xref: /freebsd/usr.bin/indent/lexi.c (revision a2f733abcff64628b7771a47089628b7327a88bd)
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 #endif
40 #include <sys/cdefs.h>
41 /*
42  * Here we have the token scanner for indent.  It scans off one token and puts
43  * it in the global variable "token".  It returns a code, indicating the type
44  * of token scanned.
45  */
46 
47 #include <err.h>
48 #include <stdio.h>
49 #include <ctype.h>
50 #include <stdlib.h>
51 #include <string.h>
52 #include <sys/param.h>
53 
54 #include "indent_globs.h"
55 #include "indent_codes.h"
56 #include "indent.h"
57 
58 struct templ {
59     const char *rwd;
60     int         rwcode;
61 };
62 
63 /*
64  * This table has to be sorted alphabetically, because it'll be used in binary
65  * search. For the same reason, string must be the first thing in struct templ.
66  */
67 struct templ specials[] =
68 {
69     {"_Bool", 4},
70     {"_Complex", 4},
71     {"_Imaginary", 4},
72     {"auto", 10},
73     {"bool", 4},
74     {"break", 9},
75     {"case", 8},
76     {"char", 4},
77     {"complex", 4},
78     {"const", 4},
79     {"continue", 12},
80     {"default", 8},
81     {"do", 6},
82     {"double", 4},
83     {"else", 6},
84     {"enum", 3},
85     {"extern", 10},
86     {"float", 4},
87     {"for", 5},
88     {"global", 4},
89     {"goto", 9},
90     {"if", 5},
91     {"imaginary", 4},
92     {"inline", 12},
93     {"int", 4},
94     {"long", 4},
95     {"offsetof", 1},
96     {"register", 10},
97     {"restrict", 12},
98     {"return", 9},
99     {"short", 4},
100     {"signed", 4},
101     {"sizeof", 2},
102     {"static", 10},
103     {"struct", 3},
104     {"switch", 7},
105     {"typedef", 11},
106     {"union", 3},
107     {"unsigned", 4},
108     {"void", 4},
109     {"volatile", 4},
110     {"while", 5}
111 };
112 
113 const char **typenames;
114 int         typename_count;
115 int         typename_top = -1;
116 
117 /*
118  * The transition table below was rewritten by hand from lx's output, given
119  * the following definitions. lx is Katherine Flavel's lexer generator.
120  *
121  * O  = /[0-7]/;        D  = /[0-9]/;          NZ = /[1-9]/;
122  * H  = /[a-f0-9]/i;    B  = /[0-1]/;          HP = /0x/i;
123  * BP = /0b/i;          E  = /e[+\-]?/i D+;    P  = /p[+\-]?/i D+;
124  * FS = /[fl]/i;        IS = /u/i /(l|L|ll|LL)/? | /(l|L|ll|LL)/ /u/i?;
125  *
126  * D+           E  FS? -> $float;
127  * D*    "." D+ E? FS? -> $float;
128  * D+    "."    E? FS? -> $float;    HP H+           IS? -> $int;
129  * HP H+        P  FS? -> $float;    NZ D*           IS? -> $int;
130  * HP H* "." H+ P  FS? -> $float;    "0" O*          IS? -> $int;
131  * HP H+ "."    P  FS  -> $float;    BP B+           IS? -> $int;
132  */
133 static char const *table[] = {
134     /*                examples:
135                                      00
136              s                      0xx
137              t                    00xaa
138              a     11       101100xxa..
