**** BEGIN LOGGING AT Tue Sep 05 02:59:56 2006 Sep 05 03:11:51 03justinp 07org.oe.oz354x * rc4f8223c... 10/ (1 packages/efl/e/fix-configure.patch): e: drop unused patch Sep 05 06:45:18 hrw@bitbake:~/devel/oe/org.openembedded/packages/rox$ mtn dif .|diffstat rox-filer_2.5.bb | 36 +++++++++++++++++++++++++++++++++++- 1 file changed, 35 insertions(+), 1 deletion(-) Sep 05 06:45:21 hrw@bitbake:~/devel/oe/org.openembedded/packages/rox$ mtn dif .|diffstat rox-filer_2.5.bb | 36 +++++++++++++++++++++++++++++++++++- 1 file changed, 35 insertions(+), 1 deletion(-) Sep 05 06:45:25 rox-filer_2.5.bb | 36 +++++++++++++++++++++++++++++++++++- Sep 05 06:45:25 1 file changed, 35 insertions(+), 1 deletion(-) Sep 05 06:45:38 something for koen ;D Sep 05 06:45:41 cu Sep 05 07:07:57 good morning all Sep 05 07:33:58 03koen 07org.oe.dev * r2af9caaa... 10/ (1 packages/povray/povray_3.6.1.bb): povray: remove bogus section Sep 05 07:38:38 morning Sep 05 07:38:54 hey XorA Sep 05 07:43:11 good morning ! Sep 05 07:49:30 hi Sep 05 07:49:39 something for koen: Sep 05 07:49:42 rox-filer_2.5.bb | 36 +++++++++++++++++++++++++++++++++++- Sep 05 07:49:44 1 file changed, 35 insertions(+), 1 deletion(-) Sep 05 07:49:45 yo hrw|work Sep 05 07:50:00 hey hrw|work Sep 05 07:50:24 koen: I can push that change but you will have to resolve problems Sep 05 07:50:32 sure Sep 05 07:51:38 * koen takes a break from looking at ifdef ZAURUS crud Sep 05 07:51:44 * XorA grns Sep 05 07:52:18 XorA: the libgnomeui you pushed need gtk 2.10.x Sep 05 07:52:36 I think it's time for a preferred-gtk28-versions.inc Sep 05 07:53:07 and a sane way to get PV from another package inside a .bb Sep 05 07:53:08 koen: arse, and lower than that wont buld wth gtk2.10.x Sep 05 07:53:23 XorA: yep, it sucks Sep 05 07:53:37 koen: ah well, at least I will have gnome swankiness Sep 05 07:53:40 I just default_pref -2'd it locally Sep 05 07:53:46 koen: feel free to pull its preference down Sep 05 07:54:00 koen: I can jack it up again in local.conf Sep 05 07:54:55 koen: will 1.14.X build on gtk2.8/2.6? Sep 05 07:55:05 * koen has no idea Sep 05 07:55:21 koen: as a simple cp of 2.15.X will work, I had that originally, but 2.14 is not gtk 2.10 compat Sep 05 07:56:26 03hrw 07org.oe.dev * rb91a2f03... 10/ (1 packages/rox/rox-filer_2.5.bb): rox-filer: package ALL needed infos. do_install() taken from Debian package which I maintained years ago Sep 05 07:59:30 03koen 07org.oe.dev * r5fca8a92... 10/ (1 packages/gnome/libgnomeui_2.15.2.bb): libgnomeui 2.15.2: lower default pref since it need gtk 2.10, which also has a negative default-pref Sep 05 07:59:39 morning folks Sep 05 08:00:35 hrw|work: you missed a '}' Sep 05 08:00:36 fsck.. I forgot (again) to test usbnet from tosa to host... Sep 05 08:00:38 hey ade|desk Sep 05 08:00:42 koen: ah. Sep 05 08:02:01 and for() is wrong Sep 05 08:03:46 koen: glib and gtk+ dirs desperately need cleaned, we need to know what people actually use for distros and dump the junk Sep 05 08:04:03 XorA: writing a mail to OE right now :) Sep 05 08:04:42 NOTE: package rox-filer-2.5-r1: task do_package: completed Sep 05 08:05:29 I have this: http://rafb.net/paste/results/3tESfJ83.html Sep 05 08:05:51 koen: I am wlling to take MAINTAINER for 2.10 series as I use it personally, the rest I would prefer someone smarter Sep 05 08:06:56 koen: wrong Sep 05 08:08:20 koen: pushed fixes Sep 05 08:09:30 anyone else here have their vim act really oddly when it comes to indentation in bitbake files? Sep 05 08:09:33 i think we need an indent plugin Sep 05 08:09:38 03hrw 07org.oe.dev * rb13a3adc... 10/ (1 packages/rox/rox-filer_2.5.bb): rox-filer: fixed do_install() typos Sep 05 08:09:47 to handle the shell indentation in shell parts of the file, python in python, etc Sep 05 08:10:47 i guess autoindent works ... sort of Sep 05 08:11:02 autoindent screw up pasting stuff Sep 05 08:11:08 +s Sep 05 08:11:19 thats what "set paste" is for :P Sep 05 08:11:31 :set paste .. paste .. :set nopaste Sep 05 08:11:33 edit more Sep 05 08:11:54 * koen notes Sep 05 08:14:14 XorA: we'd need to split up OE and have bbcollections like 'gtk-2.6' 'gtk-2.8' 'gtk-2.10', etc Sep 05 08:14:41 at the expense of some duplication Sep 05 08:16:32 We need to get versioned depends working Sep 05 08:16:42 The new taskData code should make that easier Sep 05 08:16:44 that as well Sep 05 08:17:01 * koen needs to finish packaged-staging as well Sep 05 08:17:06 morning all btw Sep 05 08:17:13 ~curse statistic and chance course Sep 05 08:17:16 RP: gtk+ stuff as koen says needs a way to say give me gtk+ ${PV} that I am using Sep 05 08:17:19 good morning RP Sep 05 08:17:22 i always wanted to see a more generic mechanism. i depend upon a package whose PROVIDES contains X, and whose PV is Y Sep 05 08:17:26 May the fleas of a thousand camels infest your most sensitive regions, statistic and chance course ! Sep 05 08:17:31 and let the system figure out which packages satisfy that condition Sep 05 08:18:17 XorA: Why do you need to know the PV though? Sep 05 08:18:38 XorA: The PV is something bitbake should be worrying about, not the .bb files Sep 05 08:18:43 RP: http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-September/000177.html Sep 05 08:19:14 "do_install for packages dependant on gtk+ breaks since ${libdir}/gtk-2.0/2.4.0/ is now ${libdir}/gtk-2.0/2.10.0/" Sep 05 08:19:30 that's why you need PV Sep 05 08:19:41 RP: because gtk+ installs in /usr/lib/${PV} Sep 05 08:20:06 XorA: ${LIBV} Sep 05 08:20:51 koen: I suppose you could dump ${LIBV} into a file in staging Sep 05 08:20:57 If a package needs 2.10, you'd just mark it DEPENDS = "gtk+ (>= 2.10.0)" Sep 05 08:20:59 mikearthur: dude Sep 05 08:21:09 hey mikearthur Sep 05 08:21:27 Figuring out PV can be done from staging Sep 05 08:21:34 RP: the problem is packages build against either version happilly, but need to know where to put their results Sep 05 08:21:55 RP: we are talking gtk+ plugins and themes here Sep 05 08:22:00 * hrw|work want PV available too Sep 05 08:22:06 XorA: and they can't detect which version they're building against? Sep 05 08:22:27 'you are running kernel 2.6.17-3.5.4.2-rc2-r47 instead of 2.6.17-3.5.4.2-rc2-r49' Sep 05 08:22:33 RP: the packages can detect it fine, it's do_install that breaks Sep 05 08:23:03 Good Morning Sep 05 08:23:19 I'd like to declare a package obsolete Sep 05 08:23:25 koen: So we need some function to set a variable based on the contents of staging Sep 05 08:23:26 XorA: option #2 is to add libdir/gtk-2.0/2.10 to all FILES Sep 05 08:23:29 Having PV for other packages available will lead to nightmares and shouldn't be needed Sep 05 08:23:59 The .bb file for emelfm is horribly outdated, and will not build. Sep 05 08:24:14 B_Lizzard: patches accepted Sep 05 08:24:15 B_Lizzard: submit updated version via bugtracker? Sep 05 08:24:24 The SRC_URI leads to a website which has long expired Sep 05 08:24:32 morning guys Sep 05 08:24:35 koen, Xora Sep 05 08:24:37 B_Lizzard: submit updated version Sep 05 08:24:40 hi mikearthur Sep 05 08:25:04 I tried to build a new one, but it fails in compilation... I'm not all that experienced with .bb files. Sep 05 08:25:09 But I'll try again Sep 05 08:25:22 hey hrw|work Sep 05 08:25:29 B_Lizzard: we like new people who learn .bb :-) Sep 05 08:25:34 :) Sep 05 08:25:47 The .bb file should be pretty much the same Sep 05 08:26:03 Dependancies haven't changed, as far as I know Sep 05 08:26:23 My only problem is front-porting that patch Sep 05 08:26:34 Anyways, I'll try Sep 05 08:27:08 Do I test new .bb files with "bitbake -b ../whatever" ? Sep 05 08:27:16 koen: actually in gtk+ .pc file s the line gtk_binary_version=2.10.0 we could use that Sep 05 08:27:59 That sounds ideal - load that into a variable and use it in the do_installs Sep 05 08:28:11 XorA: reply to my mail with that :) Sep 05 08:28:14 and save RP grey hair :-) Sep 05 08:28:26 koen: is that there for != 2.10.X Sep 05 08:29:25 morning Sep 05 08:32:06 koen: as an aside, gtk+ and glib updated here, just need to test them on device Sep 05 08:33:15 Crofton: you around? i need to pick your brain :) Sep 05 08:34:49 damn, sf.net is being glacially slow this month Sep 05 08:44:00 XorA, koen: FWIW, 2.8.9 sets gtk_binary_verson=2.4.0 Sep 05 08:44:16 RP: that sounds good Sep 05 08:44:29 anyway need to go interview for a replacement for me, back soon Sep 05 08:47:38 03koen 07org.oe.dev * rc770964c... 10/ (1 packages/tesseract packages/tesseract/tesseract_1.0.bb): tesseract: add 1.0, tesseract is a OCR engine Sep 05 08:54:05 RP: so parsing the pc sounds like a good plan Sep 05 08:54:59 koen: yes Sep 05 08:57:33 Someone added rox! Sep 05 08:57:36 Thanks Sep 05 08:58:18 03koen 07org.oe.dev * rf5ac8923... 