**** BEGIN LOGGING AT Sun Oct 05 02:59:57 2008
Oct 05 05:53:30 <Ken_Young> The gdk documentation says that most of the gdk font routines, such as gdk_font_load, are deprecated.   But the documentation does not say what new code should use instead of those calls.   Does anyone know how new fonts should be found and loaded in gdk/gtk on a Freerunner?  Are we supposed to be using pango?
Oct 05 10:12:12 <rwhitby> morning
Oct 05 10:12:20 <Ken_Young> Is there anyone online who understands how to select different X11 fonts on the Freerunner?
Oct 05 10:12:33 <quickdev> Hey
Oct 05 10:12:54 <Ken_Young> rwhitby, Good morning.
Oct 05 10:13:00 <rwhitby> g'day
Oct 05 10:14:00 * rwhitby checks on the fso autobuilder to see where it's up to ...
Oct 05 11:02:03 <rwhitby> mwester: http://moko.mwester.net/download/uImage-gta01-g78199ea0_mwester-stable.bin is linked to from the qtopia extended release page, but doesn't exist.
Oct 05 11:02:24 <rwhitby> similarly for http://moko.mwester.net/download/uImage-gta02-g78199ea0_mwester-stable.bin
Oct 05 11:02:49 <rwhitby> lpotter: got any links for kernels compatible with qtextended gta01 and gta02 rootfs?
Oct 05 12:22:21 <mwester> rwhitby:  The links on the qtopia extended page just below (the dl.htm) links should work -- I had to update the kernels to fix the GTA02 APM status.  If I could figure out how to create a custom 404 page on that hosting site, I'll do that so it offers a bit more info... 
Oct 05 12:23:25 <mwester> yeah!!  FSO-unstable built finally, don't check anything in anybody, I need to do a little work while it actually builds! ;) :-D
Oct 05 12:26:12 <emdete> mwester: does gdk-pixbuf build again?
Oct 05 12:31:38 <alphaone> mwester: When can I start commiting again? :-)
Oct 05 12:34:44 <mwester> emdete: ./fso-unstable/tmp/staging/armv4t-angstrom-linux-gnueabi/usr/include/gtk-2.0/gdk-pixbuf   <-- was created, so I assume there's an ipkg somewhere :-D
Oct 05 12:35:04 <emdete> alphaone: i attached some patches with #167 but probably introduced some typos/errors because i can't test it in fso. can you take a look?
Oct 05 12:35:05 <mwester> alphaone:  :p  immediately, as long as it all works!
Oct 05 12:35:21 <emdete> mwester: strange, it does not build here after a make clean`
Oct 05 12:35:33 <alphaone> mwester: Hmm, that could be a problem ;-)
Oct 05 12:35:36 <alphaone> emdete: Thanks, will do
Oct 05 12:36:36 <emdete> alphaone: i wanted to have 3 clean patches but... sry
Oct 05 12:55:33 <emdete> alphaone: those patches are just thought as a starting point for getting the ubx a stand-alone module...
Oct 05 12:56:04 <alphaone> emdete: Yeah, but some of that is not going to happen I fear...
Oct 05 12:56:19 <emdete> alphaone: which ones?
Oct 05 12:56:25 <alphaone> Like the ubx specific dbus calls will stay in the ubxdevice.
Oct 05 12:56:45 <alphaone> But what I though I could do is separate the ubx parsing stuff from the ubx device completely
Oct 05 12:56:54 <emdete> that's more than ugly, in the protocol layer shouldnt be a dbus dependency
Oct 05 12:59:08 <alphaone> But the generic gpsdevice shouldn't know anything about ubx specific stuff, garmin specific stuff, sirf specific stuff, ...
Oct 05 12:59:18 <alphaone> That's even uglier
Oct 05 13:00:17 <alphaone> Initially I had the idea of separating ubx parsing from the actual ubx device, but I didn't have any compelling reasons at that time
Oct 05 13:01:16 <alphaone> I mean you don't really need or want all the handle_FOO methods either, do you?
Oct 05 13:01:49 <emdete> or you make those debug methods more abstrakt and allow that for others too...
Oct 05 13:02:12 <alphaone> Yes, for some things it makes sense
Oct 05 13:02:18 <alphaone> For others it doesn't
Oct 05 13:03:01 <alphaone> imho
Oct 05 13:03:16 <emdete> alphaone: then they are just empty
Oct 05 13:04:44 <emdete> alphaone: you have UBX in a class hierarchie but the channel itself not... wouldnt it be better (for "plugging" daemons together) to have both as member? does fso still support udp/gllin/gta01 btw?
