**** BEGIN LOGGING AT Wed Aug 31 02:59:56 2005 Aug 31 03:00:02 lardman|work: I found that corgi/shepherd id problem and oe has been updated - a patch got lost... Aug 31 03:01:14 RP: yes I saw that, thanks, I've just updated and am rebuilding now Aug 31 03:02:42 koen|slug: what's with the gcc .incs - what are they? Aug 31 03:03:00 share metadate for gcc_someversion.bb Aug 31 03:03:07 shared* Aug 31 03:05:54 I mean, why do I want to look at them, to ensure that octave will build for others? Aug 31 03:06:37 lardman|work: your post on oesf Aug 31 03:07:07 ah, it all becomes clear - not so good on the first working day of the week (which it is for me) ;) Aug 31 03:12:16 koen|slug: Does libxsettings-client_0.14.bb still need the libgpewidget dependency? Aug 31 03:13:12 RP: no idea Aug 31 03:14:06 I think it was just for gpe_do_install which seems to be no longer needed... Aug 31 03:15:27 heh.. korganizer lack beam option ;( Aug 31 03:22:04 hi mickey|tv Aug 31 03:24:15 morning Aug 31 03:25:48 I've just added Zaurus hinge sensor support for the latest 2.6 kernel to Opie. It's untested yet though Aug 31 03:29:13 hi mickeyl Aug 31 03:29:46 mickeyl: Great - I'll try and get a spitz opie-image building so I can test it Aug 31 03:29:55 ok Aug 31 03:30:24 ~lart BT Aug 31 03:30:24 * ibot takes out a seltzer bottle and sprays BT in the face. You know, one of those old-school seltzer bottles clowns have? Yeah those. Anyway, consider yourself spritzed Aug 31 03:30:46 mickeyl: If it works, I'll nuke the c7x0 hack for the old method Aug 31 03:31:00 *nod* Aug 31 03:31:35 RP: Is GPE better than OPIE about the touchscreen reflection thing? Aug 31 03:31:47 RP: don't forget to tell JustinP Aug 31 03:32:03 since opie isn't the onlything using the hack Aug 31 03:33:23 righto Aug 31 03:33:25 BT suxx when it comes to pairing two computers Aug 31 03:33:31 Noodles: Both suffer from it. ts_calibrate doesn't Aug 31 03:33:43 koen|slug: Tell JustinP what? Aug 31 03:33:54 pairing c7x0 with phone is easy, but getting two bluez working is pita Aug 31 03:34:01 RP: the new input method to call xrandr Aug 31 03:34:28 mickeyl: Ah, right. Will do. Aug 31 03:34:57 Has someone fixed minilite to work with c7x0 2.6 yet btw? Aug 31 03:35:21 minilite works with 2.6 Aug 31 03:35:45 OZ is braindead enough to ship 2.4 /dev/ entries Aug 31 03:36:10 and minilite checks for /dev/sharp_fl Aug 31 03:36:19 koen|slug: OZ images have been compatible for both. minilite is braindead for using an used device ;-) Aug 31 03:36:33 an unused device... Aug 31 03:36:48 ~curse BT Aug 31 03:36:50 May you be reincarnated as a Windows XP administrator, BT ! Aug 31 03:42:21 can someone kick some BT creators from me? Aug 31 03:42:57 hrw|work: BT is a braindead protocol from hardware up :-) Aug 31 03:44:05 I've never really understood (even after reading Bluetooth Revealed) why they had to craft such a huge stacked architecture Aug 31 03:45:03 its brainless Aug 31 03:47:13 hmm Aug 31 03:47:25 *sigh* one of these days I have to rewrite my inotify wrapper Aug 31 03:47:43 the interface of the day(tm) seems to be sysctls now Aug 31 03:47:47 syscalls even Aug 31 03:48:53 mickeyl: because the original instigators were obsessed with iso 7 layer Aug 31 03:49:16 mickeyl: BT is designed to have all 7 layers even if the layer does nothing Aug 31 03:49:45 mickeyl: and someone let them save one byte a packet on a highspeed line through making the software 4x more complex Aug 31 03:50:00 XorA: too bad that bluez or BT forgot to add one command - "pair" Aug 31 03:50:34 hrw|work: try a scatter net :-) Aug 31 03:50:52 c7x0 can browse sdp on phone and desktop, desktop can browse sdp on phone but refuse to browse on c7x0 Aug 31 03:51:18 morons Aug 31 03:51:46 hrw|work: luckilly I never worked on SDP, it ws hair pulling stuff Aug 31 03:53:16 and all that fscking work just to get beaming Aug 31 03:53:33 hex writing from printed page will be faster Aug 31 03:54:27 beaming just works(TM) with gpe Aug 31 03:54:42 to phone and back Aug 31 03:54:48 koen|slug: when devices are paired Aug 31 03:55:12 or when you setup bluez right Aug 31 03:55:32 root@c7x0:/etc/bluetooth# hciconfig Aug 31 03:55:32 hci0: Type: PCCARD Aug 31 03:55:32 BD Address: 00:E0:03:25:A1:DD ACL MTU: 340:4 SCO MTU: 64:0 Aug 31 03:55:32 UP RUNNING PSCAN ISCAN Aug 31 03:55:38 12:51 root@work:bluetooth# hciconfig Aug 31 03:55:38 hci0: Type: USB Aug 31 03:55:38 BD Address: 00:80:98:24:52:AE ACL MTU: 192:8 SCO MTU: 64:8 Aug 31 03:55:38 UP RUNNING PSCAN ISCAN Aug 31 03:56:11 config is exactly same, they l2ping each other, c7x0 can browse sdp of desktop Aug 31 03:56:36 desktop cannot sdp browse c7x0 Aug 31 03:56:46 what can be wrong here? Aug 31 03:57:06 you can't l2ping if they aren't paired afaik Aug 31 03:57:22 and if the pairing went wrong you can't browse in either direction Aug 31 03:57:35 so I think sdpd on the c7x0 is doing something wrong Aug 31 03:58:40 koen|slug: so the question is... what is borken in bluez-utils-nodbus from OE.. Aug 31 03:59:44 no idea, I use bluez-utils-dbus Aug 31 04:06:00 they differ only on depends iirc Aug 31 04:08:38 koen|slug: on both sides you have bluez 2.18? Aug 31 04:09:32 my phone and powerbook have their own BT stack Aug 31 04:10:30 and I dont have a problem with !bluez stack on other side Aug 31 04:22:24 root@c7x0:/var/tmp# ipkg install libbluetooth1_2.20-r0_armv5te.ipk bluez-utils-nodbus_2.20-r5_armv5te.ipk Aug 31 04:22:34 will test does it helps Aug 31 04:24:42 * koen|slug now sends rpjday* to /dev/null in .procmailrc Aug 31 04:30:53 koen|slug: are you sure about that post to oesf? He's trying to do a native compile on an OZ 3.5.3 machine? Surely he just needs to get soft-float output from his native toolchain? Aug 31 04:31:30 you can't mix softfloat and hardfloat apps Aug 31 04:31:49 but his toolchain comes from the OZ feed, so does his 'rom', so there should be no issue Aug 31 04:32:13 except for softfloat itself doing strange stuff Aug 31 04:32:38 yes, but it must be possible to make the native toolchain produce soft-float binaries Aug 31 04:32:51 if it doesn't so that by default...? Aug 31 04:33:01 no idea Aug 31 04:33:12 but you can make sure by using the correct options Aug 31 04:33:26 which are defined in gcc.inc under the TARGET_FPU section Aug 31 04:33:32 Okay Aug 31 04:38:12 re Aug 31 04:42:25 who maintain BlueZ in OE? (no MAINTAINER field) Aug 31 04:42:48 you just answered your question Aug 31 04:42:58 check monotone log, or bk history Aug 31 04:43:12 I know - but I can push 2.20 but it will be nearly nontested Aug 31 04:43:25 attach it to bugzilla Aug 31 04:47:57 ok - bug #277 Aug 31 05:11:30 hrw|work: please just push them with -1 default preference Aug 31 05:11:47 i'll test and remove that then Aug 31 05:11:53 (if everything still works) Aug 31 05:12:41 mickeyl: added them into bugtracker - get them, test and push if work Aug 31 05:13:10 k Aug 31 05:13:34 we may eventually remove the nondbus and dbus seperation anyway Aug 31 05:16:47 what for libdbus under opie? Aug 31 05:17:41 opie doesn't listen to dbus, but it doesn't hurt to install it either. I only split the packages because of space issues - since dbus pulls in glib. Aug 31 05:20:03 and we are already out of space on 16M Aug 31 05:20:14 mickeyl: nslu2 uses bluez too Aug 31 05:20:34 koen|slug: yeah, that's _pretty_ cool. Aug 31 05:20:56 so having a dbus/nondbus split is good Aug 31 05:21:04 and slug has 8M only iirc Aug 31 05:21:10 ah. right Aug 31 05:21:53 but why sdpd does not work I dont know Aug 31 05:23:02 can you remove bluez-dbus from your gpe installation and try with bluez-nondbus? I suspect that to be a configuration issue rather than a dbus/nondbus thing Aug 31 05:23:05 i may be wrong though Aug 31 05:24:02 try restarting /etc/init.d/bluetooth Aug 31 05:24:15 and/or add services by hand via sdptool Aug 31 05:26:46 I'm too tired of BT - will do next attempt to get it working in next month or later Aug 31 05:26:53 it suxx Aug 31 05:29:33 hrw|work: slug doesn't install much packages to flash though... Aug 31 05:32:46 NAiL: but it can Aug 31 05:32:51 I go lunch Aug 31 05:36:33 ~bon appetit Aug 31 05:36:36 [bon appetit] smacznego. Guten Appetit. Eet Smakelijk. God Appetitt Aug 31 05:51:17 query: the "getting started" instructions still refer extensively to bitbake.....is there anything newer ? Aug 31 05:52:07 the GettingStarted wiki is the most accurate document Aug 31 05:53:01 I see.....but it refers to bitbake.....I take it there are substantial differences in procedure ? Aug 31 05:53:20 ? Aug 31 05:53:22 it refers to both Aug 31 05:53:30 there is nothing newer than bitbake Aug 31 05:53:40 you don't mean bitkeeper by any chance? Aug 31 05:53:54 it's just that bitbake seems to confuse most people so GettingStarted has a big section on it Aug 31 05:53:54 oops...correct....bitkeeper then Aug 31 05:55:37 but it does refer to bitbake as a build tool......I am just trying to deconfuse myself and get started, so to dpeak Aug 31 05:55:40 speak Aug 31 05:56:01 bitbake is a buildtool Aug 31 05:56:08 OE is the metadata to feed it Aug 31 05:58:05 and bk (bitkeeper) is no longer used, and has been replaced by "Monotone", whatever that is ? Aug 31 05:58:44 yes Aug 31 05:59:52 OK....now I'm getting somewhere....:-) Aug 31 06:00:40 the other required tools and versions are still the same ? (ex subversion, python, et al) Aug 31 06:02:33 more or less, yeah. having the latest svn and python doesn' hurt Aug 31 06:02:51 hm.....perhaps subversion is no longer required, as it seems to only have been used to get bitkeeper Aug 31 06:03:23 all of that I still have set up....so I'll leave that as is.....that's been working ok Aug 31 06:03:32 03mickeyl 07org.oe.dev * r9d7a75c8... 