This compiler doesn't recognize the end-of-comment sequence `*/' if
it immediately follows non-ASCII characters.
* src/autofit/afscript.h: Ensure whitespace before `*/'.
* src/truetype/ttinterp.c (Ins_SHPIX): Allow SHPIX to move points in
the twilight zone. Otherwise, treat SHPIX the same as DELTAP.
Unbreaks various fonts such as older versions of Rokkitt and DTL
Argo T Light that would glitch severly after calling ALIGNRP after a
blocked SHPIX.
Suggested by Alexei. `UL' is only needed for 16bit compilers, but
it seems noone is using this anymore (and we no longer test whether
FreeType compiles in such an environment). Otherwise, it is easy to
add the postfix to the `AF_UNICODE_RANGE' macro.
The convergence of Bézier flatteners is fast with the deviation
from straight line being assymptotically cut 4-fold on each bisection.
This justifies smaller bisection stack size.
* src/smooth/ftgrays.c (gray_TWorker): Remove common `bez_stack'.
(gray_render_conic): Create and use conic `bez_stack'. Move back the
band analysis from...
(gray_conic_to): ... here.
(gray_render_cubic): Create and use cubic `bez_stack'. Move back the
band analysis from...
(gray_cubic_to): ... here.
(gray_move_to): Updated.
* src/autofit/afblue.dat: Add blue zone data for Armenian.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Armenian standard characters.
* src/autofit/afranges.c: Add Armenian data.
* src/autofit/afstyles.h: Add Armenian data.
This essentially moves the Gurmukhi script from the `Indic' hinter to
the `Latin' hinter.
* src/autofit/afblue.dat: Add blue zone data for Gurmukhi.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Gurmukhi standard characters and move
data out of AF_CONFIG_OPTION_INDIC block.
* src/autofit/afranges.c: Move Gurmukhi data out of
AF_CONFIG_OPTION_INDIC block.
* src/autofit/afstyles.h: Update Gurmukhi data; in particular, use
AF_WRITING_SYSTEM_LATIN.
This patch prepares data structures and the like.
See added comments in `ttinterp.h' for more information on this and
the following commits in the series.
* devel/ftoption.h, include/freetype/config/ftoption.h
(TT_CONFIG_OPTION_SUBPIXEL_HINTING): Assign values to differentiate
between subpixel versions.
(TT_SUPPORT_SUBPIXEL_HINTING_INFINALITY,
TT_SUPPORT_SUBPIXEL_HINTING_MINIMAL): New macros.
* include/freetype/ftttdrv.h (TT_INTERPRETER_VERSION_40): New macro.
* include/freetype/internal/tttypes.h (TT_FaceRec): Updated.
* src/truetype/ttinterp.h (TT_ExecContextRec): Define new fields
`subpixel_hinting_lean', `vertical_lcd_lean',
`backwards_compatibility', `iupx_called', iupy_called', and
`grayscale_cleartype' for new hinting mode.
* src/truetype/ttdriver.c (tt_property_set): Handle v38 and v40
interpreters conditionally.
* src/truetype/ttgload.c (TT_Hint_Glyph): Save phantom points unless
in v38 backwards compatibility mode.
Updated.
(compute_glyph_metrics): Add v38 backwards compatibility mode
constraint for adjusting advance widths.
Updated.
(tt_loader_init): Handle new flags `subpixel_hinting_lean',
`grayscale_cleartype', and `vertical_lcd_lean'.
Updated.
(tt_get_metrics, TT_Process_Simple_Glyph, TT_LOADER_SET_PP):
Updated.
* src/truetype/ttobjs.c (tt_driver_init): Conditionally set
default interpreter version number.
* src/truetype/ttsubpix.c, src/truetype/ttsubpix.h: Updated.
* include/freetype/config/ftstdlib.h (FT_LONG_MIN): New macro.
* src/cff/cffparse.c (cff_parse_font_matrix): Use largest scaling
value of all matrix coefficients to scale matrix.
* src/cff/cffobjs.c (cff_face_init): Use `matrix->yx' member for
matrix normalization if `matrix->yy' is zero.
The previous fix for #46372 misunderstood a composite glyph referring
same component twice as a recursive reference. See the discussion
http://lists.gnu.org/archive/html/freetype/2016-05/msg00000.html
Thanks to Khaled Hosny for finding this issue.
* src/truetype/ttgload.c (ft_list_get_node_at): A function to get
the i-th node from FT_List. (load_truetype_glyph): In the traversal
scan of the reference tree in the composite glyph, we clear the
nodes filled by previous sibling chain.
We never dereference `face_id', and some implementations might use a
running number instead of a pointer. Additionally, disallowing
value zero was undocumented.
* src/cache/ftccmap.c (FTC_CMapCache_Lookup), src/cache/ftcmanag.c
(FTC_Manager_LookupFace, FTC_Manager_RemoveFaceID): Remove test for
`face_id'.
A single decrement counter of segments to draw, instead of an array,
contains all the information necessary to decide when to split and
when to draw a conic segment. The number of splits before each draw is
equal to the number of trailing zeros in the counter.
* src/smooth/ftgrays.c (gray_TWorker): Remove `lev_stack'.
(gray_render_conic): Updated to use decrement counter of segments.
Otherwise some properties could be set to arbitrary values, which is
harmless, but querying could give wrong positive results.
