xref: /netbsd-src/usr.bin/indent/lexi.c (revision 8ecbf5f02b752fcb7debe1a8fab1dc82602bc760)
1 /*	$NetBSD: lexi.c,v 1.17 2019/10/19 15:44:31 christos Exp $	*/
2 
3 /*-
4  * SPDX-License-Identifier: BSD-4-Clause
5  *
6  * Copyright (c) 1985 Sun Microsystems, Inc.
7  * Copyright (c) 1980, 1993
8  *	The Regents of the University of California.  All rights reserved.
9  * All rights reserved.
10  *
11  * Redistribution and use in source and binary forms, with or without
12  * modification, are permitted provided that the following conditions
13  * are met:
14  * 1. Redistributions of source code must retain the above copyright
15  *    notice, this list of conditions and the following disclaimer.
16  * 2. Redistributions in binary form must reproduce the above copyright
17  *    notice, this list of conditions and the following disclaimer in the
18  *    documentation and/or other materials provided with the distribution.
19  * 3. All advertising materials mentioning features or use of this software
20  *    must display the following acknowledgement:
21  *	This product includes software developed by the University of
22  *	California, Berkeley and its contributors.
23  * 4. Neither the name of the University nor the names of its contributors
24  *    may be used to endorse or promote products derived from this software
25  *    without specific prior written permission.
26  *
27  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
28  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
29  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
30  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
31  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
32  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
33  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
34  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
35  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
36  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
37  * SUCH DAMAGE.
38  */
39 
40 #if 0
41 #ifndef lint
42 static char sccsid[] = "@(#)lexi.c	8.1 (Berkeley) 6/6/93";
43 #endif /* not lint */
44 #endif
45 
46 #include <sys/cdefs.h>
47 #ifndef lint
48 #if defined(__NetBSD__)
49 __RCSID("$NetBSD: lexi.c,v 1.17 2019/10/19 15:44:31 christos Exp $");
50 #elif defined(__FreeBSD__)
51 __FBSDID("$FreeBSD: head/usr.bin/indent/lexi.c 337862 2018-08-15 18:19:45Z pstef $");
52 #endif
53 #endif
54 
55 /*
56  * Here we have the token scanner for indent.  It scans off one token and puts
57  * it in the global variable "token".  It returns a code, indicating the type
58  * of token scanned.
59  */
60 
61 #include <err.h>
62 #include <stdio.h>
63 #include <ctype.h>
64 #include <stdlib.h>
65 #include <string.h>
66 #include <sys/param.h>
67 
68 #include "indent_globs.h"
69 #include "indent_codes.h"
70 #include "indent.h"
71 
72 struct templ {
73     const char *rwd;
74     int         rwcode;
75 };
76 
77 /*
78  * This table has to be sorted alphabetically, because it'll be used in binary
79  * search. For the same reason, string must be the first thing in struct templ.
80  */
81 struct templ specials[] =
82 {
83     {"_Bool", 4},
84     {"_Complex", 4},
85     {"_Imaginary", 4},
86     {"auto", 10},
87     {"bool", 4},
88     {"break", 9},
89     {"case", 8},
90     {"char", 4},
91     {"complex", 4},
92     {"const", 4},
93     {"continue", 12},
94     {"default", 8},
95     {"do", 6},
96     {"double", 4},
97     {"else", 6},
98     {"enum", 3},
99     {"extern", 10},
100     {"float", 4},
101     {"for", 5},
102     {"global", 4},
103     {"goto", 9},
104     {"if", 5},
105     {"imaginary", 4},
106     {"inline", 12},
107     {"int", 4},
108     {"long", 4},
109     {"offsetof", 1},
110     {"register", 10},
111     {"restrict", 12},
112     {"return", 9},
113     {"short", 4},
114     {"signed", 4},
115     {"sizeof", 2},
116     {"static", 10},
117     {"struct", 3},
118     {"switch", 7},
119     {"typedef", 11},
120     {"union", 3},
121     {"unsigned", 4},
122     {"void", 4},
123     {"volatile", 4},
124     {"while", 5}
125 };
126 
127 const char **typenames;
128 int         typename_count;
129 int         typename_top = -1;
130 
131 /*
132  * The transition table below was rewritten by hand from lx's output, given
133  * the following definitions. lx is Katherine Flavel's lexer generator.
