FSO kernel configuration [Was: Re: Trouble Building Kernel]

[shawnzier]

On Thu, Sep 11, 2008 at 01:23:48AM +0800, Andy Green wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> Somebody in the thread at some point said:
> 
> |> Any recent OM kernel package should have this already.  I don't
> |> know what the deal is with Debian and which config they use.  The
> |> ones that OM use to generate the packages are defconfig-gta01 and
> |> defconfig-gta02, these are set to cook modules for g_ether and the
> |> other gadgets.
> |>
> | I'm not sure this is the case.
> |
> | I am using the latest fso-unstable images from
> | http://shr.bearstech.com
> 
> Well I did say "OM kernel package".
> 
I'm sorry, I thought FSO was an openmoko distribution. I just assumed that 
because Mickey seems to be the leader on that project and he had an 
@openmoko.org address. 
> | I'm sshing over the usb networking connection and lsmod gives me
> | nothing. Shouldn't the usb networking module show up in lsmod if it
> | is compiled as a module?
> 
> Yes, it's a pretty good test.
> 
> | Should I file a bug for this at trac.freesmartphone.org? Is this an
> | oversight or did they intend to build their kernel/rootfs this way?
> | Maybe the autobuilder is setup to use defconfig-2.6.24 instead?
> 
> I have no idea about FSO intention about the kernel or which config they
> like to use.  defconfig-2.6.24 is nice because you can just update the
> monolithic kernel and forget about modules for most cases, but the
> defconfig-gta02 that makes more modules is best for packaging since it
> gives the flexibility about gadget mode you want.
> 
> I would bring it up with them, maybe it's like it is right now because
> nobody mentioned about different config possibility being useful.
> 
Ok. I guess I should email Mickey directly then or should I post to the devel 
list? Thanks for the help thus far.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Trouble Building Kernel

[shawnzier]

On Wed, Sep 10, 2008 at 10:57:47PM +0800, Andy Green wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> Somebody in the thread at some point said:
> 
> | Yeah. I'm not trying to make things difficult for myself. I just
> | wanted g_ether as a module so I could use the freerunner as a usb
> | flash drive, which requires removing the g_ether module.
> 
> Any recent OM kernel package should have this already.  I don't know
> what the deal is with Debian and which config they use.  The ones that
> OM use to generate the packages are defconfig-gta01 and defconfig-gta02,
> these are set to cook modules for g_ether and the other gadgets.
> 
I'm not sure this is the case.

I am using the latest fso-unstable images from http://shr.bearstech.com 

I'm sshing over the usb networking connection and lsmod gives me nothing. 
Shouldn't the usb networking module show up in lsmod if it is compiled as a 
module?

Should I file a bug for this at trac.freesmartphone.org? Is this an oversight 
or did they intend to build their kernel/rootfs this way? Maybe the autobuilder 
is setup to use defconfig-2.6.24 instead?


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Trouble Building Kernel

[shawnzier]

On Wed, Sep 10, 2008 at 07:18:36PM +0800, Andy Green wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> Somebody in the thread at some point said:
> | On Tue, Sep 09, 2008 at 03:38:41PM +0800, Andy Green wrote: Somebody
> | in the thread at some point said: | I wanted to build a kernel to get
> | some drivers as modules instead of built-in. I followed | |
> | http://wiki.openmoko.org/wiki/Toolchain
> |
> | | + BRANCH='(no'
> |
> | There doesn't seem to be an active branch in git?  Try making one
> | active.
> |> OK. That worked. I've built the kernel with no errors. Now how do I
> |> build the modules? I would have to copy these to the freerunner
> |> before flashing the new kernel, from what I understand.
> 
> The modules are all built already as part of the process, but you need
> to meddle around with make modules_install to get them marshalled
> somewhere and scp them across.  Or tar them and pipe them into untar via
> ssh or somesuch.
Yeah. I tried this using the SYSROOT environment variable, but I think the 
versioning is messed up. I'll try to figure out what I did wrong and I'll post 
back to the list when I find an answer. I think what might have happened is I 
used the build script and tried to do the make modules_install part in another 
shell. Since you're setting temporary environment variables in the build 
script, this probably wouldn't work. Stupid rookie mistake.
> 
> If you use the defconfig-2.6.24 .config then all the critical drivers
> for normal use (ethernet over USB, bt, sound, etc) are already in the
> monolithic kernel.  
Yeah. I'm not trying to make things difficult for myself. I just wanted g_ether 
as a module so I could use the freerunner as a usb flash drive, which requires 
removing the g_ether module.

