xref: /netbsd-src/usr.bin/indent/pr_comment.c (revision 7d62b00eb9ad855ffcd7da46b41e23feb5476fac)
1 /*	$NetBSD: pr_comment.c,v 1.128 2022/05/09 21:41:49 rillig Exp $	*/
2 
3 /*-
4  * SPDX-License-Identifier: BSD-4-Clause
5  *
6  * Copyright (c) 1985 Sun Microsystems, Inc.
7  * Copyright (c) 1980, 1993
8  *	The Regents of the University of California.  All rights reserved.
9  * All rights reserved.
10  *
11  * Redistribution and use in source and binary forms, with or without
12  * modification, are permitted provided that the following conditions
13  * are met:
14  * 1. Redistributions of source code must retain the above copyright
15  *    notice, this list of conditions and the following disclaimer.
16  * 2. Redistributions in binary form must reproduce the above copyright
17  *    notice, this list of conditions and the following disclaimer in the
18  *    documentation and/or other materials provided with the distribution.
19  * 3. All advertising materials mentioning features or use of this software
20  *    must display the following acknowledgement:
21  *	This product includes software developed by the University of
22  *	California, Berkeley and its contributors.
23  * 4. Neither the name of the University nor the names of its contributors
24  *    may be used to endorse or promote products derived from this software
25  *    without specific prior written permission.
26  *
27  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
28  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
29  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
30  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
31  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
32  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
33  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
34  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
35  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
36  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
37  * SUCH DAMAGE.
38  */
39 
40 #if 0
41 static char sccsid[] = "@(#)pr_comment.c	8.1 (Berkeley) 6/6/93";
42 #endif
43 
44 #include <sys/cdefs.h>
45 #if defined(__NetBSD__)
46 __RCSID("$NetBSD: pr_comment.c,v 1.128 2022/05/09 21:41:49 rillig Exp $");
47 #elif defined(__FreeBSD__)
48 __FBSDID("$FreeBSD: head/usr.bin/indent/pr_comment.c 334927 2018-06-10 16:44:18Z pstef $");
49 #endif
50 
51 #include <assert.h>
52 #include <stdio.h>
53 #include <string.h>
54 
55 #include "indent.h"
56 
57 static void
58 com_add_char(char ch)
59 {
60     if (1 >= com.l - com.e)
61 	buf_expand(&com, 1);
62     *com.e++ = ch;
63 }
64 
65 static void
66 com_add_delim(void)
67 {
68     if (!opt.star_comment_cont)
69 	return;
70     const char *delim = " * ";
71     buf_add_range(&com, delim, delim + 3);
72 }
73 
74 static void
75 com_terminate(void)
76 {
77     if (1 >= com.l - com.e)
78 	buf_expand(&com, 1);
79     *com.e = '\0';
80 }
81 
82 static bool
83 fits_in_one_line(int com_ind, int max_line_length)
84 {
85     for (const char *p = inp_p(); *p != '\n'; p++) {
86 	assert(*p != '\0');
87 	assert(inp_line_end() - p >= 2);
88 	if (!(p[0] == '*' && p[1] == '/'))
89 	    continue;
90 
91 	int len = ind_add(com_ind + 3, inp_p(), p);
92 	len += ch_isblank(p[-1]) ? 2 : 3;
93 	return len <= max_line_length;
94     }
95     return false;
96 }
97 
98 static void
99 analyze_comment(bool *p_may_wrap, bool *p_break_delim,
100     int *p_adj_max_line_length)
101 {
102     int adj_max_line_length;	/* Adjusted max_line_length for comments that
103 				 * spill over the right margin */
104     bool break_delim = opt.comment_delimiter_on_blankline;
105     int com_ind;
106 
107     adj_max_line_length = opt.max_line_length;
108     bool may_wrap = true;
109 
110     if (ps.curr_col_1 && !opt.format_col1_comments) {
111 	may_wrap = false;
112 	break_delim = false;
113 	com_ind = 0;
114 
115     } else {
116 	if (inp_peek() == '-' || inp_peek() == '*' ||
117 		token.e[-1] == '/' ||
118 		(inp_peek() == '\n' && !opt.format_block_comments)) {
119 	    may_wrap = false;
120 	    break_delim = false;
121 	}
122 
123 	/*
124 	 * XXX: This condition looks suspicious since it ignores the case
125 	 * where the end of the previous comment is still in 'com'.
