+++ This bug was initially created as a clone of Bug #958550 +++

Description of problem:
In virt-viewer, View -> Displays -> deselect the first display (Main display in Windows) of 4 displays and I have no usable functionality of the remaining 3 windows. I then select the Display 1 to return and it returns to a black screen. The ONLY way I was able to fix this, is if I resized Display 1 to anything different than it's original position.

Version-Release number of selected component (if applicable):
RHEV-M 3.2 (si14) and W2K8R2 client.
spice-client-msi-3.2-10 (mingw-virt-viewer 0.5.3-25)
vdagent-win-0.1-17 to Win7 guest

How reproducible:
100%

Steps to Reproduce:
1. See Above
2.
3.
  
Actual results:
The main window (Windows config) remains on the display that is closed. You need to select the display to come back to have any functionality.

Expected results:
The main window (Windows config) should be switched over to the next usable screen. If it the only screen remaining, it should act like a File -> Exit.

Additional info:

--- Additional comment from Marc-Andre Lureau on 2013-05-02 09:07:27 EDT ---

First, what is it that you call "no functionality"? The fact that the windows start menu and taskbar isn't visible? Or input is disabled? or display is frozen or black?

When I disable the first display, the other displays are still functionnal.

But there are 2 issues:

1) for legacy reasons, virt-viewer doesn't disable the monitor #0 (so the "main" guest display isn't updated to a visible/enabled one), this can trivially be changed in virt-viewer by removing the if (nth != 0) condition

2) agents don't support disabling first monitors very well.
 - Windows agent doesn't support sparse monitors: bug 953973
 - Linux agent raises X error when disabling & enabling first monitor

--- Additional comment from Marc-Andre Lureau on 2013-05-02 09:10:02 EDT ---

Since there is a workaround and this is not a main feature, and also because the extra display are still usable, I suggest the severity be lowered and the bug moved to 3.3

--- Additional comment from Marian Krcmarik on 2013-05-02 09:38:39 EDT ---

(In reply to comment #1)
> First, what is it that you call "no functionality"? The fact that the
> windows start menu and taskbar isn't visible? Or input is disabled? or
> display is frozen or black?
In my case, simply I lose primary monitor so I cannot see taskbar and the start menu indeed, as well as The primary monitor is not disabled so I lose all the windows placed on the monitor or anything which opens on primary monitor.
> 
> When I disable the first display, the other displays are still functionnal.
> 
> But there are 2 issues:
> 
> 1) for legacy reasons, virt-viewer doesn't disable the monitor #0 (so the
> "main" guest display isn't updated to a visible/enabled one), this can
> trivially be changed in virt-viewer by removing the if (nth != 0) condition
> 
> 2) agents don't support disabling first monitors very well.
>  - Windows agent doesn't support sparse monitors: bug 953973
>  - Linux agent raises X error when disabling & enabling first monitor

--- Additional comment from Bill Sanford on 2013-05-02 10:16:29 EDT ---

I get the same as Marian. That is why I said in expected behavior: "The main window (Windows config) should be switched over to the next usable screen. If it the only screen remaining, it should act like a File -> Exit."

--- Additional comment from Bill Sanford on 2013-05-02 10:38:25 EDT ---

With Windows, you cannot "Disconnect this display" (The Main display) unless you make another display the main display. The fact that virt-viewer closes the first display without making another active display the "Main" display, is the issue here.

--- Additional comment from Marc-Andre Lureau on 2013-05-02 11:34:20 EDT ---

(In reply to comment #5)
> With Windows, you cannot "Disconnect this display" (The Main display) unless
> you make another display the main display. The fact that virt-viewer closes
> the first display without making another active display the "Main" display,
> is the issue here.

it's not the role of virt-viewer or spice to make a display "main", the agent/OS already guarantees that.

--- Additional comment from Bill Sanford on 2013-05-02 12:15:37 EDT ---

You're right. However, it is virt-viewer's role to keep multi-display functional. Before you deselect the display (That is the main display in the OS), maybe v-v can check and error if you are trying to close the "Main display" of the OS?

--- Additional comment from Marc-Andre Lureau on 2013-05-02 12:30:07 EDT ---

(In reply to comment #7)
> You're right. However, it is virt-viewer's role to keep multi-display
> functional. Before you deselect the display (That is the main display in the
> OS), maybe v-v can check and error if you are trying to close the "Main
> display" of the OS?

It doesn't have this knowledge, not do I think it should have.

This bug is all about the display 0 not being disabled by virt-viewer, imho. The fact that the display is main or not is irrelevant here, as it is more a guest concept than a hw config one.

--- Additional comment from Bill Sanford on 2013-05-02 13:01:23 EDT ---

At the minimum, we need to get this to the Doc folks. I have set the flag for release note.

--- Additional comment from Marc-Andre Lureau on 2013-05-09 07:05:02 EDT ---

patch sent to ML for case 1.
https://www.redhat.com/archives/virt-tools-list/2013-May/msg00017.html

Adding sparse monitor support bug dependency

--- Additional comment from Marc-Andre Lureau on 2013-05-09 07:24:41 EDT ---

Adding 1st monitor crash linux vdagent bug dependency