256
|
1 |
// Copyright 2018 The Go Authors. All rights reserved. |
|
2 |
// Use of this source code is governed by a BSD-style |
|
3 |
// license that can be found in the LICENSE file. |
|
4 |
|
|
5 |
package text |
|
6 |
|
|
7 |
import ( |
|
8 |
"bytes" |
|
9 |
"fmt" |
|
10 |
"io" |
|
11 |
"strconv" |
|
12 |
"unicode/utf8" |
|
13 |
|
|
14 |
"google.golang.org/protobuf/internal/errors" |
|
15 |
) |
|
16 |
|
|
17 |
// Decoder is a token-based textproto decoder. |
|
18 |
type Decoder struct { |
|
19 |
// lastCall is last method called, either readCall or peekCall. |
|
20 |
// Initial value is readCall. |
|
21 |
lastCall call |
|
22 |
|
|
23 |
// lastToken contains the last read token. |
|
24 |
lastToken Token |
|
25 |
|
|
26 |
// lastErr contains the last read error. |
|
27 |
lastErr error |
|
28 |
|
|
29 |
// openStack is a stack containing the byte characters for MessageOpen and |
|
30 |
// ListOpen kinds. The top of stack represents the message or the list that |
|
31 |
// the current token is nested in. An empty stack means the current token is |
|
32 |
// at the top level message. The characters '{' and '<' both represent the |
|
33 |
// MessageOpen kind. |
|
34 |
openStack []byte |
|
35 |
|
|
36 |
// orig is used in reporting line and column. |
|
37 |
orig []byte |
|
38 |
// in contains the unconsumed input. |
|
39 |
in []byte |
|
40 |
} |
|
41 |
|
|
42 |
// NewDecoder returns a Decoder to read the given []byte. |
|
43 |
func NewDecoder(b []byte) *Decoder { |
|
44 |
return &Decoder{orig: b, in: b} |
|
45 |
} |
|
46 |
|
|
47 |
// ErrUnexpectedEOF means that EOF was encountered in the middle of the input. |
|
48 |
var ErrUnexpectedEOF = errors.New("%v", io.ErrUnexpectedEOF) |
|
49 |
|
|
50 |
// call specifies which Decoder method was invoked. |
|
51 |
type call uint8 |
|
52 |
|
|
53 |
const ( |
|
54 |
readCall call = iota |
|
55 |
peekCall |
|
56 |
) |
|
57 |
|
|
58 |
// Peek looks ahead and returns the next token and error without advancing a read. |
|
59 |
func (d *Decoder) Peek() (Token, error) { |
|
60 |
defer func() { d.lastCall = peekCall }() |
|
61 |
if d.lastCall == readCall { |
|
62 |
d.lastToken, d.lastErr = d.Read() |
|
63 |
} |
|
64 |
return d.lastToken, d.lastErr |
|
65 |
} |
|
66 |
|
|
67 |
// Read returns the next token. |
|
68 |
// It will return an error if there is no valid token. |
|
69 |
func (d *Decoder) Read() (Token, error) { |
|
70 |
defer func() { d.lastCall = readCall }() |
|
71 |
if d.lastCall == peekCall { |
|
72 |
return d.lastToken, d.lastErr |
|
73 |
} |
|
74 |
|
|
75 |
tok, err := d.parseNext(d.lastToken.kind) |
|
76 |
if err != nil { |
|
77 |
return Token{}, err |
|
78 |
} |
|
79 |
|
|
80 |
switch tok.kind { |
|
81 |
case comma, semicolon: |
|
82 |
tok, err = d.parseNext(tok.kind) |
|
83 |
if err != nil { |
|
84 |
return Token{}, err |
