-
Notifications
You must be signed in to change notification settings - Fork 10
/
Copy pathmlparse.c
4472 lines (3994 loc) · 163 KB
/
mlparse.c
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
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/* mlparse.c implements the parser declared in mlparse.h
*
* If you want to figure out how this works, you're probably best off
* compiling the program at the bottom of this file and running it over some
* data (need to #define MLPARSE_TEST).
*
* there are two main components to understanding the code,
* understanding the buffering and understanding the state machine
* (if you don't know what a state machine is, go find out).
*
* The whole business of parsing SGML/XML/HTML is complicated by the fact that
* almost nobody writes it correctly. As such, we can't just accept structures
* like open tags and open comments as correct, we need to verify that they will
* actually close. This requires us to examine characters further ahead in the
* stream than we are currently parsing. Since our interface doesn't guarantee
* that they are available in advance, we have to be prepared to buffer a
* certain number of characters internally. This complicates things a lot,
* because we now sometimes have two sources of characters (internal and
* external buffers) and sometimes have two seperate positions (current parsing
* position and current look-ahead position) in the character stream that we
* have to maintain. Throughout the code i have referred to this look-ahead
* process as 'peek'ing. This process can only be performed forward until we
* run out of buffer space, so state->lookahead limits the amount of characters
* that can be passed over this way. In the peek'ing states, the number of
* characters peek'd over is recorded in the state->count variable, which MUST
* record all characters passed over, so we know that they'll fit in the buffer.
* (note that i actually break this slightly for the eref peek, but we only
* allow very short erefs). The buffer is implemented by the state->buf
* (current buffer), state->tmpbuf (spare buffer), state->pos
* (position in current buffer) and state->end (end of data in current buffer)
* variables within the state structure. state->pos and state->end dictate how
* much data is in the buffer, and must both be placed at the start of the
* buffer if no buffered data is available (simplifies the buffering process).
* state->tmpbuf is used for miscellaneous purposes, including copying buffered
* data into it to make it contiguous. This occurs if we don't have enough
* space in the current buffer for text that we want to add to it. At the end
* of each peek'ing process we want to ensure that the text we have looked over
* is either all in the internal buffer, or all in the external buffer (and
* hence contiguous). This means that in states where we know that we've
* already parsed the data we are looking at, we don't have to worry about
* testing for the end of the stream until we hit the end delimiter. Note that
* although buffering is necessary for the correct operation of the parser, the
* assumption is that large buffers are available and that buffering will be
* fairly rare. Thus our optimisation efforts are directed toward the case
* where buffering is extremely rare, ensuring only that buffering works when
* required. Non-stripping operation is likewise supported but not considered
* important for performance purposes.
*
* Another set of design decisions had to be made about what constitutes markup,
* especially in the cases where it doesn't quite conform to the standards.
* Particular situations that cropped up were what to accept in a tag name,
* (we accept the XML/SGML name character set) where to allow whitespace in
* structural elements (tags: anywhere but straight after '<', comments: not in
* '<!--' or '-->', cdata: not in '<![' or ']]>', entity references: not
* allowed). I beleive that valid SGML tags aren't allowed whitespace
* immediately after the start tag, although its a little hard to tell from
* casual inspection of the productions.
*
* The state machine is implemented as a set of states (declared in enum
* mlparse_states) and a set of flags (in enum mlparse_flags). The flags are
* supposed to be orthogonal to the states, so that pretty much any set of flags
* can occur in any state (this isn't true of FLAG_COMMENT and FLAG_CDATA, but
* close enough). The JUMP macro implements a mapping between states and flags
* and the bits of code that implement operations for that state. In this way,
* different, specialised bits of code can handle the same state when different
* flags are in effect, as happens to the word parsing states when FLAG_COMMENT
* or FLAG_CDATA is on. This could also be used to optimise away some of the
* overhead of buffering/stripping (both of which weren't in the previous
* version of this parser) by specialising the most used states so that they
* don't have to use conditionals to determine whether stripping/buffering are
* on. However, states that are connected by direct goto statements (pretty
* much anything not seperated by peek states, which must use JUMP statements to
* exit given the uncertainty of where they came from) have to be consistent in
* their assumptions of what flags are enabled. One technique that i have used
* to reduce the bulk of the code is to use macro's and indirect blocks of code
* that appear as 'transitional' states (marked with trans_). The macros are
* fairly self-evident (and commented), but in some cases the inline replacement
* of macros was bulking up the code a lot. In these situations i expanded the
* macros in only a few places in the code (typically in a trans_ block of
* code), using local variables in the trans union to pass parameters, and then
* used a stub macro to marshall parameters and jump to the indirect macro
* location.
*
* A major design decision in the code was to allow 'floating' states, that act
* sort of like functions in that you can jump to them from anywhere in the code
* and they return through (an extremely) limited stack mechanism. The stack is
* the state->next_state variable, which specifies how the 'floating' states are
* to exit. The 'floating' states are those that detect/process entity
* references/tags/other entities, since they can occur in a number of contexts,
* and keeping track of exactly what the context is without this technique is
* extremely difficult. However, since only one next_state variable is
* available, only one 'floating' state can be active at a time. This caused
* some minor issues with the end sentence detection, which i initially tried to
* implement as a 'floating' state, but couldn't as that raised scenarios when i
* wanted two floating states active at one time (like 'word. ', where the
* transition from end sentence state to entity reference state causes us to
* kill the next_state value that the end sentence state relies upon.
