Re: fso-gsmd: incorrect GSM channel release on Openmoko devices
On Mon, Nov 07, 2011 at 08:28:42PM +0100, Simon Busch wrote: On 06.11.2011 19:11, Sebastian Reichel wrote: Hi, This is a forward of Debian Bug #647277 [0]: The first time I request the GSM resource, it works well. Then I release the resource to save battery; after that, every GSM resource request will fail leaving the resource unusable. I have to restart fsogsmd to recover. The error message from fso-gsmd is: [ERROR] libfsotransport 0710:2: Can't allocate channel #2 from MUX: Channel 2 is already taken. [ERROR] TiCalypsoModem 4C: Can't open main channel; open returned false [ERROR] DBusServiceDevice : Can't open modem [0] http://bugs.debian.org/647277 Can you please do two things: 1. Create a bug report in our trac system (see [1]) ok, but I can't reach fso.org at the moment. 2. Which version of FSO is used? As we're not releasing our components so often I think you're using some git revision to build the debian packages from. fso-gsmd:2011-09-16 libfsobasics:2011-08-04 libfsoframework: 2011-08-05 libfsoresource: 2011-08-05 libfsotransport: 2011-08-05 libgsm0710mux: 2011-08-05 -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: Debian on Freerunner, install.sh fails
On Sat, Apr 23, 2011 at 11:33:48AM +0200, Frank Jäger wrote: Hallo, I tried to install a Debian on the uSD-Card of my Freerunner. I followed: http://wiki.debian.org/DebianOnFreeRunner#Current_Status_of_Installation_via_install.sh This end up with: P: Retrieving libc6 E: Couldn't download pool/main/e/eglibc/libc6_2.11.2-13_armel.deb! perhaps a temporary problem, or a misconfigured mirror? It loads perfectly fine from [1]. [1] http://ftp.de.debian.org/debian/pool/main/e/eglibc/libc6_2.11.2-13_armel.deb -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [PATCH 2/2] libfsobasics: ReactorChannel: add new constructor
On Fri, Aug 27, 2010 at 05:56:56PM +0200, Simon Busch wrote: On 27.08.2010 15:12, Dr. Michael Lauer wrote: Applied both patches. Sorry for the delay and thanks! thanks for applying it :) They are needed for N900 proximity plugin and will also be needed for some gpio based buttons of the N900. Could you give us a summary about the state of N900 support? I talked with Sebastion already about his efforts on getting FSO support for the N900. The big blocker was as it should in most devices: The modem. There was some logic missing which enables the modem. But now there some code in ofono which we can port to a lowlevel_n900 plugin for fsogsmd which does all the magic. It only writes to some sysfs entries. This should be only a very small part of the integration work. It's not yet in ofono, but available as patch on their mailinglist. Putting this aside you are right :) The biggest thing for getting N900 modem support in FSO is the protocol. It's called isi or phonet, I don't really now. But there is code too in ofono which we can extract and use in fsogsmd. Maybe Sebastian can tell us more in detail how complex this is. ok, let me start with explaining the phonet/isi name: Phonet is an abstract protocol from the Linux kernel: http://lwn.net/Articles/303274/ ISI (Intelligent service interface) is the protocol spoken by the modem of the N900, which is phonet + definition of appended data. The next big thing for the N900 was the battery, but as I found out there is now a script which does the battery handling. Don't know which features it supports actually, but it activates the battery via i2c. Maybe nothing we really want in directly in FSO. actually it is not about the battery, but just about charging the battery. Reading the battery status is already supported by FSO after loading the bq27x00_battery kernel driver. The current script charges the battery until it's full and exit. The specs for the charging chip are publically available from TI by the way. About the other components I don't know, but Sebastion has made a great overview which components has which status actually. You will find it here: https://elektranox.org/n900/status/index.html You can see what is needed for fsodeviced here: https://elektranox.org/n900/todo.html As Sebastian living only some steps away from me, he and me will meet in the near future and talk a little bit about hacking in FSO. Maybe I can help him to bring N900 support to FSO. I'll definitively need some help for the fsogsmd plugin. I'll contact you via xmpp for a date :) Could you give us a summary about the state of N900 support? For the modem I made good progress in the last days after my chat with Simon: I wrote myself a small skript which does the sysfs modem enabling and then concentrated on the stuff happening afterwards. I was able to enter my SIM's PIN and do different Network related stuff (e.g. getting a list of available providers). See my test program here: https://git.ring0.de/libisi/tree/test/test.vala I have not yet tried to integrate this into fsogsmd and for now my plans are to improve the library: https://git.ring0.de/libisi/plain/TODO Chances are I'll get a N900 soon and I can help with a couple of things in FSO2. cool :) -- Sebastian [1] https://elektranox.org/n900/status/index.html [2] https://elektranox.org/n900/todo.html signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
[PATCH 3/3] avoid code duplication
From: Sebastian Reichel s...@ring0.de Signed-off-by: Sebastian Reichel s...@ring0.de --- libfsobasics/fsobasics/utilities.vala | 20 +++- 1 files changed, 11 insertions(+), 9 deletions(-) diff --git a/libfsobasics/fsobasics/utilities.vala b/libfsobasics/fsobasics/utilities.vala index ddf8398..683c4d8 100644 --- a/libfsobasics/fsobasics/utilities.vala +++ b/libfsobasics/fsobasics/utilities.vala @@ -525,26 +525,28 @@ namespace FsoFramework { namespace Async { private char[] buffer; private bool rewind; -public ReactorChannel( int fd, owned ActionFunc actionfunc, size_t bufferlength = 512 ) +private init( int fd, owned ActionFunc actionfunc, size_t bufferlength = 512 ) { assert( fd -1 ); -channel = new GLib.IOChannel.unix_new( fd ); -watch = channel.add_watch( GLib.IOCondition.IN | GLib.IOCondition.HUP, onActionFromChannel ); this.fd = fd; this.actionfunc = actionfunc; -this.rewind = false; buffer = new char[ bufferlength ]; } -public ReactorChannel.rewind( int fd, owned ActionFunc actionfunc, size_t bufferlength = 512 ) +public ReactorChannel( int fd, owned ActionFunc actionfunc, size_t bufferlength = 512 ) { -assert( fd -1 ); +init( fd, actionfunc, bufferlength ); +watch = channel.add_watch( GLib.IOCondition.IN | GLib.IOCondition.HUP, onActionFromChannel ); channel = new GLib.IOChannel.unix_new( fd ); +this.rewind = false; +} + +public ReactorChannel.rewind( int fd, owned ActionFunc actionfunc, size_t bufferlength = 512 ) +{ +init( fd, actionfunc, bufferlength ); watch = channel.add_watch( GLib.IOCondition.IN | GLib.IOCondition.PRI | GLib.IOCondition.HUP, onActionFromChannel ); -this.fd = fd; -this.actionfunc = actionfunc; +channel = new GLib.IOChannel.unix_new( fd ); this.rewind = true; -buffer = new char[ bufferlength ]; } public int fileno() -- 1.7.1 ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
[PATCH 1/2] libfsobasics: ReactorChannel: add sysfs capability
From: Sebastian Reichel s...@ring0.de * Add rewind option * watch PRI in addition to IN This is needed to watch some sysfs files on the N900, for example this one: /sys/devices/platform/gpio-switch/proximity/state Signed-off-by: Sebastian Reichel s...@ring0.de --- libfsobasics/fsobasics/utilities.vala | 10 +++--- 1 files changed, 7 insertions(+), 3 deletions(-) diff --git a/libfsobasics/fsobasics/utilities.vala b/libfsobasics/fsobasics/utilities.vala index 1a2008b..21bbcab 100644 --- a/libfsobasics/fsobasics/utilities.vala +++ b/libfsobasics/fsobasics/utilities.vala @@ -523,14 +523,16 @@ namespace FsoFramework { namespace Async { private GLib.IOChannel channel; private ActionFunc actionfunc; private char[] buffer; +private bool rewind; -public ReactorChannel( int fd, owned ActionFunc actionfunc, size_t bufferlength = 512 ) +public ReactorChannel( int fd, owned ActionFunc actionfunc, size_t bufferlength = 512, bool rewind = false ) { assert( fd -1 ); channel = new GLib.IOChannel.unix_new( fd ); -watch = channel.add_watch( GLib.IOCondition.IN | GLib.IOCondition.HUP, onActionFromChannel ); +watch = channel.add_watch( GLib.IOCondition.IN | GLib.IOCondition.PRI | GLib.IOCondition.HUP, onActionFromChannel ); this.fd = fd; this.actionfunc = actionfunc; +this.rewind = rewind; buffer = new char[ bufferlength ]; } @@ -559,10 +561,12 @@ namespace FsoFramework { namespace Async { return false; } -if ( ( condition IOCondition.IN ) == IOCondition.IN ) +if ( ( ( condition IOCondition.IN ) == IOCondition.IN ) || + ( ( condition IOCondition.PRI ) == IOCondition.PRI ) ) { assert( fd != -1 ); assert( buffer != null ); +if( rewind ) Posix.lseek(fd, 0, Posix.SEEK_SET); ssize_t bytesread = Posix.read( fd, buffer, buffer.length ); actionfunc( buffer, bytesread ); return true; -- 1.7.1 ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [pkg-fso-maint] Debian FSO packaging graph
On Sun, Aug 01, 2010 at 03:14:02PM -0400, Jonas Smedegaard wrote: Hi all, I now updated the dependency graph for Debian packages: * dotting fso-gsmd (only in experimental) * adding/undotting libphone-ui-shr and phoneuid Please do tell me if this graph is helpful to others than myself in understanding the relationship and status of FSO and SHR packaging. And of course tell me if I made some mistakes in it. That's right so far; phoneui-apps are added once upstream made it's licensing clear (I filled some bugs in their trac). The arrow phonefsod - phoneui-apps can be removed though. There will be only an indirect dependency phoneui-apps - phoneuid - phonefsod. For fso-gsmd the more important fact is, that experimental does not build any armel packages :( I'm currently trying to find a solution for this. [1] http://wiki.debian.org/Teams/DebianFSO?action=AttachFiledo=gettarget=pkgdeps.png -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [pkg-fso-maint] investigating FTBFS of libfsotransport on sparc
On Thu, Jun 10, 2010 at 08:59:00PM +0200, Heiko Stübner wrote: the newer libfsotransport upload FTBFS on sparc and I'm not really sure whom I should pester with it or what changes are appropriate. Hi Heiko, First of all you should have CC'd smartphones-userland. Mickey is interested in this stuff ;) I just pointed him to your mail: 21:25 mickey Bitte antworte mal für mich, ich würde das mit dem FAST_BAUD conditional (via autotools --disable-fast-baud) unterstützen 21:25 mickey Sehe keinen Grund, SPARC nicht zu supperten 21:26 mickey err, support'en For the non German speaking in this mailinglist: He would add --disable-fast-baud to the configure script to solve the SPARC problem. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: fso-abyss