10/ (1 conf/distro/include/angstrom.inc): angstrom: set correct mail address Sep 05 08:58:36 thanks koen Sep 05 09:10:04 B_Lizzard: rox-filer... argh Sep 05 09:10:09 :) Sep 05 09:10:18 It should work well on the jornada Sep 05 09:10:54 I had the strangest problem right now Sep 05 09:11:11 rox-filer can be usefull but only after ripping about 25% of code out Sep 05 09:11:25 Whenever I tried building rxvt, I would run out of memory Sep 05 09:11:52 installing it after compiling it also brought my system down... Sep 05 09:12:16 Very strange Sep 05 09:13:23 Why is that with ROX, hrw? Sep 05 09:13:44 I don't like the program myself, but it shouldn't be that useless Sep 05 09:14:35 B_Lizzard: when I click on folder icon I want to go inside NOT run application which is in this folder Sep 05 09:14:50 with rox-filer I will not know will it open folder or run application Sep 05 09:15:10 Well, it should spice up your computer experience Sep 05 09:15:23 "Experience the unexpected" Sep 05 09:15:30 B_Lizzard: I used rox-filer for ~2 years Sep 05 09:15:34 each day Sep 05 09:15:42 :D Sep 05 09:15:55 and I had own fork of it: rox-filer-hrw which had some things removed/added Sep 05 09:15:58 That's a lot of unwanted actions there, eh? Sep 05 09:15:59 who does all the shiney OE vector art, for the bootsplash and the like? Sep 05 09:16:13 Where is that fork now? Sep 05 09:17:25 B_Lizzard: it died ~3 years ago when I upgraded machine Sep 05 09:17:43 ... Sep 05 09:17:57 That was the end of rox for you, eh? Sep 05 09:18:45 mikearthur: there was a contest Sep 05 09:18:54 koen: who won? Sep 05 09:19:16 http://openzaurus.org/wordpress/2006/05/18/graphics-contest-winner/ Sep 05 09:20:59 nice work Sep 05 09:21:31 B_Lizzard: yes Sep 05 09:23:51 where is "do_compile()" defined? Sep 05 09:24:43 T0mW: classes/base.bbclass and other classes Sep 05 09:25:06 I've a kernel mkdep that is dumping because of long string Sep 05 09:26:24 koen: so, I can take that class definition, copy it into the .bb and thus override it? Sep 05 09:26:38 yes Sep 05 09:26:43 nod Sep 05 09:30:55 any bitbake hackers here? Sep 05 09:31:32 no Sep 05 09:31:48 zecke: pssst Sep 05 09:33:16 Then I want to file a bug against http://www.openembedded.org/the-people-behind-openembedded :-) Sep 05 09:33:18 IRC: zecke Sep 05 09:33:19 +Bitbake hacker Sep 05 09:33:21 -Bitbake hacker Sep 05 09:34:50 hi likewise Sep 05 09:35:06 hrw|work: hi there Sep 05 09:36:05 ~curse my network at work Sep 05 09:36:10 May you be reincarnated as a Windows XP administrator, my network at work ! Sep 05 09:36:28 any objections to using http://www.openembedded.org/~koen/OE/OE-logo.eps for the OE t-shirts? Sep 05 09:36:58 the 'source' svg is in the same dir Sep 05 09:38:20 koen: fix the svg first though :-) Sep 05 09:38:52 in bitbake, myFilterProvider() calls findBestProvider() with 2 args instead of the required 3, it seems. Sep 05 09:40:20 XorA: firefox messed up the text, put in inkscape it's fine Sep 05 09:40:22 koen: make those ---- lines a bit ligher Sep 05 09:40:57 hrw|work: I made them darker to have better contrast on the white shirt Sep 05 09:41:01 ah Sep 05 09:41:13 and thicker as well Sep 05 09:43:01 koen: when I loaded in inkscape I found the text had wierd attribs, when I fixed those, firefox dsplays it right Sep 05 09:43:11 koen: when I did the Take Off WIth OE logo Sep 05 09:43:16 :) Sep 05 09:43:43 the printer wants eps or cdr, so weird stuff in the svg isn't really a problem Sep 05 09:47:51 RP: so a package we know needs gtk+ 2.10 is libgnomeui, do you want me to put a version DEPEND in there and then try and rebuild libgnomeui and see if bitbake has kittens? Sep 05 09:50:35 kergoth: patch.bbclass is broken in regard to quilt-native.bbclass Sep 05 09:50:45 any got a bitbake revision where -g works? Sep 05 09:50:54 what is -g? Sep 05 09:51:01 dependency graph Sep 05 09:51:15 it isn't working in 1.6ß Sep 05 09:51:27 zecke: hm? Sep 05 09:51:29 zecke: thanks. Sep 05 09:51:45 kergoth: patches are not applied in quilt-native breaking my OSX builld Sep 05 09:51:50 works just fine here. Sep 05 09:52:01 kergoth: which revision are you using? Sep 05 09:52:02 kergoth: patches are not applied! even if it says so Sep 05 09:52:03 quilt-native uses the 'patch' patchtree, not 'quilt' Sep 05 09:52:15 likewise: isn't it working in bitbake1.6? Sep 05 09:52:22 kergoth: I see PATCHTOOL = "patch" Sep 05 09:52:26 yes. Sep 05 09:52:32 which is as it should be Sep 05 09:52:37 kergoth: yes Sep 05 09:52:39 so its calling out 'patch' to apply the patches Sep 05 09:52:45 kergoth: and bitbake says patches are applied Sep 05 09:52:51 zecke: bitbake -g is not working in rev 581 Sep 05 09:52:52 kergoth: but the patch is not applied! Sep 05 09:52:53 are you on a bsd box, non-gnu patch? Sep 05 09:53:02 then maybe patch's syntax differs Sep 05 09:53:05 read patch.bbclass Sep 05 09:53:19 kergoth: both patches are not applied, they were applied before -> regression Sep 05 09:53:21 i know nothing about bsd Sep 05 09:53:32 and i cannot reproduce or test on a bsd box Sep 05 09:53:34 kergoth: note: Both patches are not applied Sep 05 09:53:35 -> fix it yourself Sep 05 09:53:42 yes, i heard you the first 3 times you said that Sep 05 09:53:48 kergoth: My fix would be to disapprove your version Sep 05 09:54:04 you using an oddball box isnt my problem Sep 05 09:54:15 go ahead and disprove it if you want to continue to use a weak implementation Sep 05 09:56:04 kergoth: Hehe. You say you don't know the cmd syntax for BSD patch... that is fine Sep 05 09:56:19 kergoth: not checking the return value of the invocation of patch is another storry Sep 05 09:56:30 kergoth: having a regression is another storry Sep 05 09:57:29 it does check the return value of patch. read the code, and come back when you have a clue about what its doing Sep 05 09:57:30 kergoth: and you are sure that patches are applied on Linux/GNU patch? Sep 05 09:58:41 kergoth: hehe, a little bit of documentation would make things more easy Sep 05 09:58:43 good point, though i thought i checked that Sep 05 09:58:56 dont assume you know what the code is doing when you obviously dont, then Sep 05 09:59:08 return values are checked. Sep 05 09:59:24 kergoth: you assumed that the syntax was wrong, I assumed that BSD patch would set an error code ;) Sep 05 09:59:30 Chain-Of-Assumption Sep 05 09:59:47 quite Sep 05 10:00:23 time for the ass out of you and me joke Sep 05 10:00:36 PATCHTOOL == patch will use PatchTree? Sep 05 10:01:27 you were right, patch isnt erroring, but it isnt applying anything