*
* The code could be improved in many ways:
* - specialise some states (especially the toplevel word detection states)
* for stripping and non-buffering for performance
* - allow multi-character entity references (not necessary at the moment,
* since we only allow HTML and numeric erefs, which only reference a single
* character). This could be implemented by loading st->count with the
* number of escaped characters and counting down in escape states. This
* would probably require that esc becomes a flag as discussed below.
* - do a better job of detection of entity references (using a trie,
* probably) and numeric character decoding (anything other than strtol).
* Same thing applies to recognition of marked sections. This will become
* much more important if you want to recognise custom entity references.
* - most of the CASE macros should be moved to a more central bit of code,
* since they are useful for more than just here. It would also be an
* improvement if the characters were turned into numbers so that it will
* work identically on non-ASCII machines.
* - further character analysis (maybe?). This would complicate things, but
* better output might be obtained by breaking words across characters like
* '/' (which seems to unequivocally represent an alternative in english
* text) and by allowing (double) quotes between sentence ending punctuation
* and whitespace (which i tried but removed to reduce complexity).
* - escape (almost certainly) and end sentence (maybe) should be flags. They
* aren't because of lack of foresight by me and the fact that turning end
* sentence into a flag could contention over the next_state variable as
* discussed above.
* - declarations could be handled properly, which would assist in comment
* resolution stuff. It isn't particularly hard either, just a bit more
* work. However, theres a bit of a conflict between SGML and XML here, in
* that XML expects comments to be a single entity. The SGML comment is
* more general, and solves some problems with whitespace in comment
* start/end. Somewhat harder along the same lines is to allow DTD
* declarations using the <!DTD [ ]> syntax, since it involves detecting
* tags within tags.
* - mark EOF as a flag, so that user only has to call EOF once.
* Unfortunately this may require a lot more in the way of states to handle
* properly.
* - sort out what should/shouldn't be filtered from parameter names (+
* implicit values)
* - self-ending tags (i.e. <p/>) are truncated for the end tag return if the
* name goes over wordlen chars
* - explicit end-tag notification (simple to do, just not done because i
* didn't want to make it less efficient)
*
* A few more random notes about the code:
* - length has to be assigned before using RETURN(). I've assigned it
* immediately before use so you can use grep to validate this (a frequent
* source of bugs not to assign length before RETURN()).
*
* FIXME: include (at least conceptual) state diagrams
*
* written nml 2004-04-10
*
*/
/* XXX: should decouple buffer size and lookahead so fill_buffer can scale up */
/* test: */
/* XXX: vary buffer size and compare outputs */
#include "firstinclude.h"
#include "mlparse.h"
#include "ascii.h"
#include "def.h"
#include "str.h"
#include <assert.h>
#include <ctype.h>
#include <errno.h>
#include <limits.h>
#include <stdlib.h>
#include <string.h>
/* the maximum number of characters that we recognise entity references over */
#define MAXENTITYREF 8
/* flags that store additional state about the parsing */
enum mlparse_flags {
FLAG_NONE = 0, /* no flags */
FLAG_STRIP = 1, /* flag indicating that junk can be
* stripped out while parsing. Note that
* strip is never set in the flags, its
* passed by the user at every invocation,
* so it can change between each call. */
FLAG_BUFFER = 2, /* indicates that bytes have been stored in
* the internal buffer */
FLAG_COMMENT = 4, /* indicates that we're in a comment */
FLAG_CDATA = 8, /* indicates that we're in a cdata sec. */
/* entry to allow us to mask in/out flags easily */
FLAG_ALL = FLAG_STRIP | FLAG_BUFFER | FLAG_COMMENT | FLAG_CDATA,
FLAG_BITS = 4 /* number of flag bits (MUST keep this up
* to date) */
};
/* states that we can be in when we enter the parser */
enum mlparse_states {
STATE_ERR = 0, /* error state */
