xref: /openbsd/usr.bin/indent/pr_comment.c (revision 043fbe51)
1 /*	$OpenBSD: pr_comment.c,v 1.7 2009/10/27 23:59:39 deraadt Exp $	*/
2 
3 /*
4  * Copyright (c) 1980, 1993
5  *	The Regents of the University of California.
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. Neither the name of the University nor the names of its contributors
19  *    may be used to endorse or promote products derived from this software
20  *    without specific prior written permission.
21  *
22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  */
34 
35 #include <err.h>
36 #include <stdio.h>
37 #include <stdlib.h>
38 #include "indent_globs.h"
39 
40 /*
41  * NAME:
42  *	pr_comment
43  *
44  * FUNCTION:
45  *	This routine takes care of scanning and printing comments.
46  *
47  * ALGORITHM:
48  *	1) Decide where the comment should be aligned, and if lines should
49  *	   be broken.
50  *	2) If lines should not be broken and filled, just copy up to end of
51  *	   comment.
52  *	3) If lines should be filled, then scan thru input_buffer copying
53  *	   characters to com_buf.  Remember where the last blank, tab, or
54  *	   newline was.  When line is filled, print up to last blank and
55  *	   continue copying.
56  *
57  * HISTORY:
58  *	November 1976	D A Willcox of CAC	Initial coding
59  *	12/6/76		D A Willcox of CAC	Modification to handle
60  *						UNIX-style comments
61  *
62  */
63 
64 /*
65  * this routine processes comments.  It makes an attempt to keep comments from
66  * going over the max line length.  If a line is too long, it moves everything
67  * from the last blank to the next comment line.  Blanks and tabs from the
68  * beginning of the input line are removed
69  */
70 
71 void
pr_comment(void)72 pr_comment(void)
73 {
74     int         now_col;	/* column we are in now */
75     int         adj_max_col;	/* Adjusted max_col for when we decide to
76 				 * spill comments over the right margin */
77     char       *last_bl;	/* points to the last blank in the output
78 				 * buffer */
79     char       *t_ptr;		/* used for moving string */
80     int         unix_comment;	/* tri-state variable used to decide if it is
81 				 * a unix-style comment. 0 means only blanks
82 				 * since / *, 1 means regular style comment, 2
83 				 * means unix style comment */
84     int         break_delim = comment_delimiter_on_blankline;
85     int         l_just_saw_decl = ps.just_saw_decl;
86     /*
87      * int         ps.last_nl = 0;	 true iff the last significant thing
88      * we've seen is a newline
89      */
90     int         one_liner = 1;	/* true iff this comment is a one-liner */
91     adj_max_col = max_col;
92     ps.just_saw_decl = 0;
93     last_bl = 0;		/* no blanks found so far */
94     ps.box_com = false;		/* at first, assume that we are not in
95 				 * a boxed comment or some other
96 				 * comment that should not be touched */
97     ++ps.out_coms;		/* keep track of number of comments */
98     unix_comment = 1;		/* set flag to let us figure out if there is a
99 				 * unix-style comment ** DISABLED: use 0 to
100 				 * reenable this hack! */
101 
102     /* Figure where to align and how to treat the comment */
103 
104     if (ps.col_1 && !format_col1_comments) {	/* if comment starts in column
105 						 * 1 it should not be touched */
106 	ps.box_com = true;
107 	ps.com_col = 1;
108     }
109     else {
110 	if (*buf_ptr == '-' || *buf_ptr == '*' || *buf_ptr == '\n') {
111 	    ps.box_com = true;	/* a comment with a '-', '*' or newline
112 				 * immediately after the / * is assumed to be
113 				 * a boxed comment */
114 	    break_delim = 0;
115 	}
116 	if ( /* ps.bl_line && */ (s_lab == e_lab) && (s_code == e_code)) {
117 	    /* klg: check only if this line is blank */
118 	    /*
119 	     * If this (*and previous lines are*) blank, don't put comment way
120 	     * out at left
121 	     */
122 	    ps.com_col = (ps.ind_level - ps.unindent_displace) * ps.ind_size + 1;
123 	    adj_max_col = block_comment_max_col;
124 	    if (ps.com_col <= 1)
125 		ps.com_col = 1 + !format_col1_comments;
126 	}
127 	else {
128 	    int    target_col;
129 	    break_delim = 0;
130 	    if (s_code != e_code)
131 		target_col = count_spaces(compute_code_target(), s_code);
