xref: /netbsd/usr.bin/indent/io.c (revision 6550d01e)
1 /*	$NetBSD: io.c,v 1.14 2009/04/12 11:09:49 lukem Exp $	*/
2 
3 /*
4  * Copyright (c) 1980, 1993
5  *	The Regents of the University of California.  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. Neither the name of the University nor the names of its contributors
16  *    may be used to endorse or promote products derived from this software
17  *    without specific prior written permission.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  */
31 
32 /*
33  * Copyright (c) 1976 Board of Trustees of the University of Illinois.
34  * Copyright (c) 1985 Sun Microsystems, Inc.
35  * All rights reserved.
36  *
37  * Redistribution and use in source and binary forms, with or without
38  * modification, are permitted provided that the following conditions
39  * are met:
40  * 1. Redistributions of source code must retain the above copyright
41  *    notice, this list of conditions and the following disclaimer.
42  * 2. Redistributions in binary form must reproduce the above copyright
43  *    notice, this list of conditions and the following disclaimer in the
44  *    documentation and/or other materials provided with the distribution.
45  * 3. All advertising materials mentioning features or use of this software
46  *    must display the following acknowledgement:
47  *	This product includes software developed by the University of
48  *	California, Berkeley and its contributors.
49  * 4. Neither the name of the University nor the names of its contributors
50  *    may be used to endorse or promote products derived from this software
51  *    without specific prior written permission.
52  *
53  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
54  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
55  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
56  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
57  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
58  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
59  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
60  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
61  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
62  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
63  * SUCH DAMAGE.
64  */
65 
66 #include <sys/cdefs.h>
67 #ifndef lint
68 #if 0
69 static char sccsid[] = "@(#)io.c	8.1 (Berkeley) 6/6/93";
70 #else
71 __RCSID("$NetBSD: io.c,v 1.14 2009/04/12 11:09:49 lukem Exp $");
72 #endif
73 #endif				/* not lint */
74 
75 #include <ctype.h>
76 #include <err.h>
77 #include <stdarg.h>
78 #include <stdio.h>
79 #include <stdlib.h>
80 #include <string.h>
81 #include "indent_globs.h"
82 
83 int     comment_open;
84 static  int paren_target;
85 
86 void
87 dump_line(void)
88 {				/* dump_line is the routine that actually
89 				 * effects the printing of the new source. It
90 				 * prints the label section, followed by the
91 				 * code section with the appropriate nesting
92 				 * level, followed by any comments */
93 	int     cur_col, target_col;
94 	static int not_first_line;
95 
96 	target_col = 0;
97 	if (ps.procname[0]) {
98 		if (troff) {
99 			if (comment_open) {
100 				comment_open = 0;
101 				fprintf(output, ".*/\n");
102 			}
103 			fprintf(output, ".Pr \"%s\"\n", ps.procname);
104 		}
105 		ps.ind_level = 0;
106 		ps.procname[0] = 0;
107 	}
108 	if (s_code == e_code && s_lab == e_lab && s_com == e_com) {
109 		if (suppress_blanklines > 0)
110 			suppress_blanklines--;
111 		else {
112 			ps.bl_line = true;
113 			n_real_blanklines++;
114 		}
115 	} else
116 		if (!inhibit_formatting) {
117 			suppress_blanklines = 0;
118 			ps.bl_line = false;
119 			if (prefix_blankline_requested && not_first_line) {
120 				if (swallow_optional_blanklines) {
121 					if (n_real_blanklines == 1)
122 						n_real_blanklines = 0;
123 				} else {
124 					if (n_real_blanklines == 0)
125 						n_real_blanklines = 1;
126 				}
127 			}