On Tue, Jun 08, 2010 at 12:15:25PM +0200, Jonas Smedegaard wrote: On Tue, Jun 08, 2010 at 11:16:25AM +0200, Dr. Michael Lauer wrote: Am 08.06.2010 um 09:11 schrieb Heiko Stuebner: does someone still work on abyss or is it completly dead? Well, since fso-abyss is just a standalone version of libgsm0710mux, which I still maintain, we can consider it being alive, although somewhat deprecated, since we no longer use the pty redirection in fsogsmd. I am confused. Which alternative package provide the functionality of fso-abyss and is also based on libgsm0710mux? fsogsmd, which would be Debian package 'fso-gsmd'. Perhaps something not yet officially in Debian? Or perhaps package dependencies are wrong somewhere? fso-gsmd is not yet packaged in Debian, since it needs a more recent Vala version than available in Debian. Perhaps look at the dependency graph at http://wiki.debian.org/Teams/DebianFSO#FSO.2B-SHRstack and clarify how libgsm0710mux is possible to be present if not throuhg fso-abyss, and we can try derive from such explanation if we need to upgrade/add packages, fix dependencies or I perhaps I just need to fix my understanding of all this :-) I have already prepared fso-gsmd, it will be available in Debian main once a new Vala version has been uploaded. FYI: One has been released yesterday ;) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: fso-frameworkd - ogpsd as a gpsd client
On Thu, May 13, 2010 at 10:18:15PM +0100, Al Johnson wrote: On Thursday 13 May 2010, Dr. Michael Lauer wrote: Am 13.05.2010 um 00:08 schrieb Al Johnson: On Wednesday 12 May 2010, Enrico Zini wrote: On Wed, May 12, 2010 at 11:43:41AM +0200, Michael 'Mickey' Lauer wrote: Note that this will also remove the 'A' of our AGPS, i.e. we will not be able to upload any ephemeris anymore to improve location. AFAIK will also limit precision -- or is gpsd able to speak UBX these days? It is indeed able to speak UBX. Among other things they are looking at putting the 'A' into our AGPS now that they have a new protocol with room for it [1]. Since it's up for discussion it would probably be good to bring up any experiences, preferences or requirements now. DBUS support is deprecated and being dropped in v3.0 by the end of June [2]. Argh. Ok, from FSO point of view we are forced to work on a bridge or roll our own then :/ They're delegating the dbus side of things to geoclue [1]. Geoclue have a 'provider' for gpsd [2] that probably provides the bridge we want already. We still have the Usage daemon handling power state, so all we really need is to get almanac/ephemeris handling into gpsd. Is there a compelling reason to roll our own rather than work with these projects? [1] http://gpsd.berlios.de/future.html [2] http://www.freedesktop.org/wiki/Software/GeoClue/Providers#Gpsd I like that idea, especially because geoclue is already prepared to locate by information like gsm cell data. It should be enough to write an FSO plugin :) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [debian] 2010-03-25 installation from scratch
There is also recent upstream activity with a new maintainer/developer at git.openmoko.org. Some of it could be considered to be brought to Debian, although we're moving to cornucopia / frameworkd 2.0. Hi, I'm not a big fan of zhone, but you may be interested in the fact, that mickey currently writes zhone2 in vala using the cornucopia daemons: http://git.freesmartphone.org/?p=aurora.git;a=summary -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [fso-gsmd] N900 support
On Sat, Mar 20, 2010 at 10:45:02AM +0200, Timo Jyrinki wrote: 2010/3/19 Sebastian Reichel elektra...@gmail.com: That sounds interesting :) If I got it right you plan to create a new Distribution for N900 using the normal Debian archive and an additional repository containing the stuff from Nokia? Indeed interesting, moebian sounds quite a bit like pkg-fso team+repo for Neo, but worked upon outside of Debian infrastructure and probably containing some proprietary bits. Eventually, feel free to join pkg-fso team at Debian when there is no (proprietary) Nokia stuff needed anymore. I'd be extremely happy if in a year or two one can switch from FreeRunner to N900, running pure Debian. Of course we first have to reach this pure Debian stage with FreeRunner :) The kernel is the only required thing from pkg-fso repository. One or two new kernel hackers would be quite welcome to the Openmoko world to work on remaining upstreaming... -Timo Hi, as I said kernel seems to be less a problem for N900. It seems they commited most of their drivers mainline. So far I saw, that the display driver is missing. You may ask luca about the status of the Debian kernel status for N900. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
[fso-gsmd] N900 support
Hi, I recently bought myself an Nokia N900. I checked out what would be needed to run FSO on it. You can find the details in [1]. Apart from that I started to restructure the isimodem driver from ofono into a library with vapi bindings [2], generated valadoc pages are here: [3]. As you can see the code makes heavy use of asynchronious functions. Do you think the library can be used for fso-gsmd like this? I will continue adding USSD and Device then. [1] http://wiki.freesmartphone.org/index.php/Hardware/N900 [2] https://git.ring0.de/libisi/ [3] http://vala.ring0.de/libisi/index.htm -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [Shr-Devel] Applying for GSoC 2010
On Thu, Mar 11, 2010 at 05:23:15PM +0100, Dr. Michael Lauer wrote: Hi folks, After our successful GSoC in 2007 and 2008 as Openmoko -- and us not being on board in 2009, I decided to try apply again as a mentoring organization this year. Please toss some ideas to http://wiki.freesmartphone.org/index.php/GSoC_Ideas I'll keep you posted on the results. Cheers, :M: Hi Mickey, I received my N900 some days ago and plan to add support for it in FSO. Some points: * Audio chip is TWL4030 (same as palm pre according to your wiki) * Accelerometer is LIS302DL (same as freerunner) * Vibrator is controlled via led interface (like on the freerunner) * GPS is connected via gsm modem - I haven't checked the details yet * There's a proximity and light sensor - How to handle these in FSO? * What to do about the cameras (video/photo)? What about tagging photos (geotag, time, whatever)? Cameras are Video4Linux devices on N900 btw. * There's an FM receiver, but I guess we can ignore this one for now. We need a multimedia daemon for this, which could also handle DVB-H on other devices. * There's an FM transmitter. There should be support for it in the new audio manager. It is available as /dev/radio0 on n900. * Keyboard LEDs, status LED (exported via led interface on sysfs) * Button Keyboard Support - I did not check this yet * Video Cable Detection - I did not check this yet either So I guess N900 support is mostly about fso-gsmd. Today I checked about any docs about GSM in N900 - Nokia commited their own IP-like kernel interface for this (Phonet). There is an existing driver for ofono to check how communication works and get information about the magic bytes. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [PATCH] frameworkd battery status reporting
On Sun, Feb 28, 2010 at 11:50:59AM +, Neil Jerram wrote: On 28 February 2010 02:54, Sebastian Reichel elektra...@gmail.com wrote: I'm pretty sure the power stuff works with fso-deviced, because the SHR idle screen uses them and I get correct feedback from it. Hi Sebastian, Thanks for your interest in this. The problem (in some sense) may be that I'm using openmoko-panel-plugin. Does Debian still install this by default? yes I've put my o-p-p log up at http://www.ossau.uklinux.net/opp-20100227-2300.log. I guess that the relevant parts of it are those that say: ERROR:root:error while calling DeviceIface.GetPower() - assuming 0 (org.freedesktop.DBus.Error.UnknownMethod: Method GetPower with signature on interface org.freesmartphone.Device.PowerControl doesn't exist ) INFO:root:entering dbusReInit of usb INFO:root:entering spreadDBusFailure INFO:root:entering init for battery INFO:root:error on setting up and probing dbus-interface (org.freedesktop.DBus.Error.UnknownMethod: Method GetInfo with signature on interface org.freesmartphone.Device.PowerSupply doesn't exist ) INFO:root:assuming hardware is GTA01 .. This is fsodeviced: s...@freerunner ~ % mdbus -s org.freesmartphone.odeviced \ /org/freesmartphone/Device/PowerSupply \ org.freesmartphone.Device.PowerSupply.GetInfo {'type': 'aggregate'} s...@freerunner ~ % mdbus -s org.freesmartphone.odeviced \ /org/freesmartphone/Device/PowerSupply \ org.freesmartphone.Device.PowerSupply.GetCapacity 89 When you mention SHR, do you mean SHR in Debian, or the complete SHR distro? I never saw a running SHR distro. So of course I mean the SHR packages in Debian ;) There are commits at git.freesmartphone.org that look like they add the missing methods, and I would guess that the complete SHR distro is using an fsodeviced that includes these. Alternatively, maybe the SHR idle screen does not rely on these methods? As I said I'm talking about Debian and SHR idle screen uses them - I'm pretty sure because there the rectangle was empty too, until I switched from odeviced to fsodeviced. Currently ousaged is still the default in Debian, but I plan to change this with one of the next fso-config updates, so it would be nice to get some feedback from you :) Do you really mean ousaged here? FWIW, I've been using fsousaged (instead of ousaged) for a long time now - actually I can't even remember when and why I switched - and it seems fine. uhm no, I meant odeviced, sorry. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [PATCH] frameworkd battery status reporting
On Sat, Feb 27, 2010 at 11:44:50PM +, Neil Jerram wrote: On 27 February 2010 19:04, Neil Jerram neiljer...@googlemail.com wrote: So I'll try switching to fsodeviced too; if that solves the problem, I agree that it's probably not worth changing the old Python code. Well, the Debian version of fsodeviced seems not to work straightforwardly yet. It dates from 10th Jan and appears not to provide the GetPower method of org.freesmartphone.Device.PowerControl, and the GetInfo method of org.freesmartphone.Device.PowerSupply. Also, FWIW, with fsodeviced instead of odeviced, and with notification-daemon installed, I get lots of spurious notification popups about resource state changes (when nothing is really changing). So, I'm going back to odeviced for now, but will try fsodeviced again after it's been updated. Neil I'm pretty sure the power stuff works with fso-deviced, because the SHR idle screen uses them and I get correct feedback from it. But AFAIK the behaviour slightly changed between odeviced and fso-deviced. You can easily debug available methods/objects with mdbus. Currently ousaged is still the default in Debian, but I plan to change this with one of the next fso-config updates, so it would be nice to get some feedback from you :) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: cross compiling e17 package for moko