>Using the build script means you won't get old
> modules from another kernel confused with modules usable on this kernel
> either because the /lib/modules path is unique by git head hash.  So you
> can typically run a rootfs just with the monolithic kernel from that
> build even DFU'd in.
> 
> If you routinely want the full module set, you can build using OE recipe
> to make OE package... but don't ask me how to do it :-)
Yeah. I think I've already taken up enough of your time with dumb questions. 
I'd rather you spend it writing kernel patches.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Glad to see Gentoo-openmoko distro

[shawnzier]

On Tue, Sep 09, 2008 at 07:25:59AM -0700, Rodney Myers wrote:
> On Sep 9, 2008, at 7:05 AM, [EMAIL PROTECTED] wrote:
>
>> On Tue, Sep 09, 2008 at 02:51:14PM +0100, Rui Miguel Silva Seabra wrote:
>>> On Tue, Sep 09, 2008 at 01:00:36PM +0100, Al Johnson wrote:
 On Tuesday 09 September 2008, Rui Miguel Silva Seabra wrote:
> I hope you guys don't mind killing your phone or uSD's memory sooner
> than everyone else :)

 Why? Are you making some unwarranted assumptions about using Gentoo on
 embedded systems?
>>>
>>> No unwarranted assumption, the only argument you could throw back at me 
>>> is
>>> an argument against using Gentoo instead of any .deb or .rpm based
>>> distribution ;)
>>>
>>> Please, no distro-war, you're free to kill your flash devices sooner ;)
>> Why would it kill the flash?
>
> With all of the thrashing, while it's attempting to compile everything?
>
> I know that I have gentoo on a MythTV box, and when installing anything 
> heavy, forget using it for 1/2 to 3/4 of a day.
See Dennis's reply. 



> ___
> Openmoko community mailing list
> community@lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Trouble Building Kernel

[shawnzier]

On Tue, Sep 09, 2008 at 03:38:41PM +0800, Andy Green wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> Somebody in the thread at some point said:
> | I wanted to build a kernel to get some drivers as modules instead of
> built-in. I followed
> |
> | http://wiki.openmoko.org/wiki/Toolchain
> 
> | + BRANCH='(no'
> 
> There doesn't seem to be an active branch in git?  Try making one active.
OK. That worked. I've built the kernel with no errors. Now how do I build the 
modules? I would have to copy these to the freerunner before flashing the new 
kernel, from what I understand.
> 
> - -Andy
> -BEGIN PGP SIGNATURE-
> Version: GnuPG v1.4.9 (GNU/Linux)
> Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org
> 
> iEYEARECAAYFAkjGJ/wACgkQOjLpvpq7dMpanACghoNEQWgZGUxu9lDE8Qp2qnR/
> q40An3KIrU3Sx1ySw24s1ai12DOd5Y5Z
> =yBC0
> -END PGP SIGNATURE-
> 
> ___
> Openmoko community mailing list
> community@lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Glad to see Gentoo-openmoko distro

[shawnzier]

On Tue, Sep 09, 2008 at 02:51:14PM +0100, Rui Miguel Silva Seabra wrote:
> On Tue, Sep 09, 2008 at 01:00:36PM +0100, Al Johnson wrote:
> > On Tuesday 09 September 2008, Rui Miguel Silva Seabra wrote:
> > > I hope you guys don't mind killing your phone or uSD's memory sooner
> > > than everyone else :)
> > 
> > Why? Are you making some unwarranted assumptions about using Gentoo on 
> > embedded systems? 
> 
> No unwarranted assumption, the only argument you could throw back at me is
> an argument against using Gentoo instead of any .deb or .rpm based
> distribution ;)
> 
> Please, no distro-war, you're free to kill your flash devices sooner ;)
Why would it kill the flash?

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Trouble Building Kernel

[shawnzier]

On Tue, Sep 09, 2008 at 03:38:41PM +0800, Andy Green wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> Somebody in the thread at some point said:
> | I wanted to build a kernel to get some drivers as modules instead of
> built-in. I followed
> |
> | http://wiki.openmoko.org/wiki/Toolchain
> 
> | + BRANCH='(no'
> 
> There doesn't seem to be an active branch in git?  Try making one active.
I'll adjust the wiki once i figure out the command. Maybe you can help me 
along. Should I use 

git-checkout -b mystable origin/stable  ?

Excuse my ignorance, but I'm brand new to git.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Trouble Building Kernel

[shawnzier]

I wanted to build a kernel to get some drivers as modules instead of built-in. 
I followed

http://wiki.openmoko.org/wiki/Toolchain

When I went to build the kernel, I had to change the 'build' script to point to 
the toolchain. After that was fixed, I got this error when running the 'build' 
script.

