**** BEGIN LOGGING AT Wed Sep 14 02:59:56 2005 Sep 14 03:29:33 AvengerMoJo: The 2.6 keymap for the cxx00 works and always has done. There was some init issues which meant it didn;t load but that is all fixed Sep 14 03:34:27 RP: really :) that's cool :P Sep 14 03:36:06 pb_: SRC_URI, and provide some kind of do_munge() <--- this will run before configure right Sep 14 03:36:59 AvengerMoJo: add "addtask do_munge() after do_patch()" into recipe Sep 14 03:37:11 hrw|work: thanks Sep 14 03:37:18 exact syntax in many examples Sep 14 03:37:24 ok :) Sep 14 03:37:29 kdepimpi/kdepimpi-base.inc is one of them Sep 14 03:37:39 hrw|work: thanks alot Sep 14 03:38:11 np Sep 14 03:40:14 hrw|work: where should i put the patch.tarball? Sep 14 03:40:25 its your patches? Sep 14 03:40:27 hrw|work: should that be src/ Sep 14 03:40:33 hrw|work: not really Sep 14 03:40:51 hrw|work: the driver add on to the pcmcia package Sep 14 03:41:00 could be from the web Sep 14 03:41:02 AvengerMoJo: when you add it into SRC_URI then it will be fetched and unpacked into ${WORKDIR} Sep 14 03:41:53 ... hrw|work what if I want to unpack it inside the {WORKDIR}/pcmcia-cs/ Sep 14 03:42:34 change archive Sep 14 03:42:58 oh Sep 14 03:43:21 so I should forget the patch but rebuild the source package you mean Sep 14 03:43:45 keep in mind that your do_munge() task can do anything.. it can cd ${WORKDIR}/pcmcia-cs/ and run patch from it etc Sep 14 03:44:21 hrw|work: ah ... so I just don't get the patch tarball in file and extra it from src directly Sep 14 03:44:26 hrw|work: thanks Sep 14 03:45:59 hrw|work: is there any label for src ? like {SOURCEDIR} Sep 14 03:46:32 WORKDIR is a place where your source is unpacked Sep 14 03:51:06 hrw|work: thanks Sep 14 04:13:26 hi Sep 14 04:15:12 hi dirk Sep 14 04:17:17 do13: hi Sep 14 04:17:34 do13: how can I help? (tosa kernel development) Sep 14 04:18:40 hi hrw|work, zecke_ Sep 14 04:18:58 hi dirk Sep 14 04:20:29 zecke_: most of the open points (charging) needs a new WM9712 driver Sep 14 04:20:39 hi RP Sep 14 04:23:16 oh man bitbakes debugging capabilities suck... Sep 14 04:23:59 zecke_: and USB host (internal wifi) is also unsupported :-) Sep 14 04:25:59 zecke_: what do you like to do? Sep 14 04:26:42 do13: everything sounds like fun. I guess I will tackle charging Sep 14 04:26:54 do13: is that the same chip as the c7xx has? Sep 14 04:27:53 zecke_: it's another chip Sep 14 04:28:29 zecke_: The tosa uses the AC97 audio codec for reading the battery status Sep 14 04:30:06 zecke_: This is a multifunctional device: http://www.wolfsonmicro.com/products/touchscreen/adcs/WM9712/ Sep 14 04:31:15 do13: headphone detection is done by that chip as well? Sep 14 04:32:10 zecke_: yep Sep 14 04:34:09 well, then I will tackle that Sep 14 04:34:18 do13: do you live in berlin/germany? Sep 14 04:36:12 zecke_: nope. I'am from hessen (between kassel and fulda) Sep 14 04:36:43 I'm sorry ;) Sep 14 04:39:05 Doesn't matter. Berlin is only 400Km away:) Sep 14 04:39:29 is fulda on the border? Sep 14 04:40:08 koen: no. It's central in germany. Sep 14 04:41:28 koen: In the past it was near the border (DDR). Sep 14 04:42:05 aha Sep 14 04:43:53 damn I looked at the Lineo Sourcecode... I feel raped Sep 14 04:46:32 zecke_: which file? Sep 14 04:46:45 do13: tosa*.c Sep 14 04:47:38 zecke_: clearly you are not a man to learn from the mistakes of others Sep 14 04:48:35 pb_: others are on crack anyway Sep 14 04:50:04 Why do you think they started doing crack? Sep 14 04:50:16 They saw the Lineo code Sep 14 04:53:58 I wand dtrace badly Sep 14 04:57:27 pb_: specially I seldomly learn from my own mistakes :} Sep 14 04:58:21 zecke_: I can annoy less people now Sep 14 04:58:34 zecke_: my nano arrived :) Sep 14 04:59:26 hehe does it run linux yet Sep 14 05:00:21 heh Sep 14 05:00:34 no, and I haven't added support for it to OE neither Sep 14 05:02:43 slacker Sep 14 05:08:47 koen: can you build billiardz with libsdl-x11? Sep 14 05:10:28 hrw|work: I suspect billiardz is a real qte app Sep 14 05:11:33 koen: currently I have "aliens billiardz freedroid gemdropx powermanga prboom rott scummvm vectoroids xmame" linked with libsdl-qpe - need to test which one are pure sdl ones Sep 14 05:12:00 prboom rott scummvm are pure sdl apps Sep 14 05:12:56 koen: when I asked Al to run Linux on her iPOD... Sep 14 05:13:15 she gave you a very funny look? Sep 14 05:15:58 koen: she could not get, why one would destroy perfectly working systems... Sep 14 05:16:14 koen: I tried her talking into dual booting Sep 14 05:16:44 32mb ram and a 75MHz chip Sep 14 05:16:54 did you mention it can play videos with linux? Sep 14 05:17:39 koen: freedroid is pure too Sep 14 05:17:52 Depends: libsdl-image-1.2-0 (>= 1.2.3), libpng12-0 (>= 1.2.8), libz1 (>= 1.2.3), libjpeg62 (>= 6b), libsdl-x11 (>= 1.2.7), libc6 (>= 2.3.5+cvs20050627), libgcc1 (>= 3.4.4) Sep 14 05:18:16 koen: but recording audio, and play more games Sep 14 05:18:25 zecke_: the ipods even have a *dual* arm7tdmi Sep 14 05:18:27 koen: my ibook maybe arrvies on friday Sep 14 05:18:41 koen: 1gb ram, 80gb harddisk Sep 14 05:18:44 *yum* Sep 14 05:19:02 nice Sep 14 05:19:18 almost like my powerbook Sep 14 05:19:40 koen: if the 12" powerbook would have a superdrive... Sep 14 05:19:50 mine has Sep 14 05:19:58 koen: you can write dvd? Sep 14 05:20:01 yes Sep 14 05:20:15 wtf, why does the website claims it is not possible... Sep 14 05:20:19 dvd-r and dvd-rw Sep 14 05:20:43 and my powerbook is over a