**** BEGIN LOGGING AT Sat Nov 19 02:59:57 2011 Nov 19 06:47:53 DocScrutinizer: hey hey :) Nov 19 09:05:19 JaMa: --disable-pam is not implemented correctly in elsa Nov 19 09:05:26 it does pam stuff unconditionally :/ Nov 19 09:30:49 heyho Nov 19 12:12:34 SHR: 03Martin.Jansa 07shr-chroot * r4c94f130de0a 10/ (173 files in 23 dirs): system upgrade Nov 19 12:12:35 SHR: 03Martin.Jansa 07shr-chroot * rcd01c4017242 10/ (837 files in 24 dirs): system upgrade Nov 19 12:28:50 mrmoku: I was reading elsa sources and the PAM support is not really opt-in :/ I've tried to fix it by more #ifdefs but some functionality like autologin is missing without PAM.. Nov 19 12:28:58 mrmoku: I'll poke cedric to fix it :) Nov 19 12:29:39 JaMa|Wrk: elsa is a new login manager for e17? Nov 19 12:29:50 yes Nov 19 12:30:22 with autologin as root it could be used as xserver-nodm-init replacement Nov 19 12:30:34 ah ok Nov 19 12:30:36 and later as normal dm when we get there :) Nov 19 12:30:40 :) Nov 19 12:30:45 we shoul Nov 19 12:30:46 d Nov 19 12:30:59 but it should be faster than xinit :) Nov 19 12:53:49 SHR: 03Martin.Jansa 07shr-makefile * r1a717f86b23d 10/Makefile: Makefile: add OE_CLASSIC_ENABLED option to disable OE-classic updates (shr-unstable, shr-testing) Nov 19 12:56:12 Good morning people! Nov 19 13:00:09 morning :) Nov 19 13:00:52 angelox|away: heyho Nov 19 13:01:11 freesmartphone.org: 03morphis 07cornucopia * r061f77e53a19 10/fsogsmd/src/plugins/modem_samsung/ (callhandler.vala mediators_call.vala): (log message trimmed) Nov 19 13:01:11 freesmartphone.org: fsogsmd: modem_samsung: fix two bugs in call status sync logic Nov 19 13:01:11 freesmartphone.org: * we need to be sure we don't access an array with invalid indexes (which the modem can Nov 19 13:01:11 freesmartphone.org: send if it wants to); so we have to check each call id if it's valid for the range of Nov 19 13:01:11 freesmartphone.org: our internal call array Nov 19 13:01:11 freesmartphone.org: * there are some problems with the GLib.ArrayList -> simple array conversion methods. Nov 19 13:01:11 freesmartphone.org: Somehow it doesn't transfer the content of array items correctly. We're using now a Nov 19 13:01:12 freesmartphone.org: 03morphis 07cornucopia * r15d9782c0e1e 10/fsoaudiod/ (25 files in 3 dirs): Nov 19 13:01:39 hey morphis! hey JaMa|Wrk! Nov 19 13:01:46 * angelox|laptop thinks he waked up too late Nov 19 13:02:08 freesmartphone.org: fsoaudiod: conf: palm_pre: import necessary audio scripts we need for audio routing Nov 19 13:02:08 freesmartphone.org: Signed-off-by: Simon Busch Nov 19 13:02:08 freesmartphone.org: 03morphis 07cornucopia * rc2555f3b61de 10/fsogsmd/src/plugins/modem_samsung/utils.vala: Nov 19 13:02:08 freesmartphone.org: fsogsmd: modem_samsung: only provide provider and network display name when value is not null Nov 19 13:02:08 freesmartphone.org: This fixes a problem with SHR as phonefsod requests a network status update already when Nov 19 13:02:08 freesmartphone.org: SIM card is not in active state. This lets fsogsmd segfault causes it tries to add a null Nov 19 13:02:09 freesmartphone.org: value to the hash map. Nov 19 13:02:09 freesmartphone.org: Signed-off-by: Simon Busch Nov 19 13:12:01 angelox|laptop: yes.. I did 150km in nfs during breakfast and now back at work and I thought I waked up late :) Nov 19 13:12:55 JaMa|Wrk: nfs,the game? :) Nov 19 13:16:47 yes :) Nov 19 13:16:52 shift 2 Nov 19 13:17:42 I was