**** BEGIN LOGGING AT Sat Aug 26 02:59:57 2006 Aug 26 07:12:39 good morning all Aug 26 07:38:23 morning all Aug 26 07:40:34 hey RP Aug 26 07:40:47 morning RP and koen Aug 26 07:40:51 RP: wanna sign the petition to make days have 36 hours? Aug 26 07:40:54 hey lamikr Aug 26 07:41:03 koen: That would be useful :) Aug 26 07:42:43 Yes, and I am wondering why no-one is never teching anything usefull in the schoold or universities. Like soldering workshop for newbie ipaq hackers Aug 26 07:43:04 we have soldering workshops Aug 26 07:43:14 but then again, I do EE, not CS :) Aug 26 07:43:30 EE? Aug 26 07:43:31 lamikr: do you plan on attending FOSDEM next year? Aug 26 07:43:37 Electrical Engineering Aug 26 07:43:44 Yes I will Aug 26 07:43:50 cool Aug 26 07:44:39 I was thought of as odd at school as I did welding evening classes at the local college... Aug 26 07:44:41 Well, I am also more like mathematics and software, but http://openwince.sourceforge.net/jtag/iPAQ-3600/ got me really excited Aug 26 07:45:35 That would be a nice practice before getting to real stuff (= trying to find jtag pins in h6300...) Aug 26 07:46:42 Good thing for the end of summer is that fosdem is some months nearer Aug 26 07:56:28 ~spitz Aug 26 07:56:31 ~husky Aug 26 07:56:32 hmm... spitz is the Sharp SL-C3000, or a dog Aug 26 07:56:34 it has been said that husky is Sharp Zaurus SL-C760 or a dog Aug 26 07:56:40 ~borzoi Aug 26 07:56:42 methinks borzoi is the Sharp SL-C3100, or a dog (see http://www.dogbreedinfo.com/borzoi.htm ) Aug 26 07:56:57 morning all Aug 26 07:57:00 ~3200 Aug 26 07:57:20 ~terrier Aug 26 07:57:21 terrier is probably Sharp Zaurus SL-C3200, or a dog Aug 26 08:36:24 grrr Aug 26 08:36:26 my AP is disliking my switch Aug 26 10:41:29 morning guys Aug 26 10:46:05 good morning mickeyl Aug 26 10:56:29 * mickeyl fighting with lexware financial office atm. Aug 26 10:56:32 darn, I'm a developer, not a bookkeeper! Aug 26 11:01:17 welcome to corporate life ;P Aug 26 11:01:31 *nod* Aug 26 11:01:48 the days i've spent dealing with the purchasing department... Aug 26 11:02:06 it's crazy. i'm just starting as a freelancer and i almost spent 2 weeks on buerocratic stuff now Aug 26 11:02:10 ibot: spell buerocratic Aug 26 11:02:12 possible spellings for buerocratic: bureaucratic bureaucrat bureaucratize bureaucrats bureaucrat's Aug 26 11:02:39 that's french for you ;) Aug 26 11:03:22 yea Aug 26 11:04:51 mickeyl: can bitbake print all PREFERRED_PROVIDERS? Aug 26 11:06:15 isn't bitbake -s doing that and more? Aug 26 11:06:53 * koen checks again Aug 26 11:07:19 for some reason bitbake trunk isn't allowing me to build diet-x11 as provider for virtual/libx11 Aug 26 11:08:36 mickeyl: it isn't Aug 26 11:09:30 ah Aug 26 11:09:36 somethings overriding it Aug 26 11:09:40 * koen puts it in local.conf Aug 26 11:15:43 ]WTF? Aug 26 11:15:44 NOTE: package pango-1.10.2-r0: task do_package: completedNOTE: package pango-1.10.2: completedNOTE: DBG: Active builds decreased to 2NOTE: package pango-1.14.0-r0: task do_install: completedNOTE: package pango-1.14.0: completed Aug 26 11:18:21 sounds like multithreaded bitbake Aug 26 11:19:02 it is Aug 26 11:19:08 and it's building 2 versions of pango Aug 26 11:19:57 that's not good Aug 26 11:21:15 morning Aug 26 11:22:28 ah Aug 26 11:24:00 PACKAGES_DYNAMIC strikes again Aug 26 11:31:53 03koen 07org.oe.dev * r989c601b... 10/ (1 packages/pango/pango_1.14.0.bb): pango 1.14: add appropriate PACKAGES_DYNAMIC to stop bitbake trunk from building pango 1.10 to satisfy dependencies on pango-module-* Aug 26 12:01:24 hi all Aug 26 12:01:42 gmane.org again carries the OE-devel list. Aug 26 12:12:32 france: Could it be that bugzilla.handhelds.org is not sending out passwords? Have been waiting for mine for about two days. I requested it to be resent just about a minute ago but did not receive anything yet. I checked spam folder. Aug 26 12:20:15 ~seen france Aug 26 12:21:05 france is currently on #gpe #familiar #oe #opie #handhelds.org. Has said a total of 26 messages. Is idling for 18h 41m 33s, last said: 'thank u :-D'. Aug 26 12:22:18 just re-installed OE database and monotone (back at college computer). Do I need to add key, or is it detected under ?home/.monotone/keys ? Aug 26 12:22:47 Laibsch: wrong channel Aug 26 12:23:06 Kristoffer: if you are using >0.25 it will be detected Aug 26 12:24:55 oki thx Aug 26 13:24:23 hi, I hope I'm in the right channel :). I was wanting to develop apps for openzaurus on my SL5500. I read through the wiki GettingStarted page, and I've got