Oct 05 13:06:17 <alphaone> emdete: The channels have a hierarchy, it's just very flat. :-)
Oct 05 13:06:33 <alphaone> Everything is a subclass of gpschannel
Oct 05 13:06:58 <alphaone> And yes, we need a gta01channel that takes care of starting and stopping gllin
Oct 05 13:07:15 <alphaone> for gta02 a normal serialchannel is enough
Oct 05 13:07:51 <emdete> isnt everything a subclass of gpsdevice, not channel? ...and channel is a member?
Oct 05 13:07:56 <alphaone> gta01channel would be a udpchannel that also manages gllin
Oct 05 13:08:00 <alphaone> no
Oct 05 13:08:10 <alphaone> a gpsdevice has a gpschannel, yes
Oct 05 13:08:21 <alphaone> but Serialchannel is a gpschannel
Oct 05 13:08:28 <alphaone> It's all in gpschannel.py
Oct 05 13:08:31 <emdete> ...und UBX... is a child of GPSDevice
Oct 05 13:08:35 <alphaone> I think 4 classes
Oct 05 13:08:39 <alphaone> yes
Oct 05 13:08:59 <alphaone> and GPSDevice -> UBXDevice -> GTA02Device
Oct 05 13:09:04 <emdete> what do you mean with "all" - i meant the dbus object itself
Oct 05 13:09:16 <emdete> yes, thats what i was refering to
Oct 05 13:09:31 <alphaone> Hmm, what's your point?
Oct 05 13:09:46 <alphaone> the channels are just so I can have different inputs
Oct 05 13:10:16 <emdete> woulnt it be better (for plugging) if you would have a dbus-object GPSSomeThing that has two member, the channel and the device?
Oct 05 13:10:20 <alphaone> i.e. I can use a FileChannel to open a dump of NMEA and test the NMEADevice
Oct 05 13:10:33 <emdete> shure, i know that
Oct 05 13:10:46 <alphaone> Or I can use the same NMEADevice with the GTA01Channel to have it use gllin
Oct 05 13:10:54 <emdete> but you can't plug in another interpreter on the fly
Oct 05 13:10:59 <alphaone> I don't think so, no
Oct 05 13:11:14 <alphaone> Why would you want to do that?
Oct 05 13:11:18 <emdete> that's my point (pyneo suffers from the same limitation :D )
Oct 05 13:11:28 <alphaone> Well, yes
Oct 05 13:11:46 <emdete> just put 2 settings into the config and get those objects at runtime. derivating classes isn't at runtime
Oct 05 13:11:50 <alphaone> Supporting the Gypsy way (opening different devices) is on my radar
Oct 05 13:11:56 <alphaone> But not really high :-)
Oct 05 13:12:19 <emdete> pyneo does so already and has more than one dbus object implementing that if
Oct 05 13:12:41 <alphaone> It's already configurable
Oct 05 13:12:52 <alphaone> in /etc/frameworkd.conf
Oct 05 13:13:03 <emdete> ...and loads a complete differetn class tree, yes?
Oct 05 13:14:38 <alphaone> emdete: Just look at factory.py
Oct 05 13:16:02 <emdete> i did already, shure ;) i love factory-factories...
Oct 05 13:28:23 <GNUtoo-desktop> hello, i missed a call with FSO(testing version) and I did an AT+CPBS="ME" then AT+CPBR=? then AT+CPBR=1,10 all with the framework gsm debug...and it didn't give me the missed call
Oct 05 13:30:54 <mickey|tw> if you want to record missed calls, you need to write that functionality
Oct 05 13:31:07 <mickey|tw> listen for signals, then write it somewhere
Oct 05 13:31:28 <mickey|tw> we didn't do that yet since we have no PIM storage where we could save the data
Oct 05 13:31:35 <mickey|tw> emdete: seen abraxa lately?
Oct 05 13:32:06 <emdete> mickey|tw: not irl but on im, shure. he is back in germany. 
Oct 05 13:32:48 <mickey|tw> any idea whether he had a chance to work more on pimd?
Oct 05 13:41:08 <GNUtoo-desktop> ok thanks a lot...so we can't save missed calls on the sim?
Oct 05 13:41:24 <GNUtoo-desktop> because ME is not writable?
Oct 05 13:41:32 <GNUtoo-desktop> s/ME/MC
Oct 05 13:42:53 <mickey|tw> well, usually smartphones handle this not by using the SIM
Oct 05 13:43:07 <emdete> mickey|tw: i think so, but he is working on some arm project at the univeristy... dont know exactly how far pimd is right now
Oct 05 13:43:25 <mickey|tw> emdete: hmm, ok. will ping him when I'm back in .de
Oct 05 13:43:26 <emdete> mickey|tw: anyway this storage should be filled by the modem as far as i know...
Oct 05 13:43:33 <GNUtoo-desktop> ah ok...my phone(not smart) did this with the sim...