139              r   11ee0001101lbuuxx.a.pp
140              t.01.e+008bLuxll0Ll.aa.p+0
141     states:  ABCDEFGHIJKLMNOPQRSTUVWXYZ */
142     ['0'] = "CEIDEHHHIJQ  U  Q  VUVVZZZ",
143     ['1'] = "DEIDEHHHIJQ  U  Q  VUVVZZZ",
144     ['7'] = "DEIDEHHHIJ   U     VUVVZZZ",
145     ['9'] = "DEJDEHHHJJ   U     VUVVZZZ",
146     ['a'] = "             U     VUVV   ",
147     ['b'] = "  K          U     VUVV   ",
148     ['e'] = "  FFF   FF   U     VUVV   ",
149     ['f'] = "    f  f     U     VUVV  f",
150     ['u'] = "  MM    M  i  iiM   M     ",
151     ['x'] = "  N                       ",
152     ['p'] = "                    FFX   ",
153     ['L'] = "  LLf  fL  PR   Li  L    f",
154     ['l'] = "  OOf  fO   S P O i O    f",
155     ['+'] = "     G                 Y  ",
156     ['.'] = "B EE    EE   T      W     ",
157     /*       ABCDEFGHIJKLMNOPQRSTUVWXYZ */
158     [0]   = "uuiifuufiuuiiuiiiiiuiuuuuu",
159 };
160 
161 static int
162 strcmp_type(const void *e1, const void *e2)
163 {
164     return (strcmp(e1, *(const char * const *)e2));
165 }
166 
167 int
168 lexi(struct parser_state *state)
169 {
170     int         unary_delim;	/* this is set to 1 if the current token
171 				 * forces a following operator to be unary */
172     int         code;		/* internal code to be returned */
173     char        qchar;		/* the delimiter character for a string */
174 
175     e_token = s_token;		/* point to start of place to save token */
176     unary_delim = false;
177     state->col_1 = state->last_nl;	/* tell world that this token started
178 					 * in column 1 iff the last thing
179 					 * scanned was a newline */
180     state->last_nl = false;
181 
182     while (*buf_ptr == ' ' || *buf_ptr == '\t') {	/* get rid of blanks */
183 	state->col_1 = false;	/* leading blanks imply token is not in column
184 				 * 1 */
185 	if (++buf_ptr >= buf_end)
186 	    fill_buffer();
187     }
188 
189     /* Scan an alphanumeric token */
190     if (isalnum((unsigned char)*buf_ptr) ||
191 	*buf_ptr == '_' || *buf_ptr == '$' ||
192 	(buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) {
193 	/*
194 	 * we have a character or number
195 	 */
196 	struct templ *p;
197 
198 	if (isdigit((unsigned char)*buf_ptr) ||
199 	    (buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) {
200 	    char s;
201 	    unsigned char i;
202 
203 	    for (s = 'A'; s != 'f' && s != 'i' && s != 'u'; ) {
204 		i = (unsigned char)*buf_ptr;
205 		if (i >= nitems(table) || table[i] == NULL ||
206 		    table[i][s - 'A'] == ' ') {
207 		    s = table[0][s - 'A'];
208 		    break;
209 		}
210 		s = table[i][s - 'A'];
211 		CHECK_SIZE_TOKEN(1);
212 		*e_token++ = *buf_ptr++;
213 		if (buf_ptr >= buf_end)
214 		    fill_buffer();
215 	    }
216 	    /* s now indicates the type: f(loating), i(integer), u(nknown) */
217 	}
218 	else
219 	    while (isalnum((unsigned char)*buf_ptr) ||
220 	        *buf_ptr == BACKSLASH ||
221 		*buf_ptr == '_' || *buf_ptr == '$') {
222 		/* fill_buffer() terminates buffer with newline */
223 		if (*buf_ptr == BACKSLASH) {
224 		    if (*(buf_ptr + 1) == '\n') {
225 			buf_ptr += 2;
226 			if (buf_ptr >= buf_end)
227 			    fill_buffer();
228 			} else
229 			    break;
230 		}
231 		CHECK_SIZE_TOKEN(1);
232 		/* copy it over */
233 		*e_token++ = *buf_ptr++;
234 		if (buf_ptr >= buf_end)
235 		    fill_buffer();
236 	    }
237 	*e_token = '\0';
238 
239 	if (s_token[0] == 'L' && s_token[1] == '\0' &&
240 	      (*buf_ptr == '"' || *buf_ptr == '\''))
241 	    return (strpfx);
242 
243 	while (*buf_ptr == ' ' || *buf_ptr == '\t') {	/* get rid of blanks */
244 	    if (++buf_ptr >= buf_end)
245 		fill_buffer();
246 	}
247 	state->keyword = 0;
248 	if (state->last_token == structure && !state->p_l_follow) {
249 				/* if last token was 'struct' and we're not
250 				 * in parentheses, then this token