|
85 |
} |
|
86 |
} |
|
87 |
d.lastToken = tok |
|
88 |
return tok, nil |
|
89 |
} |
|
90 |
|
|
91 |
const ( |
|
92 |
mismatchedFmt = "mismatched close character %q" |
|
93 |
unexpectedFmt = "unexpected character %q" |
|
94 |
) |
|
95 |
|
|
96 |
// parseNext parses the next Token based on given last kind. |
|
97 |
func (d *Decoder) parseNext(lastKind Kind) (Token, error) { |
|
98 |
// Trim leading spaces. |
|
99 |
d.consume(0) |
|
100 |
isEOF := false |
|
101 |
if len(d.in) == 0 { |
|
102 |
isEOF = true |
|
103 |
} |
|
104 |
|
|
105 |
switch lastKind { |
|
106 |
case EOF: |
|
107 |
return d.consumeToken(EOF, 0, 0), nil |
|
108 |
|
|
109 |
case bof: |
|
110 |
// Start of top level message. Next token can be EOF or Name. |
|
111 |
if isEOF { |
|
112 |
return d.consumeToken(EOF, 0, 0), nil |
|
113 |
} |
|
114 |
return d.parseFieldName() |
|
115 |
|
|
116 |
case Name: |
|
117 |
// Next token can be MessageOpen, ListOpen or Scalar. |
|
118 |
if isEOF { |
|
119 |
return Token{}, ErrUnexpectedEOF |
|
120 |
} |
|
121 |
switch ch := d.in[0]; ch { |
|
122 |
case '{', '<': |
|
123 |
d.pushOpenStack(ch) |
|
124 |
return d.consumeToken(MessageOpen, 1, 0), nil |
|
125 |
case '[': |
|
126 |
d.pushOpenStack(ch) |
|
127 |
return d.consumeToken(ListOpen, 1, 0), nil |
|
128 |
default: |
|
129 |
return d.parseScalar() |
|
130 |
} |
|
131 |
|
|
132 |
case Scalar: |
|
133 |
openKind, closeCh := d.currentOpenKind() |
|
134 |
switch openKind { |
|
135 |
case bof: |
|
136 |
// Top level message. |
|
137 |
// Next token can be EOF, comma, semicolon or Name. |
|
138 |
if isEOF { |
|
139 |
return d.consumeToken(EOF, 0, 0), nil |
|
140 |
} |
|
141 |
switch d.in[0] { |
|
142 |
case ',': |
|
143 |
return d.consumeToken(comma, 1, 0), nil |
|
144 |
case ';': |
|
145 |
return d.consumeToken(semicolon, 1, 0), nil |
|
146 |
default: |
|
147 |
return d.parseFieldName() |
|
148 |
} |
|
149 |
|
|
150 |
case MessageOpen: |
|
151 |
// Next token can be MessageClose, comma, semicolon or Name. |
|
152 |
if isEOF { |
|
153 |
return Token{}, ErrUnexpectedEOF |
|
154 |
} |
|
155 |
switch ch := d.in[0]; ch { |
|
156 |
case closeCh: |
|
157 |
d.popOpenStack() |
|
158 |
return d.consumeToken(MessageClose, 1, 0), nil |
|
159 |
case otherCloseChar[closeCh]: |
|
160 |
return Token{}, d.newSyntaxError(mismatchedFmt, ch) |
|
161 |
case ',': |
|
162 |
return d.consumeToken(comma, 1, 0), nil |
|
163 |
case ';': |
|
164 |
return d.consumeToken(semicolon, 1, 0), nil |
|
165 |
default: |
|
166 |
return d.parseFieldName() |
|
167 |
} |
|
168 |
|
|
169 |
case ListOpen: |
|
170 |
// Next token can be ListClose or comma. |
|
171 |
if isEOF { |
|
172 |
return Token{}, ErrUnexpectedEOF |
|
173 |
} |
|
174 |
switch ch := d.in[0]; ch { |
|
175 |
case ']': |
|
176 |
d.popOpenStack() |
|
177 |
return d.consumeToken(ListClose, 1, 0), nil |