either. doesnt error on either linux or bsd, just doesnt do anything. sorry for assuming it was a bsd problem Sep 05 10:01:29 XorA: sorry, I don't get that (language issue) Sep 05 10:01:50 zecke: when you assume you make an "ass" out of "u" and "me" Sep 05 10:02:00 zecke: "ass" being a donkey Sep 05 10:02:43 hahah, patchtree is just doing a dry run Sep 05 10:02:47 * kergoth slaps forehead Sep 05 10:02:56 kergoth: ;) Sep 05 10:03:23 shellcmd.append('--dry-run'), but that is okay? Sep 05 10:03:33 check if it will work and then patch for real? Sep 05 10:05:07 XorA: Testing versioned DEPENDS like that would be good :) Sep 05 10:05:18 * RP wonders how many kittens bitbake will have :) Sep 05 10:06:22 kergoth: thanks for looking into it Sep 05 10:06:34 zecke: yep, it just never actually did it for real. it only reran it when 1) dry run failed and 2) force was set Sep 05 10:07:04 zecke: just need to remove the --dry-run from the arg list and rerun runcmd, outside of that try block Sep 05 10:08:03 after looking at quality of my connection I should rather do /nick hrw|offline then hrw|work.. Sep 05 10:08:05 np, glad we're actually testing things. i shouldve unit tested it better Sep 05 10:08:09 sorry about that Sep 05 10:08:14 ;) Sep 05 10:08:24 kergoth: now we are at illegal arguments ;) Sep 05 10:08:28 hehe Sep 05 10:08:41 arguing is fun, sometimes i do it just for the sake of arguing :P Sep 05 10:08:59 :) Sep 05 10:09:04 yup, it was fun Sep 05 10:09:11 I'm waiting for the food delivery anyway Sep 05 10:11:03 kergoth: will you look into pushing a proper fix? Sep 05 10:11:35 zecke: doing so as we speak, even simplifies that code Sep 05 10:11:42 i'll look into adding some docstrings Sep 05 10:12:16 kergoth: can you say something about strippath? is that the old pnum? Sep 05 10:12:39 yup, more aptly named, since -p strips / seperated blocks off of the paths to the files Sep 05 10:12:49 s,/,'/', Sep 05 10:13:04 kergoth: do I need to update SRC_URIs for that? Sep 05 10:16:35 nope, old syntax is used, just internally its named differently Sep 05 10:16:42 okay :) Sep 05 10:16:45 the patch syntax in src_uri needs to be revamped later anyway Sep 05 10:16:49 one step at a time Sep 05 10:16:52 heh Sep 05 10:17:01 patch: option requires an argument -- p (not your problem) Sep 05 10:17:06 Command Error: exit status: 2 Output: Sep 05 10:17:11 and you check error codes ;) Sep 05 10:17:22 yeah, mine just did that too, some other bug :( looking into it Sep 05 10:17:54 patch['strippath'] must be None Sep 05 10:18:10 I have seen at least one 'default' value for that in the code Sep 05 10:18:18 but admitingly I have no clue about the code Sep 05 10:18:19 it isnt though, i just printed patch, its '1' Sep 05 10:18:22 * kergoth confusd Sep 05 10:18:28 * kergoth prints the shell command being run Sep 05 10:18:51 NOTE: shellcmd is patch < /home/kergoth/code/oe/oe/packages/quilt/files/install.patch -p 1 Sep 05 10:18:54 hmm Sep 05 10:19:47 * kergoth grumbles Sep 05 10:23:53 RP: ERROR: Nothing provides dependency (>= Sep 05 10:24:08 ~seen Crofton Sep 05 10:24:17 crofton is currently on #casualti (13h 16m 11s) #ol (13h 16m 11s) #oe (13h 16m 11s) #elinux (13h 16m 11s). Has said a total of 1 messages. Is idling for 8h 33m 11s, last said: 'and?'. Sep 05 10:25:47 hmm Sep 05 10:25:50 kergoth: weirdo Sep 05 10:26:06 XorA: that's what I suspected Sep 05 10:26:23 XorA: versioned RDEPENDS used to work, but I think bitbake 1.6 will dislike them as well Sep 05 10:26:36 koen: that was 1.6 Sep 05 10:26:39 koen: versioned RDEPENDS work fine Sep 05 10:26:54 RP: that's good to hear! Sep 05 10:27:03 We just need to add the filtering code around bitbakes use of DEPENDS Sep 05 10:27:30 koen: What bitbake doesn't like is an RDEPENDS it can't evalute which was the problem with the dyamically genereted kernel one Sep 05 10:27:53 *food* Sep 05 10:28:41 zecke: Would you think fixing up DEPENDS handing to drop version numbers would be a good change in the stable versions? Sep 05 10:30:28 hi mickeyl Sep 05 10:30:50 morning Sep 05 10:31:30 hey mickeyl Sep 05 10:31:37 * koen heads to the printer Sep 05 10:31:39 later all Sep 05 10:31:43 hi mickeyl Sep 05 10:35:43 morning all Sep 05 10:39:14 hi lrg Sep 05 10:40:42 hi liam Sep 05 10:41:00 * lrg wishes he still had a borzoi for testing Sep 05 10:42:05 lrg: You can't persuade wolfson? :) Sep 05 10:42:40 lrg: I'll generate that log... Sep 05 10:43:38 hey liam Sep 05 10:43:53 lrg: I plan to buy terrier next month Sep 05 10:43:58 RP: I should be getting a better test harness built that will drive i2c over 30cm wires. This way I'll be able to hook up the mainstone to the test boards Sep 05 10:44:09 XorA: excellent news Sep 05 10:44:27 ~borzoi Sep 05 10:44:29 i guess borzoi is the Sharp SL-C3100, or a dog (see http://www.dogbreedinfo.com/borzoi.htm ) Sep 05 10:44:49 XorA: I'll be looking for someone to fix mplayer for the cxx00 :) Sep 05 10:46:19 RP: btw is there possibility for accellerated X on pxafb? Sep 05 10:47:12 XorA: accelerated YUV video and accelerated cursor are the only possibilities Sep 05 10:47:24 RP: sucks compared to c7x0 then Sep 05 10:47:26 and the latter is pretty useless with a touchscreen Sep 05 10:47:59 XorA: From that PoV, yes. The lack of the docs on the w100 equal the playing field a lot though Sep 05 10:48:31 RP: the little that MArdy put together makes w100 a much nicer chip Sep 05 10:48:46 RP: ts done amazing things wth gtk+ and firefox rendering Sep 05 10:49:19 XorA: That's good to know. I'll have to have a look at the patch :) Sep 05 10:49:59 lrg: Which bits of the log are you interested in? I can't seem to fit the whole thing in dmesg :-/ Sep 05 10:51:51 RP: back in the days I was developing X driver for ST, most important operation was bitblt :-) Sep 05 10:53:00 XorA: I can imagine. Sadly the pxa270 has no such operation :-/ Sep 05 10:53:09 ~lart pxa270 Sep 05 10:53:10 RP: the latter bits, e.g. its runtime data Sep 05 10:53:10 * ibot nabs the moon and broadsides pxa270 with the sea of tranquility Sep 05 10:54:31 ~lart intel for good measuer Sep 05 10:54:31 * ibot slaps intel upside and over the head with one freakishly huge killer whale named hugh for good measuer Sep 05 10:57:24 lrg: sent Sep 05 10:57:26 morning Sep 05 10:58:24 hi Sep 05 10:58:59 re Sep 05 10:59:32 could anyone here tar.gz the workdir of gconf-dbus and orbi2-native for me? Sep 05 11:02:20 RP: thanks Sep 05 11:03:10 what does "re" mean? Sep 05 11:03:47 hi CoreDump|home Sep 05 11:04:00 CoreDump|home: gconf-dbus default OZ version? Sep 05 11:04:08 hrw|work: yeah Sep 05 11:04:26 ok Sep 05 11:05:34 good morning Sep 05 11:07:58 | checking for GCONF_DBUS... configure: error: You need either ORBit or D-BUS for IPC Sep 05 11:08:22 YES YES Sep 05 11:08:28 muahaha Sep 05 11:08:47 I _knew_ it was a bug in autoconf Sep 05 11:08:52 ;D Sep 05 11:09:07 one down, one more to go Sep 05 11:09:25 orbit2-native-2.10.2-r0 Sep 05 11:11:37 hrw|work: I wonder how you can build gpe-image when that package doesn't compile on your box Sep 05 11:19:38 03clarson 07org.oe.dev * r7e16c38e... 10/ (1 classes/patch.bbclass): patch.bbclass: fix issue encountered by zecke, where PatchTree was only doing a --dry-run, never actually applying the patch. Only quilt-native in oe was using that. Sep 05 11:23:39 CoreDump|home: I'm able to build them Sep 05 11:23:44 CoreDump|home: from feeds ;D Sep 05 11:23:50 hrw|work: lol Sep 05 11:23:53 CoreDump|home: | configure.in:124: error: possibly undefined macro: PKG_CONFIG_MIN_VERSION Sep 05 11:23:59 hrw|work: yeeesssssssss Sep 05 11:24:00 CoreDump|home: thats orbit2-native Sep 05 11:24:10 so my buildroot isn't f*cked