xref: /freebsd/usr.bin/indent/lexi.c (revision 6ef644f5889afbd0f681b08ed1a2f369524af83e)
1 /*-
2  * SPDX-License-Identifier: BSD-4-Clause
3  *
4  * Copyright (c) 1985 Sun Microsystems, Inc.
5  * Copyright (c) 1980, 1993
6  *	The Regents of the University of California.  All rights reserved.
7  * All rights reserved.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions
11  * are met:
12  * 1. Redistributions of source code must retain the above copyright
13  *    notice, this list of conditions and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  * 3. All advertising materials mentioning features or use of this software
18  *    must display the following acknowledgement:
19  *	This product includes software developed by the University of
20  *	California, Berkeley and its contributors.
21  * 4. Neither the name of the University nor the names of its contributors
22  *    may be used to endorse or promote products derived from this software
23  *    without specific prior written permission.
24  *
25  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35  * SUCH DAMAGE.
36  */
37 
38 #if 0
39 #ifndef lint
40 static char sccsid[] = "@(#)lexi.c	8.1 (Berkeley) 6/6/93";
41 #endif /* not lint */
42 #endif
43 #include <sys/cdefs.h>
44 /*
45  * Here we have the token scanner for indent.  It scans off one token and puts
46  * it in the global variable "token".  It returns a code, indicating the type
47  * of token scanned.
48  */
49 
50 #include <err.h>
51 #include <stdio.h>
52 #include <ctype.h>
53 #include <stdlib.h>
54 #include <string.h>
55 #include <sys/param.h>
56 
57 #include "indent_globs.h"
58 #include "indent_codes.h"
59 #include "indent.h"
60 
61 struct templ {
62     const char *rwd;
63     int         rwcode;
64 };
65 
66 /*
67  * This table has to be sorted alphabetically, because it'll be used in binary
68  * search. For the same reason, string must be the first thing in struct templ.
69  */
70 struct templ specials[] =
71 {
72     {"_Bool", 4},
73     {"_Complex", 4},
74     {"_Imaginary", 4},
75     {"auto", 10},
76     {"bool", 4},
77     {"break", 9},
78     {"case", 8},
79     {"char", 4},
80     {"complex", 4},
81     {"const", 4},
82     {"continue", 12},
83     {"default", 8},
84     {"do", 6},
85     {"double", 4},
86     {"else", 6},
87     {"enum", 3},
88     {"extern", 10},
89     {"float", 4},
90     {"for", 5},
91     {"global", 4},
92     {"goto", 9},
93     {"if", 5},
94     {"imaginary", 4},
95     {"inline", 12},
96     {"int", 4},
97     {"long", 4},
98     {"offsetof", 1},
99     {"register", 10},
100     {"restrict", 12},
101     {"return", 9},
102     {"short", 4},
103     {"signed", 4},
104     {"sizeof", 2},
105     {"static", 10},
106     {"struct", 3},
107     {"switch", 7},
108     {"typedef", 11},
109     {"union", 3},
110     {"unsigned", 4},
111     {"void", 4},
112     {"volatile", 4},
113     {"while", 5}
114 };
115 
116 const char **typenames;
117 int         typename_count;
118 int         typename_top = -1;
119 
120 /*
121  * The transition table below was rewritten by hand from lx's output, given
122  * the following definitions. lx is Katherine Flavel's lexer generator.
123  *
124  * O  = /[0-7]/;        D  = /[0-9]/;          NZ = /[1-9]/;
125  * H  = /[a-f0-9]/i;    B  = /[0-1]/;          HP = /0x/i;
126  * BP = /0b/i;          E  = /e[+\-]?/i D+;    P  = /p[+\-]?/i D+;
127  * FS = /[fl]/i;        IS = /u/i /(l|L|ll|LL)/? | /(l|L|ll|LL)/ /u/i?;
128  *
129  * D+           E  FS? -> $float;
130  * D*    "." D+ E? FS? -> $float;
131  * D+    "."    E? FS? -> $float;    HP H+           IS? -> $int;
132  * HP H+        P  FS? -> $float;    NZ D*           IS? -> $int;
133  * HP H* "." H+ P  FS? -> $float;    "0" O*          IS? -> $int;
134  * HP H+ "."    P  FS  -> $float;    BP B+           IS? -> $int;
135  */
136 static char const *table[] = {
137     /*                examples:
138                                      00
139              s                      0xx
140              t                    00xaa
141              a     11       101100xxa..
