xref: /freebsd/usr.bin/ul/ul.c (revision 0b8224d1cc9dc6c9778ba04a75b2c8d47e5d7481)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
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 #include <err.h>
33 #include <locale.h>
34 #include <stdio.h>
35 #include <stdlib.h>
36 #include <string.h>
37 #include <termcap.h>
38 #include <unistd.h>
39 #include <wchar.h>
40 #include <wctype.h>
41 
42 #define	IESC	'\033'
43 #define	SO	'\016'
44 #define	SI	'\017'
45 #define	HFWD	'9'
46 #define	HREV	'8'
47 #define	FREV	'7'
48 #define	MAXBUF	512
49 
50 #define	NORMAL	000
51 #define	ALTSET	001	/* Reverse */
52 #define	SUPERSC	002	/* Dim */
53 #define	SUBSC	004	/* Dim | Ul */
54 #define	UNDERL	010	/* Ul */
55 #define	BOLD	020	/* Bold */
56 
57 static int	must_use_uc, must_overstrike;
58 static const char
59 	*CURS_UP, *CURS_RIGHT, *CURS_LEFT,
60 	*ENTER_STANDOUT, *EXIT_STANDOUT, *ENTER_UNDERLINE, *EXIT_UNDERLINE,
61 	*ENTER_DIM, *ENTER_BOLD, *ENTER_REVERSE, *UNDER_CHAR, *EXIT_ATTRIBUTES;
62 
63 struct	CHAR	{
64 	char	c_mode;
65 	wchar_t	c_char;
66 	int	c_width;	/* width or -1 if multi-column char. filler */
67 } ;
68 
69 static struct	CHAR	sobuf[MAXBUF]; /* static output buffer */
70 static struct	CHAR	*obuf = sobuf;
71 static int	buflen = MAXBUF;
72 static int	col, maxcol;
73 static int	mode;
74 static int	halfpos;
75 static int	upln;
76 static int	iflag;
77 
78 static void usage(void) __dead2;
79 static void setnewmode(int);
80 static void initcap(void);
81 static void reverse(void);
82 static int outchar(int);
83 static void fwd(void);
84 static void initbuf(void);
85 static void iattr(void);
86 static void overstrike(void);
87 static void flushln(void);
88 static void filter(FILE *);
89 static void outc(wint_t, int);
90 
91 #define	PRINT(s)	if (s == NULL) /* void */; else tputs(s, 1, outchar)
92 
93 int
main(int argc,char ** argv)94 main(int argc, char **argv)
95 {
96 	int c;
97 	const char *termtype;
98 	FILE *f;
99 	char termcap[1024];
100 
101 	setlocale(LC_ALL, "");
102 
103 	termtype = getenv("TERM");
104 	if (termtype == NULL || (argv[0][0] == 'c' && !isatty(1)))
105 		termtype = "lpr";
106 	while ((c = getopt(argc, argv, "it:T:")) != -1)
107 		switch (c) {
108 		case 't':
109 		case 'T': /* for nroff compatibility */
110 			termtype = optarg;
111 			break;
112 		case 'i':
113 			iflag = 1;
114 			break;
115 		default:
116 			usage();
117 		}
118 
119 	switch (tgetent(termcap, termtype)) {
120 	case 1:
121 		break;
122 	default:
123 		warnx("trouble reading termcap");
124 		/* FALLTHROUGH */
125 	case 0:
126 		/* No such terminal type - assume dumb */
127 		(void)strcpy(termcap, "dumb:os:col#80:cr=^M:sf=^J:am:");
128 		break;
129 	}
130 	initcap();
131 	if ((tgetflag("os") && ENTER_BOLD == NULL ) ||
132 	    (tgetflag("ul") && ENTER_UNDERLINE == NULL && UNDER_CHAR == NULL))
133 		must_overstrike = 1;
134 	initbuf();
135 	if (optind == argc)
136 		filter(stdin);
137 	else for (; optind<argc; optind++) {
138 		f = fopen(argv[optind],"r");
139 		if (f == NULL)
140 			err(1, "%s", argv[optind]);
141 		else
142 			filter(f);
143 	}
144 	if (obuf != sobuf) {
145 		free(obuf);
146 	}
147 	exit(0);
148 }
149 
150 static void
usage(void)151 usage(void)
152 {
153 	fprintf(stderr, "usage: ul [-i] [-t terminal] [file ...]\n");
154 	exit(1);
155 }
156 
157 static void
filter(FILE * f)158 filter(FILE *f)
159 {
160 	wint_t c;
161 	int i, w;
162 	int copy;
163 
164 	copy = 0;
165 
166 	while ((c = getwc(f)) != WEOF) {
167 		if (col == buflen) {
168 			if (obuf == sobuf) {
169 				obuf = NULL;
170 				copy = 1;
171 			}
172 			obuf = realloc(obuf, sizeof(*obuf) * 2 * buflen);
173 			if (obuf == NULL) {
174 				obuf = sobuf;
175 				break;
176 			} else if (copy) {
177 				memcpy(obuf, sobuf, sizeof(*obuf) * buflen);
178 				copy = 0;
