**** BEGIN LOGGING AT Sat Feb 04 10:59:57 2006 Feb 04 11:37:48 03koen 07org.oe.oz354fam083 * r30f9edfd... 10/packages/xkbd/ (xkbd-0.8.15/fix-equalsign.patch xkbd_0.8.15.bb): xkbd: add patch to fix equal sign Feb 04 11:37:52 03koen 07org.oe.dev * r85c82ab6... 10/packages/xkbd/ (xkbd-0.8.15/fix-equalsign.patch xkbd_0.8.15.bb): xkbd: fix equal sign, closes hh.org bug #1512 Feb 04 12:33:10 morning all Feb 04 12:33:27 hi RP Feb 04 12:33:28 hey RP Feb 04 12:39:12 03koen 07org.oe.oz354fam083 * r5f304035... 10/packages/fontconfig/fontconfig_2.2.95.bb: Feb 04 12:39:12 fontconfig: add RREPLACES on libfontconfig1 so ipkg won't kill your install on upgrading Feb 04 12:39:12 * see hh.org bug #1515 Feb 04 12:54:41 * Philippe is back (gone 14:03:20) Feb 04 13:08:19 morning folks Feb 04 13:08:29 hi ade Feb 04 13:10:00 hi pb_ how is cambs today, i've been in birmingham the past few days, my wife tells me its rather cold back home Feb 04 13:10:36 it is indeed quite cold around here. I think it snowed a little bit yesterday and the day before. Feb 04 13:12:18 it can't have snowed too much, debbie's school didn't close, sadly Feb 04 13:12:22 hi all. Feb 04 13:12:51 i'm trying to use gpe-aerial, but it keeps on trying to use eth1 instead of wlan0... Feb 04 13:13:04 anyone knows where this can come from ? Feb 04 13:13:18 same error when i try to use prismstumbler Feb 04 13:13:33 yet my car uses hostap Feb 04 13:13:54 your car uses hostap? eleet. Feb 04 13:14:01 lol Feb 04 13:14:27 did i say something wrong ? Feb 04 13:14:52 alan_: i think you missed a 'd' on car Feb 04 13:15:01 hehe ^^ Feb 04 13:15:55 alan_: having never used gpe-aerial i cant really help you, sorry Feb 04 13:16:01 have you tried the folk in #gpe ? Feb 04 13:16:14 ade|desk: ok, thaks anyway Feb 04 13:16:24 yeah, florian would be your resource for that Feb 04 13:16:32 going to try on #gpe, that is a good idea Feb 04 13:49:29 hey guys, how are things Feb 04 13:57:48 hey mickeyl Feb 04 13:58:15 hail mickeyl Feb 04 14:02:25 03mickeyl 07org.oe.dev * rb6e92585... 10/packages/wlan-ng/wlan-ng-modules_0.2.2.bb: Feb 04 14:02:25 wlan-ng-modules: set default preference of version 0.2.2 to -1 for tosa, Feb 04 14:02:25 since 0.2.1-pre is the most recent version that works there Feb 04 14:03:00 hi mickeyl Feb 04 14:03:04 hey RP Feb 04 14:03:11 mickeyl: freaking busy :} Feb 04 14:03:15 terribly Feb 04 14:03:37 no, seriously. I have to work on a presentation of 5 years of work until the 16th february Feb 04 14:03:51 no idea how to structure that in a sane way Feb 04 14:04:01 only 45 minutes and there will be a lot of professors attending Feb 04 14:04:03 scary... Feb 04 14:04:18 oh right, your phd is reaching its culmination? Feb 04 14:04:30 mickeyl: oh yeah... a not the smart idea would be to use one slide Feb 04 14:04:43 yeah, slowly. it's not the public defence yet, but it's the pre-finishing lecture Feb 04 14:04:45 mickeyl: a big cloud and print void into it :} Feb 04 14:04:51 heh Feb 04 14:05:30 I guess you should just have some slides about openzaurus. Feb 04 14:05:39 "I didn't do much research, but I made this really cool operating system." Feb 04 14:05:47 hahahaha Feb 04 14:05:56 that's true Feb 04 14:06:04 :D Feb 04 14:06:20 mickeyl: good luck (seriously). What about sorting thins chronologically? Feb 04 14:06:25 unfortunately i don't think i'll receive a PhD for that :) Feb 04 14:07:17 zecke: yeah, i guess i'll start with motivating my research due to advances in mobile distributed networks - devices, comm.technologies, ubicom. then going over to why middleware should be adaptive in such contexts and then show my adaptive architecture. Feb 04 14:08:09 actually, i'll handle to include OpenEmbedded, OpenZaurus and Opie in the part about "evaluation tools" :) Feb 04 14:09:03 My final project at Uni for a years work was "Why my professor was incorrect to set this as a project as its physically impossible to do" :-/ Feb 04 14:09:32 d'oh Feb 04 14:10:15 Hi all Feb 04 14:10:18 hey dirk Feb 04 14:10:43 Hey mickeyl Feb 04 14:11:21 I found that out about 10 minutes after starting the project with nothing more complicated with a multimeter. Still had to spend a year doing the project anyway proving it was impossible :) Feb 04 14:11:23 hi dirk Feb 04 14:11:48 Hi Richard Feb 04 14:12:00 RP: did he believe you in the end? Feb 04 14:12:11 The project brief did get adapted a bit to give me other things to do... Feb 04 14:14:05 mickeyl: Yes, and he supported the conclusion and said I would be fine if I could prove it :) Feb 04 14:14:17 excellent :) Feb 04 14:14:23 by the way, any of you ever did something with latex-beamer? if not, consider it for your next presentation. it rocks. As an old powerpoint presentation guy, I really love to work with beamer. Feb 04 14:14:42 (It was about resistance mapping of solar cell materials like CdTe) Feb 04 14:14:53 mickeyl: I have started using it Feb 04 14:15:11 mickeyl: do you have a good overview on which themes are available Feb 04 14:15:34 zecke: i printed the chapter about "how to change looks" in