-
Notifications
You must be signed in to change notification settings - Fork 7
/
interpreter.go
376 lines (329 loc) · 10.2 KB
/
interpreter.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
// Package psinterpreter implement a Postscript interpreter
// required to parse .CFF files, and Type1 and Type2 Charstrings.
// This package provides the low-level mechanisms needed to
// read such formats; the data is consumed in higher level packages,
// which implement `PsOperatorHandler`.
// It also provides helpers to interpret glyph outline descriptions,
// shared between Type1 and CFF font formats.
package psinterpreter
import (
"encoding/binary"
"errors"
"fmt"
"math"
"strconv"
)
var (
// ErrInterrupt signals the interpreter to stop early, without erroring.
ErrInterrupt = errors.New("interruption")
errInvalidCFFTable = errors.New("invalid ps instructions")
errUnsupportedCFFVersion = errors.New("unsupported CFF version")
errUnsupportedRealNumberEncoding = errors.New("unsupported real number encoding")
be = binary.BigEndian
)
const (
// psArgStackSize is the argument stack size for a PostScript interpreter.
// 5176.CFF.pdf section 4 "DICT Data" says that "An operator may be
// preceded by up to a maximum of 48 operands". 5177.Type2.pdf Appendix B
// "Type 2 Charstring Implementation Limits" says that "Argument stack 48".
// T1_SPEC.pdf 6.1 Encoding as a limitation of 24.
psArgStackSize = 48
// Similarly, Appendix B says "Subr nesting, stack limit 10".
psCallStackSize = 10
maxRealNumberStrLen = 64 // Maximum length in bytes of the "-123.456E-7" representation.
)
// PsContext is the flavour of the Postcript language.
type PsContext uint32
const (
TopDict PsContext = iota // Top dict in CFF files
PrivateDict // Private dict in CFF files
Type2Charstring // Charstring in CFF files
Type1Charstring // Charstring in Type1 font files
)
type ArgStack struct {
Vals [psArgStackSize]int32
// Effecive size currently in use. The first value to
// pop is at index Top-1
Top int32
}
// Uint16 returns the top level value as uint16,
// without popping the stack.
func (a *ArgStack) Uint16() uint16 { return uint16(a.Vals[a.Top-1]) }
// Float return the top level value as a real number (which is stored as its binary representation),
// without popping the stack.
func (a *ArgStack) Float() float32 {
return math.Float32frombits(uint32(a.Vals[a.Top-1]))
}
// Pop returns the top level value and decrease `Top`
// It will panic if the stack is empty.
func (a *ArgStack) Pop() int32 {
a.Top--
return a.Vals[a.Top]
}
// Clear clears the stack
func (a *ArgStack) Clear() { a.Top = 0 }
// PopN check and remove the n top levels entries.
// Passing a negative `numPop` clears all the stack.
func (a *ArgStack) PopN(numPop int32) error {
if a.Top < numPop {
return fmt.Errorf("invalid number of operands in PS stack: %d", numPop)
}
if numPop < 0 { // pop all
a.Top = 0
} else {
a.Top -= numPop
}
return nil
}
// Machine is a PostScript interpreter.
// A same interpreter may be re-used using muliples `Run` calls.
type Machine struct {
localSubrs [][]byte
globalSubrs [][]byte
instructions []byte
callStack struct {
vals [psCallStackSize][]byte // parent instructions
top int32 // effecive size currently in use
}
ArgStack ArgStack
parseNumberBuf [maxRealNumberStrLen]byte
ctx PsContext
}
// SkipBytes skips the next `count` bytes from the instructions, and clears the argument stack.
// It does nothing if `count` exceed the length of the instructions.
func (p *Machine) SkipBytes(count int32) {
if int(count) >= len(p.instructions) {
return
}
p.instructions = p.instructions[count:]
p.ArgStack.Clear()
}
func (p *Machine) hasMoreInstructions() bool {
if len(p.instructions) != 0 {
return true
}
for i := int32(0); i < p.callStack.top; i++ {
if len(p.callStack.vals[i]) != 0 {
return true
}
}
return false
}
// 5176.CFF.pdf section 4 "DICT Data" says that "Two-byte operators have an
// initial escape byte of 12".
const escapeByte = 12
// Run runs the instructions in the PostScript context asked by `handler`.
// `localSubrs` and `globalSubrs` contains the subroutines that may be called in the instructions.
func (p *Machine) Run(instructions []byte, localSubrs, globalSubrs [][]byte, handler PsOperatorHandler) error {
p.ctx = handler.Context()
p.instructions = instructions
p.localSubrs = localSubrs
p.globalSubrs = globalSubrs
p.ArgStack.Top = 0
p.callStack.top = 0
for len(p.instructions) > 0 {
// Push a numeric operand on the stack, if applicable.
if hasResult, err := p.parseNumber(); hasResult {
if err != nil {
return err
}
continue
}
// Otherwise, execute an operator.
b := p.instructions[0]
p.instructions = p.instructions[1:]
// check for the escape byte
escaped := b == escapeByte
if escaped {
if len(p.instructions) <= 0 {
return errInvalidCFFTable
}
b = p.instructions[0]
p.instructions = p.instructions[1:]
}
err := handler.Apply(PsOperator{Operator: b, IsEscaped: escaped}, p)
if err == ErrInterrupt { // stop cleanly
return nil
}
if err != nil {
return err
}
}
return nil
}
// See 5176.CFF.pdf section 4 "DICT Data".
