**** BEGIN LOGGING AT Sat Aug 27 02:59:57 2005 Aug 27 03:24:54 03koen 07org.oe.dev * r32552511... 10/packages/device-mapper/ (2 files in 2 dirs): Device-mapper: Add it, courtesy Dan J Williams Aug 27 03:26:52 I'm trying to build tslib but I cannot download it from oesources.org, and the cvs server listed isn't accessable. What should I do? Aug 27 03:29:13 hound, if you want I can upload a snapshot from yesterday somewhere. Aug 27 03:29:37 you just need to copy it in your sources dir and mke the do_fetch stamp Aug 27 03:34:04 03koen 07org.oe.dev * r725067b9... 10/packages/lvm2/lvm2_2.01.09.bb: LVM2: add it, courtesy Dan J Williams Aug 27 03:34:08 03koen 07org.oe.dev * r7858230c... 10/packages/x11/x11-common/etc/X11/ (h2200.xmodmap Xinit.d/12keymap): x11-common: add h2200 keymap, Courtesy Matt Reimer, closes #266 Aug 27 03:34:12 03koen 07org.oe.dev * ref89247d... 10/packages/mdadm/ (4 files in 2 dirs): mdadm: add it, courtesy Dan J Williams, closes #265 Aug 27 03:41:24 Philippe, that would be great if you don't mind. Aug 27 03:41:32 ok. wait a sec Aug 27 03:42:45 hound, get it here http://handhelds.org/~philippe/ Aug 27 03:43:13 you probabely need to change the date in the name of the tarball to reflect today Aug 27 03:43:45 CVSDATE_tslib=20050824 in local.conf Aug 27 03:44:15 03koen 07org.oe.dev * r2287d15b... 10/ (5 files in 3 dirs): h6300: add machine description and kernel, courtesy 'lamikr', closes #255 Aug 27 03:44:17 koen, idd that works too :-) Aug 27 03:45:15 hey thanks a lot! gonna go try it now. Aug 27 03:47:16 thinking of it a fallback mode in OE would be good... nothing for today? try a previous date until you find one... Aug 27 03:58:15 what happens in monotone when you do an update? I mean when it's verifying all the new data it takes hours. That's how it's supposed to be though right? Aug 27 03:59:07 update or pull? Aug 27 04:06:11 morning Aug 27 04:08:04 update I believe. Aug 27 04:08:13 I pull, then update everytime. Aug 27 04:09:26 monotone pull is cpu bound Aug 27 04:09:34 update should be disk IO bound Aug 27 04:12:34 I got the impression they were cryptographically signed, and before applying them they were all verified. There is a notice that says, THIS MAY TAKE AWHILE. and they weren't lying :) Aug 27 04:14:05 oh it is, I just read up on it on the monotone website. I guess I should do pull's more often so that I don't end up with a large que of stuff Aug 27 04:14:12 yeah Aug 27 04:14:25 but over here update is >10 times faster as pull Aug 27 04:16:38 ah ok. Aug 27 04:23:44 hey kergoth_ Aug 27 04:33:08 morning Aug 27 04:59:57 mickeyl: ping? Aug 27 05:12:38 * koen proposes a BIG FLASHING WARNING when the fetcher notices upstream is down Aug 27 05:13:20 koen: It could even recommend setting CVSDATE to a previously known good date :) Aug 27 05:13:28 * reenoo_ propose a $USER=rday check Aug 27 05:15:03 :) Aug 27 05:15:42 I mean, he can't be that dumb, can he? Aug 27 05:16:25 I said the same thing a few days ago Aug 27 05:16:50 If a machine conf file were to change the MACHINE variable, how much carnage would result? :) Aug 27 05:20:54 I'm beginning to see a case for having a larger number of machine names but the name used for overrides only need be part of a much smaller set... Aug 27 05:25:09 we've been through this before. the problem in that has always been that there is at least one case where the original machine names is needed and then you end up adding a dedicated set of overrides etc. Aug 27 05:25:13 good morning, btw. Aug 27 05:25:30 in general, there is a need to have some granularity between MACHINE and ARCH Aug 27 05:25:38 