xref: /freebsd/usr.bin/indent/lexi.c (revision c4f02a891fe62fe1277c89859922804ea2c27bcd)
1 /*
2  * Copyright (c) 1985 Sun Microsystems, Inc.
3  * Copyright (c) 1980, 1993
4  *	The Regents of the University of California.  All rights reserved.
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  * 3. All advertising materials mentioning features or use of this software
16  *    must display the following acknowledgement:
17  *	This product includes software developed by the University of
18  *	California, Berkeley and its contributors.
19  * 4. Neither the name of the University nor the names of its contributors
20  *    may be used to endorse or promote products derived from this software
21  *    without specific prior written permission.
22  *
23  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33  * SUCH DAMAGE.
34  */
35 
36 #if 0
37 #ifndef lint
38 static char sccsid[] = "@(#)lexi.c	8.1 (Berkeley) 6/6/93";
39 #endif /* not lint */
40 #endif
41 #include <sys/cdefs.h>
42 __FBSDID("$FreeBSD$");
43 
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 "indent_globs.h"
56 #include "indent_codes.h"
57 #include "indent.h"
58 
59 #define alphanum 1
60 #define opchar 3
61 
62 struct templ {
63     const char *rwd;
64     int         rwcode;
65 };
66 
67 struct templ specials[1000] =
68 {
69     {"switch", 1},
70     {"case", 2},
71     {"break", 0},
72     {"struct", 3},
73     {"union", 3},
74     {"enum", 3},
75     {"default", 2},
76     {"int", 4},
77     {"char", 4},
78     {"float", 4},
79     {"double", 4},
80     {"long", 4},
81     {"short", 4},
82     {"typdef", 4},
83     {"unsigned", 4},
84     {"register", 4},
85     {"static", 4},
86     {"global", 4},
87     {"extern", 4},
88     {"void", 4},
89     {"goto", 0},
90     {"return", 0},
91     {"if", 5},
92     {"while", 5},
93     {"for", 5},
94     {"else", 6},
95     {"do", 6},
96     {"sizeof", 7},
97     {"const", 9},
98     {"volatile", 9},
99     {0, 0}
100 };
101 
102 char        chartype[128] =
103 {				/* this is used to facilitate the decision of
104 				 * what type (alphanumeric, operator) each
105 				 * character is */
106     0, 0, 0, 0, 0, 0, 0, 0,
107     0, 0, 0, 0, 0, 0, 0, 0,
108     0, 0, 0, 0, 0, 0, 0, 0,
109     0, 0, 0, 0, 0, 0, 0, 0,
110     0, 3, 0, 0, 1, 3, 3, 0,
111     0, 0, 3, 3, 0, 3, 0, 3,
112     1, 1, 1, 1, 1, 1, 1, 1,
113     1, 1, 0, 0, 3, 3, 3, 3,
114     0, 1, 1, 1, 1, 1, 1, 1,
115     1, 1, 1, 1, 1, 1, 1, 1,
116     1, 1, 1, 1, 1, 1, 1, 1,
117     1, 1, 1, 0, 0, 0, 3, 1,
118     0, 1, 1, 1, 1, 1, 1, 1,
119     1, 1, 1, 1, 1, 1, 1, 1,
120     1, 1, 1, 1, 1, 1, 1, 1,
121     1, 1, 1, 0, 3, 0, 3, 0
122 };
123 
124 int
125 lexi(void)
126 {
127     int         unary_delim;	/* this is set to 1 if the current token
128 				 * forces a following operator to be unary */
129     static int  last_code;	/* the last token type returned */
130     static int  l_struct;	/* set to 1 if the last token was 'struct' */
131     int         code;		/* internal code to be returned */
132     char        qchar;		/* the delimiter character for a string */
133 
134     e_token = s_token;		/* point to start of place to save token */
135     unary_delim = false;
136     ps.col_1 = ps.last_nl;	/* tell world that this token started in
137 				 * column 1 iff the last thing scanned was nl */
