**** BEGIN LOGGING AT Thu Apr 28 02:59:58 2011 Apr 28 06:19:35 Good morning. Apr 28 06:20:26 moin Apr 28 06:37:10 I'm following the SHR Wiki to make changes to SHR. I've editted files in ~/shr-unstable/shr-apps/libframeworkd-phonegui-efl. What do I need to 'bitbake' to incorporate these changes? Apr 28 06:37:55 ChristW: wow... there is still libframeworkd-phonegui-efl in the wiki? really? Apr 28 06:38:42 No, that one is not in the wiki. But it's the one I'm trying to change... Not the correct one? Apr 28 06:39:02 libframeworkd-phonegui-efl is no more... since years :) Apr 28 06:39:21 I'm trying to add a button to the 'dialer' screen... Apr 28 06:39:23 ChristW: what part do you want to change? Apr 28 06:39:24 ahh, ok Apr 28 06:39:37 then you need to change libphone-ui-shr Apr 28 06:40:29 ChristW: http://git.shr-project.org/git/?p=libphone-ui-shr.git;a=blob;f=src/view/dialer-view.c;h=afddb66aebb3eb48eebb78df0b94b12cb96699cf;hb=d88017138fd69ba492dbd6e7872d814ee495d49c Apr 28 06:40:39 Ah, ok, but that one is not retrieved when I 'git clone http://shr.bearstech.com/repo/shr.git shr-apps'... Apr 28 06:40:58 So, is the Wiki out of date, there, as well? Apr 28 06:41:02 yup Apr 28 06:41:05 :/ Apr 28 06:41:09 http://shr-project.org/trac/wiki/Getting%20started%20developing%20SHR Apr 28 06:41:13 * ChristW Grumbles... Apr 28 06:41:19 * mrmoku looks Apr 28 06:54:59 ChristW: ok, changed it to be more correct Apr 28 06:55:06 not very extensive but at least correct :) Apr 28 06:55:49 I'll check, thanks for updating! Apr 28 07:31:16 heyho Apr 28 07:34:34 mickeyl: I solved the ssh problem ... Apr 28 07:34:56 root password was empty but we have to allow the login with an empty password ... Apr 28 07:47:48 Blech. 'Parsing recipies' takes ages when you make a stupid error... Apr 28 07:47:56 Is there a way to get it to build faster? Apr 28 07:48:44 (at least, I know it's compiling _my_ version!) Apr 28 07:52:43 ChristW: there is Apr 28 07:53:01 BB_NUMBER_THREADS = "2" Apr 28 07:53:12 set this in your local.conf Apr 28 07:53:31 ChristW: when you only want to compile one recipe Apr 28 07:53:41 us 'bitbake -b s/us/use/ Apr 28 07:53:46 morphis meant: use 'bitbake -b I recall seeing 'bitbake -i' somewhere in the Wiki, for 'interactive' mode. But, AFAIK, that fails.... Apr 28 07:55:25 I never used the interactive mode Apr 28 07:56:21 It isn't listed in the help either. Apr 28 07:57:52 then it seems that someone removed it from bitbake Apr 28 08:04:49 Too bad :-( Apr 28 08:04:54 * ChristW starts praying... Apr 28 08:05:05 About to install my self-built .ipk... Apr 28 08:08:31 Do I need to restart a service after uploading libphone-ui-shr before I see changes? Apr 28 08:08:50 I'm starting the dialer, but I don't see any of the changes I thought i made. Apr 28 08:09:09 ChristW: you have to restart phoneuid Apr 28 08:10:17 Thanks, but I still don't see anything... Let me do that again with a more 'obvious' change. Apr 28 08:11:55 ChristW: another possible error cause is that it did not install it... I often oversee opkg telling me that it does not install it because the version did not change Apr 28 08:12:05 you have to install it with -force-reinstall in that case Apr 28 08:17:47 I caught that one, it's listed in the Wiki. Apr 28 08:20:52 mickeyl: ping Apr 28 08:37:58 Hum. I rebooted and _still_ don't see any changes (hey, you gotta try something!). That's weird... Apr 28 08:39:13 ChristW: maybe add a logger ... to be sure your changes are taken Apr 28 08:44:47 Well, I changed the 'Click to open contactlist' to 'xxxClick to open contactlist' and I don't see that... Apr 28 08:45:31 Is there a way to 'unpack' the .ipk and look if I see that string in the program? Apr 28 08:58:15 There are 4 .ipk files generated, a libphone-ui-shr, -dbg, -dev and -static. I _presume_ I need to copy the extensoin-less .ipk to my FR> Apr 28 09:01:02 Ok, I now see my 'xxx' change. Good. Now I need to figure out why I don't see my other changes ;-) Apr 28 09:02:40 :) Apr 28 09:02:57 efl/elm is a beast :P Apr 28 09:03:09 When I add a button, do I need to specify the size/layout somewhere? Apr 28 09:06:20 Ah, in data/dialer.edc... Apr 28 09:10:14 morning Apr 28 09:11:21 Morning. Apr 28 09:12:11 mickey|office: good morning Apr 28 09:15:02 mickey|office: I am currently building aurora-fb-image Apr 28 09:16:14 ah, nice! Apr 28 09:16:17 <[Rui]> http://www.amazon.de/gp/product/B003KZ4JCC/ref=oss_product Apr 28 09:16:20 <[Rui]> good price? Apr 28 09:16:53 <[Rui]> 7 neu ab EUR 163,00 4 gebraucht ab EUR 129,00 Apr 28 09:17:03 <[Rui]> I assume gebraucht is used :) Apr 28 09:21:28 it is Apr 28 09:21:30 hmm Apr 28 09:21:43 for a plus it's ok i guess Apr 28 09:21:56 [Rui]: it's ok as you pay a lot more when you go to o2 and buy a new one Apr 28 09:22:21 but the picture shows a Pre and not a Pre Plus Apr 28 09:22:38 the Pre Plus don't have the button on the bottom Apr 28 09:23:32 <[Rui]> weird Apr 28 09:24:03 mickey|office: what were your steps to try qt embedded with the touchscreen? Apr 28 09:26:20 export QT_MOUSE_PROTO=LinuxInput: Apr 28 09:26:57 cd /usr/bin/qtopia/demos/declarative/calculator; qmlviewer -qws -fullscreen -frameless calculator.qml Apr 28 09:27:04 (from memory...) Apr 28 09:27:12 no Apr 28 09:27:15 it's QWS_MOUSE_PROTO Apr 28 09:27:38 ok Apr 28 09:28:28 will try when the image is ready Apr 28 09:34:23 <[Rui]> morphis: @jwildeboer from Red Hat bought one