openembedded set up I think, but I'm not quite sure where to go from here. Can someone point me in the right direction? Aug 26 13:25:28 edit local.conf Aug 26 13:25:36 to include correct machine and distro Aug 26 13:25:47 then bitbake bootstrap-image Aug 26 13:27:05 ok, that's running now Aug 26 13:27:21 I'm guessing it's not as simple as just recompiling an existing app with a cross-compiler then? Aug 26 13:27:31 well Aug 26 13:27:45 did you check if the app you want to compile already has a bb file? Aug 26 13:28:00 basically, you need to write a bb file for the app Aug 26 13:28:01 well the app I want to compile doesn't exist yet, I'm going to write it Aug 26 13:28:11 if it is autoconf based, it is pretty easy Aug 26 13:28:48 use autotools, and it will be easy to make the bb file Aug 26 13:29:04 ok I can do that Aug 26 13:29:09 once you get the hang of it, it is easy Aug 26 13:29:17 getting the hang of it is the hard part Aug 26 13:29:25 that's always the way Aug 26 13:29:34 I did see an article on bb files in the wiki Aug 26 13:29:40 I'll have a read through that Aug 26 13:29:53 I just started by cpying bb files that worked :) Aug 26 13:30:07 I need to run now, but I am usually about weekdays Aug 26 13:30:22 thanks for your help Aug 26 13:30:33 your welcome Aug 26 13:30:37 good luck Aug 26 13:30:44 thanks :) Aug 26 13:55:19 03koen 07org.oe.dev * r011c3c34... 10/ (1 packages/busybox/busybox_1.2.1.bb): busybox 1.2.1: replace mv with a non-gnu-ish cp Aug 26 14:04:53 hi all Aug 26 14:05:23 anyone an idea what I've to add to depends if I miss "qmutex.h" in stage? Aug 26 14:26:27 03koen 07org.oe.dev * r64e3f453... 10/ (1 packages/db/db-native_4.3.29.bb): db-native: make sure do_package() is empty Aug 26 14:38:39 koen : hi ! If you find some free time have a look at #1327 #1328 #1346 #1347 Aug 26 14:40:18 koen: Is #1359 still valid now you found the problem with the pango.bb files? Aug 26 14:41:09 RP: yes Aug 26 14:41:41 RP: I can't think of a case where one would build two versions of a PN Aug 26 14:42:19 koen: Unless bitbake thought that was the only way to fulfil a set of dependencies... Aug 26 14:43:04 koen: I agree its not very obvious but the bitbake behaviour was right - we perhaps should wanr if we notice it happening? Aug 26 14:45:12 if bitbake thinks that it should halt and error out Aug 26 14:45:12 (in the OE case) Aug 26 14:45:12 and about all the 'file exists' stuff on the list: try a clean rebuild Aug 26 14:47:05 koen: I'm not sure about halt and error but warning would be nice... Aug 26 14:47:16 'file exists' stuff on list? Aug 26 14:47:32 the busybox problem (among others) Aug 26 14:49:25 Is there a reason that cdrtools isn't in packages? There's a -native version. (Looking to see if there are issues with it before trying to make the bb file) Aug 26 14:50:06 not that I'm aware off Aug 26 14:50:13 try asking eFfeM-log when he gets back Aug 26 14:50:23 okey dokey. Thanks. Aug 26 14:51:59 Koen, did you just fix the busybox build issue with the mv errors? Like an hour ago? Thanks! Aug 26 14:52:30 RP: can we exclude RRECOMMENDS from the depgraph? Aug 26 14:52:59 koen: Which depgraph? Aug 26 14:53:36 the one which bitbake complains about circular deps Aug 26 14:53:49 'bitbake xmms' if you want to know what I'm talking about Aug 26 14:54:12 The problem is debian.bbclass which needs to rewrite anything in a RRECOMMENDS Aug 26 14:54:19 (the same problem as RDEPENDS) Aug 26 14:55:37 *sigh* Aug 26 14:55:51 ~lart people complaining about bluetooth Aug 26 14:55:51 * ibot puts people complaining about bluetooth through a wood chipper Aug 26 14:56:18 If I remove bitbake's rreccommends handling, we'll then have bug where rrecommends don't get rewritten by debian.bbclass under certain circumstances Aug 26 14:56:30 Basically, I can't win :-( Aug 26 14:57:24 actually you can Aug 26 14:57:52 by having a do_package_renaming task before do_package Aug 26 14:58:09 hmm Aug 26 14:58:13 that won't work anymore Aug 26 14:58:19 now we stage after packaging Aug 26 14:59:11 If we split packaging into two parts, we probably could get something to work Aug 26 14:59:49 I have code to enhance the pkgmaps Aug 26 15:00:26 03pH5 07org.oe.dev * r295aed22... 