**** BEGIN LOGGING AT Sat Apr 22 09:59:56 2006 Apr 22 10:00:26 hehe Apr 22 10:00:37 better than reading my mind? Apr 22 10:10:21 nOOb question: how to make OE build an arm cross-toolchain ? I don't have personnal or pre-built cross-toolchain, just did 'bitbake task-boostrap' but it fails saying 'arm-linux-gcc-2.95: No such file or directory' ? Apr 22 10:11:16 good morning zecke, mickeyl Apr 22 10:13:36 cyrilRomain: You need to read the documentation on Zaurus Kernels - old 2.4 ones need a 2.95.3 gcc installed Apr 22 10:14:31 I know I could use a pre-built cross-toochain. Does it means I cannot use OE to build it ? Apr 22 10:15:43 cyrilRomain: Not the 2.95.3 one Apr 22 10:16:00 RP: ok, thanks Apr 22 10:17:47 RP: To build a 2.6 kernel, there is not so much information. Where to configure to use 2.6 kernel ? Do I also need a pre-built toolchain OE cannot build ? Apr 22 10:18:36 cyrilRomain: For 2.6 kernels, OE will use the toolchain its already built Apr 22 10:20:06 cyrilRomain: Which machine are you building for? Apr 22 10:20:25 RP: ok, I see. I'm trying to build 'collie' openzaurus, but I don't know where can I tell OE to build a 2.6 kernel (sorry for thoses noob questions, but I promisse to update the wiki in return) Apr 22 10:21:02 cyrilRomain: 2.6 support for collie is extremely limited :-( Apr 22 10:21:43 cyrilRomain: You'd set ZKERNEL_VERSION = "2.6" in your local.conf (as documented in the sample local.conf) but I'll warn you its not useable Apr 22 10:23:22 RP: is there a better 2.6 support for toza ? Apr 22 10:24:00 RP: or for armqemu ? Apr 22 10:25:41 cyrilRomain: The best 2.6 zaurus support is for the cxx00 and c7x0 devices. Tosa support is in the beta stage. qemuarm is a pseudo machine - a virtual one designed to run under the qemu emulator Apr 22 10:26:22 RP: ok thank you very much :) Apr 22 10:51:09 <_law_> hi all Apr 22 10:53:52 hi law Apr 22 11:00:08 do we have any sane solution for the sdl packages? the ones destined for opie should include a .desktop file at the appropriate location (/opt/QtPalmtop/foo/bar), but as far as I can see, this is the only real difference left. should we ship 'em with the non-opie ones or should we rather (automatically, through sdl.bbclass) add a -opie that DEPENDS on and includes the necessary desktop file? Apr 22 11:03:33 hey Apr 22 11:04:09 think it's posible to run the pda build stuff for x86 Apr 22 11:04:21 and if so how would a system with 8 meg of ram fair, mabye 12 Apr 22 11:04:48 I have a problem with bb: it tries to fetch http://www.oesources.org/source/current/quilt_cvs.savannah.nongnu.org_VER_0_39_.tar.gz while the correct name is quilt_savannah.nongnu.org_VER_0_39_.tar.gz How can I fix this please ? Apr 22 11:05:40 AlexTe: ? Apr 22 11:05:40 AlexTe: cvs gets fetched afterwards Apr 22 11:07:10 well I dont know actually it's the 1st time I use bb. I followed these instructions http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted Apr 22 11:07:27 HTTP request sent, awaiting response... 404 Not Found Apr 22 11:07:40 AlexTe: and? Apr 22 11:07:56 AlexTe: oesources.org is a service, it has not all files one ever wants Apr 22 11:08:11 AlexTe: 404 -> cvs co Apr 22 11:08:11 mickeyl: the latter would be fine, though if the files are small it might be easiest just to ship them unconditionally. I don't think it would be fatal for non opie users to have a few extraneous .desktop files on their machines. Apr 22 11:08:32 NOTE: Fetch cvs://anonymous@cvs.savannah.nongnu.org/cvsroot/quilt;method=pserver;module=quilt;tag=VER_0_39 Apr 22 11:08:32 NOTE: Task failed: Fetch failed: quilt Apr 22 11:08:34 pb_: changed your mind? It was fatal last time we discussed it Apr 22 11:08:41 mickeyl: or, of course, you could teach opie to read the .desktop files from /usr/share/applications Apr 22 11:08:41 cvs seems to fail too Apr 22 11:08:50 AlexTe: well don't sadow your error then? Apr 22 11:08:57 zecke: really? I don't remember ever having thought that, though I could just have forgotten Apr 22 11:08:57 AlexTe: which version of OE and BitBake? Apr 22 11:09:16 pb_: you know a /opt/Qtopia/apps/Games directory gets created ;) Apr 22 11:09:46 Anyone got any objections to me splitting badblocks into a separate package from e2fsprogs and adding a RDEPEND? Apr 22 11:10:17 pb_: yeah, but that would require a substantial change and I'm not prepared to touch opie at this time Apr 22 11:10:23 zecke: yeah. I'm certainly not thrilled about that prospect, but neither am I so worried about it that it seems to require heroic efforts to avoid. Apr 22 11:10:37 mickeyl: right, fair enough Apr 22 11:10:40 bb 1.5.0 and I think I have yesterday OE.db snapshot Apr 22 11:10:56 AlexTe: why do you think a unstable branch is what you want? Apr 22 11:11:05 AlexTe: http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted <- ;) Apr 22 11:13:12 well I'm trying to build something for oz 3.5.4.1 so which branch do you suggest to start with ? Apr 22 11:13:42 am I right that to assume that you guys eliminated the necessity to -Dmain=SDL_main ? Apr 22 11:14:07 AlexTe: not BitBake unstable Apr 22 11:14:21 AlexTe: use the BitBake stable branch, as the GettingStarted page now says Apr 22 11:14:31 mickeyl: right, if you have a gcc > 3.3 Apr 22 11:14:42 zecke: excellent. i'll just assume that for .dev then Apr 22 11:15:25 ok I'll try with 1.4 thanks ! Apr 22 11:16:07 AlexTe: :) Apr 22 11:16:48 AlexTe: the wiki was updated this morning to point to bitbake-1.4 Apr 22 11:18:30 BitBake 1.5 now prints a warning when used :} Apr 22 11:22:33 03nail 07org.oe.dev * r8205afa2... 10/packages/e2fsprogs/e2fsprogs_1.38.bb: e2fsprogs: Split badblocks into separate package and add RDEPENDS Apr 22 11:25:16 03nail 07org.oe.dev * r492a2865... 10/conf/distro/ (debianslug.conf openslug.conf): Openslug/Debianslug: Add e2fsprogs-badblocks to rootfs Apr 22 11:26:57 It works thanks guys! Apr 22 11:27:03 AlexTe: you are welcome :) Apr 22 11:27:05 AlexTe: get back to us Apr 22 11:27:54 mickeyl, email? Apr 22 11:29:12 Pendalar: you have access to the OpenZaurus project now. That should give you all permissions you need to rsync/etc. Apr 22 11:29:41 pb_,zecke: well... i go for the clean separation Apr 22 11:29:46 * mickeyl writes an appropriate sdl bbclass Apr 22 11:29:55 * reenoo yawns Apr 22 11:29:58 morning Apr 22 11:29:59 yo reenoo Apr 22 11:30:19 is it same paths as before? Apr 22 11:30:26 hey mickeyl Apr 22 11:30:27 Pendalar: iirc yes. Apr 22 11:30:33 home/o/op/openzaurus or something Apr 22 11:30:36 home/projects/o/op/openzaurus or so Apr 22 11:30:40 ya Apr 22 11:30:40 without projects Apr 22 11:30:49 hmm... sdl bbclass? what's that going to do? Apr 22 11:31:00 ok I'll begin looking into fixing it today Apr 22 11:31:19 reenoo: that will automatically add a -opie package that ships the opie specific stuff and rdepends on Apr 22 11:31:46 mickeyl: okay, very good Apr 22 11:31:47 hi REdOG Apr 22 11:31:51 er, hi reenoo Apr 22 11:31:58 mickeyl: ah, nice Apr 22 11:32:28 mickeyl: speaking of sdl... is libsdl-mixer intended to link against libvorbis+libogg? Apr 22 11:33:09 i don't think so or else i would have listed it in DEPENDS Apr 22 11:33:16 does it pick up those? Apr 22 11:33:22 yeah Apr 22 11:33:30 file it into the bugtracker and I'll take care of it eventually Apr 22 11:37:40 is anyone running the latest bitbake (rev 466) on FC5. I'm running into some python problems - http://pastebin.com/674998 Apr 22 11:38:01 use the 1.4 branch, liam Apr 22 11:38:08 head is considered unstable now Apr 22 11:38:30 (see channel topic :) Apr 22 11:38:31 mickeyl, cheers - I've not been here for about 2 weeks and it shows. Apr 22 11:38:35 np Apr 22 11:38:58 lrg: that is a weird error Apr 22 11:39:21 svn co svn://svn.berlios.de/bitbake/branches/bitbake-1.4/ bitbake/ Apr 22 11:39:34 mickeyl: one bitbake too many Apr 22 11:39:38 zecke: yep, I'm no python expert though - I figured it was some FC5 issue... Apr 22 11:39:48 lrg: I think you will get the same error with bitbake-1.4 Apr 22 11:39:53 zecke: sure? that was copied from gettingstarted Apr 22 11:39:57 hi Apr 22 11:40:01 hey hrw Apr 22 11:40:03 mickeyl: http://pastebin.com/675003 Apr 22 11:40:04 * lrg tries 1.4 Apr 22 11:40:06 hrw: openzaurus.berlios.de is broken Apr 22 11:40:08 I need help with serial cable Apr 22 11:40:11 hey Marcin Apr 22 11:40:23 mickeyl: its