just this week. Apr 28 09:34:57 <[Rui]> so I guess it is really a palm pre plus Apr 28 09:36:19 mickey|office: btw. whats qws and qwt? Apr 28 09:36:50 QWS, so the history says, is the Qt-embedded Windowing System Apr 28 09:37:02 Qwt is a widget library (Qt Scientific Widgets IIRC) Apr 28 09:37:08 ah ok Apr 28 09:37:15 so two things we don't need Apr 28 09:37:22 heh Apr 28 09:37:28 QWS is embedded in Qt/Embedded Apr 28 09:37:34 without qws no framebuffer Apr 28 09:37:39 ah ok Apr 28 09:37:51 so it's the graphic engine behind qt for the framebuffer Apr 28 09:37:52 ok Apr 28 09:39:49 Hurray! I have added a button to my 'dialer' screen and it shows up! Apr 28 09:40:15 Now I 'only' need to program its functionality ;-) Apr 28 09:41:19 mickey|office: and whats with lighthouse, wasn't it about making the graphics rendering in qt exchangable? Isn't there a lighthouse framebuffer plugin? Apr 28 09:43:04 lighthouse? first time i hear that Apr 28 09:44:02 ah Apr 28 09:44:05 http://qt.gitorious.org/qt/lighthouse Apr 28 09:44:09 Conceptually it is “Qt for Embedded Linux without QWS”. Apr 28 09:44:14 yeah Apr 28 09:44:26 or http://developer.qt.nokia.com/wiki/Getting_Started_With_Lighthouse Apr 28 09:45:00 that's interesting for QNX, WinCE, iOS, etc. Apr 28 09:45:24 the android port of Qt is made with Lighthouse Apr 28 09:45:32 yeah, or that Apr 28 09:45:37 http://labs.qt.nokia.com/2011/02/28/necessitas/ Apr 28 09:46:11 fun Apr 28 09:46:32 i'm not very fond of cross-platform development for mobile devices... Apr 28 09:46:36 but fun anyways Apr 28 09:47:13 :) Apr 28 09:47:16 http://qt.gitorious.org/+qt-developers/qt/lighthouse/trees/master/src/plugins/platforms/directfb Apr 28 09:47:19 http://qt.gitorious.org/+qt-developers/qt/lighthouse/trees/master/src/plugins/platforms/fb_base Apr 28 09:47:29 there are plugin for directfb and basic framebuffer Apr 28 09:47:56 at it's called linux_fb Apr 28 09:48:19 will try it when the build is ready Apr 28 09:48:29 yeah, still no too interesting i guess, since that way we introduce an additional abstraction in the path Apr 28 09:48:51 and since we already have the qws, i don't know what should be good for on framebuffer Apr 28 09:49:05 unless they're ripping it out and only support lighthouse at some point Apr 28 09:49:45 jepp Apr 28 09:53:36 there's one interesting thing Apr 28 09:54:00 it may have better support for concurrent processes Apr 28 09:54:03 "This means that Lighthouse is now part of the code base that is going to be released as Qt 4.8. Just to be clear: this does not necessarily mean that Lighthouse will be an officially supported platform for 4.8; that is a separate decision not made by the Lighthouse project team." Apr 28 09:54:13 *nod Apr 28 09:54:14 * Apr 28 09:56:54 will try it as it sounds interesting and a much simpler approach than qws Apr 28 09:57:24 k Apr 28 09:58:13 I saw even a simple windowing system made with lighthouse Apr 28 09:58:55 http://qt.gitorious.org/qt-labs/devdays-windowsystem-server Apr 28 10:01:11 hmm, impressive, but i'm not sure whether it's really interesting Apr 28 10:01:20 if you introduce windowing, you might as well use X11 Apr 28 10:04:11 Food! Apr 28 10:04:39 mickey|office: touchscreen works for me Apr 28 10:04:58 mickey|office: yes, I found it only interesting :) Apr 28 10:05:00 hmm, ok Apr 28 10:05:16 touchscreen is not great but it works with the calculator Apr 28 10:05:27 ok, it was way off for me Apr 28 10:05:37 will try again next time Apr 28 10:05:59 accuracy is not good Apr 28 10:06:03 but it's usable Apr 28 10:06:29 ok Apr 28 10:06:34 when you try again use the image here: http://amethyst.openembedded.net/~morphis/aurora/images/palmpre/ Apr 28 10:06:56 hmm, ok, updating a whole image is always cumbersome for me Apr 28 10:07:04 since the fso-installer does not like my device Apr 28 10:07:18 so i have to do it manually Apr 28 10:08:12 then again, i can build a kernel that boots from NFS Apr 28 10:08:22 that's going to make it faster to experiment with new images Apr 28 10:09:18 ah, one thing that came to my mind yesterday... Apr 28 10:09:25 just a crazy idea... Apr 28 10:09:37 would you mind reintroducing python for a while? Apr 28 10:09:54 in the image? Apr 28 10:09:55 (just to get the basic QML stuff going asap) Apr 28 10:09:58 hehe Apr 28 10:09:59 no Apr 28 10:10:03 or rather yes Apr 28 10:10:06 but also in our platform Apr 28 10:10:23 NFS should be possible Apr 28 10:10:31 I can build a kernel for you Apr 28 10:10:37 that supports nfs Apr 28 10:10:53 that'd be cool Apr 28 10:10:55 ye,s please Apr 28 10:11:03 hm, default kernel already supports nfs Apr 28 10:11:17 the important things are Apr 28 10:11:23 but what about python, for what do you want it to use? Apr 28 10:11:23 kernel level ip autoconfiguration Apr 28 10:11:25 and Apr 28 10:11:31 enable-rootfs-on-nfs Apr 28 10:11:36 which are often not ticked Apr 28 10:11:48 http://wiki.freesmartphone.org/index.php/Palm_Pre:NFS-Root Apr 28 10:12:16 Heinervdm already tried NFS with the Palm Pre Apr 28 10:12:31 and thats the way people are booting android on their pre devices ... Apr 28 10:12:40 (python) i was wondering whether we should write the 1st pass of the business logic (i.e. what's not in QML and not in JS) in Python to get us to a state where people can do more experiments w/ QML while they have a working baseline. Apr 28 10:13:30 hm Apr 28 10:13:36 just to unblock us