**** BEGIN LOGGING AT Sun Jan 22 02:59:57 2006 Jan 22 03:01:24 good morning all Jan 22 03:01:53 good morning koen ;) Jan 22 03:01:58 hey gremlin[it] Jan 22 03:02:23 RP: do you need a sshv1 or sshv2 key for git? Jan 22 03:02:38 luke-jr_: I've never heard of the c760 having an inbuild mic before... Jan 22 03:02:53 koen: sshv2 please although I can manage with v1 :) Jan 22 03:03:06 hi Jan 22 03:03:43 koen: we would need to talk about mirroring OZ from ewi to mirrors Jan 22 03:04:27 hi hrw Jan 22 03:04:40 morning all btw :) Jan 22 03:04:56 RP: how can I get git tree from ewi? Jan 22 03:05:23 hi RP, koen, hrw Jan 22 03:05:30 hey Bernardo & hrw Jan 22 03:05:32 hrw: You have git installed? Jan 22 03:05:32 hi Jose Jan 22 03:05:39 RP: git, cogito Jan 22 03:06:10 want to test qgit on something other then kernel Jan 22 03:06:18 hrw: Something like "git clone http://ewi546.ewi.utwente.nl/OE/other_scms/git/org.openembedded.dev someplace" Jan 22 03:06:46 ~lart Debian maintainer of git for lack of deps Jan 22 03:06:46 * ibot teaches Debian maintainer of git the basics, including how to RTM for lack of deps Jan 22 03:06:49 hrw: and/or send me an ssh key and I can give you access to oegit.rpsys.net (anonymous access isn't sorted yet) Jan 22 03:07:01 /usr/bin/git-clone: line 26: curl: command not found Jan 22 03:07:18 RP: ok Jan 22 03:07:48 fetching. Jan 22 03:08:57 RP: sent Jan 22 03:10:27 RP is oegit.rpsys.net a unidirectional or bidirectional mirror of monotone? Jan 22 03:10:39 RP: how long it takes to fetch git tree first time? Jan 22 03:11:20 hrw: However long it takes you to download ~150MB of data Jan 22 03:12:16 rwhitby: Unidirectional unless zecke can talk his conversion program into running the other way Jan 22 03:12:45 ok, so is anyone relying on that repo, or can it be used for testing random pushes? Jan 22 03:13:01 RP: ok Jan 22 03:13:03 rwhitby: I'll probably add an experimental branch we can mess around with commits yet continue to import history Jan 22 03:13:15 got it Jan 22 03:13:20 RP: ok, may I send you an ssh key? Jan 22 03:13:27 rwhitby: feel free Jan 22 03:14:05 rwhitby: It only contains .dev at the moment but we're working on importing the other branches Jan 22 03:14:34 rwhitby: I'm not too worried about breaking it as worst case we throw it again and start properly if we do switch to git Jan 22 03:14:53 RP: I only need oe.dev, and zecke has converted o.n-l.dev Jan 22 03:15:23 what is the recommended version of git? Jan 22 03:15:31 koen@bitbake:~/OE/git$ git --version Jan 22 03:15:31 git version 1.0.6 Jan 22 03:16:37 rwhitby: We really need to merge o.n-l.dev and the oz/fam release branch into one git repository but I'm working on this Jan 22 03:16:53 koen: I'm using a snapshot from 20060118 Jan 22 03:17:12 03jbowler 07org.oe.dev * rbf04667b... 10/packages/linux/ (11 files in 2 dirs): ixp4xx-kernel: merge in changes from CVS to 2.6.15.1 Jan 22 03:17:13 RP: o.n-l.dev is not analogous to o.oe.dev - it's the nslu2-linux MasterMakefile and about 7 other files only. Jan 22 03:17:34 nslu2-linux uses org.oe.dev directly for OE files Jan 22 03:18:18 we also have an org.nslu2-linux.bitbake monotone branch, in which we keep the bitbake that matches the head of org.openembedded.dev Jan 22 03:18:39 (so you can pull both with the one monotone pull and two updates) Jan 22 03:19:15 rwhitby: Ah, I understand. They sound like they might be separate repositorys then Jan 22 03:19:39 (in git terms) Jan 22 03:20:07 yes, org.nslu2-linux.dev would be a separate repository from org.openembedded.dev Jan 22 03:20:21 (and zecke has done that) Jan 22 03:20:33 http://ewi546.ewi.utwente.nl/OE/other_scms/git/org.nslu2-linux.dev/ Jan 22 03:20:37 ok, so it looks like I just need to merge the oz/fam branch in... Jan 22 03:20:59 yeah, all nslu2-linux stuff is always merged. Jan 22 03:24:26 someone tried xfce-image on device recently? Jan 22 03:25:11 hrw: I only tried the xfce stuff inside gpe-image Jan 22 03:26:52 Greetings ! Jan 22 03:28:15 I am trying to build oe for x86 but it looks like glibc is broken during the toolchain compile Jan 22 03:28:59 tried a newer glibc as was suggested in some messages but still does not work. any ideas ? Jan 22 03:30:42 e-image is totally broken in branch Jan 22 03:32:16 hrw, koen, rwhitby : Keys should all be added and working Jan 22 03:32:25 RP: thanks Jan 22 03:32:58 is there a git frasebook to make monotone command to git commands? Jan 22 03:33:05 Ifaistos: it's difficult to make suggestions without knowing what the actual problem is. Jan 22 03:33:05 koen: Not yet :) Jan 22 03:33:17 koen: We need to start one :) Jan 22 03:34:26 pb: that's true :p Jan 22 03:34:29 koen: I'll make a start on a page Jan 22 03:35:03 should i post the build results here or is there a flood bot ? Jan 22 03:35:19 ~pastebin Jan 22 03:35:26 pastebin is, like, a place to paste your stuff without flooding the channel - try http://pastebin.com/, or http://pastebin.ca Jan 22 03:37:11 http://pastebin.com/517254 Jan 22 03:37:36 03hrw 07org.oe.oz354fam083 * rb0b92ccb... 10/packages/dbh/dbh_1.0-18.bb: Jan 22 03:37:36 dbh: fix staging Jan 22 03:37:36 taken from .dev Jan 22 03:38:06 RP: thx Jan 22 03:38:24 Ifaistos: I think 20050917 only works if you select nptl rather than linuxthreads Jan 22 03:39:55 i found a message related to this Jan 22 03:39:58 http://www.nabble.com/building-OE-on-Fedora-Core-4-t360150.html Jan 22 03:41:07 so in the process i made my first .bb :) Jan 22 03:41:15 Ifaistos: heh Jan 22 03:41:25 try adding GLIBC_ADDONS = "nptl" to your local.conf, see if that helps Jan 22 03:41:55 might need GLIBC_EXTRA_OECONF = "--with-tls" Jan 22 03:41:55 as well Jan 22 03:44:26 let me give it a try... Jan 22 03:46:34 hrw, koen, rwhitby: http://oe.handhelds.org/cgi-bin/moin.cgi/GitScmTrial?action=show Jan 22 03:47:14 I've added some example commands there. Keep in mind I don't understand git myself ;-) Jan 22 03:48:44 Eventually, I'll probably switch to git-shell as the default shell on the server but for now, I'll leave it as bash Jan 22 03:49:01 back shortly... Jan 22 03:51:32 are people using git-core or cogito ? Jan 22 03:52:04 unable to write sha1 filename ./objects/13/7a7a8a0b842a6391312642ba8058a42463cf68: Permission denied Jan 22 03:52:08 fatal: failed to write object Jan 22 03:52:24 when doing git push ssh://koen@oegit.rpsys.net/home/oegit --all Jan 22 03:55:42 RP, koen: added a shortcut to the wiki page for pulling from different remotes Jan 22 03:55:56 e.g.: git pull rpsys Jan 22 03:55:58 cool, thanks Jan 22 03:56:54 that pulls rpsys's repo into an rpsys branch Jan 22 03:57:13 Not 100% sure about the "Pull: " line Jan 22 03:58:55 btw guys. I made some notes on what packages you need on amd64/ubuntu to get oe going. Jan 22 03:59:38 should i put them on the wiki or is it too specific ? Jan 22 04:00:02 the wiki would be fine Jan 22 04:01:48 RP: any idea why the push would fail? Jan 22 04:03:39 RP: I suspect I lack write permissions to /home/oegit Jan 22 04:09:17 koen: You are a member of the group and the group should have write access... Jan 22 04:10:59 koen: Ah, my push means the g+w bit was lost. I think we need to do something with the umask Jan 22 04:13:01 pb: it now fails while build glibc-intermediate Jan 22 04:13:06 http://pastebin.com/517282 Jan 22 04:13:10 RP: there's something about +x on the update hook Jan 22 04:13:16 let me find it ... Jan 22 04:15:18 yep, chmod +x /home/oegit/.git/hooks/update Jan 22 04:17:00 See Advanced Shared Repository Management in http://www.kernel.org/pub/software/scm/git/docs/tutorial.html Jan 22 04:19:09 hmm - that might be post-update instead Jan 22 04:21:07 Ifaistos: oh dear. you'll have to inspect the configure script to find out what version of gcc it wants. Jan 22 04:21:12 it might be happier with 3.4 rather than 4.0 Jan 22 04:25:52 pb : thanks. it looks like it :( Jan 22 04:31:24 RP: ok, first push seems to have worked :) Jan 22 04:31:33 RP: you wouldn't have any gitweb running, would you? Jan 22 04:33:02 koen: not yet. I have that in mind though Jan 22 04:34:54 koen: I think I just need to add umask 775 to the .profile's of users... Jan 22 04:35:04 hi zecke Jan 22 04:35:09 one more question :) Jan 22 04:35:25 the BBFILES line should be like this ? Jan 22 04:35:29 BBFILES:= ${HOME}/oe/org.openembedded.dev/packages/*/*.bb Jan 22 04:35:43 looks about right, yeah Jan 22 04:35:51 this is for the local.conf Jan 22 04:36:56 is there a way to scan only arch specific stuff ? now its something like 3000+ files ... Jan 22 04:37:47 no, not really Jan 22 04:38:50 i think i need to switch to a type of coffee that takes longer to make :) Jan 22 04:39:29 heh Jan 22 04:43:02 ok, permissions should be ok now Jan 22 04:45:11 let's say that i get oe to build for the x86. How should the patches be applied upstream so that they don;t break other archs ? Jan 22 04:45:36 basicaly how do i make them arch specific Jan 22 04:46:37 grrr error again : NOTE: bb.build.FuncFailed:("can't read from file '%s' (%s)", ('/mnt/hda/oe/build/tmp/staging/h3600-linux/kernel/kernel-abiversion', )) while evaluating: Jan 22 04:46:37 ${@base_read_file('/mnt/hda/oe/build/tmp/staging/h3600-linux/kernel/kernel-abiversion')} Jan 22 04:48:23 re Jan 22 04:48:33 hrw: wb Jan 22 04:48:39 hi pH5 hrw Jan 22 04:49:11 mhhh i don't have any kernel-abyversion file !!! Jan 22 04:52:06 mhhh i don't have any kernel-abiversion file !!! Jan 22 04:52:16 !!!!! Jan 22 04:52:17 pb___: Don't know, don't care Jan 22 04:52:29 tch Jan 22 04:52:32 * pb___ shoots cdbot Jan 22 04:53:06 iihihhii :) Jan 22 04:54:00 what i can do ... i alreasy update bitbake and oe :( ... Jan 22 04:55:02 dunno. it sounds like kernel.bbclass must be busted in some way. Jan 22 04:55:45 i never touch it :( :( Jan 22 04:59:25 It sounds like some module is attempting to build before the kernel itself Jan 22 05:00:55 all the modules (theoretically) depend on virtual/kernel, so that shouldn't be possible Jan 22 05:04:16 hi Jan 22 05:04:30 hi CoreDump|home Jan 22 05:07:28 hi everyone Jan 22 05:08:57 * CoreDump|home yawns Jan 22 05:11:49 i completelly wipe out tmp dir .. so i'm sure i don't have something strange .... Jan 22 05:15:53 last packege build is linux-hotplug ... Jan 22 05:16:14 cu Jan 22 05:16:18 but i can't figure out what is the package that fail :( Jan 22 05:21:14 mickey|dinner: I have my flight booked Jan 22 05:21:42 zecke: cool Jan 22 05:22:47 zecke: rock. what time do you arrive? Jan 22 05:23:10 pb___: 22h on friday :} Jan 22 05:23:31 zecke: in time to enjoy some Leffe :) Jan 22 05:24:23 okay, very good. at Charleroi, or some other airport? Jan 22 05:24:38 good question Jan 22 05:25:48 pb___: Brussel International Airport (BRU) Jan 22 05:26:30 ah cool Jan 22 05:26:53 BRU is much nearer to the city centre than CRL Jan 22 05:29:15 so, yeah, you should arrive early enough that mickeyl will not yet have consumed all the Leffe Jan 22 05:32:19 03nail 07org.oe.dev * r897657fc... 10/packages/zd1211/ (10 files in 3 dirs): zd1211: Remove old driver, add new (maintained) driver Jan 22 05:39:18 03pH5 07org.oe.dev * rf9b46277... 