**** BEGIN LOGGING AT Sun Nov 13 02:59:56 2011 Nov 13 07:07:06 lindi-: .ubi is right http://wiki.openmoko.org/wiki/Ubifs Nov 13 09:28:00 angelox|zzZZzz: hey, ayt? Nov 13 09:29:23 angelox|zzZZzz: i guess it's time for a test sip call ;) Nov 13 10:40:24 hello Nov 13 10:40:47 I think I was asking about this some time ago, already, but can't remember a thing... Nov 13 10:41:17 I'd like to build everything needed for phone functionality on debian, how to go about it? Nov 13 10:41:24 fpu_: from the sources? Nov 13 10:41:31 yes Nov 13 10:41:53 gsmd, frameworkd, deviced, all that stuff + shr apps Nov 13 10:41:57 fpu_: do you know there's already prepacked FSO from the pkg-fso-maint folks? Nov 13 10:42:39 fpu_: shr apps might be hard to build Nov 13 10:42:57 yes, I installed it even some time ago. But it's some old version and not only it's impossible to get shr apps working due to some e17 bug, it's also too old to support suspend/resume on my device, for example Nov 13 10:43:57 PaulFertser: I don't even know where their source is to be honest Nov 13 10:44:00 fpu_: i can warn you that building everything from the sources is a pain, even if you use distcc. Nov 13 10:44:15 fpu_: i'm using FSO from git myself, so i know what i'm talking about. Nov 13 10:44:30 fpu_: everything FSO is at git.freesmartphone.org Nov 13 10:44:48 fpu_: shr is at git.shr-project.org, e17 is svn.enlightment.org iirc. Nov 13 10:45:00 I'll need e17 sources? Nov 13 10:45:03 fpu_: autoreconf -i takes helluva time on FR. Nov 13 10:45:26 fpu_: if you want bleeding edge e17... Nov 13 10:45:34 fpu_: the solution is to work with getting the new stuff to debian :) Nov 13 10:45:58 there is e17-dev package in debian, I hope that'll be enough Nov 13 10:46:10 fpu_: you might want to practice on amd64 first :) Nov 13 10:46:40 lindi-: I have i386, is that a problem? Nov 13 10:47:06 fpu_: well generally slower? ;) Nov 13 10:47:24 I have 1st gen atom N270 :-) Nov 13 10:47:27 That'll be fun Nov 13 10:47:45 fpu_: an option would be to use qemubuilder or some really fast ARM device as Sheevaplug or something. Nov 13 10:47:59 hmm Nov 13 10:48:19 I actually have debian running right now on HTC HD2, 1ghz snapdragon Nov 13 10:48:21 fpu_: you'll also want an additional swap space on FR if you plan building FSO there. And i really recommend distcc too. Nov 13 10:48:22 fpu_: I'd look at the bugs in the existing packages and try to fix them one at a time :) Nov 13 10:48:45 PaulFertser: sorry but what does FR mean? Nov 13 10:48:49 fpu_: freerunner Nov 13 10:49:08 oh yes :-) well my phone is faster Nov 13 10:49:51 lindi-: the only real bug is shr-apps dependence on some old e17 libs, if I try to symlink new ones in place of them - I get black space instead of UI Nov 13 10:50:06 but also fsodeviced is too old to support everything I want on my hardware Nov 13 10:50:12 so yeah, source seems like the only way Nov 13 10:50:39 PaulFertser: is there that much code? Nov 13 10:50:52 SHR apps are overrated ;) i use opimd-cli plus my Emacs UI ;) Nov 13 10:51:06 fpu_: symlinking sounds wrong too Nov 13 10:51:30 we'll think about GUI later, for now I need fso daemons running Nov 13 10:51:41 fpu_: FSO2 will take you a day to compile anyway, it takes quite some time to figure out the proper sequence of building stuff, and autoreconf + configure takes much time etc etc. Vala is not the fastest compiler either. Nov 13 10:51:47 PaulFertser: where to begin? What to git first? Nov 13 10:52:08 fpu_: http://git.freesmartphone.org/?p=cornucopia.git;a=summary http://git.freesmartphone.org/?p=framework.git;a=summary Nov 13 10:52:41 fpu_: fso-deviced seems to be from 2011-08-05 Nov 13 10:53:33 hmm, when was android suspend/resume support added? Nov 13 10:54:01 PaulFerster: so, "git clone git://git.freesmartphone.org/cornucopia.git" shall be my first command? Nov 13 10:56:36 fpu_: i guess so, but beware, will take your whole sunday. Nov 13 10:57:09 downloading or compiling? Nov 13 10:59:39 fpu_: compiling and figuring out all the configuration etc. Nov 13 11:00:15 ok Nov 13 11:00:18 it's one-time affair anyway Nov 13 11:00:34 I only wish there was some easy way later to package it into .deb, even dirty ones Nov 13 11:17:01 fpu_: why not package it from the start? Nov 13 11:17:13 fpu_: just take the existing packaging for fso-deviced and upgrade to new upstream version Nov 13 11:17:27 fpu_: if it works you can then send a pull request to the maintainer Nov 13 11:18:08 lindi-: do you mean apt-get source, then replace source code by git, then build? Nov 13 11:18:33 fpu_: for example Nov 13 11:19:01 fpu_: but the packaging is already in git, you want to use that Nov 13 11:19:04 fpu_: http://git.debian.org/?p=pkg-fso/fso-deviced.git;a=summary