| DocScrutinizer | lbt_away: finally it is Nokia who want *us* to do sth *they* profit from - they want us to help them sell phones. So they damn better do all for making this a pleasant ride | 00:00 |
|---|---|---|
| rm_you | -mcpu=cortex-a8 -mfpu=neon -mfloat-abi=softfp | 00:00 |
| rm_you | is that still accurate for n950? | 00:00 |
| dm8tbr | wasn't it hardfloat now? (haven't compiled anything yet) | 00:01 |
| *** TSCHAKeee has quit IRC | 00:01 | |
| *** antman8969 has quit IRC | 00:02 | |
| *** TSCHAKeee has joined #harmattan | 00:02 | |
| DocScrutinizer | dm8tbr: the fun bit is .deb for N900 install and run :-o | 00:06 |
| DocScrutinizer | despite some always claimed this was *impossible* | 00:07 |
| DocScrutinizer | seems there's a bug or patch in elf loader that allows to load softfp as well as hardfp | 00:08 |
| DocScrutinizer | or harmattan *IS* softfp, and thus not at al compatible with meegoCE | 00:09 |
| dm8tbr | well as long as there isn't any floats being passed around it will just work | 00:09 |
| dm8tbr | e.g. if the binary is static it will work just fine | 00:10 |
| DocScrutinizer | that's what I thought as well, but I've been told so many times it can't I almost believed it | 00:10 |
| dm8tbr | if the binary makes only non-float calls to libraries, everything should be fine | 00:11 |
| dm8tbr | it's only if it passes floats it will likely blow up | 00:11 |
| lcuk | dm8tbr, DocScrutinizer - the libliqbase library and then apps appear to work | 00:16 |
| lcuk | and I am sure that somewhere I use floating point numbers | 00:16 |
| lcuk | the same was also occuring for maemo4/5 transition | 00:18 |
| *** Scifig has joined #harmattan | 00:25 | |
| Damion3 | have you tried simply running the n900 mplayer? | 00:43 |
| DocScrutinizer | how would I find out about soft/hardFP property of an ELF? | 00:59 |
| DocScrutinizer | file(1) seems not to show this | 00:59 |
| DocScrutinizer | obj* ? | 01:00 |
| Clint | it's not an ELF property | 01:01 |
| rm_you | Damion3: no, haven't tried? | 01:03 |
| dm8tbr | http://www.trimslice.com/forum/viewtopic.php?p=762&sid=34d281b00e6f52107467153f5f5e37d1#p762 that might help | 01:04 |
| *** epage has quit IRC | 01:05 | |
| rm_you | so lets see, what do i run in scratchbox to build packages? i know it's dpkg-buildpackage, but i feel like there was something involving fakeroot and some other args i needed to pass | 01:08 |
| * GeneralAntilles facepalms. | 01:09 | |
| GeneralAntilles | Can't get hex codes out of DigitalColor Meter anymore. | 01:09 |
| GeneralAntilles | Die, Apple, die. | 01:09 |
| rm_you | >_> | 01:09 |
| rm_you | i mean you can just convert them in another app, but i guess that's annoying | 01:10 |
| *** smoku has left #harmattan | 01:10 | |
| GeneralAntilles | Yeah, previously I could hover over a pixel and get the hex value in the clipboard. | 01:11 |
| GeneralAntilles | Somebody had to work to remove this feature. | 01:11 |
| Damion3 | rm_you: almost everything I've tried has worked | 01:20 |
| rm_you | hrm | 01:24 |
| rm_you | so, what | 01:24 |
| rm_you | just add the repo for fremantle and install mplayer? | 01:24 |
| lcuk | GeneralAntilles, retina screens are not compatible with single pixels :P | 01:24 |
| GeneralAntilles | lcuk, I wish my desktop had that pixel density. | 01:24 |
| * lcuk nods | 01:25 | |
| lcuk | me too | 01:25 |
| DocScrutinizer | http://paste.debian.net/123950/ native less binary vs alien N900 working find binary | 01:27 |
| GeneralAntilles | If they could just double it. | 01:27 |
| SpeedEvil | GeneralAntilles: It's depressing. | 01:27 |
| GeneralAntilles | 3840 x 2160 | 01:27 |
| GeneralAntilles | Then just pixel-double bitmaps when needed. | 01:27 |
| SpeedEvil | GeneralAntilles: I want ~150dpi or so - 200 would be ideal 27" or so 4:3 (or 1:1) display | 01:28 |
| DocScrutinizer | esp remarkable is the missing >>Tag_ABI_VFP_args: VFP registers<< in fremantle's find | 01:31 |
| DocScrutinizer | please note I have not the faintest idea about all that FP stuff, just stating the obvious | 01:33 |
| Damion3 | rm_you: that might clash with newer libs for video. I'd use develsh so you can execute stuff and scp mplayer and libs to user's homedir. the stuff I've run so far has just been the odd binary scped like iwconfig, qemu and rsync | 01:37 |
| rm_you | hrmrm | 01:39 |
| *** meegoexperts_mob has quit IRC | 01:39 | |
| rm_you | does maemo-optify exist? | 01:40 |
| DocScrutinizer | dm8tbr: thanks | 01:41 |
| *** rcg has quit IRC | 01:41 | |
| DocScrutinizer | rm_you: for harmattan? doesn't make any sense you'd think | 01:45 |
| DocScrutinizer | rm_you: nevertheless optified .deb from fremantle install to /opt, creating it when needed | 01:47 |
| rm_you | err k | 01:50 |
| rm_you | so i can remove the line in this debian config that runs maemo-optify? | 01:50 |
| rm_you | bbl dinner | 01:51 |
| *** antman8969 has joined #harmattan | 02:24 | |
| GeneralAntilles | Who's going to remind me not to forget MWKN tomorrow? | 02:27 |
| DocScrutinizer | GeneralAntilles: this irc client lacks proper scripting capabilities :-) I could try to set up a script with /timer in xchat :-) | 02:34 |
| GeneralAntilles | I've set an alarm on both phones, but I still worry. | 02:34 |
| GeneralAntilles | I've forgotten more than once even with Jaffa reminding me. | 02:34 |
| DocScrutinizer | I'll try to recall to ask you about it in ~18h | 02:35 |
| GeneralAntilles | Thanks | 02:35 |
| *** lynxis has joined #harmattan | 02:40 | |
| lynxis | hi | 02:41 |
| *** crevetor has quit IRC | 02:41 | |
| *** rcg has joined #harmattan | 02:48 | |
| *** rcg has quit IRC | 02:49 | |
| *** antman8969 has quit IRC | 02:50 | |
| rm_you | MWKN? | 03:13 |
| *** Termana|rdlBNC has joined #harmattan | 03:46 | |
| *** epage has joined #harmattan | 03:47 | |
| *** lynxis has quit IRC | 03:50 | |
| *** Termana has quit IRC | 03:51 | |
| *** rm_you has quit IRC | 03:51 | |
| *** bemasc has quit IRC | 03:51 | |
| *** rm_you has joined #harmattan | 03:52 | |
| *** rm_you has quit IRC | 03:52 | |
| *** rm_you has joined #harmattan | 03:52 | |
| *** bemasc has joined #harmattan | 03:52 | |
| *** antman8969 has joined #harmattan | 04:10 | |
| *** M4rtinK2 has quit IRC | 04:25 | |
| *** Termana|rdlBNC is now known as Termana | 04:30 | |
| Termana | morning | 04:31 |
| *** lynxis has joined #harmattan | 04:41 | |
| *** NIN101 has quit IRC | 04:49 | |
| rm_you | ha | 04:50 |
| rm_you | OK | 04:51 |
| rm_you | got mplayer compiled and installed | 04:51 |
| rm_you | it seems to be working | 04:51 |
| rm_you | odd bugs when i launch it from the terminal... but i can make a guio | 04:51 |
| rm_you | *gui to fix that | 04:51 |
| rm_you | it is using 60% or so CPU but seems to be playing better than the builtin T_T | 04:51 |
| Termana | rm_you, I don't know when you asked it, but it's sitting unanswered in my log - MWKN = Maemo Weekly News = mwkn.net | 04:52 |
| rm_you | ah | 04:52 |
| lynxis | somebody else here, who trying to press out some sound from fm-transceiver ? | 04:54 |
| *** wicket64 has quit IRC | 04:54 | |
| GAN900 | rm_you, what?! | 04:56 |
| GAN900 | You don't read mwkn.net? | 04:56 |
| rm_you | no? | 04:56 |
| GAN900 | I've wasted my life. | 04:56 |
| rm_you | hrm | 04:56 |
| rm_you | trying to figure out how to benchmark "video playback smoothness" so i can accurately compare between mplayer and video-suite | 04:57 |
| Termana | GAN900, lol | 04:57 |
| *** antman8969 has quit IRC | 04:57 | |
| SpeedEvil | rm_you: Does playback over composite work? | 04:59 |
| rm_you | composite? | 04:59 |
| rm_you | XV? | 04:59 |
| SpeedEvil | no - the video out | 04:59 |
| rm_you | oh | 04:59 |
| rm_you | have not tried | 04:59 |
| rm_you | device keeps going to sleep while mplayer plays video tho :( | 05:01 |
| rm_you | ok so | 05:01 |
| rm_you | lesson: don't make the device downscale | 05:02 |
| rm_you | it's a bad idea | 05:02 |
| rm_you | this is slightly concerning: | 05:06 |
| rm_you | [AO OSS] audio_setup: Can't open audio device /dev/dsp: No such file or directory | 05:06 |
| rm_you | but i think it's ok, since it then goes on to use Pulse | 05:06 |
| rm_you | hrmrmrm | 05:06 |
| rm_you | 23% CPU playing this video... | 05:07 |
| rm_you | need to compare playback quality to video-suite (still haven't decided HOW) | 05:07 |
| rm_you | video suite seems to be using ... umm.... the same CPU, across qmafw and pulseaudio :/ | 05:08 |
| rm_you | ~22% on qmafw, and 15% on pulseaudio | 05:09 |
| rm_you | ~30% on mplayer, and 9% on pulseaudio | 05:10 |
| rm_you | soooooo basically the exact same usage (mplayer does more with the audio before passing it to pulse? no idea) | 05:11 |
| rm_you | the package is here if anyone else wants to try: | 05:11 |
| rm_you | http://www.sheeplauncher.net/debs/ | 05:11 |
| rm_you | did everyone go to sleep? | 05:11 |
| rm_you | SpeedEvil: I'll test playback over composite now -- it works with video-suite, i assume, right? | 05:12 |
| rm_you | SpeedEvil: just tested, both mplayer and video-suite seem to play back exactly the same over composite | 05:20 |
| DocScrutinizer | I'm interested in the audio playback part | 05:22 |
| *** npm_ has joined #harmattan | 05:23 | |
| *** npm has quit IRC | 05:23 | |
| DocScrutinizer | I spent some 2 or 3 hours yesterday to scrutiize each available executable on my mission to find what's the new play-sound command on harmattan now. Did only find aplay | 05:23 |
| *** npm_ has quit IRC | 05:25 | |
| rm_you | DocScrutinizer: well you can install it, it works fine | 05:26 |
| rm_you | hrm, wish someone here could tell me what I need to do to switch from old Qt stuff in Python to the new Qt lib (Qtside?) | 05:26 |
| rm_you | err PySide | 05:26 |
| epage | What do you mean by make the switch? | 05:27 |
| rm_you | like, i have a python app that uses the old PyQt | 05:28 |
| rm_you | i need to reconfigure it to use pyside | 05:28 |
| rm_you | is it as easy as switching the imports? | 05:28 |
| rm_you | or do i have to do more conversion | 05:28 |
| epage | Its not too bad, I can collect up my notes | 05:28 |
| epage | It depends, there is PyQt's v2 APIs that PySide is modeled after but you had to opt-in to them. | 05:29 |
| DocScrutinizer | rm_you: install what? play-sound? | 05:29 |
| *** npm_ has joined #harmattan | 05:29 | |
| epage | I hacked up a python module to make it easy to switch back and forth between PySide and PyQt (especially for platforms with only one or the other) | 05:29 |
| epage | https://github.com/epage/DialCentral/blob/master/src/util/qt_compat.py | 05:29 |
| rm_you | hrm | 05:29 |
| rm_you | well | 05:30 |
| rm_you | it seems easy as this is a simple program | 05:30 |
| rm_you | so lets see if just changing lib names will work | 05:30 |
| epage | pyqtSignal -> Signal | 05:30 |
| epage | pyqtSlot -> Slot | 05:30 |
| epage | pyqtProperty -> Property | 05:30 |
| epage | Plus QString and QVariant were abolished | 05:30 |
| DocScrutinizer | no more QStrings? AAAWWW! | 05:31 |
| rm_you | k | 05:31 |
| rm_you | __pyqtSignals__ = ( "shutdown" ) | 05:31 |
| rm_you | what happens to that | 05:31 |
| epage | You can get the string/variant change in PyQt by doing the following before any of your pyqt imports | 05:32 |
| epage | import sip | 05:32 |
| epage | sip.setapi('QString', 2) | 05:32 |
| epage | sip.setapi('QVariant', 2) | 05:32 |
| epage | I've never used __pyqtySignals__ before | 05:32 |
| DocScrutinizer | the thing that brought most fun to your life every day. Just binned :-S | 05:32 |
| epage | I always declare the signals as class member variables | 05:32 |
| rm_you | hrm | 05:32 |
| epage | For example: | 05:32 |
| rm_you | self.emit(QtCore.SIGNAL("shutdown")) | 05:32 |
| epage | https://github.com/epage/DialCentral/blob/master/src/session.py | 05:32 |
| epage | self.signalName.emit(params) | 05:32 |
| epage | new style signal/slot API | 05:33 |
| rm_you | err so | 05:33 |
| epage | that old one I think has been discouraged for a while | 05:33 |
| rm_you | self.shutdown.emit() ? | 05:33 |
| epage | yes | 05:33 |
| DocScrutinizer | now that sounds weird | 05:33 |
| epage | It sounds natural in python | 05:33 |
| * DocScrutinizer should shut up | 05:33 | |
| epage | Also PySide requirs you to be strict about your chaining to super when calling __init__ | 05:34 |
| rm_you | Depends: python2.5, python2.5-qt4-dbus, python2.5-qt4-gui, python2.5-qt4-webkit, python2.5-qt4-network, python2.5-gobject | 05:34 |
| rm_you | what do i do with this line | 05:34 |
| rm_you | do i not NEED to depend on any of this anymore? | 05:34 |
| rm_you | well | 05:34 |
| rm_you | the QT bits | 05:34 |
| epage | My deps went from | 05:35 |
| epage | python2.5-qt4-core, python2.5-qt4-gui, python2.5-qt4-maemo5 | 05:35 |
| epage | to | 05:35 |
| epage | python-pyside.qtgui, python-pyside.qtcore, python-pyside.qtmaemo5 | 05:35 |
| *** npm_ has quit IRC | 05:35 | |
| rm_you | kk | 05:35 |
| rm_you | thanks | 05:35 |
| rm_you | and i guess 2.6 | 05:35 |
| rm_you | instead of 2.5 | 05:35 |
| epage | I stopped specifying the python version | 05:36 |
| epage | packages provide both names | 05:36 |
| *** leinir has quit IRC | 05:37 | |
| rm_you | Depends: python2.5, python2.5-qt4-dbus, python2.5-qt4-gui, python2.5-qt4-webkit, python2.5-qt4-network, python2.5-gobject | 05:38 |
| rm_you | Depends: python, python-dbus, python-pyside.qtgui, python-pyside.qtwebkit, python-pyside.qtnetwork, python-gobject | 05:38 |
| rm_you | i think that works | 05:38 |
| epage | your dbus one looks off | 05:39 |
| rm_you | well | 05:39 |
| *** npm_ has joined #harmattan | 05:39 | |
| rm_you | i did: | 05:39 |
| rm_you | apt-cache search python | 05:39 |
| rm_you | on the device | 05:39 |
| rm_you | err | 05:39 |
| rm_you | apt-cache search python-pyside | 05:39 |
| rm_you | and dbus isn't in the list | 05:39 |
| rm_you | but there is a python-dbus... | 05:40 |
| epage | that is probably gobject based | 05:40 |
| rm_you | attempting to install it | 05:40 |
| rm_you | we'll see if it runs | 05:40 |
