So the good news is that with the 2.1.9-style sigma,
FT 2.1.10pre is much better than FT 2.1.9 for FT_Glyph_Stroke(),
and no worse than FT 2.1.9 for FT_Glyph_StrokeBorder().
David wrote me that his sigma change tried to fix a peculiar case --
he no longer has the original data, unfortunately.
CVSROOT:/cvsroot/freetype
Module name:freetype2
Branch:
Changes by: Werner LEMBERG [EMAIL PROTECTED] 05/06/01 15:30:44
Modified files:
. : ChangeLog
src/base : ftstroke.c
Log message:
* src/base/ftstroke.c
I find that the function ftc_face_node_compare() use return FT_BOOL(
node-face_id == face_id ); to determine whether a node exists in the MRU
list.
But the face_id is just a pointer, then if I write the following code, it
will not work as I expect:
* code begin **
Hello Anstinus,
But the face_id is just a pointer, then if I write the
following code, it
will not work as I expect:
[code removed]
The ftFace1 and ftFace2 will have the same value!
An important design point of the cache subsystem is that a
FTC_FaceID value (i.e. the address contained
Thanks for the answer.
more question here:
You don't need to cache your My_FaceID_Rec objects. Just build a list
of them, corresponding to the list of available faces you have, and
that's all. The lifecycle of FT_Face objects is handled automatically
by the cache sub-system.
Consider the