somehow in mood to play some racing game after yesterday's competition who will go to Essen faster :) Nov 19 13:18:02 haha ;) Nov 19 13:18:31 JaMa|Wrk: ok, I started to add #ifdefs too :-P Nov 19 13:19:13 i can play just the NFS MW and NFS Hot Pursuit :P Nov 19 13:19:16 mrmoku: I've started by removing elsa_pam.h without HAVE_PAM to easily track where it was used Nov 19 13:19:38 mrmoku: and it ended almost everywhere in elsa.c and elsa-session shutdown Nov 19 13:20:23 angelox|laptop: I had average speed 170km/h so I hope that nobody will try to beat it on real autobahns in .de :) Nov 19 13:21:46 well autobahns are more straight than the track I was driving on.. so probably easy to beat without traffic Nov 19 13:22:18 JaMa|Wrk: and the velocity limit ? :) Nov 19 13:23:09 JaMa|Wrk: I surrounded all calls to pam with #ifdef HAVE_PAM Nov 19 13:23:39 JaMa|Wrk: http://paste.pocoo.org/show/510162/ Nov 19 13:23:48 it built fine... had no time to test it yet Nov 19 13:23:56 * mrmoku at a clients site doing daywork :/ Nov 19 13:59:53 angelox|laptop: we need to talk about the settings app Nov 19 14:03:25 angelox|laptop: I reworked the Header component so it does not include the back and apply button anymore Nov 19 14:03:32 morphis: ok Nov 19 14:03:35 angelox|laptop: but we need to give the app a more generic structure Nov 19 14:03:40 as you hard coded each page Nov 19 14:03:43 morphis: i had do that but you pushed first.. Nov 19 14:03:51 ah ok Nov 19 14:03:53 sorry for that Nov 19 14:04:02 np :) Nov 19 14:04:15 we need some generel settings components like Nov 19 14:04:23 a list with the config items you already did Nov 19 14:05:19 like ListItem { type: ConfigType.BOOLEAN, name: "Enable flight mode", description: "Activates the flight mode for your device" } Nov 19 14:05:31 angelox|laptop: so you have different types for each config element Nov 19 14:05:46 like BOOLEAN, LIST, TEXT, INT, DOUBLE Nov 19 14:05:55 hmm Nov 19 14:06:20 then the component reponsible for displaying the ListItem takes the values and displays the correct view for it Nov 19 14:06:30 i'm also thinking in do only one List for all settings instead one for each sub-menu Nov 19 14:06:36 no Nov 19 14:06:42 let's group the settings option Nov 19 14:06:50 ah ok Nov 19 14:06:56 we can do and then filter by group Nov 19 14:07:15 I will hack together some example you can use to start with Nov 19 14:08:08 morphis: ok; thanks Nov 19 14:08:21 no problem Nov 19 14:12:35 morphis: what you want is sth like this? : http://cdumez.blogspot.com/2010/11/heterogeneous-list-model-in-qml.html Nov 19 14:13:05 morphis, hi Nov 19 14:13:49 angelox|laptop: yes Nov 19 14:13:51 GNUtoo|Desktop: heyho Nov 19 14:14:45 morphis, did you contact mark brown? Nov 19 14:15:25 GNUtoo|Desktop: no Nov 19 14:15:58 GNUtoo|Desktop: he is still not in IRC and I didn't have time to write him a mail Nov 19 14:16:25 angelox|laptop: http://pastie.org/2888355 Nov 19 14:18:08 angelox|laptop: some like this we need as a model Nov 19 14:18:17 then we have a ConfigList component which uses the model Nov 19 14:18:40 we can stay with several pages for each group for now Nov 19 14:18:44 when will you have time to write a mail? Nov 19 14:18:49 but we should try to make this a geneirc thing too Nov 19 14:18:56 GNUtoo|Desktop: I hope this evening Nov 19 14:19:04 GNUtoo|Desktop: or you have time to do so? Nov 19 14:19:14 GNUtoo|Desktop: call is getting working btw. Nov 19 14:19:32 