10/packages/qmake/ (qmake-native_1.07a.bb qmake-native_2.00a.bb): qmake: build 1.07a from QtE 3.3.4, build 2.00a from Qt/X11 4.0.1 Aug 31 06:03:59 OSS542: svn is used to get bitbake Aug 31 06:04:33 and packages in subversion in OE Aug 31 06:04:56 I see.... so I'd better leave that as is also.....:-) Aug 31 06:05:33 I just have to rip out bitkeeper and get Monotone in place of it Aug 31 06:05:41 and I should be good to go Aug 31 06:05:59 yes Aug 31 06:06:07 as described in GettingStarted Aug 31 06:19:24 hey Aug 31 06:20:22 hey Aug 31 06:22:20 since the nslu2-linux project started using monotone I've been having merging problems. Here's a new one: Aug 31 06:22:21 monotone: trying 3-way merge Aug 31 06:22:21 monotone: misuse: unable to resolve file conflict 'packages/kdepimpi/kdepimpi_2.1.18.bb' -> 'packages/kdepimpi/kdepimpi_2.2.2.bb' vs. 'packages/kdepimpi/kdepimpi_2.2.3.bb' Aug 31 06:22:51 I mean, I made some changes, but nothing but nslu2 related stuff. Aug 31 06:22:53 VoodooZ_Work: mt --lca merge Aug 31 06:22:59 tried that. Aug 31 06:23:14 same result. Aug 31 06:23:14 install a merge tool and merge it by hand Aug 31 06:23:22 or do a explicitit_merge Aug 31 06:23:46 well, that's what it did before. it would call meld like 10 times for files I've never touched. Aug 31 06:24:16 I don't mind fixing it manually but I wish I could understand what's wrong so to avoid it. Aug 31 06:24:42 VoodooZ_Work: do pull, update, then and only then commit Aug 31 06:25:08 RP, thanks. Aug 31 06:25:37 The problem, or so I've been told, is that I make private changes to my repo and never push them. Aug 31 06:25:40 pull, merge, update Aug 31 06:25:44 RP: that still doesn't help Aug 31 06:25:56 I'm sure I tried that in the past anyways. Aug 31 06:25:58 VoodooZ_Work: Ah, if its private changes, that gets tricky :-/ Aug 31 06:26:03 RP: if you want to propagate to another branch you have to merge everything again Aug 31 06:26:04 yeah. Aug 31 06:26:11 F'n monotone. Aug 31 06:26:17 :) Aug 31 06:26:21 * koen wants empty merges in monotone Aug 31 06:26:36 I never thought I'd miss BK. Aug 31 06:27:14 I just work off the one branch so I guess I don't see these problems. All my local changes aren't committed... Aug 31 06:27:16 RP: I've flashed the new kernel and can confirm it sees all of the memory (though I imagine someone's already told you as much) Aug 31 06:27:36 lardman|work: They haven't. Its nice to know though, thanks :) Aug 31 06:28:03 lardman|work: PResumably it correctly says its a Shepherd in /proc/cpuinfo now - that's the critical bit Aug 31 06:28:11 RP, yeah. I'm running out of ideas and I'm that close to just ditching the whole thing. Every update is a pain in the neck and takes me away of real dev work. Aug 31 06:28:35 VoodooZ_Work: Keep you local changes as a patch set using quilt? Aug 31 06:28:47 last resort is to stick with the tarball release but then I have merge my changes everytime there's a new release. Aug 31 06:29:23 VoodooZ_Work: with quilt you can then "quilt pop -a; monotone update; quilt push -a" Aug 31 06:29:45 RP, somebody suggested that a while ago but it seemed convoluted. Perhaps I'll have to check it out again. Aug 31 06:30:03 * RP likes quilt Aug 31 06:30:38 so the POP function would "remove" the changes and push would reapply them? Aug 31 06:30:49 VoodooZ_Work: Yes, and the -a means all patches Aug 31 06:31:08 ok, sounds easy enough. but what about the actual changes? Aug 31 06:31:22 Do you have to do backflips everytime you change a file? Aug 31 06:31:40 VoodooZ_Work: write a mail to the monotone list Aug 31 06:31:42 "quilt new new_patch.patch; quilt add some_file_you_want_to_change" Aug 31 06:31:53 complaining over here obviously doesn't fix the issues Aug 31 06:32:15 koen, I agree but I was just looking for alternatives. Aug 31 06:32:35 RP, thanks for the help. much appreciated. Aug 31 06:32:39 "quilt refresh" will update the patches. quilt will let you know if you forget :) Aug 31 06:32:50 perfect! Aug 31 06:33:41 RP: yes it does Aug 31 06:34:32 * mickeyl recommends patcher Aug 31 06:34:46 lardman|work: excellent Aug 31 06:35:29 mickeyl, patcher you say? Aug 31 06:36:05 VoodooZ_Work: It does a similar thing. I find quilt slightly better, mickeyl likes patcher :) Aug 31 06:36:31 anythiung that saves me work is welcome. the more transparent the better. Aug 31 06:36:47 I'm looking into both. Aug 31 06:38:57 I was reading about unionfs and was wondering If that could come in handy here. ummm. Aug 31 06:39:33 it will... Aug 31 06:39:39 ... as soon as it is stable Aug 31 06:40:03 still it OOPses too much - that is even on well tested x86 boxes Aug 31 06:40:50 I see. Aug 31 06:42:02 RP: what do you think to disabling the ide part of Connect+ card for now? Aug 31 06:43:01 XorA: If it doesn't work, its fair game Aug 31 06:44:03 Good evening. Aug 31 06:44:13 hi to all Aug 31 06:44:37 RP: mine hasnt been working at all until I turned off that ID in the ide-cs table Aug 31 06:45:06 RP: ide and wifi seem to tread on each other Aug 31 06:45:51 XorA: I can understand that - feel free to add a patch to the kernel to do that Aug 31 06:46:18 (at the bottom of the patch list, clearly marked as a hack) Aug 31 06:46:54 RP: ok will do Aug 31 06:50:43 hmm, if-pre-up/hostap-fw-load doesnt bother checking that the interface is a wireless one before trying to firmware it Aug 31 06:57:48 RP: done Aug 31 06:58:12 03xora 07org.oe.dev * r4e126c45... 10/packages/linux/ (2 files in 2 dirs): linux-openzaurus-2.6.13-rc5-mm1: add a patch to remove the SanDisk ConnectPlus card from the ide-cs driver. As both the ide and hostap drivers trying to access this card at same time is not currently working. Aug 31 06:58:58 ~curses edje Aug 31 06:59:02 ~curse edje Aug 31 06:59:03 May you be reincarnated as a Windows XP administrator, edje ! Aug 31 07:00:24 heh.. I just had to start build of rc6-mm2-r19.. now have to merge with rc5-mm1-r20 Aug 31 07:00:41 hrw|work: the patch is simple Aug 31 07:00:45 hrw|work: one liner Aug 31 07:01:12 XorA: I want to follow RP current kernel as much as possible Aug 31 07:04:29 NOTE: package linux-openzaurus-2.6.13-rc6-mm2-r20: task do_patch: completed Aug 31 07:04:54 hi pb Aug 31 07:06:22 monotone list unknown lists 6452 of 13200 files from metadata ;( Aug 31 07:06:46 found why..uf Aug 31 07:07:00 hi hrw|work Aug 31 07:13:58 pb_: Can you tell me if libxsettings-client_0.14.bb still need the libgpewidget dependency? Aug 31 07:14:58 I'm also guessing that nokia770_init should have an RDEPENDS="hotplug" rather than a DEPENDS="hotplug-dbus" ? Aug 31 07:23:54 03cbrake 07org.oe.dev * r9178fa6f... 10/ (7 files in 4 dirs): devkitidp-pxa255: rename vibren-pxa255idp machine to devkitidp-pxa255 Aug 31 07:31:08 morning Aug 31 07:31:19 hi kergoth Aug 31 07:31:52 hi kergoth Aug 31 07:32:34 * kergoth cracks his neck, then yawns Aug 31 07:57:01 RP: I don't think so. Aug 31 07:57:12 regardng the nokia770 stuff, you'd need to check with Florian. Aug 31 07:58:38 pb_: you know offhand if the handhelds.org admin interface is working atm? i need to change the address i'm forwarding my mail to Aug 31 07:59:20 kergoth: send me a note, while I am siting in front of the admin box Aug 31 07:59:30 k Aug 31 07:59:33 thanks Aug 31 07:59:40 welcome Aug 31 07:59:58 france_: I will msg you the same Aug 31 08:00:05 :-) Aug 31 08:00:06 hi france_ Aug 31 08:00:09 pb_: hi Aug 31 08:05:02 kergoth: done, test mail sent. Aug 31 08:05:22 hrw|work: done Aug 31 08:06:21 france_: worked, thanks Aug 31 08:06:30 kergoth: welcome. Aug 31 08:07:24 anyone recently "rm -rf build/tmp" and find that glibc fails with an "expand: input line is too long" error? Aug 31 08:07:45 Zero_Chaos: I sense you are on gentoo Aug 31 08:08:01 * kergoth boots a current oe image in uml Aug 31 08:08:03 koen: give the man a cookie, is there a workaround? Aug 31 08:08:19 yes Aug 31 08:08:23 * koen searches for it Aug 31 08:08:28 ~logs Aug 31 08:08:31 apt/ibot/jbot/purl all log to http://ibot.rikers.org// where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/.html.gz, or updated "nightly" Aug 31 08:09:29 17:41.18 koen|slug 19:38 < peteru-home> Arghh!!! Effing gentoo patches just cover up the broken expand command by patching the testsuite to not run tests on