# configuration written to .config
#
+ VERSION=
+ '[' -d .git ']'
++ git show --pretty=oneline
++ cut '-d ' -f1
++ cut -b1-16
++ head -n1
+ HEAD=ca19d156400f8179
++ git branch
++ grep '^*'
++ cut '-d ' -f2
+ BRANCH='(no'
+ VERSION='-(no_ca19d156400f8179'
+ make -j5 ARCH=arm 'EXTRAVERSION=-(no_ca19d156400f8179'
  CHK include/linux/version.h
  SYMLINK include/asm-arm/arch -> include/asm-arm/arch-s3c2410
/bin/sh: -c: line 0: syntax error near unexpected token `('
/bin/sh: -c: line 0: `echo 2.6.24-(no_ca19d156400f8179 > 
include/config/kernel.release'
make: *** [include/config/kernel.release] Error 2
make: INTERNAL: Exiting with 6 jobserver tokens available; should be 5!
+ exit 1


Anybody have any idea what might be wrong here?


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Glad to see Gentoo-openmoko distro

[shawnzier]

On Tue, Sep 09, 2008 at 08:59:22AM +0800, Dennis.Yxun wrote:
> HI Devs:
> 
> As a die-hard Gentoo fans, I'd be happy to see someone
> 
> deliver a Gentoo openmoko distribution which leverage the greatness of
> portage.
> 
> Thank you, It's really great!
> 
>I'm downloading stage now, but due to my poor network,
> 
> the download speed is quite low, and It about takes me 5Ds to finish the
> stage1 tarball~
> 
> so I'm thinking about build the system all by myself.
> 
>I've have some experience of setting up the cross-compiler tools in
> Gentoo,
> 
> but find it's not easy to build the base system... I may try this later, and
> post here.
> 
> Proposal, Can we setup an openmoko-gentoo overlay, so people who are
> interested
> 
> to this may build on it. And also people can contribute same ebuilds back.

I'm a gentoo user too. Would love to see something done with it. I think 
someone is already working on this. I saw someone talking about it a few days 
ago on gentoo-embedded. Here's the thread.

http://archives.gentoo.org/gentoo-embedded/msg_92e6f8eb918aa1d5277481ae3c205138.xml



___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

FSO and g_ether as module [Was: Re: usb mass media storage]

[shawnzier]

On Fri, Sep 05, 2008 at 01:11:30PM +0800, Andy Green wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> Somebody in the thread at some point said:
> 
> | Where did you get that kernel image? I haven't been able to find it
> | anywhere. Is there any other way to get those commands?
> 
> Any recent kernel package should have these modules.
I am using fso-unstable. The g_ether module must be compiled in because lsmod 
gives me nothing. I tried using a kernel from the daily builds
http://downloads.openmoko.org/daily/

but usb networking won't work. It tries to configure networking on boot and 
fails with "no such device". Is there any way I can get g_ether compiled as a 
module with FSO? 




___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: usb mass media storage

[shawnzier]

On Fri, Sep 05, 2008 at 01:04:52PM +0200, pHilipp Zabel wrote:
> On Fri, Sep 5, 2008 at 4:41 AM, Dale Maggee <[EMAIL PROTECTED]> wrote:
> > Petr Vanek wrote:
> >> I have spent a while googling this up and have no wiki skills but
> >> someone might find this useful:
> >>
> >> to make FR act as a memory stick, here is what my script does:
> >>
> >>
> >> #!/bin/sh
> >> /etc/init.d/networking stop
> >> rmmod g_ether
> >> modprobe g_file_storage file=/dev/mmcblk0p1
> >>
> >> --
> >> Petr Vaněk
> >> http://biodynamika.cz
> >>
> > very cool, but before I test it out (potentially breaking my usb
> > networking in the process), I have a couple of questions:
> >
> > 1. will this persist over a reboot, or will it revert back to being an
> > ethernet gadget when I reboot?
> 
> It's not persistent unless you add something like this to the init
> scripts. After rebooting the ethernet gadget will be back.
> 
> > 2. is the following sufficient to switch it back?
> >
> > #!/bin/sh
> > rmmod g_file_storage
> > modprobe g_ether
> > ifup usb0
> >
> > (assuming I replaced '/etc/init.d/networking stop' with 'ifdown usb0' as
> > recommended by Daniel)
> >
> > 3. can somebody point me to a resource for the g_file_storage module
> > where I can learn more about it? I did a couple of quick googles but
> > didn't see anything promising. specifically I'd like to know more about
> > the file parameter - I assume that in this scenario our "usb stick" is
> > using partition/device information from /dev/mmcblk0p1, and will be
> > partitioned in the same way as the SD card? what If I wanted to have my
> > home directory accessible in "usb stick" mode?
> 
> I suggest to read the comment in the source code, it's quite detailed:
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=drivers/usb/gadget/file_storage.c;hb=HEAD
> 
> The file parameter determines the raw backing store for the storage
> device, so if you
> have file=/dev/mmcblk0p1, its contents will directly appear as
> /dev/sdx on the host. As mmcblk0p1 doesn't contain a partition table,
> so won't sdx. If you want to export the whole device, use
> file=/dev/mmcblk0, you'll get the partition table in /dev/sdx and the
> kernel will parse it and create /dev/sdx1 (containing the contents of
> /dev/mmcblk0p1) etc.
> 
> The host has direct access to the block device, so it is important
> that /dev/mmcblk0p1 is not mounted on both the phone and the host at
> the same time (unless it's mounted ro on both).