134  *
135  * O  = /[0-7]/;        D  = /[0-9]/;          NZ = /[1-9]/;
136  * H  = /[a-f0-9]/i;    B  = /[0-1]/;          HP = /0x/i;
137  * BP = /0b/i;          E  = /e[+\-]?/i D+;    P  = /p[+\-]?/i D+;
138  * FS = /[fl]/i;        IS = /u/i /(l|L|ll|LL)/? | /(l|L|ll|LL)/ /u/i?;
139  *
140  * D+           E  FS? -> $float;
141  * D*    "." D+ E? FS? -> $float;
142  * D+    "."    E? FS? -> $float;    HP H+           IS? -> $int;
143  * HP H+        P  FS? -> $float;    NZ D*           IS? -> $int;
144  * HP H* "." H+ P  FS? -> $float;    "0" O*          IS? -> $int;
145  * HP H+ "."    P  FS  -> $float;    BP B+           IS? -> $int;
146  */
147 static char const *table[] = {
148     /*                examples:
149                                      00
150              s                      0xx
151              t                    00xaa
152              a     11       101100xxa..
153              r   11ee0001101lbuuxx.a.pp
154              t.01.e+008bLuxll0Ll.aa.p+0
155     states:  ABCDEFGHIJKLMNOPQRSTUVWXYZ */
156     ['0'] = "CEIDEHHHIJQ  U  Q  VUVVZZZ",
157     ['1'] = "DEIDEHHHIJQ  U  Q  VUVVZZZ",
158     ['7'] = "DEIDEHHHIJ   U     VUVVZZZ",
159     ['9'] = "DEJDEHHHJJ   U     VUVVZZZ",
160     ['a'] = "             U     VUVV   ",
161     ['b'] = "  K          U     VUVV   ",
162     ['e'] = "  FFF   FF   U     VUVV   ",
163     ['f'] = "    f  f     U     VUVV  f",
164     ['u'] = "  MM    M  i  iiM   M     ",
165     ['x'] = "  N                       ",
166     ['p'] = "                    FFX   ",
167     ['L'] = "  LLf  fL  PR   Li  L    f",
168     ['l'] = "  OOf  fO   S P O i O    f",
169     ['+'] = "     G                 Y  ",
170     ['.'] = "B EE    EE   T      W     ",
171     /*       ABCDEFGHIJKLMNOPQRSTUVWXYZ */
172     [0]   = "uuiifuufiuuiiuiiiiiuiuuuuu",
173 };
174 
175 static int
176 strcmp_type(const void *e1, const void *e2)
177 {
178     return (strcmp(e1, *(const char * const *)e2));
179 }
180 
181 int
182 lexi(struct parser_state *state)
183 {
184     int         unary_delim;	/* this is set to 1 if the current token
185 				 * forces a following operator to be unary */
186     int         code;		/* internal code to be returned */
187     char        qchar;		/* the delimiter character for a string */
188 
189     e_token = s_token;		/* point to start of place to save token */
190     unary_delim = false;
191     state->col_1 = state->last_nl;	/* tell world that this token started
192 					 * in column 1 iff the last thing
193 					 * scanned was a newline */
194     state->last_nl = false;
195 
196     while (*buf_ptr == ' ' || *buf_ptr == '\t') {	/* get rid of blanks */
197 	state->col_1 = false;	/* leading blanks imply token is not in column
198 				 * 1 */
199 	if (++buf_ptr >= buf_end)
200 	    fill_buffer();
201     }
202 
203     /* Scan an alphanumeric token */
204     if (isalnum((unsigned char)*buf_ptr) ||
205 	*buf_ptr == '_' || *buf_ptr == '$' ||
206 	(buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) {
207 	/*
208 	 * we have a character or number
209 	 */
210 	struct templ *p;
211 
212 	if (isdigit((unsigned char)*buf_ptr) ||
213 	    (buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) {
214 	    char s;
215 	    unsigned char i;