ok Nov 19 14:19:44 about my problem: Nov 19 14:19:51 your speaker is better now? Nov 19 14:19:56 morphis: hmm... i'm getting the idea Nov 19 14:19:57 no Nov 19 14:20:01 it's still broken Nov 19 14:20:04 hm, thats bet Nov 19 14:20:09 s/bet/bad/ Nov 19 14:20:11 morphis meant: hm, thats bad Nov 19 14:20:11 it only works at low power Nov 19 14:20:20 high power => krrr Nov 19 14:20:34 and the sound is bad in replicant too or just with SHR? Nov 19 14:21:01 both Nov 19 14:21:11 hm Nov 19 14:21:13 but.... Nov 19 14:21:17 ? Nov 19 14:21:23 *the alarm is not that loud Nov 19 14:21:30 so the alarm work Nov 19 14:21:39 angelox|laptop: I would appreciate if you can do the settings app first before the launcher Nov 19 14:21:46 ok Nov 19 14:21:46 *I guess there is another speaker for telephony Nov 19 14:21:48 right? Nov 19 14:22:01 *headphones work under replicant Nov 19 14:22:14 there is a front and a back speaker Nov 19 14:22:19 which one is broken for you? Nov 19 14:22:23 back Nov 19 14:22:28 ok Nov 19 14:22:35 so not a problem when making calls Nov 19 14:22:40 indeed Nov 19 14:22:50 I wonder if I should try to replace it or not Nov 19 14:22:51 I thought the front speaker is broken Nov 19 14:23:04 I never used the front speaker Nov 19 14:23:10 so I cannot tell Nov 19 14:23:16 it looks as you can lift the plastic plate above the speaker Nov 19 14:24:24 I couldn't remove the plastic that is arround the speaker Nov 19 14:24:31 altough I opened the phone etc... Nov 19 14:24:34 removed screw Nov 19 14:24:46 *screws Nov 19 14:24:54 and the electronics etc... Nov 19 14:25:18 hm Nov 19 14:25:39 there is a plastic arround the speaker Nov 19 14:25:48 I coun't remove it Nov 19 14:25:57 altough I could detach the whole Nov 19 14:29:19 you can try Nov 19 14:29:25 but maybe we should ask broonie first Nov 19 14:29:27 I've already tried Nov 19 14:29:30 yes Nov 19 14:30:10 before we do any further steps (means raparing your speaker and getting the new alsa codec driver working) Nov 19 14:30:39 I've already tried to repair the speaker Nov 19 14:31:31 ok Nov 19 14:31:44 but maybe he has some good ideas how we can work around in software Nov 19 14:32:48 morphis: do i use a new branch for settings-rework ? Nov 19 14:32:54 angelox|laptop: yes please Nov 19 14:33:07 angelox|laptop: so I can review and we merge from time to time to master if things are stable Nov 19 14:33:16 morphis: ok Nov 19 14:33:21 thats what I am doing here too with call management and components-cleanup Nov 19 14:35:26 angelox|laptop: it's ok we talk another time about the launcher thing? Nov 19 14:36:35 morphis: sure,but if today,i'll be away in two hours :S Nov 19 14:38:46 ook Nov 19 14:40:41 angelox|laptop: btw. a documentation of the components we use you can find here: http://doc.qt.nokia.com/qtquick-components-symbian-1.1/index.html Nov 19 14:41:16 morphis: thanks! i'll bookmark it Nov 19 14:41:30 angelox|laptop: ok Nov 19 15:00:20 morphis: in your example; these ConfigType's are enumerations,or they can be strings? Nov 19 15:00:34 no, they are enumerations Nov 19 15:00:38 you define them in C++ Nov 19 15:00:42 and bind them to QML Nov 19 15:01:01 ok Nov 19 15:01:13 but we can take them as strings for now if it is easier for you Nov 19 15:03:26 let me try the enums first,if it get harder,i use strings Nov 19 15:04:22 anyway it shouldn't be hard :) Nov 19 15:05:31 does QML actually work at a decent speed without hardware