| rm_you | worst case, aegis kills my device :P | 05:40 |
| rm_you | hrm | 05:41 |
| rm_you | so what do i have to do to get an icon to show up | 05:41 |
| rm_you | apparently the old .desktop doesn't work | 05:41 |
| *** npm_ has quit IRC | 05:43 | |
| TSCHAKeee | is it possible for GridView items to have different row or column spans? | 05:43 |
| epage | .desktop files changed locations | 05:43 |
| *** npm_ has joined #harmattan | 05:44 | |
| TSCHAKeee | oh flow, nevermind | 05:44 |
| rm_you | epage: know where they went? | 05:46 |
| rm_you | also, is there a new equivilent to "/proc/component_version" to get device ID? | 05:47 |
| epage | standard desktop location | 05:47 |
| epage | http://wiki.meego.com/Porting_Fremantle_Applications_to_Harmattan#Change_Paths_for_Desktop_Files_etc. | 05:47 |
| *** lynxis has quit IRC | 05:49 | |
| *** npm_ has quit IRC | 05:50 | |
| *** npm_ has joined #harmattan | 05:52 | |
| rm_you | TypeError: PySide.QtGui.QWidget isn't a direct base class of Panda | 05:53 |
| rm_you | hrm | 05:53 |
| rm_you | ah Panda is defined here... | 05:53 |
| rm_you | class Panda(QtGui.QMainWindow): | 05:54 |
| rm_you | so somehow that doesn't work the same way | 05:54 |
| *** Scifig has quit IRC | 05:54 | |
| rm_you | QtGui.QWidget.__init__(self, parent) | 05:54 |
| rm_you | that's what you were talking about isn't it | 05:54 |
| rm_you | init is more strict about super calls | 05:54 |
| epage | Yup, that is bad code, PyQt should have never allowed it | 05:54 |
| epage | Yes | 05:54 |
| rm_you | hrm | 05:55 |
| epage | Its an easy change | 05:55 |
| rm_you | ah | 05:55 |
| rm_you | mind telling me what it is? :P | 05:56 |
| * rm_you hopes you're already typing | 05:56 | |
| * rm_you doesn't actually see what's wrong with this | 05:56 | |
| epage | QtGui.QMainWindow.__init__(self, parent) instead | 05:57 |
| * DocScrutinizer ponders if such editing couldn't get automatized | 05:57 | |
| epage | Imagine the code was pure-python rather than calling into PySide | 05:57 |
| DocScrutinizer | didn't really pan out for qt3->qt4, but maybe this one is simpler | 05:57 |
| epage | If C inherits from B inherits from A | 05:58 |
| epage | If C calls A.__init__(self) then B's __init__ is never called and never sets up its member variables | 05:58 |
| rm_you | k | 05:59 |
| rm_you | DocScrutinizer: i think possibly | 05:59 |
| rm_you | i did it all with SED | 05:59 |
| rm_you | so | 05:59 |
| rm_you | :P | 05:59 |
| *** npm_ has quit IRC | 05:59 | |
| DocScrutinizer | you know I should hate harmattan, if only for still using messybox - your mentioning sed just made me recall my last experience with messybox sed not supporting the carefully crafted minimally more complex than bare basics editing commands | 06:03 |
| *** npm_ has joined #harmattan | 06:03 | |
| rm_you | libmafw.so.0 is apparently not current | 06:05 |
| rm_you | wonder what the lib is now | 06:05 |
| DocScrutinizer | meanwhile I think there's not one single command in busybox that's *not* broken and missing capabilities of the original | 06:06 |
| rm_you | errrrm | 06:06 |
| rm_you | do we not use libmafw now? | 06:06 |
| rm_you | ah | 06:09 |
| rm_you | libqmafw | 06:09 |
| DocScrutinizer | *cough* | 06:09 |
| * DocScrutinizer is no friend of the whole mafw concept | 06:10 | |
| DocScrutinizer | mafw is to blame for the annoyance that we got not a single gapfree mp3player on fremantle | 06:11 |
| rm_you | T_T | 06:11 |
| rm_you | how do I get all the symbols available in a lib | 06:12 |
| rm_you | self.__mafw = ctypes.CDLL("libqmafw.so.0") | 06:14 |
| rm_you | registry_p = self.__mafw.mafw_registry_get_instance() | 06:14 |
| rm_you | that explodes | 06:14 |
| rm_you | obviously not ABI compatible | 06:15 |
| rm_you | sooo i need to see what functions the new one DOES provide | 06:15 |
| GAN900 | Strange newbish question for you: What's involved with porting a Qt application to Components? | 06:22 |
| GAN900 | Or less specifically, is that a feasible approach to anything? | 06:22 |
| epage | If you have a good view/model separation it'll work well | 06:24 |
| epage | I'll be porting or rewriting all of my applications to use QML from QWidget and GTK | 06:25 |
| rm_you | anyone know how to get at MAFW from python? | 06:26 |
| GAN900 | epage, how's it stack up against GTK -> Hildon? | 06:31 |
| epage | Hildon was modifications or additions to GTK, QML is completely different | 06:31 |
| epage | I'm still learning it. it has some rough spots when first learning but fixes most of the annoyances I had with QWidget | 06:32 |
| epage | I'm heading off though, goodnight | 06:32 |
| *** hardaker has joined #harmattan | 06:55 | |
| *** antman8969 has joined #harmattan | 07:02 | |
| *** epage has quit IRC | 07:11 | |
| *** eman has joined #harmattan | 07:21 | |
| *** DocScrutinizer has quit IRC | 07:33 | |
| *** DocScrutinizer has joined #harmattan | 07:33 | |
| rm_you | GAAAH | 07:56 |
| rm_you | so close | 07:56 |
| rm_you | need to get MAFW working | 07:56 |
| hiemanshu | gah, how to import 700 .vcf files to the contacts app? | 07:59 |
| antman8969 | you could import them into your n900 and then use bluetooth to get them into the n950 :) | 08:08 |
| hiemanshu | well, the n900 has them in sync with my computer (storing them as .vcf files) | 08:09 |
| antman8969 | theres a feature in the n950 that connects to your n900 and imports with bluetooth | 08:10 |
| antman8969 | but idk why they don't have an option to import files directly | 08:10 |
| antman8969 | but whatever | 08:10 |
| hiemanshu | well I used the instructions posted by Jaffa, worked just fine | 08:15 |
| antman8969 | rm_you, did you get the obs working at all? | 08:33 |
| rm_you | no, refocused | 08:33 |
| rm_you | got mplayer to build finally anyway | 08:34 |
| antman8969 | on scratchbox? | 08:34 |
| rm_you | yep | 08:34 |
| antman8969 | got it to install on device too? | 08:34 |
| rm_you | yes | 08:34 |
| rm_you | it works | 08:34 |
| antman8969 | nice | 08:34 |
| antman8969 | you should host the deb somewhere | 08:35 |
| rm_you | http://www.sheeplauncher.net/debs/ | 08:35 |
| antman8969 | that works | 08:35 |
| rm_you | hopefully will have working Pandora client soon | 08:39 |
| *** eman has quit IRC | 08:40 | |
| rm_you | this would be so much easier if i could just modify the python code on the device without packaging/reinstalling a deb every time | 08:44 |
| *** hardaker has quit IRC | 08:48 | |
| rm_you | YEEEEEEAAAAH | 08:48 |
| rm_you | no volume control yet but fuck it | 08:48 |
| rm_you | antman8969: you use pandora? | 08:51 |
| rm_you | need a tester | 08:51 |
| *** TSCHAKeee has quit IRC | 09:13 | |
| tomma | achipa, what should be the difference with harmattan-meego-arm and harmattan-nokia-meego sysroots in Madde ? | 09:19 |
| tomma | other has support for boostable config, other has headers for shareuiinterface but no lib, other has lib for it but no headers... | 09:24 |
| *** mtd_ has joined #harmattan | 09:25 | |
| antman8969 | sry was afk | 09:25 |
| antman8969 | you got it built? | 09:26 |
| antman8969 | rm_you | 09:26 |
| mtd_ | hi guys, I've just got my N950 and been reading some interesting backlog of this channel, and I was wondering if there was a wiki site or place to collect tips about how to install bash and other useful hacking tools...? | 09:26 |
| *** mtd_ is now known as mtd | 09:26 | |
| * mtd so far has found http://wiki.meego.com/Migrating_from_N900_to_N950 | 09:28 | |
| antman8969 | lol thers a lot to say. Without having to say everything... you should learn about the Meego OBS build service, and http://wiki.meego.com/Build_Infrastructure/Community_Builder/ITP | 09:29 |
| antman8969 | is the libs ppl have built so far | 09:29 |
| antman8969 | you know how to get root access? | 09:29 |
| mtd | antman8969: thanks for the OBS tip | 09:30 |
| mtd | antman8969: just ran devel-su to change the root password :) | 09:31 |
| *** TSCHAKeee has joined #harmattan | 09:31 | |
| mtd | antman8969: do you know if anyone has documented their workflow / development process on the N950 so I can learn about it? | 09:32 |
| mtd | antman8969: ISTR jabba doing it, will have to find the link. | 09:32 |
| antman8969 | I know I havn't been.... Just been in the IRC the entire time | 09:32 |
| antman8969 | if you run into anything feel free to ask of course | 09:32 |
| mtd | antman8969: heh ok, will lurk here (always a good strategy) | 09:33 |
| mtd | antman8969: thanks | 09:33 |
| antman8969 | you should expect trouble with dealing with libraries, aegis and obs | 09:33 |
| antman8969 | yep | 09:33 |
| mtd | antman8969: good to know -- will read up on them | 09:33 |
| * mtd can't browse wiki.meego.com pages from N950 because "SSL handshake failed"?! | 09:35 | |
| antman8969 | lol hmm | 09:36 |
| mtd | err, log in, sorry | 09:36 |
| mtd | brosing is ok, but of course I'd like to be able to eidt pages form the n950... | 09:36 |
| mtd | will see what I can do | 09:36 |
| * mtd bet it's meego.com's GoDaddy CA | 09:37 | |
| antman8969 | was there something in specific that you were trying to find otuj? | 09:37 |
| mtd | is there a way to run apps from the command line so one can see stderr spew? | 09:37 |
| mtd | antman8969: nope, just saving the migrating from n900 -> n950 page so I could browse it while I head out | 09:38 |
| antman8969 | a qt app? | 09:38 |
| antman8969 | out of curiosity | 09:38 |
| mtd | antman8969: in general; what got me thinking about it was the browser | 09:38 |
| *** Sicelo has quit IRC | 09:38 | |
| mtd | antman8969: its error message was not helpful | 09:38 |
| *** Sicelo has joined #harmattan | 09:39 | |
| antman8969 | well just like anythingelse really | 09:39 |
| antman8969 | just the path to the binary should do | 09:39 |
| antman8969 | assuming you are user, not root | 09:39 |
| antman8969 | like /opt/app/bin/app | 09:39 |
| mtd | antman8969: ah, ok, nothing like N900 run_standalone.sh or anything | 09:39 |
| antman8969 | well the run_standalone.sh was only if you needed the theme info and you were root | 09:40 |
| antman8969 | there is an invoker | 09:40 |
| mtd | antman8969: oh, I didn't know that...thanks. | 09:40 |
| antman8969 | let me get you the invoker link | 09:40 |
| mtd | antman8969: ah ok, that I can google :) | 09:40 |
| mtd | antman8969: oh sure, link would be extra nice :) | 09:40 |
| antman8969 | http://harmattan-dev.nokia.com/unstable/beta/api_refs/xml/daily-docs/applauncherd/qmlboost.html | 09:41 |
| antman8969 | harmattan likes to launch apps using the invoker | 09:41 |
| antman8969 | loads faster | 09:41 |
| antman8969 | of course, you don't HAVE to | 09:41 |
| antman8969 | and if you do then the only place your exec will work is harmattan lol | 09:41 |
| antman8969 | not sure if meego uses the invoker | 09:41 |
| *** npm_ has quit IRC | 09:42 | |
| mtd | antman8969: hehe ok | 09:42 |
| antman8969 | rm_you gj with mplayer... uplaoding pandora? | 09:42 |
| *** npm has joined #harmattan | 09:42 | |
| mtd | antman8969: I'll root around in the fs for a while, then hopefully can ask better questions (I have nothing in /opt, for example) | 09:42 |
| mtd | antman8969: thanks again | 09:43 |
| antman8969 | np, I think the best way to learn is to just start developing | 09:43 |
| antman8969 | solve problems as you run into them | 09:43 |
| antman8969 | not tyring to pre solve all problems lol | 09:44 |
| antman8969 | the goal is to produce applications imo | 09:44 |
| mtd | antman8969: yup - though I find messing with the device is a good way to find out interesting corners | 09:46 |
| mtd | antman8969: point taken about apps though | 09:46 |
| mtd | antman8969: should be fun | 09:47 |
| antman8969 | gl | 09:47 |
| *** antman8969_ has joined #harmattan | 10:00 | |
| *** antman8969_ has quit IRC | 10:00 | |
| *** antman8969_ has joined #harmattan | 10:00 | |
| *** antman8969 has quit IRC | 10:01 | |
| *** antman8969_ has quit IRC | 10:01 | |
| *** antman8969 has joined #harmattan | 10:02 | |
| hiemanshu | antman8969: AFAIK meego-ce uses invoker too | 10:06 |
| antman8969 | do you know if it uses it in a different way at all? | 10:08 |
| antman8969 | i was trying to build my harmattan app for meego 1.2 and it could not find MApplicationWindow.h | 10:09 |
| antman8969 | which is what harmattan was using | 10:09 |
| antman8969 | or maybe mdeclarativecache | 10:09 |
| antman8969 | either way, coudln't find the right header on meego | 10:09 |
| antman8969 | rm_you harmattan works great, using over wifi right now | 10:14 |
| antman8969 | n950 speakers are kinda crappy | 10:14 |
| hiemanshu | antman8969: you need libmeegotouch and such installed for that | 10:22 |
| *** rcg has joined #harmattan | 10:23 | |
| hiemanshu | http://meego.gitorious.org/meegotouch/libmeegotouch | 10:23 |
| hiemanshu | antman8969: ^ | 10:23 |
| antman8969 | wouldn't the meego targets on obs have that installed? | 10:23 |
| hiemanshu | you might have to include it in build required | 10:24 |
| hiemanshu | s/d/s/ | 10:24 |
| infobot | hiemanshu meant: you might have to incluse it in build required | 10:24 |
| antman8969 | did lol, in the spec file as pckgconfig(QtDeclarative) | 10:24 |
| hiemanshu | errr, lol | 10:24 |
| antman8969 | I assumed thats where it would be | 10:24 |
| antman8969 | lol @ infobot | 10:25 |
| hiemanshu | antman8969: http://meego.gitorious.org/meegotouch/libmeegotouch/blobs/master/src/include/MApplicationWindow | 10:26 |
| antman8969 | #include <MDeclarativeCache> | 10:28 |
| antman8969 | http://harmattan-dev.nokia.com/unstable/beta/api_refs/xml/daily-docs/applauncherd/qmlboost.html | 10:29 |
| antman8969 | thats the class that this nokia tutorial recommends | 10:29 |
| antman8969 | looking for that now.. maybe its just a convenience class | 10:29 |
| *** smoku has joined #harmattan | 10:58 | |
| *** antman8969 has quit IRC | 11:24 | |
| rm_you | got mplayer and pyRadio (pandora client) compiled for Harmattan now | 12:06 |
| rm_you | woot :P | 12:06 |
