Annotation of src/usr.bin/lex/flexdef.h, Revision 1.1.1.1
1.1 deraadt 1: /* flexdef - definitions file for flex */
2:
3: /*-
4: * Copyright (c) 1990 The Regents of the University of California.
5: * All rights reserved.
6: *
7: * This code is derived from software contributed to Berkeley by
8: * Vern Paxson.
9: *
10: * The United States Government has rights in this work pursuant
11: * to contract no. DE-AC03-76SF00098 between the United States
12: * Department of Energy and the University of California.
13: *
14: * Redistribution and use in source and binary forms are permitted provided
15: * that: (1) source distributions retain this entire copyright notice and
16: * comment, and (2) distributions including binaries display the following
17: * acknowledgement: ``This product includes software developed by the
18: * University of California, Berkeley and its contributors'' in the
19: * documentation or other materials provided with the distribution and in
20: * all advertising materials mentioning features or use of this software.
21: * Neither the name of the University nor the names of its contributors may
22: * be used to endorse or promote products derived from this software without
23: * specific prior written permission.
24: * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
25: * WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
26: * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
27: */
28:
29: /* @(#) $Header: /a/cvsroot/src/usr.bin/lex/flexdef.h,v 1.9 1995/05/05 05:35:21 jtc Exp $ (LBL) */
30:
31: #include <stdio.h>
32: #include <ctype.h>
33:
34: #include "config.h"
35:
36: #ifdef __TURBOC__
37: #define HAVE_STRING_H 1
38: #define MS_DOS 1
39: #ifndef __STDC__
40: #define __STDC__ 1
41: #endif
42: #pragma warn -pro
43: #pragma warn -rch
44: #pragma warn -use
45: #pragma warn -aus
46: #pragma warn -par
47: #pragma warn -pia
48: #endif
49:
50: #ifdef HAVE_STRING_H
51: #include <string.h>
52: #else
53: #include <strings.h>
54: #endif
55:
56: #ifdef HAVE_SYS_TYPES_H
57: #include <sys/types.h>
58: #endif
59:
60: #ifdef HAVE_MALLOC_H
61: #include <malloc.h>
62: #endif
63:
64: #ifdef STDC_HEADERS
65: #include <stdlib.h>
66: #endif
67:
68: /* As an aid for the internationalization patch to flex, which
69: * is maintained outside this distribution for copyright reasons.
70: */
71: #define _(String) (String)
72:
73: /* Always be prepared to generate an 8-bit scanner. */
74: #define CSIZE 256
75: #define Char unsigned char
76:
77: /* Size of input alphabet - should be size of ASCII set. */
78: #ifndef DEFAULT_CSIZE
79: #define DEFAULT_CSIZE 128
80: #endif
81:
82: #ifndef PROTO
83: #if __STDC__
84: #define PROTO(proto) proto
85: #else
86: #define PROTO(proto) ()
87: #endif
88: #endif
89:
90: #ifdef VMS
91: #ifndef __VMS_POSIX
92: #define unlink remove
93: #define SHORT_FILE_NAMES
94: #endif
95: #endif
96:
97: #ifdef MS_DOS
98: #define SHORT_FILE_NAMES
99: #endif
100:
101:
102: /* Maximum line length we'll have to deal with. */
103: #define MAXLINE 2048
104:
105: #ifndef MIN
106: #define MIN(x,y) ((x) < (y) ? (x) : (y))
107: #endif
108: #ifndef MAX
109: #define MAX(x,y) ((x) > (y) ? (x) : (y))
110: #endif
111: #ifndef ABS
112: #define ABS(x) ((x) < 0 ? -(x) : (x))
113: #endif
114:
115:
116: /* ANSI C does not guarantee that isascii() is defined */
117: #ifndef isascii
118: #define isascii(c) ((c) <= 0177)
119: #endif
120:
121:
122: #define true 1
123: #define false 0
124: #define unspecified -1
125:
126:
127: /* Special chk[] values marking the slots taking by end-of-buffer and action
128: * numbers.
129: */
130: #define EOB_POSITION -1
131: #define ACTION_POSITION -2
132:
133: /* Number of data items per line for -f output. */
134: #define NUMDATAITEMS 10
135:
136: /* Number of lines of data in -f output before inserting a blank line for
137: * readability.
138: */
139: #define NUMDATALINES 10
140:
141: /* transition_struct_out() definitions. */
142: #define TRANS_STRUCT_PRINT_LENGTH 14
143:
144: /* Returns true if an nfa state has an epsilon out-transition slot
145: * that can be used. This definition is currently not used.
146: */
147: #define FREE_EPSILON(state) \
148: (transchar[state] == SYM_EPSILON && \
149: trans2[state] == NO_TRANSITION && \
150: finalst[state] != state)
151:
152: /* Returns true if an nfa state has an epsilon out-transition character
153: * and both slots are free
154: */
155: #define SUPER_FREE_EPSILON(state) \
156: (transchar[state] == SYM_EPSILON && \
157: trans1[state] == NO_TRANSITION) \
158:
159: /* Maximum number of NFA states that can comprise a DFA state. It's real
160: * big because if there's a lot of rules, the initial state will have a
161: * huge epsilon closure.