the beamer userguide on a color laser printer. Feb 04 14:15:44 all stock themes can be seen there Feb 04 14:16:16 i settled on Frankfurt (sic) Feb 04 14:16:17 heh Feb 04 14:16:55 hehe Feb 04 14:16:57 None of the machines at Uni would have run anything like that. We were forced to use powerpoint :-( Feb 04 14:17:15 and FORTRAN :-/ Feb 04 14:17:24 oh the agony :) Feb 04 14:17:47 I hate fortran with a passion... Feb 04 14:18:28 i chose stochastics instead of numerics in university. i didn't want to learn fortran Feb 04 14:18:40 heh Feb 04 14:18:49 good choice ;-) Feb 04 14:20:04 I spent the lab sessions proving C was more efficient at everything to the lab demonstrators (and asking them really awkward questions) :) Feb 04 14:20:27 :D Feb 04 14:21:29 backus wouldn't agree, but then again, he quit CS in 1991 Feb 04 14:22:46 * mickeyl goes to wash dishes Feb 04 14:25:53 It's hard to imagine that C is actually _more_ efficient than Fortran for many numerical tasks, though it's certainly true that (since C99) it's no longer significantly less efficient. Feb 04 14:26:22 nowadays, there isn't a great deal to choose between them: you can write decent compilers and libraries for both languages. Feb 04 14:40:10 I suspect the fortran compiler was elderly and the C one a lot more modern. That didn't stop me winding the demonstrators up though ;-) Feb 04 16:13:33 03koen 07org.oe.oz354fam083 * rde179bce... 10/packages/dbus/ (dbus/no-static.patch dbus_0.23.4.bb): dbus: add patch to stop dbus-daemon-1 being built staticly, closes hh.org bug #1494 Feb 04 16:13:38 03koen 07org.oe.dev * r8b63d518... 10/packages/dbus/dbus/no-static.patch: dbus: add patch to stop dbus-daemon-1 being built staticly, closes hh.org bug #1494 Feb 04 16:15:13 cedric: ping? Feb 04 16:36:35 is there a way to force bitbake to remake the zImage despite stamps or anything else (so I can force it to create a 64-0 and a 32-32)? Feb 04 16:37:28 bitbake -b -c compile -f Feb 04 16:37:33 or bitbake -i Feb 04 16:37:35 parse Feb 04 16:37:42 rebuild virtual/kernel Feb 04 16:37:51 good evening Feb 04 16:37:52 koen: sweet, thanks Feb 04 16:37:58 Hi Koen Feb 04 16:38:27 Could i ask you to give a lecture in Utrecht on the slug day Feb 04 16:38:42 http://www.nllgg.nl/bijeenkomsten_2006_slugdag Feb 04 16:40:01 marceln: sure, I'll be attending anyway Feb 04 16:40:18 Thanks. Feb 04 16:40:51 I hoped you could give a talk about bitbake and monotome? Feb 04 16:41:32 I could handle that Feb 04 16:42:02 That just needs a translation of my sheets for FOSDEM :) Feb 04 16:42:25 Ok. I will place your name there :) Feb 04 16:43:50 koen: it rebuilt the kernel, but didn't package it into a zImage... what do I type next? Feb 04 16:44:20 it didn't package the zImage? Feb 04 16:44:22 weird Feb 04 16:44:34 you want to check with mickeyl or hrw|gone Feb 04 16:45:06 koen: thanks Feb 04 16:45:12 mickeyl: hrw|gone ping Feb 04 16:51:25 do13: Do you have an updated defconfig for tosa? Feb 04 17:00:40 03rpurdie 07org.oe.dev * r7740b6ed... 10/ (4 files in 3 dirs): linux-oz-2.6 (2.6.15): Update led patches. Refresh c7x0 and poodle defconfigs. Update sharpsl_pm patch from Dirk. Feb 04 17:00:44 03rpurdie 07org.oe.dev * r606273dc... 10/packages/linux/ (13 files in 2 dirs): linux-openzaurus-2.6.14-git3: Remove as outdated now Feb 04 17:10:41 okay, how do I make bitbake build all of the collie images. I used to not define COLLIE_RAMDISK_SIZE and COLLIE_MEMORY_SIZE and it would build all the zImages.... now it does not. Feb 04 17:23:59 RP: not yet. I've integrating new TC6393 drivers. OHCI :) Feb 04 17:25:40 do13: Ok, that sounds good :) Feb 04 17:26:02 do13: I'm just trying to tidy up the OE kernel a bit. Some of the defconfigs are suffering bitrot :-/ Feb 04 17:27:04 do13: Did you have a look at my altered version of your corgi_bl_generic patch? Feb 04 17:30:38 RP: corgi_bl_generic-r2.patch? Feb 04 17:30:47 yes Feb 04 17:31:03 RP: So it works in my recent kernel. Feb 04 17:33:37 ok. I might push that and the backlight changes to the fb developers Feb 04 17:34:00 Where I need to go through several maintainers, I need to start moving now for 2.6.17 :) Feb 04 17:34:21 I'll also submit the keyboard driver changes Feb 04 17:35:07 RP: I'll look at this and integrate this into the tosa driver Feb 04 17:36:15 It was mainly changing the scan rate from 100ms to 50ms as it solves the nasty key reptition issues everyone complained about Feb 04 17:36:29 03koen 07org.oe.oz354fam083 * r52aa4399... 10/packages/led/led_cvs.bb: led: add led from h3600-utils, close hh.org bug #1310 Feb 04 17:36:34 03koen 07org.oe.dev * rb0704e33... 10/packages/led/ (files/nokernelheader.patch led_cvs.bb): led: add led from h3600-utils (adapted from .oz354fam083) Feb 04 17:36:38 03pH5 07org.oe.dev * r389a6b49... 10/packages/linux/handhelds-pxa-2.6/ipaq-pxa270/defconfig: handhelds-pxa-2.6_cvs: update hx4700 defconfig for 2.6.15-hh0 Feb 04 17:36:58 RP: You can push this patch: http://www.do13.de/oz/temp/tosa_fstab.patch Feb 04 17:44:01 do13: pushed Feb 04 18:06:57 03rpurdie 07org.oe.dev * r952b15f5... 