Re: [ft-devel] Fast bisecting BBox_Cubic_Check

[Werner LEMBERG]

> This is as close as I could get to "actual" bboxes.  The attached
> files delta0 and delta1 are for the original and new algorithms
> respectively.  The first bbox is the actual one, followed by the
> computed one, followed by the differences.  Admittedly, the original
> code does a better job, but it is not ideal.

:-) So my concerns were not completely unfounded...

> The new code is a bit worse but faster. The problem is in the
> rounding.  I can potentially scale up the outlines temporarily to
> deal with this problem.

I really wonder whether it makes sense to speed up the exact bbox
computation if this causes a loss of precision.  For a really quick
guess the user can always compute the cbox, which is in most cases the
right thing because both the TrueType and PostScript specs says that
there should be points at all extremas.


    Werner