167 * @param fac the factory which created this Scanner |
63 * @param fac the factory which created this Scanner |
168 * @param input the input, might be modified |
64 * @param input the input, might be modified |
169 * @param inputLength the size of the input. |
65 * @param inputLength the size of the input. |
170 * Must be positive and less than or equal to input.length. |
66 * Must be positive and less than or equal to input.length. |
171 */ |
67 */ |
172 protected Scanner(ScannerFactory fac, char[] input, int inputLength) { |
68 protected Scanner(ScannerFactory fac, CharBuffer buf) { |
173 this(fac); |
69 this(fac, new JavaTokenizer(fac, buf)); |
174 eofPos = inputLength; |
|
175 if (inputLength == input.length) { |
|
176 if (input.length > 0 && Character.isWhitespace(input[input.length - 1])) { |
|
177 inputLength--; |
|
178 } else { |
|
179 char[] newInput = new char[inputLength + 1]; |
|
180 System.arraycopy(input, 0, newInput, 0, input.length); |
|
181 input = newInput; |
|
182 } |
|
183 } |
|
184 buf = input; |
|
185 buflen = inputLength; |
|
186 buf[buflen] = EOI; |
|
187 bp = -1; |
|
188 scanChar(); |
|
189 } |
70 } |
190 |
71 |
191 /** Report an error at the given position using the provided arguments. |
72 protected Scanner(ScannerFactory fac, char[] buf, int inputLength) { |
192 */ |
73 this(fac, new JavaTokenizer(fac, buf, inputLength)); |
193 private void lexError(int pos, String key, Object... args) { |
|
194 log.error(pos, key, args); |
|
195 token = ERROR; |
|
196 errPos = pos; |
|
197 } |
74 } |
198 |
75 |
199 /** Report an error at the current token position using the provided |
76 protected Scanner(ScannerFactory fac, JavaTokenizer tokenizer) { |
200 * arguments. |
77 this.tokenizer = tokenizer; |
201 */ |
78 tokens = fac.tokens; |
202 private void lexError(String key, Object... args) { |
79 token = prevToken = DUMMY; |
203 lexError(pos, key, args); |
|
204 } |
80 } |
205 |
81 |
206 /** Convert an ASCII digit from its base (8, 10, or 16) |
|
207 * to its value. |
|
208 */ |
|
209 private int digit(int base) { |
|
210 char c = ch; |
|
211 int result = Character.digit(c, base); |
|
212 if (result >= 0 && c > 0x7f) { |
|
213 lexError(pos+1, "illegal.nonascii.digit"); |
|
214 ch = "0123456789abcdef".charAt(result); |
|
215 } |
|
216 return result; |
|
217 } |
|
218 |
|
219 /** Convert unicode escape; bp points to initial '\' character |
|
220 * (Spec 3.3). |
|
221 */ |
|
222 private void convertUnicode() { |
|
223 if (ch == '\\' && unicodeConversionBp != bp) { |
|
224 bp++; ch = buf[bp]; |
|
225 if (ch == 'u') { |
|
226 do { |
|
227 bp++; ch = buf[bp]; |
|
228 } while (ch == 'u'); |
|
229 int limit = bp + 3; |
|
230 if (limit < buflen) { |
|
231 int d = digit(16); |
|
232 int code = d; |
|
233 while (bp < limit && d >= 0) { |
|
234 bp++; ch = buf[bp]; |
|
235 d = digit(16); |
|
236 code = (code << 4) + d; |
|
237 } |
|
238 if (d >= 0) { |
|
239 ch = (char)code; |
|
240 unicodeConversionBp = bp; |
|
241 return; |
|
242 } |
|
243 } |
|
244 lexError(bp, "illegal.unicode.esc"); |
|
245 } else { |
|
246 bp--; |
|
247 ch = '\\'; |
|
248 } |
|
249 } |
|
250 } |
|
251 |
|
252 /** Read next character. |
|
253 */ |
|
254 private void scanChar() { |
|
255 ch = buf[++bp]; |
|
256 if (ch == '\\') { |
|
257 convertUnicode(); |
|
258 } |
|
259 } |
|
260 |
|
261 /** Read next character in comment, skipping over double '\' characters. |
|
262 */ |
|
263 private void scanCommentChar() { |
|
264 scanChar(); |
|
265 if (ch == '\\') { |
|
266 if (buf[bp+1] == '\\' && unicodeConversionBp != bp) { |
|
267 bp++; |
|
268 } else { |
|
269 convertUnicode(); |
|
270 } |
|
271 } |
|
272 } |
|
273 |
|
274 /** Append a character to sbuf. |