132 	    else {
133 		target_col = 1;
134 		if (s_lab != e_lab)
135 		    target_col = count_spaces(compute_label_target(), s_lab);
136 	    }
137 	    ps.com_col = ps.decl_on_line || ps.ind_level == 0 ? ps.decl_com_ind : ps.com_ind;
138 	    if (ps.com_col < target_col)
139 		ps.com_col = ((target_col + 7) & ~7) + 1;
140 	    if (ps.com_col + 24 > adj_max_col)
141 		adj_max_col = ps.com_col + 24;
142 	}
143     }
144     if (ps.box_com) {
145 	buf_ptr[-2] = 0;
146 	ps.n_comment_delta = 1 - count_spaces(1, in_buffer);
147 	buf_ptr[-2] = '/';
148     }
149     else {
150 	ps.n_comment_delta = 0;
151 	while (*buf_ptr == ' ' || *buf_ptr == '\t')
152 	    buf_ptr++;
153     }
154     ps.comment_delta = 0;
155     *e_com++ = '/';		/* put '/ *' into buffer */
156     *e_com++ = '*';
157     if (*buf_ptr != ' ' && !ps.box_com)
158 	*e_com++ = ' ';
159 
160     *e_com = '\0';
161     if (troff) {
162 	now_col = 1;
163 	adj_max_col = 80;
164     }
165     else
166 	now_col = count_spaces(ps.com_col, s_com);	/* figure what column we
167 							 * would be in if we
168 							 * printed the comment
169 							 * now */
170 
171     /* Start to copy the comment */
172 
173     while (1) {			/* this loop will go until the comment is
174 				 * copied */
175 	if (*buf_ptr > 040 && *buf_ptr != '*')
176 	    ps.last_nl = 0;
177 	CHECK_SIZE_COM;
178 	switch (*buf_ptr) {	/* this checks for various spcl cases */
179 	case 014:		/* check for a form feed */
180 	    if (!ps.box_com) {	/* in a text comment, break the line here */
181 		ps.use_ff = true;
182 		/* fix so dump_line uses a form feed */
183 		dump_line();
184 		last_bl = 0;
185 		*e_com++ = ' ';
186 		*e_com++ = '*';
187 		*e_com++ = ' ';
188 		while (*++buf_ptr == ' ' || *buf_ptr == '\t');
189 	    }
190 	    else {
191 		if (++buf_ptr >= buf_end)
192 		    fill_buffer();
193 		*e_com++ = 014;
194 	    }
195 	    break;
196 
197 	case '\n':
198 	    if (had_eof) {	/* check for unexpected eof */
199 		printf("Unterminated comment\n");
200 		*e_com = '\0';
201 		dump_line();
202 		return;
203 	    }
204 	    one_liner = 0;
205 	    if (ps.box_com || ps.last_nl) {	/* if this is a boxed comment,
206 						 * we don't ignore the newline
207 						 */
208 		if (s_com == e_com) {
209 		    *e_com++ = ' ';
210 		    *e_com++ = ' ';
211 		}
212 		*e_com = '\0';
213 		if (!ps.box_com && e_com - s_com > 3) {
214 		    if (break_delim == 1 && s_com[0] == '/'
215 			    && s_com[1] == '*' && s_com[2] == ' ') {
216 			char       *t = e_com;
217 			break_delim = 2;
218 			e_com = s_com + 2;
219 			*e_com = 0;
220 			if (blanklines_before_blockcomments)
221 			    prefix_blankline_requested = 1;
222 			dump_line();
223 			e_com = t;
224 			s_com[0] = s_com[1] = s_com[2] = ' ';
225 		    }
226 		    dump_line();
227 		    CHECK_SIZE_COM;
228 		    *e_com++ = ' ';
229 		    *e_com++ = ' ';
230 		}
231 		dump_line();
232 		now_col = ps.com_col;
233 	    }
234 	    else {
235 		ps.last_nl = 1;
236 		if (unix_comment != 1) {	/* we not are in unix_style
237 						 * comment */
238 		    if (unix_comment == 0 && s_code == e_code) {
239 			/*
240 			 * if it is a UNIX-style comment, ignore the
241 			 * requirement that previous line be blank for
242 			 * unindention
243 			 */
244 			ps.com_col = (ps.ind_level - ps.unindent_displace) * ps.ind_size + 1;
245 			if (ps.com_col <= 1)
246 			    ps.com_col = 2;
247 		    }
248 		    unix_comment = 2;	/* permanently remember that we are in
249 					 * this type of comment */
250 		    dump_line();
251 		    ++line_no;
252 		    now_col = ps.com_col;
253 		    *e_com++ = ' ';
254 		    /*
255 		     * fix so that the star at the start of the line will line
256 		     * up
257 		     */
258 		    do		/* flush leading white space */
259 			if (++buf_ptr >= buf_end)
260 			    fill_buffer();
261 		    while (*buf_ptr == ' ' || *buf_ptr == '\t');
262 		    break;
263 		}
264 		if (*(e_com - 1) == ' ' || *(e_com - 1) == '\t')
265 		    last_bl = e_com - 1;
266 		/*
267 		 * if there was a space at the end of the last line, remember
268 		 * where it was
269 		 */
270 		else {		/* otherwise, insert one */
271 		    last_bl = e_com;
272 		    CHECK_SIZE_COM;
273 		    *e_com++ = ' ';
274 		    ++now_col;