216 
217 	    for (s = 'A'; s != 'f' && s != 'i' && s != 'u'; ) {
218 		i = (unsigned char)*buf_ptr;
219 		if (i >= nitems(table) || table[i] == NULL ||
220 		    table[i][s - 'A'] == ' ') {
221 		    s = table[0][s - 'A'];
222 		    break;
223 		}
224 		s = table[i][s - 'A'];
225 		CHECK_SIZE_TOKEN(1);
226 		*e_token++ = *buf_ptr++;
227 		if (buf_ptr >= buf_end)
228 		    fill_buffer();
229 	    }
230 	    /* s now indicates the type: f(loating), i(integer), u(nknown) */
231 	}
232 	else
233 	    while (isalnum((unsigned char)*buf_ptr) ||
234 	        *buf_ptr == BACKSLASH ||
235 		*buf_ptr == '_' || *buf_ptr == '$') {
236 		/* fill_buffer() terminates buffer with newline */
237 		if (*buf_ptr == BACKSLASH) {
238 		    if (*(buf_ptr + 1) == '\n') {
239 			buf_ptr += 2;
240 			if (buf_ptr >= buf_end)
241 			    fill_buffer();
242 			} else
243 			    break;
244 		}
245 		CHECK_SIZE_TOKEN(1);
246 		/* copy it over */
247 		*e_token++ = *buf_ptr++;
248 		if (buf_ptr >= buf_end)
249 		    fill_buffer();
250 	    }
251 	*e_token = '\0';
252 
253 	if (s_token[0] == 'L' && s_token[1] == '\0' &&
254 	      (*buf_ptr == '"' || *buf_ptr == '\''))
255 	    return (strpfx);
256 
257 	while (*buf_ptr == ' ' || *buf_ptr == '\t') {	/* get rid of blanks */
258 	    if (++buf_ptr >= buf_end)
259 		fill_buffer();
260 	}
261 	state->keyword = 0;
262 	if (state->last_token == structure && !state->p_l_follow) {
263 				/* if last token was 'struct' and we're not
264 				 * in parentheses, then this token
265 				 * should be treated as a declaration */
266 	    state->last_u_d = true;
267 	    return (decl);
268 	}
269 	/*
270 	 * Operator after identifier is binary unless last token was 'struct'
271 	 */
272 	state->last_u_d = (state->last_token == structure);
273 
274 	p = bsearch(s_token,
275 	    specials,
276 	    sizeof(specials) / sizeof(specials[0]),
277 	    sizeof(specials[0]),
278 	    strcmp_type);
279 	if (p == NULL) {	/* not a special keyword... */
280 	    char *u;
281 
282 	    /* ... so maybe a type_t or a typedef */
283 	    if ((opt.auto_typedefs && ((u = strrchr(s_token, '_')) != NULL) &&
284 	        strcmp(u, "_t") == 0) || (typename_top >= 0 &&
285 		  bsearch(s_token, typenames, typename_top + 1,
286 		    sizeof(typenames[0]), strcmp_type))) {
287 		state->keyword = 4;	/* a type name */
288 		state->last_u_d = true;
289 	        goto found_typename;
290 	    }
291 	} else {			/* we have a keyword */
292 	    state->keyword = p->rwcode;
293 	    state->last_u_d = true;
294 	    switch (p->rwcode) {
295 	    case 7:		/* it is a switch */
296 		return (swstmt);
297 	    case 8:		/* a case or default */
298 		return (casestmt);
299 
300 	    case 3:		/* a "struct" */
301 		/* FALLTHROUGH */
302 	    case 4:		/* one of the declaration keywords */
303 	    found_typename:
304 		if (state->p_l_follow) {
305 		    /* inside parens: cast, param list, offsetof or sizeof */
306 		    state->cast_mask |= (1 << state->p_l_follow) & ~state->not_cast_mask;
307 		}
308 		if (state->last_token == period || state->last_token == unary_op) {
309 		    state->keyword = 0;
310 		    break;
311 		}
312 		if (p != NULL && p->rwcode == 3)
313 		    return (structure);
314 		if (state->p_l_follow)
315 		    break;