year old now Sep 14 05:20:44 anyway I actually have only written two DVDs in my entire life... Sep 14 05:22:00 I still plan to backup my divx collection to data-dvd Sep 14 05:22:56 zecke_: http://www.portalplayer.com/products/documents/5022_Brief_Mar05.pdf Sep 14 05:24:04 nice Sep 14 05:24:18 even SDIO Sep 14 05:24:56 wifiPOD :) Sep 14 05:27:16 wow Sep 14 05:27:21 thats a meaty little processor Sep 14 05:28:28 morning Sep 14 05:28:34 hey chouimat Sep 14 05:28:46 koen: funny chip, but a small BGA package Sep 14 05:29:44 can someone try to build scummvm with libsdl-x11? Sep 14 05:29:55 yeah Sep 14 05:30:08 ~lart scummvm people and their buildsystem Sep 14 05:30:09 * ibot throws a AN/M-8 smoke grenade at scummvm people and their buildsystem Sep 14 05:30:38 ~lart php for lack of multibyte support Sep 14 05:30:39 * ibot teaches php the basics, including how to RTM for lack of multibyte support Sep 14 05:32:50 hrw|work: scummvm works with sdl-x11, but doesn't build for some reason Sep 14 05:32:55 I still need to look into that Sep 14 05:32:55 ok Sep 14 05:33:06 koen: can you rebuild xmame with sdl-x11? Sep 14 05:33:08 lol http://www.askmen.com/jokes/current/index.html Sep 14 05:37:00 chouimat: quite funny Sep 14 05:38:15 pb_: Am I wrong or can one only have two items in a BBCOLLECTION? Sep 14 05:40:52 zecke_: you are wrong Sep 14 05:41:12 pb_: *puh* I love to be wrong Sep 14 05:43:02 heh Sep 14 05:44:49 pb_: the debugging capabilities of bitbake really suck Sep 14 05:46:48 lunch time - then meta-sdl commit, then opie-image/meta-opie* Sep 14 05:47:51 LOL "The primary purpose of the DATA statement is to give names to constants; instead of referring to pi as 3.141592653589793 at every appearance, the variable PI can be given that value with a DATA statement and used instead of the longer form of the constant. This also simplifies modifying the program, should the value of pi change . -- FORTRAN manual for Xerox Computers" Sep 14 05:47:56 zecke_: sure do Sep 14 06:04:44 chouimat, lol Sep 14 06:06:12 shame on you all for not fixing RJDay's bugs as he reports them! Sep 14 06:06:25 'bugs' Sep 14 06:07:15 bugs, dilusions, whatever... :) Sep 14 06:08:11 dilution? that sounds hard to fix. Sep 14 06:11:20 'help, stress, multiple heads in monotone' Sep 14 06:15:45 pb_: heh Sep 14 06:16:05 pb_: I blame the german beer for my general lack of mental capacities today Sep 14 06:18:01 hehe Sep 14 06:18:08 CP|Laptop: what brand? Sep 14 06:24:27 A local brewery - Waldschlosschen, I think Sep 14 06:25:26 heh.. libsdl-native does not build ;( Sep 14 06:25:46 | SDL_cpuinfo.c:107: error: PIC register '%ebx' clobbered in 'asm' Sep 14 06:26:17 CP|Laptop: you're in germany? Sep 14 06:29:21 koen: yep Sep 14 06:30:08 cool Sep 14 06:30:22 I bet in dresden at fab13 Sep 14 06:30:23 Dresden Sep 14 06:30:23 CP|Laptop: staying here for longer? Sep 14 06:30:30 Till Sunday Sep 14 06:30:42 koen: heh - its fab30 and fab36.. :) Sep 14 06:30:59 koen: the fabs are named by the year of the company they were built Sep 14 06:31:01 at least I had the '3' right :) Sep 14 06:39:14 how much modules does our perl build generate? Sep 14 06:40:13 700 or 800 Sep 14 06:40:32 so I have to wait some time.. Sep 14 06:40:38 did the dependency script get pushed? Sep 14 06:40:50 dont remember Sep 14 06:43:14 855 packages with modules Sep 14 06:45:23 frozen-bubble is BROKEN anyway Sep 14 06:45:42 for every configuration? Sep 14 06:45:48 is so -> move to non-working/ Sep 14 06:46:33 its marked as it - I dont have machine with enviroment where I can test it Sep 14 06:49:04 I'm adding sdl stuff into meta-sdl Sep 14 06:49:26 task-sdl-games, task-sdl-libs, task-sdl-emulators Sep 14 06:50:34 and it will contain only stuff which does not need libsdl-qpe Sep 14 06:52:49 presumably that's a pretty short list Sep 14 06:53:04 task-sdl-games = "freedroid lbreakout2 openttd prboom quake1 rott supertux uqm xmame" Sep 14 06:53:07 task-sdl-emulators = "dgen-sdl e-uae frodo snes9x" Sep 14 06:54:55 thats all now Sep 14 06:55:39 ops. forgot about abuse Sep 14 07:06:08 03hrw 07org.oe.dev * r940210b3... 10/packages/meta/meta-sdl.bb: Sep 14 07:06:08 meta-sdl: added more stuff which use SDL and not SDL/QPE Sep 14 07:06:08 - task-sdl-games contain all SDL games Sep 14 07:06:08 - task-sdl-emulators with all SDL emulators Sep 14 07:06:08 - task-sdl-libs contain libs without base libsdl (all libs depend on it already) Sep 14 07:17:14 beware - opie-image/meta-opie comes.. Sep 14 07:18:49 03hrw 07org.oe.dev * r3986317b... 10/packages/meta/ (4 files): (log message trimmed) Sep 14 07:18:49 meta-opie & opie-image restructure (bug#307) Sep 14 07:18:49 - new metapackage: meta-opie-all which contain non-OPIE stuff for QPE enviroment Sep 14 07:18:49 - meta-opie contain only OPIE stuff (plus konqueror-embedded and qpdf2 for bigflash devices) Sep 14 07:18:49 - moved opie-collections.inc functionality into meta-opie Sep 14 07:18:49 - created many task-packages for all opie packages Sep 14 07:18:50 how do we handle installation of .desktop files? Sep 14 07:18:51 - building opie-image from defined tasks Sep 14 07:19:29 zecke_: opie.bbclass or manual (if using palmtop.bbclass) Sep 14 07:19:45 no special mechanism, as far as I know; just "cp opie.desktop ${datadir}/applications" Sep 14 07:19:53 zecke_: install -m 644 foo.desktop ${D}${datadir}/applications Sep 14 07:19:57 hrw|work: with regard to the SDL games? Sep 14 07:20:16 hrw|work: some used to install