* src/cff/cffdrivr.c (cff_property_set) [hinting-engine],
* src/truetype/ttdriver.c (tt_property_set) [interpreter-version]:
Only allow defined values.
This essentially moves the Gujarati script from the `Indic' hinter to
the `Latin' hinter.
* src/autofit/afblue.dat: Add blue zone data for Gujarati.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Gujarati standard characters and move
data out of AF_CONFIG_OPTION_INDIC block.
* src/autofit/afranges.c: Move Gujarati data out of
AF_CONFIG_OPTION_INDIC block.
* src/autofit/afstyles.h: Update Gujarati data; in particular, use
AF_WRITING_SYSTEM_LATIN.
Georgian is problematic, since `uppercase' forms of Mkhedruli
(called Mtavruli) are not yet defined in Unicode, which means that
proper blue zones can't be defined. However, there is already a
proposal submitted to Unicode; see
http://www.unicode.org/L2/L2016/16034-n4707-georgian.pdf
Additionally, due to historical reasons, Unicode treats Khutsuri as
the same script as Mkhedruli, and so does OpenType. However, since
the two scripts have completely different shapes it can happen that
blue zones differ considerably. The tag `geok' used here (derived
from ISO 15924) to differentiate the two scripts is not an OpenType
tag in use. If we now have a font that contains both glyphs for
Mkhedruli and Khutsuri, and it uses OpenType features for both also,
HarfBuzz unavoidably treats all glyphs as `geor'. As a consequence,
blue zones for `geok' are not used for glyphs involved in the
OpenType features.
An issue not yet resolved is which OpenType feature should be used
to access Mtavruli glyph shapes; right now, FreeType doesn't set up
support for them, but it is easy to add them later on as soon as
more information is available.
* src/autofit/afblue.dat: Add blue zone data for Georgian.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Georgian standard characters.
* src/autofit/afranges.c: Add Georgian data.
* src/autofit/afstyles.h: Add Georgian data.
Even if the dummy hinter is used as the handler for `none' (which
doesn't use blue zones), it is more consistent than the old value
(which was 0), pointing to Arabic...
* src/autofit/afblue.dat: Add `AF_BLUE_STRINGSET_NONE'.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afstyles.h (none_dflt): Use AF_BLUE_STRINGSET_NONE.
* src/pfr/pfrsbit.c (pfr_lookup_bitmap_data): Replace `left',
`right', and `middle' with `min', `max', and `mid' as used in other
FreeType binary search code.
(pfr_load_bitmap_metrics): Fix invalid left shift.
We did a binary search for a charcode without ensuring that the
searched data is ordered. Validating the order is now done lazily,
this is, the first access to a bitmap glyph triggers the order check
in the corresponding bitmap strike.
* src/pfr/pfrtypes.h (PFR_BitmapFlags): New values
`PFR_BITMAP_VALID_CHARCODES' and `PFR_BITMAP_CHARCODES_VALIDATED'.
* src/pfr/pfrsbit.c (pfr_lookup_bitmap_data): Make `flags' argument
a pointer. Handle new PFR_BITMAP_XXX flags.
(pfr_slot_load_bitmap): Updated.
Extra items are indicated with different bit positions.
* src/pfr/pfrtypes.h (PFR_GlyphFlags): Replace
`PFR_GLYPH_EXTRA_ITEMS' with `PFR_GLYPH_SIMPLE_EXTRA_ITEMS' and
`PFR_GLYPH_COMPOUND_EXTRA_ITEMS'.
* src/pfr/pfrgload.c (pfr_glyph_load_simple,
pfr_glyph_load_compound): Use them.
* src/pfr/pfrsbit.c (pfr_load_bitmap_metrics): Correctly handle
signed nibbles.
(pfr_slot_load_bitmap): Correctly exit frame in case of error.
Fix invalid left shifts.
* src/smooth/ftgrays.c (gray_render_conic, gray_render_cubic): Move
band clipping from here.
(gray_conic_to, gray_cubic_to): ... to here.
(gray_rander_line, gray_render_scanline): Initialize variables closer
to their use.
We probably did not notice this as all fonts we tested had only
tuple_coords[i] be +1 or -1 for non-intermediate tuples.
* src/truetype/ttgxvar.c (ft_var_apply_tuple): Implement it.
The goal is to avoid integer overflows in the rendering algorithms.
The limit is chosen arbitrarily at some 2^18 pixels, which should be
enough for modern devices including printers.
* src/base/ftoutln.c (FT_Outline_Render): Check CBox and reject
enormous outlines.
* src/autofit/aflatin.c (af_latin_compute_stem_width): Add argument
to pass difference between hinted and unhinted position of base
point; use this to adjust the stem width depending on the PPEM so
that it doesn't become too large under certain circumstances.
Update all callers using value 0 for this argument except...
(af_latin_align_linked_edge): Pass position delta of base point to
`af_latin_compute_stem_width'.
This essentially moves the Sinhala script from the `Indic' hinter to
the `Latin' hinter.
* src/autofit/afblue.dat: Add blue zone data for Sinhala.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Sinhala standard character and move data
out of AF_CONFIG_OPTION_INDIC block.
* src/autofit/afranges.c: Move Sinhala data out of
AF_CONFIG_OPTION_INDIC block.