which is != DISTRO Aug 27 05:25:39 hey mickeyl Aug 27 05:25:52 sirfred|lunch: pong Aug 27 05:26:08 mickeyl: morning Aug 27 05:26:48 reenoo_: good comment on my RFC. If you guys already had something like that in pre-OE familiar I wonder why no one of you did attempt to recreate it in OE ... Aug 27 05:26:53 mickeyl: How about MACHINE=machine unless the machine conf file wants to use some different common name? Aug 27 05:27:23 you can reset MACHINE to something else in the machine.conf Aug 27 05:27:29 we do the same in the distro conf's Aug 27 05:27:29 mickeyl: mostly lack of time Aug 27 05:27:38 but it doesn't solve the general problem of grouping machines Aug 27 05:27:53 familiar 0.7.2 has some more nice tweaks which aren't in OE Aug 27 05:28:11 (rotation keymaps for X being a nice example) Aug 27 05:28:20 mickeyl: It would allow you to group machines - you could create things like pxa-2.6 for example... Aug 27 05:28:21 koen: we should make this a priority after releasing the next set of distros Aug 27 05:28:40 RP: yes, but where is the sense of adding a group without the group being an OVERRIDE ? Aug 27 05:28:52 mickeyl: yah, personally I've never got around to it since the h3800 has enough flash to fit a full task-bootstrap in there Aug 27 05:29:28 mickeyl: Personally, I'd like to see a third override group but everyone else seems against it. Overloading MACHINE seems the only alternative Aug 27 05:29:54 mickeyl: some h3600 (the real one) users were going to work on that but as you see haven't done so Aug 27 05:30:20 I'm afraid we have to solve that before the next set of releases Aug 27 05:30:24 at least for the Opie image Aug 27 05:30:42 opie getting to bloated? Aug 27 05:30:44 * koen runs Aug 27 05:30:46 hah Aug 27 05:30:51 i can do without bluetooth Aug 27 05:31:06 (personally) Aug 27 05:31:29 the problem started with the truetype fonts Aug 27 05:31:37 for small sizes, qpf's are better Aug 27 05:31:41 (resource wise) Aug 27 05:33:04 is there a way in opie to make window fullscreen by default Aug 27 05:33:21 resizing window *very* time sucks donkey ass Aug 27 05:33:24 ~lart ljp Aug 27 05:33:25 * ibot pulls out a ClueBat (tm) and thwaps ljp Aug 27 05:33:28 don't worry Aug 27 05:33:30 just rebuild Aug 27 05:33:31 he broke it Aug 27 05:33:34 and i fixed it Aug 27 05:34:01 rebuild libqpe Aug 27 05:34:23 Ok, how about a new MACHCLASS override? This could be of benefit in that for distrubution, we could just have a smaller number of root filesystems, escpecially for devices which don't build the kernel into the rootfs Aug 27 05:34:48 ~bitchslap lilo Aug 27 05:34:51 * ibot beats the sh*t out of lilo Aug 27 05:35:54 RP: how would we distinguish between rootfs'es? Aug 27 05:36:20 koen: We have a set of machines where we don't need to... Aug 27 05:36:57 The same rootfs will run on c7x0, spitz, borzoi, akita and hx2750 under 2.6... Aug 27 05:37:29 which would be wrong, since the hx2750 is a small screen device Aug 27 05:38:05 koen: true Aug 27 05:38:06 a MACHCLASS is most appreciated since it fixes at least stuff like: Aug 27 05:38:36 The idea would be to try and get rid of all the identical machine overrides... Aug 27 05:38:40 SRC_URI_append_spitz = "file://kernel-keymap.patch;patch=1 file://kernel-keymap-corgi.patch;patch=1 file://kernel-keymap-CXK. Aug 27 05:38:40 patch;patch=1 " Aug 27 05:38:40 SRC_URI_append_akita = "file://kernel-keymap.patch;patch=1 file://kernel-keymap-corgi.patch;patch=1 file://kernel-keymap-CXK. Aug 27 05:38:41 patch;patch=1 " Aug 27 05:38:41 SRC_URI_append_borzoi = "file://kernel-keymap.patch;patch=1 file://kernel-keymap-corgi.patch;patch=1 