162: */
163: #define INITIAL_MAX_DFA_SIZE 750
164: #define MAX_DFA_SIZE_INCREMENT 750
165:
166:
167: /* A note on the following masks. They are used to mark accepting numbers
168: * as being special. As such, they implicitly limit the number of accepting
169: * numbers (i.e., rules) because if there are too many rules the rule numbers
170: * will overload the mask bits. Fortunately, this limit is \large/ (0x2000 ==
171: * 8192) so unlikely to actually cause any problems. A check is made in
172: * new_rule() to ensure that this limit is not reached.
173: */
174:
175: /* Mask to mark a trailing context accepting number. */
176: #define YY_TRAILING_MASK 0x2000
177:
178: /* Mask to mark the accepting number of the "head" of a trailing context
179: * rule.
180: */
181: #define YY_TRAILING_HEAD_MASK 0x4000
182:
183: /* Maximum number of rules, as outlined in the above note. */
184: #define MAX_RULE (YY_TRAILING_MASK - 1)
185:
186:
187: /* NIL must be 0. If not, its special meaning when making equivalence classes
188: * (it marks the representative of a given e.c.) will be unidentifiable.
189: */
190: #define NIL 0
191:
192: #define JAM -1 /* to mark a missing DFA transition */
193: #define NO_TRANSITION NIL
194: #define UNIQUE -1 /* marks a symbol as an e.c. representative */
195: #define INFINITY -1 /* for x{5,} constructions */
196:
197: #define INITIAL_MAX_CCLS 100 /* max number of unique character classes */
198: #define MAX_CCLS_INCREMENT 100
199:
200: /* Size of table holding members of character classes. */
201: #define INITIAL_MAX_CCL_TBL_SIZE 500
202: #define MAX_CCL_TBL_SIZE_INCREMENT 250
203:
204: #define INITIAL_MAX_RULES 100 /* default maximum number of rules */
205: #define MAX_RULES_INCREMENT 100
206:
207: #define INITIAL_MNS 2000 /* default maximum number of nfa states */
208: #define MNS_INCREMENT 1000 /* amount to bump above by if it's not enough */
209:
210: #define INITIAL_MAX_DFAS 1000 /* default maximum number of dfa states */
211: #define MAX_DFAS_INCREMENT 1000
212:
213: #define JAMSTATE -32766 /* marks a reference to the state that always jams */
214:
215: /* Maximum number of NFA states. */
216: #define MAXIMUM_MNS 31999
217:
218: /* Enough so that if it's subtracted from an NFA state number, the result
219: * is guaranteed to be negative.
220: */
221: #define MARKER_DIFFERENCE (MAXIMUM_MNS+2)
222:
223: /* Maximum number of nxt/chk pairs for non-templates. */
224: #define INITIAL_MAX_XPAIRS 2000
225: #define MAX_XPAIRS_INCREMENT 2000
226:
227: /* Maximum number of nxt/chk pairs needed for templates. */
228: #define INITIAL_MAX_TEMPLATE_XPAIRS 2500
229: #define MAX_TEMPLATE_XPAIRS_INCREMENT 2500
230:
231: #define SYM_EPSILON (CSIZE + 1) /* to mark transitions on the symbol epsilon */
232:
233: #define INITIAL_MAX_SCS 40 /* maximum number of start conditions */
234: #define MAX_SCS_INCREMENT 40 /* amount to bump by if it's not enough */
235:
236: #define ONE_STACK_SIZE 500 /* stack of states with only one out-transition */
237: #define SAME_TRANS -1 /* transition is the same as "default" entry for state */
238:
239: /* The following percentages are used to tune table compression:
240:
241: * The percentage the number of out-transitions a state must be of the
242: * number of equivalence classes in order to be considered for table
243: * compaction by using protos.
244: */
245: #define PROTO_SIZE_PERCENTAGE 15
246:
247: /* The percentage the number of homogeneous out-transitions of a state
248: * must be of the number of total out-transitions of the state in order
249: * that the state's transition table is first compared with a potential
250: * template of the most common out-transition instead of with the first
251: * proto in the proto queue.
252: */
253: #define CHECK_COM_PERCENTAGE 50
254:
255: /* The percentage the number of differences between a state's transition
256: * table and the proto it was first compared with must be of the total
257: * number of out-transitions of the state in order to keep the first
258: * proto as a good match and not search any further.
259: */
260: #define FIRST_MATCH_DIFF_PERCENTAGE 10
261:
262: /* The percentage the number of differences between a state's transition
263: * table and the most similar proto must be of the state's total number
264: * of out-transitions to use the proto as an acceptable close match.
265: */
266: #define ACCEPTABLE_DIFF_PERCENTAGE 50
267:
268: /* The percentage the number of homogeneous out-transitions of a state
269: * must be of the number of total out-transitions of the state in order
270: * to consider making a template from the state.
271: */
272: #define TEMPLATE_SAME_PERCENTAGE 60
273:
274: /* The percentage the number of differences between a state's transition
275: * table and the most similar proto must be of the state's total number
276: * of out-transitions to create a new proto from the state.