* src/autofit/afstyles.h: Update Sinhala data; in particular, use
AF_WRITING_SYSTEM_LATIN.
An example that gets better rendered is glyph `uusignTaml' (glyph
index 2286) in font `FreeSerif.ttf' (Version 0412.2263) at 22ppem.
* src/autofit/aflatin.c (af_latin_hints_compute_segments): Properly
handle segments where the last point of the first segment is
identical to the first point in the second one. This can happen for
malformed fonts or spikes. We either merge the new segment with the
previous one (both segments point into the same direction), or we
discard the shorter segment if they point into different directions.
* src/autofit/aflatin.c (af_latin_hints_compute_segments): Change
some local variable names to better differentiate between values
along a segment and values orthogonal to it.
In complex glyph shapes, the original logic was too simple to cater
for situations that would actually need something similar to PS Hint
masks. This fix should alleviate the worst cases.
* src/autofit/aflatin.c (af_latin_hint_edges): Don't allow
complete disappearance of stems.
This essentially moves the Tamil script from the `Indic' hinter to
the `Latin' hinter.
* src/autofit/afblue.dat: Add blue zone data for Tamil.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Tamli standard character and move data
out of AF_CONFIG_OPTION_INDIC block.
* src/autofit/afranges.c: Move Tamil data out of
AF_CONFIG_OPTION_INDIC block.
* src/autofit/afstyles.h: Update Tamil data; in particular, use
AF_WRITING_SYSTEM_LATIN.
This essentially moves the Malayalam script from the `Indic' hinter
to the `Latin' hinter.
* src/autofit/afblue.dat: Add blue zone data for Malayalam.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Malayalam standard characters and move
data out of AF_CONFIG_OPTION_INDIC block.
* src/autofit/afranges.c: Move Malayalam data out of
AF_CONFIG_OPTION_INDIC block.
* src/autofit/afstyles.h: Update Malayalam data; in particular, use
AF_WRITING_SYSTEM_LATIN.
Now all glyphs of MM CFFs like `ITCGaramondMM-It.otf' can be
displayed.
* src/cff/cffgload.c (cff_decoder_parse_charstrings) <cff_op_store,
cff_op_load>: Partially implement it.
* src/cff/cffparse.c (cff_parser_init): Add new parameter to pass
the number of Multiple Master axes.
Update all callers.
(cff_parse_multiple_master): Get number of axes.
(cff_parser_run) <opcode 31>: Updated.
* src/cff/cffparse.h: Updated.
(CFF_ParserRec): Add `num_axes' field.
* src/cff/cffload.c: Updated.
* src/cff/cfftypes.h (CFF_FontRecDictRec): Add `num_axes' field.
We need this to properly trace Multiple Master CFFs, which contain
two SIDs that are charstrings.
This commit makes FreeType also show the last SID, omitted
previously due to a bug.
* src/cff/cfftypes.h (CFF_FontRec): Add `string_pool_size' field.
* src/cff/cffload.c (cff_index_get_pointers): Add argument to return
the pool size.
Update all callers.
* src/cff/cffobjs.c (cff_face_init) [FT_DEBUG_LEVEL_TRACE]: Directly
access `cff->strings' to display the non-default strings.
The new code only displays the first master in the font.
* src/cff/cffgload.c (cff_decode_parse_charstrings): Add new
parameter to allow function calls from dictionaries also.
<cff_op_blend>: Partially implement it.
Update all callers.
* src/cff/cffgload.h: Updated.
* src/cff/cffparse.c (cff_parser_init): Add new parameter to pass the
number of Multiple Master designs.
Update all callers.
(cff_parse_multiple_master): New function to rudimentarily parse
operator.
(cff_parser_run): Handle `T2' operator.
* src/cff/cffparse.h: Updated.
(CFF_ParserRec): Add `num_designs' field.
* src/cff/cffload.c: Updated.
* src/cff/cfftoken.h: Handle `MultipleMaster' operator.
* src/cff/cfftypes.h (CFF_FontRecDictRec): Add `num_designs' field.
* src/sfnt/sfobjs.c (sfnt_init_face): Don't handle `fvar' table for
MM CFFs.
This essentially moves the Kannada script from the `Indic' hinter to
the `Latin' hinter.
* src/autofit/afblue.dat: Add blue zone data for Kannada.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Kannada standard characters and move
data out of AF_CONFIG_OPTION_INDIC block.
* src/autofit/afranges.c: Move Kannada data out of
AF_CONFIG_OPTION_INDIC block.
* src/autofit/afstyles.h: Update Kannada data; in particular, use
AF_WRITING_SYSTEM_LATIN.
Patch taken from ttfautohint, commit
071ae2c00e0d67f9d19418f4fade1c23d27dc185.
There were two bugs.
- We now use non-standard script tags like `khms' for special
purposes. However, HarfBuzz maps such tags to `DFLT', and
without this commit the associated lookups were incorrectly
assigned to the non-standard tags.
- Let's assume we have a Bengali font, and the font's `DFLT'
script tag handles the necessary lookups for Bengali, too.
Without this commit, the `DFLT' lookups were assigned to
ttfautohint's default script (usually `latn') before the
standard lookups for Bengali were handled.