file://kernel-keymap-CXK. Aug 27 05:38:44 patch;patch=1 " Aug 27 05:38:46 that's braindead Aug 27 05:39:07 mickeyl: I have this in mind - I'm just not expressing it very well... Aug 27 05:39:21 righto Aug 27 05:39:27 so SRC_URI_append_spitz = "file://kernel-keymap.patch;patch=1 file://kernel-keymap-corgi.patch;patch=1 file://kernel-keymap-CXK. Aug 27 05:39:27 patch;patch=1 " Aug 27 05:39:27 SRC_URI_append_akita = "file://kernel-keymap.patch;patch=1 file://kernel-keymap-corgi.patch;patch=1 file://kernel-keymap-CXK. Aug 27 05:39:27 patch;patch=1 " Aug 27 05:39:27 SRC_URI_append_borzoi = "file://kernel-keymap.patch;patch=1 file://kernel-keymap-corgi.patch;patch=1 file://kernel-keymap-CXK. Aug 27 05:39:30 patch;patch=1 " Aug 27 05:39:32 oops Aug 27 05:39:34 sorry Aug 27 05:39:43 so, if we add a MACHINECLASS, we have one more override Aug 27 05:39:51 with a priority right between MACHINE and ARCH Aug 27 05:40:03 that would reduce a lot of repetition Aug 27 05:40:30 i know we should be careful with adding more OVERRIDES but I definitly see a need for this one Aug 27 05:40:55 This is what I've thought for a long time... Aug 27 05:41:02 mickeyl: there's no need to introduce a new (global) variable for that. MACHINE.conf can include mickeyl-favourite-machines.conf, and that .conf file can bash OVERRIDES any way it likes. Aug 27 05:42:26 so you SRC_URI_append_spitz = "file://kernel-keymap.patch;patch=1 file://kernel-keymap-corgi.patch;patch=1 file://kernel-keymap-CXK. Aug 27 05:42:26 patch;patch=1 " Aug 27 05:42:26 SRC_URI_append_akita = "file://kernel-keymap.patch;patch=1 file://kernel-keymap-corgi.patch;patch=1 file://kernel-keymap-CXK. Aug 27 05:42:27 patch;patch=1 " Aug 27 05:42:27 SRC_URI_append_borzoi = "file://kernel-keymap.patch;patch=1 file://kernel-keymap-corgi.patch;patch=1 file://kernel-keymap-CXK. Aug 27 05:42:30 patch;patch=1 " Aug 27 05:42:32 argh Aug 27 05:42:34 bbl Aug 27 05:42:44 heh Aug 27 05:42:50 pb_: The advantage of doing it explicitly is its easier to see what's going on... Aug 27 05:43:32 I think the question pb_ wants to indicate is: is this a essentially zaurus thing or not Aug 27 05:43:52 s/a essentialy/essentially a/ Aug 27 05:44:13 would the ipaq devices benefit from such a class? Aug 27 05:44:31 no, not really. Aug 27 05:44:47 ok, so it's just because sharp is releasing so many devices which are pretty similar Aug 27 05:44:52 right. Aug 27 05:45:03 in that case we may go with a zaurus only override Aug 27 05:46:13 A pxa rootfs is esentially the same regardless of the device though. Some differences like the screensize are problematic but most other things can be dealt with at runtime... Aug 27 05:47:26 I don't really understand what you're saying. If two devices can use the same rootfs, they should -- pretty much by definition -- be the same MACHINE (e.g. ipaq-pxa-2.6). Aug 27 05:47:49 conversely, if they need to be different MACHINEs, that must mean that at least some packages are built differently for the two cases, and hence the rootfs must be different. Aug 27 05:48:19 wait, you're simplifying a bit because you guys have good bootloaders Aug 27 05:48:29 If you have a whole load of MACHINEs that behave identically, the easy answer is just to get rid of them. Aug 27 05:49:20 pb_: The hx4700 illustrates the problem - it needed its own kernel as it was pxa27x based but is an otherwise identical machine type to the ipaq-pxa-2.6... Aug 27 05:49:42 I can't use the hx4700 machine easily for the hx2750 (also pxa27x) becuase I need a different kernel Aug 27 05:50:14 that's because you haven't updated the hx2750 bits on