**** BEGIN LOGGING AT Sun May 28 02:59:56 2006 May 28 03:21:20 hey guys May 28 03:22:07 is that BREW sdk java based ? May 28 03:23:37 Brew is C++, IIRC.. May 28 03:24:18 NAbyss: apparently it comes only with an .exe installer May 28 03:25:09 CSMan: BREW.. you're referring to the qualcomm j2me competitor, aren't you? May 28 03:25:24 qualcomm's yeah May 28 03:25:47 Yeah.. it seems to be heavily win32/c++ targeted May 28 03:26:08 there is j2me thou May 28 03:26:30 and of course ... OE =D May 28 03:27:07 Yep May 28 03:43:28 anyone care to check out an mtn patch script before I release it to the ML(s) May 28 03:43:29 ? May 28 03:43:32 njs: around? May 28 03:44:46 http://oe.reversefold.com/mtnpatch.txt May 28 03:52:27 what are you patching for? May 28 03:53:57 just a cmdline tool? May 28 03:54:57 03lenehan 07org.oe.dev * r8dcc1ce2... 10/packages/gpsd/ (gpsd.inc gpsd_2.28.bb): May 28 03:54:57 gpsd 2.28: Stop from using X headers from /usr/include by passing in the X May 28 03:54:57 include and library directory to the configure script. Closes bug #1039. May 28 03:55:01 03lenehan 07org.oe.dev * rb09b3e3c... 10/packages/gpsd/ (gpsd_2.6.bb gpsd_2.32.bb): gpsd: Add 2.32 and remove 2.6, as request by mickey in bug #1039. May 28 03:58:33 hey RP, you about? May 28 03:59:43 hey emte, whats your real name? May 28 04:02:00 03lenehan 07org.oe.dev * r3ce8bd8b... 10/site/sh4-linux-uclibc: site/sh4-linux-uclibc: Add required entries to get openssh to build on sh4 with uclibc. May 28 04:05:09 mithro, my real name ... which part? May 28 04:05:43 any i guess May 28 04:06:14 emte: this is a script to parse a patch output by "mtn diff" and run the necessary mtn commands to make it happen May 28 04:06:20 emte: a poor man's cherrypicker May 28 04:09:03 nigel May 28 04:23:45 JustinP: Seem to do what it's meant to. Might be nice if it also output a patch command if there are embedded patches. And a -R mode to reverse all the commands might be handy... May 28 04:28:23 v8jlene: good point May 28 04:57:51 * emte wonders why mithro wanted a name May 28 04:58:03 just to match it with any emails May 28 04:58:16 emails going ? May 28 04:58:45 i cant remember if my nmae is tagged to my email address on the list or not ... May 28 04:58:50 name* May 28 04:59:44 ah yeah it sort of does May 28 05:01:39 should only be a handful or so from me tho, i try to resist the urg to write too many. i have a bad habit of writting things before i think them through throughly May 28 05:06:25 strange .... May 28 05:06:35 i just had a beer, where did it go? May 28 05:07:30 03lenehan 07org.oe.dev * r655b63aa... 10/site/sh4-linux-uclibc: site/sh4-linux-uclibc: Add required entries to get sudo to build for sh4 with uclibc. May 28 05:17:12 03lenehan 07org.oe.dev * rcb0c84fd... 10/site/sh4-linux-uclibc: site/sh4-linux-uclibc: Add required entries to get vim to build for sh4 with uclibc. May 28 06:04:09 * emte wonders if anyone here has contacts with an FOLED manuf May 28 06:07:34 make that TOLED May 28 06:12:48 * v8jlene googles TOLED May 28 06:18:16 TOLED® transparent and top-emitting OLED May 28 06:18:28 http://www.universaldisplay.com/toled.htm May 28 06:18:31 if you want a link May 28 06:30:39 emte: Yeah, found it in google quickly. I was familiar with OLED, but never heard of TOLED, and T threw me off ;) May 28 07:03:13 * v8jlene is away: Doing other things... May 28 07:54:03 good morning May 28 08:23:11 morning ll May 28 08:23:15 all May 28 08:31:51 morning all May 28 08:33:57 hi RP May 28 08:38:34 morning RP - how'd the slug going? got serial on it yet? May 28 08:38:49 s/how'd/how's/ May 28 08:55:00 rwhitby: The timing has worked out badly - very busy at work and other priorities at home - You'll note I've not been around as much. I have a serial interface chip sitting in front of me - just need to remember how to solder :) May 28 09:12:31 hey RP May 28 09:12:51 RP: what makes merging pokey changes back into OE hard? May 28 09:18:08 koen: hey May 28 09:18:15 koen: can we get line wraps on? May 28 09:19:34 ~botmail for kergoth Thanks for the info, but I think it should have gone to mickeyl, koen, or whoever asked for it :) May 28 09:25:33 hehe, i forgot who asked, so just picked one at random :P May 28 09:25:51 kergoth: that is fine ;) May 28 09:26:06 kergoth: I will try to dispatch it to [mickeyl,hrw,koen,schurig] May 28 09:26:50 hi kergoth May 28 09:27:16 morning May 28 09:27:22 hi cyril May 28 09:27:41 hi woglinde May 28 09:30:03 hi May 28 09:30:10 cyrilRomain: progress on the ipkg front? May 28 09:30:15 zecke, thanks May 28 09:30:44 zecke: well 1 error in the valgrind report May 28 09:31:04 zecke: but don't know where it comes from May 28 09:31:04 night kergoth May 28 09:31:25 zecke: I file the bug with the valgrind log May 28 09:32:54 cyrilRomain: use the hh.org bugzilla May 28 09:34:06 zecke: yep May 28 09:34:16 good :) May 28 09:34:21 mithro: time mainly :-/ May 28 09:34:31 hi rp May 28 09:34:33 kergoth`zzz: actually it was wiki.oz.org and hrw probably asked for it :) May 28 09:34:42 hehe May 28 09:34:43 so next paper for tommrow May 28 09:43:05 * cyrilRomain waits for hh.org buzgilla registration mail ... May 28 09:44:14 kergoth`zzz: oe/website/html/dl/RealPlayer10GOLD.rpm May 28 09:44:18 is that your player? May 28 09:50:10 http://forge.novell.com/pipermail/microsuse-users/2006-May/000009.html May 28 09:50:22 if their mail daemon would not be so broken one