xref: /freebsd/usr.bin/indent/pr_comment.c (revision 1d386b48)
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[] = "@(#)pr_comment.c	8.1 (Berkeley) 6/6/93";
41 #endif /* not lint */
42 #endif
43 
44 #include <sys/cdefs.h>
45 #include <err.h>
46 #include <stdio.h>
47 #include <stdlib.h>
48 #include <string.h>
49 #include "indent_globs.h"
50 #include "indent_codes.h"
51 #include "indent.h"
52 /*
53  * NAME:
54  *	pr_comment
55  *
56  * FUNCTION:
57  *	This routine takes care of scanning and printing comments.
58  *
59  * ALGORITHM:
60  *	1) Decide where the comment should be aligned, and if lines should
61  *	   be broken.
62  *	2) If lines should not be broken and filled, just copy up to end of
63  *	   comment.
64  *	3) If lines should be filled, then scan thru input_buffer copying
65  *	   characters to com_buf.  Remember where the last blank, tab, or
66  *	   newline was.  When line is filled, print up to last blank and
67  *	   continue copying.
68  *
69  * HISTORY:
70  *	November 1976	D A Willcox of CAC	Initial coding
71  *	12/6/76		D A Willcox of CAC	Modification to handle
72  *						UNIX-style comments
73  *
74  */
75 
76 /*
77  * this routine processes comments.  It makes an attempt to keep comments from
78  * going over the max line length.  If a line is too long, it moves everything
79  * from the last blank to the next comment line.  Blanks and tabs from the
80  * beginning of the input line are removed
81  */
82 
83 void
84 pr_comment(void)
85 {
86     int         now_col;	/* column we are in now */
87     int         adj_max_col;	/* Adjusted max_col for when we decide to
88 				 * spill comments over the right margin */
89     char       *last_bl;	/* points to the last blank in the output
90 				 * buffer */
91     char       *t_ptr;		/* used for moving string */
92     int         break_delim = opt.comment_delimiter_on_blankline;
93     int         l_just_saw_decl = ps.just_saw_decl;
94 
95     adj_max_col = opt.max_col;
96     ps.just_saw_decl = 0;
97     last_bl = NULL;		/* no blanks found so far */
98     ps.box_com = false;		/* at first, assume that we are not in
99 					 * a boxed comment or some other
100 					 * comment that should not be touched */
101     ++ps.out_coms;		/* keep track of number of comments */
102 
103     /* Figure where to align and how to treat the comment */
104 
105     if (ps.col_1 && !opt.format_col1_comments) {	/* if comment starts in column
106 						 * 1 it should not be touched */
107 	ps.box_com = true;
108 	break_delim = false;
109 	ps.com_col = 1;
110     }
111     else {
112 	if (*buf_ptr == '-' || *buf_ptr == '*' ||
113 	    (*buf_ptr == '\n' && !opt.format_block_comments)) {
114 	    ps.box_com = true;	/* A comment with a '-' or '*' immediately
115 				 * after the /+* is assumed to be a boxed
116 				 * comment. A comment with a newline
117 				 * immediately after the /+* is assumed to
118 				 * be a block comment and is treated as a
119 				 * box comment unless format_block_comments
120 				 * is nonzero (the default). */
121 	    break_delim = false;
122 	}
123 	if ( /* ps.bl_line && */ (s_lab == e_lab) && (s_code == e_code)) {
124 	    /* klg: check only if this line is blank */
125 	    /*
126 	     * If this (*and previous lines are*) blank, dont put comment way
127 	     * out at left
128 	     */
129 	    ps.com_col = (ps.ind_level - opt.unindent_displace) * opt.ind_size + 1;
130 	    adj_max_col = opt.block_comment_max_col;
131 	    if (ps.com_col <= 1)
132 		ps.com_col = 1 + !opt.format_col1_comments;
133 	}
134 	else {
135 	    int target_col;
136 	    break_delim = false;
137 	    if (s_code != e_code)
138 		target_col = count_spaces(compute_code_target(), s_code);
139 	    else {
140 		target_col = 1;
141 		if (s_lab != e_lab)
142 		    target_col = count_spaces(compute_label_target(), s_lab);
143 	    }
144 	    ps.com_col = ps.decl_on_line || ps.ind_level == 0 ? opt.decl_com_ind : opt.com_ind;
145 	    if (ps.com_col <= target_col)
146 		ps.com_col = opt.tabsize * (1 + (target_col - 1) / opt.tabsize) + 1;
147 	    if (ps.com_col + 24 > adj_max_col)
148 		adj_max_col = ps.com_col + 24;
149 	}
150     }
151     if (ps.box_com) {
152 	/*
153 	 * Find out how much indentation there was originally, because that
154 	 * much will have to be ignored by pad_output() in dump_line(). This
155 	 * is a box comment, so nothing changes -- not even indentation.
