**** BEGIN LOGGING AT Wed Aug 10 23:59:56 2005 Aug 11 00:45:34 what flavours of linux run on zaurus? Aug 11 00:53:14 morning Aug 11 01:04:20 monring XorA, koen et al. Aug 11 01:06:13 good morning all Aug 11 01:10:28 korn ade|desk koen Aug 11 01:38:23 morning koen, XorA, ade|desk :) Aug 11 01:52:11 how do you specify a certain svn revision for a uclibc build ? Aug 11 01:54:25 try with ;date= or ;revision= Aug 11 01:54:44 you can check the fetcher python scripts for details Aug 11 01:54:57 cheers Aug 11 02:15:37 good morning Aug 11 02:53:27 hello all Aug 11 02:54:02 hey startoe Aug 11 06:15:55 rwhitby: did the nslu2 finish another tinderbuild? Aug 11 06:16:20 koen: it does one every 15 minutes. why? Aug 11 06:17:25 apache on ewi is taking all cputime Aug 11 06:17:56 so I wondered if someone is doing something cgi related Aug 11 06:18:16 and that's either viewmtn or tinder Aug 11 06:19:01 I need someone who knows something about disassembling things... Aug 11 06:19:50 RP: are you ready to announce yet? Aug 11 06:20:20 koen: No, I now have a plan. You'll know by the weekend. Watch for the commits ;-) Aug 11 06:20:35 This disassembly is related... Aug 11 06:20:52 I have discovered something most annoying :-/ Aug 11 06:21:00 * koen resorts to pestering mallum about it Aug 11 06:25:16 RP: is writing collie SD drivers :-) Aug 11 06:28:27 RP: are you working on the collie SD- driver? Aug 11 06:28:33 XorA: Wrong. I wish I could comment further though... :) Aug 11 06:28:58 hmmm, picsel browser then Aug 11 06:29:20 RP: damn you, Im going to be on GPRS for the weekend :-) Aug 11 06:29:51 RP: Best tool for this job is IDA pro Aug 11 06:30:13 XorA: ..and i'll be on honeymoon for two weeks .. so i have to wait even longer :( Aug 11 06:30:21 ade|desk: congrats Aug 11 06:30:31 XorA: thanks Aug 11 06:30:34 ade|desk: congrats Aug 11 06:31:31 * ade|desk pokes RP ... go on tell us, we wont tell a soul Aug 11 06:31:41 ade|desk: my gongrats too... Aug 11 06:31:43 chher koen Aug 11 06:31:48 do13: I have a problem with some rouge offsets which I doubt IDA Pro would help with :-/ Aug 11 06:31:53 s/chher/cheers Aug 11 06:31:59 ade|desk: Congrats. I'm not saying though :) Aug 11 06:32:06 s/gongrats/congrats Aug 11 06:38:26 XorA: Disassembling the collie SD- driver shouldn't be so hard. I played with this some weeks ago. Aug 11 06:38:42 Sharp camera driver :-) Aug 11 06:39:01 oh, wait, the c-guys SDIO driver :-D Aug 11 06:39:07 do13: I'd agree. It should be quite straight forward if you know what you're doing... Aug 11 06:39:34 XorA: I'd suspect some of the latest sharp 2.4 kernels already have SDIO in them... Aug 11 06:39:54 RP: but not a driver for the 802.11 cards :-) Aug 11 06:40:39 XorA: True. I wish we could get some specs... Aug 11 06:41:28 RP:With some lines of code I did this trace:http://www.do13.de/openzaurus/temp/collie1.txt But the timing isn't complete. Aug 11 06:43:31 do13: That's quite interesting. We should probably put some pages together somewhere with that kind of info on them. I have all the ATI stuff to share as well... Aug 11 06:45:54 RP:Agreed. Aug 11 06:53:08 i am getting the error : http://pastebin.ca/19789 when i try to make an oe image Aug 11 06:57:45 03koen 07org.oe.dev * r366c47d0... 10/packages/gnome/gnome-mime-data_2.4.2.bb: make gnome-mime-info depend on shared-mime-info Aug 11 06:57:49 03koen 07org.oe.dev * r5381d172... 10/packages/gnome/gnome-mime-data_2.4.2.bb: make gnome-mini-data also depend on mime-info Aug 11 07:00:21 cant make out why it is trying to find the file : oe/openembedded/packages/maemo/gconf-osso/gconf-daemon-dbus-oe.sh' Aug 11 07:00:34 i dont have that file on my system Aug 11 07:00:45 anyone got this error before ? Aug 11 07:01:20 are your bitbake and OE up to date? Aug 11 07:01:40 no Aug 11 07:01:51 update those first Aug 11 07:02:10 actually, i had downloaded my bitbake and oe a couple of months back and it worked for me. Aug 11 07:02:26 so i am using the same OE packages and bitbake Aug 11 07:02:35 should not it work ? Aug 11 07:02:58 I wouldn't expect it to work Aug 11 07:03:04 now it seems the sources have changed and it does not work ? Aug 11 07:03:12 and reporting bugs against ancient versions is kind of pointless Aug 11 07:03:44 is there any solution other than updating my packages and bitbake Aug 11 07:04:30 I'm affraid if you debug it it will turn out that it was already fixed and you've wasted your time Aug 11 07:06:17 which package used gconf ? Aug 11 07:06:22 gconf2_2.6.4-3.1osso13.tar.gz Aug 11 07:09:53 ok Aug 11 07:23:45 hi Aug 11 07:26:20 hi CoreDump|home Aug 11 07:36:36 are there any tar balls available to install oe and bitbake other than downloading them using montone Aug 11 07:37:37 nope Aug 11 07:38:28 we do have a snapshot of a database online Aug 11 07:53:14 earlier i had got bk-exports and bitbake.tar.bz2 Aug 11 07:53:41 my system is behind a fire wall. is there any way i can get these latest files ? Aug 11 07:53:50 it would be useful for people like me Aug 11 07:54:02 16:38 < koen> we do have a snapshot of a database online Aug 11 07:54:40 see http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted#head-1c5adf3b1f1467ec9bcb48b5a34c18632557c3c4 Aug 11 07:57:05 ok Aug 11 07:57:09 will get back to u Aug 11 07:57:13 take care Aug 11 08:19:43 03nail 07org.oe.nslu2-linux * r0b40ae52... 