could reply... May 28 09:53:10 it is true that non-mainstream stuff suffers bitrot in OE May 28 09:53:42 we don't have much heavy uclibc users May 28 09:53:53 s/much/many/ May 28 09:53:59 koen: well he is using crosstool from within buildroot to build a toolchain May 28 09:54:04 koen: well that is bitrot :} May 28 09:54:10 true May 28 09:57:37 hm May 28 09:57:53 maybe I should try uclibc for simpad May 28 09:58:13 woglinde: http://www.albaticket.de/18.0.html May 28 09:58:47 need to ask for money first *cough* May 28 09:58:58 zecke but you are student May 28 09:59:10 woglinde: and disabled, so getting two reductions? May 28 09:59:14 no May 28 09:59:24 woglinde: that sucks May 28 09:59:43 woglinde: fucked up students have even less money than students May 28 10:00:12 zecke so you are consider to watch the final today? May 28 10:00:39 woglinde: http://www.albaticket.de/25.0.html May 28 10:00:47 woglinde: I'm considering it May 28 10:00:51 block N was my normal block May 28 10:01:29 during EuroCup I was there May 28 10:01:37 or W May 28 10:01:42 or they changed it May 28 10:01:58 yes block W May 28 10:02:06 woglinde: in the east... how... May 28 10:02:22 woglinde: we can order tickets online? May 28 10:02:33 hm not for today I think May 28 10:02:36 oh well... what happens if someone carries a knife and stabs people again? May 28 10:02:46 woglinde: you can May 28 10:02:47 thats not handball May 28 10:03:13 but oyu have to phone my wife to get permission May 28 10:03:22 lol May 28 10:03:32 woglinde: I can watch it in TV as well May 28 10:03:53 and I have no telephone May 28 10:04:03 yo are lier May 28 10:07:06 btw are there logs for this channel? May 28 10:07:51 ~logs May 28 10:07:55 apt/ibot/jbot/purl all log to http://ibot.rikers.org// where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/.html.gz, or updated "nightly" May 28 10:07:56 zecke hm May 28 10:08:33 thanks May 28 10:08:40 woglinde: is she about to give birth soon(tm)? May 28 10:08:48 4 weeks May 28 10:10:20 woglinde: BTW: Ritter has a Origami device... May 28 10:10:29 woglinde: it is bulky, has a bad screen signal May 28 10:10:35 woglinde: and it always needs power... May 28 10:11:45 zecke I know May 28 10:11:59 woglinde: okay, I will be leaving to the gym May 28 10:12:07 bye zecke May 28 10:12:13 woglinde: no decision? May 28 10:12:16 (yet) May 28 10:12:22 I hope you come with your car May 28 10:12:29 to where? May 28 10:12:31 and two tickets May 28 10:12:36 and pick me up May 28 10:12:37 hehe May 28 10:12:40 I won't try to get a parking spot at the stadium May 28 10:13:03 woglinde: well, anyway you are not allowed to leave May 28 10:18:42 03koen 07org.oe.packaged-staging * r90279e90... 10/README: README: add the current plan May 28 10:21:05 whats an Origami device? May 28 10:22:49 mithro: those new windows tablet thingies May 28 10:23:06 800x480, windows XP, heavy, expensive, etc May 28 10:23:40 http://www.microsoft.com/windowsxp/umpc/default.mspx <- the propaganda May 28 10:28:30 yeah battery last for 2 hours if you are lucky May 28 10:28:51 ist like the simpad May 28 10:28:55 *g* May 28 10:29:00 its May 28 10:30:14 well as long as it lasts longer than a typuical pr0n dvd May 28 10:30:59 good news everyone May 28 10:31:02 I've tested 2.6 on my poodle, that's the first recent kernel that was able to suspend :)) May 28 10:31:43 2.4 doesn't want to suspend _at all_ :) May 28 10:32:42 ixoth nice May 28 10:33:10 now i should probably test USB networking May 28 10:33:33 should work with gadet driver May 28 10:34:34 yes, modprobe g_ether works ok... now i should configure my host... May 28 10:39:48 morning May 28 10:40:03 hey kergoth`zzz, i think it was hrw who asked to have wiki.openzaurus.org (not oe :)) pointed to that site May 28 10:40:52 hi mickeyl May 28 10:41:15 since http://openzaurus.berlios.de/ contains the OZ wiki now May 28 10:59:57 hey mithro May 28 10:59:59 ehm May 28 11:00:02 hey mickeyl|breakfas May 28 11:00:17 hey koen :) May 28 11:00:22 :) May 28 11:03:50 koen: even though I didn't get a SoC, I'm really glad you did - the staging stuff is sorely needed May 28 11:04:32 I know May 28 11:06:19 OE really needs stuff like that May 28 11:06:37 some fundamental issues require some man months of work May 28 11:09:16 koen: yeah May 28 11:09:31 If i had just lost to you then it would have been okay :) May 28 11:14:56 03koen 07org.oe.packaged-staging * r40ef60b4... 10/classes/packaged-staging.bbclass: packaged-staging.bbclass: adapt to the new plan May 28 11:15:00 hmmm May 28 11:15:02 still on the ipgk segfault in fread libc2.4 STOP first ipkg segfault seen ~ correspond to the last arm-longlong.patch changes (include stdlib/mp instead of sysdeps/generic/mp) STOP Does anyone think it could be the cause ? May 28 11:15:08 I'm starting to be like zecke now May 28 11:15:15 where did I put the other commits? May 28 11:15:34 koen lol May 28 11:17:21 s/mp/mp_clz_tab.c/ May 28 11:18:42 cyrilRomain: could be May 28 11:23:07 ah, they must be on my linux workstation May 28 11:35:19 hum... iirc someone had trouble with floating point these day (?) is it related to gmp ? the arm-lonlong patch includes stdlib/mp_clz_tab.c which includes gmp .... May 28 11:35:30 How can I extract a source package under a different different directory name? (Problem at hand: I have a gdk-pixbuf.bb, which extracts gtk+: patches/ are in gdk-pixbuf, source in gtk+, patching fails). May 