that the info was revealed at GUADEC today May 31 16:34:15 zecke: I'll get you one once we get a workable kitchensync release :) May 31 16:34:17 treke: so I guess Miguel will buy 400 units for Ximian May 31 16:34:46 hehe May 31 16:35:01 * zecke opens the drawer and takes the GOLD-Master KitchenSync CD May 31 16:35:08 hehe May 31 16:35:18 hiding there all this time May 31 16:35:36 Use not before 21xx May 31 16:36:07 can anyone tell me for certain how to stop initrd from being padded for collie? removing --pad didn't help May 31 16:36:15 anyway regardless of a Nokia 770, kitchensync must be stable SOON May 31 16:36:22 Zero_Chaos: why do you want to? May 31 16:36:30 ok good nite guys May 31 16:36:32 Zero_Chaos: the image must be padded for it to work May 31 16:36:34 Zenight May 31 16:36:38 treke: so I can see the acctual size of my image May 31 16:36:43 zecke: night May 31 16:36:45 goodnight zecke, thanks May 31 16:36:56 treke: I'll look into Sub Todos later this week May 31 16:37:01 cool May 31 16:37:08 i'll try to get a simplified test case May 31 16:37:16 'night all May 31 16:37:18 rather than a giant list of todos and subtodos May 31 16:38:04 treke: so can you tell me how to stop that initrd from being padded/truncated? May 31 16:38:19 remove --pad= from collie.conf May 31 16:42:09 treke: should I remove --eraseblock as well? May 31 16:42:25 no May 31 16:42:30 just --pad May 31 16:42:59 treke: can I put that in local.conf or will that not override collie.conf May 31 16:43:54 you should be able to override it by putting it in local.conf after including collie.conf May 31 16:44:11 not specifying a --pad doesnt mean the image wont be padded May 31 16:44:18 it just means it will be padded to the nearest block May 31 16:45:10 treke: if I'm trying to find the REAL unpadded size of my image, how do I do it? remove --pad=XXXX and just put --pad? May 31 16:45:22 no May 31 16:45:30 it will still be padded to a block May 31 16:45:55 treke: so what is the best way to figure out the unpadded size? (or the best approximation May 31 16:46:13 take the size of the padded image May 31 16:46:40 it's somewhere between paddedimagesize and (paddedimagesize - blocksize) May 31 16:47:08 treke: so if I remove --pad=XXX then it will be about as close as I can get, right? May 31 16:47:13 yup May 31 16:48:52 treke: awesome, thanks. And my image should be no bigger than the fully padded size - 1 meg otherwise something about garbage collection blah blah blah, right? May 31 16:50:00 I don't know there. I know there used to be a bug that would cause the fs to go readonly if it was full, but I thought it was fixed May 31 16:50:20 in general full filesystems are bad though :) May 31 16:50:29 treke: good to know, my image will likely be rather large, I'll let you know what happens :-) May 31 16:58:14 treke: it's still padded..... any idea's? I don't want to have to modify something that will mess up "bk pull" May 31 16:58:35 Zero_Chaos: We know it will be padded May 31 16:58:43 it should be padded to the nearest block May 31 16:59:11 treke: it is the exact same size as before I removed --pad= May 31 16:59:14 our images are pretty full by default, so that may well be the same as what it has to be padded to May 31 16:59:41 treke: I made a lot of modifications, no way the size didn't change May 31 16:59:51 treke: I should be way over in all honesty May 31 17:00:50 it looks like the block size is 128k May 31 17:00:56 if I'm remembering this right May 31 17:03:09 treke: does it truncate the image as well (with a command different than --pad=) maybe I'm over the size? May 31 17:03:44 not that I know of May 31 17:04:24 what size is it? May 31 17:05:28 bbimage is what creates the image May 31 17:06:31 I made this one yesterday as "bitbake opie-image" 14680064 May 30 05:01 opie-image-collie-20050530031722.rootfs.jffs2 May 31 17:07:01 and I made this one today as "bitbake zc-image" 14680064 May 31 19:55 opie-image-collie-20050531234344.rootfs.jffs2 May 31 17:07:19 looks like I'm being overridden perhaps? or truncated. May 31 17:08:46 treke: do you think I can make a ~/src/build/conf/machine/collie2.conf and change my localconf to use collie2? May 31 17:09:04 that might stop me from being overridden May 31 17:09:10 just edit collie.conf May 31 17:09:21 then fx it when you're done May 31 17:09:26 treke: won't that interfere with "bk pull"? May 31 17:09:37 you're going to be changing it back anyways May 31 17:09:41 true... May 31 17:09:49 treke: thanks May 31 17:10:11 without padding, it wont work, so once you figure out what's going on size wise, you'll need to repad May 31 17:12:10 treke: yeah, that's the idea, I want to include some more important things for me, while removing things I don't need and praying that there is enough room. May 31 17:12:13 14680064 May 31 19:55 opie-image-collie-20050531234344.rootfs.jffs2 May 31 17:12:59 simple solution May 31 17:13:07 make your life a whole lot easier May 31 17:13:25 build an image with just a basic opie May 31 17:13:31 then install what you want May 31 17:13:37 then get a list of installed packages May 31 17:13:43 and create a new image bb May 31 17:13:58 no fighting with padding or anything May 31 17:14:04 you know it will all fit May 31 17:16:11 treke: if this doesn't work, I'll do that, good idea May 31 17:18:44 cool May 31 17:19:07 anything to avoid rebuilding images :) May 31 17:19:10 slow and tedious May 31 17:19:52 true May 31 18:59:01 OpenZaurus had 3 different kernels for Cxx0 on GPE... how are they selected in OE? May 31 19:06:40 it's a long time that there are no update in http://oesources.org/snapshots/ May 31 19:13:20 robert: it's not being updated May 31 20:48:15 snapshots are no longer updated. you can use the open source tools to access the tree May 31 20:52:24 ggilbert: could you tell me how to do in detail? May 31 20:54:59 no May 31 20:55:17 I haven't a clue how you do it May 31 20:56:40 just that the tools now exist May 31 20:56:48 bk? May 31 20:57:11 bitmover.com has a link to the tool May 31 21:00:55 the wiki has commands May 31 21:27:26 * france is back (gone 14:57:34) May 31 21:41:55 can anyone explain what's up with bitbake, I think there is something wrong, but I'm not sure, it might be me. I try to bitbake opie-image and I get this traceback... May 31 21:42:04 anyone else have this problem? May 31 21:46:00 Zero_Chaos: i didn't get this problem, but it failed on getting openzaurus-sa_2.4.18-rmk7-pxa3-embedix May 31 21:46:45 I completely deleted my tmp and I tryed "bitbake opie-image" and got this: May 31 21:46:46 NOTE: package makedevs-1.0.0: completed May 31 21:46:46 Traceback (most recent call last): May 31 21:46:46 File "/home/zaurus/src/bitbake/bin/bitbake", line 715, in ? May 31 21:46:46 cooker.cook( args[1:] ) May 31 21:46:53 and a lot more like that.... May 31 22:43:36 i build glibc ,and get error: May 31 22:43:37 /usr/local/arm/2.95.3/lib/gcc-lib/arm-linux/2.95.3/libgcc.a(__dummy.o): In function `__dummy': May 31 22:43:38 | __dummy.o(.text+0x0): multiple definition of `__dummy' May 31 23:26:59 * france is away: Away **** ENDING LOGGING AT Tue May 31 23:59:56 2005