**** BEGIN LOGGING AT Sat Jul 30 23:59:56 2005 Jul 31 02:15:18 moin Jul 31 02:16:22 moin greg Jul 31 02:26:13 morning all Jul 31 03:18:29 morning Jul 31 03:18:33 good afternoon all Jul 31 03:18:50 morning koen, CoreDump|home Jul 31 03:19:04 * CoreDump|home plays with xqt and icewm Jul 31 03:19:36 * RP can't see a way to solve tje corgi sound init order when using the new i2c code :-/ Jul 31 03:19:37 CoreDump|home: that is just so backwards :) Jul 31 03:20:12 koen: :) I find it kinda funny to have a X session inside a QT session Jul 31 03:21:10 x/qt proves that qt/e is 'compatible' and that 'no xserver is needed' Jul 31 03:22:05 hmm? Jul 31 03:22:22 I'm being sarcastic Jul 31 03:22:26 ah Jul 31 03:22:27 :) Jul 31 03:22:50 :) Jul 31 03:23:18 looks like icewm can't use the qpf fonts hmm Jul 31 03:23:25 * CoreDump|home build xfonts-xorg Jul 31 03:24:27 * koen is so happy opie is dumping qpf Jul 31 03:24:46 is it? Jul 31 03:25:05 OT: what hotplug script will mount my CF card? Jul 31 03:25:07 at least for !collie Jul 31 03:25:12 or how did it normally get mounted? Jul 31 03:25:34 cardmgr voodoo? Jul 31 03:25:52 yeah, cardmgr does that, in 2.4 at least Jul 31 03:28:45 ah right Jul 31 03:29:04 zecke: holy cow, we are in a bad mood, are we not? Jul 31 03:30:17 CoreDump|home: actually not at all, you should have seen what he queried me Jul 31 03:31:55 I saw that :) But you are giving him hell over basically nothing heh (mental disorder lol) Jul 31 03:32:29 CoreDump|home: "YOu won I'm going back to Win CE" was the friendliest thing he msged me Jul 31 03:32:31 * koen must've missed something Jul 31 03:32:46 CoreDump|home: a) it is not a competition b) I did not compete with him Jul 31 03:32:56 CoreDump|home: c) I'm happy if he would use a bug tracker Jul 31 03:33:25 CoreDump|home: d) the problem would have been solved now (editing one number in one file) Jul 31 03:33:33 koen: not really, we are talking about someone in #opie.de Jul 31 03:33:39 aha Jul 31 03:34:34 CoreDump|home: is mickeyl pcmcia applet known to work? Jul 31 03:34:44 zecke: to d) I have a unique view for that as I can't code C myself. It is at time _hard_, very very hard to find just this line. Even simple things take me hours and hours to do in C Jul 31 03:34:56 zecke: dunno, but it should at least compile Jul 31 03:35:14 CoreDump|home: I thought that was the point of C Jul 31 03:35:17 * treke|home ducks Jul 31 03:35:18 CoreDump|home: yes but we're here for guidance Jul 31 03:35:23 treke|home: :) Jul 31 03:36:04 damn our text editor is broken... we do not even have a undo redo... Jul 31 03:36:20 zecke: undo/redo is bloat Jul 31 03:36:30 s/bloat/fluff Jul 31 03:36:47 fluff is for the weak Jul 31 03:36:51 i wish we had something like tkcEditor Jul 31 03:37:12 heh Jul 31 03:38:12 mr gordon called me and nilsf 'jihadists' for saying maemo is using X and he should mentioning qt/e on the maemo-dev list Jul 31 03:38:26 /should/should stop/ Jul 31 03:38:58 * CoreDump|home kinda likes fluff :) Jul 31 03:39:22 ~lart xfonts-xorg for checking out the entire X.org tree Jul 31 03:39:37 'embedded' should not be the same as 'feature depleted' Jul 31 03:39:52 indeed Jul 31 03:39:52 How can I see if the ide scripts are getting ran? Jul 31 03:39:53 ipkg install ttf-bitstream-vera Jul 31 03:40:06 that should pull in fontconfig Jul 31 03:40:15 I hope xqt uses fontconfig Jul 31 03:40:39 or is it still in the 'corefonts ghetto'? Jul 31 03:41:21 koen: I don't wan't to use the feeds, want to get the stuff by compiling it locally. And maybe - just maybe - I'll do a xqt CardFS packe Jul 31 03:41:48 http://w3.mecanica.upm.es/~smuelas/debianxqt.html <- way cool Jul 31 03:43:30 why not install debian and use qvfb for the 2 or 3 sharp apps you want to use? Jul 31 03:44:29 because IMO a GUI like Opie (or GPE for that matter) is better suited for a PDA'ish device Jul 31 03:44:55 it is Jul 31 03:45:22 I have pocketworkstation running, but where's the point in installing binary stuff if you can just build it from source w/ OE :) Jul 31 03:45:27 I just got confused by icewm Jul 31 03:46:51 well, the "vnc" way of PW is kinda neat. Jul 31 03:47:50 maybe xqt is not the best way to go, but a VNC server in the background accessed by Opie VNC or Xvnc Jul 31 03:50:29 just install opie-x11 :) Jul 31 03:50:41 heh Jul 31 03:50:41 no strange indirections needed anymore Jul 31 03:50:56 I can' Jul 31 03:51:13 I can't get the stupid monotone thing to work, so I downloaded the oe tarball. Jul 31 03:51:21 What do I set my BBPATH to? Jul 31 03:51:51 I have /arrummzen/zaurus/openembdded-oz3.5.3 Jul 31 03:53:46 And it gets better, bitbake has not man page! Jul 31 03:53:58 arrummzen2: http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted#head-18960073be31b1f55cb2343d065bf13901c99c52 Jul 31 03:54:19 CoreDump|home: Do you edit ide.opts or does it work for you automatically? Jul 31 03:54:33 I have been reading that stupid (I want to ues a stronger word) Getting Started page and it doesn't work! Jul 31 03:54:50 zecke: I'm still using the old applet Jul 31 03:55:25 and it works out of the box FWIW Jul 31 03:57:07 CoreDump|home: It was probably a mix of partition, formatation and whatever magic... Jul 31 03:57:45 The problems are followed, FIRST: I do NOT have write access to /. SECOND: I can NOT get monotone to work so I HAD to download the tar file. Jul 31 03:58:02 