Oct 05 13:43:45 <GNUtoo-desktop> s/phone/old phone/
Oct 05 13:44:56 <mickey|tw> let me check
Oct 05 13:45:01 * mickey|tw selects MC
Oct 05 13:50:11 * GNUtoo-desktop must go cycling
Oct 05 13:50:18 <GNUtoo-desktop> (bycicle)
Oct 05 13:50:42 <GNUtoo-desktop> s/bycicle/bicycle
Oct 05 13:53:30 <mickey|tw> heh
Oct 05 13:53:32 <mickey|tw> it works
Oct 05 13:53:33 <mickey|tw> cool
Oct 05 13:53:38 <mickey|tw> you always learn something new
Oct 05 13:53:43 <emdete> shure ;)
Oct 05 13:53:58 <emdete> mickey|tw: did you ever change sim while the neo is powered on?
Oct 05 13:55:40 <TeXitoi> hi
Oct 05 13:57:05 <mickey|tw> no, i don't think that's advisable
Oct 05 13:57:23 <mickey|tw> although the calypso has support for that, i don't think we have connected a SIM socket that supports that
Oct 05 13:58:49 <mwester> You'd have to remove the battery first, which would also be inadvisable, I guess.
Oct 05 14:00:12 <mickey|tw> well
Oct 05 14:00:19 <mickey|tw> gsm powers down on removing the battery
Oct 05 14:00:20 <mickey|tw> which is ok
Oct 05 14:00:25 <alphaone> mickey|tw: I changed SIM - the modem is off then anyway
Oct 05 14:00:31 <mickey|tw> but i don't know how all that is electrically connected
Oct 05 14:00:36 <alphaone> Went okay so far
Oct 05 14:00:55 <mickey|tw> still i wouldn't advise it :)
Oct 05 14:01:05 <mickey|tw> for mere mortals :)
Oct 05 14:01:27 <mwester> And especially don't do it on a GTA01; you'll let the magic smoke out of the regulator.
Oct 05 14:25:20 <sigius> Hi, i'd like to be able to use fingertyping in a terminal, so not using a stylus. To do so id like to have a typical cellphone keyboard as a transparent overlay on top of the terminal (or other application). Is such a thing available or being worked on already ??
Oct 05 14:26:44 <lindi-> sigius: true transparency is bit difficult
Oct 05 14:27:02 <lindi-> sigius: but i have done some prototyping, do you know python?
Oct 05 14:27:29 <sigius> lindi-,  hardly, I played around with it a bit
Oct 05 14:27:33 <lindi-> sigius: C?
Oct 05 14:27:49 <sigius> lindi-, yes
Oct 05 14:28:44 <lindi-> sigius: darcs get http://iki.fi/lindi/darcs/keyboard is a simple virtual keyboard prototype in C. it grabs the "mouse" so that it can see when you touch the screen but it does not draw anything over other apps unless you touch the screen
Oct 05 14:28:48 <LionKMP> Hi, opkg-target update displays 5 404 errors in toolchain, why?
Oct 05 14:29:13 <LionKMP> what are the correct sources and where to configure them? Example:  * Failed to download http://downloads.openmoko.org/repository/testing/noarch/Packages.gz, error 404
Oct 05 14:32:05 <sigius> lindi-, Ok, cool i'll check that out. So it shows the keyboard when you are touching the screen and then the underlying app (say terminal) when you release ? Is that correct ?
Oct 05 14:33:12 <lindi-> sigius: it shows only one letter at the time at the moment
Oct 05 14:33:46 <sigius> lindi-, are you still working on this at the moment or did you abandon it ?
Oct 05 14:33:48 <lindi-> sigius: drawing the whole keyboard every time you touch the screen would have been too slow with xserver-xorg-video-fbdev. now that i have the faster xserver-xglamo server it might be bit faster
Oct 05 14:33:56 <lindi-> sigius: it was a quick hack
Oct 05 14:34:07 <sigius> lindi-, ok i see. 
Oct 05 14:34:39 <lindi-> sigius: mostly just to figure out what things need to be taken into account
Oct 05 14:34:48 <lindi-> sigius: i'm currently using matchbox-keyboard with custom layout. it has some issues that are hard to fix
Oct 05 14:35:23 <lindi-> sigius: 1) the keyboard layout can not define absolute coordinates. i have some useless padding here and there and i can't figure out how to remove it
Oct 05 14:35:47 <sigius> lindi-, i'm think one could make the keyboard purely virtual (i.e. not have it draw anything at all to screen, but just read and interpret screen taps.) 
Oct 05 14:36:01 <sigius> s/think/thinking/
Oct 05 14:36:03 <lindi-> sigius: 2) it can't be configured to handle mod1-mod2-C differently from mod2-mod1-C
Oct 05 14:36:38 <lindi-> sigius: since ideally i would have a only two modifiers and could use it to produce different variations like C, mod1-C, mod2-C, mod1-mod2-C, mod2-mod1-C
Oct 05 14:37:01 <lindi-> sigius: you could but it is bit hard to learn to type with that
Oct 05 14:37:19 <sigius> lindi-, sorry I do not get the mod1-mdo2 bit. 
