xref: /freebsd/usr.bin/indent/io.c (revision eb69d1f144a6fcc765d1b9d44a5ae8082353e70b)
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[] = "@(#)io.c	8.1 (Berkeley) 6/6/93";
41 #endif /* not lint */
42 #endif
43 
44 #include <sys/cdefs.h>
45 __FBSDID("$FreeBSD$");
46 
47 #include <ctype.h>
48 #include <err.h>
49 #include <stdio.h>
50 #include <stdlib.h>
51 #include <string.h>
52 #include "indent_globs.h"
53 #include "indent.h"
54 
55 int         comment_open;
56 static int  paren_target;
57 static int pad_output(int current, int target);
58 
59 void
60 dump_line(void)
61 {				/* dump_line is the routine that actually
62 				 * effects the printing of the new source. It
63 				 * prints the label section, followed by the
64 				 * code section with the appropriate nesting
65 				 * level, followed by any comments */
66     int cur_col,
67                 target_col = 1;
68     static int  not_first_line;
69 
70     if (ps.procname[0]) {
71 	if (troff) {
72 	    if (comment_open) {
73 		comment_open = 0;
74 		fprintf(output, ".*/\n");
75 	    }
76 	    fprintf(output, ".Pr \"%s\"\n", ps.procname);
77 	}
78 	ps.ind_level = 0;
79 	ps.procname[0] = 0;
80     }
81     if (s_code == e_code && s_lab == e_lab && s_com == e_com) {
82 	if (suppress_blanklines > 0)
83 	    suppress_blanklines--;
84 	else {
85 	    ps.bl_line = true;
86 	    n_real_blanklines++;
87 	}
88     }
89     else if (!inhibit_formatting) {
90 	suppress_blanklines = 0;
91 	ps.bl_line = false;
92 	if (prefix_blankline_requested && not_first_line) {
93 	    if (swallow_optional_blanklines) {
94 		if (n_real_blanklines == 1)
95 		    n_real_blanklines = 0;
96 	    }
97 	    else {
98 		if (n_real_blanklines == 0)
99 		    n_real_blanklines = 1;
100 	    }
101 	}
102 	while (--n_real_blanklines >= 0)
103 	    putc('\n', output);
104 	n_real_blanklines = 0;
105 	if (ps.ind_level == 0)
106 	    ps.ind_stmt = 0;	/* this is a class A kludge. dont do
107 				 * additional statement indentation if we are
108 				 * at bracket level 0 */
109 
110 	if (e_lab != s_lab || e_code != s_code)
111 	    ++code_lines;	/* keep count of lines with code */
112 
113 
114 	if (e_lab != s_lab) {	/* print lab, if any */
115 	    if (comment_open) {
116 		comment_open = 0;
117 		fprintf(output, ".*/\n");
118 	    }
119 	    while (e_lab > s_lab && (e_lab[-1] == ' ' || e_lab[-1] == '\t'))
120 		e_lab--;
121 	    *e_lab = '\0';
122 	    cur_col = pad_output(1, compute_label_target());
123 	    if (s_lab[0] == '#' && (strncmp(s_lab, "#else", 5) == 0
124 				    || strncmp(s_lab, "#endif", 6) == 0)) {
125 		char *s = s_lab;
126 		if (e_lab[-1] == '\n') e_lab--;
127 		do putc(*s++, output);
128 		while (s < e_lab && 'a' <= *s && *s<='z');
129 		while ((*s == ' ' || *s == '\t') && s < e_lab)
130 		    s++;
131 		if (s < e_lab)
132 		    fprintf(output, s[0]=='/' && s[1]=='*' ? "\t%.*s" : "\t/* %.*s */",
133 			    (int)(e_lab - s), s);
134 	    }
135 	    else fprintf(output, "%.*s", (int)(e_lab - s_lab), s_lab);
136 	    cur_col = count_spaces(cur_col, s_lab);
137 	}
138 	else
139 	    cur_col = 1;	/* there is no label section */
140 
141 	ps.pcase = false;
142 
143 	if (s_code != e_code) {	/* print code section, if any */
144 	    char *p;
145 
146 	    if (comment_open) {
147 		comment_open = 0;
148 		fprintf(output, ".*/\n");
149 	    }
150 	    target_col = compute_code_target();
