Re: [ft-devel] a satisfactory fix for the cubic spline bug

[GRAHAM ASHER]

Some reasons this hasn't made very many waves;

1. My impression is that nearly all actual fonts are TrueType these days. 
TrueType does't use cubic splines. It uses quadratics only.

2. Font designers very rarely use 'exciting' cubic splines. Splines with 
control 
points on both sides of the line hardly ever occur in glyphs, and I would 
expect 
self-crossing curves never to occur.

We must of course be correct for all curves, but we need optimise heavily only 
for the common case where p0, p1, p2 and p3 have r-coordinates (using Hain's 
notation) in that order, and s-coordinates with the same sign.

Graham




----- Original Message ----
From: Werner LEMBERG <address@hidden>
To: address@hidden
Cc: address@hidden; address@hidden
Sent: Tuesday, 31 August, 2010 14:29:49
Subject: Re: [ft-devel] a satisfactory fix for the cubic spline bug


> If I was managing this project, I would certainly consider this an
> unacceptable approach / sequence of events.

Usually, I'm quite conservative, and Graham has provided a lot of
useful additions to the FreeType code.  I trust him to fix that in due
course.

> Better to keep the inefficiency than (a) break the functionality,
> and (b) end up with code that is not obviously correct.

Uh, oh, you are extremely exaggerating.  This is not helpful.  Until
the bug report we weren't aware of problems, and Graham's change is in
since two months already, and the code is probably run by millions of
people...

It's easy to revert if no satisfying solution can be found.

Thanks for offering help, though.


    Werner