Lot of bugfix and enhancement

[Genva]

I submitted a patch to SF here, please give it a try.
The patch do:
1. Support 8 direction image render, (This feature
support in OpenGL and Irrlicht renderer is out-of-date.
Hit: support in OpenGL is easy, support in Irrlicht is
a little hard)
2. Better negative image offset support in frame window
3. Removed unnecessary destructor for simple class
(ColourRect, Rect, colour, Vector2, Vector3, etc)
4. Added/Moved some useful function of 'System' for
working with other application seamless.
5. System::getTargetWindow: Doesn't pass the mouse
message to child when exist a capture window
(Otherwise, if drag mouse fast when resizing/moving the
window, the mouse message will dispatch to another
window, this behavior isn't what we need).
6. Optimize overall colour rect fill, compute
sub-rectangle only if need.

[CrazyEddie]

Thanks for the patch.

I have very briefly looked over the modifications, and in general everything looks reasonable, though will obviously require more testing due to the number and range of the changes; I'll try to get this sorted out next week.

The ability to specify the orientation of images is a very nice addition

Where new parameters have been inserted mid-signature, it's likely that I'll move these to the end of the list and default them; I won't make such a breaking change without a very good reason (which generally means either a major bug fix or a "CrazyEddie executive decision" )

Could you elaborate a little on the expected use of the System::injectMouseLeaves method?

Thanks again,

CE.

[Genva]

Could you elaborate a little on the expected use of the System::injectMouseLeaves method?

I write an in-game editor using the OGRE/CEGUI, and I want it can work with other Windows application. I setup mouse tracking use TrackMouseEvent() and respond WM_MOUSELEAVE message, When I receive the WM_MOUSELEAVE message, call to System::injectMouseLeaves() make sure hovering window (a button for examples) known they should be change to normal state (This is require when hover text colour difference with normal text colour, or other appearance that based on hover/normal state).

[edit]The orientation of image support I write about two month ago, but I submit the patch the day before yesterday, so the 'orientation' parameter is place before the 'quad_split_mode' [/edit]

[Genva]

Another Bugfix for Window::moveToBack posted to SF, fixed the bug in Window::moveToBack cause the window insert itself as its' children.

[CrazyEddie]

Genva wrote:
Another Bugfix for Window::moveToBack posted to SF, fixed the bug in Window::moveToBack cause the window insert itself as its' children.

Thanks for this one This fix will be in CVS later today. How this has gone undetected for almost 12 months I'll never know