[cairo] Rotated glyph inked area problem
behdad at behdad.org
Tue Sep 14 21:17:32 PDT 2010
Thanks for the message.
At this time, I think there's still room for much experimentation before one
can denounce this deficiency as a FreeType shortcoming. Sure, FreeType could
do a better job at rasterization, but then again, users like cairo are always
free to do their own rasterization, and jumping at "fixing" FreeType before
fully understanding the problem may be a mistake. I hope that the work I'm
currently doing (an always-device-independent text rendering system) will help
us better know what we want from FreeType in the future. But even that leaves
cases that FreeType as it is today would still be desirable for the
On 09/14/10 23:52, mpsuzuki at hiroshima-u.ac.jp wrote:
> Dear Ian and Behdad,
> Thank you for the description the cairo developers
> found for FreeType. Also I relieved that you're not
> saying as FT is so buggy and cairo decided to start
> new font rasterizer.
> About 2 years ago, I worked to improve FT sources for
> 16bit & 64bit platforms, I found several ILP32-centric APIs
> and hard to fix them without breaking API compatibility.
> At that time, I didn't have good reason to introduce
> incompatible APIs and cannot guess when I can fix.
> Now I think cairo developers have shown some good
> concrete examples for FT developers to consider the
> introduction of new/incompatible APIs in future releases.
> On Tue, 14 Sep 2010 13:17:56 -0400
> Behdad Esfahbod <behdad at behdad.org> wrote:
>> On 09/14/10 13:11, Ian Britten wrote:
>>>> Sorry for asking a stupid question. This issue is caused
>>>> by the bug(s) in FreeType's scaler, rasterizer etc? Or,
>>>> caused by API limitation? Behdad's comment in 2 years ago
>>>> told that 26.6 fixed type is insufficient precision for
>>>> the purpose.
>> It's a limitation of FreeType's 26.6 design. And is going to stay that way
>> for the foreseeable future AFAICT. FreeType is simply not the best rasterizer
>> at small sizes.
More information about the cairo