|
275 */ |
|
276 private void putChar(char ch) { |
|
277 if (sp == sbuf.length) { |
|
278 char[] newsbuf = new char[sbuf.length * 2]; |
|
279 System.arraycopy(sbuf, 0, newsbuf, 0, sbuf.length); |
|
280 sbuf = newsbuf; |
|
281 } |
|
282 sbuf[sp++] = ch; |
|
283 } |
|
284 |
|
285 /** Read next character in character or string literal and copy into sbuf. |
|
286 */ |
|
287 private void scanLitChar() { |
|
288 if (ch == '\\') { |
|
289 if (buf[bp+1] == '\\' && unicodeConversionBp != bp) { |
|
290 bp++; |
|
291 putChar('\\'); |
|
292 scanChar(); |
|
293 } else { |
|
294 scanChar(); |
|
295 switch (ch) { |
|
296 case '0': case '1': case '2': case '3': |
|
297 case '4': case '5': case '6': case '7': |
|
298 char leadch = ch; |
|
299 int oct = digit(8); |
|
300 scanChar(); |
|
301 if ('0' <= ch && ch <= '7') { |
|
302 oct = oct * 8 + digit(8); |
|
303 scanChar(); |
|
304 if (leadch <= '3' && '0' <= ch && ch <= '7') { |
|
305 oct = oct * 8 + digit(8); |
|
306 scanChar(); |
|
307 } |
|
308 } |
|
309 putChar((char)oct); |
|
310 break; |
|
311 case 'b': |
|
312 putChar('\b'); scanChar(); break; |
|
313 case 't': |
|
314 putChar('\t'); scanChar(); break; |
|
315 case 'n': |
|
316 putChar('\n'); scanChar(); break; |
|
317 case 'f': |
|
318 putChar('\f'); scanChar(); break; |
|
319 case 'r': |
|
320 putChar('\r'); scanChar(); break; |
|
321 case '\'': |
|
322 putChar('\''); scanChar(); break; |
|
323 case '\"': |
|
324 putChar('\"'); scanChar(); break; |
|
325 case '\\': |
|
326 putChar('\\'); scanChar(); break; |
|
327 default: |
|
328 lexError(bp, "illegal.esc.char"); |
|
329 } |
|
330 } |
|
331 } else if (bp != buflen) { |
|
332 putChar(ch); scanChar(); |
|
333 } |
|
334 } |
|
335 |
|
336 private void scanDigits(int digitRadix) { |
|
337 char saveCh; |
|
338 int savePos; |
|
339 do { |
|
340 if (ch != '_') { |
|
341 putChar(ch); |
|
342 } else { |
|
343 if (!allowUnderscoresInLiterals) { |
|
344 lexError("unsupported.underscore.lit", source.name); |
|
345 allowUnderscoresInLiterals = true; |
|
346 } |
|
347 } |
|
348 saveCh = ch; |
|
349 savePos = bp; |
|
350 scanChar(); |
|
351 } while (digit(digitRadix) >= 0 || ch == '_'); |
|
352 if (saveCh == '_') |
|
353 lexError(savePos, "illegal.underscore"); |
|
354 } |
|
355 |
|
356 /** Read fractional part of hexadecimal floating point number. |
|
357 */ |
|
358 private void scanHexExponentAndSuffix() { |
|
359 if (ch == 'p' || ch == 'P') { |
|
360 putChar(ch); |
|
361 scanChar(); |
|
362 skipIllegalUnderscores(); |
|
363 if (ch == '+' || ch == '-') { |
|
364 putChar(ch); |
|
365 scanChar(); |
|
366 } |
|
367 skipIllegalUnderscores(); |
|
368 if ('0' <= ch && ch <= '9') { |
|
369 scanDigits(10); |
|
370 if (!allowHexFloats) { |
|
371 lexError("unsupported.fp.lit", source.name); |
|
372 allowHexFloats = true; |
|
373 } |
|
374 else if (!hexFloatsWork) |
|
375 lexError("unsupported.cross.fp.lit"); |
|
376 } else |
|
377 lexError("malformed.fp.lit"); |
|
378 } else { |
|
379 lexError("malformed.fp.lit"); |
|
380 } |
|
381 if (ch == 'f' || ch == 'F') { |
|
382 putChar(ch); |
|
383 scanChar(); |
|
384 token = FLOATLITERAL; |
|
385 } else { |
|
386 if (ch == 'd' || ch == 'D') { |
|
387 putChar(ch); |
|
388 scanChar(); |
|
389 } |
|
390 token = DOUBLELITERAL; |
|
391 } |
|
392 } |
|
393 |
|
394 /** Read fractional part of floating point number. |
|
395 */ |
|
396 private void scanFraction() { |
|
397 skipIllegalUnderscores(); |
|
398 if ('0' <= ch && ch <= '9') { |
|
399 scanDigits(10); |
|
400 } |
|
401 int sp1 = sp; |
|
402 if (ch == 'e' || ch == 'E') { |
|
403 putChar(ch); |
|
404 scanChar(); |
|
405 skipIllegalUnderscores(); |
|
406 if (ch == '+' || ch == '-') { |
|
407 putChar(ch); |
|