lardman and bronson toy not mine Apr 22 11:40:30 mickeyl: this would mean self.cachedir is None (for lrg) Apr 22 11:40:32 reenoo: excellent Apr 22 11:40:57 I connected serial to zaurus (which send kernel logs to serial) and how to get it on desktop? minicom or other tool? Apr 22 11:41:50 hrw: once the stty params are ok, a cat should be sufficient for display only Apr 22 11:41:50 * zecke heads to the britspot festival soon Apr 22 11:42:07 hrw: i use securecrt for that (windows *cough*) Apr 22 11:43:16 hi hrw Apr 22 11:43:28 mickeyl: ok. will look Apr 22 11:43:40 and my debian isntall looks a bit fscked here ;( Apr 22 11:43:41 zecke: there is not one bitbake too many, because svn will otherwise create the 'bitbake-1.4' output directory instead of just 'bitbake' Apr 22 11:44:06 cyrilRomain: right, this looked like /bitbake-1.4/bitbake in mickeys post Apr 22 11:44:25 zecke: I'm getting the same error with 1.4 :( Apr 22 11:44:53 lrg: okay, help me to help you then ;) Apr 22 11:45:00 reenoo: do you have plans to sync .oz354fam083 changes to .dev? Apr 22 11:45:03 hey pb_ (sorry, got distracted by the sdl stuff) Apr 22 11:45:14 lrg: start with putting the company credit card in the slot ;) Apr 22 11:45:28 zecke: heh ;) Apr 22 11:45:32 lrg: /home/liam/oe/bitbake/lib/bb/cache.py line 52 Apr 22 11:45:52 hrw: no concrete plans, no Apr 22 11:46:27 zecke, opened in vi Apr 22 11:46:48 lrg: add a print self.cachedir Apr 22 11:46:54 lrg: before the os.path.join line Apr 22 11:48:18 03rw 07org.oe.oz354fam083 * r60713f19... 10/ (14 files in 9 dirs): libsdl-mixer: explicitely disable ogg/vorbis support (libraries not in DEPENDS)packages/libsdl/libsdl-mixer_1.2.6.bb Apr 22 11:48:28 zecke, None Apr 22 11:48:59 * pb_ going to the office to reboot broken servers Apr 22 11:49:01 bbiab Apr 22 11:49:17 lrg: do you have a CACHE line in conf/local.conf? Apr 22 11:50:13 zecke, no Apr 22 11:50:25 lrg: feel free to file a bug Apr 22 11:50:31 lrg: and add a CACHE line Apr 22 11:50:54 zecke, will do. Shall I add a line like CACHE=tmp/cache ? Apr 22 11:51:10 lrg: I have CACHE = "${TMPDIR}/cache" or such Apr 22 11:51:23 zecke, ok, will try Apr 22 11:51:35 lrg: hmm Apr 22 11:51:44 lrg: your bitbake.conf should have that line :} Apr 22 11:53:03 lrg: is your BBPATH set right? Apr 22 11:53:15 zecke, the bitbake version I have from HEAD and 1.4 does not have this line Apr 22 11:53:37 zecke, I pretty sure - It worked before I went to CELF Apr 22 11:53:42 lrg: it should be in org.oe.dev/conf/bitbake.conf Apr 22 11:54:00 zecke, ah Apr 22 11:54:08 * lrg checks Apr 22 11:54:35 03rw 07org.oe.oz354fam083 * r9ee63270... 10/ (14 files in 9 dirs): disapproval of revision '60713f19e6d894061ca4d58a2a86420c17a0bd72' Apr 22 11:55:00 zecke, yes got it. Apr 22 11:55:18 lrg: echo $BBPATH Apr 22 11:55:57 hi Apr 22 11:56:04 reenoo: i hope "no concrete" plans is the answer because you want to focus on getting a release out. It'd be a bit sad if you would no longer care about .dev - essentially we had a fork then. Apr 22 11:56:11 s/had/would have/ Apr 22 11:56:19 zecke: /home/liam/oe/build:/home/liam/org.oe.dev Apr 22 11:56:26 zecke, spotted it Apr 22 11:56:31 mickeyl: we have a fork already. two of them even Apr 22 11:56:32 BBPATH is wrong Apr 22 11:56:49 reenoo: no we don't. there are differences between a fork and branches Apr 22 11:56:50 I can't remember changing this Apr 22 11:57:14 zecke, I'll fix this path and try again. Apr 22 11:57:31 lrg: anyway you have found a bug ;) Apr 22 11:57:52 zecke, Thanks for your help :) Apr 22 11:58:01 mickeyl: well. in SCM terms branches are something you can merge stuff back and forth between. monotone doesn't appear to support that Apr 22 11:58:33 zecke, I'm being called away atm , I'll try again in about 30mins and file the bug. Apr 22 11:59:38 reenoo: right, but please don't use the SCM incapability as an excuse. it looks like others are managing to keep divergence at a minimum. Apr 22 12:00:09 * zecke goes to www.britspot.de now Apr 22 12:00:13 have fun zec Apr 22 12:00:14 ke Apr 22 12:00:34 mickeyl: I'm not using it as an excuse. it's the exact reason why my stuff is in .oz354fam083 only Apr 22 12:01:17 mickeyl: doing that type of thing manually is doing the same work twice basically which is not going to happen with my schedule in mind Apr 22 12:02:31 i see. can we expect you to push things to .dev when the release pressure is over or do you plan to keep the fork longer? Apr 22 12:03:06 hi Apr 22 12:05:13 hi Bernardo Apr 22 12:06:02 if I connect nullmodem cable to both serial ports in pc and run two serial terminals then input from first should appear in second and viceversa - right? Apr 22 12:07:23 yes Apr 22 12:07:38 now I have situation when if I send to /dev/tts/1 then it is shown on /dev/tts/0 but not vice versa... broken cable? Apr 22 12:07:48 if it is nullmodem, if (using hw handshake) has the full complement of lines Apr 22 12:08:02 or one expecting a different handshake Apr 22 12:08:13 are both expecting xon/xoff? Apr 22 12:08:15 Bernardo: both tools configured in same way Apr 22 12:08:21 what if you turn the cable around? Apr 22 12:08:36 Bernardo: currently flow control off totally - will switch to rts Apr 22 12:08:47 try turning the cable around first Apr 22 12:08:51 ok Apr 22 12:09:09 should show on tts/1 but not on tts/0 if it is really broken Apr 22 12:10:53 Bernardo: same situation when cable exchanged - gtkterm1 (tts1) send to gtkterm2 (tts0) Apr 22 12:11:46 nevermind which flow control used Apr 22 12:12:20 looks like broken pc then... what if you try just a cat from the command line Apr 22 12:12:24 ? Apr 22 12:13:57 Bernardo: cat>/dev/tts/1 works Apr 22 12:14:11 Bernardo: output shown in tts/0 Apr 22 12:14:27 and cat > /dev/tts/0 ? Apr 22 12:15:38 nothing ;( Apr 22 12:16:11 so, either one of the ports has a different config, or the pc hw is broken Apr 22 12:16:44 not a stray getty on one of those terminals? Apr 22 12:17:58 could be Apr 22 12:18:04 no getty on serial Apr 22 12:18:38 XorA|gone: I noticed some activity on the abi-plugin bugreport :) Apr 22 12:20:21 hm.. now which one is which.. Apr 22 12:21:12 koen|work: yeah, someone has read it :-) Apr 22 12:21:32 03rw 07org.oe.oz354fam083 * r2f663806... 10/packages/libsdl/libsdl-mixer_1.2.6.bb: libsdl-mixer: explicitely disable ogg/vorbis support (libraries not in DEPENDS) Apr 22 12:23:59 * koen|work creates an account at the abi bugzilla Apr 22 12:24:26 * koen|work gets back to work Apr 22 12:25:36 koen|work: bugzilla really needs something like openid :-) Apr 22 12:27:14 * mickeyl goes to attend a marriage Apr 22 12:27:19 mickeyl: enjoy Apr 22 12:27:20 ~curse sharp for tcx3106 Apr 22 12:27:22 thanks Apr 22 12:27:27 May the fleas of a thousand camels infest your most sensitive regions, sharp for tcx3106 ! Apr 22 12:28:06 decided to ignore fact that homemade serial cable for Z exist on my desk Apr 22 12:28:31 what's the svn url for the stable bitbake branch again? Apr 22 12:28:57 rwhitby: svn co svn://svn.berlios.de/bitbake/branches/bitbake-1.4 bitbake :) Apr 22 12:29:46 cyrilRomain: thx - I'm updating the nslu2-linux MasterMakefile. Apr 22 12:41:34 I have to go - cu Apr 22 12:49:00 Hmm - we still get the depends problems even with bitbake 1.4 :-( Apr 22 12:49:19 what are the "depends problems"? Apr 22 12:49:38 ERROR: Nothing provides runtime dependency python Apr 22 12:49:38 NOTE: Apr 22 12:49:38 NOTE: no buildable providers for ipkg-utils Apr 22 12:50:12 Used to work. Then sometime after the big RDEPENDS changes a number of weeks or months ago it stopped working. Apr 22 12:50:30 koen|work referred to it as the bitbake heisenbug Apr 22 12:50:50 yes, we've tried a complete rebuild after blowing away tmp Apr 22 12:53:06 I'll try blowing away tmp/cache since we've just moved to the stable bitbake 1.4 Apr 22 12:54:38 I can provide a bitbake -v ipkg-utils pastebin if that would help Apr 22 12:55:49 http://pastebin.ca/50550 Apr 22 12:58:46 any insight would be appreciated. this is blocking an nslu2-linux release cause we can't build the package feeds due to this problem. Apr 22 12:59:17 rwhitby: I've just blown away my tmp for oz354x because of a similar problem with libgpg-error0 Apr 22 12:59:48 Bernardo|away: we spent four days rebuilding our package feeds in the hope it would solve the problem. it didn't :-( Apr 22 