10/packages/nail/ (files/nail.spec.diff nail_11.21.bb): Update nail to use /usr/sbin/sendmail instead of /usr/lib/sendmail Aug 11 08:23:01 Does somebody know a good wiki about how to install a bootloader on compact flash for x86 using a compact flash ide adapter ? Aug 11 08:26:12 anyone heard of a company called JTon? Aug 11 08:32:59 XorA Taiwanese ? Aug 11 08:33:14 gerwinin: yes Aug 11 08:33:56 gerwinin: itshould act like a normal hd Aug 11 08:37:13 Xora , what do they sell ? Aug 11 08:37:34 Xora , I sent lately a couple of ipaq with opie to some taiwanese firms Aug 11 08:40:04 koen: i noticed that EXTRA_IMAGECMD_h3900_jffs2 does not work. it should be like this: EXTRA_IMAGECMD_jffs2_h3900 Aug 11 08:40:08 Xora can you gime there website Aug 11 08:40:22 koen: in machine configs Aug 11 08:40:50 aquadran: hmmm, is that with a recent bitbake? Aug 11 08:41:01 if so I should hassle zecke for breaking tha Aug 11 08:41:08 t Aug 11 08:41:09 koen: yes Aug 11 08:41:22 could you file a bug for that? Aug 11 08:41:53 koen: i copied that into h1910 from h3900 too unfornatelly Aug 11 08:41:59 koen: yes, i can Aug 11 08:44:50 gerwinin: I dont have it, we are quoting for some work and apparently they use gcc for smartcard work, but I was wanting to poke around but cant find them in google Aug 11 08:45:10 gerwinin: Im guessing they do PDA/Phones Aug 11 08:49:35 03nail 07org.oe.nslu2-linux * rbf63a8b3... 10/packages/postfix/postfix_2.0.20.bb: Update postfix to use u-a for /usr/sbin/sendmail (Conflict with ssmtp) Aug 11 08:51:42 XorA let me have a look at globalsources Aug 11 08:55:03 * ade|desk starts packing up .. see you folks ~September Aug 11 08:55:57 ade|desk: enjoy the honeymoon Aug 11 08:56:16 cheers Aug 11 08:58:41 time to go home Aug 11 08:58:46 Xora the name on the invoice is often not the name Aug 11 08:58:52 they use Aug 11 10:09:32 03jbowler 07org.oe.nslu2-linux * rb41e1e0c... 10/packages/meta/ucslugc-packages.bb: Aug 11 10:09:32 ucslugc, UCSLUGC_PACKAGES: make various prerequesite packages explicit in Aug 11 10:09:32 the build list, therefore if the dependent package is removed the Aug 11 10:09:32 prerequisite packages are still built. Aug 11 10:23:02 03justinp 07org.oe.dev * raa3fea21... 10/packages/altboot/altboot.bb: Aug 11 10:23:02 Fix typo in postinst for spitz, add more robust checking Aug 11 10:23:02 Add a postrm for spitz (untested) Aug 11 10:23:37 CoreDump|home: hope you don't mind Aug 11 10:24:08 CoreDump|home: the postinst for spitz should work correctly now Aug 11 10:24:15 * JustinP reboot his Z to make absolutely sure Aug 11 10:24:52 np, fire away :) Aug 11 10:26:15 hmmm, seems to be missing hdd images menu option Aug 11 10:27:58 argh...it's broken still...I wonder if I have a backup... Aug 11 10:37:56 JustinP: is the current "entrance" supposed to compile? Aug 11 10:38:51 CoreDump|home: I'm not sure. I have yet to try compiling the E stuff mickeyl checked in. :-| I think he got it to compile. Aug 11 10:39:13 CoreDump|home: Could you check in the hdd3 menu option I sent you for altboot? Aug 11 10:40:07 sure Aug 11 10:42:02 03justinp 07org.oe.dev * r0fbc5bc9... 10/packages/altboot/files/spitz/altboot.cfg: CoreDump was right, the REAL_INIT is init.sysvinit Aug 11 10:52:33 CoreDump|home: nm, I found a copy I can check in Aug 11 10:53:18 was about to do it :) Aug 11 10:54:04 oh, well if you have it ready go ahead Aug 11 10:58:07 done Aug 11 10:58:52 thanks Aug 11 11:00:07 np Aug 11 11:00:39 forgot the file when i synced your spitz changes :\ Aug 11 11:02:08 did you push it? Aug 11 11:02:30 yeah Aug 11 11:04:25 trying again... Aug 11 11:08:36 ok, this time the sync got through Aug 11 11:09:04 Trying to build a standard J7xx kernel (arm based), but a vanilla kernel doesnt get me anywhere. Anyone got a cvs link or a good page (irc?) Aug 11 11:12:07 03coredump 07org.oe.dev * rc2a4ba06... 10/packages/altboot/files/spitz/altboot-menu/25-bootHDD3: altboot: Add forgotten "boot hdd3" option for spitz Aug 11 11:15:32 hey bipolar Aug 11 11:16:51 03justinp 07org.oe.dev * rccee5bf1... 10/packages/altboot/files/spitz/altboot-menu/00-Default: Add /dev/loop0 to list of modules to be created Aug 11 11:19:22 03justinp 07org.oe.dev * raf7da440... 10/packages/altboot/altboot.bb: Bump PR Aug 11 11:59:48 What is the "right" way to install libraries and include files into the staging area now? Aug 11 12:35:55 koen: I've got a set of fixes for gcc/binutils which ensure that 'TARGET_CC_ARCH' (or HOST_CC_ARCH, as appropriate) always gets to the right place - at present it doesn't get to some things (compiling libgcc for example). Aug 11 12:36:48 go ahead and commit them Aug 11 12:37:32 I think CosmicPenguin will be vey happy with that Aug 11 12:37:53 ohhhh.. :) Aug 11 12:38:10 Anyone who wants to do -mthumb-interwork will be - it's impossible without them ;-) Aug 11 12:38:42 :) Aug 11 12:38:44 The changes are pretty well tested - I get compiler warnings (at least) if the flags are wrong. Tested building cross and native. Aug 11 12:39:02 jbowler: you manage to bulid gcc-cross-sdk? Aug 11 12:39:26 Trying to build a j7xx kernel (and booting with linexec) but my own built kernels doesnt give any output (but the linexec one does). Im using the same config so im guessing its either the toolchain or kernel version (any framebuffer patches out there?) Aug 11 12:39:36 I've only done this for gcc3-build* (I didn't try 4). I