QPE/Qtopia/Opie .desktop files Sep 14 07:20:20 zecke_: there manual (if at all) Sep 14 07:20:29 fd.o has no patent on the name .desktop ;) Sep 14 07:21:30 hrw|work: I can imagine X guys do not want /opt/QtPalmtop/apps/Games/foogame.desktop in their package Sep 14 07:22:07 zecke_: and vice-versa... opie users may dont want /usr/share/applications/games/foogame.desktop Sep 14 07:22:14 right Sep 14 07:22:24 /usr/share/applications/games/foogame.desktop Sep 14 07:22:28 ehm Sep 14 07:22:35 /usr/share/applications/foogame.desktop Sep 14 07:22:39 okay back to the question Sep 14 07:22:39 koen: I dont remember exact place Sep 14 07:22:43 how do we handle that? Sep 14 07:22:57 .desktops go in /usr/share/applications/ Sep 14 07:23:15 I'm fairly sure that's what the standard says Sep 14 07:23:39 koen: doc.tt.com says otherwise Sep 14 07:23:48 zecke_: until we dont have libsdl-qpe working with libsdl-x11 I dont know does it would be good to have both places in package Sep 14 07:23:51 and since .desktop files have a 'Section' field, there isn't any need for subdirs Sep 14 07:23:51 koen: and that is the Mobile Platform with Momemtum ;) Sep 14 07:23:55 heh Sep 14 07:24:45 zecke_: tt should either use .palmtop files or adhere to the standard Sep 14 07:24:47 What I do not want to see. Break Opie SDL games (it is broken if the user has no icon to start it) Sep 14 07:24:55 in favor of a semi working X11 integration Sep 14 07:25:15 koen: sorry, AFAIK KDE invented .desktop and QPE used a KDE2 version of these files Sep 14 07:25:36 how hard would it be to teach opie to scan /usr/share/applications? Sep 14 07:26:06 koen: hey man that would be incompatible ;) Sep 14 07:26:12 heh Sep 14 07:26:37 koen: seriously I know people using OE to build Qtopia based ROMs Sep 14 07:26:54 and? Sep 14 07:27:12 you would break perfectly working bbfiles Sep 14 07:27:29 'perfectly' isn't right Sep 14 07:27:41 perfectly from Opie/Qtopia POV Sep 14 07:27:47 yeah Sep 14 07:28:05 and I know people how build X based images Sep 14 07:28:26 these SDL apps were not meant to be used by X people ;) Sep 14 07:28:28 images are better, since they are writeable, so I win :P Sep 14 07:28:39 zecke_: scummvm, rott? Sep 14 07:28:46 koen: yeah ;) Sep 14 07:29:08 scummvm does not build already with sdl/x11 ;) Sep 14 07:29:30 hrw|work: I just convinced it to build Sep 14 07:29:36 ok Sep 14 07:30:49 zecke_: from a neutral POV we should have .bbs that use virtual/libsdl and install .desktops in a neutral place Sep 14 07:31:03 change the first neutral into OE Sep 14 07:32:50 zecke_: a quick fix would be to install it in /usr/share/applications and let the postinst check if /opt/qtopia exists and copy it there Sep 14 07:33:44 koen: /usr/share/applications it not neutral ;) Sep 14 07:34:26 install then into /umba/umba/cyk/ Sep 14 07:35:21 please do not get me wrong, I'm all for virtual/libsdl but we should not break stuff for Opie users Sep 14 07:35:44 teach opie to read /usr/share/applications Sep 14 07:35:57 or implement said postinst Sep 14 07:36:14 thats need hacking opie launcher... auc... Sep 14 07:37:31 koen: the 'Opie' .desktop files present do not have a "Section" key Sep 14 07:37:45 koen: so they're not even suited for fd.o compilant launchers/menus Sep 14 07:38:10 then the postinst method seems to be the winner Sep 14 07:38:28 as long as opie doesn't crash on .desktop files that do contain 'Section' Sep 14 07:38:54 TT is arrogant enough to do something like that IMO Sep 14 07:38:57 koen: the question is what happens if Opie gets installed after SDL games are already installed Sep 14 07:39:12 koen: sorry but at this point fd.o did not even exist Sep 14 07:39:26 koen: and they do no Desktop Stuff, they have KDE for that Sep 14 07:39:45 so they should use .palmtop files :P Sep 14 07:39:57 .desktop files are so bloated Sep 14 07:40:18 desktop icons are fluff anyway Sep 14 07:40:29 where PREFERRED_PROVIDER_libgpewidget and PREFERRED_PROVIDER_gnome-vfs should be defined? in distro conf or in preffered-gpe-versions? Sep 14 07:40:44 neither Sep 14 07:40:57 local.conf only? Sep 14 07:41:11 the libgpewidget one, yes Sep 14 07:41:55 koen: so any distro can result in mix of gpe and hildon libs? Sep 14 07:42:06 yes Sep 14 07:42:17 suxx Sep 14 07:42:19 if that distro wants to ship a gpe image and a hildon-image Sep 14 07:42:39 ok Sep 14 07:43:30 still no news from CELF Sep 14 07:43:33 *sigh* Sep 14 07:43:58 koen: we still have the bug with virtual/libsdl and the dependencies ending up in libsdl-qpe or -x11? Sep 14 07:45:02 I think so Sep 14 07:45:12 I can even make it depend on both Sep 14 07:45:44 see http://bugs.openembedded.org/show_bug.cgi?id=115 Sep 14 07:46:41 I think we can't avoid special-casing sdl in OE Sep 14 07:47:01 koen: we can Sep 14 07:49:30 yay! the miniconf is on! Sep 14 07:50:53 koen: wouldn't it work if the packages depend on libsdl? (or virtual/libsdl) Sep 14 07:50:56 so now i have to figure out how to get you all here :P Sep 14 07:51:06 koen: and both X11 and QPE provide these? Sep 14 07:51:23 why do we need to special case SDL? Sep 14 07:51:36 SDL_main should DIE! Sep 14 07:51:46 zecke_: they already do, but shlibs sees the qpe and x11 libs Sep 14 07:51:49 mithro: stamp on it, kill it Sep 14 07:52:35 mithro: SDL_main is an empty stub anyway Sep 14 07:52:55 mithro: and apps use sdl-config for their flags, and there -Dmain=sdl_main should not turn up Sep 14 07:53:21 zecke_: SDL_main is the only thing stopping