Sep 05 11:24:25 CoreDump|home: maybe yours and hrws are messed up Sep 05 11:24:51 koen: possibly, I tried w/ debian SID _and_ debian testing Sep 05 11:24:54 that the old pkgconfig remove patch problem Sep 05 11:25:35 XorA: I read about it in the bugtacker but there was no indication how it was solved Sep 05 11:26:12 is it not solved by removng file://pkg.m4 from pkgconfig_*.bb Sep 05 11:27:49 XorA: hmmm, I'll try. I was under the impression that the "fix" was already pushed into .oz as well Sep 05 11:28:11 CoreDump|home: I havent ever bult from branch Sep 05 11:31:04 XorA: didn't help Sep 05 11:31:28 and removing it from staging as well Sep 05 11:31:50 * XorA assume he would rebuld pkgconfig-natve Sep 05 11:32:00 and pkgconfig Sep 05 11:32:07 hmm wait Sep 05 11:32:17 koen: just did that Sep 05 11:32:48 XorA: -native is more or less == include pkgconfig.bb Sep 05 11:33:13 CoreDump|home: but you have to rebuld them to restage a differend pkg.m4 Sep 05 11:33:41 and hoping that fixes your mmediate problem Sep 05 11:33:43 did that =) Sep 05 11:34:05 hold on, I removed the file from staging Sep 05 11:34:29 XorA: http://angstrom-distribution.org/unstable/images/20060905/c7x0/Angstrom-gpe-image-test-20060904-c7x0.rootfs.tar.bz2 Sep 05 11:35:28 and of course http://www.angstrom-distribution.org/repo/?action=details&pnm=xserver-kdrive-w100 Sep 05 11:35:29 koen: fxes shm? Sep 05 11:35:30 * florian_kc looks at the GIT documentation... *sigh* Sep 05 11:35:32 XorA: koen woohoooo Sep 05 11:35:38 that worked! Sep 05 11:35:59 XorA: if you login as root, yes Sep 05 11:36:31 koen: http://www.angstrom-distribution.org/repo/?action=details&pnm=virtual%2Fxserver Sep 05 11:36:38 koen: do we need to mount shmfs? Sep 05 11:37:04 * koen <- shm n00b Sep 05 11:37:11 CoreDump|home: can you make proper changes in metadata and push them? Sep 05 11:37:19 koen: on my desktop tmpfs on /dev/shm type tmpfs (rw) Sep 05 11:37:24 hrw|work: yeah, will do Sep 05 11:37:31 thx CoreDump|home Sep 05 11:37:44 I have to thank you guys =) Sep 05 11:37:45 hrw|work: its a hack not a proper change :-( Sep 05 11:37:54 anyway, gotta go to work, later all Sep 05 11:38:56 now its time to recreate 3.5.4.2-rc2 opie images Sep 05 11:39:41 hrw|work: could you rebuild libmpeg2 and scummvm as well? Sep 05 11:39:50 that should close another set of bugs :) Sep 05 11:39:52 koen: can Sep 05 11:40:15 koen: ~lart shlibs probably Sep 05 11:40:32 koen: do you stage also shlibs in packaged-staging? Sep 05 11:40:55 shlibs code is still the same Sep 05 11:41:29 koen: if I will populate staging from packages what will happen to staging/shlibs/ data? Sep 05 11:42:59 that will stay they Sep 05 11:43:04 OE still generates it Sep 05 11:43:13 * koen adds that path to tarX Sep 05 11:50:19 echo 'BROKEN=1'>>libmpeg2*.bb Sep 05 11:54:11 ~lart lrg for sending RP the wrong patch Sep 05 11:54:11 * ibot rm -rf's lrg for sending RP the wrong patch Sep 05 11:55:16 CoreDump|afk: I started fresh build of gpe-image Sep 05 11:59:12 lrg: ~lart me is good too Sep 05 11:59:18 03hrw 07org.oe.oz354x * ra0969173... 10/ (1 packages/libmpeg2/libmpeg2_0.4.0b.bb): libmpeg2: do not link with X11 libs Sep 05 12:08:10 0/cr Sep 05 12:15:10 03hrw 07org.oe.oz354x * r9264df9d... 10/ (5 files in 4 dirs): Sep 05 12:15:10 rox-filer: add 2.5 Sep 05 12:15:10 * compiles, packages, but not tested on device Sep 05 12:15:10 applied changes from 70693973b64b18c9083bac36a2f7b2d374f3eba8 Sep 05 12:15:10 through d392153f540fb53432b333f95c079d221f6570f1 Sep 05 12:15:15 03hrw 07org.oe.oz354x * r5b7a4da9... 10/ (1 packages/rox/rox-filer_2.5.bb): rox-filer: added do_install() from .dev Sep 05 12:18:08 how can I see what brings in "gconf" as a dependency? Sep 05 12:18:38 likewise: bitbake -n target? Sep 05 12:20:19 I think it's -g Sep 05 12:20:31 Or is that the other way around? Sep 05 12:20:46 B_Lizzard: but -g doesn't work for me... bitbake 1.6 here Sep 05 12:21:43 It should... I use 1.6 and it does Sep 05 12:22:52 B_Lizzard: could you check the SVN revision of your bitbake? Sep 05 12:24:08 03hrw 07org.oe.oz354x * r7c2dd5e8... 10/ (1 packages/scummvm/scummvm_0.6.1b.bb): scummvm: bump PR due to rebuild which close #1386 Sep 05 12:24:37 Actually, it doesn't Sep 05 12:24:56 The option is there, but bitbake crashes Sep 05 12:25:47 gah Sep 05 12:26:20 anyone know, off hand, what I need to assign in the device struct before calling a register_driver? Sep 05 12:29:05 B_Lizzard: myFilterProvider only providing two args to findBestProvider() ? Same error here. I failed to fix it. Sep 05 12:29:24 hmm Sep 05 12:30:09 Yes, same error Sep 05 12:30:15 Try updating? Sep 05 12:30:38 One question, has anyone used gvim-7.0 ? Sep 05 12:30:49 The OE version, of course Sep 05 12:31:07 B_Lizzard: my bitbake is one hour old or so Sep 05 12:31:27 likewise Sep 05 12:31:33 Pun intended Sep 05 12:31:39 yes?? :-) Sep 05 12:31:43 :) Sep 05 12:31:50 Persistence of Vision(tm) Ray Tracer Version 3.6.1 (arm-angstrom-linux-gnueabi-g++ -march=armv5te -mtune=xscale 4.1.1 @ arm-angstrom-linux-gnueabi) Sep 05 12:33:20 if a .bb does "inherit gnome", which .bb file does it pull in? Sep 05 12:33:50 it's like doing bitbake gnome, I think Sep 05 12:34:40 likewise: classes/gnome.bbclass Sep 05 12:35:34 which does DEPENDS += "gnome-common" Sep 05 12:37:46 ROX-filer seems to work on my hx Sep 05 12:39:59 koen: thanks. Yes, gnome.bbclass inherits gconf. This is fairly recent (sometime in my last months of graduating :-)) Sep 05 12:41:00 I suspect http://www.openembedded.org/viewmtn/revision.psp?id=33001ff22f95feed0938740b8c095071b9518441 Sep 05 12:41:54 kone: could you (again...) tell me how you search these changes so quickly? Sep 05 12:42:33 koen: not kone :-) Sep 05 12:44:26 likewise: I have my news reader subscriber to http://www.openembedded.org/aggregator/rss Sep 05 12:44:53 so I type 'gconf' in the search box and it will give me a zillion hits from planet.gnome.org and a few from OE :) Sep 05 12:45:03 subscribed* Sep 05 12:45:50 koen: openembedded.org does not allow me to search in the commits, right? Sep 05 12:46:44 nope, and viewmtn lacks that functionality Sep 05 12:46:53 I should poke grahame for that Sep 05 12:47:02 since he's busy rewriting viewmtn Sep 05 12:51:28 morning Sep 05 12:52:41 any versioning gurus? Sep 05 12:52:41 lrg: ah :) Sep 05 12:53:31 mithro: http://www.openembedded.org/wiki/VersioningPolicy ? Sep 05 12:53:33 RP: looks like I've bzipped the wrong file. :( Sep 05 12:54:25 i was pondering how you handle cvs packages Sep 05 12:54:43 lrg: That means you might have a fixed version after all though? :) Sep 05 12:55:56 RP: yes, i think so. I noticed the debug was the same as yesterday. I'm also in the office now and the good patch is on my PC at home. I'm now recreating it Sep 05 12:57:50 is the OE VersionPolicy compatible with the debian policy Sep 05 13:00:36 mithro: not quite but I think that we should follow debian one Sep 05 13:01:14 i'm pondering how to version this deb Sep 05 13:12:20 PreviousUpstreamVersion+cvsSRCDATE iirc Sep 05 13:13:41 Why the hell is all my RAM used up when I compile rxvt? Sep 05 13:13:55 A bug, maybe? Sep 05 13:14:35 mithro: I didn't think the two were incompatible but I could be wrong. We were forced into a certain style of naming to ensure ipkg upgrades work Sep 05 13:17:05 yep, ipkg doesn't have epochs Sep 05 13:17:38 koen: also does not have ~ Sep 05 13:17:51 Debian allow 1.0~pre3-r9 Sep 05 13:34:49 lrg: which i.mx devboard do you have? Sep 05 13:36:01 koen: MX21ADS and MX31ADS Sep 05 13:38:23 it seems freescale allows one to sample the i.mx31 SoC Sep 05 13:38:32 NOTE: package glibc-2.3.5+cvs20050627-r3: task do_package: started Sep 05 13:38:46 tomorrow I will get gpe-image (if everything will go ok) Sep 05 13:38:59 $3,990.00 !?!?! Sep 05 13:39:29 and it isn't even leadfree Sep 05 13:40:59 RP: can you explain why this change was necessary? http://www.openembedded.org/viewmtn/revision.psp?id=33001ff22f95feed0938740b8c095071b9518441 Sep 05 13:41:44 bigd0g, you rang? Sep 05 13:42:14 I have some .bb files in a different dir than org.openembedded.dev..... how do I tell bitbake to use my packages if both possible Sep 05 13:42:16 ? Sep 05 13:42:17 Crofton: hiya ... been doing some research, think i might've figured it out ... but i'll ask Sep 05 13:42:35 ok Sep 05 13:42:39 Crofton: you have preferred resources for flashing the omap5912 osk board? Sep 05 13:42:51 preferred resources? Sep 05 13:43:02 tools / guides / etc Sep 05 13:43:15 he newbies guide works pretty good Sep 05 13:43:34 Hi! Sep 05 13:43:41 yeah, it wasn't giving me what i wanted... but maybe it will now that i've spent the better part of a day learning about how it works Sep 05 13:43:43 pavlix: bbcollections Sep 05 13:43:56 heh Sep 05 13:43:57 i'm interfacing via kermit over serial and am about to update uboot Sep 05 13:43:59 that happens Sep 05 13:44:06 ok Sep 05 13:44:09 pavlix: http://www.openembedded.org/user-manual&dpage=ch04s03 Sep 05 13:44:12 pavlix: BBPATH, BBFILES. but lazy me just symlinks into org.oe.dev/packages/ ;-) Sep 05 13:44:13 you have the tftp stuff going? Sep 05 13:44:35 tftp works, but i only have 1 ethernet on this computer... so it's either use kermit or tftp + no way to look stuff up Sep 05 13:44:55 ah Sep 05 13:44:59 I use tftp Sep 05 13:45:07 it seems everyone does ;) Sep 05 13:45:12 not familiar with pasing stuff over serial Sep 05 13:45:23 probably for a good resaon :) Sep 05 13:45:24 but... i figured out kermit with some digging Sep 05 13:45:38 apparently Ethernet is faster... who would've thought? Sep 05 13:45:45 :D Sep 05 13:45:45 yeah Sep 05 13:45:58 ok, I'll be in and out for a bit Sep 05 13:46:04 alrighty Sep 05 13:46:06 i'll keep you posted Sep 05 13:46:15 i was about to boot with the latest uboot from memory Sep 05 13:46:22 now just have to burn it to rom Sep 05 13:46:27 and pray i don't fry anything Sep 05 13:47:19 03mickeyl 07org.oe.dev * rc6ec5439... 10/ (1 packages/hnb packages/hnb/hnb_1.9.17+1.9.18pre7.bb): add hnb, a nice text-mode outliner Sep 05 13:49:16 btw, psokolovsky, I am now trying to build for wrt54, which should be very close to asus wl500g Sep 05 13:49:27 but uclibc fails to build Sep 05 13:49:33 pavlix: cool! Sep 05 13:49:56 psokolovsky: even openwrt is said to work on the asus Sep 05 13:50:00 pavlix: hm, oops. did you try to use prebuilt one for starters? Sep 05 13:50:05 * koen looks at conf/machine/wrt54.conf Sep 05 13:50:16 pavlix: yes, it should Sep 05 13:50:18 * chouimat look at koen Sep 05 13:50:41 hey chouimat Sep 05 13:50:43 I've been using oleg's firmware though Sep 05 13:51:07 pavlix: me too currently at it ;-) Sep 05 13:51:34 what is the best way to make changes into bb files so I can easily report it back Sep 05 13:51:55 shall I modify files in org.openembedded.dev? or use my own directory? Sep 05 13:52:25 pavlix: to backport fixes, inside org.openembedded.dev Sep 05 13:52:29 if you do it in org.oe.dev you can use 'mtn diff' to make patches for the bugtracker Sep 05 13:52:39 pavlix: to create new stuff use BitBake collections Sep 05 13:52:45 koen: good then Sep 05 13:52:55 zecke: bitbake collections? Sep 05 13:53:03 pavlix: check the BitBake manual :) Sep 05 13:58:15 zecke: you neam... http://www.openembedded.org/user-manual&dpage=ch04s03 ? Sep 05 13:58:48 pavlix: I meant http://bitbake.berlios.de/manual, but berlios has issues... Sep 05 14:01:00 and a poor documentation ;( Sep 05 14:02:49 Genesis: you still believe in documentation? ... the only place I saw good one is in microsoft sdk ... Sep 05 14:05:14 Genesis: the only good thing about the documentaion is: if you send a patch it will be on the site within the next 30 minutes Sep 05 14:05:45 (max 30 minutes after commit) Sep 05 14:05:51 hehe Sep 05 14:06:45 hi zecke mickeyl Sep 05 14:06:57 I could make it 'instantly', but os.system() is ugly Sep 05 14:07:27 option #2 is to hook it into the ciabot Sep 05 14:08:44 yes, berlios has issues :-) Sep 05 14:10:06 * chouimat hates forum.openwrt.org ... completely useless Sep 05 14:10:30 cheers chouimat Sep 05 14:10:38 it's about time we release our own WRT distro :/ Sep 05 14:10:59 hmm... we are on topic Sep 05 14:11:02 mickeyl: yup ... Sep 05 14:11:20 do we have a read arm bigendian toolchain around? Sep 05 14:11:21 mickeyl: who will work on it? Sep 05 14:11:31 'read'? Sep 05 14:11:34 florian_kc: yes, the nslu2 is BE Sep 05 14:11:37 florian_kc: the NSLu2 is BE Sep 05 14:11:50 (it can do LE as well) Sep 05 14:11:53 hrw|work: i would once we have anything that boots :) Sep 05 14:12:02 cool - thats a IXP4xx cpu? Sep 05 14:12:08 yes Sep 05 14:12:10 florian_kc: yes Sep 05 14:12:26 florian_kc: wget http://www.nslu2-linux.org/Makefile Sep 05 14:12:27 * koen should flash openslug 3.10 on his loft Sep 05 14:12:28 florian_kc: make Sep 05 14:12:28 hmm cool Sep 05 14:12:31 hrw|work: give me good documentation on how to add a) new board b) no distro and I will also ask for c) a new girlfriend (hot chick) and I will be happy ;) Sep 05 14:12:55 many thanks Sep 05 14:13:06 chouimat: hot solely depends on your taste Sep 05 14:13:14 chouimat: and maybe try boys as well? Sep 05 14:13:18 zecke: Speaking of patches, you asked to ping you regarding svn fetcher, http://bugs.openembedded.org/show_bug.cgi?id=1367 . but it's not top priority of course. Sep 05 14:13:30 florian_kc: * you need the intel IXP libs as well (click through download) Sep 05 14:13:37 psokolovsky: stop bitching ;) Sep 05 14:13:50 zecke: lol ;-) Sep 05 14:13:58 psokolovsky: how long will you be around? Sep 05 14:14:01 zecke: the need for those might soon go away Sep 05 14:14:16 chouimat: do you not have 1000s of hot russian babes waitng for you? Sep 05 14:14:27 where??? Sep 05 14:14:28 zecke: for couple of hours at least Sep 05 14:14:43 mickeyl: mickeyl? Sep 05 14:14:48 *cough* Sep 05 14:15:05 mickeyl: I'm single, you are not Sep 05 14:15:06 zecke: that was the other part of me Sep 05 14:15:34 XorA: nope ... Sep 05 14:15:43 hmm, it seems doing anything with bitbake after changing bb files is quite slow Sep 05 14:15:47 chouimat: maybe its only me gets those emals then :-) Sep 05 14:16:25 XorA: I get somes but you have to pay before being able to do the test drive ... Sep 05 14:16:53 mickeyl: wrt54 kernel has lot of patches iirc Sep 05 14:16:56 anyway brb need to go buy some food (store were closed yesterday here) Sep 05 14:17:04 * chouimat is away: shopping Sep 05 14:17:41 chouimat: though you should know, except in rare cases "hot chicks" and computers are incompatible Sep 05 14:19:17 hrw|work: *nod* but we could "just" drag them in can't we? Sep 05 14:19:25 hrw|work: didn't you have an OE image booting once? Sep 05 14:19:27 or who was that Sep 05 14:19:36 mickeyl: br1 Sep 05 14:19:40 oh bruno Sep 05 14:19:41 righto Sep 05 14:19:49 and he cannot work on wrt54 Sep 05 14:19:50 he's gone to a world trip since over a year Sep 05 14:20:11 and he work(ed) for meshcube Sep 05 14:20:24 the meshcube will be soon maintained again Sep 05 14:20:28 the new maintainer contact me Sep 05 14:20:35 his key has been added Sep 05 14:20:39 nice Sep 05 14:20:46 zecke: ok, let me try... Sep 05 14:21:03 mickeyl: oh, good news Sep 05 14:23:47 mickeyl: wrt54 development req wrt with serial ports added and jtag Sep 05 14:24:39 when you're