138     ps.last_nl = false;
139 
140     while (*buf_ptr == ' ' || *buf_ptr == '\t') {	/* get rid of blanks */
141 	ps.col_1 = false;	/* leading blanks imply token is not in column
142 				 * 1 */
143 	if (++buf_ptr >= buf_end)
144 	    fill_buffer();
145     }
146 
147     /* Scan an alphanumeric token */
148     if (chartype[(int)*buf_ptr] == alphanum || (buf_ptr[0] == '.' && isdigit(buf_ptr[1]))) {
149 	/*
150 	 * we have a character or number
151 	 */
152 	const char *j;		/* used for searching thru list of
153 				 *
154 				 * reserved words */
155 	struct templ *p;
156 
157 	if (isdigit(*buf_ptr) || (buf_ptr[0] == '.' && isdigit(buf_ptr[1]))) {
158 	    int         seendot = 0,
159 	                seenexp = 0,
160 			seensfx = 0;
161 	    if (*buf_ptr == '0' &&
162 		    (buf_ptr[1] == 'x' || buf_ptr[1] == 'X')) {
163 		*e_token++ = *buf_ptr++;
164 		*e_token++ = *buf_ptr++;
165 		while (isxdigit(*buf_ptr)) {
166 		    CHECK_SIZE_TOKEN;
167 		    *e_token++ = *buf_ptr++;
168 		}
169 	    }
170 	    else
171 		while (1) {
172 		    if (*buf_ptr == '.') {
173 			if (seendot)
174 			    break;
175 			else
176 			    seendot++;
177 		    }
178 		    CHECK_SIZE_TOKEN;
179 		    *e_token++ = *buf_ptr++;
180 		    if (!isdigit(*buf_ptr) && *buf_ptr != '.') {
181 			if ((*buf_ptr != 'E' && *buf_ptr != 'e') || seenexp)
182 			    break;
183 			else {
184 			    seenexp++;
185 			    seendot++;
186 			    CHECK_SIZE_TOKEN;
187 			    *e_token++ = *buf_ptr++;
188 			    if (*buf_ptr == '+' || *buf_ptr == '-')
189 				*e_token++ = *buf_ptr++;
190 			}
191 		    }
192 		}
193 	    while (1) {
194 		if (!(seensfx & 1) &&
195 			(*buf_ptr == 'U' || *buf_ptr == 'u')) {
196 		    CHECK_SIZE_TOKEN;
197 		    *e_token++ = *buf_ptr++;
198 		    seensfx |= 1;
199 		    continue;
200 		}
201         	if (!(seensfx & 2) &&
202 			(*buf_ptr == 'L' || *buf_ptr == 'l')) {
203 		    CHECK_SIZE_TOKEN;
204 		    if (buf_ptr[1] == buf_ptr[0])
205 		        *e_token++ = *buf_ptr++;
206 		    *e_token++ = *buf_ptr++;
207 		    seensfx |= 2;
208 		    continue;
209 		}
210 		break;
211 	    }
212 	}
213 	else
214 	    while (chartype[(int)*buf_ptr] == alphanum || *buf_ptr == BACKSLASH) {
215 		/* fill_buffer() terminates buffer with newline */
216 		if (*buf_ptr == BACKSLASH) {
217 		    if (*(buf_ptr + 1) == '\n') {
218 			buf_ptr += 2;
219 			if (buf_ptr >= buf_end)
220 			    fill_buffer();
221 			} else
222 			    break;
223 		}
224 		CHECK_SIZE_TOKEN;
225 		/* copy it over */
226 		*e_token++ = *buf_ptr++;
227 		if (buf_ptr >= buf_end)
228 		    fill_buffer();
229 	    }
230 	*e_token++ = '\0';
231 	while (*buf_ptr == ' ' || *buf_ptr == '\t') {	/* get rid of blanks */
232 	    if (++buf_ptr >= buf_end)
233 		fill_buffer();
234 	}
235 	ps.its_a_keyword = false;
236 	ps.sizeof_keyword = false;
237 	if (l_struct) {		/* if last token was 'struct', then this token
238 				 * should be treated as a declaration */
239 	    l_struct = false;
240 	    last_code = ident;
241 	    ps.last_u_d = true;
242 	    return (decl);
243 	}
244 	ps.last_u_d = false;	/* Operator after identifier is binary */
245 	last_code = ident;	/* Remember that this is the code we will
246 				 * return */
247 
248 	/*
249 	 * This loop will check if the token is a keyword.