|
178 |
case ',': |
|
179 |
return d.consumeToken(comma, 1, 0), nil |
|
180 |
default: |
|
181 |
return Token{}, d.newSyntaxError(unexpectedFmt, ch) |
|
182 |
} |
|
183 |
} |
|
184 |
|
|
185 |
case MessageOpen: |
|
186 |
// Next token can be MessageClose or Name. |
|
187 |
if isEOF { |
|
188 |
return Token{}, ErrUnexpectedEOF |
|
189 |
} |
|
190 |
_, closeCh := d.currentOpenKind() |
|
191 |
switch ch := d.in[0]; ch { |
|
192 |
case closeCh: |
|
193 |
d.popOpenStack() |
|
194 |
return d.consumeToken(MessageClose, 1, 0), nil |
|
195 |
case otherCloseChar[closeCh]: |
|
196 |
return Token{}, d.newSyntaxError(mismatchedFmt, ch) |
|
197 |
default: |
|
198 |
return d.parseFieldName() |
|
199 |
} |
|
200 |
|
|
201 |
case MessageClose: |
|
202 |
openKind, closeCh := d.currentOpenKind() |
|
203 |
switch openKind { |
|
204 |
case bof: |
|
205 |
// Top level message. |
|
206 |
// Next token can be EOF, comma, semicolon or Name. |
|
207 |
if isEOF { |
|
208 |
return d.consumeToken(EOF, 0, 0), nil |
|
209 |
} |
|
210 |
switch ch := d.in[0]; ch { |
|
211 |
case ',': |
|
212 |
return d.consumeToken(comma, 1, 0), nil |
|
213 |
case ';': |
|
214 |
return d.consumeToken(semicolon, 1, 0), nil |
|
215 |
default: |
|
216 |
return d.parseFieldName() |
|
217 |
} |
|
218 |
|
|
219 |
case MessageOpen: |
|
220 |
// Next token can be MessageClose, comma, semicolon or Name. |
|
221 |
if isEOF { |
|
222 |
return Token{}, ErrUnexpectedEOF |
|
223 |
} |
|
224 |
switch ch := d.in[0]; ch { |
|
225 |
case closeCh: |
|
226 |
d.popOpenStack() |
|
227 |
return d.consumeToken(MessageClose, 1, 0), nil |
|
228 |
case otherCloseChar[closeCh]: |
|
229 |
return Token{}, d.newSyntaxError(mismatchedFmt, ch) |
|
230 |
case ',': |
|
231 |
return d.consumeToken(comma, 1, 0), nil |
|
232 |
case ';': |
|
233 |
return d.consumeToken(semicolon, 1, 0), nil |
|
234 |
default: |
|
235 |
return d.parseFieldName() |
|
236 |
} |
|
237 |
|
|
238 |
case ListOpen: |
|
239 |
// Next token can be ListClose or comma |
|
240 |
if isEOF { |
|
241 |
return Token{}, ErrUnexpectedEOF |
|
242 |
} |
|
243 |
switch ch := d.in[0]; ch { |
|
244 |
case closeCh: |
|
245 |
d.popOpenStack() |
|
246 |
return d.consumeToken(ListClose, 1, 0), nil |
|
247 |
case ',': |
|
248 |
return d.consumeToken(comma, 1, 0), nil |
|
249 |
default: |
|
250 |
return Token{}, d.newSyntaxError(unexpectedFmt, ch) |
|
251 |
} |
|
252 |
} |
|
253 |
|
|
254 |
case ListOpen: |
|
255 |
// Next token can be ListClose, MessageStart or Scalar. |
|
256 |
if isEOF { |
|
257 |
return Token{}, ErrUnexpectedEOF |
|
258 |
} |
|
259 |
switch ch := d.in[0]; ch { |
|
260 |
case ']': |
|
261 |
d.popOpenStack() |
|
262 |
return d.consumeToken(ListClose, 1, 0), nil |
|
263 |
case '{', '<': |
|
264 |
d.pushOpenStack(ch) |
|
265 |
return d.consumeToken(MessageOpen, 1, 0), nil |
|
266 |