acceleration? Nov 19 15:08:04 i haven't tried with these lot of code on a real device,but some months ago it worked finely and beautifully Nov 19 15:39:38 lindi-: yes Nov 19 15:39:55 morphis: got some example application? Nov 19 15:40:00 try aurora Nov 19 15:40:16 but I only tried on nexus s and palm pre Nov 19 15:40:18 not gta02 Nov 19 15:40:49 is there a release of aurora already or is it in git only? Nov 19 15:41:03 there is release Nov 19 15:41:09 but only for palm pre Nov 19 15:41:29 http://amethyst.openembedded.net/~morphis/aurora/2011-08-beta1/ Nov 19 15:41:48 morphis: ah so you talk about aurora distribution Nov 19 15:41:57 mrmoku: I was looking for an application Nov 19 15:43:07 mrmoku: or does that question even make sense? Nov 19 15:45:30 there is no aurora distribution anymore Nov 19 15:45:36 aurora distro and shr has merged Nov 19 15:45:47 so you can install aurora on any device where SHR runs Nov 19 15:45:52 just build aurora-daemon Nov 19 15:45:57 and then install on the device Nov 19 15:46:10 mrmoku: deamon sounds like non-GUI stuff though? Nov 19 15:46:18 no Nov 19 15:46:21 ok Nov 19 15:46:23 it's the complete UI Nov 19 15:46:30 morphis: got a link to a source tarball? Nov 19 15:46:35 no Nov 19 15:46:36 use git Nov 19 15:46:44 there is no release of the source until now Nov 19 15:47:12 but you can use HEAD as we're trying to keep the master branch stable Nov 19 15:47:22 git://git.freesmartphone.org/aurora.git Nov 19 15:47:24 that one? Nov 19 15:48:53 SHR: 03Martin.Jansa 07shr-makefile * rb6bb72dc7644 10/conf/shr-core/local.conf: local.conf: remove rm_work with explanation why it doesn't work very well with oe-core Nov 19 15:50:02 hmm, unfortunately that seems to build depend on libfso-qt Nov 19 15:51:42 lindi-: you have openembedded setup on your pc? Nov 19 15:51:49 morphis: nope Nov 19 15:51:57 ok Nov 19 15:52:03 morphis: I have this debian thing :) Nov 19 15:52:17 but you're right the libfso-qt dependency is unnecessary today Nov 19 15:52:32 I was hoping to just have some source tarball of a released application that uses QML Nov 19 15:52:41 just to evaluate QML performance Nov 19 15:53:24 you want to try it on which device? Nov 19 15:53:52 freerunner, my laptop and my desktop Nov 19 15:53:58 and over xpra :) Nov 19 15:54:35 morphis: I tested git://gitorious.org/qt-in-use/qt-in-industrial-embedded.git Nov 19 15:54:41 morphis: and that was maybe 2 fps Nov 19 15:54:50 freesmartphone.org: 03morphis 07aurora * r65c06b20a0f3 10/aurora-daemon/ (configure.ac src/bin/Makefile.am): Nov 19 15:54:50 freesmartphone.org: aurora-daemon: drop dependency on libfso-qt Nov 19 15:54:50 freesmartphone.org: Signed-off-by: Simon Busch Nov 19 15:54:50 mrmoku: can you try that on palm? Nov 19 15:55:00 mrmoku: it doesn't depend on anything else than qml Nov 19 15:55:06 I can Nov 19 15:55:06 (/qt4) Nov 19 15:56:02 but not now, later Nov 19 15:56:55 freesmartphone.org: 03morphis 07aurora * ra2c5b3a604c7 10/aurora-daemon/src/bin/cellularservice.h: Nov 19 15:56:55 freesmartphone.org: aurora-daemon: remove unused libfso-qt include directives Nov 19 15:56:55 freesmartphone.org: Signed-off-by: Simon Busch Nov 19 16:00:18 mrmoku: btw, libfso-qt fails to build: http://paste.debian.net/146279/ Nov 19 16:00:21 morphis: in a try with the enums,i get: type: ConfigType.INT ( ListElement: cannot use script for property value) Nov 19 16:00:52 angelox|laptop: how did you bind them to QML? Nov 19 16:01:07 