277: */
278: #define NEW_PROTO_DIFF_PERCENTAGE 20
279:
280: /* The percentage the total number of out-transitions of a state must be
281: * of the number of equivalence classes in order to consider trying to
282: * fit the transition table into "holes" inside the nxt/chk table.
283: */
284: #define INTERIOR_FIT_PERCENTAGE 15
285:
286: /* Size of region set aside to cache the complete transition table of
287: * protos on the proto queue to enable quick comparisons.
288: */
289: #define PROT_SAVE_SIZE 2000
290:
291: #define MSP 50 /* maximum number of saved protos (protos on the proto queue) */
292:
293: /* Maximum number of out-transitions a state can have that we'll rummage
294: * around through the interior of the internal fast table looking for a
295: * spot for it.
296: */
297: #define MAX_XTIONS_FULL_INTERIOR_FIT 4
298:
299: /* Maximum number of rules which will be reported as being associated
300: * with a DFA state.
301: */
302: #define MAX_ASSOC_RULES 100
303:
304: /* Number that, if used to subscript an array, has a good chance of producing
305: * an error; should be small enough to fit into a short.
306: */
307: #define BAD_SUBSCRIPT -32767
308:
309: /* Absolute value of largest number that can be stored in a short, with a
310: * bit of slop thrown in for general paranoia.
311: */
312: #define MAX_SHORT 32700
313:
314:
315: /* Declarations for global variables. */
316:
317: /* Variables for symbol tables:
318: * sctbl - start-condition symbol table
319: * ndtbl - name-definition symbol table
320: * ccltab - character class text symbol table
321: */
322:
323: struct hash_entry
324: {
325: struct hash_entry *prev, *next;
326: char *name;
327: char *str_val;
328: int int_val;
329: } ;
330:
331: typedef struct hash_entry **hash_table;
332:
333: #define NAME_TABLE_HASH_SIZE 101
334: #define START_COND_HASH_SIZE 101
335: #define CCL_HASH_SIZE 101
336:
337: extern struct hash_entry *ndtbl[NAME_TABLE_HASH_SIZE];
338: extern struct hash_entry *sctbl[START_COND_HASH_SIZE];
339: extern struct hash_entry *ccltab[CCL_HASH_SIZE];
340:
341:
342: /* Variables for flags:
343: * printstats - if true (-v), dump statistics
344: * syntaxerror - true if a syntax error has been found
345: * eofseen - true if we've seen an eof in the input file
346: * ddebug - if true (-d), make a "debug" scanner
347: * trace - if true (-T), trace processing
348: * nowarn - if true (-w), do not generate warnings
349: * spprdflt - if true (-s), suppress the default rule
350: * interactive - if true (-I), generate an interactive scanner
351: * caseins - if true (-i), generate a case-insensitive scanner
352: * lex_compat - if true (-l), maximize compatibility with AT&T lex
353: * do_yylineno - if true, generate code to maintain yylineno
354: * useecs - if true (-Ce flag), use equivalence classes
355: * fulltbl - if true (-Cf flag), don't compress the DFA state table
356: * usemecs - if true (-Cm flag), use meta-equivalence classes
357: * fullspd - if true (-F flag), use Jacobson method of table representation
358: * gen_line_dirs - if true (i.e., no -L flag), generate #line directives
359: * performance_report - if > 0 (i.e., -p flag), generate a report relating
360: * to scanner performance; if > 1 (-p -p), report on minor performance
361: * problems, too
362: * backing_up_report - if true (i.e., -b flag), generate "lex.backup" file
363: * listing backing-up states
364: * C_plus_plus - if true (i.e., -+ flag), generate a C++ scanner class;
365: * otherwise, a standard C scanner
366: * long_align - if true (-Ca flag), favor long-word alignment.
367: * use_read - if true (-f, -F, or -Cr) then use read() for scanner input;
368: * otherwise, use fread().
369: * yytext_is_array - if true (i.e., %array directive), then declare
370: * yytext as a array instead of a character pointer. Nice and inefficient.
371: * do_yywrap - do yywrap() processing on EOF. If false, EOF treated as
372: * "no more files".
373: * csize - size of character set for the scanner we're generating;
374: * 128 for 7-bit chars and 256 for 8-bit
375: * yymore_used - if true, yymore() is used in input rules
376: * reject - if true, generate back-up tables for REJECT macro
377: * real_reject - if true, scanner really uses REJECT (as opposed to just
378: * having "reject" set for variable trailing context)
379: * continued_action - true if this rule's action is to "fall through" to
380: * the next rule's action (i.e., the '|' action)
381: * in_rule - true if we're inside an individual rule, false if not.
382: * yymore_really_used - whether to treat yymore() as really used, regardless
383: * of what we think based on references to it in the user's actions.