128 			while (--n_real_blanklines >= 0)
129 				putc('\n', output);
130 			n_real_blanklines = 0;
131 			if (ps.ind_level == 0)
132 				ps.ind_stmt = 0;	/* this is a class A
133 							 * kludge. dont do
134 							 * additional statement
135 							 * indentation if we are
136 							 * at bracket level 0 */
137 
138 			if (e_lab != s_lab || e_code != s_code)
139 				++code_lines;	/* keep count of lines with
140 						 * code */
141 
142 
143 			if (e_lab != s_lab) {	/* print lab, if any */
144 				if (comment_open) {
145 					comment_open = 0;
146 					fprintf(output, ".*/\n");
147 				}
148 				while (e_lab > s_lab && (e_lab[-1] == ' ' || e_lab[-1] == '\t'))
149 					e_lab--;
150 				cur_col = pad_output(1, compute_label_target());
151 				if (s_lab[0] == '#' && (strncmp(s_lab, "#else", 5) == 0
152 					|| strncmp(s_lab, "#endif", 6) == 0)) {
153 					char   *s = s_lab;
154 					if (e_lab[-1] == '\n')
155 						e_lab--;
156 					do
157 						putc(*s++, output);
158 					while (s < e_lab && 'a' <= *s && *s <= 'z');
159 					while ((*s == ' ' || *s == '\t') && s < e_lab)
160 						s++;
161 					if (s < e_lab)
162 						fprintf(output, s[0] == '/' && s[1] == '*' ? "\t%.*s" : "\t/* %.*s */",
163 						    (int)(e_lab - s), s);
164 				} else
165 					fprintf(output, "%.*s", (int)(e_lab - s_lab), s_lab);
166 				cur_col = count_spaces(cur_col, s_lab);
167 			} else
168 				cur_col = 1;	/* there is no label section */
169 
170 			ps.pcase = false;
171 
172 			if (s_code != e_code) {	/* print code section, if any */
173 				char   *p;
174 
175 				if (comment_open) {
176 					comment_open = 0;
177 					fprintf(output, ".*/\n");
178 				}
179 				target_col = compute_code_target();
180 				{
181 					int     i;
182 
183 					for (i = 0; i < ps.p_l_follow; i++)
184 						if (ps.paren_indents[i] >= 0)
185 							ps.paren_indents[i] = -(ps.paren_indents[i] + target_col);
186 				}
187 				cur_col = pad_output(cur_col, target_col);
188 				for (p = s_code; p < e_code; p++)
189 					if (*p == (char) 0200)
190 						fprintf(output, "%d", target_col * 7);
191 					else
192 						putc(*p, output);
193 				cur_col = count_spaces(cur_col, s_code);
194 			}
195 			if (s_com != e_com) {
196 				if (troff) {
197 					int     all_here = 0;
198 					char   *p;
199 
200 					if (e_com[-1] == '/' && e_com[-2] == '*')
201 						e_com -= 2, all_here++;
202 					while (e_com > s_com && e_com[-1] == ' ')
203 						e_com--;
204 					*e_com = 0;
205 					p = s_com;
206 					while (*p == ' ')
207 						p++;
208 					if (p[0] == '/' && p[1] == '*')
209 						p += 2, all_here++;
210 					else
211 						if (p[0] == '*')
212 							p += p[1] == '/' ? 2 : 1;
213 					while (*p == ' ')
214 						p++;
215 					if (*p == 0)
216 						goto inhibit_newline;
217 					if (comment_open < 2 && ps.box_com) {
218 						comment_open = 0;
219 						fprintf(output, ".*/\n");
220 					}
221 					if (comment_open == 0) {
222 						if ('a' <= *p && *p <= 'z')
223 							*p = *p + 'A' - 'a';
224 						if (e_com - p < 50 && all_here == 2) {
225 							char   *follow = p;
226 							fprintf(output, "\n.nr C! \\w\1");
227 							while (follow < e_com) {
228 								switch (*follow) {
229 								case '\n':
230 									putc(' ', output);
231 								case 1:
232 									break;
233 								case '\\':
234 									putc('\\', output);
235 								default:
236 									putc(*follow, output);
237 								}
238 								follow++;
239 							}
240 							putc(1, output);
241 						}
242 						fprintf(output, "\n./* %dp %d %dp\n",
243 						    ps.com_col * 7,
244 						    (s_code != e_code || s_lab != e_lab) - ps.box_com,
245 						    target_col * 7);
246 					}
247 					comment_open = 1 + ps.box_com;
248 					while (*p) {
249 						if (*p == BACKSLASH)
250 							putc(BACKSLASH, output);
251 						putc(*p++, output);
252 					}
253 				} else {	/* print comment, if any */
254 					int     target = ps.com_col;
255 					char   *com_st = s_com;
