Acroread does the same.
* src/truetype/ttgload.c (TT_Process_Composite_Glyph): Call
`Update_Max' to adjust size of instructions array if necessary and
add a rough safety check.
(load_truetype_glyph): Save `loader->byte_len' before recursive
call.
* src/truetype/ttinterp.h, src/truetype/ttinterp.c (Update_Max):
Declare it as FT_LOCAL.
Initialize phantom points before calling the incremental interface
to update glyph metrics.
* src/truetype/ttgload.c (tt_get_metrics_incr_overrides)
[FT_CONFIG_OPTION_INCREMENTAL]: New function, split off from...
(tt_get_metrics): This.
Updated.
(load_truetype_glyph): Use tt_get_metrics_incr_overrides.
After long discussion, we now consider the character width vector
(wx,wy) returned by the `sbw' Type 1 operator as being part of *one*
direction only. For example, if you are using the horizontal
writing direction, you get the horizontal and vertical components of
the advance width for this direction. Note that OpenType and CFF fonts
don't have such a vertical component; instead, the GPOS table can be
used to generate two-dimensional advance widths (but this isn't
handled by FreeType).
* include/freetype/ftincrem.h (FT_Incremental_MetricsRec): Add
`advance_v' field to hold the vertical component of the advance
value.
* src/truetype/ttgload.c (tt_get_metrics), src/cff/cffgload.c
(cff_slot_load), src/type1/t1gload.c
(T1_Parse_Glyph_And_Get_Char_String), src/cid/cidgload.c
(cid_load_glyph): Use it.
Reported by Jeremy Manson <jeremy.manson@gmail.com>.
src/truetype/ttgload.c (load_truetype_glyph): Add parameter to
quickly load the glyph header only.
Update all callers.
(tt_loader_init): Add parameter to quickly load the `glyf' table
only.
Update all callers.
(TT_Load_Glyph): Compute linear advance values for embedded bitmap
glyphs too.
The calculation of `vertBearingX' is not defined in the OTF font
spec so FreeType does a `best effort' attempt. However, this value
is defined in the PDF and PostScript specs, and that algorithm is
better than the one FreeType currently uses:
FreeType: Use the middle of the bounding box as the X coordinate
of the vertical origin.
Adobe PDF spec: Use the middle of the horizontal advance vector as
the X coordinate of the vertical origin.
FreeType's algorithm goes wrong if you have a really small glyph
(like the full-width, circle-like dot at the end of the sentence, as
used in CJK scripts) with large bearings. With the FreeType
algorithm this dot gets centered on the baseline; with the PDF
algorithm it gets the correct location (in the top right). Note
that this is a serious issue, it's like printing the dot at the end
of a Roman sentence at the center of the textline instead of on the
baseline like it should. So i believe the PDF spec's algorithm
should be used in FreeType as well.
The `vertBearingY' value for such small glyphs is also very strange
if no `vmtx' information is present, since the height of the bbox is
not representable for the height of the glyph visually (the
whitespace up to the baseline is part of the glyph). The fix also
includes some code for a better estimate of `vertBearingY'.
* src/base/ftobjs.c (ft_synthesize_vertical_metrics): `vertBearingX'
is now calculated as described by the Adobe PDF Spec. Estimate for
`vertBearingY' now works better for small glyphs completely above or
below the baseline into account.
* src/cff/cffgload.c (cff_slot_load): `vertBearingX' is now
calculated as described by the Adobe PDF Spec. Vertical metrics
information was always ignored when FT_CONFIG_OPTION_OLD_INTERNALS
was not defined.
* src/truetype/ttgload.c (compute_glyph_metrics): `vertBearingX' is
now calculated as described by the Adobe PDF Spec.
Normally, the SCANMODE instruction (if present) to set the drop-out
mode in a TrueType font is located in the `prep' table only and thus
valid for all glyphs. However, there are fonts like `pala.ttf'
which additionally contain this instruction in the hinting code of
some glyphs (but not all). As a result it can happen that a
composite glyph needs multiple drop-out modes for its subglyphs
since the rendering state gets reset for each subglyph.
