Re: [Devel] FT_CONFIG_OPTION_USE_CMAPS?

[David Turner]

Hello Tom,

Wenlin Institute a écrit :
> 
> Hi,
> 
> With FreeType 2.0.6, calling FT_New_Face() with any .ttf font would
> succeed, and magically almost always resulted in
> face->charmap->encoding == ft_encoding_unicode, even for an
> ASCII/Latin1 font.
> 
> Now with 2.1.0, although FT_New_Face() still returns FT_Err_Ok, it
> gives face->charmap == NULL, even for Unicode fonts, and
> FT_Select_Charmap(face, ft_encoding_unicode) returns
> FT_Err_Invalid_Argument. This is on MacOS X, by the way, but does not
> appear to involve ftmac.c.
>
Please try the current CVS sources. I've just commited several fixes
to this problem, and everything should be fine by now..
 
> The old (2.0.6) behavior can be restored by #undef
> FT_CONFIG_OPTION_USE_CMAPS in ftoption.h, where there is a mysterious
> comment, "Define this if you want generic cmap support."
> 
> Presumably FT_CONFIG_OPTION_USE_CMAPS is an improvement, or it
> wouldn't exist, but what is the improvement and how does one take
> advantage of it?
> 
> If FT_CONFIG_OPTION_USE_CMAPS is defined, is it still possible to
> determine whether a font is Unicode-compatible, in the sense that
> FT_Get_Char_Index() will interpret the charcode as Unicode, and this
> won't result in a random glyph?
>
FT_CONFIG_OPTION_USE_CMAPS is an on-going effort to increase the
felxibility of charmaps within FreeType. It's not really the culprit
here.


Regards,

- David Turner
- The FreeType Project  (www.freetype.org)