while we're working on the ValaQt bridge Apr 28 10:14:00 but then we need to call python logic from C++ Apr 28 10:14:10 thankfully that's already done Apr 28 10:14:12 as we have to write the qml components in C++ Apr 28 10:14:14 ah ok Apr 28 10:14:22 pyqt supports QML as well Apr 28 10:14:22 then why not? Apr 28 10:15:02 *nod* that way we could e.g. do the statusbar in a couple of hours Apr 28 10:15:10 and lateron exchange that with the compiled code Apr 28 10:15:17 while keeping the QML and everything else Apr 28 10:15:54 (it would also allow us to get a bit more quickly up to speed wrt. the possibilities we have when connecting QML and the business logic) Apr 28 10:16:26 whats with pyside? Apr 28 10:16:38 it seems to be the latest way to do qt stuff with python Apr 28 10:18:14 hmm Apr 28 10:18:22 i have to take a look. politically i don't like pyside Apr 28 10:18:34 as i have a warm place in my heart for riverbankcomputing.co.uk Apr 28 10:18:45 but i will try to set this aside and evaluate both side-by-side Apr 28 10:20:26 pyside came to me as a big unnecessary reinvention of the wheel, when PyQt was up and running flawlessly for years Apr 28 10:22:43 ok, as I read it pyside is generated automatically Apr 28 10:22:50 and pyqt is handwritten, right? Apr 28 10:23:23 pyside is officially supported by qt Apr 28 10:24:13 you can't handwrite bindings for such a huge library as qt, Phil Thompson invented sip for binding Qt, which is a generator as well, but one with Qt in mind that allows manual additions, much like a vapi Apr 28 10:24:26 ok Apr 28 10:24:40 and sip is used for pyqt? Apr 28 10:24:43 yes Apr 28 10:25:13 i don't know anything about performance comparisons or binding completeness. it's been 5 years since my last working with pyqt Apr 28 10:26:03 ok Apr 28 10:29:09 http://www.pyside.org/docs/pyside/PySide/QtDeclarative/QDeclarativeExtensionPlugin.html Apr 28 10:29:18 we can write the qml plugins even in python Apr 28 10:29:47 and corresponding http://www.riverbankcomputing.co.uk/static/Docs/PyQt4/html/qml.html Apr 28 10:29:49 :) Apr 28 10:30:33 http://www.riverbankcomputing.co.uk/static/Docs/PyQt4/html/qdeclarativeextensionplugin.html Apr 28 10:31:12 right. i guess the bottom line is we can do both with both Apr 28 10:31:23 so i have to take a couple of hours evaluating them Apr 28 10:31:37 ok Apr 28 10:31:43 do we have OE support for both? Apr 28 10:31:45 pyqt has a crazy buildsystem which always took me ages Apr 28 10:31:58 if pyside is simpler to tame, then it would be a benefit Apr 28 10:32:12 pyqt is in oe, i did that many many years ago Apr 28 10:32:17 but i'm not sure how uptodate it is Apr 28 10:33:24 ok Apr 28 10:34:02 it's 4.4.3, latest version is 4.8.x Apr 28 10:34:02 ls Apr 28 10:34:10 hmm, that's ancient Apr 28 10:34:22 but pyside seems to be absent compeltely Apr 28 10:34:48 jepp Apr 28 10:34:50 look at http://pysnippet.blogspot.com/2011/03/pyside-and-qml.html Apr 28 10:34:59 QtDeclarative.qmlRegisterType(Message, "utils", 1, 0, "Message") Apr 28 10:35:02 thats what we need Apr 28 10:35:32 yeah, sweet Apr 28 10:39:19 mickey|office: maybe thats interesting for our qt-vala-bridge: http://qt.gitorious.org/pyside/apiextractor Apr 28 10:54:55 http://developer.qt.nokia.com/wiki/Differences_Between_PySide_and_PyQt is somewhat helpful Apr 28 10:56:29 there seems to be no recipes for pyside Apr 28 11:17:47 ok, if we want to use lighthouse with qt we have to switch to _git version Apr 28 11:18:17 mickey|office: and python-pyqt failed to build Apr 28 11:24:34 morphis: can you still build on amethyst? bitbake errors out for me on every command Apr 28 11:26:30 hmm, some invalid recipe is confusing it Apr 28 11:57:53 mickey|office: I am building in /local/pkg/oe/aurora Apr 28 12:10:13 mickey|office: will you try to get python-pyqt running again in OE? Apr 28 12:21:29 hi mrmoku Apr 28 12:26:42 mrmoku, I'll add charging into fsodeviced.conf Apr 28 12:29:59 http://pastie.org/private/ixcrg88nrjxfpqa6vzj63q Apr 28 12:30:13 mickeyl, is that ok? Apr 28 12:30:20 we talked about it many times Apr 28 12:30:23 and it was never done Apr 28 12:38:15 commit message is: Apr 28 12:38:22 fsodeviced: add battery charging to the nokia900 fsodeviced.conf Apr 28 12:39:17 hi leviathan Apr 28 12:45:25 hi GNUtoo Apr 28 12:45:45 GNUtoo: (charging) yeah, very good... I always forgot :/ Apr 28 12:45:52 so I push? Apr 28 12:45:52 mickey|office: ping Apr 28 12:45:55 GNUtoo: yup Apr 28 12:46:22 the thing I don't know is if we need the standard kernel26 powersupply Apr 28 12:47:45 freesmartphone.org: 03GNUtoo 07cornucopia * rc283e66dc940 10/fsodeviced/conf/nokia_n900/fsodeviced.conf: Apr 28 12:47:46 freesmartphone.org: fsodeviced: add battery charging to the nokia900 fsodeviced.conf Apr 28 12:47:46 freesmartphone.org: Signed-off-by: Denis 'GNUtoo' Carikli Apr 28 12:48:05 mrmoku: pong Apr 28 12:48:28 GNUtoo: cool, thanks. Apr 28 12:48:54 now that it's pushed I've to ask DocScrutinizer a question: what happens if you run 2 charge.sh.txt at the same time? Apr 28 12:48:59 kernel26_powersupply is somewhat helpful, there is a device coming up, but it doesn't contain charging information. I'd hope that GNUtoo could merge the code with his other battery stuff Apr 28 12:49:18 about that Apr 28 12:49:22 I wonder what to do Apr 28 12:49:25 GNUtoo: not recommended, but "works" Apr 28 12:49:36 DocScrutinizer, so no damages.... Apr 28 