10/ (1 packages/avahi/avahi_0.6.13.bb): avahi: add 0.6.13 Aug 26 15:00:41 koen: in the packaged-staging branch? Aug 26 15:00:43 do_write_ipk_list in http://www.openembedded.org/repo/org.openembedded.packaged-staging/classes/packaged-staging.bbclass Aug 26 15:00:47 RP: yes Aug 26 15:04:49 koen: Is there somewhere I can read about the current status of packaged staging? Aug 26 15:05:16 yes! Aug 26 15:05:17 http://www.openembedded.org/~koen/OE/SoC2006/ Aug 26 15:08:18 Interesting. Aug 26 15:08:42 I'm not sure "This should be safe for a multithreaded bitbake, given that no two do_stages access staging at the same time." is quite right though? Aug 26 15:11:20 hi guys Aug 26 15:11:20 it should read "do_stage may be the only task running" Aug 26 15:11:24 hey tkp Aug 26 15:11:31 hi tkp Aug 26 15:11:46 koen: So we need to add some locking in? :) Aug 26 15:12:05 yes Aug 26 15:12:26 and a clutter based gui by mallum Aug 26 15:13:00 I'm trying to get my target system mounted as read-only, but /var needs to be a tmpfs. The default oe fstab looks like it is supposed to do this, and I'm running with the default initscripts, yet /var does not appear to actually be a tmps atall Aug 26 15:13:19 We need a way to mark a task as exclusive... Aug 26 15:13:21 tkp: busybox 1.2.1? Aug 26 15:13:29 koen: yes Aug 26 15:13:43 tkp: install the 'real' mount from util-linux Aug 26 15:13:43 is that a problem? Aug 26 15:13:50 really? Aug 26 15:13:55 so this is a known issue? Aug 26 15:13:56 yes Aug 26 15:14:03 ahhh Aug 26 15:14:16 mount -a got broken Aug 26 15:14:21 so with that installed, /var will actually be a tmpfs then Aug 26 15:14:22 ahh Aug 26 15:14:23 cool... Aug 26 15:14:33 I've been puzzeling over this for days ow! Aug 26 15:14:52 I think pH5 or Xora provided that solution Aug 26 15:15:05 ok, I'll give it a go... Aug 26 15:16:40 could you file a bug about that? Aug 26 15:17:21 hi all ... can one of you give me a hint why I do not have libqte-mt - but only libqte in staging; thought that .dev is mt? Aug 26 15:17:31 sure Aug 26 15:17:32 what do I have to do to get libqte-mt as well? Aug 26 15:18:00 sell your soul to the devil^h^h^h trolltech Aug 26 15:20:46 keon: do you have some reservations against tolltech? Aug 26 15:21:47 bitbake -c clean task-stv Aug 26 15:21:55 oops... wrong window! Aug 26 15:39:05 koen; What do you want to tell us with http://bugs.openembedded.org/show_bug.cgi?id=1353#c4 ? Aug 26 15:40:14 Laibsch: read the mailinglist archives and irc logs Aug 26 15:40:49 Laibsch: I don't think people will feel inclined to repeat themselves for the fourth or fifth time Aug 26 15:48:50 koen: I do not have time to read ALL of the IRC logs. I guess "bluetooth openembedded" is not going to lead to results fast. Bug 1353 is the only bug report in the tracker about this. Aug 26 15:48:50 koen: It also seems not totally unfounded. As such, I can only see a totally unwarranted "plonk" with no further explanation as childish, unconstructive behaviour. Aug 26 15:48:53 koen: If you want the openzaurus-ja people to contribute this is NOT the way to go. Your comment might have already done GREAT damage. Aug 26 15:49:15 I have warned you guys several times to not act with Japanese people as you do among yourselves! Aug 26 15:50:00 s/"bluetooth/googling for "bluetooth/ Aug 26 15:50:08 OMG THERE IS STUFF IN THE ROOTFS LOLZORS Aug 26 15:50:14 I DONT WANT STUFF Aug 26 15:50:42 we've been over 'discussions' like that *over* and *over* again Aug 26 15:51:02 koen: You think you are funny? I think some of the discussion were also about the collie rootfs being crammed. Aug 26 15:51:22 and? Aug 26 15:51:22 As I said, I do not follow all discussions here or on the ML, but did not come across this. Aug 26 15:51:38 So, either do not comment or comment constructively. Aug 26 15:51:50 Is it so hard to be un-geeky, unarrogant? Aug 26 15:52:03 heh Aug 26 15:52:09 read what your saying Aug 26 15:52:16 and think about it for a moment Aug 26 15:53:08 OK Aug 26 15:53:10 Now what? Aug 26 15:53:49 stop being arrogant and insulting Aug 26 15:53:54 I did think about what I wrote. Did you about your stuff? Aug 26 15:54:00 I am being arrogant? Aug 26 15:54:08 then ask your distribution maintainer why he's shipping what's he's shipping Aug 26 15:54:56 This is not an #openzaurus issue. Aug 26 15:55:11 it is Aug 26 15:55:22 And I can only laugh at your suggestion I am being arrogant. Aug 26 15:55:35 Then go ahead and read #openzaurus topic. Aug 26 15:55:44 read the topic over here Aug 26 15:56:01 "this is still not a distribution support channel" Aug 26 15:57:53 This is not about distribution support. I know you (not you personally) always like to play this game about "we are not responsible, go elsewhere" if you run out of arguments or motivation for discussion (-> bug tracker, IRC). If you are serious about this, why did you comment