10/packages/blueprobe/ (blueprobe-0.17/hx4700.patch blueprobe_0.17.bb): blueprobe: add 0.17 Jan 22 05:45:09 RP: did I just push some refs? Jan 22 05:48:31 zecke: I just noticed that I can't say 'qt/e sucks' anymore Jan 22 05:48:47 I now have to say 'qtopia-core sucks' Jan 22 05:50:36 koen: "X [Windowing System] is braindamaged" Steve Jobs at USENIX conference :} Jan 22 05:51:20 zecke: yes Jan 22 05:51:35 zecke: Well, some objects and you updates master Jan 22 05:51:38 RP: did you have any luck with merging the branch Jan 22 05:52:06 wow Jan 22 05:52:10 git is way to chatty Jan 22 05:52:18 zecke: I've not tried again yet Jan 22 05:52:32 RP: what is master and what is origin? Jan 22 05:52:48 zecke: I'm fighting gitweb atm. The machine doesn't have perl :) Jan 22 05:53:35 zecke: I think origin is a pointer to where the tree originally came from (who its parent is) Jan 22 05:53:47 how do I switch branches? Jan 22 05:54:22 zecke: checkout -b xyz I think Jan 22 05:57:19 * zecke needs to learn latexbeamer now Jan 22 05:59:18 RP: could you give me the objectid of the branch point again? Jan 22 06:12:31 zecke: 1e2bf5270c6510c14e523733a278873a3f00b19f Jan 22 06:12:46 zecke: http://oe.handhelds.org/cgi-bin/moin.cgi/GitScmTrial?action=show Jan 22 06:12:52 zecke: Add notes :) Jan 22 06:13:04 RP: doing a checkout for the branch at the branch point Jan 22 06:17:45 and please add your experiences to http://oe.handhelds.org/cgi-bin/moin.cgi/GitScmTrial Jan 22 06:18:45 zecke: what scms can we test atm? Jan 22 06:19:59 koen: darcs as well Jan 22 06:20:06 03nail 07org.oe.dev * rf17d3087... 10/packages/qc-usb/qc-usb-messenger_1.1.bb: qc-usb-messenger: Upstream upgrade (0.8 to 1.1) Jan 22 06:20:21 zecke: is ghc buildable in OE? Jan 22 06:20:25 koen: if mithro finds a way to resolve the tunnel issue I will have a darcs server as well Jan 22 06:20:31 koen: dunno Jan 22 06:20:32 zecke: if not, we can skip the test all together Jan 22 06:20:43 zecke: whats the tunnel issue? Jan 22 06:21:07 mithro: the tunnel didn't get back up? Jan 22 06:21:09 zecke: there's no point in testing stuff we can't use Jan 22 06:21:19 (we = nslu2 in this case) Jan 22 06:21:31 koen: well we can add it... Jan 22 06:21:32 zecke: the tunnel appears to be up Jan 22 06:22:07 the uml appears dead however Jan 22 06:22:13 RP: merging seems hard as well Jan 22 06:22:23 mithro: yes it is halted or such Jan 22 06:22:52 what cause it to halt? Jan 22 06:22:59 did you resize your terminal? Jan 22 06:23:11 UML's dont like it when you do that Jan 22 06:23:19 LOL Jan 22 06:24:57 http://oegit.rpsys.net:8000/cgi-bin/gitweb.cgi Jan 22 06:26:01 hmmm Jan 22 06:26:06 RP: hmm Jan 22 06:26:08 git uses a bogus author Jan 22 06:26:22 koen: well it doesn't handle the SIG that is sent on terminal resize Jan 22 06:26:42 no way all time same error ... what u need to trace this issue ? Jan 22 06:27:03 zecke: do you want to kill it and restart? Jan 22 06:27:10 mithro: will do Jan 22 06:30:57 btw darcs is getting git backend support.... Jan 22 06:34:16 mithro: I need to create some slides first (I need to turn in them tomorrow) Jan 22 06:36:13 RP: what is the difference between git fetch and pull? Jan 22 06:36:40 zecke: git fetch sends a mail to Larry to lart him Jan 22 06:36:46 zecke: pull = fetch+merge? Jan 22 06:37:13 * RP -> food Jan 22 06:37:14 koen: what is git lart-larry then Jan 22 06:37:36 zecke: the same, but with a signed-off by: line Jan 22 06:41:01 please take a look : http://pastebin.com/517412 Jan 22 06:42:33 hmm Jan 22 06:43:12 is there any documentation for keylaunch / keylaunchrc beside the readme? Jan 22 06:45:38 got it, nm Jan 22 06:49:58 * CoreDump|home looks at xserver-common and shudders Jan 22 06:50:32 heh Jan 22 06:50:52 what the hell, couldn't make it harder to at xmodmaps Jan 22 06:51:20 now I need to check gpe.bbclass just to find out where the damn thing is downloaded from Jan 22 06:52:02 and every xmodmap for every device is installed into the rootfs? Jan 22 06:52:27 * CoreDump|home sighs Jan 22 06:53:34 CoreDump|home: add the keymap and adjust Xinit.d/12keymap Jan 22 06:53:40 that isn't too hard, right? Jan 22 06:54:55 it isn't hard but it's annyoing to read a bbclass just to find out how to do such a trivial thing Jan 22 06:55:03 ? Jan 22 06:55:06 the .bb is basically empty Jan 22 06:55:22 adds somepatches, no do_install etc Jan 22 06:55:35 why would you need to read a .bbclass? Jan 22 06:55:46 RP: do you have limited bandwidth? Jan 22 06:56:09 hi Jan 22 06:56:11 zecke: no, he _is_ the internet Jan 22 06:56:26 please look at the .bb... it's useless in its current form if you do not want to redo it completly Jan 22 06:57:09 everything is done through "inherit" gpe Jan 22 06:57:21 that's the point of classes Jan 22 06:58:05 the easiest way is to make a modmap and send it to Florian Jan 22 06:58:08 told ya, need to study the bbclass to do anything with that bb Jan 22 06:58:19 pb___: question about GPE bugs in OE bugtracker: should reporters be pointed to GPE bugtracker and bugs closed as INVALID or should those bugs be re-assigned to gpe@handhelds.org or other way? Jan 22 06:58:44 hrw: half of them are matchbox bugs and should go into o-hand bugzilla Jan 22 06:58:46 CoreDump|home: bbclasses are to make recipes simploer Jan 22 06:58:55 I understand that Jan 22 06:59:14 but unless you know the bbclass inside out, they make them harder to modify / adjust Jan 22 06:59:22 koen: yep - but now I ask about gpe bugs like #627 Jan 22 06:59:26 which was my point =) Jan 22 07:00:05 CoreDump|home: that's not true, you want to add a patch, so you don't needs to read the .bbclass, just a SRC_URI += Jan 22 07:00:17 hrw: I don't recommend closing them as INVALID, since this would be unfriendly to the reporters. Assigning them to the person who maintains the appropriate .bb files is probably the right answer. Jan 22 07:00:33 heh Jan 22 07:00:34 ok - best way Jan 22 07:00:41 not according to reenoo Jan 22 07:01:02 but I won't make any further comments on that Jan 22 07:01:24 what would be even better is finding out how to transfer them to the handhelds.org bugzilla, but I don't know how to do that. Jan 22 07:01:41 koen: I want to add a new file, not a patch against an existing one. While this _could_ be done with a patch, this is surely the worst way to do it in OE. Not to mention that the bbclass installs every xmodmap for every machine into every rootfs=) Jan 22 07:01:54 ~lart canonical for not opening malone Jan 22 07:01:55 * ibot wallops canonical with a main rotation server that needs rehubbing. It won't take long for not opening malone Jan 22 07:02:23 anyways, I'll write the xmodmap and send it to florian as you suggested Jan 22 07:02:32 hrw: #627 is actually an OZ bug Jan 22 07:02:44 CoreDump|home: "SRC_URI += ..." is also the correct way to add new files. They don't have to be patches. Jan 22 07:03:02 koen: marking valid bugs as invalid is impolite by definition. and again, stop talking shit behind my back Jan 22 07:03:15 SRC_URI is useless if you do not know how do_install works in the bbclass Jan 22 07:03:26 RP: how do I do a merge without knowing the revs? Jan 22 07:03:40 that's true, but the .bbclass isn't a large or complicated file. It doesn't seem unreasonable that you need to read it. Jan 22 07:03:48 re after rebot Jan 22 07:04:29 Makes thing more complicated for outsiders. That was the whole point I was trying to make. Jan 22 07:04:38 koen: how to resolve that? Jan 22 07:04:51 fair enough Jan 22 07:04:58 hrw: take a look at the familiar-version.bb Jan 22 07:05:10 koen: will do Jan 22 07:05:17 hi, have problems in bitbake bootstrap-image in org.oe.oz354fam083 Jan 22 07:05:25 NOTE: Task failed: Fetch failed: http://www.openswan.org/download/old/openswan-2/openswan-2.2.0-kernel-2.4-klips.patch.gz;patch=1 Jan 22 07:05:25 NOTE: package collie-kernel-24-8-2.4.18-rmk7-pxa3-embedix-r22: task do_fetch: failed Jan 22 07:05:32 hrw: or stop OE from putting "openembedded' in /etc/issue / motd /something Jan 22 07:05:35 our busybox need love... "ping: permission denied (are you root?)" Jan 22 07:05:38 any ideas? Jan 22 07:06:02 koen: I have "OpenZaurus 3.5.4 \n \l" in /etc/issue Jan 22 07:06:14 hrw: yeah, I noticed that yesterday as well. I suspect the right answer is probably to stop using busybox ping, in fact. Jan 22 07:07:02 heh. indeed. you don't want an suid root busybox Jan 22 07:07:18 pb___: looks like Jan 22 07:07:31 * reenoo back to work Jan 22 07:07:32 I'll open bug for it Jan 22 07:07:37 thanks Jan 22 07:08:15 ping being broken should probably be a blocking bug for familiar 0.8.3 Jan 22 07:08:30 it works if you are root Jan 22 07:08:37 so it's not completely broken Jan 22 07:09:57 true, but gpe-mininet (which doesn't run as root) requires a working ping Jan 22 07:10:24 and, generally, users wouldn't expect to need to be root in order to use ping. Jan 22 07:10:45 RP: hmm merging sucks with git (at least ATM) Jan 22 07:11:00 Bug 629 has been added to the database Jan 22 07:12:07 thanks Jan 22 07:12:12 what are the "few other tools" you mentioned? Jan 22 07:12:27 koen: looked into familiar-version. OPIE sysinfo looks at /etc/familiar-version /etc/oz-version /etc/issue so why not gpe sysinfo? :) Jan 22 07:12:48 pb___: I'm checking now