What if it was mounted ro on the phone side and rw on the host side? Would that 
cause problems?


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Om2008.8: execute script on incoming call

[shawnzier]

On Thu, Sep 04, 2008 at 11:04:38AM +0200, Michael 'Mickey' Lauer wrote:
> Thanks for posting the fixed one, I pushed it into the examples
> directory.
Cool. Glad you could find it useful.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Om2008.8: execute script on incoming call

[shawnzier]

On Tue, Sep 02, 2008 at 12:04:03AM +0200, Michael 'Mickey' Lauer wrote:
> Am Montag, 1. September 2008 23:18:02 schrieb [EMAIL PROTECTED]:
> > Should the else statement be at the same indent as the if?
> 
> No, the else refers to the try/except clause.
> 
> >   DBusException: org.freedesktop.DBus.Error.UnknownMethod: Method
> > "Release" with signature "i" on interface "org.freesmartphone.GSM.Call"
> > doesn't exist
> 
> Oops, it's probably called ReleaseCall. If in doubt about signatures, use 
> mdbus to introspect.

Finally figured it out. 

obj = bus.get_object( "org.freesmartphone.ogsmd", 
"/org/freesmartphone/Device/GSM" ) 

should be

obj = bus.get_object( "org.freesmartphone.ogsmd", 
"/org/freesmartphone/GSM/Device" )

I also had to import subprocess and dbus.mainloop.glib

Here is the corrected python code in full:

import dbus 

   
import dbus.mainloop

   
import dbus.mainloop.glib   

   
import gobject  

   
import subprocess   

   


   


   
actions = { \   

   
  "+491002":"/usr/bin/foo1",

  
  "+491002":"/usr/bin/foo2",

   
  "+491003":"/usr/bin/foo3" }   

   


   
def onCallStatus( index, status, properties ):  

   
if status == "incoming":

   
try:

   
 action = actions[properties["peer"]]   

   
except KeyError:

   
 pass 

Re: Om2008.8: execute script on incoming call

[shawnzier]

On Tue, Sep 02, 2008 at 12:04:03AM +0200, Michael 'Mickey' Lauer wrote:
> Am Montag, 1. September 2008 23:18:02 schrieb [EMAIL PROTECTED]:
> > Should the else statement be at the same indent as the if?
> 
> No, the else refers to the try/except clause.
> 
> >   DBusException: org.freedesktop.DBus.Error.UnknownMethod: Method
> > "Release" with signature "i" on interface "org.freesmartphone.GSM.Call"
> > doesn't exist
> 
> Oops, it's probably called ReleaseCall. If in doubt about signatures, use 
> mdbus to introspect.
I think you were right the first time. 

mdbus -s org.freesmartphone.frameworkd /org/freesmartphone/GSM/Device


[METHOD]org.freesmartphone.GSM.Call.Activate( i:index )
[METHOD]org.freesmartphone.GSM.Call.ActivateConference( i:index )
[METHOD]org.freesmartphone.GSM.Call.HoldActive()
[METHOD]org.freesmartphone.GSM.Call.Initiate( s:number, s:type_ )
[METHOD]org.freesmartphone.GSM.Call.ListCalls()
[METHOD]org.freesmartphone.GSM.Call.Release( i:index )
[METHOD]org.freesmartphone.GSM.Call.ReleaseAll()
[METHOD]org.freesmartphone.GSM.Call.ReleaseHeld()


It's still giving me that same error though. I'm probably misunderstanding 
something badly. Like I said, I'm brand new to python, dbus and programming in 
general. I'm just trying to get a simple program together that I can study and 
hopeully expand on.