here:(17:09:18) koen: ~lart people complaining about bluetooth ? Aug 26 15:57:58 This is pathetic! Aug 26 15:58:29 you are complaining that OZ ships 'too much' Aug 26 15:58:46 I am discussing Aug 26 15:58:48 how is that not OZ related? Aug 26 15:59:08 like I said: "then ask your distribution maintainer why he's shipping what's he's shipping" Aug 26 15:59:45 I can keep telling you what to do and you can keep insulting me if you want Aug 26 15:59:52 I ask YOU to either comment constructively or refrain from commenting. I am saying "please" and I am being friendly. The rest of it is being discussed in the BTS. Aug 26 16:00:21 I have NEVER even closely insulted you. I have been friendly unlike you. Plonk is not friendly. Aug 26 16:00:44 Especially towards somebody who I guess is a newbie in the OE BTS. Aug 26 16:00:49 I'm finding you very insulting Aug 26 16:01:13 OK, than I apologize for coming across like that. Aug 26 16:01:32 That is not my intention. I want to deescalate, not get ppl angry. Aug 26 16:01:57 Again, I kindly and politely ask you to change. Aug 26 16:07:26 Laibsch: my c700 has bluetooth Aug 26 16:07:42 via a cf bluetooth card Aug 26 16:08:28 following your bluetooth suggestion hrw should remove bt, wifi, usb, sd, cf and more Aug 26 16:08:52 "it doesn't come with an sd card, so why waste space with drivers for it?" Aug 26 16:09:00 (to use a slippery slope argument) Aug 26 16:09:10 koen: I think we have exchanged opinions. It would be nice if you can comment constructively in the future. Aug 26 16:09:31 I have commented constructively in such discussions in the past Aug 26 16:09:46 like I said, I don't feel inclined to repeat my self over and over again Aug 26 16:09:57 Then at least give an URL instead of *plonk* or don't comment. Please. Aug 26 16:10:29 Plonk is for stubborn, dumb ppl. Takeshi has neither proven to be stubborn nor dumb. Aug 26 16:10:43 Thus your comment is unwarranted. Aug 26 16:10:49 EOD for me. Aug 26 16:34:17 Laibsch: prefixing everything with 'oz-ja' seems stubborn and wastefull to me Aug 26 16:34:25 but I already made that clear Aug 26 16:35:42 sorry if this is a mistake on my part, but I'm having trouble running 'bitbake bootstrap-image'. When walking the dependencies, it starts to build hostap-modules-0.4.7. However that then encounters an error with a missing makefile. I'm not entirely sure what's going on, because I can see the makefile in the .tar.gz file. Does anyone know what might be wrong? Aug 26 16:36:53 koen: 1) I agree that is bad practice, see the comment I just made to bug 1362 about an hour ago. It is not stubborn since AFAIK they have not yet been told and have not practized this recently anyways. 2) If that is your motivation you plonked the wrong person. 3) you are becoming argumentative, trying to defend a lost position ;-) Aug 26 16:37:34 hi Aug 26 16:37:39 03rpurdie 07org.oe.dev * r63decb6a... 10/ (11 files in 5 dirs): qemu: Add arm nptl patch (as used in scratchbox) to fix EABI locale generation, add version 0.8.2, apply patches to some fixed verisons Aug 26 16:38:21 RP: yay! Aug 26 16:38:58 I installed gcc on zaurus with openzaurus 3.5.4 and I get this error Aug 26 16:38:59 libc.so: file format not recognized; treating as linker script Aug 26 16:39:16 For reference, testing qemu with EABI locale generation with an old CVS date worked fine. Testing with todays failed. Aug 26 16:39:42 03rpurdie 07org.oe.dev * r34964bec... 10/ (1 classes/base.bbclass): base.bbclass: Change FILESPATH handling to allow it to be appended/prepended to. Should be backwards compatible. Aug 26 16:40:39 koen|away: Maybe we can both cool down and get back to business ;-) Takeshi has also included some Japanese translation files for GPE which I believe is one of your pet projects. Can you maybe upload them and close bug 1356? That would be great and get us out of this unproductive mode ;-) Aug 26 16:46:31 Laibsch: translation files can only be uploaded by gpe people Aug 26 16:46:43 Laibsch: I have showed florian the bug a few days ago Aug 26 16:46:45 OK, I see. Aug 26 16:47:07 Laibsch: http://handhelds.org/hypermail/gpe/55/5536.html Aug 26 16:47:09 I just sent you a mail 5 seconds ago. Trying to calm things down. Aug 26 16:48:08 So I guess opening a bug on handhelds.org is not necessary. Aug 26 16:48:45 * koen curses linksys once more Aug 26 16:58:11 ~lart linksys Aug 26 16:58:19 doh! Aug 26 16:58:20 * ibot whips linksys with a wet and grimy noodle just because Aug 26 16:58:34 ~emulate stewie Aug 26 16:58:36 It's like an orgy