13:02:47 ouch Apr 22 13:05:47 well, it was time for a spring clean anyway. Apr 22 13:12:32 rwhitby: What is rdepending on python? Apr 22 13:13:30 ipkg-utils, presumably Apr 22 13:14:16 yep Apr 22 13:15:06 We need to improve the logging levels in the bitbake dependency code as this isn't clear from the log... Apr 22 13:15:24 rwhitby: I get the same error here - I'll look into it Apr 22 13:15:33 RP: thx Apr 22 13:15:43 RP: are you in oepnZaurus as well ? Apr 22 13:15:57 s/oepn/Open/ Apr 22 13:19:27 RP: is there still a chance for a feature request to make bitbake more usefull in group environment ? Apr 22 13:19:56 uv1: We have a bugzilla which takes feature requests? Apr 22 13:20:09 uv1: We're open to ideas... Apr 22 13:20:39 RP first I want to discuss the idea ... Apr 22 13:21:21 uv1: I'm listening although I might or might not be the best person to discuss it with Apr 22 13:22:18 If you have a group of developers working on one distro. One is the stable tree maintainer, severals are working on modification, which may crash stability ... Apr 22 13:22:51 Not all of them of course want to build the toolchain and the rest of the basic distro fully. Apr 22 13:23:42 In stead of that, a "global" stable staging / cache / stamps area would be helpfull, which could be expanded by lets call it private staging / cache / stams areas. Apr 22 13:24:21 uv1: The preferred solution is to put staging under packaging control Apr 22 13:24:36 uv1: You can the provide bits of staging, precompiled Apr 22 13:25:02 RP: Would that mean, that you can develop a package with higher versioning without effecting the staging area ? Apr 22 13:25:37 uv1: In your local copy, you'd replace the version in staging with the higher version Apr 22 13:26:08 But building new copies of the respository would be cheap as you'd be able to do it from prebuilt packages Apr 22 13:26:28 But how would the local copy be created. Is that a full copy of the staging area ? Apr 22 13:27:13 How is it handled, that if one e.g. builds a new uitilty or library, that the new one is not linked against stable packages, if a stable distro is build ? Apr 22 13:27:30 Someone would have built your stable base distro along with a set of packages that were installed to make staging Apr 22 13:27:55 Someone else would take those packages and install them to create a copy of staging, then work on top of that Apr 22 13:28:41 Obviously if you replace libraries, you'd have to rebuild any packages depending on those libraries to link against the replacements Apr 22 13:28:56 RP: You mean, you can copy the staging area by installing the stuff inside into a local copy ? Apr 22 13:30:04 uv1: I don't follow :-/ Apr 22 13:31:07 RP: would a "cp -a global/staging local" where local is you private build dir work ? I assume, that at least stamps would be missing ? Apr 22 13:31:47 You'd need cross and stamps as well. The other issue is that staging has paths hardcoded into it Apr 22 13:32:04 RP: Thats what I am after ... Apr 22 13:32:08 you'd also need deploy/ipk Apr 22 13:32:29 uv1: Is should be possible to script something to change the paths Apr 22 13:32:30 Has libqpe1 been removed lately? Apr 22 13:32:49 uv1: The ultimate aim is to have this all done with package management Apr 22 13:33:06 I think it could be managed to have $STAGING (and other required variables) handled by bitbake like a PATH variable!... Apr 22 13:33:34 Try the first PATH, thet try the next ... Use always the newest found version. Apr 22 13:33:45 uv1: I doubt it I'm afraid. Think about things like .la files Apr 22 13:34:34 This could help to reduce the harddisk usage and would not force users to rebuild everything just to try a new package. Apr 22 13:35:19 uv1: It is planned. Its just difficult and needs someone to spend time on it Apr 22 13:35:23 RP: .la files ? You mean -L, -I ? Could be handled the same way, only the order must be inverted. Apr 22 13:35:58 Search the new buidl dir first and then the next elder ! Apr 22 13:36:09 uv1: grep your staging directory for references to its path and you'll see what I'm taking about Apr 22 13:37:12 RP: hrgh. Thats bad ... Apr 22 13:38:03 RP: BTW another idea is to use unionfs. Got unionfs working now in principle, but had problems as well ... Apr 22 13:39:15 mounted my boxer build dir to a new empty dir , changed conf/local.conf by changing to collie, but still defconfig-cx7 was search for. Any ideas about that ? cache ? Apr 22 13:40:19 uv1: For different machines use the multimachine include Apr 22 13:40:58 RP: That was just a test to get the above mentioned things working ... Apr 22 13:41:28 * rwhitby has to learn about multimachine to see if it can reduce the number of nslu2-linux build directories ... Apr 22 13:41:31 I was just curios, where bitbake gets the defconfig-c7x stuff from. Apr 22 13:42:10 RP: BTW How to get bitbake 1.4 via svn. Apr 22 13:42:21 uv1: It shouldn't have been getting that from anywhere if the machine was set to collie Apr 22 13:43:10 RP: Correct, so I was wondering ... OK will give it another try with bitbake 1.4. That speeds things up really ... Apr 22 13:43:15 uv1: 1.4 is on a branch which you probably have to checkout. I don't know the extact name. I think someone updated GettingStarted Apr 22 13:43:37 RP: Will check Apr 22 13:43:48 uv1: svn co svn://svn.berlios.de/bitbake/branches/bitbake-1.4 bitbake :) Apr 22 13:45:35 cyrilRomain: Thanks a lot. Apr 22 14:41:40 rwhitby: Bitbake is right in saying nothing provides the runtime "python". You probably need to add RPROVIDES_python-core = "python" to python-*-manifest.inc Apr 22 14:42:00 rwhitby: That will make it work - I'll check its the correct fix with mickey|bbl Apr 22 15:05:23 Hhm. How to start a script in the target build dir before configure ? How to add parameters to the configure call. Apr 22 15:18:11 uv1: do_configure_prepend () {} in you bb for first question Apr 22 15:18:22 uv1: EXTRA_OECONF = "params" for second Apr 22 15:25:49 mickey|bbl, BerliOS's database machine has been down for almost 4 days now. Apr 22 15:39:02 XorA|gone: Thanks will try ... Apr 22 15:46:41 Quick question. Does OE support Sharp SL-C3200. Have just seen that device at trisoft homepage. Sounds really good. Any users here, who have experience ? Apr 22 15:47:17 the terrier should work with the same rootfs and kernel as the 3100 (borzoi). Apr 22 15:47:38 I didn't actually notice anyone owning one tho Apr 22 15:47:55 CoreDump|home: Any things still broken for borzoi ? Apr 22 15:48:17 CoreDump|home: borzoi has a Harddisk too ? Apr 22 15:48:19 hmmm not that I knew of, maybe usb <-> windows client Apr 22 15:48:33 windows sucks anyway ;-) Apr 22 15:48:35 yes borzoi has a smaller MD, that's all Apr 22 15:48:40 uv1: ;) Apr 22 15:48:55 What is the HD used for ? Apr 22 15:49:00 rootfs Apr 22 15:49:09 and /home AFAIK Apr 22 15:49:19 can't afford one =) Apr 22 15:49:40 Do we have a running debian in oe for that ? Apr 22 15:49:51 the flash is not used if that matters Apr 22 15:50:21 CoreDump|home: not used means really nothing on ? We boot from HD ? Apr 22 15:50:21 uv1: OE builds distributions. You could probably hack up a real deal debian-like distro, yes. Apr 22 15:50:34 uv1: yes, boots from HDD Apr 22 15:50:59 RP: thanks for the mail. dealing with that atm. Apr 22 15:51:33 CoreDump|home: Building a debian distro from scratch using OE would cost quite some time ;-) Apr 22 15:51:51 indeed Apr 22 15:52:25 CoreDump|home: Just have seen a OZ doc page descring a parallel GPE / OPIE install. Would that still be possible with my boxer ? Apr 22 15:53:12 sure, even works on collie Apr 22 15:53:37 but it's rather slow due to the limited RAM Apr 22 15:54:18 03mickeyl 07org.oe.dev * r429227e9... 