10/packages/base-files/ (base-files/tosa/fstab base-files_3.0.14.bb): base-files: Tweak tosa 2.6 mmc fstab options (from Dirk Opfer) Feb 04 18:15:21 hi Feb 04 18:21:13 hi CoreDump|home Feb 04 18:26:00 RP: :) Feb 04 18:27:53 RP: what do you think about the suspend/resume issues, are they due to me building images on amd64 host? Feb 04 18:28:04 or maybe vairent / broken zaurus hardware Feb 04 18:44:38 perhaps the suspend somehow relies on rootfs being on hda1 Feb 04 18:55:16 shadows: It shouldn't depends on hda1 Feb 04 18:55:30 shadows: At some point I'll share a precompiled kernel and you can test with that Feb 04 18:59:26 Ciao all Feb 04 19:00:34 hey Pigi Feb 04 19:01:12 hey koen! . It seems you are having some nice time testing ipkg these days ;) heh Feb 04 19:03:30 Pigi: status after using -force-overwrite on the failing packages: http://dominion.kabel.utwente.nl/koen/pda/files/status.gz Feb 04 19:05:13 that's even more strange ! your status file says xkbd is installed ..... Feb 04 19:05:25 in version 0.8.15-r1 Feb 04 19:06:58 RP: that would be wonderful, thank you Feb 04 19:07:12 yes, I did a ipkg install xkbd fontconfig -force-reinstall -force-overwrite to fix the install Feb 04 19:10:43 shadows: do you need a recent C3000 2.6 kernel? mine's up to date except for RP's last push I think Feb 04 19:11:16 what I don't understand is why you needed the force-reinstall and overwrite. This usually means that info files, and status are there. So xkbd seems that it was installed. Feb 04 19:11:24 even more strange...... Feb 04 19:11:43 Pigi: does ipkg update the status file right after uninstalling? Feb 04 19:12:00 or does it does somehting like 'failed on something, not writing status file'? Feb 04 19:12:14 koen, no. It does it at the end of the whole upgrade cycle. Feb 04 19:13:28 koen, sometimes it could even happens, for the "failed on something, not writing status file".... Feb 04 19:14:56 it did write something, since it refused to ipkg upgrade again Feb 04 19:15:20 koen, do you think it could be reproducible in some way ? Feb 04 19:15:49 I think so Feb 04 19:16:23 that's a good news..... Btw it seems that it could/should be really hard that it don't write out the status file.... Feb 04 19:16:41 make 3 new ipkgs (a, b and c), make c depend on b and make sure ipkg errors out with b Feb 04 19:16:55 (xkbd depends on fontconfig) Feb 04 19:17:10 JustinP: my hardware (C3000) is failing to resume from an extended time of being suspended Feb 04 19:17:26 JustinP: this means i am thinking my hardware may be faulty or different Feb 04 19:17:30 Pigi: I added a Replaces: libfontconfig1 in fontconfig r4 Feb 04 19:17:50 i would like to verify that my hardware can resume for an extended time period and then resume without troube Feb 04 19:17:51 which does broke the install ? Feb 04 19:18:23 fontconfig broke, since it overlapped with libfontconfig1 Feb 04 19:18:57 ah ok. Feb 04 19:19:23 it broke in 'configuring fontconfig' Feb 04 19:19:59 shadows: what kernel? Feb 04 19:20:11 shadows: I have had similar issues with 2.4...... Feb 04 19:21:28 Pigi: and a few ipkg versions back an error in 1 package caused nothing to be done Feb 04 19:21:55 Pigi: ipkg install x y z (with z being faulty) would not install x and y Feb 04 19:22:01 koen: that sounds rather dubious. libfontconfig1 is the correct name of the package. Feb 04 19:22:02 dunno if that's been fixed Feb 04 19:22:10 if you are encountering a fontconfig.ipk, that is probably a bug Feb 04 19:22:23 but, furthermore, file overlaps should cause ipkg to fail during unpacking, not during configuring. Feb 04 19:22:53 the latest (and greatest?) fontconfig produces fontconfig*.ipk Feb 04 19:22:58 koen, I'm not sure that your last sentence is correct, as the upgrade has ever worked this way Feb 04 19:23:23 Pigi: that last problem was in ipkg install Feb 04 19:23:41 koen: then the latest fontconfig is, regrettably, broken Feb 04 19:24:27 in particular, that will break any existing binaries with dependencies on libfontconfig1 (i.e. almost all of them) Feb 04 19:24:33 I'm even unsure that the install was working this way, but I could be wrong. And btw we were talink on upgrade..... Feb 04 19:27:33 pb_: is stuff in /etc supposed to be part of (lib)fontconfig or fontconfig-utils? Feb 04 19:27:58 specically: Feb 04 19:27:58 ./etc/fonts/fonts.conf Feb 04 19:27:59 ./etc/fonts/fonts.dtd Feb 04 19:27:59 ./etc/fonts/local.conf Feb 04 19:28:20 local.conf shouldn't exist at all anymore; the other two files should be in libfontconfig1 Feb 04 19:29:06 this should be the local.conf fontconfig ships by default Feb 04 19:29:41 oh, right, the one that just contains comments? Feb 04 19:29:51 in that case, that should be in libfontconfig1 also Feb 04 19:29:55 that one Feb 04 19:31:49 Packaged contents of fontconfig into /data/build/koen/OE/build/tmp/familiar/deploy/ipk/libfontconfig1_2.2.95-r4_arm.ipk Feb 04 19:32:00 very good Feb 04 19:32:10 the AUTO_LIBNAME_PKGS = "fontconfig-utils" in the .bb is causing troubles Feb 04 19:32:46 that sounds