in my mouth! Aug 26 17:03:28 * koen rebuils glibc to test qemu/eabi Aug 26 17:03:53 pH5: did you ever have problems with your ts on the hx? Aug 26 17:04:16 hmmm Aug 26 17:04:22 it seems evdev is busted Aug 26 17:07:47 koen: how? I works here with udev097 and 2.6.16-hh5 as of two weeks ago. Aug 26 17:08:27 pH5: all evdev things seem to be broken, so I think something went bad during compiling Aug 26 17:09:03 koen: with the new kernel? I yet have to try 2.6.17 Aug 26 17:09:07 did genirq get merged in 2.6.17 or 2.6.18? Aug 26 17:09:15 pH5: 2.6.16-hh5 Aug 26 17:11:17 koen: I think genirq is only 2.6.18, but I'm not sure. Aug 26 17:12:41 2.6.17 had some other bugfixes for ARM which I'd like to try Aug 26 17:12:50 2.6.17 is running nice on my c700 Aug 26 17:13:38 03koen 07org.oe.packaged-staging * r87498bc6... 10/ (1 docs/packaged-staging.xml): docs/packaged-staging.xml: clear up concerns about p-s + bitbake-mt Aug 26 17:20:26 Do_install is failing on util-linux-2.12r-r6 for an openslug-image build. It is trying to create a symbolic link that already exists (swapoff.util-linux) Aug 26 17:27:42 koen: genirq is 2.6.18 Aug 26 17:27:48 ok Aug 26 17:28:04 NOTE: preparing tree for binary locale generation Aug 26 17:28:04 NOTE: generating locale es_NI (UTF-8) Aug 26 17:28:05 NOTE: generating locale es_NI (ISO-8859-1) Aug 26 17:28:06 Who was that phantom answerer Aug 26 17:28:07 sweetness Aug 26 17:28:27 RP: localegen works now, thanks Aug 26 17:28:33 03koen 07org.oe.dev * r6a4cba9a... 10/ (3 files in 3 dirs): angstrom: enable binary locale generation and add visbility stuff to cxxflags Aug 26 17:29:01 and thanks to suihkulokki of course Aug 26 17:29:08 koen: I'm going to move the SRCDATE for qemu forwards a bit - my tests show it still works :) Aug 26 17:29:13 hi all Aug 26 17:29:24 RP: angstrom has it fixed at 0.8.2 Aug 26 17:29:43 koen: ok, you're fine then :) Aug 26 17:29:58 koen: The date corresponds to 0.8.2 effectively Aug 26 17:30:29 does anyone know a cas programm which i can use on a pda? is there a cas package in oe? Aug 26 17:30:30 RP: I broadened sane-srcdates a while back Aug 26 17:30:36 mr_nice_slacker: giac ? Aug 26 17:30:43 koen: do I have to change anything to try this, or will just rebuilding quemu and glibc be enough? Aug 26 17:31:01 koen: I noticed Aug 26 17:31:06 pH5: mtn up ; rebuild qemu-native, rebuild glibc Aug 26 17:31:20 * pH5 does Aug 26 17:31:23 qemu-native should rebuild due to PR/PV changes Aug 26 17:31:23 koen: thx i will try it Aug 26 17:33:33 RP: I'd like to make an asoc driver for the magician (uda1380 on i2s). is there any newbie-friendly documentation? Aug 26 17:33:39 03rpurdie 07org.oe.dev * r3e5c0728... 10/ (4 files in 4 dirs): qemu cvs: Update SRCDATE to 20060723 Aug 26 17:35:27 pH5: I believe Liam has added Documentation to the kernel which should be included in the patch Aug 26 17:36:11 pH5: New ASoC drivers would be very welcome - ask me/Liam (lrg) if you have any questions Aug 26 17:38:46 pH5: iirc h2200 uses uda1380 as well Aug 26 17:38:57 pH5: liam had plans to write a driver for that Aug 26 17:40:17 A uda1380 driver exists but I'm not sure its been tested on any hardware so far Aug 26 17:42:08 03rpurdie 07org.oe.dev * r73cf0a31... 10/ (3 files in 3 dirs): syslog-ng: Add modified syslog-ng.conf which enables kernel logging by default Aug 26 17:46:11 RP, koen: thanks. Aug 26 17:50:30 03rpurdie 07org.oe.dev * r5b128406... 10/ (4 files in 3 dirs): patch: Add patches to enable unifed reject files and a global reject file (from debian) Aug 26 17:53:43 grrrr Aug 26 17:53:44 LC_ADDRESS: field `lang_ab' must not be defined Aug 26 18:05:34 does collie kernel 2.6 support the SD/MMC yet? Aug 26 18:05:41 hm, qemu-native still doesn't like gcc4? Aug 26 18:05:59 Zero_Chaos, I wish. Aug 26 18:06:05 thought so... Aug 26 18:06:07 thanks Aug 26 18:06:49 pH5: it doesn't Aug 26 18:06:57 pH5: I switched back to gcc3 on my host Aug 26 18:07:13 pH5: but interestingly it does compile with gcc4.1.1/eabi Aug 26 18:08:25 heh Aug 26 18:08:36 F 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.4.1, dynamically linked (uses shared libs), for GNU/Linux 2.4.1, not stripped Aug 26 18:09:14 koen: I'm seeing that lang_ab error too :-/ Aug 26 18:11:08 koen: I think we need to mark gez_ER and bn_DB as broken Aug 26 18:11:18 yeah Aug 26 18:11:30 screw the ethiopians Aug 26 18:14:27 RP: and gez_ET by the looks of it Aug 26 18:14:55 koen: I meant gez_ET and bn_BD :-/ Aug 26 18:15:13 gez_E{R,T} both have lang_ab Aug 26 18:15:36 I didn't see an issue with