talking about wrt54... Sep 05 14:24:54 do you know how to get uclibc compile for it? Sep 05 14:25:45 pavlix: DISTRO=wrtoe or generic-uclibc Sep 05 14:27:01 hrw|work: thanks, I'll try Sep 05 14:29:21 do13: you around? Sep 05 14:29:29 hrw|work: I used DISTRO=wrt54oe and MACHINE=wrt54 Sep 05 14:29:57 do openwrt use openembedded? Sep 05 14:30:13 mikearthur: no Sep 05 14:31:53 pavlix: thats also good combo Sep 05 14:32:10 mikearthur: no, they use buildroot iirc and now try soemthing new Sep 05 14:32:23 hrw|work: didn't work... I'm now updating oe Sep 05 14:32:27 hrw|work: really? why? do we have to upload a new bootloader? Sep 05 14:33:20 hrw|work: where do I find DISTRO config for wrtoe? Sep 05 14:33:35 i see only wrt54oe Sep 05 14:33:58 hrw|work: good combo, but not working :-( Sep 05 14:34:44 mickeyl: what if we create image which will fsck machine? Sep 05 14:34:54 pavlix: I used it to build packages for my wrt54gs Sep 05 14:35:53 hrw|work: my device is actually asus wl500gp but... by now I didn't manage to compile anything reasonable Sep 05 14:35:55 hrw|work: hmm... how should that happen? Sep 05 14:36:06 hrw|work: it should be mostly compatible Sep 05 14:36:42 RP: you around? Sep 05 14:38:15 mickeyl: typo in some startup script, module/kernel mismatch, etc Sep 05 14:38:28 ... and? Sep 05 14:38:35 flash new image w/ bootloader Sep 05 14:38:47 it's cumbersome but i don't see where it's dangerous Sep 05 14:38:59 as longs as we don't overwrite nvram / bootloader Sep 05 14:39:09 yeah Sep 05 14:39:34 but if the network doesn't come up it is dangerous enough for $newby-user Sep 05 14:39:56 :-) Sep 05 14:40:09 sure thing. but... we are talking experts here, aren't we? Sep 05 14:40:18 I used to think so Sep 05 14:40:51 I think OE has gotten too simple to set up ;) Sep 05 14:40:55 * koen hides Sep 05 14:41:38 bug 59955: OE is too easy to use! It must suck! Sep 05 14:41:42 * chouimat is back. Sep 05 14:41:50 :-) Sep 05 14:42:10 if it was too easy to set up, I'd be already putting it in the device Sep 05 14:43:05 * chouimat need to update the firmware on his wrtsl54gs Sep 05 14:43:43 just a question about wrt... do I need wrt-imagetools-native if I don't want to create an image? Sep 05 14:47:31 Is there a way to get psyco-like speedups in a 64 bit environment? Pypy is mentioned at the psyco web site but doesn't look ready yet Sep 05 14:48:28 unfortunately not. the key to more parsing speed is the c-parser, there's nothing that comes close Sep 05 14:49:22 i suppose I don't Sep 05 14:49:30 didn't mithro and zecke have that almost working a while ago? Sep 05 14:49:42 koen: define almost Sep 05 14:49:51 koen: it doesn't segfault anymore Sep 05 14:49:56 :) Sep 05 14:50:00 chouimat: wrtsl... nice toy Sep 05 14:50:02 koen: and using pyrex looks like a good idea Sep 05 14:50:06 zecke: 'parses the testcases in svn' Sep 05 14:50:11 pavlix: i guess you don't probably that comes from a wrong/old dependency. Sep 05 14:50:23 koen: no, it can't but it is closer than before Sep 05 14:50:29 mickeyl: looks lie a bug in depencencies, yes Sep 05 14:50:33 koen: currently it needs 16 seconds to parse the OE tree Sep 05 14:50:41 koen: but it doesn't add tasks Sep 05 14:50:59 mickeyl: I put it in ASSUME_PROVIDED Sep 05 14:54:30 hrw|work: yup ... but openwrt sucks on it so I use dd-wrt :( Sep 05 14:54:39 we don't need to stinking tasks Sep 05 14:54:53 not openopenwrt? Sep 05 14:54:59 or freewrt? Sep 05 14:55:04 or netwrt? Sep 05 14:55:23 hrw|work: and I use it as my router ... on my old p200mmx I got 300KB/s with the wrtsl I get 425KB/s :) Sep 05 14:55:33 koen don't know them :) Sep 05 14:56:29 or darWRTin Sep 05 14:56:37 * koen stops with the bsd jokes Sep 05 14:56:54 koen www.freewrt.org Sep 05 14:57:16 chouimat: I installed openwrt and it was ok for me so did not searched for other. Sep 05 14:57:52 chouimat: does dd-wrt has panel to manage bandwidth and priorities for services? (cbq or how it was called) Sep 05 14:59:14 hrw|work: only basic QoS Sep 05 14:59:33 hrw|work: maybe it somethinfg better now .. Sep 05 14:59:45 chouimat: I'm a DD-WRT man too Sep 05 15:01:13 what has dd-wrt over openwrt? Sep 05 15:01:51 wow Sep 05 15:01:52 http://www.dd-wrt.com/shop/catalog/product_info.php?products_id=33&osCsid=3ad472edd8017f450fdda3e907c60313 Sep 05 15:01:59 * koen adds to wishlist Sep 05 15:02:26 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Sep 05 15:09:15 * mrz80 is back (gone 00:06:48) Sep 05 15:24:34 bah... so frustrating. i had file transfer over serial via kermit working like 2-hrs ago Sep 05 15:24:39 now it doesn't work Sep 05 15:24:47 must be setting a parameter incorrectly Sep 05 15:24:47 grr Sep 05 15:31:16 cu Sep 05 15:36:23 what's wrong, I'm building uclibc.... Sep 05 15:36:38 and it's complaining it cannot find crt1.o Sep 05 15:37:38 there's org.openembedded.dev/tmp/work/uclibc-0.9.28-r6/image/usr/lib/crt1.o Sep 05 15:41:20 grmbl!!! Sep 05 15:44:09 anyone know where i can find a kermit expert ? Sep 05 15:44:17 'kermit' as in c-kermit Sep 05 15:44:41 bigd0g: Sesamee Street? Sep 05 15:45:38 chouimat: what's that? Sep 05 15:50:41 pavlix: nothing ... just one of the guy I'm working with ... you know energy sucking type of person ... and he's the manager of the project Sep 05 15:51:25 chouimat: :-) Sep 05 15:52:10 pavlix: he's a electrical engineer and he doesn't know anything about programming so imagine the fun I have Sep 05 15:52:12 chouimat: I'm loosing all my energy on broken oe's uclibc Sep 05 15:52:43 chouimat, : i've a ton of this here :) Sep 05 15:53:22 pavlix, : and me on the initscrips Sep 05 15:53:35 and busybox issue Sep 05 15:53:37 Genesis: you're several steps farther Sep 05 15:54:38 Genesis: i just can't enjoy the fun of building a single working thing Sep 05 15:55:09 Genesis: spent a day explaining to them why I decided to use oe instead of home make buildsystem ... so after this I got in a pub and got really drunk ... Sep 05 15:56:36 i spend my time to justify the choice of OE instead of buildroot Sep 05 15:56:58 and when i see the initscript and few bb files , i lost my argues :( Sep 05 15:58:44 Genesis: I had to do that one too ... and since I'm the only programmer on the team I told them I choose what I want to work with and if they weren't happy I will simply quit the project since I won't see any money for at least a year, and I also keep the copyright of all the code I will do ... so if they piss me too much I can do whatever I want with it ;) Sep 05 15:58:46 and it's my first time with OE , it's a bit complex because we're on dead line ): Sep 05 15:59:14 right chouimat :) Sep 05 15:59:26 chouimat: I'm an EE without coding skills as well :) Sep 05 15:59:43 koen but you don't piss me off ;) Sep 05 16:00:25 ^^ Sep 05 16:02:44 Genesis: and since the hardware team is in quebec city and I'm in ottawa ... they allways find the right moment to call me ... aka lunch or me with someone ... Sep 05 16:02:48 someone experiment 2 sessions of kdrive with 2 differents fb controller and touchscreen ? Sep 05 16:04:47 I saw some references to xinerama in the new xorg kdrive Sep 05 16:05:11 but I don't know how it will handle 2 screens with one input device Sep 05 16:05:15 * chouimat decides to make some coffee before going back to work on this project managed by a paranoid guy who just found Jesus Sep 05 16:05:16 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Sep 05 16:06:08 koen, : the stupid client doesnt want a xinerama mod Sep 05 16:06:37 "we care about swithching windows between the screen etc..." Sep 05 16:06:46 "the impossible, yesterday and for free" Sep 05 16:07:01 you're right Sep 05 16:07:45 and worse of that , we shall work with ugly broken kernel from freescale Sep 05 16:10:05 well. i'm giving up for today... i'll battle with this some more tomorrow Sep 05 16:10:09 pz, guys Sep 05 16:10:53 OE 1 / bigd0g 0 ... Sep 05 16:11:27 same for me Sep 05 16:11:29 a+ Sep 05 16:11:43 03justinp 07org.oe.dev * r4e738615... 