On Thu, Feb 25, 2010 at 09:22:25PM +0100, arne anka wrote: to answer your question w/o any political implications: if you got a plain x86 box, your best bet would probably be pbuilder, qemubuilder or cowbuilder (don't ask, what the differences are, i still fail to grasp the concept). apt-cross is pretty much dead and emdebian incomplete and rather confusing, imo, that is. pbuilder - build a debian package in a chroot (chroot = will build for the arch of the box pbuilder runs on) qemubuilder - feels like pbuilder, but does not chroot into an image, but uses a qemu image (you can build for other arches with this, but it's slow of course) cowbuilder - I never used this one, but afaik it's used for building for multiple targets by running pbuilder multiple times with different setups. so why building in a chroot? First you make sure the list of packages in Build-Depends is complete (because if not build will fail) and second you may run Debian stable but want to build for unstable. So actually none of these crosscompiles ;) dpkg-cross is used for this, but don't ask me how to do it - I never tried. If you work on e17 Debian packages you should get in contact with Albin Tonnerre lu...@debian.org - The current maintainer of e17. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: cross compiling e17 package for moko
On Thu, Feb 25, 2010 at 10:04:59PM +0100, Steffen Möller wrote: arne anka wrote: So actually none of these crosscompiles ;) dpkg-cross is used for this, but don't ask me how to do it - I never tried. p/qemubuilder does. you need to set up it for armel, and iirc the debian on freerunner wiki has a description, how to do that. i got it once working on an x86 box but lost that boy due a new job. with amd64, qemubuilder does still nnot work. qemubuilder (to my understanding) compiles to arm on an arm platform, just that this arm platform is not a real one but virtualised - not cross. You can have an arm compiler running on x86 or other platforms != arm - cross compilation. I may have missed the start of this thread, can anybody run the qemubuilder on amd64 by now? Cheers, Steffen @arne anka: as Steffen said qemubuilder is not a crosscompiling setup, but native compiling on a virtual system. Cross-compiling would be if you have an x86-compiler, which generates armel bytecode. In qemubuilder you have an emulator and an armel-compiler generating armel bytecode. Crosscompiling is much faster than qemubuilder, but some packages may not build correctly (e.g. when they try to test the application by running it) Apart from this I have a qemubuilder setup on an amd64 box - It exists since Mai 2008 actually :D s...@mars ~ % cat /etc/armelrc ARCH=armel MEMORY_MEGS=256 KERNEL_IMAGE=/var/cache/pbuilder/armel/vmlinuz-2.6.24-1-versatile BASEPATH=/var/cache/pbuilder/armel/base.qemu MIRRORSITE=http://ftp.de.debian.org/debian DISTRIBUTION=sid s...@mars ~ % ls -lh /var/cache/pbuilder/armel/ insgesamt 802M -rw--- 1 root root 3,1G 15. Nov 01:20 base.qemu -rw-r--r-- 1 root root 1,4M 24. Mai 2008 vmlinuz-2.6.24-1-versatile -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [pkg-fso-maint] [debian] phoneuid-messgaes crashes
On Sun, Feb 21, 2010 at 03:36:11AM +0100, arne anka wrote: I have absolutly no clue what may cause this crash and you did not include the python traceback. well, after a reboot today, frameworkd died again and this time i tracked it down. as usual, the root cause is insufficient handling of non-ascii, this time in subsystems/opimd/type_manager.py at line 67 def make_comp_value(self, field_type, field_value, from_query = False): if field_type=='phonenumber': if from_query: return re.escape(normalize_number(field_value)) # +'$' else: return normalize_number(field_value) else: -return str(field_value) +try: +return str(field_value) +except: +return unicode(field_value) Thanks for tracing it :) I tried a bit with the interactive python shell and I think it's enough to always return unicode(field_value) - without the exception handling. I just uploaded an updated package to pkg-fso containing this fix. P.S.: I CC'd smartphones-userland, so that it can be fixed upstream. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: Debian fso-framework
On Thu, Jan 21, 2010 at 12:04:56PM +0100, A.A. wrote: 2010/1/21 arne anka openm...@ginguppin.de Segmentation fault when did you upgrade last? could be the issue i described a while ago: the new libphone-utils0 made my frameworkd segfault, too. version 0.1+git20091220-2 didn't work, downgrading to the one before (0.1+git20091114-1) seemed to fix that. dpkg -i libphone-utils0_0.1+git20091114-1_armel.deb I should have increased the version number of libphone-utils, but I didn't know they changed some internal stuff. I will upload a newer fso-framework later, which fixes the problem. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [Debian] Re: Touchscreen not working
On Mon, Jan 11, 2010 at 11:03:15AM -0600, Steven Jones wrote: Well, I tried downgrading hal to 0.5.11, which was the only previous version (to 0.5.14) I could find in aptitude, but this didn't work, and in the process, I pretty much lost the ability to boot into zhone. A debian login prompt on the screen with no windows. When I tried to install again using the install script, it fails with an error of broken packages in the fso phase. I am not sure what else to try, but I thought I would pass along my update. install script doesn't work currently because xserver-xorg-video-tslib is not yet updated (it has been compiled for the old X server). You can try installing the remaining stuff by hand (you need to add testing to sources.list to get an old Xorg and probably aptitude for error resolving :)) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: Newer kernel Debian package
On Fri, Jan 08, 2010 at 01:36:13PM +0200, Timo Jyrinki wrote: 2010/1/6 Sebastian Reichel elektra...@gmail.com: But if somebody sends me a matching .config file AND the link to the correct git repository I can make a kernel package with make-kpkg. http://git.openmoko.org/?p=kernel.git;a=shortlog;h=andy-tracking + the debug pre-empt disabling which has just been discussed on the kernel list and improves performance considerably: http://users.tkk.fi/~tajyrink/moko/kernel_20100108_nodebug_nopreempt/disable_preempt_and_debug.diff Then just use the gta02_moredrivers_defconfig for .config. Out of curiosity, where do you do the kernel building? I simply have the OE cross-compilation environment which I use to compile the kernel for my Debian, since I haven't had time to tinker other options (besides native building, but kernel is a bit big to compile on FreeRunner). I'm using a 1 Ghz ARM machine [1] for native package building :) I will have a look @ the kernel package when I'm done with SHR stuff. P.S.: scp'ing to alioth works again for me. It was the fault of my bashrc file which was supposed to start zsh. Instead it resulted in failing scp :( [1] http://en.wikipedia.org/wiki/SheevaPlug -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: Current state, what is happening and how is it going?
On Sat, Dec 12, 2009 at 01:00:45AM +0100, Bram Neijt wrote: Dear list members, Last time I tried anything with my OpenMoko phone was about 5 months ago. None of the distributions where able to make phone calls reliable (most common problem being the phone going to sleep during a call), and general usage (adding contacts etc) was terrible. As I read it, Om2009 is dead (was a terrible interface anyway) and now the main development is going into enlightenment? I also read that the freedesktop deamon is going to be used to keep power usage down? I'm not sure how to use the power it has yet, but that seems good. Anyway, I am now wondering how Debian is doing on the openmoko. Anybody working on the interface or would installing Debian still leave me with the FreeDesktop testing application? (zhone ?) Generally my question would be: how is it going, anybody still working on the OpenMoko hardware? Greets, Bram Hi, We're still on it. I think you mixed up FreeDesktop and FSO aka Freesmartphone :) In Debian you can currently install either zhone or SHR. But SHR is quite a mess, because it's an deprecated version of the software. I'm the maintainer of those package, but updating to a more recent version is blocked until e17 packages in Debian are updated. I also read that the freedesktop deamon is going to be used to keep power usage down? I don't know where you heard something like this, but in FSO they are currently moving from one big Python script to one binary for each service. This code is more efficient (no interpreting!) and thus less CPU time is needed for the same thing. This may save a bit power, but more importantly it's faster :) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: fsouid in debian?
On Sat, Dec 05, 2009 at 03:25:37PM +0100, arne anka wrote: On Sat, 05 Dec 2009 15:17:52 +0100, Sebastian Krzyszkowiak seba.d...@gmail.com wrote: On Sat, Dec 5, 2009 at 15:13, arne anka openm...@ginguppin.de wrote: just found out about the package phoneui-apps -- but it depends on fsouid. there's no such package in debian and google doesn't know it either. so, what is fsouid and where do i find it? I think packager did a mistake - it should be phoneuid (or phonefsod) :P still no dice. but i better take that to pkg-fso, i guess. For those of you just reading this list - I was not able to package the full new shr stack, as the e17 packages in Debian are too old. I need to wait for updated e17 packages ;) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: fsouid in debian?
On Sat, Dec 05, 2009 at 06:52:23PM +0100, Sebastian Krzyszkowiak wrote: So you should be able to package libphone-ui, phoneuid, phonefsod - just libphone-ui-shr will fail, as it's the only one library which depends on some e17 stuff ;) That's correct, but it doesn't make any sense, if the stuff does not work :) On Sat, Dec 5, 2009 at 18:44, arne anka openm...@ginguppin.de wrote: so, what exactly does work w/o e17? i'd guess the apps themselves (messages, contacts, dialer) need e17 do display anything, right? When using libphone-ui-shr, which uses e17 libraries to display UI - yes. But you can write and install your own UI frontend, using GTK+, Qt, ncurses or even something different :P In short: you will see absolutly nothing without this package - it contains the functions drawing the UI. I stopped uploading stuff to avoid too much confusion about new vs old SHR stuff in the repository. I hope the new e17 package will come soon :) -- Sebastian R. signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: How to use fso-usaged?
On Fri, Dec 04, 2009 at 12:38:03AM +0100, Gilles Filippini wrote: arne anka a écrit , Le 03/12/2009 23:03: Doesn't work out of the box here. Zhone says: Failed to read authentication status yeah, common issue. just start zhone over again. and again. after a while it will catch on. my personal best is 5 times so far. It's even worse. I need to restart several times both frameworkd + zhone, and it may - rarely - catch on. I don't have this problem at all with my old frameworkd.conf (attached) for which I need maximum 2 tries at launching zhone (when I don't wait enough after [re]starting frameworkd). Thanks, _Gilles. [ogsmd] # choose your muxer, available types are: gsm0710muxd [default], fso-abyss ti_calypso_muxer = fso-abyss Can you try fso-abyss with the new provided configuration? Currently it still uses gsm0710muxd as default. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: How to use fso-usaged?