I tried to properly set BBPATH myself, but it DOES NOT work. Jul 31 03:58:10 I tried BBPATH=/mnt/hdb/arrummzen/openembedded-oz3.5.3/build Jul 31 03:58:19 CoreDump|home: do you've entries in fstab? Jul 31 03:58:28 build contains local.conf Jul 31 03:58:55 arrummzen2: that's VERY SAD Jul 31 03:59:08 What is? Jul 31 03:59:17 that it DOES NOT work for you. Jul 31 03:59:17 zecke: yes, /dev/hda1 -> /media/cf Jul 31 03:59:21 Be more specific please Jul 31 03:59:23 CoreDump|home: okay Jul 31 04:00:24 CoreDump|home: that explains it Jul 31 04:01:23 yo zecke Jul 31 04:03:52 zecke: err, well, I _do_ have a hack in place to mount different partitions, but hda1 should work w/o it Jul 31 04:05:55 hi pb_ Jul 31 04:06:00 hi rp Jul 31 04:06:40 * zecke rages against linux in regards to removable stuff (usb, pcmcia..) Jul 31 04:07:01 arrummzen2: reading the page isn't enough, you have to actually perform the steps outlined in there Jul 31 04:07:31 And when the steps don't work? Jul 31 04:07:40 you didn't say that Jul 31 04:07:58 it DOES NOT work if *not* descriptive enough Jul 31 04:08:03 we are not telepathic Jul 31 04:08:07 pb_: Any tips on how a i2c client could signal to some sound driver code that its init'd and is therefore ready to be accessed? Jul 31 04:08:29 koen: HOW MANY fingers am I holding UP ? Jul 31 04:08:36 I'm favouring a function pointer in the i2c driver which the platform init code sets to some function which only registers that platforms sound device once the i2c client has been setup. Is there a neater way you can see? Jul 31 04:08:36 koen: monotone couldn't fetch the code. Jul 31 04:08:43 rwhitby-away: ehm... 6? Jul 31 04:08:47 I got an error about getting 5 and expecting 4. Jul 31 04:09:00 koen: yep, you're not telepathic :-) Jul 31 04:09:04 rwhitby-away: :) Jul 31 04:09:15 arrummzen2: you do have monotone 0.21, right? Jul 31 04:09:51 holy shit, xorg is _still_ checking out Jul 31 04:09:58 Hmmm. I was thinking the above approach would work with modules but it wont :-/ Jul 31 04:10:02 !lart xfonts-xorg Jul 31 04:10:02 * cdbot stabs xfonts-xorg Jul 31 04:10:53 RP: I'm not exactly clear on what you mean. Can you give me a bit more context? Jul 31 04:11:01 koen: monotone-0.18 is the latest I think. Jul 31 04:11:03 CoreDump|home: minimo's do_fetch has been running for hours now Jul 31 04:11:19 "). If you run into a problem at this stage, check to make sure you have at least version 20 of monotone installed" Jul 31 04:11:34 straight from GettingStarted Jul 31 04:12:30 koen: check its actually running do_fetch, my minimo runs do_compile instead and pkg-config is actually running forever Jul 31 04:12:55 ah yes Jul 31 04:13:00 ~stab minimo Jul 31 04:13:02 install monotone installed .18... Jul 31 04:13:27 pb_: I'm writing an i2c driver+client for the wm8731 in corgi. I need to find some way of letting the main corgi sound driver access it. The problem is there don't seem to be any hooks that will let me control the init order. I want to avoid the situation where corgi_snd tries to access the wm8731 but the i2c code hasn't got itself setup yet... Jul 31 04:14:14 All this needs to be done in a none platform and none arm sepcific way to be ideal... Jul 31 04:14:55 RP: using a notifier chain? (like lcd and backlight class?) Jul 31 04:15:47 arrummzen2: While attempting to build monotone, did you get succeed with it? You need at least 0.20, I am itself using monotone 0.21 for which sources are available in the http://venge.net/monotone For building it you need first build boost, for which there are step by step instructions in chapter 1.1 of http://venge.net/monotone/INSTALL Just copy/paste lines from there one by one and press enter, so you can get boost required by monotone build. Jul 31 04:17:20 I supose ill have to install .21 by hand. Jul 31 04:17:40 zecke: The notifier chain only exists in the fb code. I'd rather not have to write one for the sound system but it has a certain appeal :) Jul 31 04:17:41 This is the first time install failed to get the latest copy... Jul 31 04:17:49 the .deb worked for me on debian/unstable Jul 31 04:17:57 RP: ah, hm, right. I was about to say "wouldn't the corgi glue code just use i2c_get_client to find the wm8731 driver", but that function seems to have gone away. Jul 31 04:18:23 arrummzen2: Yes, that was what I did and with that I have been able to follow getting started guide. Jul 31 04:18:38 pb_: That also doesn't help me an init time. If the client doesn't exist yet, how would you wait until it does and get called back? Jul 31 04:18:48 RP: it will return NULL if it doesn't exist. Jul 31 04:19:18 which your driver would respond to by returning -ENODEV or something. Jul 31 04:20:14 Watch your for 2 year old childs. My MMC card just vanished from the USB card reader and the only explanation is my child... :-( Jul 31 04:20:31 lamikr: hope he does eat them Jul 31 04:20:34 doesnt Jul 31 04:20:36 pb_: I think my corgi_snd driver can't complete its probe with out sending things to the wm8731. Whether that can be worked around remains to be seen but its how the code stands at the moment Jul 31 04:20:52 RP: why's that? surely it would only need to talk to the codec at open() time? Jul 31 04:21:21 xorAgone: May really be possible, when asking where she put it, she is just repeating my words like a parrot :-) Jul 31 04:22:00 mount /dev/small/child/mmc /media/card :-) Jul 31 04:22:07 RP: doesn't the notifier chain come from the networking code? Jul 31 04:22:44 zecke: Not in the backlight/lcd drivers. It may be based on something similar - I don't know... Jul 31 04:23:24 pb_: I'm not 100% sure why its wants the early access. The existing code isn't entirely readable - I'm trying to replace chunks of it in stages... Jul 31 04:23:37 ah, I see Jul 31 04:24:44 Step 1 is to bin all the i2c code and make a nicer driver but I can't see a neat way to interface the two. hh.org's code looks to have a similar problem :-/ Jul 31 04:25:44 I guess it'd be worth writing to the mailing list. Maybe Giorgio will have some ideas. Jul 31 04:27:06 If I force the i2c driver to be built in, I can see a way to do it... Jul 31 04:33:21 * koen has a stab at upgrading gnome to 2.10 Jul 31 04:33:46 I guess I could use module_notifier to work around the situation when its built as a module. Hardly neat though :-/ Jul 31 04:39:40 * pb_ goes back to creating pcb footprints for a bit Jul 31 04:40:24 Anyone got a good recipe for working out the diff between the result of your merge, and what you had before ? Jul 31 04:40:53 does cat revision work on merges? Jul 31 04:41:04 rwhitby-away: I was wondering about the same thing last week Jul 31 04:42:07 I just had to do a merge, and I have no way of knowing if I did it right or not. So I'm going to have to blow away my db and start again, for fear of reverting a change someone else made. Jul 31 04:43:25 luckily I had no outstanding local changes Jul 31 04:45:11 03koen 07org.oe.dev * ra96c9d73... 10/packages/gnome/ (3 files in 2 dirs): update libbonobo to 2.10.0 Jul 31 04:46:42 actually, just knowing which is left and which is right when I do a propagate would probably suffice. Jul 31 04:50:13 Can somebody help me a bit on my way with fixing http://bugs.openembedded.org/show_bug.cgi?id=192 Jul 31 04:50:21 03koen 07org.oe.dev * r76e9e706... 10/packages/gnome/libgnome_2.10.1.bb: update libgnome to 2.10.1 Jul 31 05:13:28 nick rwhitby Jul 31 05:13:49 * rwhitby does some experiments to work out which is left and right for a propagate ... Jul 31 05:19:49 ~lart the xqt project for not having source tarballs Jul 31 05:20:01 CoreDump|home: hehe Jul 31 05:20:34 CoreDump|home: you might need to patch some brain damage of Xqt as well (if you're not on a C-xxx device) Jul 31 05:20:37 by the looks of it they don't even have tagged their sources. Which is great since the lastest version is broken Jul 31 05:24:49 plan B was to use fbvncserver instead - which doesn't compile heh Jul 31 05:27:46 plan C: use seperate consoles for qt/e and X Jul 31 05:28:21 make gpe ditch atd and use opie-alarm Jul 31 05:29:18 oh /me broke a gadget Jul 31 05:29:22 * koen should poke mccarthy for his patches to atd Jul 31 05:29:54 * zecke searches food Jul 31 05:31:09 koen: I did that in the past w/ GPE and it sucked Jul 31 05:31:56 plan D: raise venture capital and pay people to port opie to qt4/x11 Jul 31 05:32:01 :) Jul 31 05:37:44 Where can I see which kernels are available for my built ? Jul 31 05:43:24 ah well, I managed to compile fbvncserver by patching out the offending bits. Not very elegant =) Jul 31 05:44:01 gerwinin: dunno. only with the bitbake shell Jul 31 05:44:18 could someone please verify that bitbake fbvncserver b0rks? Jul 31 05:46:02 CoreDump|home: typed it in Jul 31 05:46:10 thx Jul 31 05:47:22 CoreDump|home: libvncserver built Jul 31 05:47:42 rfbAlwaysShared? Jul 31 05:47:42 build here, too =) now comes the interesting part Jul 31 05:49:30 CoreDump|home: fbvncserver fails with two errors Jul 31 05:49:36 CoreDump|home: I'm trying version 0.9.5 now Jul 31 05:49:51 zecke: 0.9.5 fails to patch Jul 31 05:51:26 03koen 07org.oe.dev * r200ebaff... 10/packages/gnome/gnome-mime-data_2.4.2.bb: update gnome-mime-data to 2.4.2 Jul 31 05:51:38 and after fixing that fails on the same missing elements Jul 31 05:52:09 hehe Jul 31 05:52:14 CoreDump|home: who maintains fbvncserver? Jul 31 05:52:48 good question Jul 31 05:53:21 I'll patch the 2 errors out from 0.9.4 Jul 31 05:55:06 03koen 07org.oe.dev * r824eb36b... 10/packages/gnome/ (libbonoboui_2.10.0.bb libgnomecanvas_2.10.2.bb): Update libbonoboui and libgnomecanvas to 2.10.x Jul 31 06:12:02 CoreDump|home: did we upgrade to a newer libvncserver? Jul 31 06:12:29 latest is 0.7.1 (which we use) Jul 31 06:14:50 gi all Jul 31 06:14:58 oh Jul 31 06:15:09 s/gi/hi Jul 31 06:15:43 hi florian Jul 31 06:18:30 koen: That request from tapo reminded me to the nasty locale issue we still have in OE. Jul 31 06:21:38 03ccsmart 07org.oe.nslu2-linux * r2692c20b... 10/packages/initscripts/ (3 files in 2 dirs): Migrate per package configuration files in drop in directory. Jul 31 06:23:38 florian: locale-base-xx? Jul 31 06:23:50 what's pulling that in anyways Jul 31 06:24:36 CoreDump|home: no the localegen oom issue (and the lack of gpe locale desriptions) Jul 31 06:24:57 florian: we are talking about the same thing :) Jul 31 06:29:00 * koen has a h5550 with 128MB ram Jul 31 06:29:13 * koen asks "what oom?" Jul 31 06:29:38 CoreDump|home: ehem.. okay :-) Jul 31 06:29:43 koen: out of memory Jul 31 06:30:27 koen: does localegen succeed on it when it is run the first time oe starts? Jul 31 06:37:04 florian: I think koen was just boasting about his 128MB RAM. Jul 31 06:37:17 :) Jul 31 06:37:34 it's true that, if you are an eleet 128MB-using h4x0r, you can run localegen with impunity. Jul 31 06:37:45 pb_: hi Jul 31 06:37:58 aquadran: hello Jul 31 06:38:06 pb_: heh... okay Jul 31 06:38:32 * florian steals koen half of his RAM and let him look again Jul 31 06:38:39 heh Jul 31 06:44:50 more than 64MB RAM is fluff anyways heh Jul 31 06:45:13 running localgen with 64MB is a bit of a roulette Jul 31 06:51:09 03rwhitby 07org.oe.nslu2-linux * racedaeff... 