pretty bogus. who added that? Feb 04 19:33:02 fontconfig-utils is the only package you _don't_ want renamed. Feb 04 19:34:58 as far as the branch goes, you probably want to just remove the offending line. Feb 04 19:35:31 for -dev, I guess you should identify the culprit and make them fix it properly Feb 04 19:36:23 JustinP: linux-openzaurus-2.6.15r2 Feb 04 19:36:47 it will resume if i suspend for less than 2 minutes Feb 04 19:36:49 http://pastebin.com/538930 Feb 04 19:36:57 RP: could you fix fontconfig in .dev? Feb 04 19:37:01 JustinP: otherwise, it does not resume. Feb 04 19:40:30 shadows: hmmm..I haven't used 2.6 much myself. I hope to test more soon.... Feb 04 19:44:11 shadows: but I've had my Z suspended with 2.6 for most of a day now...let's see.... Feb 04 19:46:12 JustinP: everyone except me has this working Feb 04 19:47:47 koen: I'll look at it. I suspect there are some subtle bugs in the AUTO_LIBNAME_PKGS as my conclusion last time was the logic was inverted Feb 04 19:48:10 i.e. you needed specify the names of the packages you didn't want auto-named Feb 04 19:53:23 shadows: well, I turned it on, then it turned itself off again (?) then turning it on again worked Feb 04 19:53:30 shadows: want to try the image I'm running? Feb 04 19:53:35 please Feb 04 19:53:46 rootfs tarball and kernel please, for c3000? Feb 04 19:53:58 the cavet is i'm running hda6 as my rootfs Feb 04 19:54:05 shadows: it's e-image-core, BTW, my new gpe/opie images aren't done yet Feb 04 19:54:08 hehe Feb 04 19:54:11 :) Feb 04 19:54:22 well...I'd have to recompile the kernel for that, right? Feb 04 19:54:28 oh Feb 04 19:54:34 you changed back to default partitioning? Feb 04 19:54:44 no, but I'm running from hda1 for 2.6 right now Feb 04 19:54:50 okay Feb 04 19:54:53 I just backed it up and overwrote it Feb 04 19:55:19 that's not a trouble to change the rootfs args for the kernel is it? Feb 04 19:55:24 koen, in your previous example obviosly package c should depends on b (new-revision) isn't it ? Feb 04 19:55:29 i mean, it should be a fairly straighforward and documentable process Feb 04 19:55:36 without recompiling Feb 04 19:55:41 can I? Feb 04 19:55:47 i would think so Feb 04 19:55:54 what format is the zImage Feb 04 19:56:02 ::shrug:: Feb 04 19:56:25 03koen 07org.oe.oz354fam083 * r40829548... 10/packages/fontconfig/fontconfig_2.2.95.bb: fontconfig: remove troublesome AUTO_LIBNAME_PKGS = "fontconfig-utils" to avoid issues with the debian renamer Feb 04 19:56:27 let me worry about that detail, i suppose Feb 04 19:56:31 Pigi: yes, a,b and c are all new revisions of installed packages, with one package depending on a 'bad' package Feb 04 19:57:00 ok Feb 04 19:58:30 shadows: np, recompile is easy Feb 04 20:00:50 although not too fast....poor little celeron.... Feb 04 20:01:08 patching Feb 04 20:01:58 heh. my athlon64 3000+ xp churns through the kernel recompile in a reasonable amount of time. it still takes a long long time to do a whole system build Feb 04 20:02:11 i'd like to eliminate all bugs pertaining to amd64 build hosts Feb 04 20:02:21 hello ! Feb 04 20:02:27 shadows just watch out for the sqlite build problem on amd64, thats the only build problem i've run into Feb 04 20:02:35 Pigi: I have reproduced it Feb 04 20:02:50 simcop2387: documented in the bug tracker, finally :) Feb 04 20:03:11 good, as I have some trouble in upgrade, but can't reproduce the same behaviour you say. Feb 04 20:03:16 Does anyone now why libgcc_s.so.1 is not added when creating bootstrap-image ? Feb 04 20:03:35 as a result init fails... :( Feb 04 20:03:53 Ifaistos: what distro, machine, kernel, etc? Feb 04 20:03:57 hmm i don't think i had that problem Feb 04 20:04:22 epia / familiar unstable Feb 04 20:05:51 local.conf -> http://pastebin.com/538969 Feb 04 20:06:29 Pigi: http://handhelds.org/scap/port.756.png Feb 04 20:07:16 Pigi: the link to the status file is in the bugreport Feb 04 20:08:44 well, xkbd is installed ( at least for status file ) Feb 04 20:09:19 libfontconfig is not-installed ( as expected by the screenshot ) Feb 04 20:09:28 the last case only dealt with fontconfig vs libfontconfig1 Feb 04 20:09:45 Pigi: if it's not installed, then why does 'ipkg upgrade' says 'nothing to do'? Feb 04 20:10:18 "ipkg upgrade" only upgrades packages that are installed. Feb 04 20:10:34 right Feb 04 20:11:02 I think upgrade fails because of a wrong check in status .... Feb 04 20:11:12 it *was* installed before the 'pkg upgrade' Feb 04 20:11:50 sorry, koen. Who ( or even what ) was installed before ? fontconfig or libfontconfig ? Feb 04 20:11:59 both Feb 04 20:12:22 libfontconfig has a newer version in the feed Feb 04 20:12:56 well, fontconfig is still installed ( according to status ) Feb 04 20:13:23 yes, since that one wasn't an upgrade candidate Feb 04 20:13:25 in his 2.2.95-r4 version. Feb 04 20:14:11 and libfontconfig ( which fail, because of "not-installed" ) state has a 2.2.95-r5 version. Feb 04 20:14:26 yes, and an r3 installed before the upgrade Feb 04 20:14:41 2.2.95-r3 to be exact Feb 04 20:15:21 so, r3 