156 	 *
157 	 * The comment we're about to read usually comes from in_buffer,
158 	 * unless it has been copied into save_com.
159 	 */
160 	char *start;
161 
162 	start = buf_ptr >= save_com && buf_ptr < save_com + sc_size ?
163 	    sc_buf : in_buffer;
164 	ps.n_comment_delta = 1 - count_spaces_until(1, start, buf_ptr - 2);
165     }
166     else {
167 	ps.n_comment_delta = 0;
168 	while (*buf_ptr == ' ' || *buf_ptr == '\t')
169 	    buf_ptr++;
170     }
171     ps.comment_delta = 0;
172     *e_com++ = '/';		/* put '/' followed by '*' into buffer */
173     *e_com++ = '*';
174     if (*buf_ptr != ' ' && !ps.box_com)
175 	*e_com++ = ' ';
176 
177     /*
178      * Don't put a break delimiter if this is a one-liner that won't wrap.
179      */
180     if (break_delim)
181 	for (t_ptr = buf_ptr; *t_ptr != '\0' && *t_ptr != '\n'; t_ptr++) {
182 	    if (t_ptr >= buf_end)
183 		fill_buffer();
184 	    if (t_ptr[0] == '*' && t_ptr[1] == '/') {
185 		if (adj_max_col >= count_spaces_until(ps.com_col, buf_ptr, t_ptr + 2))
186 		    break_delim = false;
187 		break;
188 	    }
189 	}
190 
191     if (break_delim) {
192 	char       *t = e_com;
193 	e_com = s_com + 2;
194 	*e_com = 0;
195 	if (opt.blanklines_before_blockcomments && ps.last_token != lbrace)
196 	    prefix_blankline_requested = 1;
197 	dump_line();
198 	e_com = s_com = t;
199 	if (!ps.box_com && opt.star_comment_cont)
200 	    *e_com++ = ' ', *e_com++ = '*', *e_com++ = ' ';
201     }
202 
203     /* Start to copy the comment */
204 
205     while (1) {			/* this loop will go until the comment is
206 				 * copied */
207 	switch (*buf_ptr) {	/* this checks for various spcl cases */
208 	case 014:		/* check for a form feed */
209 	    CHECK_SIZE_COM(3);
210 	    if (!ps.box_com) {	/* in a text comment, break the line here */
211 		ps.use_ff = true;
212 		/* fix so dump_line uses a form feed */
213 		dump_line();
214 		last_bl = NULL;
215 		if (!ps.box_com && opt.star_comment_cont)
216 		    *e_com++ = ' ', *e_com++ = '*', *e_com++ = ' ';
217 		while (*++buf_ptr == ' ' || *buf_ptr == '\t')
218 		    ;
219 	    }
220 	    else {
221 		if (++buf_ptr >= buf_end)
222 		    fill_buffer();
223 		*e_com++ = 014;
224 	    }
225 	    break;
226 
227 	case '\n':
228 	    if (had_eof) {	/* check for unexpected eof */
229 		printf("Unterminated comment\n");
230 		dump_line();
231 		return;
232 	    }
233 	    last_bl = NULL;
234 	    CHECK_SIZE_COM(4);
235 	    if (ps.box_com || ps.last_nl) {	/* if this is a boxed comment,
236 						 * we dont ignore the newline */
237 		if (s_com == e_com)
238 		    *e_com++ = ' ';
239 		if (!ps.box_com && e_com - s_com > 3) {
240 		    dump_line();
241 		    if (opt.star_comment_cont)
242 			*e_com++ = ' ', *e_com++ = '*', *e_com++ = ' ';
243 		}
244 		dump_line();
245 		if (!ps.box_com && opt.star_comment_cont)
246 		    *e_com++ = ' ', *e_com++ = '*', *e_com++ = ' ';
247 	    }
248 	    else {
249 		ps.last_nl = 1;
250 		if (*(e_com - 1) == ' ' || *(e_com - 1) == '\t')
251 		    last_bl = e_com - 1;
252 		/*
253 		 * if there was a space at the end of the last line, remember
