**** BEGIN LOGGING AT Fri Aug 27 02:59:57 2010 Aug 27 05:38:37 PaulF: ping Aug 27 05:40:36 general linux question: when I have a process with swapped out mempages (possibly more than avail RAM total) and I do a swapoff, what will happen to such process? Aug 27 05:41:36 kill -9 ? Aug 27 05:42:07 or similar fatal signal? Aug 27 05:58:33 DocScrutinizer51: pong Aug 27 05:58:44 moinmoin Aug 27 05:59:08 [2010-08-27 07:56:02] DocScrutinizer51: most likely, they'll segfault or have future malloc()'s die Aug 27 05:59:10 [2010-08-27 07:56:36] future malloc()'s? the ones in the discarded memepages? :-P Aug 27 06:03:53 DocScrutinizer2: I suspect that oom killer will be called in that case Aug 27 06:04:32 hmm, sounds sensible Aug 27 06:17:27 DocScrutinizer51: but oom killer kills the process chosen on some heueristics, not the one that had its pages swapped out. Aug 27 06:18:35 that's what I thought oom does, yes. But maybe it has an 'API' for IPC to kswapd or whatever Aug 27 06:19:50 but then, why shouldn't kswapd talk to kernel directly, to kill -9 the process when the process' allocated swap pages can't get swapped in Aug 27 06:20:35 or kill -SIGSEGV which probably is more of the right signal for such situation Aug 27 06:21:19 DocScrutinizer51: ok, looks like swapoff process is the worst (badness-wise), so it's always getting killed first. Aug 27 06:21:43 in such case the swapped in pages could go directly to a coredump file, even with filled RAM Aug 27 06:22:06 PaulF: rotfl Aug 27 06:23:04 DocScrutinizer: it kills with SIGKILL Aug 27 06:23:11 PaulF: doesn't match my test results - on maemo a swapoff took minutes but finally succeeded Aug 27 06:23:20 mhm Aug 27 06:23:34 so it's -9 positively :-D Aug 27 06:23:45 DocScrutinizer: and btw some process can be made immune to oom-killer by setting a special variable. Aug 27 06:23:55 (i mean you can make any process immune to that) Aug 27 06:24:02 yep Aug 27 06:24:07 got it Aug 27 06:26:53 DocScrutinizer: when you do a "swapoff" it raises a special swapoff syscall and that sets special PF_SWAPOFF flag for the current process and this guarantees that oom-killer will kill it first. At least that's how i understand. Aug 27 06:28:05 so, from a short glance to http://kerneltrap.org/Linux/Swapoff_Performance, I gather kswapd will try to swap in pages sequentially, by sacrificing buffers and other cruft in RAM, and kills arbitrary processes when ever it runs into the situation the current page to swap in and owned by the process doesn't fit into RAM. By killing the process there will be arbitrary number of RAM pages get freed, and kswapd continues with next swap Aug 27 06:28:07 page until all swap is cleaned Aug 27 06:28:51 PaulF: so I'd expect swapoff itself setting this oom-protection flag :-D Aug 27 06:30:19 otherwise this whole concept would result in a guaranteed fail every time, as swapoff is asking oom to kill swapoff Aug 27 06:30:37 DocScrutinizer: nothing in there about killing Aug 27 06:32:40 swapoff can easily use up all memory, so kill those first. Aug 27 06:32:51 (kernel source comment) Aug 27 06:33:30 that would mean swapoff fails in almost all situations Aug 27 06:34:33 well that might be the desired behaviour Aug 27 06:35:21 actually I'd expect this behaviour