275 		}
276 	    }
277 	    ++line_no;		/* keep track of input line number */
278 	    if (!ps.box_com) {
279 		int         nstar = 1;
280 		do {		/* flush any blanks and/or tabs at start of
281 				 * next line */
282 		    if (++buf_ptr >= buf_end)
283 			fill_buffer();
284 		    if (*buf_ptr == '*' && --nstar >= 0) {
285 			if (++buf_ptr >= buf_end)
286 			    fill_buffer();
287 			if (*buf_ptr == '/')
288 			    goto end_of_comment;
289 		    }
290 		} while (*buf_ptr == ' ' || *buf_ptr == '\t');
291 	    }
292 	    else if (++buf_ptr >= buf_end)
293 		fill_buffer();
294 	    break;		/* end of case for newline */
295 
296 	case '*':		/* must check for possibility of being at end
297 				 * of comment */
298 	    if (++buf_ptr >= buf_end)	/* get to next char after * */
299 		fill_buffer();
300 
301 	    if (unix_comment == 0)	/* set flag to show we are not in
302 					 * unix-style comment */
303 		unix_comment = 1;
304 
305 	    if (*buf_ptr == '/') {	/* it is the end!!! */
306 	end_of_comment:
307 		if (++buf_ptr >= buf_end)
308 		    fill_buffer();
309 
310 		if (*(e_com - 1) != ' ' && !ps.box_com) {	/* insure blank before
311 								 * end */
312 		    *e_com++ = ' ';
313 		    ++now_col;
314 		}
315 		if (break_delim == 1 && !one_liner && s_com[0] == '/'
316 			&& s_com[1] == '*' && s_com[2] == ' ') {
317 		    char       *t = e_com;
318 		    break_delim = 2;
319 		    e_com = s_com + 2;
320 		    *e_com = 0;
321 		    if (blanklines_before_blockcomments)
322 			prefix_blankline_requested = 1;
323 		    dump_line();
324 		    e_com = t;
325 		    s_com[0] = s_com[1] = s_com[2] = ' ';
326 		}
327 		if (break_delim == 2 && e_com > s_com + 3
328 			 /* now_col > adj_max_col - 2 && !ps.box_com */ ) {
329 		    *e_com = '\0';
330 		    dump_line();
331 		    now_col = ps.com_col;
332 		}
333 		CHECK_SIZE_COM;
334 		*e_com++ = '*';
335 		*e_com++ = '/';
336 		*e_com = '\0';
337 		ps.just_saw_decl = l_just_saw_decl;
338 		return;
339 	    }
340 	    else {		/* handle isolated '*' */
341 		*e_com++ = '*';
342 		++now_col;
343 	    }
344 	    break;
345 	default:		/* we have a random char */
346 	    if (unix_comment == 0 && *buf_ptr != ' ' && *buf_ptr != '\t')
347 		unix_comment = 1;	/* we are not in unix-style comment */
348 
349 	    *e_com = *buf_ptr++;
350 	    if (buf_ptr >= buf_end)
351 		fill_buffer();
352 
353 	    if (*e_com == '\t')	/* keep track of column */
354 		now_col = ((now_col - 1) & tabmask) + tabsize + 1;
355 	    else if (*e_com == '\b')	/* this is a backspace */
356 		--now_col;
357 	    else
358 		++now_col;
359 
360 	    if (*e_com == ' ' || *e_com == '\t')
361 		last_bl = e_com;
362 	    /* remember we saw a blank */
363 
364 	    ++e_com;
365 	    if (now_col > adj_max_col && !ps.box_com && unix_comment == 1 && e_com[-1] > ' ') {
366 		/*
367 		 * the comment is too long, it must be broken up
368 		 */
369 		if (break_delim == 1 && s_com[0] == '/'
370 			&& s_com[1] == '*' && s_com[2] == ' ') {
371 		    char       *t = e_com;
372 		    break_delim = 2;
373 		    e_com = s_com + 2;
374 		    *e_com = 0;
375 		    if (blanklines_before_blockcomments)
376 			prefix_blankline_requested = 1;
377 		    dump_line();
378 		    e_com = t;
379 		    s_com[0] = s_com[1] = s_com[2] = ' ';
380 		}
381 		if (last_bl == 0) {	/* we have seen no blanks */
382 		    last_bl = e_com;	/* fake it */
383 		    *e_com++ = ' ';
384 		}
385 		*e_com = '\0';	/* print what we have */
386 		*last_bl = '\0';
387 		while (last_bl > s_com && last_bl[-1] < 040)
388 		    *--last_bl = 0;
389 		e_com = last_bl;
390 		dump_line();
391 
392 		*e_com++ = ' ';	/* add blanks for continuation */
393 		*e_com++ = ' ';
394 		*e_com++ = ' ';
395 
396 		t_ptr = last_bl + 1;
397 		last_bl = 0;
398 		if (t_ptr >= e_com) {
399 		    while (*t_ptr == ' ' || *t_ptr == '\t')
400 			t_ptr++;
401 		    while (*t_ptr != '\0') {	/* move unprinted part of
402 						 * comment down in buffer */
403 			if (*t_ptr == ' ' || *t_ptr == '\t')
404 			    last_bl = e_com;
405 			*e_com++ = *t_ptr++;
406 		    }
407 		}
408 		*e_com = '\0';
409 		now_col = count_spaces(ps.com_col, s_com);	/* recompute current
410 								 * position */
411 	    }
412 	    break;
413 	}
414     }
415 }
416