408 scanChar(); |
|
409 } |
|
410 skipIllegalUnderscores(); |
|
411 if ('0' <= ch && ch <= '9') { |
|
412 scanDigits(10); |
|
413 return; |
|
414 } |
|
415 lexError("malformed.fp.lit"); |
|
416 sp = sp1; |
|
417 } |
|
418 } |
|
419 |
|
420 /** Read fractional part and 'd' or 'f' suffix of floating point number. |
|
421 */ |
|
422 private void scanFractionAndSuffix() { |
|
423 this.radix = 10; |
|
424 scanFraction(); |
|
425 if (ch == 'f' || ch == 'F') { |
|
426 putChar(ch); |
|
427 scanChar(); |
|
428 token = FLOATLITERAL; |
|
429 } else { |
|
430 if (ch == 'd' || ch == 'D') { |
|
431 putChar(ch); |
|
432 scanChar(); |
|
433 } |
|
434 token = DOUBLELITERAL; |
|
435 } |
|
436 } |
|
437 |
|
438 /** Read fractional part and 'd' or 'f' suffix of floating point number. |
|
439 */ |
|
440 private void scanHexFractionAndSuffix(boolean seendigit) { |
|
441 this.radix = 16; |
|
442 Assert.check(ch == '.'); |
|
443 putChar(ch); |
|
444 scanChar(); |
|
445 skipIllegalUnderscores(); |
|
446 if (digit(16) >= 0) { |
|
447 seendigit = true; |
|
448 scanDigits(16); |
|
449 } |
|
450 if (!seendigit) |
|
451 lexError("invalid.hex.number"); |
|
452 else |
|
453 scanHexExponentAndSuffix(); |
|
454 } |
|
455 |
|
456 private void skipIllegalUnderscores() { |
|
457 if (ch == '_') { |
|
458 lexError(bp, "illegal.underscore"); |
|
459 while (ch == '_') |
|
460 scanChar(); |
|
461 } |
|
462 } |
|
463 |
|
464 /** Read a number. |
|
465 * @param radix The radix of the number; one of 2, j8, 10, 16. |
|
466 */ |
|
467 private void scanNumber(int radix) { |
|
468 this.radix = radix; |
|
469 // for octal, allow base-10 digit in case it's a float literal |
|
470 int digitRadix = (radix == 8 ? 10 : radix); |
|
471 boolean seendigit = false; |
|
472 if (digit(digitRadix) >= 0) { |
|
473 seendigit = true; |
|
474 scanDigits(digitRadix); |
|
475 } |
|
476 if (radix == 16 && ch == '.') { |
|
477 scanHexFractionAndSuffix(seendigit); |
|
478 } else if (seendigit && radix == 16 && (ch == 'p' || ch == 'P')) { |
|
479 scanHexExponentAndSuffix(); |
|
480 } else if (digitRadix == 10 && ch == '.') { |
|
481 putChar(ch); |
|
482 scanChar(); |
|
483 scanFractionAndSuffix(); |
|
484 } else if (digitRadix == 10 && |
|
485 (ch == 'e' || ch == 'E' || |
|
486 ch == 'f' || ch == 'F' || |
|
487 ch == 'd' || ch == 'D')) { |
|
488 scanFractionAndSuffix(); |
|
489 } else { |
|
490 if (ch == 'l' || ch == 'L') { |
|
491 scanChar(); |
|
492 token = LONGLITERAL; |
|
493 } else { |
|
494 token = INTLITERAL; |
|
495 } |
|
496 } |
|
497 } |
|
498 |
|
499 /** Read an identifier. |
|
500 */ |
|
501 private void scanIdent() { |
|
502 boolean isJavaIdentifierPart; |
|
503 char high; |
|
504 do { |
|
505 if (sp == sbuf.length) putChar(ch); else sbuf[sp++] = ch; |
|
506 // optimization, was: putChar(ch); |
|
507 |
|
508 scanChar(); |
|
509 switch (ch) { |
|
510 case 'A': case 'B': case 'C': case 'D': case 'E': |
|
511 case 'F': case 'G': case 'H': case 'I': case 'J': |
|
512 case 'K': case 'L': case 'M': case 'N': case 'O': |
|
513 case 'P': case 'Q': case 'R': case 'S': case 'T': |
|
514 case 'U': case 'V': case 'W': case 'X': case 'Y': |
|
515 case 'Z': |
|
516 case 'a': case 'b': case 'c': case 'd': case 'e': |
|
517 case 'f': case 'g': case 'h': case 'i': case 'j': |
|
518 case 'k': case 'l': case 'm': case 'n': case 'o': |
|
519 case 'p': case 'q': case 'r': case 's': case 't': |
|
520 case 'u': case 'v': case 'w': case 'x': case 'y': |
|
521 case 'z': |
|
522 case '$': case '_': |
|
523 case '0': case '1': case '2': case '3': case '4': |
|
524 case '5': case '6': case '7': case '8': case '9': |
|
525 case '\u0000': case '\u0001': case '\u0002': case '\u0003': |
|