for 'normal swapoff invocation' while the way I thought it works should be `swapoff --force` Aug 27 06:35:32 (I know there's no such option` Aug 27 06:35:35 ) Aug 27 06:37:32 I feel tempted to write a miniprog with a malloc(MEGAMUCHMEM), then do a swapoff and se what's happening to that miniprog Aug 27 06:37:59 DocScrutinizer: http://groups.google.com/group/fa.linux.kernel/msg/aa910a454c24c7d2 Aug 27 06:38:40 blaerrgh, non-public Aug 27 06:39:16 ? Aug 27 06:39:28 google login screen Aug 27 06:39:46 DocScrutinizer: i'm surely not logged in Aug 27 06:39:51 Using emacs-w3m here Aug 27 06:40:31 If there's not enough freeable memory for what's out on swap, swapoff Aug 27 06:40:31 cannot very well complete. Either it can hang while other processes Aug 27 06:40:32 get killed by the OOM killer once swapoff has filled memory (as 2.4), Aug 27 06:40:32 until there's enough memory free to take in what's still needed from Aug 27 06:40:32 swap; or the OOM killer can kill it off (as 2.6). I much prefer the Aug 27 06:40:33 2.6 behaviour - unlike many processes, swapoff can safely be restarted. Aug 27 06:40:36 So the admin can then choose what else to kill, or add replacement Aug 27 06:40:38 swap, then try the original swapoff again. Aug 27 06:40:40 forwards to https://www.google.com/accounts/ServiceLogin?service=groups2&passive=1209600&continue=http://groups.google.com/group/fa.linux.kernel/msg/aa910a454c24c7d2&followup=http://groups.google.com/group/fa.linux.kernel/msg/aa910a454c24 here Aug 27 06:41:11 I guess that's something about your personal cookies probably. Aug 27 06:42:58 (cookies) mhm :-S --- (quote) you're the real crack, that's exactly the answer to my question Aug 27 06:47:40 PaulF: news from quim? Aug 27 06:48:57 DocScrutinizer: zero news. Aug 27 06:49:13 DocScrutinizer: i assume other folks haven't heard anything from him/nokia either. Aug 27 06:49:34 yup :-S Aug 27 06:50:31 actually even worse than last time :-(( Aug 27 06:51:58 maybe we should start some bitching thread somewhere Aug 27 06:53:25 I do not think bitching will be of any good. They're already doing us a favor. Aug 27 06:55:07 answering right way on http://talk.maemo.org/showthread.php?p=798828#post798828 is tempting though Aug 27 06:55:40 ""sorry, the whole team is waiting for Nokia moving..."" Aug 27 08:38:30 radekp: hi, what about some testing again? :) Aug 27 09:43:10 anarsoul: gotta take care of children now, i hope i will have some time after lunch - if they will sleep Aug 27 09:46:02 radekp: ok Aug 27 09:47:18 radekp: I have idea to try decrease card clock, maybe that will help to use DMA, however it decreases performance... Aug 27 09:48:02 anarsoul: oki, i'll tell when i have some time Aug 27 09:48:57 radekp: ok, I'll publish patch and provide link ('cos I'm leaving for dinner soon) Aug 27 09:50:03 radekp: here it is: http://downloads.tuxfamily.org/linuxrx1950/patches/random/s3cmci_decrease_clock.patch Aug 27 09:50:14 should be applied on top of 3 previous patches... Aug 27 09:51:15 anarsoul: oki Aug 27 12:06:01 JaMa|GoNe: ok, so i have taken a fresh image and tested xrandr and the same behavior remains... input and output are 90% off... the only chance is in updates, so i am opkg update..upgrade right now... Aug 27 12:06:47 it