Has the guy that started this thread (Ole) had any luck getting this example to 
work?

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Om2008.8: execute script on incoming call

[shawnzier]

On Mon, Sep 01, 2008 at 08:47:55PM -0400, Shawn Zier wrote:
> On Thu, Aug 28, 2008 at 01:18:29AM +0200, Michael 'Mickey' Lauer wrote:
> > def onCallStatus( index, status, properties ):
> > if status == "incoming":
> > try:
> >  action = actions[properties["number"]]
> > except KeyError:
> >  pass
> > else:
> >   obj = bus.get_object( "org.freesmartphone.ogsmd", 
> > "/org/freesmartphone/Device/GSM" ) 
> >   callInterface = dbus.Interface( obj, 
> > "org.freesmartphone.GSM.Call" )
> >   callInterface.Release( index )
> >   subprocess.Popen( action, shell=True )
> > 
> 
> I think the 
> 
> action = actions[properties["number"]]
> 
> line is wrong. Should it be
> 
> action = action[properties["peer"]]
> 
> Here is the dbus signal when I call the Neo. Phone number has been scrubbed.
> 
>  [SIGNAL]org.freesmartphone.GSM.Call.CallStatusfrom :1.5 
> /org/freesmartphone/GSM/Device
>  (dbus.Int32(1), dbus.String(u'incoming'), 
> dbus.Dictionary({dbus.String(u'status'): dbus.String(u'incoming', 
> variant_level=1), dbus.String(u'peer'): dbus.String(u'XXX', 
> variant_level=1), dbus.String(u'direction'): dbus.String(u'incoming', 
> variant_level=1), dbus.String(u'line'): dbus.Int32(0, variant_level=1)}, 
> signature=dbus.Signature('sv')))
> 
> 
> I changed it to my suggestion and it didn't get the exception. Now I am 
> getting an exception on 
> 
> subprocess.Popen( action, shell=True )
> 
> ERROR:dbus.connection:Exception in handler for D-Bus signal:
> Traceback (most recent call last):
>   File "/usr/lib/python2.5/site-packages/dbus/connection.py", line 214, in 
> maybe_handle_message
>   self._handler(*args, **kwargs)
>   File "oncallscript.py", line 17, in onCallStatus
>   subprocess.Popen( action, shell=True )
>   NameError: global name 'subprocess' is not defined
> 
> I found something about the python-subprocess package on the wiki. Do I need 
> this? It says it's not available in official repos. Would it be safe to use 
> ScaredyCats repo?
> 
> http://wiki.openmoko.org/wiki/Python
> 
> Thanks for any more help you can give me.

Replying to my own post. I installed the python-subrocess package from the 
fso-testing repository. Seems to work fine. 

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Om2008.8: execute script on incoming call

[shawnzier]

On Thu, Aug 28, 2008 at 01:18:29AM +0200, Michael 'Mickey' Lauer wrote:
> def onCallStatus( index, status, properties ):
> if status == "incoming":
> try:
>  action = actions[properties["number"]]
> except KeyError:
>  pass
> else:
>   obj = bus.get_object( "org.freesmartphone.ogsmd", 
> "/org/freesmartphone/Device/GSM" ) 
> callInterface = dbus.Interface( obj, 
> "org.freesmartphone.GSM.Call" )
>   callInterface.Release( index )
>   subprocess.Popen( action, shell=True )
> 

I think the 

action = actions[properties["number"]]

line is wrong. Should it be

action = action[properties["peer"]]

Here is the dbus signal when I call the Neo. Phone number has been scrubbed.

 [SIGNAL]org.freesmartphone.GSM.Call.CallStatusfrom :1.5 
/org/freesmartphone/GSM/Device
 (dbus.Int32(1), dbus.String(u'incoming'), 
dbus.Dictionary({dbus.String(u'status'): dbus.String(u'incoming', 
variant_level=1), dbus.String(u'peer'): dbus.String(u'XXX', 
variant_level=1), dbus.String(u'direction'): dbus.String(u'incoming', 
variant_level=1), dbus.String(u'line'): dbus.Int32(0, variant_level=1)}, 
signature=dbus.Signature('sv')))


I changed it to my suggestion and it didn't get the exception. Now I am getting 
an exception on 

subprocess.Popen( action, shell=True )