| rm_you | pyRadio is bugged right now, it doesn't automatically advance to the next track -- i know exactly why, just haven't gotten to fixing that part yet. tomorrow. | 12:07 |
| rm_you | for now you can always just advance it manually :P | 12:07 |
| dm8tbr | stuff is on COBS? | 12:16 |
| *** smoku has left #harmattan | 12:21 | |
| *** trx has quit IRC | 12:22 | |
| *** CaCO3 has joined #harmattan | 12:23 | |
| *** smoku has joined #harmattan | 12:24 | |
| *** trx has joined #harmattan | 12:26 | |
| rm_you | mine is in my own repo :/ | 12:35 |
| rm_you | can't get OBS to work right | 12:35 |
| rm_you | just using SB right now | 12:35 |
| Damion3 | rm_you: you build mplayer to use the hw assisting libs of the n950 ? | 12:35 |
| rm_you | no, can't | 12:35 |
| rm_you | but it runs about as well anyway | 12:35 |
| rm_you | needed mplayer to get pyRadio working | 12:36 |
| Damion3 | from what it shows cpuwise? | 12:36 |
| rm_you | erm | 12:36 |
| rm_you | i did comparisons earlier but it seemed like | 12:36 |
| Damion3 | need a battery load test | 12:36 |
| rm_you | video-suite (builtin) was ~20 CPU and pulse-audio was ~15 CPU | 12:36 |
| rm_you | mplayer was ~30% CPU and pulse-audio was ~9% CPU | 12:37 |
| rm_you | so slightly more total | 12:37 |
| Damion3 | n the same video? | 12:37 |
| rm_you | yes | 12:37 |
| Damion3 | okay | 12:37 |
| rm_you | mplayer handles the audio better somehow | 12:37 |
| rm_you | but the video worse (no hardware) | 12:37 |
| rm_you | that was my conclusion | 12:37 |
| rm_you | i would love someone to test that could do better data collection | 12:38 |
| Damion3 | it's odd it's only 10% | 12:38 |
| Damion3 | almost like the video you picked wasn't using accel anyway | 12:38 |
| Damion3 | 10% difference that is | 12:38 |
| rm_you | hrm | 12:38 |
| rm_you | well | 12:38 |
| rm_you | i tested on other videos | 12:39 |
| rm_you | but most of my other videos get serious lag on both | 12:39 |
| rm_you | so i didn't really consider those valid tests | 12:39 |
| Damion3 | so with your mplayer the n950 can how play significantly more formats that before | 12:39 |
| Damion3 | as for just copying binaries over to test, you can use develsh | 12:40 |
| *** leinir has joined #harmattan | 12:41 | |
| rm_you | yeah | 12:43 |
| rm_you | ok i lied, i am fixing the bug in pyRadio today | 12:44 |
| rm_you | i just realized how this works :P | 12:44 |
| rm_you | i had to write a new backend for it | 12:44 |
| rm_you | using mplayer | 12:44 |
| rm_you | :) | 12:44 |
| rm_you | since MAFW is &*#@^ed | 12:44 |
| rm_you | i will go fix MAFW later | 12:45 |
| rm_you | once someone figures out WTF | 12:45 |
| rm_you | and maybe makes a python lib for it | 12:45 |
| rm_you | anyway if you want to test | 12:46 |
| rm_you | http://www.sheeplauncher.net/debs/ | 12:46 |
| rm_you | pandora requires mplayer | 12:46 |
| Tronic | The telephone app only displays black. Is anyone else seeing this bug? | 12:48 |
| Tronic | Reboot fixed it the first time it happened but now it is broken again. | 12:49 |
| Tronic | Kind of annoying because you cannot hangup or do anything else. Even after the app is closed the call won't terminate. | 12:52 |
| *** NIN101 has joined #harmattan | 12:54 | |
| Damion3 | sounds like the proximity detection is wrong | 13:01 |
| Damion3 | earwax gunk over the sensor? | 13:02 |
| *** razvaway has joined #harmattan | 13:02 | |
| *** razvaway has left #harmattan | 13:02 | |
| Damion3 | 13:03 | |
| MohammadAG | <rm_you> i will go fix MAFW later | 13:05 |
| MohammadAG | it's closed in harmattan | 13:05 |
| rm_you | right | 13:05 |
| rm_you | it can be closed | 13:05 |
| MohammadAG | it is | 13:05 |
| rm_you | i can still see the symbols it provides and hook them with python-ctypes :P | 13:05 |
| rm_you | it being closed just makes it more guesswork | 13:06 |
| MohammadAG | you can use fremantle's python bindings | 13:06 |
| rm_you | nope | 13:06 |
| MohammadAG | yes | 13:06 |
| rm_you | completely rewritten lib | 13:06 |
| MohammadAG | same DBus APIs | 13:06 |
| rm_you | didn't use DBUS? | 13:06 |
| MohammadAG | mafw? always did | 13:06 |
| rm_you | this one hooked libmafw.so.0 with ctypes | 13:06 |
| rm_you | O_o | 13:06 |
| MohammadAG | then port libmafw | 13:06 |
| rm_you | lol | 13:06 |
| rm_you | srsly? | 13:07 |
| MohammadAG | it's just a dbus wrapper | 13:07 |
| MohammadAG | yes | 13:07 |
| rm_you | hrm | 13:07 |
| MohammadAG | libmafw0 is a C wrapper API | 13:07 |
| MohammadAG | in the end, it calls DBus | 13:07 |
| MohammadAG | MafwGstRenderer is on Harmattan, I guess it's slightly improved | 13:08 |
| rm_you | where's the mafw source? | 13:08 |
| MohammadAG | fremantle? apt-get source, harmattan? Nokia | 13:09 |
| rm_you | lol | 13:09 |
| Damion3 | does anyone else find the wifi flakey? I have multiple AP with same SSIDs/PSK as my house has many walls. When it roams over, only small packets continue to work but browsing breaks and ssh hangs after a moment unless I detach and reattach. | 13:09 |
| MohammadAG | I'd guess libmafw0 should build fine on Harmattan's SDK | 13:09 |
| rm_you | kk trying it | 13:10 |
| Damion3 | it's definitely accocating to the new/near AP | 13:10 |
| Tronic | Damion3: No dirt on the sensor but maybe it misses an event. I'll try playing with the sensor the next time that happens. | 13:10 |
| MohammadAG | I hope my N950 ships tomorrow :/ | 13:11 |
| Damion3 | Tronic: does using the power button turn the screen on? | 13:11 |
| rm_you | err how do i get root in scratchbox | 13:12 |
| Tronic | THe screen is on, only the app is black. | 13:12 |
| Tronic | I can taskswitch etc. normally. | 13:12 |
| *** gri has joined #harmattan | 13:13 | |
| MohammadAG | rm_you, fakeroot | 13:13 |
| MohammadAG | fakeroot dpkg-buildpackage -b | 13:13 |
| rm_you | umm yes | 13:13 |
| rm_you | but i need ACTUAL root | 13:13 |
| rm_you | not for building packages | 13:13 |
| Damion3 | oh, that's not proximity, as that turns off the screen completely | 13:13 |
| MohammadAG | you don't then :P | 13:13 |
| rm_you | FFFFFFFFFFF | 13:14 |
| rm_you | trying to apt-get install something | 13:14 |
| rm_you | and... | 13:14 |
| rm_you | Setting up xml-core (0.13+maemo11+0m6) ... | 13:14 |
| MohammadAG | why do you need actual root? | 13:14 |
| rm_you | /scratchbox/tools/bin/chown: changing ownership of `/usr/local/share/xml/misc': Operation not permitted | 13:14 |
| MohammadAG | used fakeroot? | 13:14 |
| rm_you | no? why would fakeroot help here? | 13:14 |
| MohammadAG | fakeroot apt-get install xml-core | 13:14 |
| rm_you | wow srsly?! | 13:15 |
| rm_you | why would fakeroot help | 13:15 |
| rm_you | it does | 13:15 |
| rm_you | but WHY | 13:15 |
| MohammadAG | Because. | 13:15 |
| rm_you | it's not a "pretend" operation | 13:15 |
| rm_you | it's a real operation | 13:15 |
| MohammadAG | old versions of scratchbox didn't need it | 13:16 |
| MohammadAG | you could just apt-get install something | 13:16 |
| MohammadAG | Hathor(?) needs fakeroot | 13:16 |
| rm_you | T_T | 13:17 |
| gri | Is there a command which lets me exeute something in scratchbox from outside? I remember "sb <command>" from somewhere, but it does not exist | 13:17 |
| rm_you | well anyway, looks good except it's dying on some weird error now | 13:17 |
| MohammadAG | gri, that would be madde | 13:18 |
| rm_you | autoreconf: automake failed with exit status: 1 | 13:18 |
| MohammadAG | I doubt scratchbox has something that does the same | 13:18 |
| gri | Damn | 13:18 |
| MohammadAG | why? there should be an error | 13:18 |
| rm_you | no other useful info presented | 13:18 |
| gri | I run Qt Creator from inside the scratchbox and I hate to enter the commandlines every time ... a desktop shortcut would be much better | 13:18 |
| rm_you | MohammadAG: http://pastie.org/2262966 | 13:19 |
| rm_you | strange | 13:22 |
| rm_you | thought maybe it had to do with it trying to build gtk-docs but disabled those and still dying | 13:23 |
| rm_you | :( | 13:23 |
| rm_you | got it | 13:30 |
| rm_you | some docs stuff hardcoded in | 13:30 |
| rm_you | i think | 13:30 |
| rm_you | nope still dying | 13:31 |
| rm_you | ah well it compiled | 13:39 |
| rm_you | but says it's an invalid installation package | 13:39 |
| *** lbt_away is now known as lbt | 13:44 | |
| rm_you | ok got libmafw.so installed | 13:51 |
| rm_you | but where does libmafw-shared.so come from!? | 13:51 |
| rm_you | T_T | 13:51 |
| rm_you | found it | 13:54 |
| * rm_you sighs | 13:54 | |
| rm_you | wish you'd been around before i wasted like 3 hours writing an mplayer backend from scratch T_T | 13:58 |
| rm_you | lolno | 14:05 |
| rm_you | now it segfaults :P | 14:05 |
| RST38h | Yoooohoooo | 14:08 |
| RST38h | N950 is here | 14:08 |
| RST38h | H.A.P.P.I.N.E.S.S | 14:11 |
| *** deimos has joined #harmattan | 14:12 | |
| RST38h | SDK! SDK! | 14:17 |
| * RST38h restrains himself | 14:17 | |
| thp | RST38h: :p | 14:25 |
| thp | RST38h: so will we see tasty emulators for harmattan soon? :) | 14:25 |
| RST38h | thp:Definitely | 14:26 |
| RST38h | thp:theonly thing that pains me is that all this beauty is about to be squandered... or not... | 14:28 |
| *** spenap has joined #harmattan | 14:30 | |
| *** smoku has joined #harmattan | 14:31 | |
| *** smoku has quit IRC | 14:31 | |
| DocScrutinizer | meh | 15:06 |
| RST38h | Does this let me disable Aegis? http://pastebin.com/C5jbvj7N | 15:12 |
| gri | DocScrutinizer: Do you know if aegis protects .desktop files? | 15:21 |
| DocScrutinizer | not for sure, but I'd dare betting money against it | 15:21 |
| thp | gri: aegis protects only files for which it has hashes | 15:22 |
| DocScrutinizer | RST38h: you CAN NOT disable aegis, our devices miss the R&D cert | 15:22 |
| DocScrutinizer | ...is what $Nokia said | 15:23 |
| gri | thp: So it has no hashes for .desktop files? | 15:24 |
| gri | or does it depend? | 15:24 |
| DocScrutinizer | the question is: what is the official way to find out about it? | 15:24 |
| DocScrutinizer | probably not to edit a file and watch it go "Yor Device Got Problems, please try to reflash" :-S | 15:26 |
| RST38h | Doc: Will they include the R&D certificate the next time? | 15:26 |
| DocScrutinizer | Nokia, by implementing and arming Aegis the way they did, has created a playground fu | 15:27 |
| DocScrutinizer | full off APM mines for us | 15:27 |
| RST38h | There was no other way :) | 15:27 |
| RST38h | Have you ever seen a DRM "solution" that would not be either 1) ineffective or 2) a fuckup or 1+2? | 15:28 |
| DocScrutinizer | fine, so when you win a holiday in Afghanistan in the mine fields, you cheer and pack your things? | 15:28 |
| RST38h | I run for the hills. | 15:29 |
| DocScrutinizer | I honestly tried NOT to start and rant about fsckdup Aegis/TC/MSSF for at least 3h after getting up - seems I didn't make it | 15:29 |
| RST38h | But the art of this is how to make it to the end of your life without winning these =) | 15:29 |
| RST38h | BTW, Doc, are there already some repos with the third party software that I can add to N950? | 15:30 |
| DocScrutinizer | RST38h: you just won and accepted a holiday in the minefields - it's called N950&Aegis and AIUI sits on your desk | 15:30 |
| DocScrutinizer | RST38h: check out this: http://ageofikon.com/prh/index.php?Action=list&System=./&Arg=mc&Section=&Repo=0 | 15:31 |
| DocScrutinizer | there's rzr's repo, plus a few others. There's NO central repo for random developer untested rocks yet | 15:33 |
| DocScrutinizer | packrat gets a lot of them though | 15:33 |
| DocScrutinizer | obviously not for apt-get but hey :-D | 15:33 |
| RST38h | Ah, rm_me, good chap, added the Harmattan repos to the Packrat =)) | 15:34 |
| DocScrutinizer | btw I seem to have read a mail that suggests today won't work anything | 15:34 |
| RST38h | Given that N950 seemingly misses App Manager,how do I add repos? Directly to sources.list? | 15:34 |
| dm8tbr | there is a directory like sources.list.d or sthg like that | 15:35 |
| DocScrutinizer | sth like that | 15:35 |
| RST38h | ok | 15:35 |
| macmaN | btw | 15:35 |
| macmaN | has anyone tried copying authorized_keys into ~user/.ssh? | 15:36 |
| RST38h | freaking shit everyone has got his ownrepo now | 15:36 |
| RST38h | exactly what admins tried to avoid with maemo5 | 15:36 |
| macmaN | you are in for a surprise with that one, if you havent | 15:36 |
| RST38h | how touching... | 15:36 |
| * DocScrutinizer realizes he managed since >11years to not touch proprietary or evil crap like $redmond or TC, and now has a device with Aegis on his desk AT HOME!!1!! :-o (goes throwing up) | 15:37 | |
| DocScrutinizer | macmaN: I'm doing that all the time | 15:38 |
| *** M4rtinK2 has joined #harmattan | 15:38 | |
| DocScrutinizer | to /root though | 15:38 |
| dm8tbr | RST38h: premature conclusions :) | 15:39 |
| DocScrutinizer | macmaN: what's the surprise with $USERHOME ? | 15:39 |
| DocScrutinizer | aegis another time :-o | 15:39 |
| dm8tbr | RST38h: there will be a repository, alike extras where those community packages will be promoted to. at least that's my understanding how the whole story of apps.meego.com looks | 15:40 |
| DocScrutinizer | dm8tbr: damn, we need extras-devel_just_worse - for NOW | 15:40 |
| DocScrutinizer | then otoh installing from an untrusted source to find you can't execute doesn't make much sense either | 15:41 |
| dm8tbr | DocScrutinizer: talk to lbt, step up to help with this | 15:41 |
| DocScrutinizer | dm8tbr: I'm the one certified idiot about repos and packaging here | 15:42 |
| DocScrutinizer | you couldn't find anybody less qualified | 15:42 |
| *** lynxis has joined #harmattan | 15:42 | |
| dm8tbr | DocScrutinizer: the story is fairly new to many people, still it needs volunteers | 15:43 |
| DocScrutinizer | I'm just qualified to ask the NOOB questions, and whine about how messed up all this is | 15:43 |
| DocScrutinizer | and I'm having a hard time supporting THIS particular mix of linux and TC | 15:45 |