526 case '\u0004': case '\u0005': case '\u0006': case '\u0007': |
|
527 case '\u0008': case '\u000E': case '\u000F': case '\u0010': |
|
528 case '\u0011': case '\u0012': case '\u0013': case '\u0014': |
|
529 case '\u0015': case '\u0016': case '\u0017': |
|
530 case '\u0018': case '\u0019': case '\u001B': |
|
531 case '\u007F': |
|
532 break; |
|
533 case '\u001A': // EOI is also a legal identifier part |
|
534 if (bp >= buflen) { |
|
535 name = names.fromChars(sbuf, 0, sp); |
|
536 token = keywords.key(name); |
|
537 return; |
|
538 } |
|
539 break; |
|
540 default: |
|
541 if (ch < '\u0080') { |
|
542 // all ASCII range chars already handled, above |
|
543 isJavaIdentifierPart = false; |
|
544 } else { |
|
545 high = scanSurrogates(); |
|
546 if (high != 0) { |
|
547 if (sp == sbuf.length) { |
|
548 putChar(high); |
|
549 } else { |
|
550 sbuf[sp++] = high; |
|
551 } |
|
552 isJavaIdentifierPart = Character.isJavaIdentifierPart( |
|
553 Character.toCodePoint(high, ch)); |
|
554 } else { |
|
555 isJavaIdentifierPart = Character.isJavaIdentifierPart(ch); |
|
556 } |
|
557 } |
|
558 if (!isJavaIdentifierPart) { |
|
559 name = names.fromChars(sbuf, 0, sp); |
|
560 token = keywords.key(name); |
|
561 return; |
|
562 } |
|
563 } |
|
564 } while (true); |
|
565 } |
|
566 |
|
567 /** Are surrogates supported? |
|
568 */ |
|
569 final static boolean surrogatesSupported = surrogatesSupported(); |
|
570 private static boolean surrogatesSupported() { |
|
571 try { |
|
572 Character.isHighSurrogate('a'); |
|
573 return true; |
|
574 } catch (NoSuchMethodError ex) { |
|
575 return false; |
|
576 } |
|
577 } |
|
578 |
|
579 /** Scan surrogate pairs. If 'ch' is a high surrogate and |
|
580 * the next character is a low surrogate, then put the low |
|
581 * surrogate in 'ch', and return the high surrogate. |
|
582 * otherwise, just return 0. |
|
583 */ |
|
584 private char scanSurrogates() { |
|
585 if (surrogatesSupported && Character.isHighSurrogate(ch)) { |
|
586 char high = ch; |
|
587 |
|
588 scanChar(); |
|
589 |
|
590 if (Character.isLowSurrogate(ch)) { |
|
591 return high; |
|
592 } |
|
593 |
|
594 ch = high; |
|
595 } |
|
596 |
|
597 return 0; |
|
598 } |
|
599 |
|
600 /** Return true if ch can be part of an operator. |
|
601 */ |
|
602 private boolean isSpecial(char ch) { |
|
603 switch (ch) { |
|
604 case '!': case '%': case '&': case '*': case '?': |
|
605 case '+': case '-': case ':': case '<': case '=': |
|
606 case '>': case '^': case '|': case '~': |
|
607 case '@': |
|
608 return true; |
|
609 default: |
|
610 return false; |
|
611 } |
|
612 } |
|
613 |
|
614 /** Read longest possible sequence of special characters and convert |
|
615 * to token. |
|
616 */ |
|
617 private void scanOperator() { |
|
618 while (true) { |
|
619 putChar(ch); |
|
620 Name newname = names.fromChars(sbuf, 0, sp); |
|
621 if (keywords.key(newname) == IDENTIFIER) { |
|
622 sp--; |
|
623 break; |
|
624 } |
|
625 name = newname; |
|
626 token = keywords.key(newname); |
|
627 scanChar(); |
|
628 if (!isSpecial(ch)) break; |
|
629 } |
|
630 } |
|
631 |
|
632 /** |
|
633 * Scan a documention comment; determine if a deprecated tag is present. |
|
634 * Called once the initial /, * have been skipped, positioned at the second * |
|
635 * (which is treated as the beginning of the first line). |
|
636 * Stops positioned at the closing '/'. |
|
637 */ |
|
638 @SuppressWarnings("fallthrough") |
|
639 private void scanDocComment() { |
|
640 boolean deprecatedPrefix = false; |
|
641 |
|
642 forEachLine: |
|
643 while (bp < buflen) { |
|
644 |
|
645 // Skip optional WhiteSpace at beginning of line |
|
646 while (bp < buflen && (ch == ' ' || ch == '\t' || ch == FF)) { |
|