expand. Aug 31 08:09:32 17:41.28 mickeyl heh Aug 31 08:09:35 17:42.03 koen|slug people how emerged world and use OE might have some troubles Aug 31 08:09:48 the solution is in gentoo bugzilla Aug 31 08:10:45 * Zero_Chaos begins to search Aug 31 08:11:23 stupid slow internet connection.... Aug 31 08:15:17 is it odd that I get this error and I don't even have app-text/expander installed? Aug 31 08:16:05 coreutils 2.3.0/expand Aug 31 08:16:29 ahh Aug 31 08:17:19 I can't seem to find it on bugs.gentoo.org.... can you help me out :-) Aug 31 08:17:48 peteru-home said: it's in gentoo bugzilla now Aug 31 08:17:53 that's all I know Aug 31 08:18:02 you could try to change the locale Aug 31 08:18:33 or I could downgrade coreutils one version.... Aug 31 08:18:54 I'm ~x86.... maybe I should send that back to stable branch Aug 31 08:30:30 root@powerlink:~# ssh localhost Aug 31 08:30:30 Host 'localhost' is not in the trusted hosts file. Aug 31 08:30:30 (fingerprint md5 94:34:d1:63:26:dc:a4:c9:07:98:80:7a:9c:94:18:4f) Aug 31 08:30:30 Do you want to continue connecting? (y/n) Aug 31 08:30:30 y Aug 31 08:30:31 Segmentation fault Aug 31 08:30:32 hehe Aug 31 08:30:38 dropbear cant ssh into itself? :P Aug 31 08:31:09 kergoth: You seem to have a knack for breaking things :) Aug 31 08:31:41 thats whats weird, i really dont.. i can never reproduce people's problems, always Just Works Aug 31 08:31:45 at least until recently ;) Aug 31 08:31:52 kergoth: maybe the developers though that type of thing was too adult to allow? Aug 31 08:32:13 RP: he must be working for microsoft :) Aug 31 08:32:30 kergoth: You've become a user? ;-) Aug 31 08:32:33 * RP hides Aug 31 08:32:37 * kergoth hangs head in shame Aug 31 08:33:22 Its probably OE/bitbake missing you ;-) Aug 31 08:33:31 hehe Aug 31 08:33:41 ah good, we do have vlan in oe Aug 31 08:34:22 hmm, wonder if there's /etc/network/interfaces integration for vlan, or if i'll have to set up the pre-up bits myself Aug 31 08:35:19 mickeyl: ping Aug 31 08:35:26 kergoth: having the real ifupdown is recommended Aug 31 08:35:47 * kergoth nods Aug 31 08:35:48 that will save you from a few 'wtf?' moments Aug 31 08:36:22 hmm... Aug 31 08:37:16 RP: backlight values has changed in latest kernel? I got black screen when started opie and after some time it goes lighter (just to go black after press anyway) Aug 31 08:38:06 RP: maximal value (in opie atleast) == black screen Aug 31 08:38:36 hrw|work: what device is that on? Aug 31 08:38:44 Zero_Chaos: c7x0 - husky Aug 31 08:39:05 * Zero_Chaos breathes a sigh of relief for his collie... Aug 31 08:39:09 hrw|work: Ah, I think I suddenly understand a problem. Yes, they've changed and I need to fix something Aug 31 08:39:50 Zero_Chaos: my collie sits whole days in cradle and wait for 3.5.4 pre-release Aug 31 08:40:29 hrw|work: I've been slowing makeing an image with kismet/aircrack and simialar tools hopeing for 3.5.4 so my image is more stable :-) Aug 31 08:40:42 sectect-gpe-image? Aug 31 08:41:31 koen: I enjoy building it myself, but out of curiousity, where can I find that? Aug 31 08:41:50 bitbake sectest-gpe-image Aug 31 08:41:59 see packages/meta/ Aug 31 08:42:04 hmm... I'll check it out, thanks Aug 31 08:43:17 aw, the oe vlan package doesnt include the ifupdown integration Aug 31 08:43:28 debian and ubuntu have the scripts for integration.. just a matter of pulling them in Aug 31 08:44:22 koen: WOW, that meta-sectest is rather impressive, thanks for the info. Aug 31 08:44:43 * koen should add the logo to the image Aug 31 08:45:12 this is going to be the splashscreen: http://dominion.kabel.utwente.nl/koen/pda/files/sectest-logo.pdf Aug 31 08:48:07 koen: that is sweet Aug 31 08:55:50 cu all Aug 31 09:19:29 koen: do you know if gpe-sectest will fit on a collie? Aug 31 09:20:02 Zero_Chaos: sure. Use an SD card Aug 31 09:20:37 JustinP: it's an image, I don't think I can flash to SD Aug 31 09:21:00 make OE build a ext2 or tar image Aug 31 09:21:51 Zero_Chaos: OE can build any kind of image you want. It does bz2 as well as jffs2 by default Aug 31 09:22:02 koen: and what do I load that then? Aug 31 09:22:13 don't I still have to flash it onto the Z? Aug 31 09:22:14 Zero_Chaos: I mean move root to SD. You can make a simple pivotboot image for the flash and have the main root on SD Aug 31 09:22:18 I always wondered what the bz2 was Aug 31 09:22:23 Zero_Chaos: see the pivotboot image used for the C3000 Aug 31 09:23:24 * Zero_Chaos is so confused... Aug 31 09:24:06 JustinP: I built an e-image-core, but it's still a bit too large - any ideas on ways to reduce the disk size of the fonts (which appear to be the largest component)? Aug 31 09:24:22 lardman|work: the fonts? don't you mean the themes? Aug 31 09:24:38 JustinP: No, def. the fonts Aug 31 09:24:52 JustinP: Hang on, I'll pastebin it all Aug 31 09:25:43 JustinP: Or not, I forgot I'd built a couple of gpe-images - the rootfs has gone Aug 31 09:25:58 JustinP: One of the fonts was ~12MB and another was ~7MB! Aug 31 09:26:09 wowie Aug 31 09:26:42 I'll just run bitbake e-image-core again and post the du & ls output or the rootfs Aug 31 09:35:00 lardman|work: interesting...I thought all this time it was the themes... Aug 31 09:35:16 fireflysung.ttf and kochi-gothic.ttf seem to be the big space killers Aug 31 09:36:06 Yep, those are the ones Aug 31 09:36:18 and they are pretty huge too Aug 31 09:37:00 My initrd.bin came out at 37MB which is too big as I only have ~32MB for it on my C750 Aug 31 09:38:18 I understand Aug 31 09:38:39 If anyone else is interested, here's the breakdown: http://pastebin.ca/21742 Aug 31 09:39:53 JustinP: Do you have any idea what those fonts are used for? Coule something be used instead, or the fonts trimmed (why would they be so large anyway?)? Aug 31 09:40:21 I have no idea Aug 31 09:40:25 on any of thos epoints ;-) Aug 31 09:40:31 I'll look around Aug 31 09:40:41 cool, thanks Aug 31 09:41:43 I worked more than 30 hours last weekend so I didn't have any time for OE ;-) Aug 31 09:42:08 lardman|work: which package do they come from ? e-wm ? Aug 31 09:42:40 mickeyl: I don;t know, how do I find out? grep my tmp/work directory? Aug 31 09:42:45 nah Aug 31 09:42:54 ipkg Aug 31 09:43:00 ipkg --whatprovides or so Aug 31 09:43:18 on the build machine? Aug 31 09:43:21 sure Aug 31 09:43:27 ah, ok Aug 31 09:43:35 or, much fsater, check /usr/lib/ipkg/info/*.list Aug 31 09:43:52 yep, e-wm Aug 31 09:43:59 mm, http://www.expansys-usa.com/product.asp?code=114076 nifty. i should pick those up Aug 31 09:44:08 I don't see them referenced in any files in the e-em source, though... Aug 31 09:44:13 e-wm from the looks of it Aug 31 09:44:35 if they're not referenced, just yank them from the rootfs, build an image and check if it still works Aug 31 09:45:00 how do I build the image without the roofs being re-populated? Aug 31 09:45:19 call the image command manually Aug 31 09:45:45 bitbake -c image e-image-core ? Aug 31 09:45:56 no no Aug 31 09:46:07 do the same what bitbake would do Aug 31 09:46:08 mkfs.jffs2 Aug 31 09:46:20 oh right Aug 31 09:46:35 or even better Aug 31 09:46:40 edit e-image Aug 31 09:46:54 IMAGE_PREPROCESS_COMMAND = "rm ${IMAGE_ROOTFS}/foo/bar" Aug 31 09:47:07 then just bitbake the image again Aug 31 09:47:10 okay, that sounds like a plan Aug 31 09:47:11 or even better...patch the makefile Aug 31 09:47:16 * JustinP is going to do that Aug 31 10:00:27 * JustinP did that Aug 31 10:02:04 is anyone else having problems building gconf-dbus? "/opt/oe/build/tmp/staging/arm-linux/lib/libpangox-1.0.so: undefined reference to `_pango_engine_shape_covers'" Aug 31 10:02:10 I've rebuilt pango but to no avail Aug 31 10:02:27 try removing the builddir and start from scratch Aug 31 10:02:50 which builddir? gconf-dbus? build/tmp? Aug 31 10:03:15 build/tmp/work/gconf-dbus* Aug 31 10:03:34 phew Aug 31 10:03:46 I was hoping you'd say that Aug 31 10:04:34 no, TMP_DIR Aug 31 10:04:47 I was hoping you wouldn't say that :-( Aug 31 10:05:34 Ok - so a bitbake question - is it possible to set a variable in a fetch class for use by a .bb? Aug 31 10:06:04 mreimer: just cleaning and rebuilding everything that depends on cairo (gtk and gpe) worked for me Aug 31 10:06:14 eh pango Aug 31 10:06:17 thanks pH5 Aug 31 10:06:40 is there a bitbake command to rebuild all packages that depend on a given package? Aug 31 10:09:23 mreimer: If there is, I'd like to know. If your computer is reasonably fast, perhaps you should do as koen says - might cost more time, but at least not yours. Aug 31 10:09:39 pH5: PIII 450MHz :-( Aug 31 10:10:45 oh. Aug 31 10:15:13 This is the problem with OE... it's to easy to build a dist, but to hard to figure out a name for the dist :) Aug 31 10:16:05 'floppy the happy penguin' Aug 31 10:16:51 floppy... just gets me thinking about the old Amiga demo flossy :) Aug 31 10:17:08 hmmm...whatw as I thinking...e-image-core doesn't work anyway.... Aug 31 10:19:07 Gonna list a few names that we have come up with... Maybe it can