binary compatibility between the to libSDL Sep 14 07:53:49 i think they should however install themselves as libsdl-qt.so and do a "update-alternative" for libsdl.so Sep 14 07:53:51 mithro: I doubt that ;) Sep 14 07:54:00 (the SDL_main thing) Sep 14 07:54:29 at least there is technically no reason to depend on SDL_main Sep 14 07:54:34 zecke_: -Dmain=sdl_main still exists in libsdl-qpe :-( Sep 14 07:54:34 zecke_: didn't you fix the sdl_main thing months ago? Sep 14 07:54:51 XorA: be more precise? Sep 14 07:55:00 pb_: yeah I'm using constructors and destructors Sep 14 07:55:03 ink we can't avoid special-casing sdl in OE Sep 14 07:55:06 right Sep 14 07:55:16 zecke_: the sdl-config generated by libsdl-qpe still outputs -Dmain=SDL_Main Sep 14 07:55:19 koen: why? Sep 14 07:55:21 http://bugs.openembedded.org/show_bug.cgi?id=115 Sep 14 07:55:39 mithro: I pressed copy after the line scrolled up Sep 14 07:55:52 koen: libsdl-qte and libsdl-x11 SHOULD be binary compatible Sep 14 07:55:59 if they arn't then thats the problem Sep 14 07:56:15 even if they are, shlibs gets in the way Sep 14 07:56:39 since sdl is linked to either x11 or qte Sep 14 07:56:43 koen: hrm? Sep 14 07:56:50 shlibs shouldn't care about that. Sep 14 07:57:01 if it does, it's a bug. file a report in bugzilla and I will look at it. Sep 14 07:57:01 don't they link against libsdl.so.0? Sep 14 07:57:21 pb_: isn't that what #115 is about? Sep 14 07:57:46 pb_: when 2 packages provide the same library shlibs depends on the first one it finds, -qpe being alphabetically higher I think it gets picked Sep 14 07:58:48 koen: hm, bugs.openembedded.org won't resolve for me for some reason. What does #115 say? Sep 14 07:59:14 pb_: we still do -Dmain=SDL_main Sep 14 07:59:21 zecke_: ah, then we suck Sep 14 07:59:29 hi Sep 14 07:59:32 pb_: specially I do ;) Sep 14 07:59:35 pb_: yes we do Sep 14 07:59:38 pb_: http://pastebin.ca/22906 Sep 14 07:59:45 pb_: but fixing sdl-config should be minor... Sep 14 07:59:50 anyone use arm-linux-g++ 3.3.2 for building qte and opie before? Sep 14 07:59:54 koen: heh, pastebin.ca is even worse. that site _never_ works for me. Sep 14 08:00:10 ah, here we go, now bugs.openembedded.org is working Sep 14 08:00:11 pb_: but statically linking an almost empty SDL_main should not harm Sep 14 08:00:13 yeah, pastebin.ca pretends to do ipv6 I think Sep 14 08:00:28 ah, right Sep 14 08:00:35 * AvengerMoJo having linking error at the end .... lot of symbol not find Sep 14 08:00:36 I wondered what was wrong with it. Sep 14 08:00:39 pb_: http://pastebin.com/363577 Sep 14 08:00:47 thanks Sep 14 08:02:01 it sounds like that's just related to the -Dmain thing, nothing to do with shlibs. Sep 14 08:02:52 so, assuming that Ãœberhacker Zecke can fix the thing with sdl-config, it should be in good shape. Sep 14 08:03:21 Good evening Sep 14 08:04:04 pb_: but even linking SDL_main should not harm? Sep 14 08:04:12 pb_: as it probably only calls main() Sep 14 08:04:41 bitbaking prboom now Sep 14 08:07:15 c Sep 14 08:17:19 well... Sep 14 08:21:16 works? Sep 14 08:22:14 hrw|work: 1.) -Dmain... is reported Sep 14 08:22:26 hrw|work: 2.) -lSDLmain gets added as well Sep 14 08:22:32 hrw|work: 3.) I do not have a SDLmain Sep 14 08:23:30 'aha' Sep 14 08:25:37 I don't want to know how it even works.. Sep 14 08:26:50 kopete lack one important option... 'ignore messages from people not in contactlist' Sep 14 08:43:29 good morning kergoth Sep 14 08:46:38 zecke_: mm, yes, true Sep 14 08:46:55 since SDL_main is part of libSDL.a, it doesn't impact the binary interface to libSDL.so Sep 14 08:47:07 so, I don't know what the dude in that bug was concerned about Sep 14 08:48:13 hi kergoth Sep 14 08:51:09 pb_: anyway it is more nice, to patch sdl-config as well Sep 14 08:51:15 yeah Sep 14 09:30:14 strace echo "2" > /sys/class/backlight/corgi_bl/brightness is interesting Sep 14 09:30:27 It sends the 2 happily and then loops sending the "\n" Sep 14 09:34:05 cu Sep 14 09:43:15 on a normal system anyone have an opinion on what tmpfs should be set to for shm? Sep 14 09:43:30 1gb seems rather excessive Sep 14 09:44:14 Is it a 2GB machine? Sep 14 09:44:19 yup Sep 14 09:44:26 I think it defaults to half of main mem. Sep 14 09:44:33 seems default is 1/2 your ram Sep 14 09:44:51 was thinking about cutting it back to 256 Sep 14 09:45:00 Is it an issue? Sep 14 09:45:22 would rather use the ram for other things ... Sep 14 09:45:35 as far as i can figure out its only X thats using it Sep 14 09:45:46 and rarely Sep 14 09:46:17 emte: It only takes as much space as it needs. The 1GB is the maximum it can grow to Sep 14 09:46:35 hmm Sep 14 09:46:51 maybe there is a bug in my driver then Sep 14 09:47:02 its allocated the full gb and isnt sharing Sep 14 09:47:18 er doesnt change size Sep 14 09:47:48 i just never see amything more than a few MB using it Sep 14 09:48:25 It will list 1GB as available - it won't use it though Sep 14 09:48:43 yeah that seems to be my problem because it has Sep 14 09:49:01 tmpfs 1015M 16K 1015M 1% /dev/shm Sep 14 09:49:11 mtab entry Sep 14 09:50:18 Time to roll back to the hotel Sep 14 09:51:29 i wish somewhere it would list a reasonable value for shm ... Sep 14 09:53:22 emte: It won't have allocated that amount in memory Sep 14 09:57:27 Mem: 2076812k total, 1672108k used, Sep 14 09:57:32 it has Sep 14 09:58:05 unless my system has decided to use 1.6gb ram idling Sep 14 10:09:35 seems a lot of issues with shm and tmpfs ... Sep 