142              r   11ee0001101lbuuxx.a.pp
143              t.01.e+008bLuxll0Ll.aa.p+0
144     states:  ABCDEFGHIJKLMNOPQRSTUVWXYZ */
145     ['0'] = "CEIDEHHHIJQ  U  Q  VUVVZZZ",
146     ['1'] = "DEIDEHHHIJQ  U  Q  VUVVZZZ",
147     ['7'] = "DEIDEHHHIJ   U     VUVVZZZ",
148     ['9'] = "DEJDEHHHJJ   U     VUVVZZZ",
149     ['a'] = "             U     VUVV   ",
150     ['b'] = "  K          U     VUVV   ",
151     ['e'] = "  FFF   FF   U     VUVV   ",
152     ['f'] = "    f  f     U     VUVV  f",
153     ['u'] = "  MM    M  i  iiM   M     ",
154     ['x'] = "  N                       ",
155     ['p'] = "                    FFX   ",
156     ['L'] = "  LLf  fL  PR   Li  L    f",
157     ['l'] = "  OOf  fO   S P O i O    f",
158     ['+'] = "     G                 Y  ",
159     ['.'] = "B EE    EE   T      W     ",
160     /*       ABCDEFGHIJKLMNOPQRSTUVWXYZ */
161     [0]   = "uuiifuufiuuiiuiiiiiuiuuuuu",
162 };
163 
164 static int
165 strcmp_type(const void *e1, const void *e2)
166 {
167     return (strcmp(e1, *(const char * const *)e2));
168 }
169 
170 int
171 lexi(struct parser_state *state)
172 {
173     int         unary_delim;	/* this is set to 1 if the current token
174 				 * forces a following operator to be unary */
175     int         code;		/* internal code to be returned */
176     char        qchar;		/* the delimiter character for a string */
177 
178     e_token = s_token;		/* point to start of place to save token */
179     unary_delim = false;
180     state->col_1 = state->last_nl;	/* tell world that this token started
181 					 * in column 1 iff the last thing
182 					 * scanned was a newline */
183     state->last_nl = false;
184 
185     while (*buf_ptr == ' ' || *buf_ptr == '\t') {	/* get rid of blanks */
186 	state->col_1 = false;	/* leading blanks imply token is not in column
187 				 * 1 */
188 	if (++buf_ptr >= buf_end)
189 	    fill_buffer();
190     }
191 
192     /* Scan an alphanumeric token */
193     if (isalnum((unsigned char)*buf_ptr) ||
194 	*buf_ptr == '_' || *buf_ptr == '$' ||
195 	(buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) {
196 	/*
197 	 * we have a character or number
198 	 */
199 	struct templ *p;
200 
201 	if (isdigit((unsigned char)*buf_ptr) ||
202 	    (buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) {
203 	    char s;
204 	    unsigned char i;
205 
206 	    for (s = 'A'; s != 'f' && s != 'i' && s != 'u'; ) {
207 		i = (unsigned char)*buf_ptr;
208 		if (i >= nitems(table) || table[i] == NULL ||
209 		    table[i][s - 'A'] == ' ') {
210 		    s = table[0][s - 'A'];
211 		    break;
212 		}
213 		s = table[i][s - 'A'];
214 		CHECK_SIZE_TOKEN(1);
215 		*e_token++ = *buf_ptr++;
216 		if (buf_ptr >= buf_end)
217 		    fill_buffer();
218 	    }
219 	    /* s now indicates the type: f(loating), i(integer), u(nknown) */
220 	}
221 	else
222 	    while (isalnum((unsigned char)*buf_ptr) ||
223 	        *buf_ptr == BACKSLASH ||
224 		*buf_ptr == '_' || *buf_ptr == '$') {
225 		/* fill_buffer() terminates buffer with newline */
226 		if (*buf_ptr == BACKSLASH) {
227 		    if (*(buf_ptr + 1) == '\n') {
228 			buf_ptr += 2;
229 			if (buf_ptr >= buf_end)
230 			    fill_buffer();
231 			} else
232 			    break;
233 		}
234 		CHECK_SIZE_TOKEN(1);
235 		/* copy it over */
236 		*e_token++ = *buf_ptr++;
237 		if (buf_ptr >= buf_end)
238 		    fill_buffer();
239 	    }
240 	*e_token = '\0';
241 
242 	if (s_token[0] == 'L' && s_token[1] == '\0' &&
243 	      (*buf_ptr == '"' || *buf_ptr == '\''))
244 	    return (strpfx);
245 
246 	while (*buf_ptr == ' ' || *buf_ptr == '\t') {	/* get rid of blanks */
247 	    if (++buf_ptr >= buf_end)
248 		fill_buffer();
249 	}
250 	state->keyword = 0;
251 	if (state->last_token == structure && !state->p_l_follow) {
252 				/* if last token was 'struct' and we're not
253 				 * in parentheses, then this token