151 	    {
152 		int i;
153 
154 		for (i = 0; i < ps.p_l_follow; i++)
155 		    if (ps.paren_indents[i] >= 0)
156 			ps.paren_indents[i] = -(ps.paren_indents[i] + target_col);
157 	    }
158 	    cur_col = pad_output(cur_col, target_col);
159 	    for (p = s_code; p < e_code; p++)
160 		if (*p == (char) 0200)
161 		    fprintf(output, "%d", target_col * 7);
162 		else
163 		    putc(*p, output);
164 	    cur_col = count_spaces(cur_col, s_code);
165 	}
166 	if (s_com != e_com) {
167 	    if (troff) {
168 		int         all_here = 0;
169 		char *p;
170 
171 		if (e_com[-1] == '/' && e_com[-2] == '*')
172 		    e_com -= 2, all_here++;
173 		while (e_com > s_com && e_com[-1] == ' ')
174 		    e_com--;
175 		*e_com = 0;
176 		p = s_com;
177 		while (*p == ' ')
178 		    p++;
179 		if (p[0] == '/' && p[1] == '*')
180 		    p += 2, all_here++;
181 		else if (p[0] == '*')
182 		    p += p[1] == '/' ? 2 : 1;
183 		while (*p == ' ')
184 		    p++;
185 		if (*p == 0)
186 		    goto inhibit_newline;
187 		if (comment_open < 2 && ps.box_com) {
188 		    comment_open = 0;
189 		    fprintf(output, ".*/\n");
190 		}
191 		if (comment_open == 0) {
192 		    if ('a' <= *p && *p <= 'z')
193 			*p = *p + 'A' - 'a';
194 		    if (e_com - p < 50 && all_here == 2) {
195 			char *follow = p;
196 			fprintf(output, "\n.nr C! \\w\1");
197 			while (follow < e_com) {
198 			    switch (*follow) {
199 			    case '\n':
200 				putc(' ', output);
201 			    case 1:
202 				break;
203 			    case '\\':
204 				putc('\\', output);
205 				/* add a backslash to escape the '\' */
206 			    default:
207 				putc(*follow, output);
208 			    }
209 			    follow++;
210 			}
211 			putc(1, output);
212 		    }
213 		    fprintf(output, "\n./* %dp %d %dp\n",
214 			    ps.com_col * 7,
215 			    (s_code != e_code || s_lab != e_lab) - ps.box_com,
216 			    target_col * 7);
217 		}
218 		comment_open = 1 + ps.box_com;
219 		while (*p) {
220 		    if (*p == BACKSLASH)
221 			putc(BACKSLASH, output);
222 		    putc(*p++, output);
223 		}
224 	    }
225 	    else {		/* print comment, if any */
226 		int target = ps.com_col;
227 		char *com_st = s_com;
228 
229 		target += ps.comment_delta;
230 		while (*com_st == '\t')	/* consider original indentation in
231 					 * case this is a box comment */
232 		    com_st++, target += tabsize;
233 		while (target <= 0)
234 		    if (*com_st == ' ')
235 			target++, com_st++;
236 		    else if (*com_st == '\t')
237 			target = tabsize * (1 + (target - 1) / tabsize) + 1, com_st++;
238 		    else
239 			target = 1;
240 		if (cur_col > target) {	/* if comment can't fit on this line,
241 					 * put it on next line */
242 		    putc('\n', output);
243 		    cur_col = 1;
244 		    ++ps.out_lines;
245 		}
246 		while (e_com > com_st && isspace(e_com[-1]))
247 		    e_com--;
248 		(void)pad_output(cur_col, target);
249 		fwrite(com_st, e_com - com_st, 1, output);
250 		ps.comment_delta = ps.n_comment_delta;
251 		++ps.com_lines;	/* count lines with comments */
252 	    }
253 	}
254 	if (ps.use_ff)
255 	    putc('\014', output);
256 	else
257 	    putc('\n', output);
258 inhibit_newline:
259 	++ps.out_lines;
260 	if (ps.just_saw_decl == 1 && blanklines_after_declarations) {
261 	    prefix_blankline_requested = 1;
262 	    ps.just_saw_decl = 0;
263 	}
264 	else
265 	    prefix_blankline_requested = postfix_blankline_requested;