On Tue, Dec 01, 2009 at 12:00:57PM +1100, Jonathan Schultz wrote: The latest version has been in use by SHR for quite a while now and seems pretty stable. Yes and it works well for me too so far. The only problem is this is the first I've heard about it. Would it be possible for fso-config-gta02 to be kept updated? Jonathan It will be included in the next fso-config package. I will upload a new version to pkg-fso repository today. I will be asking for sponsoring once you tell me the new package is working :) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: Configuration files
On Sun, Nov 15, 2009 at 10:06:53PM +0300, Nikita V. Youshchenko wrote: as you might have seen more and more FSO 2.0 subsystems are ready for prime time, so we have to think about how configuration works for FSO 2.0. Until now I used the same method as in FSO 1.0, which is having it lumped all together in one config file. For FSO 2.0 I'm leaning towards changing this. The benefit I see from one configuration file per subsystem is: *) promoting the fine granular usage of individual FSO subsystems, *) enable packagers to ship the config with the subsystem, if they want, *) a third reason which i forgot while writing this mail. What do you think? Hi, from the Debian point of view I can support this idea, because it would make packaging easier :) How do you plan to start the different subsystems, if each subsystem uses it's own files? Will there be a init.d script for each subsystem? If I undrstand things correctly, mentioned subsystems are actually dbus services, and as such, could be started on demand bu dbus daemon itself. This may be not true for a subsystem that takes long to initialize, in which turn this subsystem should be started at system startup. Nikita This may work for services, which can be started on demand, e.g. ogpsd, but some of them are supposed to run @ system startup, e.g. I want my display auto-disabled without requesting the event handler via DBus. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [debian] fso-frameworkd backtrace
On Sat, Oct 17, 2009 at 12:42:44AM +0400, Nikita V. Youshchenko wrote: Hi I'm getting this in frameworkd stdout: Traceback (most recent call last): File /usr/lib/pymodules/python2.5/framework/patterns/tasklet.py, line 178, in send value = self.generator.send(value) File /usr/lib/pymodules/python2.5/framework/subsystems/opimd/pimb_sim_messages_fso.py, line 287, in load_entries self.process_all_entries(entries) File /usr/lib/pymodules/python2.5/framework/subsystems/opimd/pimb_sim_messages_fso.py, line 247, in process_all_entries self.process_single_entry([[message[0][0]],message[0][1],message[0] [2],message[0][3],message[0][4]]) File /usr/lib/pymodules/python2.5/framework/subsystems/opimd/pimb_sim_messages_fso.py, line 139, in process_single_entry entry_id = self._domain_handlers['Messages'].register_entry(self, entry) File /usr/lib/pymodules/python2.5/framework/subsystems/opimd/pimd_messages.py, line 324, in register_entry message_id = GenericDomain.register_entry(self, backend, message_data) File /usr/lib/pymodules/python2.5/framework/subsystems/opimd/pimd_generic.py, line 647, in register_entry entry.import_fields(entry_data, backend.name) File /usr/lib/pymodules/python2.5/framework/subsystems/opimd/pimd_generic.py, line 171, in import_fields compare_value = self.make_comp_value(field_value) File /usr/lib/pymodules/python2.5/framework/subsystems/opimd/pimd_generic.py, line 140, in make_comp_value if str(field_value).startswith('tel:'): UnicodeEncodeError: 'ascii' codec can't encode characters in position 0-4: ordinal not in range(128) Could somebody please help to debug this? Hi, You may want to get in contact with Sebastian Krzyszkowiak, who is the one doing most of the commits on opimd at the moment. I added a CC, because I don't know if he is subscribed to the list. P.S.: I hope to find some time to update the FSO/SHR packages + config package in pkg-fso next week (university started, so I don't have as much time as before). The good thing is, that I now own my own 1.2Ghz armel build system [1], so I don't have to use the slow qemubuilder :) [1] http://en.wikipedia.org/wiki/SheevaPlug -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: python-elementary and paroli
On Sun, Oct 11, 2009 at 12:58:45PM +0200, A.A. wrote: Hi, I have installed the python-elementary with apt-get, my sources.list is: deb http://ftp.it.debian.org/debian unstable main contrib deb http://pkg-fso.alioth.debian.org/debian unstable main but when i try tu run paroli with: ../paroli/scripts# DISPLAY=:0 ./paroli it return: pwd: /root/Programmi/paroli PYTHONPATH=core \ exec python -OO core/paroli/__init__.py --cfgfile /etc/paroli/paroli.fso.cfg Traceback (most recent call last): File core/paroli/__init__.py, line 38, in module import elementary File /usr/lib/pymodules/python2.5/elementary/__init__.py, line 19, in module from c_elementary import Object, Window, Background, Icon, \ ImportError: /usr/lib/pymodules/python2.5/elementary/c_elementary.so: undefined symbol: elm_spinner_value_get marte:~/Programmi/paroli/scripts# cat /etc/apt/sources.list deb http://ftp.it.debian.org/debian unstable main contrib deb http://pkg-fso.alioth.debian.org/debian unstable main whereis the problem? The python-elementary version in pkg-fso is depending on a more recent elementary version, which is not yet in Debian. I guess I should have removed it, so that nobody tries to use it :/ -- Sebastian Reichel signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: fso-frameworkd issues
On Thu, Oct 08, 2009 at 08:18:16AM +1100, Jonathan Schultz wrote: All I can say is that, after a couple of days, frameworkd has been working reliably so far with Zhone. zhone doesn't work for me either. It stops with 'Failed to read authentication status' and never connects to the GSM network. I take it there must be something suspect in my framework configuration. Can someone attach a complete working configuration file? Cheers, Jonathan Hi, here is the one I'm currently using on my phone. You need to install fso-usaged if you want to use it. The file shipped with the new gta02 configration package has some more changes, but this one should work for you. Please note, that it does not use fso-abyss at the moment! -- Sebastian [frameworkd] # indicates this configuration version, do not change version = 1 # the default log_level, if not specified per module # available log levels are: DEBUG, INFO, WARNING, ERROR, CRITICAL log_level = WARNING # the global log_destination. Uncomment to enable #log_to = stderr log_to = file #log_to = syslog # if logging to a file, specify the destination log_destination = /tmp/frameworkd.log # persistance format, one of pickle, yaml persist_format = pickle rootdir = ../etc/freesmartphone:/etc/freesmartphone:/usr/etc/freesmartphone [odeviced] # set 1 to disable a module disable = 0 [odeviced.kernel26] # set 1 to disable a module disable = 0 # poll capacity once every 5 minutes # (usually, you do not have to change this) capacity_check_timeout = 300 # set 0 to disable FB_BLANK ioctl to blank framebuffer # (if you have problems on Openmoko GTA02) fb_blank = 1 [odeviced.audio] # set 1 to disable a module disable = 0 # set directory where the alsa audio scenarios are stored scenario_dir = /usr/share/openmoko/scenarios # set default scenario loaded at startup default_scenario = stereoout [odeviced.idlenotifier] # set 1 to disable a module disable = 0 # don't read from accellerometers for GTA02 ignoreinput = 2,3 # configure timeouts (in seconds) here. A value of 0 # means 'never fall into this state' (except programatically) idle = 10 idle_dim = 20 idle_prelock = 12 lock = 2 suspend = 0 [odeviced.input] # set 1 to disable a module disable = 0 # don't read from accellerometers for GTA02 ignoreinput = 2,3 # format is keyname,type,input device keycode,report held seconds in addition to press/release report1 = AUX,key,169,1 report2 = POWER,key,116,1 report3 = USB,key,356,0 report4 = HEADSET,switch,2,0 [odeviced.powercontrol-neo] # set 1 to disable a module disable = 0 # disable accelerometer since it uses LOADS of CPU resources [odeviced.accelerometer] disable = 1 accelerometer_type = gta02 [ogsmd] # set 1 to disable a module disable = 0 # choose your modem type, available types are: ti_calypso, freescale_neptune, singleline, muxed4line, option, ... modemtype = ti_calypso # if you have a ti_calypso, you can choose the deep sleep mode. Valid values are: never, adaptive (default), always #ti_calypso_deep_sleep = adaptive ti_calypso_deep_sleep = always # if you have a ti_calypso, you can choose the dsp mode for audio enhancement. Valid values are: #short-aec: Short Echo Cancellation (max) #long-aec: Long Echo Cancellation (max) #long-aec:6db: Long Echo Cancellation (-6db) #long-aec:12db: Long Echo Cancellation (-12db) #long-aec:18db: Long Echo Cancellation (-18db) #nr: Noise Reduction (max) #nr:6db: Noise Reduction (-6db) #nr:12db: Noise Reduction (-12db) #nr:18db: Noise Reduction (-18db) #aec+nr: Long Echo Cancellation (max) plus Noise Reduction (max) [default] #none: No audio processing. ti_calypso_dsp_mode = aec+nr #ti_calypso_muxer = fso-abyss [ogpsd] # set 1 to disable a module disable = 0 # possible options are NMEADevice, UBXDevice, GTA02Device, EtenDevice device = GTA02Device # possible options are SerialChannel, GllinChannel, UDPChannel, FileChannel channel = SerialChannel # For UDPChannel the path defines the port to listen to path = /dev/ttySAC1 [ousaged] # set 1 to disable a module disable = 1 # choose whether resources should be disabled at startup, at shutdown, always (default), or never. sync_resources_with_lifecycle = always [opreferencesd] # set 1 to disable a module disable = 0 log_level = WARNING rootdir = ../etc/freesmartphone/opreferences:/etc/freesmartphone/opreferences:/usr/etc/freesmartphone/opreferences [oeventsd] # set 1 to disable a module disable = 0 log_level = WARNING rules_file = ../etc/freesmartphone/oevents/rules.yaml:/etc/freesmartphone/oevents/rules.yaml:/usr/etc/freesmartphone/oevents/rules.yaml [opimd] # set 1 to disable a module disable = 0 contacts_default_backend = CSV-Contacts messages_default_backend = SIM-Messages-FSO calls_default_backend = SQLite-Calls dates_default_backend = SQLite-Dates notes_default_backend = SQLite-Notes tasks_default_backend = SQLite-Tasks contacts_merging_enabled = 1 messages_default_folder = Unfiled messages_trash_folder = Trash
Re: fso-frameworkd issues
On Tue, Oct 06, 2009 at 02:00:18PM +1100, Jonathan Schultz wrote: Failed to handle dbus error: Resource Device is not enabled, current status is 'enabling', 81 (dbus-glib-error-quark), code 32 ** (process:1605): DEBUG: power_up_antenna_callback() Failed to handle dbus error: Resource Device is not enabled, current status is 'enabling', 81 (dbus-glib-error-quark), code 32 ** (process:1605): DEBUG: sim_ready_status_callback(status=1074515968) ** (process:1605): DEBUG: sim ready Failed to handle dbus error: Resource Device is not enabled, current status is 'enabling', 81 (dbus-glib-error-quark), code 32 ** (process:1605): DEBUG: get_messagebook_info_callback() Segmentation fault Hi, can you please check if you installed the newest ophonekitd package available? The functions, which need the ressource enabled, should be called by a callback once the ressource is enabled. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: fso-frameworkd issues
On Tue, Oct 06, 2009 at 02:00:18PM +1100, Jonathan Schultz wrote: Failed to handle dbus error: Resource Device is not enabled, current status is 'enabling', 81 (dbus-glib-error-quark), code 32 ** (process:1605): DEBUG: power_up_antenna_callback() Failed to handle dbus error: Resource Device is not enabled, current status is 'enabling', 81 (dbus-glib-error-quark), code 32 ** (process:1605): DEBUG: sim_ready_status_callback(status=1074515968) ** (process:1605): DEBUG: sim ready Failed to handle dbus error: Resource Device is not enabled, current status is 'enabling', 81 (dbus-glib-error-quark), code 32 ** (process:1605): DEBUG: get_messagebook_info_callback() Segmentation fault I just saw there are some other interesting changes, which may help in this case, in SHR's git. I will update the package later to include those, too. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: fso-frameworkd issues
Hi, I haven't noticed any discussion of this on this list so I wonder if noone else has experienced the same issues with fso-frameworkd 0.9.5.9+git20090920-1 using ophonekit. You can blame me for the issues, I updated fso package in pkg-fso and packaged ophonekitd ;) Most of the discussion about fso-frameworkd is currently being done in pkg-fso-maint mailinglist. The main problem is, that I have *not* updated fso-config packages so far. The most obvious symptom is that it just stops working whenever I have an incoming text message. Sometimes it responds to SIGINT by performing the actions that it should have done when the text message arrived, and then proceeds to completely hang until I send it a SIGKILL. Other times it shuts down cleanly, but without performing the actions on SIGINT. A debug log show no obvious sign of anything going astray. I'm almost done with the config packages. I guess your problems are the result of a bad configuration. I tried running syslog to see if anything showed up there but there was nothing. Could anyone suggest where I might look next to see what's going on? if you installed custom rules you could try using the old default configuration. At least on my phone the combination of old default configuration and getting a text message worked. If you want all hints frameworkd can give you set log to stderr, log level to INFO and start frameworkd via (root) shell. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: Running the Framework on Ubuntu laptop?