Oct 05 14:37:35 <sigius> lindi-, should take some getting used to,yes
Oct 05 14:37:44 <dos1> hello
Oct 05 14:37:47 <lindi-> sigius: like ctrl-alt-c and alt-ctrl-c could produce different symbol
Oct 05 14:38:02 <lindi-> sigius: on normal keyboard you usually don't want this but with touchscreen it would make sense
Oct 05 14:38:45 <sigius> lindi-, ok i see
Oct 05 14:38:53 <dos1> where can i find the most actual 2007.2 sources? (i am interested in openmoko-dialer2 and phone-kit)
Oct 05 14:39:25 <lindi-> dos1: mokomakefile is easiest way to find it
Oct 05 14:39:54 <lindi-> dos1: it will fetch bitbake recipes from different repos. those bitbake recipes then refer to the source code of the apps
Oct 05 14:49:50 <sigius> lindi-, btw what is darcs ?
Oct 05 14:50:25 <sigius> is that darcs the cm tool ?
Oct 05 14:55:03 <lindi-> sigius: yep
Oct 05 14:55:15 <lindi-> sigius: apt-get install darcs
Oct 05 14:57:25 <sigius> lindi-, right. I gather the keyboard (in main.c) you have in mind is a regular qwerty, not a cellphone keyboard, correct ?
Oct 05 14:57:46 <lindi-> sigius: making the layout configurable is a lot of work
Oct 05 15:02:58 <sigius> lindi-, How do you build this, i.e. using what toolchain etc ?
Oct 05 15:03:19 <lindi-> sigius: i build on FR, with qemubuilder or inside qemu
Oct 05 15:03:32 <sigius> FR ?
Oct 05 15:03:39 <lindi-> freerunner
Oct 05 15:04:01 <LionKMP> hah, easy solution. I also tried that but there is no gcc in Om 2008.8 :)
Oct 05 15:04:08 <sigius> right :) , ofcourse
Oct 05 15:04:09 <LionKMP> (i know I know, debian :) )
Oct 05 15:04:47 <LionKMP> may I ask one. Toolchain question. ---prefix=/someting/ does not work, because toolchain sets all destination directories. Hardcoded. 
Oct 05 15:04:52 <LionKMP> What to do?
Oct 05 15:05:23 <sigius> LionKMP, rebuild the toolchain if thats at all possible
Oct 05 15:05:30 <LionKMP> for testing it would be easier to do --prefix=/media/card 
Oct 05 15:06:12 <LionKMP> sigius: it's only a shell script which makes all those "--execdir=" and many more, so I could modify that script, but I don't see if there is a hidden reason to do that way.
Oct 05 15:06:14 <lindi-> just to speed things up i am planning to setup cross-compiling distcc with qemubuilder
Oct 05 15:06:43 <LionKMP> actually "om-conf" sets all the output dirs that hardcoded way.
Oct 05 15:07:19 <LionKMP> otherwise I could do "make install" on my host pc and tar-gz the result e.g. /media/card/testprogram 
Oct 05 15:08:08 <LionKMP> lindi-: you seems to be choosig the easiest way all the time. Cool :)
Oct 05 15:10:37 <LionKMP> checking for gtk-config... no
Oct 05 15:10:48 <LionKMP> is there a way to build this in toolchain ^^^^
Oct 05 15:11:15 <lindi-> (apt-get install libgtk1.2-dev..)
Oct 05 15:11:26 <LionKMP> :-D
Oct 05 15:11:46 <LionKMP> hm, host os? 
Oct 05 15:11:49 <LionKMP> or kidding?
Oct 05 15:11:59 <LionKMP> I don't think it wants my PC's gtk-config
Oct 05 15:12:11 <lindi-> LionKMP: no no but i do it natively on am
Oct 05 15:12:13 <lindi-> arm
Oct 05 15:12:15 <LionKMP> ok
Oct 05 15:12:34 <LionKMP> yes, that's the easy way. but I try using tool chains.