266 	postfix_blankline_requested = 0;
267     }
268     ps.decl_on_line = ps.in_decl;	/* if we are in the middle of a
269 					 * declaration, remember that fact for
270 					 * proper comment indentation */
271     ps.ind_stmt = ps.in_stmt & ~ps.in_decl;	/* next line should be
272 						 * indented if we have not
273 						 * completed this stmt and if
274 						 * we are not in the middle of
275 						 * a declaration */
276     ps.use_ff = false;
277     ps.dumped_decl_indent = 0;
278     *(e_lab = s_lab) = '\0';	/* reset buffers */
279     *(e_code = s_code) = '\0';
280     *(e_com = s_com = combuf + 1) = '\0';
281     ps.ind_level = ps.i_l_follow;
282     ps.paren_level = ps.p_l_follow;
283     if (ps.paren_level > 0)
284 	paren_target = -ps.paren_indents[ps.paren_level - 1];
285     not_first_line = 1;
286 }
287 
288 int
289 compute_code_target(void)
290 {
291     int target_col = ps.ind_size * ps.ind_level + 1;
292 
293     if (ps.paren_level)
294 	if (!lineup_to_parens)
295 	    target_col += continuation_indent
296 		* (2 * continuation_indent == ps.ind_size ? 1 : ps.paren_level);
297 	else {
298 	    int w;
299 	    int t = paren_target;
300 
301 	    if ((w = count_spaces(t, s_code) - max_col) > 0
302 		    && count_spaces(target_col, s_code) <= max_col) {
303 		t -= w + 1;
304 		if (t > target_col)
305 		    target_col = t;
306 	    }
307 	    else
308 		target_col = t;
309 	}
310     else if (ps.ind_stmt)
311 	target_col += continuation_indent;
312     return target_col;
313 }
314 
315 int
316 compute_label_target(void)
317 {
318     return
319 	ps.pcase ? (int) (case_ind * ps.ind_size) + 1
320 	: *s_lab == '#' ? 1
321 	: ps.ind_size * (ps.ind_level - label_offset) + 1;
322 }
323 
324 
325 /*
326  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
327  *
328  * All rights reserved
329  *
330  *
331  * NAME: fill_buffer
332  *
333  * FUNCTION: Reads one block of input into input_buffer
334  *
335  * HISTORY: initial coding 	November 1976	D A Willcox of CAC 1/7/77 A
336  * Willcox of CAC	Added check for switch back to partly full input
337  * buffer from temporary buffer
338  *
339  */
340 void
341 fill_buffer(void)
342 {				/* this routine reads stuff from the input */
343     char *p;
344     int i;
345     FILE *f = input;
346 
347     if (bp_save != NULL) {	/* there is a partly filled input buffer left */
348 	buf_ptr = bp_save;	/* do not read anything, just switch buffers */
349 	buf_end = be_save;
350 	bp_save = be_save = NULL;
351 	if (buf_ptr < buf_end)
352 	    return;		/* only return if there is really something in
353 				 * this buffer */
354     }
355     for (p = in_buffer;;) {
356 	if (p >= in_buffer_limit) {
357 	    int size = (in_buffer_limit - in_buffer) * 2 + 10;
358 	    int offset = p - in_buffer;
359 	    in_buffer = realloc(in_buffer, size);
360 	    if (in_buffer == NULL)
361 		errx(1, "input line too long");
362 	    p = in_buffer + offset;
363 	    in_buffer_limit = in_buffer + size - 2;
364 	}
365 	if ((i = getc(f)) == EOF) {
366 		*p++ = ' ';
367 		*p++ = '\n';
368 		had_eof = true;
369 		break;
370 	}
371 	*p++ = i;
372 	if (i == '\n')
373 		break;
374     }
375     buf_ptr = in_buffer;
376     buf_end = p;
377     if (p - in_buffer > 2 && p[-2] == '/' && p[-3] == '*') {
378 	if (in_buffer[3] == 'I' && strncmp(in_buffer, "/**INDENT**", 11) == 0)
379 	    fill_buffer();	/* flush indent error message */