hh.org cvs Aug 27 05:50:50 koen: We have a perfectly working hx2750 kernel in OE... Aug 27 05:51:22 RP: well, the hx4700 only needs a separate kernel due to a bug in the kernel. But, leaving that aside, if it does need a different kernel and the kernel is in the rootfs, then the rootfs must be different. Aug 27 05:51:25 which is non obvious since it's called 'linux-openzaurus' Aug 27 05:51:27 mickeyl: The hx2750 has the problem specifically becuase it doesn't have a good bootloader :-/ Aug 27 05:52:50 koen: I'm happy with a rename. I just can't think of one that covers everything... Aug 27 05:53:20 pb_: That doesn't mean having duplicate machine overrides everywhere is a good thing... Aug 27 05:54:19 RP: how many files contain specific overrides for the hx4700? Aug 27 05:55:26 * pb_ going home now Aug 27 06:02:38 What happened to the ipaq-pxa machine? Are theye no pxa255 based ipaqs? Aug 27 06:02:55 h3900 Aug 27 06:04:14 h3900 support h39xx h5xxx h22xx aximx5 and asus mypal Aug 27 06:04:46 with both 2.4 and 2.6 kernels Aug 27 06:05:20 With the 2.4/2.6 kernel selection done totally differently to the Zaurus. Great. Aug 27 06:10:33 I think the best sign of a need for machine class groupings is the shear number of "common" .conf files in the machine directory tbh... Aug 27 06:11:20 But I shall go back to biting my tounge :) Aug 27 06:12:00 RP: that's because of the way the clamshells handle 2.4/2.6 Aug 27 06:27:22 I'm not quite sure what "shear number" you're thinking of. In my conf/machine directory there are a total of three files with "common" in the name, which doesn't strike me as an overwhelming number. Aug 27 06:28:39 and none of the three have associated OVERRIDES, so it isn't very obvious what replacing them with some kind of machine class grouping would buy you. Aug 27 06:29:08 pb__: When you include all the zaurus-2.4 and zaurus 2.6 ones, the number increases a bit which is why I said "common" and not *common* Aug 27 06:29:36 I agree the Zaurus handles this badly. I also think these common file shouldn't be in the machine directory at all Aug 27 06:30:28 Yeah, that's probably true. If you wanted to move them to a conf/machine/include subdirectory or somethin, that would be fine. Aug 27 06:31:07 I'd suggest a conf/machclass directory ;-) Aug 27 06:36:17 mickeyl: Hi. I think I have a working suspend/resume driver (even for the suspend-from-opie case). Aug 27 06:38:23 sirfred: excellent :) Aug 27 06:38:49 RP: Thanks. :). Finally, I think I won't need that cleanup function we talked about. Aug 27 06:39:19 sirfred: That's good to know. I suspect we'll eventualy have one anyway but it would be better not to need it Aug 27 06:39:19 pretty cool. how did you do it? Aug 27 06:40:17 mickeyl: Well, after seeing that the qcop messages reaching the driver were the same in the two cases, I started to suspect that the driver suspend/resume procedure itself was the problem. Aug 27 06:40:54 mickeyl: I was just experimenting, and found that calling some other detach/attach aticore functions, it started to work. Aug 27 06:41:01 cool Aug 27 06:41:35 mickeyl: Anyway, the processEvents before the QCopEnvelope in odevice_zaurus.cpp are recommended. Aug 27 06:41:59 mickeyl: Without that, the messages are not reaching "on time" to the driver, anyway, now is working with the new suspend/resume implementation. ;-) Aug 27 06:44:46 mickeyl: Do you want a patch from me for the odevice_zaurus.cpp part, also? It's just a pair of qApp->processEvents() after the QCopEnvelope stuff Aug 27 06:45:05 sirfred: yes, please Aug 27 06:45:39 if that works out there's only one little