didn't build gcc-cross-sdk. Aug 11 12:40:08 gcc-cross-sdk was what was breaking apart for me Aug 11 12:40:53 Because of missing TARGET_CC_ARCH? Aug 11 12:41:58 Because of misdirected, most likely Aug 11 12:42:10 Line 117 of gcc-cross-sdk_3.3.4.bb needs ${TARGET_CC_ARCH} at the end, that might be enough, but I suspect it needs my fix for FLAGS_FOR_TARGET as well. Aug 11 12:42:49 quite possibly Aug 11 12:42:55 it also needs to be updated for a newer version Aug 11 12:43:29 Also I did the changes on 3.4.4 - I changed the -build.inc files, in a way which is safe for the older .bb's, but didn't change the .bb's themselves. Aug 11 12:45:22 Still, the .bb changes are all of the form I gave above - simply add {TARGET,HOST}_CC_ARCH to 'CC_FOR_TARGET' and 'GCC_FOR_TARGET' (the latter is for libgcc compilation). Aug 11 12:45:59 There's a uclibc fix too - it has the same problem. Aug 11 12:46:16 I bet thats it Aug 11 12:46:22 when are you going to have these things checked in? Aug 11 12:47:03 In the next few minutes - I'll check in {binutils,gcc,uclibc} to org.openembedded.nslu2-linux then propagate back to dev (that might take a while ;-) Aug 11 12:47:25 jbowler: don't forget --lca Aug 11 12:47:36 What is --lca? Aug 11 12:47:47 latest common ancestor Aug 11 12:47:53 a monotone 0.22 feature Aug 11 12:48:00 to keep the merging hydra at bay Aug 11 12:48:12 Ah ha... I haven't upgraded yet. Aug 11 12:48:29 monotone 0.22 also has a smarter netsync Aug 11 12:48:53 But no protocol change, I assume - 'cause my vanille sync still works ok. Aug 11 12:49:15 correct Aug 11 12:49:26 although I don't know if mickeyl changed yet Aug 11 12:49:50 ewi546 and dominion use 0.22 Aug 11 12:51:05 any tricks to upgrading monotone? Aug 11 12:51:56 not really Aug 11 13:15:35 03justinp 07org.oe.dev * rc05b7d0d... 10/packages/ppp/ppp_2.4.3.bb: Add a patch to run ranlib on 2 libraries (bug 172) Aug 11 13:16:19 kergoth: musicpd (mpd): this has both --enable-mpd-mad (which causes a local libmad to be built) and --with-mad-libraries=${STAGING_LIBDIR} - which implies you intended to use the OE version. That looks like a mistake, ok if I change it to not have --enable-mpd-mad? Aug 11 13:16:37 (Same for id3tag) Aug 11 13:23:21 jbowler: yes, go for it Aug 11 13:24:38 03jbowler 07org.oe.nslu2-linux * r073419be... 10/packages/ (14 files in 5 dirs): (log message trimmed) Aug 11 13:24:38 This set of changes ensures that TARGET_CC_ARCH is passed reliably to all Aug 11 13:24:38 packages in a build. The change fixes problems in the following packages: Aug 11 13:24:38 binutils{,-cross}_2.15.94.0.1 Aug 11 13:24:38 HOST_CC_ARCH was not being passed to HOST_CC. The problem does Aug 11 13:24:39 not seem to be present in binutils* 2.16, however this version Aug 11 13:24:41 is still used for uClibc builds. Aug 11 13:24:58 03jbowler 07org.oe.nslu2-linux * rde3c1699... 10/packages/libmad/libmad_0.15.0b.bb: Aug 11 13:24:58 libmad: thumb only (so no PR change): remove the erroneous use of 32 bit Aug 11 13:24:58 ARM assembler when compiling thumb Aug 11 13:25:26 03jbowler 07org.oe.nslu2-linux * rccb298ba... 10/packages/libvorbis/libvorbis_1.0.1.bb: Aug 11 13:25:26 libvorbis: thumb only (no PR change): prevent all optimisation to hide an Aug 11 13:25:26 apparent compiler bug (register spilling) when compiling thumb code. Aug 11 13:25:31 03jbowler 07org.oe.nslu2-linux * r99bf5683... 10/packages/db/db_4.3.27.bb: Aug 11 13:25:31 db-4: thumb only (no PR bump): prevent the erroneous use of 32 bit arm Aug 11 13:25:31 assembly when compiling thumb instructions. Aug 11 13:25:55 ::sigh::.....we should not be using 2 different branches for core development.... Aug 11 13:26:35 I suppose with the --lca option it's better but using that option could quite easily lead to other merge problems Aug 11 13:27:08 I did the merge. Aug 11 13:27:24 But yes, I agree completely. It seemed like a good idea at the start :-( Aug 11 13:27:47 it makes much more sense to me to do it all in .dev then propagate out to the branch Aug 11 13:27:57 I don't believe we gain any advantage from the split simply because both sets of developers work on the core. Aug 11 13:28:00 only distro/machine specific things should be in the other branches Aug 11 13:28:24 And only then if a branch is really required to freeze the core for a time. Aug 11 13:28:28 exactly. All of the changes to nslu2 are propagated back into .dev anyway Aug 11 13:29:05 IMHO we should only propagate *out* from .dev. Aug 11 13:29:31 Hum... a complely separate something keeps on deleting 'iconv.h' from staging (when I rebuild gcc I think). I have to go in and rebuild libiconv afterware. Aug 11 13:30:28 JustinP: yes - .dev is 'unstable', so most changes happen there. Changes propagate in only infrequently - that's the way the monotone guys did it with the netsync and botan branches. Aug 11 13:30:32 "rebuild" implies it was built in the first place. if you're using glibc, it isnt. you're using the iconv in glibc proper Aug 11 13:30:54 kergoth: I'm using uclibc Aug 11 13:31:18 So that seems to be the problem, gcc doesn't build it (it has the USING_UCLIBC define), but it does remove it? Aug 11 13:32:09 what makes you think gcc has anything to do with it? Aug 11 13:32:18 if I have two conflicing FILES sets is it set which will take precedence? Say the first has /usr/share/ecore/themes and another has /usr/share/ecore