This concludes the changes needed to add Type 1 support.
* src/psaux/psintrp.c: Update includes.
(cf2_interpT2CharString) <cf2_escSEAC>: Implement this similarly to
implied seac for CFF.
* src/psaux/t1decode.c (t1_lookup_glyph_by_stdcharcode_ps): New
function to look up the glyph index.
* src/psaux/psft.c (cf2_getT1SeacComponent,
cf2_freeT1SeacComponent): New functions to get the charstrings for
seac components.
* src/psaux/t1decode.h, src/psaux/psft.h: Update declarations.
Move `CF2_Font' instance to `PS_Decoder'. This is the context for
the interpreter and since it is currently stored in `CFF_Font', is
unavailable in Type 1 mode.
* include/freetype/internal/psaux.h (T1_Decoder, PS_Decoder): New
`cf2_instance' field.
* src/psaux/psdecode.c (ps_decoder_init): Copy `cf2_instance' to
`PS_Decoder'.
* src/psaux/t1decode.c (t1_decoder_done): Add finalization code.
* src/psaux/psft.c (cf2_decoder_parse_charstrings): Update accesses.
Reported as
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=935
* src/psaux/t1decode.c (t1_decoder_parse_charstrings)
<callothersubr>: Since there is not a single flex operator but a
series of subroutine calls, malformed fonts can call arbitrary other
operators after the start of a flex, possibly adding points. For
this reason we have to check the available number of points before
inserting a point.
* src/cff/cffload.c (FT_fdot14ToFixed): Fix casting.
(cff_blend_doBlend): Don't left-shift negative numbers.
Handle 5-byte numbers byte by byte to avoid alignment issues.
* src/cff/cffparse.c (cff_parse): Handle 5-byte numbers byte by byte
to avoid alignment issues.
* src/cid/cidload (cid_read_subrs): Do nothing if we don't have any
subrs.
* src/psaux/t1decode.c (t1_decode_parse_charstring): Fix tracing.
* src/tools/glnames.py (main): Put `DEFINE_PSTABLES' guard around
definition of `ft_get_adobe_glyph_index'.
* src/psnames/pstables.h: Regenerated.
* src/psnames/psmodule.c: Inlude `pstables.h' twice to get both
declaration and definition.
* src/truetype/ttgxvar.c (FT_fdot14ToFixed, FT_intToFixed): Fix
casting.
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.
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.
The functions in this patch *do* return non-trivial errors that must
be taken care of.
* src/autofit/afloader.c (af_loader_load_g), src/base/ftobjs.c
(FT_Render_Glyph_Internal), src/base/ftoutln.c (FT_Outline_Render),
src/cff/cffgload.c (cff_decoder_parse_charstrings) <cff_op_endchar>,
src/psaux/psobjs.c (ps_parser_load_field_table), src/psaux/t1decode
(t1_decoder_parse_charstrings) <op_endchar>, src/truetype/ttgload.c
(load_truetype_glyph <subglyph loop>, tt_loader_init,
TT_Load_Glyph), src/truetype/ttgxvar.c (TT_Set_MM_Blend),
src/truetype/ttobjs.c (tt_size_run_fpgm, tt_size_run_prep): Do it.
* src/base/ftcalc.c (FT_DivFix): Use unsigned values for
computations which use the left shift operator and convert to signed
as the last step.
* src/base/fttrigon.c (ft_trig_prenorm, FT_Vector_Rotate,
FT_Vector_Length, FT_Vector_Polarize): Ditto.
* src/cff/cffgload.c (cff_decoder_parse_charstrings): Simplify.
* src/cff/cffload.c (cff_subfont_load): Fix constant.
* src/cff/cffparse.c (cff_parse_integer, cff_parse_real, do_fixed,
cff_parse_fixed_dynamic): Use unsigned values for computations which
use the left shift operator and convert to signed as the last step.
* src/cid/cidload.c (cid_get_offset): Ditto.
* src/psaux/psconv.c (PS_Conv_ToFixed): Ditto.