| DocScrutinizer | I'm still waiting what happens, but I'm definitely closer to sending back the device than to investing much effort to help Nokia make Aegis work | 15:46 |
| dm8tbr | I don't think whining is the constructive thing to do. | 15:47 |
| DocScrutinizer | the most radical of my FOSS fighter friends already frown at me | 15:47 |
| DocScrutinizer | dm8tbr: there's no other constructive thing to do with TC than to refuse to touch the whole thing | 15:48 |
| DocScrutinizer | each sound advise about TC is "don't buy it, and tell all the people you can reach to also not buy it" | 15:49 |
| RST38h | What is TC? | 15:50 |
| DocScrutinizer | Nokia promised there was a way to switch of aegis. There's *not* | 15:50 |
| *** harbaum has joined #harmattan | 15:50 | |
| DocScrutinizer | ~aegis | 15:50 |
| infobot | well, aegis is http://www.developer.nokia.com/Community/Wiki/Harmattan:Developer_Library/Developing_for_Harmattan/Harmattan_security/Security_guide , or "The purpose of this framework is: ... to make sure that the platform meets the requirements set by third party software that requires a safe execution environment.", or http://en.wikipedia.org/wiki/Trusted_Computing#Criticism | 15:50 |
| RST38h | ah | 15:50 |
| RST38h | not nokia, its engineers promised | 15:50 |
| RST38h | engineers do not decide things at nokia | 15:50 |
| dm8tbr | DocScrutinizer: *sigh*, if you didn't realize, devloper mode enables you to declare aegis capabilities yourself for almost every possible capability there is | 15:51 |
| DocScrutinizer | so I'm really sorry about the engineers at Nokia and the brilliant work they did with harmattan, but as it's now I'm not temped to help make Aegis a success story so it seems I can't help with harmattan either | 15:52 |
| DocScrutinizer | dm8tbr: show me how to declare capabilities for http://repo.pub.meego.com/home:/djszapi/MeeGo_1.2_Harmattan/armel/mc_4.7.0.9-2_armel.deb to make mc run under root and I'm all yours | 15:53 |
| DocScrutinizer | don't tell me "you `only` have to compile it in your local SDK" | 15:54 |
| DocScrutinizer | show me how to set capabilities so I can edit arbitrary files on device without ending at "Reflash please" | 15:56 |
| *** hardaker has joined #harmattan | 15:56 | |
| dm8tbr | DocScrutinizer: what's wrong with that package as it is? | 15:56 |
| DocScrutinizer | how to do friggin normal thinks like kill -9 $random-pid | 15:56 |
| DocScrutinizer | dm8tbr: aiui it has no manifest | 15:56 |
| DocScrutinizer | things* | 15:57 |
| dm8tbr | then it should end up with default capabilities that shouldn't be too restrictive. if it needs aditional capabilities then it needs a manifest | 15:57 |
| DocScrutinizer | dm8tbr: the default capabilities *ARE WAY TOO* restrictive | 15:58 |
| dm8tbr | DocScrutinizer: show me please | 15:58 |
| DocScrutinizer | in developer mode the default should be: no restrictions at all | 15:58 |
| DocScrutinizer | dm8tbr: install that mc.deb, try to run as root | 15:58 |
| DocScrutinizer | simple as that | 15:58 |
| dm8tbr | DocScrutinizer: besides, you are tyring to shoehorn the 'default unix/linux system with no security at all besides UID' way onto this device, this does NOT work. period. | 15:59 |
| DocScrutinizer | then this device is crap. period. | 15:59 |
| dm8tbr | this is not your average linux box as you end up with after flipping in a ubuntu CD | 15:59 |
| dm8tbr | please stop trolling, ktx | 15:59 |
| DocScrutinizer | I'm not trolling | 16:00 |
| DocScrutinizer | this system is restricted to a level where even the so called developer mode doesn't allow BASIC things to be done - not even by and esp not by root - that are working on every other sane system | 16:01 |
| dm8tbr | it's simple: if you are a jackass and don't even try to understand that there is an additional set of restrictions that do not necessarily pose a problem to you if you understand them and work with them, then I don't see why I should even try and help/feed you. | 16:02 |
| dm8tbr | abandon the notion of root | 16:02 |
| dm8tbr | abandon what you know about system security on linux, this is different and it's not as restrictive as you think | 16:02 |
| thp | DocScrutinizer: actually you can install normal .debs and run them - as normal user "user" | 16:03 |
| thp | (i.e. install them as "root", run them as "user" | 16:03 |
| dm8tbr | trying to kick down the door doesn't work when it's reinforced. but hey why would you, someone didn't lock it at all | 16:03 |
| *** harbaum has quit IRC | 16:03 | |
| dm8tbr | so instead of trying to be joe-root-rambo try to work with it | 16:04 |
| DocScrutinizer | damn, I don't care about semantics. Give me permissions to run arbitrary binaries, generated by dd if=/dev/random, and I'm happily taking the responsibility of the result. Tell me what friggin switches I need to throw to enable this. I fthere is not way to enable this (and it obviously IS like this) then goodbye harmattan | 16:04 |
| Elleo | DocScrutinizer: --relaxed-exec should allow you to run arbitrary binaries (as the devel user, not as root) | 16:05 |
| DocScrutinizer | as metioned above there IS a way to disable aegis, just it's locked on our developer(!) devices | 16:05 |
| *** CaCO3 has quit IRC | 16:05 | |
| Elleo | DocScrutinizer: in the develsh use aegis-developer-mode --relaxed-exec | 16:05 |
| DocScrutinizer | Elleo: yes | 16:05 |
| DocScrutinizer | Elleo: exactly | 16:05 |
| DocScrutinizer | and running mc as user, user permissions apply | 16:06 |
| dm8tbr | DocScrutinizer: in all honesty, I suggest you send back your device to Nokia, someone else would make much better use of it. | 16:06 |
| macmaN | DocScrutinizer: lol you tried it yes? total facepalm moment there | 16:06 |
| DocScrutinizer | dm8tbr: from a nokia perspective your statement for sure is 100% true | 16:07 |
| DocScrutinizer | macmaN: I tried what? edit a 644 file as user? you're not suggesting this would work as soon as I run mc as user? | 16:09 |
| macmaN | no | 16:10 |
| DocScrutinizer | if I tried to run mc as root? sure I did | 16:10 |
| dm8tbr | DocScrutinizer: I'm saying this from a community perspective. All you are ever doing is yacking and being a jackass about aegis, this has a negative impact on the atmosphere here and on other people. In contrast there are obviously people that manage to get things working and develop successfully for the device. | 16:10 |
| macmaN | devel-su; cp /root/.ssh/authorized_keys ~user/.ssh | 16:10 |
| DocScrutinizer | macmaN: aah that, no I didn't | 16:10 |
| macmaN | maybe you could try that for verification? | 16:11 |
| DocScrutinizer | macmaN: I asked what's the result | 16:11 |
| macmaN | ahh | 16:11 |
| DocScrutinizer | I'm not eager to do a reflash | 16:11 |
| macmaN | nah, nothing that serious | 16:11 |
| DocScrutinizer | k | 16:11 |
| macmaN | permission denied or operation not permitted, dont remember | 16:11 |
| macmaN | about to ssh in there though, so i can recheck | 16:11 |
| DocScrutinizer | macmaN: that's what I'm saying and get bashed for that by dm8tbr | 16:12 |
| DocScrutinizer | root is not root on this device | 16:12 |
| macmaN | ah right. | 16:12 |
| macmaN | ok i didnt read through the discussion with dm8tbr | 16:12 |
| DocScrutinizer | RM680-22-6_PR_RM680:~# cp /root/.ssh/authorized_keys ~user/.ssh/ | 16:14 |
| DocScrutinizer | cp: can't create '/home/user/.ssh/authorized_keys': Permission denied | 16:14 |
| DocScrutinizer | RM680-22-6_PR_RM680:~# ls -ld ~user/.ssh/ | 16:14 |
| DocScrutinizer | drwxr-xr-x 2 user users 4096 Jul 24 15:14 /home/user/.ssh/ | 16:14 |
| Tronic | DocScrutinizer: root cannot touch user's .ssh | 16:15 |
| dm8tbr | are you root or user? | 16:15 |
| DocScrutinizer | root | 16:15 |
| dm8tbr | that won't work | 16:15 |
| Tronic | DocScrutinizer: Use temporary file, then use user to install the file to .ssh | 16:15 |
| dm8tbr | 13:02:47< dm8tbr> abandon what you know about system security on linux, this is different and it's not as restrictive as you think | 16:15 |
| DocScrutinizer | sigh | 16:15 |
| Elleo | DocScrutinizer: I don't get a permission denied when running mc as root via devel-su (unlike when logged in as a normal root user) | 16:15 |
| DocScrutinizer | I'm off finding a DHL package | 16:15 |
| Elleo | although it still doesn't actually run for me since I don't have libslang2 | 16:16 |
| Tronic | Dunno which security system is preventing this (aegis, selinux, whatnot) but it is not UNIX permissions and being root doesn't help you there. | 16:16 |
| macmaN | Tronic: whatsup dawg :> | 16:16 |
| dm8tbr | approaching harmattan with the mindset of 'I'm root I can do everything' is the single most dangerous thing to do | 16:16 |
| Tronic | macmaN: I'm too low on caffeine but still trying to force myself to do some development. | 16:17 |
| Tronic | At the moment I am trying to get meego-terminal built. | 16:17 |
| macmaN | Tronic: dayjob or fun stuff :> | 16:17 |
| Tronic | "fun" | 16:17 |
| dm8tbr | meego terminal could use some url-magic and c&p magic :-/ | 16:18 |
| Tronic | Yes, and proper colors, getting rid of that theme systems and whatnot. | 16:20 |
| Tronic | I guess the development version already has some fixed beyond what comes with the phone. | 16:20 |
| macmaN | Tronic: oh youre saying user can modify ~user/.ssh eh | 16:21 |
| Tronic | macmaN: user can, root cannot. | 16:21 |
| macmaN | yep right. ok gonna go try | 16:21 |
| DocScrutinizer | btw it's NOT harmattan that is crippled. Harmattan is a proper linux system and just fine. It's harmattan+aegis+**the_control_nokia_grants_by_their_configuration_of_agies** that makes a I'm_root mindset dangerous | 16:23 |
| DocScrutinizer | and for nothing | 16:23 |
| DocScrutinizer | as we DO NOT NEED aegis in devel-mode (or open mode, or whatever it's called) | 16:23 |
| thp | DocScrutinizer: actually, aegis does fulfill a good purpose. protecting *your data* from applications that run as user "user" (i.e. apps that you install from the repos) | 16:24 |
| DocScrutinizer | for checking compatibility of our elaborates we got "normal mode" where I'm completely fine with aegis kicking my ass for trying to write to ~user as ~root | 16:24 |
| thp | although I'm not sure if this is still the case (but I assume apps would get permission denied when trying to access contact data, etc.) | 16:24 |
| DocScrutinizer | thp: on a DEVELOPER device I'm not supposed to have any valuable data | 16:25 |
| DocScrutinizer | If I had, than how BAD is aegis forcing a reflash every other day? | 16:26 |
| thp | i actually haven't run into the "aegis bricks the system" situation yet | 16:26 |
| thp | but then again, i don't edit files in installed packages on my ubuntu system either | 16:27 |
| DocScrutinizer | maybe you're not running self developed (python) scripts on your ubuntu either? | 16:28 |
| DocScrutinizer | or you simply don't need manifests to request tokens to make those self written python scripts run on ubuntu? | 16:28 |
| DocScrutinizer | aegis has definitely zero benefit for a developer during develpment. It's only helpful in final phase when testing for user system compliance | 16:29 |
| DocScrutinizer | and of course users (sic!) only developing apps that can run under user account completely, won't run into any aegis limitations usually | 16:31 |
| thp | DocScrutinizer: but then, i'd like to have my development environment match the final deployment environment as much as possible | 16:32 |
| Damion3 | actually I found I can now iwconfig essid BLAH as "user", which I'm pretty sure needs root on other systems. Is aegis helping here? | 16:32 |
| thp | otherwise i'd run into complications down the road when my assumptions don't hold anymore in an aegis-protected environment | 16:32 |
| DocScrutinizer | thp: then you're still free to switch to "normal mode" | 16:32 |
| Damion3 | also user can chown stuff as arbitrary users including stuff to root | 16:32 |
| thp | DocScrutinizer: switching modes is only possible via reflash afaik | 16:32 |
| DocScrutinizer | thp: switching modes basically impossible afaik | 16:33 |
| DocScrutinizer | it allegedly *is* possible if your device got R&D certificate | 16:33 |
| Damion3 | DocScrutinizer: what exactly happens when you echo 0 to the virtual file in the /proc/sys/security/something file ? | 16:33 |
| DocScrutinizer | I.E. Nokians *can* disable aegis | 16:34 |
| Damion3 | that works on mine and I don't know why I'd have a special nokia r&d version | 16:34 |
| thp | DocScrutinizer: then again, you can build .debs, make sure they are signed and request the right permissions and off you go | 16:34 |
| DocScrutinizer | thp: that would work for a self built cp, kill, mv ? | 16:35 |
| thp | i can't see a reason why it would not work | 16:35 |
| Damion3 | thp: that was what I assumed, it's only slightly more of a pain and helps you test what it would be like as a user rather than littering your system via ssh with stuff that has made it work that doesn't replicate on other phones | 16:35 |
| Damion3 | by turning on relaxed exec I can faff about a bit, but then go back to normal and try and sort a .deb properly. | 16:36 |
| thp | DocScrutinizer: that's also why i have little problems with my python stuff - /usr/bin/python is hashed, so i can run it and then as arguments any python script that i want | 16:36 |
| DocScrutinizer | Damion3: I'm not interested in Nokia teaching me to protect myself from own generic mess | 16:36 |
| thp | DocScrutinizer: then use the n900 or some other open device instead of the n950 | 16:37 |
| Damion3 | DocScrutinizer: personally I hate the faff of packaging and normally mess on-device, certainly did a lot of that on the n900 | 16:37 |
| DocScrutinizer | thp: I actually do | 16:37 |
| Damion3 | I got a long way, but it also means my stuff is balls and I've not bothered putting a Qt GUI on anything (or hildon before). Which means it's not useful for many people | 16:38 |
| thp | DocScrutinizer: great :) and as a side project, you could try to contact the right nokians and petition them to provide aegis-less images (as qole does) | 16:38 |