10/ (1 contrib/mtnpatch.py mtn:execute true): mtnpatch: resurrect mtnpatch Sep 05 16:14:18 Hello Sep 05 16:14:39 hey sirfred Sep 05 16:15:10 koen: hi Sep 05 16:15:18 sirfred: http://ewi546.ewi.utwente.nl/tmp/libw100_svn.bb Sep 05 16:15:29 koen: hey! Sep 05 16:15:33 but I suspect you already have something like that :) Sep 05 16:15:39 koen: No, I haven't. Sep 05 16:16:42 koen: I'm changing libw100 just now to make it multiprocess safe. Sep 05 16:17:19 koen: But as it's still not used from any other project, I didn't setup a bb. Sep 05 16:17:50 koen: I'm thinking about dropping the old qte aticore based driver and make a new one based in libw100. Sep 05 16:18:25 koen: But I don't know if it actually deserves the effort. Sep 05 16:18:38 Xw100 is doing nicely on my hx4700 Sep 05 16:18:41 (w3220) Sep 05 16:19:03 although xw100 doesn't use libw100 Sep 05 16:19:16 koen: No, anyway Mardy shared the same information. Sep 05 16:19:42 koen: Well, I created libw100 as a reference library. Sep 05 16:20:10 koen: Just to test things. But now, I don't know what's better to say you the truth. Sep 05 16:20:52 koen: Centralize all the accel stuff in a library, that could be shared by different projects, or just use it as a reference, and then implement different drivers for any use (X, Qte, SDL,..) Sep 05 16:21:24 koen: I don't know how compatible w3220 and w100 are, thought. Sep 05 16:21:41 koen: It seems that at least basic raster operations are compatible. Sep 05 16:21:50 xw100 works, so I'm happy :) Sep 05 16:22:23 That's great. :) Sep 05 16:22:31 RP: ping Sep 05 16:23:06 ~seen RP Sep 05 16:23:23 hi sirfred Sep 05 16:23:27 rp is currently on #gpe #oe #openzaurus #handhelds.org. Has said a total of 150 messages. Is idling for 3h 8m 52s, last said: 'mithro: I didn't think the two were incompatible but I could be wrong. We were forced into a certain style of naming to ensure ipkg upgrades work'. Sep 05 16:23:33 RP: Hi. :) Sep 05 16:23:43 RP: Did you have time to see my mail+patch for w100fb ? Sep 05 16:24:02 sirfred: I did although I haven't had time to stufy exactly what you were doing there Sep 05 16:24:22 RP: Well, a lot of dangerous things, of course. :) Sep 05 16:24:47 RP: I feel (perhaps placebo effect) the scroll a little better. Sep 05 16:24:53 sirfred: We need to find out exactly what the framebuffer sync call requires Sep 05 16:25:06 i.e. what its there to ensure Sep 05 16:25:17 RP: In my opinion is just to ensure that no hardware operations are running. Sep 05 16:25:32 RP: It's needed to start an operation in non hardware mode. Sep 05 16:26:06 RP: If you don't wait for the graphics engine to be idle, your software operation can conflict and/or mix with the running hardware one. THe software operations are inherently synchronous. Sep 05 16:26:14 sirfred: I think its also to ensure that all outstanding data is in sync so userspace access to the framebuffer sees the correct data Sep 05 16:26:52 RP: Don't know about that. Sep 05 16:27:26 RP: But when we didn't have hardware acceleration, we didn't need a sync implementation, do we? Sep 05 16:27:42 sirfred: no Sep 05 16:28:35 RP: So, what have we changed? That now, before every cfb_imageblit we are calling w100fb_sync Sep 05 16:28:48 RP: That's not needed , in my opinion. Sep 05 16:29:19 RP: Only when the last operation was in hardware, we should actually assure that the graphic engine is idle to start the next software operation. Sep 05 16:29:34 later Sep 05 16:30:53 RP: It's a theory, of course. But I've been using this modified framebuffer driver some days, and have had no problems. Sep 05 16:31:45 RP: Unfortunately, I've failed trying to compile oprofile. Waiting for the stars to be aligned to have all the required dependencies compiled. Sep 05 16:34:09 sirfred: I agreed we can probably optimise it, we just need to make sure of that :) Sep 05 16:34:37 RP: Of course. Sep 05 16:34:42 sirfred: oprofile should compile ok - where did you have problems? Were you using OE? Sep 05 16:35:04 RP: Yes. well, with glibc, and other dependencies. Now, it seems that glibc is compiling fine. Sep 05 16:35:13 RP: So, let's see this time. Sep 05 16:35:24 sirfred: Ah, right. glibc can be problematic :) Sep 05 16:35:35 RP: Is it difficult to use it? Sep 05 16:41:53 mickeyl: Have you tried adding IPython to oe ? Sep 05 16:42:03 bbl Sep 05 16:42:05 sirfred: Not really. There is one perverse error message to watch out for - basicaly make sure to install the binutils symlinks package or it will say your vmlinux file is invalid Sep 05 16:42:58 RP: Thanks for the advise Sep 05 16:46:10 sirfred: I should really fix that message :) Sep 05 16:46:22 about the 'bitbake sanitze ' (see http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-September/000159.html) Sep 05 16:46:31 here is a first script attempt to sanitize bb files: http://rafb.net/paste/results/oO8gTW27.html Sep 05 16:46:51 you need also odict.py downloadable here: http://www.voidspace.org.uk/python/odict.html Sep 05 16:47:18 feedbacks are welcome :) Sep 05 16:56:15 koen@bitbake:~/OE/monotone/org.openembedded.dev$ ./sanitize.py packages/gpe-shield/gpe-shield_0.9.bb Sep 05 16:56:16 : No such file or directory Sep 05 16:56:23 cyrilRomain|bbl: any idea on that? Sep 05 16:57:44 koen: hum .. no Sep 05 16:59:06 ah Sep 05 16:59:12 python sanitize.py packages/gpe-shield/gpe-shield_0.9.bb works Sep 05 16:59:54 ah, ok, it's #!/usr/bin/env python that don't work for you Sep 05 17:02:52 cyrilRomain: I also added RRECOMMENDS and RSUGGESTS to the styleguide Sep 05 17:03:09 ok Sep 05 17:04:26 koen: as I sais in the mail, I think I will make a list of common variables used in .bb, so that you (or mickeyl, or...) could sort them and add them to the wiki Sep 05 17:04:38 s/sais/said Sep 05 17:05:51 koen: feel free to tell me on what packages you find big problem with my script Sep 05 17:06:07 s/problem/problems/ even Sep 05 17:08:22 03koen 07org.oe.dev * rcf8c1ed4... 10/ (15 files in 5 dirs): gpe-shield: test run of Cyril's sanitizer and drop obsolete versions Sep 05 17:25:09 * mrz80 is back (gone 01:19:53) Sep 05 17:29:17 koen: be care that this version of sanitize removes comments that are outside "" or {} :-/ Sep 05 18:32:02 * chouimat is away: stuff Sep 05 18:38:55 kergoth: I still do not know the code, but I fear it is tries to apply the same patch twice ;) Sep 05 18:39:03 kergoth: this can be seen with quilt-native Sep 05 18:40:38 NOTE: package quilt-native-0.45-r0: task do_patch: started Sep 05 18:40:38 NOTE: Applying patch 'install.patch' Sep 05 18:40:38 NOTE: package quilt-native-0.45-r0: task do_patch: completed Sep 05 18:40:49 i dont know what you mean Sep 05 18:40:50 kergoth: hehe Sep 05 18:41:07 kergoth: I will raise some bucks so you can afford a BSD ;) Sep 05 18:41:29 when force=True, it calls patch once. when force=False or None, it calls it once with dry-run, once without Sep 05 18:42:01 kergoth: I apply install.patch and 'non-gnu.patch' for quilt-native Sep 05 18:42:14 kergoth: and after "NOTE: Applying patch 'non-gnu.patch'" Sep 05 18:42:19 kergoth: I receive a "Reversed (or previously applied) patch detected! Assume -R? [n] " Sep 05 18:42:37 kergoth: which IMHO indicates that instead of applying non-gnu.patch it applies install.patch again Sep 05 18:43:15 kergoth: I will try to look into this