face->cid_stream so that we can deallocate it safely.
Make the PS parser more tolerant w.r.t. non-standard font data. In
general, an error is only reported in case of a syntax error; a
wrong type is now simply ignored (if possible). To be independent
of the order of various MM-specific keywords, the parse_shared_dict
routine has been removed -- the PS parser is now capable to skip
this data. It no longer fails on parsing e.g.
dup /WeightVector exch def
Since the token following /WeightVector isn't `[' (starting an
array) it is simply ignored.
* include/freetype/fterrdef.h: Define `FT_Err_Ignore' (0xA2) as a
new internal error value.
* src/type1/t1load.c (parse_blend_axis_types,
parse_blend_design_positions, parse_blend_design_map): Return
T1_Err_Ignore if no proper array is following the keyword.
(parse_weight_vector): Use T1_ToTokenArray, initializing `blend'
structure, if necessary.
Return T1_Err_Ignore if no proper array is following the keyword.
(parse_shared_dict): Removed.
(parse_encoding): Set parser->root.error to return T1_Err_Ignore
if no result can be obtained.
Check for errors before accessing `elements' array.
(t1_keywords): Remove /shareddict.
(parse_dict): Reset error if t1_load_keyword returns T1_Err_Ignore.
Set keyword_flag only in case of success.
Check error code if skipping an unrecognized token.
(T1_Open_Face) [!T1_CONFIG_OPTION_NO_MM_SUPPORT]: Call T1_Done_Blend
if blend commands haven't set up a proper MM font.
* src/psaux/psobjs.c (ps_parser_load_field_table): Remove special
code for synthetic fonts.
Return PSaux_Err_Ignore if no proper value has been found.
<cff_op_endchar>: Preserve glyph width before calling
cff_operator_seac.
* src/cff/cffgload.c (cff_decoder_parse_charstrings): Handle special
first argument for `hintmask' and `cntrmask' operators also.
`hardcode_libdir_flag_spec', and `wl'.
* builds/unix/configure: Regenerated.
* builds/unix/freetype-config.in: Make --prefix and --exec-prefix
actually work.
Report a proper --rpath (or -R) value for --libs argument if a
shared library has been built.
* docs/CHANGES: Updated.
outside of /Subrs or /CharStrings. This can happen if there is
additional code manipulating those two arrays so that FreeType
doesn't recognize them properly.
(T1_Open_Face): Improve an error message.
Remove CR/LF on many files.
* src/cache/*, include/freetype/cache/*: fixing a bug after heavy
testing. The current sources are now "release candidates" for the
final version of the cache sub-system
* Jamfile: updating "refdoc" target, and adding "autohint" to the
list of modules to build. Both the autohinter and autofitter will be
built by default. But which one will be used is determined by
the content of "ftmodule.h"
* src/autofit/*: much updates, but the code is still buggy as hell.
Aargh..
accent offset.
Update code similarly to the seac support for Type 1 fonts.
(cff_decoder_parse_charstrings) <cff_op_endchar>: Fix magnitude
of accent offset.
Don't hint glyphs twice if seac is emulated.
<cff_op_flex>: Assign correct point tags.
* docs/CHANGES: Updated.
* src/type1/t1parse.c (T1_Get_Private_Dict): Use FT_MEM_MOVE, not
FT_MEM_COPY, for copying the private dict.
* src/type1/t1load.c (parse_subrs): Assign number of subrs only
in first run.
(parse_charstrings): Parse /CharStrings in second run without
assigning values.
(parse_dict): Skip all /CharStrings arrays but the first. We need
this for non-standard fonts like `Optima' which have different
outlines depending on the resolution. Note that there is no
guarantee that we get fitting /Subrs and /CharStrings arrays; this
can only be done by a real PS interpreter.
started with `StartData' in CID-keyed Type 1 fonts.
* include/freetype/internal/t1types.h (CID_FaceRec): Add new
members `binary_data' and `cid_stream'.
* src/cid/cidload.c (cid_read_subrs): Use `face->cid_stream'.
(cid_hex_to_binary): New auxiliary function.
(cid_face_open): Add new argument `face_index' to return quickly
if less than zero. Updated all callers.
Call `cid_hex_to_binary', then open and assign memory stream to
`face->cid_stream' if `parser->binary_length' is non-zero.
* src/cid/cidload.h: Updated.
* src/cid/cidobjs.c (cid_face_done): Free `binary_data' and
`cid_stream'.