func (p *Machine) parseNumber() (hasResult bool, err error) {
number := int32(0)
switch b := p.instructions[0]; {
case b == 28:
if len(p.instructions) < 3 {
return true, errInvalidCFFTable
}
number, hasResult = int32(int16(be.Uint16(p.instructions[1:]))), true
p.instructions = p.instructions[3:]
case b == 29 && p.ctx != Type2Charstring:
if len(p.instructions) < 5 {
return true, errInvalidCFFTable
}
number, hasResult = int32(be.Uint32(p.instructions[1:])), true
p.instructions = p.instructions[5:]
case b == 30 && p.ctx != Type2Charstring && p.ctx != Type1Charstring:
// Parse a real number. This isn't listed in 5176.CFF.pdf Table 3
// "Operand Encoding" but that table lists integer encodings. Further
// down the page it says "A real number operand is provided in addition
// to integer operands. This operand begins with a byte value of 30
// followed by a variable-length sequence of bytes."
s := p.parseNumberBuf[:0]
p.instructions = p.instructions[1:]
loop:
for {
if len(p.instructions) == 0 {
return true, errInvalidCFFTable
}
by := p.instructions[0]
p.instructions = p.instructions[1:]
// Process by's two nibbles, high then low.
for i := 0; i < 2; i++ {
nib := by >> 4
by = by << 4
if nib == 0x0f {
f, err := strconv.ParseFloat(string(s), 32)
if err != nil {
return true, errInvalidCFFTable
}
number, hasResult = int32(math.Float32bits(float32(f))), true
break loop
}
if nib == 0x0d {
return true, errInvalidCFFTable
}
if len(s)+maxNibbleDefsLength > len(p.parseNumberBuf) {
return true, errUnsupportedRealNumberEncoding
}
s = append(s, nibbleDefs[nib]...)
}
}
case b < 32:
// not a number: no-op.
case b < 247:
p.instructions = p.instructions[1:]
number, hasResult = int32(b)-139, true
case b < 251:
if len(p.instructions) < 2 {
return true, errInvalidCFFTable
}
b1 := p.instructions[1]
p.instructions = p.instructions[2:]
number, hasResult = +int32(b-247)*256+int32(b1)+108, true
case b < 255:
if len(p.instructions) < 2 {
return true, errInvalidCFFTable
}
b1 := p.instructions[1]
p.instructions = p.instructions[2:]
number, hasResult = -int32(b-251)*256-int32(b1)-108, true
case b == 255 && (p.ctx == Type2Charstring || p.ctx == Type1Charstring):
if len(p.instructions) < 5 {
return true, errInvalidCFFTable
}
number, hasResult = int32(be.Uint32(p.instructions[1:])), true
p.instructions = p.instructions[5:]
}
if hasResult {
if p.ArgStack.Top == psArgStackSize {
return true, errInvalidCFFTable
}
p.ArgStack.Vals[p.ArgStack.Top] = number
p.ArgStack.Top++
}
return hasResult, nil
}
const maxNibbleDefsLength = len("E-")
// nibbleDefs encodes 5176.CFF.pdf Table 5 "Nibble Definitions".
var nibbleDefs = [16]string{
0x00: "0",
0x01: "1",
0x02: "2",
0x03: "3",
0x04: "4",
0x05: "5",
0x06: "6",
0x07: "7",
0x08: "8",
0x09: "9",
0x0a: ".",
0x0b: "E",
0x0c: "E-",
0x0d: "",
0x0e: "-",
0x0f: "",
}
// subrBias returns the subroutine index bias as per 5177.Type2.pdf section 4.7
// "Subroutine Operators".
func subrBias(numSubroutines int) int32 {
if numSubroutines < 1240 {
return 107
}
if numSubroutines < 33900 {
return 1131
}
return 32768
}
// CallSubroutine calls the subroutine, identified by its index, as found
// in the instructions (that is, before applying the subroutine biased).
// `isLocal` controls whether the local or global subroutines are used.
// No argument stack modification is performed.
func (p *Machine) CallSubroutine(index int32, isLocal bool) error {
subrs := p.globalSubrs
if isLocal {
subrs = p.localSubrs
}
// no bias in type1 fonts
if p.ctx == Type2Charstring {
index += subrBias(len(subrs))
}
if index < 0 || int(index) >= len(subrs) {
return fmt.Errorf("invalid subroutine index %d (for length %d)", index, len(subrs))
}
if p.callStack.top == psCallStackSize {
return errors.New("maximum call stack size reached")
}
// save the current instructions
p.callStack.vals[p.callStack.top] = p.instructions
p.callStack.top++
// activate the subroutine
p.instructions = subrs[index]
return nil
}
// Return returns from a subroutine call.
func (p *Machine) Return() error {
if p.callStack.top <= 0 {
return errors.New("no subroutine has been called")
}
p.callStack.top--
// restore the previous instructions
p.instructions = p.callStack.vals[p.callStack.top]
return nil
}
// PsOperator is a postcript command, which may be escaped.
type PsOperator struct {
Operator byte
IsEscaped bool
}
func (p PsOperator) String() string {
if p.IsEscaped {
return fmt.Sprintf("2-byte operator (12 %d)", p.Operator)
}
return fmt.Sprintf("1-byte operator (%d)", p.Operator)
}
// PsOperatorHandler defines the behaviour of an operator.
type PsOperatorHandler interface {
// Context defines the precise behaviour of the interpreter,
// which has small nuances depending on the context.
Context() PsContext
// Apply implements the operator defined by `operator` (which is the second byte if `escaped` is true).
//
// Returning `ErrInterrupt` stop the parsing of the instructions, without reporting an error.
// It can be used as an optimization.
Apply(operator PsOperator, state *Machine) error
}