ERROR:dbus.connection:Exception in handler for D-Bus signal:
Traceback (most recent call last):
  File "/usr/lib/python2.5/site-packages/dbus/connection.py", line 214, in 
maybe_handle_message
  self._handler(*args, **kwargs)
  File "oncallscript.py", line 17, in onCallStatus
  subprocess.Popen( action, shell=True )
  NameError: global name 'subprocess' is not defined

I found something about the python-subprocess package on the wiki. Do I need 
this? It says it's not available in official repos. Would it be safe to use 
ScaredyCats repo?

http://wiki.openmoko.org/wiki/Python

Thanks for any more help you can give me.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Om2008.8: execute script on incoming call

[shawnzier]

On Thu, Aug 28, 2008 at 01:18:29AM +0200, Michael 'Mickey' Lauer wrote:
> Am Donnerstag 28 August 2008 01:10:25 schrieb Michael 'Mickey' Lauer:
> > def onCallStatus( index, status, properties ):
> > if status == "incoming":
> > try:
> >  action = actions[properties["number"]]
> > except KeyError:
> >  pass
> > else:
> >   subprocess.Popen( action, shell=True )
> 
> Bummer, forgot to hangup here... use that:
> 
> def onCallStatus( index, status, properties ):
> if status == "incoming":
> try:
>  action = actions[properties["number"]]
> except KeyError:
>  pass
> else:
>   obj = bus.get_object( "org.freesmartphone.ogsmd", 
> "/org/freesmartphone/Device/GSM" ) 
> callInterface = dbus.Interface( obj, 
> "org.freesmartphone.GSM.Call" )
>   callInterface.Release( index )
>   subprocess.Popen( action, shell=True )
> 
Should the else statement be at the same indent as the if? I am really new to 
programming in general and this is the first python script I've tried to work 
with, so forgive me if I'm ignorant to something. I've tried it both ways. If I 
try it with the else at the same indent as the if, nothing happens until I 
hangup from the phone I'm calling from, when I get an exception from dbus. If I 
try it the way you have here, nothing happens at all. I tried putting some 
print statements in after the 

obj = bus.get_object( "org.freesmartphone.ogsmd", 
"/org/freesmartphone/Device/GSM" ) 

and it never gets to that print statement when the else statement is at the 
indent level in your snippet.

The exception I get when I put the else statement at the same level as the if 
statement is (This only happens when I manually hang up from either the calling 
end or the receiving (neo) end:

ROR:dbus.connection:Exception in handler for D-Bus signal:
Traceback (most recent call last):
File "/usr/lib/python2.5/site-packages/dbus/connection.py", line 214, in 
maybe_handle_message
  self._handler(*args, **kwargs)
File "oncallscript.py", line 17, in onCallStatus
  callInterface.Release( index )
File "/usr/lib/python2.5/site-packages/dbus/proxies.py", line 68, in __call__
  return self._proxy_method(*args, **keywords)
File "/usr/lib/python2.5/site-packages/dbus/proxies.py", line 140, in __call__
  **keywords)
File "/usr/lib/python2.5/site-packages/dbus/connection.py", line 607, in 
call_blocking
  message, timeout)
  DBusException: org.freedesktop.DBus.Error.UnknownMethod: Method "Release" 
with signature "i" on interface "org.freesmartphone.GSM.Call" doesn't exist

I am using vanilla FSO milestone 2 with no updates. If anybody has any ideas, 
please let me know. 


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Repository and Images

[shawnzier]

On Wed, Aug 27, 2008 at 07:13:58AM -0400, Geoff Ruscoe wrote:
> On Tue, Aug 26, 2008 at 9:07 AM, Julian Chu <[EMAIL PROTECTED]> wrote:
> 
> > Hi Community,
> >
> >
> >   *) A new 2008.8 stable image
> >
> >  [2]http://downloads.openmoko.org/releases/Om2008.8-update/
> >
> >  Oh, it's daily build.
> >
> >   *) The repository of 2008.8
> >  After the building process complete, we also sync the repository[3]
> >  People can upgrade from there.
> >
> >  [3] http://downloads.openmoko.org/repository/Om2008.8/
> >
> >   *) Base Images and repository
> >  [5] http://downloads.openmoko.org/repository/testing/
> >
> 
> Okay I think these details above sound great, but when I tried them out they
> did not go as expected.  Maybe we can get some clarification on the correct
> path for a user (like me :-)).
> 
> Anyway,
> Starting point 2008.08 w/ Zeche's testing repos -- everything up to date.
> Get this mail installed the testing image
> created:
> testing-all-feed.conf
> testing-armv4t-feed.conf
> testing-om-gta02-feed.conf
> all pointing to the testing repository listed above (5)
> 
> did the opkg update opkg upgrade
> This did lots of stuff ... I was getting excited about the new code

It also should have asked you if you want to update certain config
files. For some reason, with these repositories, they try to overwrite
your /etc/opk/*.conf files. It does ask you though, and you can see a
diff of the files by pressing D when it asks. 