has been removed, and r5 failed on install. Feb 04 20:15:28 yes Feb 04 20:15:54 I know why ( maybe )... Feb 04 20:16:02 Is that 0.99.157, isn't it ? Feb 04 20:16:05 this is with .157 Feb 04 20:26:51 pb_, koen: Do we know offhand who wrote debian.bbclass? Feb 04 20:26:54 koen, this trouble could be mitigated with 158, but I suspect another bug too.... Feb 04 20:32:00 whats a good program for gaim to use to play waves on OSS? Feb 04 20:32:05 or esd for that matter Feb 04 20:32:12 perhaps sox Feb 04 20:32:18 someone know why KEY_CONTACTS became KEY_SEND on kernel 2.6.15 ? Feb 04 20:32:19 sox includes the 'play' command Feb 04 20:32:25 RP: either me or Chris, I think Feb 04 20:32:34 gremlin[it]: ask RP Feb 04 20:32:41 or maybe mickey, but I don't think so Feb 04 20:32:47 tried sox it doesn't want to play the waves from gaim for some reason, complains about size being 254 in the header Feb 04 20:33:00 koen unfortunally atm I have too much headache to look at it. I'll give a look at this tomorrow Feb 04 20:34:14 koen, if you have some spare time test 158 anyway as the trouble I'm having now is with Depends: part of this bug, not Replace Feb 04 20:34:43 RP ... ;) Feb 04 20:34:44 later all Feb 04 20:36:34 gremlin[it]: I not noticed that change - you'll probably have to look it up in the changelogs Feb 04 20:36:51 i'll go :) Feb 04 20:38:43 anyone know if alsa works on a collie? Feb 04 20:39:12 pb_: The logic for handling AUTO_LIBNAME_PKG in debian.bbclass in puzzling me. I've just tried setting it to "fontconfig", "libfontconfig" and "fontconfig-utils". The latter two cause no renaming. The former case causes all packages to get renamed Feb 04 20:40:11 ah yeah, it would do Feb 04 20:40:37 AUTO_LIBNAME_PKGS won't help with the particular issue that you have. You'll have to set PKG_foo directly. Feb 04 20:41:16 That's a totally broken design then Feb 04 20:41:44 well, no, it's operating as intended. It just happens not to be intended to solve the problem that you're facing. Feb 04 20:42:13 I mean that having to set PKG_foo directly is a totally broken design Feb 04 20:42:29 Why? Feb 04 20:43:01 or, perhaps more to the point, what exactly is it that you feel is broken? Feb 04 20:43:44 OE shouldn't be favouring any one naming scheme Feb 04 20:43:58 Nor does it. Feb 04 20:44:26 Setting PKG_foo = would stop any other renaming scheme from renaming that package Feb 04 20:44:44 Indeed it would, but your point seemed to be that you wanted that package to not be renamed. Feb 04 20:45:33 Irrespective of that, if you want to implement output-side filtering for debian.bbclass then you are very welcome to do so. Feb 04 20:45:46 RP -> http://handhelds.org/cgi-bin/cvsweb.cgi/linux/kernel26/include/linux/input.h.diff?r1=1.19&r2=1.20&f=h Feb 04 20:45:54 but there is no comment ;( Feb 04 20:45:56 In this case debian.bbclass gets it wrong. This is a problem with debian.bbclass though, not the packages metadata Feb 04 20:47:51 gremlin[it]: Was this part of the change to 2.6.15 for hh.org? Its possible some code got lost in the changeover... Feb 04 20:48:03 That might be true, though "gets it wrong" is somewhat subjective. Debian.bbclass is operating as it was intended to, but it's entirely possible that this is not the right thing for every package. Feb 04 20:49:06 I don't understand your remarks about "a totally broken design" and "favouring any one naming scheme". Feb 04 20:50:17 RP really don't know ... i never touched include/linux/input.h ... i just notice that KEY_CONTACTS disappear ... and with the same value (231) now there is KEY_SEND ... 2.6.13-hh1 and i think this problem shlud happen on others ipaq not only h3600 i'm trying ... Feb 04 20:50:40 pb_: Say I want my own naming scheme where every package must be appended with the letter 'a'. Can OE do that if we set PKG_foo = "xxx" in random bb files - it can't Feb 04 20:50:43 2.6.13-hh1 worked ... and i ... Feb 04 20:51:16 RP: indeed not. But, presumably, for the same reason that you don't want libfontconfig-utils renamed here, you also wouldn't want it renamed there. Feb 04 20:52:14 pb_: The ultimate aim is to enable that to be allowed Feb 04 20:52:44 or that's the only sane way to proceed anyway, otherwise we'll never have a meta-sdk again Feb 04 20:54:23 gremlin[it]: I think it will have just accidently got lost in the upgrade Feb 04 20:55:17 RP which ipaq u kow work (at least compile) 2.6.15 ? so i try to figure out if key have really changed name ... Feb 04 20:56:15 gremlin[it]: At a guess hx4700 but I don't use hh.org kernels Feb 04 20:56:31 RP: so, if you _do_ want it renamed, what's the issue? Feb 04 20:56:58 tried to upgrade ipkg from .155 to .158 in order to get a fix for a bug that was irritating me. however if I create ipkg_0.99.158.bb the command bb ipkg still builds 155 Feb 04 20:57:06 what am I doing wrong Feb 04 20:57:28 ahh ok RP ... Feb 04 20:57:39 eFfeM: openslug-packages is getting in the way? Feb 04 20:57:50 eFfeM: you could try bitbake -b ipkg_0.99.158.bb Feb 04 20:58:04 pb_: My point is OE shouldn't care. A