Aug 11 13:32:28 Because I build binutils, gcc and uclibc and it disappeared. Aug 11 13:32:35 and? Aug 11 13:33:03 Hum, you suspect uclibc... ok, easy enough to find out. Aug 11 13:33:11 first of all, if you're targeting linux-uclibc and you have nls support enabled for the build, you should go into the uclibc build and ensure that its nls support is _not_ enabled Aug 11 13:33:15 because it isnt ready for primetime Aug 11 13:33:40 (just set USE_NLS = no in the uclibc .bb, which will ensure that the PROVIDES is updated correctly) Aug 11 13:33:50 if you do that, libiconv will automatically be built when necessary Aug 11 13:33:58 rather than everything expecting the c library to have included it Aug 11 13:35:40 USE_NLS = "no" in conf/distro/ucslugc.conf Aug 11 13:35:44 going from memory here, its been months since i've touched any of this Aug 11 13:35:48 then why do you give a shit about iconv.h? Aug 11 13:35:54 Ah, hum, ?= "no" - that's always scary. Aug 11 13:36:00 if something isnt obeying the fact that USE_NLS is no, its broken Aug 11 13:36:01 period Aug 11 13:36:12 I don't care at all about iconv, but musicpd does... Aug 11 13:36:19 then its build isnt obeying our USE_NLS Aug 11 13:36:22 hehe, musicpd, yay :-) Aug 11 13:36:22 and again, is broken. Aug 11 13:36:30 does it use gettext? make sure the .bb inehrits gettext. Aug 11 13:36:33 heh Aug 11 13:36:41 Ok - I'll fix one or the other. Aug 11 13:36:50 (Either my distro or mpd) Aug 11 13:36:52 erm, gettext is intl not iconv, but .. Aug 11 13:40:20 JustinP: did it again I think (two heads). You're right, we need to fix this. Aug 11 13:40:38 03jbowler 07org.oe.nslu2-linux * rc7dbf3bd... 10/packages/musicpd/mpd_0.11.2.bb: Aug 11 13:40:38 mpd was building, but not necessarily using, it's own local cache of the Aug 11 13:40:38 libmad and libid3tag source. Removing the --enable-mpd-* settings causes Aug 11 13:40:38 it not to build the local source (and it uses the versions in staging). Aug 11 13:40:42 03justinp 07org.oe.dev * r4cabc1f4... 10/classes/efl.bbclass: themes dir should be included in the main package Aug 11 13:40:56 jbowler: would you like to merge or shall I? Aug 11 13:41:06 I've just done it. Aug 11 13:41:13 2 heads would still have happened with 1 branch. Aug 11 13:41:14 i havent touched oe since the switch to monotone Aug 11 13:41:21 hmm Aug 11 13:41:30 it's the propagates back and forth that worry me Aug 11 13:42:04 monotone-viz 'Everything' - two branches is just a formal way of having two heads without the need to merge. Aug 11 13:42:20 yeah, I know Aug 11 13:42:24 So it just allows people to be lazy about merging ;-) Aug 11 13:42:47 in the past the nslu2 folks had their own branch and did _all_ their work there, and trhat was periodically merged manually upstream to ensure they didnt break anything common. in the past, they didnt do much modificaiton of any core stuff, so that was appropriate. clearly that's changed, and perhaps hte policies should as well Aug 11 13:42:52 heh Aug 11 13:43:23 That's close to what I think - that we're doing more core work. Aug 11 13:43:29 and no offense, but in the past i've had to really keep a close eye on it, smacking people in #nslu2-linux over the head with the "no more ugly hacks, do it fucking right, damnit" stick Aug 11 13:43:43 oooh Aug 11 13:43:53 kergoth: we miss your cluebat at times Aug 11 13:43:56 hehe Aug 11 13:44:27 evening Aug 11 13:44:27 hey mickeyl Aug 11 13:44:47 hey Aug 11 13:45:19 JustinP: i don't think we want to include themes in the main package Aug 11 13:45:26 because then the whole shlib renaming doesn't take place Aug 11 13:45:31 which is quite unfortunate Aug 11 13:45:45 mickeyl: sorry I'm not sure what you mean by that Aug 11 13:45:53 RDEPEND on the default theme? Aug 11 13:45:58 but I *did* hit a case where an app didn't work because the theme wasn't shipped Aug 11 13:46:00 evas -> libevas0 Aug 11 13:46:12 this only works when there are only libraries in a package Aug 11 13:46:15 oh, so the $SRCNAME is incorrect you're saying? Aug 11 13:46:28 i think it would be better to say PACKAGE += themes Aug 11 13:46:34 and then make the lib RRECOMMEND the theme package Aug 11 13:46:37 I think there's only 1 theme in any of these packages anyway Aug 11 13:47:00 03mickeyl 07org.oe.dev * rc2d75b77... 10/packages/xstroke/ (xstroke_0.6.bb xstroke_cvs.bb): xstroke needs xpm Aug 11 13:47:05 03mickeyl 07org.oe.dev * r089971ad... 10/packages/linux/linux-openzaurus-2.6.13-rc5-mm1/add-elpp-stuff.patch: refresh elpp patch for 2.6.13-rc5-mm1 Aug 11 13:47:09 03mickeyl 07org.oe.dev * rf4976c9f... 10/packages/e17/entice_20050810.bb: fix entice dependencies Aug 11 13:47:17 mickeyl: did you get my botmail? Aug 11 13:47:26 no Aug 11 13:47:33 ibot: botmail read Aug 11 13:47:59 koen: i see Aug 11 13:48:08 we need more patches to Makefile.in's Aug 11 13:48:26 or a "remove libpng-dev from your host" warning Aug 11 13:48:27 all that INCLUDES += -I$(includedir) is to blame Aug 11 13:48:33 (in general) Aug 11 13:49:01 mickeyl: I know for sure that some of the programs depend on the themes being installed...they won't even start without it Aug 11 13:49:04 find ${S} -name Makefile | xargs sed -i s:'-I$(includedir)':'-I.':g Aug 11 13:49:15 mickeyl: you *do* need a default theme to use this stuff.... Aug 11 13:49:39 JustinP: so RDEPEND or RRECOMMEND on the theme Aug 11 13:50:08 koen: I suppose Aug 11 13:50:57 koen: I'll just have to figure out what needs what again... Aug 11 13:53:31 03jbowler 07org.oe.nslu2-linux * r2860daf5... 