/* states based on words */
STATE_TOPLEVEL = 1, /* not in any HTML entities or words */
STATE_WORD = 2, /* in a word */
STATE_WORD_ENDS = 3, /* in a word, after sentence-ending
* punctuation */
STATE_PUNC = 6, /* in a word, after punctuation */
STATE_PUNC_ENDS = 7, /* in a word, after sentence ending
* punctuation, after punctuation*/
STATE_ACRONYM = 8, /* in an acronym, after capital letter */
STATE_ACRONYM_LETTER = 9, /* in an acronym, after a period */
STATE_SPACE = 10, /* in whitespace */
/* states based on tags */
STATE_TAG_PEEK_FIRST = 20, /* looking ahead to validate a tag,
* expecting first letter of tag name (or
* '!' or '?' or '/' depending on what it
* is) */
STATE_TAG_PEEK_NAME = 21, /* looking ahead to validate a tag,
* parsing the tag name after first
* char */
STATE_TAG_PEEK_NAME_CONT = 22, /* looking ahead to validate a tag,
* parsing tag name after its exceeded
* wordlen characters */
STATE_TAG_PEEK = 23, /* looking ahead to validate a tag,
* after the tag name */
STATE_TAG_ENTRY = 24, /* in a tag from start once
* validated */
STATE_TAG_NAME_CONT = 25, /* in a tag name after
* continuation */
STATE_TAG_NAME_SELFEND = 26, /* in a tag that may self-end during
* name */
STATE_TAG = 27, /* in a tag after name */
STATE_TAG_SELFEND = 28, /* in a tag that may self-end */
/* states based on parameters/values */
STATE_PARAM = 30, /* in a parameter name */
STATE_PARAM_EQ = 31, /* after parameter name */
STATE_PVAL_FIRST = 32, /* at the first char of a pval */
STATE_PVAL = 33, /* in a whitespace-delimited pval */
STATE_PVAL_PUNC = 35, /* in a whitespace-delimited pval, after
* puncation */
STATE_PVAL_SELFEND = 36, /* in a whitespace-delimited pval, after
* a / may self-end the tag */
STATE_PVAL_QUOT = 37, /* in a quote-delimited pval */
STATE_PVAL_QUOT_WORD = 38, /* in a word in a quote-delimited pval */
STATE_PVAL_QUOT_PUNC = 40, /* in a word in a quote-delimited pval,
* after punctuation */
STATE_PVAL_QUOT_SPACE = 41, /* in whitespaec in a quote-delimited
* pval */
STATE_PVAL_DQUOT = 42, /* */
STATE_PVAL_DQUOT_WORD = 43, /* these states are equivalent to */
STATE_PVAL_DQUOT_PUNC = 45, /* double-quote delimited pvals */
STATE_PVAL_DQUOT_SPACE = 46, /* */
/* states based on entity references */
STATE_EREF_PEEK_FIRST = 50, /* looking ahead to validate an eref, on
* first character after '&' */
STATE_EREF_NUM_PEEK_FIRST = 52, /* looking ahead to validate a numeric
* eref, on first char after '#' */
STATE_EREF_NUM_PEEK = 53, /* looking ahead to validate a numeric
* eref */
STATE_EREF_HEX_PEEK = 54, /* looking ahead to validate a
* hexadecimal numeric eref */
STATE_EREF_PEEK = 55, /* looking ahead to validate a
* non-numeric eref */
/* states based on comment or cdata sections */
STATE_DECL_PEEK = 70, /* looking ahead to validate a
* declaration, after '!' char */
STATE_COMMENT_PEEK_FIRST = 71, /* looking ahead to validate a
* comment, after first '-' char */
STATE_COMMENT_PEEK = 72, /* looking ahead to validate a
* comment */
/* these two states just validate --> ending on comment (one for each char
* except the first) */
STATE_COMMENT_PEEK_END_FIRST = 73,
STATE_COMMENT_PEEK_END = 74,
/* these three states just look out for <!-- within comment (one for each
* char except the first) */
STATE_COMMENT_PEEK_BADEND_FIRST = 75,
STATE_COMMENT_PEEK_BADEND_SECOND = 76,
STATE_COMMENT_PEEK_BADEND = 77,
STATE_COMMENT_ENTRY = 78, /* parsing a validated comment */
STATE_MARKSEC_PEEK_LABEL_FIRST = 79, /* looking ahead to validate a
* marked section, after first '[' */
STATE_MARKSEC_PEEK_LABEL = 80, /* looking ahead to validate a marked
* section, during its label
* (i.e. 'CDATA') */
STATE_MARKSEC_PEEK_LABEL_LAST = 81, /* looking ahead to validate a marked
* section, after end of label */
STATE_MARKSEC_PEEK = 82, /* looking ahead to validate a marked
* section */
STATE_MARKSEC_PEEK_END_FIRST = 83, /* looking ahead to validate a marked
* section, after first ']' end
* char */
STATE_MARKSEC_PEEK_END = 84, /* looking ahead to validate a marked
* section, after second ']' end
* char */
STATE_CDATA_ENTRY = 85, /* parsing a validated cdata marked
* section */
/* these two states just look for --> within a cdata/comment section */
STATE_CCDATA_END_COMMENT_FIRST = 86,
STATE_CCDATA_END_COMMENT = 87,