ER - I think some just have typos in Aug 26 18:15:47 google gives some info about debian patches to fix them Aug 26 18:16:25 -GLIBC_BROKEN_LOCALES = "sid_ET tr_TR mn_MN" Aug 26 18:16:26 +GLIBC_BROKEN_LOCALES = "sid_ET tr_TR mn_MN gez_ET bn_DB" Aug 26 18:16:33 like that? Aug 26 18:16:46 s/DB/BD/ Aug 26 18:17:03 We could fix up the locale files ;-) Aug 26 18:17:16 but marking them as broken is probably easier Aug 26 18:17:34 more patches to glibc? Aug 26 18:17:37 * koen runs Aug 26 18:17:49 ours is already patched to the max Aug 26 18:18:03 koen: remember you're talking to the oz kernel maintainer ;-) Aug 26 18:18:10 ah, right Aug 26 18:18:21 you can always add more patches :) Aug 26 18:19:24 I remember a discussion on L-A-K about sane bootloaders Aug 26 18:19:36 What about it? Aug 26 18:19:37 kergoth suggested that people could apply a single patch locally Aug 26 18:19:54 beewoolie said kergoth was too used to OE :) Aug 26 18:20:08 that and quilt Aug 26 18:20:33 You'll note qemu is starting to get a bit patched with more probably coming soon :) Aug 26 18:22:04 :) Aug 26 18:22:33 that wouldn't have anything todo with scratchbox using OEtoolchains, wouldn't it? Aug 26 18:23:12 Not yet although that was noted and is on my todo list Aug 26 18:23:21 or perhaps kergoth's :) Aug 26 18:23:28 :) Aug 26 18:24:17 Has anyone any notes on using distcc on a Z to use the processing power on their desktop btw? Aug 26 18:24:29 you might have noticed the abundance of "chroot" in the socdoc Aug 26 18:24:40 RP: the pdax people had some notes Aug 26 18:24:53 RP: basically you give the toolchain a weirdname Aug 26 18:25:13 and install the cross version with a similar weirdname on your desktop Aug 26 18:25:38 in OE speak "don't use gcc-symlinks" Aug 26 18:25:48 ah, right :) Aug 26 18:25:58 you had me worried for a minute :) Aug 26 18:26:29 the distcc in OE has the avahi patch applied Aug 26 18:26:31 I'm trying to do 'bitbake perl', but I get an error during the do_configure - 'sed couldn't file "Binary"'. It seems to be grepping through the whole perl source tree for a pattern, and doing a sed on any file that matches - one file comes up and grep says "Binary file matches", and so the for loop thinks there are 3 files, Binary, file and matches to pass to sed. Any idea what I can do to fix this? Aug 26 18:26:47 RP: I'd love a 'bitbake cluster-image' :) Aug 26 18:27:00 koen: As would I ;-) Aug 26 18:27:26 * koen has enough arm toys to power a small country Aug 26 18:27:59 Its worrying when you have more cpus that fingers within arms reach :-/ Aug 26 18:28:20 * koen moves to the couch Aug 26 18:28:28 worrying over :) Aug 26 18:30:13 RP: Somebody suggested that icecc was better suited. Just inherit icecc Aug 26 18:30:28 Laibsch: I want to do the opposite to that ;-) Aug 26 18:30:36 * RP -> food - back shortly Aug 26 18:31:01 icecc will distribute the toolchain, which would break for heterogenous hosts Aug 26 18:31:43 we want to exploit the dumb behaviour of distcc :) Aug 26 18:31:59 RP: I think it is you who is introducing multithreading to OE, aren't you? Maybe you can use an openmosix cluster. Aug 26 18:32:21 koen: OK, I guess I am not quite following what you guys are trying to do. Aug 26 18:33:02 Compile on the Z for ix86? Reverse cross-compile? Aug 26 18:33:05 compiling on your favourite toy with the help of your desktop Aug 26 18:33:18 no, native compile with some aid Aug 26 18:33:21 target is arm or ix86? Aug 26 18:33:28 OK Aug 26 18:33:28 arm Aug 26 18:47:01 generating locales takes quite some time Aug 26 18:47:12 * koen should lobby for a faster buildmachine Aug 26 18:47:24 zecke: moin Aug 26 18:47:54 morsche Aug 26 18:49:46 Laibsch: I am responsible for bitbake multithreading, yes. It doesn't scale from one machine yet (but at least its more possible now) Aug 26 18:49:48 hi zecke Aug 26 18:50:14 RP: make bitbake a plan9 app :) Aug 26 18:51:00 Ojective-C Aug 26 18:51:29 RP: could you commit the BROKEN_LOCALE change, I have a ton of local diffs in that file :( Aug 26 18:51:43 I feel like puking... Aug 26 18:55:50 koen: Ok, once I confirm we have all the broken ones Aug 26 18:56:17 good idea Aug 26 18:59:22 Is there a reason that bitbake is always recompiling glibc? AFAICT it seems to be the same version and release so I do not understand why it is being rebuilt. Aug 26 18:59:48 it's only compiles glibc once over here Aug 26 19:00:03 hm... Aug 26 19:00:50 RP: spi-inc.org works now Aug 26 19:01:38 Laibsch: i've seen that Aug 26 19:01:50 the other day every time i ran