647 scanCommentChar(); |
|
648 } |
|
649 |
|
650 // Skip optional consecutive Stars |
|
651 while (bp < buflen && ch == '*') { |
|
652 scanCommentChar(); |
|
653 if (ch == '/') { |
|
654 return; |
|
655 } |
|
656 } |
|
657 |
|
658 // Skip optional WhiteSpace after Stars |
|
659 while (bp < buflen && (ch == ' ' || ch == '\t' || ch == FF)) { |
|
660 scanCommentChar(); |
|
661 } |
|
662 |
|
663 deprecatedPrefix = false; |
|
664 // At beginning of line in the JavaDoc sense. |
|
665 if (bp < buflen && ch == '@' && !deprecatedFlag) { |
|
666 scanCommentChar(); |
|
667 if (bp < buflen && ch == 'd') { |
|
668 scanCommentChar(); |
|
669 if (bp < buflen && ch == 'e') { |
|
670 scanCommentChar(); |
|
671 if (bp < buflen && ch == 'p') { |
|
672 scanCommentChar(); |
|
673 if (bp < buflen && ch == 'r') { |
|
674 scanCommentChar(); |
|
675 if (bp < buflen && ch == 'e') { |
|
676 scanCommentChar(); |
|
677 if (bp < buflen && ch == 'c') { |
|
678 scanCommentChar(); |
|
679 if (bp < buflen && ch == 'a') { |
|
680 scanCommentChar(); |
|
681 if (bp < buflen && ch == 't') { |
|
682 scanCommentChar(); |
|
683 if (bp < buflen && ch == 'e') { |
|
684 scanCommentChar(); |
|
685 if (bp < buflen && ch == 'd') { |
|
686 deprecatedPrefix = true; |
|
687 scanCommentChar(); |
|
688 }}}}}}}}}}} |
|
689 if (deprecatedPrefix && bp < buflen) { |
|
690 if (Character.isWhitespace(ch)) { |
|
691 deprecatedFlag = true; |
|
692 } else if (ch == '*') { |
|
693 scanCommentChar(); |
|
694 if (ch == '/') { |
|
695 deprecatedFlag = true; |
|
696 return; |
|
697 } |
|
698 } |
|
699 } |
|
700 |
|
701 // Skip rest of line |
|
702 while (bp < buflen) { |
|
703 switch (ch) { |
|
704 case '*': |
|
705 scanCommentChar(); |
|
706 if (ch == '/') { |
|
707 return; |
|
708 } |
|
709 break; |
|
710 case CR: // (Spec 3.4) |
|
711 scanCommentChar(); |
|
712 if (ch != LF) { |
|
713 continue forEachLine; |
|
714 } |
|
715 /* fall through to LF case */ |
|
716 case LF: // (Spec 3.4) |
|
717 scanCommentChar(); |
|
718 continue forEachLine; |
|
719 default: |
|
720 scanCommentChar(); |
|
721 } |
|
722 } // rest of line |
|
723 } // forEachLine |
|
724 return; |
|
725 } |
|
726 |
|
727 /** The value of a literal token, recorded as a string. |
|
728 * For integers, leading 0x and 'l' suffixes are suppressed. |
|
729 */ |
|
730 public String stringVal() { |
|
731 return new String(sbuf, 0, sp); |
|
732 } |
|
733 |
|
734 /** Read token. |
|
735 */ |
|
736 public void nextToken() { |
|
737 |
|
738 try { |
|
739 prevEndPos = endPos; |
|
740 sp = 0; |
|
741 |
|
742 while (true) { |
|
743 pos = bp; |
|
744 switch (ch) { |
|
745 case ' ': // (Spec 3.6) |
|
746 case '\t': // (Spec 3.6) |
|
747 case FF: // (Spec 3.6) |
|
748 do { |
|
749 scanChar(); |
|
750 } while (ch == ' ' || ch == '\t' || ch == FF); |
|
751 endPos = bp; |
|
752 processWhiteSpace(); |
|
753 break; |
|
754 case LF: // (Spec 3.4) |
|
755 scanChar(); |
|
756 endPos = bp; |
|
757 processLineTerminator(); |
|
758 break; |
|
759 case CR: // (Spec 3.4) |
|
760 scanChar(); |
|
761 if (ch == LF) { |
|
762 scanChar(); |
|
763 } |
|
764 endPos = bp; |
|
765 processLineTerminator(); |
|
766 break; |
|
767 case 'A': case 'B': case 'C': case 'D': case 'E': |
|
768 case 'F': case 'G': case 'H': case 'I': case 'J': |
|
769 case 'K': case 'L': case 'M': case 'N': case 'O': |
|
770 case 'P': case 'Q': case 'R': case 'S': case 'T': |
|
771 case 'U': case 'V': case 'W': case 'X': case 'Y': |
|
772 case 'Z': |
|
773 case 'a': case 'b': case 'c': case 'd': case 'e': |
|
774 case 'f': case 'g': case 'h': case 'i': case 'j': |
|
775 case 'k': case 'l': case 'm': case 'n': case 'o': |
|
776 case 'p': case 'q': case 'r': case 's': case 't': |
|