/* these two states just look for ]]> within a cdata/comment section */
STATE_CCDATA_END_CDATA_FIRST = 88,
STATE_CCDATA_END_CDATA = 89,
STATE_CCDATA_RECOVERY = 90, /* recover from false cdata/comment
* end detection */
/* escape states, which correspond to the normal states except that the next
* character is escaped */
STATE_TOPLEVEL_ESC = 100,
STATE_WORD_ESC = 101,
STATE_WORD_ENDS_ESC = 102,
STATE_PUNC_ESC = 105,
STATE_PUNC_ENDS_ESC = 106,
STATE_SPACE_ESC = 107,
STATE_ACRONYM_ESC = 108,
STATE_ACRONYM_LETTER_ESC = 109,
STATE_PVAL_QUOT_ESC = 110,
STATE_PVAL_QUOT_WORD_ESC = 111,
STATE_PVAL_QUOT_PUNC_ESC = 113,
STATE_PVAL_QUOT_SPACE_ESC = 114,
STATE_PVAL_DQUOT_ESC = 115,
STATE_PVAL_DQUOT_WORD_ESC = 116,
STATE_PVAL_DQUOT_PUNC_ESC = 118,
STATE_PVAL_DQUOT_SPACE_ESC = 119,
STATE_EOF = 120 /* end of file state */
};
struct mlparse_state {
enum mlparse_states state; /* current state of parser */
enum mlparse_flags flags; /* additional state flags */
unsigned int len; /* length of entity in state */
enum mlparse_states next_state; /* next state to go to (where
* applicable) */
unsigned int count; /* chars we've processed in current
* state */
unsigned int wordlen; /* maximum length of a word */
unsigned int lookahead; /* maximum length to look ahead for
* end tags */
unsigned int buflen; /* length of the internal buffer */
const char *pos; /* current position in buffer */
const char *end; /* first position past end of
* buffer */
char *buf; /* buffer for lookahead */
char *tagbuf; /* buffer to store tag names in */
unsigned int tagbuflen; /* length of stuff in tagbuf */
char *tmpbuf; /* temporary buffer (size of buf) */
char erefbuf[MAXENTITYREF + 1]; /* entity reference buffer */
unsigned int errline; /* line error occurred on (debugging
* tool) */
};
int mlparse_new(struct mlparse *space, unsigned int wordlen,
unsigned int lookahead) {
if (wordlen <= 1) {
return 0;
}
/* XXX: note slight overallocation of buf for fudge factor purposes:
* sometimes the maths are a bit funny, so we allow a few extra characters
* on the end just in case */
if ((space->state = malloc(sizeof(*space->state)))
&& (space->state->buflen = lookahead + 2 * MAXENTITYREF)
&& (space->state->buf = malloc(space->state->buflen + 1))
&& (space->state->tmpbuf = malloc(space->state->buflen + 1))
&& (space->state->tagbuf = malloc(wordlen + 1))) {
space->state->wordlen = wordlen;
space->state->lookahead = lookahead;
space->state->tagbuf[0] = '\0';
space->state->erefbuf[0] = '\0';
space->state->buf[space->state->buflen] = '\0';
mlparse_reinit(space);
} else if (space->state) {
if (space->state->buf) {
if (space->state->tmpbuf) {
free(space->state->tmpbuf);
}
free(space->state->buf);
}
free(space->state);
return 0;
}
return 1;
}
void mlparse_reinit(struct mlparse *p) {
/* reset variables */
p->state->pos = p->state->end = p->state->buf;
p->state->state = STATE_TOPLEVEL; /* start in toplevel state */
p->state->flags = FLAG_NONE;
p->state->errline = 0;
return;
}
unsigned long int mlparse_buffered(struct mlparse *parser) {
if (parser->state->flags & FLAG_BUFFER) {
return parser->state->end - parser->state->pos;
} else {
return 0;
}
}
const char *mlparse_buffer(struct mlparse *parser) {
return parser->state->pos;
}
void mlparse_delete(struct mlparse *parser) {
free(parser->state->tagbuf);
free(parser->state->buf);
free(parser->state->tmpbuf);
free(parser->state);
}
unsigned int mlparse_err_line(struct mlparse *parser) {
return parser->state->errline;
}
/* macro to combine state values with flag values by shifting state values */
#define COMB(stateval) ((stateval) << (FLAG_BITS))
/* stub macro to jump to state state_ (modified by flags_). unbuf_ is boolean,
* indicating whether to allow jump to trans_unbuf_label */
#define JUMP(state_, flags_, unbuf_) \
trans.jump.state = (state_); \
trans.jump.flags = (flags_); \
if (unbuf_) { \
goto trans_jump_unbuf_label; \
} else if (1) { \
goto trans_jump_label; \
} else
/* macro to manage jump table, which maps the state/flag variables to chunks of
* code that are executed when they are encountered. However, this has to be
* consistent with the direct gotos between states for proper operation.
* (please don't change this stuff unless you *really* know what you're
* doing) */
#define ACTUAL_JUMP(state_, flags, unbuf_) \