Nov 13 11:19:05 hmm Nov 13 11:25:25 I'm still not sure how it all fits together. What you mean is, "apt-get source" gits that repo? Nov 13 11:26:02 But yes, that would be a lot easier. Just rebuild fsodeviced (to get features I need), and then shr-apps Nov 13 11:26:06 instead of everything from scratch Nov 13 11:26:48 fpu_: "apt-get source" will warn you that the packaging is in git Nov 13 11:27:04 fpu_: so you want to git clone that instead Nov 13 11:28:51 okay. It starts to make some sense Nov 13 11:28:59 I'm going now but promise to bug you later, thanks guys Nov 13 11:33:24 Good Morning Nov 13 11:33:55 PaulFertser: Just waked up,when you called me,i was sleeping deeply :) Nov 13 11:44:47 angelox|laptop: hey hey :) good morning there Nov 13 12:49:37 JaMa|Wrk, hi, now I've the same issue on SHR: Nov 13 12:49:39 | /bin/sh: /home/gnutoo/port4/embedded/oe-core/oetmps/shr-eglibc/sysroots/x86_64-linux/usr/share/build-1/x86_64-linux-libtool: No such file or directory Nov 13 12:49:52 NOTE: package apr-util-native-1.3.12-r1: task do_compile: Failed Nov 13 12:53:52 is it the same as on SHR trac? Nov 13 12:53:59 no idea Nov 13 12:54:01 should be fixed today Nov 13 12:54:03 ok Nov 13 12:54:09 where's the track? Nov 13 12:54:21 http://trac.shr-project.org/trac/ticket/1544 Nov 13 12:54:21 *what'sthe track # Nov 13 12:54:26 ok thanks Nov 13 12:54:42 yes it's the same Nov 13 12:54:54 would be great if you find out why you don't have SHELL in there Nov 13 12:54:54 I'll update then Nov 13 12:55:03 because with shr-chroot it's there Nov 13 12:55:10 at least on shr buildhost and my host Nov 13 12:55:32 and it would be better to say why it's needed to push it upstream to oe-core Nov 13 12:56:03 $ echo $SHELL => /bin/bash Nov 13 12:56:40 hmm I mean in rules.mk Nov 13 12:56:52 which is stored by apr-native Nov 13 12:58:42 I have SHELL: Nov 13 12:58:49 http://www.pastie.org/private/xxdxlpw8vntlcb5ouhuh3w Nov 13 12:59:39 the whole file: http://www.pastie.org/private/qj28yvmrbq04shazz7vw Nov 13 13:01:16 what file is this? Nov 13 13:01:40 /home/gnutoo/port4/embedded/oe-core/oetmps/shr/work/x86_64-linux/apr-util-native-1.3.12-r1/apr-util-1.3.12/build/rules.mk Nov 13 13:01:55 doesn't look like ${STAGING_DATADIR_NATIVE}/apr/apr_rules.mk Nov 13 13:02:02 no it's not Nov 13 13:02:12 GNUtoo: but that's after those sed calls in do_configure, right? Nov 13 13:02:14 I missunderstood you when you asked my rules.mk Nov 13 13:02:20 no Nov 13 13:02:22 that's before Nov 13 13:02:41 ah so even before replacing it with ${STAGING_DATADIR_NATIVE}/apr/apr_rules.mk Nov 13 13:02:54 yes I didn't run a new build yet Nov 13 13:03:20 hmm sorry I'm confused by what you say.. Nov 13 13:03:39 ok Nov 13 13:03:45 basically you told me that: Nov 13 13:03:53 hmm I mean in rules.mk Nov 13 13:04:03 if there is LIBTOOL=$(SHELL) /home/gnutoo/port4/embedded/oe-core/oetmps/shr/sysroots/x86_64-linux/usr/bin/x86_64-linux-libtool Nov 13 13:04:04 so I missunderstood you and went into workdir Nov 13 13:04:08 in /home/gnutoo/port4/embedded/oe-core/oetmps/shr/work/x86_64-linux/apr-util-native-1.3.12-r1/apr-util-1.3.12/build/rules.mk Nov 13 13:04:09 looking for rules.mk Nov 13 13:04:24 then it wouldn't fail with | /bin/sh: /var/lib/jenkins/jobs/shr-core-n900/workspace/shr-core/tmp/sysroots/x86_64-linux/usr/share/build-1/x86_64-linux-libtool: No such file or directory Nov 13 13:04:36 ok let me look Nov 13 13:04:45 JaMa|Wrk: hmm... there is systemctl-native... does that mean hosts systemctl would work in the arm rootfs? Nov 13 13:05:01 GNUtoo: btw there is shr in your paste and shr-eglibc in another Nov 13 13:05:11 http://www.pastie.org/private/rqjypv6h8nlbkodmmyuda is /home/gnutoo/port4/embedded/oe-core/oetmps/shr/sysroots/x86_64-linux/usr/share/apr/apr_rules.mk Nov 13 13:05:13 GNUtoo: better read last commit in meta-smartphone Nov 13 13:05:26 ok Nov 13 13:06:22 weird.. you don't have even $(apr_builddir) so last patch wont work for you imho Nov 13 13:06:26 my error is: Nov 13 13:06:28 | /bin/sh: /home/gnutoo/port4/embedded/oe-core/oetmps/shr-eglibc/sysroots/x86_64-linux/usr/share/build-1/x86_64-linux-libtool: No such file or directory Nov 13 13:07:03 I've that dir: Nov 13 13:07:05 /home/gnutoo/port4/embedded/oe-core/oetmps/shr-eglibc/sysroots/x86_64-linux/usr/share/build-1/ Nov 13 13:07:14 but there is only that inside: Nov 13 13:07:15 apr_rules.mk make_exports.awk make_var_export.awk mkdir.sh Nov 13 13:07:27 no libtool.... Nov 13 13:07:51 that's why that sed is replacing it with STAGING_BINDIR_NATIVE Nov 13 13:07:57 ok Nov 13 13:08:07 can I try your commit then? Nov 13 13:08:19 no read last meta-smartphone commit first Nov 13 13:08:49 and do you have TMPDIR = shr instead of default TMPDIR = tmp ? Nov 13 13:09:01 yes I do Nov 13 13:09:16 tmpdir is /home/gnutoo/port4/embedded/oe-core/oetmps/shr/ Nov 13 13:09:31 I'd like to move it but I guess I can't Nov 13 13:09:46 (to /home/gnutoo/embedded/...) Nov 13 13:09:52 ok that explains first part, but then you have to rebuild at least python apr-native etc.. because sstate does not work in this case (as reported) Nov 13 13:10:06 GNUtoo: no your tmpdir now is /home/gnutoo/port4/embedded/oe-core/oetmps/shr-eglibc Nov 13 13:10:25 ah ouch Nov 13 13:10:32 how that did happen? Nov 13 13:10:49 should I rebuild everything from scratch then? Nov 13 13:10:49 14:04:07 < JaMa|Wrk> GNUtoo: better read last commit in meta-smartphone Nov 13 13:10:51 did you? Nov 13 13:10:57 yes Nov 13 13:11:02 meta-shr: drop TCLIBCAPPEND, add it to local.conf if you don't want to rebuild now Nov 13 13:11:09 11f5fab0c4e382c99e2fb89de5121a93d7031816 right? Nov 13 13:11:11 ahh I start to understand Nov 13 13:11:17 yes Nov 13 13:11:28 TCLIBCAPPEND = "" was in shr.conf before Nov 13 13:11:33 I didn't get the context Nov 13 13:11:44 now I do Nov 13 13:11:51 now it's gone and you need to put it to local.conf if you don't want to rebuild _now_ Nov 13 13:12:02 yes I understand now Nov 13 13:12:09 http://git.shr-project.org/git/?p=shr-makefile.git;a=blobdiff;f=conf/shr-core/local.conf;h=05ad42277644bbb0e4091f1b663333e732e2bcf0;hp=dd1bf9d6b22ab42a1a6d9712d9f61eda031d24a8;hb=add788602ec621598fc4b17fd778a44f1ec0beec;hpb=24c4ca22169abeee6cb5d509bfb7ecc3a38eeb55 Nov 13 13:12:25 mrmoku: no idea :/ Nov 13 13:12:36 * JaMa|Wrk back to work Nov 13 13:12:42 JaMa|Wrk: ok, will try that with next image :) Nov 13 13:15:30 JaMa|Wrk, btw if sstate is deleted is it an issue? Nov 13 13:18:54 and can I share sstate dir between distros? Nov 13 13:20:23 not an issue (it just won't be used to populate sysroot or whatever and will build if needed) Nov 13 13:20:35 -Martinp23- [Global Notice] Hi folks. Over the next 30 mins or so I'll be doing some network upgrades. This will be in two stages. First, there will be a huge netsplit, followed by smaller netsplits. About 2700 users will lose connections as a result. I'll send messages directly to users on servers which will be restarted in a moment (zelazny, kornbluth). Thanks for your patience and have a great sunday. Nov 13 13:20:51 ok thanks a lot Nov 13 13:20:54 and probably the same between distros (wont save anything as it will build different checksum for different distro) Nov 13 13:25:27 JaMa|Wrk: http://packages.qa.debian.org/q/qi/news/20111113T104710Z.html Nov 13 13:31:53 lindi-: nice Nov 13 13:32:02 lindi-: debian is using ubifs by default? Nov 13 13:32:11 or how is it possible to dual boot now? Nov 13 13:32:14 JaMa|Wrk: no, ext3 Nov 13 13:32:16 JaMa|Wrk: on SD Nov 13 13:32:27 JaMa|Wrk: well previously you needed to use SHR's qi to boot SHR Nov 13 13:32:48 ah.. but you can still install SHR jffs2 image Nov 13 13:32:56 JaMa|Wrk: well you can't boot that :) Nov 13 13:33:07 and that worked before this patch Nov 13 13:33:15 ah true Nov 13 13:33:32 bit misleading entry indeed then, sorry Nov 13 13:33:59 ok.. I was just wondering how this patch is making dual booting possible.. as it's just changing what image you can dual boot with Nov 13 13:34:54 maybe we should say that ubi is now default for SHR :) Nov 13 13:35:33 but I'm not using nand and if I have to then I'll probably keep jffs2 to boot it with NOR u-boot :/ Nov 13 13:35:54 argh Nov 13 13:35:55 as skipping 3 uSD partitions with AUX and qi is _hard_ :) Nov 13 13:36:08 JaMa|Wrk: I guess next version of qi is going to support specifying commandline Nov 13 13:36:25 yes that would resolve this Nov 13 13:37:09 there's this uboot-env "partition" :) Nov 13 13:38:36 JaMa|Wrk: anyways, support for two boot options with one partition and initrd are the most important fixes for me personally Nov 13 13:38:44 I can finally repartition my SD :) Nov 13 13:39:27 lindi-: but that's read only by normal u-boot not NOR, isn't it? Nov 13 13:39:40 JaMa|Wrk: true Nov 13 13:39:57 JaMa|Wrk: but I don't think we want to support NOR u-boot for booting? Nov 13 13:40:06 lindi-: I can change env even for NOR u-boot but only in runtime and that's not much faster than removing uSD and booting nand with qi :) Nov 13 13:40:18 JaMa|Wrk: for one, NOR u-boot does not help you during resume Nov 13 13:40:30 JaMa|Wrk: you need to have a working bootloader in NAND Nov 13 13:40:32 lindi-: but that's what I use to boot from NAND jffs2 image Nov 13 13:40:46 14:34:49 < JaMa|Wrk> as skipping 3 uSD partitions with AUX and qi is _hard_ :) Nov 13 13:40:50 ^ because of this Nov 13 13:40:53 JaMa|Wrk: hrm Nov 13 13:41:01 JaMa|Wrk: improving qi usability is also on my list Nov 13 13:41:11 