buglet left, the tiny pixmaps Aug 27 06:45:39 mickeyl: OK. I just need to sync patcher with the mess I've created testing all this. Aug 27 06:45:43 hehe Aug 27 06:45:47 mickeyl: Yes, I remember. Aug 27 06:45:50 mickeyl: .) Aug 27 06:46:16 mickeyl: I was looking into that a while ago. It's incredible but the bug goes away just disabling accelerated drawline. Aug 27 06:46:50 mickeyl: The problem is that I suspect that those pixmaps are not drawing using drawline, It seems an ugly collateral effect. Aug 27 06:48:17 hmm Aug 27 06:48:20 let me take a look Aug 27 06:48:28 This is starting to be more demanding than my actual work. ;-) Aug 27 06:48:38 i know exactly what you mean :D Aug 27 06:49:32 see here. Aug 27 06:49:49 p.eraseRect( 0, 0, w, h ); Aug 27 06:49:49 QPointArray a; Aug 27 06:49:49 a.setPoints( 3, 0, 1, w-1, 1, h-2, h-1 ); Aug 27 06:49:49 p.setBrush( color1 ); Aug 27 06:49:49 p.drawPolygon( a ); Aug 27 06:49:53 p.eraseRect( 0, 0, w, h ); Aug 27 06:49:53 QPointArray a; Aug 27 06:49:53 a.setPoints( 3, 0, 1, w-1, 1, h-2, h-1 ); Aug 27 06:49:55 p.setBrush( color1 ); Aug 27 06:49:57 p.drawPolygon( a ); Aug 27 06:50:01 ~lart my cutn'paste Aug 27 06:50:02 * ibot hauls my cutn'paste up by the scruff of the neck and spanks him until he waddles Aug 27 06:50:16 they are drawn using lines Aug 27 06:51:38 It's using drawPolygon. Aug 27 06:51:39 Interesting. Aug 27 06:51:51 That's a really valuable information. Aug 27 06:52:08 see QSpinBox::updateButtonSymbols() Aug 27 06:52:16 in src/kernel/widgets/qspinbox.cpp : 983 ff Aug 27 06:52:21 mickeyl: OK. I'm going to setup the patch for suspend/resume and then try to fix this. Aug 27 06:52:26 excellent Aug 27 06:52:29 thanks Aug 27 06:52:36 mickeyl: np Aug 27 06:53:04 * mickeyl removes one nice-to-have in the list of things blocking 3.5.4 Aug 27 07:17:09 anyone using hentges image? Aug 27 07:17:23 pocketppc: see topic, 3rd item Aug 27 07:19:23 koen: ok... i need some help , which channel should i go? Aug 27 07:19:37 #openzaurus or something Aug 27 07:19:59 ok.. thanx... i'm new here... Aug 27 07:20:24 #hentges.net may be even more appropriate Aug 27 07:20:25 heh Aug 27 07:21:04 mickeyl: thanx Aug 27 07:26:34 any idea.. why the battery meter keep showing 100%? Aug 27 07:27:21 sorry.. for non developer question.. i post some questions on openzaurus and hentges.net but no reply.. Aug 27 07:38:43 I have all the packages compiled for an opie image, but when it tries to do_rootfs it fails when trying to execute makedevs because it can't find it. I see it in ./tmp/work/makedevs-1.0.0-r2/install/makedevs/sbin/makedevs, where should it be? Aug 27 08:09:51 03ph5 07org.oe.dev * r6719efa4... 10/packages/gtk+/ (gtk+-2.8.0/pkg-config-pangoft2.patch gtk+_2.8.0.bb): Aug 27 08:09:51 gtk+-2.8.0: Build fix (Bug #267) Aug 27 08:09:51 - This helps gtk+ to find the freetype includes during build. Aug 27 08:28:07 03koen 07org.oe.dev * r7441878b... 10/packages/meta/openslug-packages.bb: openslug-packages: re-add ifupdown which seems to have gotten lost in on of the merges Aug 27 08:46:26 I must be missing a package. the makdevs binary compiled is for ARM. no debian packages however include makedevs. Aug 27 08:46:52 isn't makedev just a script? Aug 27 08:46:59 probably Aug 27 08:47:00 err Aug 27 08:47:03 lemme look Aug 27 08:47:17 nah it's a binary Aug 27 08:47:28 on debian the /dev/MAKEDEV is a script. that's not what do_rootfs is calling though Aug 27 09:12:19 koen: libgcc*_arm.ipk shouldn't have any dependencies on non-arm packages Aug 27 09:12:55 it doesn't have it Aug 27 09:13:24 it says "hey, I'm finding gcc-cross-locale-fr" "Damn, it's