256 
257 					target += ps.comment_delta;
258 					while (*com_st == '\t')
259 						com_st++, target += 8;	/* ? */
260 					while (target <= 0)
261 						if (*com_st == ' ')
262 							target++, com_st++;
263 						else
264 							if (*com_st == '\t')
265 								target = ((target - 1) & ~7) + 9, com_st++;
266 							else
267 								target = 1;
268 					if (cur_col > target) {	/* if comment cant fit
269 								 * on this line, put it
270 								 * on next line */
271 						putc('\n', output);
272 						cur_col = 1;
273 						++ps.out_lines;
274 					}
275 					while (e_com > com_st
276 					&& isspace((unsigned char)e_com[-1]))
277 						e_com--;
278 					cur_col = pad_output(cur_col, target);
279 					if (!ps.box_com) {
280 						if (star_comment_cont
281 						&& (com_st[1] != '*'
282 						    || e_com <= com_st + 1)) {
283 							if (com_st[1] == ' '
284 							&&  com_st[0] == ' '
285 							&&  e_com > com_st + 1)
286 								com_st[1] = '*';
287 							else
288 								fwrite(" * ",
289 								com_st[0] == '\t'
290 								? 2
291 								: com_st[0]=='*'
292 								? 1
293 								: 3, 1, output);
294 						}
295 					}
296 					fwrite(com_st,
297 					    e_com - com_st, 1, output);
298 					ps.comment_delta = ps.n_comment_delta;
299 					cur_col = count_spaces(cur_col, com_st);
300 					++ps.com_lines;	/* count lines with
301 							 * comments */
302 				}
303 			}
304 			if (ps.use_ff)
305 				putc('\014', output);
306 			else
307 				putc('\n', output);
308 	inhibit_newline:
309 			++ps.out_lines;
310 			if (ps.just_saw_decl == 1 && blanklines_after_declarations) {
311 				prefix_blankline_requested = 1;
312 				ps.just_saw_decl = 0;
313 			} else
314 				prefix_blankline_requested = postfix_blankline_requested;
315 			postfix_blankline_requested = 0;
316 		}
317 	ps.decl_on_line = ps.in_decl;	/* if we are in the middle of a
318 					 * declaration, remember that fact for
319 					 * proper comment indentation */
320 	ps.ind_stmt = ps.in_stmt & ~ps.in_decl;	/* next line should be
321 						 * indented if we have not
322 						 * completed this stmt and if
323 						 * we are not in the middle of
324 						 * a declaration */
325 	ps.use_ff = false;
326 	ps.dumped_decl_indent = 0;
327 	*(e_lab = s_lab) = '\0';/* reset buffers */
328 	*(e_code = s_code) = '\0';
329 	*(e_com = s_com) = '\0';
330 	ps.ind_level = ps.i_l_follow;
331 	ps.paren_level = ps.p_l_follow;
332 	paren_target = -ps.paren_indents[ps.paren_level - 1];
333 	not_first_line = 1;
334 }
335 
336 int
337 compute_code_target(void)
338 {
339 	int     target_col = ps.ind_size * ps.ind_level + 1;
340 
341 	if (ps.paren_level) {
342 		if (!lineup_to_parens)
343 			target_col += continuation_indent * ps.paren_level;
344 		else {
345 			int     w;
346 			int     t = paren_target;
347 
348 			if ((w = count_spaces(t, s_code) - max_col) > 0
349 			    && count_spaces(target_col, s_code) <= max_col) {
350 				t -= w + 1;
351 				if (t > target_col)
352 					target_col = t;
353 			} else
354 				target_col = t;
355 		}
356 	} else
357 		if (ps.ind_stmt)
358 			target_col += continuation_indent;
359 	return target_col;
360 }
361 
362 int
363 compute_label_target(void)
364 {
365 	return
366 	ps.pcase ? (int) (case_ind * ps.ind_size) + 1
367 	: *s_lab == '#' ? 1
368 	: ps.ind_size * (ps.ind_level - label_offset) + 1;
369 }
370 
371 
372 /*
373  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
374  *
375  * All rights reserved
376  *
377  *
378  * NAME: fill_buffer
379  *
380  * FUNCTION: Reads one block of input into input_buffer
381  *
382  * HISTORY: initial coding 	November 1976	D A Willcox of CAC 1/7/77 A
383  * Willcox of CAC	Added check for switch back to partly full input
384  * buffer from temporary buffer
385  *
386  */
387 void
388 fill_buffer(void)
389 {				/* this routine reads stuff from the input */
390 	char   *p;
391 	int     i;
392 	FILE   *f = input;
393 	char   *n;
394 