28 11:35:43 cyrilRomain: yes, me on a x86_64 machine. May 28 11:35:55 cyrilRomain: Had to add a site variable to get gmp to compile. May 28 11:36:24 likewise: S = "${WORKDIR}/thediryouwnat" ? May 28 11:36:30 want* May 28 11:36:30 likewise: but a compilation issue, not an execution one ? (i'm tracking down a ipkg segfault) May 28 11:37:09 cyrilRomain: that was a compilation issue, I have not seen execution failures yet (and I think I have used a bit of floating point apps) May 28 11:37:33 likewise: ok, glad to read that May 28 11:37:49 koen: I tried that already, but good to see that should be it. May 28 11:37:49 * cyrilRomain should track the bug elsewhere... May 28 11:44:34 koen: Using "S =" I can match the source dir to the tar dir. I want to rename the untarred dir (as it clashes with full gtk+), so I have to override the do_unpack() method. May 28 11:45:46 sounds like you answered your own question :) May 28 11:45:54 do_unpack() { May 28 11:45:56 foo May 28 11:45:56 } May 28 11:46:55 koen: I think I am starting to get the hang of OE :-) (And now I will forget everything I knew about make :-)) May 28 11:47:00 :) May 28 11:51:29 whoa May 28 11:51:38 8 new survey entries since yesterday May 28 12:08:57 <[g2]> koen ping May 28 12:09:04 pong May 28 12:09:07 <[g2]> morning May 28 12:09:12 morning May 28 12:09:18 <[g2]> heh ChanServ May 28 12:09:27 <[g2]> damn tab complete ! May 28 12:09:28 hey [g2] you are the loft guy? May 28 12:09:34 <[g2]> mithro I am :) May 28 12:10:01 <[g2]> koen I've got a custom ubuntu that runs OE from the livecd May 28 12:10:10 <[g2]> it's Dapper as of yesterday May 28 12:10:16 [g2]: cool May 28 12:10:27 my workstation it ubdating right now May 28 12:10:44 <[g2]> I'd like to torrent it May 28 12:10:55 <[g2]> it runs most any OE build May 28 12:11:11 a torrent would be a good idea May 28 12:11:29 <[g2]> so I'd like to have a very simple Makefile for openezx and maybe something else May 28 12:12:33 how big is it? May 28 12:12:41 do you think we could do something a bit smaller? May 28 12:13:06 <[g2]> mithro it could be customized but its really not worth it May 28 12:13:25 <[g2]> openoffice is removed and a couple extra things are added May 28 12:13:47 <[g2]> the stuff needed to run OE, minicom, xchat etc May 28 12:14:02 <[g2]> I think the image is around 525MB May 28 12:14:37 <[g2]> there are only 2 things really missing in my mind May 28 12:14:55 who was the qemu specialist? The darn thing doesn't compile on Collie in .oz :\ May 28 12:15:28 <[g2]> 1) the OE meta-data and simple structure May 28 12:15:32 03koen 07org.oe.packaged-staging * r403feaa6... 10/packages/update-rc.d/update-rc.d_0.7.bb: Add update-rc.d May 28 12:15:37 03koen 07org.oe.packaged-staging * rd9745682... 10/classes/packaged-staging.bbclass: packaged-staging.bbclass: merge uncommitted stuf from my other workstation May 28 12:15:42 <[g2]> 2) being able to do a "TORAM" boot May 28 12:16:03 <[g2]> 1) is handled easily with the persistent storage option May 28 12:16:13 <[g2]> dunno if 2 is implemented in Dapper yet May 28 12:19:18 TORAM boot? May 28 12:19:32 <[g2]> mithro yeah May 28 12:19:46 whats that? May 28 12:20:09 put rootfs in ram so cd is removeabe May 28 12:20:18 <[g2]> in Knoppix there's a TORAM boot where the compressed rootfs is in ram for speed May 28 12:20:32 <[g2]> and the CD is removeable as ade|desk pointed out :) May 28 12:20:40 ahh May 28 12:37:21 is anyone else having problems with qemu-nate b0rking with http://openembedded.pastebin.com/742709? May 28 12:37:29 s/nate/native May 28 12:38:06 CoreDump|home: I have it assume-provided since it doesn't build and debian has 0.8.x May 28 12:38:34 doesn't build for me, thatis May 28 12:38:39 ahh May 28 12:38:57 well, it build fine and now b0rks on do-stage heh May 28 12:39:04 and it built before here May 28 12:39:40 do you have a test in configure.ac to find the correct strip in your toolchain ? May 28 12:40:54 ade|desk: I've even linked strip to /bin, I believe the problem is that qemu-img is not found and install is complaining about a missing source binary May 28 12:41:58 nm, qemu-img is there alright May 28 12:43:13 ibot: botmail for zecke: would this work: http://dominion.kabel.utwente.nl/koen/pda/files/glibc-no-isystem.patch ? May 28 12:48:16 heh, if I remove the "-s" from "install" in the makefiles it works fine. May 28 12:50:28 sounds like your test for @STRIP@ is foobar May 28 12:50:58 hmm May 28 12:54:41 * chouimat is back. May 28 12:56:36 strip is in the search path and "install" still doesn't find it grrr May 28 12:57:27 and ran manually it works of course May 28 12:59:01 CoreDump|home, yea, i had to do that in my scripts for a few packages, the hosts install runs just plain strip, not sure if oe adds ${target_alias}- to the prefix of the tools May 28 13:00:09 JoeSchmo: qemu-native is a native package, the distro strip works fine. It is just that it isn't found for whatever reason May 28 13:00:45 ah May 28 13:01:15 I've run the install command in the same shell env where it breaks in the makefile...it works manually May 28 13:01:52 CoreDump|home: is there a bit in configure.ac , AC_CHECK_TOOL(STRIP, strip, : ) ? May 28 13:02:24 install -s probably wouldn't be affected by that May 28 13:02:40 ade|desk: it's not a typical configure script =) May 28 13:02:51 but yeah, it is found May 28 13:06:58 koen: could you post your ASSUME_PROVIDED line please? May 