on device Jan 22 07:12:57 okay, cool Jan 22 07:12:58 hrw: gpe-conf looks at something that says 'OpenEmbedded' I guess Jan 22 07:13:00 pb___: traceroute for example Jan 22 07:13:13 hrw: but shipping an oz-versions would be a good idea anyway Jan 22 07:14:27 actually, it looks like gpe-conf says "OpenEmbedded" if /etc/version exists (irrespective of its contents). Jan 22 07:14:32 that seems a little bit broken. Jan 22 07:14:41 heh Jan 22 07:14:51 indeed Jan 22 07:15:26 it's overridden by /etc/familiar-version or /etc/debian_version, which is why it wouldn't be showing up on Familiar, but I think the bug is in gpe-conf. Jan 22 07:15:46 so florian lied ;) Jan 22 07:16:52 well, he might have been mistaken. or, for that matter, it is quite possible that he was correct and I am mistaken. Jan 22 07:18:25 bb.build.FuncFailed: ("can't read from file '%s' (%s)", ('/mnt/hda/oe/build/tmp/staging/h3600-linux/kernel/kernel-abiversion', )) Jan 22 07:18:51 i don't have build/staging/h3600-linux ;( :( ... Jan 22 07:19:49 that is very unfortunate Jan 22 07:20:31 yes ... Jan 22 07:21:55 which is kernel for famliar-unstable hh37.4 or hh40 ? (kernel 2.4) Jan 22 07:22:26 lol Jan 22 07:22:53 was playing with X on my Z when I needed to shut the Xserver down. Guess what I did? Press Ctrl-Alt-BS on my latop *sigh* Jan 22 07:23:46 CoreDump|home: ;)) Jan 22 07:23:47 heh Jan 22 07:23:59 gremlin[it]: 40, I think Jan 22 07:24:15 koen: gpe-sysinfo is in which gpe-tool contained? Jan 22 07:24:23 hrw: gpe-conf Jan 22 07:24:32 k Jan 22 07:24:38 gremlin[it]: 37.4 for sa, hh39 for pxa Jan 22 07:27:49 * RP returns Jan 22 07:28:04 zecke: The upstream link from the git server is only 256MBit Jan 22 07:28:24 RP: Mbit?!?! Jan 22 07:28:35 KBit :) Jan 22 07:28:39 aaaah Jan 22 07:30:03 so people should pull from ewi (100Mbit) ? Jan 22 07:31:43 I have fucked my tree again :} Jan 22 07:33:04 the git tree? Jan 22 07:33:16 koen: I want a git-unpull Jan 22 07:33:21 that doesn't sound too encouraging Jan 22 07:33:34 koen: the issue is if you start a pull and needs manual merging Jan 22 07:33:41 koen: you have almost lost Jan 22 07:33:47 that doesn't sound too encouraging Jan 22 07:34:06 please add that to the experiences Jan 22 07:34:13 koen: will do Jan 22 07:34:47 zecke: Is this the oz/fam branch merge? Jan 22 07:35:19 zecke: git-revert; git-prune should fix that is it doesn't do what was intended. Jan 22 07:35:40 RP: but why is there no interactive merge tool? Jan 22 07:36:02 I think we're missing something about doing branching deep in the past. Its not exactly something we'd be doing all the time either... Jan 22 07:36:22 zecke: Good question. I'm sure there will be a way to handle it... Jan 22 07:37:00 RP: we should have a oz354 branch now in git Jan 22 07:37:15 zecke: You managed to import it? Jan 22 07:37:59 dunno, I did -s ours on the pull Jan 22 07:38:31 zecke: Its certainly there... Jan 22 07:38:58 * CoreDump|home hugs NFS root Jan 22 07:39:12 now I would like to see the gitk output... Jan 22 07:39:21 anyway if we switch SCM I need to hack tailor... Jan 22 07:41:23 * pb___ heads to the office to reconfigure the phones again Jan 22 07:41:24 bbl Jan 22 07:55:46 ~curse my desktop machine Jan 22 07:55:51 May the fleas of a thousand camels infest your most sensitive regions, my desktop machine ! Jan 22 08:50:11 RP: ping Jan 22 08:50:58 hey mithro Jan 22 08:54:08 automatic cherry pick failed :} Jan 22 08:55:12 doh Jan 22 08:58:22 03hrw 07org.oe.oz354fam083 * r52a13468... 10/packages/linux/ (4 files): openzaurus, opensimpad 2.4 kernels: fixed SRC_URI for openswan patch Jan 22 09:01:11 zecke_, I am puzzling out apache and ssl this morning Jan 22 09:01:27 Hopefully I have a svn repo later. Jan 22 09:01:34 Do I need to worry about svk? Jan 22 09:03:22 Crofton: no ;) Jan 22 09:03:31 Crofton: it just works on top of svn Jan 22 09:04:07 zecke_: you have 20 minutes with any problems Jan 22 09:04:14 plus, http://www.cafepress.com/thousandparsec/ Jan 22 09:05:03 mithro: sadly I'm busy with preparing the slides... Jan 22 09:08:05 anyone got a link to the new OZ "upgrades" folder? Jan 22 09:10:33 zecke_, right answer :) Jan 22 09:10:49 RP: want to get another sd/mmc subsystem sources? Jan 22 09:11:09 zecke_: how are you getting to FOSDEM, plane? Jan 22 09:12:45 i would like to bring a poster to FOSDEM, but I don't have anyway to print it before then, do you think you would be able to print one of for me and bring it? Jan 22 09:14:31 03hrw 07org.oe.dev * ra51640b1... 10/packages/linux/ (4 files): openzaurus, opensimpad 2.4 kernels: fixed SRC_URI for openswan patch Jan 22 09:16:19 welp i'm going to head off to greece now! Jan 22 09:16:22 have fun everyone! Jan 22 09:18:02 zecke_: i'll send you an email about the poster Jan 22 09:23:39 mithro: don't get lost! Jan 22 09:27:19 http://oegit.rpsys.net:8000/cgi-bin/gitweb.cgi?p=.git;a=commit;h=25c0148f30400c0bcf4408f13856d34857adc81f Jan 22 09:27:23 that was an easy pick Jan 22 09:28:47 hrw: I'm always interested in that... Jan 22 09:29:40 RP: sashz (pdaX) worked on GPL mmc system from montavista - my 1G card is working with it on c760 (under 2.4.20) Jan 22 09:30:45 RP: http://home.hrw.one.pl/tmp/mmc-pdaXrom-corgi-2.4.20.tar.bz2 Jan 22 09:31:02 hrw: 403 Jan 22 09:31:18 RP: reload Jan 22 09:32:38 RP: md5sum of my 1G sd card is same on Z and desktop Jan 22 09:34:32 hrw: got it, thanks. Comparing the two versions isn't going to be easy :-( Jan 22 09:34:51 RP: I know Jan 22 09:35:10 RP: I plan to write about it to lakml Jan 22 09:35:29 hrw: Good idea. rmk might be able to see the problem more easily than me Jan 22 09:35:36 sure Jan 22 09:35:42 RP: if not him then Pierre Jan 22 09:35:58 hrw: Yes, Pierre seems keen on mmc :) Jan 22 09:36:23 hrw: I'll have a look through it if I get a chance but I'm not sure when that'll be :-( Jan 22 09:36:31 ok Jan 22 09:39:13 RP: the amount of strcmp and strcpy in git makes me sick... will be fun to review Jan 22 09:41:00 RP: want Cc: or lakml is enough for you? Jan 22 09:46:40 zecke_: not good :-( Jan 22 09:46:56 hrw: lakml is fine, thanks Jan 22 09:46:58 RP: sent Jan 22 09:47:05 hrw: cc pierre though Jan 22 09:47:11 RP: of course Jan 22 09:47:19 cc: pierre, rmk, sashz Jan 22 09:47:29 RP: git cherry-pick is nice Jan 22 09:47:47 hrw: sounds good :) Jan 22 09:48:01 zecke_: Yes, I thought that might be useful Jan 22 09:48:07 zecke_: rebase looks fun as well Jan 22 09:50:19 RP: I used git-cat-file blob, git-file-ls --unmerged and git-update-index manually... Jan 22 10:17:59 03nail 07org.oe.dev * r208f8ee4... 10/packages/wpa-supplicant/ (6 files in 2 dirs): wpa-supplicant: Upstream upgrade to 0.4.7, add Zydas zd1211-support Jan 22 10:18:03 03nail 07org.oe.dev * re8db3dc6... 10/ (2 files in 2 dirs): Add wpa-supplicant to slugos feed Jan 22 10:18:38 rwhitby, jbowler-zzz: what's your plan for EABI? Jan 22 10:20:07 after 0.8.3 is out I should try to add ekiga to OE Jan 22 10:26:00 koen: my current assumption is that EABI might fix the thumb stuff, so I will ignore it for a while (because if it does it will be very broken initially). Jan 22 10:26:59 I'm tempted at the moment to stop doing anything with uclibc/thumb and just try to fix glibc/arm. Jan 22 10:27:04 jbowler-zzz: Let's see what the debian people find out when they switch Jan 22 10:27:23 My guess is that they will switch back PDQ Jan 22 10:27:54 Doesn't EABI require gcc 4.x? Jan 22 10:28:34 either 4.1 or a patched 3.4, yes. Jan 22 10:30:03 Ah... patched 3.4 might be ok for them, but I got the impression that 4.x was actually not compiling a lot of code without patching. Jan 22 10:30:24 including gcc 3.4 itself :) Jan 22 10:30:41 4.0 has been the default in debian for a while, I think Jan 22 10:31:32 a few packages don't build with it, but it isn't a large enough number to lose any serious amount of sleep over Jan 22 10:31:32 rwhitby seemed to be using it, so maybe they've overcome the problems - I guess on the packages which actually work on ARM matter. Jan 22 10:31:46 s/guess on/guess only/ Jan 22 10:31:47 jbowler-zzz meant: rwhitby seemed to be using it, so maybe they've overcome the problems - I guess only the packages which actually work on ARM matter. Jan 22 10:32:18 Worth a try then, if it doesn't compile gcc it gets my vote Jan 22 10:32:42 tch. I do wish riker would turn that "s/" thing off again, it's very annoying. Jan 22 10:34:21 Which glibc version has EABI support? Jan 22 10:34:32 cvs > 20051214 Jan 22 10:35:22 So that would be 2.3.5+cvs20051214? Jan 22 10:35:25 http://dominion.kabel.utwente.nl/koen/pda/files/angstrom-eabi-r7.diff.bz2 was my latest attempt to add it to OE Jan 22 10:36:30 that predates the glibc cvstree change Jan 22 10:36:38 moving the cvsdate to januari should fix that Jan 22 10:36:42 * pb_ going home now Jan 22 10:36:43 bbiab Jan 22 10:36:47 cu Jan 22 10:38:05 Is that on a branch in the OE servers? Jan 22 10:38:54 zecke_: oegit shouldn't have an origin branch, just a master :-/ Jan 22 10:38:59 no, a diff against .dev (20051216) Jan 22 10:39:29 RP: did I push an origin? Jan 22 10:39:33 :} Jan 22 10:41:43 zecke_: No, I created it when I cloned ewi :) Jan 22 10:41:56 ah hehe Jan 22 10:42:03 git branch -d origin Jan 22 10:42:56 jbowler: I had hoped it that eabi stuff would be easy to add and apply it to .dev Jan 22 10:43:04 zecke_: The idea is you never develop on a branch you pull from. by default git clone creates an orgin and a master. When fetching, it pulls from the remote master into origin, then applies those changes to the local master... Jan 22 10:44:18 origin is gone. That was surprisingly painless :) Jan 22 10:46:24 RP: I've got a set of patches which make an ixp4xx LED driver and then have three board level patches to use this for the LEDs on the specific boards (loft - an avila, nslu2 and nas100d, the latter not tested yet). Jan 22 10:46:50 Seems to all work fine - current NSLU2 monotone head builds use it for the startup stuff. Jan 22 10:47:03 * koen anxiously awaits his loft Jan 22 10:47:13 Oh, in fact, all slugos builds do so long as there is an LED labelled 'ready' Jan 22 10:47:44 hmm. time to watch some kind of movie Jan 22 10:47:49 Patches are in: packages/linux/ixp4xx-kernel/2.6.15/9[46]*.patch Jan 22 10:47:57 jbowler: Excellent :). I need to make some small tweaks to the led patches, then I'll repost to LKML and try to work out what mainline are going to do... Jan 22 10:48:09 s/9[46]/9[56]/ Jan 22 10:48:10 jbowler meant: Patches are in: packages/linux/ixp4xx-kernel/2.6.15/9[56]*.patch Jan 22 10:49:06 RP: you are welcome to add those patches if you wish. The 96* patches are dependent on board level patches we have already, but the 95 (generic) one should be clean. Jan 22 10:49:25 RP: let input use led system.. Jan 22 10:49:34 I did modified the led_device name field to be (const char*) (from (char*)) so that I could use the LED resource name. Jan 22 10:50:13 Hi all Jan 22 10:50:20 hi dirk Jan 22 10:50:22 hey do13 Jan 22 10:50:33 Hi koen, hrw Jan 22 10:50:34 hi do13 Jan 22 10:50:39 Hi Richard Jan 22 10:50:45 jbowler: Were there any other changes? Jan 22 10:51:01 No Jan 22 10:51:32 jbowler: If they apply to code in mainline and you're happy for them to be submitted, it makes the patches looks slightly less zaurus biased :) Jan 22 10:51:34 A change to include/linux/leds.h, a new leds-ixp4xx-gpio.c in drivers/leds and the appropriate Makefile/Kconfig Jan 22 10:52:02 I'm happy - they've got the Signed-off-by line in already Jan 22 10:52:15 excellent - I'm just monotone pulling :) Jan 22 10:53:36 I considered, but didn't implement, a generic "struct resource gpio_resource" so that different GPIO pin users wouldn't grab the same pin, the resource stuff is currently 'blind' at present (i.e. the 'parent' pointer remains NULL). Jan 22 10:53:57 I didn't want to start making changes to the ixp4xx core code. Jan 22 10:54:50 That is a bit ugly in theory on the pxa devices as well. In practice, the drivers are so hardware specific, I've never seen a conflict... Jan 22 10:55:22 Except for the serial port mess... Jan 22 10:56:33 Yes, the primary reason for using a resource is just to pass in the information from the board saying which pins are connected to which LEDs (and whether they are ACTIVE_HIGH or ACTIVE_LOW). Jan 22 10:57:07 ixp4xx boards differ so much in the connections that the approach you took would result in massive code duplication Jan 22 10:58:28 There is a fair bit of duplication between corgi and spitz. I decided the two drivers weren't worth trying to make common code for though... Jan 22 10:59:09 Right, and it avoided having to define a way to pass the information through platform_data Jan 22 11:01:15 There's still some data-space duplication in each machine description because each one has a separate platform_device (though only one), but that's inherent in the architecture. Jan 22 11:03:28 I'm going to remove the buzzer "LED" from 96-nslu2-leds.patch - it works (with timer and frequency about 10) but it duplicates functionality in another module (one which had got compiled out). Jan 22 11:03:45 RP: I really begin to hate the git merging skills... Jan 22 11:03:52 jbowler: Looking at the code, is the error handling in the probe function correct? If a registation fails, you return without unregistering any previously registered LEDs? Jan 22 11:04:14 Look at the call immediately before the 'return rc' Jan 22 11:04:37 It sets 'name' on the one which failed to NULL then calls remove which iterates through everything just added. Jan 22 11:04:48 Saves code over a while --i loop Jan 22 11:05:34 ok. time to make some OZ branding work for opie/gpe Jan 22 11:05:45 jbowler: Yes, I see :) Jan 22 11:06:28 jbowler: It looks good :) Jan 22 11:08:46 RP: and then one understands it... git merging... Jan 22 11:09:24 bbl Jan 22 11:09:26 guy i found the reason of my problem, a package is build before kernel ...but it need kernel :( :( :( ... Jan 22 11:09:49 RP: I just realised that it would be possible to use command line options too in the board level, just allocate (struct resources[16]) with empty names and fill in the names from the command line, led:ready=3 for example. Jan 22 11:10:04 Or led:3=ready Jan 22 11:10:18 i build kernel by hand (bitbake -b ) and i go further .... now another issue but i try Jan 22 11:12:55 * hrw goes away - that fscking desktop crashes too often Jan 22 11:12:56 jbowler: You could do that. That isn't much use on something like the zaurus but it might work for you... Jan 22 11:13:41 RP: it works badly for NSLU2 (where the command line can't be changed) and well on Loft, where it can and where there is a general purpose GPIO header on the board. Jan 22 11:14:15 Well, something for the future - it's not necessary for me atm Jan 22 11:15:22 We can't change the CMDLINE easily on the zaurus without recompiling... Jan 22 11:16:39 jbowler: One other tiny thing - you've used : in the jumper names for loft. I was aiming to allow : as a separator in the name to dnote properties like led colour Jan 22 11:17:06 RP: ok - I used '-' elsewhere so I'll change that. Jan 22 11:17:42 In fact, I'll remove the J8 stuff, it was only in there because I had a bug with multiple resources which was easier to debug on Loft (there is only one LED on loft) Jan 22 11:17:55 And the command line option is better Jan 22 11:17:58 It was that or have a ton of get attributes that returned things like colour which looked like a waste of space to me Jan 22 11:18:33 RP: hmm git-merge-index calls git-merge-recursive with the wrong arguments Jan 22 11:18:49 git-merge-index -o is not working Jan 22 11:19:18 zecke_: On your machine or on oegit? Jan 22 11:19:29 RP: you should probably not attempt to push the Loft patch, because I think it has got to the point where Deepak will want it outside ixdp425-setup.c Jan 22 11:19:29 on my machine ;) Jan 22 11:19:30 or everywhere? :) Jan 22 11:19:43 RP: just for people that are too used to GNU utils Jan 22 11:19:57 jbowler: I'll just take the general driver and leave the machine specific bits to you Jan 22 11:19:58 RP: and the nslu2 and nas100d patches prove that the code works on multiple ARM systems. Jan 22 11:20:12 evening Jan 22 11:20:16 hail mickeyl Jan 22 11:20:19 jbowler: is the nas100d a recommendable device? Jan 22 11:20:21 hi pb_ Jan 22 11:20:23 Right - that's what I assumed, I figured you might want to show an example of the driver use though. Jan 22 11:20:24 mickeyl: hey, I vae my flight booked Jan 22 11:20:29 zecke_: very good Jan 22 11:20:36 jbowler: Do you mind if I import the const leds.h change into my main led patches, just so it looks tidier? Jan 22 11:20:38 mickeyl: I believe it is discontinued, and mine doesn't boot yet. Jan 22 11:20:46 hi mickeyl Jan 22 11:20:49 RP: no problem Jan 22 11:20:56 zecke_: you'll go straight to Rue d'espagne and we can handle your luggage then, ok? Jan 22 11:21:01 Roy D'espagne even Jan 22 11:21:03 hi rp Jan 22 11:21:15 jbowler: oh, ok. thanks. i guess the 200d is too expensive for me, it's interesting hardware though Jan 22 11:21:49 hrw|gone: ping Jan 22 11:22:29 Hi mickeyl Jan 22 11:23:13 hey do13, how are things? Jan 22 11:23:14 mickeyl: they discounted the price a lot in the US, then it was worth buying Jan 22 11:23:37 ah yeah Jan 22 11:23:47 like the SIMpad. it was a joke @ 1000EUR, but cool @ 300 Jan 22 11:24:07 mickeyl: I will try to get there then... maybe I need to get a palm+tomtom to find the way Jan 22 11:24:38 RP: okay the merging capabilities are promising, not too well documented but not worse than csv Jan 22 11:24:48 <<<<<<< .merge_file_V91P3t Jan 22 11:24:48 h3600 | h3900 | h1940 | h6300 | h2200 | ipaq-pxa270) Jan 22 11:24:48 ======= Jan 22 11:24:48 h3600 | h3900 | h1940 | h6300 | h2200 | ipaq-pxa270 | blueangel) Jan 22 11:24:49 >>>>>>> .merge_file_7CAIfO Jan 22 11:24:50 zecke_: if you can afford the 25 EUR just take a taxi from the airport to the grand place. the roy d'espagne is directly there Jan 22 11:25:22 mickeyl: yeah will do so... FOSDEM will be my birthday present Jan 22 11:25:33 zecke_: ah right, what's the exact date? Jan 22 11:26:01 mickeyl: I believe NAiL saw a rebadged version from someone else (not sure about this), that might imply IOMega sold the design to someone, but of course the price then goes back up... Jan 22 11:26:03 somewhen this year ;) Jan 22 11:26:14 ljp: How could I run Opie on Qtopia 1.7? Jan 22 11:26:42 zecke_: heh. tell me :) Jan 22 11:26:49 ljp: I would really appreciate it, you once look at library... Jan 22 11:26:53 mickeyl: 26th of January Jan 22 11:27:06 ah that soon Jan 22 11:27:08 I will get 14 :} Jan 22 11:27:12 heh no way Jan 22 11:27:18 24 ? Jan 22 11:27:29 23 Jan 22 11:27:34 hmm *sigh* Jan 22 11:27:38 all these young folks Jan 22 11:27:39 *fucking old* Jan 22 11:27:42 hah Jan 22 11:27:43 oops Jan 22 11:28:21 zecke_: That looks better. Given it can create files for comparision for a merge, hacking git to launch meld on the files shouldn't be impossible... Jan 22 11:28:37 RP: hehe can you hack shell scripts? Jan 22 11:29:02 zecke_: yes ;-) Jan 22 11:29:06 woot Jan 22 11:29:30 maybe we should open a OE git tree... Jan 22 11:29:57 You mean an anonymous one? Jan 22 11:30:15 RP: OE version of 'git' Jan 22 11:30:27 ah, right :) Jan 22 11:30:29 RP: to use lua scripting to call meld, kdiff3 and such Jan 22 11:30:54 mickeyl: I have bought a book about Death March Projects Jan 22 11:30:55 zecke_: I suspect if we wrote nice patches, it would get accepted into git mainline Jan 22 11:31:11 RP: they are welcome to pull from our tree Jan 22 11:31:23 mickeyl: bussy Jan 22 11:32:00 Summary: Merging OZ3.5.4 with HEAD is not working automatically (yet) Jan 22 11:32:06 mickeyl: Did you booked your flight to Brüssel? Jan 22 11:32:33 do13: yep. done. I'll arrive Friday at 20:00 and will depart Sonday at 19:00 Jan 22 11:33:02 mickeyl: Which airline? Jan 22 11:33:34 do13: Lufthansa, LH 4582 / LH 4587 Jan 22 11:34:00 oh, i'll arrive 17:15 on friday... leaves me more time to checkin Jan 22 11:34:41 do13: 194,87 EUR in total. Jan 22 11:35:16 might get something cheapear with other airlines, but I support the local economy Jan 22 11:35:17 heh Jan 22 11:36:11 mickeyl: Yep I found the same price. Jan 22 11:37:46 mickeyl: Maybe I can book the same flight. Which hotel did you booked? Jan 22 11:38:00 do13: astridhotel.be Jan 22 