254 				 * should be treated as a declaration */
255 	    state->last_u_d = true;
256 	    return (decl);
257 	}
258 	/*
259 	 * Operator after identifier is binary unless last token was 'struct'
260 	 */
261 	state->last_u_d = (state->last_token == structure);
262 
263 	p = bsearch(s_token,
264 	    specials,
265 	    sizeof(specials) / sizeof(specials[0]),
266 	    sizeof(specials[0]),
267 	    strcmp_type);
268 	if (p == NULL) {	/* not a special keyword... */
269 	    char *u;
270 
271 	    /* ... so maybe a type_t or a typedef */
272 	    if ((opt.auto_typedefs && ((u = strrchr(s_token, '_')) != NULL) &&
273 	        strcmp(u, "_t") == 0) || (typename_top >= 0 &&
274 		  bsearch(s_token, typenames, typename_top + 1,
275 		    sizeof(typenames[0]), strcmp_type))) {
276 		state->keyword = 4;	/* a type name */
277 		state->last_u_d = true;
278 	        goto found_typename;
279 	    }
280 	} else {			/* we have a keyword */
281 	    state->keyword = p->rwcode;
282 	    state->last_u_d = true;
283 	    switch (p->rwcode) {
284 	    case 7:		/* it is a switch */
285 		return (swstmt);
286 	    case 8:		/* a case or default */
287 		return (casestmt);
288 
289 	    case 3:		/* a "struct" */
290 		/* FALLTHROUGH */
291 	    case 4:		/* one of the declaration keywords */
292 	    found_typename:
293 		if (state->p_l_follow) {
294 		    /* inside parens: cast, param list, offsetof or sizeof */
295 		    state->cast_mask |= (1 << state->p_l_follow) & ~state->not_cast_mask;
296 		}
297 		if (state->last_token == period || state->last_token == unary_op) {
298 		    state->keyword = 0;
299 		    break;
300 		}
301 		if (p != NULL && p->rwcode == 3)
302 		    return (structure);
303 		if (state->p_l_follow)
304 		    break;
305 		return (decl);
306 
307 	    case 5:		/* if, while, for */
308 		return (sp_paren);
309 
310 	    case 6:		/* do, else */
311 		return (sp_nparen);
312 
313 	    case 10:		/* storage class specifier */
314 		return (storage);
315 
316 	    case 11:		/* typedef */
317 		return (type_def);
318 
319 	    default:		/* all others are treated like any other
320 				 * identifier */
321 		return (ident);
322 	    }			/* end of switch */
323 	}			/* end of if (found_it) */
324 	if (*buf_ptr == '(' && state->tos <= 1 && state->ind_level == 0 &&
325 	    state->in_parameter_declaration == 0 && state->block_init == 0) {
326 	    char *tp = buf_ptr;
327 	    while (tp < buf_end)
328 		if (*tp++ == ')' && (*tp == ';' || *tp == ','))
329 		    goto not_proc;
330 	    strncpy(state->procname, token, sizeof state->procname - 1);
331 	    if (state->in_decl)
332 		state->in_parameter_declaration = 1;
333 	    return (funcname);
334     not_proc:;
335 	}
336 	/*
337 	 * The following hack attempts to guess whether or not the current
338 	 * token is in fact a declaration keyword -- one that has been
339 	 * typedefd
340 	 */
341 	else if (!state->p_l_follow && !state->block_init &&
342 	    !state->in_stmt &&
343 	    ((*buf_ptr == '*' && buf_ptr[1] != '=') ||
344 		isalpha((unsigned char)*buf_ptr)) &&
345 	    (state->last_token == semicolon || state->last_token == lbrace ||
346 		state->last_token == rbrace)) {
347 	    state->keyword = 4;	/* a type name */
348 	    state->last_u_d = true;
349 	    return decl;
350 	}
351 	if (state->last_token == decl)	/* if this is a declared variable,
352 					 * then following sign is unary */
353 	    state->last_u_d = true;	/* will make "int a -1" work */
354 	return (ident);		/* the ident is not in the list */
355     }				/* end of processing for alpanum character */
356 
357     /* Scan a non-alphanumeric token */
358 
359     CHECK_SIZE_TOKEN(3);		/* things like "<<=" */
360     *e_token++ = *buf_ptr;		/* if it is only a one-character token, it is
361 				 * moved here */
362     *e_token = '\0';
363     if (++buf_ptr >= buf_end)