10/ (4 files in 4 dirs): Aug 11 13:53:31 Change the NSLU2 kernel build configuration method to allow the Aug 11 13:53:31 configuration override to be specified in the machine.conf rather than Aug 11 13:53:31 hardwiring -mno-thumb-interwork into nslu2-kernel.inc Aug 11 13:53:35 03jbowler 07org.oe.nslu2-linux * ra02108b1... 10/conf/distro/ucslugc.conf: Aug 11 13:53:35 Set -mthumb-interwork on ucslugc builds to allow building (and maybe Aug 11 13:53:35 someday running) of thumb exectuables. Aug 11 13:53:54 damn.... Aug 11 13:54:42 or not ;-) Aug 11 13:54:47 Different branches - so it's safe at the moment Aug 11 13:55:10 03justinp 07org.oe.dev * r07cf6fbd... 10/classes/efl.bbclass: Switch to creating a -theme package instead of including themes in the main package Aug 11 13:55:17 can you approve or deny #225 ? Aug 11 13:55:48 it's a bit odd, but I'm pretty out of ideas on that Aug 11 13:56:38 perhaps raster can comment on that Aug 11 13:57:40 mickeyl: Same error here (h3900, familiar) Aug 11 13:57:42 mickeyl: that's interesting...does it make your builds bail? Aug 11 13:58:51 yes Aug 11 13:59:34 i can't build entrance, for example Aug 11 13:59:42 * koen confirms that Aug 11 13:59:42 my local branches builds don't fail. I got edje all working Aug 11 13:59:56 you still using cvs builds of everything? Aug 11 14:00:00 yeah Aug 11 14:00:01 could be an upstream thing then Aug 11 14:00:07 well, my local branch... Aug 11 14:00:14 I'm checking out the stuff in .dev right now Aug 11 14:00:24 ok, perhaps we should switch from the fixed releases back to cvs Aug 11 14:00:28 but if possible, not free floating please Aug 11 14:00:42 i think it's better to use fixed dates in $PV and set CVSDATE=${PV} Aug 11 14:00:42 ~kill floating cvs .bbs Aug 11 14:00:42 * ibot shoots a pseudoproton gun at floating cvs .bbs Aug 11 14:01:15 sure, makes sense Aug 11 14:01:22 *if* that's the problem Aug 11 14:01:34 I *may* have a patch for it or something...like I said, I'm trying to check into it Aug 11 14:01:40 evas is bailing on me, though Aug 11 14:01:50 autotools_do_configure: command not found Aug 11 14:02:10 woohoo! http://www.livejournal.com/users/anholt/22710.html?thread=29366#t29366 Aug 11 14:04:33 did I break efl.bbclass somehow? It looks ok.... Aug 11 14:04:55 autotools_do_configure: command not found looks like a fundamental problem with BB not finding its classes Aug 11 14:05:01 yeah Aug 11 14:06:34 Crofton: the recommended way is to do it manually - install the needed headers and binaries via 'install -m ...' and install the libraries via oe_libinstall Aug 11 14:07:06 Crofton: calling oe_runmake install during staging is not recommended. it would pollute our STAGING_DIR and until we have per-package staging this would be a problem Aug 11 14:07:26 mickeyl: isn't autotools_stage_includes the recommended thing? Aug 11 14:07:28 jbowler: I'll give .dev until the morning so I can have a full two hours to pull down changes - hopefully your gcc stuff wil lbe in there by then Aug 11 14:07:33 jbowler: thanks, by the way Aug 11 14:07:45 koen: i don't like that, because it calls oe_runmake install Aug 11 14:08:06 damn it... Aug 11 14:08:20 mickeyl: to a different DESTDIR Aug 11 14:08:32 yes. that's cool for the packages which obey DESTDIR Aug 11 14:08:33 CosmicPenguin: I sort of understand what's going on inside there now, so if there is still a problem (there probably still will be in 3.3) then I might know where to look... Aug 11 14:09:10 thanks mickeyl, I am messing with some bb files and that is what it did Aug 11 14:09:18 I was hoping something better had come along :) Aug 11 14:09:39 per-package staging would rock. Aug 11 14:10:00 we could safely call oe_runmake install then and just pick forget about name clashes etc. Aug 11 14:10:01 mickeyl: use multimachine.inc to build native stuff Aug 11 14:10:11 that would be the cleanest solution Aug 11 14:10:32 kergoth: bitbake-ng should have better tracing and debugging support ;) Aug 11 14:10:55 automatic binary search for CVSDATES :) Aug 11 14:11:32 automatic upstream larting Aug 11 14:11:54 and *finally* the promised pizza fetching Aug 11 14:11:58 heh Aug 11 14:12:32 hehe Aug 11 14:12:44 ok...I have no idea why it's not inheriting autotools... Aug 11 14:12:48 koen: Pizza is overrated Aug 11 14:12:59 JustinP: use -b and -DDD Aug 11 14:13:03 zecke_: yeah shoarma is much better Aug 11 14:13:23 koen: anything not forcing me to stay +3 days in bed, is better Aug 11 14:13:51 zecke_: I discovered my girlfriend doesn't like vla Aug 11 14:13:55 zecke_: more vla for me :) Aug 11 14:14:09 --snoopje Aug 11 14:16:05 DEBUG: CONF file 'classes/efl.bbclass' not found Aug 11 14:16:20 it's...right...there.... Aug 11 14:17:18 what the.... Aug 11 14:17:28 ugh Aug 11 14:17:31 damn it Aug 11 14:17:37 * JustinP starts over from scratch again Aug 11 14:18:26 BBFILES was right but BBPATH wasn't.... Aug 11 14:25:56 l8er Aug 11 14:32:22 What would cause "/etc/init.d/pcmcia start" to stop ide? Aug 11 14:33:41 JustinP: does it execute cardctl resume or such? and this upsets ide-cs? Aug 11 14:34:17 zecke_: I don't know. All I know is something has causes PCMCIA *start* to try to *stop* ide, which remounts my root read-only. Aug 