384: * reject_really_used - same for REJECT
385: */
386:
387: extern int printstats, syntaxerror, eofseen, ddebug, trace, nowarn, spprdflt;
388: extern int interactive, caseins, lex_compat, do_yylineno;
389: extern int useecs, fulltbl, usemecs, fullspd;
390: extern int gen_line_dirs, performance_report, backing_up_report;
391: extern int C_plus_plus, long_align, use_read, yytext_is_array, do_yywrap;
392: extern int csize;
393: extern int yymore_used, reject, real_reject, continued_action, in_rule;
394:
395: extern int yymore_really_used, reject_really_used;
396:
397:
398: /* Variables used in the flex input routines:
399: * datapos - characters on current output line
400: * dataline - number of contiguous lines of data in current data
401: * statement. Used to generate readable -f output
402: * linenum - current input line number
403: * out_linenum - current output line number
404: * skelfile - the skeleton file
405: * skel - compiled-in skeleton array
406: * skel_ind - index into "skel" array, if skelfile is nil
407: * yyin - input file
408: * backing_up_file - file to summarize backing-up states to
409: * infilename - name of input file
410: * outfilename - name of output file
411: * did_outfilename - whether outfilename was explicitly set
412: * prefix - the prefix used for externally visible names ("yy" by default)
413: * yyclass - yyFlexLexer subclass to use for YY_DECL
414: * do_stdinit - whether to initialize yyin/yyout to stdin/stdout
415: * use_stdout - the -t flag
416: * input_files - array holding names of input files
417: * num_input_files - size of input_files array
418: * program_name - name with which program was invoked
419: *
420: * action_array - array to hold the rule actions
421: * action_size - size of action_array
422: * defs1_offset - index where the user's section 1 definitions start
423: * in action_array
424: * prolog_offset - index where the prolog starts in action_array
425: * action_offset - index where the non-prolog starts in action_array
426: * action_index - index where the next action should go, with respect
427: * to "action_array"
428: */
429:
430: extern int datapos, dataline, linenum, out_linenum;
431: extern FILE *skelfile, *yyin, *backing_up_file;
432: extern const char *skel[];
433: extern int skel_ind;
434: extern char *infilename, *outfilename;
435: extern int did_outfilename;
436: extern char *prefix, *yyclass;
437: extern int do_stdinit, use_stdout;
438: extern char **input_files;
439: extern int num_input_files;
440: extern char *program_name;
441:
442: extern char *action_array;
443: extern int action_size;
444: extern int defs1_offset, prolog_offset, action_offset, action_index;
445:
446:
447: /* Variables for stack of states having only one out-transition:
448: * onestate - state number
449: * onesym - transition symbol
450: * onenext - target state
451: * onedef - default base entry
452: * onesp - stack pointer
453: */
454:
455: extern int onestate[ONE_STACK_SIZE], onesym[ONE_STACK_SIZE];
456: extern int onenext[ONE_STACK_SIZE], onedef[ONE_STACK_SIZE], onesp;
457:
458:
459: /* Variables for nfa machine data:
460: * current_mns - current maximum on number of NFA states
461: * num_rules - number of the last accepting state; also is number of
462: * rules created so far
463: * num_eof_rules - number of <<EOF>> rules
464: * default_rule - number of the default rule
465: * current_max_rules - current maximum number of rules
466: * lastnfa - last nfa state number created
467: * firstst - physically the first state of a fragment
468: * lastst - last physical state of fragment
469: * finalst - last logical state of fragment
470: * transchar - transition character
471: * trans1 - transition state
472: * trans2 - 2nd transition state for epsilons
473: * accptnum - accepting number
474: * assoc_rule - rule associated with this NFA state (or 0 if none)
475: * state_type - a STATE_xxx type identifying whether the state is part
476: * of a normal rule, the leading state in a trailing context
477: * rule (i.e., the state which marks the transition from
478: * recognizing the text-to-be-matched to the beginning of
479: * the trailing context), or a subsequent state in a trailing
480: * context rule
481: * rule_type - a RULE_xxx type identifying whether this a ho-hum
482: * normal rule or one which has variable head & trailing
483: * context
484: * rule_linenum - line number associated with rule
485: * rule_useful - true if we've determined that the rule can be matched
486: */
487:
488: extern int current_mns, current_max_rules;
489: extern int num_rules, num_eof_rules, default_rule, lastnfa;
490: extern int *firstst, *lastst, *finalst, *transchar, *trans1, *trans2;
491: extern int *accptnum, *assoc_rule, *state_type;
492: extern int *rule_type, *rule_linenum, *rule_useful;
493:
494: /* Different types of states; values are useful as masks, as well, for
495: * routines like check_trailing_context().
496: */
497: #define STATE_NORMAL 0x1
498: #define STATE_TRAILING_CONTEXT 0x2
499:
500: /* Global holding current type of state we're making. */
501:
502: extern int current_state_type;
503:
504: /* Different types of rules. */
505: #define RULE_NORMAL 0
506: #define RULE_VARIABLE 1
507:
508: /* True if the input rules include a rule with both variable-length head
509: * and trailing context, false otherwise.