777 case 'u': case 'v': case 'w': case 'x': case 'y': |
|
778 case 'z': |
|
779 case '$': case '_': |
|
780 scanIdent(); |
|
781 return; |
|
782 case '0': |
|
783 scanChar(); |
|
784 if (ch == 'x' || ch == 'X') { |
|
785 scanChar(); |
|
786 skipIllegalUnderscores(); |
|
787 if (ch == '.') { |
|
788 scanHexFractionAndSuffix(false); |
|
789 } else if (digit(16) < 0) { |
|
790 lexError("invalid.hex.number"); |
|
791 } else { |
|
792 scanNumber(16); |
|
793 } |
|
794 } else if (ch == 'b' || ch == 'B') { |
|
795 if (!allowBinaryLiterals) { |
|
796 lexError("unsupported.binary.lit", source.name); |
|
797 allowBinaryLiterals = true; |
|
798 } |
|
799 scanChar(); |
|
800 skipIllegalUnderscores(); |
|
801 if (digit(2) < 0) { |
|
802 lexError("invalid.binary.number"); |
|
803 } else { |
|
804 scanNumber(2); |
|
805 } |
|
806 } else { |
|
807 putChar('0'); |
|
808 if (ch == '_') { |
|
809 int savePos = bp; |
|
810 do { |
|
811 scanChar(); |
|
812 } while (ch == '_'); |
|
813 if (digit(10) < 0) { |
|
814 lexError(savePos, "illegal.underscore"); |
|
815 } |
|
816 } |
|
817 scanNumber(8); |
|
818 } |
|
819 return; |
|
820 case '1': case '2': case '3': case '4': |
|
821 case '5': case '6': case '7': case '8': case '9': |
|
822 scanNumber(10); |
|
823 return; |
|
824 case '.': |
|
825 scanChar(); |
|
826 if ('0' <= ch && ch <= '9') { |
|
827 putChar('.'); |
|
828 scanFractionAndSuffix(); |
|
829 } else if (ch == '.') { |
|
830 putChar('.'); putChar('.'); |
|
831 scanChar(); |
|
832 if (ch == '.') { |
|
833 scanChar(); |
|
834 putChar('.'); |
|
835 token = ELLIPSIS; |
|
836 } else { |
|
837 lexError("malformed.fp.lit"); |
|
838 } |
|
839 } else { |
|
840 token = DOT; |
|
841 } |
|
842 return; |
|
843 case ',': |
|
844 scanChar(); token = COMMA; return; |
|
845 case ';': |
|
846 scanChar(); token = SEMI; return; |
|
847 case '(': |
|
848 scanChar(); token = LPAREN; return; |
|
849 case ')': |
|
850 scanChar(); token = RPAREN; return; |
|
851 case '[': |
|
852 scanChar(); token = LBRACKET; return; |
|
853 case ']': |
|
854 scanChar(); token = RBRACKET; return; |
|
855 case '{': |
|
856 scanChar(); token = LBRACE; return; |
|
857 case '}': |
|
858 scanChar(); token = RBRACE; return; |
|
859 case '/': |
|
860 scanChar(); |
|
861 if (ch == '/') { |
|
862 do { |
|
863 scanCommentChar(); |
|
864 } while (ch != CR && ch != LF && bp < buflen); |
|
865 if (bp < buflen) { |
|
866 endPos = bp; |
|
867 processComment(CommentStyle.LINE); |
|
868 } |
|
869 break; |
|
870 } else if (ch == '*') { |
|
871 scanChar(); |
|
872 CommentStyle style; |
|
873 if (ch == '*') { |
|
874 style = CommentStyle.JAVADOC; |
|
875 scanDocComment(); |
|
876 } else { |
|
877 style = CommentStyle.BLOCK; |
|
878 while (bp < buflen) { |
|
879 if (ch == '*') { |
|
880 scanChar(); |
|
881 if (ch == '/') break; |
|
882 } else { |
|
883 scanCommentChar(); |
|
884 } |
|
885 } |
|
886 } |
|
887 if (ch == '/') { |
|
888 scanChar(); |
|
889 endPos = bp; |
|
890 processComment(style); |
|
891 break; |
|
892 } else { |
|
893 lexError("unclosed.comment"); |
|
894 return; |
|
895 } |
|
896 } else if (ch == '=') { |
|
897 name = names.slashequals; |
|
898 token = SLASHEQ; |
|
899 scanChar(); |
|
900 } else { |
|
901 name = names.slash; |
|
902 token = SLASH; |
|
903 } |
|
904 return; |
|
905 case '\'': |
|
906 scanChar(); |
|
907 if (ch == '\'') { |
|
908 lexError("empty.char.lit"); |
|
909 } else { |
|
910 if (ch == CR || ch == LF) |
|
911 lexError(pos, "illegal.line.end.in.char.lit"); |
|
912 scanLitChar(); |
|
913 if (ch == '\'') { |
|
914 scanChar(); |
|
915 token = CHARLITERAL; |
|
916 } else { |
|
917 lexError(pos, "unclosed.char.lit"); |
|
918 } |
|
919 } |
|
920 return; |
|
921 case '\"': |
|