On Fri, Sep 25, 2009 at 12:45:36PM +0100, Arigead wrote: Sebastian Reichel wrote: On Fri, Sep 25, 2009 at 09:05:53AM +0100, Arigead wrote: Hello all, Hope this question ain't too stupid. I'd like to try and run the FSO Framework on my ubuntu laptop. More specifically I'd like to play with the PIM part of the framework. I'm hoping to play with it and write code to interface with the PIM Engine with DBus calls. I've done the basic sudo python setup.py install which seems to have run fine but executing frameworkd tells me that there is no frameworkd.conf Could somebody point me in the right direction here. Can I simply copy the conf file from a freerunner across to the ubuntu laptop. That sounds a bit too simple. Any help or advice would be gratefully appreciated. John Hi, Just copy the default frameworkd.conf to /etc/frameworkd.conf and disable all subsystems except of opimd. Btw. you may want to install the Debian packages in Ubuntu instead of installing FSO without any packaging control system. If so you can find the package here: http://pkg-fso.nomeata.de/sid/fso-frameworkd You will also need some dependencies, which are linked on the page. Dependencies, which are not linked can be found in the Debian repository and most likley in the Ubuntu repository, too. -- Sebastian Thanks a million for that response Sebastian. It makes sense to install everything from deb packages so I started that but unfortunatley encountered an error with an install of python-phoneutils it's complaining that /usr/share/python-support/python-phoneutils.public is not a directory which is fair comment as it's not a directory. I'm a bit suspicious of a package that's installing python-phoneutils having a problem with python-phoneutils ;-) John mh maybe ubuntu handles python-support packages in a different way than Debian does. I don't have the time to check that, since there are still remaining FSO packages, which need to be cleaned (to get into Debian) or updated. You could try building the package yourself from the source. You can download the source package for python-phoneutils from the repository, too. Afterwards you search a Tutorial in the WWW (there are plenty) and build your package :) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Wed, Sep 16, 2009 at 07:16:10PM +0200, arne anka wrote: starting ophonekitd from Xsession.d seems not to work becaus eof timeouts and atm even starting from .xsession or manually has issues. I switched to Xsession.d and it's working for me (tm). debian-gta02:~$ DISPLAY=:0 ophonekitd ** (process:11425): DEBUG: Configuration file read ** (process:11425): DEBUG: Entering ecore loop ** (process:11425): DEBUG: Phonegui initiated ** (process:11425): DEBUG: Entering glib main loop ** (process:11425): DEBUG: /org/shr/ophonekitd/Usage ** (process:11425): DEBUG: Trying to get the system bus ** (process:11425): DEBUG: Adding signals. ** (process:11425): DEBUG: Added sim AuthStatus. ** (process:11425): DEBUG: Added sim ReadyStatus. ** (process:11425): DEBUG: Added sim IncomingStoredMessage. ** (process:11425): DEBUG: Added call CallStatus. ** (process:11425): DEBUG: Added device Idle Notifier State. ** (process:11425): DEBUG: Added network IncomingUssd. ** (process:11425): DEBUG: Connected to the buses ** (process:11425): DEBUG: list_resources() ** (process:11425): DEBUG: list_resources_callback() ** (process:11425): DEBUG: Resource TEST available ** (process:11425): DEBUG: Resource CPU available ** (process:11425): DEBUG: Resource Display available ** (process:11425): DEBUG: Resource GSM available ** (process:11425): DEBUG: Request GSM resource ** (process:11425): DEBUG: Initiated elementary ** (process:11425): DEBUG: Initiated etk ** (process:11425): DEBUG: request_resource_callback() ** (process:11425): DEBUG: call ogsmd_device_set_antenna_power() Failed to handle dbus error: Resource Device is not enabled, current status is 'enabling', 80 (dbus-glib-error-quark), code 32 ** (process:11425): DEBUG: power_up_antenna_callback() Failed to handle dbus error: Resource Device is not enabled, current status is 'enabling', 80 (dbus-glib-error-quark), code 32 ** (process:11425): DEBUG: sim_ready_status_callback(status=1074499584) ** (process:11425): DEBUG: sim ready Failed to handle dbus error: Resource Device is not enabled, current status is 'enabling', 80 (dbus-glib-error-quark), code 32 ** (process:11425): DEBUG: get_messagebook_info_callback() Segmentation fault Looks like ophonekitd assumes the modem is up without checking it, while the modem has not been enabled completly. I guess some wait loops are needed arround ogsmd_device_set_antenna_power similar to the one being used when enabling the ressouces. From my logfile: ** (process:1865): DEBUG: list_resources() ** (process:1865): DEBUG: list_resources_callback() ** (process:1865): DEBUG: GSM not available, try again in 5s I guess it's time to annoy the SHR developers. Can you please ask them via IRC or their mailing list? You can find connection data for both @ http://www.shr-project.org/trac/wiki/Communicate I will package a new version at once when there is a solution for the race condition. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Thu, Sep 17, 2009 at 11:52:43AM +0200, arne anka wrote: I guess it's time to annoy the SHR developers. Can you please ask them via IRC or their mailing list? You can find connection data for both @ http://www.shr-project.org/trac/wiki/Communicate well, it happend what i expected -- though with considerable delay: You are not allowed to post to this mailing list, and your message has been automatically rejected. If you think that your messages are being rejected in error, contact the mailing list owner at shr-devel-ow...@lists.shr-project.org. there's not much i can do right now, i don't intend to add yet another list to the ever growing number of mails and i have atm no time to listen on irc if someone is saying something sometime. I just sent a mail to shr-devel containing your log. I didn't send the one of joachim, though, because it appears to be the frameworkd, which is faulty on his phone. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Tue, Sep 15, 2009 at 11:38:41PM +, Joachim Breitner wrote: Hi, Am Mittwoch, den 16.09.2009, 01:33 +0200 schrieb Joachim Breitner: But then, I have the same problem with zhone, it seems. I’ll have a closer look tomorrow. frameworkd says: 2009.09.16 01:37:35 ogsmd.modems.abstract.channel WARNING UNHANDLED INTERMEDIATE: OK and zhone says: Antenna power OK. Registering to network now. Registering to network SIM is ready: 1 retrieving phonebook... SIM is ready: 1 retrieving messagebook... IDLE STATE = idle_prelock IDLE STATE = lock org.freesmartphone.GSM.Device.Timeout: device did not answer within 30 seconds has anyone seen this before? It used to work without problems... Greetings, Joachim I think you should try out fso-abyss [1], which heiko did package for us (thanks!). It's a replacement for the gsm muxers, which made the GSM part of my phone much more stable than before :) [1] http://pkg-fso.nomeata.de/sid/fso-abyss -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Thu, Sep 17, 2009 at 06:00:22PM +0200, Joachim Breitner wrote: Hi, Am Donnerstag, den 17.09.2009, 15:21 +0200 schrieb Sebastian Reichel: I think you should try out fso-abyss [1], which heiko did package for us (thanks!). It's a replacement for the gsm muxers, which made the GSM part of my phone much more stable than before :) [1] http://pkg-fso.nomeata.de/sid/fso-abyss I would, but it seems after the last tests it did not shut down cleanly, so I’ll have to find a way to revive the battery first :-( Greetings, Joachim Have a look at [1]. It happened to my GTA02v5, too before the framework learned to shut it down before it's deep discharged. For me the following procedure worked most times: * remove battery * plug in wall charger * reinsert battery * press power [1] http://wiki.openmoko.org/wiki/Neo_FreeRunner_Hardware_Issues#Can.27t_boot_with_discharged_or_missing_battery_.28FIXED.29 signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Thu, Sep 17, 2009 at 06:21:00PM +0200, Klaus 'mrmoku' Kurzmann wrote: Am Donnerstag 17 September 2009 15:13:44 schrieb Sebastian Reichel: On Thu, Sep 17, 2009 at 11:52:43AM +0200, arne anka wrote: I guess it's time to annoy the SHR developers. Can you please ask them via IRC or their mailing list? You can find connection data for both @ http://www.shr-project.org/trac/wiki/Communicate well, it happend what i expected -- though with considerable delay: You are not allowed to post to this mailing list, and your message has been automatically rejected. If you think that your messages are being rejected in error, contact the mailing list owner at shr-devel-ow...@lists.shr-project.org. there's not much i can do right now, i don't intend to add yet another list to the ever growing number of mails and i have atm no time to listen on irc if someone is saying something sometime. I just sent a mail to shr-devel containing your log. I didn't send the one of joachim, though, because it appears to be the frameworkd, which is faulty on his phone. hopefully I fixed this. For the fix you have to update libframeworkd-glib and ophonekitd. thanks, I will create new packages in some minutes :) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Wed, Sep 16, 2009 at 09:20:30AM +0200, Joachim Breitner wrote: Hi, Am Mittwoch, den 16.09.2009, 07:22 +0400 schrieb Nikita V. Youshchenko: Why did dependences allow installation with obsolete libframeworkd-glib0? It is a bug in libframeworkd-glib0 package then, it has broken shlibs data. I don’t think I ever uploaded this package, and since I used a different version scheme, my version number happened to be larger than Sebastians. Nothing to worry about. Greetings, Joachim Yes, it's like that. Actually you did upload it to pkg-fso, but I guessed nobody uses it anyway, since there was no application depending on it. So I just removed it instead of naming the new package 1:0.1+git20090801. Sorry for the inconvenience. I used the version scheme used for all the other packages in pkg-fso by the way. I think the shlibs file is ok: elektra...@mars ~/tmp/ophonekitd % dpkg --info ophonekitd_0.1+git20090914-2_armel.deb | grep Depends Depends: libc6 (= 2.4), libdbus-1-3 (= 1.0.2), libdbus-glib-1-2 (= 0.78), libframeworkd-glib0 (= 0.1+git20090801), libframeworkd-phonegui0 (= 0.1+git20090802), libglib2.0-0 (= 2.14.0), debconf (= 0.5) | debconf-2.0, fso-frameworkd, libframeworkd-phonegui-efl | libframeworkd-phonegui-implementation - -- Sebastian -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iEYEARECAAYFAkqw1k0ACgkQH0JwilpTmKjLrQCfVg4LEfdWjPjWyDQO8Z4G3k70 SWUAniF9kuVQF7OCno0HFkSbYKf/sLRf =2lhX -END PGP SIGNATURE- ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