364 	fill_buffer();
365 
366     switch (*token) {
367     case '\n':
368 	unary_delim = state->last_u_d;
369 	state->last_nl = true;	/* remember that we just had a newline */
370 	code = (had_eof ? 0 : newline);
371 
372 	/*
373 	 * if data has been exhausted, the newline is a dummy, and we should
374 	 * return code to stop
375 	 */
376 	break;
377 
378     case '\'':			/* start of quoted character */
379     case '"':			/* start of string */
380 	qchar = *token;
381 	do {			/* copy the string */
382 	    while (1) {		/* move one character or [/<char>]<char> */
383 		if (*buf_ptr == '\n') {
384 		    diag2(1, "Unterminated literal");
385 		    goto stop_lit;
386 		}
387 		CHECK_SIZE_TOKEN(2);
388 		*e_token = *buf_ptr++;
389 		if (buf_ptr >= buf_end)
390 		    fill_buffer();
391 		if (*e_token == BACKSLASH) {	/* if escape, copy extra char */
392 		    if (*buf_ptr == '\n')	/* check for escaped newline */
393 			++line_no;
394 		    *++e_token = *buf_ptr++;
395 		    ++e_token;	/* we must increment this again because we
396 				 * copied two chars */
397 		    if (buf_ptr >= buf_end)
398 			fill_buffer();
399 		}
400 		else
401 		    break;	/* we copied one character */
402 	    }			/* end of while (1) */
403 	} while (*e_token++ != qchar);
404 stop_lit:
405 	code = ident;
406 	break;
407 
408     case ('('):
409     case ('['):
410 	unary_delim = true;
411 	code = lparen;
412 	break;
413 
414     case (')'):
415     case (']'):
416 	code = rparen;
417 	break;
418 
419     case '#':
420 	unary_delim = state->last_u_d;
421 	code = preesc;
422 	break;
423 
424     case '?':
425 	unary_delim = true;
426 	code = question;
427 	break;
428 
429     case (':'):
430 	code = colon;
431 	unary_delim = true;
432 	break;
433 
434     case (';'):
435 	unary_delim = true;
436 	code = semicolon;
437 	break;
438 
439     case ('{'):
440 	unary_delim = true;
441 
442 	/*
443 	 * if (state->in_or_st) state->block_init = 1;
444 	 */
445 	/* ?	code = state->block_init ? lparen : lbrace; */
446 	code = lbrace;
447 	break;
448 
449     case ('}'):
450 	unary_delim = true;
451 	/* ?	code = state->block_init ? rparen : rbrace; */
452 	code = rbrace;
453 	break;
454 
455     case 014:			/* a form feed */
456 	unary_delim = state->last_u_d;
457 	state->last_nl = true;	/* remember this so we can set 'state->col_1'
458 				 * right */
459 	code = form_feed;
460 	break;
461 
462     case (','):
463 	unary_delim = true;
464 	code = comma;
465 	break;
466 
467     case '.':
468 	unary_delim = false;
469 	code = period;
470 	break;
471 
472     case '-':
473     case '+':			/* check for -, +, --, ++ */
474 	code = (state->last_u_d ? unary_op : binary_op);
475 	unary_delim = true;
476 
477 	if (*buf_ptr == token[0]) {
478 	    /* check for doubled character */
479 	    *e_token++ = *buf_ptr++;
480 	    /* buffer overflow will be checked at end of loop */
481 	    if (state->last_token == ident || state->last_token == rparen) {
482 		code = (state->last_u_d ? unary_op : postop);
483 		/* check for following ++ or -- */
484 		unary_delim = false;
485 	    }
486 	}
487 	else if (*buf_ptr == '=')
488 	    /* check for operator += */
489 	    *e_token++ = *buf_ptr++;
490 	else if (*buf_ptr == '>') {
491 	    /* check for operator -> */
492 	    *e_token++ = *buf_ptr++;
493 	    if (!opt.pointer_as_binop) {
494 		unary_delim = false;
495 		code = unary_op;
496 		state->want_blank = false;
497 	    }
498 	}
499 	break;			/* buffer overflow will be checked at end of
500 				 * switch */
501 
502     case '=':
503 	if (state->in_or_st)
504 	    state->block_init = 1;
505 	if (*buf_ptr == '=') {/* == */
506 	    *e_token++ = '=';	/* Flip =+ to += */
507 	    buf_ptr++;
508 	    *e_token = 0;
509 	}
510 	code = binary_op;
511 	unary_delim = true;
512 	break;
513 	/* can drop thru!!! */
514 
515     case '>':