510: */
511: extern int variable_trailing_context_rules;
512:
513:
514: /* Variables for protos:
515: * numtemps - number of templates created
516: * numprots - number of protos created
517: * protprev - backlink to a more-recently used proto
518: * protnext - forward link to a less-recently used proto
519: * prottbl - base/def table entry for proto
520: * protcomst - common state of proto
521: * firstprot - number of the most recently used proto
522: * lastprot - number of the least recently used proto
523: * protsave contains the entire state array for protos
524: */
525:
526: extern int numtemps, numprots, protprev[MSP], protnext[MSP], prottbl[MSP];
527: extern int protcomst[MSP], firstprot, lastprot, protsave[PROT_SAVE_SIZE];
528:
529:
530: /* Variables for managing equivalence classes:
531: * numecs - number of equivalence classes
532: * nextecm - forward link of Equivalence Class members
533: * ecgroup - class number or backward link of EC members
534: * nummecs - number of meta-equivalence classes (used to compress
535: * templates)
536: * tecfwd - forward link of meta-equivalence classes members
537: * tecbck - backward link of MEC's
538: */
539:
540: /* Reserve enough room in the equivalence class arrays so that we
541: * can use the CSIZE'th element to hold equivalence class information
542: * for the NUL character. Later we'll move this information into
543: * the 0th element.
544: */
545: extern int numecs, nextecm[CSIZE + 1], ecgroup[CSIZE + 1], nummecs;
546:
547: /* Meta-equivalence classes are indexed starting at 1, so it's possible
548: * that they will require positions from 1 .. CSIZE, i.e., CSIZE + 1
549: * slots total (since the arrays are 0-based). nextecm[] and ecgroup[]
550: * don't require the extra position since they're indexed from 1 .. CSIZE - 1.
551: */
552: extern int tecfwd[CSIZE + 1], tecbck[CSIZE + 1];
553:
554:
555: /* Variables for start conditions:
556: * lastsc - last start condition created
557: * current_max_scs - current limit on number of start conditions
558: * scset - set of rules active in start condition
559: * scbol - set of rules active only at the beginning of line in a s.c.
560: * scxclu - true if start condition is exclusive
561: * sceof - true if start condition has EOF rule
562: * scname - start condition name
563: */
564:
565: extern int lastsc, *scset, *scbol, *scxclu, *sceof;
566: extern int current_max_scs;
567: extern char **scname;
568:
569:
570: /* Variables for dfa machine data:
571: * current_max_dfa_size - current maximum number of NFA states in DFA
572: * current_max_xpairs - current maximum number of non-template xtion pairs
573: * current_max_template_xpairs - current maximum number of template pairs
574: * current_max_dfas - current maximum number DFA states
575: * lastdfa - last dfa state number created
576: * nxt - state to enter upon reading character
577: * chk - check value to see if "nxt" applies
578: * tnxt - internal nxt table for templates
579: * base - offset into "nxt" for given state
580: * def - where to go if "chk" disallows "nxt" entry
581: * nultrans - NUL transition for each state
582: * NUL_ec - equivalence class of the NUL character
583: * tblend - last "nxt/chk" table entry being used
584: * firstfree - first empty entry in "nxt/chk" table
585: * dss - nfa state set for each dfa
586: * dfasiz - size of nfa state set for each dfa
587: * dfaacc - accepting set for each dfa state (if using REJECT), or accepting
588: * number, if not
589: * accsiz - size of accepting set for each dfa state
590: * dhash - dfa state hash value
591: * numas - number of DFA accepting states created; note that this
592: * is not necessarily the same value as num_rules, which is the analogous
593: * value for the NFA
594: * numsnpairs - number of state/nextstate transition pairs
595: * jambase - position in base/def where the default jam table starts
596: * jamstate - state number corresponding to "jam" state
597: * end_of_buffer_state - end-of-buffer dfa state number
598: */
599:
600: extern int current_max_dfa_size, current_max_xpairs;
601: extern int current_max_template_xpairs, current_max_dfas;
602: extern int lastdfa, *nxt, *chk, *tnxt;
603: extern int *base, *def, *nultrans, NUL_ec, tblend, firstfree, **dss, *dfasiz;
604: extern union dfaacc_union
605: {
606: int *dfaacc_set;
607: int dfaacc_state;
608: } *dfaacc;
609: extern int *accsiz, *dhash, numas;
610: extern int numsnpairs, jambase, jamstate;
611: extern int end_of_buffer_state;
612:
613: /* Variables for ccl information:
614: * lastccl - ccl index of the last created ccl
615: * current_maxccls - current limit on the maximum number of unique ccl's
616: * cclmap - maps a ccl index to its set pointer
617: * ccllen - gives the length of a ccl
618: * cclng - true for a given ccl if the ccl is negated
619: * cclreuse - counts how many times a ccl is re-used
620: * current_max_ccl_tbl_size - current limit on number of characters needed
621: * to represent the unique ccl's
622: * ccltbl - holds the characters in each ccl - indexed by cclmap
623: */
624:
625: extern int lastccl, *cclmap, *ccllen, *cclng, cclreuse;
626: extern int current_maxccls, current_max_ccl_tbl_size;