922 scanChar(); |
|
923 while (ch != '\"' && ch != CR && ch != LF && bp < buflen) |
|
924 scanLitChar(); |
|
925 if (ch == '\"') { |
|
926 token = STRINGLITERAL; |
|
927 scanChar(); |
|
928 } else { |
|
929 lexError(pos, "unclosed.str.lit"); |
|
930 } |
|
931 return; |
|
932 default: |
|
933 if (isSpecial(ch)) { |
|
934 scanOperator(); |
|
935 } else { |
|
936 boolean isJavaIdentifierStart; |
|
937 if (ch < '\u0080') { |
|
938 // all ASCII range chars already handled, above |
|
939 isJavaIdentifierStart = false; |
|
940 } else { |
|
941 char high = scanSurrogates(); |
|
942 if (high != 0) { |
|
943 if (sp == sbuf.length) { |
|
944 putChar(high); |
|
945 } else { |
|
946 sbuf[sp++] = high; |
|
947 } |
|
948 |
|
949 isJavaIdentifierStart = Character.isJavaIdentifierStart( |
|
950 Character.toCodePoint(high, ch)); |
|
951 } else { |
|
952 isJavaIdentifierStart = Character.isJavaIdentifierStart(ch); |
|
953 } |
|
954 } |
|
955 if (isJavaIdentifierStart) { |
|
956 scanIdent(); |
|
957 } else if (bp == buflen || ch == EOI && bp+1 == buflen) { // JLS 3.5 |
|
958 token = EOF; |
|
959 pos = bp = eofPos; |
|
960 } else { |
|
961 lexError("illegal.char", String.valueOf((int)ch)); |
|
962 scanChar(); |
|
963 } |
|
964 } |
|
965 return; |
|
966 } |
|
967 } |
|
968 } finally { |
|
969 endPos = bp; |
|
970 if (scannerDebug) |
|
971 System.out.println("nextToken(" + pos |
|
972 + "," + endPos + ")=|" + |
|
973 new String(getRawCharacters(pos, endPos)) |
|
974 + "|"); |
|
975 } |
|
976 } |
|
977 |
|
978 /** Return the current token, set by nextToken(). |
|
979 */ |
|
980 public Token token() { |
82 public Token token() { |
981 return token; |
83 return token; |
982 } |
84 } |
983 |
85 |
984 /** Sets the current token. |
86 public Token prevToken() { |
985 * This method is primarily used to update the token stream when the |
87 return prevToken; |
986 * parser is handling the end of nested type arguments such as |
|
987 * {@code List<List<String>>} and needs to disambiguate between |
|
988 * repeated use of ">" and relation operators such as ">>" and ">>>". Noting |
|
989 * that this does not handle arbitrary tokens containing Unicode escape |
|
990 * sequences. |
|
991 */ |
|
992 public void token(Token token) { |
|
993 pos += this.token.name.length() - token.name.length(); |
|
994 prevEndPos = pos; |
|
995 this.token = token; |
|
996 } |
88 } |
997 |
89 |
998 /** Return the current token's position: a 0-based |
90 public void nextToken() { |
999 * offset from beginning of the raw input stream |
91 prevToken = token; |
1000 * (before unicode translation) |
92 token = tokenizer.readToken(); |
1001 */ |
|
1002 public int pos() { |
|
1003 return pos; |
|
1004 } |
93 } |
1005 |
94 |
1006 /** Return the last character position of the current token. |
95 public Token split() { |
1007 */ |
96 Token[] splitTokens = token.split(tokens); |
1008 public int endPos() { |
97 prevToken = splitTokens[0]; |
1009 return endPos; |
98 token = splitTokens[1]; |
|
99 return token; |
1010 } |
100 } |
1011 |
101 |
1012 /** Return the last character position of the previous token. |
102 public LineMap getLineMap() { |
1013 */ |
103 return tokenizer.getLineMap(); |
1014 public int prevEndPos() { |
|
1015 return prevEndPos; |
|
1016 } |
104 } |
1017 |
105 |
1018 /** Return the position where a lexical error occurred; |
|
1019 */ |
|
1020 public int errPos() { |
106 public int errPos() { |
1021 return errPos; |
107 return tokenizer.errPos(); |
1022 } |
108 } |
1023 |
109 |
1024 /** Set the position where a lexical error occurred; |
|
1025 */ |
|
1026 public void errPos(int pos) { |
110 public void errPos(int pos) { |
1027 errPos = pos; |
111 tokenizer.errPos(pos); |
1028 } |