250 	 */
251 	for (p = specials; (j = p->rwd) != 0; p++) {
252 	    const char *q = s_token;	/* point at scanned token */
253 	    if (*j++ != *q++ || *j++ != *q++)
254 		continue;	/* This test depends on the fact that
255 				 * identifiers are always at least 1 character
256 				 * long (ie. the first two bytes of the
257 				 * identifier are always meaningful) */
258 	    if (q[-1] == 0)
259 		break;		/* If its a one-character identifier */
260 	    while (*q++ == *j)
261 		if (*j++ == 0)
262 		    goto found_keyword;	/* I wish that C had a multi-level
263 					 * break... */
264 	}
265 	if (p->rwd) {		/* we have a keyword */
266     found_keyword:
267 	    ps.its_a_keyword = true;
268 	    ps.last_u_d = true;
269 	    switch (p->rwcode) {
270 	    case 1:		/* it is a switch */
271 		return (swstmt);
272 	    case 2:		/* a case or default */
273 		return (casestmt);
274 
275 	    case 3:		/* a "struct" */
276 		if (ps.p_l_follow)
277 			break;	/* inside parens: cast */
278 		/*
279 		 * Next time around, we may want to know that we have had a
280 		 * 'struct'
281 		 */
282 		l_struct = true;
283 
284 		/*
285 		 * Fall through to test for a cast, function prototype or
286 		 * sizeof().
287 		 */
288 	    case 4:		/* one of the declaration keywords */
289 		if (ps.p_l_follow) {
290 		    ps.cast_mask |= 1 << ps.p_l_follow;
291 
292 		    /*
293 		     * Forget that we saw `struct' if we're in a sizeof().
294 		     */
295 		    if (ps.sizeof_mask)
296 			l_struct = false;
297 
298 		    break;	/* inside parens: cast, prototype or sizeof() */
299 		}
300 		last_code = decl;
301 		return (decl);
302 
303 	    case 5:		/* if, while, for */
304 		return (sp_paren);
305 
306 	    case 6:		/* do, else */
307 		return (sp_nparen);
308 
309 	    case 7:
310 		ps.sizeof_keyword = true;
311 	    default:		/* all others are treated like any other
312 				 * identifier */
313 		return (ident);
314 	    }			/* end of switch */
315 	}			/* end of if (found_it) */
316 	if (*buf_ptr == '(' && ps.tos <= 1 && ps.ind_level == 0) {
317 	    char *tp = buf_ptr;
318 	    while (tp < buf_end)
319 		if (*tp++ == ')' && (*tp == ';' || *tp == ','))
320 		    goto not_proc;
321 	    strncpy(ps.procname, token, sizeof ps.procname - 1);
322 	    ps.in_parameter_declaration = 1;
323 	    rparen_count = 1;
324     not_proc:;
325 	}
326 	/*
327 	 * The following hack attempts to guess whether or not the current
328 	 * token is in fact a declaration keyword -- one that has been
329 	 * typedefd
330 	 */
331 	if (((*buf_ptr == '*' && buf_ptr[1] != '=') || isalpha(*buf_ptr) || *buf_ptr == '_')
332 		&& !ps.p_l_follow
333 	        && !ps.block_init
334 		&& (ps.last_token == rparen || ps.last_token == semicolon ||
335 		    ps.last_token == decl ||
336 		    ps.last_token == lbrace || ps.last_token == rbrace)) {
337 	    ps.its_a_keyword = true;
338 	    ps.last_u_d = true;
339 	    last_code = decl;
340 	    return decl;
341 	}
342 	if (last_code == decl)	/* if this is a declared variable, then
343 				 * following sign is unary */
344 	    ps.last_u_d = true;	/* will make "int a -1" work */
345 	last_code = ident;
346 	return (ident);		/* the ident is not in the list */
347     }				/* end of procesing for alpanum character */
348 
349     /* Scan a non-alphanumeric token */
350 