12:49:46 no way to damage device this way Apr 28 12:49:53 mickey|office, if the kernel changes too often we might do it in userspace Apr 28 12:50:03 else we can do it in kernel space Apr 28 12:50:11 I'd never have published it if it were so dangerous Apr 28 12:50:16 ok Apr 28 12:50:25 since you already have the code, i'd say just merge it with the existing power supply Apr 28 12:50:38 mickeyl, the code for what? Apr 28 12:50:47 sq2700 or how it's called Apr 28 12:50:51 bqq27200 Apr 28 12:50:56 yep Apr 28 12:50:58 kernel driver then Apr 28 12:51:11 I guess eventually one will stop/break due to the other occupyng the I2C bus+device Apr 28 12:51:22 mickey|office: where do we want to end up the cmtspeech component? Apr 28 12:51:27 ok Apr 28 12:51:47 mrmoku: for the sake of simplicity i see this in fsogsmd modem plugin Apr 28 12:51:48 mickeyl, hmmm maybe I should send a (contreversed) patch upstream Apr 28 12:51:51 for that Apr 28 12:51:59 platform support for bq27200 Apr 28 12:52:01 mickey|office: alternative would be a fsoaudiod plugin? Apr 28 12:52:09 meego won't like it Apr 28 12:52:23 charger is bq24150 Apr 28 12:52:38 largely unrelated to bq27200 fuel gauge Apr 28 12:52:44 ok with mickeyl we were talking about the fuel gauge Apr 28 12:52:54 the battery % indicator Apr 28 12:53:19 * mrmoku notes that isi has an energy resource too :) Apr 28 12:53:19 btw i've another device with bq27200 Apr 28 12:53:28 and the gauge is very inexact Apr 28 12:53:34 yes, the BB5 also checks battery state Apr 28 12:53:34 it has uboot+xload Apr 28 12:53:57 I eared there was a way to recalibrate it Apr 28 12:53:59 DocScrutinizer: would that be a viable way too? Apr 28 12:54:08 I don't find it anymore tough Apr 28 12:54:16 I've never looked into it, but maybe yes Apr 28 12:54:22 ok Apr 28 12:55:10 NB the BB5 fw probably is running another ncarnation of bme, so as well is a strange obscure guestimate about battery charge Apr 28 12:55:35 mrmoku: yes, probably. Apr 28 12:56:02 DocScrutinizer: hmm, ok Apr 28 12:56:12 BB5 is used in a lot of featurephones Apr 28 12:56:20 what is BB5? Apr 28 12:56:22 much like calypso Apr 28 12:56:28 cellmo Apr 28 12:56:30 ah ok Apr 28 12:57:24 * mrmoku goes back to implement the next gpsd subblock dissector :) Apr 28 12:57:37 it even has a complete battery charger circuit that's not used in N900 Apr 28 12:57:52 mrmoku: yay Apr 28 12:58:01 so the value of battery state info from ISI is questionable Apr 28 12:59:12 DocScrutinizer: ok, will drop that idea then :) Apr 28 13:00:17 GNUtoo: as for the actual charging, adding the i2c code to the existing charging plugin might be a good idea as well... Apr 28 13:00:25 s/plugin/power supply/ Apr 28 13:00:27 mickey|office meant: GNUtoo: as for the actual charging, adding the i2c code to the existing charging power supply might be a good idea as well... Apr 28 13:01:44 ok Apr 28 13:02:26 would you be interested in a proper kernel side implementation of a bq24150 API? Apr 28 13:03:12 wow Apr 28 13:03:18 sure Apr 28 13:03:40 specially if it hits upstream Apr 28 13:03:49 mompl, I have a 50% done spec for /sys/bq24150/nodes Apr 28 13:04:12 no way this might be upstream, you'd have to push it :-D Apr 28 13:04:40 np Apr 28 13:05:06 for maemo I cancelled work on that, as it conflicts with closed bme methos of directly accessing the I2C interface to bq24150 Apr 28 13:05:49 ok Apr 28 13:05:52 this led to breaking bme so charging didn't work when that API / kernel module was loaded Apr 28 13:06:13 you don't have a closed blob bme you need to take care of Apr 28 13:07:03 indeed Apr 28 13:07:10 mompl, I'n pastebin the halfdone text Apr 28 13:07:41 it has a bit strange tab settings than need to be kept so table makes sense Apr 28 13:08:55 ok Apr 28 13:09:11 halfdone....what is incomplete? Apr 28 13:12:40 *lacking Apr 28 13:15:41 http://maemo.cloud-7.de/bq24150-sysnode.spec.txt Apr 28 13:16:20 GNUoo|n900: a few registers are lacking, and it's not too difficult to imagine how they should look Apr 28 13:16:57 ok Apr 28 13:17:02 if you actually plan to implement a kernel module based on this spec, I'd finish it in no time Apr 28 13:17:28 ok Apr 28 13:18:47 I must think with mickeyl if we go userspace or kernelspace Apr 28 13:19:59 but if my/the code hit upstream(that is to say, I submit for upstream) it could be worth to go kernelspace Apr 28 13:21:39 honsetly, there is not the slightest doubt here that this needs to go into kernel space. it's about one of the most vital functions in every electronic device and only the paranoid thought of losing intellectual property has been responsible for the crazy idea of bringing this up to userspace. Apr 28 13:21:48 can't say any more to that, really. Apr 28 13:22:06 ack Apr 28 13:22:09 I know that but.... Apr 28 13:22:26 meego kernel won't have our patches Apr 28 13:22:39 and it's rather a question of rebase/maintainibility Apr 28 13:22:57 meego kernel will have proprietary meego-patchhed closed version of bme Apr 28 13:23:09 there is no doubt that in in theory it should go in the kernel Apr 28 13:23:34 meego kernel doesn't care. so be it. they're free to go a technically worse way, they've been doing this in more than one place ;) Apr 28 13:23:35 <[Rui]> GNUtoo: who cares about meego kernel, just git am linus :) Apr 28 13:24:03 as those cowards prefer to *use* upstream or closed Nokia shit, rather than tackling the root issue and create