10/packages/sdlperl/sdl-perl_1.20.3.bb: sdl-perl: remove bogus usage of a build class Apr 22 15:54:23 03mickeyl 07org.oe.dev * rd68918ea... 10/packages/python/ (python_2.3.4.bb python_2.4.3.bb): python 2.3.4, 2.4.3: make python-core provide the python runtime Apr 22 15:54:36 rprovide, that is Apr 22 15:54:38 CoreDump|home: So really both are started ? Apr 22 15:54:57 correct Apr 22 15:55:20 IMO not very useful Apr 22 15:55:45 CoreDump|home: How does the Terrier boot from HD. Bootmanager ? Apr 22 15:56:07 nope, kernel 2.6 boots straight into HDD Apr 22 15:56:28 no hacks involved Apr 22 15:56:31 CoreDump|home: So that device would not be better (same amount of RAM). Apr 22 15:57:13 but faster CPU Apr 22 15:57:29 if you've got cash to burn, go for it =) Apr 22 15:57:45 BTW : Just noticed, that my RPM patches are not yet in ;-(( Apr 22 15:58:08 ~lart rpm hell Apr 22 15:58:09 * ibot teaches rpm hell the basics, including how to RTM Apr 22 15:58:19 CoreDump|home: Terrier - May be ;-)) Apr 22 15:59:11 I got a c1000, basically a c3xxxx w/o HDD. Very nice machine Apr 22 15:59:57 Thats like my Cl860. Also works fine, but the HD would be nice instead of the SD card ! Apr 22 16:01:16 * CoreDump|home nods Apr 22 16:01:34 Collected errors: Apr 22 16:01:34 | ERROR: Cannot satisfy the following dependencies for task-bootstrap: Apr 22 16:01:34 But does not show any failing depends. Apr 22 16:01:58 scroll up Apr 22 16:02:40 Oops. Lots ... Apr 22 16:03:16 Lots of update-modules entries though. Any ideas ? Apr 22 16:03:19 meh inittab is hard-wired to spawn only one VT Apr 22 16:03:46 not really, didn't try to build .dev in ages Apr 22 16:04:16 search for "ERROR" Apr 22 16:05:05 mickeyl: Thanks Apr 22 16:05:38 uv1: The spitz machine gives working images for borzoi and terrier. As far as OE is concerned, they're all the same Apr 22 16:06:13 RP: Thanks for info Apr 22 16:06:34 uv1: I have a spitz and love the microdrive ;-) Apr 22 16:08:11 bitbake 1.4.1 rocks ! Why wasnt the speed not reached before ? Apr 22 16:09:03 s/wasnt/was/ Apr 22 16:10:09 uv1: We just found a more efficient way to do things... Apr 22 16:10:57 RP: do we have MACHINE redirectors yet? Apr 22 16:10:59 like, for the husky Apr 22 16:11:04 RP: If you once need a job in good old germany ;-) Apr 22 16:11:49 mickeyl: They just error out and tell you to use c7x0 or spitz. I did wonder what would happen if they simply said MACHINE="c7x0"... Apr 22 16:12:02 mickeyl: I added one for terrier Apr 22 16:12:06 RP: nah, we want things to happen explicitly :) Apr 22 16:12:11 RP: thanks, very good Apr 22 16:12:24 borzoi as well? Apr 22 16:12:42 mickeyl: borzoi was its own machine type for a while. I purged all references to it :) Apr 22 16:12:53 cool Apr 22 16:13:06 uv1: Thanks for the offer but I'm happy where I am ;-) Apr 22 16:13:12 btw., I might get a borzoi donated if things go well Apr 22 16:13:33 which would mean I could send my spitz to another developer Apr 22 16:13:37 mickeyl: We're not doing badly for devices atm :) Apr 22 16:13:38 lets see and cross fingers Apr 22 16:13:44 RP: no kidding :) Apr 22 16:14:57 mickeyl, You were asking about openzaurus.berlios.de earlier... Apr 22 16:15:08 mickeyl: I'm curious to see what Sharp bring out next... Apr 22 16:15:22 mickeyl: The new generation of PXA might give us some headaches... Apr 22 16:15:24 RP: May be you were too efficient ? If I want to build one of the unmet dependencies packages (e.g. busybox) nothing happens. So I assume that busybox was not at all build. So somethings may be wrong with dependencies! Apr 22 16:15:38 bronson: thanks. seen your reply + mail Apr 22 16:15:53 ah, good. Apr 22 16:15:59 I hope it gets resolved soon! Apr 22 16:16:02 yep. me too Apr 22 16:16:26 uv1: Instead of assuming it didn't build, why not check? Apr 22 16:17:42 NOTE: build 200604221822: completed Apr 22 16:17:42 Build statistics: Apr 22 16:17:42 Attempted builds: 0 Apr 22 16:17:57 RP: no ipk available. bitbake busybox returns just with the above mentioned message. Apr 22 16:18:16 uv1: Is it in tmp/work/busybox ? Apr 22 16:18:33 which subdir of work ? Apr 22 16:18:42 RP: we really need a reverse lookup Apr 22 16:18:48 for attributes Apr 22 16:18:58 mickeyl: How do you mean? Apr 22 16:19:03 i'm getting crazy looking for where the final say of an assignment is being defined Apr 22 16:19:09 i.e. FOO = "bar" Apr 22 16:19:14 where does it come from? Apr 22 16:19:19 a conf? a bb? which one? Apr 22 16:19:43 mickeyl: That would be useful. It starts to sound like depencency tracking... Apr 22 16:20:15 RP: It is in armv5te-linux! Apr 22 16:20:54 uv1: That's where it would be if you were compiling for an armv5te machine. So busybox did build at some point in the past Apr 22 16:21:06 uv1: Are you sure the .ipk files don't exist? Apr 22 16:21:24 RP: It was a fresh build (cleaned away tmp this morning). Apr 22 16:21:57 RP: no run.do_package file though. Apr 22 16:22:43 uv1: Does tmp/stamps/armv5te-linux/busybox-1.01-r9.do_package exist? Apr 22 16:22:50 or something similar? Apr 22 16:23:21 RP: yep Apr 22 16:23:51 uv1: That means it should have packaged and the .ipk files should exist. Bitbake is there quite correct in thinking it already built Apr 22 16:24:49 RP: but definitely it isnt. BTW I started this morning with a bitbake 1.5.0 version and due to mickeys hint changed to 1.4.1 when kernel could not build. Apr 22 16:24:57 RP: May be thats the reason ? Apr 22 16:25:40 uv1: That's possible - bitbake 1.5.0 is totally broken Apr 22 16:25:56 OK, try again ;-) Sorry have to go off for an hour or so. See you later ... Apr 22 16:28:13 hmm Apr 22 16:28:27 r459 doesn't recognize my setting PREFERRED_PROVIDER Apr 22 16:28:30 this is very odd Apr 22 16:28:33 r458 even Apr 22 16:29:59 mickeyl: :-/ Apr 22 16:30:34 * mickeyl tries some revs older Apr 22 16:30:45 any known good before 458? Apr 22 16:31:09 mickeyl: 429 was a good one if I remember rightly Apr 22 16:31:13 thanks Apr 22 16:31:14 trying Apr 22 16:31:27 mickeyl: I'm concerened 458 doesn't work though. What are you trying to do? Apr 22 16:32:01 I can't get SD working on poddle under 2.6 :-/ Apr 22 16:35:21 RP: I have three providers for virtual/libqte2 Apr 22 16:35:42 NOTE: multiple providers are available (qte-mt-static, qte, qte-mt); Apr 22 16:35:55 NOTE: consider defining PREFERRED_PROVIDER_virtual/libqte2 Apr 22 16:36:04 and in preferred-opie-versions.inc I have Apr 22 16:36:36 PREFERRED_PROVIDERS += "virtual/libqte2:qt-mt" Apr 22 16:36:36 PREFERRED_PROVIDER_virtual/libqte2 = "qt-mt" Apr 22 16:36:43 to be sure i even used both syntaxes Apr 22 16:36:55 but no matter what i do, It always choses qte-mt-static Apr 22 16:36:58 sorry, lunch Apr 22 16:37:03 dinner, even Apr 22 16:37:08 well, OZ patches are not in http://www.rpsys.net/openzaurus/patches/ anymore but in http://www.rpsys.net/openzaurus/patches/archive/. Apr 22 16:37:18 mickey|dinner: Did it have a problem building one of the other versions in the past? Apr 22 16:37:19 RP: (note that you can't reproduce that for now since I have some unapplied patches atm.) Apr 22 16:37:54 mickey|dinner: The output of -D -D -D might say more about what it considered doing... Apr 22 16:38:02 Should I report in a offcial bur report ? It seems that just packages/linux/linux-openzaurus.inc has to be updated Apr 22 16:38:35 cyrilRomain: What needs updating? Apr 22 16:39:13 RP: in packages/linux/linux-openzaurus.inc s:openzaurus/patches:openzaurus/patches/archive: Apr 22 16:39:38 cyrilRomain: Which branch are you using? Apr 22 16:40:20 RP: for such problem I never know who to warn: OE team or OZ people to update location ? Apr 22 16:40:44 RP: oz354fam083 Apr 22 16:41:01 cyrilRomain: It depends on the branch. I don't think that problem exists in .dev or the OZ release branch (oz354x) Apr 22 16:41:25 right, so given that branch is now for familar 0.8.4, I can't touch it and its a familar bug Apr 22 16:41:58 ok, maybe should I switch to oz354x then, since it's to build OZ for tosa Apr 22 16:42:13 cyrilRomain: You will have a lot more success with that branch doing that Apr 22 16:42:38 RP: well, didn't know, thanks for the info :) Apr 22 16:43:18 RP: I'll try. Do you know where can I report the problem to Familiar people ? Is there a channel ? Apr 22 16:43:34 cyrilRomain: Basically, use .dev if you want the cutting edge, oz354x for more stable openzaurus and oz354fam083 for familar (ipaqs) Apr 22 16:43:56 cyrilRomain: Since familar doesn't support Zaurus devices, I doubt they'll care :-/ Apr 22 16:45:10 well thanks, this is such info that lack in wiki. I'll update it soon to make this more obvious to OE noob like me :) Apr 22 16:46:04 cyrilRomain: oz354fam083 was used for OpenZaurus 3,5,4 but that version of Openzaurus didn't support 2.6 kernels on many of the zaurus models Apr 22 16:48:15 RP: ok, switching to .dev in progress ;) Apr 22 16:52:30 mickey|dinner: if you wrote "qt-mt" while it was expecting "qte-mt", this behaviour would be