won't change.. there is IMHO wrong patch in xserver.. nobody fixed it yet Aug 27 12:07:06 do you want me to search for ML reply where I explained it a bit? Aug 27 12:07:33 http://shr-users-discussions.2691941.n2.nabble.com/xserver-xorg-video-glamo-td5121088.html Aug 27 12:08:16 sure, gladly Aug 27 12:08:26 i mean, gladly will read up Aug 27 12:12:57 JaMa|GoNe: oh, grat as the xinput setting found by the user called Rainer are working Aug 27 12:13:08 i had been trying some combinations of this too but had no luck Aug 27 12:14:02 so if we have no more wsod i can finally test Danielle's mokowm ecore input method, where everytime keyboard is called, the screen is rotated and keybord resized to landscape Aug 27 12:15:20 JaMa|GoNe, are we using glib 2.24 in shr-u? I'll need 2.25 for emtooth. (gdbus needs 2.25) Aug 27 12:15:32 vanous: the only problem is that randr-support-1.7.0.patch applies wrong rotation to xinput.. fixing those constants should fix it for all Aug 27 12:15:47 pespin: iirc yes, but rather check Aug 27 12:16:08 JaMa|GoNe, building now an shr-u image for htc-artemis to try it out, and it seems so ;) Aug 27 12:16:10 libglib-2.0-0 - 2.24.1-r0.5 Aug 27 12:16:16 JaMa|GoNe: this would be good, but requires a coder with time :( Aug 27 12:16:16 libdbus-glib-1-2 - 0.86-r2.1.5 Aug 27 12:16:28 JaMa|GoNe, gdbus != libdbus-glib Aug 27 12:16:53 JaMa|GoNe, gdbus is gona replace libdbus-glib Aug 27 12:17:11 and libdbus Aug 27 12:17:14 in #vala they encouraged me to start using gdbus Aug 27 12:17:46 FSO2 will move to gdbus once the remaining missing features are in Aug 27 12:17:55 vanous: but then it can save time in runtime (no manual xinput calls) :) Aug 27 12:18:17 mickey|, hi Aug 27 12:18:25 mickey|, here's my todo list: Aug 27 12:18:25 vanous: if you can found right xinput rotations for xrandr rotations.. then it's just one patch, right url http://git.openembedded.org/cgit.cgi/openembedded/tree/recipes/xorg-xserver/xserver-xorg-1.9.0/randr-support.patch Aug 27 12:18:28 JaMa|GoNe: well, i would love to help, but my non existend knowledge cannot help :( Aug 27 12:18:29 push gps config Aug 27 12:18:35 fsotldt.conf Aug 27 12:18:40 JaMa|GoNe: will look Aug 27 12:18:43 new frameworkd.conf Aug 27 12:18:47 use gpsd? Aug 27 12:18:54 * vanous has had a hell of dayjob work these days Aug 27 12:19:05 and then keyboard and then GUI Aug 27 12:19:09 about keyboard Aug 27 12:19:09 JaMa|GoNe, yes, I'm waiting for a codegen bug to be fixed in vala for gdbus. Then new emtooth-vala will have same features as old emtooth. Aug 27 12:19:30 GNUtoo|laptop: ok, cool Aug 27 12:19:37 mickey|, where should that go: GTK_IM_MODULE=xim Aug 27 12:19:48 profile.d Aug 27 12:19:52 or something like that Aug 27 12:19:55 which is in /etc ? Aug 27 12:19:59 yep Aug 27 12:20:01 ok Aug 27 12:20:04 vanous: here too :/ Aug 27 12:20:06 because there are lots of profiles Aug 27 12:20:06 mickey|, are you planning to implement the bluetoothd call on bluetooth power up in little time from now? Aug 27 12:20:08 JaMa|GoNe: but i am very interested :) Aug 27 12:20:12 there in one in /root too Aug 27 12:20:21 s#/root/#home/root# Aug 27 12:20:32 pespin: yes. i have started to think about it and will implement a first versoin soon Aug 27 12:20:43 vanous: I was