395 	if (bp_save != 0) {	/* there is a partly filled input buffer left */
396 		buf_ptr = bp_save;	/* dont read anything, just switch
397 					 * buffers */
398 		buf_end = be_save;
399 		bp_save = be_save = 0;
400 		if (buf_ptr < buf_end)
401 			return;	/* only return if there is really something in
402 				 * this buffer */
403 	}
404 	for (p = in_buffer;;) {
405 		if (p >= in_buffer_limit) {
406 			int     size = (in_buffer_limit - in_buffer) * 2 + 10;
407 			int     offset = p - in_buffer;
408 			n = (char *) realloc(in_buffer, size);
409 			if (n == 0)
410 				errx(1, "input line too long");
411 			in_buffer = n;
412 			p = in_buffer + offset;
413 			in_buffer_limit = in_buffer + size - 2;
414 		}
415 		if ((i = getc(f)) == EOF) {
416 			*p++ = ' ';
417 			*p++ = '\n';
418 			had_eof = true;
419 			break;
420 		}
421 		*p++ = i;
422 		if (i == '\n')
423 			break;
424 	}
425 	buf_ptr = in_buffer;
426 	buf_end = p;
427 	if (p[-2] == '/' && p[-3] == '*') {
428 		if (in_buffer[3] == 'I' && strncmp(in_buffer, "/**INDENT**", 11) == 0)
429 			fill_buffer();	/* flush indent error message */
430 		else {
431 			int     com = 0;
432 
433 			p = in_buffer;
434 			while (*p == ' ' || *p == '\t')
435 				p++;
436 			if (*p == '/' && p[1] == '*') {
437 				p += 2;
438 				while (*p == ' ' || *p == '\t')
439 					p++;
440 				if (p[0] == 'I' && p[1] == 'N' && p[2] == 'D' && p[3] == 'E'
441 				    && p[4] == 'N' && p[5] == 'T') {
442 					p += 6;
443 					while (*p == ' ' || *p == '\t')
444 						p++;
445 					if (*p == '*')
446 						com = 1;
447 					else {
448 						if (*p == 'O') {
449 							if (*++p == 'N')
450 								p++, com = 1;
451 							else
452 								if (*p == 'F' && *++p == 'F')
453 									p++, com = 2;
454 						}
455 					}
456 					while (*p == ' ' || *p == '\t')
457 						p++;
458 					if (p[0] == '*' && p[1] == '/' && p[2] == '\n' && com) {
459 						if (s_com != e_com || s_lab != e_lab || s_code != e_code)
460 							dump_line();
461 						if (!(inhibit_formatting = com - 1)) {
462 							n_real_blanklines = 0;
463 							postfix_blankline_requested = 0;
464 							prefix_blankline_requested = 0;
465 							suppress_blanklines = 1;
466 						}
467 					}
468 				}
469 			}
470 		}
471 	}
472 	if (inhibit_formatting) {
473 		p = in_buffer;
474 		do
475 			putc(*p, output);
476 		while (*p++ != '\n');
477 	}
478 }
479 /*
480  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
481  *
482  * All rights reserved
483  *
484  *
485  * NAME: pad_output
486  *
487  * FUNCTION: Writes tabs and spaces to move the current column up to the desired
488  * position.
489  *
490  * ALGORITHM: Put tabs and/or blanks into pobuf, then write pobuf.
491  *
492  * PARAMETERS: current		integer		The current column target
493  * 	       target		integer		The desired column
494  *
495  * RETURNS: Integer value of the new column.  (If current >= target, no action is
496  * taken, and current is returned.
497  *
498  * GLOBALS: None
499  *
500  * CALLS: write (sys)
501  *
502  * CALLED BY: dump_line
503  *
504  * HISTORY: initial coding 	November 1976	D A Willcox of CAC
505  *
506  */
507 int
508 pad_output(int current, int target)
509 {
510 	int     curr;		/* internal column pointer */
511 	int     tcur;
512 
513 	if (troff)
514 		fprintf(output, "\\h'|%dp'", (target - 1) * 7);
515 	else {
516 		if (current >= target)
517 			return (current);	/* line is already long enough */
518 		curr = current;
519 		while ((tcur = ((curr - 1) & tabmask) + tabsize + 1) <= target) {
520 			putc('\t', output);
521 			curr = tcur;
522 		}
523 		while (curr++ < target)
524 			putc(' ', output);	/* pad with final blanks */
525 	}
526 	return (target);
527 }
528 /*
529  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
530  *
531  * All rights reserved
532  *
533  *
534  * NAME: count_spaces
535  *
536  * FUNCTION: Find out where printing of a given string will leave the current
537  * character position on output.