| DocScrutinizer | thp: I probably should ask konttori why he was to keen to give me a device. I was sceptic from beginning about it | 16:38 |
| Damion3 | mediatomb, EyeFi daemon, squid (for lazy non NAT adhoc tethering), qemu-user for fast x86 running. All of these are useless to users | 16:38 |
| DocScrutinizer | thp: that's a very rrasonable suggestion | 16:38 |
| DocScrutinizer | thanks | 16:38 |
| DocScrutinizer | (no sarcasm) | 16:39 |
| thp | ;) | 16:39 |
| Damion3 | thp: does the boot loader only boot signed images? | 16:39 |
| thp | Damion3: from what i've heard, yes | 16:39 |
| Damion3 | well that's a pain | 16:39 |
| thp | or rather, the device won't allow flashing of an unsigned image | 16:39 |
| DocScrutinizer | that's it | 16:40 |
| thp | (although i'm not sure if you could e.g. flash an unsigned rootfs, but not flash an unsigned kernel, etc..) | 16:40 |
| DocScrutinizer | Stskeeps looks into that, it would be the ultimate showstopper for meegoCE | 16:40 |
| thp | but again, there's nothing technically stopping nokia from providing a signed image that has aegis disabled | 16:40 |
| Damion3 | s/there's nothing/there's only management who control everything eng would like to do/ | 16:41 |
| DocScrutinizer | thp: nothing is technically stopping Nokia from giving us a R&D cert | 16:41 |
| DocScrutinizer | as R&D is what we do | 16:42 |
| DocScrutinizer | ;-D | 16:42 |
| thp | so next step to achieve your goal is to learn how to talk to management and come up with a list of advantages that giving R&D certs/signed images would have | 16:42 |
| DocScrutinizer | hmm, sounds like an unpleasant and daunting but necessary task, yes | 16:43 |
| thp | easier said than done, but still "give is R&D certs!" is not as good an argument as "i could do thing A, thing B and awesome thing C with that device and make it much cooler/attractive to buying customers if only I had some R&D certs" | 16:43 |
| thp | (with "R&D certs" being a placeholder for whatever it is that you really need to get your work done) | 16:44 |
| DocScrutinizer | alas I'm notorious to bitch even at a metric of FOSS community, mgmt never likes my statements, even when written in best available sense | 16:44 |
| Damion3 | thp: of my project plans, most are going to be quite a pain as things stand: http://forum.meego.com/showpost.php?p=26823&postcount=32 | 16:44 |
| DocScrutinizer | then OTOH Nokia has a long record of ignoring 100% of "user" requests like these, so I'm seriously convinced I'd waste my time | 16:46 |
| thp | Damion3: afaik tethering will be supported by joikuspot (not floss, but gratis) | 16:47 |
| Damion3 | DocScrutinizer: I've been told they'll listen to my account manager as they value the partnership my company has. I'll forumate an email in the next few days | 16:47 |
| DocScrutinizer | and AIUI Nokia not even has the inhouse competence/manpower to do anything about it anymore | 16:47 |
| thp | DocScrutinizer: probably yes | 16:47 |
| Damion3 | thp: it's AdHoc, so useless. I tried it | 16:47 |
| *** xerxes2 has quit IRC | 16:48 | |
| thp | Damion3: what's the problem with adhoc vs. infrastructure? | 16:48 |
| Damion3 | Before I edited a startup script (just a commented out line, but it'll have altered the hash and i didn't know about aegis at the time), thus aegis forcing it to reflash. The images I was on some release candidate firmware images, they had angrybirds and other quite impressive games and joikuspot's tether | 16:49 |
| thp | DocScrutinizer: and compared to fremantle, most of the ux is closed on harmattan, so no chance for the community to do improvements there down the road (ignoring the aegis issues for now) | 16:50 |
| Damion3 | thp: I have simless android devices and a Sony PSP and I think the iPad also - none of these work with adhoc | 16:50 |
| thp | Damion3: not even if adding the connection manually? | 16:50 |
| Damion3 | thp: you can root an android device and MAKE it work, PSP definitely not. Can't recall the ipad | 16:51 |
| Damion3 | thp: sure my linux based laptop doesn't care, so it's not useless. Although I already have ppp support working and that's better for battery of both devices! | 16:51 |
| DocScrutinizer | thp: ack. Still enough to mess around with. Hostmode a nogo without a bit of access at *real* root level, unless I'm going to build gentoo native on this thing, as a devlopment environment. And then I'm sure the result definitly won't fit into user aegis enabled systems | 16:52 |
| thp | Damion3: i see. that's because the other devices are even more evil :p | 16:52 |
| Damion3 | thp: On kernel 2.6.32, infrastructure mode became trivial VS h/ware specific/hard | 16:52 |
| Damion3 | thp: I'nm not disagreeing with that, just stating a point | 16:52 |
| Damion3 | thp: I'm not sure what posessed them to just kludge joikuspot rather than to do htis properly | 16:53 |
| Damion3 | HostAP and associated stuff used to rely on atheros chipset and some others. I don't actually know what it is you do to use the new 80211.ko in its new fancy mode, I just know it is explained in a few one liners in the kernel release | 16:54 |
| DocScrutinizer | thp: simple brightness applet, load applet, custom profiles... all needs a fair bit of system access, and probably community was smart enough to plumb it to harmattan, not sure though how aegis during development could help to speed up the process | 16:55 |
| Damion3 | there is little point in my finding out what is needed if I can't load a kernel module | 16:55 |
| Damion3 | DocScrutinizer: flashlight app too | 16:55 |
| DocScrutinizer | yes | 16:55 |
| DocScrutinizer | probably | 16:55 |
| DocScrutinizer | anyway, enough frustration for now. Maybe c u l8r | 16:57 |
| thp | DocScrutinizer: otoh, i can see this improving the architecture of said apps. as opposed to some apps/plugins just running as root, they actually have to create a small utility app/service that provides the information in a proper way | 16:57 |
| thp | and most developers don't fix the architecture once they have got something working on their dev machine | 16:57 |
| DocScrutinizer | thp: moot argument, as I'm *not* most developers, and I don't need forced lessons from nokia how to optimize my development processes | 16:58 |
| DocScrutinizer | those who like this kind of guidance are free to develop under normal mode | 16:58 |
| RST38h | Mem[||||||||||##*********300/985MB] | 17:00 |
| RST38h | Swp[ 0/255MB] | 17:00 |
| DocScrutinizer | I'm not eager to learn to live with aegis, only to find this being all obsolete knowledge in max 6 months | 17:00 |
| RST38h | Yeah, baby, let us put Maemo5 on it! | 17:00 |
| DocScrutinizer | and now bbl | 17:00 |
| *** CaCO3 has joined #harmattan | 17:02 | |
| DocScrutinizer | meh, PS: not willing to live with aegis *for development*, as that obviously means start to learn all your unix knowhow from scratch again. I'd be willing to cope with aegis after the PoC or even prototype is running just fine | 17:02 |
| DocScrutinizer | to do a cp ~root/ssh ~user/.ssh I don't need to learn to do this under aegis, as that'S not part of my final product, it's a part of the development process, and I'm not going to train for development under aegis, regarding that aspect | 17:04 |
| DocScrutinizer | as I system architect I'm rather confident my design is so sane I'd rather expect a bug in aegis than in my project's architecture if the final integration to aegis reveals problems :-) | 17:05 |
| RST38h | Doc: btw mc does not install | 17:07 |
| RST38h | cannot install libslang | 17:07 |
| DocScrutinizer | RST38h: that's the one from djszapi, yes. Sorry | 17:07 |
| RST38h | yeah | 17:08 |
| DocScrutinizer | lemme checkif I find the one of achipa again | 17:08 |
| RST38h | why doesn't he make libslang2 available? | 17:08 |
| Damion3 | DocScrutinizer: it's similar to nfs mounting with rootsquash | 17:08 |
| Damion3 | liberal use of chmod 777 for a moment before cp and then chmod 700 after (as user) will work. And as far as I can see su -c user chmod will work as root anyway | 17:08 |
| DocScrutinizer | RM680-22-6_PR_RM680:~# history | 17:09 |
| DocScrutinizer | -sh: history: not found | 17:09 |
| DocScrutinizer | FSCK messybox! | 17:09 |
| DocScrutinizer | RM680-22-6_PR_RM680:~# bash | 17:09 |
| DocScrutinizer | -sh: bash: Operation not permitted | 17:09 |
| RST38h | try h | 17:09 |
| RST38h | BTW, looks like N950 apt-get simply can't resolve dependencies | 17:10 |
| RST38h | It never tries installing depends, even if they are available | 17:10 |
| RST38h | maybe I am wrong though... | 17:10 |
| * RST38h has to talk rm_you into merging Harmattan and pre-Harmattan versions of Packrat | 17:11 | |
| RST38h | No need to have two really | 17:11 |
| DocScrutinizer | RST38h: h: not found. Sorry it scrolled out of the 100lines histrory anyway. So I'm sorry I forgot the URL for wget to install the *Real* (still flawed) mc | 17:11 |
| DocScrutinizer | RST38h: check chanlog | 17:12 |
| RST38h | mgedmin to the rescue! | 17:12 |
| DocScrutinizer | bye | 17:12 |
| * mtd answers own question from a few hours ago: this page seems to be a place to collect useful info for N950 devkit program participants: http://wiki.meego.com/N950_landing_page | 17:17 | |
| RST38h | OMG nano comes preinstalled! | 17:20 |
| mtd | RST38h: meh :). Wake me up when a real editor is on there. | 17:23 |
| RST38h | mtd: vi is there, I know that is what you want B)~ | 17:25 |
| mtd | RST38h: nope, I'm waiting for emacs :) | 17:26 |
| * mtd will be waiting for some time, I expect | 17:26 | |
| RST38h | feh | 17:49 |
| *** rcg has quit IRC | 17:59 | |
| Stskeeps | n.q | 18:15 |
| Stskeeps | p | 18:15 |
| Stskeeps | qa65.]qq | 18:16 |
| hiemanshu | Stskeeps: ? | 18:18 |
| *** Scifig has joined #harmattan | 18:33 | |
| *** CaCO3 has quit IRC | 18:34 | |
| macmaN | heh, so does anyone else's n950 lock screen drop a number from the clock? | 18:37 |
| macmaN | time is now 1_:36 | 18:37 |
| macmaN | it is an exercise to the user to figure out exactly what time it is | 18:37 |
| mtd | macmaN: no, mine looks fine | 18:38 |
| macmaN | right, mine did too for the first few weeks | 18:38 |
| macmaN | just noticed it yesterday | 18:38 |
| mtd | macmaN: weird | 18:39 |
| macmaN | no idea how it gets this way | 18:39 |
| frals | its a bug in that build | 18:39 |
| *** seif has quit IRC | 18:39 | |
| tomma | macmaN, it works after boot | 18:40 |
| macmaN | yep | 18:40 |
| tomma | what number was it? | 18:40 |
| macmaN | it actually works intermittently again i think without boot as well | 18:40 |
| macmaN | it's 18:38 right now, i think 8 therefore, but not sure if thats the only one | 18:40 |
| tomma | mine dropped number 3 | 18:40 |
| macmaN | lol | 18:41 |
| tomma | only 3 | 18:41 |
| macmaN | thats funny | 18:41 |
| *** hardaker has quit IRC | 18:52 | |
| *** spenap has quit IRC | 18:55 | |
| hiemanshu | macmaN: yeah, mine dropped 4, DocScrutinizer's dropped a 9 once | 18:59 |
| macmaN | haha | 19:00 |
| macmaN | i think its part of the n9 youtube video game | 19:00 |
| macmaN | "find the missing numbers!" | 19:00 |
| macmaN | and i also know how their sending this information from everybody's n950 | 19:01 |
| macmaN | RM680-22-6_PR_RM680:/var/log# cat pasms.log | 19:01 |
| macmaN | [2011-07-06T15:35:06] void PhoneActivation::smsSendFinished(bool, QString): success = "true" | 19:01 |
| macmaN | lol | 19:01 |
| macmaN | ok i really need to configure ssh keepalive for n950 | 19:02 |
| macmaN | this thing just drops wifi whenever it feels like it | 19:02 |
| hiemanshu | just waiting for a new image | 19:03 |
| macmaN | are you holding your breath like i am? | 19:03 |
| hiemanshu | nope | 19:03 |
| macmaN | if yes there could be breaking news about at least two suffocated meego developers soon | 19:03 |
| *** spenap has joined #harmattan | 19:06 | |
| *** TSCHAKeee has quit IRC | 19:10 | |
| M4rtinK2 | unlike with the N900 I have no problems with ssh connections dying on the N950 | 19:19 |
| M4rtinK2 | through wifi, even when it is not plugged in... | 19:19 |
| macmaN | M4rtinK2: could depend on the client. ServerAliveInterval 30 in .ssh/config seems to work | 19:19 |
| macmaN | now im trying to figure out why i cannot ssh user@n950 | 19:20 |
| M4rtinK2 | on the N900 I need to have ping running or the connection soon stops working | 19:20 |
| macmaN | how do i restart sshd | 19:20 |
| M4rtinK2 | is connect as root and then just do su user... | 19:20 |
| macmaN | isnt it normal in a sane world to do it the other way around? | 19:21 |
| macmaN | ok what kind of init system is this /etc/init/*.conf | 19:22 |
| macmaN | upstart or something? | 19:22 |
| macmaN | its not systemd right | 19:22 |
| Clint | yeah, upstart | 19:23 |
| macmaN | oh it is upstart, right | 19:23 |
| macmaN | well crap i dont know anything about it yet | 19:23 |
| macmaN | hmmm sqlite3 would be handy on n950 | 19:35 |
| M4rtinK2 | macmaN: well, Pythons sqlite3 is there (I just checked) | 19:41 |
| Mek | isn't there a normal sqlite3 package too you can easily install? | 19:42 |
| macmaN | apt-cache search sqlite doesnt give me anything | 19:42 |
| Mek | hmm, okay, must be in the internal repo only then.. sorry | 19:43 |
| macmaN | oh yeah, somehow the dpkg db gets locked also | 19:43 |
| macmaN | Updating desktop entries... Done | 19:43 |
| macmaN | E: Could not get lock /var/lib/dpkg/lock - open (11: Resource temporarily unavailable) | 19:43 |
| macmaN | E: Unable to lock the administration directory (/var/lib/dpkg/), is another process using it? | 19:43 |
| macmaN | only reboot releases it, or something that i dont know about | 19:43 |
| macmaN | heh | 19:46 |
| DocScrutinizer | kill $someprocess - somebody mentioned it here some days ago. Sorry forgot which process exactly | 19:46 |
| macmaN | File "/usr/lib/python2.6/curses/__init__.py", line 33, in initscr | 19:46 |
| macmaN | fd=_sys.__stdout__.fileno()) | 19:46 |