We now have the following order while searching for covered
glyph indices.
special features of scripts (e.g. `sups' for Cyrillic)
Unicode mappings of scripts
remaining features of scripts (especially important for Indic
scripts)
default features of default script
* src/autofit/afshaper.c, src/autofit/afshaper.h
(af_shaper_get_coverage): Add boolean parameter to indicate default
script.
Update all callers.
* src/autofit/afglobal.c (af_face_globals_compute_style_coverage):
Fix search order for coverages.
* src/base/ftlcdfil.c (FT_Library_SetLcdFilterWeights): Enable filter
in addition to setting weights.
(FT_Library_SetLcdFilter): Clean out FT_FORCE_LIGHT_LCD_FILTER and
FT_FORCE_LEGACY_LCD_FILTER.
* include/freetype/ftlcdfil.h: Documentation update.
Such macro names are reserved for both C and C++.
* src/cache/ftccache.h: s/_FTC_FACE_ID_HASH/FTC_FACE_ID_HASH/.
Update all callers.
(FTC_CACHE_LOOKUP_CMP): Replace `_XXX' with `XXX_'.
* src/cache/ftcmru.c (FTC_MRULIST_LOOKUP_CMP): Ditto.
Such macro names are reserved for both C and C++.
* include/freetype/ftimage.h, src/raster/ftraster.c,
src/smooth/ftgrays.c, src/smooth/ftgrays.h:
s/_STANDALONE_/STANDALONE_/.
* src/cff/cf2intrp.c (cf2_interpT2CharString) [cf2_cmdEXTENDEDNMBR,
default]: `|' is not guaranteed to be processed from left to right
by the compiler. However, the code repeatedly calls
`cf2_buf_readByte' to get the arguments to `|' ... Fix this.
This improves rendering of scripts like Bengali or Devanagari.
* src/autofit/afhints.c (af_axis_hints_new_edge): Add parameter to
pass top-to-bottom hinting flag. This makes the function sort edges
in descending vertical position.
* src/autofit/afhints.c: Updated.
* src/autofit/aflatin.c (af_latin_hints_compute_edges,
af_latin_hint_edges): Use `top_to_bottom_hinting' flag.
* src/autofit/afcjk.c (af_cjk_hints_compute_edges),
src/autofit/aflatin2.c (af_latin2_hints_compute_edges): Updated.
Right now, it does nothing.
* src/autofit/afscript.h: Add another parameter to `SCRIPT',
specifying hinting direction.
* src/autofit/afglobal.c, scr/autofit/afglobal.h,
src/autofit/afpic.c, src/autofit/afranges.h, src/autofit/afshaper.c,
src/autofit/aftypes.h: Extend `SCRIPT' definitions.
We now use a hash to map from subr indices to array elements holding
the subroutines, if necessary.
* include/freetype/internal/t1types.h: Include FT_INTERNAL_HASH_H.
(T1_FontRec): Add `subrs_hash' field.
* include/freetype/internal/psaux.h: Include FT_INTERNAL_HASH_H.
(T1_DecoderRec): Add `subrs_hash' field.
* src/type1/t1load.h (T1_LoaderRec): Add `subrs_hash' field.
* src/type1/t1driver.c: Include FT_INTERNAL_HASH_H.
(t1_ps_get_font_value) [PS_DICT_SUBR]: Look up hash if necessary.
* src/type1/t1load.c: Include FT_INTERNAL_HASH_H.
(parse_subrs): Use hash for subr indices that exceed the allocated
number of subr slots.
(t1_init_loader): Remove unnecessary code.
(t1_done_loader, T1_Open_Face): Updated.
* src/type1/t1gload.c (T1_Compute_Max_Advance, T1_Get_Advances,
T1_Load_Glyph): Updated.
* src/type1/t1objs.c (T1_Face_Done): Updated.
* src/psaux/t1decode.c: Include FT_INTERNAL_HASH_H.
(t1_decoder_parse_charstrings) [op_callsubr]: Look up hash if
necessary.
* src/cid/cidgload.c (cid_load_glyph): Updated.
* src/base/fthash.c (hash_num_lookup, hash_num_compare): New
functions.
(ft_hash_init): Add argument to select between number and string
hashing.
(ft_hash_num_insert, ft_hash_num_lookup): New functions.
* include/freetype/internal/fthash.h: Updated.
* src/bdf/bdflib.c (_bdf_parse_start): Updated.
* include/freetype/internal/fthash.c (FT_Hash_LookupFunc,
FT_Hash_CompareFunc, FT_Hash_FreeFunc): New typedefs.
(FT_HashRec): Add `lookup', `compare', and `free' fields.
* src/base/fthash.c (hash_str_lookup, hash_str_compare,
hash_str_free): New functions.
(ft_hash_init): Set function pointers.
(hash_bucket, ft_hash_free): Use them.
We want to support both an integer and a string key later on.
* include/freetype/internal/fthash.h (FT_Hashkey): New union.
(FT_HashnodeRec): Updated.
(ft_hash_insert, ft_hash_lookup): Renamed to ...
(ft_hash_str_insert, ft_hash_str_lookup): ... this.
* src/base/fthash.c (hash_bucket): Updated.
(ft_hash_insert, ft_hash_lookup): Renamed to ...
(hash_insert, hash_lookup): ... this.
(ft_hash_str_insert, ft_hash_str_lookup): New wrapper functions.
* src/bdf/bdflib.c: Updated.