11:38:22 do13: I added the two patches of yours (bl and pm) to the oe kernel tree btw. My c3000 still appears to be working although I haven't had a chance to full test yet Jan 22 11:38:26 do13: 400 meters from the Grand Place - a good location for the evenings so the way to bed is not far :) Jan 22 11:38:50 mickeyl: thanks Jan 22 11:38:54 np Jan 22 11:41:42 RP: OK. If the bl works we can push this patch. Jan 22 11:42:03 jbowler: do you know if anyone has sound working with the nslu yet? I'm still looking for a custom built audio streamer device in my home Jan 22 11:42:39 RP: The pm needs more changes. I have a newer patch in my tree. Jan 22 11:43:12 do13: Can you make sure to reference include/asm-arm and not include/asm? Jan 22 11:44:22 RP: oops. Jan 22 11:47:14 no way ... i got at end the same error --> ERROR: Nothing provides (2.4.19-rmk6-pxa1-hh37) Jan 22 11:47:35 do13: Feel free to push the bl patch (and add a signed-off-by line from me). I've just double checked the c3000 and its fine Jan 22 11:48:10 do13: I've been caught out by the symlinks before ;-) Jan 22 11:48:14 mickeyl: do you know where I might find some pre-release 3.5.4 gpe images for my shepherd? Jan 22 11:48:58 zecke, you about? Jan 22 11:51:36 RP: Thanks. I'll do this next weekend. This week I'am on a business trip. Jan 22 11:51:48 Crofton: yes Jan 22 11:52:50 https://ossie-dev.mprg.org/repos/oe/ Jan 22 11:52:59 need to work out access controls though Jan 22 11:53:25 Crofton: good start Jan 22 11:54:11 I'm going to remember the acess stuff now Jan 22 11:54:11 pb__: I think you'd need to ask hrw|gone where he's got the test images stashed :) Jan 22 11:54:20 okay Jan 22 11:54:27 hrw|gone: where have you got the test images stashed? Jan 22 11:54:28 http://ossie-dev.mprg.org:8080/timeline Jan 22 11:54:29 pb_: i think hrw did some.... now where is the url... Jan 22 11:54:36 zecke, can you make a login here? Jan 22 11:54:49 The password setup stuff isn;t ssl though Jan 22 11:54:50 RP: I'm having kdiff3 called from merge-one-file Jan 22 11:54:52 pb__: I'd guess somewhere on ewi546... Jan 22 11:55:13 Crofton: I got the listing without any pw Jan 22 11:55:18 zecke_: I wondered about that... Jan 22 11:55:49 yeah, but I suspect you can't write :) Jan 22 11:58:22 RP: ah, thanks Jan 22 11:58:59 pb__: I'm not sure that's for public release but I doubt hrw will mind you trying one ;-) Jan 22 11:59:16 google know everything :) Jan 22 11:59:20 heh Jan 22 11:59:26 okay, excellent Jan 22 11:59:37 maybe I should try the opie images as well, just to see what life is like on the other side Jan 22 11:59:49 Crofton: should I try to commit some stuff? Jan 22 12:00:34 can someone remind me what the magic key sequence is to flash a shepherd? Jan 22 12:00:57 well, you can try Jan 22 12:01:04 but I think something is broke ... Jan 22 12:01:08 I am trying now Jan 22 12:01:57 Crofton: which password? Jan 22 12:02:09 Merging back to mirror source https://ossie-dev.mprg.org/repos/oe. Jan 22 12:02:10 Authentication realm: OSSIE Jan 22 12:02:10 Password for 'ich': Jan 22 12:03:47 RP: do you know how to discard a cset? Jan 22 12:04:22 zecke, you can make a username at http://ossie-dev.mprg.org:8080/timeline Jan 22 12:04:42 don't use a good pw if you are paranoid as the setup isn't over ssl Jan 22 12:04:53 the rest of it should be good though Jan 22 12:06:13 Crofton: how do I know if this pw is valid? Jan 22 12:06:26 Crofton: I'm getting a 500 Internal Server Error Jan 22 12:06:30 smae here Jan 22 12:06:47 give me a few Jan 22 12:06:51 minutes Jan 22 12:07:23 hmm I need to do an assignment now Jan 22 12:07:27 heh Jan 22 12:07:30 so do I Jan 22 12:07:36 progress report crap for some pref Jan 22 12:07:41 er professor Jan 22 12:07:47 I'll get repo running first Jan 22 12:07:53 very close Jan 22 12:09:31 thats better, now I just forbidden :) Jan 22 12:10:22 ah, here we go Jan 22 12:10:24 * pb__ flashing Jan 22 12:11:24 Crofton: right Jan 22 12:12:24 now I just need to figure out per directory access file voodoo Jan 22 12:12:37 hm, this doesn't look good Jan 22 12:12:44 I still have an upside down "gpe 2.6" splash screen Jan 22 12:13:26 oh, heh, forgot to rename the rootfs Jan 22 12:18:05 ah, now I seem to be completely hosed Jan 22 12:18:09 Crofton: 500 error Jan 22 12:19:29 kernel panic :-( Jan 22 12:19:37 ah but you can read :) Jan 22 12:21:01 zecke_: I'm not sure Jan 22 12:24:54 Crofton: forbidden again Jan 22 12:24:59 yeah Jan 22 12:25:07 give me a little while Jan 22 12:25:13 don't hurry Jan 22 12:25:17 need to figure out where the issue lies Jan 22 12:25:20 heh Jan 22 12:25:27 I will SWING a bit Jan 22 12:25:33 cool Jan 22 12:26:04 I have it working on a non ssl machine .. Jan 22 12:29:57 can someone w/ commit access please enable IPv6 and SSL in curl? Jan 22 12:33:00 03coredump 07org.oe.dev * rb6b0df14... 10/packages/linux/ (4 files in 2 dirs): Jan 22 12:33:01 Akita / Borzoi / Spitz: Jan 22 12:33:01 - defconfig changed to include iptables, tun and ipv6 support. (Closing Bug #455) Jan 22 12:33:01 - Kernel tested on Akita, no problems but ipv6 and tun modules are auto-installed into the rootfs Jan 22 12:34:22 pb__: http://ewi546.ewi.utwente.nl/tmp/hrw/oz3.5.4-test-images/ Jan 22 12:34:27 morning Jan 22 12:34:35 hrw: thanks Jan 22 12:34:53 I tried the gpe-image from there, but my test was... unsuccessful. Jan 22 12:35:02 hrw: morning. what's the time frame for 3.5.4? Let me know when I can tag and do the world build. I did it two weeks ago, but since then many changes hit oz354fam083 Jan 22 12:35:46 mickeyl: this changeset which coredump just did need propagate Jan 22 12:36:38 mickeyl: I have patch for gpe-conf to show 'distribution:openzaurus', have to add openzaurus-version package (which just add /etc/openzaurus-version), patch for libopie so opie-sysinfo also use /etc/openzaurus-version Jan 22 12:37:12 mickeyl: there are some bugs in bugtracker which would be nice to get closed - like 'composite in gpe' one Jan 22 12:37:49 pb__: Any idea what the problem was? Jan 22 12:38:15 RP: "VFS: Unable to mount root fs on unknown-block(31,2)" Jan 22 12:38:16 mickeyl: and we have to decide about mirroring - it was easy on sf.net where anyone can get ssh account and run own software Jan 22 12:38:39 http://bugs.openembedded.org/show_bug.cgi?id=620 Jan 22 12:38:53 GPE has a nifty little tool for that ^ ^ Jan 22 12:39:13 CoreDump|home: on you shots I saw opie-sh script for it.... Jan 22 12:39:17 your Jan 22 12:39:27 heh right Jan 22 12:39:34 DEPENDS: opie-sh Jan 22 12:39:49 I have similar one for 2.6 usb-gadget control Jan 22 12:39:55 and cpufreq Jan 22 12:40:04 cpufreq? neat! Jan 22 12:40:21 hmm... i don't know why all insist on /etc/$DISTRO-version. I thought using /etc/issue was enough *shrug* Jan 22 12:40:30 pb__: Did you use the rootfs.tar.bz2 file or the initrd.bin file as initrd.bin? Jan 22 12:40:51 RP: ah, whoops, the .tar.bz2 Jan 22 12:40:54 I guess that was a dim plan Jan 22 12:41:44 mickeyl: I was thinking like that too - our images get oe-version which gpe-conf/sysinfo takes to tell 'disttro: OE' Jan 22 12:41:51 pb__: We've all done it :) Jan 22 12:42:11 heh Jan 22 12:42:14 * pb__ try again with the right file Jan 22 12:42:35 heh.. I aired this room and cpu goes to 40C.. soon will be 55C anyway Jan 22 12:43:13 mickeyl: for your interest, the list of outstanding familiar 0.8.3 bugs is at http://handhelds.org/~bugzilla/showdependencytree.cgi?id=1334 Jan 22 12:43:19 I guess some of those probably apply to oz as well Jan 22 12:44:37 1459 and newer for sure Jan 22 12:45:09 pb_: i see. thanks. yeah. Jan 22 12:45:17 "useless distinction between "Utilities" and "Other"." Indeed heh Jan 22 12:46:52 is there a magic sysrq key on modern zaurus kernels? Jan 22 12:47:06 (or some other way to reboot a crashed machine without taking the battery out) Jan 22 12:47:25 1364 is fixed Jan 22 12:47:29 python-datetime Jan 22 12:47:38 rock Jan 22 12:47:59 pb__: right japanese+home == sysrq Jan 22 12:48:21 I had a "clever plan" for avoiding 1385 by generating the locales at build time using qemu, but unfortunately it seems that qemu's arm emulation is not equal to that task Jan 22 12:48:23 so, such Jan 22 12:48:25 suck, even Jan 22 12:48:38 hmm yeah that's unfortunate. what options do we have left then? Jan 22 12:48:54 hrw: thanks Jan 22 12:49:07 pb__: np Jan 22 12:49:13 it says "SysRq : Resetting", but... it doesn't seem to actually be resetting Jan 22 12:49:21 so, that's not quite as great as I had hoped Jan 22 12:49:22 heh Jan 22 12:49:29 pb__: sometimes it does not work (resetting) Jan 22 12:49:33 ah Jan 22 12:49:37 mickeyl: using NSLU2 as an music server is pretty straightforward (mpd?), people have discussed usb audio output devices on the nslu2-linux mailing list - I think they work, I'm not sure (I haven't followed the discussions). I believe rwhitby has also looked at things which have built in audio output capability. Jan 22 12:50:26 mickeyl: no easy options that I can think of. Jan 22 12:50:33 jbowler: i guess i could just buy a usb audio device - their cheap these days Jan 22 12:50:55 obviously one could either (a) fix qemu so it works, or (b) fix localedef to use less ram, but neither of those sound very straightforward Jan 22 12:51:00 pb__: Each iteration of zaurus does different things in the bootloader upon a watchdog reset (which is what the kernel does to reset by default). c760 works, c700 was a bit weird and I've never had a c750 to see that that does Jan 22 12:51:35 pb__: what does not work with qemu? Jan 22 12:51:43 zecke_: qemu segfaults Jan 22 12:52:02 oh Jan 22 12:52:14 yeah Jan 22 12:52:44 where i can look to solve this issue ??? ERROR: Nothing provides (2.4.19-rmk6-pxa1-hh37) ERROR: dependency (2.4.19-rmk6-pxa1-hh37) (for alsa-driver) not satisfied Jan 22 12:52:48 mickeyl: could you take care about WE18 and wireless-tools? 