| macmaN | error: setupterm: could not find terminal | 19:46 |
| macmaN | trying to get https://github.com/hut/ranger running | 19:47 |
| DocScrutinizer | macmaN: lsof might help | 19:47 |
| macmaN | # htop | 19:48 |
| macmaN | Error opening terminal: xterm-256color. | 19:48 |
| macmaN | ok i hear that | 19:48 |
| macmaN | ok just regular xterm works | 19:48 |
| macmaN | ouch i guess not, messed up my terminal | 19:49 |
| macmaN | DocScrutinizer: lsof or fuser dont give anything for /var/lib/dpkg/lock | 19:54 |
| DocScrutinizer | :nod: same here | 19:54 |
| macmaN | https://bugreports.qt.nokia.com//browse/QTCREATORBUG-5380 ok this looks like something | 19:56 |
| macmaN | ok what the heck is /usr/lib/sdk-connectivity-tool/devrootsh | 19:56 |
| DocScrutinizer | ""<Mek> hmm, okay, must be in the internal repo only then.. sorry"" just makes me sad | 19:57 |
| Stskeeps | hiemanshu: n9, pocket | 19:57 |
| macmaN | say whaaat? | 19:58 |
| macmaN | is that a subtle hint about n9 in your pocket? or are you just happy to see us | 19:58 |
| Stskeeps | errn | 19:58 |
| Stskeeps | n950 | 19:58 |
| DocScrutinizer | obviously Nokia isn't inclined to really support the community regarding development. QML apps yes, but please don't touch the system, like using SQLite or anything | 19:58 |
| Stskeeps | sorry, i get those confused | 19:58 |
| macmaN | hmm damn could use wgetpaste on device | 20:00 |
| DocScrutinizer | basically QML is our dalvik, and we're not supposed to leave the sandbox | 20:00 |
| macmaN | so it turns out everyone has their dalvik to bear | 20:01 |
| Elleo | DocScrutinizer: there's nothing stopping you from compiling sqlite and using it | 20:01 |
| DocScrutinizer | yeah | 20:01 |
| DocScrutinizer | cool stuff | 20:01 |
| Elleo | it's literally the work of minutes with OBS just to link the existing debian packages up and just let it get on with building it | 20:01 |
| Elleo | when it's working correctly ;) | 20:02 |
| macmaN | if someone showed me a demo of how to do that i would | 20:03 |
| macmaN | obs sounds ufo to me | 20:03 |
| DocScrutinizer | see, 2 Nokians tried to "simply build" mc, and none of the both mc.deb really works. And it seems it tok them both more than just a few minutes to fail | 20:03 |
| DocScrutinizer | achipa's mc.deb actually works, but again only as non-root user | 20:05 |
| frals | "like using SQLite or anything" <- whats the problem with using SQLite? | 20:05 |
| DocScrutinizer | maybe I'm missing the instructions how to make it work under root, but then again that's not my fault | 20:06 |
| achipa | which nokians ? | 20:06 |
| Elleo | DocScrutinizer: did you try via devel-su from develsh rather than logging in as root? | 20:07 |
| DocScrutinizer | frals: ""<Mek> hmm, okay, must be in the internal repo only then.. sorry"" is the first annoyance with it | 20:07 |
| macmaN | killall -KILL pkgmgrd does the trickz | 20:07 |
| DocScrutinizer | achipa: aren't you a "Nokian" | 20:07 |
| Elleo | it started for me like that (as opposed to giving a permission denied), albeit not completetly since I don't have libslang2 which it requires | 20:07 |
| DocScrutinizer | ? | 20:07 |
| achipa | I'm not sure I failed. It took me <5min to port it from Maemo | 20:08 |
| achipa | but who is the other ? | 20:08 |
| DocScrutinizer | djszapi | 20:08 |
| achipa | he's not a nokian | 20:08 |
| DocScrutinizer | ooh | 20:08 |
| DocScrutinizer | anyway today I got more prodding to not bother about aegis anymore, as it | 20:09 |
| DocScrutinizer | s not welcome here | 20:09 |
| DocScrutinizer | so I'll stop about it | 20:09 |
| macmaN | YES | 20:10 |
| macmaN | ranger works | 20:10 |
| macmaN | try that for file manager DocScrutinizer | 20:11 |
| macmaN | if youre a vim person, you should dig it | 20:11 |
| macmaN | ok at least .ashrc works | 20:15 |
| DocScrutinizer | macmaN: thanks for the hint, but I think I made up my mind I'll stay with the unix toolbox I'm used to - that's after all why I spent money on Nokia devices running maemo. If that toolbox is obsolete, I'm not going to learn Nokia's world of system administration, I'll rather go for the real thing then if I'd really consider learning all new from scratch. | 20:15 |
| mtd | macmaN: not sure if it's related, but after setting authorized_keys ssh'ing in as user stopped asking for a password once I set a password for the user and symlinked authorized_keys2 -> authorized_keys | 20:16 |
| macmaN | mtd: hmmmm. it's asking for user@ password, even though i have authorized_keys. perhaps i should set one. also .ashrc seems to be ignored for Terminal hrm. where do i set my aliases then. | 20:16 |
| macmaN | DocScrutinizer: right. well im sure it'll get sorted out. just a holdover suggestion until then. | 20:17 |
| *** conny has joined #harmattan | 20:17 | |
| DocScrutinizer | macmaN: I'll wait for the day it gets sorted, until then I'm not wasting any more time on that | 20:18 |
| mtd | macmaN: yeah, I had the same symptom. I set the password and symlinked authorized_keys2 -> authorized_keys (IIRC at one point I'm sure this was necessary) | 20:18 |
| conny | Hi guys! | 20:18 |
| mtd | macmaN: one of those two worked for me, it seems. | 20:18 |
| conny | Anyone knows how to the the highlight color in Harmattan QML Components? | 20:18 |
| mtd | macmaN: I could ssh in as root w/o a password (using pub/priv key) before I could ssh in as user, and I did that w/o the a_k2 symlink, so I guess the password setting was the trick (I had set a root password long before) | 20:19 |
| conny | ...how to set the highlight color .... I mean | 20:19 |
| macmaN | mtd: symlinking alone isnt fixing it. i was wondering how i can get sshd to log into syslog. | 20:19 |
| mtd | macmaN: yeah I didn't get that far. I just set a password for "user". | 20:20 |
| macmaN | mtd: yep will try that next | 20:20 |
| mtd | macmaN: perhaps you could kill sshd and run it as user in -d mode | 20:20 |
| mtd | should QtCreator-compiled binaries run on the n950 w/o any aegis fiddling? | 20:21 |
| * mtd is doing the QtCreator tutorial and it works in the simulator, but gives a strage error when I try to run the binary on the N590. | 20:22 | |
| macmaN | conny: no deep qt devs around right now it seems | 20:23 |
| conny | macmaN: shouldn't be that deep ;) | 20:23 |
| mtd | ok, I lied, when I tried to run the binary _using_QtCreator_ it fails. It runs with a butt-ugly theme when I run it from the command line as "user". | 20:23 |
| npm | what code do i use to find out whether my application is in the "foreground"? (and not one of the deselected/inactive apps) | 20:24 |
| macmaN | 11 auth.info;authpriv.*;remoteauth.crit /var/log/secure.log | 20:24 |
| macmaN | hrm yet sshd doesnt log anything | 20:24 |
| npm | macmaN if you're having ssh problems read the ssh manpage carefully and set the permissions for all your files correctly | 20:25 |
| *** hardaker has joined #harmattan | 20:25 | |
| macmaN | npm: maybe i need manually touch secure.log | 20:25 |
| macmaN | trying that | 20:25 |
| npm | -rw------- 1 user users 744 Nov 16 2010 id_dsa | 20:25 |
| npm | -rw------- 1 user users 951 Nov 16 2010 id_rsa | 20:26 |
| npm | -rw------- 1 user users 536 Nov 16 2010 identity | 20:26 |
| lynxis | btw. qtcreator hates password protected keys | 20:26 |
| mtd | npm: don't forget dir permissions | 20:26 |
| npm | drwx------ 2 user users 4096 Jul 19 09:28 .ssh | 20:26 |
| npm | it'll fail silently if those are wrong | 20:26 |
| npm | and never go to logs | 20:27 |
| macmaN | npm: it will if you have LogLevel DEBUG or something higher than INFO i think. ive run into the same thing on desktop. | 20:27 |
| SpeedEvil | Also - permissions on .ssh and your user directoy are important | 20:27 |
| macmaN | yes true, thats why i am trying to get sshd log to show up in secure.log like its supposed to | 20:27 |
| macmaN | but i dont know how to restart upstart daemons | 20:27 |
| *** rcg has joined #harmattan | 20:27 | |
| macmaN | /etc/init.d/sshd apparently is not hip enough anymore | 20:28 |
| DocScrutinizer | stop <service>;slepp 2;start <service> | 20:28 |
| npm | "5. You may not (or allow others to) enable any logic or protocols in the Product that are disabled when delivered to You by Nokia." | 20:28 |
| DocScrutinizer | sleep* | 20:28 |
| DocScrutinizer | npm: haha, yeah a exceptionally funny one | 20:29 |
| macmaN | :) | 20:29 |
| macmaN | wow start and stop? really :) | 20:29 |
| SpeedEvil | npm: So you can't turn it on. | 20:29 |
| macmaN | that's actually... nice | 20:29 |
| DocScrutinizer | Imnetioned that before anybody (except jaffa?) got his device | 20:30 |
| npm | sorry, i just read that prior to looking into enabling twitter | 20:30 |
| npm | and it stopped me. back to working on apps :-) | 20:30 |
| mtd | npm: how does one enable twitter? | 20:30 |
| DocScrutinizer | quim suggested to basically ignore it | 20:30 |
| npm | i want that in writing | 20:31 |
| *** harbaum has joined #harmattan | 20:31 | |
| mtd | npm: why? | 20:31 |
| DocScrutinizer | npm: you'll face same bashing I got, when mentioning it | 20:31 |
| macmaN | aha initctl is the heart of upstart | 20:31 |
| npm | as i realluy want to keep my n950 -- it's amazing and wonderful and i'm like a 2001 monkey to the monolith about it | 20:31 |
| achipa | npm: http://library.developer.nokia.com/topic/MeeGo_1.2_Harmattan_API/html/guide/html/Developer_Library_Application_development_framework_Application_lifecycle_a92f.html | 20:32 |
| achipa | you got fore/background/active/etc stuff there | 20:32 |
| npm | i have plenty to do anyways, so it'll keep me on track :-) | 20:32 |
| macmaN | ooo initctl list | 20:32 |
| achipa | mtd: is that a qwidget application ? | 20:33 |
| DocScrutinizer | macmaN: check start and stop cmds, they are just symlinks to initctrl | 20:33 |
| macmaN | DocScrutinizer: yep yep | 20:33 |
| DocScrutinizer | macmaN: also man start gives you initctl manpage | 20:33 |
| macmaN | n950 doesnt have man | 20:33 |
| mtd | achipa: not sure - it's the BatteryMonitor tutorial app from QtCreator. Let me check. | 20:33 |
| DocScrutinizer | LOL | 20:33 |
| npm | achipa: thanks! | 20:33 |
| DocScrutinizer | forgot | 20:33 |
| macmaN | i dont run upstart anywhere else | 20:33 |
| macmaN | gentoo all the way | 20:33 |
| macmaN | man.die.net i guesses | 20:34 |
| macmaN | # initctl restart syslogd | 20:34 |
| macmaN | initctl: Unknown instance: | 20:34 |
| achipa | mtd: battery monitor or battery indicator ? | 20:35 |
| npm | achipa: wow i even had that one bookmarked but never thought that "application lifecycle" had something to do with it | 20:35 |
| macmaN | nevermind i think i somehow stopped it before already | 20:35 |
| macmaN | hrm i needz tmux on this thing | 20:36 |
| macmaN | didnt someone build a tmux package | 20:36 |
| DocScrutinizer | macmaN: http://ageofikon.com/prh/ , or google | 20:37 |
| *** deimos_ has joined #harmattan | 20:37 | |
| macmaN | http://repo.pub.meego.com/home:/rzr/MeeGo_1.2_Harmattan_Maemo.org_MeeGo_1.2_Harmattan_standard/armel/ | 20:37 |
| DocScrutinizer | good luck | 20:37 |
| *** deimos_ has left #harmattan | 20:37 | |
| *** deimos_ has quit IRC | 20:37 | |
| achipa | mtd: if you are referring to http://doc.qt.nokia.com/qtcreator-2.2/creator-mobile-example.html close the project and forget it ASAP | 20:37 |
| macmaN | uhh pack rat? what the heck, (c) 2008 how come ive never seen this before | 20:38 |
| DocScrutinizer | macmaN: note that apparently fremantle .deb install and run just fine under certain circumstances | 20:38 |
| npm | "However, applications must not update the UI when it is not visible to the end user." ... but i kinda liked seeing the video playing back all small while listening and doing other apps.... | 20:38 |
| npm | ( ^^ http://library.developer.nokia.com/index.jsp?topic=/MeeGo_1.2_Harmattan_API/html/guide/html/Developer_Library_Application_development_framework_Application_lifecycle_a92f.html ) | 20:38 |
| npm | hopefully that one is safe to ignore | 20:38 |
| npm | i just want to disable preventing screenblanking when app is in background | 20:38 |
| mtd | achipa: yeah, that one. What's not applicable about it? | 20:39 |
| achipa | npm: visible != minimized | 20:39 |
| DocScrutinizer | npm: an app IS visible but not active in taskswitcher | 20:39 |
| npm | true. it is technically still "visible" when updating in the app chooser | 20:39 |
| npm | gotta get my terminology rigjht | 20:40 |
| mtd | macmaN: looks like a nice repo | 20:40 |
| macmaN | so the only way to do bookmarking in device is to put them in apps right | 20:40 |
| DocScrutinizer | npm: the page achipa pointed you at *clearly* states exactly tis case | 20:41 |
| DocScrutinizer | macmaN: alas yes | 20:41 |
| achipa | mtd: it refers to the use of qwidgets on mobiles, which has been deprecated since last ~October | 20:41 |
| DocScrutinizer | strange concept | 20:41 |
| mtd | achipa: ah, ok. Is there a good replacement? | 20:42 |
| macmaN | can't install, invalid installation package | 20:42 |
| macmaN | at least straight from browser | 20:42 |
| macmaN | lets try dpkg | 20:42 |
| DocScrutinizer | macmaN: that's why I wish "good luck" :-D | 20:42 |
| DocScrutinizer | macmaN: won't pan out | 20:43 |
| npm | hmmm... is there a C++ infc for platformWindow.viewMode etc | 20:43 |
| achipa | mtd: QML (+components) | 20:43 |
| DocScrutinizer | it will install but not work when you force it via dpkg -i | 20:43 |
| npm | as i don't want to bother my qml code w/ stuff like that (and the display blanker's already in C++) | 20:43 |
| DocScrutinizer | macmaN: been there done that, 2 times ;-) | 20:44 |
| DocScrutinizer | well, actually that's not true, Nokia's bash .deb was "invalid format" but runs under user nevertheless | 20:45 |
| DocScrutinizer | so it's just useless as a generic shell :-P | 20:45 |
| mtd | achipa: thanks. It's a shame there is nothing with "QML" in the QtCreator initial "Welcome" screen, and that the "Creating a Mobile Application" section that is is so wrong :( | 20:46 |