11 14:34:43 zecke_: It has ot be something I did to it since a fresh image works. I just don't know what to check for... Aug 11 14:34:56 hehe Aug 11 14:34:59 no idea then sorry Aug 11 14:35:14 also unmounts my media dir Aug 11 14:35:59 also....my "lo" isn't set up...I have to mnually ifconfig it... Aug 11 14:37:57 * JustinP has obviously really fscked up his Z... Aug 11 14:38:51 z fscker! Aug 11 14:41:33 ah, the joys of building glibc....so much time for self contemplation Aug 11 14:53:20 oh the wonders of packaging glibc, the multitude of locales to choose from Aug 11 14:54:16 200+ Aug 11 14:55:51 hi zecke_ Aug 11 15:00:08 die zecke die ;) Aug 11 15:00:42 "No, no, it's German for The, Bart, The!" Aug 11 15:01:57 hahah Aug 11 15:04:42 JustinP: I always wondered how that episode looked liked (untranslated) Aug 11 15:04:50 kergoth: re the "no more ugly hacks, do it * right, damnit' - that was definitely true for Unslung, but I hope the OE core devs are seeing #nslu2-linux "do it right" for OpenSlug (which is not constrained by Linksys compatibility, and can therefore do things the "OE way"). Aug 11 15:07:50 03freyther 07org.oe.dev * ra0a93c4c... 10/packages/uicmoc/uicmoc3-native_3.3.2.bb: (log message trimmed) Aug 11 15:07:50 packages/uicmoc/uicmoc3-native_3.3.2.bb: Aug 11 15:07:50 -If the package is named uicmoc, install uic and moc Aug 11 15:07:50 do not stage static libaries and do to stage the header files. Aug 11 15:07:50 There is a reason we have qt2 and qt3 dir in staging and this Aug 11 15:07:51 is not because uicmoc3 can stage its Qt3 headers into STAGING_INCDIR Aug 11 15:07:51 god damn, am I the first building both Qt3/X11 and Qt2E in one run... Aug 11 15:07:53 directly. Aug 11 15:08:41 ? Aug 11 15:08:54 I don't know what eigher of those things is Aug 11 15:08:57 either Aug 11 15:10:24 JustinP: Re: the separate org.openembedded.nslu2-linux branch - what's the proposal? Aug 11 15:11:37 rwhitby: basically that all core development should be done on the org.oe.dev branch exclusively and that *only* distro specific changes should be done in the nslu2 branch Aug 11 15:11:55 rwhitby: this would mean that we don't have to propagate back and forth, only one way (theoretically) Aug 11 15:12:44 ok, what that would mean is that we remove the nslu2-linux branch all together, cause we intentionally don't have distro specific changes which are not able to be incorporated into the main repo. Aug 11 15:12:50 I'm not really sure this *needs* to be done since the --lca option has been added to monotone, but it seems that a bunch of "core" development gets done in nslu2 and gets merged back into .dev Aug 11 15:12:55 and the concern about the propagating is ? Aug 11 15:13:17 (is it that changes can be silently reverted?) Aug 11 15:13:29 yeah, all the merging stuff Aug 11 15:13:41 It's just very difficult to propagate because of the delayed merge effect. Aug 11 15:13:52 I especially saw problems when I was propagating from .dev to my own local branch Aug 11 15:14:18 03nail 07org.oe.nslu2-linux * re3962f6a... 10/packages/ctorrent/ (10 files in 2 dirs): Updated ctorrent to use the complete extended ctorrent patch set Aug 11 15:14:18 It's pretty clear what is going on from monotone-viz - most of the core OE guys have 'short' merges (short path to ancestor), but NAiL and I have really long ones... Aug 11 15:14:20 whenever nslu2 propagated back into .dev my own propagate would all of a sudden start complaining about nslu2 files Aug 11 15:14:44 when I hadn't ever touched them Aug 11 15:14:53 granted, this is "fixed" by usig --lca.... Aug 11 15:16:16 if the nslu2 guys keep merging the same as the propagating then the merges will probably not really be any different Aug 11 15:16:34 but it just makes sense to use one branch since we're keeping the two in sync anyway Aug 11 15:16:40 JustinP, jbowler: ok. If we remain on a separate server (which we will) will that still present a similar problem even if we're using the same branch, or can we just sync that with the official OE server (BTW, when is that coming on line?) regularly? Aug 11 15:17:10 Multiple servers cause no problem. Aug 11 15:17:25 OE has multiple - any OE dev can sync to ewi or vanille. Aug 11 15:17:49 I and koen (at least) automatically sync all the servers together. Aug 11 15:18:21 ok, so the final aspect is commit permissions. Currently we have people with non-@nslu2-linux.org and non-@oe.org id's about to commit to the nslu2-linux branch. How do we handle that? Aug 11 15:18:32 ~lart E source code for having inconsistent whitespace Aug 11 15:18:33 * ibot eats E source code's liver with some fava beans and a nice chianti for having inconsistent whitespace Aug 11 15:18:39 jbowler: correct, I sync the 4 known public repos Aug 11 15:20:21 The non nslu2-linux.org commit indirectly to .dev - the delay before NAiL or I propagate really isn't a whole lot of protection... Aug 11 15:21:14 I'm happy with any decision you guys make up Aug 11 15:22:03 if it's really 2 seperate communities and all we can leave it as-is for now Aug 11 15:22:30 It just seemed to me that since the 2 branches don't actually diverge it doesn't make sense to keep them seperate Aug 11 15:24:07 well, oe.org has a requirement for developers to register an @oe.org address before getting commit privs, right? Aug 11 15:25:22 rwhitby: jbowler : I'm highly