if (1) { \
switch (COMB(state_) | (flags) | strip) { \
case FLAG_STRIP | COMB(STATE_SPACE): \
case COMB(STATE_TOPLEVEL): \
case FLAG_STRIP | COMB(STATE_TOPLEVEL): goto toplevel_label; \
\
case COMB(STATE_WORD_ENDS): \
case FLAG_STRIP | COMB(STATE_WORD_ENDS): goto word_ends_label; \
\
case COMB(STATE_WORD): \
case FLAG_STRIP | COMB(STATE_WORD): goto word_label; \
\
case COMB(STATE_PUNC_ENDS): \
case FLAG_STRIP | COMB(STATE_PUNC_ENDS): goto punc_ends_label; \
\
case COMB(STATE_PUNC): \
case FLAG_STRIP | COMB(STATE_PUNC): goto punc_label; \
\
case COMB(STATE_ACRONYM): \
case FLAG_STRIP | COMB(STATE_ACRONYM): goto acronym_label; \
\
case COMB(STATE_ACRONYM_LETTER): \
case FLAG_STRIP | COMB(STATE_ACRONYM_LETTER): \
goto acronym_letter_label; \
case COMB(STATE_SPACE): goto space_label; \
\
case COMB(STATE_TAG_PEEK_FIRST): \
case FLAG_STRIP | COMB(STATE_TAG_PEEK_FIRST): \
goto tag_peek_first_label; \
\
case COMB(STATE_TAG_PEEK_NAME): \
case FLAG_STRIP | COMB(STATE_TAG_PEEK_NAME): \
goto tag_peek_name_label; \
\
case COMB(STATE_TAG_PEEK_NAME_CONT): \
case FLAG_STRIP | COMB(STATE_TAG_PEEK_NAME_CONT): \
goto tag_peek_name_cont_label; \
\
case COMB(STATE_TAG_PEEK): \
case FLAG_STRIP | COMB(STATE_TAG_PEEK): \
goto tag_peek_label; \
\
case COMB(STATE_TAG_ENTRY): \
case FLAG_STRIP | COMB(STATE_TAG_ENTRY): \
goto tag_entry_label; \
\
case COMB(STATE_TAG_NAME_CONT): \
case FLAG_STRIP | COMB(STATE_TAG_NAME_CONT): \
goto tag_name_cont_label; \
\
case COMB(STATE_TAG_NAME_SELFEND): \
case FLAG_STRIP | COMB(STATE_TAG_NAME_SELFEND): \
goto tag_name_selfend_label; \
\
case COMB(STATE_TAG): \
case FLAG_STRIP | COMB(STATE_TAG): \
goto tag_label; \
\
case COMB(STATE_TAG_SELFEND): \
case FLAG_STRIP | COMB(STATE_TAG_SELFEND): \
goto tag_selfend_label; \
\
case COMB(STATE_PARAM): \
case FLAG_STRIP | COMB(STATE_PARAM): \
goto param_label; \
\
case COMB(STATE_PARAM_EQ): \
case FLAG_STRIP | COMB(STATE_PARAM_EQ): \
goto param_eq_label; \
\
case COMB(STATE_PVAL_FIRST): \
case FLAG_STRIP | COMB(STATE_PVAL_FIRST): \
goto pval_first_label; \
\
case COMB(STATE_PVAL): \
case FLAG_STRIP | COMB(STATE_PVAL): \
goto pval_label; \
\
case COMB(STATE_PVAL_PUNC): \
case FLAG_STRIP | COMB(STATE_PVAL_PUNC): \
goto pval_punc_label; \
\
case COMB(STATE_PVAL_SELFEND): \
case FLAG_STRIP | COMB(STATE_PVAL_SELFEND): \
goto pval_selfend_label; \
\
case COMB(STATE_PVAL_QUOT): \
case FLAG_STRIP | COMB(STATE_PVAL_QUOT): \
case FLAG_STRIP | COMB(STATE_PVAL_QUOT_SPACE): \
goto pval_quot_label; \
\
case COMB(STATE_PVAL_QUOT_WORD): \
case FLAG_STRIP | COMB(STATE_PVAL_QUOT_WORD): \
goto pval_quot_word_label; \
\
case COMB(STATE_PVAL_QUOT_PUNC): \
case FLAG_STRIP | COMB(STATE_PVAL_QUOT_PUNC): \
goto pval_quot_punc_label; \
\
case COMB(STATE_PVAL_QUOT_SPACE): \
goto pval_quot_space_label; \
\
case COMB(STATE_PVAL_DQUOT): \
case FLAG_STRIP | COMB(STATE_PVAL_DQUOT): \
goto pval_dquot_label; \
\
case COMB(STATE_PVAL_DQUOT_WORD): \
case FLAG_STRIP | COMB(STATE_PVAL_DQUOT_WORD): \
goto pval_dquot_word_label; \
\
case COMB(STATE_PVAL_DQUOT_PUNC): \
case FLAG_STRIP | COMB(STATE_PVAL_DQUOT_PUNC): \
goto pval_dquot_punc_label; \
\
case COMB(STATE_PVAL_DQUOT_SPACE): \
case FLAG_STRIP | COMB(STATE_PVAL_DQUOT_SPACE): \
goto pval_dquot_space_label; \
\
case COMB(STATE_EREF_PEEK_FIRST): \
case FLAG_STRIP | COMB(STATE_EREF_PEEK_FIRST): \
goto eref_peek_first_label; \
\
case COMB(STATE_EREF_NUM_PEEK_FIRST): \
case FLAG_STRIP | COMB(STATE_EREF_NUM_PEEK_FIRST): \
goto eref_num_peek_first_label; \
\
case COMB(STATE_EREF_NUM_PEEK): \
case FLAG_STRIP | COMB(STATE_EREF_NUM_PEEK): \
goto eref_num_peek_label; \
\
case COMB(STATE_EREF_HEX_PEEK): \
case FLAG_STRIP | COMB(STATE_EREF_HEX_PEEK): \
goto eref_hex_peek_label; \
\
case COMB(STATE_EREF_PEEK): \
case FLAG_STRIP | COMB(STATE_EREF_PEEK): \
goto eref_peek_label; \
\
case FLAG_STRIP | COMB(STATE_SPACE_ESC): \
case COMB(STATE_TOPLEVEL_ESC): \
case FLAG_STRIP | COMB(STATE_TOPLEVEL_ESC): goto toplevel_esc_label; \
\
case COMB(STATE_WORD_ESC): \
case FLAG_STRIP | COMB(STATE_WORD_ESC): goto word_esc_label; \
\
case COMB(STATE_WORD_ENDS_ESC): \
case FLAG_STRIP | COMB(STATE_WORD_ENDS_ESC): \
goto word_ends_esc_label; \
\
case COMB(STATE_PUNC_ESC): \
case FLAG_STRIP | COMB(STATE_PUNC_ESC): goto punc_esc_label; \
\
case COMB(STATE_PUNC_ENDS_ESC): \