14 10:09:46 lot of forum arguments about it Sep 14 10:10:56 i guess all i can do is lock it down to 256 and see if POSIX IPC complains Sep 14 11:32:40 hey mickeyl Sep 14 11:32:53 mickeyl: did you get contacted by any of the CELF people? Sep 14 11:34:19 hmm .. Sep 14 11:34:25 my ipaq locked up :( Sep 14 11:36:26 * koen is getting a bit impatient Sep 14 11:36:48 * emte gives koen decaf Sep 14 11:39:16 hi all, is there a channel better suited to support questions? Sep 14 11:42:17 good evening Sep 14 11:42:22 koen: no, i haven't been contacted yet Sep 14 11:43:16 * mickeyl dives into tslib calibration for ecore Sep 14 11:46:38 mickeyl: how did you test the tslib patch on a simpad? Sep 14 11:47:41 koen: by handing a copy of the binary over to my friend Christian Sep 14 11:48:03 (who is mainly using his SIMpad as a IR transmitter) Sep 14 11:48:21 it doesn't look like raster is too interested though :/ Sep 14 11:48:24 not a single reply Sep 14 11:49:15 the people on the familiar list are raving about the eem demo Sep 14 11:49:29 heh Sep 14 11:49:31 lemmings, cliff Sep 14 11:49:35 *nod* Sep 14 11:49:42 there is virtually _no_ code behind that Sep 14 11:49:50 all raster was doing was show off an edje Sep 14 11:50:25 it's quite cool he's able to do that with virtually no code Sep 14 11:51:30 that's true. edje is a pretty interesting approach to interfaces Sep 14 11:51:44 s/no code/no code outside the edje definition/ Sep 14 11:52:00 * koen jumps and waves to get some attention for http://bugs.openembedded.org/show_bug.cgi?id=326 Sep 14 11:52:42 also note http://www.linuxdevices.com/news/NS3625053852.html and the connection to CELF Sep 14 11:53:07 * koen wants some paid codemonkeys to work on OE Sep 14 11:53:17 * mickeyl as well Sep 14 11:54:21 bleh Sep 14 11:54:49 I just realize I'll have to update all maemo stuff in ~16 days from now Sep 14 11:55:41 heh Sep 14 11:55:55 mickeyl: it seems you share my dislike for dots in overrides :) Sep 14 11:56:33 yeah Sep 14 11:56:59 are they supposed to send us the codes anytime soon? Sep 14 11:57:09 yes Sep 14 11:57:12 mickeyl, raster is no longer awnswering emails Sep 14 11:57:24 he is trying to get the textblock2 stuff done Sep 14 11:57:55 as for eem, that demo os almost 2yrs old Sep 14 11:57:57 is* Sep 14 11:58:08 it doent even work correctly anymore Sep 14 11:59:06 yeah. it doesn't matter, ecore-fb will have to be rewritten anyway. it's much too ipaq-h3600 centric Sep 14 11:59:48 didnt know justinp even had an ipaq Sep 14 12:00:00 ? Sep 14 12:00:10 mickeyl: is there anyone who knows how to fix the udev init scripts the right way? Sep 14 12:00:10 his files Sep 14 12:00:15 * pH5 points to http://bugs.openembedded.org/show_bug.cgi?id=325#c2 Sep 14 12:00:28 pH5: that would be you ;) Sep 14 12:00:34 <--- to busy atm. Sep 14 12:00:43 could you do that? Sep 14 12:01:14 okokok. I just don't really know how it's supposed to work. Sep 14 12:01:24 start udevd directly from the init script? Sep 14 12:01:45 pH5: not much to know besides a) udevd should be started by the script (either directly or via udevsend) and b) echo "" >/proc/sys/kernel/hotplug Sep 14 12:02:15 and then we need to see which hotplug rules we have left to convert Sep 14 12:02:20 great. that sounds easier than I thought. Sep 14 12:02:40 i think there is a simple init script in the latest udev distribution Sep 14 12:02:45 just don't use the lsb one, it's outdated Sep 14 12:03:43 mickeyl: what's your opinion on teaching opie to read /usr/share/applications for .desktop files Sep 14 12:03:54 ? Sep 14 12:04:54 zecke and I had a discussion about that since pure sdl apps put it in /opt/qtopia, which is unacceptable for non-qt environments Sep 14 12:07:06 I'm a bit reluctant to do _any_ changes to Opie from now on. I think it may be better to add a symlink during postinstall if an Opie installation is detected Sep 14 12:07:38 that's sort of with what I came up with Sep 14 12:08:27 I consider Opie really really frozen now, unless for critical bugfixes Sep 14 12:08:59 it's not urgent Sep 14 12:09:06 just something to keep in mind Sep 14 12:13:11 righto Sep 14 12:15:31 another thought: bump the opie PV to 2.1.1 Sep 14 12:15:39 I bet that would confuse a lot of people Sep 14 12:29:04 ah well, i have that odd sense of version numbers reflecting the state of programs ... Sep 14 12:29:12 ... and i can't really justify 2.x.x for Opie :D Sep 14 12:38:20 mickeyl: any ideas on how to name the branch? Sep 14 12:41:06 org.openembedded.oz354fam083 Sep 14 12:41:17 or vice versa Sep 14 12:42:19 sounds ok Sep 14 12:42:41 I'm feeling lazy, so shall we branch in the weekend? Sep 14 12:44:03 *nod* that'll be sufficient Sep 14 12:51:49 ping :-/ Sep 14 13:07:47 For reference, the sysfs/backlight bug we saw is fixed in 2.6.13-mm3. 2.6.14-rc1+git(today) also works with some patches :) Sep 14 13:08:05 I'll upgrade to 2.6.13-mm3 tomorrow Sep 14 13:14:31 cool Sep 14 13:15:48 excellent Sep 14 13:17:08 hmm i have a calibration problem. let's say i calibrate @ rot270. i guess i need to rotate the screen points back to 0 in order to work with the values got by ts_read_raw, right? Sep 14 13:36:52 03koen 07org.oe.dev * r4b441ac3... 