JaMa|Wrk: did I mention the idea to you yet? Nov 13 13:41:20 only about cmdline Nov 13 13:41:33 JaMa|Wrk: the idea is that if you hit AUX then qi stops booting Nov 13 13:41:41 JaMa|Wrk: and then using AUX lets you cycle through boot options Nov 13 13:41:42 but sorry .. I have to work now.. will read it later Nov 13 13:41:48 JaMa|Wrk: bit like u-boot but without LCD Nov 13 13:41:57 no problem Nov 13 13:42:24 will it vibrate or blink AUX to let you know which option/partition is selected? Nov 13 13:43:03 JaMa|Wrk: the idea is that when it is at boot option 1 it will blink once and then be off for 1 second Nov 13 13:43:11 JaMa|Wrk: and when it is at boot option 2 it will blink twice Nov 13 13:43:29 so you can just count the blinks Nov 13 13:45:26 about QI, could we bump it in SHR Nov 13 13:49:01 I guess the issue is that we need to refresh the shr patch who changes the cmdline Nov 13 13:49:46 GNUtoo: that patch has been merged upstream too Nov 13 13:50:01 ok Nov 13 13:50:33 GNUtoo: we need to revert it Nov 13 13:50:45 ah? Nov 13 13:51:02 GNUtoo: and use require qi-ubi.bb from qi.bb which will apply the revert Nov 13 13:51:25 revert what patch exactly? the shr patches? Nov 13 13:51:27 GNUtoo: now it's vice-versa.. qi.bb is "upstream" and qi-ubi applied ubifs patch on top of it Nov 13 13:51:39 freesmartphone.org: 03angelo 07aurora * rbcf89905e7c5 10/aurora-daemon/src/bin/ (application.cpp application.h fbevents.cpp fbevents.h): aurora-daemon: use local fbFakeKey function instead using from a external class Nov 13 13:51:52 qi.bb will apply patch which reverts last patch in upstream == ubifs Nov 13 13:52:05 argh Nov 13 13:52:24 ok lol Nov 13 13:52:25 GNUtoo: or use SRCREV just before this patch and keep it as it's for now Nov 13 13:52:54 ok, the problem is that I've other things to do right now Nov 13 13:53:00 hi morphis Nov 13 13:53:09 morphis, did you see my mail? Nov 13 13:53:31 heyho Nov 13 13:53:37 GNUtoo: yes I saw your mail this morning Nov 13 13:53:39 great work! Nov 13 13:53:47 GNUtoo: I'll do it when I find time Nov 13 13:54:00 JaMa|Wrk, ok nice thanks a lot Nov 13 13:54:18 * JaMa|Wrk hopes that it will be this year :) Nov 13 13:54:41 GNUtoo: you are already any steps further? Nov 13 13:55:29 GNUtoo: btw I got audio working Nov 13 13:55:37 morphis, nice, how did you do it? Nov 13 13:58:19 GNUtoo: I did first "aplay test.wav" and then I set the output path to speaker with amixer Nov 13 13:58:36 as you already said the settings are lost when the alsa connection is closed Nov 13 13:58:40 yes Nov 13 13:58:41 ok Nov 13 13:58:44 so you have to reset it on every playback Nov 13 13:59:15 yes Nov 13 13:59:19 so for any further steps, how much work is left to get the linux-next kernel working? Nov 13 13:59:23 do you want me to fix it? Nov 13 13:59:37 I think it's a lot of work Nov 13 13:59:41 ok Nov 13 14:00:13 I already thought about some way to get audio "working" a little bit faster but I htink somebody people will not like it Nov 13 14:02:26 morphis: hey! Nov 13 14:02:39 angelox|laptop: heyho Nov 13 14:02:59 morphis: did you ping me few nights ago? Nov 13 14:03:26 angelox|laptop: yes Nov 13 14:03:33 angelox|laptop: we need to talk about aurora Nov 13 14:04:15 morphis: in which context exactly? Nov 13 14:06:52 about further development Nov 13 14:07:07 as I changed a lot an for example your settings app needs adjustments to work again Nov 13 14:07:54 ok Nov 13 14:08:40 adjustments in the dbus communication,isn't? Nov 13 14:11:59 angelox|laptop: yes but a lot more Nov 13 14:14:16 morphis: ok; i've saw you did a lot of great work,i'll check what need to be done at that side of application Nov 13 14:14:27 morphis: i have whole day free for talking btw ;) Nov 13 14:14:42 angelox|laptop: ok, I will ping you later as I have some other things to do right now Nov 13 14:15:50 morphis: ok Nov 13 14:16:12 freesmartphone.org: 03angelo 07aurora * r20b2ed792200 10/aurora-daemon/src/bin/Makefile.am: aurora-daemon: do not forget to remove from Makefile fbevents class Nov 13 14:19:07 JaMa|Wrk: what do you think alsa-plugins would fit in best: meta-oe or oe-core? Nov 13 14:21:03 morphis: are all dependencies already in oe-core? Nov 13 14:21:10 I think s Nov 13 14:21:11 o Nov 13 14:21:31 alsa-lib pulseaudio ffmpeg Nov 13 14:21:34 morphis: but even if they are I don't think it's "core" enough, so probably meta-oe would be better Nov 13 14:21:41 ok Nov 13 14:21:47 koen is always complaining that oe-core is too big already Nov 13 14:21:50 we can move it to oe-care later if someody wants Nov 13 14:21:54 ok :) Nov 13 14:26:43 GNUtoo: what do you think, can we fix the alsa problem with pulseaudio for the first time? Nov 13 14:29:49 ??? Nov 13 14:29:53 what