594, 595 and check 538? Jan 22 12:52:58 hrw: sure. i'm on it as we speak Jan 22 12:53:55 great Jan 22 12:54:35 gremlin[it]: Looks like a bug in alsa-driver - it should probably be something like kernel-image-2.4.19-rmk6-pxa1-hh37 Jan 22 12:54:59 03mickeyl 07org.oe.dev * r655180ca... 10/packages/wireless-tools/ (wireless-tools_28-pre6.bb wireless-tools_28-pre13.bb): Jan 22 12:54:59 wireless-tools: upgrade 28pre6 to 28pre13 Jan 22 12:54:59 the header includes have been removed, so the fixheader patch is no longer necessary Jan 22 12:55:04 excellent, now it really is booting Jan 22 12:55:10 * pb__ waits patiently for S99configure to run Jan 22 12:55:19 I wonder what takes so long in there. Jan 22 12:55:20 gremlin[it]: alsa-driver is probably at fault with your earlier build error as well Jan 22 12:55:21 mickeyl: add it into branch Jan 22 12:56:16 what's our policy again? do i need to add it to our meta-bug or can i commit right away? Jan 22 12:56:41 commit Jan 22 12:56:47 excellent Jan 22 12:56:53 * mickeyl does that Jan 22 12:57:03 heh Jan 22 12:57:11 mickeyl: it's your branch, I guess you can do what you like Jan 22 12:57:27 mickeyl: I think its just us mere mortals that need to use the meta-bug :) Jan 22 12:57:36 hehe Jan 22 12:57:40 always trying to be polite here ): Jan 22 12:57:41 :) Jan 22 12:57:52 hehe Jan 22 12:58:27 -20 degree tomorrow... I think I'm staying home Jan 22 12:58:40 zecke_: -22 here but I have to go to work... Jan 22 12:58:44 RP uhhh the .bb file for alsa driver is quite incomprensible for me ... :( ... Jan 22 12:59:02 03jbowler 07org.oe.dev * r90e5eaef... 10/packages/linux/ (4 files in 2 dirs): ixp4xx-kernel: update LEDs patches, add NSLU2 beeper back in to 2.6.15.1 Jan 22 12:59:31 heh.. forgot about few commits... Jan 22 12:59:38 hm, I see that oz enables the matchbox compositor out of the box Jan 22 12:59:45 bbi1h Jan 22 12:59:50 I'm not sure that's a great plan; it introduces a few nasty artifacts Jan 22 13:00:38 gremlin[it]: It inherits module so it should get the dependencies right :-/ Jan 22 13:01:04 03mickeyl 07org.oe.oz354fam083 * ra5db6230... 10/packages/wireless-tools/ (wireless-tools_28-pre6.bb wireless-tools_28-pre13.bb): Jan 22 13:01:04 wireless-tools: upgrade 28pre6 to 28pre13 Jan 22 13:01:04 (from .dev) Jan 22 13:01:18 pb__: you sure that it does not left from previous installation? Jan 22 13:01:33 ah, maybe it was Jan 22 13:02:06 pb__: thats why collie is my favorite fresh install platform - no /home partition Jan 22 13:02:12 heh, right Jan 22 13:02:31 I forgot /home isn't wiped on the shepherd Jan 22 13:02:42 pb__: it is on ipaq? Jan 22 13:02:52 some of them, yeah Jan 22 13:03:13 iirc, only h3970, h5500 and hx4700 have a persistent /home Jan 22 13:03:31 oh, and h5400 Jan 22 13:04:56 also, the lcd on my shepherd has an interesting "shimmer" effect now. I don't remember seeing that on older versions. Jan 22 13:05:31 echo 1 > /sys/device/platform/w100fb/fastpllclk or whatever it is... Jan 22 13:05:52 * RP -> food Jan 22 13:07:49 yeah, that does make it better Jan 22 13:08:18 it's still there, but not really noticeable anymore unless you're looking for it. That's probably how it was before. Jan 22 13:08:58 flashing collie with more up-to-date image (but not quite up-to-date anyway) Jan 22 13:10:34 zecke, try commiting now Jan 22 13:10:59 hrw: it would be nice to have sylpheed back in the shepherd gpe-image, if there's room for it Jan 22 13:11:25 pb__: how big rootfs is there? Jan 22 13:11:32 Crofton: I'm pushing now... Jan 22 13:11:48 hrw: I have 4264kb free on mine Jan 22 13:12:01 out of 25600kb total, apparently Jan 22 13:12:07 Crofton: but I will need to rm it... Jan 22 13:12:18 and again users will complain about full rootfs... Jan 22 13:12:27 thats fine Jan 22 13:12:34 I ahve a bogus dir in now Jan 22 13:13:55 hrw: how much space do they need to be available before they quit complaining? Jan 22 13:14:39 pb__: dont know - I was looking only at collie where tried to get atleast 1M Jan 22 13:15:26 zecke, can you confirm you have write access? I may need to restart server Jan 22 13:15:31 ah. well, I imagine you should still have at least a couple of megs free after installing sylpheed. it's not _that_ big. Jan 22 13:15:38 Crofton: I'm comitting to it Jan 22 13:15:43 pb__: ok Jan 22 13:15:46 oko Jan 22 13:15:47 Crofton: so I can confirm it Jan 22 13:15:49 ok Jan 22 13:16:03 hrw: incidentally, you could win some space back by ditching openswan, which doesn't even seem to run with the kernel that you ship. Jan 22 13:16:03 just checking if I need to restart server after i update svn access file Jan 22 13:16:32 including ipsec out of the box is awesome, but in this day and age you would probably be better off with racoon and 26sec. Jan 22 13:16:49 pb__: ipsec is gone from images already (but not from test ones) Jan 22 13:16:53 I have killed two birds with one stone today :) Jan 22 13:17:19 hrw: ah, ipsec is removed altogether? that's a shame. Jan 22 13:17:30 pb__: blame koen Jan 22 13:17:44 I was planning to remove it only from collie Jan 22 13:18:29 btw.. who maintain gpe-image/meta-gpe? I added sylpheed there but dont know does it was correct way Jan 22 13:19:17 florian, probably Jan 22 13:20:16 Crofton: I try to svk smergee everything into your repo Jan 22 13:20:22 ok Jan 22 13:20:23 Crofton: this will be a good test for svk as well Jan 22 13:20:49 Crofton: it is way slower than git... (as one can't simply copy the 'revs') Jan 22 13:20:55 ok Jan 22 13:21:22 have you committed yet? I don't see anything ... Jan 22 13:21:34 so this might take some time, if not I did something wrong Jan 22 13:21:39 ok Jan 22 13:21:40 Crofton: the first rev is 34 mb in size Jan 22 13:21:53 and then I hope 3??? will follow Jan 22 13:21:57 we have space and bw Jan 22 13:22:30 ok Jan 22 13:22:41 I see ssl_access_log messages :) Jan 22 13:22:43 * zecke_ has bricked his device Jan 22 13:23:32 with svn+svk, we could use svk to push to main oe repos Jan 22 13:23:38 but use svn locally? Jan 22 13:25:15 pb__: sylpheed is small - added Jan 22 13:25:34 atleast in branch Jan 22 13:25:38 Crofton: otherway around Jan 22 13:25:52 hmm Jan 22 13:25:55 Crofton: you mirror the central OE server and sync it into your depot Jan 22 13:26:03 hi Pigi Jan 22 13:26:08 Crofton: then you do svk cp //oe-central //my-oe-work Jan 22 13:26:10 Ciao all Jan 22 13:26:12 hi hrw Jan 22 13:26:26 Pigi: which ipkg you suggest to use for stable: .154 or .156? Jan 22 13:26:28 but then people can use svn againest my-oe-work? Jan 22 13:26:28 Crofton: you check out //my-oe-work and commits, moves and then you do 'push' Jan 22 13:26:59 Crofton: they could, if you make the svn revs available again Jan 22 13:27:06 I think 156 should be ok, but if pb_confirm the new mods, then I will release a new 157 that would speed up a bit something. Jan 22 13:27:06 Crofton: svk uses 'svn' to store the data Jan 22 13:27:10 basically, we use svn locally and it would be a win for me if I could hide the svk part :) Jan 22 13:27:33 Crofton: then use svk only for merging ;) Jan 22 13:27:38 yeah Jan 22 13:27:48 Crofton: svk smerge //local-group //remote-group Jan 22 13:27:57 Crofton: svk smerge //remote-group //local-group Jan 22 13:29:01 besides hosting the server, I may have to try this Jan 22 13:29:26 although my opinion shouldn't carry too much weight Jan 22 13:29:33 Pigi: ok Jan 22 13:29:44 Crofton: all opinions are equally weighted Jan 22 13:30:20 I just need to scm stuff to calm down, the monotone "blow ups" have me out of sync :( Jan 22 13:30:36 and I spend more time than I should dealing with scm issues than work :( Jan 22 13:30:55 each time we release OZ/Fam we change scm Jan 22 13:30:56 :( Jan 22 13:31:08 pb_ ping Jan 22 13:31:08 monotone might win :) Jan 22 13:31:35 hopefully it gets faster :) Jan 22 13:32:00 hi pigi Jan 22 13:32:23 hi pb__ ! Did you got some time to test the ipkg modified ? Jan 22 13:32:41 I tested it a bit, and it seems fine so far Jan 22 13:33:24 good. I will release it then for the stable image .... Jan 22 13:33:30 the win from the svn point of view is that I already can deal with it pretty good Jan 22 13:33:51 of course, I was figuring out monotone ok also Jan 22 13:33:52 zecke_: we need to patch opie-aboutapplet... it tell about opie v1.2 not v1.2.1 ;) Jan 22 13:34:10 Pigi: okay, great Jan 22 13:34:30 hi Pigi Jan 22 13:34:35 hi gremlin[it] Jan 22 13:36:08 we should have ~1.5M free on collie now Jan 22 13:37:03 maybe this is an issue for u -> ERROR: Nothing provides (2.4.19-rmk6-pxa1-hh37) Jan 22 13:37:04 ERROR: dependency (2.4.19-rmk6-pxa1-hh37) (for alsa-driver) not satisfied Jan 22 13:37:09 Pigi Jan 22 13:37:51 should it be an error for me ? can I ask you why ? Jan 22 13:37:58 s/error/issue Jan 22 13:39:17 maybe if is an issue of alsa-driver ... i try to take a look to the .bb file and it state u are the manteiner Jan 22 13:39:19 :) Jan 22 13:40:15 gremlin[it]: this sounds like it is an issue with bitbake, maybe you still have an old version without the explode_deps code? Jan 22 13:40:35 pH5, i did svn update this morning Jan 22 13:40:37 mickey|bbl: nas100d is in the kernel, but the RTC and IDE support are not there yet. We're waiting for hardware to be delivered to dwery for the IDE support, and the RTC support depends on the new RTC subsystem. Jan 22 13:41:50 well, I used to be the "importer" for alsa, sometime ago, but after that someone other took the bb maintaing. I will try to give a look at it, but I'm not sure I will have much free time in the next days Jan 22 13:42:45 ah ok Pigi ... don't worry :) Jan 22 13:42:53 np Jan 22 13:51:13 hrw, 0.99.156 is out. Jan 22 13:51:18 s/156/157 Jan 22 13:53:19 Pigi: its suggested version to normal use? Jan 22 13:56:09 anyone here have an H5000 series ipaq? Jan 22 13:57:38 hm Jan 22 13:58:14 why would the LCD of my ipaq work with the old familiar 0.8.2 images, but not with the newest images made from OE? Jan 22 13:58:33 kernel changed? Jan 22 13:59:11 what is the newest kernel? Jan 22 13:59:12 heh.. looks like I fscked one pacakge Jan 22 13:59:46 because the kernel of the workable image is 2.4.19-rmk6-pxa1-hh40 Jan 22 13:59:47 03hrw 07org.oe.oz354fam083 * r37a4377a... 10/packages/meta/gpe-image.bb: gpe-image: added sylpheed for bigflash devices Jan 22 14:00:09 zecke_, still ticking along Jan 22 14:00:28 http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=8aba06fecc0a39a0ccb881cbff7993804b834529 - someone has idea how to make it work as I expect? Jan 22 