10/packages/libgpevtype/libgpevtype_cvs.bb: packages/libgpevtype/libgpevtype_cvs.bb: add (disabled) floating cvs version of libgpevtype Sep 14 13:45:11 Folks, 1st build after quite some time glibc-2.3.5+cvs20050627 fails with undefined reference to `_errno... (see http://pastebin.ca/22933). Did you see that in your builds? Sep 14 14:15:13 mickeyl: correct :-/ Sep 14 14:41:35 mmh, k Sep 14 14:42:00 * mickeyl opens up a linar algebra book Sep 14 14:42:05 linear, even Sep 14 14:46:09 mickeyl: forgetten the basic stuff that fast? Sep 14 14:47:10 ljp: Congrats on the new baby Sep 14 14:47:27 * koen quickly gets his book on linear algebra in case some starts asking questions Sep 14 14:51:15 *strike* Sep 14 14:51:17 * mickeyl dances a jig Sep 14 14:51:27 *phew* that took longer than it should've taken :/ Sep 14 14:51:32 guess my math is getting rusty Sep 14 14:52:12 don't you have a scientific calculator? Sep 14 14:52:19 * mickeyl points to his head Sep 14 14:52:37 hmmm, on collie i don't even need to take a median of a couple of samples. jitter is low enough. guess i need to check on other models Sep 14 14:53:46 I just put new batteries into my ti92+.. can't live without it anymore.. Sep 14 14:53:53 heh Sep 14 14:53:58 calculator weenies Sep 14 14:54:24 http://handhelds.org/scap/port.9517.png Sep 14 14:54:35 is that ugly or what? :))) Sep 14 14:54:56 it is :) Sep 14 14:55:12 thanks for confirming. my graphical talent converges to 0 Sep 14 14:56:05 your graphical talent can be expressed as f(x) = 0 :p Sep 14 14:56:25 hah Sep 14 14:56:34 for what values of x ? Sep 14 14:56:44 lim x-->0 Sep 14 14:56:50 mickeyl: Its a lovely feeling when you do that on a device or piece of software for the first time :) Sep 14 14:56:50 *nod* Sep 14 14:57:18 RP: indeed it is :) Sep 14 14:57:57 * Delta__ looks at his black screen..hm.. Sep 14 14:58:14 now that we have touchscreen, i can start adding sanity in the ecore_fb keyboard driver Sep 14 14:58:26 surprise surprise... it contains a hacked map for h3600 :) Sep 14 14:58:40 Delta__: when I started uni calculators were forbidden at all math exams Sep 14 14:59:15 guess I'll make it use the input system from the start Sep 14 14:59:56 mickeyl: an excellent start. That will enable most 2.6 devices to use it... Sep 14 15:00:26 koen, same here, but we needed to use them on the other subjects, it's handy to see a curve of the fourier transformation you calculate.. Sep 14 15:00:42 mickeyl: all values of x :p Sep 14 15:00:54 I could draw that faster as the calculater 2 years ago :) Sep 14 15:01:23 hmm now let's see if efl features changing rotation on-the-fily Sep 14 15:01:31 fly, even Sep 14 15:01:39 alright. gotta run. Sep 14 15:01:46 cu treke Sep 14 15:01:50 one of these days I'll remember to put #oe back on autojoin :p Sep 14 15:02:28 hey, my calculator has a 68000 on 16 mhz :) Sep 14 15:02:34 We were allowed calculators at uni but only simple ones... Sep 14 15:02:58 RP: I have to add graphing calculators were *mandatory* at highschool Sep 14 15:03:11 luckily my ti83 broke down Sep 14 15:03:54 nowadays people all have laptops here, using matlab.. Sep 14 15:05:12 koen: They weren't mandatory for me but strongly advised as otherwise, everyone else had an advantaged. They are banned at A-Level now as well Sep 14 15:05:56 Delta__: I still don't own a laptop. I bought husky instead ;-) Sep 14 15:06:43 cu Sep 14 15:08:13 does husky have a keyboard? I tried irssi on ssh on my palm from the couch, typing with the pen on the display , hitting the letters is so hard :) Sep 14 15:08:36 Delta__: It has - its a clamshell Sep 14 16:08:23 is there a way to force tslib to reinitialize without ts_close and ts_open again? i.e. after calibration Sep 14 16:08:52 hey people Sep 14 16:09:06 mickeyl: how much would it cost you to get the LCA2006? Sep 14 16:09:35 mithro: what's the nearest airport? Sep 14 16:10:29 BNE, MEL, or SYD? Sep 14 16:11:19 it's in New Zealand! Sep 14 16:11:26 :) Sep 14 16:11:52 hmm Sep 14 16:12:12 * mickeyl confesses his obvious lack of geographic understanding Sep 14 16:12:53 linux.conf.au is Australia's national Linux conference. linux.conf.au 2006 will be held in Dunedin, New Zealand from the 23rd - 28th January 2006 at the University of Otago. Sep 14 16:13:14 mickeyl: yes an "Australian" conference is been held in New Zealand Sep 14 16:13:21 don't try to understand it Sep 14 16:13:46 I dread to think how many changes I'd need to get to New Zealand... Sep 14 16:13:55 btw anyone here going to the Linux Kongress? Sep 14 16:14:00 hmm Sep 14 16:14:00 wow Sep 14 16:14:06 EXPENSIVE Sep 14 16:14:15 around 2000 EUR Sep 14 16:14:41 uhm no, it's even worse Sep 14 16:16:53 mickeyl, 2000 EUR to just get in, or what? Sep 14 16:17:19 looks like the cheapest two way is ~3000EUR Sep 14 16:17:22 * mickeyl shakes head Sep 14 16:17:50 oh...plane ticket Sep 14 16:17:51 haha Sep 14 16:18:06 yeah Sep 14 16:18:12 absolutely unbelievable Sep 14 16:18:17 i was seriously confused... Sep 14 16:18:24 ya...thats incredibly stepp Sep 14 16:18:26 *steep Sep 14 16:18:37 wtf 3000EUR? Sep 14 16:18:50 how much does it cost to get to Sydney? Sep 14 16:19:00 where are you flying from? Sep 14 16:19:27 oh hell ya...I am king... Sep 14 16:19:41 i'm working on adding WPA support to the at76c503a driver.. Sep 14 16:19:57 and I finally got it to spit out the wpa_ie...which is the first of many steps... Sep 14 16:20:08 but as this is my first major programming challenge ever... Sep 14 16:20:14 this first step is damn good Sep 14 16:21:55 mickeyl: For me it'd be £5,518 :) Sep 14 16:23:11 RP: what about just to Sydney? Sep 14 16:23:11 That's about 8,200EUR! Sep 14 16:24:41 mithro: £1,400 which makes me wonder about this website's sanity... Sep 14 16:26:09 the trip is about 36 hours :-/ Sep 14 16:27:43 well i can get you from sydney to Dunedin for like $400 AUS Sep 14 16:28:12 mithro: I guess its trying to get all the flights to connect though Sep 14 16:29:04 how much is 1,400 in EURO? Sep 14 16:29:34 2000EUR ish Sep 14 16:30:52 okay Sep 14 16:31:11 i'm guess it's going to be about 2000EUR for anyone around europe Sep 14 16:32:15 I'm a bit off the major air routes - would have to fly to London before going anywhere so yes, as long as they're near a major international airport, that'd be about right Sep 14 16:32:42 * mithro guesses pb is in bed atm Sep 14 16:43:26 mithro: I'd guess so. Its also where I'm heading ;-). 