FreeType collects the hinted outlines from all subglyphs, then it
sends the data to the rasterizer. It also sends the drop-out mode
-- after hinting has been applied -- and here is the error: It sends
the drop-out mode of the last subglyph only; drop-out modes of all
other subglyphs are lost.
This patch fixes the problem; it adds a second, alternative
mechanism to pass the drop-out mode: For each contour, the
rasterizer now checks the first `tags' array element. If bit 2 is
set, bits 5-7 contain the contour's drop-out mode, overriding the
global drop-out mode.
* include/freetype/ftimage.h (FT_CURVE_TAG_HAS_SCANMODE): New macro.
* src/truetype/ttgload.c (TT_Hint_Glyph): Store drop-out mode in
`tags[0]'.
* src/raster/ftraster.c (Flow_Up, Overshoot_Top, Overshoot_Bottom):
Use bits 3-5 instead of 0-2.
(New_Profile): Set the drop-out mode in the profile's `flags' field.
(Decompose_Curve): Check `tags[0]' and set `dropOutControl' if
necessary.
(Vertical_Sweep_Drop, Horizontal_Sweep_Drop,
Horizontal_Gray_Sweep_Drop, Draw_Sweep): Use the profile's drop-out
mode.
nor bitmaps can be considered as containing space `glyphs' only.
* src/truetype/ttpload.c (tt_face_load_loca): Handle the case where
a `glyf' table is missing.
* src/truetype/ttgload.c (load_truetype_glyph): Abort if we have no
`glyf' table but a non-zero `loca' entry.
(tt_loader_init): Handle missing `glyf' table.
* src/base/ftobjs.c (FT_Load_Glyph): Undo change 2008-12-05.
* src/sfnt/sfobjs.c (sfnt_load_face): A font with neither outlines
nor bitmaps is scalable.
include/freetype/config/ftheader.h, include/freetype/freetype.h,
src/base/Jamfile, src/base/rules.mk, src/cff/cffdrivr.c,
src/cff/cffgload.c, src/cff/cffgload.h, src/truetype/ttdriver.c,
src/truetype/ttgload.h, src/truetype/ttgload.c, src/type1/t1driver.c,
src/type1/t1gload.h, src/type1/t1gload.c:
Add a new header named FT_ADVANCES_H declaring some new APIs
to extract the advances of one or more glyphs without necessarily
loading their outlines. Also provide 'fast loaders' for the
TrueType, Type1 and CFF font drivers (more to come later)
* autogen.sh: add checks for minimum version of the 'autotools'
stuff.
Pass dropout rules from the TT bytecode interpreter to the
rasterizer; temporarily this is enabled only if
`USE_SCAN_CONVERSION_RULES' is defined.
* include/freetype/ftimage.h (FT_OUTLINE_SMART_DROPOUTS,
FT_OUTLINE_EXCLUDE_STUBS): New flags for for FT_Outline.
* src/raster/ftraster.c (Vertical_Sweep_Drop, Horizontal_Sweep_Drop,
Horizontal_Gray_Sweep_Drop): Use same mode numbers as given in the
OpenType specification.
Fix mode 4 computation.
(Render_Glyph, Render_Gray_Glyph): Handle new outline flags.
* src/truetype/ttgload.c (TT_Load_Glyph)
[USE_SCAN_CONVERSION_RULES]: Convert scan conversion mode to
FT_OUTLINE_XXX flags.
* src/truetype/ttinterp.c (Ins_SCANCTRL): Enable ppem check.
of the graphics state for subglyphs to...
(TT_Hint_Glyph): This function.
Hinting instructions for a composite glyph apparently refer to the
just hinted subglyphs, not the unhinted, unscaled outline. This
seems to fix Savannah bugs #20973 and (at least partially) #23310.
expanded). This is needed for options like
--includedir='${prefix}/include'.