lindi-: i got that too,but i did a "touch ChangeLog" and worked :) Nov 19 16:01:23 morphis: app->rootContext()->setContextProperty("ConfigType", cT); Nov 19 16:01:40 ok Nov 19 16:01:44 there is something more Nov 19 16:01:52 and that is the wrong way Nov 19 16:02:10 btw,cT is: ConfigType cT; Nov 19 16:02:13 hmm Nov 19 16:02:54 yes Nov 19 16:02:56 thats wrong Nov 19 16:03:03 you don't need to bind an instance to QML Nov 19 16:03:05 just the type Nov 19 16:03:09 look at this: Nov 19 16:03:10 http://git.freesmartphone.org/?p=aurora.git;a=blob;f=aurora-daemon/src/components/plugin.cpp;h=6dc343d72fe3f06daf62e5c73ed8d77600667525;hb=HEAD Nov 19 16:03:14 and http://git.freesmartphone.org/?p=aurora.git;a=blob;f=aurora-daemon/src/components/plugin.h;h=b0682367400c9272f577807027c4e1c9162980b2;hb=HEAD Nov 19 16:03:21 there are enums bound to QML Nov 19 16:03:39 you need to do it the same way Nov 19 16:04:56 ah yes,i saw some examples of that Nov 19 16:04:57 thx Nov 19 16:06:44 no problem Nov 19 16:10:05 morphis: what is const char *uri? Nov 19 16:11:07 it's the uri of the plugin you register the component for Nov 19 16:11:14 in this case Aurora.Components Nov 19 16:12:32 freesmartphone.org: 03morphis 07cornucopia * re8668e0fe3c0 10/fsogsmd/src/plugins/modem_samsung/ (4 files): Nov 19 16:12:32 freesmartphone.org: fsogsmd: modem_samsung: last bits to get call handling working Nov 19 16:12:32 freesmartphone.org: Signed-off-by: Simon Busch Nov 19 16:19:53 morphis: did that way,but still getting ListElement: cannot use script for property value (simple try console.log(ConfigType.INT) works finely returning 3 value) Nov 19 16:20:36 hm Nov 19 16:20:56 maybe it's not possible to use enums in this case Nov 19 16:21:20 so using it's integers ? Nov 19 16:21:39 strings Nov 19 16:22:16 ok Nov 19 18:08:49 lindi-: no idea what libfso-qt is? Nov 19 18:09:18 lindi-: ah, you ment morphis probably :) Nov 19 18:36:28 ah yes Nov 19 18:46:26 JaMa|Wrk, can I add 20ae2e1c88c1dbc13d4baa9a609403c8957cf4f5 to meta-smartphone's master? Nov 19 18:46:33 base-files: move PRINC to meta-shr and bump it Nov 19 18:46:59 then I'll bump it again to add a better /etc/fstab to gta02 Nov 19 18:55:05 yes Nov 19 18:57:08 ok thanks Nov 19 18:57:17 I'll do that after eating if you don't do it before Nov 19 18:58:56 hmm food, good idea Nov 19 18:59:22 bitbake food Nov 19 18:59:46 opkg update #feed/food update Nov 19 19:02:16 what xorg driver can we use inside qemuarm? Nov 19 19:04:15 make package/food/compile V=99 Nov 19 19:04:27 * PaulFertser 's compiling OpenWrt as you can guess ;) Nov 19 19:04:57 :) Nov 19 19:05:38 Are you all folks getting gta04? ;) Nov 19 19:09:29 * JaMa|Wrk don't know.. don't want to sacrifice gta02 yet Nov 19 20:03:35 GNUtoo: pushed Nov 19 20:04:12 SHR: 03Martin.Jansa 07meta-smartphone * r86542e48f9ab 10/ (2 files in 2 dirs): base-files: move PRINC to meta-shr Nov 19 20:04:17 SHR: 03Martin.Jansa 07meta-smartphone * rb266f9c72fb1 10/meta-openmoko/recipes-bsp/qi/ (8 files in 2 dirs): qi: upgrade to latest SRCREV Nov 19 20:22:06 JaMa|Wrk: any news from cedric? Nov 19 20:22:25 everytime I check #edevelop he's not there :/ Nov 19 20:33:30 JaMa|Wrk, ok thanks, I'm back, I'll push an /etc/fstab with: Nov 19 20:33:36 /dev/mmcblk0p1 /media/card auto defaults,sync 0 0 Nov 19 20:33:41 is it ok? Nov 19 20:35:21 mrmoku: no news yet :/ Nov 19 20:35:25 GNUtoo: seems so Nov 19 20:35:36 ok Nov 19 20:35:44 JaMa|Wrk: it does not