alsa problem? Nov 13 14:30:10 you mean the reset of stuff at .close? Nov 13 14:30:45 btw fsogsmd segfaults Nov 13 14:31:34 or loops at least Nov 13 14:31:44 I can try to fix the reset of paths during close if you want Nov 13 14:32:30 and when calling: Nov 13 14:32:31 fsogsmd: symbol lookup error: /usr/lib/cornucopia/modules/fsogsm/modem_samsung.so: undefined symbol: samsung_ipc_call_outgoing_message_setup Nov 13 14:32:55 should I cleansstate and rebuild the samsung lib? Nov 13 14:35:45 morphis, ^^^ Nov 13 14:36:13 GNUtoo: the possible to use dmix Nov 13 14:36:22 ah ok Nov 13 14:36:23 s/possible/possibility/ Nov 13 14:36:23 morphis meant: GNUtoo: the possibility to use dmix Nov 13 14:36:32 hmmm Nov 13 14:36:41 I am think about the following: Nov 13 14:36:42 you mean that I should stop working on mainlining? Nov 13 14:36:46 no Nov 13 14:36:49 ok Nov 13 14:36:56 just to get something working for now Nov 13 14:36:58 it means: Nov 13 14:37:06 - we use pulseaudio to access hw:0 Nov 13 14:37:09 mainlining is long and will take quite some time(won't be ready for december meeting) Nov 13 14:37:16 yes Nov 13 14:37:38 ok Nov 13 14:37:39 - pulseaudio does software mixing and offers a alsachannel normal alsa software can use Nov 13 14:37:46 ah ok Nov 13 14:37:51 could be a possibility Nov 13 14:38:02 thats how Palm/HP is doing it on the palmpre Nov 13 14:38:08 ok Nov 13 14:38:12 it's not the best way but we can try Nov 13 14:38:24 ok Nov 13 14:38:28 and we can use scenario files as pulseaudio will keep the connction to hw:0 open all time Nov 13 14:38:40 also ask the other people what they think about pulse Nov 13 14:38:50 I think some people are rather anti-pulse Nov 13 14:38:56 it will just installed for the nexus s Nov 13 14:39:01 as machine dependency Nov 13 14:39:10 ok Nov 13 14:39:35 I don't love it to have pulse in our images Nov 13 14:39:48 maybe I should try compat wireless again... Nov 13 14:39:51 as it is in normal situations unneeded Nov 13 14:40:01 but we don't have a normal situation Nov 13 14:40:01 indeed Nov 13 14:40:04 GNUtoo: yes please Nov 13 14:41:41 GNUtoo: you already took a look at the dhd driver from Palm/HP? Nov 13 14:41:58 not yet Nov 13 14:42:05 I only looked at palm pre wifi driver Nov 13 14:42:11 (for palm) Nov 13 14:51:41 ok Nov 13 14:55:11 else I looked at other wifi drivers like the htc dream Nov 13 15:04:21 GNUtoo: ok pulseaudio works and multiple connections too Nov 13 15:04:30 ok Nov 13 15:04:30 nice Nov 13 15:04:34 but sounds gets very noisy when playing multiple files Nov 13 15:04:39 ah ok Nov 13 15:04:46 did you fine tune the params? Nov 13 15:04:52 like hw:0 Nov 13 15:04:53 etc... Nov 13 15:05:16 load-module module-alsa-sink device=hw:0 mmap=1 sink_name=pcm_output fragment_size=4096 tsched=0 Nov 13 15:05:16 load-module module-alsa-source device=hw:0 mmap=1 source_name=pcm_input fragment_size=4096 tsched=0 Nov 13 15:05:16 set-default-sink pcm_output Nov 13 15:05:16 set-default-source pcm_input Nov 13 15:05:16 load-module module-native-protocol-unix Nov 13 15:05:21 thats my system.pa Nov 13 15:05:26 ok Nov 13 15:06:38 GNUtoo: http://pastie.org/2856994 Nov 13 15:06:40 morphis, did you look at crespo modem power management? Nov 13 15:06:55 paulk: no, not yet Nov 13 15:07:10 ok Nov 13 15:07:18 ah connection died.... Nov 13 15:07:25 what's that? Nov 13 15:07:34 maybe the reset of the paths Nov 13 15:08:32 no Nov 13 15:08:38 it's because I killed one of the clients Nov 13 15:08:46 ok Nov 13 15:54:22 morphis, the thing to bitbake is pulseaudio? Nov 13 16:11:55 lindi-: btw, why did you change only gta02 in qi.. most of those changes would apply to gta01 too Nov 13 16:12:36 JaMa|Wrk: because he doesn't want to deal with gta01 users of qi of which there's no one known in the wild and so the Debian package doesn't claim the support :) Nov 13 16:14:38 ok :) Nov 13 16:15:21 morphis, still that: Nov 13 16:15:22 [ 1303.852553] mmc2:0001:2: pending IRQ with no handler Nov 13 16:15:31 it's very very long to add wifi Nov 13 16:15:39 since you have to fix compat wireless stuff Nov 13 16:15:47 and once you did it doesn't work Nov 13 16:15:55 because compat wireless is too old maybe Nov 13 16:16:11 with more recent compat wireless it lacks the backporting of a function Nov 13 16:16:16 so I cannot use it Nov 13 16:16:31 hmmm Nov 13 16:16:42 we're jumping back and forward between mainline and old kernel Nov 13 16:16:53 so we should talk about that Nov 13 16:17:03 but you still want to wait for december I guess Nov 13 16:17:23 at the end I will wait for december before doing anything I guess Nov 13 16:31:16 JaMa|Wrk: who is creating /var/run and /var/lock links to volatile? Nov 13 16:35:58 meta/files/fs-perms.txt:${localstatedir}/run