xref: /netbsd/usr.bin/indent/io.c (revision bf9ec67e)
1 /*	$NetBSD: io.c,v 1.11 2002/05/26 22:53:38 wiz Exp $	*/
2 
3 /*
4  * Copyright (c) 1980, 1993
5  *	The Regents of the University of California.  All rights reserved.
6  * Copyright (c) 1976 Board of Trustees of the University of Illinois.
7  * Copyright (c) 1985 Sun Microsystems, Inc.
8  * All rights reserved.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. All advertising materials mentioning features or use of this software
19  *    must display the following acknowledgement:
20  *	This product includes software developed by the University of
21  *	California, Berkeley and its contributors.
22  * 4. Neither the name of the University nor the names of its contributors
23  *    may be used to endorse or promote products derived from this software
24  *    without specific prior written permission.
25  *
26  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
27  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
28  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
29  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
30  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
31  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
32  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
33  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
34  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
35  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
36  * SUCH DAMAGE.
37  */
38 
39 #include <sys/cdefs.h>
40 #ifndef lint
41 #if 0
42 static char sccsid[] = "@(#)io.c	8.1 (Berkeley) 6/6/93";
43 #else
44 __RCSID("$NetBSD: io.c,v 1.11 2002/05/26 22:53:38 wiz Exp $");
45 #endif
46 #endif				/* not lint */
47 
48 #include <ctype.h>
49 #include <err.h>
50 #include <stdarg.h>
51 #include <stdio.h>
52 #include <stdlib.h>
53 #include <string.h>
54 #include "indent_globs.h"
55 
56 int     comment_open;
57 static  int paren_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, target_col;
67 	static int not_first_line;
68 
69 	target_col = 0;
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 	} else
89 		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 				} else {
97 					if (n_real_blanklines == 0)
98 						n_real_blanklines = 1;
99 				}
100 			}
101 			while (--n_real_blanklines >= 0)
102 				putc('\n', output);
103 			n_real_blanklines = 0;
104 			if (ps.ind_level == 0)
105 				ps.ind_stmt = 0;	/* this is a class A
106 							 * kludge. dont do
107 							 * additional statement
108 							 * indentation if we are
109 							 * at bracket level 0 */
110 
111 			if (e_lab != s_lab || e_code != s_code)
112 				++code_lines;	/* keep count of lines with
113 						 * code */
114 
115 
116 			if (e_lab != s_lab) {	/* print lab, if any */
117 				if (comment_open) {
118 					comment_open = 0;
119 					fprintf(output, ".*/\n");
120 				}
121 				while (e_lab > s_lab && (e_lab[-1] == ' ' || e_lab[-1] == '\t'))
122 					e_lab--;
123 				cur_col = pad_output(1, compute_label_target());
124 				if (s_lab[0] == '#' && (strncmp(s_lab, "#else", 5) == 0
125 					|| strncmp(s_lab, "#endif", 6) == 0)) {
126 					char   *s = s_lab;
127 					if (e_lab[-1] == '\n')
128 						e_lab--;
129 					do
130 						putc(*s++, output);
131 					while (s < e_lab && 'a' <= *s && *s <= 'z');
132 					while ((*s == ' ' || *s == '\t') && s < e_lab)
133 						s++;
134 					if (s < e_lab)
135 						fprintf(output, s[0] == '/' && s[1] == '*' ? "\t%.*s" : "\t/* %.*s */",
136 						    (int)(e_lab - s), s);
137 				} else
