Looks like you found a new bug in the graph control. The bug happens when you change the opacity of a bitmap plot without having assigned it a bitmap.
The usual scenario is for a bitmap to be assigned to the plot at the moment that the plot is created, by passing the bitmap file to the PlotBitmap function. But it's also possible to create the plot without a bitmap, and to set it later from a handle, which is what you are doing, quite correctly. The bug happens when you change the opacity of this bitmap plot without having assigned it the bitmap handle, via the SetCtrlBitmap function.
I've created a bug report for this (534492). Until it's fixed, you should avoid this situation by setting the opacity in tandem with the bitmap handle. Something like this:
GetCtrlVal (panelHandle, PANEL_TEST_ERROR, &ErrorOn);
if (ErrorOn==0)
{
SetCtrlBitmap (panelHandle, PANEL_GRAPH, ImgPlotHandle, PG_RwyImgID);
SetPlotAttribute (panelHandle, PANEL_GRAPH, ImgPlotHandle, ATTR_PLOT_OPACITY, 125);
}
As for the flashing problem, that's because of a limitation of the ATTR_REFRESH_GRAPH attribute. This attribute only disables updates for the plotting functions. It does not disable updates for the SetCtrlBitmap function. That's where the flashing is coming from. The only way I can think of to work around this limitation is to always plot from file, and avoid using SetCtrlBitmap altogether. Something like this:
DeleteGraphPlot (panelHandle, PANEL_GRAPH, -1, VAL_DELAYED_DRAW);
SetCtrlAttribute (panelHandle, PANEL_GRAPH, ATTR_REFRESH_GRAPH, 0);
GetCtrlVal (panelHandle, PANEL_TEST_ERROR, &ErrorOn);
if (ErrorOn==0)
{
ImgPlotHandle = PlotBitmap (panelHandle, PANEL_GRAPH,
0.0, 0.0,
0.0, 0.0, DirAndFile);
SetPlotAttribute (panelHandle, PANEL_GRAPH, ImgPlotHandle, ATTR_PLOT_OPACITY, 125);
}
GetCtrlVal (panelHandle, PANEL_RECT_LENGTH, &RectLength);
PlotRectangle (panelHandle, PANEL_GRAPH, 10.0, 20.0, RectLength-10,
40.0, VAL_RED, VAL_RED);
SetCtrlAttribute (panelHandle, PANEL_GRAPH, ATTR_REFRESH_GRAPH, 1);
Note a couple of other changes I made: I got rid of the RefreshGraph call because it's not necessary. Instead, I simply turn ATTR_REFRESH_GRAPH back on when I'm done plotting (turning it on results in an implicit refresh). It's more intuitive to use the attribute this way, so that you don't have to keep turning it off without ever turning it back on. Also, I moved it below the DeleteGraphPlot, since this attribute does not affect what happens during deletion, which is determined solely by the last parameter you pass to the delete function.