keeping this one as easy fix for someone who would like to join development :) Aug 27 12:20:53 hehe Aug 27 12:20:57 OK Aug 27 12:20:58 mickey|, great. tell me if I need to do some additional call to fso or similar in my app ;) Aug 27 12:21:00 let me try Aug 27 12:21:01 vanous: because I don't care about rotation and it's imho easy enough to test/fix Aug 27 12:21:09 i can do dayjob after family is sleeping too :) Aug 27 12:21:43 * vanous never used rotation due to wsod but this input method is great :) Aug 27 12:21:54 uhm why make image is building udev? afaik there's no udev in shr-u now :S Aug 27 12:22:09 mickey|, so I add that to the xim package, but only for SHR? Aug 27 12:22:29 s/xim/gtk-immodule-xim/ Aug 27 12:22:29 GNUtoo|laptop meant: mickey|, so I add that to the gtk-immodule-xim package, but only for SHR? Aug 27 12:23:11 pespin: runtime depends (image content) != build time depends Aug 27 12:23:19 GNUtoo|laptop: no idea, might be worth to ask on the list whether it's interesting for all users of gtk Aug 27 12:24:02 JaMa|GoNe, ok good to know :) Aug 27 12:24:13 basically here's the issue Aug 27 12:24:21 xmodmap dream.xmodmap Aug 27 12:24:27 and you maps some keys Aug 27 12:24:38 like / ? - ' " etc... Aug 27 12:24:45 but that doesn't work in GTK Aug 27 12:24:46 pespin: if some recipe is building optional subpackage with udev support then you need udev built before this recipe, but it won't be installed in image unless you install that subpackage manually Aug 27 12:25:03 JaMa|GoNe: hi. if I want to ask a question to OE which list do I need to write? openembedded-devel or openembedded-users ? Aug 27 12:25:03 opkg install gtk-immodule-xim Aug 27 12:25:14 + GTK_IM_MODULE=xim does the trick Aug 27 12:25:28 but there is already a GTK_IM_MODULE Aug 27 12:25:40 GarthPS: depends on question :) but I guess openembedded-devel Aug 27 12:25:43 the one of matchbox for the software keyboard Aug 27 12:25:45 so I wonder Aug 27 12:26:02 I check for the keyobard machine feature Aug 27 12:26:03 JaMa|GoNe: some help an a recipe that does not build Aug 27 12:26:11 GarthPS: -devel Aug 27 12:26:14 and add that Aug 27 12:26:14 anyway I'll send a patch Aug 27 12:26:16 JaMa|GoNe: thx Aug 27 12:26:18 GarthPS: still automoc? Aug 27 12:26:49 JaMa|GoNe: yeap but I did not go really futher as I miss an understanding of the ecosystem Aug 27 12:27:12 JaMa|GoNe: my mail is nearly ready with all the details Aug 27 12:27:34 GarthPS: check tinderbox if someone actually built that recipe lately Aug 27 12:28:53 JaMa|GoNe: nope just me failling :) Aug 27 12:30:20 JaMa|GoNe: do I need to subscrib to oe-devel list for my mail?(is it preferable?) Aug 27 12:31:45 I always subscribe to ML before writting there.. so probably yes (and unsubscribe later if it has too much traffic for you to handle) Aug 27 12:32:22 it's usually reasonable flow of often usefull info Aug 27 12:32:31 ok thx Aug 27 12:40:35 mickey|, btw how to handle the lack of udev in gpsd? Aug 27 12:40:52 because according to JaMa it's an image feature Aug 27 12:41:18 I think it should be distro which read machine, for SHR Aug 27 12:41:23 for instance: Aug 27 12:41:27 where is the problem in udev not being present for gpsd? Aug 27 12:41:39 DISTRO SHR: if machine has <2.6.32 => udev Aug 27 12:41:43 else no udev Aug 27 12:41:52 udev is a dep of gpsd Aug 27 12:42:02 i wonder why Aug 27 12:42:04 gpsd->gpsd-config->gpsd-udev -> udev Aug 27 12:42:09 or something like that Aug 27 12:42:14 it's because of the dev node Aug 27 12:42:16 /dev/gps Aug 27 12:42:21 or something like that Aug 27 12:42:26 I'll check Aug 27 12:43:51 anyway Aug 27 12:43:58 anytime I install gpsd it installs udev Aug 27 12:44:31 Depends: udev, gpsd-conf Aug 27 12:44:39 in: opkg info gpsd-udev Aug 27 12:46:10 RRECOMMENDS_${PN} = "gpsd-conf gpsd-udev" Aug 27 12:46:11 ok Aug 27 12:46:13 that's the reason Aug 27 12:46:29 does BAD_RECOMMENDATION work now? Aug 27 12:46:35 I don't know Aug 27 12:46:47 I don't even know well what it is, but the name is obvious Aug 27 12:47:00 conf/distro/angstrom-2010.x.conf:# BAD_RECOMMENDATIONS = "avahi-daemon avahi-autoipd" Aug 27 12:47:03 conf/distro/angstrom-2010.x.conf:# Note that BAD_RECOMMENDATIONS is a feature of rootfs_ipk.bbclass, not angstrom Aug 27 12:47:30 ok Aug 27 12:47:50 so if it's a feature of rootfs Aug 27 12:48:04 what does it do exactly? Aug 27 12:48:06 if it's set Aug 27 12:48:19 skips installing RRECOMMENDS Aug 27 12:48:22 hmmm Aug 27 12:48:34 so if you really don't need gpsd-udev in image you can "blacklist" it Aug 27 12:48:36 ok Aug 27 12:48:36 what does something like this does: Aug 27 12:49:09 if (IMAGE_DEV_MANAGER = "udev") BAD_RECOMMANDATION += gpsd-udev ? Aug 27 12:49:17 how could something like that work? Aug 27 12:49:23 using @ python of course Aug 27 12:49:35 here it should be fixed.. http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=2e6dfcb63de90d073a022a01690e850e9d5a66e8 Aug 27 12:49:47 I'll explain better Aug 27 12:49:59 in SHR we have om-gta01,om-gta02,htcdream Aug 27 12:50:07 all don't have udev Aug 27 12:50:10 but palm pre will come Aug 27 12:50:14 and will have udev Aug 27 12:50:23 because it has a 2.6.24 kernel Aug 27 12:50:28 and doesn't have devtmpfs Aug 27 12:50:38 so how to handle that? Aug 27 12:50:47 2 options: Aug 27 12:50:52 in that case udev should never be a direct dependency, but rather $(IMAGE_DEV_MANAGER) Aug 27 12:50:53 you can use machine override for IMAGE_DEV_MANAGER in shr-image.inc Aug 27 12:50:58 the same for BAD_RECOMMANDATION Aug 27 12:51:12 yes but it's an image setting Aug 27 12:51:14 not a distro one Aug 27 12:51:17 or a machine onw Aug 27 12:51:33 image is afail always machine arch Aug 27 12:51:42 so using machine overrid in image recipe won't be problem Aug 27 12:51:56 I know, hmmm how to explain that Aug 27 12:52:04 image should be an image setting Aug 27 12:52:14 I mean you could have an image with udev and one without Aug 27 12:52:32 here we're setting something in a package Aug 27 12:52:40 I'm not sure that it would even work Aug 27 12:52:52 because maybe IMAGE_DEV_MANAGER would only be set for the image Aug 27 12:52:52 what package are you talking about? Aug 27 12:52:59 gpsd-udev Aug 27 12:53:12 or gpsd.inc Aug 27 12:53:20 put it in recipes/images/shr-image.inc Aug 27 12:53:21 gpsd.inc has: Aug 27 12:53:23 RRECOMMENDS_${PN} = "gpsd-conf gpsd-udev" Aug 27 12:53:24 that's whole idea Aug 27 12:53:40 read again this: Aug 27 12:53:40 what if you do: Aug 27 12:53:41 14:49:42 < JaMa|GoNe> you can use machine