for i686" Aug 27 09:13:58 I've had the same problem a few months ago Aug 27 09:14:13 Depends: gcc-cross-locale*, libc6 (>= 2.3.5+cvs20050627) Aug 27 09:14:33 don't say it doesn't Aug 27 09:15:03 gcc-cross-locale* itself isn't non-ARM Aug 27 09:15:21 you can't a priori say what arch it is Aug 27 09:16:47 I'm sorry. but that's not true Aug 27 09:17:07 all gcc-cross-locale* packages are _i686 here Aug 27 09:17:30 they are armeb here Aug 27 09:17:37 it depends on your build host Aug 27 09:17:48 which is something ipkg on the device doesn't know about Aug 27 09:18:12 regardless. it's a bug that needs to be fixed Aug 27 09:18:45 yeah Aug 27 09:18:55 our i18n deps are the wrong way around Aug 27 09:19:14 that's a different story Aug 27 09:19:26 the same Aug 27 09:19:47 if every package doesn't depends on it's locales, this problem doesn't exist Aug 27 09:20:02 no. here we're talking about an _arm package having dependencies on _$HOST_ARCH packages Aug 27 09:20:53 and which host-arg package? Aug 27 09:20:58 exactly, the locales Aug 27 09:21:16 incidently, yeah Aug 27 09:21:33 but not all locale package have the _$HOST_ARCH problem Aug 27 09:21:38 packages even Aug 27 09:22:48 gcc-cross is simply a special case that needs special care Aug 27 09:27:03 ah had to build makedev-native for it to work. Aug 27 09:37:59 reenoo_: would a special case in package_ipk.bbclass solve it? Aug 27 09:38:17 sed -i s:-cross:: control ? Aug 27 09:42:33 koen: like I said I don't have the time to look into fixing it right now Aug 27 09:53:53 [cc]smart: @openembedded.org mail addresses don't work Aug 27 09:54:15 [cc]smart: so the bugzilla bugmail will get lost Aug 27 09:54:31 anyone happen to know of a good struct pointer example for C ? Aug 27 09:56:33 what u mean emte ? Aug 27 09:57:08 double koen !!! Aug 27 09:57:26 03ccsmart 07org.oe.dev * r6a83bbd0... 10/packages/sysvinit/sysvinit/rcS: Based on OE bug #244, thanks to Pfudd. Aug 27 09:58:01 big endian koen :) Aug 27 09:58:23 have u buy also a nslu2 ? Aug 27 09:58:37 yes Aug 27 09:58:41 who is Peter Urbanec ? Aug 27 09:59:02 peteru(-home) from #openslug Aug 27 09:59:12 mhh what can be a reasonable price for a h3700 .. used obviously ... Aug 27 09:59:21 rwhitby told him to send his keys to us Aug 27 09:59:29 ok Aug 27 09:59:49 [cc]smart: http://article.gmane.org/gmane.comp.handhelds.openembedded/5878 Aug 27 10:08:38 gremlin[it], i dont quite remember how to do it and what the scope is Aug 27 10:12:59 anyone took a look into initng yet? Aug 27 10:20:17 mickeyl: Do you know if I need to make any change in my program makefiles to use ttf? I have some test programs for pre-ttf qpe and they don't show any font now. Aug 27 10:21:29 ttf is handled by opie-freetype plugin. if that's installed and the fonts are registered in $QTDIR/fonts/fontdir then they should show up in any program Aug 27 10:21:42 ... at least all programs that use the standard font handling Aug 27 10:21:51 perhaps they do something hardcoded... Aug 27 10:21:54 always possible Aug 27 10:22:38 mickeyl: The only thing I make is: QPushButton *button2 = new QPushButton( "Lines", this ); Aug 27 10:23:14 is the default font installed? Aug 27 10:23:43 How can I know that? Aug 27 10:23:51 heh Aug 27 10:23:54 ls ? Aug 27 10:24:02 go to dinner Aug 27 10:24:09 check qpe.conf for the default font Aug 27 10:24:15 then look in fontdir if it's registered Aug 27 10:24:22 the look if it actually is in the file system where fontdir tells Aug 27 10:24:26 ok Aug 27 10:28:56 03mickeyl 07org.oe.dev * r23b5b6fd... 10/packages/qte/ (qte-2.3.10/c7x0-w100-accel.patch qte_2.3.10.bb): qte-2.3.x: update c7x0-w100 accelleration patch courtesy Manuel Teira Aug 27 10:29:01 03mickeyl 07org.oe.dev * rafe5b334... 