* src/bdf/bdf.h: Include FT_INTERNAL_HASH_H.
(hashnode, hashtable): Removed.
(bdf_font_t): Use `FT_HashRec' type for `proptbl'.
* src/bdf/bdflib.c: Remove all hash functions.
Update code for new hash structure and function names.
This essentially moves the Bengali script from the `Indic' hinter to
the `Latin' hinter.
* src/autofit/afblue.dat: Add blue zone data for Bengali.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Bengali standard characters and move
data out of AF_CONFIG_OPTION_INDIC block.
* src/autofit/afranges.c: Move Bengali data out of
AF_CONFIG_OPTION_INDIC block.
* src/autofit/afstyles.h: Update Bengali data; in particular, use
AF_WRITING_SYSTEM_LATIN.
The BDF specification only allows decimal numbers, no octal or
hexidecimal decoding is needed.
* src/bdf/bdflib.c (_bdf_atoul, _bdf_atol, _bdf_atous,
_bdf_atos): Remove unused code and parameters.
Update all callers.
(odigits): Remove.
* src/base/ftobjs.c (open_face_PS_from_sfnt_stream):
`ft_lookup_PS_in_sfnt_stream' returns offset and and length from
user supplied data. Use of this these values must be checked.
This was removed while rewriting the HarfBuzz interface.
* src/autofit/afglobal.h (AF_FaceGlobalsRec): Add `hb_buf' field to
hold internal HarfBuzz buffer, needed for feature comparison.
* src/autofit/afglobal.c (af_face_globals_new,
af_face_globals_free): Initialize and destroy `hb_buf'.
* src/autofit/afshaper.c (af_shaper_get_cluster): Compare character
(cluster) with and without applied feature.
* src/autofit/aflatin.c (af_latin_metrics_init_blues): Fix tracing
message.
This feature is mainly for Khmer: The idea is to avoid a clash
between the top of subscript glyphs and the bottom of normal
baseline glyphs.
This only works for character clusters mapped to multiple glyphs.
* src/autofit/afblue.dat: Add subscript top blue zone for Khmer.
* src/autofit/afblue.hin (AF_BLUE_PROPERTY_LATIN_SUB_TOP): New
macro.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/aflatin.h (AF_LATIN_IS_SUB_TOP_BLUE,
AF_LATIN_BLUE_SUB_TOP): New macros.
* src/autofit/aflatin.c (af_latin_metrics_init_blues): Handle new
blue zone property.
Update tracing messages.
(af_latin_metrics_scale_dim): Handle new blue zone property.
(af_latin_hints_compute_blue_edges): Updated.
We split Khmer into two auto-hinter scripts: `Khmer' (`khmr') and
`Khmer symbols' (`khms', U+19E0-U+19FF).
* src/autofit/afblue.dat: Add blue zone data for Khmer.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Khmer standard characters.
* src/autofit/afranges.c: Add Khmer data.
* src/autofit/afstyles.h: Add Khmer data.
Scripts like Khmer have blue zones that can't be directly
represented by Unicode characters. Instead, it is necessary to let
HarfBuzz convert character clusters into proper glyph representation
forms, then deriving the blue zone information from the resulting
glyphs.
* src/autofit/hbshim.c, src/autofit/hbshim.h: Replaced by...
* src/autofit/afshaper.c, src/autofit/afshaper.h: ... these two new
files, providing a new API to access HarfBuzz.
The new API manages a HarfBuzz buffer with `af_shaper_buf_create'
and `af_shaper_buf_destroy'. The buffer receives a UTF8 encoded
string with function `af_shaper_get_cluster', and the resulting
glyph data (indices, advance widths, vertical offsets) can be
iteratively accessed with function `af_shaper_get_elem'.
* src/autofit/afcjk.c (af_cjk_metrics_init_widths,
af_cjk_metrics_init_blues, af_cjk_metrics_check_digits): Updated.
* src/autofit/aflatin.c (af_latin_metrics_init_widths,
af_latin_metrics_init_blues, af_latin_metrics_check_digits):
Updated.
* include/freetype/internal/fttrace.h: s/afharfbuzz/afshaper/.
* src/autofit/afglobal.c: s/hbshim.h/afshaper.h/.
(af_face_globals_compute_style_coverage): Updated.
* src/autofit/afglocal.h: s/hbshim.h/afshaper.h/.
* src/autofit/autofit.c: s/hbshim.c/afshaper.c/.
* src/autofit/Jamfile, src/autofit/rules.mk (AUTOF_DRV_SRC):
Updated.
This is more flexible; additionally, it would allow character
clusters.
* src/autofit/aftypes.h (SCRIPT, AF_DEFINE_SCRIPT_CLASS): Updated.
(AF_ScriptClassRec): Replace `standard_char[123]' with
`standard_charstring'.
* src/autofit/afscript.h: Replace last three character arguments
of the `SCRIPT' calls with a string parameter, holding the standard
characters (in UTF-8 encoding) separated with spaces.
* src/autofit/afglobal.c, src/autofit/afglobal.h,
src/autofit/afpic.c, src/autofit/afranges.c, src/autofit/hbshim.c
(SCRIPT): Updated.
* src/autofit/afcjk.c (af_cjk_metrics_init_widths),
src/autofit/aflatin.c (af_latin_metrics_init_widths): Updated.