Hi Joachim, On Tue, Sep 15, 2009 at 12:04:45PM +0200, Joachim Breitner wrote: Actually, your package ships /etc/X11/Xsession.d/80ophonekitd which means that (at least when using the default nodm setup), ophonekitd is started automatically. Not sure if we really want that, some people might prefer zhone or pyneo and just happen to install ophonekit by accident. Anyways, this file has severe bug: It calls exit 0 at the end. But since these files are sourced (not executed), this ends the X session immediately. I suggest you just remove this file for now. I can add a /etc/default/ophonekitd with AUTOSTART=yes, which is checked before the start. The file could be configured via debconf. Unfortunately, ophonekitd crashes on me. The crashes occur at different points, sometimes after ** (process:2463): DEBUG: Entering glib main loop No idea what may fail that early in the progress :( sometimes after ** (process:2429): DEBUG: Adding signals. For me shr-dialer crashed at this point when I was packaging it. In my case the real bug was missing configuration file for libphone-utils, but it's now in the package. The code which prints those lines is from libframeworkd-glib btw. ** (process:2509): DEBUG: Initiated etk ** (process:2509): DEBUG: call status handler called, id: 0, status: 1074028672 GLib-CRITICAL **: g_hash_table_lookup: assertion `hash_table != NULL' failed aborting... I didn't get this one before - No idea what it may be and where it occurs. No useful backtraces from gdb. This is due to all those abstraction libraries, I had problems, too, when packaging shr-dialer. gdb seems to be unusable with shr stuff. Also shr-messages fails on me: $ shr-messages ** (process:2553): DEBUG: Trying to get the system bus ** (process:2553): DEBUG: Adding signals. ** (process:2553): DEBUG: Entering ecore loop ** (process:2553): DEBUG: phonegui_messages_show() ** (process:2553): DEBUG: ogsmd_sim_retrieve_phonebook_callback() Failed to handle dbus error: Resource Device is not enabled, current status is 'disabled', 73 (dbus-glib-error-quark), code 32 and the shr-dialer fails when I try to call someone. Obviously they don’t have proper error handling in case ophonekitd is not running. Actually it seems they have no error handling at all. The error in shr-messages is obviously, that GSM ressource from FSO is not enabled, so it can't get what it needs. You should have checked ophonekitd.log, too because it tells if it couldn't enable ressources. Maybe you just didn't wait long enough between starting ophonekitd and calling shr-dialer/shr-messages. If you use illume as window manager you can use them once the gsm icon shows, that GSM is connected. shr-dialer, shr-contacts and shr-messages are very stupid applications btw. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Tue, Sep 15, 2009 at 05:51:00PM +1000, Jonathan Schultz wrote: finally there are more or less working shr packages in the pkg-fso repository :) Wow, that looks fantastic! I had one problem that I've been beating my head against the wall solving, but I found it eventually in the file /etc/X11/Xsession.d which seems to have a syntax error in it that made X fall over when it started. In any case I got this message repeatedly in my ~/.xsession-errors file. /etc/X11/Xsession.d/80ophonekitd: 6: [[: not found I think there should only be one bracket rather than two around the expression. But I don't want ophonekitd starting up automatically with X anyway so I moved the file out of the way. I don't know if there is a place to log bugs in ophonekitd so thought I'd post it to this list to save others wasting their time finding this bug. actually you sent the mail to me and not to the list. I CC'd the list this time and quoted your complete mail. In short: I will have a look at the ophonekitd package now and provide a better Xsession.d with configurable autostart. Read my mail to Joachim for more details. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Tue, Sep 15, 2009 at 01:41:46PM +0200, arne anka wrote: besides the /etc/X11/Xsession.d/80ophonekitd issue joachim already mentioned (saving my day btw) this will be fixed with the next upload, I'm writing the debconf stuff atm. all windows are far to small. pin input, dialer, messages use normal sized font, but the buttons are tiny (and the labels too). the windows are maybe 80px wide and 120px high. same goes for call announcement, though the font is far smaller, allowing the whole number to be seen. windows are not resizable and have no scrollbars, messages are illegible. i don't use e or illume or whatever, but lxde as wm. mh in illume they are fullscreen except of the illume bar. The windows are actually resized in libframeworkd-phonegui-efl/src/util/window.c - window_init(); I'm not that used to enlightenment code, so I can't tell you more at the moment. ophonekitd spits out a lot of DEBUG messages (** (process:6089): DEBUG: idle notifier state handler called, id 4) --is it possible to change the level to, say, WARN or so? AFAIK not, they are just using g_debug() calls at the moment. So you have the choice between all or nothing ( /dev/null 21). But I can patch away the idle notifier state handler message :P -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
I just uploaded a new version of ophonekitd. Changes: * Fix /etc/X11/XSession.d/80ophonekitd * Provide /etc/default/ophonekitd for enabling ophonekit autostart via XSession * Add debconf for configuring ophonekitd autostart * Patch away idle notifier debug message for a clean logfile * Don't append to logfile, but create a new one and a backup of the old one * Provide a dbus configuration file with permissions for group=users, so that users don't need to add a custom file, but just have to be in the users group. The default setting for autostart is 'false', so existing installations won't be destroyed by installing ophonekitd. If you want to switch to SHR you can just do dpkg-reconfigure ophonekitd and select 'yes' :) Apart from that I extended the dbus configuration file to give all users in group 'users' the permission to start the ophonekitd daemon. So it should be enough to add your user to the users group instead of adding your own configuration file. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Tue, Sep 15, 2009 at 05:24:26PM +0200, Joachim Breitner wrote: Hi again, Am Dienstag, den 15.09.2009, 17:01 +0200 schrieb Joachim Breitner: Anywas, I get the crashes at quite random points (commands executed more or less directly after each other): After looking at the output of strace -f for a few times, I think there are three separate problems: $ ophonekitd [..] GLib-CRITICAL **: g_hash_table_lookup: assertion `hash_table != NULL' failed aborting... Aborted Ok, this one I have not yet seen under strace... I guess this one comes directly from ophonekitd. ~ $ ophonekitd [..] ** (process:1985): DEBUG: Entering glib main loop Segmentation fault And segfaults at other points in time: These actually come from a second thread (presumably the gui thread, as it does call stuff from the EFL), and due to concurrency, the segfault does not always occur at the same time. And sometimes, the main thread seems to have enough time to hit the above or below problem. ~ $ ophonekitd ** (process:2004): DEBUG: Configuration file read [..] ** (process:2004): DEBUG: call ogsmd_device_set_antenna_power() ophonekitd: symbol lookup error: ophonekitd: undefined symbol: ogsmd_sim_get_sim_ready This is an interesting error message. In my version the symbol is exported correctly: elektra...@freerunner /usr/lib % rabin -s libframeworkd-glib.so.0.0.0 | grep ogsmd_sim_get_sim_ready address=0x842c offset=0x842c size=0144 bind=GLOBAL type=FUNC name=ogsmd_sim_get_sim_ready_callback address=0x77a4 offset=0x77a4 size=0164 bind=GLOBAL type=FUNC name=ogsmd_sim_get_sim_ready What's the result of `ldd /usr/bin/ophonekitd` on your box? is the right shared object used? I should add that I did _not_ update my frameworkd to any of Heiko’s new packages. I'm using frameworkd package from debian main, too and it is working for me, so I guess this is not the problem ;) P.S.: does anybody else have these problems? -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Tue, Sep 15, 2009 at 07:52:42PM +0200, arne anka wrote: mh in illume they are fullscreen except of the illume bar. The windows are actually resized in libframeworkd-phonegui-efl/src/util/window.c - window_init(); there it seems to be set to 480x600 ... what number of dpi is set at your fr? elektra...@freerunner ~ % grep NODM_X_OPTIONS /etc/default/nodm NODM_X_OPTIONS='-nolisten tcp -dpi 96' -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: SHR @ Debian
On Tue, Sep 15, 2009 at 07:50:52PM +0200, arne anka wrote: I just uploaded a new version of ophonekitd. Changes: great! * Provide a dbus configuration file with permissions for group=users, so that users don't need to add a custom file, but just have to be in the users group. i seem to recall a group messagebus being introduced a whiel ago for access to frameworkd via dbus. wouldn't that be more appropriate? Actually it would be nice to have a feature user=current_X_user for the dbus conf files, but AFAIK there's no such feature :( Also as far as I know no group has been found for access to frameworkd so far. But please note, that both daemons are very different from each other. ophonekitd does not touch your hardware or suspend your phone, but just shows you the GUIs and sends calls to frameworkd. There are only two things, which kept me from giving access to *: * other users may spam you with GUIs * other users may start a ophonekitd server, which will probably conflict with the first one. IMHO they made a mistake when they put ophonekitd on the system bus. I think it would be better using the session bus. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