10/packages/linux/ (2 files in 2 dirs): Added the missing-exports patch from Matta to the OpenSlug kernel, to resolve the missing hotplug_path and add_input_randomness symbols. Jul 31 06:51:36 is this a bug that localegen takes up that much of memory? Jul 31 06:57:53 for our purposes, certainly Jul 31 06:58:03 as far as Uli Drepper is concerned, probably not. Jul 31 07:01:21 oh well... that's the maintainer? Jul 31 07:01:45 mr Drepper is also known as mr glibc Jul 31 07:02:14 Ah... that's why that name sounds familiar :-) Jul 31 07:03:04 Can we maybe run this at build time on the build machine? Jul 31 07:04:34 hmm Jul 31 07:04:41 anyone used Trac before? Jul 31 07:04:54 emte: opensync.org Jul 31 07:04:56 its pretty , i'lladmit that much Jul 31 07:05:50 there we go Jul 31 07:06:49 wasn't there a "donations" page at openzaurus.org somewhere? Jul 31 07:07:45 pb_: It looks like the early accesses to the wm8731 were to play null audio for two seconds in a kernel thread as the kernel boots. My delete key is getting lots of exercise :) Jul 31 07:08:01 ah got it Jul 31 07:08:18 * RP shudders at the sharp code Jul 31 07:10:33 RP: cute Jul 31 07:11:21 florian: not easily, because the locale archive is a non portable binary format Jul 31 07:11:23 RP: and someone got paid for it as well Jul 31 07:11:54 florian: it's also one monolithic file containing many locales, so you would have a problem if you wanted to install more locales later. Jul 31 07:13:14 zecke: How anyone got paid for any of the Sharp code is a big mystery... Jul 31 07:13:31 pb_: That sounds nasty... Jul 31 07:13:32 RP: I know people getting paid for ripping drivers from hh.org :} Jul 31 07:13:47 florian: glibc is simply not well suited for our purposes Jul 31 07:15:04 Zecke: I wish I was that person Jul 31 07:15:46 zecke: maybe, but imho we don't really have a choice and i think it might be possible to improve the situation. Jul 31 07:15:58 * koen looks at the timezone corrected http://ewi546.ewi.utwente.nl/OE_qa/OpenEmbeddedBuild/status.html Jul 31 07:16:02 zecke: congrats! Jul 31 07:16:54 Koen can you add x86 opie-image ? Jul 31 07:17:21 no, but you can :) Jul 31 07:17:32 Koen : Okay Jul 31 07:18:44 koen: the bars for Unslung are too long Jul 31 07:20:32 nice ... i update gaim and debian removed all my kde support ... Jul 31 07:20:47 ~hail debian Jul 31 07:21:44 yeah, it wouldnt be so bad but i was using pikdev Jul 31 07:21:55 emte: libaspell ;) Jul 31 07:22:00 yahoo, MMC card found! Back to resolving why gpe-login still crashes with h6300. Jul 31 07:22:09 ah would that be the culprit Jul 31 07:22:23 stupid thing, i can spell bad enough on my own Jul 31 07:26:33 zecke: I updated bitbake and monotone but I keep on having the error as described in bug 192 can you have a look and indicate where I need to start to fix this bug ? Jul 31 07:29:39 currently not Jul 31 07:31:01 aargh Jul 31 07:31:06 ~lart scummvm people Jul 31 07:31:37 reinventing autotools Jul 31 07:32:19 * CoreDump|home dissects pdaXrom Jul 31 07:34:19 03nail 07org.oe.nslu2-linux * r84ca6b48... 10/packages/ (3 files in 3 dirs): Added client programs for controlling the the mpd music player from the commandline, courtesy of Hannes Reich Jul 31 07:34:23 03nail 07org.oe.nslu2-linux * r9bfa38ef... 10/packages/at76c503a/at76c503a-modules_cvs.bb: Fixed compilation for 2.4 and 2.6 (.o vs .ko). Fixes OE bug #190 Jul 31 07:34:27 03nail 07org.oe.nslu2-linux * r17e420c5... 10/packages/portmap/portmap_5-9.bb: Added portmap-utils RDEPENDS so nfs works out of the box. Fixes slugbug #229 Jul 31 07:36:54 If I mix my own bb files in the org.openembedded.dev directory, and do a mt update, are files still there after the update? Jul 31 07:37:51 micropal: yes. monotone even starts up meld if needed if you modify existing files. Jul 31 07:39:08 03nail 07org.oe.nslu2-linux * r17cd385f... 10/packages/ (3 files in 3 dirs): disapproval of revision 84ca6b48d0e2960ee88c5a4d6c227b9cc53e4d46 Jul 31 07:41:24 since i broke pikdev anyway ... lets see how a system upgrade fares ... Jul 31 07:41:29 ok, thanks. Jul 31 07:49:34 could someone update the defconfig for oz-2.6.12 and add these for the cfu1: Jul 31 07:49:35 CONFIG_USB_SL811_HCD=m Jul 31 07:49:35 CONFIG_USB_SL811_CS=m Jul 31 07:50:01 I meant oz-2.6.13-rc3-mm3 Jul 31 07:58:10 i take it they are all asleep in #openzaurus? Jul 31 08:00:07 looks like Jul 31 08:02:18 B0Ti: could you make that a enhancement req in bugzilla, so it doesn't get lost? Jul 31 08:03:45 up late raster? Jul 31 08:04:01 emte: aye squiddie Jul 31 08:04:05 will be napping soon Jul 31 08:04:26 thats good, you need all your energy for your strike Jul 31 08:04:29 :) Jul 31 08:04:34 ahahaha Jul 31 08:12:53 hey kergoth Jul 31 08:13:17 hi Jul 31 08:13:26 hi hrw|wrt Jul 31 08:13:27 hey hrw|wrt Jul 31 08:13:44 clientu Jul 31 08:13:46 ops Jul 31 08:13:55 koen: filed as #194 Jul 31 08:13:56 monotone/client still preferred 0.20? Jul 31 08:14:24 0.21 if you know how eplicit_merge works :) Jul 31 08:14:35 B0Ti: Thanks Jul 31 08:14:38 ah.. Jul 31 08:17:33 03nail 07org.oe.nslu2-linux * rf99b0f81... 