In the reference, we show the section's title enclosed in single
quotes.
* src/tools/docmaker/formatter.py (Formatter::__init__): Collect
section names as identifiers.
* src/tools/docmaker/tohtml.py (section_title_header): Split into...
(section_title_header1, section_title_header2): ... these two
strings.
(HtmlFormatter::make_block_url, make_html_word, html_source_quote):
Handle sections.
(HtmlFormatter::section_enter): Updated to add `id' HTML attribute.
We need this to handle equally named properties in different
modules.
* src/tools/docmaker/content.py (re_identifier),
src/tools/docmaker/sources.py (re_crossref): Allow `foo[bar]'.
* src/tools/docmaker/tohtml.py (HtmlFormatter::make_html_word,
HtmlFormatter::index_exit, HtmlFormatter::section_enter,
HtmlFormatter::block_enter): Handle `foo[bar]'.
This does the same as `FT_LCD_FILTER_LEGACY'.
See
https://bugs.freedesktop.org/show_bug.cgi?id=92981
for the reasoning.
* include/freetype/ftlcdfil.h (FT_LcdFilter): New value
`FT_LCD_FILTER_LEGACY1'.
* src/base/ftlcdfil.c (FT_Library_SetLcdFilter): Use it.
Both the native CFF hinter and the auto-hinter now have a very
similar rendering style.
* include/freetype/freetype.h: Mention that FT_LOAD_TARGET_LIGHT no
longer implies FT_LOAD_FORCE_AUTOHINT.
* include/freetype/ftmodapi.h (FT_MODULE_DRIVER_HINTS_LIGHTLY): New
macro.
* include/freetype/internal/ftobjs.h (FT_DRIVER_HINTS_LIGHTLY): New
macro.
* src/cff/cffdrivr.c (cff_driver_class): Use it.
* src/base/ftobjs.c (FT_Load_Glyph): Update auto-hinter selection
logic.
* src/autofit/aflatin.h (AF_LatinBlueRec): Add `ascender' and
`descender' fields.
* src/autofit/aflatin.c (af_latin_metrics_init_blues): Collect
ascender and descender data for blue zones.
(af_latin_metrics_scale_dim): Reject vertical scaling values that
change the result by more than two pixels.
* include/freetype/internal/tttypes.h (TT_LoaderRec): New field
`composites'.
* src/truetype/ttgload.c: Include FT_LIST_H.
(load_truetype_glyph): Add composite subglyph index to a list;
abort if index is already in list.
(tt_loader_init): Updated.
(tt_loader_done): New function.
(TT_Load_Glyph): Call `tt_loader_done'.
This is a crude adaption of the original `cf2_computeDarkening'
function.
* src/autofit/afloader.c (af_intToFixed, af_fixedToInt,
af_floatToFixed): New macros, taken from `cf2fixed.h'.
(af_loader_compute_darkening): New function.
* src/autofit/afloader.h: Updated.
We need the computed standard horizontal and vertical widths for the
emboldening calculation. This method provides a convenient way to
extract it from writing-system-specific metrics structures, which
all script definitions must implement.
* src/autofit/aftypes.h (AF_WritingSystem_GetStdWidthsFunc): New
function type.
(AF_WritingSystemClassRec): New member `style_metrics_getstdw'.
(AF_DEFINE_WRITING_SYSTEM_CLASS): Updated.
* src/autofit/afcjk.c (af_cjk_get_standard_width): New function.
(af_cjk_writing_system_class): Updated.
* src/autofit/afdummy.c (af_dummy_writing_system_class): Updated.
* src/autofit/afindic.c (af_cjk_get_standard_width): New function.
(af_indic_writing_system_class): Updated.
* src/autofit/aflatin.c (af_latin_get_standard_width): New function.
(af_indic_writing_system_class): Updated.
* src/autofit/aflatin.c (af_latin_get_standard_width): New function.
(af_indic_writing_system_class): Updated.
Actual code follows in a later commit.
* include/freetype/ftautoh.h: Document `no-stem-darkening' and
`darkening-parameters'.
* src/autofit/afmodule.h: New fields `no_stem_darkening' and
`darken_params'.
* src/autofit/afmodule.c (af_property_set, af_property_get):
Handle them.
(af_autofitter_init): Initialize them.
Currently, libFuzzer only supports mutation of a single file. We
circumvent this problem by using an uncompressed tar archive as
multiple-file input for the fuzzer.
This patch enables tests of `FT_Attach_Stream' and AFM/PFM parsing;
a constructed tarball should contain a font file as the first
element, and files to be attached as further elements.
* src/tools/ftfuzzer/ftfuzzer.cc: Include libarchive headers.
(archive_read_entry_data, parse_data): New functions.
(LLVMFuzzerTestOneInput): Updated.
* src/tools/ftfuzzer/ftmutator.cc: New file, providing a custom
mutator for libFuzzer that can mutate tarballs in a sensible way.
* src/autofit/afhints.c (af_print_idx, af_get_segment_index,
af_get_edge_index): New functions.
(af_glyph_hints_dump_points): Remove unnecessary `|', `[', and `]'.
Add segment and edge index for each point.
Slightly change printing order of some elements.
Don't print `-1' but `--' for missing elements.
(af_glyph_hints_dump_segments, af_glyph_hints_dump_edges): Remove
unnecessary `|', `[', and `]'.