251 				 * should be treated as a declaration */
252 	    state->last_u_d = true;
253 	    return (decl);
254 	}
255 	/*
256 	 * Operator after identifier is binary unless last token was 'struct'
257 	 */
258 	state->last_u_d = (state->last_token == structure);
259 
260 	p = bsearch(s_token,
261 	    specials,
262 	    sizeof(specials) / sizeof(specials[0]),
263 	    sizeof(specials[0]),
264 	    strcmp_type);
265 	if (p == NULL) {	/* not a special keyword... */
266 	    char *u;
267 
268 	    /* ... so maybe a type_t or a typedef */
269 	    if ((opt.auto_typedefs && ((u = strrchr(s_token, '_')) != NULL) &&
270 	        strcmp(u, "_t") == 0) || (typename_top >= 0 &&
271 		  bsearch(s_token, typenames, typename_top + 1,
272 		    sizeof(typenames[0]), strcmp_type))) {
273 		state->keyword = 4;	/* a type name */
274 		state->last_u_d = true;
275 	        goto found_typename;
276 	    }
277 	} else {			/* we have a keyword */
278 	    state->keyword = p->rwcode;
279 	    state->last_u_d = true;
280 	    switch (p->rwcode) {
281 	    case 7:		/* it is a switch */
282 		return (swstmt);
283 	    case 8:		/* a case or default */
284 		return (casestmt);
285 
286 	    case 3:		/* a "struct" */
287 		/* FALLTHROUGH */
288 	    case 4:		/* one of the declaration keywords */
289 	    found_typename:
290 		if (state->p_l_follow) {
291 		    /* inside parens: cast, param list, offsetof or sizeof */
292 		    state->cast_mask |= (1 << state->p_l_follow) & ~state->not_cast_mask;
293 		}
294 		if (state->last_token == period || state->last_token == unary_op) {
295 		    state->keyword = 0;
296 		    break;
297 		}
298 		if (p != NULL && p->rwcode == 3)
299 		    return (structure);
300 		if (state->p_l_follow)
301 		    break;
302 		return (decl);
303 
304 	    case 5:		/* if, while, for */
305 		return (sp_paren);
306 
307 	    case 6:		/* do, else */
308 		return (sp_nparen);
309 
310 	    case 10:		/* storage class specifier */
311 		return (storage);
312 
313 	    case 11:		/* typedef */
314 		return (type_def);
315 
316 	    default:		/* all others are treated like any other
317 				 * identifier */
318 		return (ident);
319 	    }			/* end of switch */
320 	}			/* end of if (found_it) */
321 	if (*buf_ptr == '(' && state->tos <= 1 && state->ind_level == 0 &&
322 	    state->in_parameter_declaration == 0 && state->block_init == 0) {
323 	    char *tp = buf_ptr;
324 	    while (tp < buf_end)
325 		if (*tp++ == ')' && (*tp == ';' || *tp == ','))
326 		    goto not_proc;
327 	    strncpy(state->procname, token, sizeof state->procname - 1);
328 	    if (state->in_decl)
329 		state->in_parameter_declaration = 1;
330 	    return (funcname);
331     not_proc:;
332 	}
333 	/*
334 	 * The following hack attempts to guess whether or not the current
335 	 * token is in fact a declaration keyword -- one that has been
336 	 * typedefd
337 	 */
338 	else if (!state->p_l_follow && !state->block_init &&
339 	    !state->in_stmt &&
340 	    ((*buf_ptr == '*' && buf_ptr[1] != '=') ||
341 		isalpha((unsigned char)*buf_ptr)) &&
342 	    (state->last_token == semicolon || state->last_token == lbrace ||
343 		state->last_token == rbrace)) {
344 	    state->keyword = 4;	/* a type name */
345 	    state->last_u_d = true;
346 	    return decl;
347 	}
348 	if (state->last_token == decl)	/* if this is a declared variable,
349 					 * then following sign is unary */
350 	    state->last_u_d = true;	/* will make "int a -1" work */
351 	return (ident);		/* the ident is not in the list */
352     }				/* end of processing for alpanum character */
353 
354     /* Scan a non-alphanumeric token */
355 
356     CHECK_SIZE_TOKEN(3);		/* things like "<<=" */
357     *e_token++ = *buf_ptr;		/* if it is only a one-character token, it is
358 				 * moved here */
359     *e_token = '\0';
360     if (++buf_ptr >= buf_end)