trusting you (web of trust) Aug 11 15:25:54 if you trust your developers enough to let them directly commit to oe.dev I'm doing the same Aug 11 15:26:01 obviously the same applies to Nail Aug 11 15:26:14 rwhitby: 'our' key ids aren't email addressess Aug 11 15:26:18 (yet) Aug 11 15:26:47 I trust our current developers. But we use @nslu2-linux.org specifically for core team members only. Aug 11 15:27:15 and other devs just use an email address which relates to their IRC nick Aug 11 15:28:41 so we might want the non-core nslu2-linux devs to register with @oe and then have a common trust regime Aug 11 15:30:02 zecke_: then you get to vet non-core nslu2-linux developers Aug 11 15:30:25 who own openembedded.org? Aug 11 15:30:32 michaelo or kergoth ? Aug 11 15:31:01 koen: some called christopher larson Aug 11 15:31:05 at digi.com Aug 11 15:31:08 that would be kergoth :p Aug 11 15:31:11 digi.com? haha Aug 11 15:31:18 man Aug 11 15:31:24 my monitor at my new job has burnin :( Aug 11 15:31:26 hehe Aug 11 15:31:27 * rwhitby has always wondered about the seeming inability for the current OE core devs here to create monotone.openembedded.org etc. Aug 11 15:31:29 damn small companies Aug 11 15:31:46 kergoth: heh, I didn't think you could get that on monitors these days Aug 11 15:31:53 Registrant Email:clarson@digi.com Aug 11 15:32:02 its an old one, i can read the text Aug 11 15:32:13 someone let a windows installer cd burn in Aug 11 15:32:19 lol Aug 11 15:32:20 apparently one of hte choices was windows 95 with visual studio Aug 11 15:32:23 hehe Aug 11 15:32:24 whoa Aug 11 15:32:37 well, windows sure does take a long time to install, but still Aug 11 15:35:50 kergoth: what are you doing? Aug 11 15:37:00 hrm, I wonder why dosemu isn't working on here Aug 11 15:37:51 hacking on network appliances at http://www.astrocorp.com/ Aug 11 15:37:54 oooh. I wonder if my treo can order papa johns Aug 11 15:38:27 * zecke_ smells the smell of garlic bread - but only in his mind Aug 11 15:38:31 zecke_, jbowler: current list of non-core nslu2-linux devs is: DaKa2, ccsmart, ingeba and kolla. Aug 11 15:40:36 so we could ask them to get @oe.org id's from #oe ... Aug 11 15:41:19 rwhitby: I'm not upto decide, but I would not mind that Aug 11 15:41:41 zecke_: who's the decision maker? Aug 11 15:42:09 rwhitby: mickeyl, kergoth, pb_ (probably schurig as well) Aug 11 15:42:35 yah, mickey or kergoth Aug 11 15:42:46 I don't even have write access myself, so I don't think I am really qualified to make decisions about others Aug 11 15:42:47 heh Aug 11 15:43:11 pb doesnt have write access to monotone? Aug 11 15:43:12 wtf Aug 11 15:43:31 and what is the process and criteria for someone to get an @oe.org id approved? Aug 11 15:43:54 rwhitby: you get approved by someone, generate your key, and send it on to mickey and koen Aug 11 15:44:08 heh Aug 11 15:44:19 never mind me, that statement had no worthwhile content Aug 11 15:44:24 ok, so basically the same as us, with different names :-) Aug 11 15:44:37 probably the same, yeah Aug 11 15:44:41 JustinP: but you know it's probably accurate Aug 11 15:44:49 you get approved by being approved :) Aug 11 15:45:19 rwhitby: probably easiest to have you guys approve someone and have them send their key to you, then you send it on to mickey and koen Aug 11 15:47:12 JustinP: that would work Aug 11 15:49:18 of course I just got commit access in the past few days, so I'm certainly not a decision maker ;-) Aug 11 15:53:01 later all Aug 11 16:00:50 03justinp 07org.oe.dev * ra0401b9f... 10/classes/efl.bbclass: Change theme package to themes instead Aug 11 16:39:25 when I run "monotone update" I get "misuse: multiple candidates remain after selection", how do I deal with this? Aug 11 16:39:48 run "monotone heads" Aug 11 16:39:53 does it show more than 1 head? Aug 11 16:40:25 yes, justin and freyther Aug 11 16:40:58 how do I fix? Aug 11 16:41:23 pull again Aug 11 16:42:08 my update.sh script pulls and then updates, and it fails on the update Aug 11 16:42:32 oops, looks like I didn't push my merge... Aug 11 16:42:34 update again Aug 11 16:42:43 LoL Aug 11 16:42:53 so then you are justin :-) Aug 11 16:42:56 yep Aug 11 16:43:03 do I have to pull again, or just update? Aug 11 16:43:14 freyther caused the multiple heads, though Aug 11 16:43:16 pull Aug 11 16:43:20 thanks Aug 11 16:43:33 03justinp 07org.oe.dev * r544313c3... 10/packages/efl/e_20050810.bb: Add edje to e depends Aug 11 16:44:00 I am just getting around to setting up monotone, and it has given me nothing but trouble... the main server never works for me.... Aug 11 16:44:22 what version do you have? Aug 11 16:44:38 of monotone? Aug 11 16:44:51 0.22 Aug 11 16:45:30 me too Aug 11 16:45:34 Zero_Chaos: never works, or is not contactable? Aug 11 16:45:38 pulling from monotone.vanille.de should work Aug 11 16:46:04 connection time out Aug 11 16:46:34 of course, now that I complain, it works fine :-P Aug 11 16:48:29 thanks for the help gentlemen Aug 11 16:48:51 np Aug 11 16:53:53 ahh, to see bitbake parsing .bb files again... it's so nice. Aug 11 16:54:11 Anyone here know if I build openzaurus-3.5.4 right now if I get the updated kernel? Aug 11 16:54:17 you should Aug 11 16:54:24 that's what I thought, thanks Aug 11 17:07:11 uhm... ${STAGING_BINDIR} is supposed to point to the staging//bin dir right? Why is it pointing to staging//bin ? Aug 11 17:09:23 wait a second Aug 11 17:09:35 NAiL: doesn't it point to host/bin Aug 11 17:10:40 STAGING_BINDIR = "${STAGING_DIR}/${BUILD_SYS}/bin" Aug 11 17:11:06 becuase binaries for STAGING_BINDIR should be executable on the system you built them Aug 11 17:11:22 *sigh*. Of course. Aug 11 17:11:26 for example tools like uic/moc belong there Aug 11 17:11:50 anyway I always mix up host and target ... Aug 11 17:11:53 good nite Aug 11 17:11:55 nite Aug 11 17:11:57 thanks Aug 11 17:34:56 03justinp 07org.oe.dev * r57a54192... 