xref: /netbsd-src/usr.bin/indent/io.c (revision 122b5006ee1bd67145794b4cde92f4fe4781a5ec)
1 /*	$NetBSD: io.c,v 1.108 2021/10/30 11:49:38 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[] = "@(#)io.c	8.1 (Berkeley) 6/6/93";
42 #endif
43 
44 #include <sys/cdefs.h>
45 #if defined(__NetBSD__)
46 __RCSID("$NetBSD: io.c,v 1.108 2021/10/30 11:49:38 rillig Exp $");
47 #elif defined(__FreeBSD__)
48 __FBSDID("$FreeBSD: head/usr.bin/indent/io.c 334927 2018-06-10 16:44:18Z pstef $");
49 #endif
50 
51 #include <ctype.h>
52 #include <stdarg.h>
53 #include <stdio.h>
54 #include <string.h>
55 
56 #include "indent.h"
57 
58 static int paren_indent;
59 static bool suppress_blanklines;
60 
61 static void
62 output_char(char ch)
63 {
64     fputc(ch, output);
65     debug_vis_range("output_char '", &ch, &ch + 1, "'\n");
66 }
67 
68 static void
69 output_range(const char *s, const char *e)
70 {
71     fwrite(s, 1, (size_t)(e - s), output);
72     debug_vis_range("output_range \"", s, e, "\"\n");
73 }
74 
75 static inline void
76 output_string(const char *s)
77 {
78     output_range(s, s + strlen(s));
79 }
80 
81 static int
82 output_indent(int old_ind, int new_ind)
83 {
84     int ind = old_ind;
85 
86     if (opt.use_tabs) {
87 	int tabsize = opt.tabsize;
88 	int n = new_ind / tabsize - ind / tabsize;
89 	if (n > 0)
90 	    ind -= ind % tabsize;
91 	for (int i = 0; i < n; i++) {
92 	    fputc('\t', output);
93 	    ind += tabsize;
94 	}
95     }
96 
97     for (; ind < new_ind; ind++)
98 	fputc(' ', output);
99 
100     debug_println("output_indent %d", ind);
101     return ind;
102 }
103 
104 static int
105 dump_line_label(void)
106 {
107     int ind;
108 
109     while (lab.e > lab.s && ch_isblank(lab.e[-1]))
110 	lab.e--;
111     *lab.e = '\0';
112 
113     ind = output_indent(0, compute_label_indent());
114 
115     if (lab.s[0] == '#' && (strncmp(lab.s, "#else", 5) == 0
116 	    || strncmp(lab.s, "#endif", 6) == 0)) {
117 	const char *s = lab.s;
118 	if (lab.e[-1] == '\n')
119 	    lab.e--;
120 	do {
121 	    output_char(*s++);
122 	} while (s < lab.e && 'a' <= *s && *s <= 'z');
123 
124 	while (s < lab.e && ch_isblank(*s))
125 	    s++;
126 
127 	if (s < lab.e) {
128 	    if (s[0] == '/' && s[1] == '*') {
129 		output_char('\t');
130 		output_range(s, lab.e);
131 	    } else {
132 		output_string("\t/* ");
133 		output_range(s, lab.e);
134 		output_string(" */");
135 	    }
136 	}
137     } else
138 	output_range(lab.s, lab.e);
139     ind = indentation_after(ind, lab.s);
140 
141     ps.is_case_label = false;
142     return ind;
143 }
144 
145 static int
146 dump_line_code(int ind)
147 {
148 
149     int target_ind = compute_code_indent();
150     for (int i = 0; i < ps.p_l_follow; i++) {
151 	if (ps.paren_indents[i] >= 0) {
152 	    int paren_ind = ps.paren_indents[i];
153 	    /* XXX: the '+ 1' smells like an off-by-one error. */
154 	    ps.paren_indents[i] = (short)-(paren_ind + target_ind + 1);
155 	    debug_println(
156 		"setting paren_indents[%d] from %d to %d for column %d",
157 		i, paren_ind, ps.paren_indents[i], target_ind + 1);
158 	}
159     }
160 
161     ind = output_indent(ind, target_ind);
162     output_range(code.s, code.e);
163     return indentation_after(ind, code.s);
164 }
165 
166 static void
167 dump_line_comment(int ind)
168 {
169     int target_ind = ps.com_ind;
170     const char *p = com.s;
171 
172     target_ind += ps.comment_delta;
173 
174     /* consider original indentation in case this is a box comment */
175     for (; *p == '\t'; p++)
176 	target_ind += opt.tabsize;
177 
178     for (; target_ind < 0; p++) {
179 	if (*p == ' ')
180 	    target_ind++;
181 	else if (*p == '\t')
182 	    target_ind = next_tab(target_ind);
183 	else {
184 	    target_ind = 0;
185 	    break;
186 	}
187     }
188 
189     /* if comment can't fit on this line, put it on the next line */
190     if (ind > target_ind) {
191 	output_char('\n');
192 	ind = 0;
193 	ps.stats.lines++;
194     }
195 
196     while (com.e > p && isspace((unsigned char)com.e[-1]))
197 	com.e--;
198 
199     (void)output_indent(ind, target_ind);
200     output_range(p, com.e);
201 
202     ps.comment_delta = ps.n_comment_delta;
203     ps.stats.comment_lines++;
204 }
205 
206 /*
207  * Write a line of formatted source to the output file. The line consists of
208  * the label, the code and the comment.