351     *e_token++ = *buf_ptr;		/* if it is only a one-character token, it is
352 				 * moved here */
353     *e_token = '\0';
354     if (++buf_ptr >= buf_end)
355 	fill_buffer();
356 
357     switch (*token) {
358     case '\n':
359 	unary_delim = ps.last_u_d;
360 	ps.last_nl = true;	/* remember that we just had a newline */
361 	code = (had_eof ? 0 : newline);
362 
363 	/*
364 	 * if data has been exhausted, the newline is a dummy, and we should
365 	 * return code to stop
366 	 */
367 	break;
368 
369     case '\'':			/* start of quoted character */
370     case '"':			/* start of string */
371 	qchar = *token;
372 	if (troff) {
373 	    e_token[-1] = '`';
374 	    if (qchar == '"')
375 		*e_token++ = '`';
376 	    e_token = chfont(&bodyf, &stringf, e_token);
377 	}
378 	do {			/* copy the string */
379 	    while (1) {		/* move one character or [/<char>]<char> */
380 		if (*buf_ptr == '\n') {
381 		    printf("%d: Unterminated literal\n", line_no);
382 		    goto stop_lit;
383 		}
384 		CHECK_SIZE_TOKEN;	/* Only have to do this once in this loop,
385 					 * since CHECK_SIZE guarantees that there
386 					 * are at least 5 entries left */
387 		*e_token = *buf_ptr++;
388 		if (buf_ptr >= buf_end)
389 		    fill_buffer();
390 		if (*e_token == BACKSLASH) {	/* if escape, copy extra char */
391 		    if (*buf_ptr == '\n')	/* check for escaped newline */
392 			++line_no;
393 		    if (troff) {
394 			*++e_token = BACKSLASH;
395 			if (*buf_ptr == BACKSLASH)
396 			    *++e_token = BACKSLASH;
397 		    }
398 		    *++e_token = *buf_ptr++;
399 		    ++e_token;	/* we must increment this again because we
400 				 * copied two chars */
401 		    if (buf_ptr >= buf_end)
402 			fill_buffer();
403 		}
404 		else
405 		    break;	/* we copied one character */
406 	    }			/* end of while (1) */
407 	} while (*e_token++ != qchar);
408 	if (troff) {
409 	    e_token = chfont(&stringf, &bodyf, e_token - 1);
410 	    if (qchar == '"')
411 		*e_token++ = '\'';
412 	}
413 stop_lit:
414 	code = ident;
415 	break;
416 
417     case ('('):
418     case ('['):
419 	unary_delim = true;
420 	code = lparen;
421 	break;
422 
423     case (')'):
424     case (']'):
425 	code = rparen;
426 	break;
427 
428     case '#':
429 	unary_delim = ps.last_u_d;
430 	code = preesc;
431 	break;
432 
433     case '?':
434 	unary_delim = true;
435 	code = question;
436 	break;
437 
438     case (':'):
439 	code = colon;
440 	unary_delim = true;
441 	break;
442 
443     case (';'):
444 	unary_delim = true;
445 	code = semicolon;
446 	break;
447 
448     case ('{'):
449 	unary_delim = true;
450 
451 	/*
452 	 * if (ps.in_or_st) ps.block_init = 1;
453 	 */
454 	/* ?	code = ps.block_init ? lparen : lbrace; */
455 	code = lbrace;
456 	break;
457 
458     case ('}'):
459 	unary_delim = true;
460 	/* ?	code = ps.block_init ? rparen : rbrace; */
461 	code = rbrace;
462 	break;
463 
464     case 014:			/* a form feed */
465 	unary_delim = ps.last_u_d;
466 	ps.last_nl = true;	/* remember this so we can set 'ps.col_1'
467 				 * right */
468 	code = form_feed;
469 	break;
470 
471     case (','):
472 	unary_delim = true;
473 	code = comma;
474 	break;
475 
476     case '.':
477 	unary_delim = false;
478 	code = period;
479 	break;
480 
481     case '-':
482     case '+':			/* check for -, +, --, ++ */
483 	code = (ps.last_u_d ? unary_op : binary_op);