bitbake it decided to rebuild glibc and gtk Aug 26 19:01:58 doesnt seem to happen now though, so *shrug* Aug 26 19:02:09 kergoth`zzz: Are you compiling from different machines? Aug 26 19:02:48 did all my builds on this box Aug 26 19:03:10 I was wondering if maybe the reason is that one computer is a Pentium III laptop with 686 kernel, the other has a Via C3 CPU and 386 kernel. Aug 26 19:03:19 kergoth: OK. Aug 26 19:03:32 I hope it goes away soon for me, too. ;-) Aug 26 19:04:30 hehe Aug 26 19:06:04 RP: so far no new broken locales Aug 26 19:06:16 * koen watches some more locales scroll by Aug 26 19:06:18 *slowly* Aug 26 19:15:07 koen: same here... Aug 26 19:16:13 Laibsch, if you have floating CVS dates it will recompile everyday Aug 26 19:17:12 emte: The CVS date is from 2005. Does that matter? I already compiled today. I do not know what triggers this. Aug 26 19:19:43 People have bumped PR on glibc recently Aug 26 19:22:13 RP: te_IN Aug 26 19:27:43 gd evening Aug 26 19:28:29 hey greentux Aug 26 19:29:13 somebody read about the trolltech greenphone? Aug 26 19:29:46 is that the one that only makes calls with gcc 2.95.3? Aug 26 19:30:21 i dont know, i even dont know about avaibility and so Aug 26 19:30:31 i heard something about kernel 2.4 :( Aug 26 19:33:44 oh, baby is calling... mom Aug 26 19:39:13 * koen wonders in what order the locales are getting processed Aug 26 19:46:45 I get a message about CROSS COMPILE Badness when compiling frozen-bubble. As far as I can tell, that's a problem with the package, is that right? Aug 26 19:56:27 baldurk: yes Aug 26 20:18:03 time to implement ALC/LCT Aug 26 20:18:28 is that a good thing? Aug 26 20:18:45 Kerwood: something I wanted to do for some time Aug 26 20:18:56 Asynchronous Layered Coding (ALC) Aug 26 20:19:02 whew! Aug 26 20:19:18 RFCs 3450, 3926 and two more... Aug 26 20:21:44 wow, finally some net transport types talking about using FEC for once, instead of relying on timeouts and resends. 9_9 Aug 26 20:22:20 Kerwood: well, I have not yet taken the exam in the linear algebra class... so my understanding on FECs is bad... Aug 26 20:22:33 BUT: e.g. we use MPE and MPE-FEC to transport IP streams Aug 26 20:22:47 and if a kernel passes UDP packets to userspace the checksum must be correct Aug 26 20:23:11 and I wonder how a FEC can help there... only with some simple bit flips... Aug 26 20:23:25 but I wonder why the Reed Solomon should have failed (MPE-FEC) Aug 26 20:26:01 zecke: Well, from what I concluded from an initial scan of the RFCs, they're trying to create a transport protocol which accommodates different content, but in the same stream Aug 26 20:26:38 Kerwood: the aim is to create a reliable uni-directional 'file' delivery Aug 26 20:27:24 its just a digital differential singaling protocol is it not? Aug 26 20:27:39 zecke: my read is that they aim to do that in a way which doesn't fail if the content is time-sensitive Aug 26 20:28:21 zecke: and that the delivery is asymmetric i.e., multicast Aug 26 20:28:34 Kerwood: the way it is used is for broadcast networks like DVB Aug 26 20:28:41 ah Aug 26 20:28:42 Kerwood: timely manner is just another nice thing Aug 26 20:29:42 at any rate, it's about damn time Aug 26 20:29:42 TCP/IP was never "streamlined" and was never meant to be Aug 26 20:29:42 but we can't keep relying on faster and faster pipes... Aug 26 20:31:57 Forward Error Correction? Aug 26 20:32:09 koen: yes Aug 26 20:32:28 zecke, is updating bitbake a good idea today :) ? been getting some weird errors Aug 26 20:32:56 zecke: I still have to pass ControlEngineering :( Aug 26 20:34:11 Kristoffer: I need to admit I'm drowned in real-life+work I don't know Aug 26 20:34:26 Kristoffer: using BitBake 1.6.x and BitBake 1.4.4 is the most safe thing you can do Aug 26 20:34:33 Oki, will update to that then Aug 26 20:34:44 Kristoffer: for BitBake trunk, ask our MVP RP about the state Aug 26 20:34:58 trunk works pretty well Aug 26 20:35:24 but, as always, caveat emptor Aug 26 20:38:00 * koen remembers his spare OSX tree Aug 26 20:42:53 03koen 07org.oe.dev * rbf8790b0... 10/ (1 packages/glibc/glibc_2.4.bb): glibc 2.4: mask out more broken locales Aug 26 20:52:47 Is there a problem with the SD driver in oz-3.5.4.2? Or perhaps I borked my build Aug 26 20:53:13 Didn't see any bugs listed... Aug 26 21:02:48 koen: Did your build complete? Aug 26 21:02:55 * RP is still waiting :-/ Aug 26 21:02:56 RP: yes Aug 26 21:03:50 RP: the next commit should interest you as well Aug 26 21:03:52 Kristoffer: trunk is unstable. I'm happy enough to have it tested but if it breaks, I can only deal with it when I have time ;-) Aug 26 21:05:14 koen: I also had issues with gez_ER Aug 26 21:05:26 crap Aug 26 21:05:30 I edited that one Aug 26 21:06:01 koen: For some reason it did them in a different order for me - I came accross it last :-/ Aug 26 21:06:38 03koen 07org.oe.dev * re756d424... 