209  */
210 static void
211 output_line(char line_terminator)
212 {
213     static bool first_line = true;
214 
215     ps.procname[0] = '\0';
216 
217     if (code.s == code.e && lab.s == lab.e && com.s == com.e) {
218 	if (suppress_blanklines)
219 	    suppress_blanklines = false;
220 	else
221 	    blank_lines_to_output++;
222 
223     } else if (!inhibit_formatting) {
224 	suppress_blanklines = false;
225 	if (blank_line_before && !first_line) {
226 	    if (opt.swallow_optional_blanklines) {
227 		if (blank_lines_to_output == 1)
228 		    blank_lines_to_output = 0;
229 	    } else {
230 		if (blank_lines_to_output == 0)
231 		    blank_lines_to_output = 1;
232 	    }
233 	}
234 
235 	for (; blank_lines_to_output > 0; blank_lines_to_output--)
236 	    output_char('\n');
237 
238 	if (ps.ind_level == 0)
239 	    ps.ind_stmt = false;	/* this is a class A kludge. don't do
240 					 * additional statement indentation if
241 					 * we are at bracket level 0 */
242 
243 	if (lab.e != lab.s || code.e != code.s)
244 	    ps.stats.code_lines++;
245 
246 	int ind = 0;
247 	if (lab.e != lab.s)
248 	    ind = dump_line_label();
249 	if (code.e != code.s)
250 	    ind = dump_line_code(ind);
251 	if (com.e != com.s)
252 	    dump_line_comment(ind);
253 
254 	output_char(line_terminator);
255 	ps.stats.lines++;
256 
257 	if (ps.just_saw_decl == 1 && opt.blanklines_after_decl) {
258 	    blank_line_before = true;
259 	    ps.just_saw_decl = 0;
260 	} else
261 	    blank_line_before = blank_line_after;
262 	blank_line_after = false;
263     }
264 
265     ps.decl_on_line = ps.in_decl;	/* for proper comment indentation */
266     ps.ind_stmt = ps.in_stmt && !ps.in_decl;
267     ps.decl_indent_done = false;
268 
269     *(lab.e = lab.s) = '\0';	/* reset buffers */
270     *(code.e = code.s) = '\0';
271     *(com.e = com.s = com.buf + 1) = '\0';
272 
273     ps.ind_level = ps.ind_level_follow;
274     ps.paren_level = ps.p_l_follow;
275 
276     if (ps.paren_level > 0) {
277 	/* TODO: explain what negative indentation means */
278 	paren_indent = -ps.paren_indents[ps.paren_level - 1];
279 	debug_println("paren_indent is now %d", paren_indent);
280     }
281 
282     first_line = false;
283 }
284 
285 void
286 dump_line(void)
287 {
288     output_line('\n');
289 }
290 
291 void
292 dump_line_ff(void)
293 {
294     output_line('\f');
295 }
296 
297 int
298 compute_code_indent(void)
299 {
300     int target_ind = opt.indent_size * ps.ind_level;
301 
302     if (ps.paren_level != 0) {
303 	if (!opt.lineup_to_parens) {
304 	    if (2 * opt.continuation_indent == opt.indent_size)
305 		target_ind += opt.continuation_indent;
306 	    else
307 		target_ind += opt.continuation_indent * ps.paren_level;
308 
309 	} else if (opt.lineup_to_parens_always) {
310 	    /*
311 	     * XXX: where does this '- 1' come from?  It looks strange but is
312 	     * nevertheless needed for proper indentation, as demonstrated in
313 	     * the test opt-lpl.0.