expected. Apr 22 16:53:12 regrettably you don't get any diagnostic if you tell bitbake to prefer a package that either doesn't exist or doesn't provide the virtual in question Apr 22 16:59:01 uh no! Apr 22 16:59:04 thanks pb_ Apr 22 16:59:07 * mickey|dinner slaps forehead Apr 22 17:08:41 mickeyl: heh Apr 22 17:08:53 if [ -n "`quilt applied`" ]; then quilt pop -a -R -f || exit 1; fi Apr 22 17:09:19 I'm having a problem with that line... It's killing my build, and spits out: Usage: quilt pop [-afRqv] [num|patch] Apr 22 17:18:16 Can someone paste their do_patchcleancmd() function from "log.do_patchcleancmd" for gnu-config-native-0.1cvs20050701-r4 ? Apr 22 17:20:03 i have a workaround, not a fix Apr 22 17:20:05 INHERIT += "patcher" Apr 22 17:20:07 into your local.cof Apr 22 17:20:09 local.conf, even Apr 22 17:20:15 then you won't have to deal with quilt at all Apr 22 17:30:08 Hmm... Said the patch didn't apply cleanly, I'm going to clear out the tmp and try again Apr 22 17:32:55 hey Apr 22 17:33:51 hi zecke Apr 22 17:35:59 03mwester 07org.oe.dev * r3d32d5cf... 10/packages/busybox/ (slingbox-1.00/defconfig slingbox_1.00.bb): slingbox: add losetup to defconfig Apr 22 17:52:37 Got it :). Poodle has two SD power control GPIOs, one of which is undocumented and previously unknown :) Apr 22 17:53:49 Nasty :) Apr 22 17:54:47 Don't sharp have to release the (working) source for 2.4? Or are they doing the binary blob module thing? Apr 22 17:55:11 NA|Gone: The mmc/sd driver is binary only Apr 22 17:55:47 Hooray! It progresses further in the build, but it gets shot down with: cannot stat `i686-cygwin_nt-libtool': No such file or directory Apr 22 17:55:54 RP: Ah, smeg.. Apr 22 17:56:45 NA|Gone: Thankfully for most models they used the controller in the PXA processor which we have docs for. Figure out the power controls and we can make it work... Apr 22 17:56:56 collie isn't so easy though Apr 22 17:57:07 * NA|Gone nods.. Apr 22 17:57:16 I thought some guy signed the nda and was going to write drivers for it Apr 22 17:57:37 Is there a reason they don't release the details? Apr 22 17:57:41 That would be kergoth who never found the time :-/ Apr 22 18:27:50 ~praise RP for working reboot Apr 22 18:27:52 All hail RP for working reboot! Apr 22 18:29:05 huh? Apr 22 18:29:51 zach: reboot / halt was b0rked w/ kernel 2.6 on sl-cxxxx Apr 22 18:30:20 Ah... Well congrats on it being fixed. =) Apr 22 18:30:38 heh Apr 22 18:32:03 * zach shakes fist at Cygwin Apr 22 18:32:43 What's better to build on, FreeBSD or Cygwin? Apr 22 18:33:09 Because I'm just about ready to through in the towel with Cygwin Apr 22 18:33:23 throw* =) Hehe Apr 22 18:33:24 zach: both need work Apr 22 18:34:40 All I really want is to apply a patch to kphonpi, and see if it works with sipphone.com like the author claims, because then my Z would become my WiFi phone =) Apr 22 18:34:48 kphonepi* Apr 22 18:35:03 zach: use CoLinux, VMware or such ;) Apr 22 18:35:17 zach: and probably the SDK from a OpenZaurus release Apr 22 18:35:45 I am using the sdk from the Z release, well, not the toolset, since that's Linux only Apr 22 18:35:54 but I'm using the rest of the files Apr 22 18:37:04 no cross compiler? Apr 22 18:37:48 the cross compiler in the sdk is the linux build Apr 22 18:38:33 zach: well, to phrase it like this Apr 22 18:38:39 zach: you are on your own for now Apr 22 18:38:49 Heh, yeah... Apr 22 18:38:52 I figured that Apr 22 18:39:16 zach: Linux is the best tested and known working host Operating System Apr 22 18:39:48 Yeah... I'll have to see if I can get the vmware images working then Apr 22 18:42:41 zach: I'm a BSD (FreeBSD, OSX) user myself, but use OE only on Linux Apr 22 18:43:21 zekoZeko, Yeah, I've got a FreeBSD server that's out in the garage, I was thinking of building with that, but it seems like too much of a hassle Apr 22 18:43:42 I guess I'll just try and get the vmware images to work, and build OE in one of those... Apr 22 18:43:50 The VMWare player is free, right? Apr 22 18:46:18 yes it is Apr 22 18:46:24 'free of charge' that is Apr 22 18:47:10 good enough Apr 22 18:55:20 Thanks for the help Apr 22 19:08:36 CoreDump|home: DId the RTC survive? Apr 22 19:09:07 good question as I didn't bother to set it in the first place heh Apr 22 19:09:12 will let you know Apr 22 19:09:28 CoreDump|home: I should really test it myself - just haven't had time :) Apr 22 19:09:48 i know what you mean Apr 22 19:10:04 CoreDump|home: is initial parsing still slower for you? Apr 22 19:10:12 If someone beats me too it, fine. If not, I'll get there :) Apr 22 19:10:42 zecke: I didn't dare to use 1.4 w/ .oz354x yet Apr 22 19:10:52 CoreDump|home: ah that is fine Apr 22 19:11:09 CoreDump|home: I think hrw did Apr 22 19:11:16 CoreDump|home: I also have been Apr 22 19:11:43 I guess I'll wait a week or two =) Apr 22 19:23:04 ~lart SUSE Apr 22 19:23:04 * ibot overclocks SUSE until SUSE burns out Apr 22 19:25:09 RP: the RTC just survived a reboot. Apr 22 20:00:27 03rw 07org.oe.oz354fam083 * rea560a88... 10/packages/linux/handhelds-pxa_2.4.19-rmk6-pxa1-hh41.bb: handhelds-pxa: use -hh41 by default Apr 22 20:21:31 Trying to build openzaurus-unstable Kernel 2.6 for tosa with .dev branch + bitbake 1.4 : OE fails to build after parsing finished "ERROR: Openembedded's config sanity checker detected a potential misconfiguration." :-/ I double-checked my local.conf but it looks correct, any ideas ? Apr 22 20:22:46 cyrilRomain: what is the potential misconfiguration it claims? Apr 22 20:23:03 zecke: don't know. I do not have more info Apr 22 20:23:44 cyrilRomain: sure? Apr 22 20:23:50 zecke: I tried bitbake -v -D task-bootstrap but except some "DEBUG: executing anonymous function: incompatible with host arm-linu" for few packages, I don't see other error Apr 22 20:24:16 arm-linu ? Apr 22 20:24:21 cyrilRomain: new kind of OS? Apr 22 20:25:04 zecke: I work in a chrooted environment a Gentoo with gcc4.1 and glic2.4 Apr 22 20:25:20 cyrilRomain: at least in your above message Apr 22 20:25:25 cyrilRomain: a 'x' is missing Apr 22 20:25:33 cyrilRomain: which indicates a wrong TARGET_OS Apr 22 20:25:48 cyrilRomain: anyway, if you fail to pastebin stuff, I can't help Apr 22 20:26:22 zecke: no, just a bad copy/paste :) Apr 22 20:26:39 cyrilRomain: right, and this wastes my time Apr 22 20:26:44 ~pastebin Apr 22 20:26:45 [pastebin] a place to paste your stuff without flooding the channel - try http://pastebin.com/, or http://pastebin.ca, or http://channels.debian.net/paste, or for #oe use http://oe.pastebin.com Apr 22 20:26:58 soo easy, why are making people their lives harder than necessary Apr 22 20:27:13 zecke: ok, sorry for that. I did'nt know, thanks Apr 22 20:28:23 cyrilRomain: also read what it writes "ยท bb.fatal("Openembedded's config sanity checker detected a potential guration.\nEither fix the cause of this error or at your own risk disable the ecker (see sanity.conf).\n%s" % msg)" Apr 22 20:28:39 cyrilRomain: so it tells you how to disable the heuristic Apr 22 20:28:48 hmm Apr 22 20:28:49 cyrilRomain: and it should print a message with the reason.. Apr 22 20:28:57 a fresh opie-image in .dev doesn't look too god Apr 22 20:29:03 mickeyl: hey Apr 22 20:29:08 mickeyl: what is missing/broken? Apr 22 20:29:17 open("/dev/fb0", O_RDWR) = 4 Apr 22 20:29:17 ioctl(4, FBIOGET_VSCREENINFO, 0xbeab7ff4) = 0 Apr 22 20:29:17 mmap2(NULL, 0, PROT_READ|PROT_WRITE, MAP_SHARED, 4, 0) = -1 EINVAL (Invalid argument) Apr 22 20:29:23 and then it crashes Apr 22 20:29:40 oh Apr 22 20:30:01 Error: failed to map framebuffer. Transformed driver cannot connect. Apr 22 20:30:11 dunno whether this is Apr 22 20:30:14 a) recent kernel Apr 22 20:30:16 b) qt-mt Apr 22 20:30:37 c) 2.3.12 Apr 22 20:30:49 Ok, I got theVMWare Player up and working, are there any extra steps to use the oz sdk instead of having bitbake create all of it for you? Apr 22 20:31:16 mickeyl: looks like too many variables to solve automatically Apr 22 20:31:26 i'm afraid so Apr 22 20:31:27 zecke: hey I cannot understand: I did a 'bitbake -v -D task-bootstrap > temp' to dump temp into pastebin, but bitbake know downloads... Apr 22 20:31:47 zecke: guess i'm going back to 2.3.10 for now and then do it step by step *sigh* Apr 22 20:31:57 zach: I'm not a OpenZaurus guy. I think the SDK is meant for people that do not want to use