default: |
|
267 |
return d.parseScalar() |
|
268 |
} |
|
269 |
|
|
270 |
case ListClose: |
|
271 |
openKind, closeCh := d.currentOpenKind() |
|
272 |
switch openKind { |
|
273 |
case bof: |
|
274 |
// Top level message. |
|
275 |
// Next token can be EOF, comma, semicolon or Name. |
|
276 |
if isEOF { |
|
277 |
return d.consumeToken(EOF, 0, 0), nil |
|
278 |
} |
|
279 |
switch ch := d.in[0]; ch { |
|
280 |
case ',': |
|
281 |
return d.consumeToken(comma, 1, 0), nil |
|
282 |
case ';': |
|
283 |
return d.consumeToken(semicolon, 1, 0), nil |
|
284 |
default: |
|
285 |
return d.parseFieldName() |
|
286 |
} |
|
287 |
|
|
288 |
case MessageOpen: |
|
289 |
// Next token can be MessageClose, comma, semicolon or Name. |
|
290 |
if isEOF { |
|
291 |
return Token{}, ErrUnexpectedEOF |
|
292 |
} |
|
293 |
switch ch := d.in[0]; ch { |
|
294 |
case closeCh: |
|
295 |
d.popOpenStack() |
|
296 |
return d.consumeToken(MessageClose, 1, 0), nil |
|
297 |
case otherCloseChar[closeCh]: |
|
298 |
return Token{}, d.newSyntaxError(mismatchedFmt, ch) |
|
299 |
case ',': |
|
300 |
return d.consumeToken(comma, 1, 0), nil |
|
301 |
case ';': |
|
302 |
return d.consumeToken(semicolon, 1, 0), nil |
|
303 |
default: |
|
304 |
return d.parseFieldName() |
|
305 |
} |
|
306 |
|
|
307 |
default: |
|
308 |
// It is not possible to have this case. Let it panic below. |
|
309 |
} |
|
310 |
|
|
311 |
case comma, semicolon: |
|
312 |
openKind, closeCh := d.currentOpenKind() |
|
313 |
switch openKind { |
|
314 |
case bof: |
|
315 |
// Top level message. Next token can be EOF or Name. |
|
316 |
if isEOF { |
|
317 |
return d.consumeToken(EOF, 0, 0), nil |
|
318 |
} |
|
319 |
return d.parseFieldName() |
|
320 |
|
|
321 |
case MessageOpen: |
|
322 |
// Next token can be MessageClose or Name. |
|
323 |
if isEOF { |
|
324 |
return Token{}, ErrUnexpectedEOF |
|
325 |
} |
|
326 |
switch ch := d.in[0]; ch { |
|
327 |
case closeCh: |
|
328 |
d.popOpenStack() |
|
329 |
return d.consumeToken(MessageClose, 1, 0), nil |
|
330 |
case otherCloseChar[closeCh]: |
|
331 |
return Token{}, d.newSyntaxError(mismatchedFmt, ch) |
|
332 |
default: |
|
333 |
return d.parseFieldName() |
|
334 |
} |
|
335 |
|
|
336 |
case ListOpen: |
|
337 |
if lastKind == semicolon { |
|
338 |
// It is not be possible to have this case as logic here |
|
339 |
// should not have produced a semicolon Token when inside a |
|
340 |
// list. Let it panic below. |
|
341 |
break |
|
342 |
} |
|
343 |
// Next token can be MessageOpen or Scalar. |
|
344 |
if isEOF { |
|
345 |
return Token{}, ErrUnexpectedEOF |
|
346 |
} |
|
347 |
switch ch := d.in[0]; ch { |
|
348 |
case '{', '<': |
|
349 |
d.pushOpenStack(ch) |
|
350 |
return d.consumeToken(MessageOpen, 1, 0), nil |
|
351 |
default: |
|
352 |
return d.parseScalar() |
|
353 |
} |
|
354 |
} |
|
355 |
} |
|