10/packages/apmd/apmd_3.2.2.bb: apmd: install the apm wrapper script only on Zaurus models using 2.4-embedix Aug 27 11:16:00 mickey_away: your monotone broke at vanille.de Aug 27 11:17:14 mickey_away: and restart.sh doesn't work (unmatched quote) Aug 27 11:27:08 Is there an akita owner around who fancies trying a 2.6 kernel? (or a borzoi owner but I guess they're a bit more rare) Aug 27 11:35:10 http://www.rpsys.net/openzaurus/temp/zImage-akita is anyone appears. I've also sent details to the oz-dev list. I'll be back later Aug 27 13:06:49 * RP returns Aug 27 13:08:50 mickeyl: monotone.vanille.de seems to be down Aug 27 13:10:34 again? Aug 27 13:10:48 *sigh* Aug 27 13:11:02 restarted Aug 27 13:11:21 03koen 07org.oe.dev * r989366a7... 10/conf/distro/openslug-packages.conf: openslug-packages.conf: add ifupdown so it will be parsed too Aug 27 13:11:22 it's working again Aug 27 13:21:26 03koen 07org.oe.dev * rbc98a6d4... 10/packages/gpe-bluetooth/ (gpe-bluetooth_0.42.bb gpe-bluetooth_0.43.bb): gpe-bluetooth: add 0.43 and blow away the broken 0.42 Aug 27 13:21:28 why is blueprobe in x11/network? Aug 27 13:21:30 03koen 07org.oe.dev * rb9d8eb40... 10/conf/distro/preferred-gpe-versions-2.7.inc: preferred-gpe-versions-2.7.inc: gpe-bluetooth 0.42 -> 0.43 Aug 27 13:21:34 it only depends on libc6 Aug 27 13:23:39 koen: because most of the SECTION entries are wrong after CoreDump|busy's SECTION adding vendetta Aug 27 13:23:59 hmmm Aug 27 13:24:11 what would be the correct one? Aug 27 13:24:58 dunno. probably base Aug 27 13:26:23 bluez is in console Aug 27 13:26:32 * koen is a bit lost in section-land Aug 27 13:27:59 what is the difference between base and console? Aug 27 13:29:17 I guess, console is probably an openzaurus'ism to split packages by non-gui/opie/x11 Aug 27 13:29:53 but there's never been any type of section definition even though I asked for such a thing Aug 27 13:30:05 anybody against SECTION="base" for blueprobe? Aug 27 13:30:59 no, sounds reasonable Aug 27 13:31:48 and it's changed Aug 27 13:32:33 i think the difference was that console is nothing system relevant Aug 27 13:32:46 for instance, sysvinit is base Aug 27 13:32:49 03koen 07org.oe.dev * r7d61c557... 10/packages/blueprobe/blueprobe_0.13.bb: blueprobe: change section to 'base' Aug 27 13:32:58 whereas bluez-utils is console Aug 27 13:33:04 next question: Aug 27 13:33:12 why do we have BROKEN=1 packages? Aug 27 13:33:30 shouldn't they be either in nonworking/ or have EXCLUDE_FROM_WORLD? Aug 27 13:33:33 to make them not being considered in a WORLD build Aug 27 13:34:06 i don't know why we have three categories Aug 27 13:34:09 but two are good Aug 27 13:34:18 BROKEN means - broken for _some_ configurations Aug 27 13:34:27 nonworking means broken in all configurations Aug 27 13:34:45 EXCLUDE_FROM_WORLD... hmm... ah right, we need to exclude some things from world which are not broken Aug 27 13:34:59 but they taint packages which are later build Aug 27 13:35:05 for instance due to shlib oddities Aug 27 13:35:07 bbl, tv time Aug 27 13:35:10 Googling for "Borzoi Zaurus" has hh.org, oz.org and my kernel page as the top hits :) Aug 27 14:07:31 yeah, BROKEN predates EXCLUDE_FROM_WORLD, but (although the effect is identical) they signify slightly different things. Aug 27 14:08:26 I'm always slightly amused when I succesfully build a 'BROKEN' package Aug 27 14:08:38 without making changes Aug 27 14:08:56 like mickeyl said, BROKEN doesn't mean "unbuildable in every situation". Aug 27 