myself soon(tm) Sep 05 18:43:19 grr, damnit Sep 05 18:45:30 finally OE compiles again and i can check if my macbook shutsdown randomly *again* Sep 05 18:45:58 ouch Sep 05 18:51:30 zecke: what's the state of your qcop-to-x11 thingie? Sep 05 18:51:32 kergoth: any idea if it could apply the patch twice? Sep 05 18:51:51 or 'zecke's qcop substitution API' as mickey|sports calls it Sep 05 18:52:09 koen: from a point of view from today. Use d-bus Sep 05 18:52:16 zecke: it is. i dont know why, but it is. investigating Sep 05 18:52:32 koen: and this Zecke COP is not endian aware :( Sep 05 18:52:41 zecke: I'm toying with the idea of building opie/x11 Sep 05 18:52:55 zecke: or atleast opie-pim Sep 05 18:53:06 koen: let us wait for the day Trolltech releases Qtopia4 GPL Sep 05 18:53:11 since the multisync plugin seems to be getting usefull Sep 05 18:53:32 building busybox is ridiculous Sep 05 18:53:58 [16:36] mickeyl: gpl'ing qtopia4 could have had a refreshening effect on development Sep 05 18:53:58 [16:36] mickeyl: _if_ it would have happened last year Sep 05 18:54:13 koen: that is from 2004? Sep 05 18:54:21 heh Sep 05 18:54:27 no from this afternoon Sep 05 18:54:59 koen: isn't the o-hand evo stuff getting along quite nicely? Sep 05 18:55:11 it is Sep 05 18:55:27 koen: now OpenSync should just be able to utilize d-bus over a ssh link? Sep 05 18:55:28 nokia is using it in their products Sep 05 18:55:41 koen: or isn't it running a e-d-s? Sep 05 18:55:49 zecke: dbus over ssh ... ewwww Sep 05 18:55:52 it's running e-d-s Sep 05 18:55:56 * chouimat is away: Gone away for now. Sep 05 18:56:06 * chouimat is away: Gone away for now. Sep 05 18:56:09 chouimat: I tunnel everything over ssh Sep 05 18:56:14 chouimat: I love to do ppp over ssh Sep 05 18:56:21 * chouimat is away: re Sep 05 18:56:32 zecke: http://svn.o-hand.com/view/misc/trunk/sync/ Sep 05 18:56:46 zecke: I'm using jabber for thing like dbus over ssh ;) Sep 05 18:57:16 koen: how far is chris? Sep 05 18:57:21 no idea Sep 05 18:57:31 koen: Evo2<->Evo2 sync should work perfectly Sep 05 18:57:41 koen: e.g. no messing with matching attributes Sep 05 18:57:56 yes, but the juice is in a nice gui and autodiscovery Sep 05 18:58:12 koen: but for Opie on X11, I would be glad if you would start it (again) Sep 05 18:58:32 koen: or port it to Qt4 or Gtk+ directly Sep 05 19:07:29 zecke: I think I'm play with the loft instead of porting opie to something sane Sep 05 19:09:08 insert 'going to' somewhere above Sep 05 19:16:49 does 'RRECOMMEND' exists (without the 'S' at the end) ? Sep 05 19:18:05 I guess not although some .bb file use 'RRECOMMEND =' Sep 05 19:18:06 cyrilRomain: if you have seen it, yes Sep 05 19:18:15 cyrilRomain: it has no semantic though Sep 05 19:18:18 zecke: ah, ok Sep 05 19:19:06 zecke: by 'no semantic' you mean it should be 'RRECOMMENDS' so that bitbake takes it into account, right ? Sep 05 19:19:56 cyrilRomain: right :) Sep 05 19:20:13 zecke: ok, so maybe it is just a typo in some .bb Sep 05 19:25:32 zecke: there ya go, should work now Sep 05 19:25:36 stupid, stupid bug Sep 05 19:25:43 kergoth: don't worry Sep 05 19:25:56 kergoth: it needed a moron to discover it Sep 05 19:26:35 no, it needed someone to actually test more than one patch in an src_uri with PatchTree :P Sep 05 19:26:48 kergoth: I checked my bank account and I could afford creating a qemu image with a BSD of your choiche Sep 05 19:26:58 kergoth: and even pay for the digital transfer Sep 05 19:27:04 03clarson 07org.oe.dev * r0c337b44... 10/ (1 classes/patch.bbclass): patch.bbclass: Fix bug in PatchTree.Import resulting in new patches being imported -before- the current patch rather than -after-. Sep 05 19:27:08 kergoth: if you are interested, give me a bing Sep 05 19:31:13 "grep -r 'RRECOMMEND' * | grep -v 'RRECOMMENDS' | grep packa" from org.openembedded.dev gives the list of packages that use 'RRECOMMEND'. Just in case someone wants to update those .bb :) Sep 05 19:31:18 zecke: s/moron/Deutsche Gründlichkeit/ ? Sep 05 19:31:59 lol Sep 05 19:39:23 cyrilRomain: I fixed the fuse stuff, thanks for spotting it Sep 05 19:39:50 * koen nudges eFfeM-log to add the S to RECOMMEND(S) in pvr-mci*.bb Sep 05 19:40:15 03koen 07org.oe.dev * r93f38c27... 10/ (5 files in 2 dirs): fuse: fix RRECOMMEND(S) Sep 05 19:42:39 packages/gpe-buttonbox/gpe-buttonbox_0.5.bb and packages/gpe-buttonbox/gpe-buttonbox_cvs.bb needs s/DESCIPTION/DESCRIPTION/ Sep 05 19:43:40 and packages/tilibs/libticalcs_4.5.9.bb needs s/DPENDS/DEPENDS/ Sep 05 19:44:04 cyrilRomain: feel free to write a bittest for this stupidity Sep 05 19:44:57 ah Sep 05 19:45:13 * cyrilRomain takes a deeper look at bittest Sep 05 19:45:59 zecke: should the sanity check be part of bittest (in near future) ? Sep 05 19:46:22 cyrilRomain: no Sep 05 19:46:31 cyrilRomain: bittest should test the meta data itself Sep 05 19:46:38 heh Sep 05 19:46:41 cyrilRomain: where sanity checks if the configuration is sane Sep 05 19:46:47 it seems all the typos are in stuff I added Sep 05 19:46:49 zecke: ok Sep 05 19:46:51 * koen hides in shame Sep 05 19:47:08 koen: haha Sep 05 19:47:25 koen: if the dutch beer wouldn't be light beer I would suggest to stop to drink Sep 05 19:49:11 zecke: so I guess my script is misnamed (sanity.py). It just sort variables and routines used in a .bb to follow the Style Guidelines Sep 05 19:49:23 anyway ... Sep 05 19:49:39 cyrilRomain: ah, I thought you meant sanity.bbclass Sep 05 19:49:51 zecke: no, about the sanity script Sep 05 19:49:53 cyrilRomain: folding sanity.py into a bittest module would be appreciated Sep 05 19:50:08 zecke: ah ok, I'll have a look Sep 05 19:50:37 zecke: since my script is in python, this will be easy to integrate I think :) Sep 05 19:51:02 * cyrilRomain plays with bittest now Sep 05 19:58:36 03clarson 07org.oe.dev * r331c7967... 10/ (1 classes/base.bbclass): Add 'rebuild' task to base.bbclass, as it's quite useful and simple to add. Sep 05 19:58:44 03koen 07org.oe.dev * r727e7196... 10/ (3 files in 2 dirs): gpe-buttonbox: fix embarrasing typos Sep 05 19:58:58 koen: a list of variables found in .bb files that are not in the Style Guidelines: http://rafb.net/paste/results/kFihqd62.html . Thanks in advance to insert them into the wiki and to tell me the one to discard. Sep 05 20:00:05 could you send that list to the mailinglist? Sep 05 20:00:20 koen: ok, I'll do a more complete list then Sep 05 20:00:55 koen: because the script dumped too many unknown variables Sep 05 20:37:13 giel: my ALC/FLUTE client and server is shaping Sep 05 20:37:57 giel: I hope to crash DoS MAD/MCL/Papageno with my server as I can do all TOI and TSI sizes, create custom Extension Headers, Wrap Around FEC Instance ID... Sep 05 20:57:11 zecke: nice! i'm wondering how the S3 stack holds up against that Sep 05 21:48:43 night Sep 05 23:41:38 DOes anyone have any experience with OE on Mac OS X? Sep 06 00:14:51 RP: problems with 2.6.17. ALSA is crapping out on resume. current playback hangs, then dies, and (with mpd) next song starts Sep 06 00:17:09 RP: my SD card is also not coming up since I flashed 2.6.17. I will try rebooting to see if it helps Sep 06 00:17:28 CoreDump|afk: altboot's sound is failing with 2.6.17.... Sep 06 00:18:25 * CoreDump|afk mumbles something about b0rked alsa drivers and heads to bed Sep 06 00:22:13 RP: I/O errors on many blocks in dmesg after boot......rebooting doesn't help either problem Sep 06 00:22:23 CoreDump|afk: I figured...just thought I'd let you know Sep 06 00:25:14 ~lart these darn kernel updates Sep 06 00:25:14 * ibot judo chops these darn kernel updates **** ENDING LOGGING AT Wed Sep 06 02:59:56 2006