their own charger Apr 28 13:25:49 the problem is that even if we're not meego, we use their kernel because it has lots of not-yet upstream patches Apr 28 13:26:14 and since theses patches won't be there Apr 28 13:26:39 my question was is it worth the possible tons of rebase on top of meego? Apr 28 13:26:48 *meego kernel Apr 28 13:27:25 for the linus kernel the solution is to submit upstream... Apr 28 13:27:39 that can be long but I think it's worth Apr 28 13:28:35 tons? Apr 28 13:28:41 we're talking about a charging driver... Apr 28 13:28:52 how many patches do you want to apply? :D Apr 28 13:29:08 DocScrutinizer: any idea what 'residual bers' might be? Apr 28 13:29:41 * mrmoku once in a while wants to understand one of the chinese things he is dissecting :P Apr 28 13:30:43 I seem to have heard about BER Apr 28 13:30:43 that's why I have to evaluate the work needed Apr 28 13:30:51 can't recall where Apr 28 13:31:07 ok :) Apr 28 13:31:13 nand ? Apr 28 13:31:18 no, GPRS Apr 28 13:31:26 hmm Apr 28 13:31:31 it's one of the QOS99 enums Apr 28 13:31:34 yeah, possible Apr 28 13:31:46 block errors Apr 28 13:31:55 isi? Apr 28 13:31:58 yup Apr 28 13:32:02 some CRC thing iirc Apr 28 13:32:13 because I recall having seen it somewhere Apr 28 13:32:20 maybe in fsogsmd logs or isi code Apr 28 13:33:12 probably isi code Apr 28 13:35:02 residual ber... Restfehlerrate Apr 28 13:35:08 something RF Apr 28 13:36:56 yeah, probably used to calculate a somewhat meaningful "signal strength" value in userland, as this isn't exactly signal *stength* but rather *quality* Apr 28 13:38:53 you notice that when upstream is obscured, the signal "strength" display of your cellphone suddenly drops from good to zero as soon as your phone tries to transmit some data to the BTS Apr 28 13:59:12 mrmoku, last meego kernel is kernel-adaptation-n900-2.6.37-38.1.src.rpm and ours is from kernel-adaptation-n900-2.6.37-38.1.src.rpm Apr 28 13:59:17 we should update Apr 28 13:59:22 maybe I'll work on that Apr 28 14:00:10 yep Apr 28 14:00:33 hmm, btw. 2.6.27-38.1 is equal to 2.6.37-38.1 Apr 28 14:00:39 so what are you going to do? :D Apr 28 14:09:57 28 and 38 Apr 28 14:09:58 sorry Apr 28 14:10:10 import them into fso git Apr 28 14:10:15 and then I'll see Apr 28 14:10:51 you don't want to use some of the brainfscked meego kernel bits Apr 28 14:11:12 e.g the braindamaged lis3lv02 kernel driver Apr 28 14:12:27 ah ? Apr 28 14:12:46 what's the alternative? Apr 28 14:12:48 mainline? Apr 28 14:12:56 what do mainline lack? Apr 28 14:13:17 pretty please reconsider this decission. meego kernel for N900 is quite borked in some respects. Maemo kernel (i.e. the Nokia patches on top of mainline kernel) is way better and and more mature, e.g. regarding power management Apr 28 14:13:48 so you want us to use 2.6.28 maemo? Apr 28 14:14:04 the patches should apply on .38 as well Apr 28 14:14:06 I was for it for a long time unfortunately mickeyl was against it Apr 28 14:14:09 ok Apr 28 14:14:21 if they don't it's probably not too hard to adapt them Apr 28 14:14:28 but then isi changed between both kernel Apr 28 14:14:47 which prevent us from using stock maemo kernel Apr 28 14:14:55 :nod: Apr 28 14:15:15 for instance I'm not sure forwarding would work on 28 Apr 28 14:15:17 so just use a hybrid patchset on top of mainline kernel Apr 28 14:15:23 ah ok Apr 28 14:15:27 lots of work then? Apr 28 14:15:43 time is very scarse for me Apr 28 14:16:41 I don't even know what's lacking in mainline Apr 28 14:17:16 but I like a lot the mainline idea Apr 28 14:17:27 I'm just not sure it's the right moment to use it Apr 28 14:18:29 did dvfs hit mainline? Apr 28 14:18:35 well, all I can say is meego kernel == won't yield a proper working phone Apr 28 14:19:04 ok Apr 28 14:20:31 meego guys postponed basic design decisions for a time after first working implemetation. This is BS as you can't change e.g sensor APIs after you're done with building a framework using the borked API Apr 28 14:20:57 ouch Apr 28 14:22:04 power management? doesn't matter - we'll care about that minor detail after the system is running in first PoC version Apr 28 14:22:45 which is a notion that leads to using kernel driver for lis302 acellerometer that's POLLING Apr 28 14:23:01 polling....ouch ouch !!!! Apr 28 14:23:16 DocScrutinizer: adapting the maemo patches is _no easy_ job Apr 28 14:23:23 indeed Apr 28 14:23:30 now after they implemented their QtMob and sensorfw middleware on that, how will they ever change that kernel api? Apr 28 14:23:35 quite everything changed in the arm part of mainline meanwhile Apr 28 14:23:42 the first thing that I really need to take a decision is what's lacking in mainline Apr 28 14:23:51 s/mainline/linux-next Apr 28 14:24:30 mrmoku, yes, it is still possible bug require *a lot of time*, which I haven't Apr 28 14:24:46 I have also: Apr 28 14:24:49 *real life Apr 28 14:24:54 *replicant Apr 28 14:24:57 *buglabs Apr 28 14:25:05 (community involvement) Apr 28 14:25:30 not to mention that the n900 has also other areas of work Apr 28 14:25:41 such as fso and modem Apr 28 14:25:42 quite some bits hit mainline meanwhile I think Apr 28 14:25:46 yes Apr 28 14:25:53 so first thing would be indeed to check what's missing Apr 28 14:26:50 also perfect is the enemy of good Apr 28 14:27:07 maybe let's just aim at good for now? Apr 28 14:31:26 maybe my aim for minimal work make me not see stuff, tough... Apr 