179 			}
180 			bzero((char *)(obuf + buflen), sizeof(*obuf) * buflen);
181 			buflen *= 2;
182 		}
183 		switch(c) {
184 		case '\b':
185 			if (col > 0)
186 				col--;
187 			continue;
188 
189 		case '\t':
190 			col = (col+8) & ~07;
191 			if (col > maxcol)
192 				maxcol = col;
193 			continue;
194 
195 		case '\r':
196 			col = 0;
197 			continue;
198 
199 		case SO:
200 			mode |= ALTSET;
201 			continue;
202 
203 		case SI:
204 			mode &= ~ALTSET;
205 			continue;
206 
207 		case IESC:
208 			switch (c = getwc(f)) {
209 
210 			case HREV:
211 				if (halfpos == 0) {
212 					mode |= SUPERSC;
213 					halfpos--;
214 				} else if (halfpos > 0) {
215 					mode &= ~SUBSC;
216 					halfpos--;
217 				} else {
218 					halfpos = 0;
219 					reverse();
220 				}
221 				continue;
222 
223 			case HFWD:
224 				if (halfpos == 0) {
225 					mode |= SUBSC;
226 					halfpos++;
227 				} else if (halfpos < 0) {
228 					mode &= ~SUPERSC;
229 					halfpos++;
230 				} else {
231 					halfpos = 0;
232 					fwd();
233 				}
234 				continue;
235 
236 			case FREV:
237 				reverse();
238 				continue;
239 
240 			default:
241 				errx(1, "unknown escape sequence in input: %o, %o", IESC, c);
242 			}
243 			continue;
244 
245 		case '_':
246 			if (obuf[col].c_char || obuf[col].c_width < 0) {
247 				while (col > 0 && obuf[col].c_width < 0)
248 					col--;
249 				w = obuf[col].c_width;
250 				for (i = 0; i < w; i++)
251 					obuf[col++].c_mode |= UNDERL | mode;
252 				if (col > maxcol)
253 					maxcol = col;
254 				continue;
255 			}
256 			obuf[col].c_char = '_';
257 			obuf[col].c_width = 1;
258 			/* FALLTHROUGH */
259 		case ' ':
260 			col++;
261 			if (col > maxcol)
262 				maxcol = col;
263 			continue;
264 
265 		case '\n':
266 			flushln();
267 			continue;
268 
269 		case '\f':
270 			flushln();
271 			putwchar('\f');
272 			continue;
273 
274 		default:
275 			if ((w = wcwidth(c)) <= 0)	/* non printing */
276 				continue;
277 			if (obuf[col].c_char == '\0') {
278 				obuf[col].c_char = c;
279 				for (i = 0; i < w; i++)
280 					obuf[col + i].c_mode = mode;
281 				obuf[col].c_width = w;
282 				for (i = 1; i < w; i++)
283 					obuf[col + i].c_width = -1;
284 			} else if (obuf[col].c_char == '_') {
285 				obuf[col].c_char = c;
286 				for (i = 0; i < w; i++)
287 					obuf[col + i].c_mode |= UNDERL|mode;
288 				obuf[col].c_width = w;
289 				for (i = 1; i < w; i++)
290 					obuf[col + i].c_width = -1;
291 			} else if ((wint_t)obuf[col].c_char == c) {
292 				for (i = 0; i < w; i++)
293 					obuf[col + i].c_mode |= BOLD|mode;
294 			} else {
295 				w = obuf[col].c_width;
296 				for (i = 0; i < w; i++)
297 					obuf[col + i].c_mode = mode;
298 			}
299 			col += w;
300 			if (col > maxcol)
301 				maxcol = col;
302 			continue;
303 		}
304 	}
305 	if (ferror(f))
306 		err(1, NULL);
307 	if (maxcol)
308 		flushln();
309 }
310 
311 static void
flushln(void)312 flushln(void)
313 {
314 	int lastmode;
315 	int i;
316 	int hadmodes = 0;
317 
318 	lastmode = NORMAL;
319 	for (i = 0; i < maxcol; i++) {
320 		if (obuf[i].c_mode != lastmode) {
321 			hadmodes++;
322 			setnewmode(obuf[i].c_mode);
323 			lastmode = obuf[i].c_mode;
324 		}
325 		if (obuf[i].c_char == '\0') {
326 			if (upln)
327 				PRINT(CURS_RIGHT);
328 			else
329 				outc(' ', 1);
330 		} else
331 			outc(obuf[i].c_char, obuf[i].c_width);
332 		if (obuf[i].c_width > 1)
333 			i += obuf[i].c_width - 1;
334 	}
335 	if (lastmode != NORMAL) {
336 		setnewmode(0);
337 	}
338 	if (must_overstrike && hadmodes)
339 		overstrike();
340 	putwchar('\n');
341 	if (iflag && hadmodes)
342 		iattr();
343 	(void)fflush(stdout);
344 	if (upln)
345 		upln--;
346 	initbuf();
347 }
348 
349 /*
350  * For terminals that can overstrike, overstrike underlines and bolds.
