Local Pi Display

Dave

Fair enough.

I use WinSCP on a windows machine and luckily it remembers for me.

mark

Dave, you can make an entry in the file /etc/hosts. The system uses this as a look-up to translate system-names to addresses (you still need to enter the system name).

     How about creating a command-file on the desktop that  contains the connect command ? Just a click to run that.  

     If you have a forward-facing camera, you could create a program that watches for left-winks, then runs the connect-script  (not too sure about the details of that......)

This thread might be helpful:

http://www.raspberrypi.org/phpBB3/viewtopic.php?f=28&t=49311

coder27 wrote:

 

This thread might be helpful:

http://www.raspberrypi.org/phpBB3/viewtopic.php?f=28&t=49311

Thanks for the link.  I've had my problems both when I use the rasp-config to initiate the display manager (start in GUI) and when I've remotely sent "startx" via RDP.  I have tried multiple variations of export DISPLAY=:0.0

One thing that confuses me is that when I run the command "w" to find out what other users are on the raspberry, I usually only get one:

      21:29:10 up 13 min,  1 user,  load average: 0.02, 0.12, 0.14
      USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU WHAT
      pi       pts/3    :10:S.0          21:21    0.00s  0.62s  0.05s w

so I'm not sure what DISPLAY I want to export to.  Again, I've tried [almost] all combinations of 0, 1, 10 with and without S.

At times I do see two users when I run "w" but right now I can't duplicate that.

I think learning how to use DISPLAY is the optimal solution.  In the meantime, I've found that I can link the Pi and my Mint desktop using the "screen" program.  This convoluted approach allows me to set up a "screen" on the Pi, RDP to the Pi from the Mint to control the Pi, connect to the Pi "screen" through ssh&screen and have the Atrix display the Pi through the desktop. Maybe Mr. Smith is right and using a Pi shouldn't require thinking but I'm sure my explanations do.

At least I think this approach will work.  I have an A model Pi with only a single USB connector so I can't have the keyboard and wireless at the same time.  And when I try to switch from keyboard to wireless, the Pi reboots.

I've got to get my USB hub back from my nephew.  When I got the Atrix, I thought I didn't need it anymore and he said he did.

Dave Guthridge wrote:

 

coder27 wrote:

 

This thread might be helpful:

http://www.raspberrypi.org/phpBB3/viewtopic.php?f=28&t=49311

Thanks for the link.  I've had my problems both when I use the rasp-config to initiate the display manager (start in GUI) and when I've remotely sent "startx" via RDP.  I have tried multiple variations of export DISPLAY=:0.0

One thing that confuses me is that when I run the command "w" to find out what other users are on the raspberry, I usually only get one:

      21:29:10 up 13 min,  1 user,  load average: 0.02, 0.12, 0.14
      USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU WHAT
      pi       pts/3    :10:S.0          21:21    0.00s  0.62s  0.05s w

 

so I'm not sure what DISPLAY I want to export to.  Again, I've tried [almost] all combinations of 0, 1, 10 with and without S.

I suggest trying to set DISPLAY to "127.0.0.1:0.0".  127.0.0.1 is the usual "localhost" which lets you to IP to yourself (or something like that).  You can verify this value of "localhost" with "cat /etc/hosts".

If that doesn't work, bring up X windows using the RasPi keyboard and whatever HDMI display you have (I guess the Lapdock).  Then in a terminal window enter "echo $DISPLAY".  This displays the value of DISPLAY when running from a local keyboard, for example ":0" on my RasPi.  Try setting DISPLAY to that value for remote access.

I read the thread only partially, but do you wish direct graphical apps onto the local display (i.e. the HDMI connected monitor) that are invoked remotely (via SSH)?

Apologies if I've misunderstood.

If that's what you want to do, I believe this link may help (I did it a few months back - not on an RPI, and I think this was the link that helped - I could be wrong!).

xhost, xauth and of course the export DISPLAY all ring a bell.

When I did it, I just wanted to bring up an xclock or something on the local display, but invoked from a remote connection which was either SSH or serial from memory - not local USB keyboard.

Apologies if it is the wrong link - it was still in my browser cache, I should have bookmarked it at the time :-( I just don't use GUI programs often enough in Linux to be 100% sure - I'm happy with SSH into Linux servers.

EDIT: Actually, it may be this link that helped me, this was in my browser history too. I remember reading the xroach bit in the text, so I think this is the page.

I've got several workstations that are "headless" in that they don't have a keyboard or mouse attached. But I AM using the display.

What, at least by default, you don't want is that "random" guys on the network can tap into your session if you log in locally. So several steps are taken to prevent that from happening if you log into the local display/keyboard/mouse.

First authentication is used: The login manager pulls a big random number out of a hat and tells the server: "Only allow programs that know THIS number to connect". When a succesful username/password is provided that user is logged in and the random number is copied over to HIS homedir, allowing programs under his uid to grab the number from the file and show the server that they know that number. I assume the number is changed once you log out.