link volatile/run Nov 13 16:36:32 meta/recipes-core/initscripts/initscripts-1.0/volatiles:d root root 0755 /var/volatile/run none Nov 13 16:36:38 ok Nov 13 16:36:59 JaMa|Wrk: for systemd /var/run and /var/lock linking to volatile are pointless and harmfull Nov 13 16:37:21 systemd mounts them to tmpfs under /run anyway Nov 13 16:40:05 bbl Nov 13 16:42:39 hello Nov 13 16:45:52 mrmoku: meta-angstrom is using own fs-perms.txt to fix that Nov 13 17:05:34 I must be missing something stupid...where is fso-deviced source in fso git? Nov 13 17:05:58 cornucopia repo Nov 13 17:06:56 you're right, it's in the tree Nov 13 17:07:25 I was looking elsewhere, git noob Nov 13 17:07:28 :-) Nov 13 17:18:24 GNUtoo: some interesting information: HP Pre 3 has a wm8994 build in too Nov 13 17:18:33 GNUtoo: and it's 2.6.32 based Nov 13 17:19:10 ok Nov 13 17:19:15 I will try it's driver Nov 13 17:19:20 ok Nov 13 17:19:27 I cannot work on everything at once Nov 13 17:19:36 no problem Nov 13 17:19:37 mainline and android kernel Nov 13 17:19:39 I will do it Nov 13 17:19:41 ok Nov 13 17:19:49 should be easy Nov 13 17:19:56 also remove -Werror whille you're at it Nov 13 17:20:13 from driver/net/wireless/bcm4329/Makefile Nov 13 17:20:55 ok Nov 13 17:25:52 GNUtoo: why should I remove -Werror? Nov 13 17:25:57 it does not compile with? Nov 13 17:27:44 then I'll do it Nov 13 17:27:50 it compiles only fine in oe Nov 13 17:27:59 outside oe it fails Nov 13 17:28:23 btw is it possible to call with the nexus S? Nov 13 17:28:29 *to make a call Nov 13 17:29:56 you mean with FSO? Nov 13 17:32:16 no Nov 13 17:32:21 it segfaults atm Nov 13 17:33:17 GNUtoo: I removed the -Werror already Nov 13 17:34:23 ok Nov 13 17:34:33 yes I meant with FSO Nov 13 17:34:36 ok Nov 13 17:35:07 for me it doesn't exactly segfault but rather vala binding doesn't match so it doesn't find some symbols Nov 13 17:36:27 yes Nov 13 17:36:32 I saw some seconds ago Nov 13 17:36:36 /usr/sbin/fsogsmd: symbol lookup error: /usr/lib/cornucopia/modules/fsogsm/modem_samsung.so: undefined symbol: samsung_ipc_call_outgoing_message_setup Nov 13 17:36:40 will fix it Nov 13 17:37:01 but let me first try the wm8994 audio codec from pre 3 sources Nov 13 18:02:13 GNUtoo: hm ok, it does not compile Nov 13 18:03:49 ok Nov 13 18:03:54 what's the error? Nov 13 18:04:06 because it was predictable that it didn't compile Nov 13 18:04:14 since the kernel versions are different Nov 13 18:05:33 http://pastie.org/2857720 Nov 13 18:06:42 hmmm no idea Nov 13 18:06:50 maybe you should rather port the mainline one Nov 13 18:06:56 but you'll get even more failures Nov 13 18:07:09 I'll take a break Nov 13 18:07:12 bbl Nov 13 18:07:48 do that Nov 13 18:20:11 GNUtoo: do you remember maybe when android suspend support was added to fso-deviced? Nov 13 18:21:54 GNUtoo: ok, one step further: "fsogsmd: org.freesmartphone.InternalError: Modem told us it can not initialize call with number xxxx" Nov 13 18:24:33 GNUtoo: what exactly is the cause that plug:hw is not working with the wm8994 codec driver from samsung? Nov 13 18:30:57 morphis, are your doing calls with nexus s modem. Nov 13 18:30:58 ? Nov 13 18:31:03 (working on it?) Nov 13 18:31:16 paulk: I already implemented it and fixed a bug some minutes ago Nov 13 18:31:20 ok Nov 13 18:31:25 I will start working on it again tomorrow Nov 13 18:31:36 because I have it working (except sound) on ipc-modemctrl Nov 13 18:31:49 so maybe it can help you Nov 13 18:32:22 modem_exec_call_out Nov 13 18:32:34 and modem_response_call Nov 13 18:32:35 functions Nov 13 18:32:38 are of interest Nov 13 18:38:06 JaMa|Wrk: the debian package also does not offer gta01 qi binaries so it can't be used on gta01 Nov 13 18:38:43 paulk: great I will take a look Nov 13 18:40:07 morphis: what is needed to get launcher working again? Nov 13 18:43:30 angelox|laptop: launcher should work Nov 13 18:43:48 I am currently reworking it to use a dynamically gernated list of applications Nov 13 18:44:09 great Nov 13 18:44:09 fpu, yes look at the commits Nov 13 18:44:22 morphis, no idea about the cause Nov 13 18:44:26 morphis: because,when i tried,i got seg fault :( Nov 13 18:44:36 GNUtoo: as it get's some conversion wrong I think Nov 13 18:44:43 angelox|laptop: segfault? Nov 13 18:44:47 do you have a backtrace? Nov 13 18:45:03 GNUtoo: do you know which kernel parts are needed for plug:hw? Nov 13 18:45:30 morphis: i don't know debugging applications Nov 13 18:45:33 morphis, plughw is like plug->hw Nov 13 18:45:40 plug is in userspace Nov 13 18:45:42 I guess Nov 13 18:45:52 so no kernel access? Nov 13 18:46:16 I don't think so, plug just convert to the right format Nov 13 18:46:34 so it's the same as android is doing? Nov 13 18:46:40 with