* builds/unix/detect.mk, configure: Prevent argument expansion in
call to the (real) `configure' script.
* src/truetype/ttgload.c (load_truetype_glyph): Fix compilation if
TT_USE_BYTECODE_INTERPRETER isn't defined.
There exist CFFs which contain opcodes for the Type 1 operators
`hsbw' and `closepath' which are both invalid in Type 2 charstrings.
However, it doesn't harm to support them.
* src/cff/cffgload.c (CFF_Operator): Add `cff_op_hsbw' and
`cff_op_closepath.'
(cff_argument_counts): Ditto.
(cff_decoder_parse_charstrings): Handle Type 1 opcodes 9 (closepath)
and 13 (hsbw) which are invalid in Type 2 charstrings.
recursion depth of 1. This was the maximum value in TrueType 1.0,
and some older fonts don't set this field correctly.
* src/gxvalid/gxvmort1.c
(gxv_mort_subtable_type1_substTable_validate): Fix tracing message.
number of points in contours. Problem reported by Victor Stinner
<victor.stinner@haypocalc.com>.
(TT_Process_Simple_Glyph): Synchronize variable types.
* src/truetype/ttgload.c (tt_loader_init): Re-execute the CVT
program after a change from mono to grayscaling (and vice versa).
Use correct constant for comparison to get `exec->grayscale'.
include/freetype/internal/tttypes.h, src/base/ftgloadr.c,
src/base/ftobjs.c, src/truetype/ttgload.c, src/truetype/ttinterp.c,
src/truetype/ttobjs.c: improvements to native TrueType hinting,
this is a first try, controlled by the FIX_BYTECODE macro in
src/truetype/ttinterp.c
src/autofit/afcjk.c, src/autofit/afhints.c, src/autofit/aflatin.c,
src/autofit/aftypes.h, src/base/ftcalc.c, src/base/ftoutln.c,
src/gzip/ftgzip.c, src/psaux/psconv.c, src/truetype/ttgload.c,
src/type1/t1gload.c:
this is a major patch used to drastically improve the performance
of loading glyphs. This both speeds up loading the glypn vector
themselves and the auto-fitter.
note that we've started using inline assembler with GCC to
implement FT_MulFix, given that this function is so damn
important for the engine's performance.
the resulting speed-up is about 25%.
return error if table is missing.
Check table length in non-FT_OPTIMIZE_MEMORY'ed `tt_face_load_hmtx'.
* src/sfnt/sfobjs.c (sfnt_load_face): Take care of missing metrics
tables. The last change makes Mac bitmap-only font not load and this
fixes it.
* src/truetype/ttgload.c (load_truetype_glyph): Fix compilation error
when FT_CONFIG_OPTION_INCREMENTAL is defined.
after the tables' tags; `hdmx' is TrueType-specific and thus the code
is moved to the truetype module; `get_metrics' is moved here from the
truetype module so that the code can be shared with the cff module.
This pass involves no real changes. That is, the code is moved
verbatim mostly. The only exception is the return value of
`tt_face_get_metrics'.
* include/freetype/internal/sfnt.h, src/sfnt/rules.mk,
src/sfnt/sfdriver.c, src/sfnt/sfnt.c, src/sfnt/sfobjs.c,
src/sfnt/ttload.c, src/sfnt/ttload.h, src/sfnt/ttsbit.c,
src/sfnt/ttsbit.h, src/sfnt/ttsbit0.c: Clean up the SFNT_Interface.
* src/sfnt/ttmtx.c, src/sfnt/ttmtx.h: Metrics-related tables' loading
and parsing code is moved here.
Move `tt_face_get_metrics' here from the truetype module. The return
value is changed from `void' to `FT_Error'.
* include/freetype/internal/fttrace.h: New trace: ttmtx.
* src/truetype/ttpload.c, src/truetype/ttpload.h: `hdmx' loading and
parsing code is moved here.
New function `tt_face_load_prep' splitted from `tt_face_load_fpgm'.
`tt_face_load_fpgm' returns `FT_Err_Ok' if `fpgm' doesn't exist.
* src/cff/cffgload.c, src/cff/cffobjs.c: Update.
* src/truetype/ttgload.c, src/truetype/ttobjs.c: Update.
src/base/ftobjs.c (ft_fake_vertical_metrics): New function to fake the
vertical metrics.
* src/cff/cffgload.c, src/cid/cidgload.c, src/pcf/pcfdrivr.c,
src/type1/t1gload.c, src/winfonts/winfnt.c: Fake the vertical metrics.
The fake metrics is monotone.
* src/truetype/ttgload.c (compute_glyph_metrics): Some fixes and
formattings in vertical metrics faking. There are still rooms for
improvements (and so do the CFF module).
`load_strike_metrics' used to load the strike's metrics.
* src/sfnt/sfdriver.c, src/sfnt/ttsbit.c, src/sfnt/ttsbit.h,
src/sfnt/ttsbit0.c: New function `tt_face_load_strike_metrics'.
* src/pfr/pfrobjs.c (pfr_face_init): Set FT_Bitmap_Size correctly.
* src/winfonts/winfnt.c (FNT_Face_Init): Use `nominal_point_size' for
nominal size unless it is obviously incorrect.
* include/freetype/freetype.h (FT_Bitmap_Size): Update the comments on
FNT driver.
Introduce new size selection interface.
* include/freetype/internal/ftdriver.h (struct FT_Driver_ClassRec_):
Replace `set_char_sizes' and `set_pixel_sizes' by `request_size' and
`select_size'.
* include/freetype/freetype.h (FT_Select_Size, FT_Size_Request_Type,
FT_Size_Request, FT_Request_Size, FT_Select_Size), src/base/ftobjs.c
(FT_Select_Size, FT_Request_Size): API additions to export the new
size selection interface.
* src/base/ftobjs.c (FT_Set_Char_Size, FT_Set_Pixel_Sizes): Use
`FT_Request_Size'.
* include/freetype/internal/ftobjs.h (FT_Match_Size),
src/base/ftobjs.c (FT_Match_Size): New function to match a size
request against `available_sizes'. Drivers supporting bitmap strikes
can use this function to implement `request_size'.
* src/bdf/bdfdrivr.c, src/cid/cidobjs.c, src/cid/cidobjs.h,
src/cid/cidriver.c, src/pcf/pcfdrivr.c, src/type1/t1driver.c,
src/type1/t1objs.c, src/type1/t1objs.h, src/type42/t42drivr.c,
src/type42/t42objs.c, src/type42/t42objs.h, src/winfonts/winfnt.c:
Update to new size selection interface.
* src/cff/cffdrivr.c, src/cff/cffgload.c, src/cff/cffobjs.c,
src/cff/cffobjs.h, src/truetype/ttdriver.c, src/truetype/ttgload.c,
src/truetype/ttobjs.c, src/truetype/ttobjs.h: Update to new size
selection interface.
Make `strike_index' FT_ULong and always defined.
Use `load_strike_metrics' provided by SFNT interface.
- fixing apinames.c, adding support for Watcom and Borland compilers
- adding generation of exported symbols list to the build system, including the Unix one !!
sorry Werner, I have no time to document this in ChangeLog at the moment
ttmetrics here. TrueType fonts with only sbits always have
ttmetrics.valid set to false.
* src/truetype/ttgload.c (TT_Load_Glyph): Check that ttmetrics is
valid before loading outline glyph.
* src/cache/ftcimage.c (FTC_INode_New): Fix a memory leak.
TT_Done_Context and rename TT_Destroy_Context to TT_Done_Context with
slight changes.
* src/truetype/ttobjs.h (tt_slot_init): New function.