trigger some ideas for a few more? Aug 31 10:19:18 whats the nature of your distro? Aug 31 10:19:33 Tiny UxIXOS ( you know who you are who suggested this =) , FreeTux , Freevox , Efux / Efix - Embedded Freevo Linux , Feux / Feix - Freevo Enhanced Linux , Evux / Evix - Enhanced Video Linux Aug 31 10:20:10 haha, efux.. dirty Aug 31 10:20:15 =) Aug 31 10:20:32 Feux is also nice :) Aug 31 10:21:01 Heh Aug 31 10:21:03 Feux = fire, if i remember my english lessions correctly Aug 31 10:21:08 Then we could get shits that say Feux you! Aug 31 10:21:13 s/shits/shirts/ Aug 31 10:21:19 lol Aug 31 10:21:20 hehe Aug 31 10:21:21 =)) Aug 31 10:21:39 And little stickers for our PDAs that say "I got Feuxed! Aug 31 10:21:49 * JustinP attempts to fix the keyboard problems in e-core Aug 31 10:22:21 hehe... well, this is not really for a PDA... it's for those small Epia based mb's Aug 31 10:24:01 hmmm... decisions decisions... it's just to hard to come up with a name... Aug 31 10:24:24 NoName - The Embedded Freevo distribution =) Aug 31 10:26:47 i really cant wait to convince my work to switch to oe Aug 31 10:26:53 the setup we're using now really sucks Aug 31 10:27:28 doubt itll happen for the current product, since we cant risk a huge changej, but for the next one.. Aug 31 10:35:57 <[cc]smart> i'm on setting 2.6 kernel for building a collie image so that i could see if it compiles ok at least (can not test really but it should compile) Aug 31 10:36:19 <[cc]smart> i found : PREFERRED_PROVIDER_virtual/kernel ?= "collie-kernels-2.4-embedix" Aug 31 10:36:37 <[cc]smart> so i should define a new PREFERRED_PROVIDER_virtual/kernel on local.conf Aug 31 10:36:40 <[cc]smart> correct ? Aug 31 10:37:06 <[cc]smart> and in the case, this should then be linux-openzaurus_2.6.11 Aug 31 10:37:09 <[cc]smart> er. Aug 31 10:37:25 <[cc]smart> wrong line... this one: linux-openzaurus_2.6.13-rc5-mm1 Aug 31 10:37:30 <[cc]smart> correct ? Aug 31 10:37:38 no Aug 31 10:37:50 preferred provider takes a package name, not a version Aug 31 10:38:21 <[cc]smart> linux-openzaurus_2.6.13-rc5-mm1.bb Aug 31 10:38:59 yes, the package name of that package is linux-openzaurus, the package version is 2.6.13-rc5-mm1 Aug 31 10:39:05 _ seperates name from version Aug 31 10:39:37 <[cc]smart> but doesn't the eoriginal entry then violate this ? Aug 31 10:40:01 no. Aug 31 10:40:06 no, "collie-kernels-2.4-embedix" is a collection Aug 31 10:40:17 go look at that collie-kernels .bb. Aug 31 10:40:26 do you see a _ between 2.4-embedix and collie-kernels? Aug 31 10:40:27 i dont. Aug 31 10:40:54 <[cc]smart> hmm... don't see this in the bb file either Aug 31 10:40:59 <[cc]smart> in the name of Aug 31 10:41:18 what? Aug 31 10:41:21 <[cc]smart> collie-kernels-2.4-embedix.bb Aug 31 10:41:32 yes, the package name of that .bb is collie-kernels-2.4-embedix Aug 31 10:41:35 * JustinP boots into e-image-core Aug 31 10:41:36 the package version isnt set by the filename Aug 31 10:41:43 this really isnt that hard to understand Aug 31 10:42:01 JustinP: without the fonts? Aug 31 10:42:07 once it fscked itself....again... Aug 31 10:42:08 lardman|work: yes Aug 31 10:42:13 <[cc]smart> well, i thought ${PV} is taken from there Aug 31 10:42:23 lardman|work: however the e-image-core was broken anyway. You can't type in entrance :-| Aug 31 10:42:29 lardman|work: I'm going to try to fix that now as well Aug 31 10:42:54 JustinP: Right, my attempts to get the fonts removed have failed one way or another anyway Aug 31 10:43:10 [cc]smart: yes, PV is taken from the filename _if it is defined there_ Aug 31 10:43:12 lardman|work: hehe Aug 31 10:43:15 [cc]smart: if not, it gets set to a default Aug 31 10:43:22 [cc]smart: look at the PV = line in bitbake.conf Aug 31 10:43:51 JustinP: What machine are you building for? husky? Aug 31 10:44:17 lardman|work: I have a spitz Aug 31 10:44:25 lardman|work: BTW, here's the patch I used: http://pastebin.ca/21747 Aug 31 10:45:50 JustinP: Thanks Aug 31 10:46:25 ~husky Aug 31 10:46:26 hmm... husky is Sharp Zaurus SL-C760 or a dog Aug 31 10:47:08 JustinP: I'll give it another go tomorrow, I've got to head off, good luck with fixing Entrance Aug 31 10:47:11 Night all Aug 31 10:54:43 <[cc]smart> so what was done is to make use "collie-kernels-2.4-embedix" as a version. Aug 31 10:55:17 <[cc]smart> correct ? Aug 31 10:55:20 [cc]smart: once again, PREFERRED_PROVIDERS_ takes a package NAME Aug 31 10:55:21 NOT A VERSION Aug 31 10:55:42 perhaps you should read the bitbake documentation before asking anymore questions Aug 31 10:55:46 <[cc]smart> once again, this is the line i found: PREFERRED_PROVIDER_virtual/kernel ?= "collie-kernels-2.4-embedix" Aug 31 10:55:51 yes Aug 31 10:56:00 and as i just said, preferred provider takes a package name Aug 31 10:56:03 so what does that tell you? Aug 31 10:56:13 lets practice our logic, shall we? Aug 31 10:56:17 if preferred provider takes a package name Aug 31 10:56:24 and we're setting it to collie-kernels-2.4-embedix Aug 31 10:56:30 then collie-kernels-2.4-embedix must be a package name Aug 31 10:56:33 isnt logic fun? Aug 31 10:56:44 <[cc]smart> goan Aug 31 10:57:03 kergoth: do you have pointy ears? Aug 31 10:57:12 hehe Aug 31 11:05:18 <[cc]smart> kergoth: actually yes, you're right, me stupid. why should i at all have to get pissed off by you. don't need to hook on you, do i ? Aug 31 11:05:51 <[cc]smart> we had that. Aug 31 11:06:33 the *very first* think I tried fixed entrance....amazing Aug 31 11:06:40 so, you wouldnt set the preferred provider to linux-openzaurus_2.6.whatever, because that isnt the package name. you'd set it to linux-openzaurus, which is, and then you'd set the preferred version appropriately. Aug 31 11:06:45 JustinP: nice Aug 31 11:11:30 hey kergoth soemone is looking for a working tailor svn/k Aug 31 11:11:46 the tailor guy fixed a couple of my reported bugs Aug 31 11:11:52 need to test the current version and see how it is Aug 31 11:11:52 yeah Aug 31 11:12:21 kergoth: monotone < tailor > svk? Aug 31 11:12:43 * kergoth nods Aug 31 11:14:11 are we still keeping monotone with the new issues? Aug 31 11:14:32 i havent obviously been paying attention Aug 31 11:14:56 new issues? Aug 31 11:16:00 servers being down and wasnt there something about a miss-sync? Aug 31 11:16:24 I don't know if any new issues....it's worked fine for me for quite a while now Aug 31 11:16:33 yes, servers *can* go down.... Aug 31 11:17:20 that's why we have backup servers Aug 31 11:17:24 and I repeat: Aug 31 11:17:44 #newscm has to be *a lot* better as monotone, not 'just faster' Aug 31 11:18:26 yeah Aug 31 11:18:41 svk gives you everything monotone has, less headaches wrt branching/merging/multiple heads, better performance, and lets you use svn to check out anonymously automatically, since its svn under the hood Aug 31 11:18:43 among other things. Aug 31 11:19:04 but i remember kergoth saying what he just pasted again earlier Aug 31 11:20:01 kergoth: then float svk past oe@hh. Make sure to address all of the points in the earlier e-mail about monotone and new possibilities. Aug 31 11:20:11 lol Aug 31 11:20:25 i thought kergoth's choice was the only one that mattered Aug 31 11:20:36 JustinP: thats the plan, just need to get tailor behaving so we can retain history Aug 31 11:20:50 emte: hehe, there are a lot more active devs on oe now than there used to be :) Aug 31 11:20:55 and i'm not one of them Aug 31 11:20:57 true Aug 31 11:21:08 31 keys on the db Aug 31 11:21:15 aw cmon we all know youhave secretly been polishing bb2 Aug 31 11:21:27 of which 20 are committing on a regular basis Aug 31 11:21:32 i wish.. but i'm planning on revisiting it again soon Aug 31 11:21:34 :) Aug 31 11:22:01 reminds me i need to get bb working ... Aug 31 11:22:13 also going to do a linux firewalling/routing tool that instead of just being a shitload of iptables commands, lets you use lua to express your routing/filtering/mangling policies, and itll do the implementation for you Aug 31 11:22:21 * kergoth is happy to be all opensourcey again Aug 31 11:22:35 ah lua Aug 31 11:22:45 so you can embed it in monotone ;) Aug 31 11:22:55 ... Aug 31 11:22:59 *g* Aug 31 11:23:19 it will take you an hour to test the rules :P Aug 31 11:23:32 no it wont Aug 31 11:23:40 i know Aug 31 11:24:11 lets see .. Aug 31 11:24:49 JustinP, how hard is it going to be to strip out all your gpe stuff from ecore, evas, and edje Aug 31 11:24:59 i havent looks at your metafiles yet Aug 31 11:25:08 emte: *what*? Aug 31 11:25:22 JustinP: just /ignore emte Aug 31 11:25:27 emte: what GPE stuff are you talking about? There is *nothing* GPE related in those things... Aug 31 11:25:36 good to know Aug 31 11:25:38 save you from reading a lot of uninformed crap Aug 31 11:25:43 i dont need any of the x stuff Aug 31 11:25:44 heh Aug 31 11:25:57 not for the server side Aug 31 11:25:58 emte: GPE and X are completely different things Aug 31 11:26:28 and if you had actually looked at what's in the repo..... Aug 31 11:26:31 i am aware