Don't print `-1' but `--' for missing elements.
Problem reported by Nikolay Sivov <bunglehead@gmail.com>.
* src/sfnt/ttsbit.c (tt_face_load_strike_metrics): Avoid zero value
for `metrics->height' by applying some heuristics.
This modifies the addition of subroutine number to subroutine bias
from unsigned to signed, but does not change any results.
* src/cff/cf2ft.c (cf2_initGlobalRegionBuffer,
cf2_initLocalRegionBuffer): Change variable names from (unsigned)
`idx' to (signed) `subrNum', since it is not an index until after
the bias is added.
* src/cff/cf2ft.h: Updated.
* src/cff/cf2intrp.c (cf2_interpT2CharString) <cf2_cmdCALLSUBR>:
Updated similarly.
* src/cid/cidload.c (parse_fd_array): Check `num_dicts' against
stream size.
(cid_read_subrs): Check largest offset against stream size.
(cid_parse_dict): Move safety check to ...
(cid_face_open): ... this function.
Also test length of binary data and values of `SDBytes',
`SubrMapOffset', `SubrCount', `CIDMapOffset', and `CIDCount'.
Due to incorrect casting negative numbers were shown as very large
(positive) integers on 64bit systems.
* src/psaux/t1decode.c (t1_decoder_parse_charstrings) <op_none>:
Use division instead of shift.
* src/bdf/bdflib.c (ACMSG17): New message macro.
(_bdf_parse_t): Add member `size'.
(bdf_load_font): Set `size'.
(_bdf_parse_glyphs): Adjust `cnt' if necessary.
Similar to other font formats, this commit makes the parser no
longer check the whole PCF file but only the header and the TOC if
we just want to get the number of available faces (and a proper
recognition of the font format).
* src/pcf/pcfdrivr.c (PCF_Face_Init): Updated.
Exit quickly if face_index < 0.
* src/pcfread.c (pcf_load_font): Add `face_index' argument.
Exit quickly if face_index < 0.
* src/pcf/pcf.h: Updated.
This patch also contains various other improvements.
* src/tools/ftfuzzer/ftfuzzer.cc: Add preprocessor guard to reject
pre-C++11 compilers.
(FT_Global): New class. Use it to provide a global constructor and
destructor for the `FT_Library' object.
(setIntermediateAxis): New function to select an (arbitrary)
instance.
(LLVMFuzzerTestOneInput): Loop over all faces and named instances.
Also call `FT_Set_Char_Size'.
Use the `gvar' table size instead of the remaining bytes in the
stream.
* src/truetype/ttgxvar.h (GX_BlendRec): New field `gvar_size'.
* src/truetype/ttgxvar.c (ft_var_load_gvar): Set `gvar_size'.
(ft_var_readpackedpoints, ft_var_readpackeddeltas: New argument
`size'.
(tt_face_vary_cvt, TT_Vary_Apply_Glyph_Deltas): Updated.
These tests should mainly help avoid unnecessarily large memory
allocations in case of malformed fonts.
* src/truetype/ttgxvar.c (ft_var_readpackedpoints,
ft_var_readpackeddeltas): Check number of points against stream
size.
(ft_var_load_avar): Check `pairCount' against table length.
(ft_var_load_gvar): Check `globalCoordCount' and `glyphCount'
against table length.
(tt_face_vary_cvt): Check `tupleCount' and `offsetToData'.
Fix trace.
(TT_Vary_Apply_Glyph_Deltas): Fix trace.
Free `sharedpoints' to avoid memory leak.
This implementation renders the entire line segment at once without
subdividing it into scanlines. The main speed improvement comes from
reducing the number of divisions to just two per line segment, which
is a bare minimum to calculate cell coverage in a smooth rasterizer.
Notably, the progression from cell to cell does not itself require any
divisions at all. The speed improvement is more noticeable at larger
sizes.
* src/smooth/ftgrays.c (gray_render_line): New implementation.
=========================
Tag sources with `VER-2-6-1'.
* docs/VERSION.DLL: Update documentation and bump version number to
2.6.1.
* README, Jamfile (RefDoc), builds/windows/vc2005/freetype.vcproj,
builds/windows/vc2005/index.html,
builds/windows/vc2008/freetype.vcproj,
builds/windows/vc2008/index.html,
builds/windows/vc2010/freetype.vcxproj,
builds/windows/vc2010/index.html,
builds/windows/visualc/freetype.dsp,
builds/windows/visualc/freetype.vcproj,
builds/windows/visualc/index.html,
builds/windows/visualce/freetype.dsp,
builds/windows/visualce/freetype.vcproj,
builds/windows/visualce/index.html,
builds/wince/vc2005-ce/freetype.vcproj,
builds/wince/vc2005-ce/index.html,
builds/wince/vc2008-ce/freetype.vcproj,
builds/wince/vc2008-ce/index.html: s/2.6/2.6.1/, s/26/261/.
* include/freetype/freetype.h (FREETYPE_PATCH): Set to 1.
* builds/unix/configure.raw (version_info): Set to 18:1:12.
* CMakeLists.txt (VERSION_PATCH): Set to 1.
* src/autofit/afmodule.c [AF_DEBUG_AUTOFIT]: Ensure C linking for
dumping functions.