112 } |
1029 |
|
1030 /** Return the name of an identifier or token for the current token. |
|
1031 */ |
|
1032 public Name name() { |
|
1033 return name; |
|
1034 } |
|
1035 |
|
1036 /** Return the radix of a numeric literal token. |
|
1037 */ |
|
1038 public int radix() { |
|
1039 return radix; |
|
1040 } |
|
1041 |
|
1042 /** Has a @deprecated been encountered in last doc comment? |
|
1043 * This needs to be reset by client with resetDeprecatedFlag. |
|
1044 */ |
|
1045 public boolean deprecatedFlag() { |
|
1046 return deprecatedFlag; |
|
1047 } |
|
1048 |
|
1049 public void resetDeprecatedFlag() { |
|
1050 deprecatedFlag = false; |
|
1051 } |
|
1052 |
|
1053 /** |
|
1054 * Returns the documentation string of the current token. |
|
1055 */ |
|
1056 public String docComment() { |
|
1057 return null; |
|
1058 } |
|
1059 |
|
1060 /** |
|
1061 * Returns a copy of the input buffer, up to its inputLength. |
|
1062 * Unicode escape sequences are not translated. |
|
1063 */ |
|
1064 public char[] getRawCharacters() { |
|
1065 char[] chars = new char[buflen]; |
|
1066 System.arraycopy(buf, 0, chars, 0, buflen); |
|
1067 return chars; |
|
1068 } |
|
1069 |
|
1070 /** |
|
1071 * Returns a copy of a character array subset of the input buffer. |
|
1072 * The returned array begins at the <code>beginIndex</code> and |
|
1073 * extends to the character at index <code>endIndex - 1</code>. |
|
1074 * Thus the length of the substring is <code>endIndex-beginIndex</code>. |
|
1075 * This behavior is like |
|
1076 * <code>String.substring(beginIndex, endIndex)</code>. |
|
1077 * Unicode escape sequences are not translated. |
|
1078 * |
|
1079 * @param beginIndex the beginning index, inclusive. |
|
1080 * @param endIndex the ending index, exclusive. |
|
1081 * @throws IndexOutOfBounds if either offset is outside of the |
|
1082 * array bounds |
|
1083 */ |
|
1084 public char[] getRawCharacters(int beginIndex, int endIndex) { |
|
1085 int length = endIndex - beginIndex; |
|
1086 char[] chars = new char[length]; |
|
1087 System.arraycopy(buf, beginIndex, chars, 0, length); |
|
1088 return chars; |
|
1089 } |
|
1090 |
|
1091 public enum CommentStyle { |
|
1092 LINE, |
|
1093 BLOCK, |
|
1094 JAVADOC, |
|
1095 } |
|
1096 |
|
1097 /** |
|
1098 * Called when a complete comment has been scanned. pos and endPos |
|
1099 * will mark the comment boundary. |
|
1100 */ |
|
1101 protected void processComment(CommentStyle style) { |
|
1102 if (scannerDebug) |
|
1103 System.out.println("processComment(" + pos |
|
1104 + "," + endPos + "," + style + ")=|" |
|
1105 + new String(getRawCharacters(pos, endPos)) |
|
1106 + "|"); |
|
1107 } |
|
1108 |
|
1109 /** |
|
1110 * Called when a complete whitespace run has been scanned. pos and endPos |
|
1111 * will mark the whitespace boundary. |
|
1112 */ |
|
1113 protected void processWhiteSpace() { |
|
1114 if (scannerDebug) |
|
1115 System.out.println("processWhitespace(" + pos |
|
1116 + "," + endPos + ")=|" + |
|
1117 new String(getRawCharacters(pos, endPos)) |
|
1118 + "|"); |
|
1119 } |
|
1120 |
|
1121 /** |
|
1122 * Called when a line terminator has been processed. |
|
1123 */ |
|
1124 protected void processLineTerminator() { |
|
1125 if (scannerDebug) |
|
1126 System.out.println("processTerminator(" + pos |
|
1127 + "," + endPos + ")=|" + |
|
1128 new String(getRawCharacters(pos, endPos)) |
|
1129 + "|"); |
|
1130 } |
|
1131 |
|
1132 /** Build a map for translating between line numbers and |
|
1133 * positions in the input. |
|
1134 * |
|
1135 * @return a LineMap */ |
|
1136 public Position.LineMap getLineMap() { |
|
1137 return Position.makeLineMap(buf, buflen, false); |
|
1138 } |
|
1139 |
|
1140 } |
113 } |