138 					fprintf(output, "%.*s", (int)(e_lab - s_lab), s_lab);
139 				cur_col = count_spaces(cur_col, s_lab);
140 			} else
141 				cur_col = 1;	/* there is no label section */
142 
143 			ps.pcase = false;
144 
145 			if (s_code != e_code) {	/* print code section, if any */
146 				char   *p;
147 
148 				if (comment_open) {
149 					comment_open = 0;
150 					fprintf(output, ".*/\n");
151 				}
152 				target_col = compute_code_target();
153 				{
154 					int     i;
155 
156 					for (i = 0; i < ps.p_l_follow; i++)
157 						if (ps.paren_indents[i] >= 0)
158 							ps.paren_indents[i] = -(ps.paren_indents[i] + target_col);
159 				}
160 				cur_col = pad_output(cur_col, target_col);
161 				for (p = s_code; p < e_code; p++)
162 					if (*p == (char) 0200)
163 						fprintf(output, "%d", target_col * 7);
164 					else
165 						putc(*p, output);
166 				cur_col = count_spaces(cur_col, s_code);
167 			}
168 			if (s_com != e_com) {
169 				if (troff) {
170 					int     all_here = 0;
171 					char   *p;
172 
173 					if (e_com[-1] == '/' && e_com[-2] == '*')
174 						e_com -= 2, all_here++;
175 					while (e_com > s_com && e_com[-1] == ' ')
176 						e_com--;
177 					*e_com = 0;
178 					p = s_com;
179 					while (*p == ' ')
180 						p++;
181 					if (p[0] == '/' && p[1] == '*')
182 						p += 2, all_here++;
183 					else
184 						if (p[0] == '*')
185 							p += p[1] == '/' ? 2 : 1;
186 					while (*p == ' ')
187 						p++;
188 					if (*p == 0)
189 						goto inhibit_newline;
190 					if (comment_open < 2 && ps.box_com) {
191 						comment_open = 0;
192 						fprintf(output, ".*/\n");
193 					}
194 					if (comment_open == 0) {
195 						if ('a' <= *p && *p <= 'z')
196 							*p = *p + 'A' - 'a';
197 						if (e_com - p < 50 && all_here == 2) {
198 							char   *follow = p;
199 							fprintf(output, "\n.nr C! \\w\1");
200 							while (follow < e_com) {
201 								switch (*follow) {
202 								case '\n':
203 									putc(' ', output);
204 								case 1:
205 									break;
206 								case '\\':
207 									putc('\\', output);
208 								default:
209 									putc(*follow, output);
210 								}
211 								follow++;
212 							}
213 							putc(1, output);
214 						}
215 						fprintf(output, "\n./* %dp %d %dp\n",
216 						    ps.com_col * 7,
217 						    (s_code != e_code || s_lab != e_lab) - ps.box_com,
218 						    target_col * 7);
219 					}
220 					comment_open = 1 + ps.box_com;
221 					while (*p) {
222 						if (*p == BACKSLASH)
223 							putc(BACKSLASH, output);
224 						putc(*p++, output);
225 					}
226 				} else {	/* print comment, if any */
227 					int     target = ps.com_col;
228 					char   *com_st = s_com;
229 
230 					target += ps.comment_delta;
231 					while (*com_st == '\t')
232 						com_st++, target += 8;	/* ? */
233 					while (target <= 0)
234 						if (*com_st == ' ')
235 							target++, com_st++;
236 						else
237 							if (*com_st == '\t')
238 								target = ((target - 1) & ~7) + 9, com_st++;
239 							else
240 								target = 1;
241 					if (cur_col > target) {	/* if comment cant fit
242 								 * on this line, put it
243 								 * on next line */
244 						putc('\n', output);
245 						cur_col = 1;
246 						++ps.out_lines;
247 					}
248 					while (e_com > com_st
249 					&& isspace((unsigned char)e_com[-1]))
250 						e_com--;
251 					cur_col = pad_output(cur_col, target);
252 					if (!ps.box_com) {
253 						if (star_comment_cont
254 						&& (com_st[1] != '*'
255 						    || e_com <= com_st + 1)) {
256 							if (com_st[1] == ' '
