This is inspired by the discussion in and analysis of
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=859
* src/tools/ftfuzzer/ftfuzzer.cc (LLVMFuzzerTestOneInput): Use only
up to 20 face indices.
Use only up to 20 instance indices.
Everything is guarded with TT_CONFIG_OPTION_GX_VAR_SUPPORT.
* include/freetype/internal/tttypes.h (TT_FaceRec): New fields
`var_postscript_prefix' and `var_postscript_prefix_len'.
* src/sfnt/sfdriver.c: Include FT_TRUETYPE_IDS_H.
(sfnt_is_alphanumeric): New wrapperfunction for `ft_isalnum'.
(get_win_string, get_apple_string): Remove `const' from return
value.
(MAX_VALUE_DESCRIPTOR_LEN, MAX_PS_NAME_LEN): New macros.
(hexdigits): New array.
(sfnt_get_var_ps_name): New function, implementing Adobe TechNote
5902 to construct a PS name for a variation font instance.
(sfnt_get_ps_name): Call `sfnt_get_var_ps_name' for font instances.
* src/sfnt/sfobjs.c (sfnt_done_face): Updated.
* src/truetype/ttgxvar.c (tt_set_mm_blend): Reset
`face->postscript_name' to trigger recalculation for new instance
parameters.
Add 128bit MurmurHash 3 function.
Everything is guarded with TT_CONFIG_OPTION_GX_VAR_SUPPORT.
* src/sfnt/sfdriver.c (ROTL32): New macro.
(fmix32, murmur_hash_3_128): New functions.
* include/freetype/internal/services/svmm.h (FT_Get_Var_Blend_Func):
Add `normalizedcoords' argument.
* src/truetype/ttgxvar.h (GX_BlendRec): Add `coords' field to store
the design coordinates of the current instance.
Updated.
* src/truetype/ttgxvar.c (TT_Set_MM_Blend): Move functionality to...
(tt_set_mm_blend): ... New function.
Convert data in `normalizedcoords' array to `coords' array on
demand.
(TT_Set_Var_Design): Store argument data in `coords' array.
(TT_Get_Var_Design): Get data from `coords' array.
(tt_get_var_blend): Updated.
(tt_done_blend): Updated.
* src/cff/cffload.c, src/cff/cffload.h (cff_get_var_blend): Updated.
* src/cff/cf2ft.c (cf2_getNormalizedVector): Updated.
* src/cff/cffobjs.c (cff_face_init): Updated.
Reported as
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=759
* src/sfnt/sfobjs.c (sfnt_init_face): While setting number of
instances to zero for `CFF' fonts table, ensure that there is no
`CFF2' present also (which gets priority).
Reported as
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=738
* src/sfnt/sfobjs.c (sfnt_init_face): While setting number of
instances to zero for `CFF' fonts table, ensure that there is no `glyf'
present also (which gets priority).
As documented in the OpenType specification, an entry for the
default instance may be omitted in the named instance table. In
particular this means that even if there is no named instance table
in the font we actually do have a named instance, namely the default
instance.
For consistency, we always want the default instance in our list of
named instances. If it is missing, we try to synthesize it.
* src/sfnt/sfobjs.c (sfnt_init_face): Check whether the default
instance is in the table of named instances. Otherwise adjust
number of instances.
* src/truetype/ttgxvar.c: Include FT_TRUETYPE_IDS_H.
(TT_Get_MM_Var): Use `face->root.style_flags' as the number of named
instances.
Sythesize a named instance entry if necessary.
(tt_done_blend): Free `normalized_stylecoords'.
* src/sfnt/sfdriver.c (sfnt_ps_map): New array.
(sfnt_is_postscript): New function.
(char_type_func): New typedef.
(get_win_string, get_apple_string): Add argument to specify
character checking function.
Add argument whether argument checking failures should be reported.
Update callers.
(search_name_id): Fix return value.
* src/sfnt/sfdriver.c (sfnt_get_ps_name): Split off some
functionality into...
(IS_WIN, IS_APPLE): ... New macros.
(get_win_string, get_apple_string): ... New functions.
We need this for support of the `random' operator.
* include/freetype/ftcffdrv.h (FT_PARAM_TAG_RANDOM_SEED): New macro.
* include/freetype/internal/ftobjs.h (FT_Face_InternalRec): New
field `random_seed'.
* src/cff/cffobjs.h (CFF_DriverRec): New field `random_seed'.
This commit provides the framework, to be filled with something
useful in the next commits.
* include/freetype/freetype.h (FT_Face_Properties): Declare.
* src/base/ftobjs.c (FT_Face_Properties): New function.
Problem reported as
https://github.com/google/fonts/issues/632
The font in question (Nunito) has values 705 and 713 for the
reference and overshoot values, respectively, of the first blue
zone. Blue zone 2, however, has value 710 for both the reference
and overshoot. At 12ppem, reference and overshoot of blue zone 0
becomes 8px, while blue zone 2 becomes 9px.
A peculiarity of this font is that the tops of isolated vertical
stems like `N' have a slight overshoot also. The auto-hinter tries
to find the nearest blue zone using the *original* coordinates. For
vertical stems, this is value 713. For normal horizontal tops like
in character `E', this is value 710. Since value 713 is mapped to
8px but value 710 to 9px, `N' and similar characters are one pixel
higher than `E', which looks very bad.
This commit sanitizes blue zones to avoid such a behaviour.
* src/autofit/aflatin.c (af_latin_sort_blue): New function.
(af_latin_metrics_init_blues): Sort blue values and remove overlaps.
* src/truetype/ttgxvar.h (GX_HVarTable): Renamed to...
(GX_HVVarTable): ...This.
(GX_Blend): Add fields for `VVAR' table handling.
Other minor updates.
* src/truetype/ttgxvar.c (ft_var_load_hvar): Renamed to...
(ft_var_load_hvvar): ...This.
Handle VVAR loading also (controlled by an additional parameter).
(tt_hadvance_adjust): Renamed to...
(tt_hvadvance_adjust): ...This.
Handle application of advance height also (controlled by an
additional parameter).
(tt_hadvance_adjust, tt_vadvance_adjust): Wrappers for
`tt_hvadvance_adjust'.
* src/truetype/ttdriver.c (tt_service_metrics_variations): Updated.
The number of lowercase characters for computing the top flat blue
zone value was too small (in most cases only `x' and `z'). If one
of the two characters has a large serif, say, it can happen that
FreeType must select between two different values, having a 50%
chance to use the wrong one. As a result, rendering at larger PPEM
values could yield uneven lowercase glyph heights.
Problem reported by Christoph Koeberlin <christoph@koe.berlin>.
* src/autofit/afblue.dat (AF_BLUE_STRING_LATIN_SMALL): Replaced
with...
(AF_BLUE_STRING_LATIN_SMALL_TOP, AF_BLUE_STRING_LATIN_SMALL_BOTTOM):
... New, extended sets.
(AF_BLUE_STRINGSET_LATN): Updated.
* src/autofit/afblue.c, scr/autofit/afblue.h: Regenerated.
Werner Lemberg