10/ (3 files in 2 dirs): qemu 0.8.2: use the cross compiler to build 'qemu' Aug 26 21:07:14 koen: hmm :) Aug 26 21:07:35 qemu doesn't use autotools.... Aug 26 21:07:47 but it does do ';inherit autotools' Aug 26 21:07:50 koen: you want to check that commit - I think it has a typo Aug 26 21:08:29 the qemu one? Aug 26 21:08:42 koen: yes --cc$CC} Aug 26 21:09:18 RP: OE doesn't mind that ;) Aug 26 21:09:38 koen: It should do though ;-) Aug 26 21:11:18 03koen 07org.oe.dev * rc8c8e34a... 10/ (1 packages/glibc/glibc_2.4.bb): glibc 2.4: also mask out gez_ER Aug 26 21:21:37 03koen 07org.oe.dev * r23f08e86... 10/ (1 packages/glibc/glibc_2.4.bb): glibc 2.4: remove spurious committed Werror bit Aug 26 21:24:05 03koen 07org.oe.dev * r79739dff... 10/ (1 packages/qemu/qemu_0.8.2.bb): qemu: fix typo Aug 26 21:28:03 there, that should have fixed all my damage Aug 26 21:28:44 koen: it happens :) Aug 26 21:30:00 ~lart tons of local diffs Aug 26 21:30:01 * ibot declares tons of local diffs a moron Aug 26 21:32:12 Its the glibc locale limiting patches that I worry I'll check in at some point... Aug 26 21:35:33 * koen snuck in a DEFAULT_PREFERENCE removal Aug 26 21:36:35 * RP noticed Aug 26 22:07:55 rwhitby: hi! I haven't had time yet to answer you about yeaphone... Aug 26 22:08:17 rwhitby: still getting the HEAD environment up and running Aug 26 22:08:57 treitmayr: great - need any help? Aug 26 22:09:45 rwhitby: hmm, do I understand this corretly - the nslu2-linux repository just directly links to the oe repository? Aug 26 22:10:53 yes, monotone.nslu2-linux.org is just a monotone peer to monotone.openembedded.org, which syncs regularly. Aug 26 22:11:37 we do have other branches on m.n-l.o, but the org.openembedded.dev branch is identical (modulo changes since last sync) Aug 26 22:12:04 ok. Aug 26 22:12:53 So I'm pretty much done with linphone/yeaphone (+fixing a few memory leaks). I could either send it all to you or commit it myself Aug 26 22:14:16 I just cannot get (read) access to the nslu2-linux repository... Aug 26 22:25:25 treitmayr: what is the error you get? Aug 26 22:26:40 "mtn: network error: failed to connect: Connection refused" - did I something wrong? It worked a few weeks ago... Aug 26 22:27:05 happens eg. at "pull" Aug 26 22:27:37 treitmayr: is there another error as well? Aug 26 22:28:04 treitmayr: I get a connection refused as well Aug 26 22:28:25 treitmayr: you could pull from monotone.openembedded.org? but better wait for rwhitby to answer Aug 26 22:28:41 ok, I'll give it a try... Aug 26 22:29:04 just to make sure my network isn't screwy Aug 26 22:44:35 treitmayr: try now - the server was stuck Aug 26 22:45:10 ok, works! Thanks! Aug 26 22:48:43 How are input methods being dealt with on GPE? Is there no such thing? Aug 26 22:49:17 Or is it keyboard only? Aug 26 22:49:31 you mean aside from tslib? Aug 26 22:53:36 Laibsch: once uim+gtk or similar work, you can use the keyboard + gtk input method. onscreen keyboard/HWR/whatever can be started by a matchbox panel applet. Aug 26 22:54:34 I see. oz-ja ppl have uim working but I cannot get it compiled with "regular" OE. Aug 26 22:54:51 I guess they also work on the opie image but I am not sure about that. Aug 26 23:15:06 Laibsch: NOTE: package uim-1.1.0: completed Aug 26 23:15:28 the packaging has to be fixed up a little bit, but the build works without modification Aug 26 23:42:21 pH5: Right, it was anthy for uim that IIRC failed. Aug 26 23:42:52 And after installation on my opie image, nothing showed up in the input method selector. Aug 26 23:48:13 Can I change the machine to build for without removing everything in tmp/ ? Or do I need to clean that out? Aug 27 00:33:25 it will overwrite it all Aug 27 00:33:31 afaik Aug 27 00:33:56 unless you have multi-machine setup Aug 27 02:32:16 RP: ping Aug 27 02:32:53 ginp :PR Aug 27 02:33:29 ~herring emte Aug 27 02:33:31 * ibot whacks emte on the side of the head with a large red herring named alfred Aug 27 02:33:32 emte: how goes it? Aug 27 02:33:47 * emte eats alfred Aug 27 02:33:56 mm alright Aug 27 02:34:06 trying to figure out how to use opto-couplers Aug 27 02:34:12 u? **** ENDING LOGGING AT Sun Aug 27 02:59:57 2006