356 |
|
|
357 |
line, column := d.Position(len(d.orig) - len(d.in)) |
|
358 |
panic(fmt.Sprintf("Decoder.parseNext: bug at handling line %d:%d with lastKind=%v", line, column, lastKind)) |
|
359 |
} |
|
360 |
|
|
361 |
var otherCloseChar = map[byte]byte{ |
|
362 |
'}': '>', |
|
363 |
'>': '}', |
|
364 |
} |
|
365 |
|
|
366 |
// currentOpenKind indicates whether current position is inside a message, list |
|
367 |
// or top-level message by returning MessageOpen, ListOpen or bof respectively. |
|
368 |
// If the returned kind is either a MessageOpen or ListOpen, it also returns the |
|
369 |
// corresponding closing character. |
|
370 |
func (d *Decoder) currentOpenKind() (Kind, byte) { |
|
371 |
if len(d.openStack) == 0 { |
|
372 |
return bof, 0 |
|
373 |
} |
|
374 |
openCh := d.openStack[len(d.openStack)-1] |
|
375 |
switch openCh { |
|
376 |
case '{': |
|
377 |
return MessageOpen, '}' |
|
378 |
case '<': |
|
379 |
return MessageOpen, '>' |
|
380 |
case '[': |
|
381 |
return ListOpen, ']' |
|
382 |
} |
260
|
383 |
panic(fmt.Sprintf("Decoder: openStack contains invalid byte %c", openCh)) |
256
|
384 |
} |
|
385 |
|
|
386 |
func (d *Decoder) pushOpenStack(ch byte) { |
|
387 |
d.openStack = append(d.openStack, ch) |
|
388 |
} |
|
389 |
|
|
390 |
func (d *Decoder) popOpenStack() { |
|
391 |
d.openStack = d.openStack[:len(d.openStack)-1] |
|
392 |
} |
|
393 |
|
|
394 |
// parseFieldName parses field name and separator. |
|
395 |
func (d *Decoder) parseFieldName() (tok Token, err error) { |
|
396 |
defer func() { |
|
397 |
if err == nil && d.tryConsumeChar(':') { |
|
398 |
tok.attrs |= hasSeparator |
|
399 |
} |
|
400 |
}() |
|
401 |
|
|
402 |
// Extension or Any type URL. |
|
403 |
if d.in[0] == '[' { |
|
404 |
return d.parseTypeName() |
|
405 |
} |
|
406 |
|
|
407 |
// Identifier. |
|
408 |
if size := parseIdent(d.in, false); size > 0 { |
|
409 |
return d.consumeToken(Name, size, uint8(IdentName)), nil |
|
410 |
} |
|
411 |
|
|
412 |
// Field number. Identify if input is a valid number that is not negative |
|
413 |
// and is decimal integer within 32-bit range. |
|
414 |
if num := parseNumber(d.in); num.size > 0 { |
|
415 |
if !num.neg && num.kind == numDec { |
|
416 |
if _, err := strconv.ParseInt(string(d.in[:num.size]), 10, 32); err == nil { |
|
417 |
return d.consumeToken(Name, num.size, uint8(FieldNumber)), nil |
|
418 |
} |
|
419 |
} |
|
420 |
return Token{}, d.newSyntaxError("invalid field number: %s", d.in[:num.size]) |
|
421 |
} |
|
422 |
|
260
|
423 |
return Token{}, d.newSyntaxError("invalid field name: %s", errId(d.in)) |
256
|
424 |
} |
|
425 |
|
|
426 |
// parseTypeName parses Any type URL or extension field name. The name is |
|
427 |
// enclosed in [ and ] characters. The C++ parser does not handle many legal URL |
|
428 |