14:02:03 hi....i just got setup for devel on my ipaq h5550. Can someone tell me, what are most developers doing as far as testing/trying out new code. If i have to install it on actual device each time, this seems very tediuos. I feel like there must be an emulator, or way to run a image locally....? Jan 22 14:02:20 Crofton: yes Jan 22 14:02:52 its not using any cpu Jan 22 14:04:30 hrw: is sylpheed working for you? Jan 22 14:05:01 tyz: what do you plan to develop? Jan 22 14:05:21 generally, you can just compile and run the code on your regular computer. Jan 22 14:06:13 pb_: currently i need to work on the wlan driver, mmc in kernel. i'll probably work on things based mostly in the console, not gui Jan 22 14:07:44 argh.. desktop crashed twice in a row... Jan 22 14:09:09 for drivers, you have no option but to test on the real hardware. Jan 22 14:10:52 okay, i guess i'll just copy the new packages over and try my changes as required. I've found bugs in opie/console i'd like to fix.. is there anyway to run/boot opie locally? Jan 22 14:11:02 gremlin[it]: It looks like a version error it wants -hh37 and I suspect something else like -hh40 is built. The kernel-abiversion file in staging is probably therefore stale. Why is the question you'll have to answer Jan 22 14:12:39 CoreDump|home: sylpheed works on my c760 (atleast start) Jan 22 14:12:46 hmmmm Jan 22 14:13:01 doesnt work for me. just freezes Jan 22 14:13:28 check that you don't have the compositor enabled Jan 22 14:13:45 well, built from 354 branch Jan 22 14:14:01 so it's enabled I think Jan 22 14:14:46 *Yeah Fight Club* Jan 22 14:14:48 if you get drop shadows under pop-up menus, it's enabled Jan 22 14:15:00 * pb__ fights zecke Jan 22 14:15:30 pb_: yep, it's active then. Can it be turned off? Jan 22 14:15:53 yeah, use the "visual effects" checkbox under "look and feel" in the settings menu Jan 22 14:16:02 * zecke_ surrenders Jan 22 14:16:08 you may need to restart X after turning it off if everything goes black Jan 22 14:16:22 pb__: you look too much like an english hooligan :} Jan 22 14:16:31 zecke_: appearances are not deceptive Jan 22 14:16:33 ugh Jan 22 14:16:43 X just crapped its pants Jan 22 14:16:55 looks like this? Jan 22 14:16:56 http://www.elvisgrotto.com/fractbx.gif Jan 22 14:17:45 lol Jan 22 14:18:04 ~kill cvs Jan 22 14:18:09 * ibot shoots a inverse neutrino gun at cvs Jan 22 14:18:21 hey...what package provides the kernel? Jan 22 14:18:30 pb__: excellent. that worked, thanks! Jan 22 14:18:46 CoreDump|home: very good Jan 22 14:18:49 hrw: we should probably turn that of in branch Jan 22 14:18:56 as you may gather, the matchbox compositor is a bit buggy Jan 22 14:19:13 CoreDump|home: patches welcome. you know that I'm not gpe user Jan 22 14:19:17 indeed Jan 22 14:19:27 neither am I heh Jan 22 14:19:36 sadly it's been that way for about a year and a half and the matchbox crew don't seem to have much interest in fixing it Jan 22 14:19:56 * CoreDump|home can live w/o drop shadows Jan 22 14:20:14 slows things down anyway Jan 22 14:20:14 pb__: Is that bug in the bugzilla? Jan 22 14:20:44 RP: dunno. I think it pre-dates the establishment of the openedhand bugzilla, but it might be in the handhelds.org one. Jan 22 14:22:00 pb__: I'd suggest adding it to the OH bugzilla, then we can't claim not to know about it ;-) Jan 22 14:22:19 heh Jan 22 14:22:28 yeah, sounds like a good plan Jan 22 14:23:02 I guess I'll have to do some testing to remind myself what the actual bug is. I think it's something along the lines of matchbox just plain forgetting to draw certain windows. Jan 22 14:23:32 iirc, the reason for the sylpheed "freeze" is that it sits there waiting for you to click OK on a modal dialog, but the dialog doesn't actually show up on the screen. Jan 22 14:23:45 if you manage to guess where the OK button is and click there, it will proceed happily. Jan 22 14:23:50 pb__: That sounds sensible. The more comprehensive the bug report, the more likely the problem will get fixed... Jan 22 14:24:38 to be honest, though, the compositor doesn't bring a whole lot of benefit right now so I'm not too bothered about it. Jan 22 14:24:57 ~lart monotone revert for monotone: misuse: unknown path 'packages/opie-aboutapplet/packages/wireless-tools/wireless-tools_28-pre6.bb' Jan 22 14:24:57 * ibot changes monotone revert's permissions to 0777 and tells the world for monotone: misuse: unknown path 'packages/opie-aboutapplet/packages/wireless-tools/wireless-tools_28-pre6.bb' Jan 22 14:25:10 the only real driver for using Composite would be to get support for ARGB windows, and the matchbox compositor doesn't support those anyway. Jan 22 14:26:22 it also produces this unpleasant visual effect when popping up a menu, where the area is briefly filled with garbage before being drawn correctly. Jan 22 14:26:32 RP : only -hh37 is build ... Jan 22 14:26:36 so, all in all, best to turn it off :-} Jan 22 14:26:51 pb__: It does sound like there is a bug there so filing a bug and then changing the default options in OE sounds like the best course of action Jan 22 14:27:56 now i try to build -hh40 and see what happen ... Jan 22 14:28:02 pb__: I know our matchbox .bb files have subtle differences to the ones on OE, I don't know enough about it say whether the ones in OE are more or less correct for OE Jan 22 14:28:27 RP: it'd be interesting to see the diffs Jan 22 14:28:47 afaik, there isn't anything very wrong with the ones in OE, but that's not to say that yours aren't better Jan 22 14:29:37 actually, that's not quite true: OE builds without xsettings, which has always seemed like a bug Jan 22 14:30:11 please, what .bb file handles the kernel? Jan 22 14:30:38 LoDown: any of the ones in the linux directory Jan 22 14:30:53 LoDown, depend on machine but all is in packeges/linux dir Jan 22 14:31:30 ok...thank you Jan 22 14:32:33 pb__: The composite tag is missing for a start :) Jan 22 14:32:46 RP: git-http-fetch just segfaulted :} Jan 22 14:33:35 03hrw 07org.oe.dev * r8846b4fd... 10/packages/opie-aboutapplet/ (2 files in 2 dirs): opie-aboutapplet 1.2.1: let it show OPIE 1.2.1 instead of OPIE 1.2 (CVS backport) Jan 22 14:33:49 ok...this is really weird...older images for my ipaq built with kernel hh40 Jan 22 14:33:53 and the touchscreen worked Jan 22 14:34:10 but the new images (touchscreen not working) build with kernel hh39-r1 Jan 22 14:34:55 zecke_: oops :) Jan 22 14:35:09 pb__: http://sam.rpsys.net/matchbox.diff Jan 22 14:35:43 pb__: You'll have to filter out the CVS->SRC changes but its readable enough Jan 22 14:36:02 Greetings to all ! Jan 22 14:37:13 RP: thanks Jan 22 14:37:36 yeah, most of that looks in order. Jan 22 14:37:46 I wonder why OE uses --disable-cairo with matchbox-keyboard. Jan 22 14:37:53 pb__: I think i found why the the x86/epia do not build. its gcc 4.x Jan 22 14:38:04 Ifaistos: ah Jan 22 14:38:24 is there a way to force oe to use gcc 3.4.5 Jan 22 14:39:19 Ifaistos: I don't think we have 3.4.5. PERFERRED_VERSION_gcc-cross = "3.4.4" in your local.conf should help though. Jan 22 14:39:23 ....? I believe OE uses GCC3 internally by default.... Jan 22 14:39:39 JustinP: It's DISTRO dependent Jan 22 14:39:45 ah yes Jan 22 14:40:28 another question...where is the kernel version being set? I don't see it in machine or distro Jan 22 14:40:50 LoDown: you set preffered kernel and bitbake use newest one by default Jan 22 14:40:57 LoDown: Look for PREFERRED_PROVIDER_virtual/kernel Jan 22 14:41:00 hmm... Jan 22 14:41:10 LoDown: use PREFERRED_VERSION_virtual/kernel = "2.0.36" Jan 22 14:41:16 so...why would hh39-r1 be newer then hh40? Jan 22 14:41:52 hmmm ubuntu has 3.4.5... i'll give it a try and then try to get 3.4.4 from source to compile it if it fails... hope it does not Jan 22 14:42:16 Ifaistos: 3.4.5 should be straight forward to add Jan 22 14:42:51 i'm using breezy you can install 2.95 3.3 3.4 and default it 4 Jan 22 14:43:08 not sure on hoary if they have other vers packaged up Jan 22 14:43:21 CoreDump|home: could you look on #608? Jan 22 14:44:32 hrw: well I'm seeing the same here Jan 22 14:44:50 tyz: the 3.4 is 3.4.5 :( Jan 22 14:44:53 on collie, too Jan 22 14:45:11 CoreDump|home: heh.. Jan 22 14:45:35 sharp - I 'love' you Jan 22 14:45:58 CoreDump|home: pushed your defconfig changes to branch Jan 22 14:46:01 hrw: newest from monotone, or newest on my harddrive? Jan 22 14:46:03 ty Jan 22 14:46:27 Ifaistos: Why the need for 3.4.5? OE should just compile its own compiler and not care too much about what's on your system? Jan 22 14:46:34 LoDown: newest in OE metadata Jan 22 14:46:57 hm...ok Jan 22 14:47:48 it stops when configuring glibc-intermediate :( Jan 22 14:47:51 http://pastebin.com/517282 Jan 22 14:48:14 ahh yeh. i've built opie-image gpe-image successfully with hmm shoot it was 3.4.5 or 3.3 on ubuntu, though i tried again after recent update some things were broken. Jan 22 14:50:34 Ifaistos: Change to a 3.4 based gcc-cross-initial Jan 22 14:54:48 pb__: thanks I have now swap on nbd Jan 22 14:55:01 i am trying with PERFERRED_VERSION_gcc-cross = "3.4.4.... lets see if this works Jan 22 14:55:21 zecke_: rock Jan 22 14:55:40 pb__: now I need to see if it deadlocks Jan 22 14:59:22 hrw: if you need me to do some stuff, please e-mail me, I will be away from irc for the next 3 days Jan 22 14:59:39 koen: cya! Jan 22 14:59:42 koen: ok will remember Jan 22 14:59:49 nop... it did not work... :( Jan 22 14:59:51 cu Jan 22 14:59:56 cu Jan 22 15:00:23 Ifaistos: add PERFERRED_VERSION_gcc-cross-initial = "3.4.4" as well Jan 22 15:08:35 zecke, I guess this is faster than monotone :) Jan 22 15:13:57 Crofton: well I do not think many people will sync from ewi to your server ;) Jan 22 15:14:23 heh Jan 22 15:14:30 does it pull faster? Jan 22 15:14:45 I am just wondering what is slowing things down? Jan 22 15:15:01 Crofton: svn normal speed, svk well network io for the revs Jan 22 15:15:09 yeah Jan 22 15:15:11 Crofton: dunno, could be my uploading speed Jan 22 15:15:38 Crofton: maybe I will take my ethernet cable with to get 100mbit in uni Jan 22 15:15:40 I am on 10MBps ethernet Jan 22 15:15:51 all other links should be good for me Jan 22 15:16:51 The real