* src/bzip2/ftbzip2.c (ft_bzip2_file_fill_input), src/gzip/ftgzip.c
(ft_gzip_file_fill_input): In case of an error, adjust the limit to
avoid copying uninitialized memory.
* src/bzip2/ftbzip2.c (ft_bzip2_file_fill_output), src/gzip/ftgzip.c
(ft_gzip_file_fill_output): In case of an error, adjust the limit to
avoid copying uninitialized memory.
* src/sfnt/ttcmap.c (tt_cmap4_char_map_linear): Add code to better
skip invalid segments.
If searching the next character, provide a more efficient logic to
speed up the code.
Thanks to Danh Hong <danhhong@gmail.com> for guidance with blue zone
characters!
* src/autofit/afblue.dat: Add blue zone data for Lao.
* src/autofit/afblue.c, src/autofit/afblue.h: Regenerated.
* src/autofit/afscript.h: Add Lao standard characters.
* src/autofit/afranges.c: Add Lao data.
* src/autofit/afstyles.h: Add Lao data.
open_face_from_buffer() frees passed buffer if valid font
is not found. But if copying to the buffer is failed,
the allocated buffer should be freed within the caller.
* src/base/ftobjs.c (open_face_PS_from_sfnt_stream): Free
the buffer `sfnt_ps' if an error caused before calling
open_face_from_buffer().
(Mac_Read_sfnt_Resource): Free the buffer `sfnt_data' if
an error caused before calling open_face_from_buffer();
Suggested by Hin-Tak Leung.
* src/bzip2/ftbzip2.c (ft_bzip2_stream_io), src/gzip/ftgzip.c
(ft_gzip_stream_io), src/lzw/ftlzw.c (ft_lzw_stream_io): Do it.
While the current code in `FT_Get_Next_Char' correctly rejects
out-of-bounds glyph indices, it can be extremely slow for malformed
cmaps that use 32bit values. This commit tries to improve that.
* src/sfnt/ttcmap.c (tt_cmap8_char_next, tt_cmap12_next,
tt_cmap12_char_map_binary, tt_cmap13_next,
tt_cmap13_char_map_binary): Reject glyph indices larger than or
equal to the number of glyphs.
* src/base/ftobjs.c (FT_Get_Char_Index): Don't return out-of-bounds
glyph indices.
(FT_Get_First_Char): Updated.
* src/sfnt/ttcmap.c (tt_cmap6_char_next): Don't return character
codes greater than 0xFFFF.
(tt_cmap8_char_index): Avoid integer overflow in computation of
glyph index.
(tt_cmap8_char_next): Avoid integer overflows in computation of
both next character code and glyph index.
(tt_cmap10_char_index): Fix unsigned integer logic.
(tt_cmap10_char_next): Avoid integer overflow in computation of
next character code.
(tt_cmap12_next): Avoid integer overflows in computation of both
next character code and glyph index.
(tt_cmap12_char_map_binary): Ditto.
(tt_cmap12_char_next): Simplify.
(tt_cmap13_char_map_binary): Avoid integer overflow in computation
of next character code.
(tt_cmap13_char_next): Simplify.
* src/base/ftbase.h (FT_MAC_RFORK_MAX_LEN): Maximum length
of the resource fork for Mac OS. The resource fork larger
than 16 MB can be written but could not be handled
correctly, at least in Carbon routine.
See https://support.microsoft.com/en-us/kb/130437
* src/base/ftobjs.c (Mac_Read_POST_Resource): No need `0x'
for `%p' formatter.
* src/base/ftbase.c (Mac_Read_POST_Resource): Check the
fragment and total size of the concatenated POST resource
before buffer allocation.
(Mac_Read_sfnt_Resource): Check the declared size of
sfnt resource before buffer allocation.
* src/base/ftmac.c (read_lwfn, FT_New_Face_From_SFNT):
Check the total resource size before buffer allocation.
This patch fixes weaknesses in function `tt_face_load_font_dir'.
- It incorrectly assumed that valid tables are always at the
beginning. As a consequence, some valid tables after invalid
entries (which are ignored) were never seen.
- Duplicate table entries (this is, having the same tag) were not
rejected.
- The number of valid tables was sometimes too large, leading to
access of invalid tables.
* src/sfnt/ttload.c (check_table_dir): Add argument to return number
of valid tables.
Add another tracing message.
(tt_face_load_font_dir): Only allocate table array for valid
entries as returned by `check_table_dir'.
Reject duplicate tables and adjust number of valid tables
accordingly.
* src/pcf/pcfread.c (pcf_load_font): Take absolute values for
AVERAGE_WIDTH, POINT_SIZE, PIXEL_SIZE, RESOLUTION_X, and
RESOLUTION_Y. In tracing mode, add warnings.
* src/gzip/ftgzip.c (ft_gzip_stream_close): Avoid memory leak.
(ft_gzip_get_uncompressed_file): Correct byte order while reading
unsigned long value. Without this change, the whole optimization of
accessing small files in `FT_Stream_OpenGzip' is never executed! As
a consequence, access to PCF files in general (which are normally
small files) should be much improved now as originally intended.
* src/psaux/psobjs.c (ps_parser_skip_PS_token): If a token is
enclosed in balanced expressions, ensure that the cursor position
doesn't get larger than the current limit.
Alexei Podtelezhnikov