257 							&&  com_st[0] == ' '
258 							&&  e_com > com_st + 1)
259 								com_st[1] = '*';
260 							else
261 								fwrite(" * ",
262 								com_st[0] == '\t'
263 								? 2
264 								: com_st[0]=='*'
265 								? 1
266 								: 3, 1, output);
267 						}
268 					}
269 					fwrite(com_st,
270 					    e_com - com_st, 1, output);
271 					ps.comment_delta = ps.n_comment_delta;
272 					cur_col = count_spaces(cur_col, com_st);
273 					++ps.com_lines;	/* count lines with
274 							 * comments */
275 				}
276 			}
277 			if (ps.use_ff)
278 				putc('\014', output);
279 			else
280 				putc('\n', output);
281 	inhibit_newline:
282 			++ps.out_lines;
283 			if (ps.just_saw_decl == 1 && blanklines_after_declarations) {
284 				prefix_blankline_requested = 1;
285 				ps.just_saw_decl = 0;
286 			} else
287 				prefix_blankline_requested = postfix_blankline_requested;
288 			postfix_blankline_requested = 0;
289 		}
290 	ps.decl_on_line = ps.in_decl;	/* if we are in the middle of a
291 					 * declaration, remember that fact for
292 					 * proper comment indentation */
293 	ps.ind_stmt = ps.in_stmt & ~ps.in_decl;	/* next line should be
294 						 * indented if we have not
295 						 * completed this stmt and if
296 						 * we are not in the middle of
297 						 * a declaration */
298 	ps.use_ff = false;
299 	ps.dumped_decl_indent = 0;
300 	*(e_lab = s_lab) = '\0';/* reset buffers */
301 	*(e_code = s_code) = '\0';
302 	*(e_com = s_com) = '\0';
303 	ps.ind_level = ps.i_l_follow;
304 	ps.paren_level = ps.p_l_follow;
305 	paren_target = -ps.paren_indents[ps.paren_level - 1];
306 	not_first_line = 1;
307 }
308 
309 int
310 compute_code_target(void)
311 {
312 	int     target_col = ps.ind_size * ps.ind_level + 1;
313 
314 	if (ps.paren_level) {
315 		if (!lineup_to_parens)
316 			target_col += continuation_indent * ps.paren_level;
317 		else {
318 			int     w;
319 			int     t = paren_target;
320 
321 			if ((w = count_spaces(t, s_code) - max_col) > 0
322 			    && count_spaces(target_col, s_code) <= max_col) {
323 				t -= w + 1;
324 				if (t > target_col)
325 					target_col = t;
326 			} else
327 				target_col = t;
328 		}
329 	} else
330 		if (ps.ind_stmt)
331 			target_col += continuation_indent;
332 	return target_col;
333 }
334 
335 int
336 compute_label_target(void)
337 {
338 	return
339 	ps.pcase ? (int) (case_ind * ps.ind_size) + 1
340 	: *s_lab == '#' ? 1
341 	: ps.ind_size * (ps.ind_level - label_offset) + 1;
342 }
343 
344 
345 /*
346  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
347  *
348  * All rights reserved
349  *
350  *
351  * NAME: fill_buffer
352  *
353  * FUNCTION: Reads one block of input into input_buffer
354  *
355  * HISTORY: initial coding 	November 1976	D A Willcox of CAC 1/7/77 A
356  * Willcox of CAC	Added check for switch back to partly full input
357  * buffer from temporary buffer
358  *
359  */
360 void
361 fill_buffer(void)
362 {				/* this routine reads stuff from the input */
363 	char   *p;
364 	int     i;
365 	FILE   *f = input;
366 
367 	if (bp_save != 0) {	/* there is a partly filled input buffer left */
368 		buf_ptr = bp_save;	/* dont read anything, just switch
369 					 * buffers */
370 		buf_end = be_save;
371 		bp_save = be_save = 0;
372 		if (buf_ptr < buf_end)
373 			return;	/* only return if there is really something in
374 				 * this buffer */
375 	}
376 	for (p = in_buffer;;) {
377 		if (p >= in_buffer_limit) {
378 			int     size = (in_buffer_limit - in_buffer) * 2 + 10;
379 			int     offset = p - in_buffer;