short while ago, we had the problem that packages RDEPEND on libfontconfig-utils Feb 04 20:58:27 koen bitbake -b ipkg_0.99.158.bb works! Feb 04 20:59:11 actually I did bb -cclean ipkg then made the 158 bb file, and bb ipkg still builds the 155 version! Feb 04 20:59:21 so i'm very puzzled Feb 04 21:00:17 pb_: We also did discuss that debian.bbclass' renaming in this case was incorrect Feb 04 21:00:19 RP: indeed, and I seem to remember that we agreed at the time that there were two ways to fix that problem: either disable renaming of that subpackage (via PKG_xx) or add some extra logic to put RDEPENDS through the same PKG mapping that the subpackages themselves experience. Feb 04 21:02:15 Adding a PKG_xxx is not a fix, its a hack Feb 04 21:03:46 I'll add the PKG_ line but only as a temp fix as I strongly disagree with it :-/ Feb 04 21:05:10 Okay, very good. Please remember to add RPROVIDES for the old name as well, so as not to break old binaries. Feb 04 21:05:42 If PKG_xx offends your sensibilities that badly, I guess your best option would be to implement one of the other two plans that we've mentioned. Feb 04 21:06:20 pb_: I was only aware of one? Feb 04 21:06:46 Irrespective of that, if you want to implement output-side filtering for debian.bbclass then you are very welcome to do so. Feb 04 21:06:48 that was the first one Feb 04 21:07:01 RP: indeed, and I seem to remember that we agreed at the time that there were two ways to fix that problem: either disable renaming of that subpackage (via PKG_xx) or add some extra logic to put RDEPENDS through the same PKG mapping that the subpackages themselves experience. Feb 04 21:07:13 that (the second half of the sentence) was the second one Feb 04 21:08:33 pb_: ok, we really need both of them implementing. One to workarond bugs in debian.bbclass and the other to let things like meta-sdk work again Feb 04 21:09:42 okay, very good Feb 04 21:22:40 03pH5 07org.oe.dev * r4403312a... 10/packages/update-modules/ (update-modules-1.0/update-modules update-modules_1.0.bb): Feb 04 21:22:40 update-modules: fix module loading during first boot Feb 04 21:22:40 - Phil's patch from http://www.handhelds.org/hypermail/oe/current/6167.html Feb 04 21:22:40 - this also works when ipkg intercepts update-modules in postinst. Feb 04 21:23:11 pb_: should that also be in the release branch? Feb 04 21:23:18 koen: yes, please Feb 04 21:23:21 (that = update-modules patch) Feb 04 21:26:45 koen: I've just pushed the fontconfig changes Feb 04 21:29:34 03koen 07org.oe.oz354fam083 * rfcce7a80... 10/packages/update-modules/ (update-modules-1.0/update-modules update-modules_1.0.bb): Feb 04 21:29:34 update-modules: fix module loading during first boot Feb 04 21:29:34 - Phil's patch from http://www.handhelds.org/hypermail/oe/current/6167.html Feb 04 21:29:34 - this also works when ipkg intercepts update-modules in postinst. Feb 04 21:29:34 - taken from .dev Feb 04 21:29:38 03rpurdie 07org.oe.dev * r4b06abdf... 10/packages/fontconfig/fontconfig_2.2.95.bb: fontconfig: Correct the hack to avoid problems with debian.bbclass Feb 04 21:42:23 RP: JustinP and myself have conducted a test, and we are both C3000 owners. the same image rootfs and zimage on each of our hardware devices, on JustinP's device it suspends/resumes fine, on mine it suspends and does not resume after a while Feb 04 21:42:29 RP: what is the next step? Feb 04 21:47:04 shadows: Something on your hardware must be slightly different to JustinP's. The next step would be to get a serial lead and what the suspend/resume logs and/or start doing debugging with the LED Feb 04 21:47:28 RP: know of any serial leads that would work? i will buy one Feb 04 21:47:52 shadows: Get a Sharp one - I know the serialIO ones have issues with the C3000 Feb 04 21:48:25 shadows: When developing the suspend/resume code, I had it turning the green LED on/off deep in the suspend code to tell me where it was getting to Feb 04 21:48:49 okay. i am looking for a place in the usa that sells the sharp serial lead Feb 04 21:49:43 * JustinP got a SerialIO one....:-( Feb 04 21:49:51 and found that it has issues Feb 04 21:50:08 JustinP: same here... Feb 04 21:52:45 shadows: I think the lead is the CE-170TS Feb 04 21:53:17 it is for SL-5500, says the website. that is also workable for C3000? Feb 04 21:54:23 shadows: yes, the IO port is standard between models Feb 04 21:54:28 great Feb 04 21:54:35 now to find a retailer who carries it in-stock Feb 04 21:54:49 "parts store" has one for $118usd Feb 04 21:54:50 ouch Feb 04 21:57:02 I went for the serialio one originally as it was a lot cheaper... Feb 04 21:57:19 The Sharp ones seem to be in limited supply now Feb 04 21:57:44 yeah....sux Feb 04 21:57:57 FWIW, you can probably do the debugging you need with a serialio one - it can send data to the PC, just not send it from the PC to the handheld Feb 04 21:58:05 oh yeah Feb 04 21:58:06 heh Feb 04 21:58:12 shadows: where are you? Feb 04 21:58:19 chicago vicinity Feb 04 21:58:33 ah....I'm in the SF bay area Feb 04 21:58:36 ebay has a seller offering the cables for $50~ Feb 04 21:58:40 which is more reasonable Feb 04 21:59:51 kind of worried about getting ripped off for a cheap clone that doesn't work Feb 04 22:00:08 well the serialio one is enought for output, so you should be fine Feb 04 22:00:14 hm Feb 04 22:00:16 with that one Feb 04 22:00:29 if i'm going to spend $20 anyways, i might as well spend a little more and get the fully capable cable Feb 04 22:01:44 03eFfeM 07org.oe.dev * r8fb0e135... 