516     case '<':
517     case '!':			/* ops like <, <<, <=, !=, etc */
518 	if (*buf_ptr == '>' || *buf_ptr == '<' || *buf_ptr == '=') {
519 	    *e_token++ = *buf_ptr;
520 	    if (++buf_ptr >= buf_end)
521 		fill_buffer();
522 	}
523 	if (*buf_ptr == '=')
524 	    *e_token++ = *buf_ptr++;
525 	code = (state->last_u_d ? unary_op : binary_op);
526 	unary_delim = true;
527 	break;
528 
529     case '*':
530 	unary_delim = true;
531 	if (!state->last_u_d) {
532 	    if (*buf_ptr == '=')
533 		*e_token++ = *buf_ptr++;
534 	    code = binary_op;
535 	    break;
536 	}
537 	while (*buf_ptr == '*' || isspace((unsigned char)*buf_ptr)) {
538 	    if (*buf_ptr == '*') {
539 		CHECK_SIZE_TOKEN(1);
540 		*e_token++ = *buf_ptr;
541 	    }
542 	    if (++buf_ptr >= buf_end)
543 		fill_buffer();
544 	}
545 	if (ps.in_decl) {
546 	    char *tp = buf_ptr;
547 
548 	    while (isalpha((unsigned char)*tp) ||
549 		   isspace((unsigned char)*tp)) {
550 		if (++tp >= buf_end)
551 		    fill_buffer();
552 	    }
553 	    if (*tp == '(')
554 		ps.procname[0] = ' ';
555 	}
556 	code = unary_op;
557 	break;
558 
559     default:
560 	if (token[0] == '/' && *buf_ptr == '*') {
561 	    /* it is start of comment */
562 	    *e_token++ = '*';
563 
564 	    if (++buf_ptr >= buf_end)
565 		fill_buffer();
566 
567 	    code = comment;
568 	    unary_delim = state->last_u_d;
569 	    break;
570 	}
571 	while (*(e_token - 1) == *buf_ptr || *buf_ptr == '=') {
572 	    /*
573 	     * handle ||, &&, etc, and also things as in int *****i
574 	     */
575 	    CHECK_SIZE_TOKEN(1);
576 	    *e_token++ = *buf_ptr;
577 	    if (++buf_ptr >= buf_end)
578 		fill_buffer();
579 	}
580 	code = (state->last_u_d ? unary_op : binary_op);
581 	unary_delim = true;
582 
583 
584     }				/* end of switch */
585     if (buf_ptr >= buf_end)	/* check for input buffer empty */
586 	fill_buffer();
587     state->last_u_d = unary_delim;
588     CHECK_SIZE_TOKEN(1);
589     *e_token = '\0';		/* null terminate the token */
590     return (code);
591 }
592 
593 /* Initialize constant transition table */
594 void
595 init_constant_tt(void)
596 {
597     table['-'] = table['+'];
598     table['8'] = table['9'];
599     table['2'] = table['3'] = table['4'] = table['5'] = table['6'] = table['7'];
600     table['A'] = table['C'] = table['D'] = table['c'] = table['d'] = table['a'];
601     table['B'] = table['b'];
602     table['E'] = table['e'];
603     table['U'] = table['u'];
604     table['X'] = table['x'];
605     table['P'] = table['p'];
606     table['F'] = table['f'];
607 }
608 
609 void
610 alloc_typenames(void)
611 {
612 
613     typenames = (const char **)malloc(sizeof(typenames[0]) *
614         (typename_count = 16));
615     if (typenames == NULL)
616 	err(1, NULL);
617 }
618 
619 void
620 add_typename(const char *key)
621 {
622     int comparison;
623     const char *copy;
624 
625     if (typename_top + 1 >= typename_count) {
626 	typenames = realloc((void *)typenames,
627 	    sizeof(typenames[0]) * (typename_count *= 2));
628 	if (typenames == NULL)
629 	    err(1, NULL);
630     }
631     if (typename_top == -1)
632 	typenames[++typename_top] = copy = strdup(key);
633     else if ((comparison = strcmp(key, typenames[typename_top])) >= 0) {
634 	/* take advantage of sorted input */
635 	if (comparison == 0)	/* remove duplicates */
636 	    return;
637 	typenames[++typename_top] = copy = strdup(key);
638     }
639     else {
640 	int p;
641 
642 	for (p = 0; (comparison = strcmp(key, typenames[p])) > 0; p++)
643 	    /* find place for the new key */;
644 	if (comparison == 0)	/* remove duplicates */
645 	    return;
646 	memmove(&typenames[p + 1], &typenames[p],
647 	    sizeof(typenames[0]) * (++typename_top - p));
648 	typenames[p] = copy = strdup(key);
649     }
650 
651     if (copy == NULL)
652 	err(1, NULL);
653 }
654