316 		return (decl);
317 
318 	    case 5:		/* if, while, for */
319 		return (sp_paren);
320 
321 	    case 6:		/* do, else */
322 		return (sp_nparen);
323 
324 	    case 10:		/* storage class specifier */
325 		return (storage);
326 
327 	    case 11:		/* typedef */
328 		return (type_def);
329 
330 	    default:		/* all others are treated like any other
331 				 * identifier */
332 		return (ident);
333 	    }			/* end of switch */
334 	}			/* end of if (found_it) */
335 	if (*buf_ptr == '(' && state->tos <= 1 && state->ind_level == 0 &&
336 	    state->in_parameter_declaration == 0 && state->block_init == 0) {
337 	    char *tp = buf_ptr;
338 	    while (tp < buf_end)
339 		if (*tp++ == ')' && (*tp == ';' || *tp == ','))
340 		    goto not_proc;
341 	    strncpy(state->procname, token, sizeof state->procname - 1);
342 	    if (state->in_decl)
343 		state->in_parameter_declaration = 1;
344 	    return (funcname);
345     not_proc:;
346 	}
347 	/*
348 	 * The following hack attempts to guess whether or not the current
349 	 * token is in fact a declaration keyword -- one that has been
350 	 * typedefd
351 	 */
352 	else if (!state->p_l_follow && !state->block_init &&
353 	    !state->in_stmt &&
354 	    ((*buf_ptr == '*' && buf_ptr[1] != '=') ||
355 		isalpha((unsigned char)*buf_ptr)) &&
356 	    (state->last_token == semicolon || state->last_token == lbrace ||
357 		state->last_token == rbrace)) {
358 	    state->keyword = 4;	/* a type name */
359 	    state->last_u_d = true;
360 	    return decl;
361 	}
362 	if (state->last_token == decl)	/* if this is a declared variable,
363 					 * then following sign is unary */
364 	    state->last_u_d = true;	/* will make "int a -1" work */
365 	return (ident);		/* the ident is not in the list */
366     }				/* end of procesing for alpanum character */
367 
368     /* Scan a non-alphanumeric token */
369 
370     CHECK_SIZE_TOKEN(3);		/* things like "<<=" */
371     *e_token++ = *buf_ptr;		/* if it is only a one-character token, it is
372 				 * moved here */
373     *e_token = '\0';
374     if (++buf_ptr >= buf_end)
375 	fill_buffer();
376 
377     switch (*token) {
378     case '\n':
379 	unary_delim = state->last_u_d;
380 	state->last_nl = true;	/* remember that we just had a newline */
381 	code = (had_eof ? 0 : newline);
382 
383 	/*
384 	 * if data has been exhausted, the newline is a dummy, and we should
385 	 * return code to stop
386 	 */
387 	break;
388 
389     case '\'':			/* start of quoted character */
390     case '"':			/* start of string */
391 	qchar = *token;
392 	do {			/* copy the string */
393 	    while (1) {		/* move one character or [/<char>]<char> */
394 		if (*buf_ptr == '\n') {
395 		    diag(1, "Unterminated literal");
396 		    goto stop_lit;
397 		}
398 		CHECK_SIZE_TOKEN(2);
399 		*e_token = *buf_ptr++;
400 		if (buf_ptr >= buf_end)
401 		    fill_buffer();
402 		if (*e_token == BACKSLASH) {	/* if escape, copy extra char */
403 		    if (*buf_ptr == '\n')	/* check for escaped newline */
404 			++line_no;
405 		    *++e_token = *buf_ptr++;
406 		    ++e_token;	/* we must increment this again because we
407 				 * copied two chars */
408 		    if (buf_ptr >= buf_end)
409 			fill_buffer();
410 		}
411 		else
