xref: /freebsd/usr.bin/ul/ul.c (revision fed1ca4b719c56c930f2259d80663cd34be812bb)
1 /*
2  * Copyright (c) 1980, 1993
3  *	The Regents of the University of California.  All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
8  * 1. Redistributions of source code must retain the above copyright
9  *    notice, this list of conditions and the following disclaimer.
10  * 2. Redistributions in binary form must reproduce the above copyright
11  *    notice, this list of conditions and the following disclaimer in the
12  *    documentation and/or other materials provided with the distribution.
13  * 3. Neither the name of the University nor the names of its contributors
14  *    may be used to endorse or promote products derived from this software
15  *    without specific prior written permission.
16  *
17  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27  * SUCH DAMAGE.
28  */
29 
30 #ifndef lint
31 static const char copyright[] =
32 "@(#) Copyright (c) 1980, 1993\n\
33 	The Regents of the University of California.  All rights reserved.\n";
34 #endif /* not lint */
35 
36 #ifndef lint
37 #if 0
38 static char sccsid[] = "@(#)ul.c	8.1 (Berkeley) 6/6/93";
39 #endif
40 static const char rcsid[] =
41   "$FreeBSD$";
42 #endif /* not lint */
43 
44 #include <err.h>
45 #include <locale.h>
46 #include <stdio.h>
47 #include <stdlib.h>
48 #include <string.h>
49 #include <termcap.h>
50 #include <unistd.h>
51 #include <wchar.h>
52 #include <wctype.h>
53 
54 #define	IESC	'\033'
55 #define	SO	'\016'
56 #define	SI	'\017'
57 #define	HFWD	'9'
58 #define	HREV	'8'
59 #define	FREV	'7'
60 #define	MAXBUF	512
61 
62 #define	NORMAL	000
63 #define	ALTSET	001	/* Reverse */
64 #define	SUPERSC	002	/* Dim */
65 #define	SUBSC	004	/* Dim | Ul */
66 #define	UNDERL	010	/* Ul */
67 #define	BOLD	020	/* Bold */
68 
69 static int	must_use_uc, must_overstrike;
70 static const char
71 	*CURS_UP, *CURS_RIGHT, *CURS_LEFT,
72 	*ENTER_STANDOUT, *EXIT_STANDOUT, *ENTER_UNDERLINE, *EXIT_UNDERLINE,
73 	*ENTER_DIM, *ENTER_BOLD, *ENTER_REVERSE, *UNDER_CHAR, *EXIT_ATTRIBUTES;
74 
75 struct	CHAR	{
76 	char	c_mode;
77 	wchar_t	c_char;
78 	int	c_width;	/* width or -1 if multi-column char. filler */
79 } ;
80 
81 static struct	CHAR	obuf[MAXBUF];
82 static int	col, maxcol;
83 static int	mode;
84 static int	halfpos;
85 static int	upln;
86 static int	iflag;
87 
88 static void usage(void);
89 static void setnewmode(int);
90 static void initcap(void);
91 static void reverse(void);
92 static int outchar(int);
93 static void fwd(void);
94 static void initbuf(void);
95 static void iattr(void);
96 static void overstrike(void);
97 static void flushln(void);
98 static void filter(FILE *);
99 static void outc(wint_t, int);
100 
101 #define	PRINT(s)	if (s == NULL) /* void */; else tputs(s, 1, outchar)
102 
103 int
104 main(int argc, char **argv)
105 {
106 	int c;
107 	const char *termtype;
108 	FILE *f;
109 	char termcap[1024];
110 
111 	setlocale(LC_ALL, "");
112 
113 	termtype = getenv("TERM");
114 	if (termtype == NULL || (argv[0][0] == 'c' && !isatty(1)))
115 		termtype = "lpr";
116 	while ((c = getopt(argc, argv, "it:T:")) != -1)
117 		switch (c) {
118 		case 't':
119 		case 'T': /* for nroff compatibility */
120 			termtype = optarg;
121 			break;
122 		case 'i':
123 			iflag = 1;
124 			break;
125 		default:
126 			usage();
127 		}
128 
129 	switch (tgetent(termcap, termtype)) {
130 	case 1:
131 		break;
132 	default:
133 		warnx("trouble reading termcap");
134 		/* FALLTHROUGH */
135 	case 0:
136 		/* No such terminal type - assume dumb */
137 		(void)strcpy(termcap, "dumb:os:col#80:cr=^M:sf=^J:am:");
138 		break;
139 	}
140 	initcap();
141 	if ((tgetflag("os") && ENTER_BOLD == NULL ) ||
142 	    (tgetflag("ul") && ENTER_UNDERLINE == NULL && UNDER_CHAR == NULL))
143 		must_overstrike = 1;
144 	initbuf();
145 	if (optind == argc)
146 		filter(stdin);
147 	else for (; optind<argc; optind++) {
148 		f = fopen(argv[optind],"r");
149 		if (f == NULL)
150 			err(1, "%s", argv[optind]);
151 		else
152 			filter(f);
153 	}
154 	exit(0);
155 }
156 
157 static void
158 usage(void)
159 {
160 	fprintf(stderr, "usage: ul [-i] [-t terminal] [file ...]\n");
161 	exit(1);
162 }
163 
164 static void
165 filter(FILE *f)
166 {
167 	wint_t c;
168 	int i, w;
169 