Aug 31 11:26:38 emte i havent looks at your metafiles yet Aug 31 11:26:39 * koen emphasizes on 'uninformed' Aug 31 11:26:47 emte emte i havent looks at your metafiles yet Aug 31 11:27:09 emte: next time please try looking at the metafiles first Aug 31 11:28:22 that would be my intent after having a BB system that works Aug 31 11:28:28 <[cc]smart> let me make a guess... the growth in devs on OE comes from external projects, right ? Aug 31 11:29:11 mickeyl: any thoughts on slfb? I'm about ready to commit it.... Aug 31 11:37:47 whee 390 revs Aug 31 11:38:49 .... Aug 31 11:39:09 why did florian change all of these x11-common depends to xserver-common? There *is* no xserver-common... Aug 31 11:39:12 i think he went to sleep Aug 31 11:41:08 yes there is Aug 31 11:41:19 koen: in OE? Aug 31 11:41:25 *I* don't see one Aug 31 11:41:47 I know this is what he wanted to rename it to.. Aug 31 11:42:03 http://ftp.handhelds.org/pub/projects/gpe/source/xserver-common-1.1.tar.gz Aug 31 11:42:20 koen: but it wasn't added to OE, damn it Aug 31 11:42:30 argh Aug 31 11:42:34 * JustinP is adding it now Aug 31 11:42:48 just annoying that he changed all of the depends and didn't fix the package Aug 31 11:43:23 koen@dominion:/data/build/oe/monotone/org.openembedded.dev$ ls packages/xserver-common/ Aug 31 11:43:26 ~lart florian Aug 31 11:43:27 * ibot takes out florian with the trash Aug 31 11:43:27 xserver-common_1.1.bb Aug 31 11:43:36 ~lart justinp for being an emte Aug 31 11:43:36 * ibot cuts off justinp's head with a halberd that could have been a little bit sharper for being an emte Aug 31 11:43:50 * emte shrugges Aug 31 11:43:58 god damn it Aug 31 11:44:09 I swear I looked for it... Aug 31 11:44:10 you get buzy and dont pay attention Aug 31 11:44:10 oh well Aug 31 11:45:09 my repo is even ~2weeks out of date Aug 31 11:45:52 my repo is not out of date Aug 31 11:45:59 I just did a grep, which of course missed the filename Aug 31 11:46:16 nah, i ment in response to koen Aug 31 11:46:43 why the hell if this run-calibrate.sh in here? Aug 31 11:46:46 we cant all be as smart as him Aug 31 11:47:17 is Aug 31 11:47:45 so it only runs if you have X? Aug 31 11:47:58 eh? Aug 31 11:48:10 just a guess Aug 31 11:48:36 why would we need run-calibrate.sh, which only runs "exec xtscal" which is called directly from the clibrate Xinit script Aug 31 11:48:44 * JustinP will patch that out in a minute Aug 31 11:48:46 lunch first Aug 31 11:50:03 hmm ... i could strip my laptop down and use it for a server ... Aug 31 12:10:55 JustinP: slfb ? Aug 31 12:11:00 that simulated framebuffer thing? Aug 31 12:11:32 I think the means the the fb version of slbl Aug 31 12:11:46 hmm Aug 31 12:11:47 sharp 2.4 utils Aug 31 12:11:49 slbl is fb only Aug 31 12:11:52 actually it's console Aug 31 12:12:00 i know it , i wrote it Aug 31 12:12:11 JustinP said something about blanking the fb a while ago Aug 31 12:12:28 I guess he wrote an app to poke at the right ioctls Aug 31 12:18:30 03nail 07org.oe.dev * rdc7f6044... 10/packages/openntpd/ (files/init openntpd_3.7p1.bb): Aug 31 12:18:30 openntpd init-script Aug 31 12:18:30 Check to see if privsep dir {localstatedir}/shared/empty exists. Create if not. Aug 31 12:18:34 03justinp 07org.oe.dev * rb7600f4a... 10/packages/e17/entrance_20050810.bb: entrance: add glibc-gconv-iso8859-1 to RDEPENDS, fixes earlier issue with keyboard not working Aug 31 12:18:38 03nail 07org.oe.dev * r06f3983d... 10/packages/openntpd/files/init: Openntpd init-script. Fix directory check Aug 31 12:18:42 03justinp 07org.oe.dev * r79ddb5aa... 10/packages/e17/ (e-wm/remove-large-fonts.patch e-wm_20050810.bb): Aug 31 12:18:42 e-wm: Remove 2 large unused fonts Aug 31 12:18:43 - This makes e-image-core small enough for most devices :-) Aug 31 12:20:37 ~praise JustinP for finding the keyboard bug Aug 31 12:20:46 All hail JustinP for finding the keyboard bug! Aug 31 12:21:06 good work JustinP! What's the image size down to now? Aug 31 12:24:50 must be ~30 Aug 31 12:28:47 hi all Aug 31 12:29:08 hi pb_ Aug 31 12:29:27 mickeyl: good evening Aug 31 12:56:13 mickeyl: Just tried a 2.6 opie spitz image. tscalibration isn't working and selecting rotate from the menu doesn't do anything either. auto rotate doesn't work but that may or may not be part of the latter problem Aug 31 12:56:57 auto rotate calls the same code as the rotate menu Aug 31 12:57:39 I'll build the latest bits for my c7x0 Aug 31 12:58:20 Did zecke mention a sign error in one of the touchscreen quadrants? Aug 31 12:58:33 something like that Aug 31 12:58:46 I'll bet this is the touchscreen calibration problem... Aug 31 12:58:55 iirc he told something about the matrice not being fully calculated in opie and gpe ts calibration Aug 31 12:58:56 <[cc]smart> RP: hi. seems it'll take me far more time for test compiling than swicthing those schedulers to module. do you think i should sync these in anyways or will this simple change just be done by somebody who already does OZ ? Aug 31 12:59:37 [cc]smart: I can make the change if you prefer... Aug 31 12:59:59 mickeyl: The same thing happens with gpe... Aug 31 13:02:00 mickeyl: I think we could conciveably ship a 2.6 kernel with 3.5.4... Aug 31 13:02:15 hmm Aug 31 13:02:29 what's missing on spitz/akita/borzoi? Aug 31 13:02:43 USB UDC and sound Aug 31 13:02:53 I mean in addition to a 2.4 kernel Aug 31 13:03:13 <[cc]smart> RP: yes, think that's easiest... will stick to slug :) Aug 31 13:03:15 I'm not sure we can handle the support load for both Aug 31 13:03:34 UDC? Aug 31 13:03:40 mmm.... UDC Aug 31 13:03:47 JustinP: usb subset Aug 31 13:03:52 I thought you had a hacked sound module working. Of course...it's a *hac* still... Aug 31 13:03:56 JustinP: USB Client Aug 31 13:03:59 koen: subset for what? Aug 31 13:04:12 bbiab Aug 31 13:04:13 so host is working but it can't be a device? Aug 31 13:04:19 JustinP: I have hacked sound for c7x0 which can probably be easily extended to cover cxx00 Aug 31 13:04:36 JustinP: correct. host works Aug 31 13:05:22 mickey|bbiab: http://dioxide.randomvoids.com/~papercrane/slfb.c.txt Aug 31 13:05:34 RP: Ah, I see. Aug 31 13:05:36 I just got an email to say some of my pxa27x host changes are making mainline soon Aug 31 13:05:43 awesome :-) Aug 31 13:08:49 JustinP: I'm told I need to mention the cxx00 changes to the keyboard rotation switches Aug 31 13:09:17 JustinP: Instead of passing an EV_KEY when these change, the 2.6 kernel will be passing an EV_SW Aug 31 13:10:07 JustinP: See http://www.rpsys.net/openzaurus/patches/corgikbd_switch-r0.patch for details - it probably has implications for xrandr Aug 31 13:10:55 RP: I have never touched xranrd code. Just played with the scripts which call it Aug 31 13:10:59 xrandr Aug 31 13:11:15 xrandr also does not do auto-rotation at all Aug 31 13:11:21 GPE has never had auto-rotation AFAIK Aug 31 13:11:38 JustinP: Well, the scripts that call it could be modified to do autorotation then... Aug 31 13:11:56 I got it working by using keylaunch to watch for the hinge-key and run a script for me Aug 31 13:12:24 Ah, then you'll need to watch for EV_SW events in future... Aug 31 13:12:31 (or under 2.6) Aug 31 13:13:02 are you talking about the hinge sensor or the hinge key? Aug 31 13:13:36 The one and the same. The only way you see it under 2.6 is via EV_SW events from /dev/input/event* Aug 31 13:14:24 no, they are *not* the same Aug 31 13:14:53 the sensor is what chkhinge uses to see what way it's rotated Aug 31 13:15:03 (sorry, I don't mean to tell you what you know....) Aug 31 13:15:09 * JustinP tries again with some respect Aug 31 13:15:20 chkhinge uses the sensor to tell which way the screen is rotated Aug 31 13:15:36 the hinge key is what is pressed, just below the hinge, when the cover is closed Aug 31 13:15:51 The apm interface chkhinge uses is sharp junk and doesn't exist under 2.6 and never will. The key event that gets generated also doesn't exist Aug 31 13:16:29 .... Aug 31 13:16:48 The only way to tell how the key board is rotated is to listen for EV_SW events from the kernel input system. Under older 2.6 kernels thie used to be an EV_KEY event. It now has its own event type EV_SW Aug 31 13:16:56 ok Aug 31 13:17:00 so.... Aug 31 13:17:03 argh Aug 31 13:17:14 damn it Aug 31 13:17:34 so is there a way to find out which way the screen is rotated? Aug 31 13:18:04 Yes, but you need to listen to the input system. Aug 31 13:18:08 no Aug 31 13:18:14 I mean, say, on startup Aug 31 13:18:20 RP: he wants to know how to query the current state. its necessary to be able to do so Aug 31 13:18:26 yes Aug 31 13:18:28 thank you kergoth Aug 31 13:18:29 not just listen for changes in state Aug 31 13:18:35 Ah, in that case you ask the input system - there is an ioctl Aug 31 13:19:05 ok Aug 31 13:19:25 so this means we're going to need a special daemon to listen for this stuff.... Aug 31 13:19:39 I'm afraid so :-( Aug 31 13:19:49 what about the CE-RH2 remote? Aug 31 13:19:52 chkhinged Aug 31 13:20:07 CE-RH2 remote? Aug 