Secondly, some (most?) people seldomly use "remote display". A program running on one computer displaying on another. A very powerful and useful feature. But again: A program that can connect to the x-server would have too many powers. So it was decided that an X-server will start without the networking features enabled. (That's why I don't expect the "EXPORT DISPLAY=localhost:0.0" to work: it specifies to connect to the networking part of the protocol even though it stays local on the machine!).

To prevent all that from interfering with what I want to do, I simply start the X server:
   X -ac :2

The -ac means: Don't check authentications. And the :2 means: use the <hostname:2> as the display. This allows me to start it next to a running gui login program. On the other hand, that won't be necessary if you had to use "init 3" to get to the gui login. I then get to manually start everything I want. Personally I start:

- xsetbg -solid steelblue (I've used that background since about 1990).

- sawfish (A window manager that I've gotten used to).

- gnome-panel (a panel that I've gotten used to).

Of course I'm overriding several sensible security practises. But I trust and have-to-trust my local network. I have a firewall so that it is not connected to the internet etc. So I'm happy with this situation.

I appreciate the time peple have spent on my problem.  Obviously I have a lot to learn about Linux and X Windows.  Based on the responses, I think I need to pay more attention to the inherent security measures.  Since I am firewall protected and the only user on my network, I've ignored security.  Based on Roger's reply, I'll look into the necessity of actively defeating them.  I'll spend some time on his useful suggestions.

Last night, I was able to remotely generate a window on the Pi's local LCD display using xterm.  Unfortunately, I was not able to write to it.  I'll keep banging away at it.

It looks like I've done a poor job of explaining my desire so let me take one more shot at it.  I have a A model Pi with a USB wireless adapter.  The Pi has a local LCD display but no keyboard/mouse.  I would like to connect to the Pi remotely from a desktop, control the Pi and have the applications on the Pi be displayed in both the desktop's remote window and on the Pi's local LCD.  As I described in an earlier post, I think I have a method to do this but it requires temporary attachment of a keyboard to the Pi to initialize it. 

You should look at "xrdp".

It will allow you to use a "remote mouse and keyboard" with your 'pi.

My raspberry pi has "xrdp" in the repository. (so apt-get install xrdp will work!).

I don't know right now if you'll have to run xrdp "from scratch" i.e. without starting the "X -ac" first, or if you'll have to run that first.

I've played around with this and have gotten a solution that works for me.  Elegant it ain't but it's pretty straightforward, at least it is if you're using a Debian based desktop.  That's the only configuration I've tested.  I'm using an Atrix but I think this will work for any display local to the Pi.

"Mint" is my Mint 14 Desktop.  "Pi" is my 2012 model A Raspberry Pi.  I'm not sure of all the software requirements because I've installed so many in trying to solve this.  I think as a minimum, you'll need the Linux "screen" program on both computers, synergy on the Pi, QuickSynergy on the Mint and Remmina on the Mint.  I haven't tried to connect using a Windows desktop.

Note 1:  In my messing around, I have screwed up the Pi's ability to boot into XWindows.  So these instructions start from the Pi's text login screen.

NOTE 2: IN MY EXPLANATION THE TEXT ":0" (colon zero with NO SPACE) HAS BEEN CHANGED TO AN EMOTICON CONTRARY TO MY WISHES.  In step 3, replace the happy face with the colon symbol followed by the number zero.  The Happy Face isn't really appropriate until you successfully complete step 8.

1.  From Mint, use Remmina to connect to Pi.  I used RDP protocol.
2.  Open an LXTerminal and type:
    startx
to start an X WIndows session on the Pi.
3.  Open a new LXTerminal tab or window and type:
    xterm -display -e "screen -S mPi" &
to open the 'screen' program in an xterminal on the Pi.  You should see this on the Pi's display.  "mPi" is just a name I gave my session.  You can pick your own.  [In all my tests, the local Pi display has been and the remote has been :10. ]  The "&" at the end just lets the xterm run in the background.
4.  To get a list of available screen sockets type:
    screen -ls
The name of your session (for me "mPi") should be there.  [Unless you've started other screen sessions, it will be the only one]
5. Connect to that screen using:
    screen -x mPi
Now when you type of your desktop, you should see text in the Pi.  As a test, enter:
    ls
and you should see a directory listing on both screens.
6.  Now to get the mouse to appear on the Pi using Synergy. First, get your desktop's IP.  Mine is 192.168.1.134.  To start Synergy on the Pi:
    synergyc -- daemon 192.168.1.134
This tells the Pi that the mouse is coming from Mint.
7.  Get an identifier for your Pi.  Mine uses Avahi to announce itself as raspberrypi.  If you don't have a name for your Pi, get its IP.
Start the QuickSynergy GUI on Mint.  In the directional box you want to use, type the Pi identifier.  I type raspberrypi into the left box.  Hit the EXECUTE button.
8.  If QuickSynergy is working properly, when you push your mouse beyond the end of the Mint display in the QuickSynergy direction (left for me) the mouse appears on the Pi display. 

When I use the GPIO using Python, now the prints to the standard output appear on both displays.  That was my original intent.  The use of the mouse and keyboard has additional benefits.