Hi Mario,
On 4/11/2012 3:39 PM, Mario Torre wrote:
I also agree with Omair.
To be honest, I'm not that happy with the fix, but not because of
Omair's patch, which is very fine given the current state of things,
but because we keep adding this to each and every new WM we find,
although fixing AWT in this regard will be probably more work than
adding another exception to the list and an issue for perennial
backward compatibility...
That would be fine, indeed. However, this would require a lot of testing
with all the WMs AWT knows about to make sure no regressions are introduced.
Btw, this patch brings in another small issue:
static boolean isNonReparentingWM() {
return (XWM.getWMID() == XWM.COMPIZ_WM || XWM.getWMID() ==
XWM.LG3D_WM || XWM.getWMID() == XWM.CWM_WM);
}
I don't remember at this moment why this was needed at all, but if my
memory is correct, Compiz is a reparenting WM now. I'm not sure if
mutter is or not.
I believe there would be a bug filed already about this, since in
XDecoratedPeer.handleConfigureNotify() this would cause the newLocation
to be assigned with the relative position of the frame, not absolute
position. As long as there's no bug filed, I assume it's OK to report
Compiz as a non-reparenting WM for AWT.
--
best regards,
Anthony
Anyway, I just wanted to give Omair my +1 to this patch.
Cheers,
Mario
2012/4/11 Anthony Petrov <anthony.pet...@oracle.com>:
Hi Omair,
The analysis below sounds reasonable to me, and as I've already mentioned
I'm OK with your fix.
Let's hear what Artem says though.
PS. Perhaps it also makes sense to rewrite that comment in the
XDecoratedPeer to replace the word "bug" with something saying that this is
implemented according to the ICCCM specification with a reference to the
paragraph 4.1.5 of it?
--
best regards,
Anthony
On 4/11/2012 8:50 AM, Omair Majid wrote:
Hi Artem,
On 04/09/2012 07:10 AM, Artem Ananiev wrote:
I really hope we can drop most of the ancient WMs listed in the XWM
class (MOTIF, OPENLOOK, CDE, SAWFISH, etc) in JDK8. We know AWT/Swing
works fine on the modern WMs that conform to ICCCM and NET standards,
and I don't see any reasons to have (and add more!) workarounds for
non-conformant window managers.
I spent a little bit of time reading up on ICCCM (and X11), and here's a
summary of my findings. I am not familiar with this, so it could be that
I am making a mistake. Please correct me if I am wrong.
The problematic case (that the reproducer shows) is that when we
maximize a window by double clicking on the title bar under mutter, java
does not detect that the window has moved/changed size.
ICCCM 4.1.5 [1] states:
"""
If the window manager decides to respond to a ConfigureRequest request by:
... snip ...
- Resizing the window or changing its border width (regardless of
whether the window was also moved or restacked).
A client that has selected for StructureNotify events will receive a
real ConfigureNotify event. Note that the coordinates in this event are
relative to the parent, which may not be the root if the window has been
reparented. The coordinates will reflect the actual border width of the
window (which the window manager may have changed). The
TranslateCoordinates request can be used to convert the coordinates if
required.
The general rule is that coordinates in real ConfigureNotify events are
in the parent's space; in synthetic events, they are in the root space.
Advice to Implementors
Clients cannot distinguish between the case where a top-level window
is resized and moved from the case where the window is resized but not
moved, since a real ConfigureNotify event will be received in both
cases. Clients that are concerned with keeping track of the absolute
position of a top-level window should keep a piece of state indicating
whether they are certain of its position. Upon receipt of a real
ConfigureNotify event on the top-level window, the client should note
that the position is unknown. Upon receipt of a synthetic
ConfigureNotify event, the client should note the position as known,
using the position in this event. If the client receives a KeyPress ,
KeyRelease , ButtonPress , ButtonRelease , MotionNotify , EnterNotify ,
or LeaveNotify event on the window (or on any descendant), the client
can deduce the top-level window's position from the difference between
the (event-x, event-y) and (root-x, root-y) coordinates in these events.
Only when the position is unknown does the client need to use the
TranslateCoordinates request to find the position of a top-level window.
"""
To me, this says that when a window has been resized (by, say, double
clicking on the title bar), a real ConfigureNotify event will be sent.
The implementation (awt, in this case) should query for the coordinates
relative to the root and use them.
This is pretty much exactly what the "CDE/MWM/Metacity/Sawfish bug"
currently does. It seems like this should be the correct default
behaviour (for all window managers, including mutter).
What do you think?
Thanks,
Omair
[1] http://tronche.com/gui/x/icccm/sec-4.html#s-4.1.5