31 13:20:09 or CE-RH1 for the Cxx0 Aug 31 13:20:17 audio remote Aug 31 13:20:21 I've not written drivers for them... Aug 31 13:20:25 with 2.4 I get keypresses Aug 31 13:20:37 I assume with 2.6 I won't.... Aug 31 13:20:40 If someone wants support they'll have to write a driver Aug 31 13:20:52 I need to go out - back 15 minutes Aug 31 13:22:08 * JustinP is being strangely obtuse today...I don't know why Aug 31 13:27:14 RP: I'll try to take a look at adding a driver for the CE-RH2. I haven't switched to 2.6 yet but I really want to. The suspend/timing issues in 2.4 are really getting on my last nerves Aug 31 13:30:20 03justinp 07org.oe.dev * r657fab8a... 10/packages/ (19 files in 6 dirs): xserver-common, x11-common: Remove x11-common in favor of xserver-common, remove unneeded run-calibrate.sh from xserver-common Aug 31 13:31:30 JustinP: could you also submit those patches upstream? Aug 31 13:31:43 sad Aug 31 13:32:34 CosmicPenguin: you're sad? Aug 31 13:32:44 bitbake wants this file mtd-utils-native-20050801, and none of the sources have it :-( Aug 31 13:33:14 i only updated yesturday to get round it wanting an older version Aug 31 13:38:13 koen: I can try..... Aug 31 13:38:33 koen: that calibration script simply isn't needed...it's an extra file Aug 31 13:39:15 koen: I was experimenting with ways to pass my current Perforce changeset from the fetch.py into my .bb, but without success Aug 31 13:40:27 a perforce fetcher? Aug 31 13:40:49 yeah - we've had it for a long time now Aug 31 13:43:38 * RP returns Aug 31 14:03:05 Can anyone offhand tell me where the touchscreen calibration code used by gpe and opie is? Aug 31 14:06:25 RP: GPE uses xtscal; haven't found it in CVS yet Aug 31 14:08:34 mreimer: it's in apps/ Aug 31 14:09:01 thanks pb_ Aug 31 14:09:03 RP: http://handhelds.org/cgi-bin/cvsweb.cgi/apps/xtscal/ Aug 31 14:09:33 or http://ftp.handhelds.org/pub/projects/gpe/source/xtscal-0.6.3.tar.bz2 if you prefer tarballs Aug 31 14:09:39 RP: for Opie, see http://handhelds.org/cgi-bin/cvsweb.cgi/opie/core/apps/calibrate/ Aug 31 14:10:47 Hmm. It looks like the same code as ts_calibrate.c which works... Aug 31 14:13:51 koen, drw: That code seems to print the display and takes some readings. Is there some function somewhere it gets passed to? Aug 31 14:14:57 I did see the code in qte itself but that also doesn't seem to be the full thing... Aug 31 14:15:26 RP: what exactly are you looking for? Aug 31 14:15:56 RP: I am not familiar with the code myself, but maybe mickey or zecke could help out Aug 31 14:16:00 the opie calibrator collects values and calls out the qte calibration code, which generates the constants for the transform and emits /etc/pointercal Aug 31 14:16:11 tslib's linear module is what uses that to actually perform the transformation Aug 31 14:16:36 or maybe kergoth can help too :) Aug 31 14:17:09 kergoth: Right. Any idea offhand why opie wouldn't calibrate correctly yet ts_calibrate will? Aug 31 14:17:40 RP: opie/qte's calibration, iirc, only actually reads two points from the screen. really, really crappy algo. Aug 31 14:17:44 unless that's changed recently Aug 31 14:17:57 schurig had played with a calibrate2 to make it use the algo ts_calibrate does at one point Aug 31 14:18:32 So they take five points but only pass two of them? Aug 31 14:18:58 yeah, it only takes the others for a rough sanity check Aug 31 14:19:07 makes sure they form a 'box', if it doesnt, it keeps asking for more points Aug 31 14:19:19 which is why tapping repeatedly in the center doesnt pass calibration in opie, but does in ts_calibrate Aug 31 14:19:46 Great. So spitz must have a warped touchscreen :) Aug 31 14:20:30 well, just one that opie's calibrator isnt capable of handling. unless the touchscreen is nonlinear, the touchscreen itself isnt really to blame ;) Aug 31 14:20:51 .......how did I manage to quit the channel? Aug 31 14:21:04 carefully? Aug 31 14:22:18 ...... Aug 31 14:22:19 .... Aug 31 14:22:20 .. Aug 31 14:22:27 my e-image-core isn't working.... Aug 31 14:23:16 kergoth: Do you know what happened to the calibrate2 by any chance? Aug 31 14:23:52 iirc it was sitting out in noncore or unmaintained or something in the cvs tree Aug 31 14:24:39 but iirc, there was a licensing concern Aug 31 14:24:46 check the license of ts_calibrate Aug 31 14:26:05 kergoth: GPL'd by rmk Aug 31 14:26:17 does anybody hav this mtd-snapshot-20050801.tar.bz2 file and coulld possibly send it to me? Aug 31 14:26:52 RP: right, now, iirc, you'd have to stuff that code into qte, and qte is multilicensed. if we're going to improve the calibration algo, we should really do it in such a way that we can actually submit it upstream, not which is infected by third party gpl code Aug 31 14:27:28 dan2003: http://dioxide.randomvoids.com/~papercrane/mtd-snapshot-20050801.tar.bz2 Aug 31 14:27:34 tell me when you have it Aug 31 14:28:06 http://www.embedded.com/story/OEG20020529S0046 has a 3 point touch screen calibration algorithm that might be worth trying Aug 31 14:28:14 http://handhelds.org/cgi-bin/cvsweb.cgi/opie/development/calibrate2/ Aug 31 14:28:20 JustinP, thanks, got it Aug 31 14:28:57 RP: yeah thats it, dunno what he changed or anything though. do a diff. Aug 31 14:29:04 kergoth: Can we extend tslib to handle this? Aug 31 14:29:23 RP: itd be a substantial rework of tslib. we've talked about adding calibration to tslib for some time now Aug 31 14:29:32 would certainly be worth looking into Aug 31 14:29:53 but then we'd need to decide whether a hardcoded one is sufficient, or whether we want to support plugging in any arbitrary calibration routine.. Aug 31 14:29:55 kergoth: Initially, some functions to return the calibration data would be enough Aug 31 14:30:06 the latter would be nice, since i'd like to calibrate pressure as well Aug 31 14:30:11 RP: what do you mean? Aug 31 14:30:32 hand it the raw 5 point coordinates and get it to return the constants? Aug 31 14:30:39 Yes Aug 31 14:30:41 why isn't bitbake re-building all of my e programs.....I removed the stamps and work... Aug 31 14:30:44 argh Aug 31 14:30:49 * JustinP starts from scratch yet again Aug 31 14:30:50 again, licensing problem. we cant just take rmk's algo and stuff it into tslib without infecting tslib itself Aug 31 14:30:55 tslib is lgpl, iirc, not gpl Aug 31 14:31:04 we'd have to talk to him about it Aug 31 14:31:15 I guess we could ask him... Aug 31 14:40:06 arg, can't install an ipkg because of configure errors in another: "update-alternatives: Error: cannot register alternative lsmod to /bin/lsmod since it is already registered to /sbin/lsmod". How can I make ipkg ignore the error? Aug 31 14:40:16 I'm trying to install udev and getting that error Aug 31 14:40:59 edit the ipkg status file Aug 31 14:41:10 change 'unpacked' to 'installed' Aug 31 14:41:15 thanks :-) Aug 31 14:43:05 kergoth: I've sent an email asking him... Aug 31 14:43:33 should probably send an email for comments to linux-arm and linux-omap-open-source Aug 31 14:43:34 mreimer: "ipkg flag installed ", I think. Aug 31 14:43:42 thanks pb Aug 31 14:44:03 something like that, anyway. Aug 31 14:45:09 thanks god for kernel debug messages Aug 31 14:45:32 pxa-i2s Aug 31 14:46:18 so the audio remote is part of the scoop....it looks like.... Aug 31 14:46:23 or controlled by it or something Aug 31 14:47:21 JustinP: No, its on an io pin which goes to the max1111 accessed via the ssp interface Aug 31 14:48:41 RP: ok, then Aug 31 14:49:11 The max1111 is an adc Aug 31 14:49:30 perhaps I shouldn't be screwing around in here... Aug 31 14:49:57 You're quite safe with the max111 Aug 31 14:50:08 routines to use it are already there... Aug 31 14:50:16 You just need to start a new input driver Aug 31 14:50:39 i'm curious (trying to learn), what are you all talking about? Aug 31 14:51:11 Lo-Down: A music remote control connected to the headphone socket of spitz Aug 31 14:51:16 ~spitz Aug 31 14:51:17 hmm... spitz is the Sharp SL-C3000, or a dog Aug 31 14:51:25 ah Aug 31 14:51:32 thanks Aug 31 14:51:35 JustinP: Did you find the 2.4 driver? Aug 31 14:52:24 JustinP: drivers/char/corgi_rc.c Aug 31 14:53:20 RP: no, I found linux/drivers/sound/pxa-i2s_spitz.c Aug 31 14:53:36 JustinP: See the above file - much more interesting ;-) Aug 31 14:53:37 yay! NOTE: package gpe-image-1.0: completed Aug 31 14:53:40 RP: that's the file which has the "hp status" output that I see when I inplug the remote Aug 31 14:53:55 inplug....how did I manage that one? Aug 31 14:54:00 mreimer: excellent Aug 31 14:54:14 JustinP: That's detection of a set of headphones. corgi_rc is the remote control driver Aug 31 14:54:42 oh.... Aug 31 14:54:45 (stupid me...) Aug 31 14:54:45 They are linked, you're right - just start with corig_rc unless you want to go insane Aug 31 14:56:37 JustinP: There is an equivalent function to ssp_get_max1111_val under 2.6 corgi_ssp_max1111_get... Aug 31 14:56:46 They do the same thing Aug 31 14:57:49 ok Aug 31 14:59:55 03koen 07org.oe.dev * rb16c9e25... 