28 14:37:26 seems that if I'm right about the patches format, the dvfs was not merged Apr 28 14:38:40 for instance: Apr 28 14:38:53 linux-2.6-38-leds-lp5523-fix-circular-locking.patch means that it was merged in 38 if I'm right Apr 28 14:39:10 and dvfs has linux-2.6-patchname Apr 28 14:40:16 GNUtoo: fastest way might be to build current mainline from git and see what happens? Apr 28 14:40:52 http://www.pastie.org/1843545 Apr 28 14:40:56 all the patches Apr 28 14:40:57 no Apr 28 14:41:08 I would have to test everything by hand Apr 28 14:41:13 and it takes too much time Apr 28 14:41:25 and some stuff can only be noticed after long time Apr 28 14:41:32 for instance improper shutdown Apr 28 14:41:46 reading the commit messages of the patches is way better and faaster Apr 28 14:43:11 should I push? Apr 28 14:44:29 the MeeGo N900 (http://wiki.meego.com/ARM/N900) team would be receptive for any constructive feedback regarding the kernel as well Apr 28 14:44:31 GNUtoo: whatever you think is better for you Apr 28 14:44:46 ie. #meego-arm or meego-dev mailing list (with N900 subject prefix) Apr 28 14:45:13 now that the team was beefed up Apr 28 14:45:33 pushed the new patchset Apr 28 14:45:35 not sure though who is handling the kernel Apr 28 14:45:49 i'm in meego-arm channel too Apr 28 14:46:04 (but stskeepz probably knows, who) Apr 28 14:48:13 SHR: 03mok 07isi-wireshark-plugin * r8cbfe7da64e8 10/src/isi-call.c: work on the call dissector Apr 28 14:48:20 SHR: 03mok 07isi-wireshark-plugin * r77c01f388ed5 10/ (Makefile src/isi-gpds.c src/isi-gpds.h src/packet-isi.c): add dissector for gpds resource Apr 28 14:48:20 SHR: 03mok 07isi-wireshark-plugin * rd9ec02bb0610 10/ (5 files in 2 dirs): add isi-commgr Apr 28 14:48:23 SHR: 03mok 07isi-wireshark-plugin * r70bf9d89dc6c 10/src/isi-phoneinfo.c: work on the phoneinfo dissector Apr 28 14:48:24 SHR: 03mok 07isi-wireshark-plugin * r8dc554841680 10/ (4 files in 2 dirs): add isi-selftest Apr 28 14:48:26 SHR: 03mok 07isi-wireshark-plugin * r616e48da177a 10/src/isi-mtc.c: work on the mtc dissector Apr 28 14:49:17 Mirv: actually #meego-arm team is rather ignorant and "we use upstream" mantra centered Apr 28 14:50:27 Mirv: I.E. I brought up that topic to them several times, from very beginning to maybe last month last time. I gave up on them for now, as their main answers always are "is it upstream?" and "supply patches!" Apr 28 14:51:42 DocScrutinizer: communication is usually the hardest hurdle. initiative for other than supplying patches directly etc. could be better taken on the mailing list. Apr 28 14:52:05 honesrly I'm not interested in it anymore Apr 28 14:52:07 or maybe the IRC meetings, but it's also a bit hard to formulate the whole idea there Apr 28 14:53:00 well, then it's moot Apr 28 14:53:39 I tried it several times, others tried it on my behalf when I annoyed them with my rant in other than #meego-arm channel, the result always been the same Apr 28 14:53:45 but also I think the work of theirs shouldn't be dismissed if one is not interested in improving it either. maemo kernel is a dead end anyway, and meego n900 team at least tries to gets new stuff upstream (AFAIK) Apr 28 14:54:09 the alternative of course would be to submit better patches directly to upstream, ie. ignoring them Apr 28 14:54:27 I'm not saying like I'd know anything about meego n900 kernel, just talking in general Apr 28 14:54:54 stskeeps even put me on his ignore list Apr 28 14:55:12 heh fun Apr 28 14:55:19 ouch Apr 28 14:55:34 :( and so another foss projects have communication problems instead of co-operation Apr 28 14:56:10 there should be a software for dealing with people interaction properly Apr 28 14:58:36 for instance an improved spell checker? Apr 28 14:58:56 that would check also the content Apr 28 14:59:09 or in case like for koen for rude things Apr 28 14:59:21 and read it aloud before posting ? ( xkcd did it ) Apr 28 14:59:28 lol Apr 28 15:06:10 btw the meego phone-UX & kernel seems as at very least dead as maemo Apr 28 15:06:29 ah? Apr 28 15:06:34 do you have more details? Apr 28 15:06:43 what kernel? Apr 28 15:06:57 if it's the gitorious one it has been superseeded Apr 28 15:07:08 by a src.rmp one Apr 28 15:07:10 it has exactly same amount of successor devices based on it (zero), it hasn't the maturity of maemo kernel, and it has no real developer community Apr 28 15:07:22 well the UX is uninteresting, the UX:s are anyway reference placeholder type of things Apr 28 15:07:25 ok Apr 28 15:11:35 if by 2012 meego investment will be radically cut down, we can just hope they fix the kernel before then Apr 28 15:11:41 *shrug* Apr 28 15:12:29 well the MeeGo future is in non-Nokia anyway, but indeed there is some time at least for them also to work on proper non-maemo N900 support + support for the one future device whatever it is Apr 28 15:13:55 only God knows what kernel Nokia will use for their first and last "meego related product" they announced for 2011, at very least rumour has it they asked for allowance to use the meego brand for their maemo-meego-harmattan hybrid bastard and got denied. And maybe that "related product" is a t-shirt, who knows Apr 28 15:14:34 a sticker to place on the phone ? Apr 28 15:14:57 certainly it's going to be a mixed bag as usual, but also as usual the hw makes probably one drool compared to FreeRunner as well :) Apr 28 15:15:17 but maybe GTA04 will