it's libaudio Nov 13 18:46:44 android does writei Nov 13 18:46:50 with an ioctl Nov 13 18:46:58 basically: Nov 13 18:47:05 *mmap works with aplay Nov 13 18:47:12 *mmap doesn't work in all conditions Nov 13 18:47:20 *writei seem to work fine Nov 13 18:47:37 and alsa uses mmap? Nov 13 18:47:37 maybe restricting to certain format could do the trick Nov 13 18:47:58 "Playing WAVE 'test.wav' : Signed 16 bit Little Endian, Rate 44100 Hz, Stereo" Nov 13 18:47:59 alsa is complex and feature rich so it has many stuff Nov 13 18:48:06 aplay -M does mmap Nov 13 18:50:36 freesmartphone.org: 03angelo 07aurora * rfae9697aaac7 10/aurora-daemon/src/bin/application.cpp: aurora-daemon: fix typo Nov 13 18:50:54 GNUtoo: hm Nov 13 18:51:08 as normal audio playback without plug:hw works Nov 13 18:51:22 we need maybe to adjust the conversion in userspace Nov 13 18:51:30 maybe Nov 13 18:51:39 do you speak .asoundrc? Nov 13 18:51:51 if so try with .asounrc Nov 13 18:51:58 you can restrict the formats etc... Nov 13 18:54:56 I need to find some examples Nov 13 18:55:47 http://www.alsa-project.org/alsa-doc/alsa-lib/pcm_plugins.html Nov 13 18:57:03 JaMa|Wrk: can I push systemd fixes to jama or do you want me to push into my own branch and pull from there? Nov 13 18:58:18 morphis: helps? http://pastebin.com/H1VWiyXD Nov 13 18:58:33 morphis: that is the log i get when running launcher via aurora-daemon :( Nov 13 19:00:28 http://shr-project.org/trac/wiki/Android%20Porting%20Guide Nov 13 19:00:46 so is it fsousaged or fsodeviced responsible for suspend/resume? Nov 13 19:01:25 angelox|laptop: can you try with STATIC_RESOLUTION=1 gdb aurora-daemon Nov 13 19:01:38 fpu, fsousaged Nov 13 19:01:54 me silly, I was trying to compile wrong thing then Nov 13 19:02:16 fpu, what's your target distro? Nov 13 19:02:30 because with autorev you get all the last version if you build shr Nov 13 19:02:58 debian Nov 13 19:03:04 ok Nov 13 19:03:58 ask arround to debian people like lindi- , there are external repository for armv4 that has recent enough fsousaged Nov 13 19:04:11 morphis: whole stuff: http://pastebin.com/PcchXSBu Nov 13 19:04:14 fpu, also note that(and complete the wiki with that info): Nov 13 19:04:22 watch out for frameworkd.conf Nov 13 19:04:46 if frameworkd.conf has ousaged activated, ousaged is used instead of fsousaged Nov 13 19:04:51 which is very bad Nov 13 19:05:02 because ousaged has no android support at all Nov 13 19:05:06 hmm Nov 13 19:05:16 so pastebin /etc/frameworkd.conf Nov 13 19:05:17 is needed fsousaged version newer than 20110805? Nov 13 19:05:32 let me dig up frameworkd.conf I previously used Nov 13 19:05:49 morphis: i did include this in the code: _applicationDirectories << "/opt/fso/usr/share/aurora/system"; Nov 13 19:06:10 which is the month in 0805? Nov 13 19:06:18 08 or 05 Nov 13 19:06:24 08 I think Nov 13 19:06:29 ok Nov 13 19:06:41 angelox|laptop: where do you install aurora to? Nov 13 19:06:42 I'm not sure though :p Nov 13 19:06:46 angelox|laptop: /opt/fso? Nov 13 19:07:00 morphis: yep Nov 13 19:07:07 GNUtoo: in [ousaged] section: "disable = 1" Nov 13 19:07:12 so I guess it's good Nov 13 19:07:14 morphis: there's any problem? Nov 13 19:07:15 angelox|laptop: maybe you need to adjust build configuration with --with-qml-importdir parameter Nov 13 19:07:18 ok Nov 13 19:07:47 GNutoo: you're saying if 08 is month, it should be recent enough? Nov 13 19:08:08 morphis: phone app works finely for example,since qml import stuff is in /usr/lib/qt Nov 13 19:08:21 fpu, yes it should Nov 13 19:08:56 let me find this commit to be sure Nov 13 19:09:48 http://git.freesmartphone.org/?p=cornucopia.git;a=commit;h=1ca7deba83824d29e317b953cdfb320730303300 Nov 13 19:09:51 look at the date Nov 13 19:10:05 Mon, 27 Jun 2011 07:28:49 +0000 (09:28 +0200) Nov 13 19:10:07 ok Nov 13 19:10:20 I think they updated debian packages since my last tries Nov 13 19:10:32 And I thought I must compile it myself ]:-> Nov 13 19:10:54 ok Nov 13 19:11:13 fpu, question: what wifi driver do you use? where can I find the source of your bcm4329? Nov 13 19:12:49 https://gitorious.org/linux-on-wince-htc/linux_on_wince_htc/trees/master/drivers/net/wireless Nov 13 19:12:57 one of these bcm4329s Nov 13 19:13:03 I think 204 Nov 13 19:13:44 ok thanks a lot!!! Nov 13 19:13:46 yes, 204 Nov 13 19:14:00 I know it's 204 but I lost the git address Nov 13 19:14:04 (hdd crash) Nov 13 19:14:16 :) Nov 13 19:15:05 I tried to contact dcordes many times for that Nov 13 19:15:15 but he was not there everytime I tried Nov 13 19:16:27 I think he could help me with some things too, usb host is really a sticking point Nov 13 19:17:44 GNUtoo: I take it that you tested suspend/resume on your nexus one, is there any significant