I did give it an extra look, but stepping through the code, I didn't see a paint or invalidate.

I know this:
When I comment out the line
gdk_draw_pixmap(PDeviceContext(DestDC)^.Drawable,
PDeviceContext(DestDC)^.GC, PDeviceContext(SrcDC)^.Drawable, XSrc, YSrc, X, Y, Width, Height);
in TgtkObject.StretchBlt, the repainting doesn't happen. Of course, the glyph on the speedbutton doesn't get painted either.

I give up. :-( This is for the real developers, not for lurkers.

Vincent.

Hi,

I have been looking at the problem with the speedbuttons in win32. I haven't found a solution but I know more or less what is hapening, not the reason why.

The problem is in painting the glyph on the button face. This paint action (a copyrect on a canvas) causes gtk to emit (is that the right word) an expose-event on the form and the form get repainted, which includes repainting the speedbutton etc, etc.

Commenting out
TButtonGlyph(FGlyph).Draw(Canvas, PaintRect, Offset, Caption, FLayout, FMargin, FSpacing, FState, Transparent, (0));
in the Procedure TSpeedbutton.Paint stops the repainting.

The reason why I can't solve this problem is a huge lack of knowledge on graphics (DC etc) and gtk's interface to the graphics. You might take a look at the stretchBlt function in gtkwinapi.inc. The problem might be solved there.

I hope this helps somebody.

Regards,
Vincent Snijders.

Hmmm.... In order to get a paint message on the first exposure of a form I generated a LM_Paint message on an expose-event. At least on Linux you don't get a paint event when a form is shown for the first time (you will get an expose).

If a draw results in an expose on the win32 platform I think it might be a bug in the libs, or I don't understand what an expose event means and it doesnt work on a Linux platform

Thanks, it gives a point to look at.

Marc

+ From: Shane Miller
+
+ It could be that after I did the paint I called repaint or
+ invalidate or something. Maybe that causes it.

Hmm...

+ Are you going to look further into it? If not, I guess I can
+ look at it to see if that's the case. Problem is I do my
+ development under LINUX, not windows.

If it is a problem I can have a look, some small win32 development is possible.

Marc

It could be that after I did the paint I called repaint or invalidate or something. Maybe that causes it.

Are you going to look further into it? If not, I guess I can look at it to see if that's the case. Problem is I do my development under LINUX, not windows. Shane

>> 05/02/00 04:53PM >>>

Hi,

I have been looking at the problem with the speedbuttons in win32. I haven't found a solution but I know more or less what is hapening, not the reason why.

The problem is in painting the glyph on the button face. This paint action (a copyrect on a canvas) causes gtk to emit (is that the right word) an expose-event on the form and the form get repainted, which includes repainting the speedbutton etc, etc.

Commenting out
TButtonGlyph(FGlyph).Draw(Canvas, PaintRect, Offset, Caption, FLayout, FMargin, FSpacing, FState, Transparent, (0));
in the Procedure TSpeedbutton.Paint stops the repainting.

The reason why I can't solve this problem is a huge lack of knowledge on graphics (DC etc) and gtk's interface to the graphics. You might take a look at the stretchBlt function in gtkwinapi.inc. The problem might be solved there.

I hope this helps somebody.

Regards,
Vincent Snijders.

_________________________________________________________________ To unsubscribe: mail with
"unsubscribe" as the Subject
archives at http://www.miraclec.com/list_archives/lazarus