bitbake/OpenEmbedded Apr 22 20:32:01 zecke: I'll will restart completly from scratch to check we don't need to call bitbake twice to make it work. strange anyway Apr 22 20:32:17 zach: e.g. for you, you could set a lot of ASSUME_PROVIDED (but then I don't know the SDK) Apr 22 20:32:48 gosh I hate SUSE Apr 22 20:33:11 Crap... Why? That's the image just started using Apr 22 20:33:46 zach: do not weight my information too much Apr 22 20:34:31 RP: is their something like bk changes -L Apr 22 20:38:18 cyrilRomain: calling it twice should not be necessary Apr 22 20:38:56 zecke: I just test it again and I need... strange Apr 22 20:39:34 s/test/tested/ Apr 22 20:39:56 has anyone tried to compile zdict or zten ? Apr 22 20:40:14 e.g. pastebin the complete sanity output... Apr 22 20:45:10 What was that inherit line again? Apr 22 20:45:19 INHERIT = "patches" ? Apr 22 20:45:32 I was wondeting how difficult it would be to compile those dictionnary viewers, originally intended for sharp rom Apr 22 20:45:36 INHERIT += "patcher" Apr 22 20:45:54 AlexTe: are they in OE already? Apr 22 20:46:10 nope Apr 22 20:48:47 I just tried but there are lots of errors Apr 22 20:49:30 AlexTe: many reasons: gcc2 vs. gcc3, QtE upgrades... Apr 22 20:50:25 so I guess it's difficult then Apr 22 20:50:46 AlexTe: no, some googling Apr 22 20:50:58 AlexTe: gcc is mean, it throws out many errors, if you fix the first one Apr 22 20:51:07 AlexTe: the amount of output decreases rapidly Apr 22 20:52:39 errors seems to be about missing qtopia includes Apr 22 20:53:57 AlexTe: if it doesn't use proprietary SharpSL stuff, it should be fairly easy Apr 22 20:54:29 nothing proprietary Apr 22 20:54:41 AlexTe: should be fairly easy then Apr 22 20:54:48 not for me :) Apr 22 20:54:48 AlexTe: feel free to pastebin the error Apr 22 20:55:29 zecke: well, the pastebin of the 'bitbake -v -D task-bootstrap' without many lines of bb parsing: http://pastebin.com/675810 Apr 22 20:56:24 cyrilRomain: line 263: Please install the diffstat utilits Apr 22 20:56:33 qmake: WARNING: Failure to find: pixmap.h and qlabeledit.h Apr 22 20:56:44 cyrilRomain: file a bug report for the spelling error Apr 22 20:57:07 AlexTe: both are not Qtopia files Apr 22 20:57:47 oh what package do they belong to then? Apr 22 20:57:55 AlexTe: dunno, zdict probably Apr 22 20:57:59 zecke: An equivalent command in which scm? Apr 22 20:58:01 zecke: hum... 'utilits' ? Apr 22 20:58:10 RP: sorry, git ;) Apr 22 20:58:42 zecke: I'm afraid not to understand Apr 22 20:58:51 zecke: git-log? Apr 22 20:59:09 cyrilRomain: what does line 263 say? Apr 22 20:59:25 cyrilRomain: 'Please install diffstat' Apr 22 21:00:07 Yay, sane output from objdump after I patched it :) Apr 22 21:00:48 zecke: well, should I add it in the wiki's RequiredSoftware too ? Apr 22 21:00:58 RP: congrats on the poodle findings Apr 22 21:01:07 cyrilRomain: take a look at ASSUME_PROVIDED += in bitbake.conf Apr 22 21:01:23 cyrilRomain: and feel free to add these bits to the wiki :) Apr 22 21:01:33 mickeyl: thanks. I had to do it the same way I learnt about the shared power control on spitz :-/ Apr 22 21:02:01 mickeyl: I just don't understand why there are two GPIOs involved... :-/ Apr 22 21:02:25 zecke: well ok, thank you :) Apr 22 21:05:29 03mickeyl 07org.oe.dev * r92b6d01e... 10/packages/konqueror/ (konqueror-embedded_svn.bb konqueror-embedded_20060404.bb): konqueror-embedded 3.5: pinpoint a svndate Apr 22 21:05:33 03mickeyl 07org.oe.dev * r29d15d33... 10/classes/sdl.bbclass: Apr 22 21:05:33 sdl.bbclass: remove bogus stuff that was present before Apr 22 21:05:33 first shot at an sdl class to remove the need for seperate Apr 22 21:05:33 SDL-x11 and SDL-opie bbfiles. For now, it adds an extra package Apr 22 21:05:34 ${PN}-opie that RDEPENDS on ${PN} and adds the necessary opie .desktop glue Apr 22 21:05:38 03mickeyl 07org.oe.dev * rf0c927ee... 10/classes/palmtop.bbclass: palmtop.bbclass: virtual/libqpe -> virtual/libqpe1 Apr 22 21:05:43 03mickeyl 07org.oe.dev * rbc7db962... 10/classes/sdl.bbclass: sdl.bbclass: use ?= for APPNAME, APPIMAGE and APPDESKTOP defaults Apr 22 21:05:47 03mickeyl 07org.oe.dev * r53049d43... 10/classes/qmake.bbclass: qmake.bbclass: show the used qmake version if we bail out Apr 22 21:05:51 03mickeyl 07org.oe.dev * r065b1ce1... 10/classes/palmtop.bbclass: Apr 22 21:05:51 palmtop.bbclass: document what this class does and automatically set up CONFIG=thread, if PALMTOP_USE_MULTITHREADED_QT is set Apr 22 21:05:52 to yes Apr 22 21:05:55 03mickeyl 07org.oe.dev * r520ed650... 10/packages/libqpe/ (libqpe-opie.inc libqpe-opie_1.2.1.bb libqpe-opie_cvs.bb): libqpe-opie: make provides and depends more precise Apr 22 21:05:59 03mickeyl 07org.oe.dev * r6d9f809a... 10/packages/qte/ (10 files in 2 dirs): qte: use .inc file to produce three incarnations of every version Apr 22 21:06:03 03mickeyl 07org.oe.dev * r3a6c0a29... 10/conf/distro/ (5 files): configuration files: adjust opie versions Apr 22 21:06:21 mickeyl: did you change documentation.conf as well? Apr 22 21:06:34 oops... ah yes, of course... just not pushed :D Apr 22 21:06:46 *cough* Apr 22 21:07:05 mickeyl: weird monotone bug? Apr 22 21:07:17 no no Apr 22 21:07:19 =D Apr 22 21:08:42 * CoreDump|home thinks XFCE in OE needs some TLC Apr 22 21:09:02 no kidding Apr 22 21:09:05 it's pretty messed up Apr 22 21:09:26 and segfaulting left and right Apr 22 21:09:47 You know it might be possible to shim the collie SD code into 2.6... Apr 22 21:09:58 tell us more Apr 22 21:10:37 Run it through objdump, fixup the differences between 2.4 and 2.6, then compile the resulting assembler/fixup code... Apr 22 21:10:45 hmm Apr 22 21:10:48 sounds like a lot of work Apr 22 21:10:52 03mickeyl 07org.oe.dev * r7953d7e9... 10/conf/documentation.conf: documentation.conf: document PALMTOP_USE_MULTITHREADED_QT Apr 22 21:11:02 RP: did you change objdump to emit code that can be assembled again? Apr 22 21:11:15 zecke: No :-/ Apr 22 21:11:40 zecke: I did find a patch to fix up the external undefined symbols though Apr 22 21:11:48 CoreDump|home: xfce(-wm) doesn't like kdrive Apr 22 21:12:17 CoreDump|home: it assume 24 bit = 8 8 8 RGB, instead of 5 6 5 8 RGBA Apr 22 21:12:26 heh Apr 22 21:12:37 lol Apr 22 21:12:42 then XFCE could run on QtE Apr 22 21:13:34 wouldn't that be QFCE then? ;) Apr 22 21:13:52 XFC/E? Apr 22 21:14:02 hehe Apr 22 21:14:37 well, at least the panel works after "fixing" the icons Apr 22 21:14:37 03mreimer 07org.oe.oz354fam083 * r521efa20... 10/packages/base-files/base-files/h2200/fstab: base-files: Add an fstab for h2200. Apr 22 21:14:40 xfce should build, compile and install after my previous round of fixes Apr 22 21:14:51 koen: it does Apr 22 21:15:11 mreimer_: please don't forget .dev ;) Apr 22 21:15:26 but putting an image together is out of the question Apr 22 21:16:43 koen: mickeyl : I wan't to make Angstrom work for the SIMpad? Apr 22 21:16:50 are the policies up-to-date? Apr 22 21:17:01 zecke: please send us the bootldr + flasher files :) Apr 22 21:17:04 * mickeyl promotes zecke to Angstrom SIMpad manager Apr 22 21:17:15 you know that Angstrom is 2.6 only right? ) Apr 22 21:17:15 :) Apr 22 21:17:40 mickeyl: sets Xvfb as preferred xserver Apr 22 21:18:15 koen@bitbake:/data/build/koen/OE/build/tmp/angstrom/deploy/images$ ls Apr 22 21:18:16 akita c7x0 ep93xx h2200 ipaq-pxa270 spitz Apr 22 21:18:25 it seems the simpad 2.6 image doesn't build right now Apr 22 21:18:30 * koen checks logs Apr 22 21:18:36 koen: we are working on it Apr 22 21:19:18 * RP hopes poodle will make the list soon... Apr 22 21:19:28 zecke: which is probably caused by gaim not being able to fetch from sf.net Apr 22 21:19:54 koen: I have a git tree for SIMpad now Apr 22 21:22:09 zecke: cool Apr 22 21:22:19 zecke: at uni-berlin? Apr 22 21:22:35 koen: (non public in my hh.org account atm) Apr 22 21:22:47 koen: but I think I need to move it away, disk space on hh.org is getting low Apr 22 21:23:14 I think mickeyl and I can arrange something about hosting that git tree ;) Apr 22 21:23:39 which reminds me: I need to add all those fancy xlibs git trees to OE Apr 22 21:23:45 koen: and SIMpad is StrongARM Apr 22 21:24:09 zecke: with the current toolchain that isn't a problem Apr 22 21:24:19 our EABI doesn't need thumb Apr 22 21:24:36 and I expect the thumb issue to be fixed with gcc 4.1.1 or 4.1.2 Apr 22 21:25:12 zecke: I freed a couple gigs recently. I also think George was going to add more disk space or move stuff to different servers. Apr 22 21:25:38 reenoo: ATM it is at 96% (if I can trust df -m) Apr 22 21:26:02 zecke: the simpad kernel bails out at arch/arm/mach-pxa/eseries/e750_lcd.c Apr 22 21:26:18 zecke: so it seems a new defconfig would fix a lot of stuff Apr 22 21:26:37 koen: yes, and some other things Apr 22 21:28:32 koen: h2200 fstab in base-files pushed to .dev Apr 22 21:28:48 mreimer_: thanks Apr 22 21:28:48 koen: how are the angstrom images coming along? Apr 22 21:29:22 mreimer_: http://oe.pastebin.com/675875 Apr 22 21:29:34 03mreimer 07org.oe.dev * r0b5d6ed4... 