SHR @ Debian
Hi, finally there are more or less working shr packages in the pkg-fso repository :) To get it working you need to remove zhone from autostart first. Then you can do: apt-get install shr-dialer shr-messages shr-contacts This will install the 3 programs and ophonekitd, which is the actual program doing the hard work. Now you should put ophonekitd in your X autostart and if you are using a non root account as X user you need to give your user dbus access to ophonekitd. I did it like this: elektra...@freerunner ~ % cat .xsession #zhone ophonekitd /tmp/ophonekitd.log 21 enlightenment_start elektra...@freerunner ~ % sudo cp /etc/dbus-1/system.d/ophonekitd.conf \ /etc/dbus-1/system.d/my-ophonekitd.conf elektra...@freerunner ~ % sudo vim /etc/dbus-1/system.d/my-ophonekitd.conf elektra...@freerunner ~ % cat /etc/dbus-1/system.d/my-ophonekitd.conf busconfig policy user=elektranox allow own=org.shr.ophonekitd.Usage/ allow send_path=/org/shr/ophonekitd/Usage/ allow send_destination=org.shr.ophonekitd.Usage/ allow receive_sender=org.shr.ophonekitd.Usage/ /policy /busconfig Last but not least do not forget to edit /etc/phone-utils.conf. It needs your countries prefix for number - contact resolving when you're called. Default settings from upstream are for Germany by the way. If you are living there you don't need to change anything. If you are not familiar with SHR - it's enough to have ophonekitd running if you want to be reachable. It will open a window when you get called :) If you find bugs report them in this thread. So far I found out, that speaker mode seems to be broken, but it's 06:11 in the morning now and I want to get some sleep :) ToDo: shr-settings- package is prepared, but waiting for python-elementary libphone-utils - needs debconf system for config file ophonekitd - general cleanup general - update the shr stuff in git.debian.org -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
[Debian] intone and neon now in pkg-fso repository
Hi, I just uploaded intone and neon to the pkg-fso repository. intone is an audio mplayer frontend written in C using elementary for the GUI. You can find a screenshot at [1], upstream homepage is at [2]. neon is an image viewer written in Python using efl for the GUI. It's features are rotation, zoom, diashow, basic exif and an intuitive, fingerfriendly UI. You can find screenshots on the upstream hompage, which can be found at [3]. So from the Hall of Fame [4] only numptyphysics and orrey are missing in Debian. With numptyphysics being packaged by jedd, see Debian Bug 496586 [5]. P.S.: due to some problems with the pkg-fso repository key the packages are not yet in the Package Index. They will be once nomeata has fixed this :) -- Sebastian [1] http://wiki.openmoko.org/images/5/56/Intone-1.png [2] http://code.google.com/p/intone/ [3] http://neon.projects.openmoko.org/ [4] http://wiki.openmoko.org/wiki/Hall_of_Fame [5] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=496586 signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [debian] kernel 2.6.29 and XF86PowerOff xevent
On Wed, Aug 05, 2009 at 11:35:41PM +0200, Nicola Mfb wrote: Hi! With debian prepackaged 2.6.29 kernel I do not get anymore XF86PowerOff key press/release xevents (using glamo or fbdev is the same). All goes fine with 2.6.28, just use xev to reproduce disabling the fso suspend rule. Is that a known bug? where to report? Hi, as far as I know it's not a known bug. To report a bug under Debian please use the reportbug utility as described at [1]. Please do not forget to CC 'pkg-fso-ma...@lists.alioth.debian.org' in your bug report (you will be asked for CCs by the reportbug utility), so that pkg-fso guys get notified. -- Sebastian [1] http://wiki.debian.org/DebianOnFreeRunner#Reportingbugs signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: GSM firmware moko11, kernel and boot loader (was Re: [Debian] 2.6.29 kernel)
Hi Luca, Don't ask me into the details, I just got a hint about qi + moko11 + new kernel results in non working gsm from the fso bugtracker, tried updating qi and it worked. The FSO bug I got the information from is this one btw: http://trac.freesmartphone.org/ticket/367 -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [Pkg-e-devel] [debian/illume] digital clock?
On Sun, Jun 21, 2009 at 10:22:28PM +0200, Albin Tonnerre wrote: On Mon, Jun 22, 2009 at 12:08:56AM +0400, Nikita V. Youshchenko wrote : On most (all?) SHR snapshot, clock on illume's top panel is digital. However, on Debian/Illume, clock is analog, and too small to be usable. Any hints how to get digital clock on debian/illume? Thanks. Use the tclock module. It's not packaged, but you can find it at svn http://svn.enlightenment.org/svn/e/trunk/E-MODULES-EXTRA/tclock Regards, Is tclock what SHR is using? Well ASU and I think SHR, too just use a different theme to get a digital clock. I use a custom one, too. Debian/illume just shows you a analog clock, because this is the default theme of illume. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [Debian] Packaging new frameworkd snapshot?
Hi @ll, On Tue, 16 Jun 2009 21:21:36 +0200, Heiko Stübner wrote: currently I'm packaging some fso stuff out of cornucopia. nice :) On Tue, 16 Jun 2009 21:21:36 +0200, Heiko Stübner wrote: The handling of the misc-vapis feels strange as they will conflict with newer vala-versions when these files are included there. So input on how to handle this right would be welcome. uhm options, which come to my mind: a) ask Mickey for the earliest version of vala, which may contain the files and conflict with it. You can update your package if this version stills does not contain the files. b) let vala package conflict with misc-vapis once the files are included in the vala package. On Tue, Jun 16, 2009 at 09:44:22PM +0200, Luca Capello wrote: You replied before me, Heiko: Sebastian, for any work please start from his work. I didn't know somebody was already working on this. I will let Heiko handle this work for now and go packaging SHR stuff using the base from Joachim. On Tue, Jun 16, 2009 at 09:44:22PM +0200, Luca Capello wrote: Heiko, do you mind becoming part of the pkg-fso team? This means that you need an Alioth login http://alioth.debian.org and that everything will be hosted on http://git.debian.org/. This would make it easier to find your work, too :) BTW, now that frameworkd is in Debian main, we also need to think where to upload the new packages: we can upload directly to unstable (and fix things as they break) or to experimental, which will give the user the option to keep a working version, i.e. the old one. +1 for unstable -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
[Debian] gestures in debian
Hi, I couldn't test it so far, because I will send my Freerunner in for hardware fixes, but I just packaged the gestures daemon [1] for Debian :) Instead of providing the scripts for starting the daemon for the correct accelerometer (e.g. gesm-neo2.sh) the Debian package uses a configuration file (/etc/default/gestures-daemon), which contains a single variable DEVICE=... (either neo-top, neo-bottom or wii). If you don't specify any (or an unknown one) the daemon won't start through /etc/init.d/gestures-daemon. In difference to the opkg package gestures stuff is split into 4 packages in Debian: - the gestures daemon - the notifier daemon - the GUI manager - a package containing the libs used by all of them As a result you can install just the core daemon and ignore the notifier. If you test the package please report back if it works correctly. -- Sebastian [1] http://wiki.openmoko.org/wiki/Gestures signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [Debian] 2.6.29 kernel
Hi Luca, On Sun, Jun 14, 2009 at 03:54:32AM +0200, Luca Capello wrote: Should a note be added to the Debian install.sh script or somewhere else (Debian or Openmoko wiki)? while being at it should also be noted, that a recent qi version is needed to work together with gsm firmware moko11 and kernel 2.6.29. I had a qi install from some months ago and as a result gsm modem didn't work. Updating qi solved it. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [Debian] Packaging new frameworkd snapshot?
Hi Michael, On Sun, Jun 14, 2009 at 07:08:23PM +0200, Michael 'Mickey' Lauer wrote: fsousaged greatly improves stability of the framework (read the gory details in #435), hence we need to ship it as part of ms5.5. very nice :) fsodevice -- the successor for the device subsystem in frameworkd -- is also almost ready and will give us a major boost on latency, but we won't delay 5.5 until then. while you are at it - what's the status of fsogsmd and fsotimed? Also I couldn't find any fsogpsd. Are there any plans to write your own fsogpsd, or will you just use gypsy-daemon? If not done by somebody else at that point I'll start packaging cornucopia stuff next week after I'm finished with the gestures daemon. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [Debian] 2.6.29 kernel
Hi, On Sat, Jun 13, 2009 at 11:39:14PM +0200, Sebastian Reichel wrote: = kernel remounts filesystem because ext4 detects crappy filesystem - but I do not even use ext4 :S I just rechecked the thread I started when the last 2.6.29 kernel got out and reread Luca's answer. Adding rootfstype=ext2 to the kernel options solved the problem (since I use qi it's *NOT* default). The problem itself is described (incl. fix) at ext4 mailinglist: http://kerneltrap.org/mailarchive/linux-ext4/2009/1/20/4774074/thread -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
[Debian] Wiki Page with Tips?