126 	 *
127 	 * See test token_comment.c, keyword 'analyze_comment'.
128 	 */
129 	if (lab.s == lab.e && code.s == code.e) {
130 	    adj_max_line_length = opt.block_comment_max_line_length;
131 	    com_ind = (ps.ind_level - opt.unindent_displace) * opt.indent_size;
132 	    if (com_ind <= 0)
133 		com_ind = opt.format_col1_comments ? 0 : 1;
134 
135 	} else {
136 	    break_delim = false;
137 
138 	    int target_ind = code.s != code.e
139 		? ind_add(compute_code_indent(), code.s, code.e)
140 		: ind_add(compute_label_indent(), lab.s, lab.e);
141 
142 	    com_ind = ps.decl_on_line || ps.ind_level == 0
143 		? opt.decl_comment_column - 1 : opt.comment_column - 1;
144 	    if (com_ind <= target_ind)
145 		com_ind = next_tab(target_ind);
146 	    if (com_ind + 25 > adj_max_line_length)
147 		adj_max_line_length = com_ind + 25;
148 	}
149     }
150 
151     ps.com_ind = com_ind;
152 
153     if (!may_wrap) {
154 	/*
155 	 * Find out how much indentation there was originally, because that
156 	 * much will have to be ignored by output_complete_line.
157 	 */
158 	const char *start = inp_line_start();
159 	ps.n_comment_delta = -ind_add(0, start, inp_p() - 2);
160     } else {
161 	ps.n_comment_delta = 0;
162 	while (ch_isblank(inp_peek()))
163 	    inp_skip();
164     }
165 
166     ps.comment_delta = 0;
167     com_add_char('/');
168     com_add_char(token.e[-1]);	/* either '*' or '/' */
169 
170     /* TODO: Maybe preserve a single '\t' as well. */
171     if (inp_peek() != ' ' && may_wrap)
172 	com_add_char(' ');
173 
174     if (break_delim && fits_in_one_line(com_ind, adj_max_line_length))
175 	break_delim = false;
176 
177     if (break_delim) {
178 	if (opt.blanklines_before_block_comments &&
179 		ps.prev_token != lsym_lbrace)
180 	    out.blank_line_before = true;
181 	output_line();
182 	com_add_delim();
183     }
184 
185     *p_adj_max_line_length = adj_max_line_length;
186     *p_break_delim = break_delim;
187     *p_may_wrap = may_wrap;
188 }
189 
190 /*
191  * Copy characters from 'inp' to 'com'. Try to keep comments from going over
192  * the maximum line length. To do that, remember where the last blank, tab, or
193  * newline was. When a line is filled, print up to the last blank and continue
194  * copying.