10/packages/pvrusb2-mci/pvrusb2-mci_20060121.bb: pvrusb2-mci: added release 20060121 Feb 04 22:02:37 tnx koen, it works again Feb 04 22:14:21 * shadows bids on a serial cable Feb 04 22:58:46 Hmm. Attempting to drink what appears to be chinese beer upon inspection of the label was a mistake :-/ Feb 04 22:58:59 heh Feb 04 22:59:49 RP, avoid Finnish beer too... Feb 04 23:01:15 This next bottle appears to be Brazilian... Feb 04 23:03:10 it appears to at least be drinkable :) Feb 04 23:08:52 03philippe 07org.oe.dev * r9b98e639... 10/packages/matchbox-panel/ (2 files in 2 dirs): Feb 04 23:08:52 packages/matchbox-panel/matchbox-panel_0.9.2.bb Feb 04 23:08:52 * add usage of system-monitot-crash-fix.patch Feb 04 23:08:52 * bump release to r7 Feb 04 23:08:52 packages/matchbox-panel/matchbox-panel-0.9.2/system-monitor-crash-fix.patch Feb 04 23:08:53 * patch that fixes the segfaulting of the system monitor Feb 04 23:08:55 * based on matchbox svn (04/02/2006) Feb 04 23:16:47 * Philippe is away: visual contact - real goth bed Feb 04 23:19:47 Does anyone of you guys have any idea why libgcc_s.so.1 will not be copied from the build to the rootfs ? Feb 04 23:20:04 i can'r figure out why.... Feb 04 23:20:45 it seems to exist in several places -> http://pastebin.com/539265 Feb 04 23:21:00 that's a known OE bug Feb 04 23:21:00 but not in rootfs... Feb 04 23:21:48 you probably want to do somthing like http://ewi546.ewi.utwente.nl/tmp/viewmtn/getdiff.py?id1=bda4787c4d39c0c680a67a9406b6259986ebc180&id2=a2e3bdde97045586a369322e42c6a4e342cc87b0&fname=conf/distro/familiar.conf to work around that Feb 04 23:25:40 03koen 07org.oe.oz354fam083 * rb3ffd18d... 10/packages/matchbox-panel/ (2 files in 2 dirs): Feb 04 23:25:40 matchbox-panel: add patch to stop system monitor from crashing Feb 04 23:25:40 * mallum really, really, really needs to do new releases! Feb 05 00:42:49 RP: ping Feb 05 00:43:27 RP: what is the blocking issue for bitbake? Feb 05 00:44:35 RP: the eli* being None? If that is the case my communication sucked. I have applied this patch Feb 05 01:16:02 zecke: The stamps issue Feb 05 01:16:43 RP: ah okay Feb 05 01:16:55 cya later Feb 05 01:19:18 <[g2]> are there and minimal x86 profiles in OE that are active ? Feb 05 01:35:56 RP: how to disable the suspend function? Feb 05 01:36:05 RP: i want a temporary fix for this issue on my hardware Feb 05 01:36:47 s/fix/workaround/ Feb 05 01:36:47 shadows meant: RP: i want a temporary workaround for this issue on my hardware Feb 05 01:55:50 RP: by stamps issue do you mean that when in bitbake -i it doesn't try to rebuild deps when a dep has previously failed? Feb 05 02:13:20 JustinP: No, I mean the hardcoded tmp/stamps path in bitbake. I'm not very familiar with bitbake -i but that sounds like it could be another issue Feb 05 02:14:11 'night all Feb 05 02:14:41 'nite RP Feb 05 02:28:38 night RP Feb 05 02:40:43 how would i go about disabling the suspend function of 2.6.15-r2? Feb 05 02:40:58 or change it to simply turn on/off the backlight Feb 05 02:41:06 i don't know. it's bothering me how this crashes my Z Feb 05 02:41:27 shadows: you can set it to turn on/off the backlight in the light&power app Feb 05 02:41:37 at least, on my 5500 in opie I can Feb 05 02:44:03 shadows: yeah, it should be configurable not to suspend Feb 05 02:44:12 shadows: perhaps in the keylaunch config? Feb 05 02:44:39 oh, right Feb 05 02:45:00 it suspends from the power button before anything is loaded though Feb 05 02:45:09 so i thought that was a function of the kernel Feb 05 02:45:43 "anything"? Feb 05 02:45:47 you mean on command-line? Feb 05 02:46:34 command-line it's probably part of /etc/keymap-2.6 Feb 05 02:46:36 i mean right after init loads, before it starts going through the majority of the init scripts Feb 05 02:46:51 in X it looks like it's in /etc/keylaunchrc Feb 05 02:47:11 it's probably default in kernel, but you can remap it Feb 05 02:47:19 hm Feb 05 02:47:29 I suggest trying keylaunchrc Feb 05 02:47:41 do i bother to load up the SharpROM and see if it's locking up? Feb 05 02:47:44 you should also be able to remove APM fairly easily in the defconfig for the kernel if you want Feb 05 02:47:52 I say go for it Feb 05 02:48:00 righto Feb 05 02:48:09 should be as easy as a NAND restore, right? Feb 05 02:48:13 ;-) Feb 05 02:48:35 surrre Feb 05 02:48:48 it would have been for me until a little while ago Feb 05 02:50:13 did you overwrite your hda1? Feb 05 02:50:24 didn't you backup NAND before reflashing OZ? Feb 05 02:50:53 i actually did the restore process at http://www.xlfag.com/zaurus/ Feb 