28 13:07:46 ASSUME_PROVIDED += "diffstat-native qemu-nativ" May 28 13:07:54 ehm May 28 13:07:57 ASSUME_PROVIDED += "diffstat-native qemu-native" May 28 13:08:03 CoreDump|home, just looking at the coreutils install code, it calls execlp '("strip", "strip", name, NULL);', just as a debugging thing, i'd put a @echo $PATH above the install -s line in the Makefile May 28 13:08:12 * chouimat is away: time to clean the apartment before the ladies arrive May 28 13:08:38 koen: thanks! May 28 13:08:45 JoeSchmo: I gave up... May 28 13:09:02 :/ May 28 13:09:12 but what the hell, I'll check =) May 28 13:12:08 /usr/bin is in PATH May 28 13:14:31 if strip is /usr/bin/strip, what the heck May 28 13:14:54 yep May 28 13:15:05 install: cannot run strip: No such file or directory May 28 13:15:06 install: strip failed May 28 13:15:10 hmm May 28 13:15:46 is $cross_prefix defined as your native complier prefix ? May 28 13:15:50 the sourcefile for the install command is there May 28 13:16:41 no, it points into staging/bin May 28 13:16:57 s/bin// May 28 13:17:12 but install shouldn't give a damn about that ;) May 28 13:17:49 anyways, enough is enough, I'm using ASSUME_PROVIDED since it takes quite some time to compile in any case May 28 13:18:55 re zecke May 28 13:19:04 hey May 28 13:19:09 woglinde: may we go? May 28 13:19:24 zecke holste mich ab? May 28 13:20:08 aber wird nix May 28 13:22:19 :( May 28 13:22:29 hm May 28 13:22:34 oder willst du wirklich? May 28 13:23:10 woglinde: well, I can do my physichs assignment as well May 28 13:23:23 yepp mch das May 28 13:23:27 ich mach alp4 May 28 13:25:19 ung guck online May 28 13:25:29 ups sorrx May 28 13:25:51 hehe May 28 13:27:59 03koen 07org.oe.dev * r30f22806... 10/packages/gpe-calendar/gpe-calendar_cvs.bb: gpe-calendar: add libsoup to the DEPENDS field of the cvs version May 28 13:28:55 CoreDump|home: gpe-calendar cvs should now be able to download and update .ics calendars over webdav May 28 13:29:05 i.e. webcals, google calender, etc May 28 13:29:25 koen: libsoup was SOAP? May 28 13:29:39 zecke: no idea May 28 13:29:45 iirc evo uses libsoup as well May 28 13:29:51 well May 28 13:29:56 then it must be good :} May 28 13:29:56 * CoreDump|home somehow has the feeling that koen made a wrong tab-complete May 28 13:30:09 CoreDump|home: no, I didn't May 28 13:30:16 hmm May 28 13:30:33 I didn't play w/ calendar in ages May 28 13:30:38 CoreDump|home: I suspect that most hentges users complain to you about ko/pi being unavailable in X May 28 13:30:48 ah May 28 13:30:58 well, matter of factly they don't =D May 28 13:31:38 98% of "my" users use Opie since the GPE image is still kinda new May 28 13:32:12 CoreDump|home: qemu-native needs to DEPEND on zlib-native then May 28 13:32:47 zecke: I pretty much gave up with trying to compile it. Check the backlog ;) May 28 13:32:58 zecke: why zlib ? May 28 13:33:23 CoreDump|home: no backlog ;) May 28 13:33:47 ade|desk: that was a previous problem w/ qemu, unrelated to strip May 28 13:34:06 ah ok May 28 13:39:22 can you configure which locales are generated by glibc or is it all or nothing only? May 28 13:40:03 CoreDump|home: it configurable May 28 13:40:09 hmm May 28 13:40:37 I'm reading the bbclass but that doesn't exactly stand out May 28 13:40:50 ENABLE_BINARY_LOCALE_GENERATION ?= "0" May 28 13:40:56 that's all or nothing May 28 13:41:17 CoreDump|home: something like http://rpsys.net/openzaurus/temp/glibc_workionprogress.patch May 28 13:41:43 koen: nice May 28 13:42:20 CoreDump|home: iirc it doesn't work 100%, but you'd have to ask RP for the details May 28 13:43:03 if [ "${LIMIT_BUILT_LOCALES}" != "${LIMIT_BUILT_LOCALES}" ]; then <- pretty much disabled by default ;) May 28 13:45:23 disabled unless defined May 28 13:46:15 hi May 28 13:46:18 re hrw May 28 13:46:25 hey hrw May 28 13:48:48 koen: nice password you choosed for me;) May 28 13:49:25 :) May 28 14:00:36 koen: now I can add planet.ltg to my links section on webpage May 28 14:06:14 :) May 28 14:07:35 hmm May 28 14:07:36 :( May 28 14:08:33 koen: you may have seen it already but the first set of graph generation is done May 28 14:08:50 koen: alldepends used '---' for DEPENDS and '- - -' for RDEPENDS May 28 14:08:58 koen: and the graph is not looking too bad May 28 14:09:42 http://ewi546.ewi.utwente.nl/OE/kernel-build.png looks cure May 28 14:09:45 ehm May 28 14:09:46 cute May 28 14:10:01 mickeyl|breakfas: Do you still maintain the .bb for qpegps? May 28 14:10:08 koen: that is old May 28 14:10:14 koen: it now prints the version as well May 28 14:10:22 koen: and probably soon the provider name as well :} May 28 14:14:09 koen: and pdf output looks better (on OSX) May 28 14:16:27 er... I have a question... how much time does it take to generate binary locales using qemu-arm? May 28 14:16:53 1.5-3h May 28 14:16:59 oh, ok... May 28 14:17:02 thanks May 28 14:17:17 then I'm out to drink some tea :) May 28 14:21:49 wow OPie has ~12000 revs May 28 14:34:08 hi - i wanted to ask whether OE works for i386 May 28 14:35:14 yes May 28 14:35:34 so to say i could make my own "distro" then? May 28 14:35:57 yes May 28 14:36:11 i am just looking for a way to build a relatively small linux rootfs for my netvista, i tried debootstrap with debian sarge and the base installation is about 144mb - a bit too big for me May 28 14:36:33 MACHINE="netvista" is already in May 28 14:36:43 dont know if it?s the same netvista.. May 28 