484 	unary_delim = true;
485 
486 	if (*buf_ptr == token[0]) {
487 	    /* check for doubled character */
488 	    *e_token++ = *buf_ptr++;
489 	    /* buffer overflow will be checked at end of loop */
490 	    if (last_code == ident || last_code == rparen) {
491 		code = (ps.last_u_d ? unary_op : postop);
492 		/* check for following ++ or -- */
493 		unary_delim = false;
494 	    }
495 	}
496 	else if (*buf_ptr == '=')
497 	    /* check for operator += */
498 	    *e_token++ = *buf_ptr++;
499 	else if (*buf_ptr == '>') {
500 	    /* check for operator -> */
501 	    *e_token++ = *buf_ptr++;
502 	    if (!pointer_as_binop) {
503 		unary_delim = false;
504 		code = unary_op;
505 		ps.want_blank = false;
506 	    }
507 	}
508 	break;			/* buffer overflow will be checked at end of
509 				 * switch */
510 
511     case '=':
512 	if (ps.in_or_st)
513 	    ps.block_init = 1;
514 #ifdef undef
515 	if (chartype[*buf_ptr] == opchar) {	/* we have two char assignment */
516 	    e_token[-1] = *buf_ptr++;
517 	    if ((e_token[-1] == '<' || e_token[-1] == '>') && e_token[-1] == *buf_ptr)
518 		*e_token++ = *buf_ptr++;
519 	    *e_token++ = '=';	/* Flip =+ to += */
520 	    *e_token = 0;
521 	}
522 #else
523 	if (*buf_ptr == '=') {/* == */
524 	    *e_token++ = '=';	/* Flip =+ to += */
525 	    buf_ptr++;
526 	    *e_token = 0;
527 	}
528 #endif
529 	code = binary_op;
530 	unary_delim = true;
531 	break;
532 	/* can drop thru!!! */
533 
534     case '>':
535     case '<':
536     case '!':			/* ops like <, <<, <=, !=, etc */
537 	if (*buf_ptr == '>' || *buf_ptr == '<' || *buf_ptr == '=') {
538 	    *e_token++ = *buf_ptr;
539 	    if (++buf_ptr >= buf_end)
540 		fill_buffer();
541 	}
542 	if (*buf_ptr == '=')
543 	    *e_token++ = *buf_ptr++;
544 	code = (ps.last_u_d ? unary_op : binary_op);
545 	unary_delim = true;
546 	break;
547 
548     default:
549 	if (token[0] == '/' && *buf_ptr == '*') {
550 	    /* it is start of comment */
551 	    *e_token++ = '*';
552 
553 	    if (++buf_ptr >= buf_end)
554 		fill_buffer();
555 
556 	    code = comment;
557 	    unary_delim = ps.last_u_d;
558 	    break;
559 	}
560 	while (*(e_token - 1) == *buf_ptr || *buf_ptr == '=') {
561 	    /*
562 	     * handle ||, &&, etc, and also things as in int *****i
563 	     */
564 	    *e_token++ = *buf_ptr;
565 	    if (++buf_ptr >= buf_end)
566 		fill_buffer();
567 	}
568 	code = (ps.last_u_d ? unary_op : binary_op);
569 	unary_delim = true;
570 
571 
572     }				/* end of switch */
573     if (code != newline) {
574 	l_struct = false;
575 	last_code = code;
576     }
577     if (buf_ptr >= buf_end)	/* check for input buffer empty */
578 	fill_buffer();
579     ps.last_u_d = unary_delim;
580     *e_token = '\0';		/* null terminate the token */
581     return (code);
582 }
583 
584 /*
585  * Add the given keyword to the keyword table, using val as the keyword type
586  */
587 void
588 addkey(char *key, int val)
589 {
590     struct templ *p = specials;
591     while (p->rwd)
592 	if (p->rwd[0] == key[0] && strcmp(p->rwd, key) == 0)
593 	    return;
594 	else
595 	    p++;
596     if (p >= specials + sizeof specials / sizeof specials[0])
597 	return;			/* For now, table overflows are silently
598 				 * ignored */
599     p->rwd = key;
600     p->rwcode = val;
601     p[1].rwd = 0;
602     p[1].rwcode = 0;
603 }
604