380 			in_buffer = (char *) realloc(in_buffer, size);
381 			if (in_buffer == 0)
382 				errx(1, "input line too long");
383 			p = in_buffer + offset;
384 			in_buffer_limit = in_buffer + size - 2;
385 		}
386 		if ((i = getc(f)) == EOF) {
387 			*p++ = ' ';
388 			*p++ = '\n';
389 			had_eof = true;
390 			break;
391 		}
392 		*p++ = i;
393 		if (i == '\n')
394 			break;
395 	}
396 	buf_ptr = in_buffer;
397 	buf_end = p;
398 	if (p[-2] == '/' && p[-3] == '*') {
399 		if (in_buffer[3] == 'I' && strncmp(in_buffer, "/**INDENT**", 11) == 0)
400 			fill_buffer();	/* flush indent error message */
401 		else {
402 			int     com = 0;
403 
404 			p = in_buffer;
405 			while (*p == ' ' || *p == '\t')
406 				p++;
407 			if (*p == '/' && p[1] == '*') {
408 				p += 2;
409 				while (*p == ' ' || *p == '\t')
410 					p++;
411 				if (p[0] == 'I' && p[1] == 'N' && p[2] == 'D' && p[3] == 'E'
412 				    && p[4] == 'N' && p[5] == 'T') {
413 					p += 6;
414 					while (*p == ' ' || *p == '\t')
415 						p++;
416 					if (*p == '*')
417 						com = 1;
418 					else {
419 						if (*p == 'O') {
420 							if (*++p == 'N')
421 								p++, com = 1;
422 							else
423 								if (*p == 'F' && *++p == 'F')
424 									p++, com = 2;
425 						}
426 					}
427 					while (*p == ' ' || *p == '\t')
428 						p++;
429 					if (p[0] == '*' && p[1] == '/' && p[2] == '\n' && com) {
430 						if (s_com != e_com || s_lab != e_lab || s_code != e_code)
431 							dump_line();
432 						if (!(inhibit_formatting = com - 1)) {
433 							n_real_blanklines = 0;
434 							postfix_blankline_requested = 0;
435 							prefix_blankline_requested = 0;
436 							suppress_blanklines = 1;
437 						}
438 					}
439 				}
440 			}
441 		}
442 	}
443 	if (inhibit_formatting) {
444 		p = in_buffer;
445 		do
446 			putc(*p, output);
447 		while (*p++ != '\n');
448 	}
449 }
450 /*
451  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
452  *
453  * All rights reserved
454  *
455  *
456  * NAME: pad_output
457  *
458  * FUNCTION: Writes tabs and spaces to move the current column up to the desired
459  * position.
460  *
461  * ALGORITHM: Put tabs and/or blanks into pobuf, then write pobuf.
462  *
463  * PARAMETERS: current		integer		The current column target
464  * 	       target		integer		The desired column
465  *
466  * RETURNS: Integer value of the new column.  (If current >= target, no action is
467  * taken, and current is returned.
468  *
469  * GLOBALS: None
470  *
471  * CALLS: write (sys)
472  *
473  * CALLED BY: dump_line
474  *
475  * HISTORY: initial coding 	November 1976	D A Willcox of CAC
476  *
477  */
478 int
479 pad_output(int current, int target)
480 {
481 	int     curr;		/* internal column pointer */
482 	int     tcur;
483 
484 	if (troff)
485 		fprintf(output, "\\h'|%dp'", (target - 1) * 7);
486 	else {
487 		if (current >= target)
488 			return (current);	/* line is already long enough */
489 		curr = current;
490 		while ((tcur = ((curr - 1) & tabmask) + tabsize + 1) <= target) {
491 			putc('\t', output);
492 			curr = tcur;
493 		}
494 		while (curr++ < target)
495 			putc(' ', output);	/* pad with final blanks */
496 	}
497 	return (target);
498 }
499 /*
500  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
501  *
502  * All rights reserved
503  *
504  *
505  * NAME: count_spaces
506  *
507  * FUNCTION: Find out where printing of a given string will leave the current
508  * character position on output.
509  *
510  * ALGORITHM: Run thru input string and add appropriate values to current
511  * position.