| achipa | mtd: welcome to my world :) :( | 20:47 |
| achipa | desktop people go crazy when you say QML and mobile people go crazy when you say QWidgets... | 20:47 |
| mtd | achipa: I was asking earlier about any N950 developer tips...does everybody but me know that already ? Where should I have found that info? | 20:48 |
| macmaN | and i still dont know the difference between those | 20:48 |
| mtd | s/know/that/know that one/ | 20:48 |
| macmaN | perhaps ignorance is bliss after all | 20:48 |
| macmaN | :) | 20:48 |
| macmaN | DocScrutinizer: im not sure what youre talking about there buddy, cause tmux fired up just fine for me | 20:48 |
| DocScrutinizer | indeed I sypathize with both parties | 20:48 |
| macmaN | dpkg -i libevent...deb | 20:49 |
| macmaN | dpkg -i tmux...deb | 20:49 |
| macmaN | tmux | 20:49 |
| macmaN | party! | 20:49 |
| DocScrutinizer | macmaN: as root? or under user, with --relxed-exec? | 20:49 |
| macmaN | so far just tried root | 20:49 |
| macmaN | was busy jumping around the room to try user | 20:50 |
| DocScrutinizer | hmm, fine, so we have a third behaviour, actually a 4th when we count normal behaviour as well | 20:50 |
| macmaN | yep works fine for user too | 20:50 |
| achipa | mtd: actually, just start at http://library.developer.nokia.com/topic/MeeGo_1.2_Harmattan_API/html/guide/html/Developer_Library_Application_development_framework_ebcf.html | 20:50 |
| macmaN | hmm can i just see everything thats available in pack rat? | 20:51 |
| macmaN | it says i have to search something for empty query | 20:51 |
| DocScrutinizer | a) doesn't install under browser (diub), and doesn't work at all, b) diub and works under user account only, c) your case, diub and works nevertheless, and d) installs under browser | 20:52 |
| DocScrutinizer | this is a really intriguing text adventure | 20:52 |
| macmaN | DocScrutinizer: ok well this is interesting... | 20:53 |
| macmaN | # tmux | 20:53 |
| macmaN | -sh: tmux: Operation not permitted | 20:53 |
| DocScrutinizer | that's "normal" ;-P | 20:53 |
| macmaN | right | 20:53 |
| macmaN | then i do su - user, then it works | 20:53 |
| DocScrutinizer | yes | 20:53 |
| macmaN | inside su, i do deve-su, then it works under root | 20:53 |
| macmaN | lovely | 20:54 |
| DocScrutinizer | that's b) | 20:54 |
| macmaN | right my first attempt was inside this devel-su inside su - user | 20:54 |
| macmaN | so everything is consistent :) | 20:54 |
| * DocScrutinizer sighs and turns away | 20:54 | |
| mtd | achipa: thanks very much | 20:55 |
| DocScrutinizer | I guess the last time I sucked up this amount of useless new things in this short time was when I had to deal with VBA | 20:55 |
| DocScrutinizer | at a time when they thought translating sourcecode was a smart thing | 20:56 |
| mtd | does editing /etc/mce/mce.ini trigger aegis-lock? | 20:56 |
| DocScrutinizer | prolly not | 20:57 |
| DocScrutinizer | though on your own peril | 20:57 |
| DocScrutinizer | the question however is: HOW T*F* do we check if it's safe or not? | 20:57 |
| DocScrutinizer | and again I'm involved. KTNXBY | 20:58 |
| macmaN | :) | 20:58 |
| macmaN | you can never get enough of a good thing | 20:58 |
| *** hardaker has quit IRC | 21:00 | |
| *** seif has joined #harmattan | 21:00 | |
| seif | hey does meego CE work on n950 | 21:00 |
| DocScrutinizer | <quote IT crowd>use su - and Edit it! when it doesn't work, get a *real* OS! Then edit it</quote> | 21:01 |
| DocScrutinizer | nuff said | 21:01 |
| *** meegoexperts_mob has joined #harmattan | 21:02 | |
| seif | Damion3, huh | 21:04 |
| seif | ? | 21:04 |
| seif | DocScrutinizer, ^ | 21:04 |
| Elleo | whoever it was that was after sqlite I've just built some packages: https://build.pub.meego.com/package/show?package=sqlite3&project=home%3Aelleo they should appear in my repository whenever OBS gets around to publishing them: http://repo.pub.meego.com/home:/elleo/Harmattan/ | 21:07 |
| Elleo | macmaN: ^ think it was you? | 21:08 |
| *** lynxis has quit IRC | 21:13 | |
| *** cpscotti has quit IRC | 21:14 | |
| macmaN | Elleo: mhm it was | 21:17 |
| macmaN | thanks dawg | 21:17 |
| Elleo | no problem :) | 21:18 |
| macmaN | just doing devel-su when logging in as root will also allow you to run stuff | 21:18 |
| macmaN | at least dpkg-installed stuff.. | 21:19 |
| seif | vgrade, there? | 21:20 |
| seif | Elleo, you kinda rock :) | 21:20 |
| mtd | DocScrutinizer: I'm looking up the inode of the file-that-might-be-under-the-aegis-of-aegis in /sys/kernel/security/validator/hashlist | 21:22 |
| mtd | DocScrutinizer: that should tell me is aegis covers it. | 21:22 |
| macmaN | Elleo: can you ship an openvpn package while you're at it? | 21:22 |
| Elleo | macmaN: I'll see what I can do | 21:23 |
| macmaN | thanks, really really dying for that. ill take a look at coding up a *decent* status applet for it. | 21:24 |
| macmaN | can anyone explain why .bashrc is executed for root on login | 21:24 |
| macmaN | is that built into ash? | 21:25 |
| *** lynxis has joined #harmattan | 21:25 | |
| macmaN | at the same time i cant get anything to execute on terminal launch | 21:25 |
| macmaN | for user | 21:25 |
| macmaN | anything = .bashrc or .ashrc | 21:26 |
| macmaN | what sense does this make http://lists.maemo.org/pipermail/maemo-users/2006-January/000450.html | 21:26 |
| *** cpscotti has joined #harmattan | 21:26 | |
| macmaN | oh Just had a look at the busybox source and it says it uses the contents of environment variable ENV as a startup script so you could do | 21:28 |
| macmaN | yeah ok that works | 21:29 |
| Elleo | macmaN: openvpn may take a little longer as there are a few extra dependencies that need building first | 21:31 |
| macmaN | yeah liblzo and friends | 21:31 |
| *** antman8969 has joined #harmattan | 21:37 | |
| rm_you | RST38h: yeah | 21:43 |
| rm_you | RST38h: i just wanted to hack it into working | 21:43 |
| rm_you | RST38h: but seriously, i had to rewrite a bunch of stuff completely to get it to work | 21:43 |
| rm_you | "merging" it would require like... having two versions of every function | 21:44 |
| rm_you | it's not feasible | 21:44 |
| rm_you | if YOU want to try, go for it | 21:44 |
| rm_you | but i barely even know Perl anyway (the Crawler is perl... did you write it? I forgot who gave it to me) | 21:44 |
| *** antman8969_ has joined #harmattan | 21:45 | |
| *** antman8969_ has quit IRC | 21:46 | |
| *** antman8969_ has joined #harmattan | 21:46 | |
| *** antman8969 has quit IRC | 21:46 | |
| *** antman8969_ has quit IRC | 21:47 | |
| *** antman8969 has joined #harmattan | 21:47 | |
| rm_you | woo maybe fixed my pandora bug. probably not the right way, but it goddamn works. this is going to be battery draining :/ | 21:57 |
| *** seif has quit IRC | 21:58 | |
| *** seif has joined #harmattan | 21:58 | |
| *** harbaum has quit IRC | 22:03 | |
| *** meegoexperts_mob has quit IRC | 22:05 | |
| RST38h | rm_you: Hey, is there absolutely no way to have maemo5- and harmattan repos in the same version of packrat? | 22:06 |
| rm_you | RST38h: like i said, it's completely different | 22:06 |
| RST38h | hmm...how different? | 22:06 |
| RST38h | must still be a debian repo no? | 22:06 |
| rm_you | the logic for everything from detecting repositories to indexing them to reading the package info | 22:06 |
| rm_you | all different | 22:06 |
| rm_you | all before maemo5 pull from gronmayer.it | 22:07 |
| rm_you | harmattan pulls from a wiki | 22:07 |
| rm_you | maemo5 are true apt repos so they have a set structure | 22:07 |
| rm_you | err | 22:07 |
| rm_you | pre-maemo5 | 22:07 |
| rm_you | harmattan repos are PPAs with no structure predictable | 22:07 |
| rm_you | and then once you get to package metadata, they have completely different attributes | 22:07 |
| rm_you | also -- i don't care about anything < Harmattan anymore, and have other things i'm busy with :P i just got it working as a stopgap | 22:08 |
| RST38h | ah | 22:08 |
| RST38h | Meanwhile: The Hacker News is reporting that Android password data is being stored as plain text in its SQlite database. | 22:09 |
| rm_you | lol | 22:10 |
| macmaN | sure why not. its not like people are going to steal ur tablet. | 22:10 |
| RST38h | it is not? | 22:12 |
| *** conny has quit IRC | 22:12 | |
| *** gri_ has joined #harmattan | 22:17 | |
| *** gri has quit IRC | 22:20 | |
| *** antman8969 has quit IRC | 22:24 | |
| *** antman8969 has joined #harmattan | 22:25 | |
| *** Scifig has quit IRC | 22:26 | |
| *** Venemo has joined #harmattan | 22:32 | |
| DocScrutinizer | hey, if you're talking about passwords for services off-device, then there's actually little to no alternative to storing the passwords on device in plaintext - no matter which encoding you use for this plaintext storage | 22:34 |
| Venemo | good evening harmattanites | 22:34 |
| DocScrutinizer | it's an absolute nonsense idea to obfuscate passwords for e.g. pop3 mail accounts on client side. A simple wireshark monitoring of next time the mail client is polling the mailserver will reveal the passwords anyway | 22:37 |
| Venemo | DocScrutinizer, how're you doin'? | 22:41 |
| *** andre__ has joined #harmattan | 22:54 | |
| *** andre__ has joined #harmattan | 22:54 | |
| *** spenap has quit IRC | 22:56 | |
| *** Scifig has joined #harmattan | 22:57 | |
| *** wicket64 has joined #harmattan | 22:57 | |
| Venemo | so, is there any news in Harmattan since I was gone? | 22:58 |
| RST38h | aside from DocScrutinizier going ballistic over Aegis?No. | 23:00 |
| Damion3 | for some reaon clicking on the packrat .deb doesn't install but the binaries within those .debs ork (tmux) | 23:05 |
| Venemo | RST38h, that is not news :P | 23:05 |
| Venemo | do we have an IRC client yet? | 23:06 |
| Damion3 | a useful one? | 23:06 |
| Venemo | uummm... any one? | 23:07 |
| dm8tbr | ssh + irssi | 23:07 |
| Damion3 | I'm using irssi (albeit via ssh+screen+host), but if you mean something which talks to a persistant daemon maintaining your presence for you that you communicate with from whatever network. Then not yet | 23:08 |
| Damion3 | I suspect the maemo version of what I describe runs fine | 23:08 |
| Venemo | um, isn't irssi a terminal thingy? | 23:08 |
| Damion3 | isn't irc all text? | 23:08 |
| Damion3 | I'm on my n950 right now | 23:09 |
| Venemo | mhm | 23:09 |
| *** antman8969 has quit IRC | 23:09 | |
| *** Scifig has quit IRC | 23:11 | |
| *** Scifig has joined #harmattan | 23:12 | |
| *** djszapi has joined #harmattan | 23:12 | |
| achipa | DocScrutinizer: actually, http://library.developer.nokia.com/index.jsp?topic=/MeeGo_1.2_Harmattan_API/html/categories/signon.html gives you support for auth stuff with reasonable security. | 23:14 |
| Venemo | hey achipa, how're you? | 23:15 |
| achipa | hey, just finishing up and getting ready for late night stroll, while it's still not dark | 23:15 |
| *** spenap has joined #harmattan | 23:16 | |
| Venemo | :) | 23:16 |
| Venemo | sounds nice | 23:17 |
| DocScrutinizer | achipa: I don't think this page can give me *any* new insight about pop3 properties and protocol. I'm well aware you can have data stored "secure" on a TC device, but that's still nonsense unless you make sure nobody does e.g install a wireshark with a spy function. If you can ensure this, you also can protect files that hold passwords in plain text. | 23:17 |
| Venemo | DocScrutinizer, so, you think it'd be better storing them in plaintext? | 23:18 |
| djszapi | achipa: this documentation issue was already mentioned to the SDK team | 23:18 |
| DocScrutinizer | Venemo: sure, what's the use of complicating life of your sysadmin? | 23:18 |
| djszapi | achipa: security fw documentation. | 23:18 |
| DocScrutinizer | bye o/ | 23:19 |
| Venemo | DocScrutinizer, I agree with you about wireshark, but about all the other usecases, an encrypted password might be useful. | 23:19 |
| djszapi | achipa: and it requires a lot of resources which is probably not worth it, mainly with everybody leaving. This is mainly a user device, Nokia does not focus on developer device purposes from the software point of view. | 23:19 |
| * RST38h screams BANZAI!, goes to try the SDK | 23:20 | |
| achipa | Venemo: you could protect them. But since you won't bother, or can mess up, it's probably less fuss to do it through signon - and that one also leverages cross-application authentication | 23:20 |
| Venemo | achipa, I see | 23:20 |
| Damion3 | laterz | 23:21 |
| achipa | DocScrutinizer: and that's something your file protection won't give | 23:21 |
| *** Damion3 has left #harmattan | 23:21 | |
| DocScrutinizer | achipa: protecting against WHAT? | 23:21 |
| achipa | DocScrutinizer: against people/apps accessing/leaking the passwords, or keeping unnecessary parallel connections | 23:22 |
| djszapi | mmh: http://qole.blogspot.com/2011/07/please-remove-harmattan-platform.html | 23:23 |
| DocScrutinizer | tzz I always thought that's the duty of the app's developer | 23:23 |
| achipa | if he's the only one using it, yes (and probably fail at it) | 23:24 |
| djszapi | DocScrutinizer: you cannot write secure application on Windows, if the platform is not like that :) | 23:24 |
| achipa | but think, say, google credentials | 23:24 |
| achipa | it's used for mail, youtube, picasa, whatnot | 23:24 |
| achipa | zillion different services | 23:24 |
| DocScrutinizer | honestly there's NO increased security by whatever means as long as your protocol requires transfer of the plain text credentials at any given time | 23:24 |
| achipa | DocScrutinizer: and who says you transfer plaintext ? | 23:25 |
| DocScrutinizer | pop3 RFC says | 23:25 |
| achipa | who gives a s**t, should be SSL/TLS anyway with hashes | 23:26 |
| achipa | this is not about "I can make a use case where a 30 year old braindead protocol exposes passwords" | 23:26 |
| DocScrutinizer | pfff, still you send the plain password, just over a secure channel | 23:26 |
| achipa | err, no ? | 23:26 |