412 		    break;	/* we copied one character */
413 	    }			/* end of while (1) */
414 	} while (*e_token++ != qchar);
415 stop_lit:
416 	code = ident;
417 	break;
418 
419     case ('('):
420     case ('['):
421 	unary_delim = true;
422 	code = lparen;
423 	break;
424 
425     case (')'):
426     case (']'):
427 	code = rparen;
428 	break;
429 
430     case '#':
431 	unary_delim = state->last_u_d;
432 	code = preesc;
433 	break;
434 
435     case '?':
436 	unary_delim = true;
437 	code = question;
438 	break;
439 
440     case (':'):
441 	code = colon;
442 	unary_delim = true;
443 	break;
444 
445     case (';'):
446 	unary_delim = true;
447 	code = semicolon;
448 	break;
449 
450     case ('{'):
451 	unary_delim = true;
452 
453 	/*
454 	 * if (state->in_or_st) state->block_init = 1;
455 	 */
456 	/* ?	code = state->block_init ? lparen : lbrace; */
457 	code = lbrace;
458 	break;
459 
460     case ('}'):
461 	unary_delim = true;
462 	/* ?	code = state->block_init ? rparen : rbrace; */
463 	code = rbrace;
464 	break;
465 
466     case 014:			/* a form feed */
467 	unary_delim = state->last_u_d;
468 	state->last_nl = true;	/* remember this so we can set 'state->col_1'
469 				 * right */
470 	code = form_feed;
471 	break;
472 
473     case (','):
474 	unary_delim = true;
475 	code = comma;
476 	break;
477 
478     case '.':
479 	unary_delim = false;
480 	code = period;
481 	break;
482 
483     case '-':
484     case '+':			/* check for -, +, --, ++ */
485 	code = (state->last_u_d ? unary_op : binary_op);
486 	unary_delim = true;
487 
488 	if (*buf_ptr == token[0]) {
489 	    /* check for doubled character */
490 	    *e_token++ = *buf_ptr++;
491 	    /* buffer overflow will be checked at end of loop */
492 	    if (state->last_token == ident || state->last_token == rparen) {
493 		code = (state->last_u_d ? unary_op : postop);
494 		/* check for following ++ or -- */
495 		unary_delim = false;
496 	    }
497 	}
498 	else if (*buf_ptr == '=')
499 	    /* check for operator += */
500 	    *e_token++ = *buf_ptr++;
501 	else if (*buf_ptr == '>') {
502 	    /* check for operator -> */
503 	    *e_token++ = *buf_ptr++;
504 	    unary_delim = false;
505 	    code = unary_op;
506 	    state->want_blank = false;
507 	}
508 	break;			/* buffer overflow will be checked at end of
509 				 * switch */
510 
511     case '=':
512 	if (state->in_or_st)
513 	    state->block_init = 1;
514 	if (*buf_ptr == '=') {/* == */
515 	    *e_token++ = '=';	/* Flip =+ to += */
516 	    buf_ptr++;
517 	    *e_token = 0;
518 	}
519 	code = binary_op;
520 	unary_delim = true;
521 	break;
522 	/* can drop thru!!! */
523 
524     case '>':
525     case '<':
526     case '!':			/* ops like <, <<, <=, !=, etc */
527 	if (*buf_ptr == '>' || *buf_ptr == '<' || *buf_ptr == '=') {
528 	    *e_token++ = *buf_ptr;
529 	    if (++buf_ptr >= buf_end)
530 		fill_buffer();
531 	}
532 	if (*buf_ptr == '=')
533 	    *e_token++ = *buf_ptr++;
534 	code = (state->last_u_d ? unary_op : binary_op);
535 	unary_delim = true;
536 	break;
537 
538     case '*':
539 	unary_delim = true;
540 	if (!state->last_u_d) {
541 	    if (*buf_ptr == '=')
542 		*e_token++ = *buf_ptr++;
543 	    code = binary_op;
544 	    break;
545 	}