361 	fill_buffer();
362 
363     switch (*token) {
364     case '\n':
365 	unary_delim = state->last_u_d;
366 	state->last_nl = true;	/* remember that we just had a newline */
367 	code = (had_eof ? 0 : newline);
368 
369 	/*
370 	 * if data has been exhausted, the newline is a dummy, and we should
371 	 * return code to stop
372 	 */
373 	break;
374 
375     case '\'':			/* start of quoted character */
376     case '"':			/* start of string */
377 	qchar = *token;
378 	do {			/* copy the string */
379 	    while (1) {		/* move one character or [/<char>]<char> */
380 		if (*buf_ptr == '\n') {
381 		    diag2(1, "Unterminated literal");
382 		    goto stop_lit;
383 		}
384 		CHECK_SIZE_TOKEN(2);
385 		*e_token = *buf_ptr++;
386 		if (buf_ptr >= buf_end)
387 		    fill_buffer();
388 		if (*e_token == BACKSLASH) {	/* if escape, copy extra char */
389 		    if (*buf_ptr == '\n')	/* check for escaped newline */
390 			++line_no;
391 		    *++e_token = *buf_ptr++;
392 		    ++e_token;	/* we must increment this again because we
393 				 * copied two chars */
394 		    if (buf_ptr >= buf_end)
395 			fill_buffer();
396 		}
397 		else
398 		    break;	/* we copied one character */
399 	    }			/* end of while (1) */
400 	} while (*e_token++ != qchar);
401 stop_lit:
402 	code = ident;
403 	break;
404 
405     case ('('):
406     case ('['):
407 	unary_delim = true;
408 	code = lparen;
409 	break;
410 
411     case (')'):
412     case (']'):
413 	code = rparen;
414 	break;
415 
416     case '#':
417 	unary_delim = state->last_u_d;
418 	code = preesc;
419 	break;
420 
421     case '?':
422 	unary_delim = true;
423 	code = question;
424 	break;
425 
426     case (':'):
427 	code = colon;
428 	unary_delim = true;
429 	break;
430 
431     case (';'):
432 	unary_delim = true;
433 	code = semicolon;
434 	break;
435 
436     case ('{'):
437 	unary_delim = true;
438 
439 	/*
440 	 * if (state->in_or_st) state->block_init = 1;
441 	 */
442 	/* ?	code = state->block_init ? lparen : lbrace; */
443 	code = lbrace;
444 	break;
445 
446     case ('}'):
447 	unary_delim = true;
448 	/* ?	code = state->block_init ? rparen : rbrace; */
449 	code = rbrace;
450 	break;
451 
452     case 014:			/* a form feed */
453 	unary_delim = state->last_u_d;
454 	state->last_nl = true;	/* remember this so we can set 'state->col_1'
455 				 * right */
456 	code = form_feed;
457 	break;
458 
459     case (','):
460 	unary_delim = true;
461 	code = comma;
462 	break;
463 
464     case '.':
465 	unary_delim = false;
466 	code = period;
467 	break;
468 
469     case '-':
470     case '+':			/* check for -, +, --, ++ */
471 	code = (state->last_u_d ? unary_op : binary_op);
472 	unary_delim = true;
473 
474 	if (*buf_ptr == token[0]) {
475 	    /* check for doubled character */
476 	    *e_token++ = *buf_ptr++;
477 	    /* buffer overflow will be checked at end of loop */
478 	    if (state->last_token == ident || state->last_token == rparen) {
479 		code = (state->last_u_d ? unary_op : postop);
480 		/* check for following ++ or -- */
481 		unary_delim = false;
482 	    }
483 	}
484 	else if (*buf_ptr == '=')
485 	    /* check for operator += */
486 	    *e_token++ = *buf_ptr++;
487 	else if (*buf_ptr == '>') {
488 	    /* check for operator -> */
489 	    *e_token++ = *buf_ptr++;
490 	    if (!opt.pointer_as_binop) {
491 		unary_delim = false;
492 		code = unary_op;
493 		state->want_blank = false;
494 	    }
495 	}
496 	break;			/* buffer overflow will be checked at end of
497 				 * switch */
498 
499     case '=':
500 	if (state->in_or_st)
501 	    state->block_init = 1;
502 	if (*buf_ptr == '=') {/* == */
503 	    *e_token++ = '=';	/* Flip =+ to += */
504 	    buf_ptr++;
505 	    *e_token = 0;
506 	}
507 	code = binary_op;
508 	unary_delim = true;
509 	break;
510 	/* can drop thru!!! */
511 
512     case '>':