627: extern Char *ccltbl;
628:
629:
630: /* Variables for miscellaneous information:
631: * nmstr - last NAME scanned by the scanner
632: * sectnum - section number currently being parsed
633: * nummt - number of empty nxt/chk table entries
634: * hshcol - number of hash collisions detected by snstods
635: * dfaeql - number of times a newly created dfa was equal to an old one
636: * numeps - number of epsilon NFA states created
637: * eps2 - number of epsilon states which have 2 out-transitions
638: * num_reallocs - number of times it was necessary to realloc() a group
639: * of arrays
640: * tmpuses - number of DFA states that chain to templates
641: * totnst - total number of NFA states used to make DFA states
642: * peakpairs - peak number of transition pairs we had to store internally
643: * numuniq - number of unique transitions
644: * numdup - number of duplicate transitions
645: * hshsave - number of hash collisions saved by checking number of states
646: * num_backing_up - number of DFA states requiring backing up
647: * bol_needed - whether scanner needs beginning-of-line recognition
648: */
649:
650: extern char nmstr[MAXLINE];
651: extern int sectnum, nummt, hshcol, dfaeql, numeps, eps2, num_reallocs;
652: extern int tmpuses, totnst, peakpairs, numuniq, numdup, hshsave;
653: extern int num_backing_up, bol_needed;
654:
655: void *allocate_array PROTO((int, size_t));
656: void *reallocate_array PROTO((void*, int, size_t));
657:
658: void *flex_alloc PROTO((size_t));
659: void *flex_realloc PROTO((void*, size_t));
660: void flex_free PROTO((void*));
661:
662: #define allocate_integer_array(size) \
663: (int *) allocate_array( size, sizeof( int ) )
664:
665: #define reallocate_integer_array(array,size) \
666: (int *) reallocate_array( (void *) array, size, sizeof( int ) )
667:
668: #define allocate_int_ptr_array(size) \
669: (int **) allocate_array( size, sizeof( int * ) )
670:
671: #define allocate_char_ptr_array(size) \
672: (char **) allocate_array( size, sizeof( char * ) )
673:
674: #define allocate_dfaacc_union(size) \
675: (union dfaacc_union *) \
676: allocate_array( size, sizeof( union dfaacc_union ) )
677:
678: #define reallocate_int_ptr_array(array,size) \
679: (int **) reallocate_array( (void *) array, size, sizeof( int * ) )
680:
681: #define reallocate_char_ptr_array(array,size) \
682: (char **) reallocate_array( (void *) array, size, sizeof( char * ) )
683:
684: #define reallocate_dfaacc_union(array, size) \
685: (union dfaacc_union *) \
686: reallocate_array( (void *) array, size, sizeof( union dfaacc_union ) )
687:
688: #define allocate_character_array(size) \
689: (char *) allocate_array( size, sizeof( char ) )
690:
691: #define reallocate_character_array(array,size) \
692: (char *) reallocate_array( (void *) array, size, sizeof( char ) )
693:
694: #define allocate_Character_array(size) \
695: (Char *) allocate_array( size, sizeof( Char ) )
696:
697: #define reallocate_Character_array(array,size) \
698: (Char *) reallocate_array( (void *) array, size, sizeof( Char ) )
699:
700:
701: /* Used to communicate between scanner and parser. The type should really
702: * be YYSTYPE, but we can't easily get our hands on it.
703: */
704: extern int yylval;
705:
706:
707: /* External functions that are cross-referenced among the flex source files. */
708:
709:
710: /* from file ccl.c */
711:
712: extern void ccladd PROTO((int, int)); /* add a single character to a ccl */
713: extern int cclinit PROTO((void)); /* make an empty ccl */
714: extern void cclnegate PROTO((int)); /* negate a ccl */
715:
716: /* List the members of a set of characters in CCL form. */
717: extern void list_character_set PROTO((FILE*, int[]));
718:
719:
720: /* from file dfa.c */
721:
722: /* Check a DFA state for backing up. */
723: extern void check_for_backing_up PROTO((int, int[]));
724:
725: /* Check to see if NFA state set constitutes "dangerous" trailing context. */
726: extern void check_trailing_context PROTO((int*, int, int*, int));
727:
728: /* Construct the epsilon closure of a set of ndfa states. */
729: extern int *epsclosure PROTO((int*, int*, int[], int*, int*));
730:
731: /* Increase the maximum number of dfas. */
732: extern void increase_max_dfas PROTO((void));
733:
734: extern void ntod PROTO((void)); /* convert a ndfa to a dfa */
735:
736: /* Converts a set of ndfa states into a dfa state. */
737: extern int snstods PROTO((int[], int, int[], int, int, int*));
738:
739:
740: /* from file ecs.c */
741:
742: /* Convert character classes to set of equivalence classes. */
743: extern void ccl2ecl PROTO((void));
744:
745: /* Associate equivalence class numbers with class members. */
746: extern int cre8ecs PROTO((int[], int[], int));
747:
748: /* Update equivalence classes based on character class transitions. */
749: extern void mkeccl PROTO((Char[], int, int[], int[], int, int));
750:
751: /* Create equivalence class for single character. */
752: extern void mkechar PROTO((int, int[], int[]));
753:
754:
755: /* from file gen.c */
756:
757: extern void do_indent PROTO((void)); /* indent to the current level */
758:
759: /* Generate the code to keep backing-up information. */
760: extern void gen_backing_up PROTO((void));
