* src/sfnt/ttload.c, src/bdf/bdflib.c: removing compiler warnings
* src/type42/t42objs.c: removed the bug that prevented un-hinted
outlines to be loaded
src/base/ftobjs.c, src/objs/fttype1.c, src/sfnt/ttcmap0.c,
src/smooth/ftgrays.c: changed uses of "setjmp" and "longjmp"
to "ft_setjmp" and "ft_lonjmp". Removed direct references to
<stdio.h> and <setjmp.h> when appropriate, to eventually replace
them with a FT_CONFIG_STANDARD_LIBRARY_H. Useful for the XFree86
Font Server backend based on FT2.
- fixed the infamous bug that caused the cache to crash with large fonts.
the hash table buckets array wasn't shrinked correctly during cache
flushes..
src/cache/ftccache.i, src/cache/ftcsbits.c: adding various
experimental optimisations to the cache manager
* src/type42/t42parse.c: removing duplicate function
src/type42/t42parse.h, src/type42/t42objs.h, src/type42/t42objs.c,
src/type42/type42.c:
updated the Type42 driver by splitting it into several files since
it makes the code easier to read and maintain. Also fixed the bug
that prevented the correct display of fonts with "ftview"
FT_Get_X11_Font_Format to return an X11-compatible string describing the
font format of a given face. This was put in a new optional base source
file, corresponding to a new public header (named FT_XFREE86_H since
this function should only be used within the XFree86 font server IMO).
* include/freetype/config/ftheader.h: adding FT_XFREE86_H, though it's
not documented yet.
* include/freetype/t1tables.h, src/base/fttype1.c: adding two new APIs
named "FT_Get_PS_Font_Info" and "FT_Has_PS_Glyph_Names". This required
a new optional source in 'src/base' named "fttype1.c"
* src/base/Jamfile, src/base/rules.mk, src/base/descrip.mms: updating
build control files for the new files "ftxf86.c" and "fttype1.c" in
src/base
src/sfnt/sfobjs.c, src/sfnt/sfdriver.c, src/base/ftnames.c:
fixing the SFNT name table loader to support various buggy fonts.
it now ignores empty name entries, entries with invalid pointer
offsets and certain fonts containing tables with broken "storageOffset"
fields.
name strings are now loaded on demand, which reduces the memory
requirements for a given FT_Face tremendously (for example, the
name table of Arial.ttf is about 10Kb and contains 70 names !!)
finally, this is a _quick_ fix. The whole name table loader and
interface will be rewritten in a much more cleanly way shortly,
once CSEH have been introduced in the sources.