351  * We don't do anything with halfline ups and downs, or Greek.
352  */
353 static void
overstrike(void)354 overstrike(void)
355 {
356 	int i;
357 	wchar_t lbuf[256];
358 	wchar_t *cp = lbuf;
359 	int hadbold=0;
360 
361 	/* Set up overstrike buffer */
362 	for (i=0; i<maxcol; i++)
363 		switch (obuf[i].c_mode) {
364 		case NORMAL:
365 		default:
366 			*cp++ = ' ';
367 			break;
368 		case UNDERL:
369 			*cp++ = '_';
370 			break;
371 		case BOLD:
372 			*cp++ = obuf[i].c_char;
373 			if (obuf[i].c_width > 1)
374 				i += obuf[i].c_width - 1;
375 			hadbold=1;
376 			break;
377 		}
378 	putwchar('\r');
379 	for (*cp=' '; *cp==' '; cp--)
380 		*cp = 0;
381 	for (cp=lbuf; *cp; cp++)
382 		putwchar(*cp);
383 	if (hadbold) {
384 		putwchar('\r');
385 		for (cp=lbuf; *cp; cp++)
386 			putwchar(*cp=='_' ? ' ' : *cp);
387 		putwchar('\r');
388 		for (cp=lbuf; *cp; cp++)
389 			putwchar(*cp=='_' ? ' ' : *cp);
390 	}
391 }
392 
393 static void
iattr(void)394 iattr(void)
395 {
396 	int i;
397 	wchar_t lbuf[256];
398 	wchar_t *cp = lbuf;
399 
400 	for (i=0; i<maxcol; i++)
401 		switch (obuf[i].c_mode) {
402 		case NORMAL:	*cp++ = ' '; break;
403 		case ALTSET:	*cp++ = 'g'; break;
404 		case SUPERSC:	*cp++ = '^'; break;
405 		case SUBSC:	*cp++ = 'v'; break;
406 		case UNDERL:	*cp++ = '_'; break;
407 		case BOLD:	*cp++ = '!'; break;
408 		default:	*cp++ = 'X'; break;
409 		}
410 	for (*cp=' '; *cp==' '; cp--)
411 		*cp = 0;
412 	for (cp=lbuf; *cp; cp++)
413 		putwchar(*cp);
414 	putwchar('\n');
415 }
416 
417 static void
initbuf(void)418 initbuf(void)
419 {
420 
421 	bzero((char *)obuf, buflen * sizeof(*obuf)); /* depends on NORMAL == 0 */
422 	col = 0;
423 	maxcol = 0;
424 	mode &= ALTSET;
425 }
426 
427 static void
fwd(void)428 fwd(void)
429 {
430 	int oldcol, oldmax;
431 
432 	oldcol = col;
433 	oldmax = maxcol;
434 	flushln();
435 	col = oldcol;
436 	maxcol = oldmax;
437 }
438 
439 static void
reverse(void)440 reverse(void)
441 {
442 	upln++;
443 	fwd();
444 	PRINT(CURS_UP);
445 	PRINT(CURS_UP);
446 	upln++;
447 }
448 
449 static void
initcap(void)450 initcap(void)
451 {
452 	static char tcapbuf[512];
453 	char *bp = tcapbuf;
454 
455 	/* This nonsense attempts to work with both old and new termcap */
456 	CURS_UP =		tgetstr("up", &bp);
457 	CURS_RIGHT =		tgetstr("ri", &bp);
458 	if (CURS_RIGHT == NULL)
459 		CURS_RIGHT =	tgetstr("nd", &bp);
460 	CURS_LEFT =		tgetstr("le", &bp);
461 	if (CURS_LEFT == NULL)
462 		CURS_LEFT =	tgetstr("bc", &bp);
463 	if (CURS_LEFT == NULL && tgetflag("bs"))
464 		CURS_LEFT =	"\b";
465 
466 	ENTER_STANDOUT =	tgetstr("so", &bp);
467 	EXIT_STANDOUT =		tgetstr("se", &bp);
468 	ENTER_UNDERLINE =	tgetstr("us", &bp);