314 	     */
315 	    target_ind = paren_indent - 1;
316 
317 	} else {
318 	    int w;
319 	    int t = paren_indent;
320 
321 	    if ((w = 1 + indentation_after(t - 1, code.s) - opt.max_line_length) > 0
322 		&& 1 + indentation_after(target_ind, code.s) <= opt.max_line_length) {
323 		t -= w + 1;
324 		if (t > target_ind + 1)
325 		    target_ind = t - 1;
326 	    } else
327 		target_ind = t - 1;
328 	}
329 
330     } else if (ps.ind_stmt)
331 	target_ind += opt.continuation_indent;
332 
333     return target_ind;
334 }
335 
336 int
337 compute_label_indent(void)
338 {
339     if (ps.is_case_label)
340 	return (int)(case_ind * (float)opt.indent_size);
341     if (lab.s[0] == '#')
342 	return 0;
343     return opt.indent_size * (ps.ind_level - 2);
344 }
345 
346 static void
347 skip_blank(const char **pp)
348 {
349     while (ch_isblank(**pp))
350 	(*pp)++;
351 }
352 
353 static bool
354 skip_string(const char **pp, const char *s)
355 {
356     size_t len = strlen(s);
357     if (strncmp(*pp, s, len) == 0) {
358 	*pp += len;
359 	return true;
360     }
361     return false;
362 }
363 
364 static void
365 parse_indent_comment(void)
366 {
367     bool on;
368 
369     const char *p = inp.buf;
370 
371     skip_blank(&p);
372     if (!skip_string(&p, "/*"))
373 	return;
374     skip_blank(&p);
375     if (!skip_string(&p, "INDENT"))
376 	return;
377     skip_blank(&p);
378 
379     if (*p == '*' || skip_string(&p, "ON"))
380 	on = true;
381     else if (skip_string(&p, "OFF"))
382 	on = false;
383     else
384 	return;
385 
386     skip_blank(&p);
387     if (!skip_string(&p, "*/\n"))
388 	return;
389 
390     if (com.s != com.e || lab.s != lab.e || code.s != code.e)
391 	dump_line();
392 
393     inhibit_formatting = !on;
394     if (on) {
395 	blank_lines_to_output = 0;
396 	blank_line_after = false;
397 	blank_line_before = false;
398 	suppress_blanklines = true;
399     }
400 }
401 
402 /*
403  * Copyright (C) 1976 by the Board of Trustees of the University of Illinois
404  *
405  * All rights reserved
406  */
407 void
408 inbuf_read_line(void)
409 {
410     char *p;
411     int ch;
412     FILE *f = input;
413 
414     if (saved_inp_s != NULL) {	/* there is a partly filled input buffer left */
415 	inp.s = saved_inp_s;	/* do not read anything, just switch buffers */
416 	inp.e = saved_inp_e;
417 	saved_inp_s = saved_inp_e = NULL;
418 	debug_println("switched inp.s back to saved_inp_s");
419 	if (inp.s < inp.e)
420 	    return;		/* only return if there is really something in
421 				 * this buffer */
422     }
423 
424     for (p = inp.buf;;) {
425 	if (p >= inp.l) {
426 	    size_t size = (size_t)(inp.l - inp.buf) * 2 + 10;
427 	    size_t offset = (size_t)(p - inp.buf);
428 	    inp.buf = xrealloc(inp.buf, size);
429 	    p = inp.buf + offset;
430 	    inp.l = inp.buf + size - 2;
431 	}
432 
433 	if ((ch = getc(f)) == EOF) {
434 	    if (!inhibit_formatting) {
435 		*p++ = ' ';
436 		*p++ = '\n';
437 	    }
438 	    had_eof = true;
439 	    break;
440 	}
441 
442 	if (ch != '\0')
443 	    *p++ = (char)ch;
444 	if (ch == '\n')
445 	    break;
446     }
447 
448     inp.s = inp.buf;
449     inp.e = p;
450 
451     if (p - inp.s >= 3 && p[-3] == '*' && p[-2] == '/') {
452 	if (strncmp(inp.s, "/**INDENT**", 11) == 0)
453 	    inbuf_read_line();	/* flush indent error message */
454 	else
455 	    parse_indent_comment();
456     }
457 
458     if (inhibit_formatting)
459 	output_range(inp.s, inp.e);
460 }
461 
462 int
463 indentation_after_range(int ind, const char *start, const char *end)
464 {
465     for (const char *p = start; *p != '\0' && p != end; ++p) {
466 	if (*p == '\n' || *p == '\f')
467 	    ind = 0;
468 	else if (*p == '\t')
469 	    ind = next_tab(ind);
470 	else if (*p == '\b')
471 	    --ind;
472 	else
473 	    ++ind;
474     }
475     return ind;
476 }
477 
478 int
479 indentation_after(int ind, const char *s)
480 {
481     return indentation_after_range(ind, s, NULL);
482 }
483