10/packages/gpe-contacts/gpe-contacts_0.42.bb: gpe-contacts: add 0.42 Aug 31 14:59:59 03koen 07org.oe.dev * r3f06eda7... 10/packages/libmimedir/libmimedir_cvs.bb: libmimedir_cvs: change default_preference to 1 Aug 31 15:00:03 03koen 07org.oe.dev * rec8bcd92... 10/packages/gpe-calendar/gpe-calendar_0.65.bb: gpe-calendar: add 0.65 Aug 31 15:02:06 03koen 07org.oe.dev * r18d47398... 10/packages/gpe-beam/gpe-beam_0.2.8.bb: gpe-beam: add 0.2.8 Aug 31 15:03:30 Is there a way to get bitbake to fetch from cvs instead of oesources, for a single .bb? Aug 31 15:03:41 ? Aug 31 15:04:04 i have Compaq OHH BootLoader, Rev 2-21-13 on my simpad, and am now lost as to what methid i should use to flash an image, can i simply put the simpad-rootfs.jffs2 on a cf and boot with it inserted? surley i have to tell it i want to flash somehow Aug 31 15:04:47 mreimer: no, but it isn't a question of "instead of". It will always try to fetch from cvs if the tarball is not available in the STASH. Aug 31 15:05:31 CVS is fresher than what's in oesources Aug 31 15:06:12 ok, think i know wot to do now ;) Aug 31 15:06:22 mreimer: ah. in that case, you need to set CVSDATE to a newer date. Aug 31 15:06:30 ok, I'll try that Aug 31 15:06:53 (or CVSDATE_package) Aug 31 15:07:50 do I have to reparse, or can I just "setvar CVSDATE_blueprobe 20050901" and rebuild? didn't seem to work Aug 31 15:08:43 I suspect you have to reparse, but I'm not sure. I'm not really eleet enough to understand the shell thing. Aug 31 15:08:54 it's *nice* Aug 31 15:09:01 especially on a slow box like mine Aug 31 15:09:30 cool Aug 31 15:11:08 I guess I just don't use oe enough any more to have a need for it. Aug 31 15:12:24 heh, you too? Aug 31 15:12:47 03koen 07org.oe.dev * ra7cc3c7d... 10/conf/distro/preferred-gpe-versions-2.7.inc: preferred-gpe-versions-2.7.inc: update gpe-beam, gpe-contacts and gpe-calendar Aug 31 15:12:53 i've used ASSUME_PROVIDED = "bluez-utils-nodbus" in local.con because my machine wont build it, but now ive gotton to the stage where it wants to make the image it compains it cant find bluez-util, how do i prevent it from being included in the image? Aug 31 15:13:03 RP|brb: so should the remote be a keyboard extension or a new device or something like the hinge sensor thing? Aug 31 15:13:35 dan2003: you need to remove it from the DEPENDS and RDEPENDS of whatever you're installing Aug 31 15:13:42 k thanks Aug 31 15:13:43 dan2003: you can only ASSUME_PROVIDED tools needed for building stuff Aug 31 15:13:43 dan2003: you should make a copy of the image file and start hacking Aug 31 15:13:59 dan2003: you really shouldn't use ASSUME_PROVIDED for such things Aug 31 15:15:33 k Aug 31 15:15:57 kergoth: heh, yeah. haven't really touched oe for at least a couple of months. Aug 31 15:16:43 why is controlling a remote control over 600 lines of code? it has 7 buttons! Aug 31 15:17:05 I don't think this is the right way to learn how kernel code works.... Aug 31 15:17:20 RP: could you point me to the hinge sensor code in 2.6? I assume it's much cleaner.... Aug 31 15:22:04 JustinP, lol, that does sound just a little exsesive! Aug 31 15:22:32 JustinP: are you .. voluntarily looking at lineo code? Aug 31 15:22:34 never a good idea Aug 31 15:23:02 i hate debugging cgi scripts Aug 31 15:23:06 error 500 is my worst enemy Aug 31 15:23:13 well I want my CE-RH2 remote working on 2.6 and since I know nothing of the hardware I have to look at it to even think of adding support in 2.6 Aug 31 15:23:25 ps -ef Aug 31 15:23:27 bah Aug 31 15:23:34 of course...since I know nothing of the hardware I'm probably screwed anyway Aug 31 15:24:51 yah, I seem to be spending a lot of time in cgi hell recently. Aug 31 15:24:52 SPITZ_GPIO_AK_INT Aug 31 15:25:16 gotta figure out some way to provide a web frontend to email filtering for our sales guys. Aug 31 15:36:18 so it's written as a kernel thread watching for the remote....it looks like... Aug 31 15:48:44 JustinP: You want to write an additional input driver Aug 31 15:49:17 JustinP: http://www.rpsys.net/openzaurus/patches/spitzkbd-r1.patch is the spitz keyboard code and includes the hinge switches Aug 31 15:51:13 kergoth: Are you the offical tslib maintainer? Aug 31 15:51:24 yep Aug 31 15:51:33 douglas lowder handed it over some time ago Aug 31 15:51:42 RP: yeah, I found the hinge stuff Aug 31 15:51:48 That answers the first of rmk's questions then, I have discussed it with you :) Aug 31 15:51:51 we should tag what we have as 1.0 and create a new branch to add the new api bits Aug 31 15:51:53 hehe Aug 31 15:52:00 His other comment was prod me at the weekend... Aug 31 15:52:04 k Aug 31 15:52:13 i dont want to just throw a half assed api in Aug 31 15:52:20 need to think about a clean implementation Aug 31 15:52:24 RP: I was just seeing a minor # in the 2.4 stuff and was wondering if the remote is accessible in 2.4 as an entry in /dev somehow... Aug 31 15:52:37 kergoth: No, I appreaiate that. Getting a version 1.0 out would be good Aug 31 15:52:45 Then we can look at an api... Aug 31 15:52:54 i kindal ike the idea of having pluggable calibration Aug 31 15:53:19 for someone needing a more perfect calibration, or who has a more flaky ts, could drop in a 10 pointer or whatever Aug 31 15:54:07 Yes. With the right structures, it should be possible to have a nice api to it as well Aug 31 15:54:20 kergoth: right. or, equally, those who can manage with a 3-point calibration. Aug 31 15:54:23 * kergoth nods Aug 31 15:54:37 I keep meaning to implement that in xtscal, but, well, apathy and stuff. Aug 31 15:54:41 hehe Aug 31 15:54:48 i keep meaning to do all sorts of things.. but yeah.. Aug 31 15:55:27 girlfriend from out of town is in town till labor day, after she leaves i'll have some time to hack on things Aug 31 15:55:55 03koen 07org.oe.dev * r71288bf2... 10/packages/blueprobe/blueprobe_0.14.bb: blueprobe: add 0.14 Aug 31 16:06:16 am I crazy or is this code checking the status of the remote 100 times per second? Aug 31 16:06:26 a kernel thread... Aug 31 16:06:35 shouldn't this come through as an interrupt or something? Aug 31 16:06:45 is that lineo code? Aug 31 16:07:00 yes Aug 31 16:08:41 JustinP: Its entirely likely it does crazy things Aug 31 16:09:00 make that almost certain Aug 31 16:09:02 lineo's slogan: 'because we can' Aug 31 16:10:44 RP: so....should an interrupt be tied to some signal the chip sends or something? Aug 31 16:12:32 what is lineo code? Aug 31 16:12:58 LoDown: you don't want to know ;) Aug 31 16:13:08 LoDown: Lineo is the company that Sharp hired to do their kernel work for the Zaurii Aug 31 16:13:10 well...prolly not :-D Aug 31 16:13:13 JustinP: Although it has a device number, I don't think it exposes itself that way. Its just a red herring... Aug 31 16:13:14 ah Aug 31 16:13:27 RP: ah, ok Aug 31 16:13:34 so...they didn't follow anything that can be defined as standard linux kernel programming, i'm guessing? Aug 31 16:14:10 JustinP: There is no interrupt or chip and you probably will have to poll, at a guess at about HZ/8 Aug 31 16:15:06 JustinP: The remote control provides a voltage on the wire. The ADC converts that voltage to a value. By looking at the value, you can tell which key it is Aug 31 16:15:08 RP: So you have to hold the key for ~1/8s to get it to work? Aug 31 16:15:17 RP: yes, I understand that part Aug 31 16:15:19 Yes Aug 31 16:15:37 A normal press should see it held for at least that long Aug 31 16:16:18 ok Aug 31 16:17:01 I guess I'll just try to get some miminal code in there to display something for now Aug 31 16:17:04 hacks Aug 31 16:17:11 since I know nothing about kernel procedures as yet Aug 31 16:17:16 (next to) Aug 31 16:18:13 RP: do you have any idea how the remote turned the unit *on* in the Sharp ROM? Aug 31 16:18:40 RP: to clarify: when I had the unit suspended and the remote plugged in and pressed, say, play, the unit would turn on Aug 31 16:18:45 RP: no such luck in OZ Aug 31 16:19:33 If it doesn't work, you increase the value... Aug 31 16:20:15 I have a fair idea Aug 31 16:20:25 apm hacks? Aug 31 16:22:04 JustinP: Not so much hacks. Presumably they have some wiring that triggers gpio 0 or 1 when you press the button. That will bring the device out of suspend Aug 31 16:22:17 The code looks for the reason it was woken Aug 31 16:22:51 You can see corgi_wakeup_remocon_hook(). That says whether it was the remote control or not Aug 31 16:23:22 ok Aug 31 16:23:39 so....OZ isn't listening for the wakeup? Aug 31 16:23:53 Their kernel should just work Aug 31 16:23:55 RP: never mind, don't waste your time Aug 31 16:24:04 it's not worth digging into this Aug 31 16:24:08 Something must be broken with their apm code... Aug 31 16:24:18 (surprise surprise) Aug 31 16:24:40 FWIW, spitz runs nicely with my same apm code ;-) Aug 31 16:24:48 good good Aug 31 16:24:48 sane Aug 31 16:24:51 no suspend problems? Aug 31 16:25:00 none Aug 31 16:25:08 no "hda: lost interrupt" when resuming? Aug 31 16:25:14 nope Aug 31 16:25:24 no green