complain... but does not work either :/ Nov 19 20:36:00 starts the xserver... and ends it again Nov 19 20:41:45 SHR: 03GNUtoo 07meta-smartphone * r18b34488b52e 10/ (3 files in 3 dirs): meta-openmoko: base-files: override /etc/fstab with one that automounts the microsd Nov 19 20:45:06 lindi-, hi Nov 19 20:45:31 what's your process to load the kernel when doing kernel hacking? Nov 19 20:45:50 since there is no tftpboot....I wonder what's the most efficient way Nov 19 20:46:05 GNUtoo: kexec is fastest Nov 19 20:46:09 ok Nov 19 20:46:19 GNUtoo: but with systemtap you don't need to boot :) Nov 19 20:46:24 is there some trick using dfu Nov 19 20:46:25 ? Nov 19 20:46:38 GNUtoo: possibly Nov 19 20:47:30 GNUtoo: but just copying new kernel over scp to sd card is not that slow Nov 19 20:47:54 ok Nov 19 20:48:02 I use ubifs anyway Nov 19 20:48:10 ext3 :) Nov 19 20:49:56 lindi-, what's better? kgdb or system tap? Nov 19 20:50:41 GNUtoo: I don't think I have ever used kgdb Nov 19 20:50:48 ok Nov 19 20:50:55 GNUtoo: I've just used normal gdb with openocd with JTAG Nov 19 20:50:57 I've the debug board and I guess the callbacks are there Nov 19 20:51:02 ok Nov 19 20:51:10 afaik kgdb is something different Nov 19 20:51:11 I guess that serial is faster than openocd Nov 19 20:51:36 also it offers more breakpoints Nov 19 20:51:45 with jtag you only have hardware breakpoints, right? Nov 19 20:51:59 GNUtoo: of course you can place software breakpoints too? Nov 19 20:52:00 or did I remember badly Nov 19 20:52:04 they just are not that nice Nov 19 20:52:13 maybe it's was a special case with uboot relocation then Nov 19 20:52:16 GNUtoo: if you have never used systemtap I highly recommend you to try it out Nov 19 20:52:21 ok Nov 19 20:52:24 GNUtoo: first on your PC Nov 19 20:52:29 ah ok Nov 19 20:52:43 I think it's even possible to use it with yocto Nov 19 20:52:50 it should work with just apt-get install, no need t.. no idea about yocto Nov 19 21:03:04 yo Slyon Nov 19 21:03:06 good evening! Nov 19 21:03:24 when do you want to depart for the meeting? Nov 19 21:04:02 mrmoku, hmm when do you want to start? Nov 19 21:04:18 no plans yet Nov 19 21:04:21 I'm flexible :-) Nov 19 21:04:35 i've classes untils 15:00... but they aren't very important Nov 19 21:04:39 as I'm somehow the contact with LH I prefer to not arrive too late Nov 19 21:04:45 sure Nov 19 21:05:04 it is about 6h driving, right? Nov 19 21:05:11 something like that, yeah Nov 19 21:05:39 so we should maybe start around 12 o'clock. Nov 19 21:05:53 sounds ok to me Nov 19 21:06:29 GNUtoo: when will you arrive in Essen? Nov 19 21:07:53 I don't know exactly Nov 19 21:08:09 ok Nov 19 21:08:24 basically I take a plane to dusseldorf Nov 19 21:08:42 and then I will take the train from dusseldorf to essen Nov 19 21:09:12 the plane ticket says that I arrive at dusseldorf at 12h30 Nov 19 21:09:16 AM of course Nov 19 21:09:25 ok Nov 19 21:10:25 then we'll be late to pick you up at Essen station... Nov 19 21:11:08 mrmoku, we can drive earlier. i don't care Nov 19 21:15:23 Slyon: ok Nov 19 21:15:49 well... we still have some time to think about it :-) Nov 19 21:15:55 ouch!!!! Nov 19 21:16:01 now old man needs sleep Nov 19 21:16:09 spam allert on openmoko wiki Nov 19 21:16:32 heh Nov 19 21:16:58 alright Nov 19 21:17:35 the mainpage is full of spam Nov 19 21:17:49 and the website is very slow Nov 19 21:17:54 ouch Nov 19 21:18:13 if you look at recent changes... all