// strings. This implementation is more liberal and allows for the pattern |
|
429 |
// ^[-_a-zA-Z0-9]+([./][-_a-zA-Z0-9]+)*`). Whitespaces and comments are allowed |
|
430 |
// in between [ ], '.', '/' and the sub names. |
|
431 |
func (d *Decoder) parseTypeName() (Token, error) { |
|
432 |
startPos := len(d.orig) - len(d.in) |
|
433 |
// Use alias s to advance first in order to use d.in for error handling. |
|
434 |
// Caller already checks for [ as first character. |
|
435 |
s := consume(d.in[1:], 0) |
|
436 |
if len(s) == 0 { |
|
437 |
return Token{}, ErrUnexpectedEOF |
|
438 |
} |
|
439 |
|
|
440 |
var name []byte |
|
441 |
for len(s) > 0 && isTypeNameChar(s[0]) { |
|
442 |
name = append(name, s[0]) |
|
443 |
s = s[1:] |
|
444 |
} |
|
445 |
s = consume(s, 0) |
|
446 |
|
|
447 |
var closed bool |
|
448 |
for len(s) > 0 && !closed { |
|
449 |
switch { |
|
450 |
case s[0] == ']': |
|
451 |
s = s[1:] |
|
452 |
closed = true |
|
453 |
|
|
454 |
case s[0] == '/', s[0] == '.': |
|
455 |
if len(name) > 0 && (name[len(name)-1] == '/' || name[len(name)-1] == '.') { |
|
456 |
return Token{}, d.newSyntaxError("invalid type URL/extension field name: %s", |
|
457 |
d.orig[startPos:len(d.orig)-len(s)+1]) |
|
458 |
} |
|
459 |
name = append(name, s[0]) |
|
460 |
s = s[1:] |
|
461 |
s = consume(s, 0) |
|
462 |
for len(s) > 0 && isTypeNameChar(s[0]) { |
|
463 |
name = append(name, s[0]) |
|
464 |
s = s[1:] |
|
465 |
} |
|
466 |
s = consume(s, 0) |
|
467 |
|
|
468 |
default: |
|
469 |
return Token{}, d.newSyntaxError( |
|
470 |
"invalid type URL/extension field name: %s", d.orig[startPos:len(d.orig)-len(s)+1]) |
|
471 |
} |
|
472 |
} |
|
473 |
|
|
474 |
if !closed { |
|
475 |
return Token{}, ErrUnexpectedEOF |
|
476 |
} |
|
477 |
|
|
478 |
// First character cannot be '.'. Last character cannot be '.' or '/'. |
|
479 |
size := len(name) |
|
480 |
if size == 0 || name[0] == '.' || name[size-1] == '.' || name[size-1] == '/' { |
|
481 |
return Token{}, d.newSyntaxError("invalid type URL/extension field name: %s", |
|
482 |
d.orig[startPos:len(d.orig)-len(s)]) |
|
483 |
} |
|
484 |
|
|
485 |
d.in = s |
|
486 |
endPos := len(d.orig) - len(d.in) |
|
487 |
d.consume(0) |
|
488 |
|
|
489 |
return Token{ |
|
490 |
kind: Name, |
|
491 |
attrs: uint8(TypeName), |
|
492 |
pos: startPos, |
|
493 |
raw: d.orig[startPos:endPos], |
|
494 |
str: string(name), |
|
495 |
}, nil |
|
496 |
} |
|
497 |
|
|
498 |
func isTypeNameChar(b byte) bool { |
|
499 |
return (b == '-' || b == '_' || |
|
500 |
('0' <= b && b <= '9') || |
|
501 |
('a' <= b && b <= 'z') || |
|
502 |
('A' <= b && b <= 'Z')) |
|
503 |
} |
|
504 |
|
|
505 |
func isWhiteSpace(b byte) bool { |
|
506 |
switch b { |
|
507 |
case ' ', '\n', '\r', '\t': |
|
508 |
return true |
|
509 |
default: |
|
510 |
return false |
|
511 |
} |
|
512 |
} |
|
513 |
|
|
514 |