ease the pain Apr 28 15:15:41 if I get it and get it working as a daily phone, I should be all set for a couple of years again Apr 28 15:15:59 all rumour has it there'll be no hw kbd and a c-ts on whatever the 'meego related phone' will be, if any Apr 28 15:17:11 so definitely not attractive in any way for me at least Apr 28 15:17:41 if I'd be interested in such a design, I could have sticked to FR, or get me an iPhone Apr 28 15:19:54 and honestly I wonder why Nokia should sink money into a hw design that has no future - as Nokia themselves says there'll be no further meego devices after that. And for sure they won't earn a lot of money on that one-time-wonder Apr 28 15:22:02 I'm still having hopes for GTA-04, at least as a FR replacement... Apr 28 15:22:15 ~hail GTA04 Apr 28 15:22:16 * apt bows down to GTA04 and chants, "I'M NOT WORTHY!!" Apr 28 15:22:24 My idea would've been they open up the last few closed bits of *maemo*, and for the future just ship their top notch smartphones with whatever winP7 crap *plus a linux BSP* so it'd be easy to port maemo to the new devices Apr 28 15:23:41 no need for meego then Apr 28 15:28:57 mickey|office: do you already looked at pyqt and pyside? Apr 28 15:29:22 So, do 'the SHR-devs' have access to a GTA-04 already? Or at least, will they? Apr 28 15:31:42 instead Nokia decided to bin maemo for a supposedly superior more open approach based on that maemo-moblin merger meego. Then they stopped support for maemo first, then they stopped participation in meego at large, and finally they binned the whole idea of FOSS for elopocalypse Apr 28 15:32:30 I was surprised, to say the _very_ least, to hear that Nokia went WinMobile! Apr 28 15:32:47 1st thought: Well, that's the end of Nokia, then! Apr 28 15:33:08 which likely it is Apr 28 15:33:22 Ah well, more place for Apple and Android, then! Apr 28 15:33:38 And GTAxxx, ofcourse (wichful thinking) Apr 28 15:42:03 morphis: no, i'm at work :) Apr 28 15:42:24 i should even chat less *cough* Apr 28 15:42:31 anyways, homewards now Apr 28 16:10:34 btw I changed the nokia900 picture here: http://wiki.freesmartphone.org/index.php/HardwareComparison Apr 28 16:16:13 mickeyl, with newer kernel bq272xx works out of the box Apr 28 16:16:23 I'll test a bit and push my oe changes Apr 28 16:16:38 awesome, cool Apr 28 16:20:20 hmmm bluetooth still fails Apr 28 16:20:28 Inquiry failed: Connection timed out Apr 28 16:39:17 morphis: building pyside now (desktop) Apr 28 16:46:27 morphis: http://www.pyside.org/pyside-v0-3-benchmarks/ Apr 28 17:01:38 GNUtoo: wow, out of the box means reading battery status works? Apr 28 17:01:51 yes Apr 28 17:02:03 I'll push instead of playing with bluetooth that doesn't work Apr 28 17:04:25 well, the basic bq27200 driver worked for maemo powerkernel quite flawlessly, except it broke bme, again by blocking bme direct I2C access to the chip Apr 28 17:07:25 I seem to recall efforts to tweak the generic kernel driver behaviour so it doesn't occupy the direct I2C access exclusively by loading a dedicated chip driver like bq27200. No idea if it resulted in anything usable Apr 28 17:07:58 not sure, the bq27x00 I saw was very hackish Apr 28 17:08:16 mhm, then it's probably been that hack Apr 28 17:08:20 it contained a patched driver in order to avoid writing platform device stuff in the board machine Apr 28 17:08:21 GNUtoo: btw. where do you push to? Apr 28 17:08:26 oe Apr 28 17:08:30 good :) Apr 28 17:08:31 earlier versions should be way more clean Apr 28 17:08:36 ok Apr 28 17:08:43 brb to try it out :) Apr 28 17:19:17 mrmoku, pushed Apr 28 17:23:35 where to? Apr 28 17:25:44 ah, oe Apr 28 17:32:16 Heinervdm, hi, did you see the log I posted you here? Apr 28 17:32:31 pespin: no Apr 28 17:35:19 Heinervdm, can you look for it in log? There's some lines of text with explanations of what happened when I tried it ;) Apr 28 17:35:40 pespin: ok, i will look Apr 28 17:35:41 afair it was the day before yesterday's night Apr 28 17:37:56 found it Apr 28 17:55:09 DocScrutinizer: where/when did Nokia say there will be no further meego devices? Apr 28 17:57:32 err, sorry I haven't stored the links, but it's commonly known Nokia said meego will be put on some R&D experimental status after that "meego" related product Apr 28 17:58:12 and their future high end devices won't be based on meego but rather on winP7 now Apr 28 18:15:46 btw, someone knows if there's some project trying to port linux to blackberry? Apr 28 18:16:32 blackberry uses a proprietary O.S.? Apr 28 18:20:23 yes, fully Apr 28 18:20:32 pespin: never heard about it Apr 28 18:20:44 pespin: you want them all? ;) Apr 28 18:21:50 i tought blackberry uses WM6,so you could try haret and try a arm kernel,if blackberry runs about one.. Apr 28 18:22:18 it uses its own Java OS from waht i've read. Apr 28 18:22:22 no windows. apparantly the bootloader is pretty locked Apr 28 18:22:55 nothing a good JTAG interface can't fix though... once you found it ;) Apr 28 18:23:14 no idea about all this hehe :) Apr 28 18:23:37 I was just thinking about it because some blackberries are quite cheap Apr 28 18:24:18 I don't know if those have good cpu or much mem thought Apr 28 18:24:33 you already opened it ? Apr 28 18:24:37 to see the CPU? Apr 28 18:24:46 or it is also proprietary :( ? Apr 28 18:25:11 I have no blackberry hehe Apr 28 18:25:53 mickey|patio: hmm... do you know how gprs works on the