546 	while (*buf_ptr == '*' || isspace((unsigned char)*buf_ptr)) {
547 	    if (*buf_ptr == '*') {
548 		CHECK_SIZE_TOKEN(1);
549 		*e_token++ = *buf_ptr;
550 	    }
551 	    if (++buf_ptr >= buf_end)
552 		fill_buffer();
553 	}
554 	if (ps.in_decl) {
555 	    char *tp = buf_ptr;
556 
557 	    while (isalpha((unsigned char)*tp) ||
558 		   isspace((unsigned char)*tp)) {
559 		if (++tp >= buf_end)
560 		    fill_buffer();
561 	    }
562 	    if (*tp == '(')
563 		ps.procname[0] = ' ';
564 	}
565 	code = unary_op;
566 	break;
567 
568     default:
569 	if (token[0] == '/' && *buf_ptr == '*') {
570 	    /* it is start of comment */
571 	    *e_token++ = '*';
572 
573 	    if (++buf_ptr >= buf_end)
574 		fill_buffer();
575 
576 	    code = comment;
577 	    unary_delim = state->last_u_d;
578 	    break;
579 	}
580 	while (*(e_token - 1) == *buf_ptr || *buf_ptr == '=') {
581 	    /*
582 	     * handle ||, &&, etc, and also things as in int *****i
583 	     */
584 	    CHECK_SIZE_TOKEN(1);
585 	    *e_token++ = *buf_ptr;
586 	    if (++buf_ptr >= buf_end)
587 		fill_buffer();
588 	}
589 	code = (state->last_u_d ? unary_op : binary_op);
590 	unary_delim = true;
591 
592 
593     }				/* end of switch */
594     if (buf_ptr >= buf_end)	/* check for input buffer empty */
595 	fill_buffer();
596     state->last_u_d = unary_delim;
597     CHECK_SIZE_TOKEN(1);
598     *e_token = '\0';		/* null terminate the token */
599     return (code);
600 }
601 
602 /* Initialize constant transition table */
603 void
604 init_constant_tt(void)
605 {
606     table['-'] = table['+'];
607     table['8'] = table['9'];
608     table['2'] = table['3'] = table['4'] = table['5'] = table['6'] = table['7'];
609     table['A'] = table['C'] = table['D'] = table['c'] = table['d'] = table['a'];
610     table['B'] = table['b'];
611     table['E'] = table['e'];
612     table['U'] = table['u'];
613     table['X'] = table['x'];
614     table['P'] = table['p'];
615     table['F'] = table['f'];
616 }
617 
618 void
619 alloc_typenames(void)
620 {
621 
622     typenames = (const char **)malloc(sizeof(typenames[0]) *
623         (typename_count = 16));
624     if (typenames == NULL)
625 	err(1, NULL);
626 }
627 
628 void
629 add_typename(const char *key)
630 {
631     int comparison;
632     const char *copy;
633 
634     if (typename_top + 1 >= typename_count) {
635 	typenames = realloc((void *)typenames,
636 	    sizeof(typenames[0]) * (typename_count *= 2));
637 	if (typenames == NULL)
638 	    err(1, NULL);
639     }
640     if (typename_top == -1)
641 	typenames[++typename_top] = copy = strdup(key);
642     else if ((comparison = strcmp(key, typenames[typename_top])) >= 0) {
643 	/* take advantage of sorted input */
644 	if (comparison == 0)	/* remove duplicates */
645 	    return;
646 	typenames[++typename_top] = copy = strdup(key);
647     }
648     else {
649 	int p;
650 
651 	for (p = 0; (comparison = strcmp(key, typenames[p])) > 0; p++)
652 	    /* find place for the new key */;
653 	if (comparison == 0)	/* remove duplicates */
654 	    return;
655 	memmove(&typenames[p + 1], &typenames[p],
656 	    sizeof(typenames[0]) * (++typename_top - p));
657 	typenames[p] = copy = strdup(key);
658     }
659 
660     if (copy == NULL)
661 	err(1, NULL);
662 }
663