override for IMAGE_DEV_MANAGER in shr-image.inc Aug 27 12:53:44 14:49:47 < JaMa|GoNe> the same for BAD_RECOMMANDATION Aug 27 12:53:46 I know that Aug 27 12:54:02 I'm just not confortable about using an image setting for a distro package Aug 27 12:54:10 not because of machine thing Aug 27 12:54:19 but because image are for image Aug 27 12:54:23 and distro for distro Aug 27 12:54:36 maybe we should talk about that in #oe? Aug 27 12:54:51 sorry I don't see the problem.. Aug 27 12:55:49 I'll Aug 27 13:13:59 freesmartphone.org: 03sre 07cornucopia * r90cf8ad3b8c0 10/libfsobasics/fsobasics/utilities.vala: (log message trimmed) Aug 27 13:14:00 freesmartphone.org: libfsobasics: ReactorChannel: add sysfs capability Aug 27 13:14:00 freesmartphone.org: * Add rewind option Aug 27 13:14:00 freesmartphone.org: * watch PRI in addition to IN Aug 27 13:14:00 freesmartphone.org: This is needed to watch some sysfs files on the N900, Aug 27 13:15:14 freesmartphone.org: 03sre 07cornucopia * ra561ec741ba9 10/libfsobasics/fsobasics/utilities.vala: Aug 27 13:15:14 freesmartphone.org: avoid code duplication Aug 27 13:15:14 freesmartphone.org: Signed-off-by: Sebastian Reichel Aug 27 13:16:25 JaMa|GoNe: ok, seems i could do this. first will try to build with no modify, so i guess it is xserver-xorg_1.9.0.bb Aug 27 13:17:04 vanous: yes it's there Aug 27 13:17:10 ok Aug 27 13:17:50 my son's day name, perhaps will have to go out soon... will see, still raining here Aug 27 13:21:45 ~seen PaulF Aug 27 13:21:47 paulf is currently on #openmoko (12h 37m 24s) #openmoko-cdevel (12h 37m 24s). Has said a total of 26 messages. Is idling for 6h 28m 20s, last said: 'I do not think bitching will be of any good. They're already doing us a favor.'. Aug 27 13:41:22 JaMa|GoNe: i am having to rebuild from scratch here, as my tmp was quite old, so i can only test later... so without test here is the bit i think would work http://pastebin.com/0wFx1wUj in the changed patch http://pastebin.com/HCSkX5ig of the patch you point to me. certainly this will be a nice exercise of patching and testing... when my machine gets there... Aug 27 13:42:36 vanous: yes this looks better Aug 27 13:43:01 vanous: by that time I'll be probably really gone.. so not sure if there will be someone to apply, but you can send patch to oe-devel Aug 27 13:44:11 JaMa|GoNe: when are you leaving? if my unmodified build doesn't go through... Aug 27 13:44:59 maybe for 2 days today and then I have monday to pack my stuff and leave Aug 27 13:45:58 but I should buy some stuff first and borrow reasonable sized travelling case somewhere Aug 27 13:46:55 ok Aug 27 13:47:35 where is mrmoku, also away on holidays? haven't seen him for some time... Aug 27 13:57:44 vanous: yup iirc for 2 more weeks Aug 27 14:00:51 i see Aug 27 14:03:14 anyone with ubifs can pastebin cat /proc/diskstats please? Aug 27 14:38:40 let me retry, anyone using the nand can pastebin cat /proc/diskstats please? Aug 27 14:40:06 Is there some qi with 2-4-2 glamo timing somewhere? Aug 27 14:41:11 GarthPS: are you after faster bootup? Aug 27 14:41:33 lindi-: nope faster boot but smooth display Aug 27 14:41:42 GarthPS: then just set it at runtime after boot Aug 27 14:41:55 lindi-: how? Aug 27 14:42:24 GarthPS: the new omhacks