* src/psaux/t1decode.c (t1_decoder_parse_charstrings): Ditto.
* src/truetype/ttinterp.c (TT_MulFix14, TT_DotFix14): Ditto.
This is essentially a mechanical conversion, adding inclusion of
`FT_INTERNAL_DEBUG_H' where necessary, and providing the macros for
stand-alone compiling modes of the rasterizer modules.
To convert the remaining occurrences of FT_Err_XXX and friends it is
necessary to rewrite the code. Note, however, that it doesn't harm
if some cases are not handled since FT_THROW is a no-op.
* include/freetype/config/ftstdlib.h (FT_USHORT_MAX): New macro.
* src/base/ftbitmap.c (FT_Bitmap_Convert): Protect against invalid
value of `target->rows'.
* src/psaux/t1decode.c (t1_decoder_parse_charstrings): Add check for
flex start.
* src/raster/ftrend1.c (ft_raster1_render): Check `width' and
`height'.
* src/truetype/ttgxvar.c (TT_Vary_Get_Glyph_Deltas): Protect against
invalid values in `localpoints' array.
Found with font fuzzying.
* src/psaux/t1decode.c (t1_decoder_parse_charstrings): Check
`decoder->buildchar'.
* src/type1/t1load.c (t1_load_keyword): Check `blend->num_designs'.
When shifting right a signed value, it is not defined by the
C standard whether one gets a sign extension or not. Use a macro to
do an explicit cast from a signed short (assuming that this is
16bit) to an int.
* src/psaux/t1decode.c (Fix2Int): New macro.
Use it where appropriate.
start_point, check_points, add_point, add_point1, close_contour
macros.
Remove add_contour macro.
Return error code from t1_builder_start_point and
t1_builder_check_points when there was one (instead of returning 0).
Previously we discovered fonts which used `setcurrentpoint' to set
the initial point of a contour to 0,0. This caused FreeType to
raise an error, because the `setcurrentpoint' operator is only
supposed to be used with the results from an OtherSubr subroutine.
This was fixed by simply ignoring the error and carrying on.
Now we have found a font which uses setcurrentpoint to actually
establish a non-zero point for a contour during the course of a
glyph program. FWIW, these files may be produced by an application
called `Intaglio' on the Mac, when converting TrueType fonts to
Type 1.
The fix allows the new invalid behaviour, the old invalid behaviour
and real proper usage of the operator to work the same way as Adobe
interpreters apparently do.
(t1_decoder_parse_charstrings): Make `setcurrentpoint' use the top
two elements of the stack to establish unconditionally the current x
and y coordinates.
Make the `flex' subroutine handling (OtherSubr 0) put the current
x,y coordinates onto the stack, instead of two dummy uninitialised
values.
Issue #28226 involved a work-around for a font which used the
`setcurrentpoint' operator in an invalid way; this operator is only
supposed to be used with the result of OtherSubrs, and the font used
it directly. The supplied patch removed the block of code which
checked this usage entirely.
This turns out to be a Bad Thing. If `setcurrentpoint' is being
used correctly it should reset the flex flag in the decoder. If we
don't do this then the flag never gets reset and we omit any further
contours from the glyph (at least until we close the path or
similar).
* src/psaux/t1decode.c (t1_decoder_parse_charstrings)
<op_setcurrentpoint>: Handle `flex_state' correctly.
This fixes Savannah bug #28226.
At least two wild PostScript files of unknown provenance contain
Type 1 fonts, apparently converted from TrueType fonts in earlier
PDF versions of the files, which use the `setcurrentpoint' operator
inappropriately.
FreeType currently throws an error in this case, but Ghostscript and
Adobe Distiller both accept the fonts and ignore the problem. This
commit #ifdefs out the check so PostScript interpreters using
FreeType can render these files.
The specification says `setcurrentpoint' should only be used to set
the point after a `Subr' call, but these fonts use it to set the
initial point to (0,0). Unnecessarily so, as they correctly use an
`hsbw' operation which implicitly sets the initial point.
