Hi

I'm having exactly the same issue. What I see is also that if I select any scenario/floorplan in a web orbiter the screen changes also for others web orbiters.
This makes me think that in fact only one instance of Proxy orbiter is running.

Some times ago I posted a question related to the possibility to have multiple instances of proxy web running, and tshack909 confirmed that this is possible.

So probably we are doing it in the wrong way. Can anyone point us in the right direction?

TIA
Marco

Hi colinjones

I have similar issue with my VIA EPIA M10000 MD, but in my case sound is crappy also with LMCE.

I've made some tests and I came to the conclusion that I don't think this issue is related to KDE itself (at least in my case, but likely also in yours).
What I made was to play the very same mp3 set (locally stored) in various situations: LMCE, KDE desktop (LMCE) and KDE from a plain Kubuntu 7.10 livecd.

It turned out that when playing from KDE - livecd the sound is cristal clear, while in the remaining situations the sound is clicking/bumping (don't know how to exactly describe in english).

Note that when you fire up a Kubuntu livecd you cannot play mp3 out of the box, but you are supposed to install some restricted packages. If I remember well you have to basically get ffmpeg and libxine1.
With stock Kubuntu you simply use adept and you're done, with LMCE you are not allowed to do so because all updates are driven via script to avoid breaking some LMCE dependecies.

In particular libxine1 is heavily customized, and I think that at least in my case this somehow fights with my MD hardware.

So KDE is not the nasty one ....

What I suggest is to try similar tests in your environment, and see what comes out.

In my case I've given up, as VIA MD is a bit old and plans are to replace it with some ATOM based hardware and demote VIA to simple MythTV client.
Of course I wouldn't mind to have it working fine also under LMCE, but for the time beeing I'm clueless.

HTH
Marco

Hi Bulek

I'm facing exactly the same issue. I'm writing a GSD driver for my lighting system, and I'm trying to figure out as well how could the blinds icons be better displayed on floorplans.

I'm currently using light switch (ON/OFF) for blinds, so there are only 2 states to show and I'm trying to play a bit to find out a reasonable compromise.
At this stage I'm trying to get a yellow icon when blind is not totally closed, and a black icon when blind is totally closed. Unfortunately my lighting system seems not to return any state about blinds position, so it's a bit hard ...

I think that showing opening percentage figures besides icons could be nice, but only for not-so-complex floorplans otherwise the overall look is likely to be a mess.
Unless a new scenario category is added specifically for blinds, we will still have lights and blinds shown on the same floorplan and I'm afraid that having a lot of info on that could be confusing.

I think that simply showing different icons for different opening percentage (maybe at 25%-30% steps) could be more effective and useful from user perspective.
Actually I don't care to know whether blinds of living room are 10% open or 20 %open .... I care to know that they are partially opened but not open wide (but of course this it's me ..).

I totally agree when you say that travelling state should be shown on floorplans. Maybe a blinking icon could be more effective, but a yellow one could be good as well.

The point now is: how to do it? Is there anyone already working on this, or anyone that could give some pointers to start digging from?

Regards
Marco

well ..I think I'm still missing some bits ...

I added a Generic Web Device (with its Generic Proxy Orbiter child )  following the wiki info, I formatted the orbiter skin to fit a web pad and everything went fine.
Web pad is showing the web orbiter in the desired shape and everything is OK.

Then I repeated the very same procedure to add a second Generic Web Device (with its child as well) and formatted the related orbiter skin to fit a mobile phone.

To my surprise both web orbiters are showing mobile phone skin.

In the web browser when going to http://dcerouter/pluto-admin/weborbiter.php I put the proper user/passwd, then it comes a page where I'm supposed to choose the proper device.

No matter which one I select, (web pad or mobile) I always get the mobile orbiter, that is the last one I added to my system.

Is this the expected behaviour, or am I doing something wrong?

TIA
Marco

Hi Thom and thanks for reply.

Just wanted to make sure it was possible before possibly messing up my system.

Thanks again and regards

Marco

 ... and probably a dumb one.

Is Proxy Orbiter intended to be unique in one LMCE system or is it possible to have multiple instances of it?

I'm trying to use as web orbiters a couple of devices not directly supported (i.e. no native orbiter for them) but they are quite different in size (one is an old web tablet, the other one is a mobile phone), so a combination of skin/size good for a device looks terrible on the other one.

One solution (besides using fully supported devices, of course) could be to istantiate 2 different Web orbiters, so I can set up each of them with the proper skin.

Can anyone tell me whether this is possible?

TIA
Marco

Thanks Dan

I feel a little dumb for asking a question that could answer myself by just looking webadmin ...

Nevertheless, back to business.

I tried to use cmd 68 (Get Device Data), but there's still something wrong in the way I'm using it because I can't get params_[5] that is supposed to hold the Name of the device.

My code is:

Code:

tmpCmd = Command.new(cmd.devidto_ , -1000 , 0, 1, 68)
tmpCmd.params_[2] = cmd.devidto_.to_s
tmpCmd.params_[52] = 156.to_s
tmpCmd.params_[53] = false.to_s
SendCommand(tmpCmd)

I'm running this code from Process Receive Command for Child, so cmd.devidto_ is the device # of a child of my GSD. Basically I'm sending DCE router a command in order to get child device name.
params_[52] contains the PK_DeviceData of "Name" parameter (deduced from DeviceData table).
What I'm expecting after SendCommand is to be able to access to params_[5] and find the child name in it, but things are not going like this.