case FLAG_STRIP | COMB(STATE_PUNC_ENDS_ESC): \
goto punc_ends_esc_label; \
\
case COMB(STATE_SPACE_ESC): goto space_esc_label; \
\
case COMB(STATE_ACRONYM_ESC): goto acronym_esc_label; \
case FLAG_STRIP | COMB(STATE_ACRONYM_ESC): goto acronym_esc_label; \
\
case COMB(STATE_ACRONYM_LETTER_ESC): \
case FLAG_STRIP | COMB(STATE_ACRONYM_LETTER_ESC): \
goto acronym_letter_esc_label; \
\
case COMB(STATE_PVAL_QUOT_ESC): \
case FLAG_STRIP | COMB(STATE_PVAL_QUOT_ESC): \
case FLAG_STRIP | COMB(STATE_PVAL_QUOT_SPACE_ESC): \
goto pval_quot_esc_label; \
\
case COMB(STATE_PVAL_QUOT_WORD_ESC): \
case FLAG_STRIP | COMB(STATE_PVAL_QUOT_WORD_ESC): \
goto pval_quot_word_esc_label; \
\
case COMB(STATE_PVAL_QUOT_PUNC_ESC): \
case FLAG_STRIP | COMB(STATE_PVAL_QUOT_PUNC_ESC): \
goto pval_quot_punc_esc_label; \
\
case COMB(STATE_PVAL_QUOT_SPACE_ESC): \
goto pval_quot_space_esc_label; \
\
case COMB(STATE_PVAL_DQUOT_ESC): \
case FLAG_STRIP | COMB(STATE_PVAL_DQUOT_ESC): \
case FLAG_STRIP | COMB(STATE_PVAL_DQUOT_SPACE_ESC): \
goto pval_dquot_esc_label; \
\
case COMB(STATE_PVAL_DQUOT_WORD_ESC): \
case FLAG_STRIP | COMB(STATE_PVAL_DQUOT_WORD_ESC): \
goto pval_dquot_word_esc_label; \
\
case COMB(STATE_PVAL_DQUOT_PUNC_ESC): \
case FLAG_STRIP | COMB(STATE_PVAL_DQUOT_PUNC_ESC): \
goto pval_dquot_punc_esc_label; \
\
case COMB(STATE_PVAL_DQUOT_SPACE_ESC): \
goto pval_dquot_space_esc_label; \
\
case COMB(STATE_DECL_PEEK): \
case FLAG_STRIP | COMB(STATE_DECL_PEEK): \
goto decl_peek_label; \
\
case COMB(STATE_COMMENT_PEEK_FIRST): \
case FLAG_STRIP | COMB(STATE_COMMENT_PEEK_FIRST): \
goto comment_peek_first_label; \
\
case COMB(STATE_COMMENT_PEEK): \
case FLAG_STRIP | COMB(STATE_COMMENT_PEEK): \
goto comment_peek_label; \
\
case COMB(STATE_COMMENT_PEEK_END_FIRST): \
case FLAG_STRIP | COMB(STATE_COMMENT_PEEK_END_FIRST): \
goto comment_peek_end_first_label; \
\
case COMB(STATE_COMMENT_PEEK_END): \
case FLAG_STRIP | COMB(STATE_COMMENT_PEEK_END): \
goto comment_peek_end_label; \
\
case COMB(STATE_COMMENT_PEEK_BADEND_FIRST): \
case FLAG_STRIP | COMB(STATE_COMMENT_PEEK_BADEND_FIRST): \
goto comment_peek_badend_first_label; \
\
case COMB(STATE_COMMENT_PEEK_BADEND_SECOND): \
case FLAG_STRIP | COMB(STATE_COMMENT_PEEK_BADEND_SECOND): \
goto comment_peek_badend_second_label; \
\
case COMB(STATE_COMMENT_PEEK_BADEND): \
case FLAG_STRIP | COMB(STATE_COMMENT_PEEK_BADEND): \
goto comment_peek_badend_label; \
\
case FLAG_COMMENT | COMB(STATE_COMMENT_ENTRY): \
case FLAG_COMMENT | FLAG_STRIP | COMB(STATE_COMMENT_ENTRY): \
goto comment_entry_label; \
\
case FLAG_COMMENT | COMB(STATE_TOPLEVEL): \
case FLAG_CDATA | COMB(STATE_TOPLEVEL): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_TOPLEVEL): \
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_TOPLEVEL): \
goto ccdata_toplevel_label; \
\
case FLAG_COMMENT | COMB(STATE_WORD): \
case FLAG_CDATA | COMB(STATE_WORD): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_WORD): \
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_WORD): \
goto ccdata_word_label; \
\
case FLAG_COMMENT | COMB(STATE_PUNC): \
case FLAG_CDATA | COMB(STATE_PUNC): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_PUNC): \
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_PUNC): \
goto ccdata_punc_label; \
\
case FLAG_COMMENT | COMB(STATE_SPACE): \
case FLAG_CDATA | COMB(STATE_SPACE): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_SPACE): \
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_SPACE): \
goto ccdata_space_label; \
\
case FLAG_COMMENT | COMB(STATE_CCDATA_END_COMMENT_FIRST): \
case FLAG_CDATA | COMB(STATE_CCDATA_END_COMMENT_FIRST): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_CCDATA_END_COMMENT_FIRST):\
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_CCDATA_END_COMMENT_FIRST): \
goto ccdata_end_comment_first_label; \
\
case FLAG_COMMENT | COMB(STATE_CCDATA_END_COMMENT): \
case FLAG_CDATA | COMB(STATE_CCDATA_END_COMMENT): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_CCDATA_END_COMMENT): \
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_CCDATA_END_COMMENT): \
goto ccdata_end_comment_label; \
\
case FLAG_COMMENT | COMB(STATE_CCDATA_END_CDATA_FIRST): \
case FLAG_CDATA | COMB(STATE_CCDATA_END_CDATA_FIRST): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_CCDATA_END_CDATA_FIRST): \