___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Freerunner CAD Files - IGES ans STEP formats availalbe - Who wants to test them?

[shawnzier]

On Fri, Aug 22, 2008 at 04:52:54PM +0200, Lothar Behrens wrote:
> Am Donnerstag, den 21.08.2008, 13:55 +0200 schrieb Lothar Behrens:
> > Hi,
> > 
> > 
> > just a note: I knew about another CAD software, that will convert IGES
> > or ProE format and is Open Source.
> > 
> > 
> > See this link for more information: http://brlcad.org/
> > 
> 
> Hi,
> 
> I do currently convert these iges files into g files that could be read
> directly by mged of the BRL-CAD package. If someone is interested, I
> could upload a tgz file to anywhere if the conversion is ready.
> 
> The CAD package also was compileable without any problems on my Debian
> Etch (PPC)
> 
> Regards
> 
> Lothar

I am definitely interested. When I tried to convert them myself, the
reference planes were in the files. I am new to brl-cad and wasn't sure
if it was possible to take them out if they weren't defined as separate
primitives or whatever.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Official update feeds "testing" on downloads.openmoko.org

[shawnzier]

On Thu, Aug 21, 2008 at 09:39:42AM +0200, Olivier Berger wrote:
> Kevin Zuber <[EMAIL PROTECTED]> writes:
> 
> > Hi,
> >
> > I've just noticed on a new "testing" repository on
> > downloads.openmoko.org [1]. Are that the official update feeds for ASU
> > or another distribution? Is the merge of the two branches finished?
> > Should I use this repo instead of the "zecke" repo now?
> >
> > Thanks for answers.
> >
> > Kev
> >
> > [1] http://downloads.openmoko.org/repository/testing/
> >
> 
> I just had a look at contents of
> http://downloads.openmoko.org/repository/testing/om-gta02/distro-feed-configs_1.0-r0.01_om-gta02.ipk
> which may be the package providing the opkg feeds config for that very
> feed if things were consistent...
> 
> But the feeds definition there refer to
> http://buildhost.openmoko.org/daily-feed/ ... so I'm not so sure it's
> an operational testing upgrade feed for 2008.8/ASU.
> 
Just tried again today. The packages still don't get upgraded using
opkg ugrade. It says they do during the process, but they definitely
aren't. It's almost like I did opkg -test upgrade.

But, if I upgrade a package specifically, it works fine. I'm not about
to pick out each package and do an opkg upgrade on it. I can't figure
out what would cause this. I think zeckes feeds worked fine. I'll have
to try them again and test the same way.

Anyhow, if any one tries these feeds and it works for them, please let
me know.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Official update feeds "testing" on downloads.openmoko.org

[shawnzier]

On Thu, Aug 21, 2008 at 09:39:42AM +0200, Olivier Berger wrote:
> Kevin Zuber <[EMAIL PROTECTED]> writes:
> 
> > Hi,
> >
> > I've just noticed on a new "testing" repository on
> > downloads.openmoko.org [1]. Are that the official update feeds for ASU
> > or another distribution? Is the merge of the two branches finished?
> > Should I use this repo instead of the "zecke" repo now?
> >
> > Thanks for answers.
> >
> > Kev
> >
> > [1] http://downloads.openmoko.org/repository/testing/
> >
> 
> I just had a look at contents of
> http://downloads.openmoko.org/repository/testing/om-gta02/distro-feed-configs_1.0-r0.01_om-gta02.ipk
> which may be the package providing the opkg feeds config for that very
> feed if things were consistent...
> 
> But the feeds definition there refer to
> http://buildhost.openmoko.org/daily-feed/ ... so I'm not so sure it's
> an operational testing upgrade feed for 2008.8/ASU.
> 
To keep it from trying to point to the daily builds, just don't accept the 
changes to the config files when opkg asks you to.
-- INSERT --

I tried that feed
(http://downloads.openmoko.org/repository/testing/*) last night and got a 
really weird result. I did an
opkg update
opkg ugrade

It seemed to update a bunch of packages, but then I tried an
opkg list 

and none of the packages actually showed the newer versions. I just
checked by looking at the actual versions of the packages in the
repositories. I can't understand what would cause this exactly.