'night Sep 14 16:45:38 have a good night Sep 14 16:47:29 so mickeyl, any chance you'll pop-up to the Linux Kongress? Sep 14 17:27:58 * mithro pokes mickeyl - you fallen asleep or something? Sep 14 18:24:33 anyone here Sep 14 18:24:46 i got problem building from bitbake Sep 14 18:45:39 depnds Sep 14 18:45:50 what do you think your problem is? Sep 14 18:49:30 i setup according to what the wiki in openembedded.org .. Sep 14 18:49:42 then i bitbake gpe-image Sep 14 18:49:59 it dump error on libtool-native Sep 14 18:51:06 few reasons it might Sep 14 18:51:21 most probably is your local.conf is wrong Sep 14 18:52:10 which part of the local.conf? Sep 14 18:53:14 but the previous few packages before libtool.. have no problem.. Sep 14 18:53:22 distro or machine Sep 14 18:53:45 distro="openzaurus-3.5.4" Sep 14 18:53:50 machine="akita" Sep 14 18:53:53 anyone here handy with the gimp? Sep 14 18:56:39 mithro, yeah Sep 14 18:56:52 does 3.5.4 build for the akita? Sep 14 18:57:00 brb food Sep 14 18:58:05 what do i put for target_os Sep 14 19:03:07 hi Sep 14 19:04:16 i got libtool error when bitbake... Sep 14 19:07:14 when you building bitbake? Sep 14 19:09:58 yup.. i bitbake gpe-image Sep 14 19:10:12 it stop at libtool-native Sep 14 19:10:59 for openzaurus... Sep 14 19:11:13 do i set Distro="openzaurus" Sep 14 19:11:27 or distro="openzaurus-3.5.4" Sep 14 19:11:41 "openzaurus-3.5.4" Sep 14 19:13:02 z-akita: what is the error Sep 14 19:14:52 target_os = "linux" Sep 14 19:14:57 yup Sep 14 19:14:58 set Sep 14 19:15:20 i'm checking the work directory.. Sep 14 19:16:59 if i didn't set target_os .. will it give error? Sep 14 19:20:14 i rebuild again... Sep 14 19:24:57 emte: would you be able to add a fade to the OE logo? Sep 14 19:26:43 mithro: which logo? Sep 14 19:29:47 the one in the top corner here Sep 14 19:29:48 http://oe.handhelds.org/ Sep 14 19:30:25 i want to fade it in such a way that i can use it as a replacement for the top righthand corner of http://www.linux.conf.au/miniconfs.php Sep 14 19:30:30 s/right/left Sep 14 19:30:36 oh Sep 14 19:30:38 cool Sep 14 19:30:54 will we add bootsplash into the kernel? Sep 14 19:31:26 AvengerMoJo: it's already done - if you see the OZ release images Sep 14 19:31:40 oh the little logo Sep 14 19:31:49 that is actually bootsplash? Sep 14 19:32:20 I keep thinking that is framebuffer only Sep 14 19:35:42 AvengerMojo: can u take a look at my log ... http://pastebin.com/364151 Sep 14 19:36:26 it still give me the same error.. Sep 14 19:37:03 z-akita: remove the stamp for the libtool in oetmp/ Sep 14 19:38:13 you are using last time stamp Sep 14 19:38:22 z-akita: so it didn't reconfigure Sep 14 19:39:51 there is do_configure, do_patch, do_unpack.. do i remove all of them? Sep 14 19:40:10 no ... remove the stamp Sep 14 19:40:39 ls /oetmp/stamps/libtool-* Sep 14 19:40:54 rm move them Sep 14 19:42:01 then i bitbake gpe-image ? Sep 14 19:44:15 you can try the libtool-native Sep 14 19:46:15 mithro, i am not quite understanding what you want ... Sep 14 19:47:00 or we have differing concepts of "fade" Sep 14 19:47:35 unless ... you mean the left corner Sep 14 19:51:55 http://oe.handhelds.org/miniconf/ Sep 14 19:52:11 emte: yes i ment the left corner Sep 14 19:53:11 AvengerMojo: i've remove the libtool in stamps, and then try bitbake libtool-native... it gives me error... http://pastebin.com/364168 Sep 14 19:54:06 i'm wondering whether its the automake problem... Sep 14 19:54:34 mithro, you want a white -> color fade? Sep 14 19:54:47 yes Sep 14 19:54:48 the OE png is currently transparnet Sep 14 19:55:08 z-akita: why don't you remove this /openembedded/oetmp/work/libtool-native-1.5.10-r3/ directory all together and try it again Sep 14 19:55:29 ok... i'll try.. Sep 14 20:14:15 mithro, dcc, how does that work for you? Sep 14 20:14:30 pain in the arse gradient Sep 14 20:15:20 not quite happy with it myself, but ... Sep 14 20:18:29 emte: i think it's quite good! Sep 14 20:19:03 on scree its .. okay Sep 14 20:19:07 screen* Sep 14 20:19:36 http://oe.handhelds.org/miniconf/ Sep 14 20:30:21 i manage to get it to build... Sep 14 20:30:36 but manual... make... Sep 14 20:30:58 then i touch do_compile Sep 14 20:31:15 then bitbake again.. Sep 14 20:32:48 z-akita: I'm not sure this make any different but you can try ccache -C to clear the cache may be Sep 14 20:33:50 AvengerMojo: thank you ... Sep 14 20:36:21 np Sep 14 20:36:41 * AvengerMoJo about time to fly to hongkong Sep 14 20:36:46 see you guys later Sep 14 20:37:58 have a safe trip Sep 15 00:21:57 hi... Sep 15 00:23:19 morning Sep 15 00:24:20 the oe metadata have the bvdd driver? Sep 15 00:25:35 Looks like it does Sep 15 00:25:50 for the pxa27x-2.4.20-rmk2-embedix20050602 kernel Sep 15 00:28:52 the metadata doesn't seem to have the bvdd.c