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_CCDATA_END_CDATA_FIRST): \
goto ccdata_end_cdata_first_label; \
\
case FLAG_COMMENT | COMB(STATE_CCDATA_END_CDATA): \
case FLAG_CDATA | COMB(STATE_CCDATA_END_CDATA): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_CCDATA_END_CDATA): \
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_CCDATA_END_CDATA): \
goto ccdata_end_cdata_label; \
\
case FLAG_COMMENT | COMB(STATE_CCDATA_RECOVERY): \
case FLAG_CDATA | COMB(STATE_CCDATA_RECOVERY): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_CCDATA_RECOVERY): \
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_CCDATA_RECOVERY): \
goto ccdata_recovery_label; \
\
case COMB(STATE_CDATA_ENTRY): \
case FLAG_STRIP | COMB(STATE_CDATA_ENTRY): \
case FLAG_CDATA | COMB(STATE_CDATA_ENTRY): \
case FLAG_CDATA | FLAG_STRIP | COMB(STATE_CDATA_ENTRY): \
goto cdata_entry_label; \
\
case COMB(STATE_MARKSEC_PEEK_LABEL_FIRST): \
case FLAG_STRIP | COMB(STATE_MARKSEC_PEEK_LABEL_FIRST): \
goto marksec_peek_label_first_label; \
\
case COMB(STATE_MARKSEC_PEEK_LABEL): \
case FLAG_STRIP | COMB(STATE_MARKSEC_PEEK_LABEL): \
goto marksec_peek_label_label; \
\
case COMB(STATE_MARKSEC_PEEK_LABEL_LAST): \
case FLAG_STRIP | COMB(STATE_MARKSEC_PEEK_LABEL_LAST): \
goto marksec_peek_label_last_label; \
\
case COMB(STATE_MARKSEC_PEEK): \
case FLAG_STRIP | COMB(STATE_MARKSEC_PEEK): \
goto marksec_peek_label; \
\
case COMB(STATE_MARKSEC_PEEK_END_FIRST): \
case FLAG_STRIP | COMB(STATE_MARKSEC_PEEK_END_FIRST): \
goto marksec_peek_end_first_label; \
\
case COMB(STATE_MARKSEC_PEEK_END): \
case FLAG_STRIP | COMB(STATE_MARKSEC_PEEK_END): \
goto marksec_peek_end_label; \
\
case COMB(STATE_EOF): \
case FLAG_STRIP | COMB(STATE_EOF): \
case FLAG_COMMENT | COMB(STATE_EOF): \
case FLAG_CDATA | COMB(STATE_EOF): \
case FLAG_STRIP | FLAG_COMMENT | COMB(STATE_EOF): \
case FLAG_STRIP | FLAG_CDATA | COMB(STATE_EOF): \
case FLAG_STRIP | FLAG_COMMENT | FLAG_CDATA | COMB(STATE_EOF): \
goto eof_label; \
\
default: \
/* have to save state because trans_unbuffer_label will jump to \
* st->state after altering pos, end */ \
if (unbuf_) { \
goto trans_unbuffer_label; \
} else { \
st->state = (state_); \
st->errline = UINT_MAX; \
goto err_label; \
} \
} \
} else /* dangling else to allow trailing semi-colon */
/* macro to detect (unexpected) errors in the parsing process, leaving enough
* info around to figure out how they occurred (hopefully) */
#define CANT_GET_HERE() \
assert(0); \
st->errline = __LINE__; \
goto err_label
/* catch sentence ending punctuation (no ending colon and no first case so we
* can use case syntax) */
#define CASE_ENDS \
'.': case '?': case '!'
/* convert a character case, given that its the opposite case, if CASE_FOLDING
* has been defined */
#ifdef CASE_FOLDING
/* they want case folding */
#define TOLOWER(c) ASCII_TOLOWER(c)
#define TOUPPER(c) ASCII_TOUPPER(c)
#else
/* no case folding */
#define TOLOWER(c) c
#define TOUPPER(c) c
#endif
/* fix things up for a return where we're requesting more input */
#define RETURN_INPUT(state_) \
trans.return_input.state = (state_); \
goto trans_return_input_label
/* fix things up for return, assuming that we're working from the buffer */
#define RETURN_BUFFER(retval, state_) \
assert(st->flags & FLAG_BUFFER); \
st->pos = pos; \
st->state = (state_); \
return (retval);
/* fix things up for return, assuming that we're not working from the buffer */
#define RETURN_NOBUFFER(retval, state_) \
assert(!(st->flags & FLAG_BUFFER)); \
parser->next_in = pos; \
parser->avail_in = end - pos; \
st->state = (state_); \
return (retval);
/* fix things up for return (by altering buffer positions to reflect what we've
* parsed this call) */
#define RETURN(retval, state_) \
if (1) { \
if (st->flags & FLAG_BUFFER) { \
RETURN_BUFFER(retval, state_); \
} else { \
RETURN_NOBUFFER(retval, state_); \
} \
} else
/* macro to handle boilerplate transitions to tag parsing */
#define CASE_TAG(next_state_) \
'<': \
tmppos = pos + 1; \
tmpend = end; \
st->count = 1; \
st->next_state = (next_state_); \
goto tag_peek_first_label
/* macro to handle boilerplate transitions to entity reference parsing */
#define CASE_EREF(next_state_) \
'&': \
tmppos = pos + 1; \
tmpend = end; \
st->count = 1; \