Hi, is there any wiki page with tipps for running Debian on the Freerunner? I guess there are quite some tweaks for the Freerunner. For example I want to use my Freerunner as audio player. mplayer will use mp3lib by default for audio playback (file is 128kbit): no mplayer: system: 10% (by htop+sshd via usb) mplayer: - mplayer with mp3lib: (plays almost fluently without anything else running) system: 100% (constant) mplayer: 81% mplayer with ffmpeg: (plays fluently even with htop running) system: 50%-55% mplayer: 25% This is big difference comes due to the fact, that mp3lib is optimized for mmx/sse3/3dnow, which are cpu extensions of x86 processors. These don't exist on ARM. mp3lib also makes usage of floating point operations and the Freerunner's CPU doesn't have a FPU = very slow. ffmpeg decodecs using integers only, which is perfect in Freerunner's case :) The decoder can be changed by just adding -afm ffmpeg or adding afm=ffmpeg into ~/.mplayer/config, but for this one has to know, that it is needed. So I guess the wiki should give a hint. I also think, that a package freerunner-tweaks with some optimized configurations and docs for freerunner might be useful. I thought of a structure similar to this: /usr/share/doc/freerunner-tweaks/examples/mplayer.conf: afm=ffmpeg /usr/share/doc/freerunner-tweaks/mplayer-faster-mp3-decoding.txt: For mp3 decoding using less cpu cycles you should copy the content of /usr/share/.../examples/mplayer.conf to your mplayer config. This will change the mp3 decoding codec to ffmpeg instead of using the x86 optimized mp3lib. /usr/share/doc/freerunner-tweaks/interface-performance.txt: Wifi interface will survive package flooding with only very few cpu usage. If you do the same via USB or Bluetooth your Freerunner will be unusable until you stop the package flooding. Wifi interface is also the fastest of the 3 methods. So if you plan to transfer big files you should use wifi if possible. /usr/share/doc/freerunner-tweaks/mplayer-faster-vorbis-decoding.txt: For vorbis decding using less cpu cycles libvorbisidec should be used. But due to a Debian packaging bug (#506244) mplayer is currently not linked against it. -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [Debian] Kernel package contains binary blobs
On Fri, Mar 13, 2009 at 11:53:46AM +0100, Luca Capello wrote: FYI, I anyway uploaded the 2.6.29 kernel package to the pkg-fso repository, because it should fix some nasty bugs and various people asked for it. Please do not advise for it, it should be tested before going into the public (the pkg-fso repository still provides the 2.6.28 kernel package). I also get this in dmesg: [ 1819.345000] EXT4-fs error (device mmcblk0p2): ext4_free_data: circular indirect block detected, inode=112310, block=512 [ 1819.35] Remounting filesystem read-only But fsck looks ok: # fsck /dev/mmcblk0p2 fsck 1.41.3 (12-Oct-2008) e2fsck 1.41.3 (12-Oct-2008) /dev/mmcblk0p2 contains a file system with errors, check forced. Pass 1: Checking inodes, blocks, and sizes Pass 2: Checking directory structure Pass 3: Checking directory connectivity Pass 4: Checking reference counts Pass 5: Checking group summary information /dev/mmcblk0p2: 89775/496992 files (2.9% non-contiguous), 440685/992984 blocks -- Sebastian Reichel signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: suspend / resume
On Mon, Mar 09, 2009 at 06:58:56PM +0100, Davide wrote: Hi, I am using fso on debian. If I suspend the freerunner with apm -s no problem. But if the gsm is active then the freerunner will wakeup about in a minute ... without I am receiving a call or sms. Any idea? Where can I find the event that wake up the phone? Any log? Hi, 1. You will have to suspend the GSM modem. It would be the easiest to do this via DBus. Here is a small Python script, which will suspend the GSM modem *and* your system: - #!/usr/bin/python import dbus, sys, os, gobject # Setup mainloop and dbus from dbus.mainloop.glib import DBusGMainLoop DBusGMainLoop(set_as_default=True) mainloop = gobject.MainLoop() bus = dbus.SystemBus() def getObject( busname, objectpath ): return bus.get_object( busname, objectpath, follow_name_owner_changes=True ) def getInterface( busname, objectpath, interface ): return dbus.Interface(getObject(busname, objectpath), interface) usageiface = getInterface( org.freesmartphone.ousaged, /org/freesmartphone/Usage, org.freesmartphone.Usage ) usageiface.Suspend() - 2. You will get the wakeup reason by: $ cat /sys/class/i2c-adapter/i2c-0/0-0073/neo1973-resume.0/resume_reason | grep * In your case it should be EINT01_GSM 3. Please be aware of the fact, that the current (2.6.28) debian kernel will resume with an WSOD, because some of Openmoko's recent patches are missing. -- Sebastian Reichel signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: Vibrate
On Mon, Mar 09, 2009 at 07:03:23PM +0100, Davide wrote: How to disable the phone vibrate when receiving a call? Hi, I may be wrong, but I think this is not possible at the moment without changing the SRC of the framework. At least I couldn't find ring-vibrator-strength and message-vibrator-strength config option for the yaml rules files of the framewokrk. If you *really* want to disable you could try commenting out line 225, 264 and 310 in the follwing file: /usr/share/python-support/fso-frameworkd/framework/subsystems/oeventsd/fso_actions.py -- Sebastian Reichel signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [Debian] Please update 2.6.28 kernel
Hi Luca, The discussion about a new kernel started on the pkg-fso-maint mailing list less than two weeks ago [1] and FYI I was already planning to update the Debian kernel, also because of the WiFi problem [2]. The last problem I encounter seems to be solved now [3], so I can go on. ah ok :) I'm not sure if I'm subscribed to pkg-fso-maint yet, but I didn't read this mail. I'll check later if I'm subscribed :) I have already started working on the new kernel locally: is it OK if I package up to commit b9de904e7aadd7b37c4035fd5d6f59e5f80089b6 [4] (which is already 2.6.29) or should I pull again? The more recent commits seem mostly related to GTA03. Looks fine for me, Bluetooth PM resume bug has been fixed one commit before and Audio stuff even before that (I just read it will even save some power :)) As a general note, I strongly disagree to a additional kernel packages, as well as to any additional (i.e. snapshot) package for other FSO software. First, the pkg-fso team is already low in manpower; second, adding new packages means that debugging problems will be a bit more difficult because of differences in package versions. Probably I'm just to impatient waiting for new versions :) P.S.: I get WSOD with 2.6.28 from time to time after resume. My FR was WSOD free with 2.6.24, though. Does any of you have the same behaviour? For different problems, the best thing would be to start a new thread... I have never experienced it, have you tried if it is present with the FSO-MS5 kernel as well? I didn't test it with FSO-MS5, because I'm running Debian only on my Freerunner. Actually my last boot from internal memory was in 2008... I will wait for the 2.6.29 package and recheck with it. Perhaps I've got luck and it's solved there. Depending on the result I'll try with FSO-MS5.1, too. I hope this will fix it [1]. -- Sebastian [1] http://git.openmoko.org/?p=kernel.git;a=commit;h=83cf37799009f6a3db4fef141c798c876124954f signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
[Debian] Please update 2.6.28 kernel
Hi, I request to update the debian kernel. The current one in pkg-fso repo has problems with Bluetooth after suspend/resume if Bluetooth has been disabled before. This has been fixed by Balaji two days ago [1] and would work with a newer kernel. There are also fixes for some other problems with audio and lcm [2]. I just saw, that I looked @ andy-tracking. Perhaps we can get an additional 2.6.29 kernel for the impatient ones? :) P.S.: I get WSOD with 2.6.28 from time to time after resume. My FR was WSOD free with 2.6.24, though. Does any of you have the same behaviour? -- Sebastian [1] http://git.openmoko.org/?p=kernel.git;a=commit;h=75d17a700b015cd5e99636db035c8bd1324eeca4 [2] http://git.openmoko.org/?p=kernel.git;a=commit;h=83cf37799009f6a3db4fef141c798c876124954f signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: Default FSO system group
Hi, On Wed, Mar 04, 2009 at 12:02:50AM +0100, Luca Capello wrote: On Debian, we would like to start running applications as normal user: briefly, in Joachim's words, frameworkd D-Bus config should allow all members of a given system group to use D-Bus calls, but not own the D-Bus names [1]. nice :) Thus, we are faced with a decision: which system group should be set as the default for frameworkd access? On Debian, the system group should be picked up from the ones the base-passwd package installs (Debian Policy §9.2.1 [2]): As far as I understand §9.2.1 you are not allowed to edit /etc/group, but you are allowed to create a new group via postinst and addgroup. In fact there are some packages doing this (e.g. kvm). However, since frameworkd depends on D-Bus, the installation of the latter creates another system group: messagebus In the past, in what IIRC is the relevant thread, I fully supported the choice for the dialout system group [3], because the primary function for the FR is making phone calls. However, given the fact that frameworkd ambition is bigger than simply managing a mobile phone, I understand that the dialout system group can be seen as reductive. That is why I'm asking here :-) In my opinion the framework should get its own group. All the other groups are meant only for parts of the framework (e.g. dialout), others are to powerful (e.g. root). If we have a group just for framework we won't conflict with any other meanings of system group entries. Another option would be to split the functionality, so that dialout group is allowed to use the dial functions, but nothing else, ... But IMHO this way is to much work for developers and users. -- Sebastian Reichel signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: pkg-e and zhone packages updated in my repo
On Mon, Jan 12, 2009 at 11:07:50PM +0300, Nikita V. Youshchenko wrote: Exactly same packages have been uploaded to pkg-fso repository. Hi, I just installed and started them on my Freerunner. First: there are no keyboard information files in illume's module directory, so there are no keyboards. I downloaded them from svn and put them there, now I get the keyboards. Though the entered symbols are not forwarded to the running application :( Second: The Launcher is not working - I don't see any icons to start applications, although there are some .desktop files in /usr/share/applications. I tried to put the apps into Utility Category, which is known to work - nothing. Third: I would like to have the black illume theme from ASU(?). Can you create a package for it? -- Sebastian Reichel ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: installing Xglamo disables [Alt][Ctrl][x] to launch xterm?
On Mon, Dec 15, 2008 at 10:19:20PM +0100, Luca Capello wrote: Hi Sebastian! Well, the above is not completely correct: xserver-xglamo is a kdrive-based server while the Xglamo driver needs the normal X.Org server. Since, as you reported, the problem seems to be in xserver-xglamo not correclty implementing xtest, the Xglamo driver will not solve the problem, because it is not for the kdrive-based server. Thx, bye, Gismo / Luca Hi Luca, well I hoped it would :/ But I guess we can ask graeme to correct the xtest implementation, so that it will work =) -- Sebastian Reichel signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: installing Xglamo disables [Alt][Ctrl][x] to launch xterm?
On Sat, Dec 06, 2008 at 08:20:43PM +0100, e.waelde wrote: When installing xserver-xglamo [...] then the keyboard shortcut [Alt][Ctrl][x] to launch xterm does not work any more. Hi, this is a known bug [1] with xserver-xglamo and matchbox's keyboard. You can either use another virtual keyboard or you use xserver-xorg. I guess this bug will be fixed automatically, when the xglamo driver for xorg is ready (graeme is working on this). -- Sebastian Reichel [1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=501575 pgpBBgXw2cOEf.pgp Description: PGP signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: [PATCH] Zhone: show the name from the phonebook during call
On Wed, Nov 26, 2008 at 11:07:52PM +0100, Stefan Schmidt wrote: Applied. Thanks for it. One pain less using it as daily phone. Hi, Joachim or Luca, can one of you please update the Zhone package in Debian? IMHO this feature is a must-have ;) -- Sebastian signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: milestone4 now in Debian
On Wed, Nov 19, 2008 at 03:04:53PM +0100, Sebastian Reichel wrote: Though remoko was not able to power on bluetooth, since the DBUS paths seems to have changed. According to remoko website it's working with ms4, so a new version of this may also be needed. Just tried it with my own notebook - worked there, seems the warning, which is displayed can be ignored (but looks not very nice ;)) and Bluetooth turned off is shown when pressing CTRL+C. Which confused me. -- Sebastian Reichel signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland