ref: 12ef831fc314518bff45278008a568608501a8e4
dir: /src/psaux/cffdecode.c/
/****************************************************************************
*
* cffdecode.c
*
* PostScript CFF (Type 2) decoding routines (body).
*
* Copyright (C) 2017-2021 by
* David Turner, Robert Wilhelm, and Werner Lemberg.
*
* This file is part of the FreeType project, and may only be used,
* modified, and distributed under the terms of the FreeType project
* license, LICENSE.TXT. By continuing to use, modify, or distribute
* this file you indicate that you have read the license and
* understand and accept it fully.
*
*/
#include <freetype/freetype.h>
#include <freetype/internal/ftdebug.h>
#include <freetype/internal/ftserv.h>
#include <freetype/internal/services/svcfftl.h>
#include "cffdecode.h"
#include "psobjs.h"
#include "psauxerr.h"
/**************************************************************************
*
* The macro FT_COMPONENT is used in trace mode. It is an implicit
* parameter of the FT_TRACE() and FT_ERROR() macros, used to print/log
* messages during execution.
*/
#undef FT_COMPONENT
#define FT_COMPONENT cffdecode
#ifdef CFF_CONFIG_OPTION_OLD_ENGINE
typedef enum CFF_Operator_
{
cff_op_unknown = 0,
cff_op_rmoveto,
cff_op_hmoveto,
cff_op_vmoveto,
cff_op_rlineto,
cff_op_hlineto,
cff_op_vlineto,
cff_op_rrcurveto,
cff_op_hhcurveto,
cff_op_hvcurveto,
cff_op_rcurveline,
cff_op_rlinecurve,
cff_op_vhcurveto,
cff_op_vvcurveto,
cff_op_flex,
cff_op_hflex,
cff_op_hflex1,
cff_op_flex1,
cff_op_endchar,
cff_op_hstem,
cff_op_vstem,
cff_op_hstemhm,
cff_op_vstemhm,
cff_op_hintmask,
cff_op_cntrmask,
cff_op_dotsection, /* deprecated, acts as no-op */
cff_op_abs,
cff_op_add,
cff_op_sub,
cff_op_div,
cff_op_neg,
cff_op_random,
cff_op_mul,
cff_op_sqrt,
cff_op_blend,
cff_op_drop,
cff_op_exch,
cff_op_index,
cff_op_roll,
cff_op_dup,
cff_op_put,
cff_op_get,
cff_op_store,
cff_op_load,
cff_op_and,
cff_op_or,
cff_op_not,
cff_op_eq,
cff_op_ifelse,
cff_op_callsubr,
cff_op_callgsubr,
cff_op_return,
/* Type 1 opcodes: invalid but seen in real life */
cff_op_hsbw,
cff_op_closepath,
cff_op_callothersubr,
cff_op_pop,
cff_op_seac,
cff_op_sbw,
cff_op_setcurrentpoint,
/* do not remove */
cff_op_max
} CFF_Operator;
#define CFF_COUNT_CHECK_WIDTH 0x80
#define CFF_COUNT_EXACT 0x40
#define CFF_COUNT_CLEAR_STACK 0x20
/* count values which have the `CFF_COUNT_CHECK_WIDTH' flag set are */
/* used for checking the width and requested numbers of arguments */
/* only; they are set to zero afterwards */
/* the other two flags are informative only and unused currently */
static const FT_Byte cff_argument_counts[] =
{
0, /* unknown */
2 | CFF_COUNT_CHECK_WIDTH | CFF_COUNT_EXACT, /* rmoveto */
1 | CFF_COUNT_CHECK_WIDTH | CFF_COUNT_EXACT,
1 | CFF_COUNT_CHECK_WIDTH | CFF_COUNT_EXACT,
0 | CFF_COUNT_CLEAR_STACK, /* rlineto */
0 | CFF_COUNT_CLEAR_STACK,
0 | CFF_COUNT_CLEAR_STACK,
0 | CFF_COUNT_CLEAR_STACK, /* rrcurveto */
0 | CFF_COUNT_CLEAR_STACK,
0 | CFF_COUNT_CLEAR_STACK,
0 | CFF_COUNT_CLEAR_STACK,
0 | CFF_COUNT_CLEAR_STACK,
0 | CFF_COUNT_CLEAR_STACK,
0 | CFF_COUNT_CLEAR_STACK,
13, /* flex */
7,
9,
11,
0 | CFF_COUNT_CHECK_WIDTH, /* endchar */
2 | CFF_COUNT_CHECK_WIDTH, /* hstem */
2 | CFF_COUNT_CHECK_WIDTH,
2 | CFF_COUNT_CHECK_WIDTH,
2 | CFF_COUNT_CHECK_WIDTH,
0 | CFF_COUNT_CHECK_WIDTH, /* hintmask */
0 | CFF_COUNT_CHECK_WIDTH, /* cntrmask */
0, /* dotsection */
1, /* abs */
2,
2,
2,
1,
0,
2,
1,
1, /* blend */
1, /* drop */
2,
1,
2,
1,
2, /* put */
1,
4,
3,
2, /* and */
2,
1,
2,
4,
1, /* callsubr */
1,
0,
2, /* hsbw */
0,
0,
0,
5, /* seac */
4, /* sbw */
2 /* setcurrentpoint */
};
static FT_Error
cff_operator_seac( CFF_Decoder* decoder,
FT_Pos asb,
FT_Pos adx,
FT_Pos ady,
FT_Int bchar,
FT_Int achar )
{
FT_Error error;
CFF_Builder* builder = &decoder->builder;
FT_Int bchar_index, achar_index;
TT_Face face = decoder->builder.face;
FT_Vector left_bearing, advance;
FT_Byte* charstring;
FT_ULong charstring_len;
FT_Pos glyph_width;
if ( decoder->seac )
{
FT_ERROR(( "cff_operator_seac: invalid nested seac\n" ));
return FT_THROW( Syntax_Error );
}
adx = ADD_LONG( adx, decoder->builder.left_bearing.x );
ady = ADD_LONG( ady, decoder->builder.left_bearing.y );
#ifdef FT_CONFIG_OPTION_INCREMENTAL
/* Incremental fonts don't necessarily have valid charsets. */
/* They use the character code, not the glyph index, in this case. */
if ( face->root.internal->incremental_interface )
{
bchar_index = bchar;
achar_index = achar;
}
else
#endif /* FT_CONFIG_OPTION_INCREMENTAL */
{
CFF_Font cff = (CFF_Font)(face->extra.data);
bchar_index = cff_lookup_glyph_by_stdcharcode( cff, bchar );
achar_index = cff_lookup_glyph_by_stdcharcode( cff, achar );
}
if ( bchar_index < 0 || achar_index < 0 )
{
FT_ERROR(( "cff_operator_seac:"
" invalid seac character code arguments\n" ));
return FT_THROW( Syntax_Error );
}
/* If we are trying to load a composite glyph, do not load the */
/* accent character and return the array of subglyphs. */
if ( builder->no_recurse )
{
FT_GlyphSlot glyph = (FT_GlyphSlot)builder->glyph;
FT_GlyphLoader loader = glyph->internal->loader;
FT_SubGlyph subg;
/* reallocate subglyph array if necessary */
error = FT_GlyphLoader_CheckSubGlyphs( loader, 2 );
if ( error )
goto Exit;
subg = loader->current.subglyphs;
/* subglyph 0 = base character */
subg->index = bchar_index;
subg->flags = FT_SUBGLYPH_FLAG_ARGS_ARE_XY_VALUES |
FT_SUBGLYPH_FLAG_USE_MY_METRICS;
subg->arg1 = 0;
subg->arg2 = 0;
subg++;
/* subglyph 1 = accent character */
subg->index = achar_index;
subg->flags = FT_SUBGLYPH_FLAG_ARGS_ARE_XY_VALUES;
subg->arg1 = (FT_Int)( adx >> 16 );
subg->arg2 = (FT_Int)( ady >> 16 );
/* set up remaining glyph fields */
glyph->num_subglyphs = 2;
glyph->subglyphs = loader->base.subglyphs;
glyph->format = FT_GLYPH_FORMAT_COMPOSITE;
loader->current.num_subglyphs = 2;
}
FT_GlyphLoader_Prepare( builder->loader );
/* First load `bchar' in builder */
error = decoder->get_glyph_callback( face, (FT_UInt)bchar_index,
&charstring, &charstring_len );
if ( !error )
{
/* the seac operator must not be nested */
decoder->seac = TRUE;
error = cff_decoder_parse_charstrings( decoder, charstring,
charstring_len, 0 );
decoder->seac = FALSE;
decoder->free_glyph_callback( face, &charstring, charstring_len );
if ( error )
goto Exit;
}
/* Save the left bearing, advance and glyph width of the base */
/* character as they will be erased by the next load. */
left_bearing = builder->left_bearing;
advance = builder->advance;
glyph_width = decoder->glyph_width;
builder->left_bearing.x = 0;
builder->left_bearing.y = 0;
builder->pos_x = SUB_LONG( adx, asb );
builder->pos_y = ady;
/* Now load `achar' on top of the base outline. */
error = decoder->get_glyph_callback( face, (FT_UInt)achar_index,
&charstring, &charstring_len );
if ( !error )
{
/* the seac operator must not be nested */
decoder->seac = TRUE;
error = cff_decoder_parse_charstrings( decoder, charstring,
charstring_len, 0 );
decoder->seac = FALSE;
decoder->free_glyph_callback( face, &charstring, charstring_len );
if ( error )
goto Exit;
}
/* Restore the left side bearing, advance and glyph width */
/* of the base character. */
builder->left_bearing = left_bearing;
builder->advance = advance;
decoder->glyph_width = glyph_width;
builder->pos_x = 0;
builder->pos_y = 0;
Exit:
return error;
}
#endif /* CFF_CONFIG_OPTION_OLD_ENGINE */
/*************************************************************************/
/*************************************************************************/
/*************************************************************************/
/********** *********/
/********** *********/
/********** GENERIC CHARSTRING PARSING *********/
/********** *********/
/********** *********/
/*************************************************************************/
/*************************************************************************/
/*************************************************************************/
/**************************************************************************
*
* @Function:
* cff_compute_bias
*
* @Description:
* Computes the bias value in dependence of the number of glyph
* subroutines.
*
* @Input:
* in_charstring_type ::
* The `CharstringType' value of the top DICT
* dictionary.
*
* num_subrs ::
* The number of glyph subroutines.
*
* @Return:
* The bias value.
*/
static FT_Int
cff_compute_bias( FT_Int in_charstring_type,
FT_UInt num_subrs )
{
FT_Int result;
if ( in_charstring_type == 1 )
result = 0;
else if ( num_subrs < 1240 )
result = 107;
else if ( num_subrs < 33900U )
result = 1131;
else
result = 32768U;
return result;
}
FT_LOCAL_DEF( FT_Int )
cff_lookup_glyph_by_stdcharcode( CFF_Font cff,
FT_Int charcode )
{
FT_UInt n;
FT_UShort glyph_sid;
FT_Service_CFFLoad cffload;
/* CID-keyed fonts don't have glyph names */
if ( !cff->charset.sids )
return -1;
/* check range of standard char code */
if ( charcode < 0 || charcode > 255 )
return -1;
#if 0
/* retrieve cffload from list of current modules */
FT_Service_CFFLoad cffload;
FT_FACE_FIND_GLOBAL_SERVICE( face, cffload, CFF_LOAD );
if ( !cffload )
{
FT_ERROR(( "cff_lookup_glyph_by_stdcharcode:"
" the `cffload' module is not available\n" ));
return FT_THROW( Unimplemented_Feature );
}
#endif
cffload = (FT_Service_CFFLoad)cff->cffload;
/* Get code to SID mapping from `cff_standard_encoding'. */
glyph_sid = cffload->get_standard_encoding( (FT_UInt)charcode );
for ( n = 0; n < cff->num_glyphs; n++ )
{
if ( cff->charset.sids[n] == glyph_sid )
return (FT_Int)n;
}
return -1;
}
#ifdef CFF_CONFIG_OPTION_OLD_ENGINE
/**************************************************************************
*
* @Function:
* cff_decoder_parse_charstrings
*
* @Description:
* Parses a given Type 2 charstrings program.
*
* @InOut:
* decoder ::
* The current Type 1 decoder.
*
* @Input:
* charstring_base ::
* The base of the charstring stream.
*
* charstring_len ::
* The length in bytes of the charstring stream.
*
* in_dict ::
* Set to 1 if function is called from top or
* private DICT (needed for Multiple Master CFFs).
*
* @Return:
* FreeType error code. 0 means success.