Oct 05 15:13:02 <dos1> ok guys, who are using 2007.2 and want to set different ringtones for incoming dial and sms? ;>
Oct 05 15:13:47 * LionKMP no
Oct 05 15:14:26 <sigius> lindi-, anything else byond gcc I need to opkg to build main.c
Oct 05 15:15:31 <dos1> here - http://openmoko.opendevice.org/~dos/phone-kit - is a patched version of phone-kit
Oct 05 15:16:13 <LionKMP> wwowowow, I crosscompiled Xbill in Toolchain. :DDDD 
Oct 05 15:16:35 <dos1> in /etc/pulse/session you can set load-sample ringtone_sms /path/to/your/sms/ringtone.wav
Oct 05 15:16:51 <dos1> restart pulseaudio and X, and you have different ringtone for sms :)
Oct 05 15:18:17 <lindi-> sigius: you need package with gtk+-2.0.pc
Oct 05 15:18:30 <lindi-> sigius: libfakekey.pc
Oct 05 15:18:40 <lindi-> sigius: pkg-config
Oct 05 15:18:50 <lindi-> sigius: those three
Oct 05 15:19:33 <quickdev> Hey
Oct 05 15:20:18 <quickdev> Ainulindale, btw., I got retrieve_messagebook to work
Oct 05 15:20:57 <sigius> lindi-, Ok thanks. I hope thats all, i noticed I was even missing 'make' :)
Oct 05 15:21:06 <lindi-> sigius: oh, and make too
Oct 05 15:40:36 <quickdev> How do I make my FSO kernel suspend?
Oct 05 15:41:10 <alphaone> The question is how do you make it wake up
Oct 05 15:42:18 <quickdev> alphaone, how long do I have to press the power button to suspend it? :)
Oct 05 15:42:35 <alphaone> quickdev: I think we don't suspend at the moment
Oct 05 15:42:43 <quickdev> ah, ok
Oct 05 15:42:49 <quickdev> alphaone, kernel issue, right?
Oct 05 15:42:51 <alphaone> because resume is broken (or at least has been for a long time)
Oct 05 15:50:14 <quickdev> alphaone, is there a way to suspend from command line?
Oct 05 15:51:32 <alphaone> apm -a
Oct 05 15:51:34 <alphaone> apm -s
Oct 05 15:51:42 <alphaone> echo mem > /sys/power/state
Oct 05 15:51:44 <alphaone> should work
Oct 05 15:55:31 <quickdev> alphaone, suspend and wake up works for me. Is there a way to enable it after 3 seconds power button press? :)
Oct 05 15:58:23 <lindi-> alphaone: that does not properly configure GSM for suspend
Oct 05 15:58:39 <lindi-> quickdev: zhone does that, sure it is possible
Oct 05 15:58:45 <alphaone> lindi-: I am aware of that
Oct 05 15:58:58 <quickdev> ah, thanks for the hint
Oct 05 15:59:16 <alphaone> quickdev: You need to call Suspend (or similar) on the Usage interface
Oct 05 15:59:46 <alphaone> That should let the framework take care of suspend
Oct 05 16:03:27 <quickdev> I wonder why I haven't heard a complaint about the too small "qwerty" and the missing "Exit" button (without expanding the bar) of the illume theme.
Oct 05 16:09:51 <quickdev> alphaone, http://git.freesmartphone.org/?p=specs.git;a=blob_plain;f=html/org.freesmartphone.Usage.html;hb=HEAD - Suspend is not listed although it is used in frameworkd. I'll force mickey to add that tomorrow :)
Oct 05 16:10:13 <alphaone> quickdev: Do that :-)
Oct 05 16:10:13 <Sup3rkiddo> alphaone, the Usage interface is supposed to offer a unified way of handling other subsystems?
Oct 05 16:10:36 <alphaone> Sup3rkiddo: It is for handling resources
Oct 05 16:10:45 <alphaone> Which is not really subsystems
Oct 05 16:11:19 <Sup3rkiddo> alphaone, yes, so currently for me, I see only resources from the odeviced subsystem when I invoke ListResources
Oct 05 16:12:00 <alphaone> Sup3rkiddo: ogsmd and ogpsd should register their resources as well
Oct 05 16:12:26 <Sup3rkiddo> alphaone, ah ok, since i am running this on my laptop, thats no wonder why they are not registering
Oct 05 16:14:40 <alphaone> okay, makes sense
Oct 05 16:15:11 <quickdev> just abused a .desktop file to suspend, hehe
Oct 05 16:17:25 <alphaone> heh
Oct 05 16:18:40 <dos1> wibruje
Oct 05 16:18:47 <dos1> ups, not here :x
Oct 05 16:26:04 <quickdev> mwester, did you test suspend and the AT%SLEEP=2 patch?
Oct 05 16:28:51 <lindi-> quickdev: what's AT%SLEEP? google does not find it
Oct 05 16:30:38 <quickdev> lindi-, yesterday we found out, that the oscillating gsm bug (ticket #1024) is only appearing, when gsm goes into deep sleep. AT%SLEEP=2 disables deep sleep. mwster found that command in the enfora enabler docs, a similar device.
Oct 05 16:32:06 <lindi-> quickdev: but that should only be done during calls to save power?
Oct 05 16:32:33 <dos1> quickdev: i think this is not kernel job
Oct 05 16:33:00 <lindi-> quickdev: ah it's not about voice quality during call
Oct 05 16:33:04 <quickdev> lindi-, that should only be done on phones where the oscillating gsm bug appears. During calls, the gsm is stable. I has to be sent at the beginning to stop oscillating.