// parseIdent parses an unquoted proto identifier and returns size. |
|
515 |
// If allowNeg is true, it allows '-' to be the first character in the |
|
516 |
// identifier. This is used when parsing literal values like -infinity, etc. |
|
517 |
// Regular expression matches an identifier: `^[_a-zA-Z][_a-zA-Z0-9]*` |
|
518 |
func parseIdent(input []byte, allowNeg bool) int { |
|
519 |
var size int |
|
520 |
|
|
521 |
s := input |
|
522 |
if len(s) == 0 { |
|
523 |
return 0 |
|
524 |
} |
|
525 |
|
|
526 |
if allowNeg && s[0] == '-' { |
|
527 |
s = s[1:] |
|
528 |
size++ |
|
529 |
if len(s) == 0 { |
|
530 |
return 0 |
|
531 |
} |
|
532 |
} |
|
533 |
|
|
534 |
switch { |
|
535 |
case s[0] == '_', |
|
536 |
'a' <= s[0] && s[0] <= 'z', |
|
537 |
'A' <= s[0] && s[0] <= 'Z': |
|
538 |
s = s[1:] |
|
539 |
size++ |
|
540 |
default: |
|
541 |
return 0 |
|
542 |
} |
|
543 |
|
|
544 |
for len(s) > 0 && (s[0] == '_' || |
|
545 |
'a' <= s[0] && s[0] <= 'z' || |
|
546 |
'A' <= s[0] && s[0] <= 'Z' || |
|
547 |
'0' <= s[0] && s[0] <= '9') { |
|
548 |
s = s[1:] |
|
549 |
size++ |
|
550 |
} |
|
551 |
|
|
552 |
if len(s) > 0 && !isDelim(s[0]) { |
|
553 |
return 0 |
|
554 |
} |
|
555 |
|
|
556 |
return size |
|
557 |
} |
|
558 |
|
|
559 |
// parseScalar parses for a string, literal or number value. |
|
560 |
func (d *Decoder) parseScalar() (Token, error) { |
|
561 |
if d.in[0] == '"' || d.in[0] == '\'' { |
|
562 |
return d.parseStringValue() |
|
563 |
} |
|
564 |
|
|
565 |
if tok, ok := d.parseLiteralValue(); ok { |
|
566 |
return tok, nil |
|
567 |
} |
|
568 |
|
|
569 |
if tok, ok := d.parseNumberValue(); ok { |
|
570 |
return tok, nil |
|
571 |
} |
|
572 |
|
260
|
573 |
return Token{}, d.newSyntaxError("invalid scalar value: %s", errId(d.in)) |
256
|
574 |
} |
|
575 |
|
|
576 |
// parseLiteralValue parses a literal value. A literal value is used for |
|
577 |
// bools, special floats and enums. This function simply identifies that the |
|
578 |
// field value is a literal. |
|
579 |
func (d *Decoder) parseLiteralValue() (Token, bool) { |
|
580 |
size := parseIdent(d.in, true) |
|
581 |
if size == 0 { |
|
582 |
return Token{}, false |
|
583 |
} |
|
584 |
return d.consumeToken(Scalar, size, literalValue), true |
|
585 |
} |
|
586 |
|
|
587 |
// consumeToken constructs a Token for given Kind from d.in and consumes given |
|
588 |
// size-length from it. |
|
589 |
func (d *Decoder) consumeToken(kind Kind, size int, attrs uint8) Token { |
|
590 |
// Important to compute raw and pos before consuming. |
|
591 |
tok := Token{ |
|
592 |
kind: kind, |
|
593 |
attrs: attrs, |
|
594 |
pos: len(d.orig) - len(d.in), |
|
595 |
raw: d.in[:size], |
|
596 |
} |
|
597 |
d.consume(size) |
|
598 |
return tok |
|
599 |
} |
|
600 |
|
|
601 |
// newSyntaxError returns a syntax error with line and column information for |
|
602 |