10/packages/base-files/base-files/h2200/fstab: base-files: Add fstab for h2200. Apr 22 21:29:46 I'm busy getting gpe-image to build for all machines with angstrom Apr 22 21:30:08 opie-image builds but SIGSEGVs pretty early Apr 22 21:30:14 (4 weeks ago) Apr 22 21:30:19 zecke: well, yeah. 3.9G free though... Apr 22 21:30:52 the complete result: http://oe.pastebin.com/675876 Apr 22 21:30:53 koen: yay! let me know when you have an h2200 image ready to test Apr 22 21:31:42 mreimer_: sunday, if matlab lets me finish my bachelor thesis tomorrow afternoon Apr 22 21:31:45 ~lart matlab Apr 22 21:31:45 * ibot gives matlab a good seeing to Apr 22 21:31:51 ~lart matlab again Apr 22 21:31:51 * ibot whips out a sword and chops matlab again in half Apr 22 21:32:03 ~lart octave as well Apr 22 21:32:03 * ibot steals octave as well's mojo Apr 22 21:32:08 * zecke hands a free copy of octave to koen Apr 22 21:32:10 hmm Apr 22 21:32:12 koen: I hope your thesis goes well. Apr 22 21:32:34 mreimer_: just highlighting code and adding graphs + images Apr 22 21:32:48 koen: ah good, easy stuff Apr 22 21:32:54 indeed Apr 22 21:33:09 bbiab Apr 22 21:34:01 RP: ipaq-pxa270.conf has a few lines to rum sumtool on jffs2 images, which might be nice on c7x0 + akita as well Apr 22 21:34:25 koen: Yes please :) Apr 22 21:34:50 EXTRA_IMAGECMD_jffs2 = "; sumtool -i ${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs.jffs2 \ -o ${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs-summary.jffs2 \ -e 256KiB -p" Apr 22 21:35:16 (replace \ with \ + linebreak) Apr 22 21:35:30 koen: Slowly, we're getting some nice speedups into the images :) Apr 22 21:35:38 koen: Try prelink ;-) Apr 22 21:35:48 RP: it's on my list Apr 22 21:35:56 but first I need to get the images to build :) Apr 22 21:36:54 koen: ah. Are you planning to do regular builds of the supported machines? Apr 22 21:37:14 RP: I plan Apr 22 21:37:14 yes Apr 22 21:37:37 RP: we are slowly getting the infrastructure for such tests Apr 22 21:37:47 I keep planning to but never quite get it working :) Apr 22 21:38:07 angstrom will need *lots* of testing by *lots* of users + developers before we can do a "1.0" Apr 22 21:39:10 koen: How are the images looking at present? booting ok now? Apr 22 21:39:24 it boots in to X Apr 22 21:39:40 with copious amounts of segfaults along the way Apr 22 21:39:53 :-/ Apr 22 21:40:26 * koen sets his alarm to 6 am Apr 22 21:40:30 'night all Apr 22 21:40:34 g'night koen Apr 22 21:40:38 * mickeyl wanders off for some tv Apr 22 21:40:40 n8 koen, n8 all Apr 22 21:40:40 'night koen|sleep Apr 22 21:40:48 'night CoreDump|afk Apr 22 21:40:51 g'night CoreDump|afk Apr 22 22:14:05 03rw 07org.oe.oz354fam083 * r6b3ee943... 10/packages/gpe-plucker/ (gpe-plucker/gpe-plucker-desktop.patch gpe-plucker_0.2.bb): gpe-plucker: adjust naming in .desktop ("E-Book Reader"). part of a fix for hh.org Bug #1531 from .dev Apr 22 22:21:31 good nite Apr 22 22:21:39 night zecke Apr 22 22:27:59 'night zecke Apr 22 22:28:22 'night zecke Apr 22 22:29:35 RP: thx for the python depends fi Apr 22 22:29:36 fix Apr 22 22:30:01 Unfortunately, it brings me to the next one: Apr 22 22:30:06 rwhitby: np. The RDEPENDS changes have thrown a few bugs out the woodwork :) Apr 22 22:30:10 ERROR: Nothing provides runtime dependency glibc-gconv-iso8859-1 Apr 22 22:30:10 NOTE: no buildable providers for bogofilter Apr 22 22:30:48 Note that this one only happens in the uClibC build, not the glibc build. Apr 22 22:31:02 (UcSlugC, not OpenSlug or DebianSlug) Apr 22 22:31:35 rwhitby: The question is why is it depending on something glibc based? Apr 22 22:31:44 yeah, good question. Apr 22 22:31:47 rwhitby: Does uclibc have an equivalent? Apr 22 22:32:05 RP: dunno. jbowler is our uClibC build expert Apr 22 22:32:57 rwhitby: That .bb file has explicit RDEPENDS on two glibc related items. This could mean it can't be built with uclibc - I don't honestly know Apr 22 22:33:27 One for a uclibc person I guess :-/ Apr 22 22:33:32 yup Apr 22 22:33:43 are there generic libc versions of those depends? Apr 22 22:33:56 (this used to build, it's not a newly added package) Apr 22 22:34:44 uclibc.inc seems to PROVIDE virtual/libiconv Apr 22 22:34:52 rwhitby: I'm not aware of generic libc versions but its a bit outside my area of knowledge... Apr 22 22:35:18 rwhitby: Even if it used to build the package, did it install/work? Apr 22 22:35:34 rwhitby: If the rdepends were like this, I doubt it would have installed? Apr 22 22:35:59 RP: good question. jbowler put it in the UcSlugC packages list (which is hand-picked - it's a small subset of all OE packages) Apr 22 22:36:18 And I assume it worked/installed before the big RDEPENDS change. Apr 22 22:36:44 Out of curiosity, what are the build times like for building 1 package from scratch? Apr 22 22:36:55 But you're right - it seems either a bogofilter bug (depending on glibc-specific) or a real dependency that prevents it running on uclibc Apr 22 22:37:04 rwhitby: The big RDEPENDS change only made bitbake care about them - the RDEPENDS of a package like this is the same as before Apr 22 22:37:05 (Assuming all dependencies must also be compiled first) Apr 22 22:37:31 RP: ah, ok. I think I'll just remove it from the uclibc build until someone complains. Apr 22 22:38:26 rwhitby: Whilst it has broken things, its usually been faults in the OE metadata rather than bitbake being wrong :) Apr 22 22:39:04 RP: yep, I don't mean to give the impression that the RDEPENDS changes were bad - I'm fully aware that it's just revealing latent bugs elsewhere. Apr 22 22:39:52 howdy rwhitby Apr 22 22:40:00 hey jacques Apr 22 22:40:10 we still have multiple heads? Apr 22 22:40:19 don't think so Apr 22 22:40:42 we did about 4 hours ago Apr 22 22:40:57 The distributed-caused hydra usually only last until the next ewi sync Apr 22 22:41:01 rwhitby: np. Hopefully your project will be seeing some great simplifications from the RDEPENDS changes :) Apr 22 22:41:24 RP: nslu2-linux has much to thank bitbake and OE for. Apr 22 22:42:05 rwhitby: I think there's been benefits all around :) Apr 22 22:42:15 rwhitby, ok cool, it seems better now Apr 22 22:42:25 Its nice to know people are using the uclibc code for example Apr 22 22:42:31 In fact, it's time we donated again. RP: would you be interested in a developer NSLU2 ? Apr 22 22:42:47 (since we're using your LEDS patches) Apr 22 22:43:15 I guess I should look up exactly what an NSLU2 is! :) Apr 22 22:44:15 heh Apr 22 22:44:18 IXP420, 8MB flash, 32MB RAM, 10/100 ethernet, 2xUSB2.0 host ports, runs three OE-derived distros. Apr 22 22:44:47 (actually, four, if you include Unslung) Apr 22 22:45:51 You realise what your webpage lacks? A description of what it is and a photo! :) Apr 22 22:46:34 RP: http://www.nslu2-linux.org/wiki/Info/HomePage - first two links Apr 22 22:46:39 it's smaller than it looks Apr 22 22:47:30 http://www.nslu2-linux.org/gallery/buildfarms/IMG_0290 for size comparison against a Shuttle box Apr 22 22:47:46 (that's our build farm) Apr 22 22:49:01 rwhitby: Technically, I wouldn't follow a link that said CPUOverview or PhotosOfTheInternals to get an overview of the devcice and neither gives me an external picture :) Apr 22 22:49:12 RP: good point. Apr 22 22:49:20 But now I realise it is effectively a CPU., it makes more sense Apr 22 22:49:38 03rwhitby 07org.oe.dev * r4813ca2e... 10/packages/meta/slugos-packages.bb: slugos-packages: demoted obexftp - fails to build. Apr 22 22:49:44 They do look a lovely size :) Apr 22 22:51:07 they fit just about anywhere Apr 22 22:55:48 ooh Apr 22 22:55:50 i want one =P Apr 22 22:59:30 TheMasterMind1: they do make a good OE development target :-) Apr 22 23:01:43 yep. I don't wanna thing about where the nslu2 project would be today without oe Apr 22 23:02:59 s/g/k/ Apr 22 23:07:00 wow, that was the only g in the sentence Apr 22 23:07:01 03rwhitby 07org.oe.dev * r36623665... 