st->next_state = (next_state_); \
goto eref_peek_first_label
/* push a character onto the current word, returning MLPARSE_CONT | retval
* and maintaining state state_ if there is no room left */
#define PUSH(char_, retval, state_) \
if (st->len < st->wordlen) { \
word[st->len++] = (char_); \
} else if (1) { \
*length = st->len; \
st->len = 0; \
RETURN(MLPARSE_CONT | (retval), (state_)); \
} else
/* a macro to place characters examined by a peek state (as indicated by tmppos)
* into the buffer */
#define BUFFER() \
if (1) { \
const char *tmpstart; \
\
assert((tmppos >= parser->next_in) \
&& (tmppos <= parser->next_in + parser->avail_in)); \
\
/* figure out where to start copying, else we're in danger of \
* trying to copy stuff from the buffer into the buffer */ \
if (st->flags & FLAG_BUFFER) { \
assert((pos >= st->pos) && (pos <= st->end)); \
tmpstart = parser->next_in; \
} else { \
assert(1 && (pos >= parser->next_in) \
&& (pos <= parser->next_in + parser->avail_in)); \
tmpstart = pos; \
} \
\
if (tmppos - tmpstart) { \
/* need to copy stuff into buffer */ \
assert(st->end >= st->pos); \
assert((st->end > st->pos) \
|| ((st->buf == st->end) && (st->buf == st->pos))); \
\
if ((tmppos - tmpstart) < (st->buf + st->buflen - st->end)) { \
/* easy case, just copy chunk in to end of buffer */ \
memcpy((char *) st->end, tmpstart, (tmppos - tmpstart)); \
pos = st->pos; \
end = st->end += (tmppos - tmpstart); \
} else { \
unsigned int tmp = end - pos; \
char *tmpbuf = st->buf; \
\
/* must already be in the buffer */ \
assert((unsigned int) ((tmppos - tmpstart) + (end - pos)) \
< st->buflen); \
assert((pos >= st->pos) && (pos <= st->end)); \
\
/* copy existing into temporary buffer */ \
memcpy(st->tmpbuf, pos, tmp); \
\
/* copy new stuff in */ \
memcpy(st->tmpbuf + tmp, tmpstart, tmppos - tmpstart); \
tmp += tmppos - tmpstart; \
assert(tmp < st->buflen); \
\
/* swap buffers over */ \
pos = st->pos = st->buf = st->tmpbuf; \
end = st->end = st->pos + tmp; \
st->tmpbuf = tmpbuf; \
} \
st->flags |= FLAG_BUFFER; \
parser->next_in += (tmppos - tmpstart); \
parser->avail_in -= (tmppos - tmpstart); \
} \
} else
/* macro to return a stored value depending on state in state->next_state.
* It can return a value (setting next_state to next_state_ param) or merely
* continue if no value needs to be returned. This is the end of each of the
* peek states, where they have to return buffered values if appropriate or
* continue on into the states associated with the entity just detected. Note
* that it alters the buffer to ensure contiguity either way. */
#define RETURN_IF_STORED(next_state_) \
if (1) { \
enum mlparse_ret RETURN_STORED_ret; \
\
/* note that this only has to handle 'toplevel' states, since we \
* can't enter an entity while parsing another entity */ \
switch (st->next_state) { \
case STATE_WORD: \
case STATE_WORD_ESC: \
case STATE_WORD_ENDS: \
case STATE_WORD_ENDS_ESC: \
case STATE_PUNC: \
case STATE_PUNC_ESC: \
case STATE_PUNC_ENDS: \
case STATE_PUNC_ENDS_ESC: \
case STATE_ACRONYM: \
case STATE_ACRONYM_ESC: \
case STATE_ACRONYM_LETTER: \
case STATE_ACRONYM_LETTER_ESC: \
*length = st->len; \
RETURN_STORED_ret = MLPARSE_WORD; \
break; \
\
case STATE_SPACE: \
case STATE_SPACE_ESC: \
*length = st->len; \
RETURN_STORED_ret = MLPARSE_WHITESPACE; \
break; \
\
default: \
RETURN_STORED_ret = MLPARSE_INPUT; \
break; \
} \
\
/* ensure everything is either wholly buffered or not buffered at all \
* to ensure contiguity as we reparse over the peek'd section. Also \
* save buffer position as we might return below. */ \
if ((st->flags & FLAG_BUFFER) && (tmppos >= parser->next_in) \
&& (tmppos <= parser->next_in + parser->avail_in)) { \
/* need to buffer */ \
assert((pos >= st->pos) \
&& (pos <= st->end)); \
\
st->pos = pos; \
BUFFER(); \
} else if (!(st->flags & FLAG_BUFFER)) { \
/* nothing in buffer, no need to alter it */ \
assert((pos >= parser->next_in) \
&& (pos <= parser->next_in + parser->avail_in)); \
assert((tmppos >= parser->next_in) \
&& (tmppos <= parser->next_in + parser->avail_in)); \
\
parser->next_in = pos; \
parser->avail_in = end - pos; \
} else { \
st->pos = pos; \
} \
\
st->len = 0; \
if (RETURN_STORED_ret != MLPARSE_INPUT) { \
st->next_state = STATE_ERR; \
st->state = (next_state_); \
return (RETURN_STORED_ret); \
} \