will include "om screen glamo-bus-timings 2-4-2" command :) Aug 27 14:42:32 and what does it matter with faster bootup? Aug 27 14:42:51 GarthPS: well boot loader will be able to set this earlier Aug 27 14:43:04 ah ok Aug 27 14:43:12 lindi-: what is omhacks? Aug 27 14:44:17 GarthPS: a library and command line tool for controlling openmoko hardware Aug 27 14:45:00 lindi-: and when the new omhacks will be out then ? :) Aug 27 14:45:24 GarthPS: http://manpages.ubuntu.com/manpages/maverick/man1/om.1.html Aug 27 14:45:27 GarthPS: today Aug 27 14:47:34 lindi-: in shr-u feed or where? Aug 27 14:48:00 GarthPS: I think SHR includes it yes Aug 27 14:48:58 then we should ask JaMa|GoNe to bump it before he leaves Aug 27 14:49:44 I'd appreciate if you could test it before release Aug 27 14:49:53 just to make sure there are no silly bugs :) Aug 27 14:50:41 lindi-: this sounds cool, not having to alter the bootloader and still enjoy faster glamo Aug 27 14:51:34 vanous: yes I did it to get 25 fps 320x240 theora playback Aug 27 14:52:11 hum sweet ! with mplayer ? Aug 27 14:52:20 freesmartphone.org: 03mickey 07specs * r30e64db296e1 10/ (3 files in 3 dirs): org.freesmartphone.Resource: add GetDependencies() Aug 27 14:52:39 GarthPS: yes it required a bit of bug reporting Aug 27 14:52:45 freesmartphone.org: 03mickey 07specs * r58cd2e8db79f 10/configure.ac: bump Aug 27 14:53:05 lindi-: like what ? Aug 27 14:53:45 GarthPS: http://bugzilla.mplayerhq.hu/show_bug.cgi?id=1774 Aug 27 14:54:22 and a couple of others Aug 27 14:54:27 lindi-: :) ok thx. sure Aug 27 14:54:33 freesmartphone.org: 03mickey 07libfso-glib * r7bfd41d4c49e 10/configure.ac: bump Aug 27 14:55:11 * GarthPS being curious to compare mplayer vs vlc on the FR with glamo timings Aug 27 14:55:25 vlc 1.1 Aug 27 14:55:54 GarthPS: patches to libtheora are the main thing that are not in any distro yet Aug 27 14:56:05 everything else should be in debian at least soon Aug 27 14:56:35 lindi-, theora-dev is a desert Aug 27 14:56:41 xrmx: yep Aug 27 14:56:46 xrmx: #theora has talk Aug 27 15:00:02 mrmoku|away: ping Aug 27 15:00:34 mrmoku|away: how long takes mail from home to your holiday location? :-D Aug 27 15:00:41 freesmartphone.org: 03mickey 07cornucopia * rcaeb69fa995d 10/libfsoresource/fsoresource/dbusresource.vala: libfsoresource: catch up with extended org.freesmartphone.Usage.Resource API Aug 27 15:03:32 * GarthPS it has been a while without I have been running debian on my FR . oh no Qtmoko is a debian :) Aug 27 15:22:36 JaMa|GoNe: when i call bb with -b , it seems dependencies are ignored? Aug 27 15:23:02 yes they are ignored Aug 27 15:23:09 i see, this explains Aug 27 15:25:41 hmmm Aug 27 15:25:46 where does frameworkd.conf comes from? Aug 27 15:25:59 the frameworkd git? Aug 27 15:26:00 oe? Aug 27 15:26:11 a shr repo? Aug 27 15:28:28 mickey|zzZZzz, hi Aug 27 15:31:28 GNUtoo|laptop: frameworkd-config-shr is from shr repo Aug 27 15:35:07 JaMa|GoNe: where is it defined that xserver-xorg_1.9.0.bb is the xserver we want when xserver-xorg_git.bb requires PV = "1.8.99.906" Aug 27 15:35:42 ok Aug 27 15:36:08 JaMa|GoNe, who should I send a patch to? I'ld like to disable ogpsd for htcdream Aug 27 15:38:47 vanous: highest PV with highest DEFAULT_PREFERRENCE is used Aug 27 