14:37:16 well sounds quite good though May 28 14:38:59 what image would i choose then? essential-image? bootstrap-image? essential-machine-image? May 28 14:40:42 xterminal-image ? May 28 14:41:16 no i want to use the netvista as "thin server" May 28 14:41:49 image should be as small as possible but provide neccessary libs for mysql/apache etc. (i wanted to simply install xampp on it) May 28 14:43:15 try bootstrap-image May 28 14:43:19 ok May 28 14:43:38 distr = generic May 28 14:43:39 ? May 28 14:43:43 +o May 28 14:46:25 later all May 28 14:55:05 koen: it looks like kdrive ignores a keymapping of XF86_Switch_VT_* May 28 14:55:21 but we can achive the same goal with keylaunch... May 28 14:55:57 CoreDump|home: I personally don't have a use case for going to console while X is running May 28 14:56:06 I either fire up rxvt or kill X May 28 14:56:14 so I haven't looked into the problem May 28 14:56:18 it is useful at times May 28 14:56:43 no doubt May 28 14:56:48 I just don't use it :) May 28 14:56:51 ;) May 28 14:57:38 when not cross-building (i386 -> i386) OE will still use it?s own compiler/toolchain and so first build it, not use the available one in the system? May 28 14:58:07 would anyone have any problems with setting SUID but on chvt.console tools? May 28 14:58:17 TheCan: well unless you ASSUME it as provided May 28 14:58:39 zecke, ok just wanted to be sure - i guess it?s better it uses its own so the versions match May 28 14:59:39 cu May 28 15:00:45 is the netvista port currently maintained? May 28 15:02:23 hi, anyone with bluetooth knowledge around? I'm trying to connect to my BT headset but I can't even l2ping it although I can l2ping my phone May 28 15:02:56 switch the headset on? May 28 15:03:31 it is on, and if it is in pairing mode hcitool scan sees it but l2ping gives no reply May 28 15:03:57 if it is not in pairing mode hcitool scan does not see it May 28 15:04:19 and I think by now I have seen every l2ping error message that exists .... May 28 15:07:30 eFfeM, maybe it requires a pre-set bluetooth pin to be set and therefore pairing does not work? May 28 15:08:40 it does require a pin, but I have a pinhelper pointing to an executable shell script saying echo "PIN:0000" May 28 15:09:58 but it could well be related, I've added an echo date >>/log in the pinhelper script but it does not do anything May 28 15:10:07 that is: it is not invoked May 28 15:10:54 "generic" on i386 is broken btw - glibc-initial does not build May 28 15:16:12 hm. started hcid in non daemon mode, it does not even get to the pin request but hangs at link_key_request May 28 15:18:44 soo, I'll just assume that no-one has a problem w/ an SUID root chvt and change it.... May 28 15:26:29 found it: /var/lib/bluetooth/*/linkkeys contained a bad value. don't know how it got there, but removing the file cured it May 28 15:30:53 hi May 28 15:37:28 anyone knows which glibc is known to work with i386? May 28 15:40:50 TheCan: no May 28 15:41:00 2.3.5+cvs and 2.3.3 failed both on me :( May 28 15:41:04 now trying 2.3.2 May 28 15:41:23 oh and newest cvs also failed May 28 15:42:44 TheCan: you might need NTPL :} May 28 15:45:35 zecke, the configure script complains about gcc being too new :/ May 28 15:46:01 TheCan: using gcc4? May 28 15:46:04 yes May 28 15:46:10 TheCan: set PREFERRED_VERSION_.... May 28 15:46:15 for gcc? May 28 15:46:24 yes May 28 15:46:29 indeed i wanted to use gcc 4.1.1 thats why i set preferred version so May 28 15:46:37 didnt know old glibc could fail because of this May 28 15:46:47 on sh3 it worked at least May 28 15:46:59 i?ll give glibc 2.4 a try May 28 15:47:52 :D May 28 15:48:01 2.4 even the patches fail May 28 15:49:07 mithro: http://doc.trolltech.com/qtopia2.2/html/environment-prereq.html <- lol May 28 15:50:39 zecke: ? May 28 15:50:53 heh May 28 15:50:58 it doesn't 2.95.3 May 28 15:51:02 compiles != works May 28 15:51:03 doesn't list May 28 15:51:29 I was searching for a good analysis of jffs2 and other filesystems May 28 15:51:40 in regard to compression ratio and wanted to show to CoreDump|home May 28 16:50:20 no glibc version seems to work with gcc 4.1.1 on i386 :( May 28 16:53:12 TheCan: well May 28 16:53:22 TheCan: you build an initial gcc, then glibc, then gcc May 28 16:53:24 "patches welcome"? :D May 28 16:53:33 TheCan: you should be able to use a gcc3 for the initial one May 28 16:53:42 TheCan: compile the glibc, then compile a modern gcc4 May 28 16:54:13 zecke, setting gcc-cross and gcc to 4.1.1 and gcc-initial to 3.4.4 ? May 28 16:54:21 sorry gcc-cross-initial May 28 16:55:14 TheCan: I suck at remembering names, but yes :) May 28 16:55:20 :) thx! May 28 16:55:36 how big is the final userland be expected to be? May 28 16:55:46 like 10mb or less? May 28 16:56:05 TheCan: that depends May 28 16:56:22 using glibc, not uclibc and just the "bootstrap-image" task May 28 16:56:36 depends on the filesystem May 28 16:56:36 TheCan: ~7-9Mb IIRC May 28 16:56:45 thx May 28 16:57:03 this sounds quite great compared to debootstrapped sarge?s 144mb :D May 28 16:57:41 TheCan: http://hentges.net/tmp/screenshots/Zaurus/Akita/Misc/essential-to-boot+wifi.jpg ;) May 28 16:57:59 :) May 28 16:58:06 and yes, it booted on its own heh May 28 16:58:09 on zaurus that is? May 28 16:58:19 yes, SL-C1000 May 28 16:58:27 hmm my problem is i got somehow precompiled apache2+mysql+perl+php etc. May 28 16:59:01 it?s linked against normal glibc - is