* src/truetype/ttobjs.c (tt_driver_init): Initialize execution context
here.
(tt_slot_init): New function to create extra points for the internal
glyph loader. We then use it directly, instead of face's glyph loader,
when loading glyph.
* src/truetype/ttdriver.c: Use tt_slot_init for glyph slot
initialization.
(TT_Load_Glyph): Load flag dependencies are handled here.
Return error if size is NULL.
* src/truetype/ttgload.c: Heavy cleanup and refactor.
(TT_Hint_Glyph): New funcion to hint a zone, prepared by caller.
(TT_Process_Simple_Glyph): Use loader->pp's instead of recalculating.
Use TT_Hint_Glyph.
No need to save/restore loader->stream before/after
TT_Vary_Get_Glyph_Deltas now.
(TT_LOADER_SET_PP): New macro to calculate and set the four phantom points.
(TT_Process_Composite_Component, TT_Process_Composite_Glyph,
load_truetype_glyph): Refactor load_truetype_glyph into these three functions.
Never set exec->glyphSize to 0. (close#13107)
Forget glyph frame before calling TT_Process_Simple_Glyph.
(TT_Load_Glyph, load_sbit_image, tt_loader_init): Refactor TT_Load_Glyph into
these three functions.
Set various fields of `glyph' here, instead of in load_truetype_glyph
and compute_glyph_metrics.
removing compiler warnings (Visual C++ /W4)
* Jamfile, src/otvalid/{otvcommn.h,otvgdef.c,otvgpos.c,otvgsub.c,
otvjstf.c}: modified the code to use a different pre-processor trick
to implement the OTV_NEST1, OTV_NEST2 and OTV_NEST3 macros. The code
now compiles neatly with Visual C++. The 'otvalid' module has been
put in the Jamfile build again.
ft_glyphslot_grid_fit_metrics.
* src/truetype/ttgload.c (compute_glyph_metrics): Use
ft_glyphslot_grid_fit_metrics.
* src/cff/cffgload.c (cff_slot_load), src/cid/cidgload.c
(cid_slot_load_glyph), src/type1/t1gload.c (T1_Load_Glyph): Use
ft_glyphslot_grid_fit_metrics.
FT_Outline_Get_CBox is called twice.
* src/base/ftsynth.c (FT_GlyphSlot_Embolden): Modify metrics to more
reasonable values when emboldening outline glyphs. The theoretic
ones are unrealistic.
halved.
* src/base/ftsynth.c (FT_GlyphSlot_Embolden): Change the default
strength.
Don't increase slot->advance.y.
* include/freetype/freetype.h (FREETYPE_MINOR): Set to 2.
(FREETYPE_PATCH): Set to 0.
* builds/unix/configure.ac (version_info): Set to 9:9:3.
Currently, we are still binary compatible.
* builds/win32/visualc/index.html,
builds/win32/visualc/freetype.dsp,
builds/win32/visualc/freetype.vcproj: s/219/2110/, s/2.1.9/2.1.10/.
* builds/freetype.mk (refdoc), README, Jamfile (RefDoc):
s/2.1.9/2.1.10/.
* docs/CHANGES, docs/VERSION.DLL: Updated.
* ChangeLog: Split off older entries into...
* ChangeLog.20, ChangeLog.21: These new files.
The next release will be 2.2.0, so don't worry about source code
backwards compatibility.
* include/freetype/ftimage.h (FT_Outline_MoveToFunc,
FT_Outline_LineToFunc, FT_Outline_ConicToFunc,
FT_Outline_CubicToFunc, FT_SpanFunc, FT_Raster_RenderFunc),
include/freetype/ftrender.h (FT_Glyph_TransformFunc,
FT_Renderer_RenderFunc, FT_Renderer_TransformFunc): Decorate
parameters with `const' where appropriate.
* src/sfnt/ttsbit.c (tt_face_load_sbit_image): Compute vertBearingY
to make glyphs centered vertically.