254 		 * where it was
255 		 */
256 		else {		/* otherwise, insert one */
257 		    last_bl = e_com;
258 		    *e_com++ = ' ';
259 		}
260 	    }
261 	    ++line_no;		/* keep track of input line number */
262 	    if (!ps.box_com) {
263 		int         nstar = 1;
264 		do {		/* flush any blanks and/or tabs at start of
265 				 * next line */
266 		    if (++buf_ptr >= buf_end)
267 			fill_buffer();
268 		    if (*buf_ptr == '*' && --nstar >= 0) {
269 			if (++buf_ptr >= buf_end)
270 			    fill_buffer();
271 			if (*buf_ptr == '/')
272 			    goto end_of_comment;
273 		    }
274 		} while (*buf_ptr == ' ' || *buf_ptr == '\t');
275 	    }
276 	    else if (++buf_ptr >= buf_end)
277 		fill_buffer();
278 	    break;		/* end of case for newline */
279 
280 	case '*':		/* must check for possibility of being at end
281 				 * of comment */
282 	    if (++buf_ptr >= buf_end)	/* get to next char after * */
283 		fill_buffer();
284 	    CHECK_SIZE_COM(4);
285 	    if (*buf_ptr == '/') {	/* it is the end!!! */
286 	end_of_comment:
287 		if (++buf_ptr >= buf_end)
288 		    fill_buffer();
289 		if (break_delim) {
290 		    if (e_com > s_com + 3) {
291 			dump_line();
292 		    }
293 		    else
294 			s_com = e_com;
295 		    *e_com++ = ' ';
296 		}
297 		if (e_com[-1] != ' ' && e_com[-1] != '\t' && !ps.box_com)
298 		    *e_com++ = ' ';	/* ensure blank before end */
299 		*e_com++ = '*', *e_com++ = '/', *e_com = '\0';
300 		ps.just_saw_decl = l_just_saw_decl;
301 		return;
302 	    }
303 	    else		/* handle isolated '*' */
304 		*e_com++ = '*';
305 	    break;
306 	default:		/* we have a random char */
307 	    now_col = count_spaces_until(ps.com_col, s_com, e_com);
308 	    do {
309 		CHECK_SIZE_COM(1);
310 		*e_com = *buf_ptr++;
311 		if (buf_ptr >= buf_end)
312 		    fill_buffer();
313 		if (*e_com == ' ' || *e_com == '\t')
314 		    last_bl = e_com;	/* remember we saw a blank */
315 		++e_com;
316 		now_col++;
317 	    } while (!memchr("*\n\r\b\t", *buf_ptr, 6) &&
318 		(now_col <= adj_max_col || !last_bl));
319 	    ps.last_nl = false;
320 	    if (now_col > adj_max_col && !ps.box_com && e_com[-1] > ' ') {
321 		/*
322 		 * the comment is too long, it must be broken up
323 		 */
324 		if (last_bl == NULL) {
325 		    dump_line();
326 		    if (!ps.box_com && opt.star_comment_cont)
327 			*e_com++ = ' ', *e_com++ = '*', *e_com++ = ' ';
328 		    break;
329 		}
330 		*e_com = '\0';
331 		e_com = last_bl;
332 		dump_line();
333 		if (!ps.box_com && opt.star_comment_cont)
334 		    *e_com++ = ' ', *e_com++ = '*', *e_com++ = ' ';
335 		for (t_ptr = last_bl + 1; *t_ptr == ' ' || *t_ptr == '\t';
336 		    t_ptr++)
337 			;
338 		last_bl = NULL;
339 		/*
340 		 * t_ptr will be somewhere between e_com (dump_line() reset)
341 		 * and l_com. So it's safe to copy byte by byte from t_ptr
342 		 * to e_com without any CHECK_SIZE_COM().
343 		 */
344 		while (*t_ptr != '\0') {
345 		    if (*t_ptr == ' ' || *t_ptr == '\t')
346 			last_bl = e_com;
347 		    *e_com++ = *t_ptr++;
348 		}
349 	    }
350 	    break;
351 	}
352     }
353 }
354