it supposed to work? May 28 16:59:25 never built x86 images, dunno May 28 16:59:52 well hopefully i get glibc finally working now (2.3.5+cvs failed just right now:( ) so we will see May 28 17:01:32 btw is there some way to use "arch-dependent" patches? this would be nice May 28 17:01:51 like you define in the bb-file to apply patches only for certain architectures May 28 17:02:02 SRC_URI_ += "file://patch.patch;patch=1" May 28 17:02:09 it could be of course that some of the say ARM patches break i386 May 28 17:02:17 unlikely May 28 17:02:26 hm.. May 28 17:02:34 unless i386 uses sysdeps/arm/ May 28 17:02:54 then it?s a bit strange glibc 2.3.5+cvs compiled nice on (initial) gcc 4.1.1 with sh3 while it breaks on i386 May 28 17:06:27 arr...2.3.3 doesnt work either :( May 28 17:09:43 if 2.3.2 also breaks i?ll try without all the patches - i suppose glibc is supposed to compile cleanly without patches on i386? May 28 17:10:06 damn OpenEmbedded.biz May 28 17:10:16 ? May 28 17:10:38 name stealing? :/ May 28 17:10:38 TheCan: lol, I do not suspect any patch to break i386 build May 28 17:10:48 TheCan: start reading the error message and think :) May 28 17:10:58 TheCan: yeah name stealing. OpenEmbedded.de, OpenEmbedded.biz May 28 17:11:16 TheCan: we now have 'two' Embedded companies that are doing name stealing May 28 17:11:32 you could propably court them? May 28 17:11:50 you used this name first so you got the name rights for it May 28 17:12:03 otherwise they could register it as their trademark and then court you May 28 17:12:47 oe isnt an organisation May 28 17:12:59 TheCan: we have no trademark :} May 28 17:13:03 and there is nothing to prevent others from using the name May 28 17:13:36 zecke: I've managed to generate lots of errors on the tinderbox :) May 28 17:13:45 errors? May 28 17:13:49 and if a trademark existed it would only be valid in the country that it was registered May 28 17:13:50 hm well..i don?t remember exactly how this was unfortunately but i had once hearings about this topic :D May 28 17:13:54 zecke: well, build failures May 28 17:14:06 emte, well you can have international trademarks May 28 17:14:12 zecke: oz354x appears to be the most stable :) May 28 17:14:14 of course this costs a lot too much May 28 17:14:16 TheCan, IF you have the money May 28 17:14:35 RP: that was the goal of .oz354x :) May 28 17:15:04 koen: It doesn't help that my scripts keep screwing up... May 28 17:15:21 How are we going to handle multimachine and packages like sysvinit? May 28 17:15:46 add a new flag? May 28 17:15:49 RP: any explanations why your builds look so long? May 28 17:15:50 interesting ... they point sombrio.com to oe.biz May 28 17:15:57 MULTIMACHINE_UNSAFE? May 28 17:16:06 RP: interrupted internet connection? May 28 17:16:30 zecke: Which builds took "so long" in yor opinion? May 28 17:16:34 RP: see contrib/openzaurus for buildscripts May 28 17:16:42 they also use embeddedoss.org May 28 17:16:47 btw - why is the ipk/deploy place not seperated in archs? May 28 17:16:50 koen: I wondered about MULTIMACH_ARCH which is generted by lkooking thought each item in package and summarising the PACKAGE_ARCHs? May 28 17:17:23 TheCan: because packages are? May 28 17:17:24 RP: I'd like to know it in advance May 28 17:17:34 TheCan: _arm, _xscale _moo, _openzaurus May 28 17:17:35 koen: We can know in advance... May 28 17:17:42 zecke, yeah i know May 28 17:17:50 but wouldnt folders be better? May 28 17:17:57 TheCan: in which way? May 28 17:18:01 koen: We'd do this in an anonymous function May 28 17:18:02 RP: yes by looking at SRC_URI and stuff in FILES_PATH May 28 17:18:19 zecke, would look more clean in my eyes May 28 17:18:21 TheCan: this way, it works with existing stuff (debian, ipk) May 28 17:18:46 anyway, first final starts... May 28 17:20:06 zecke|tv: How do I stop a build that tinderbox hasn't noticed has been aborted? May 28 17:20:30 RP: later, how did you abort? May 28 17:20:43 zecke|tv: Ctrl+C to bitbake May 28 17:21:08 hi guys May 28 17:21:18 hi Bernardo May 28 17:21:23 hey Bernardo May 28 17:24:59 glibc 2.3.2 worked fine btw (unlike 2.3.3, 2.3.5cvs etc.) May 28 17:25:13 is there a chance i could commit my netvista distro file into OE? May 28 17:25:17 (once it?s finished) May 28 17:25:58 if you plan to maintain it May 28 17:26:38 yeah i thought about opening a netvista mini-distribution May 28 17:26:50 but hard to tell how much time this takes etc. May 28 17:27:31 just setup a way to do automated builds occasionally and fix quirks May 28 17:28:03 and awnswer questions about it i suppose May 28 17:28:17 yeah i thought about just putting a bb-file which allows for clean bootstrap-image compilation May 28 17:28:44 sounds good May 28 17:28:54 it?s basically i386 May 28 17:29:05 but since generic (using latest libs) seems to be broken on i386.. May 28 17:38:09 koen: How does http://www.rpsys.net/openzaurus/temp/multimachine.bbclass look? May 28 17:43:59 looks ok May 28 17:55:41 * chouimat is back. May 28 17:56:50 koen: ok to update angstrom with that? (I'll update the other distros as well) May 28 17:58:57 RP: please commit it if it works for you May 28 18:02:26 On a c7x0 running with 354rc, I have a strange alsamixer error message : inappropriate iotctl for device May 28 18:02:35 koen: I just need to change multimachine from an include into a .bbclass. I can't see a problem with that as most people will be using it from