Oct 05 16:33:34 <quickdev> oscillating = drop and reregister network many times in an hour
Oct 05 16:34:03 <LionKMP> is there a page on Track where I could type in "1024" to view that ticket?
Oct 05 16:34:15 <LionKMP> (other than URL bar)
Oct 05 16:34:26 <quickdev> http://docs.openmoko.org/trac/ticket/1024
Oct 05 16:34:43 <LionKMP> is there a page on Track where I could type in "1024" to view that ticket?
Oct 05 16:35:20 <lindi-> LionKMP: the search field at bugs.openmoko.org
Oct 05 16:35:27 <quickdev> LionKMP, enter "#1024" into the search field, hehe
Oct 05 16:36:12 <LionKMP> ah, ok thanks. I tried without #. :) thanks!
Oct 05 16:37:49 <mwester> http://moko.mwester.net/brc.html also has information on the problem.
Oct 05 16:38:43 <mwester> quickdev:  Yes, it sleeps very nicely with SLEEP set to 2
Oct 05 16:39:03 <mwester> The only noticible difference (which could be important to developers)
Oct 05 16:39:20 <mwester> is that with SLEEP set to 2, the UART never goes into the mode where it eats the first character.
Oct 05 16:39:45 <quickdev> mwester, which kernel to you use?
Oct 05 16:40:38 <mwester> I use my own as I never really know what is and is not in the official moko kernels; it's basically the offical git repo's "stable" branch with some debug turned on.
Oct 05 16:40:52 <lindi-> mwester: hrm, it's really irritating we can't look at the gsm firmware
Oct 05 16:41:32 <mwester> There was a very unfortunate commit to the stable not too long ago that broke resume on the gta02; stable has been fixed but I don't know if the change has made it into any of the distros yet.
Oct 05 16:42:50 <mwester> lindi-:  mickey||tw looked into the source code that Om has, and the %SLEEP command is not even in that source -- so nobody except TI knows that command exists.  We are coding to that GSM sometimes by sheer luck.
Oct 05 16:43:36 <lindi-> mwester: do you remember where in http://www.enfora.com/index.cgi?CONTENT_ID=11&User:LANGUAGE=en they mention AT%SLEEP?
Oct 05 16:43:46 <lindi-> mwester: or should i just wget -r && grep?
Oct 05 16:44:23 <lindi-> mwester: is it possible to get access to the binary firmware image somehow?
Oct 05 16:44:32 <lindi-> grep from that should find it?
Oct 05 16:45:46 <mwester> http://www.enfora.com/index.cgi?CONTENT_ID=1350&User:LANGUAGE=en   <--- First document listed.
Oct 05 16:46:15 <mwester> Check the ML; somebody with some fake id posted notes on how to read the GSM firmware out.
Oct 05 16:47:06 <mwester> That's really more effort in hardware and software than I want to do (even though it might have actually been easier than the trial and error process I used).
Oct 05 16:47:16 <lindi-> mwester: http://n2.nabble.com/GSM-firmware-hacking-td684093.html?
Oct 05 16:49:13 <mwester> Yeah
Oct 05 17:47:31 <GNUtoo> i'm back
Oct 05 17:53:57 <GNUtoo> mickey|tw, so what did your checks return?
Oct 05 19:23:47 <admiral0> hi
Oct 05 19:24:03 <admiral0> anybody from debian 4 FR here?
Oct 05 19:27:09 <nomeata> depends on your question :-)
Oct 05 19:53:26 <bbs> i'm getting a neo ! 
Oct 05 19:53:40 * bbs gets ready for working on a email app
Oct 05 19:53:56 <bbs> is there a preferred language for apps here? python c++ c?
Oct 05 19:58:27 <admiral0> nomeata: ihave repo problems with debian
Oct 05 19:59:38 <nomeata> admiral0: which problems exactly?
Oct 05 20:01:18 <admiral0> impossible to install zhone session
Oct 05 20:05:07 <admiral0> it just says it's broken
Oct 05 20:07:32 <admiral0> ...
Oct 05 20:08:19 <admiral0> it's exactly that http://groups.google.com/group/linux.debian.bugs.dist/browse_thread/thread/45fee418891a0b7d
Oct 05 20:08:26 <neolynx> admiral0: debian unstable is waiting for some packages
Oct 05 20:08:55 <admiral0> ah
Oct 05 20:08:55 <neolynx> python dbus stuff IIRC
Oct 05 20:09:08 <admiral0> ok
Oct 05 20:09:20 <admiral0> when will it be ok then?
Oct 05 20:10:55 <neolynx>  zhone: Depends: python-edbus but it is not going to be installed
Oct 05 20:10:55 <neolynx>          Depends: python-ecore but it is not installable
Oct 05 20:10:55 <neolynx>          Depends: python-edje but it is not going to be installed
Oct 05 20:10:55 <neolynx>          Depends: python-evas but it is not going to be installed
Oct 05 20:10:57 <neolynx> ...