761:
762: /* Generate the code to perform the backing up. */
763: extern void gen_bu_action PROTO((void));
764:
765: /* Generate full speed compressed transition table. */
766: extern void genctbl PROTO((void));
767:
768: /* Generate the code to find the action number. */
769: extern void gen_find_action PROTO((void));
770:
771: extern void genftbl PROTO((void)); /* generate full transition table */
772:
773: /* Generate the code to find the next compressed-table state. */
774: extern void gen_next_compressed_state PROTO((char*));
775:
776: /* Generate the code to find the next match. */
777: extern void gen_next_match PROTO((void));
778:
779: /* Generate the code to find the next state. */
780: extern void gen_next_state PROTO((int));
781:
782: /* Generate the code to make a NUL transition. */
783: extern void gen_NUL_trans PROTO((void));
784:
785: /* Generate the code to find the start state. */
786: extern void gen_start_state PROTO((void));
787:
788: /* Generate data statements for the transition tables. */
789: extern void gentabs PROTO((void));
790:
791: /* Write out a formatted string at the current indentation level. */
792: extern void indent_put2s PROTO((char[], char[]));
793:
794: /* Write out a string + newline at the current indentation level. */
795: extern void indent_puts PROTO((char[]));
796:
797: extern void make_tables PROTO((void)); /* generate transition tables */
798:
799:
800: /* from file main.c */
801:
802: extern void check_options PROTO((void));
803: extern void flexend PROTO((int));
804: extern void usage PROTO((void));
805:
806:
807: /* from file misc.c */
808:
809: /* Add a #define to the action file. */
810: extern void action_define PROTO(( char *defname, int value ));
811:
812: /* Add the given text to the stored actions. */
813: extern void add_action PROTO(( char *new_text ));
814:
815: /* True if a string is all lower case. */
816: extern int all_lower PROTO((register char *));
817:
818: /* True if a string is all upper case. */
819: extern int all_upper PROTO((register char *));
820:
821: /* Bubble sort an integer array. */
822: extern void bubble PROTO((int [], int));
823:
824: /* Check a character to make sure it's in the expected range. */
825: extern void check_char PROTO((int c));
826:
827: /* Replace upper-case letter to lower-case. */
828: extern Char clower PROTO((int));
829:
830: /* Returns a dynamically allocated copy of a string. */
831: extern char *copy_string PROTO((register const char *));
832:
833: /* Returns a dynamically allocated copy of a (potentially) unsigned string. */
834: extern Char *copy_unsigned_string PROTO((register Char *));
835:
836: /* Shell sort a character array. */
837: extern void cshell PROTO((Char [], int, int));
838:
839: /* Finish up a block of data declarations. */
840: extern void dataend PROTO((void));
841:
842: /* Flush generated data statements. */
843: extern void dataflush PROTO((void));
844:
845: /* Report an error message and terminate. */
846: extern void flexerror PROTO((const char[]));
847:
848: /* Report a fatal error message and terminate. */
849: extern void flexfatal PROTO((const char[]));
850:
851: /* Convert a hexadecimal digit string to an integer value. */
852: extern int htoi PROTO((Char[]));
853:
854: /* Report an error message formatted with one integer argument. */
855: extern void lerrif PROTO((const char[], int));
856:
857: /* Report an error message formatted with one string argument. */
858: extern void lerrsf PROTO((const char[], const char[]));
859:
860: /* Spit out a "#line" statement. */
861: extern void line_directive_out PROTO((FILE*, int));
862:
863: /* Mark the current position in the action array as the end of the section 1
864: * user defs.
865: */
866: extern void mark_defs1 PROTO((void));
867:
868: /* Mark the current position in the action array as the end of the prolog. */
869: extern void mark_prolog PROTO((void));
870:
871: /* Generate a data statment for a two-dimensional array. */
872: extern void mk2data PROTO((int));
873:
874: extern void mkdata PROTO((int)); /* generate a data statement */
875:
876: /* Return the integer represented by a string of digits. */
877: extern int myctoi PROTO((char []));
878:
879: /* Return character corresponding to escape sequence. */
880: extern Char myesc PROTO((Char[]));
881:
882: /* Convert an octal digit string to an integer value. */
883: extern int otoi PROTO((Char [] ));
884:
885: /* Output a (possibly-formatted) string to the generated scanner. */
886: extern void out PROTO((const char []));
887: extern void out_dec PROTO((const char [], int));
888: extern void out_dec2 PROTO((const char [], int, int));
889: extern void out_hex PROTO((const char [], unsigned int));
890: extern void out_line_count PROTO((const char []));
891: extern void out_str PROTO((const char [], const char []));
892: extern void out_str3
893: PROTO((const char [], const char [], const char [], const char []));
894: extern void out_str_dec PROTO((const char [], const char [], int));
895: extern void outc PROTO((int));
896: extern void outn PROTO((const char []));
897:
898: /* Return a printable version of the given character, which might be
899: * 8-bit.