and bvdd.h Sep 15 00:29:21 http://atty.skr.jp/WebSVN/listing.php?repname=bvdd&path=%2F&sc=0 Sep 15 00:31:12 I see a patch with bvdd in the name, that contains a bvdd.c and a bvdd.h - thats all I know, I don't use strongarm Sep 15 00:31:13 i've got an error while building openzaurus-pxa27x-2.4.20-rmk2-embedix-r15 Sep 15 00:31:28 if its not up to date, I'm sure that somebody would be happy if you submitted a new patch Sep 15 00:34:34 how do i compile the kernel in bitbake? missing arm-linux-gcc-2.95 Sep 15 00:37:24 z-akita: read teh zauruskernels wiki Sep 15 00:37:34 hey CP|Laptop Sep 15 00:37:41 Morning Sep 15 00:39:21 yup.. just found the wiki page Sep 15 00:44:30 koen: any idea how to get bvdd in OE? Sep 15 00:45:08 mplayer with bvdd be great for CXk ... Sep 15 00:45:48 update the patch or write a .bb for it Sep 15 00:45:58 and when it works attach it to bugzilla Sep 15 00:58:09 koen: whats the progress of kernel 2.6 in CXk? Sep 15 01:00:44 see http://rpsys.net/openzaurus/ Sep 15 01:02:43 morning Sep 15 01:03:47 hey XorA Sep 15 01:11:46 03koen 07org.oe.dev * r22cba8e3... 10/conf/machine/ipaq-pxa270.conf: conf/machine/ipaq-pxa270.conf: remove some cruft, needs to in meta/meta-gpe.bb anyway Sep 15 01:19:15 morning folks Sep 15 01:24:36 morning all Sep 15 01:26:03 morning Sep 15 01:38:46 RP: I id opie benchmarks yesterday on my husky with 2.6.13-mm1-r5.1 cpufreq enabled kernels Sep 15 01:38:57 will post them on my website today Sep 15 01:39:35 it would be nice to compare 2.4, 2.6 and 2.6 cpufreq Sep 15 01:40:18 koen: does the cpufreq works in CXk? Sep 15 01:40:27 I never used 2.4 on husky (except first half hour and updater) Sep 15 01:40:34 z-akita: no Sep 15 01:40:47 hrw|work: you don't want to ;-) Sep 15 01:40:54 RP: exactly Sep 15 01:41:18 hrw|work: I'll be interested to see the comparision under cpufreq - battery life should be the big gain though Sep 15 01:41:34 not tested batterylife Sep 15 01:51:54 03rpurdie 07org.oe.dev * re55a2edf... 10/packages/linux/ (5 files in 2 dirs): linuz-oz-2.6: Upgrade to 2.6.13-mm3. This fixes the sysfs problems people were seeing. Sep 15 02:04:52 http://www.hrw.one.pl/2005/09/15/cpufreq-on-zaurus/ Sep 15 02:07:25 will update Sep 15 02:07:49 updated Sep 15 02:10:16 hrw|work: Interesting. Its a near linear tradeoff Sep 15 02:25:01 hrw|work: could you add a column '2.4 400MHz' ? Sep 15 02:27:15 re Sep 15 02:27:35 koen: hmm.. probably can Sep 15 02:27:45 if my rootfs boots with 2.4 kernel Sep 15 02:29:26 c750: 490677 dhrys. 2.452 sec, 437 chars/s, 1396 gops, 132 kb/s Sep 15 02:29:34 according to the opie sysinfo I have Sep 15 02:29:42 no idea if that's 2.4 or 2.6 Sep 15 02:29:45 old results Sep 15 02:30:39 heh.. my husky will be raped by 2.4 kernel Sep 15 02:30:39 it's very annoying the list shows machines which don't have an entry Sep 15 02:32:35 hmm.. under 2.4 opie start rotated Sep 15 02:32:48 and ts is dead Sep 15 02:33:17 keyboard also does not work under opie Sep 15 02:33:27 hrw|work: You'll probably need to wipe calibration for the ts Sep 15 02:33:52 It did used to work. I guess we've broken 2.4 at some point... Sep 15 02:33:56 rebooted Sep 15 02:33:59 shrug ;-) Sep 15 02:34:07 RP: who cares about 2.4 on c7x0? Sep 15 02:34:15 hrw|work: exactly Sep 15 02:34:29 microphone users? Sep 15 02:35:06 Someday soon I will write the alsa driver Sep 15 02:36:40 RP: I bet RMK will say alsa doesn't work on ARM and reject it Sep 15 02:36:51 ok. started benchmarks Sep 15 02:38:54 koen: I doubt it will go through rmk... Sep 15 02:40:24 heh.. I cant make cf card test under 2.4 - no pcmcia-cs in rootfs Sep 15 02:41:00 Ciao all Sep 15 02:41:00 http://www.hrw.one.pl/2005/09/15/cpufreq-on-zaurus/ - now with 2.4 results Sep 15 02:41:06 hey Pigi Sep 15 02:41:07 ciao Pigi Sep 15 02:41:22 Hi koen, hrw|work Sep 15 02:41:46 I think I have fixed the bug for ipkg. Is there someone that would test it before I commit ? Sep 15 02:41:50 argh....... I cant flash to 2.6.... kernel only on CF card which I cannot access now as I dont have pcmcia-cs installed.. Sep 15 02:42:13 koen: its your fault Sep 15 02:42:19 oops Sep 15 02:42:37 hi Pigi Sep 15 02:42:48 hi pb_ ! Sep 15 02:43:06 I'll add some notes to post Sep 15 02:43:52 hrw|work: Were those results with a rotated display? I'm just wondering why the graphics ops are so much faster?! Sep 15 02:43:54 strange that 2.4 is faster at graphics Sep 15 02:44:56 RP: no - with landscape Sep 15 02:45:25 w8 a moment - I'm adding notes Sep 15 02:46:09 ~seen zecke_? Sep 15 02:46:13 RP: ram is almost twice as fast Sep 15 02:46:15 zecke_ was last seen on IRC in channel #oe, 17h 55m 1s ago, saying: 'pb_: anyway it is more nice, to patch sdl-config as well'. Sep 15 02:47:05 ram can be 44M under 2.4 kernel Sep 15 02:47:15 opie benchmarks suxx under 2.4 Sep 15 02:47:17 koen: Yes, all things considered, those benchmarks aren't too bad... Sep 15 02:47:58 and it doesn't take in account the 'latency' Sep 15 02:48:05 2.6 feels loads faster Sep 15 02:48:16 That's preempt for you... Sep 15 02:51:45 koen: add some better benchmarks into OE then so we can test Sep 15 02:52:41 nice.. 2.4-husky say that I have shepherd (proc/cpuinfo) Sep 15 02:52:57 isn't lmbench in OE? Sep 15 02:54:30 RP: I didnt realised how much improved backlight is in 2.6 until now.. you are great Sep 15 02:55:10 hrw|work: thanks :) Sep 15 02:55:49 Its done properly. Which is why we don't have alsa or leds yet ;-) **** ENDING LOGGING AT Thu Sep 15 02:59:56 2005