538  *
539  * ALGORITHM: Run thru input string and add appropriate values to current
540  * position.
541  *
542  * RETURNS: Integer value of position after printing "buffer" starting in column
543  * "current".
544  *
545  * HISTORY: initial coding 	November 1976	D A Willcox of CAC
546  *
547  */
548 int
549 count_spaces(int current, char *buffer)
550 /*
551  * this routine figures out where the character position will be after
552  * printing the text in buffer starting at column "current"
553  */
554 {
555 	char   *buf;		/* used to look thru buffer */
556 	int     cur;		/* current character counter */
557 
558 	cur = current;
559 
560 	for (buf = buffer; *buf != '\0'; ++buf) {
561 		switch (*buf) {
562 
563 		case '\n':
564 		case 014:	/* form feed */
565 			cur = 1;
566 			break;
567 
568 		case '\t':
569 			cur = ((cur - 1) & tabmask) + tabsize + 1;
570 			break;
571 
572 		case 010:	/* backspace */
573 			--cur;
574 			break;
575 
576 		default:
577 			++cur;
578 			break;
579 		}		/* end of switch */
580 	}			/* end of for loop */
581 	return (cur);
582 }
583 
584 
585 int     found_err;
586 
587 void
588 diag(int level, const char *msg, ...)
589 {
590 	va_list ap;
591 
592 	va_start(ap, msg);
593 
594 	if (level)
595 		found_err = 1;
596 	if (output == stdout) {
597 		fprintf(stdout, "/**INDENT** %s@%d: ", level == 0 ? "Warning" : "Error", line_no);
598 		vfprintf(stdout, msg, ap);
599 		fprintf(stdout, " */\n");
600 	} else {
601 		fprintf(stderr, "%s@%d: ", level == 0 ? "Warning" : "Error", line_no);
602 		vfprintf(stdout, msg, ap);
603 		fprintf(stderr, "\n");
604 	}
605 	va_end(ap);
606 }
607 
608 void
609 writefdef(struct fstate *f, int nm)
610 {
611 	fprintf(output, ".ds f%c %s\n.nr s%c %d\n",
612 	    nm, f->font, nm, f->size);
613 }
614 
615 char   *
616 chfont(struct fstate *of, struct fstate *nf, char *s)
617 {
618 	if (of->font[0] != nf->font[0]
619 	    || of->font[1] != nf->font[1]) {
620 		*s++ = '\\';
621 		*s++ = 'f';
622 		if (nf->font[1]) {
623 			*s++ = '(';
624 			*s++ = nf->font[0];
625 			*s++ = nf->font[1];
626 		} else
627 			*s++ = nf->font[0];
628 	}
629 	if (nf->size != of->size) {
630 		*s++ = '\\';
631 		*s++ = 's';
632 		if (nf->size < of->size) {
633 			*s++ = '-';
634 			*s++ = '0' + of->size - nf->size;
635 		} else {
636 			*s++ = '+';
637 			*s++ = '0' + nf->size - of->size;
638 		}
639 	}
640 	return s;
641 }
642 
643 
644 void
645 parsefont(struct fstate *f, const char *s0)
646 {
647 	const char *s = s0;
648 	int     sizedelta = 0;
649 	memset(f, 0, sizeof *f);
650 	while (*s) {
651 		if (isdigit((unsigned char)*s))
652 			f->size = f->size * 10 + *s - '0';
653 		else
654 			if (isupper((unsigned char)*s)) {
655 				if (f->font[0])
656 					f->font[1] = *s;
657 				else
658 					f->font[0] = *s;
659 			} else
660 				if (*s == 'c')
661 					f->allcaps = 1;
662 				else
663 					if (*s == '+')
664 						sizedelta++;
665 					else
666 						if (*s == '-')
667 							sizedelta--;
668 						else {
669 							fprintf(stderr, "indent: bad font specification: %s\n", s0);
670 							exit(1);
671 						}
672 		s++;
673 	}
674 	if (f->font[0] == 0)
675 		f->font[0] = 'R';
676 	if (bodyf.size == 0)
677 		bodyf.size = 11;
678 	if (f->size == 0)
679 		f->size = bodyf.size + sizedelta;
680 	else
681 		if (sizedelta > 0)
682 			f->size += bodyf.size;
683 		else
684 			f->size = bodyf.size - f->size;
685 }
686