900: */
901: extern char *readable_form PROTO((int));
902:
903: /* Write out one section of the skeleton file. */
904: extern void skelout PROTO((void));
905:
906: /* Output a yy_trans_info structure. */
907: extern void transition_struct_out PROTO((int, int));
908:
909: /* Only needed when using certain broken versions of bison to build parse.c. */
910: extern void *yy_flex_xmalloc PROTO(( int ));
911:
912: /* Set a region of memory to 0. */
913: extern void zero_out PROTO((char *, size_t));
914:
915:
916: /* from file nfa.c */
917:
918: /* Add an accepting state to a machine. */
919: extern void add_accept PROTO((int, int));
920:
921: /* Make a given number of copies of a singleton machine. */
922: extern int copysingl PROTO((int, int));
923:
924: /* Debugging routine to write out an nfa. */
925: extern void dumpnfa PROTO((int));
926:
927: /* Finish up the processing for a rule. */
928: extern void finish_rule PROTO((int, int, int, int));
929:
930: /* Connect two machines together. */
931: extern int link_machines PROTO((int, int));
932:
933: /* Mark each "beginning" state in a machine as being a "normal" (i.e.,
934: * not trailing context associated) state.
935: */
936: extern void mark_beginning_as_normal PROTO((register int));
937:
938: /* Make a machine that branches to two machines. */
939: extern int mkbranch PROTO((int, int));
940:
941: extern int mkclos PROTO((int)); /* convert a machine into a closure */
942: extern int mkopt PROTO((int)); /* make a machine optional */
943:
944: /* Make a machine that matches either one of two machines. */
945: extern int mkor PROTO((int, int));
946:
947: /* Convert a machine into a positive closure. */
948: extern int mkposcl PROTO((int));
949:
950: extern int mkrep PROTO((int, int, int)); /* make a replicated machine */
951:
952: /* Create a state with a transition on a given symbol. */
953: extern int mkstate PROTO((int));
954:
955: extern void new_rule PROTO((void)); /* initialize for a new rule */
956:
957:
958: /* from file parse.y */
959:
960: /* Build the "<<EOF>>" action for the active start conditions. */
961: extern void build_eof_action PROTO((void));
962:
963: /* Write out a message formatted with one string, pinpointing its location. */
964: extern void format_pinpoint_message PROTO((char[], char[]));
965:
966: /* Write out a message, pinpointing its location. */
967: extern void pinpoint_message PROTO((char[]));
968:
969: /* Write out a warning, pinpointing it at the given line. */
970: extern void line_warning PROTO(( char[], int ));
971:
972: /* Write out a message, pinpointing it at the given line. */
973: extern void line_pinpoint PROTO(( char[], int ));
974:
975: /* Report a formatted syntax error. */
976: extern void format_synerr PROTO((char [], char[]));
977: extern void synerr PROTO((char [])); /* report a syntax error */
978: extern void format_warn PROTO((char [], char[]));
979: extern void warn PROTO((char [])); /* report a warning */
980: extern void yyerror PROTO((char [])); /* report a parse error */
981: extern int yyparse PROTO((void)); /* the YACC parser */
982:
983:
984: /* from file scan.l */
985:
986: /* The Flex-generated scanner for flex. */
987: extern int flexscan PROTO((void));
988:
989: /* Open the given file (if NULL, stdin) for scanning. */
990: extern void set_input_file PROTO((char*));
991:
992: /* Wrapup a file in the lexical analyzer. */
993: extern int yywrap PROTO((void));
994:
995:
996: /* from file sym.c */
997:
998: /* Add symbol and definitions to symbol table. */
999: extern int addsym PROTO((register char[], char*, int, hash_table, int));
1000:
1001: /* Save the text of a character class. */
1002: extern void cclinstal PROTO ((Char [], int));
1003:
1004: /* Lookup the number associated with character class. */
1005: extern int ccllookup PROTO((Char []));
1006:
1007: /* Find symbol in symbol table. */
1008: extern struct hash_entry *findsym PROTO((register char[], hash_table, int ));
1009:
1010: extern void ndinstal PROTO((char[], Char[])); /* install a name definition */
1011: extern Char *ndlookup PROTO((char[])); /* lookup a name definition */
1012:
1013: /* Increase maximum number of SC's. */
1014: extern void scextend PROTO((void));
1015: extern void scinstal PROTO((char[], int)); /* make a start condition */
1016:
1017: /* Lookup the number associated with a start condition. */
1018: extern int sclookup PROTO((char[]));
1019:
1020:
1021: /* from file tblcmp.c */
1022:
1023: /* Build table entries for dfa state. */
1024: extern void bldtbl PROTO((int[], int, int, int, int));
1025:
1026: extern void cmptmps PROTO((void)); /* compress template table entries */
1027: extern void expand_nxt_chk PROTO((void)); /* increase nxt/chk arrays */
1028: /* Finds a space in the table for a state to be placed. */
1029: extern int find_table_space PROTO((int*, int));
1030: extern void inittbl PROTO((void)); /* initialize transition tables */
1031: /* Make the default, "jam" table entries. */
1032: extern void mkdeftbl PROTO((void));
1033:
1034: /* Create table entries for a state (or state fragment) which has
1035: * only one out-transition.
1036: */
1037: extern void mk1tbl PROTO((int, int, int, int));
1038:
1039: /* Place a state into full speed transition table. */
1040: extern void place_state PROTO((int*, int, int));
1041:
1042: /* Save states with only one out-transition to be processed later. */
1043: extern void stack1 PROTO((int, int, int, int));
1044:
1045:
1046: /* from file yylex.c */
1047:
1048: extern int yylex PROTO((void));