513     case '<':
514     case '!':			/* ops like <, <<, <=, !=, etc */
515 	if (*buf_ptr == '>' || *buf_ptr == '<' || *buf_ptr == '=') {
516 	    *e_token++ = *buf_ptr;
517 	    if (++buf_ptr >= buf_end)
518 		fill_buffer();
519 	}
520 	if (*buf_ptr == '=')
521 	    *e_token++ = *buf_ptr++;
522 	code = (state->last_u_d ? unary_op : binary_op);
523 	unary_delim = true;
524 	break;
525 
526     case '*':
527 	unary_delim = true;
528 	if (!state->last_u_d) {
529 	    if (*buf_ptr == '=')
530 		*e_token++ = *buf_ptr++;
531 	    code = binary_op;
532 	    break;
533 	}
534 	while (*buf_ptr == '*' || isspace((unsigned char)*buf_ptr)) {
535 	    if (*buf_ptr == '*') {
536 		CHECK_SIZE_TOKEN(1);
537 		*e_token++ = *buf_ptr;
538 	    }
539 	    if (++buf_ptr >= buf_end)
540 		fill_buffer();
541 	}
542 	if (ps.in_decl) {
543 	    char *tp = buf_ptr;
544 
545 	    while (isalpha((unsigned char)*tp) ||
546 		   isspace((unsigned char)*tp)) {
547 		if (++tp >= buf_end)
548 		    fill_buffer();
549 	    }
550 	    if (*tp == '(')
551 		ps.procname[0] = ' ';
552 	}
553 	code = unary_op;
554 	break;
555 
556     default:
557 	if (token[0] == '/' && *buf_ptr == '*') {
558 	    /* it is start of comment */
559 	    *e_token++ = '*';
560 
561 	    if (++buf_ptr >= buf_end)
562 		fill_buffer();
563 
564 	    code = comment;
565 	    unary_delim = state->last_u_d;
566 	    break;
567 	}
568 	while (*(e_token - 1) == *buf_ptr || *buf_ptr == '=') {
569 	    /*
570 	     * handle ||, &&, etc, and also things as in int *****i
571 	     */
572 	    CHECK_SIZE_TOKEN(1);
573 	    *e_token++ = *buf_ptr;
574 	    if (++buf_ptr >= buf_end)
575 		fill_buffer();
576 	}
577 	code = (state->last_u_d ? unary_op : binary_op);
578 	unary_delim = true;
579 
580 
581     }				/* end of switch */
582     if (buf_ptr >= buf_end)	/* check for input buffer empty */
583 	fill_buffer();
584     state->last_u_d = unary_delim;
585     CHECK_SIZE_TOKEN(1);
586     *e_token = '\0';		/* null terminate the token */
587     return (code);
588 }
589 
590 /* Initialize constant transition table */
591 void
592 init_constant_tt(void)
593 {
594     table['-'] = table['+'];
595     table['8'] = table['9'];
596     table['2'] = table['3'] = table['4'] = table['5'] = table['6'] = table['7'];
597     table['A'] = table['C'] = table['D'] = table['c'] = table['d'] = table['a'];
598     table['B'] = table['b'];
599     table['E'] = table['e'];
600     table['U'] = table['u'];
601     table['X'] = table['x'];
602     table['P'] = table['p'];
603     table['F'] = table['f'];
604 }
605 
606 void
607 alloc_typenames(void)
608 {
609 
610     typenames = (const char **)malloc(sizeof(typenames[0]) *
611         (typename_count = 16));
612     if (typenames == NULL)
613 	err(1, NULL);
614 }
615 
616 void
617 add_typename(const char *key)
618 {
619     int comparison;
620     const char *copy;
621 
622     if (typename_top + 1 >= typename_count) {
623 	typenames = realloc((void *)typenames,
624 	    sizeof(typenames[0]) * (typename_count *= 2));
625 	if (typenames == NULL)
626 	    err(1, NULL);
627     }
628     if (typename_top == -1)
629 	typenames[++typename_top] = copy = strdup(key);
630     else if ((comparison = strcmp(key, typenames[typename_top])) >= 0) {
631 	/* take advantage of sorted input */
632 	if (comparison == 0)	/* remove duplicates */
633 	    return;
634 	typenames[++typename_top] = copy = strdup(key);
635     }
636     else {
637 	int p;
638 
639 	for (p = 0; (comparison = strcmp(key, typenames[p])) > 0; p++)
640 	    /* find place for the new key */;
641 	if (comparison == 0)	/* remove duplicates */
642 	    return;
643 	memmove(&typenames[p + 1], &typenames[p],
644 	    sizeof(typenames[0]) * (++typename_top - p));
645 	typenames[p] = copy = strdup(key);
646     }
647 
648     if (copy == NULL)
649 	err(1, NULL);
650 }
651