test is how fast people can make local copies Jan 22 15:18:02 * france is back (gone 14:50:27) Jan 22 15:18:09 what operation are you doing at the moment? Jan 22 15:19:08 svk smerge -B /oe/convert_oe /oe/crofton-svn Jan 22 15:22:17 nite all Jan 22 15:22:19 RP: No luck... :( Jan 22 15:22:48 i had a look and it default to "gcc" which is 4.0.1 Jan 22 15:24:05 Ifaistos: "gcc" or "gcc-cross" (or "gcc-cross-initial") ? Jan 22 15:25:08 no i meant the it compiled gcc-initial with the default gcc Jan 22 15:25:38 wait a sec to get a copy of the build Jan 22 15:25:58 Ifaistos: and that didn't work? I guess 3.4 might not compile with 4.0 :-/ Jan 22 15:26:36 Ifaistos: You might have to make your system default to use a 3.x compiler... Jan 22 15:28:11 i think so too.... do you think if i do a symlink (gcc -> gcc-3.4.5) to a dummy dir and add it infont of the path. to wotk ? Jan 22 15:28:46 Ifaistos: I don't know. You could try... Jan 22 15:29:38 does oe have a "hard" gcc path or its uses the system path to execute ? Jan 22 15:29:51 Ifaistos: system path Jan 22 15:30:15 hmmm it might work then. I'll give it a try Jan 22 15:31:06 here is the build Jan 22 15:31:09 http://pastebin.com/518136 Jan 22 15:31:52 Crofton: https://ossie-dev.mprg.org/repos/oe/ Jan 22 15:31:57 Crofton: first revision comitted Jan 22 15:32:11 cool Jan 22 15:32:26 this is entire oe repo? Jan 22 15:32:44 Crofton: koen's initial commit Jan 22 15:33:48 just the --remoterev option wasn't too smart... Jan 22 15:34:01 heh Jan 22 15:34:19 what now? Jan 22 15:35:21 nothing ;) the commit messages are just altered Jan 22 15:35:28 Ifaistos: That build says it built "gcc-cross-initial-4.0.2". See above where I suggested you add PERFERRED_VERSION_gcc-cross-initial = "3.4.4" as well Jan 22 15:36:08 i have added it to the local.conf. should i added it anywhere else ? Jan 22 15:36:15 france: ping? Jan 22 15:36:33 03hrw 07org.oe.oz354fam083 * rc6bbbbb1... 10/packages/opie-aboutapplet/ (2 files in 2 dirs): Jan 22 15:36:33 opie-aboutapplet 1.2.1: let it show OPIE 1.2.1 instead of OPIE 1.2 (CVS backport) Jan 22 15:36:33 taken from .dev Jan 22 15:36:36 Ifaistos: No, local.conf should have made it build something else... Jan 22 15:36:37 03hrw 07org.oe.oz354fam083 * r5db97260... 10/packages/linux/ (4 files in 2 dirs): Jan 22 15:36:38 Akita / Borzoi / Spitz: Jan 22 15:36:38 - defconfig changed to include iptables, tun and ipv6 support. (Closing Bug #455) Jan 22 15:36:38 - Kernel tested on Akita, no problems but ipv6 and tun modules are auto-installed into the rootfs Jan 22 15:36:40 taken from .dev Jan 22 15:36:48 03hrw 07org.oe.oz354fam083 * rf5cec8ba... 10/packages/openswan/ (openswan-2.1.2/flags.patch openswan_2.1.2.bb): dropped openswan 2.1.2 - not fetchable from upstream Jan 22 15:36:53 03hrw 07org.oe.oz354fam083 * r27955fca... 10/packages/openswan/openswan_2.2.0.bb: openswan 2.2.0: fixed SRC_URI - close #630 Jan 22 15:36:56 03hrw 07org.oe.dev * rd051b493... 10/packages/openswan/ (openswan-2.1.2/flags.patch openswan_2.1.2.bb): dropped openswan 2.1.2 - its not fetchable from upstream Jan 22 15:36:59 03hrw 07org.oe.dev * r674f74b6... 10/packages/openswan/openswan_2.2.0.bb: openswan 2.2.0: fixed SRC_URI - close #630 Jan 22 15:37:04 03nail 07org.oe.dev * r1f08208f... 10/conf/distro/ (debianslug.conf openslug.conf): debian/openslug.conf: Add _EXTRA_RDEPENDS Jan 22 15:45:04 Crofton: could you reinitialize the database again? Jan 22 15:45:12 03mickeyl 07org.oe.dev * r7237d2d3... 10/packages/linux/ (3 files in 2 dirs): openzaurus-sa: upgrade wireless extension API to V18 Jan 22 15:45:17 03mickeyl 07org.oe.dev * r5531936b... 10/packages/linux/ (2 files): zaurus 2.4 kernels: upgrade wireless extension API to V18 Jan 22 15:45:17 hmmm Jan 22 15:45:26 like delete it and make a new one? Jan 22 15:45:34 Crofton: yes :} Jan 22 15:45:45 ok I just wanted to be certain :) Jan 22 15:46:15 done Jan 22 15:48:08 zecke, I am going home now, I'll check back in later Jan 22 15:48:20 Permission denied: Can't create directory '/var/svn/oe/db/transactions/0-1.txn': Permission denied Jan 22 15:48:23 hmm Jan 22 15:48:29 hmm Jan 22 15:48:56 try again Jan 22 15:49:19 must remember to make new repo owned by webserver ...... Jan 22 15:50:04 03tmbinc 07org.oe.dreambox * r3e2d2423... 10/packages/twisted/twisted_2.1.0.bb: twisted: add missing dependencies to python-crypto, zope-interface Jan 22 15:50:08 03tmbinc 07org.oe.dreambox * r71f0f78f... 10/packages/tuxbox/libdvbsi++.bb: libdvbsi++: upgrade to 20060120 Jan 22 15:50:13 03tmbinc 07org.oe.dreambox * ree966b47... 10/conf/distro/opendreambox-1.2.conf: opendreambox: add machine type to feed URI Jan 22 15:50:54 Crofton: thanks :) Jan 22 15:51:01 Crofton: okay now replying each revision Jan 22 15:51:56 03mickeyl 07org.oe.oz354fam083 * re643b984... 10/packages/linux/ (5 files in 2 dirs): Jan 22 15:51:56 zaurus 2.4 kernels: ugprade wireless extension API to V18 Jan 22 15:51:56 (from .dev) Jan 22 16:19:00 Hmm. rmk actually got both my email about the leds subsystem and the resent email asking if he got the first one but hasn't had time to read them yet. Jan 22 16:19:05 * RP sighs Jan 22 16:21:40 heh Jan 22 16:31:29 RP: Found what the problem was.... Jan 22 16:31:48 PREFERED was PERFERED and i copied from here Jan 22 16:32:16 now it compiles 3.4.4 Jan 22 16:40:13 Ifaistos: ah, sorry about that. I'm pleased you have it working :) Jan 22 16:41:46 no problem. :) Jan 22 16:42:16 btw once the x86 image is build how is it transfered to a CF/HD ? dd ? Jan 22 16:45:08 Ifaistos: That would depend on what kind of image it was and how your host system planned to use it Jan 22 16:45:22 target system even... Jan 22 16:45:55 i am building the epia Jan 22 16:46:36 task-bootstrap Jan 22 16:46:50 Ifaistos: I know nothing about the epia... Jan 22 16:47:18 its a small x86 board (17x17 cm) with c3 cpu Jan 22 16:47:36 kind of "mini me" pc Jan 22 16:47:57 Ifaistos: 17x17cm isn't small ;-) Jan 22 16:48:13 for pc's it is :)) Jan 22 16:49:34 Ifaistos: It depends what you mean by PC I guess. I have one the size of a 3.5" harddisk. My Zaurus is even smaller but not a PC :) Jan 22 16:50:02 lets say in can run even windows hehe Jan 22 16:50:13 Ifaistos: I'd guess you'd create a image in the form of a tar ball to extract onto the systems HD Jan 22 16:50:27 but I am just guessing... Jan 22 16:58:01 RP: Bye and many thanks for the help ! Jan 22 17:34:23 'night all Jan 22 17:55:34 zecke, any problems? Jan 22 23:32:00 03jbowler 07org.oe.dev * r6bd1bf08... 10/packages/linux/ (3 files in 2 dirs): Jan 22 23:32:00 ixp4xx-kernel: add patch for cpu activity monitor to new LEDS timer trigger in 2.6.15.1 Jan 22 23:32:00 - this adds two new triggers under a separate Kconfig setting, cpu-idle Jan 22 23:32:00 and cpu-activity. The latter lights the given LED when the cpu is Jan 22 23:32:00 active (optionally flashing or flickering it), the former makes the LED Jan 22 23:32:01 solid when the CPU is idle and flickers it when active. Jan 22 23:32:05 The patch also adds a duty_cycle setting to the triggers. Jan 22 23:32:34 RP: there's another patch in there now to implement cpu activity: packages/linux/ixp4xx-kernel/2.6.15/951-ixp4xx-leds-cpu-activity.patch Jan 22 23:33:04 I started to do it as a separate trigger, but it was duplicating all the timers code, so I made it a config option in ledtrig-timer.c Jan 23 00:02:05 I have a bootstrap-image failure. It apears to have the wrong path or enviorment variables. Jan 23 00:02:46 Anyone have a similar problem Jan 23 00:09:52 03jbowler 07org.oe.dev * rfaa6a7fa... 10/packages/linux/ (6 files in 2 dirs): ixp4xx-kernel: update to nslu2 beeper class from CVS in 2.6.15.1 Jan 23 00:10:00 To clarify the bootstrap-image .bb file dies stating that the deploy folder is not present Jan 23 00:18:34 hmm Jan 23 00:18:45 Linksys WCF54G using Agere is incorrect Jan 23 00:18:56 its a Prism 3 card Jan 23 00:19:11 unless Prism3 is known as Agere Jan 23 00:20:13 unless Jan 23 00:20:30 it uses Agere for G and Prism3 for B ... Jan 23 00:37:59 on more reading i've no clue Jan 23 00:59:17 hmm Jan 23 02:22:44 morning Jan 23 02:26:00 morning all Jan 23 02:26:40 pb__: did you send your private ssh key to RP yet? Jan 23 02:27:30 morning folks Jan 23 02:31:58 zecke: no. where do I send it? Jan 23 02:32:18 hm, wait a minute, sending my private key would be dumb Jan 23 02:32:33 pb_: http://oe.handhelds.org/cgi-bin/moin.cgi/GitScmTrial Jan 23 02:32:42 * pb_ not so easily fooled! Jan 23 02:32:42 pb_: well you can send your public one instead Jan 23 02:32:53 pb_: I'd prefer your public one :) Jan 23 02:33:11 pb_: rpurdie@rpsys.net Jan 23 02:34:22 zecke: okay, that sounds better Jan 23 02:34:57 done Jan 23 02:36:34 * zecke notes to himself, English folk is not so easily fooled - try harder Jan 23 02:36:50 zecke: next time wait until he has liquid lunch Jan 23 02:37:16 pb_: congratulation to your monotone account Jan 23 02:40:40 plop all Jan 23 02:40:52 zecke: hell! Jan 23 02:41:02 it seems I am easily fooled after all Jan 23 02:41:34 pb_: I've set it up Jan 23 02:41:56 thanks Jan 23 02:42:14 zecke: I hope you managed to plant those monotone symlinks;-) Jan 23 02:44:10 http://oe.handhelds.org/cgi-bin/moin.cgi/SvnScmTrial?action=show Jan 23 02:44:25 I patched his git binary Jan 23 02:47:39 pb_: you might want to try to pull from ewi and merge the new csets into the rpsys master Jan 23 02:47:40 morning Jan 23 02:47:56 pb_: I'm starting a git conversion of the last 30 revs soon Jan 23 02:49:58 Crofton: may anyone create a svn account on your site? Jan 23 02:54:52 http://ossie-dev.mprg.org:8080/register Jan 23 02:55:02 if someones wants a svn account for croftons server :} Jan 23 02:58:43 zecke: rock Jan 23 02:58:46 RP: the more I use plain git, the lamer I find cogito Jan 23 02:59:08 sadly I have to do some work now, but I will gather my eleet credentials and experiment with git a bit later Jan 23 02:59:13 pb_: your task: clone oegit.rpsys, branch to the oz354fam083 branch Jan 23 02:59:24 pb_: pull from ewi, push to rpsys Jan 23 02:59:43 okay Jan 23 02:59:48 "pull from ewi" meaning monotone? Jan 23 02:59:55 pb_: no ;) **** ENDING LOGGING AT Mon Jan 23 02:59:57 2006