Oct 05 20:11:14 <neolynx> you'll have to wait til the packages are uploaded to unstable
Oct 05 20:11:30 <admiral0> thank you.
Oct 05 20:12:06 <admiral0> do you know how much i'll have to wait? just info
Oct 05 20:12:18 <neolynx> I'm waiting as well :) no idea
Oct 05 20:12:52 <admiral0> does zhone-session get you a keyboard?
Oct 05 20:13:33 <lindi-> admiral0: matchbox-keyboard is the keyboard
Oct 05 20:14:16 <admiral0> ok
Oct 05 20:14:25 <admiral0> i was in trouble with midori 
Oct 05 20:14:28 <admiral0> lol
Oct 05 20:14:37 <neolynx> admiral0: maybe you can install the packages manually from http://pkg-fso.alioth.debian.org/debian/pool/main/p/python-edbus/ with dpkg
Oct 05 20:15:03 <IOhannes> hi, is there a way to read the volume an audio-signal during a call (e.g. i want to monitor the level of the outgoing signal)
Oct 05 20:15:31 <IOhannes> i would be glad with either high-level access to this data or low level access to the pcm data
Oct 05 20:15:59 <IOhannes> but i guess the audio-device will be blocked (?)
Oct 05 20:16:12 <admiral0> yeah
Oct 05 20:16:18 <lindi-> neolynx: why would that help?
Oct 05 20:16:41 <neolynx> lindi-: manually downgrade ? dont know
Oct 05 20:17:16 <lindi-> admiral0: you can workaround libcaca0 bug by using beta14 version of libcaca0
Oct 05 20:18:00 <admiral0> doh
Oct 05 20:18:13 <admiral0> i got midori working 
Oct 05 20:18:20 <admiral0> now i am a happy person
Oct 05 20:18:48 <admiral0> lol
Oct 05 20:21:02 <admiral0> as i see it here it's a dependency hell
Oct 05 20:21:31 <admiral0> libcaca0 and libcucul0 depend on each other
Oct 05 20:22:26 <lindi-> admiral0: no they don't.
Oct 05 20:23:37 <lindi-> admiral0: apt-cache show libcucul0
Oct 05 20:23:47 <admiral0> wait
Oct 05 20:23:55 <admiral0> i'm apt-get updating
Oct 05 20:24:07 <admiral0> maybe something got fucked up
Oct 05 20:25:48 <admiral0> Depends: libcaca0 (>= 0.99.beta15-1)
Oct 05 20:26:22 <lindi-> admiral0: but read the package description
Oct 05 20:26:36 <admiral0> ah k
Oct 05 20:27:16 <lindi-> admiral0: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=500492 is the bug
Oct 05 20:27:54 <lindi-> admiral0: as you can see it has been fixed and a fixed package has been built: http://buildd.debian.org/fetch.cgi?&pkg=libcaca&ver=0.99.beta15-2&arch=arm&stamp=1223227047&file=log
Oct 05 20:28:59 <lindi-> however the fixed package is not yet in ftp://ftp.debian.org/debian/pool/main/libc/libcaca
Oct 05 20:29:33 <admiral0> ok
Oct 05 20:29:35 <lindi-> admiral0: you have multiple options. 1) wait 2) build the package yourself with the patch 3) build an older version yourself
Oct 05 20:29:42 <lindi-> admiral0: i personally did 3)
Oct 05 20:30:31 <admiral0> hmm
Oct 05 20:30:36 <admiral0> i prefer 1
Oct 05 20:30:47 <admiral0> i'm not good at cross compiling
Oct 05 20:31:06 <admiral0> and i do not want dev libs on my freerunner
Oct 05 20:31:14 <lindi-> admiral0: i used qemubuilder
Oct 05 20:31:27 <lindi-> admiral0: it's really simple
Oct 05 20:31:40 <admiral0> qemubuilder?