195  */
196 static void
197 copy_comment_wrap(int adj_max_line_length, bool break_delim)
198 {
199     ssize_t last_blank = -1;	/* index of the last blank in com.buf */
200 
201     for (;;) {
202 	switch (inp_peek()) {
203 	case '\f':
204 	    output_line_ff();
205 	    last_blank = -1;
206 	    com_add_delim();
207 	    inp_skip();
208 	    while (ch_isblank(inp_peek()))
209 		inp_skip();
210 	    break;
211 
212 	case '\n':
213 	    if (had_eof) {
214 		diag(1, "Unterminated comment");
215 		output_line();
216 		return;
217 	    }
218 
219 	    last_blank = -1;
220 	    if (ps.next_col_1) {
221 		if (com.s == com.e)
222 		    com_add_char(' ');	/* force empty line of output */
223 		if (com.e - com.s > 3) {
224 		    output_line();
225 		    com_add_delim();
226 		}
227 		output_line();
228 		com_add_delim();
229 
230 	    } else {
231 		ps.next_col_1 = true;
232 		if (!(com.e > com.s && ch_isblank(com.e[-1])))
233 		    com_add_char(' ');
234 		last_blank = com.e - 1 - com.buf;
235 	    }
236 	    ++line_no;
237 
238 	    bool skip_asterisk = true;
239 	    do {		/* flush any blanks and/or tabs at start of
240 				 * next line */
241 		inp_skip();
242 		if (inp_peek() == '*' && skip_asterisk) {
243 		    skip_asterisk = false;
244 		    inp_skip();
245 		    if (inp_peek() == '/')
246 			goto end_of_comment;
247 		}
248 	    } while (ch_isblank(inp_peek()));
249 
250 	    break;		/* end of case for newline */
251 
252 	case '*':
253 	    inp_skip();
254 	    if (inp_peek() == '/') {
255 	end_of_comment:
256 		inp_skip();
257 
258 		if (break_delim) {
259 		    if (com.e - com.s > 3)
260 			output_line();
261 		    else
262 			com.e = com.s;
263 		    com_add_char(' ');
264 		}
265 
266 		if (!(com.e > com.s && ch_isblank(com.e[-1])))
267 		    com_add_char(' ');
268 		com_add_char('*');
269 		com_add_char('/');
270 		com_terminate();
271 		return;
272 
273 	    } else		/* handle isolated '*' */
274 		com_add_char('*');
275 	    break;
276 
277 	default:		/* we have a random char */
278 	    ;
279 	    int now_len = ind_add(ps.com_ind, com.s, com.e);
280 	    for (;;) {
281 		char ch = inp_next();
282 		if (ch_isblank(ch))
283 		    last_blank = com.e - com.buf;
284 		com_add_char(ch);
285 		now_len++;
286 		if (memchr("*\n\r\b\t", inp_peek(), 6) != NULL)
287 		    break;
288 		if (now_len >= adj_max_line_length && last_blank != -1)
289 		    break;
290 	    }
291 
292 	    ps.next_col_1 = false;
293 
294 	    if (now_len <= adj_max_line_length)
295 		break;
296 	    if (ch_isspace(com.e[-1]))
297 		break;
298 
299 	    if (last_blank == -1) {	/* only a single word in this line */
300 		output_line();
301 		com_add_delim();
302 		break;
303 	    }
304 
305 	    const char *last_word_s = com.buf + last_blank + 1;
306 	    size_t last_word_len = (size_t)(com.e - last_word_s);
307 	    com.e = com.buf + last_blank;
308 	    output_line();
309 	    com_add_delim();
310 
311 	    memcpy(com.e, last_word_s, last_word_len);
312 	    com.e += last_word_len;
313 	    last_blank = -1;
314 	}
315     }
316 }
317 
318 static void
319 copy_comment_nowrap(void)
320 {
321     for (;;) {
322 	if (inp_peek() == '\n') {
323 	    if (token.e[-1] == '/') {
324 		com_terminate();
325 		return;
326 	    }
327 
328 	    if (had_eof) {
329 		diag(1, "Unterminated comment");
330 		output_line();
331 		return;
332 	    }
333 
334 	    if (com.s == com.e)
335 		com_add_char(' ');	/* force output of an empty line */
336 	    output_line();
337 	    ++line_no;
338 	    inp_skip();
339 	    continue;
340 	}
341 
342 	com_add_char(inp_next());
343 	if (com.e[-2] == '*' && com.e[-1] == '/' && token.e[-1] == '*') {
344 	    com_terminate();
345 	    return;
346 	}
347     }
348 }
349 
350 /*
351  * Scan, reformat and output a single comment, which is either a block comment
352  * starting with '/' '*' or an end-of-line comment starting with '//'.
353  */
354 void
355 process_comment(void)
356 {
357     int adj_max_line_length;
358     bool may_wrap, break_delim;
359 
360     ps.just_saw_decl = 0;
361     ps.stats.comments++;
362 
363     int saved_just_saw_decl = ps.just_saw_decl;
364     analyze_comment(&may_wrap, &break_delim, &adj_max_line_length);
365     if (may_wrap)
366 	copy_comment_wrap(adj_max_line_length, break_delim);
367     else
368 	copy_comment_nowrap();
369     ps.just_saw_decl = saved_just_saw_decl;
370 }
371