led stays on when suspending? Aug 31 16:25:31 ;-) Aug 31 16:25:41 No, we don't use the green LED yet ;-) Aug 31 16:25:48 ah Aug 31 16:26:01 so no HD activity Aug 31 16:26:05 But I know why the Sharp rom will do that and how to prevent us from doing it Aug 31 16:26:06 I see Aug 31 16:26:11 ok Aug 31 16:26:12 No hdd activity yet, no Aug 31 16:26:16 minor issue ;-) Aug 31 16:26:34 well it sure makes me nervous when the HD activity LED is *ON* when the unit suspends... Aug 31 16:26:56 I know exactly why their code does that. I suspected it might Aug 31 16:27:05 makes me worried they don't shut down the microdrive before suspending...or shut it down during an operation Aug 31 16:27:11 I've read it in that much detail :-( Aug 31 16:27:17 poor RP Aug 31 16:27:26 you probably need a shrink by not Aug 31 16:27:32 The LED is not connected to the microdrive. The microdrive is powered down Aug 31 16:27:45 The light is a software hack Aug 31 16:27:56 You can make it do anything Aug 31 16:28:59 I was actually physically shaking at one point reading the code. The battery/power code is that bad :-( Aug 31 16:29:31 so....the LED has nothing to do with the hard drive activity? Aug 31 16:29:55 Its driven by the ide driver in software Aug 31 16:30:04 ah Aug 31 16:30:33 look at drivers/ide/ide-disk.c Aug 31 16:31:23 fun Aug 31 16:36:25 I'm afraid of these lines.... SCP2_REG_GPCR |= SCP2_AKIN_PULLUP Aug 31 16:36:42 it *looks* like they're altering registers by changing a global variable... Aug 31 16:36:56 :-D Aug 31 16:37:42 03koen 07org.oe.dev * rdb6937a3... 10/packages/gpe-timesheet/gpe-timesheet_0.18.bb: gpe-timesheet: add 0.18 Aug 31 16:39:35 ughh im unable to compile GLIB for my IPAQ 4705: | gconvert.c:47:2: #error GNU libiconv not in use but included iconv.h is from libiconv Aug 31 16:40:07 and I'm unable to build e-image any more Aug 31 16:40:14 JustinP: They do - Its normal scoop stuff and drivers are already written to handle it... Aug 31 16:40:18 JustinP: what DISTRO? Aug 31 16:40:19 for some reaosn it looks like it always assumes that e-wm and entrance are built.... Aug 31 16:41:04 koen: ? openzaurus-3.5.4 of course Aug 31 16:41:37 I don't understand...it started doing this earlier Aug 31 16:41:47 ehm Aug 31 16:42:00 it wasn't rebuiling *all* of my e apps/libs even though I cleared out the stamps and work Aug 31 16:42:01 JustinTodd: what DISTRO? Aug 31 16:42:14 went right on building the image with outdated ipks Aug 31 16:42:20 and now it's just not building them Aug 31 16:43:33 MACHINE = "ipaq-pxa270", DISTRO = "familiar", TARGET_OS = "linux" Aug 31 16:44:49 don't build unversioned distros Aug 31 16:45:29 ughh, gaim went wonky. did i miss anything? Aug 31 16:45:33 morning All ;) Aug 31 16:45:52 koen: perhaps we shouldn't have those files in distro? Aug 31 16:45:57 koen: distro/base or something? Aug 31 16:47:41 'night all Aug 31 16:51:39 mickey|bbiab: what did you do to efl?? Aug 31 16:51:44 argh Aug 31 16:51:48 this should be building Aug 31 16:52:03 and why the hell isn't bitbake complaining about missing dependencies? Aug 31 16:56:03 * koen just removed a big misfeautre Aug 31 16:56:43 03koen 07org.oe.dev * rf4c52b5a... 10/conf/distro/ (familiar.conf familiar-distro.conf familiar-0.8.3.conf): familiar distro files: make it harder to build unversioned distros and set preferences for iconv Aug 31 17:00:04 koen. should i be specifying a familiar version in my local conf somewhere? Aug 31 17:03:52 JustinTodd: yes Aug 31 17:03:59 familiar-VERSIONNUMBERHERE Aug 31 17:04:43 why the hell isn't EFL building... Aug 31 17:04:56 come on...someone has got to know why an image wouldn't complain about missing depends Aug 31 17:05:15 after i change that, will bitbake be smart enough to realize that some dependancies may have been built with out a familiar version number? Aug 31 17:05:15 until is tries to install them Aug 31 17:05:29 JustinTodd: probably Aug 31 17:05:31 JustinP: does OZ still include the preferred-e-versions? Aug 31 17:05:42 haha. k thx ill give it a try Aug 31 17:05:52 JustinTodd: start from scratch to be sure Aug 31 17:06:58 koen: looks like it Aug 31 17:07:03 koen: why *is* it anyway Aug 31 17:07:13 not that should matter Aug 31 17:07:23 see familiar-0.8.3.conf for why it was there Aug 31 17:07:27 the problem isn't that it's building the wrong version, the problem is that it's not building period Aug 31 17:08:26 koen: I still don't see why that would make a difference. I've been building just fine for a while now Aug 31 17:19:59 how much of a performance increase is observed when using the psyco jit compiler? Aug 31 17:24:19 uh oh Aug 31 17:24:36 somehow during the middle of me flashing the jffs2 image the transfer stopped Aug 31 17:25:01 i have tried redoing it...but it doesn't flash past where it already did...ie, it doesn't get to the verify steps and such Aug 31 17:25:31 and when i try load root again...it doesn't actually transfer anything...it just runs through and makes sure it is all there... Aug 31 17:25:42 is there a way from the boot loader i can blast that stuff out? Aug 31 17:26:00 you're way off topic. Aug 31 17:26:05 this is not a support channel Aug 31 17:26:14 if you're dealing with an ipaq, bootloader questions belong in #handhelds.org Aug 31 17:26:27 ah, mybad Aug 31 17:26:35 tsk tsk...didn't even check the room Aug 31 17:26:40 sorry :( Aug 31 18:28:37 how do I build gpe-image so that it uses a 2.6 kernel? Aug 31 23:32:37 off home l8r Sep 01 00:36:03 im trying to flash my simpad using a 4 gig microdrive, and it keeps saying Sep 01 00:36:03 vfat_find_file_entry: fname='/simpad-rootfs.jffs2' Sep 01 00:36:03 dirname='' Sep 01 00:36:03 basename='SIMPAD-ROOTFS.JFFS2' Sep 01 00:36:03 Could not find file.vfat_file_open: file not found errno=00000016 Sep 01 00:36:04 Could not open source hda1:/simpad-rootfs.jffs2 Sep 01 00:38:11 is this due to the fs proably begin fat32 its larger than 2 gig? Sep 01 00:39:08 or would mkfs use a larger cluster size and still have FAT16, im unsure wot the max fat16 fs size is now! i was under the impression it was 2ish gig Sep 01 00:48:56 morning Sep 01 01:00:14 morning all Sep 01 01:12:31 hi Sep 01 01:14:37 hi hrw|work Sep 01 01:29:35 I do 3.5.4/collie build today - will look how it work Sep 01 01:30:15 and write howto "how to make OZ lite" Sep 01 01:34:47 morning all Sep 01 01:35:43 hi ade|desk Sep 01 01:39:41 hey Sep 01 01:39:53 hi zecke Sep 01 01:41:09 morning Sep 01 01:41:17 morning XorA Sep 01 01:41:34 CoreDump|afk: ping Sep 01 01:44:49 RP: r20 kernel, charging is not working properly :-( It decides the battery is charged even though it is well below 100% Sep 01 01:45:21 RP: currently at 68% and its decided to turn charge LED off Sep 01 01:54:40 XorA: I'd suspect it is charged properly and this is a voltage reading problem... Sep 01 01:55:01 hi zecke Sep 01 01:55:33 *totally exhausted* Sep 01 01:55:57 zecke: You mentioned a problem with opie's calibration a while back - can you remember any of the details? Sep 01 01:57:08 XorA: You'll note that even with a 100% charged battery, it will drop to roughly that value if you leave is suspended for a while... Sep 01 01:57:54 XorA: I suspect the voltage readings are at fault. Someone really needs to look into it in detail - I'm at a loss as to what's wrong :-( Sep 01 02:07:30 mickey|bbiab: ping **** BEGIN LOGGING AT Thu Sep 01 02:16:41 2005 Sep 01 02:16:53 lardman use koen's org.oe.dev Sep 01 02:17:23 ade|desk: can I change servers after the db has been used? Sep 01 02:18:25 lardman: i use 'monotone pull ewi546.ewi.utwente.nl org.openembedded.dev' works for me Sep 01 02:19:17 Well it's connected and not complained yet, thanks Sep 01 02:20:14 good morning all Sep 01 02:20:27 morning koen Sep 01 02:21:42 morning koen Sep 01 02:24:46 Opie's touchscreen calibration is just plain brain dead. b=0, d=0 which is the root of the problem. I'll bet if I reversed the X and Y coordinates in the kernel driver, it would all be fine... Sep 01 02:27:28 hi koen Sep 01 02:30:19 opie 0901 should have autorotate - installing Sep 01 02:38:14 03hrw 07org.oe.dev * r6103a65a... 10/packages/debianutils/ (debianutils_2.13.2.bb debianutils_2.14.3.bb): updated debianutils to 2.14.3 (current 'sid' version) Sep 01 02:47:47 hmm.. looks like I have to reconfigure rotate support in my opie Sep 01 02:48:04 hrw|work: What's wrong with it? Sep 01 02:48:26 no autorotate here Sep 01 02:48:50 hrw|work: mickeyl updated it to the new interface recently. Maybe something got broken? :-/ Sep 01 02:48:55 and never mind which orientation I have during start it always start in landscape Sep 01 02:49:24 hrw|work: Will it rotate manually using the menu option? Sep 01 02:49:42 yes Sep 01 02:50:20 hrw|work: spitz won't even do that... Sep 01 02:54:59 03mickeyl 07org.oe.dev * r6804be27... 10/packages/meta/ (2 files): python: fix bluez typo in task packages - closes #278 Sep 01 02:55:40 mickeyl comes to live Sep 01 02:59:15 Ciao all **** ENDING LOGGING AT Thu Sep 01 02:59:56 2005