Oct 05 20:31:45 <lindi-> yeah
Oct 05 20:31:47 <lindi-> apt-get install qemubuilder
Oct 05 20:32:13 <lindi-> wget ftp://ftp.debian.org/debian/pool/main/libc/libcaca/libcaca_0.99.beta14-1.diff.gz ftp://ftp.debian.org/debian/pool/main/libc/libcaca/libcaca_0.99.beta14-1.dsc ftp://ftp.debian.org/debian/pool/main/libc/libcaca/libcaca_0.99.beta14.orig.tar.gz
Oct 05 20:32:25 <lindi-> sudo qemubuilder --build libcaca_0.99.beta14-1.dsc
Oct 05 20:32:40 <lindi-> => tada, you have caca-utils_0.99.beta14-1_armel.deb libcaca-dev_0.99.beta14-1_armel.deb libcaca0_0.99.beta14-1_armel.deb libcucul-dev_0.99.beta14-1_armel.deb libcucul0_0.99.beta14-1_armel.deb
Oct 05 20:32:50 <admiral0> wow
Oct 05 20:33:26 <lindi-> admiral0: you need debian unstable, see the instructions at http://wiki.debian.org/DebianOnFreeRunner
Oct 05 20:34:47 <lindi-> admiral0: and feel free to ask questions, we need more people who want to learn how to fix bugs
Oct 05 20:35:51 <admiral0> :D
Oct 05 20:37:00 <lindi-> admiral0: but that is the truth. i am using debian since with it i can fix bugs :) I haven't managed to even compile most other distros for openmoko
Oct 05 20:37:40 <admiral0> moko makefile requires very much to build
Oct 05 20:39:24 <admiral0> i installed iceweasel 
Oct 05 20:39:27 <admiral0> am i nuts?
Oct 05 20:39:45 <admiral0> lol
Oct 05 20:41:02 <lindi-> admiral0: mokomakefile build unfortunately often fails
Oct 05 20:41:06 <lindi-> admiral0: heh :)
Oct 05 20:41:11 <lindi-> i have used midori for now
Oct 05 20:42:15 <admiral0> it works
Oct 05 20:42:21 <admiral0> pretty well
Oct 05 20:42:34 <admiral0> i built shr one time
Oct 05 20:42:50 <admiral0> it failed 3 tasks before completion
Oct 05 20:43:24 <lindi-> admiral0: pretty well yes. qemubuilder works every time :-P
Oct 05 20:44:06 <lindi-> admiral0: qemubuilder builds the package in a clean environment and installs only packages that have been marked as Build-Depends. this way nothing in your normal system can affect the produced package
Oct 05 20:44:36 <admiral0> that's good
Oct 05 20:44:58 <lindi-> of course it is much slower than cross compilation then
Oct 05 20:45:08 <lindi-> but you only need to compile the packages you need, not everything from scratch
Oct 05 20:46:06 <admiral0> i noticed a bad thing
Oct 05 20:46:18 <admiral0> there's no wicd in the repo
Oct 05 20:50:22 <admiral0> ah ok
Oct 05 20:50:30 <admiral0> at least networkmanager
Oct 05 20:55:49 <tanner> regarding stability of FSOMakeFile, am i doing something wrong or is there a known bug? Whenever i try to build fso-{testing,unstable}-image it dies building glibc with an error from localedef which somehow can't generate locale for es_NI.
Oct 05 20:58:47 <mwester> Sounds like Qemu won't run on your system.  Is it a Fedora or Red Hat distro?
Oct 05 20:59:50 <tanner> debian unstable
Oct 05 21:01:09 <mwester> Ok.  Check the log files, and try to sort your qemu problem (there's a security change in most recent kernels that breaks qemu -- google for the sysctl change to make it work again).
Oct 05 21:01:19 <mwester> Or just disable locale generation for your build.
Oct 05 21:02:51 <tanner> ok i will try to fix it i am running a more or less vanilla 2.6.26 kernel
Oct 05 21:03:27 <mwester> many people don't bother with locale generation; it's much faster to build that way.
Oct 05 21:05:49 <tanner> ok but i guess the sysctl change is needed anyway, right?
Oct 05 21:07:28 <LionKMP> GLIBC_GENERATE_LOCALES = "en_US.UTF-8" maybe?
Oct 05 21:12:12 <tanner> in openembedded/packages/glibc/glibc.inc ?
Oct 05 21:12:25 <mwester_> No, it should be on the wiki
Oct 05 21:14:09 <mwester_> ENABLE_BINARY_LOCALE_GENERATION = "0"
Oct 05 21:14:14 <mwester_> in the local.conf file
Oct 05 21:15:12 <tanner> ok will try it, thanks
Oct 05 21:18:21 <admiral0> night everybody
Oct 05 21:45:31 <gismo> nomeata: still awake?
Oct 05 21:45:49 <nomeata> gismo: barely, I was about to leave
Oct 05 21:46:28 <gismo> nomeata: pvt :-D
Oct 05 21:46:47 <quickdev> Hey
Oct 05 21:46:47 <nomeata> http://git.debian.org/?p=pkg-fso/fso-frameworkd.git;a=commitdiff;h=f78af4c5dcf5e48a8b6465f3722f920fd8a38f95 is still needed, because the default python encodings differ on FSO’s image and Debian, IIRC
Oct 05 21:46:52 <nomeata> pvt?
Oct 05 21:47:02 <gismo> nomeata: private 
Oct 05 21:47:45 <gismo> nomeata: it's quite a long time since I've ever tested an upstream FSO ;-)
Oct 05 23:54:42 <quickdev> Ainulindale, here?
**** ENDING LOGGING AT Mon Oct 06 02:59:57 2008