10/packages/ppp/ (ppp-2.4.3/plugins-fix-CC.patch ppp_2.4.3.bb): Fix ranlib problem correctly by removing CC in plugins Makefile Aug 11 17:36:14 Damn I am stuck with a bootloader problem Aug 11 18:26:13 Does anyone know the size limit of the binary SD driver on the SL-5500? Aug 11 18:32:06 size limit? Aug 11 18:32:47 does it support 1 gig SD Cards? 2 gig? Aug 11 18:33:25 any card that conforms to the SD card spec should work Aug 11 18:34:01 I had been hearing things about it not supporting cards over 512 meg, have you ever tested Aug 11 18:34:12 No, I haven't Aug 11 18:35:18 anyone here used a card over 512meg in an SL-5500 (success or failure)? Aug 11 18:37:10 I used a 512 Aug 11 18:37:20 I'm pretty sure I heard about some others using 1GB.. Aug 11 18:38:36 I've seen some questions on the OZ mailing list about using cards over 512, and no one ever responds. Aug 11 18:44:44 no response means pretty much nothing Aug 11 18:45:05 exactly, hence my question :-) Aug 11 18:54:39 yeah my zaurus uses 1gb kingmax but have had probs with other cheaper brands Aug 11 18:55:03 ZH_n_SB: that is an SL-5500? Aug 11 18:55:08 yep Aug 11 18:55:23 ZH_n_SB: thanks for the info, I've been trying to figure that out. Aug 11 18:55:25 my boss also has a 5500 and he had probs with certin brans Aug 11 18:55:31 mickey_away: when you compiled e-wm did you have edje installed locally? It seems to need it to compile. Aug 11 18:55:35 dat's kool Aug 11 18:55:39 ZH_n_SB: do you remember which ones failed? Aug 11 18:55:49 stupid 5500 SD controller... Aug 11 18:56:01 Zero_Chaos: this is not a distribution support channel, an oz channel, or a zaurus channel. Aug 11 18:56:08 Zero_Chaos: try somewhere where its on topic Aug 11 18:56:22 kergoth: all apologies, I will desist at once Aug 11 18:57:16 the cluebat is back Aug 11 19:05:42 any1 know how to kill a tor proxy client connection? Aug 11 19:25:54 damn....edje is really broken.... Aug 11 22:17:50 does any one know what imagemagick depends on ?. i am trying to write a .bb file for it Aug 11 22:19:28 03nail 07org.oe.nslu2-linux * r8c8d23c7... 10/packages/sysvinit/ (sysvinit/openslug/inittab sysvinit_2.86.bb): Update /etc/inittab, to prepare for pbutton fix Aug 11 22:20:29 check its configure.in Aug 11 22:21:15 eJumbo: BuildRequires: bzip2-devel >= 1.0.1 Aug 11 22:21:15 BuildRequires: expat-devel >= 1.95.7 Aug 11 22:21:15 BuildRequires: freetype-devel >= 2.0.2-2 Aug 11 22:21:15 BuildRequires: gd-devel >= 2.0.15 Aug 11 22:21:27 BuildRequires: ghostscript-devel Aug 11 22:21:35 BuildRequires: graphviz-devel >= 1.12 Aug 11 22:21:43 jasper-devel >= 1.700.5 Aug 11 22:21:50 BuildRequires: jbigkit-devel Aug 11 22:21:50 BuildRequires: lcms-devel Aug 11 22:21:59 BuildRequires: libfpx-devel Aug 11 22:22:18 BuildRequires: libjpeg-devel Aug 11 22:22:18 BuildRequires: libltdl-devel Aug 11 22:22:18 BuildRequires: libpng-devel >= 1.0.8 Aug 11 22:22:18 BuildRequires: libstdc++-devel Aug 11 22:22:18 BuildRequires: libtiff-devel Aug 11 22:22:20 BuildRequires: libtool >= 2:1.5 Aug 11 22:22:23 BuildRequires: libwmf-devel >= 2:0.2.2 Aug 11 22:22:25 BuildRequires: libxml2-devel >= 2.0 Aug 11 22:22:33 luzik: Image Magick requires all these ? Aug 11 22:22:34 thats all :) Aug 11 22:23:00 do u have a .bb file with u ? Aug 11 22:23:08 eJumbo: few of this packages are required for additionaly features Aug 11 22:23:12 can u pate it in pastebin Aug 11 22:23:27 eJumbo: no i don't have bb Aug 11 22:23:27 eJumbo: you can always check the gentoo ebuild Aug 11 22:23:44 will show you all of the optional ones as well Aug 11 22:30:34 eJumbo: http://rpmfind.net//linux/RPM/PLD/dists/ac/ready/i686/ImageMagick-6.1.8-2.i686.html Aug 11 22:32:18 03justinp 07org.oe.dev * r6db0fc38... 10/packages/pcre/pcre_4.4.bb: Commit band-aid for pcre until someone finds a better solution Aug 11 22:32:53 someona build opie-image yesterday? Aug 11 23:06:39 luzik: have u compiled imagemagick ? Aug 11 23:07:08 03nail 07org.oe.nslu2-linux * rf4bc4dc4... 10/packages/linux/ (2 files in 2 dirs): Aug 11 23:07:08 Fix pbutton to signal init, so it shuts down safely (Requires upgraded sysvinit) Aug 11 23:07:08 Fix rbutton to do machine_power_off(), since reset if unreliable. Aug 11 23:07:22 urgh Aug 11 23:07:40 * NAiL notes that it didn't include "nslu2-kernel" in the commit log Aug 11 23:08:21 i get these errors : coders/xwd.c:61:25: X11/XWDFile.h: No such file or directory Aug 11 23:18:47 i get these errors : coders/xwd.c:61:25: X11/XWDFile.h: No such file or directory while compiling imagemagick Aug 11 23:19:10 please see http://pastebin.ca/19859 Aug 11 23:21:24 * XWDFile.h MIT Project Athena, X Window system window raster Aug 11 23:21:25 * image dumper, dump file format header file. Aug 11 23:24:53 how do i include the package containing XWD: image dumper **** ENDING LOGGING AT Thu Aug 11 23:59:56 2005