10/ (4 files in 4 dirs): Back out RDEPENDS Jul 31 08:21:42 03nail 07org.oe.nslu2-linux * reab45770... 10/ (4 files in 4 dirs): disapproval of revision f99b0f8157e1cebc59aada3ef6513fcf8e724c48 Jul 31 08:22:10 horrible day Jul 31 08:24:18 03nail 07org.oe.nslu2-linux * r8fb3851e... 10/packages/portmap/portmap_5-9.bb: Correct commit. Back out the RDEPENDS from portmap Jul 31 09:15:10 ping RP Jul 31 09:15:16 uv1: hi Jul 31 09:15:45 RP: Have problem to build 2.6.13 kernel (monotone pull from an hour ago). Jul 31 09:16:06 uv1: Can you be more specific? :) Jul 31 09:16:50 corgi_tspmu-r2.patch not found. Tried -r3 but had problems to patch w100_core-r1 or w100_corgi-r1. Jul 31 09:18:33 If I use archive and corgi_tspmu-r2 it works. Jul 31 09:18:50 uv1: I've moved it back out of archive Jul 31 09:19:20 I'm not sure why the r3 version isn't working. I've probably screwed up the version numbers somewhere along the line... Jul 31 09:20:37 BTW : What is the state of the CL860 suspend / resume state. Jul 31 09:20:50 uv1: It should be working as far as I know Jul 31 09:20:52 Why? Jul 31 09:21:35 With the last 2.6.12 now and then the system wakes up but does not reach opie. Instead of that it sticks on the console and DRAINS the battery ;-( Jul 31 09:21:36 It could be -mm3 is behaving a little oddly btw - I'm seeing Xfbdev hanging here and I'm not sure why... Jul 31 09:22:47 Also more often, that I saw before, the power button causes the system to stuck. Do you think, thats fixed with 2.6.13 ? Jul 31 09:23:06 Yes, the power button is fixed Jul 31 09:23:17 I've seen the opie bug but have never been able to reproduce to debug Jul 31 09:23:36 My only advice would be to see if 2.6.13 does it :-/ Jul 31 09:23:40 It could be related to the power button issues... Jul 31 09:23:48 OK, will try 2.6.13 then and inform you about the result. Jul 31 09:24:14 With the power button, keep in mind it queues the requests so if you press it 5 times and the system fails to suspend, it will remember for later 8-) Jul 31 09:24:58 But you can only press it once per second of so now... Jul 31 09:25:24 So you mean, that the bouncing keyboard is also history ? Jul 31 09:25:40 No, that's a different problem :-/ Jul 31 09:25:55 That's the input layer being a pain Jul 31 09:26:23 If I can get all the existing input fixes into mainline, I might be more inclined to look at these other issues... Jul 31 09:26:23 But isnt that a possible root cause for the power button behaviour as well ? Jul 31 09:26:36 No... Jul 31 09:27:23 Hmm : unrsolved symbols in irnet, irlan ? Jul 31 09:30:14 Building opie-image ... That'll take a while. Thanks for now and bye ! Jul 31 09:30:52 uv1: Let me know how it goes Jul 31 09:32:12 RP: Be sure ;-) Jul 31 09:35:06 SD patches have made 2.6.13-rc4-mm1 ! Jul 31 09:36:38 great Jul 31 09:37:21 RP: Is that in org.openembedded already ? Jul 31 09:37:52 ~kill irk modders Jul 31 09:39:20 uv1: We use an old version of those patches in the openzaurus kernel tree Jul 31 09:39:58 old in that I know it works and haven't bothered to keep with with things like 4 bit which we don't need (yet) Jul 31 09:40:38 RP: OK Jul 31 10:27:06 03koen 07org.oe.dev * rbda34a30... 10/packages/gpe-mini-browser/gpe-mini-browser_0.15.bb: Add gpe-mini-browser 0.15, courtesy Philippe DeSwert Jul 31 10:35:33 03nail 07org.oe.nslu2-linux * r6c6537eb... 10/packages/meta/openslug-packages.bb: Added timezones for glibc builds Jul 31 10:35:38 03nail 07org.oe.nslu2-linux * r121ddca6... 10/conf/distro/openslug-packages.conf: Forgot to add it here ;-) Jul 31 10:57:28 evening Jul 31 10:57:44 hi mickeyl Jul 31 10:58:00 RP: Syntax error in keymap file line 146. Is that known ? Jul 31 11:00:45 ping RP Jul 31 11:00:50 03jbowler 07org.oe.nslu2-linux * rebd51a9e... 10/packages/ixp425-eth/ixp425-eth_1.1.bb: Remove the unnecessary and confusing copy of the patch Jul 31 11:00:55 03jbowler 07org.oe.nslu2-linux * r214e2822... 10/ (8 files in 2 dirs): Jul 31 11:00:55 Add support for UcSlugC builds. Jul 31 11:00:55 (1) the 'standard' settings in local.conf.template (from the nslu2-linux.org Jul 31 11:00:55 master makefile) have been folded back into the three distro .conf files. Jul 31 11:00:55 (2) openslug-image now builds for any NSLU2 distro Jul 31 11:00:55 (3) ucslugc configuration added Jul 31 11:00:57 (4) more comments in the openslug and ucslugc configuration files Jul 31 11:00:59 03mickeyl 07org.oe.dev * r2c9a9f52... 