I'll try them again today and let you know what happens.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Bluetooth, WiFi and Zecke's feeds?

[shawnzier]

On Wed, Aug 20, 2008 at 01:03:52PM +0200, Christ van Willegen wrote:
> Hi,
> 
> I'm using 2008.8 with Zecke's testing and dev feeds.
> 
> At the moment, I can enable WiFi from the settings menu, but am unable
> to associate with my router (someone on IRC hinted that I sould run
> wpa_supplicant in debug mode, and it told me that it didn't get a
> timely reply from my router (WRT54G). Can anyone comment on that?)
> 
> Also, I can switch BlueTooth on and off, but 'hcitool scan' gives me
> 'Device unavailable' with both settings.
> 
> Should I create a ticket somewhere, or are these known bugs and
> (hopefully!) already ironed out?

I don't think zecke's feeds are official, so I dont know if its
appropriate to file bugs with OpenMoko.

You can always check the bug tracker though.

http://docs.openmoko.org/trac/report


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Lock version of package with opkg

[shawnzier]

On Wed, Aug 20, 2008 at 01:09:15PM +1000, Dale Maggee wrote:
> [EMAIL PROTECTED] wrote:
> > I'm using the scaredycat repo and gpsd doesn't seem to work right. Is
> > there any way to lock the old version so I can do an 'opkg upgrade'
> > without upgrading gpsd?

> I think this might be the same problem I'm having, how would I go about 
> downgrading to a working version? I'm assuming I'd use "opkg 
> -force-downgrade", but what version of gpsd is known to be working? link?

Try using cat to read the /dev/ttySAC* devices and see if you get NMEA
data from any of them first. I have a suspicion that whatever handles the 
naming of the
devices changed after an opkg upgrade to the daily feeds.

Then change your /etc/default/gpsd file so that it has whatever device
spits out NMEA data.
 
Let me know how this turns out.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Lock version of package with opkg

[shawnzier]

I'm using the scaredycat repo and gpsd doesn't seem to work right. Is
there any way to lock the old version so I can do an 'opkg upgrade'
without upgrading gpsd?

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: dfu_upload error -84 when trying to backup images

[shawnzier]

On Sat, Aug 16, 2008 at 09:36:43PM +0530, Vikas Saurabh wrote:
> I dont think this is related to the architecture. I had faced similar
> issue on core duo (i had put a comment on the wiki for the same). Btw,
> i didnt have to flash the back up image so i dont know if it actually
> worked out correctly.

So does flashing new images work? Is it just uploading that's broken?


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

dfu_upload error -84 when trying to backup images

[shawnzier]

I'm getting the following error when trying to backup rootfs:

dfu_upload error -84

I've been able to back up the other partitions, but the rootfs fails
every time at 258076672 Bytes. I think this is the size of the rootfs,
but how can I be sure?

I don't want to flash it with this image if there is missing data. I
also don't want to try to flash it with 2008.8 if I'm not sure that it
will complete. If it doesn't complete, will I be left with a brick until
I can get it to complete successfully, or does it not write until the
whole image is transferred?

The host is a Macbook Pro CoreDuo (not C2D) running Gentoo.

Has anybody had a similar experience? 


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Tangogps pre-cache?

[shawnzier]

On Sun, Aug 17, 2008 at 12:19:15AM +1000, Robert William Hutton wrote:
> Robert William Hutton wrote:
> > Before you do this, make sure you go into Config and change the cache 
> > directory to something under /media/card.  By default it's in /tmp, and 
> > you lose the contents of /tmp each time you reboot.
> 
> Hey is anyone else having a problem with their Cache Dir being reset to 
> /tmp/Maps/OSM after every restart of tangoGPS?
I had this same problem. You have to hit the "Save respository
information" button after enterning the filepath.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: Tangogps pre-cache?

[shawnzier]

On Tue, Jul 29, 2008 at 03:50:30PM -0600, -stacy wrote:
> Steven Kurylo wrote:
> > The tango gps documentation says "Additionally you can conveniently
> > pre-cache areas."  Can anyone tell me how to do this?
> > 
> > I can scroll around and have those tiles cached for off line use, but
> > a pre-cache would be great.  Downloading all the OSM tiles and all the
> > zoom levels for a 20km radios would be excellent.

Just go to the area of the map that you want to download and click on
the map. A menu should come up that has a "map download" option. Another
menu will come up that asks you what zoom level (im assuming this means
relative to the current level) that you want to download. Finally, click
OK.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community