>
Not a good sign of numerical stability...
Yep, I have thought about scaling down all the values before
computations where overflow can happen. Moreover, since the final
data is normalized, It won't have any effect on the output.
This is what I did from line segments and it did work without any
overflow. So, I will do the same for bezier curves and see how it goes.
>> Perhaps you have to subdivide the Bezier curves, using the Bezier
>> envelope as a guide?
> I thought about it as well: subdivide Bezier curves until the segments
> are flat enough. The code can be borrowed from FreeType smooth
> renderer stopping short of calculating the coverages.
I haven't thought about subdividing the curves, but I can give it a try
if scaling down doesn't work.
>>
Hmm. The last commit in this tree deletes the code for conic and
>> cubic curves in `get_min_distance`...
That's because I did not port conic and cubic to FT_Fixed yet so I
decided to remove them for now.
Thanks,
Anuj