the distro files though and as long as I change that, everything should be fine. I'm just checking it all works as expected May 28 18:02:37 which modules are necessary for the sound? May 28 18:03:18 guylhem_: That's probably best asked in #openzaurus May 28 18:03:30 oops sorry May 28 18:04:19 btw I will try to release a bb for postgres and exim which I have running on my device May 28 18:04:29 ah cool May 28 18:04:35 guylhem_: sounds good :) May 28 18:04:46 I've been wanting a postgres .bb for a while May 28 18:04:46 hmmm.... binary locale generation takes a really long time... 4h have passed, and it still generates first one :( May 28 18:05:04 compilation is easy. the problem for me will be learning how to use bb and sstuff May 28 18:05:25 koen, do you want me to send you the binaries and the configuration script by email? they run on angstrom May 28 18:06:01 guylhem_: I won't have time to test them anytime soon, so I'll wait for the .bb :) May 28 18:32:52 hmm i wonder if mtn will ever be smp aware May 28 18:33:27 ade|desk: when I asked njs he said "you can use the other cpu for watching a dvd while it's pulling" May 28 18:33:39 ade|desk, yep. That would be nice... OTOH I still have CPU power left to do something else while it eats up the other CPU May 28 18:34:06 i have 3 idle cpus :( May 28 18:34:16 watch 3 dvds at once! May 28 18:34:27 ade|desk, real or virtual? May 28 18:34:28 njs: on a headless server May 28 18:34:33 Philippe: real May 28 18:34:56 more seriously, the answer is "maybe someday, but really there's no point in even needing all that cpu, so..." May 28 18:35:03 my net connection aint good enough to stream mpeg2 over adsl May 28 18:35:13 ade|desk, cool! I only have two real, and two more thanks to hypertreading. May 28 18:35:31 njs: ins't verifying stuff inherently linear? May 28 18:35:32 ade|desk, you could recode DVD's :-) May 28 18:35:50 no dvd drive on the server either May 28 18:35:56 koen: *shrug* I don't know what exactly one would do, but e.g. it might be possible to SHA1 things in parallel May 28 18:36:20 and in the long run SHA1 should be the bottleneck May 28 18:36:26 it sure ain't the bottleneck now, though May 28 18:36:54 ade|desk, the solution is "bitbake world" ... :-) May 28 18:37:18 bitbake world while pulling, seems a bit pointless May 28 18:37:19 bootstrap-image on i386 done :) wohoo May 28 18:37:43 03coredump 07org.oe.oz354x * r26addc12... 10/packages/keylaunch/ (4 files in 4 dirs): keylaunch: Akita / Spitz: Update keymapping to allow switching to the VT from inside X with Fn+Left or Fn+Right. Also includes a new script to make chvt SUID root to allow users to switch VTs May 28 18:37:47 03coredump 07org.oe.dev * r0c120ae0... 10/packages/keylaunch/ (4 files in 4 dirs): keylaunch: Akita / Spitz: Update keymapping to allow switching to the VT from inside X with Fn+Left or Fn+Right. Also includes a new script to make chvt SUID root to allow users to switch VTs May 28 18:37:51 03rpurdie 07org.oe.dev * rdd8feb40... 10/ (6 files in 3 dirs): multimachine: Change from an include to a .bblass (and update distros using multimachine accordingly). Add code to find packages with machine specific subpackages and build these in a multimachine compatible way. May 28 18:37:53 * RP -> back later May 28 18:38:37 still having problems with gcc 4.1.0 w.r.t atexit() May 28 18:38:55 the image is 9.9mb - nice nice :) May 28 18:39:19 9.9 for a bootstrap-image .. seems large May 28 18:39:30 i386 May 28 18:39:33 not compressed image May 28 18:39:35 would expect ~5 May 28 18:39:39 i mean how much space it takes (inside) the image May 28 18:39:45 compressed it?s 1.8mb May 28 18:39:52 i just mounted it as loop device May 28 18:39:52 that better May 28 18:40:07 ade|desk: well, the bootstrap-image has too much crap in it May 28 18:40:09 glibc or uclibc ? May 28 18:40:17 CoreDump|home: indeed May 28 18:41:00 03coredump 07org.oe.oz354x * r3d1e1f91... 10/packages/gpe-login/ (files/chvt-keylaunch.patch gpe-login_0.86.bb): gpe-login: Enable switching to the VT with Fn+[Left|Right]Arrow and killing X with Fn+Esc May 28 18:41:04 03coredump 07org.oe.dev * rcd0b649e... 10/packages/gpe-login/ (files/chvt-keylaunch.patch gpe-login_0.86.bb): gpe-login: Enable switching to the VT with Fn+[Left|Right]Arrow and killing X with Fn+Esc May 28 18:43:05 ade|desk, with glibc May 28 18:43:08 hmmmm.... my i386 image takes 9.5M with uclibc and 12M wigh glibc May 28 18:43:15 how can i get the kernel now btw? May 28 18:43:28 should i manually extract it out of the ipk file? May 28 18:43:33 or is there any kernel-target? May 28 18:43:49 virtual/kernel May 28 18:43:57 and .ipkgs are .debs May 28 18:44:01 i know May 28 18:44:44 i mostly do modules as ipkg and keep the kernel as a bzImage May 28 18:45:19 hmm the kernel_image ipk is only ~700 bytes May 28 18:45:35 time to watch top gear May 28 18:45:43 allow_empty is set to 1 May 28 18:45:49 oooh top gear May 28 18:46:02 koen|tv: you can get bbc2 over there May 28 18:46:10 ade|desk: yes :) May 28 18:46:13 ~hail bbc2 May 28 18:46:15 * ibot bows down to bbc2 and chants, "I'M NOT WORTHY!!" May 28 18:51:04 omg now i built everything for hours but now can?t find the machine for which i did it :D May 28 18:55:09 greetings ! May 28 18:56:32 Ifaistos: hey, I have the ppc almost setup :} May 28 18:56:57 hallo ifaistos May 28 18:57:49 zecke : Good :) Let me know if you manage to get any graphical