14:09:05 yeah Aug 27 14:09:31 but if that's the case it should be SLIGHTLYBROKEN or MAYNOTWORK Aug 27 14:47:56 afternoon Aug 27 14:49:31 hi treke Aug 27 14:51:02 how are things doing on the hh.o front? Aug 27 14:53:24 I am getting a connection refused when trying to pull Aug 27 14:58:19 gerwinin: monotone.vanille.de is down atm. try ewi546.ewi.utwente.nl? Aug 27 15:10:21 'night all Aug 27 21:45:02 moo Aug 28 00:09:38 good morning Aug 28 00:49:22 Hi, it appears somebody forgot to merge and now there are multiple heads. Can somebody please rectify? Thank you. Aug 28 00:58:10 <[cc]smart> pulling... Aug 28 00:58:22 so ... waiting ... Aug 28 00:58:50 <[cc]smart> :) Aug 28 01:01:23 <[cc]smart> appears to be clean to me, this is the one i get: http://monotone.vanille.de/viewmtn/revision.psp?id=7d61c557a09552a762287f1693518286014137dc Aug 28 01:39:19 mickey|tv: ping Aug 28 01:46:46 Morning all Aug 28 01:47:23 hi uv1 Aug 28 01:48:25 RP: Was 3 weeks off. Did I miss something. Monotone pull;update takes 30 secs now. Was something changed ? Aug 28 01:48:53 uv1: No, its still as slow as ever... Aug 28 01:49:17 RP: Anyhow I builded the opie image and a 2.6.13 kernel. No I have probs with network and the keyboard. Fn doesnt work. Any ideas ? Aug 28 01:50:30 uv1: Sounds like the keymap didn't load? Aug 28 01:50:38 RP: Trying a pull : Connection refused now. Can you try that as well ? (monotone.vanille.de) Aug 28 01:50:39 This is c7x0? Aug 28 01:50:46 RP: C7x0 Aug 28 01:51:04 Yes, connection refused for me as well. Try the backup? Aug 28 01:51:19 RP: Which backup ? Aug 28 01:51:21 ewi546.ewi.utwente.nl Aug 28 01:51:53 monotone pull ewi546.ewi.utwente.nl "org.openembedded.{dev}" Aug 28 01:53:38 uv is monotoning utwente. Aug 28 01:53:49 * RP has just trashed everything on his husky :-/ Aug 28 01:54:21 RP: How can I go on with keymap ? Aug 28 01:55:02 uv1: There should be a file in /etc called keymap-2.6.map. Running loadkeys /etc/keymap-2.6.map should load it... Aug 28 01:55:19 If it doesn't exist, there is something wrong with your image Aug 28 01:56:20 My Fn key works as you have to hit it then hit the button to go with it, don't push them at the same time. Aug 28 01:56:25 is that your problem maybe? Aug 28 01:56:39 RP: keymap parse error line 146! I thought thats gone ? Aug 28 01:58:07 uv1: People have broken it now and again. It should be ok in the latest OE. I don't know about older versions Aug 28 01:58:47 RP: OK, will do a rebuil after monotone update. Thanks so far. Aug 28 02:00:15 hey Aug 28 02:00:25 * zecke is *busy* as hell Aug 28 02:00:33 zecke look Aug 28 02:00:40 root@h3600:/dev# uname -a Aug 28 02:00:40 Linux h3600 2.6.12-hh2 #3 Sun Aug 28 10:17:32 CEST 2005 armv4l unknown Aug 28 02:00:50 hey ;) Aug 28 02:01:05 hey zecke Aug 28 02:01:08 I have IFA work to do Aug 28 02:01:12 zecke: are you in nurnberg? Aug 28 02:01:12 gremlin[it]: congrats Aug 28 02:01:31 koen|slug: no, I was in Nürnberg Sunday to Wednesday Aug 28 02:01:42 koen|slug: now I'm busy integrating software for IFA Aug 28 02:01:51 IFA? Aug 28 02:01:55 * zecke fights fire... Aug 28 02:02:03 koen|slug: International Funk Austellung Aug 28 02:02:07 koen|slug: www.ifa-berlin.de Aug 28 02:02:11 ah cool Aug 28 02:02:20 * koen|slug is annoyed with his fever Aug 28 02:02:29 * koen|slug wanted to attend the OE meeting today Aug 28 02:02:46 :} Aug 28 02:20:39 I still get "monotone: multiple update candidates". One from koen and one from Florian. Aug 28 02:21:09 Laibsch: monotone pull Aug 28 02:23:29 * gremlin[it] would find some one pay him to work on linux ... **** ENDING LOGGING AT Sun Aug 28 02:59:56 2005