512  *
513  * RETURNS: Integer value of position after printing "buffer" starting in column
514  * "current".
515  *
516  * HISTORY: initial coding 	November 1976	D A Willcox of CAC
517  *
518  */
519 int
520 count_spaces(int current, char *buffer)
521 /*
522  * this routine figures out where the character position will be after
523  * printing the text in buffer starting at column "current"
524  */
525 {
526 	char   *buf;		/* used to look thru buffer */
527 	int     cur;		/* current character counter */
528 
529 	cur = current;
530 
531 	for (buf = buffer; *buf != '\0'; ++buf) {
532 		switch (*buf) {
533 
534 		case '\n':
535 		case 014:	/* form feed */
536 			cur = 1;
537 			break;
538 
539 		case '\t':
540 			cur = ((cur - 1) & tabmask) + tabsize + 1;
541 			break;
542 
543 		case 010:	/* backspace */
544 			--cur;
545 			break;
546 
547 		default:
548 			++cur;
549 			break;
550 		}		/* end of switch */
551 	}			/* end of for loop */
552 	return (cur);
553 }
554 
555 
556 int     found_err;
557 
558 void
559 diag(int level, char *msg,...)
560 {
561 	va_list ap;
562 
563 	va_start(ap, msg);
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 		vfprintf(stdout, msg, ap);
570 		fprintf(stdout, " */\n");
571 	} else {
572 		fprintf(stderr, "%s@%d: ", level == 0 ? "Warning" : "Error", line_no);
573 		vfprintf(stdout, msg, ap);
574 		fprintf(stderr, "\n");
575 	}
576 	va_end(ap);
577 }
578 
579 void
580 writefdef(struct fstate *f, int nm)
581 {
582 	fprintf(output, ".ds f%c %s\n.nr s%c %d\n",
583 	    nm, f->font, nm, f->size);
584 }
585 
586 char   *
587 chfont(struct fstate *of, struct fstate *nf, char *s)
588 {
589 	if (of->font[0] != nf->font[0]
590 	    || of->font[1] != nf->font[1]) {
591 		*s++ = '\\';
592 		*s++ = 'f';
593 		if (nf->font[1]) {
594 			*s++ = '(';
595 			*s++ = nf->font[0];
596 			*s++ = nf->font[1];
597 		} else
598 			*s++ = nf->font[0];
599 	}
600 	if (nf->size != of->size) {
601 		*s++ = '\\';
602 		*s++ = 's';
603 		if (nf->size < of->size) {
604 			*s++ = '-';
605 			*s++ = '0' + of->size - nf->size;
606 		} else {
607 			*s++ = '+';
608 			*s++ = '0' + nf->size - of->size;
609 		}
610 	}
611 	return s;
612 }
613 
614 
615 void
616 parsefont(struct fstate *f, char *s0)
617 {
618 	char   *s = s0;
619 	int     sizedelta = 0;
620 	memset(f, 0, sizeof *f);
621 	while (*s) {
622 		if (isdigit((unsigned char)*s))
623 			f->size = f->size * 10 + *s - '0';
624 		else
625 			if (isupper((unsigned char)*s)) {
626 				if (f->font[0])
627 					f->font[1] = *s;
628 				else
629 					f->font[0] = *s;
630 			} else
631 				if (*s == 'c')
632 					f->allcaps = 1;
633 				else
634 					if (*s == '+')
635 						sizedelta++;
636 					else
637 						if (*s == '-')
638 							sizedelta--;
639 						else {
640 							fprintf(stderr, "indent: bad font specification: %s\n", s0);
641 							exit(1);
642 						}
643 		s++;
644 	}
645 	if (f->font[0] == 0)
646 		f->font[0] = 'R';
647 	if (bodyf.size == 0)
648 		bodyf.size = 11;
649 	if (f->size == 0)
650 		f->size = bodyf.size + sizedelta;
651 	else
652 		if (sizedelta > 0)
653 			f->size += bodyf.size;
654 		else
655 			f->size = bodyf.size - f->size;
656 }
657