*/
FT_LOCAL_DEF( FT_Error )
cff_decoder_parse_charstrings( CFF_Decoder* decoder,
FT_Byte* charstring_base,
FT_ULong charstring_len,
FT_Bool in_dict )
{
FT_Error error;
CFF_Decoder_Zone* zone;
FT_Byte* ip;
FT_Byte* limit;
CFF_Builder* builder = &decoder->builder;
FT_Pos x, y;
FT_Fixed* stack;
FT_Int charstring_type =
decoder->cff->top_font.font_dict.charstring_type;
FT_UShort num_designs =
decoder->cff->top_font.font_dict.num_designs;
FT_UShort num_axes =
decoder->cff->top_font.font_dict.num_axes;
T2_Hints_Funcs hinter;
/* set default width */
decoder->num_hints = 0;
decoder->read_width = 1;
/* initialize the decoder */
decoder->top = decoder->stack;
decoder->zone = decoder->zones;
zone = decoder->zones;
stack = decoder->top;
hinter = (T2_Hints_Funcs)builder->hints_funcs;
builder->path_begun = 0;
if ( !charstring_base )
return FT_Err_Ok;
zone->base = charstring_base;
limit = zone->limit = charstring_base + charstring_len;
ip = zone->cursor = zone->base;
error = FT_Err_Ok;
x = builder->pos_x;
y = builder->pos_y;
/* begin hints recording session, if any */
if ( hinter )
hinter->open( hinter->hints );
/* now execute loop */
while ( ip < limit )
{
CFF_Operator op;
FT_Byte v;
/*********************************************************************
*
* Decode operator or operand
*/
v = *ip++;
if ( v >= 32 || v == 28 )
{
FT_Int shift = 16;
FT_Int32 val;
/* this is an operand, push it on the stack */
/* if we use shifts, all computations are done with unsigned */
/* values; the conversion to a signed value is the last step */
if ( v == 28 )
{
if ( ip + 1 >= limit )
goto Syntax_Error;
val = (FT_Short)( ( (FT_UShort)ip[0] << 8 ) | ip[1] );
ip += 2;
}
else if ( v < 247 )
val = (FT_Int32)v - 139;
else if ( v < 251 )
{
if ( ip >= limit )
goto Syntax_Error;
val = ( (FT_Int32)v - 247 ) * 256 + *ip++ + 108;
}
else if ( v < 255 )
{
if ( ip >= limit )
goto Syntax_Error;
val = -( (FT_Int32)v - 251 ) * 256 - *ip++ - 108;
}
else
{
if ( ip + 3 >= limit )
goto Syntax_Error;
val = (FT_Int32)( ( (FT_UInt32)ip[0] << 24 ) |
( (FT_UInt32)ip[1] << 16 ) |
( (FT_UInt32)ip[2] << 8 ) |
(FT_UInt32)ip[3] );
ip += 4;
if ( charstring_type == 2 )
shift = 0;
}
if ( decoder->top - stack >= CFF_MAX_OPERANDS )
goto Stack_Overflow;
val = (FT_Int32)( (FT_UInt32)val << shift );
*decoder->top++ = val;
#ifdef FT_DEBUG_LEVEL_TRACE
if ( !( val & 0xFFFFL ) )
FT_TRACE4(( " %hd", (FT_Short)( (FT_UInt32)val >> 16 ) ));
else
FT_TRACE4(( " %.5f", val / 65536.0 ));
#endif
}
else
{
/* The specification says that normally arguments are to be taken */
/* from the bottom of the stack. However, this seems not to be */
/* correct, at least for Acroread 7.0.8 on GNU/Linux: It pops the */
/* arguments similar to a PS interpreter. */
FT_Fixed* args = decoder->top;
FT_Int num_args = (FT_Int)( args - decoder->stack );
FT_Int req_args;
/* find operator */
op = cff_op_unknown;
switch ( v )
{
case 1:
op = cff_op_hstem;
break;
case 3:
op = cff_op_vstem;
break;
case 4:
op = cff_op_vmoveto;
break;
case 5:
op = cff_op_rlineto;
break;
case 6:
op = cff_op_hlineto;
break;
case 7:
op = cff_op_vlineto;
break;
case 8:
op = cff_op_rrcurveto;
break;
case 9:
op = cff_op_closepath;
break;
case 10:
op = cff_op_callsubr;
break;
case 11:
op = cff_op_return;
break;
case 12:
if ( ip >= limit )
goto Syntax_Error;
v = *ip++;
switch ( v )
{
case 0:
op = cff_op_dotsection;
break;
case 1: /* this is actually the Type1 vstem3 operator */
op = cff_op_vstem;
break;
case 2: /* this is actually the Type1 hstem3 operator */
op = cff_op_hstem;
break;
case 3:
op = cff_op_and;
break;
case 4:
op = cff_op_or;
break;
case 5:
op = cff_op_not;
break;
case 6:
op = cff_op_seac;
break;
case 7:
op = cff_op_sbw;
break;
case 8:
op = cff_op_store;
break;
case 9:
op = cff_op_abs;
break;
case 10:
op = cff_op_add;
break;
case 11:
op = cff_op_sub;
break;
case 12:
op = cff_op_div;
break;
case 13:
op = cff_op_load;
break;
case 14:
op = cff_op_neg;
break;
case 15:
op = cff_op_eq;
break;
case 16:
op = cff_op_callothersubr;
break;
case 17:
op = cff_op_pop;
break;
case 18:
op = cff_op_drop;
break;
case 20:
op = cff_op_put;
break;
case 21:
op = cff_op_get;
break;
case 22:
op = cff_op_ifelse;
break;
case 23:
op = cff_op_random;
break;
case 24:
op = cff_op_mul;
break;
case 26:
op = cff_op_sqrt;
break;
case 27:
op = cff_op_dup;
break;
case 28:
op = cff_op_exch;
break;
case 29:
op = cff_op_index;
break;
case 30:
op = cff_op_roll;
break;
case 33:
op = cff_op_setcurrentpoint;
break;
case 34:
op = cff_op_hflex;
break;
case 35:
op = cff_op_flex;
break;
case 36:
op = cff_op_hflex1;
break;
case 37:
op = cff_op_flex1;
break;
default:
FT_TRACE4(( " unknown op (12, %d)\n", v ));
break;
}
break;
case 13:
op = cff_op_hsbw;
break;
case 14:
op = cff_op_endchar;
break;
case 16:
op = cff_op_blend;
break;
case 18:
op = cff_op_hstemhm;
break;
case 19:
op = cff_op_hintmask;
break;
case 20:
op = cff_op_cntrmask;
break;
case 21:
op = cff_op_rmoveto;
break;
case 22:
op = cff_op_hmoveto;
break;
case 23:
op = cff_op_vstemhm;
break;
case 24:
op = cff_op_rcurveline;
break;
case 25:
op = cff_op_rlinecurve;
break;
case 26:
op = cff_op_vvcurveto;
break;
case 27:
op = cff_op_hhcurveto;
break;
case 29:
op = cff_op_callgsubr;
break;
case 30:
op = cff_op_vhcurveto;
break;
case 31:
op = cff_op_hvcurveto;
break;
default:
FT_TRACE4(( " unknown op (%d)\n", v ));
break;
}
if ( op == cff_op_unknown )
continue;
/* in Multiple Master CFFs, T2 charstrings can appear in */
/* dictionaries, but some operators are prohibited */
if ( in_dict )
{
switch ( op )
{
case cff_op_hstem:
case cff_op_vstem:
case cff_op_vmoveto:
case cff_op_rlineto:
case cff_op_hlineto:
case cff_op_vlineto:
case cff_op_rrcurveto:
case cff_op_hstemhm:
case cff_op_hintmask:
case cff_op_cntrmask:
case cff_op_rmoveto:
case cff_op_hmoveto:
case cff_op_vstemhm:
case cff_op_rcurveline:
case cff_op_rlinecurve:
case cff_op_vvcurveto:
case cff_op_hhcurveto:
case cff_op_vhcurveto:
case cff_op_hvcurveto:
case cff_op_hflex:
case cff_op_flex:
case cff_op_hflex1:
case cff_op_flex1:
case cff_op_callsubr:
case cff_op_callgsubr:
/* deprecated opcodes */
case cff_op_dotsection:
/* invalid Type 1 opcodes */
case cff_op_hsbw:
case cff_op_closepath:
case cff_op_callothersubr:
case cff_op_seac:
case cff_op_sbw:
case cff_op_setcurrentpoint:
goto MM_Error;
default:
break;
}
}
/* check arguments */
req_args = cff_argument_counts[op];
if ( req_args & CFF_COUNT_CHECK_WIDTH )
{
if ( num_args > 0 && decoder->read_width )
{
/* If `nominal_width' is non-zero, the number is really a */
/* difference against `nominal_width'. Else, the number here */
/* is truly a width, not a difference against `nominal_width'. */
/* If the font does not set `nominal_width', then */
/* `nominal_width' defaults to zero, and so we can set */
/* `glyph_width' to `nominal_width' plus number on the stack */
/* -- for either case. */
FT_Int set_width_ok;
switch ( op )
{
case cff_op_hmoveto:
case cff_op_vmoveto:
set_width_ok = num_args & 2;
break;
case cff_op_hstem:
case cff_op_vstem:
case cff_op_hstemhm:
case cff_op_vstemhm:
case cff_op_rmoveto:
case cff_op_hintmask:
case cff_op_cntrmask:
set_width_ok = num_args & 1;
break;
case cff_op_endchar:
/* If there is a width specified for endchar, we either have */
/* 1 argument or 5 arguments. We like to argue. */
set_width_ok = in_dict
? 0
: ( ( num_args == 5 ) || ( num_args == 1 ) );
break;
default:
set_width_ok = 0;
break;
}
if ( set_width_ok )
{
decoder->glyph_width = decoder->nominal_width +
( stack[0] >> 16 );
if ( decoder->width_only )
{
/* we only want the advance width; stop here */
break;
}
/* Consumed an argument. */
num_args--;
}
}
decoder->read_width = 0;
req_args = 0;
}
req_args &= 0x000F;
if ( num_args < req_args )
goto Stack_Underflow;
args -= req_args;
num_args -= req_args;
/* At this point, `args' points to the first argument of the */
/* operand in case `req_args' isn't zero. Otherwise, we have */
/* to adjust `args' manually. */
/* Note that we only pop arguments from the stack which we */
/* really need and can digest so that we can continue in case */
/* of superfluous stack elements. */
switch ( op )
{
case cff_op_hstem:
case cff_op_vstem:
case cff_op_hstemhm:
case cff_op_vstemhm:
/* the number of arguments is always even here */
FT_TRACE4(( "%s\n",
op == cff_op_hstem ? " hstem" :
( op == cff_op_vstem ? " vstem" :
( op == cff_op_hstemhm ? " hstemhm" : " vstemhm" ) ) ));
if ( hinter )
hinter->stems( hinter->hints,
( op == cff_op_hstem || op == cff_op_hstemhm ),
num_args / 2,
args - ( num_args & ~1 ) );
decoder->num_hints += num_args / 2;
args = stack;
break;
case cff_op_hintmask:
case cff_op_cntrmask:
FT_TRACE4(( "%s", op == cff_op_hintmask ? " hintmask"
: " cntrmask" ));
/* implement vstem when needed -- */
/* the specification doesn't say it, but this also works */
/* with the 'cntrmask' operator */
/* */
if ( num_args > 0 )
{
if ( hinter )
hinter->stems( hinter->hints,
0,
num_args / 2,
args - ( num_args & ~1 ) );
decoder->num_hints += num_args / 2;
}
/* In a valid charstring there must be at least one byte */
/* after `hintmask' or `cntrmask' (e.g., for a `return' */
/* instruction). Additionally, there must be space for */
/* `num_hints' bits. */
if ( ( ip + ( ( decoder->num_hints + 7 ) >> 3 ) ) >= limit )
goto Syntax_Error;
if ( hinter )
{
if ( op == cff_op_hintmask )
hinter->hintmask( hinter->hints,
(FT_UInt)builder->current->n_points,
(FT_UInt)decoder->num_hints,
ip );
else
hinter->counter( hinter->hints,
(FT_UInt)decoder->num_hints,
ip );
}
#ifdef FT_DEBUG_LEVEL_TRACE
{
FT_UInt maskbyte;
FT_TRACE4(( " (maskbytes:" ));
for ( maskbyte = 0;
maskbyte < (FT_UInt)( ( decoder->num_hints + 7 ) >> 3 );
maskbyte++, ip++ )
FT_TRACE4(( " 0x%02X", *ip ));
FT_TRACE4(( ")\n" ));
}
#else
ip += ( decoder->num_hints + 7 ) >> 3;
#endif
args = stack;
break;
case cff_op_rmoveto:
FT_TRACE4(( " rmoveto\n" ));
cff_builder_close_contour( builder );
builder->path_begun = 0;
x = ADD_LONG( x, args[-2] );
y = ADD_LONG( y, args[-1] );
args = stack;
break;
case cff_op_vmoveto:
FT_TRACE4(( " vmoveto\n" ));
cff_builder_close_contour( builder );
builder->path_begun = 0;
y = ADD_LONG( y, args[-1] );
args = stack;
break;
case cff_op_hmoveto:
FT_TRACE4(( " hmoveto\n" ));
cff_builder_close_contour( builder );
builder->path_begun = 0;
x = ADD_LONG( x, args[-1] );
args = stack;
break;
case cff_op_rlineto:
FT_TRACE4(( " rlineto\n" ));
if ( cff_builder_start_point( builder, x, y ) ||
cff_check_points( builder, num_args / 2 ) )
goto Fail;
if ( num_args < 2 )
goto Stack_Underflow;
args -= num_args & ~1;
while ( args < decoder->top )
{
x = ADD_LONG( x, args[0] );
y = ADD_LONG( y, args[1] );
cff_builder_add_point( builder, x, y, 1 );
args += 2;
}
args = stack;
break;
case cff_op_hlineto:
case cff_op_vlineto:
{
FT_Int phase = ( op == cff_op_hlineto );
FT_TRACE4(( "%s\n", op == cff_op_hlineto ? " hlineto"
: " vlineto" ));
if ( num_args < 0 )
goto Stack_Underflow;
/* there exist subsetted fonts (found in PDFs) */
/* which call `hlineto' without arguments */
if ( num_args == 0 )
break;
if ( cff_builder_start_point( builder, x, y ) ||
cff_check_points( builder, num_args ) )
goto Fail;
args = stack;
while ( args < decoder->top )
{
if ( phase )
x = ADD_LONG( x, args[0] );
else
y = ADD_LONG( y, args[0] );
if ( cff_builder_add_point1( builder, x, y ) )
goto Fail;
args++;
phase ^= 1;
}
args = stack;
}
break;
case cff_op_rrcurveto:
{
FT_Int nargs;
FT_TRACE4(( " rrcurveto\n" ));
if ( num_args < 6 )
goto Stack_Underflow;
nargs = num_args - num_args % 6;
if ( cff_builder_start_point( builder, x, y ) ||
cff_check_points( builder, nargs / 2 ) )
goto Fail;
args -= nargs;
while ( args < decoder->top )
{
x = ADD_LONG( x, args[0] );
y = ADD_LONG( y, args[1] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[2] );
y = ADD_LONG( y, args[3] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[4] );
y = ADD_LONG( y, args[5] );
cff_builder_add_point( builder, x, y, 1 );
args += 6;
}
args = stack;
}
break;
case cff_op_vvcurveto:
{
FT_Int nargs;
FT_TRACE4(( " vvcurveto\n" ));
if ( num_args < 4 )
goto Stack_Underflow;
/* if num_args isn't of the form 4n or 4n+1, */
/* we enforce it by clearing the second bit */
nargs = num_args & ~2;
if ( cff_builder_start_point( builder, x, y ) )
goto Fail;
args -= nargs;
if ( nargs & 1 )
{
x = ADD_LONG( x, args[0] );
args++;
nargs--;
}
if ( cff_check_points( builder, 3 * ( nargs / 4 ) ) )
goto Fail;
while ( args < decoder->top )
{
y = ADD_LONG( y, args[0] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[1] );
y = ADD_LONG( y, args[2] );
cff_builder_add_point( builder, x, y, 0 );
y = ADD_LONG( y, args[3] );
cff_builder_add_point( builder, x, y, 1 );
args += 4;
}
args = stack;
}
break;
case cff_op_hhcurveto:
{
FT_Int nargs;
FT_TRACE4(( " hhcurveto\n" ));
if ( num_args < 4 )
goto Stack_Underflow;
/* if num_args isn't of the form 4n or 4n+1, */
/* we enforce it by clearing the second bit */
nargs = num_args & ~2;
if ( cff_builder_start_point( builder, x, y ) )
goto Fail;
args -= nargs;
if ( nargs & 1 )
{
y = ADD_LONG( y, args[0] );
args++;
nargs--;
}
if ( cff_check_points( builder, 3 * ( nargs / 4 ) ) )
goto Fail;
while ( args < decoder->top )
{
x = ADD_LONG( x, args[0] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[1] );
y = ADD_LONG( y, args[2] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[3] );
cff_builder_add_point( builder, x, y, 1 );
args += 4;
}
args = stack;
}
break;
case cff_op_vhcurveto:
case cff_op_hvcurveto:
{
FT_Int phase;
FT_Int nargs;
FT_TRACE4(( "%s\n", op == cff_op_vhcurveto ? " vhcurveto"
: " hvcurveto" ));
if ( cff_builder_start_point( builder, x, y ) )
goto Fail;
if ( num_args < 4 )
goto Stack_Underflow;
/* if num_args isn't of the form 8n, 8n+1, 8n+4, or 8n+5, */
/* we enforce it by clearing the second bit */
nargs = num_args & ~2;
args -= nargs;
if ( cff_check_points( builder, ( nargs / 4 ) * 3 ) )
goto Stack_Underflow;
phase = ( op == cff_op_hvcurveto );
while ( nargs >= 4 )
{
nargs -= 4;
if ( phase )
{
x = ADD_LONG( x, args[0] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[1] );
y = ADD_LONG( y, args[2] );
cff_builder_add_point( builder, x, y, 0 );
y = ADD_LONG( y, args[3] );
if ( nargs == 1 )
x = ADD_LONG( x, args[4] );
cff_builder_add_point( builder, x, y, 1 );
}
else
{
y = ADD_LONG( y, args[0] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[1] );
y = ADD_LONG( y, args[2] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[3] );
if ( nargs == 1 )
y = ADD_LONG( y, args[4] );
cff_builder_add_point( builder, x, y, 1 );
}
args += 4;
phase ^= 1;
}
args = stack;
}
break;
case cff_op_rlinecurve:
{
FT_Int num_lines;
FT_Int nargs;
FT_TRACE4(( " rlinecurve\n" ));
if ( num_args < 8 )
goto Stack_Underflow;
nargs = num_args & ~1;
num_lines = ( nargs - 6 ) / 2;
if ( cff_builder_start_point( builder, x, y ) ||
cff_check_points( builder, num_lines + 3 ) )
goto Fail;
args -= nargs;
/* first, add the line segments */
while ( num_lines > 0 )
{
x = ADD_LONG( x, args[0] );
y = ADD_LONG( y, args[1] );
cff_builder_add_point( builder, x, y, 1 );
args += 2;
num_lines--;
}
/* then the curve */
x = ADD_LONG( x, args[0] );
y = ADD_LONG( y, args[1] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[2] );
y = ADD_LONG( y, args[3] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[4] );
y = ADD_LONG( y, args[5] );
cff_builder_add_point( builder, x, y, 1 );
args = stack;
}
break;
case cff_op_rcurveline:
{
FT_Int num_curves;
FT_Int nargs;
FT_TRACE4(( " rcurveline\n" ));
if ( num_args < 8 )
goto Stack_Underflow;
nargs = num_args - 2;
nargs = nargs - nargs % 6 + 2;
num_curves = ( nargs - 2 ) / 6;
if ( cff_builder_start_point( builder, x, y ) ||
cff_check_points( builder, num_curves * 3 + 2 ) )
goto Fail;
args -= nargs;
/* first, add the curves */
while ( num_curves > 0 )
{
x = ADD_LONG( x, args[0] );
y = ADD_LONG( y, args[1] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[2] );
y = ADD_LONG( y, args[3] );
cff_builder_add_point( builder, x, y, 0 );
x = ADD_LONG( x, args[4] );
y = ADD_LONG( y, args[5] );
cff_builder_add_point( builder, x, y, 1 );
args += 6;
num_curves--;
}
/* then the final line */
x = ADD_LONG( x, args[0] );
y = ADD_LONG( y, args[1] );
cff_builder_add_point( builder, x, y, 1 );
args = stack;
}
break;
case cff_op_hflex1:
{
FT_Pos start_y;
FT_TRACE4(( " hflex1\n" ));
/* adding five more points: 4 control points, 1 on-curve point */
/* -- make sure we have enough space for the start point if it */
/* needs to be added */
if ( cff_builder_start_point( builder, x, y ) ||
cff_check_points( builder, 6 ) )
goto Fail;
/* record the starting point's y position for later use */
start_y = y;
/* first control point */
x = ADD_LONG( x, args[0] );
y = ADD_LONG( y, args[1] );
cff_builder_add_point( builder, x, y, 0 );
/* second control point */
x = ADD_LONG( x, args[2] );
y = ADD_LONG( y, args[3] );
cff_builder_add_point( builder, x, y, 0 );
/* join point; on curve, with y-value the same as the last */
/* control point's y-value */
x = ADD_LONG( x, args[4] );
cff_builder_add_point( builder, x, y, 1 );
/* third control point, with y-value the same as the join */
/* point's y-value */
x = ADD_LONG( x, args[5] );
cff_builder_add_point( builder, x, y, 0 );
/* fourth control point */
x = ADD_LONG( x, args[6] );
y = ADD_LONG( y, args[7] );
cff_builder_add_point( builder, x, y, 0 );
/* ending point, with y-value the same as the start */
x = ADD_LONG( x, args[8] );
y = start_y;
cff_builder_add_point( builder, x, y, 1 );
args = stack;
break;
}
case cff_op_hflex:
{
FT_Pos start_y;
FT_TRACE4(( " hflex\n" ));
/* adding six more points; 4 control points, 2 on-curve points */
if ( cff_builder_start_point( builder, x, y ) ||
cff_check_points( builder, 6 ) )
goto Fail;
/* record the starting point's y-position for later use */
start_y = y;
/* first control point */
x = ADD_LONG( x, args[0] );
cff_builder_add_point( builder, x, y, 0 );
/* second control point */
x = ADD_LONG( x, args[1] );
y = ADD_LONG( y, args[2] );
cff_builder_add_point( builder, x, y, 0 );
/* join point; on curve, with y-value the same as the last */
/* control point's y-value */
x = ADD_LONG( x, args[3] );
cff_builder_add_point( builder, x, y, 1 );
/* third control point, with y-value the same as the join */
/* point's y-value */
x = ADD_LONG( x, args[4] );
cff_builder_add_point( builder, x, y, 0 );
/* fourth control point */
x = ADD_LONG( x, args[5] );
y = start_y;
cff_builder_add_point( builder, x, y, 0 );
/* ending point, with y-value the same as the start point's */
/* y-value -- we don't add this point, though */
x = ADD_LONG( x, args[6] );
cff_builder_add_point( builder, x, y, 1 );
args = stack;
break;
}
case cff_op_flex1:
{
FT_Pos start_x, start_y; /* record start x, y values for */
/* alter use */
FT_Fixed dx = 0, dy = 0; /* used in horizontal/vertical */
/* algorithm below */
FT_Int horizontal, count;
FT_Fixed* temp;
FT_TRACE4(( " flex1\n" ));
/* adding six more points; 4 control points, 2 on-curve points */
if ( cff_builder_start_point( builder, x, y ) ||
cff_check_points( builder, 6 ) )
goto Fail;
/* record the starting point's x, y position for later use */
start_x = x;
start_y = y;
/* XXX: figure out whether this is supposed to be a horizontal */
/* or vertical flex; the Type 2 specification is vague... */
temp = args;
/* grab up to the last argument */
for ( count = 5; count > 0; count-- )
{
dx = ADD_LONG( dx, temp[0] );
dy = ADD_LONG( dy, temp[1] );
temp += 2;
}
if ( dx < 0 )
dx = NEG_LONG( dx );
if ( dy < 0 )
dy = NEG_LONG( dy );
/* strange test, but here it is... */
horizontal = ( dx > dy );
for ( count = 5; count > 0; count-- )
{
x = ADD_LONG( x, args[0] );
y = ADD_LONG( y, args[1] );
cff_builder_add_point( builder, x, y,
FT_BOOL( count == 3 ) );
args += 2;
}
/* is last operand an x- or y-delta? */
if ( horizontal )
{
x = ADD_LONG( x, args[0] );
y = start_y;
}
else
{
x = start_x;
y = ADD_LONG( y, args[0] );
}
cff_builder_add_point( builder, x, y, 1 );
args = stack;
break;
}
case cff_op_flex:
{
FT_UInt count;
FT_TRACE4(( " flex\n" ));
if ( cff_builder_start_point( builder, x, y ) ||
cff_check_points( builder, 6 ) )
goto Fail;
for ( count = 6; count > 0; count-- )
{
x = ADD_LONG( x, args[0] );
y = ADD_LONG( y, args[1] );
cff_builder_add_point( builder, x, y,
FT_BOOL( count == 4 || count == 1 ) );
args += 2;
}
args = stack;
}
break;
case cff_op_seac:
FT_TRACE4(( " seac\n" ));
error = cff_operator_seac( decoder,
args[0], args[1], args[2],
(FT_Int)( args[3] >> 16 ),
(FT_Int)( args[4] >> 16 ) );
/* add current outline to the glyph slot */
FT_GlyphLoader_Add( builder->loader );
/* return now! */
FT_TRACE4(( "\n" ));
return error;
case cff_op_endchar:
/* in dictionaries, `endchar' simply indicates end of data */
if ( in_dict )
return error;
FT_TRACE4(( " endchar\n" ));
/* We are going to emulate the seac operator. */
if ( num_args >= 4 )
{
/* Save glyph width so that the subglyphs don't overwrite it. */
FT_Pos glyph_width = decoder->glyph_width;
error = cff_operator_seac( decoder,
0L, args[-4], args[-3],
(FT_Int)( args[-2] >> 16 ),
(FT_Int)( args[-1] >> 16 ) );
decoder->glyph_width = glyph_width;
}
else
{
cff_builder_close_contour( builder );
/* close hints recording session */
if ( hinter )
{
if ( hinter->close( hinter->hints,
(FT_UInt)builder->current->n_points ) )
goto Syntax_Error;
/* apply hints to the loaded glyph outline now */
error = hinter->apply( hinter->hints,
builder->current,
(PSH_Globals)builder->hints_globals,
decoder->hint_mode );
if ( error )
goto Fail;
}
/* add current outline to the glyph slot */
FT_GlyphLoader_Add( builder->loader );
}
/* return now! */
FT_TRACE4(( "\n" ));
return error;
case cff_op_abs:
FT_TRACE4(( " abs\n" ));
if ( args[0] < 0 )
{
if ( args[0] == FT_LONG_MIN )
args[0] = FT_LONG_MAX;
else
args[0] = -args[0];
}
args++;
break;
case cff_op_add:
FT_TRACE4(( " add\n" ));
args[0] = ADD_LONG( args[0], args[1] );
args++;
break;
case cff_op_sub:
FT_TRACE4(( " sub\n" ));
args[0] = SUB_LONG( args[0], args[1] );
args++;
break;
case cff_op_div:
FT_TRACE4(( " div\n" ));
args[0] = FT_DivFix( args[0], args[1] );
args++;
break;
case cff_op_neg:
FT_TRACE4(( " neg\n" ));
if ( args[0] == FT_LONG_MIN )
args[0] = FT_LONG_MAX;
args[0] = -args[0];
args++;
break;
case cff_op_random:
{
FT_UInt32* randval = in_dict ? &decoder->cff->top_font.random
: &decoder->current_subfont->random;
FT_TRACE4(( " random\n" ));
/* only use the lower 16 bits of `random' */
/* to generate a number in the range (0;1] */
args[0] = (FT_Fixed)( ( *randval & 0xFFFF ) + 1 );
args++;
*randval = cff_random( *randval );
}
break;
case cff_op_mul:
FT_TRACE4(( " mul\n" ));
args[0] = FT_MulFix( args[0], args[1] );
args++;
break;
case cff_op_sqrt:
FT_TRACE4(( " sqrt\n" ));
/* without upper limit the loop below might not finish */
if ( args[0] > 0x7FFFFFFFL )
args[0] = 46341;
else if ( args[0] > 0 )
{
FT_Fixed root = args[0];
FT_Fixed new_root;
for (;;)
{
new_root = ( root + FT_DivFix( args[0], root ) + 1 ) >> 1;
if ( new_root == root )
break;
root = new_root;
}
args[0] = new_root;
}
else
args[0] = 0;
args++;
break;
case cff_op_drop:
/* nothing */
FT_TRACE4(( " drop\n" ));
break;
case cff_op_exch:
{
FT_Fixed tmp;
FT_TRACE4(( " exch\n" ));
tmp = args[0];
args[0] = args[1];
args[1] = tmp;
args += 2;
}
break;
case cff_op_index:
{
FT_Int idx = (FT_Int)( args[0] >> 16 );
FT_TRACE4(( " index\n" ));
if ( idx < 0 )
idx = 0;
else if ( idx > num_args - 2 )
idx = num_args - 2;
args[0] = args[-( idx + 1 )];
args++;
}
break;
case cff_op_roll:
{
FT_Int count = (FT_Int)( args[0] >> 16 );
FT_Int idx = (FT_Int)( args[1] >> 16 );
FT_TRACE4(( " roll\n" ));
if ( count <= 0 )
count = 1;
args -= count;
if ( args < stack )
goto Stack_Underflow;
if ( idx >= 0 )
{
idx = idx % count;
while ( idx > 0 )
{
FT_Fixed tmp = args[count - 1];
FT_Int i;
for ( i = count - 2; i >= 0; i-- )
args[i + 1] = args[i];
args[0] = tmp;
idx--;
}
}
else
{
/* before C99 it is implementation-defined whether */
/* the result of `%' is negative if the first operand */
/* is negative */
idx = -( NEG_INT( idx ) % count );
while ( idx < 0 )
{
FT_Fixed tmp = args[0];
FT_Int i;