10/packages/unionfs/ (12 files in 5 dirs): unionfs: add make.patch for 1.0.13 and remove a couple of obsolete versions Jul 31 11:01:08 hi mickeyl Jul 31 11:01:29 uv1: that isn't know. What's reporting that? When you try to boot the image? Jul 31 11:01:40 nice, UcSlugc Jul 31 11:02:05 RP: yes. And keymap is weird. Cannot find "/" e.g. Jul 31 11:02:22 And opie does not start ;-( Screen stays blank. Jul 31 11:02:28 uv1: It would be weird if it didn't load. Blame those messing with the keymap files recently Jul 31 11:02:31 s/blank/black/ Jul 31 11:03:01 It didnt load (said the boot output) Jul 31 11:03:06 uv1: The same problem as gpe then. I'm blaiming -mm3... Jul 31 11:03:13 monotone: fd 8 (peer 66.159.209.54:44218) exchanged goodbyes and flushed output, disconnecting Jul 31 11:03:16 monotone: fatal: std::runtime_error: network error: recv failure: Connection timed out Jul 31 11:03:19 whoops Jul 31 11:03:24 wrong chan, sry Jul 31 11:03:26 The keymap issue is more to do with oe - people have been messing with that recently... Jul 31 11:03:36 mickeyl: A few bits of good news. A load of corgi updates were accepted into -mm bypassing the input maintainer Jul 31 11:03:54 RP: congrats Jul 31 11:04:06 mickeyl: The switch event type is in there so in theory set to make mainline after 2.6.13 Jul 31 11:04:22 uv is trying the old kernel ;-( Jul 31 11:05:06 uv1: change the -mm3 to -mm2 in the kernel .bb and try that. I'd try locally but I've too many editted files around Jul 31 11:05:36 mickeyl: Pierre Ossman has also got the SD drivers into -mm :) Jul 31 11:05:59 RP: woot. this is good stuff indeed Jul 31 11:06:16 RP: did we complain about the KERNEL_VERSION goof yet ? Jul 31 11:06:23 hi mickey Jul 31 11:06:32 hey pb_ Jul 31 11:06:38 hey mickeyl Jul 31 11:06:46 hi Jul 31 11:06:58 mickeyl: Others have - its a change the external modules will have to adapt to Jul 31 11:07:09 Kernel devs response is "tough" Jul 31 11:07:44 hmm Jul 31 11:07:54 is the macro gone for good? Jul 31 11:07:58 or just moved to another place? Jul 31 11:08:03 (and why... btw.) Jul 31 11:08:11 mickeyl: moved to utsname.h Jul 31 11:08:28 hmm Jul 31 11:08:38 ok, we don't have _that_ many external modules in oe Jul 31 11:08:39 because its a constant, not a variable and therefore doesn't belong in a variable file apparently Jul 31 11:09:02 speaking about kernel version... can we add PR into kernel version? 2.6.11-r20 is different then 2.6.11-r16 Jul 31 11:09:06 http://lkml.org/lkml/2005/7/31/52 Jul 31 11:09:32 hrw|wrt: we could attach it to the openzaurus release Jul 31 11:09:34 * mickeyl reads Jul 31 11:09:52 hah Jul 31 11:10:12 * hrw|wrt will update to 2.6.13-rc3-mm3 on husky... Jul 31 11:10:30 hmm.. 4M free ram left... Jul 31 11:10:36 I'm not sure -mm3 is working properly - I'd go to -mm2 Jul 31 11:10:45 or -rc4-mm1 :) Jul 31 11:10:50 sometimes i really wonder whether the oss people still have a clue about a natural balance of doing the technically right thing versus generating a lot of additional workload for a ... uhm... marginal outcome Jul 31 11:11:27 mickeyl: I know what you mean. Adrian bunk is the one you have to thank for all the elpp problems ;-) Jul 31 11:11:38 RP: how to go to mm2 ? Just changing *mm3 to *mm2 does not work ;-( Jul 31 11:11:48 ~lart adrian bunk for having to catch up with every minor release Jul 31 11:11:54 +me Jul 31 11:12:13 uv1: You mean patches don't apply or what exactly? Jul 31 11:12:19 RP: Yes Jul 31 11:12:37 RP: /home/sc/scdev/oe/sources/2.6.13-rc3-mm2-r0.patch' Jul 31 11:12:37 Resolving www.oesources.org... 192.216.230.225 Jul 31 11:12:37 Connecting to www.oesources.org|192.216.230.225|:80... connected. Jul 31 11:12:37 HTTP request sent, awaiting response... 404 Not Found Jul 31 11:12:37 20:08:25 ERROR 404: Not Found. Jul 31 11:13:01 remove cvs_tarball_stash Jul 31 11:13:36 RP: so which version of 2.6 works now on husky? Jul 31 11:14:08 re Jul 31 11:14:12 re florian Jul 31 11:14:16 hrw|wrt: 2.6.12-rc3-mm1 is known to work. 2.6.12-rc3-mm3 has something wrong Jul 31 11:14:35 ah... Jul 31 11:14:42 where do I put my 2.6.12 modules... Jul 31 11:14:46 uv1: Wait a while and I'll produce a fixed file one way or another Jul 31 11:15:07 RP: uv1 takes a shower ... Jul 31 11:15:13 uv1: You need to change more than that line. That was just one of my fixes for that release... Jul 31 11:15:15 uf... got 2.6.13-rc3-mm1 in deploy/images Jul 31 11:15:36 RP: I changed all mm3 to mm2 Jul 31 11:16:06 You can delete that 2.6.13-rc3-mm3-r0.patch line entirely Jul 31 11:16:11 (for -mm2) Jul 31 11:16:52 mickeyl: Locally, I've started work on the corgi sound driver. I have it working with the core linux i2c driver rather than Sharp's hacks Jul 31 11:17:14 * RP enjoys deleting Sharp code :) Jul 31 11:17:25 RP: alsa or oss? Jul 31 11:17:56 RP: Kernel builds, hx2750_bl-r1.patch fails to patch. Jul 31 11:18:28 Change it to -r0. Also remove the tosa_bl line Jul 31 11:19:06 (You're not building for tosa right? :) ) Jul 31 11:19:43 RP: No, Boxer instead. Jul 31 11:20:16 RP: Where can I find the keymap file ? Jul 31 11:20:28 there is no such thing as boxer... ;) Jul 31 11:20:29 RP: I mean in tmp/work Jul 31 11:20:41 Boxer = SL-C860 ? Jul 31 11:20:53 It was part of initscripts Jul 31 11:21:15 maybe it gots fscked during merging Jul 31 11:21:15 yes, but boxer = husky = c760 as far as I'm concerned ;-) Jul 31 11:21:55 I'm off for food. A -mm2 kernel is compiling. If it works, I'll move back to that later on Jul 31 11:21:55 03jbowler 07org.oe.nslu2-linux * r174865e1... 10/packages/meta/openslug-packages.bb: Remove duplicated packages. Jul 31 11:23:24 RP: The keymap file is definitely rubish line 146 : "compose '`' 'A' to '�". Is that correct ? Jul 31 11:24:56 I have older version on my Z Jul 31 11:29:28 hostap 0.3.7 fails with 2.6.13-rc3-mm1 on husky Jul 31 11:33:25 hostap_crypt: registered algorithm 'NULL' Jul 31 11:33:25 hostap_cs: 0.3.7 - 2005-02-12 (Jouni Malinen ) Jul 31 11:33:25 pcmcia: hostap_cs does misses a callback function<7>hostap_cs: setting Vcc=33 (constant) Jul 31 11:34:45 hostap_crypt: registered algorithm 'NULL' Jul 31 11:34:46 hostap_cs: 0.3.9 - 2005-06-10 (Jouni Malinen ) Jul 31 11:34:46 pcmcia: hostap_cs does misses a callback function<7>hostap_cs: setting Vcc=33 (constant) Jul 31 11:34:49 ;( Jul 31 11:36:21 ok. I have good keymap file iirc Jul 31 11:37:30 hrw|wrt: Which version of initscripts ? in r50 it seems to be OK still. Jul 31 11:37:31 RP: sorry, was getting distracted. Good news on the sound driver. May I pester you again because of the blinkin' lights? :)) Jul 31 11:37:53 http://pastebin.ca/18946 Jul 31 11:39:53 03nail 07org.oe.nslu2-linux * rc3ddcc36... 