469 	EXIT_UNDERLINE =	tgetstr("ue", &bp);
470 	ENTER_DIM =		tgetstr("mh", &bp);
471 	ENTER_BOLD =		tgetstr("md", &bp);
472 	ENTER_REVERSE =		tgetstr("mr", &bp);
473 	EXIT_ATTRIBUTES =	tgetstr("me", &bp);
474 
475 	if (!ENTER_BOLD && ENTER_REVERSE)
476 		ENTER_BOLD = ENTER_REVERSE;
477 	if (!ENTER_BOLD && ENTER_STANDOUT)
478 		ENTER_BOLD = ENTER_STANDOUT;
479 	if (!ENTER_UNDERLINE && ENTER_STANDOUT) {
480 		ENTER_UNDERLINE = ENTER_STANDOUT;
481 		EXIT_UNDERLINE = EXIT_STANDOUT;
482 	}
483 	if (!ENTER_DIM && ENTER_STANDOUT)
484 		ENTER_DIM = ENTER_STANDOUT;
485 	if (!ENTER_REVERSE && ENTER_STANDOUT)
486 		ENTER_REVERSE = ENTER_STANDOUT;
487 	if (!EXIT_ATTRIBUTES && EXIT_STANDOUT)
488 		EXIT_ATTRIBUTES = EXIT_STANDOUT;
489 
490 	/*
491 	 * Note that we use REVERSE for the alternate character set,
492 	 * not the as/ae capabilities.  This is because we are modelling
493 	 * the model 37 teletype (since that's what nroff outputs) and
494 	 * the typical as/ae is more of a graphics set, not the greek
495 	 * letters the 37 has.
496 	 */
497 
498 	UNDER_CHAR =		tgetstr("uc", &bp);
499 	must_use_uc = (UNDER_CHAR && !ENTER_UNDERLINE);
500 }
501 
502 static int
outchar(int c)503 outchar(int c)
504 {
505 	return (putwchar(c) != WEOF ? c : EOF);
506 }
507 
508 static int curmode = 0;
509 
510 static void
outc(wint_t c,int width)511 outc(wint_t c, int width)
512 {
513 	int i;
514 
515 	putwchar(c);
516 	if (must_use_uc && (curmode&UNDERL)) {
517 		for (i = 0; i < width; i++)
518 			PRINT(CURS_LEFT);
519 		for (i = 0; i < width; i++)
520 			PRINT(UNDER_CHAR);
521 	}
522 }
523 
524 static void
setnewmode(int newmode)525 setnewmode(int newmode)
526 {
527 	if (!iflag) {
528 		if (curmode != NORMAL && newmode != NORMAL)
529 			setnewmode(NORMAL);
530 		switch (newmode) {
531 		case NORMAL:
532 			switch(curmode) {
533 			case NORMAL:
534 				break;
535 			case UNDERL:
536 				PRINT(EXIT_UNDERLINE);
537 				break;
538 			default:
539 				/* This includes standout */
540 				PRINT(EXIT_ATTRIBUTES);
541 				break;
542 			}
543 			break;
544 		case ALTSET:
545 			PRINT(ENTER_REVERSE);
546 			break;
547 		case SUPERSC:
548 			/*
549 			 * This only works on a few terminals.
550 			 * It should be fixed.
551 			 */
552 			PRINT(ENTER_UNDERLINE);
553 			PRINT(ENTER_DIM);
554 			break;
555 		case SUBSC:
556 			PRINT(ENTER_DIM);
557 			break;
558 		case UNDERL:
559 			PRINT(ENTER_UNDERLINE);
560 			break;
561 		case BOLD:
562 			PRINT(ENTER_BOLD);
563 			break;
564 		default:
565 			/*
566 			 * We should have some provision here for multiple modes
567 			 * on at once.  This will have to come later.
568 			 */
569 			PRINT(ENTER_STANDOUT);
570 			break;
571 		}
572 	}
573 	curmode = newmode;
574 }
575