380 	else {
381 	    int         com = 0;
382 
383 	    p = in_buffer;
384 	    while (*p == ' ' || *p == '\t')
385 		p++;
386 	    if (*p == '/' && p[1] == '*') {
387 		p += 2;
388 		while (*p == ' ' || *p == '\t')
389 		    p++;
390 		if (p[0] == 'I' && p[1] == 'N' && p[2] == 'D' && p[3] == 'E'
391 			&& p[4] == 'N' && p[5] == 'T') {
392 		    p += 6;
393 		    while (*p == ' ' || *p == '\t')
394 			p++;
395 		    if (*p == '*')
396 			com = 1;
397 		    else if (*p == 'O') {
398 			if (*++p == 'N')
399 			    p++, com = 1;
400 			else if (*p == 'F' && *++p == 'F')
401 			    p++, com = 2;
402 		    }
403 		    while (*p == ' ' || *p == '\t')
404 			p++;
405 		    if (p[0] == '*' && p[1] == '/' && p[2] == '\n' && com) {
406 			if (s_com != e_com || s_lab != e_lab || s_code != e_code)
407 			    dump_line();
408 			if (!(inhibit_formatting = com - 1)) {
409 			    n_real_blanklines = 0;
410 			    postfix_blankline_requested = 0;
411 			    prefix_blankline_requested = 0;
412 			    suppress_blanklines = 1;
413 			}
414 		    }
415 		}
416 	    }
417 	}
418     }
419     if (inhibit_formatting) {
420 	p = in_buffer;
421 	do
422 	    putc(*p, output);
423 	while (*p++ != '\n');
424     }
425 }
426 
427 /*
428  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
429  *
430  * All rights reserved
431  *
432  *
433  * NAME: pad_output
434  *
435  * FUNCTION: Writes tabs and spaces to move the current column up to the desired
436  * position.
437  *
438  * ALGORITHM: Put tabs and/or blanks into pobuf, then write pobuf.
439  *
440  * PARAMETERS: current		integer		The current column target
441  * nteger		The desired column
442  *
443  * RETURNS: Integer value of the new column.  (If current >= target, no action is
444  * taken, and current is returned.
445  *
446  * GLOBALS: None
447  *
448  * CALLS: write (sys)
449  *
450  * CALLED BY: dump_line
451  *
452  * HISTORY: initial coding 	November 1976	D A Willcox of CAC
453  *
454  */
455 static int
456 pad_output(int current, int target)
457 			        /* writes tabs and blanks (if necessary) to
458 				 * get the current output position up to the
459 				 * target column */
460     /* current: the current column value */
461     /* target: position we want it at */
462 {
463 
464     if (troff)
465 	fprintf(output, "\\h'|%dp'", (target - 1) * 7);
466     else {
467 	int curr;		/* internal column pointer */
468 
469 	if (current >= target)
470 	    return (current);	/* line is already long enough */
471 	curr = current;
472         if (use_tabs) {
473 	    int tcur;
474 
475 	    while ((tcur = tabsize * (1 + (curr - 1) / tabsize) + 1) <= target) {
476                 putc('\t', output);
477                 curr = tcur;
478             }
479         }
480         while (curr++ < target)
481 	    putc(' ', output);	/* pad with final blanks */
482     }
483     return (target);
484 }
485 
486 /*
487  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
488  *
489  * All rights reserved
490  *
491  *
492  * NAME: count_spaces
493  *
494  * FUNCTION: Find out where printing of a given string will leave the current
495  * character position on output.
496  *
497  * ALGORITHM: Run thru input string and add appropriate values to current
498  * position.
499  *
500  * RETURNS: Integer value of position after printing "buffer" starting in column
501  * "current".