for ( i = 0; i < count - 1; i++ )
args[i] = args[i + 1];
args[count - 1] = tmp;
idx++;
}
}
args += count;
}
break;
case cff_op_dup:
FT_TRACE4(( " dup\n" ));
args[1] = args[0];
args += 2;
break;
case cff_op_put:
{
FT_Fixed val = args[0];
FT_Int idx = (FT_Int)( args[1] >> 16 );
FT_TRACE4(( " put\n" ));
/* the Type2 specification before version 16-March-2000 */
/* didn't give a hard-coded size limit of the temporary */
/* storage array; instead, an argument of the */
/* `MultipleMaster' operator set the size */
if ( idx >= 0 && idx < CFF_MAX_TRANS_ELEMENTS )
decoder->buildchar[idx] = val;
}
break;
case cff_op_get:
{
FT_Int idx = (FT_Int)( args[0] >> 16 );
FT_Fixed val = 0;
FT_TRACE4(( " get\n" ));
if ( idx >= 0 && idx < CFF_MAX_TRANS_ELEMENTS )
val = decoder->buildchar[idx];
args[0] = val;
args++;
}
break;
case cff_op_store:
/* this operator was removed from the Type2 specification */
/* in version 16-March-2000 */
/* since we currently don't handle interpolation of multiple */
/* master fonts, this is a no-op */
FT_TRACE4(( " store\n" ));
break;
case cff_op_load:
/* this operator was removed from the Type2 specification */
/* in version 16-March-2000 */
{
FT_Int reg_idx = (FT_Int)args[0];
FT_Int idx = (FT_Int)args[1];
FT_Int count = (FT_Int)args[2];
FT_TRACE4(( " load\n" ));
/* since we currently don't handle interpolation of multiple */
/* master fonts, we store a vector [1 0 0 ...] in the */
/* temporary storage array regardless of the Registry index */
if ( reg_idx >= 0 && reg_idx <= 2 &&
idx >= 0 && idx < CFF_MAX_TRANS_ELEMENTS &&
count >= 0 && count <= num_axes )
{
FT_Int end, i;
end = FT_MIN( idx + count, CFF_MAX_TRANS_ELEMENTS );
if ( idx < end )
decoder->buildchar[idx] = 1 << 16;
for ( i = idx + 1; i < end; i++ )
decoder->buildchar[i] = 0;
}
}
break;
case cff_op_blend:
/* this operator was removed from the Type2 specification */
/* in version 16-March-2000 */
if ( num_designs )
{
FT_Int num_results = (FT_Int)( args[0] >> 16 );
FT_TRACE4(( " blend\n" ));
if ( num_results < 0 )
goto Syntax_Error;
if ( num_results > num_args ||
num_results * (FT_Int)num_designs > num_args )
goto Stack_Underflow;
/* since we currently don't handle interpolation of multiple */
/* master fonts, return the `num_results' values of the */
/* first master */
args -= num_results * ( num_designs - 1 );
num_args -= num_results * ( num_designs - 1 );
}
else
goto Syntax_Error;
break;
case cff_op_dotsection:
/* this operator is deprecated and ignored by the parser */
FT_TRACE4(( " dotsection\n" ));
break;
case cff_op_closepath:
/* this is an invalid Type 2 operator; however, there */
/* exist fonts which are incorrectly converted from probably */
/* Type 1 to CFF, and some parsers seem to accept it */
FT_TRACE4(( " closepath (invalid op)\n" ));
args = stack;
break;
case cff_op_hsbw:
/* this is an invalid Type 2 operator; however, there */
/* exist fonts which are incorrectly converted from probably */
/* Type 1 to CFF, and some parsers seem to accept it */
FT_TRACE4(( " hsbw (invalid op)\n" ));
decoder->glyph_width =
ADD_LONG( decoder->nominal_width, ( args[1] >> 16 ) );
decoder->builder.left_bearing.x = args[0];
decoder->builder.left_bearing.y = 0;
x = ADD_LONG( decoder->builder.pos_x, args[0] );
y = decoder->builder.pos_y;
args = stack;
break;
case cff_op_sbw:
/* this is an invalid Type 2 operator; however, there */
/* exist fonts which are incorrectly converted from probably */
/* Type 1 to CFF, and some parsers seem to accept it */
FT_TRACE4(( " sbw (invalid op)\n" ));
decoder->glyph_width =
ADD_LONG( decoder->nominal_width, ( args[2] >> 16 ) );
decoder->builder.left_bearing.x = args[0];
decoder->builder.left_bearing.y = args[1];
x = ADD_LONG( decoder->builder.pos_x, args[0] );
y = ADD_LONG( decoder->builder.pos_y, args[1] );
args = stack;
break;
case cff_op_setcurrentpoint:
/* this is an invalid Type 2 operator; however, there */
/* exist fonts which are incorrectly converted from probably */
/* Type 1 to CFF, and some parsers seem to accept it */
FT_TRACE4(( " setcurrentpoint (invalid op)\n" ));
x = ADD_LONG( decoder->builder.pos_x, args[0] );
y = ADD_LONG( decoder->builder.pos_y, args[1] );
args = stack;
break;
case cff_op_callothersubr:
{
FT_Fixed arg;
/* this is an invalid Type 2 operator; however, there */
/* exist fonts which are incorrectly converted from */
/* probably Type 1 to CFF, and some parsers seem to accept */
/* it */
FT_TRACE4(( " callothersubr (invalid op)\n" ));
/* subsequent `pop' operands should add the arguments, */
/* this is the implementation described for `unknown' */
/* other subroutines in the Type1 spec. */
/* */
/* XXX Fix return arguments (see discussion below). */
arg = 2 + ( args[-2] >> 16 );
if ( arg >= CFF_MAX_OPERANDS )
goto Stack_Underflow;
args -= arg;
if ( args < stack )
goto Stack_Underflow;
}
break;
case cff_op_pop:
/* this is an invalid Type 2 operator; however, there */
/* exist fonts which are incorrectly converted from probably */
/* Type 1 to CFF, and some parsers seem to accept it */
FT_TRACE4(( " pop (invalid op)\n" ));
/* XXX Increasing `args' is wrong: After a certain number of */
/* `pop's we get a stack overflow. Reason for doing it is */
/* code like this (actually found in a CFF font): */
/* */
/* 17 1 3 callothersubr */
/* pop */
/* callsubr */
/* */
/* Since we handle `callothersubr' as a no-op, and */
/* `callsubr' needs at least one argument, `pop' can't be a */
/* no-op too as it basically should be. */
/* */
/* The right solution would be to provide real support for */
/* `callothersubr' as done in `t1decode.c', however, given */
/* the fact that CFF fonts with `pop' are invalid, it is */
/* questionable whether it is worth the time. */
args++;
break;
case cff_op_and:
{
FT_Fixed cond = ( args[0] && args[1] );
FT_TRACE4(( " and\n" ));
args[0] = cond ? 0x10000L : 0;
args++;
}
break;
case cff_op_or:
{
FT_Fixed cond = ( args[0] || args[1] );
FT_TRACE4(( " or\n" ));
args[0] = cond ? 0x10000L : 0;
args++;
}
break;
case cff_op_not:
{
FT_Fixed cond = !args[0];