n900? Apr 28 18:26:14 I think if SHR could run on those devices we could get more people involved because they are cheaper to acquire :) Apr 28 18:27:07 i think if Linux could run on those new phones,Linux on the world would be so big :) Apr 28 18:27:21 mickey|patio: possible traffic goes via the 'Pipe Messages' resource? Apr 28 18:28:19 DocScrutinizer: I was just wondering whether there has been some further news that I missed. the original announcement was vague enough that you can read all kinds of things into it; my own interpretation is not quite as negative... Apr 28 18:28:22 i have no idea Apr 28 18:28:28 ok, I will find out :-) Apr 28 18:28:32 :) Apr 28 18:28:47 just saw some sip registration in the payload of some messages there... Apr 28 18:28:51 so I started to wonder Apr 28 18:29:42 sigh... next dissector to implement :P Apr 28 18:30:26 activating and deactivating a context looks easy enough Apr 28 18:30:31 i don't know what happens afterwards though Apr 28 18:32:54 actually Apr 28 18:33:07 it looks like a network interface should come up Apr 28 18:33:27 mickey|patio: do you looked at the pyside thing already? Apr 28 18:33:40 morphis_: yes, it's nice, but it's going to be a nightmare to add it to OE Apr 28 18:34:31 mickey|patio: hmm... let me check if there is one in maemo Apr 28 18:34:38 mickey|patio: ok, then you discovered the same than me Apr 28 18:35:18 we need shiboken, this needs generatorrunner and apiextractor, some of them we need native, some cross Apr 28 18:35:24 fun fun fun Apr 28 18:35:24 jepp Apr 28 18:35:29 and all on top of cmake Apr 28 18:35:30 and we need qt4 native too Apr 28 18:35:37 that one we hav Apr 28 18:35:45 but it doesn't install everything necessary Apr 28 18:35:54 and it's lacking xmlpatterns Apr 28 18:36:00 hm Apr 28 18:36:33 apparantly there is no other -native package in OE that needs Qt Apr 28 18:36:52 or they all have it installed on the respective buldhosts Apr 28 18:38:19 so you don't think we get it working without a lot of work? Apr 28 18:39:05 either way it's work Apr 28 18:39:07 as pyqt is outdated Apr 28 18:39:15 the version in OE, yes Apr 28 18:39:39 i'll continue to play a bit around with pyside in OE Apr 28 18:39:57 turnarounds are very long though Apr 28 18:40:00 ah, you already started? Apr 28 18:40:02 yes Apr 28 18:40:23 on amethyst? Apr 28 18:40:28 yes Apr 28 18:40:31 mickey|patio: indeed there is gprs0 Apr 28 18:40:39 mrmoku: excellent. then it should be simple Apr 28 18:40:59 * mrmoku still wonders what that pipe thing is for... Apr 28 18:41:12 yeah, trying to implement that Apr 28 18:41:19 cool, good speed Apr 28 18:41:41 we stalled to long due to that EPOC_INFO shit... need to speed up :) Apr 28 18:42:08 yes, that was unfortunate. we lost like... 1 month or so Apr 28 18:42:25 and many grey hairs on the way *cough* Apr 28 18:43:46 yeah... the last remaining dark ones got grey on the road indeed :P Apr 28 18:47:12 i think you're right wrt. the pipe Apr 28 18:47:56 but it merely looks like the way to enable the automatic transfer between loosely coupled units on the modem Apr 28 18:48:05 much like the cmtspeech is being used to couple stuff Apr 28 18:48:13 so we don't have to shuffle bytes arond with the pipe Apr 28 18:48:21 but we need to create it and set the endpoints correctly Apr 28 18:48:47 the rest will be handled by the modem then Apr 28 18:49:22 and phonet kernel Apr 28 18:49:57 yeah, that's what the packets tell... right after starting configuration with gpsd Apr 28 18:50:01 it creates a pipe Apr 28 18:50:18 will know more in half an hour when my dissector is finished :) Apr 28 18:50:23 excellent Apr 28 19:20:24 mickey|patio: so our plan is still using python for aurora and get pyside running? Apr 28 19:24:28 as then I will start experimenting a bit more with pyside Apr 28 19:24:32 we will not use Qml anymore ? Apr 28 19:24:44 angelox_123: no Apr 28 19:25:05 we will maybe just use python instead of c++/vala Apr 28 19:25:09 to speed up development Apr 28 19:25:29 python/QML we will use Apr 28 19:25:33 hum,ok ... i have never programmed in python... Apr 28 19:25:43 python/QML... Apr 28 19:26:12 it's easy Apr 28 19:26:16 so,the graphical part will continue be using QML? Apr 28 19:26:21 yes Apr 28 19:26:34 hum,ok i'll try learn a bit about python Apr 28 19:26:35 the most things will be done in QML Apr 28 19:26:55 i have a experience with QML.. Apr 28 19:26:56 python will be just there for the parts around Apr 28 19:27:33 hum ok..thanks..i've started to make a GUI for my project using QML...so i guess i will can understand the GUI... Apr 28 19:27:42 ok Apr 28 19:28:04 thanks by information :D Apr 28 19:28:10 no problem Apr 28 19:43:45 so I am off for today Apr 28 19:43:47 bye Apr 28 19:43:56 <[Rui]> bye Apr 28 19:44:01 <[Rui]> and hi :) Apr 28 19:50:55 /whois me Apr 28 19:50:58 ops Apr 28 23:38:40 someone online? Apr 28 23:38:51 i need GPS assistance fot GTA02 Apr 28 23:39:14 after a long time without using my GTA02, i just reflashed SHR unstable Apr 28 23:39:21 but GPS is not working :( Apr 28 23:40:56 i did not test todays image so far. but it worked last week Apr 29 00:00:03 i don't have the capacitor fix Apr 29 00:00:18 darkh, do you remember the option for FFTT? and ephemerides? Apr 29 00:02:22 don't know what option you mean. but i have the gps fix, so i don't have gps fix problems **** ENDING LOGGING AT Fri Apr 29 02:59:57 2011