Re: [ft-devel] rendering differences between 2.4.2 and 2.4.4

[Werner LEMBERG]

> All arches seem to have lost weight.  It is possible that the conic
> Bezier flattener is not splitting enough.

I'm now really interested what David and Bevan are saying :-)

> You may want to fiddle with line 900 in ftgrays.c
>
>   while ( dx > ONE_PIXEL / 4 )
>
> make it
>
>   while ( dx > ONE_PIXEL / 6 )

Indeed, this makes the rendering almost equal to the previous result.
I favor this solution as a quick fix $(Q#|(B I consider this degradation
seriously enough to make a new release.  It's really astonishing that
it has escaped our checks...

> I still think that both cubic and conic flatteners should use
> essentially the same code.  Currently, the cubic solution is
> actually controlling the amount of error while doing the job,
> whereas the conic one is trying to make the best guess ahead of
> doing the job.

I currently can't remember: Your proposed code changes some months ago
handles this better?  Otherwise, do you have a fix?


    Werner