Touched up several comments, pointing up things that may be worth reflect-
ing upon in the future. git-svn-id: https://svn.code.sf.net/p/freeglut/code/trunk@379 7f0cb862-5218-0410-a997-914c9d46530a
This commit is contained in:
parent
03cf899635
commit
2e4d5de7c0
@ -126,6 +126,10 @@ static void fghReshapeWindowByHandle ( SFG_WindowHandleType handle,
|
||||
);
|
||||
}
|
||||
|
||||
/*
|
||||
* XXX Should update {window->State.OldWidth, window->State.OldHeight}
|
||||
* XXX to keep in lockstep with UNIX_X11 code.
|
||||
*/
|
||||
if( FETCH_WCB( *window, Reshape ) )
|
||||
INVOKE_WCB( *window, Reshape, ( width, height ) );
|
||||
else
|
||||
@ -561,6 +565,13 @@ void FGAPIENTRY glutMainLoopEvent( void )
|
||||
*
|
||||
* GLUT presumably does this because it generally tries to treat
|
||||
* sub-windows the same as windows.
|
||||
*
|
||||
* XXX Technically, GETWINDOW( xconfigure ) and
|
||||
* XXX {event.xconfigure} may not be legit ways to get at
|
||||
* XXX data for CreateNotify events. In practice, the data
|
||||
* XXX is in a union which is laid out much the same either
|
||||
* XXX way. But if you want to split hairs, this isn't legit,
|
||||
* XXX and we should instead duplicate some code.
|
||||
*/
|
||||
case CreateNotify:
|
||||
case ConfigureNotify:
|
||||
@ -597,10 +608,15 @@ void FGAPIENTRY glutMainLoopEvent( void )
|
||||
case Expose:
|
||||
/*
|
||||
* We are too dumb to process partial exposes...
|
||||
*
|
||||
* XXX Well, we could do it. However, it seems to only
|
||||
* XXX be potentially useful for single-buffered (since
|
||||
* XXX double-buffered does not respect viewport when we
|
||||
* XXX do a buffer-swap).
|
||||
*
|
||||
* XXX GETWINDOW( xexpose );
|
||||
* XXX fgSetWindow( window );
|
||||
* XXX glutPostRedisplay( );
|
||||
*/
|
||||
if( event.xexpose.count == 0 )
|
||||
fghRedrawWindowByHandle( event.xexpose.window );
|
||||
@ -625,6 +641,9 @@ void FGAPIENTRY glutMainLoopEvent( void )
|
||||
case VisibilityNotify:
|
||||
{
|
||||
GETWINDOW( xvisibility );
|
||||
/*
|
||||
* XXX INVOKE_WCB() does this check for us.
|
||||
*/
|
||||
if( ! FETCH_WCB( *window, WindowStatus ) )
|
||||
break;
|
||||
fgSetWindow( window );
|
||||
@ -692,7 +711,9 @@ void FGAPIENTRY glutMainLoopEvent( void )
|
||||
|
||||
/*
|
||||
* XXX For more than 5 buttons, just check {event.xmotion.state},
|
||||
* XXX rather than a host of bit-masks?
|
||||
* XXX rather than a host of bit-masks? Or maybe we need to
|
||||
* XXX track ButtonPress/ButtonRelease events in our own
|
||||
* XXX bit-mask?
|
||||
*/
|
||||
#define BUTTON_MASK \
|
||||
( Button1Mask | Button2Mask | Button3Mask | Button4Mask | Button5Mask )
|
||||
@ -789,12 +810,19 @@ void FGAPIENTRY glutMainLoopEvent( void )
|
||||
else if( pressed )
|
||||
/*
|
||||
* Outside the menu, deactivate if it's a downclick
|
||||
*
|
||||
* XXX This isn't enough. A downclick outside of
|
||||
* XXX the interior of our freeglut windows should also
|
||||
* XXX deactivate the menu. This is more complicated.
|
||||
*/
|
||||
fgDeactivateMenu( window->ActiveMenu->ParentWindow );
|
||||
|
||||
/*
|
||||
* XXX Why does an active menu require a redisplay at
|
||||
* XXX this point? If this can come out cleanly, then
|
||||
* XXX it probably should do so; if not, a comment should
|
||||
* XXX explain it.
|
||||
*/
|
||||
window->State.Redisplay = GL_TRUE;
|
||||
break;
|
||||
}
|
||||
@ -807,6 +835,9 @@ void FGAPIENTRY glutMainLoopEvent( void )
|
||||
( window->Menu[ button ] ) &&
|
||||
pressed )
|
||||
{
|
||||
/*
|
||||
* XXX Posting a requisite Redisplay seems bogus.
|
||||
*/
|
||||
window->State.Redisplay = GL_TRUE;
|
||||
fgSetWindow( window );
|
||||
fgActivateMenu( window, button );
|
||||
@ -826,7 +857,7 @@ void FGAPIENTRY glutMainLoopEvent( void )
|
||||
/*
|
||||
* Finally execute the mouse or mouse wheel callback
|
||||
*
|
||||
* XXX Use a symbolic constant, *not* "4"!
|
||||
* XXX Use a symbolic constant, *not* "4"! ("3, sire!")
|
||||
*/
|
||||
if( ( button < 3 ) || ( ! FETCH_WCB( *window, MouseWheel ) ) )
|
||||
INVOKE_WCB( *window, Mouse, ( button,
|
||||
|
Reference in New Issue
Block a user