DCE Router simply says:

Code:

Received Message from 67 (Luce Studio / Studio) to -1000 (unknown / ), type 1 id 68 Command:Get Device Data, retry none, parameters: <0x84fcbb90>
08 08/25/08 14:55:15.198   Parameter 2(PK_Device): 67 <0x84fcbb90>
08 08/25/08 14:55:15.198   Parameter 52(PK_DeviceData): 156 <0x84fcbb90>
08 08/25/08 14:55:15.198   Parameter 53(UseDefault): false <0x84fcbb90>

If I try to access to tmpCmd.params_[5] I get a runtime error, as the variable does not exist.

What the hell am I doing wrong?

TIA
Marco

Hi Dan and thanks for your reply

I got more or less your point, will try your suggestion very shortly and let you know.

By the way, you mention somehow a list of DCE Commands. Can you point me to some docs where I can find such info ? Wiki does not seem to have anything about DCE (at least anything with the needed detail), so some pointers to docs about DCE Commands would be really valuable for any developer (or anyone pretending to be one of them ... like me ...)

Thanks and regards
Marco

Hi Dan

sorry to bump an old thread .... but did you ever figure out how to access to devices names from GSD-Ruby?
And what about accessing a device Room?

TIA

Marco

It looks like they are intercepted by phone and not passed to keycode.jar

[...]

My fear is that if keycodes (that is a J2ME app) can't catch some keypress, so also JavaMO won't be able to catch them as well, and this may be due to some J2ME limitations.

This is no j2me limitation, it is the implementation on the nokia phones. SE ones pass nearly every key to java, the nokias don't. So you e.g. don't get keypresses for the red "hang up" key or the multimedia buttons.
The original mobile orbiter variant was done for the symbian app on the phone. With c++ you can catch every key on the nokias.

Glad to hear that I was wrong... and most of all that I'm not (yet) supposed to throw away my six month old E65 and my wife's brand new N82 ...

Therefore JavaMO would be pretty useless on that class of mobile phones.

We need a slightly optimized variation/skin for those. Many phones feature a touchscreen, too. Maybe we can get more keys with python on the Nokias. Posde is working on that.

br, Hari

I got the point.
Probably it's really time to seriously try to make HADesigner working on my laptop ...

Thanks a lot for your clear reply

Regards
Marco

this is mostly for sure a mapping problem with the keys. Run the keycodes application and post your keys. There is a translation method in the JavaMO main class. You could tweak that. The source compiles with standard make, you need to have a java sdk and the wireless toolkit from sun installed (tweak the paths in the makefile, wiki instructions available on the JavaMO page)

There is a bit of a backlog. The basic work for a nice JavaMO is already done. It needs better keypress handling (supporting different phones, repeated key handling). Pointer support for touchscreens is already implemented in the latest svn rev.

I'm pretty busy with Z-Wave at the moment, but sadly enough nobody feels like doing a bit of J2ME polishing...
The Bluetooth_Dongle code would also need some research. We could prerender datagrids with todays resolutions of phones.

best regards,
Hari

I downloaded keycode application and ran it on my E65, but it fails to show the keycode of those keys that are giving troubles ...
It looks like they are intercepted by phone and not passed to keycode.jar

I've googled around and it seems like there are not that much keycode - scancode applications, nor I found any keycode table for E65.
The only thing I've found is something called jbak tools, that is a native S60 applications that among others offers a keycode function. Unfortunately I could not have it working on my phone (needs to be signed, but the open signed site won't sign it ... did not figure out why ...).

So I'm pretty clueless.

My fear is that if keycodes (that is a J2ME app) can't catch some keypress, so also JavaMO won't be able to catch them as well, and this may be due to some J2ME limitations.
Therefore JavaMO would be pretty useless on that class of mobile phones.

Can my assumption be somehow correct?

TIA
Marco

Hi all
sorry to bump up this old thread, but I'm facing a similar issue (audio stopping for a fraction of second, then going on for some 30 sec, then disappearing again and so on ...) with one of my MDs although on different hardware and I was wandering whether a solution has been found.

TIA
Marco

Hi

i do NOT know Nokia tablets (n770, n800, n810), so I can only guess that they have the same VoIP client as others Nokia models.

If my assumption is correct you may have a look at this wiki page (http://wiki.linuxmce.org/index.php/Nokia_E65), where you can find step by step instructions about how to set up the VoIP client of a Nokia E65 to connect with Asterisk.

HTH
Marco

Hi Michael and thanks for your reply.

Actually I've tried to press any possible button on the phone (the central "Ok" button, all digit buttons, even the buttons on the right side of the phone) with no luck.

Anyway I think you gave me a good hint ... i.e to have a look at what lies in charonmedia.

As a starter I've noticed that here (http://svn.charonmedia.org/svn/home/hari/javamo/output/) there is a more recent JavaMO (version 0.2.0, instead of 0.1.1 pointed from wiki pages).

I've already replaced the old version with the new one on my mobile, but nothing has changed.
Next weekend I'll give a try on compiling the libBDCommon.so part and see what happens.

Thanks again

Obviously .. if you happen to recall how you got it to work, your note are always welcome 

Marco

I also have solved somehow ....

I reinstalled everything from scratch on the very same hardware, and now PXE boot is working fine.

So it seems not strictly related to hardware, but to some other strange thing that may have happened during the first install process.

My 0,02, of course.

HTH
Marco