10/packages/meta/openslug-packages.bb: Moved manpages to glibc-only section and strace back to working Jul 31 11:44:07 hrw|wrt: thanks, the r50 map is working as well. Opie works with RP's mm2 hints. Jul 31 11:44:50 03jbowler 07org.oe.nslu2-linux * rc462fc4d... 10/packages/meta/openslug-packages.bb: Merge the man/man-pages change Jul 31 11:48:58 :( this thing is really blocking me Jul 31 11:58:13 gerwinin: what was the bug number again?! Jul 31 11:58:39 New kernel, wrong modules. What is the 2.6.13-mm2 module package(s) ? Jul 31 12:00:00 192 Jul 31 12:01:23 Zecke: I renewed the epia configs if it is finsihed it should support all boards that are sold now Jul 31 12:32:42 zecke: I'm starting to run a gpe-image + opie-image tinderrebuild on ewi, like the slug people do Jul 31 12:33:48 koen|tv: good Jul 31 12:33:58 koen|tv: I need to talk with proti to 'fix' bitbake Jul 31 12:34:07 koen|tv: so we can have true incremental builds as well Jul 31 12:35:10 cool Jul 31 12:36:38 cu Jul 31 12:58:33 D'oh Jul 31 12:58:38 ? Jul 31 12:58:42 "don't use '/'" Jul 31 12:58:50 I typed that myself Jul 31 12:58:54 lol Jul 31 12:58:58 * koen|tv just removed the '/' Jul 31 13:06:13 the build are looking good Jul 31 14:00:43 03coredump 07org.oe.dev * rf82fa51c... 10/packages/sharp-binary-only/sharp-sdmmc-support-2.4.18-rmk7-pxa3-embedix/sdcontrol: sdcontrol: Fix parsing when /media/card is defined twice in fstab Jul 31 14:00:50 03coredump 07org.oe.dev * r1ff1e268... 10/packages/fbvncserver/ (files/buildfix.patch fbvncserver_0.9.4.bb): fbvncserver: compile fix Jul 31 14:00:55 03coredump 07org.oe.dev * r8dcbf324... 10/packages/sharp-binary-only/sharp-sdmmc-support_2.4.18-rmk7-pxa3-embedix.bb: Bump PR Jul 31 14:06:38 zecke: bug resolved Jul 31 14:07:44 :) Jul 31 14:07:52 gerwinin: good and sorry I'm busy ATM Jul 31 14:08:40 zecke: no problem , I am going to watch a krimi called der emittler :) Jul 31 14:08:54 zecke: in the meanwhile my image is compiling Jul 31 14:09:18 zecke: any idea why the tinderserver doesn't list coloured bars anymore? Jul 31 14:10:10 koen|tv: what is the tinderlog_oe.log saying? Jul 31 14:11:12 [Sun Jul 31 22:10:32 2005] Error in: Persistence::load_structure reading file: /data/tinderbox2/data_dir/OpenEmbeddedBuild/db/Build.Update..h3900.familiar083.opie.gpe.1133205032 can't open /data/tinderbox2/data_dir/OpenEmbeddedBuild/db/Build.Update..h3900.familiar083.opie.gpe.1133205032: No such file or directory at /data/tinderbox2/lib/Persistence/Storable.pm line 116 Jul 31 14:11:30 I guess I killed the tinderserver with my '/' adventures Jul 31 14:12:01 koen|tv: delete that update file Jul 31 14:13:20 bitbake:~/feed$ ls /data/tinderbox2/data_dir/OpenEmbeddedBuild/db/Build.DBdat Notice.DBdat Jul 31 14:14:03 oh Jul 31 14:14:09 koen|tv: call tinder.cgi manually? Jul 31 14:14:21 bitbake:~/feed$ /data/tinderbox2/cgi-bin/tinder.cgi --daemon-modebitbake:~/feed$ Jul 31 14:14:24 no output Jul 31 14:14:41 koen|tv: --daemon-mode == shut up Jul 31 14:14:59 without that it's the same Jul 31 14:15:18 koen|tv: tinder.cgi grabs all Build.* and places them in Build.DBdat Jul 31 14:15:34 koen|tv: so if there are no files (created by process_*) it will not update anything Jul 31 14:15:46 koen|tv: question is why are people not sending logs? ;) Jul 31 14:17:29 * koen|tv TINDER_BUILD = "${MACHINE} ${DISTRO}-${DISTRO_VERSION} opie gpe" Jul 31 14:18:05 koen|tv: sorry I'm busy ATM Jul 31 14:18:14 zecke: no problemo Jul 31 14:18:32 is there anyone who made scons work under OZ? Jul 31 14:18:44 koen|tv: try wget ewi/OE_qa/cgi-bin/process_builds.cgi Jul 31 14:19:55 [Sun Jul 31 23:19:39 2005] process_builds.cgi: No Tinderbox variables found in message. Jul 31 14:20:01 EnceloAMD: you mean OE? Jul 31 14:21:20 ohh Jul 31 14:21:30 koen|tv: not bad if it was your wget Jul 31 14:21:37 zecke: wget Jul 31 14:21:45 koen|tv: I think it's the same, because OZ 3.5.3 is based on OE :) Jul 31 14:21:53 ehm Jul 31 14:21:55 no Jul 31 14:22:12 OE is a *buildsystem*, OZ is a *distro* Jul 31 14:23:20 koen|tv: yes, I know this... but the OE buildsystem is responsible for what is inside an OZ 3.5.3 package... Jul 31 14:23:50 OE is also a tinderclient framework Jul 31 14:23:54 koen|tv: and python-scons OE package seems missing a lot of dependencies... Jul 31 14:24:16 EnceloAMD: yes, but OZ question are for #openzaurus Jul 31 14:25:03 ok Jul 31 14:30:53 ibot's gone Jul 31 14:36:29 koen: I think I'm the only man who want to use scons on OZ :) Jul 31 14:36:50 what's a scon? sounds like pastry Jul 31 14:37:06 subdue: some fancy autotools replacements written in python Jul 31 14:37:38 oh. do packages even recognize it? Jul 31 14:37:55 nm. I'm sure he must want it for a reason Jul 31 14:39:08 koen: it's not fancy, it's serious and it's a salvation :) Jul 31 14:40:06 Blender, Yafray and Ardour all use scons... this means it's not a toy, but a serious tool Jul 31 14:40:31 subdue: yes, I need it to port a game Jul 31 15:22:23 03coredump 07org.oe.dev * r136b69b8... 