502  *
503  * HISTORY: initial coding 	November 1976	D A Willcox of CAC
504  *
505  */
506 int
507 count_spaces_until(int cur, char *buffer, char *end)
508 /*
509  * this routine figures out where the character position will be after
510  * printing the text in buffer starting at column "current"
511  */
512 {
513     char *buf;		/* used to look thru buffer */
514 
515     for (buf = buffer; *buf != '\0' && buf != end; ++buf) {
516 	switch (*buf) {
517 
518 	case '\n':
519 	case 014:		/* form feed */
520 	    cur = 1;
521 	    break;
522 
523 	case '\t':
524 	    cur = tabsize * (1 + (cur - 1) / tabsize) + 1;
525 	    break;
526 
527 	case 010:		/* backspace */
528 	    --cur;
529 	    break;
530 
531 	default:
532 	    ++cur;
533 	    break;
534 	}			/* end of switch */
535     }				/* end of for loop */
536     return (cur);
537 }
538 
539 int
540 count_spaces(int cur, char *buffer)
541 {
542     return (count_spaces_until(cur, buffer, NULL));
543 }
544 
545 void
546 diag4(int level, const char *msg, int a, int b)
547 {
548     if (level)
549 	found_err = 1;
550     if (output == stdout) {
551 	fprintf(stdout, "/**INDENT** %s@%d: ", level == 0 ? "Warning" : "Error", line_no);
552 	fprintf(stdout, msg, a, b);
553 	fprintf(stdout, " */\n");
554     }
555     else {
556 	fprintf(stderr, "%s@%d: ", level == 0 ? "Warning" : "Error", line_no);
557 	fprintf(stderr, msg, a, b);
558 	fprintf(stderr, "\n");
559     }
560 }
561 
562 void
563 diag3(int level, const char *msg, int a)
564 {
565     if (level)
566 	found_err = 1;
567     if (output == stdout) {
568 	fprintf(stdout, "/**INDENT** %s@%d: ", level == 0 ? "Warning" : "Error", line_no);
569 	fprintf(stdout, msg, a);
570 	fprintf(stdout, " */\n");
571     }
572     else {
573 	fprintf(stderr, "%s@%d: ", level == 0 ? "Warning" : "Error", line_no);
574 	fprintf(stderr, msg, a);
575 	fprintf(stderr, "\n");
576     }
577 }
578 
579 void
580 diag2(int level, const char *msg)
581 {
582     if (level)
583 	found_err = 1;
584     if (output == stdout) {
585 	fprintf(stdout, "/**INDENT** %s@%d: ", level == 0 ? "Warning" : "Error", line_no);
586 	fprintf(stdout, "%s", msg);
587 	fprintf(stdout, " */\n");
588     }
589     else {
590 	fprintf(stderr, "%s@%d: ", level == 0 ? "Warning" : "Error", line_no);
591 	fprintf(stderr, "%s", msg);
592 	fprintf(stderr, "\n");
593     }
594 }
595 
596 void
597 writefdef(struct fstate *f, int nm)
598 {
599     fprintf(output, ".ds f%c %s\n.nr s%c %d\n",
600 	    nm, f->font, nm, f->size);
601 }
602 
603 char *
604 chfont(struct fstate *of, struct fstate *nf, char *s)
605 {
606     if (of->font[0] != nf->font[0]
607 	    || of->font[1] != nf->font[1]) {
608 	*s++ = '\\';
609 	*s++ = 'f';
610 	if (nf->font[1]) {
611 	    *s++ = '(';
612 	    *s++ = nf->font[0];
613 	    *s++ = nf->font[1];
614 	}
615 	else
616 	    *s++ = nf->font[0];
617     }
618     if (nf->size != of->size) {
619 	*s++ = '\\';
620 	*s++ = 's';
621 	if (nf->size < of->size) {
622 	    *s++ = '-';
623 	    *s++ = '0' + of->size - nf->size;
624 	}
625 	else {
626 	    *s++ = '+';
627 	    *s++ = '0' + nf->size - of->size;
628 	}
629     }
630     return s;
631 }
632 
633 void
634 parsefont(struct fstate *f, const char *s0)
635 {
636     const char *s = s0;
637     int         sizedelta = 0;
638 
639     memset(f, '\0', sizeof(*f));
640     while (*s) {
641 	if (isdigit(*s))
642 	    f->size = f->size * 10 + *s - '0';
643 	else if (isupper(*s))
644 	    if (f->font[0])
645 		f->font[1] = *s;
646 	    else
647 		f->font[0] = *s;
648 	else if (*s == 'c')
649 	    f->allcaps = 1;
650 	else if (*s == '+')
651 	    sizedelta++;
652 	else if (*s == '-')
653 	    sizedelta--;
654 	else {
655 	    errx(1, "bad font specification: %s", s0);
656 	}
657 	s++;
658     }
659     if (f->font[0] == 0)
660 	f->font[0] = 'R';
661     if (bodyf.size == 0)
662 	bodyf.size = 11;
663     if (f->size == 0)
664 	f->size = bodyf.size + sizedelta;
665     else if (sizedelta > 0)
666 	f->size += bodyf.size;
667     else
668 	f->size = bodyf.size - f->size;
669 }
670