[Bug 10522] endless WM_PAINT loop bug (update regions, queue paint count)

[wine-bugs at winehq.org]

http://bugs.winehq.org/show_bug.cgi?id=10522





--- Comment #15 from Austin Lund <austin.lund at gmail.com>  2009-07-09 00:59:10 ---
There seem to be two confounding effects here. 

a) When DispatchMessage() is called on a WM_PAINT message in wine, the WNDPROC
is called and then _after_ it has done what it does, if parts of the window are
still marked as invalid they are updated using GetUpdateRgn(..,..,TRUE).  This
doesn't seem to happen in windows.  The WM_NCPAINT and WM_ERASEBACKGROUND
messages seem to be processed in a WM_PAINT message _before_ the WNDPROC is
called.

b) When BeginPaint() is called and a WM_NCPAINT message is processed within it,
the window is marked as invalid in wine when BeginPaint() returns.  In windows
a BeginPaint() seems to mark everything as clean even if WM_NCPAINT called
RedrawWindow(..,..,..,0x401).

-- 
Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email
Do not reply to this email, post in Bugzilla using the
above URL to reply.
------- You are receiving this mail because: -------
You are watching all bug changes.