15:39:08 vanous: both xserver-xorg_1.9.0.bb and xserver-xorg_git.bb have D_P == 1 for shr, so 1.9.0 is used because of newer Aug 27 15:39:18 GNUtoo|laptop: shr-devel Aug 27 15:39:22 ok,thanks Aug 27 15:39:52 JaMa|GoNe, ok thanks Aug 27 15:44:45 GNUtoo|laptop: hi Aug 27 15:45:04 GNUtoo|laptop: lindi- was absolutely right. he sent me a patch about gpsd support a while ago Aug 27 15:45:05 hi Aug 27 15:45:19 indeed Aug 27 15:45:20 GNUtoo|laptop: I won't be at home until sunday at night Aug 27 15:45:24 ok Aug 27 15:45:29 but the patch has some issue Aug 27 15:45:38 it doesn't work anymore Aug 27 15:45:41 GNUtoo|laptop: so I won't be able to look at it and discuss with it what changes may be needed Aug 27 15:45:42 because of gpsd api change Aug 27 15:46:15 GNUtoo|laptop: I just wanted to avoid you spend time on this to face an almost ready patch 2 days later... Aug 27 15:46:54 onen|openBmap, ok thanks a lot Aug 27 15:47:02 I'll spend time on the following changes instead: Aug 27 15:47:08 switch htcdream to gpsd Aug 27 15:47:09 GNUtoo|laptop: ttyl then. have a nich weekend Aug 27 15:47:14 fix keyboard handling Aug 27 15:47:18 fix GUI Aug 27 15:47:27 thanks a lot Aug 27 15:53:25 so is the fso gps handling being changed and gpsd will be the future again? Aug 27 15:53:41 i mean protocol wise, not resource management wise Aug 27 15:57:35 resource will be the same Aug 27 15:57:44 but the gypsy protocol will go away Aug 27 15:57:53 for something better Aug 27 15:57:55 morphis, hi Aug 27 15:58:06 hey GNUtoo|laptop Aug 27 15:59:07 GNUtoo|laptop: how's the GPS work going? Aug 27 15:59:31 mickeyl autotoolized the resource allocator Aug 27 15:59:43 then because ogpsd is deprecated Aug 27 15:59:43 I won't use it Aug 27 15:59:49 I will go the gpsd way Aug 27 15:59:56 ? Aug 27 16:00:02 I will use plain gpsd Aug 27 16:00:06 never really looked into it Aug 27 16:00:10 instead of ogpsd Aug 27 16:00:21 that works with fsotdld Aug 27 16:00:21 anyway ogpsd is from frameworkd Aug 27 16:00:26 no Aug 27 16:00:31 fsotdld is not finished Aug 27 16:00:35 it has no GPGGA parsing Aug 27 16:00:40 only time parsing Aug 27 16:00:41 GPGGA? Aug 27 16:00:50 GPGGA = position fix Aug 27 16:00:53 ah ok Aug 27 16:01:05 and what about implementing it in fsotdld? Aug 27 16:01:15 maybe the better way Aug 27 16:01:22 GNUtoo|laptop: does gpsd support clients over dbus? Aug 27 16:01:29 no Aug 27 16:01:33 only tcp...but Aug 27 16:01:36 there is a newer api Aug 27 16:01:39 that could be interesting Aug 27 16:02:23 can gpsd be used with fsotdld? Aug 27 16:02:30 no Aug 27 16:02:32 oh, so what about signalling - position/course change? Aug 27 16:02:58 morphis, I mean plain gpsd and fsotdld conflicts Aug 27 16:03:07 but maybe fsotdld can speak gpsd Aug 27 16:03:09 I don't know Aug 27 16:03:11 hm Aug 27 16:03:18 taltked with mickey about that? Aug 27 16:04:25 http://www.freedesktop.org/wiki/Software/GeoClue Aug 27 16:04:28 indeed Aug 27 16:04:34 he told me to go gpsd for now Aug 27 16:04:45 until he finishes fsotdld Aug 27 16:04:53 it will take a month minimum Aug 27 16:04:56 to finish fsotdld Aug 27 16:05:07 geoclue could be a good api Aug 27 16:05:16 I didn't understand if mickeyl will use it or not Aug 27 16:05:23 but It makes sense to use it Aug 27 16:05:32