* src/cid/cidparse.c (cid_parser_new): Check arguments to
`StartData' and set parser->binary_length accordingly.
* src/cid/cidparse.h (CID_Parser): New member `binary_length'.
* src/cid/cidgload.c (cid_load_glyph): Use `face->cid_stream'.
* docs/CHANGES: Updated.
include/freetype/config/ftstdlib.h (ft_atoi): Replaced with...
(ft_atol): This.
* src/base/ftdbgmem.c: s/atol/ft_atol/.
* src/type42/t42drivr.c: s/ft_atoi/ft_atol/.
FTC_FaceID.
* src/cff/cffdrivr.c (cff_get_cmap_info): Call sfnt module's TT CMap
Info service function if the cmap comes from sfnt. Return 0 if the
cmap is sythesized in cff module.
Formatting; updating copyright.
src/autohint/ahhint.c, src/base/ftgloadr.c,
src/base/ftglyph.c, src/base/ftoutln.c,
src/base/ftstroke.c, src/cff/cffload.c, src/truetype/ttgload.c,
src/truetype/ttinterp.c:
introducing the new FT_ARRAY_COPY and FT_ARRAY_MOVE macros
to make copying arrays easier
Steve Hartwell <shspamsink@comcast.net>.
* src/cff/cffdrivr.c: Include FT_SERVICE_TT_CMAP_H.
(cff_services): Added an entry for FT_SERVICE_ID_TT_CMAP.
(cff_get_cmap_info): New function.
(cff_service_get_cmap_info) New entry for cff_services.
* src/sfnt/ttcmap0.c: Exit loop after a format match has been found.
Suggested by Steve Hartwell <shspamsink@comcast.net>.
FT_INTERNAL_OBJECTS_H.
* src/base/ftstroke.c (FT_Outline_GetInsideBorder,
FT_Outline_GetOutsideBorder): s/or/o/ to make it compile with
C++ compilers.
* src/cache/ftcmru.c, include/freetype/cache/ftcmru.h:
s/select/selection/ to avoid compiler warning.
* src/cff/cffload.h: s/select/ftselect/ to avoid potential
compiler warning.
Formatting.
include/freetype/cache/ftcmru.h,
include/freetype/cache/ftcglyph.h,
src/cache/ftcbasic.c,
src/cache/ftccache.c,
src/cache/ftccmap.c,
src/cache/ftcmanag.c:
additional speed optimization to the cache sub-system. It is
now up to 70% faster than the one in the previous table
release (i.e. 2.1.7). Note that the API did slightly change
though.
to describe the `charset' field in FT_WinFNT_HeaderRec.
* src/winfonts/winfnt.c (FNT_Face_Init): Set encoding to
FT_ENCODING_NONE except for FT_WinFNT_ID_MAC.
* include/freetype/freetype.h (FT_Encoding): Improve comment,
based on work by Detlef Wrkner <TetiSoft@apg.lahn.de>.
* docs/CHANGES: Updated.
include/freetype/cache/ftcmanag.h,
include/freetype/cache/ftccache.h,
include/freetype/cache/ftcmanag.h,
include/freetype/cache/ftcmru.h (added),
include/freetype/cache/ftlru.h (removed),
include/freetype/cache/ftcsbits.h,
include/freetype/cache/ftcimage.h,
include/freetype/cache/ftcglyph.h,
src/cache/ftcmru.c,
src/cache/ftcmanag.c,
src/cache/ftccache.c,
src/cache/ftcglyph.c,
src/cache/ftcimage.c,
src/cache/ftcsbits.c,
src/cache/ftccmap.c,
src/cache/ftcbasic.c (added),
src/cache/ftclru.c (removed):
*Complete* rewrite of the cache sub-system to "solve" the
following points:
- all public APIs have been moved to FT_CACHE_H, everything
under "include/freetype/cache" is only needed by client
applications that want to implement their own caches
- a new function named FTC_Manager_RemoveFaceID to deal
with the uninstallation of FaceIDs
- the image and sbit cache are now abstract classes, that
can be extended much more easily by client applications
- better performance in certain areas. Further optimizations
to come shortly anyway...
- the FTC_CMapCache_Lookup function has changed its signature,
charmaps can now only be retrieved by index
- FTC_Manager_Lookup_Face => FTC_Manager_LookupFace
FTC_Manager_Lookup_Size => FTC_Manager_LookupSize (still in
private header for the moment)
Werner Lemberg