Code review comment for lp:~vanvugt/compiz/fix-751605

> + * (1) It's confusing to the user to auto-magically move a window
> + * between monitors when they didn't ask for it. So don't.
I agree 100%.

> + * (2) In the worst case where the window can't go small enough to fit
> + * the output, they can simply move it with Alt+drag, Alt+F7 or
> + * expo.
>
True also.

> Reading the comments of bug 751605 it seems like the /real/ gripe is with the
> fact that the grid plugin displays the window maximization preview on the
> first monitor and then the maximization happens on the second. I think, just
> brainstorming a little that there are some better solutions to this problem:
>
This is a big problem as the action taken should always follow the preview displayed.
Currently this works correctly, btw if the windows' x- or y-size is larger than the largest resolution of your monitors it is resized on the correct one via grid.
To test this simply open a window, resize it to a huge x or y resolution (larger than the respective highest resolution of your monitors) and then activate the grid on the smaller monitor - it will recale and go fullscreen correctly.

> a) Don't allow maximization of windows that might be bigger than one of the
> available monitors

Bad solution IMHO. Compiz should do what the user orders him to do (via grid for example).

> b) Allow maximization of windows when they are already on the monitor that
> they can be maximized on (though this could be a little tricky to implement -
> and could look a bit weird if a window suddenly gains a maximize button as a
> result of moving it between the monitors)

I cannot notice a problem with making a window smaller when maximizing it. We should make sure though that the size gets correctly restored afterwards (which is not happening in all cases, but that is another story...)

> c) Display the grid preview in the place where the window is about to go.
>

This should always be the case, but it would feel strange if I want to activate the grid on the smaller monitor and the preview shows me it would go fullscreen on the larger one...

> Is it possible that we can consider those alternatives? Trust me - if you've
> ever seen what happens when you maximize a window to a smaller size then its
> minimize size the result is really really bad.

You can easily test what happens then by creating a window larger than your largest monitor's x and y resolution and then maximize it on your smaller monitor (via grid for example). I cannot notice anything bad...