| DocScrutinizer | the secure cannel ends on your device | 23:26 |
| achipa | look, this is about services in general. If someone wants to use plaintext POP3, their problem | 23:27 |
| kimju | pop3 with client side storage on phone sounds so wrong.. imap(s) and server side storage of messages is much better (and secure) idea | 23:27 |
| DocScrutinizer | and even when you negotiate a asymmetric set of credentials, you still have to store the credentials locally, and the only way to keep this secure is to store them inside TPM | 23:28 |
| kimju | I actually like the idea of phone providing secure storage for passwords etc.. | 23:28 |
| *** djszapi has left #harmattan | 23:28 | |
| achipa | DocScrutinizer: and that's what signon does | 23:29 |
| *** gri_ has quit IRC | 23:30 | |
| DocScrutinizer | kimju: achipa: OT - the fact that *some* app needs to know about a plain text of some credential makes it moot to try and crypt the credential on the device that runs the app | 23:30 |
| achipa | DocScrutinizer: no apps need to know, that's what I'm saying | 23:31 |
| DocScrutinizer | there's *always* a way to run the app under gdb, or run wireshark, or find any other way to snoop the credential | 23:31 |
| achipa | DocScrutinizer so that's why you curse aegis that much ? :) | 23:31 |
| DocScrutinizer | achipa: sorry, I'm bored about this discussion really | 23:31 |
| DocScrutinizer | and I kinda promised I won't bitch about aegis anymore, I will just ignore it whenever I run into it. "doesn't work? oh well, doesn't work then" | 23:32 |
| achipa | The bottom line is - POP3 is a non issue | 23:33 |
| achipa | APOP is pretty much supported by all major mail providers | 23:33 |
| Elleo | macmaN: openvpn has built, I had to modify it slightly to disable the PAM plugin, hope that isn't an issue for you | 23:33 |
| *** djszapi|win has joined #harmattan | 23:33 | |
| rcg | when talking about aegis.. afaik the secure storage is "secured" by some some "token" that it is "given" to the app, e.g. the app id, via the aegis manifest.. so what prevents me from writing another app that requests the same token in its manifest/uses the same app id and access that "secure" storage? | 23:33 |
| achipa | (though why one would want to go POP3 in the first place is beyond me) | 23:33 |
| Elleo | macmaN: should be in my repo soon (whenever OBS transfers it) | 23:33 |
| rcg | *-one some | 23:34 |
| macmaN | Elleo: thank you sir :> | 23:34 |
| Elleo | no problem :) | 23:34 |
| djszapi|win | Elleo: are you a Harmattan packager ? | 23:34 |
| achipa | rcg: source - your key signs the package. He cannot claim to have come from com.foocorp if he lacks your keys | 23:34 |
| DocScrutinizer | haha ok OT is OT as everything is a nonissue when you decide it's a nonissue just by avoiding to use it. Doesn't change my statement about storage of credentials and nonsense to crypt them | 23:34 |
| achipa | rcg: and even the app cannot 'read' the credentials, that's the idea of signon | 23:35 |
| achipa | rcg: it's a one way trip, once it's in there, only the auth plugin uses it and gives back a handle or whatever | 23:35 |
| rcg | achipa: ic.. am not talking about signon.. just asking about the concept behing aegis storage | 23:35 |
| macmaN | Elleo: do you have your own repo appended to sources.list.d or how do you do it? | 23:35 |
| Elleo | djszapi|win: well I've been putting together packages for Harmattan, dunno if I'd call myself a "Harmattan packager" though ;) | 23:36 |
| Elleo | macmaN: yeah just wget http://mikeasoft.com/~mike/elleo.list in to /etc/apt/sources.list.d/ | 23:36 |
| achipa | rcg: yes, as said, that's regulated through source - you can provide tokens from/to predefined sources, not 'everybody' | 23:36 |
| spenap | Elleo, thanks: I got my package building in OBS :) | 23:37 |
| Elleo | spenap: awesome :) | 23:37 |
| djszapi|win | Ellea: do you happen to have an idea how to make kdelibs work OBS ? :) It has not been working for 2 weeks or so | 23:37 |
| Elleo | spenap: is that the space invaders one? | 23:37 |
| spenap | Elleo, nope, I found that googling for the debhelper message error | 23:37 |
| spenap | I'm writing butaca, this application which provides movie information using themovie database | 23:38 |
| Elleo | djszapi|win: is it a specific package that's stalling things or is nothing being sent to build? | 23:38 |
| spenap | (this free approach to IMDB) | 23:38 |
| Elleo | spenap: ah, cool | 23:38 |
| *** deimos has quit IRC | 23:38 | |
| djszapi|win | Elleo: sometimes, rpc timeout - I guess that is mic2 related. Sometimes, it completely compiles the whole kdelibs, but it got stuck before creating the debian files. | 23:38 |
| Elleo | djszapi|win: I don't really know that much about OBS (or KDE stuff) I just have some basic debian packaging experience and have been playing with OBS for a couple of weeks | 23:38 |
| spenap | you can check it out if you're interested: it's (it should be) quite usable right now | 23:38 |
| djszapi|win | The build time is about 3-5 hours for it on OBS. | 23:39 |
| Elleo | djszapi|win: yeah, the rpc timeout stuff is a problem for all Harmattan builds it seems | 23:39 |
| rcg | achipa: hmm ic.. so simply said: in a "real world" use case.. that would prevent someone else from creating an app with the same "token" as my app to access my secured aegis storage? | 23:39 |
| Elleo | seems pretty intermitent too | 23:39 |
| djszapi|win | Elleo: I am not getting it for all | 23:39 |
| Elleo | djszapi|win: nah, it happens seemingly randomly on different packages | 23:39 |
| Elleo | I just trigger rebuilds on packages that get stuck until it works | 23:40 |
| achipa | DocScrutinizer: you don't make sense. the fact that ONE authentication method of ONE protocol travels through the net unprotected means nothing from a platform standpoint (and all the other services and auth methods that ARENT plaintext or want to be leaked) | 23:40 |
| rcg | but there is nothing which would allow the user to enter his own passphrase to secure an aegis storage? | 23:40 |
| djszapi|win | Elleo: I am not getting it for random packages, it is all the time kdelibs. | 23:40 |
| Elleo | I think only x-fade or lbt are really in a position to work out what's actually going on there | 23:40 |
| djszapi|win | Elleo: kdelibs built fine 2 weeks ago locally | 23:40 |
| lbt | mmm | 23:40 |
| DocScrutinizer | achipa: you don't get it | 23:40 |
| achipa | rcg: correct | 23:41 |
| djszapi|win | I have never could manage it on OBS. | 23:41 |
| Elleo | djszapi|win: perhaps there's some sort of race condition going on with very large things like kdelibs? | 23:41 |
| djszapi|win | Elleo: I was guessing more times about the memory | 23:41 |
| Elleo | with the package not finishing transfering before the build vm starts trying to do stuff? | 23:41 |
| djszapi|win | 6 GB really seems low enough | 23:41 |
| achipa | DocScrutinizer: sorry, I'm too dumb | 23:41 |
| Elleo | but I can only make random uninformed guesses, I don't really know anything much about OBS's internals | 23:41 |
| DocScrutinizer | achipa: I'm not talking about traveling the net unprotected, I'm talking about vulnerabilities ob your local device are same category for a file with plain password and for an app decrypting a crypted password to send it over SSL right away | 23:42 |
| djszapi|win | Elleo: I have 4 GB memory and it consumes a lot locally for kdelibs. | 23:42 |
| Venemo | how is the USB host mode coming? | 23:42 |
| DocScrutinizer | not at all Venemo | 23:42 |
| djszapi|win | Elleo: with a build server like that where there are more paralel processes. I can merely imagine it is not enough | 23:42 |
| Venemo | DocScrutinizer :( | 23:42 |
| DocScrutinizer | and now bye for good | 23:42 |
| Venemo | bye | 23:43 |
| achipa | DocScrutinizer: as said, in a normal case, your app does not get to decrypt the password, it stays on the plugin/signon side of the wall | 23:43 |
| Elleo | djszapi|win: perhaps, I'd have thought it'd start building though (until it ran out of RAM and/or swap) rather than just giving rpc timeouts straight off | 23:43 |
| macmaN | Elleo: | 23:45 |
| macmaN | # sqlite3 | 23:45 |
| macmaN | SQLite header and source version mismatch | 23:45 |
| macmaN | 2010-12-07 20:14:09 a586a4deeb25330037a49df295b36aaf624d0f45 | 23:45 |
| macmaN | 2011-06-23 19:49:22 4374b7e83ea0a3fbc3691f9c0c936272862f32f2 | 23:45 |
| macmaN | interesting | 23:45 |
| achipa | DocScrutinizer: think about google. If you make a youtube app, you use the google signon features, and your app does NOT see the google credentials | 23:45 |
| Elleo | macmaN: odd, that's just built straight from the debian package without any modification | 23:45 |
| *** djszapi|win_ has joined #harmattan | 23:45 | |
| * lbt just upped the memory / job on one of the workers from 4Gb to 6Gb | 23:46 | |
| lbt | what we should do is ask people with 'large' pacages to notify us and we can direct them to a beefy worker | 23:46 |
| lbt | but clearly we're not at that point yet | 23:46 |
| djszapi|win_ | lbt: it blocks the whole KDE development process for N9(50). It is somewhat high priority bug | 23:47 |
| macmaN | Elleo: | 23:47 |
| macmaN | The following packages have unmet dependencies. | 23:47 |
| macmaN | openvpn: Depends: openssl-blacklist (>= 0.4) but it is not installable | 23:47 |
| macmaN | Depends: openvpn-blacklist but it is not installable | 23:47 |
| Elleo | macmaN: ah okay, I'll just build those quickly | 23:48 |
| Elleo | those are just runtime dependencies rather than build deps so I didn't run across them before | 23:48 |
| Elleo | if they're just blacklists they should be pretty simple to build though | 23:48 |
| *** djszapi|win has quit IRC | 23:48 | |
| *** djszapi|win_ is now known as djszapi|windows | 23:50 | |
| lbt | djszapi|windows: yep... I've been working all weekend and X-fade has had a *gasp* holiday ... there's very little dedicated resource and a *lot* going on. We wish we had more time to help. Really. | 23:51 |
| macmaN | Elleo: holding my breath now :> | 23:51 |
| lbt | can you keep an eye on the build and see if it works or fails on any particular worker? kdelibs and the rpc timeout | 23:51 |
| kimju | lbt, would be nice if obs project/package config had "I need X amount of resources to build" fields.. | 23:52 |
| djszapi|windows | lbt: I am not blaming you, do not read it that way. I am just trying to put its priority clear. | 23:52 |
| lbt | djszapi|windows: yeah ... I guess I just feel responsible.... I want it to work better :) | 23:52 |
| djszapi|windows | lbt: How can I help with it ? My KDE dudes are waiting for some KDE apps development on this gadget :p | 23:52 |
| lbt | kimju: *nod* ... in general it's only for exceptional cases - and admins do that | 23:53 |
| lbt | djszapi|windows: I will be trying to get more workers online next week - high prio for MeeGo IT | 23:53 |
| djszapi|windows | I agree. It would not be nice to expose. | 23:53 |
| djszapi|windows | lbt: I think proper error messages would help also a lot | 23:54 |
| lbt | yes - OBS is actually very cool - but there are a *lot* of rough areas. | 23:54 |
| djszapi|windows | lbt: for instance, I can ping you when the compilation finished and maybe you can monitor the memory usage, I do not know... | 23:54 |
| lbt | hmm... | 23:54 |
| djszapi|windows | me as a Harmattan developer, I do not care about how much OBS rocks :) | 23:55 |
| djszapi|windows | if it does not work for my use case ;) | 23:55 |
| lbt | :) | 23:55 |
| djszapi|windows | but yes, it is a cool tool, I agree. | 23:55 |
| djszapi|windows | all my respect =) | 23:55 |
| lbt | What would be nice is a user level script that reports memory usage | 23:56 |
| lbt | atop springs to mind | 23:56 |
| djszapi|windows | lbt: the problem is that oom would make a boooommm | 23:56 |
| djszapi|windows | so that might be also not a good guess. | 23:56 |
| Venemo | so, why don't we just make a harmattan-extras on repository.maemo.org? | 23:56 |
| djszapi|windows | Venemo: why not on OBS ? | 23:56 |
| djszapi|windows | since that is what we discussed last time. | 23:57 |
| Venemo | it seems to me that the OBS doesn't exactly work. | 23:57 |
| lbt | Venemo: clearly that's not the case | 23:57 |
| djszapi|windows | lbt: 6 GB is not enough in my opinion. | 23:57 |
| Venemo | all I heard about it was complaints, so I didn't even bother with it myself | 23:57 |
| djszapi|windows | lbt: is there any chance to expand that ? | 23:57 |
| lbt | djszapi|windows: yes... you can have 32Gb ... and 1 job at a time | 23:58 |
| djszapi|windows | Venemo: your idea is totally irrelevant | 23:58 |
| Venemo | thanks djszapi|windows | 23:58 |
| Elleo | Venemo: it works well for me (most of the time) | 23:58 |
| djszapi|windows | Venemo: you would like to have a repository independently from OBS building. | 23:58 |
| Venemo | what do you mean, most of the time? | 23:58 |
| lbt | Venemo: it's the usual problem... if you don't help us fix it, it won't get better. | 23:58 |
| djszapi|windows | Venemo: it is really irrelevant to creating a repository | 23:58 |
| Elleo | there are intermittent problems now and then | 23:58 |
| lbt | There are times (right now) when it has worse patches | 23:58 |
| Elleo | sometimes vms don't talk to the system (and you just have to wait until it gets reattempted) | 23:59 |
| Venemo | lbt, I'm incompetent about OBS, so I can't help you fix it | 23:59 |
| lbt | the N900 CE project has been using it for months w/o any real issues | 23:59 |
| Elleo | but for the most part it works really nicely | 23:59 |
| lbt | Venemo: docs, tutorials... | 23:59 |
| DocScrutinizer | Venemo: see? that's why it won't get better any time soon :-P | 23:59 |
| lbt | Venemo: careful bug reports, analysis | 23:59 |
| djszapi|windows | It does not really make sense to create something different than OBS if there are quite a few packagers there with ready made packages. | 23:59 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!