Fix two WINDOW scaling problems (GitHub Issue #113):

[countingpine]

• coordinates should be multiplied by e.g. 'width - 1', not just 'width'.
• CINT() macro was not rounding positive values properly

Note: CINT() is also used in https://github.com/freebasic/fbc/blob/master/src/gfxlib2/gfx_circle.c#L96 and https://github.com/freebasic/fbc/blob/master/src/gfxlib2/gfx_draw.c#L247, so we need to make sure that they were not relying on the buggy rounding.

[countingpine]

I think this change yields a positive effect in the rounding of DRAW, shown in this code:

screen 13
draw "bm160, 90 TA5" & string(50, "U")
draw "bm160,110 TA5" & string(50, "D")
draw "bm150,100 TA5" & string(50, "L")
draw "bm170,100 TA5" & string(50, "R")
sleep

Where previously there was a small "hook" at the start of the up/right spokes, this now no longer happens. (This is because it now starts in the middle of the "step" instead of the start/end.)
Before:

After:

[countingpine]

Similar improvements can be seen when drawing arcs with Circle. Arcs are quite ugly anyway (no Bresenham), so I don't think the implementer would have cared much about the rounding method.


Note: in this case the patch only changes the rounding on positive x,y (the right/top sections of the circle).

I conclude that the patch is sound, and fixes less-than-ideal behaviour in other use cases, where the bug has not been accounted for.

[dkl]

Even though I don't know the Draw code, the patch seemed logical. Especially the change to CINT(), which changes the casting (if I understood that correctly...). It makes sense to cast to int after the + 0.5, instead of before, for rounding.