spam Nov 19 21:18:22 apart from lindi's upstreaming page Nov 19 21:18:23 :( Nov 19 21:18:33 nah, old man needs sleep anyway now :) Nov 19 21:18:36 gnight all Nov 19 21:18:51 gnight! Nov 19 21:19:05 GNUtoo, i restored the main page Nov 19 21:19:57 good night mrmoku Nov 19 21:21:03 ok thanks Nov 19 21:29:48 SHR: 03lukasmaerdian 07meta-smartphone * r60d5c442675d 10/meta-fso/recipes-freesmartphone/freesmartphone/frameworkd_git.bb: frameworkd: update SRCREV and bump PR Nov 19 21:39:04 lindi-, how to add cmdline arguments to QI? Nov 19 21:39:18 GNUtoo: when booting from nand? Nov 19 21:39:21 yes Nov 19 21:39:25 GNUtoo: not supported Nov 19 21:39:31 GNUtoo: see "man qi" Nov 19 21:39:42 hmmm Nov 19 21:39:43 oh hmm, the man page is not in upstream yet Nov 19 21:40:00 because now I've 179 1 31165952 mmcblk0p1 every time Nov 19 21:40:08 solution in 2 passes: Nov 19 21:40:12 *parts Nov 19 21:40:16 first part: Nov 19 21:40:34 use the shr-2.6.39-nodrm-mci-wip branch Nov 19 21:40:38 second part: Nov 19 21:40:55 glamo_mci.sd_max_clk=1000000 Nov 19 21:41:08 but I need to pass that arg to glamo_mci..... Nov 19 21:41:26 and it need to be possible for the user to define that Nov 19 21:41:46 right now the easiest thing to do it to tell the user to recompile QI Nov 19 21:41:59 with that added to the cmdline arguments in QI Nov 19 21:42:09 recompiling the kenrel is not a good solution since it changes often Nov 19 21:42:57 lindi-, even if I add that to each microsd card? Nov 19 21:43:04 GNUtoo: yes I know this is an annoying design decision in qi Nov 19 21:43:04 for instance /boot/append-om-gta02 Nov 19 21:43:16 GNUtoo: you can workaround this with microsd to some degree yes Nov 19 21:43:19 I've a vfat card unforunately Nov 19 21:43:37 GNUtoo: but I'm planning to offer a better solution, nobody just seems to understand what that should be :) Nov 19 21:44:04 I've no microsoft windows at all but I've fvat for 2 reasons: Nov 19 21:44:09 *compatibility with android Nov 19 21:44:17 *the card expects vfat Nov 19 21:44:52 (I mean it's a 32G card and I don't want to break it with ext) Nov 19 21:45:18 (there is a FTL on theses cards and it expects vfat) Nov 19 21:45:36 FTL = flash translation layer, the thing that makes it appear as block device Nov 19 21:46:28 lindi-, was it supported before? Nov 19 21:46:44 GNUtoo: FAT? no. command line arguments for booting from nand? no Nov 19 21:47:05 GNUtoo: a lot of people run ext* on SD cards. I would not worry about wear Nov 19 21:47:17 GNUtoo: the price of the SD card will go down faster than that :) Nov 19 21:47:50 it's a 32G card and when I bought it it costed near 100E Nov 19 21:47:55 so not cheap at all Nov 19 21:48:11 yeah Nov 19 21:48:15 I only use 8 GB Nov 19 21:48:22 and I've a 16G that is broken for an unknown reason Nov 19 21:48:30 so I'm extra caucious now Nov 19 21:49:39 what about u-boot? Nov 19 21:50:22 it's a lot more flexible Nov 19 21:50:25 uboot supports changing the cmdline arguments Nov 19 21:50:46 unfortunately efforts to get openmoko u-boot patches mainline did not work that well Nov 19 21:50:54 ok Nov 19 21:50:55 ouch Nov 19 21:50:58 what happened? Nov 19 21:53:19 GNUtoo: well it's simply too much work and nobody is interested Nov 19 21:53:33 ok Nov 19 21:53:38 GNUtoo: so I switched to qi for now since that I was able to upload to debian Nov 19 21:54:04 uploading u-boot with openmoko patches would not have worked :) Nov 19 21:55:03 yes