enviroment running on it May 28 18:57:49 mickeyl|breakfas: http://handhelds.org/~bugzilla/show_bug.cgi?id=650 May 28 18:57:59 woglinde : Hi ! May 28 18:58:39 i think i found a bug in ipkg.... May 28 18:58:56 if i try to remove a package with dependancies i get the usual erros May 28 18:59:26 root@OWMNR:~# ipkg remove quagga May 28 18:59:26 Nothing to be done May 28 18:59:26 Collected errors: May 28 18:59:26 Package quagga is depended upon by packages: May 28 18:59:26 quagga-bgpd May 28 18:59:27 These might cease to work if package quagga is removed. May 28 18:59:29 You can force removal of this package with -force-depends. May 28 18:59:31 You can force removal of this package and its dependents May 28 18:59:33 with -force-removal-of-dependent-packages or -recursive May 28 18:59:35 or by setting option force_removal_of_dependent_packages May 28 18:59:37 in ipkg.conf. May 28 18:59:57 if i use ipkg remove quagga -V 0 in order to get only errors.... nothing is displayed May 28 19:01:02 according to the ipkg help -V 0 should return errors only... but it returns nothing May 28 19:01:12 any ideas ? May 28 19:46:17 I'm going to bed now May 28 19:46:18 cya May 28 19:46:28 cya May 28 19:46:52 n8 zecke May 28 19:50:14 hi May 28 19:51:02 hi Bernardo May 28 19:51:15 hey Bernardo May 28 19:51:17 I've been unable to build oz354x since I switched my host from mandriva to kubuntu dapper, alway fails in konq-embed with "no rule to make target '-lpcre'" May 28 19:51:30 and pcre built without problems May 28 19:51:48 any idea? I've blasted tmp/* and rebuilt twice May 28 19:52:10 that you installed pcre-dev? May 28 19:52:21 or whatever ubuntu calls it May 28 19:52:27 libpcre3-dev iirc May 28 19:52:51 native? I don't think so, let me check May 28 19:53:12 ok, installing it now May 28 19:53:17 thanks koen May 28 19:56:24 yep, now it built May 28 20:02:46 where do i normally configure static network in OE? which files i mean? May 28 20:23:39 Thecan : /etc/network/interafces May 28 20:23:48 yeah i found it already, thx May 28 20:23:48 Thecan : /etc/network/interfaces May 28 20:24:00 i just see stupid xampp is precompiled for i686 :( May 28 20:24:04 lets see if it works on the netvista May 28 20:36:44 hmm no apache in OE ? May 28 20:54:09 * Philippe is away: visual contact - melancholic dreams May 28 21:52:08 he kergoth May 28 21:53:53 hey May 28 23:31:58 * v8jlene is back (gone 16:28:49) May 28 23:59:28 Why do I get "ERROR: Nothing provides runtime dependency libz1" for something that has RDEPENDS="python-compression"... and why if I remove the rdepends, build, add the rdepends back I no longer get the error ;( May 29 00:33:41 because it caches it May 29 00:34:49 and i suspect python-compression redepends on zlib May 29 00:35:07 which is a compression library May 29 00:38:23 emte: Yes, python-compression RDEPENDS on libz1. I think zlib should have an RPROVIDES libz1, which it currently doesn't. May 29 00:39:07 who is zlibs maintainer? May 29 00:39:36 MAINTAINER = "Chris Larson " May 29 00:40:21 it does not need an RPROVIDES. thats all automatic May 29 00:40:28 the problem isnt likely to be with the zlib .bb, but elsewhere May 29 00:44:26 kergoth: If I put RDEPENDS = "libz1" in a recipe then I get "ERROR: Nothing provides runtime dependency libz1". Should this RDEPENDS be something else? May 29 00:44:35 i suspect that your probably not building a v1 xlib May 29 00:44:40 zlib* May 29 00:45:41 emte: The ipkg's are there - even building all the zlibs recipe manually doesn't change this issue... May 29 00:56:53 what versions tho May 29 00:58:25 as well you should note that making an ipkg is not the same as being staged May 29 00:59:33 v8jlene, RDEPENDS get added automatically. you dont have to add that at all. May 29 01:00:18 emte: Yeah, I'm release that. Note that this is just a runtime dependency, everything actually builds and works fine if I remove it - except I need to manually install python-compression then... May 29 01:02:17 kergoth: Yeah, I know that's what's meant to happen ;) I'm trying to figure out why it isn't and/or what's wrong with it... May 29 01:03:55 i am curious, why are you specifying an rdepends if you actually rdepend on python-compression? May 29 01:04:06 rdepends for zlib* May 29 01:04:07 If I take a package, lets says nano for example, and add RDEPENDS="libz1" then you get "ERROR: Nothing provides runtime dependency libz1". May 29 01:04:59 emte: I RDEPENDS python-compression, python-compression RDEPENDS python-zlib and python-zlib RDEPENDS libz1 May 29 01:05:05 why are you adding RDEPENDS=libz1? you should not be. May 29 01:05:32 kergoth: I'm not, that's python-zlib. That was one of my questions, is python-zlib wrong? May 29 01:06:05 hard to say. is your ipk actually named libz1, or is it named zlib1g? May 29 01:06:30 libz1_1.2.3-r0_sh4.ipk May 29 01:06:37 no idea then. May 29 01:06:40 the ipkg is named libz1 but i think it actually provides libz.so.1 May 29 01:07:07 doesnt matter May 29 01:07:09 lookint at my system that is what i expect May 29 01:07:24 ah May 29 01:07:28 anyway bbiab May 29 01:29:13 kergoth: Would removing libz1 from python-zlib RDEPENDS be the correct fix? The python-zlib.shlibdeps is automatically getting a "libz1 (>= 1.2.3)" entry, so there should be no need to try and explicitly declare that dependency? And nothing else explicitly RDEPEDS on libz1. **** ENDING LOGGING AT Mon May 29 02:59:56 2006