10/packages/prboom/ (6 files in 5 dirs): prboom: add prboom-devconfig for device specific default configuration. Changed prboom.sh to honor default config Jul 31 16:12:44 03daka 07org.oe.nslu2-linux * rb1872f89... 10/conf/distro/openslug-native.conf: Initial distro for compiling nativly for openslug on the nslu2. Needs some work, but basically works. Jul 31 16:14:13 ah, awesome Jul 31 16:42:52 03nail 07org.oe.nslu2-linux * r0f708a1a... 10/conf/distro/openslug-packages.conf: Added groff and udev Jul 31 17:08:13 hm, what's the proper thing todo when i change a distribution/machine specific config file? - changing the PR of the whole package sounds bad to me, as it forces everybody to rebuild that package, for most of them without any changes. Jul 31 17:16:45 Where can I download boost? Jul 31 17:17:44 www.boost.org is down. Jul 31 17:19:29 Bug 183 fixed Jul 31 17:19:53 Don't know how to commit the patch will ask somebody tomorrow Jul 31 17:20:17 Never mind me, I found boost on source forge. Jul 31 17:22:42 tmbinc: last time I asked the answer was to bump the PR anyway Jul 31 17:24:24 ok Jul 31 17:26:40 gerwinin: ntp stuff? Jul 31 17:26:49 gerwinin: try setting another ntp version more recent Jul 31 17:30:57 couldn't PRs and the whole remote-version-comparing (i.e. ipkg's) be avoided by just using build dates for package versions? (last time, but i admit that's been a while, ipkg didn't supported multiple versions in one feed at all - so that wouldn't make any difference then) and local (i.e. package) versions by looking at depending file's hashes (and comparing them with "last known ones", to detect differences, maybe cached with some inode lookup Jul 31 17:32:34 local versions could of course be still used to determine which version should be built - but the version used would be always the most recently build one, which i would like for my workflow Jul 31 17:33:41 Zecke: the problem was with gcc Jul 31 17:34:01 Zecke: I applied a patch to the most recent version and than it worked Jul 31 17:41:41 How do I install boost, Boost does not yet have a standardized installation process. - http://www-eleves-isia.cma.fr/documentation/BoostDoc/boost_1_29_0/more/download.html Jul 31 17:41:50 So you use a library without an installation method. Jul 31 17:42:05 Where do I need to move the boost directory so that ./configure for monotone will find it? Jul 31 17:46:20 And just so you know, the string "boost" doesn't not occure once in the GettingStarted OR RequiredSoftware pages. Jul 31 17:50:53 don't you use a distribution which maintains this stuff for you? monotone is a quite standard (well, ..) application, not at all OE specific Jul 31 17:51:55 If someone would make a distribution with 1/2 the features I need I would use it, but I appear to be the ONLY person in the world who wants a distributed diskless cluster. Jul 31 17:54:56 and you can't build this on top of a usual distribution? Jul 31 17:55:34 (not that i want to start a "my distro is better than yours", but debian is certainly better than nothing at all ;) Jul 31 17:56:04 Im experimenting with using gentoo on top of my substructure, Isn't gentoo a binnary distro? Jul 31 17:56:14 I can't use binnarys compiled for other distros. Jul 31 17:56:47 But I still have some issues to work out with gentoo. Jul 31 17:56:52 ah, ok Jul 31 17:57:08 Which is why I spent a few hours last night building my own cross compiling tool chain. Jul 31 17:57:16 what exactly are you doing, btw? (just out of interest) Jul 31 17:58:47 I am ensureing that the information I would with is very difficult to gain unauthorized access to. Jul 31 17:59:55 I don't use a standard mass storage system like most Linux systems, so it is not possible for me to just install one. Jul 31 18:00:37 / is on a distributed and encrypted datastore. (/dev/loop129 on this machine). Jul 31 18:00:50 I got boost installed now. Jul 31 18:06:27 03jbowler 07org.oe.nslu2-linux * r8f3f305e... 10/conf/distro/ucslugc.conf: Fix the UcSlugC feeds (they were not being set). Jul 31 19:32:27 nail: no need to re-add mpc and ncmpc, they're already in the musicpd folder. I got them added a little while ago. Jul 31 19:32:59 JustinP: They're not there now, are they? Jul 31 19:33:21 I thought I backed out that change Jul 31 19:33:36 when I discovered that they were already in the musicpd folder Jul 31 19:34:01 How long will it take for monotone to pull oe? Jul 31 19:34:42 it appears to have frozen. Jul 31 20:27:24 NAiL: not sure yet, I just saw the commit Jul 31 20:27:54 It was never meant to be committed actually Jul 31 20:28:16 arrummzen2: not likely. If you're doing an initial pull without a snapshot it will be hours. Pulls from a week ago could also take 1hour+ Jul 31 20:28:29 NAiL: I see. Just making sure Jul 31 21:01:53 JustinP, Ok. Jul 31 21:03:51 Does OE build the kernel or just the applications? Jul 31 21:04:02 Do Zauruses use initrds? Jul 31 21:05:29 It tells me Revs Written = 318. How many are their total (before its done)? Jul 31 22:17:45 arrummzen2: it builds *everything*. Zaurii have to be flashed with kernels and the boot image. "Revs in" is the final # but it may stop before it gets there. **** ENDING LOGGING AT Sun Jul 31 23:59:56 2005