I understand Nov 19 21:55:51 GNUtoo: where would you want qi to read commandline options for nand kernel? Nov 19 21:56:12 file in ubifs Nov 19 21:56:44 GNUtoo: that'd require adding ubifs support Nov 19 21:56:48 GNUtoo: I doubt we can do that Nov 19 21:57:58 ok Nov 19 21:58:02 hmmm Nov 19 21:58:21 and adding vfat support is also out of the question I guess Nov 19 21:58:45 I've an idea Nov 19 21:58:48 in-memory Nov 19 21:58:59 that would be not-so-good idea Nov 19 21:59:12 that would work that way Nov 19 21:59:19 s/memory/nand/ Nov 19 21:59:25 the user boots Nov 19 21:59:39 he uses an uboot uttiliy to change cmdline Nov 19 21:59:45 that writes it to a nand partition Nov 19 21:59:52 and he reboots Nov 19 22:00:07 the problem is if it doesn't boot anymore because of a cmdline argument Nov 19 22:00:10 for instance bad root= Nov 19 22:05:23 GNUtoo: qi could read u-boot_env partition in nand Nov 19 22:07:24 yes Nov 19 22:07:27 something like that Nov 19 22:07:44 altough beware with the conflicts with uboot in nor Nov 19 22:07:50 maybe NOR uboot read that too Nov 19 22:09:36 GNUtoo: it doesn't Nov 19 22:09:45 ok nice Nov 19 22:19:30 lindi-, btw what music player do you use? Nov 19 22:19:41 GNUtoo: mpd + ario Nov 19 22:19:51 * GNUtoo will look at ario Nov 19 22:20:58 http://ario-player.sourceforge.net/screenshots.php seem to be aimed at huge screens Nov 19 22:21:25 GNUtoo: yep it's not completely touchscreen friendly Nov 19 22:21:31 haven't found anything better that'd be in debian Nov 19 22:21:38 ok Nov 19 22:21:46 in SHR we have intone and enjoy Nov 19 22:22:04 and hmm, intone is in debian too Nov 19 22:22:09 but I eared someone wrote an mpd client for small screens (320x240) in sdl Nov 19 22:22:11 haven't just tried Nov 19 22:22:35 I'd like to remind you guys that it was conscious design decision in Qi *not* to read any "env" partition... Nov 19 22:22:54 antrik: yeah Nov 19 22:23:04 but then, it's not the only design decision in Qi that I consider questionable... Nov 19 22:23:13 antrik: but we can change the design Nov 19 22:23:18 *heard Nov 19 22:23:34 though frakly, the more you turn Qi into another uboot, the more pointless it gets... Nov 19 22:23:47 antrik: I think we discussed this already :) Nov 19 22:23:59 * GNUtoo tought about turning u-boot in QI Nov 19 22:24:06 but that won't work for lindi- Nov 19 22:24:17 (too much work for upstreaming that) Nov 19 22:24:54 lindi-: yes, and I still don't see how a GTA02-only bootloader with (mostly) distinct codebase is better than an GTA02-only uboot fork Nov 19 22:25:10 antrik: well it is in debian as I said :) Nov 19 22:25:46 antrik: if you can upload u-boot with openmoko patches then I have no problem reconsidering this Nov 19 22:26:06 well, I would have to become a Debian developer first ;-) Nov 19 22:26:59 possibly Nov 19 22:28:11 alternatively just upstream the u-boot patches, will probably take 6 months :) Nov 19 22:29:00 quite frankly, I don't see the problem. if you give the forked uboot a different name, you don't even have to mention that it's a fork. I don't see how this would be harder to get in than Qi... Nov 19 22:50:32 antrik: well the copyright file will mention that it is a fork :) Nov 19 22:53:42 sure... but having forks in Debian is not exactly unheard of :-) Nov 19 22:54:16 not making it too explicit should be enough though to get around pointless arguments I'd guess **** ENDING LOGGING AT Sun Nov 20 02:59:58 2011