* src/truetype/ttgload.c (compute_glyph_metrics): Compute
vertBearingY to make glyphs centered vertically.
Fix some bugs in vertical metrics:
. loader->pp3.y and loader->pp4.y are in 26.6 format, not in font
units.
. As we use the glyph's cbox to calculate the top bearing now
there iss no need to adjust `top'.
* src/otvalid/otvcommn.h (OTV_OPTIONAL_TABLE): Use FT_UShort to be
in sync with OTV_OPTIONAL_OFFSET. Reported by YAMATO Masatake.
* docs/release: Update.
* src/autofit/*: Add copyright messages.
Formatting.
* src/autofit/afhints.c (af_glyph_hints_done): Don't use
`AF_Dimension' but `int' for loop counter.
* src/autofit/aflatin.c (af_latin_metrics_init_widths): Don't use
`AF_Dimension' but `int' for loop counter.
Use proper enumeration value for `render_mode'.
(af_latin_metrics_scale_dim): Don't shadow variables.
(af_latin_hints_compute_segments): Use proper cast for `major_dir'
and `segment_dir'.
(af_latin_align_linked_edge, af_latin_hint_edges): Fix arguments of call to
`af_latin_compute_stem_width'.
(af_latin_hints_apply): Don't use `AF_Dimension' but `int' for loop
counter.
* src/base/ftdbgmem.c (ft_mem_table_get_source, FT_DumpMemory): Use
proper cast for memory allocation.
* src/cff/cffdrivr.c (cff_get_kerning): Use proper cast for
initialization of `sfnt'.
* src/sfnt/sfdriver.c: Include `ttkern.h'.
* src/sfnt/ttkern.c (tt_face_get_kerning): Don't shadow variables.
* src/truetype/ttgload.c: Include `ttpload.h'.
* src/truetype/ttpload.c (tt_face_load_loca) [FT_OPTIMIZE_MEMORY]:
Remove redundant variable.
drastically reduce the heap usage of FreeType, especially in the case
of memory-mapped files. The idea is to avoid loading and decoding tables
in the heap, and instead access the raw data whenever possible (i.e.
when it doesn't compromise performance).
This had several impacts: first, opening vera.ttf uses a ridiculous amount
of memory (when the FT_Library footprint is accounted for), until you start
loading glyphs. Even then, you'll save at least 20 Kb compared to the non
optimized case. performance of various operations, including open/close
has also been dramatically improved.
More optimisations to come. The auto-hinter eats memory like crazy? This
must be stopped...
src/truetype/ttpload.c, src/truetype/ttpload.h, src/truetype/ttgload.c:
added the temporary configuration FT_OPTIMIZE_MEMORY to control various
optimizations used to reduce the heap footprint of memory-mapped TrueType
files.
* src/truetype/ttpload.c (tt_face_load_loca, tt_face_get_location,
tt_face_done_loca): when FT_OPTIMIZE_MEMORY is set, the locations table
is read directly from memory-mapped streams, instead of being decoded
into the heap.
* src/truetype/ttpload.c: only load the CVT and fpgm tables when the
bytecode interpreter is compiled in.
* src/truetype/ttgload.c (TT_Process_Simple_Glyph): Don't shift
points vertically before hinting.
* docs/CHANGES: Updated.
* src/cache/ftcglyph.c (FTC_GNode_UnselectFamily,
FTC_GCache_Lookup): A new try to fix comparison with zero.
overflow.
* src/sfnt/ttload.c (tt_face_load_hdmx): Protect against bad input.
Don't use FT_QNEW_ARRAY but FT_NEW_ARRAY to make deallocation work
in case of failure.
* src/sfnt/ttsbit.c (Load_SBit_Range): Check range intervals.
(tt_face_load_sbit_strikes): Allocate `strike_sbit_ranges' after
frame test.
* src/truetype/ttgload.c (TTLoad_Simple_Glyph): Add assertion for
`flag'.