FT_TRACE4(( " not\n" ));
args[0] = cond ? 0x10000L : 0;
args++;
}
break;
case cff_op_eq:
{
FT_Fixed cond = ( args[0] == args[1] );
FT_TRACE4(( " eq\n" ));
args[0] = cond ? 0x10000L : 0;
args++;
}
break;
case cff_op_ifelse:
{
FT_Fixed cond = ( args[2] <= args[3] );
FT_TRACE4(( " ifelse\n" ));
if ( !cond )
args[0] = args[1];
args++;
}
break;
case cff_op_callsubr:
{
FT_UInt idx = (FT_UInt)( ( args[0] >> 16 ) +
decoder->locals_bias );
FT_TRACE4(( " callsubr (idx %d, entering level %ld)\n",
idx,
zone - decoder->zones + 1 ));
if ( idx >= decoder->num_locals )
{
FT_ERROR(( "cff_decoder_parse_charstrings:"
" invalid local subr index\n" ));
goto Syntax_Error;
}
if ( zone - decoder->zones >= CFF_MAX_SUBRS_CALLS )
{
FT_ERROR(( "cff_decoder_parse_charstrings:"
" too many nested subrs\n" ));
goto Syntax_Error;
}
zone->cursor = ip; /* save current instruction pointer */
zone++;
zone->base = decoder->locals[idx];
zone->limit = decoder->locals[idx + 1];
zone->cursor = zone->base;
if ( !zone->base || zone->limit == zone->base )
{
FT_ERROR(( "cff_decoder_parse_charstrings:"
" invoking empty subrs\n" ));
goto Syntax_Error;
}
decoder->zone = zone;
ip = zone->base;
limit = zone->limit;
}
break;
case cff_op_callgsubr:
{
FT_UInt idx = (FT_UInt)( ( args[0] >> 16 ) +
decoder->globals_bias );
FT_TRACE4(( " callgsubr (idx %d, entering level %ld)\n",
idx,
zone - decoder->zones + 1 ));
if ( idx >= decoder->num_globals )
{
FT_ERROR(( "cff_decoder_parse_charstrings:"
" invalid global subr index\n" ));
goto Syntax_Error;
}
if ( zone - decoder->zones >= CFF_MAX_SUBRS_CALLS )
{
FT_ERROR(( "cff_decoder_parse_charstrings:"
" too many nested subrs\n" ));
goto Syntax_Error;
}
zone->cursor = ip; /* save current instruction pointer */
zone++;
zone->base = decoder->globals[idx];
zone->limit = decoder->globals[idx + 1];
zone->cursor = zone->base;
if ( !zone->base || zone->limit == zone->base )
{
FT_ERROR(( "cff_decoder_parse_charstrings:"
" invoking empty subrs\n" ));
goto Syntax_Error;
}
decoder->zone = zone;
ip = zone->base;
limit = zone->limit;
}
break;
case cff_op_return:
FT_TRACE4(( " return (leaving level %ld)\n",
decoder->zone - decoder->zones ));
if ( decoder->zone <= decoder->zones )
{
FT_ERROR(( "cff_decoder_parse_charstrings:"
" unexpected return\n" ));
goto Syntax_Error;
}
decoder->zone--;
zone = decoder->zone;
ip = zone->cursor;
limit = zone->limit;
break;
default:
FT_ERROR(( "Unimplemented opcode: %d", ip[-1] ));
if ( ip[-1] == 12 )
FT_ERROR(( " %d", ip[0] ));
FT_ERROR(( "\n" ));
return FT_THROW( Unimplemented_Feature );
}
decoder->top = args;
if ( decoder->top - stack >= CFF_MAX_OPERANDS )
goto Stack_Overflow;
} /* general operator processing */
} /* while ip < limit */
FT_TRACE4(( "..end..\n" ));
FT_TRACE4(( "\n" ));
Fail:
return error;
MM_Error:
FT_TRACE4(( "cff_decoder_parse_charstrings:"
" invalid opcode found in top DICT charstring\n"));
return FT_THROW( Invalid_File_Format );
Syntax_Error:
FT_TRACE4(( "cff_decoder_parse_charstrings: syntax error\n" ));
return FT_THROW( Invalid_File_Format );
Stack_Underflow:
FT_TRACE4(( "cff_decoder_parse_charstrings: stack underflow\n" ));
return FT_THROW( Too_Few_Arguments );
Stack_Overflow:
FT_TRACE4(( "cff_decoder_parse_charstrings: stack overflow\n" ));
return FT_THROW( Stack_Overflow );
}
#endif /* CFF_CONFIG_OPTION_OLD_ENGINE */
/**************************************************************************
*
* @Function:
* cff_decoder_init
*
* @Description:
* Initializes a given glyph decoder.
*
* @InOut:
* decoder ::
* A pointer to the glyph builder to initialize.
*
* @Input:
* face ::
* The current face object.
*
* size ::
* The current size object.
*
* slot ::
* The current glyph object.
*
* hinting ::
* Whether hinting is active.
*
* hint_mode ::
* The hinting mode.
*/
FT_LOCAL_DEF( void )
cff_decoder_init( CFF_Decoder* decoder,
TT_Face face,
CFF_Size size,
CFF_GlyphSlot slot,
FT_Bool hinting,
FT_Render_Mode hint_mode,
CFF_Decoder_Get_Glyph_Callback get_callback,
CFF_Decoder_Free_Glyph_Callback free_callback )
{
CFF_Font cff = (CFF_Font)face->extra.data;
/* clear everything */
FT_ZERO( decoder );
/* initialize builder */
cff_builder_init( &decoder->builder, face, size, slot, hinting );
/* initialize Type2 decoder */
decoder->cff = cff;
decoder->num_globals = cff->global_subrs_index.count;
decoder->globals = cff->global_subrs;
decoder->globals_bias = cff_compute_bias(
cff->top_font.font_dict.charstring_type,
decoder->num_globals );
decoder->hint_mode = hint_mode;
decoder->get_glyph_callback = get_callback;
decoder->free_glyph_callback = free_callback;
}
/* this function is used to select the subfont */
/* and the locals subrs array */
FT_LOCAL_DEF( FT_Error )
cff_decoder_prepare( CFF_Decoder* decoder,
CFF_Size size,
FT_UInt glyph_index )
{
CFF_Builder *builder = &decoder->builder;
CFF_Font cff = (CFF_Font)builder->face->extra.data;
CFF_SubFont sub = &cff->top_font;
FT_Error error = FT_Err_Ok;
FT_Service_CFFLoad cffload = (FT_Service_CFFLoad)cff->cffload;
/* manage CID fonts */
if ( cff->num_subfonts )
{
FT_Byte fd_index = cffload->fd_select_get( &cff->fd_select,
glyph_index );
if ( fd_index >= cff->num_subfonts )
{
FT_TRACE4(( "cff_decoder_prepare: invalid CID subfont index\n" ));
error = FT_THROW( Invalid_File_Format );
goto Exit;
}
FT_TRACE3(( " in subfont %d:\n", fd_index ));
sub = cff->subfonts[fd_index];
if ( builder->hints_funcs && size )
{
FT_Size ftsize = FT_SIZE( size );
CFF_Internal internal = (CFF_Internal)ftsize->internal->module_data;
/* for CFFs without subfonts, this value has already been set */
builder->hints_globals = (void *)internal->subfonts[fd_index];
}
}
decoder->num_locals = sub->local_subrs_index.count;
decoder->locals = sub->local_subrs;
decoder->locals_bias = cff_compute_bias(
decoder->cff->top_font.font_dict.charstring_type,
decoder->num_locals );
decoder->glyph_width = sub->private_dict.default_width;
decoder->nominal_width = sub->private_dict.nominal_width;
decoder->current_subfont = sub;
Exit:
return error;
}
/* END */