power consumption over android suspend (couple of mA?) Nov 13 19:18:25 why is usb host that important? Nov 13 19:18:30 why not using bluetooth instead? Nov 13 19:18:51 I didn't mesure the power consumation.... Nov 13 19:18:55 GNUtoo: usb storage is big thing to me Nov 13 19:18:58 but we do like android.... Nov 13 19:19:05 so it shouldn't change that much Nov 13 19:19:11 ah ok Nov 13 19:19:45 also, I don't own bluetooth keyboard :-) Nov 13 19:20:18 the good ones seem expensive Nov 13 19:20:26 ok Nov 13 19:30:39 2011-11-13T19:28:31.570312Z [ERROR] DBusServiceDevice <>: Can't find modem for modem_type singleline; corresponding modem plugin loaded? Nov 13 19:32:32 fpu, do you have the singleline plugin? Nov 13 19:32:49 it seems no Nov 13 19:32:56 morphis_, I'll try a last time to make wifi work with bcm4329_204 Nov 13 19:33:12 I'm just wondering why it isn't in .deb package and how to get it Nov 13 19:33:37 but that's secondary Nov 13 19:33:54 http://pastie.org/2858097 Nov 13 19:34:05 how can I test suspend/resume from cli? Nov 13 19:34:13 there's error but maybe it will work Nov 13 19:34:34 no idea Nov 13 19:34:43 pastebin your config too Nov 13 19:36:30 http://pastie.org/2858110 Nov 13 19:36:34 fsousaged.conf Nov 13 19:37:30 hmmm no idea Nov 13 19:37:34 ask arround here Nov 13 19:40:30 I rebooted and now I think it instantly tries to suspend :-) Nov 13 19:40:34 no, sorry Nov 13 19:40:37 that was just lag Nov 13 20:11:15 searching and searching and nothing Nov 13 20:11:32 please someone look at this (fsousaged), I'm going to bed, goodnight all Nov 13 20:11:35 http://pastie.org/2858097 Nov 13 20:11:44 let me look Nov 13 20:11:56 ah I don't know Nov 13 20:12:13 fpu_away, you'll save frustration by porting/improving SHR instead Nov 13 20:12:57 SHR lacks many packages I need, it's just no full "desktop" distro Nov 13 20:13:27 This one is especially frustrating because last time I checked, fousaged tried to suspend all right Nov 13 20:13:41 black screen was due to lack of android support Nov 13 20:13:58 and now it doesn't even try because of some screw-up god knows where Nov 13 20:14:23 ok Nov 13 20:23:32 morphis_, still there? Nov 13 20:32:05 GNUtoo: yes Nov 13 20:32:36 morphis_, I ported 2.6.32 bcm4329 to 35 Nov 13 20:32:46 and it works? Nov 13 20:32:52 last time I tried I failed because I didn't see that WEXT was still there Nov 13 20:33:00 it just changed name Nov 13 20:33:02 ok Nov 13 20:33:08 let me finish it Nov 13 20:33:15 ok Nov 13 20:33:18 mrmoku: for now please push to jama branches and I'll try to not overwrite those changes (as I do rebase & push -f usually) Nov 13 20:33:22 it compiles but It surely lack some stuff Nov 13 20:33:25 I'm checking it all Nov 13 20:33:36 bbl Nov 13 20:34:44 lol I'm comparing the driver(dhd_linux.c) and it did the same workarround than me(mc_count issue) Nov 13 20:34:53 s/workarround/fix/ Nov 13 20:34:54 GNUtoo meant: lol I'm comparing the driver(dhd_linux.c) and it did the same fix than me(mc_count issue) Nov 13 20:50:19 morphis_: i've discovered the launcher problem Nov 13 20:50:54 morphis_: someway, aurora-daemon breaks if not found an info file for each application that will be launched Nov 13 20:53:57 it makes the phone reboot Nov 13 20:54:12 freesmartphone.org: 03angelo 07aurora * r983363bfb39e 10/aurora-daemon/src/system/launcher/ (Makefile.am launcher.info): aurora-daemon: add info file for applications launcher Nov 13 20:54:25 http://www.pastie.org/private/hacmmlk9jngw7rfkg5u1hq Nov 13 20:54:28 I've to look more Nov 13 21:01:15 http://www.pastie.org/private/mg3bhojt8cfzuexjzhr1uw #better Nov 13 21:02:11 http://www.pastie.org/2858515 #with rfkill Nov 13 21:02:27 ah was bluetooth rfkill Nov 13 21:02:28 hmmm Nov 13 21:02:40 anyway the only change was passing the right paths for the firmware Nov 13 21:26:06 freesmartphone.org: 03angelo 07aurora * r9b02e9fcde9a 10/aurora-daemon/ (configure.ac src/applications/app-settings/Makefile.am): aurora-daemon: add settings pages to Makefile Nov 13 21:55:54 freesmartphone.org: 03angelo 07aurora * rc28245a5efcc 10/aurora-daemon/src/applications/app-settings/ (12 files in 5 dirs): aurora-daemon: update aurora components importing in app-settings Nov 13 22:05:43 GNUtoo: great work! does scanning work now? Nov 13 22:07:57 no Nov 13 22:08:00 not yet Nov 13 22:08:15 it doesn't find the card totally yet Nov 13 22:08:23 I'm merging the 2 drivers Nov 13 22:09:25 ah ok Nov 13 22:09:28 I need to go now Nov 13 22:09:29 gn8 Nov 14 00:30:01 freesmartphone.org: 03angelo 07aurora * rc637241f01c8 10/aurora-daemon/src/bin/ (application.cpp virtualkeyboard.cpp): aurora-daemon: first bits to get virtual keyboard working again **** ENDING LOGGING AT Mon Nov 14 02:59:57 2011