// current position. |
|
603 |
func (d *Decoder) newSyntaxError(f string, x ...interface{}) error { |
|
604 |
e := errors.New(f, x...) |
|
605 |
line, column := d.Position(len(d.orig) - len(d.in)) |
|
606 |
return errors.New("syntax error (line %d:%d): %v", line, column, e) |
|
607 |
} |
|
608 |
|
|
609 |
// Position returns line and column number of given index of the original input. |
|
610 |
// It will panic if index is out of range. |
|
611 |
func (d *Decoder) Position(idx int) (line int, column int) { |
|
612 |
b := d.orig[:idx] |
|
613 |
line = bytes.Count(b, []byte("\n")) + 1 |
|
614 |
if i := bytes.LastIndexByte(b, '\n'); i >= 0 { |
|
615 |
b = b[i+1:] |
|
616 |
} |
|
617 |
column = utf8.RuneCount(b) + 1 // ignore multi-rune characters |
|
618 |
return line, column |
|
619 |
} |
|
620 |
|
|
621 |
func (d *Decoder) tryConsumeChar(c byte) bool { |
|
622 |
if len(d.in) > 0 && d.in[0] == c { |
|
623 |
d.consume(1) |
|
624 |
return true |
|
625 |
} |
|
626 |
return false |
|
627 |
} |
|
628 |
|
|
629 |
// consume consumes n bytes of input and any subsequent whitespace or comments. |
|
630 |
func (d *Decoder) consume(n int) { |
|
631 |
d.in = consume(d.in, n) |
|
632 |
return |
|
633 |
} |
|
634 |
|
|
635 |
// consume consumes n bytes of input and any subsequent whitespace or comments. |
|
636 |
func consume(b []byte, n int) []byte { |
|
637 |
b = b[n:] |
|
638 |
for len(b) > 0 { |
|
639 |
switch b[0] { |
|
640 |
case ' ', '\n', '\r', '\t': |
|
641 |
b = b[1:] |
|
642 |
case '#': |
|
643 |
if i := bytes.IndexByte(b, '\n'); i >= 0 { |
|
644 |
b = b[i+len("\n"):] |
|
645 |
} else { |
|
646 |
b = nil |
|
647 |
} |
|
648 |
default: |
|
649 |
return b |
|
650 |
} |
|
651 |
} |
|
652 |
return b |
|
653 |
} |
|
654 |
|
260
|
655 |
// errId extracts a byte sequence that looks like an invalid ID |
|
656 |
// (for the purposes of error reporting). |
|
657 |
func errId(seq []byte) []byte { |
|
658 |
const maxLen = 32 |
|
659 |
for i := 0; i < len(seq); { |
|
660 |
if i > maxLen { |
|
661 |
return append(seq[:i:i], "…"...) |
|
662 |
} |
|
663 |
r, size := utf8.DecodeRune(seq[i:]) |
|
664 |
if r > utf8.RuneSelf || (r != '/' && isDelim(byte(r))) { |
|
665 |
if i == 0 { |
|
666 |
// Either the first byte is invalid UTF-8 or a |
|
667 |
// delimiter, or the first rune is non-ASCII. |
|
668 |
// Return it as-is. |
|
669 |
i = size |
|
670 |
} |
|
671 |
return seq[:i:i] |
|
672 |
} |
|
673 |
i += size |
|
674 |
} |
|
675 |
// No delimiter found. |
|
676 |
return seq |
|
677 |
} |
256
|
678 |
|
|
679 |
// isDelim returns true if given byte is a delimiter character. |
|
680 |
func isDelim(c byte) bool { |
|
681 |
return !(c == '-' || c == '+' || c == '.' || c == '_' || |
|
682 |
('a' <= c && c <= 'z') || |
|
683 |
('A' <= c && c <= 'Z') || |
|
684 |
('0' <= c && c <= '9')) |
|
685 |
} |