05 02:51:06 then i modified the partition table Feb 05 02:51:46 * JustinP shrugs Feb 05 02:52:07 if you left your hda1 and 2 alone then a NAND restore from Sharp should get it back up Feb 05 02:52:20 I mean a NAND restore from your Sharp ROM Feb 05 02:53:11 well the recovery process outlines changing your partitions such that a restore will always work Feb 05 02:53:48 * JustinP shrugs again Feb 05 02:54:01 there were very few people with a C3000 when I got mine Feb 05 02:54:08 (at least that I knew of) Feb 05 02:54:15 same Feb 05 02:54:20 I did it all on my own Feb 05 02:54:27 and then helped you ;-) Feb 05 02:54:59 I just repartitioned with a hacked parted Feb 05 02:55:13 as you may remember from....what, a year ago now Feb 05 03:00:31 i.e. I've never followed anyone else's process ;-) Feb 05 03:07:24 yeah Feb 05 03:07:27 that was a while ago, wow Feb 05 03:27:32 hmm Feb 05 03:28:08 It's been a long while since the nslu2 monotone server has been synced with the OE ones. Any specific reason for this? Feb 05 03:38:11 NAiL: cosmic drift in the interstellar time alignment Feb 05 03:38:25 NAiL: actually i don't know Feb 05 03:38:28 Been reading BOFH excuses? ;) Feb 05 03:47:07 NAiL: ;) Feb 05 03:47:54 funkin' up the cirricular breeze, joe Feb 05 04:47:00 CoreDump|afk: what do you think of an altboot-image? Feb 05 04:47:15 CoreDump|afk: I think I have one which works (not tested, though...) Feb 05 05:30:44 argh!! Feb 05 05:30:54 oh well... Feb 05 06:10:24 RP: SharpROM suspends/resumes fine. now i'm confused. Feb 05 06:21:47 RP: I had to patch drivers/leds/ledtrig-timer.c to handle frequencies <1 - I made it resolve up to three decimal places. The patch is in packages/linux/ixp4xx-kernel/2.6.15/950-leds-timer.patch Feb 05 06:44:38 anyone run kernel 2.6 on collie? Feb 05 06:46:11 and what is "monotone update -rjan2006prebreakage"? I don't do that, and I don't notice any breakage Feb 05 06:46:18 my collie died before 2.6 was started :-| Feb 05 06:46:41 OE was unstable for a time while some changes occurred Feb 05 06:46:56 it's fairly stable now but may be unstable again soon Feb 05 06:47:06 ahh, so that doens't really apply now, but may again soon.... Feb 05 06:47:16 it still sort of applies Feb 05 06:47:21 I've notice some annoying changes myself :/ Feb 05 06:47:26 and it doesn't apply if you use the oz354fam083 branch Feb 05 06:47:30 only dev Feb 05 06:47:31 annoying? Feb 05 06:47:46 you mean....breakage? Feb 05 06:47:54 such as the strip stuff? Feb 05 06:48:01 and module change swhich broke things? Feb 05 06:48:03 yeah, all the stuff in packages/meta that builds opie-image changed Feb 05 06:48:04 :-/ Feb 05 06:48:09 yep Feb 05 06:48:16 part of the breakage Feb 05 06:48:20 but it's sorted out Feb 05 06:48:20 I noticed the strip, but that fixed quick, no worries there. Feb 05 06:48:30 needed to be moved around Feb 05 06:48:54 reorganized Feb 05 06:49:39 I don't mind breakage while making progress. It's the complete changes that have happened to opie-image stuff since I started making my own images.... and then I have to start over Feb 05 06:49:45 that is annoying Feb 05 06:49:54 it happens Feb 05 06:50:00 I know Feb 05 06:50:30 that's when you get into the core and start being able to check your changes in so you don't have to maintain your own ;-) Feb 05 06:51:02 JustinP: no one would want my image, well at least not most. Feb 05 06:51:16 are you using your own bbfile? Feb 05 06:51:28 more maintainable that way Feb 05 06:51:39 JustinP: I rape opie-image as clean as possible, not frills, no bluez, no orinoco, no irda. only 802.11b hacking stuff Feb 05 06:51:55 JustinP: yeah, but when they make huge changes, it messes me up. Feb 05 06:51:58 I suggest making a copy then Feb 05 06:52:19 if you make your own with just the progs you want it shouldn't be a problem for you Feb 05 06:52:28 only if you depend on tasks which change... Feb 05 06:52:34 but it shouldn't happen again any time soon Feb 05 06:52:58 JustinP: well, I depended on something that I didn't notice, and somehow my image got mest Feb 05 06:53:11 JustinP: my sd suddently stopped working Feb 05 09:48:26 RP: don't suppose you're around.... Feb 05 10:47:14 * Philippe is back (gone 11:30:31) Feb 05 10:53:40 JustinP: I am now Feb 05 10:54:35 couldn't find the backlight... Feb 05 10:54:37 I found it Feb 05 10:54:51 my WiFi card is also not working...I get a timeout when I insert it Feb 05 10:56:58 RP: i just reverted to a full restore of my c3000 to the sharp rom, and suspend/restore works fine Feb 05 10:57:11 it is not likely a hardware problem Feb 05 10:58:21 shadows: Ok, that doesn't really surprise me. My code is much more likely to be at fault. I just don't know where and how :-/ Feb 05 10:59:05 here's a useless hint, the power charge light works while the devices is suspended, and eventually stops working when the device refuses to resume Feb 05 10:59:43 shadows: That's actually very useful to know - that means the code is suspending and resuming in some form **** ENDING LOGGING AT Sun Feb 05 10:59:59 2006