10/packages/meta/slugos-packages.bb: slugos-packages: Demote bogofilter for uclibc-based builds - it has an explicit glibc dependency. Apr 22 23:07:04 the honor is on our side. OE would still be much more handheld-centric if it wasn't for the nslu2 and dreambox projects Apr 22 23:07:43 i just wish someone would take care a bit for the WRT Apr 22 23:07:49 * mickey|tv appends to TODO Apr 22 23:07:54 I guess that's what they call synergy :-) Apr 22 23:08:14 mickey|tv: how is your slug going? Apr 22 23:08:35 People are finding OE and it is getting used by more and more people, some of them quite influential... Apr 22 23:08:39 rwhitby: pretty good. i had the need for some standard usage so I upgraded to unslung 6.8 beta Apr 22 23:09:30 rwhitby: next thing is buying an usb audio box and trying to connect a usb-wlan thing to it to have it streaming audio from my pc to the living room Apr 22 23:09:38 (via wlan) Apr 22 23:09:41 I've always considered oe undervalued. If it were a public corporation I would be buying lots of stock. Apr 22 23:09:55 rwhitby: are there known problems with usb hubs? Apr 22 23:10:01 mickey|tv: the zd1211 is best supported for unslung Apr 22 23:10:06 jacques: The problem is the barrier to entry - documentation Apr 22 23:10:29 mickey|tv: For audio I have a squeezebox Apr 22 23:10:47 rwhitby: I have a dlink dwl150 laying around here which i need to use :) Apr 22 23:10:54 and a slimp3 for that matter... Apr 22 23:10:57 yeah, more docs would be good Apr 22 23:11:07 RP: wow these are nice Apr 22 23:12:09 mickey|tv: I did some development work on them a few years ago. Its a lovely idea... Apr 22 23:12:26 mickey|tv: Funnily enough I fixed their caching problems at the time :) Apr 22 23:12:35 heh, cool Apr 22 23:13:13 I remain impressed they've done the software side all in perl and support windows, mac and linux Apr 22 23:14:30 that piece of hardware rocks Apr 22 23:14:35 pretty cool design Apr 22 23:14:42 what does it run as OS? Apr 22 23:15:17 mickey|tv: That's the bad bit - Its proprietry to the chip in the device :-( Apr 22 23:16:03 mickey|tv: Its basically their own software on top of a cpu vendor supplied TCP/IP stack Apr 22 23:16:22 RP: i see. cool anyway. hmm... 299 USD Apr 22 23:16:23 * mickey|tv tempted Apr 22 23:17:02 I do like mine and their software is nice as well - try it as you can run a software client with it Apr 22 23:18:20 can it play from samba disks containing mp3s as well? Apr 22 23:20:28 ah, nm Apr 22 23:20:28 it can Apr 22 23:20:54 mickey|tv: Its server driven so it can play anything the server can access Apr 22 23:21:30 pretty cool indeed. i wonder why i didn't hear about that earlier Apr 22 23:22:46 I knew about them long before I got invloved with OZ/OE... Apr 22 23:23:04 I'd see if you can find a picture of the orignal SLIMP3... Apr 23 01:28:45 After I build a package, how do I rebuild it? Apr 23 01:29:04 I applied an external patch to the source, and want to rebuild, but it just says it's already built Apr 23 01:35:55 If you are in the BB interactive session you can do rebuild Apr 23 01:37:07 You could also do bb -f Apr 23 01:37:51 I tried that Apr 23 01:37:55 That didn't rebuild it Apr 23 01:38:01 But the interactive one did Apr 23 01:38:36 Unfortunately, it deleated the source I just modified... So how do I go about editing the source, and getting it to not delete it? Apr 23 01:38:44 deleted* Apr 23 01:38:58 Ah, that's what I figured you wanted to do Apr 23 01:43:44 zach: bitbake -b path/to/bb/file.bb -c compile -f Apr 23 01:43:50 that will force the compile step only Apr 23 01:43:55 Thanks Apr 23 01:44:04 you use a second packages dir and override the repo Apr 23 01:44:10 then if you do bitbake -b file.bb, it should notice that the compile step has a newer timestamp than the ones after it and redo the ones after it Apr 23 01:44:31 or that Apr 23 01:44:43 that's how i do it anyway. once it works create a .patch and stick it in SRC_URI Apr 23 05:12:17 Hrm. Apr 23 05:12:22 NOTE: package osb-jscore-0.5.0+cvs20060212-r0: task do_fetch: failed Apr 23 05:12:55 Does anyone else have incessant problems with sf.net anoncvs? Apr 23 05:34:52 I just managed to build an patch kphonepi, and it works... But it seems like the QtE stuff is all wrong... It looks different (The theme doesn't seem to be applied to it), and there's no buttons in the title bar Apr 23 05:35:00 and* Apr 23 05:46:45 Necronom, everyone on the planet does Apr 23 06:09:44 How do I package something I've built? I built with bitbake -b my.bb -c compile -f, but how do I build the .ipk? Apr 23 06:11:04 zach: Don't tell it to run a specific task, just let it do its thing Apr 23 06:11:23 Then it will run the "package" task :p Apr 23 06:11:44 treke|home, the problem is I modified the source, and don't want to clean the source, and reextract, and build Apr 23 06:12:01 which is what happens when I call rebuild Apr 23 06:12:13 then run the package task Apr 23 06:13:57 Hmm Apr 23 06:14:05 That's not listed when I run interactive, but it works Apr 23 06:14:39 ah. I didnt even know there was an interactive mode :p Apr 23 06:49:27 Morning all. Apr 23 07:10:24 03rwhitby 07org.oe.dev * rcf432ace... 10/packages/busybox/busybox-1.01/slugos/sysctl.conf: busybox/slugos: Reinstated the fix to sysctl.conf which seems to have been reverted by someone. Apr 23 07:13:13 Problems with the latest .dev build (pull from 22.04.2006) and my c860. AlsaMixer does not provide left and right mixer any more. The device has no sound ;-( Any ideas ? Apr 23 07:42:56 Hi all Apr 23 07:43:00 03rwhitby 07org.oe.dev * rf57df095... 10/packages/linux/ (ixp4xx-kernel/2.6.16/defconfig ixp4xx-kernel_2.6.16.bb): ixp4xx-kernel: Put required modules for usb storage back into the kernel. Apr 23 07:43:54 Have problems with alsamixer on my cl860. No sound at all after updateing to 2.6.16 kernel (.dev feed from 22.04.2006) Apr 23 07:44:01 Anyone here who can help ? Apr 23 08:01:29 Have problems with alsamixer on my cl860. No sound at all after updateing to 2.6.16 kernel (.dev feed from 22.04.2006) Apr 23 08:01:30 Anyone here who can help ? Apr 23 08:52:50 Morning Apr 23 09:03:07 03rwhitby 07org.oe.dev * r21e5c1bd... 10/packages/linux/ (2 files in 2 dirs): ixp4xx-kernel: Fixed the name of the disk led on the nas100d. Apr 23 09:23:04 morning Apr 23 09:26:00 morning Coredump Apr 23 09:31:10 RP: 4th projects: http://venge.net/monotone/wiki/SummerOfCode2006, the perl cvs thingie :) Apr 23 09:33:03 koen: Interesting although I'd prefer svn :) Apr 23 09:39:10 Trac with a monotone backend would be nice. Especially if it supported the nice trac changeset viewing. Apr 23 09:40:40 Hmm. First patch from pdaxrom - reboot handling for the cxx00. Doesn't quite do what sashz told me and isn't quite right :-/ Apr 23 09:42:18 RP: does the latest LEDS timer trigger use delay_on/delay_off or frequency/duty ? Apr 23 09:42:25 RP: Dunno, rebooting appears to work fine Apr 23 09:42:59 RP: didn't you fix that a week ago already? Apr 23 09:43:36 koen: Yes, and properly ;-) Apr 23 09:43:41 rwhitby: delay_on/delay_off Apr 23 09:43:59 rwhitby: Otherwise the conversions had too many nasty corner cases Apr 23 09:44:18 RP: thx - looks like I've got an nslu2-linux leds script to modify then. Do you have a nice formula to convert freq/duty to delay_on/delay_off ? Apr 23 09:45:07 rwhitby: There was one in the driver... :) Apr 23 09:45:19 delay_on = duty and delay of = 1/(2 * freq) ? Apr 23 09:46:44 delay_on = (50 - (50 - duty))  1000 / 50  freq and delay_off (50 + (50 -duty))  1000 /50  freq Apr 23 09:49:02 er, (duty - 50) in the inner brackets even... Apr 23 09:49:17 I was wondering about that :) Apr 23 09:52:11 writings essays suck Apr 23 09:53:27 maybe I should apply for some SoC stuff :P Apr 23 09:54:15 mithro: please do Apr 23 09:55:22 sadly last year my 4 applications got rejected :/ Apr 23 09:55:53 ah, that's too bad Apr 23 09:57:00 i might still apply for some of the OE/monotone ones as it would mean I could spend quite a bit more time on them **** ENDING LOGGING AT Sun Apr 23 09:59:56 2006