170 	while ((c = getwc(f)) != WEOF && col < MAXBUF) switch(c) {
171 
172 	case '\b':
173 		if (col > 0)
174 			col--;
175 		continue;
176 
177 	case '\t':
178 		col = (col+8) & ~07;
179 		if (col > maxcol)
180 			maxcol = col;
181 		continue;
182 
183 	case '\r':
184 		col = 0;
185 		continue;
186 
187 	case SO:
188 		mode |= ALTSET;
189 		continue;
190 
191 	case SI:
192 		mode &= ~ALTSET;
193 		continue;
194 
195 	case IESC:
196 		switch (c = getwc(f)) {
197 
198 		case HREV:
199 			if (halfpos == 0) {
200 				mode |= SUPERSC;
201 				halfpos--;
202 			} else if (halfpos > 0) {
203 				mode &= ~SUBSC;
204 				halfpos--;
205 			} else {
206 				halfpos = 0;
207 				reverse();
208 			}
209 			continue;
210 
211 		case HFWD:
212 			if (halfpos == 0) {
213 				mode |= SUBSC;
214 				halfpos++;
215 			} else if (halfpos < 0) {
216 				mode &= ~SUPERSC;
217 				halfpos++;
218 			} else {
219 				halfpos = 0;
220 				fwd();
221 			}
222 			continue;
223 
224 		case FREV:
225 			reverse();
226 			continue;
227 
228 		default:
229 			errx(1, "unknown escape sequence in input: %o, %o", IESC, c);
230 		}
231 		continue;
232 
233 	case '_':
234 		if (obuf[col].c_char || obuf[col].c_width < 0) {
235 			while (col > 0 && obuf[col].c_width < 0)
236 				col--;
237 			w = obuf[col].c_width;
238 			for (i = 0; i < w; i++)
239 				obuf[col++].c_mode |= UNDERL | mode;
240 			if (col > maxcol)
241 				maxcol = col;
242 			continue;
243 		}
244 		obuf[col].c_char = '_';
245 		obuf[col].c_width = 1;
246 		/* FALLTHROUGH */
247 	case ' ':
248 		col++;
249 		if (col > maxcol)
250 			maxcol = col;
251 		continue;
252 
253 	case '\n':
254 		flushln();
255 		continue;
256 
257 	case '\f':
258 		flushln();
259 		putwchar('\f');
260 		continue;
261 
262 	default:
263 		if ((w = wcwidth(c)) <= 0)	/* non printing */
264 			continue;
265 		if (obuf[col].c_char == '\0') {
266 			obuf[col].c_char = c;
267 			for (i = 0; i < w; i++)
268 				obuf[col + i].c_mode = mode;
269 			obuf[col].c_width = w;
270 			for (i = 1; i < w; i++)
271 				obuf[col + i].c_width = -1;
272 		} else if (obuf[col].c_char == '_') {
273 			obuf[col].c_char = c;
274 			for (i = 0; i < w; i++)
275 				obuf[col + i].c_mode |= UNDERL|mode;
276 			obuf[col].c_width = w;
277 			for (i = 1; i < w; i++)
278 				obuf[col + i].c_width = -1;
279 		} else if ((wint_t)obuf[col].c_char == c) {
280 			for (i = 0; i < w; i++)
281 				obuf[col + i].c_mode |= BOLD|mode;
282 		} else {
283 			w = obuf[col].c_width;
284 			for (i = 0; i < w; i++)
285 				obuf[col + i].c_mode = mode;
286 		}
287 		col += w;
288 		if (col > maxcol)
289 			maxcol = col;
290 		continue;
291 	}
292 	if (ferror(f))
293 		err(1, NULL);
294 	if (maxcol)
295 		flushln();
296 }
297 
298 static void
299 flushln(void)
300 {
301 	int lastmode;
302 	int i;
303 	int hadmodes = 0;
304 
305 	lastmode = NORMAL;
306 	for (i = 0; i < maxcol; i++) {
307 		if (obuf[i].c_mode != lastmode) {
308 			hadmodes++;
309 			setnewmode(obuf[i].c_mode);
310 			lastmode = obuf[i].c_mode;
311 		}
312 		if (obuf[i].c_char == '\0') {
313 			if (upln)
314 				PRINT(CURS_RIGHT);
315 			else
316 				outc(' ', 1);
317 		} else
318 			outc(obuf[i].c_char, obuf[i].c_width);
319 		if (obuf[i].c_width > 1)
320 			i += obuf[i].c_width - 1;
321 	}
322 	if (lastmode != NORMAL) {
323 		setnewmode(0);
324 	}
325 	if (must_overstrike && hadmodes)
326 		overstrike();
327 	putwchar('\n');
328 	if (iflag && hadmodes)
329 		iattr();
330 	(void)fflush(stdout);
331 	if (upln)
332 		upln--;
333 	initbuf();
334 }
335 
336 /*
337  * For terminals that can overstrike, overstrike underlines and bolds.
338  * We don't do anything with halfline ups and downs, or Greek.
339  */
340 static void
341 overstrike(void)
342 {
343 	int i;
344 	wchar_t lbuf[256];
345 	wchar_t *cp = lbuf;
346 	int hadbold=0;
347 
348 	/* Set up overstrike buffer */
349 	for (i=0; i<maxcol; i++)
350 		switch (obuf[i].c_mode) {
351 		case NORMAL:
352 		default:
353 			*cp++ = ' ';
354 			break;
355 		case UNDERL:
356 			*cp++ = '_';
357 			break;
358 		case BOLD:
359 			*cp++ = obuf[i].c_char;
360 			if (obuf[i].c_width > 1)
361 				i += obuf[i].c_width - 1;
362 			hadbold=1;
363 			break;
364 		}
365 	putwchar('\r');
366 	for (*cp=' '; *cp==' '; cp--)
367 		*cp = 0;
368 	for (cp=lbuf; *cp; cp++)
369 		putwchar(*cp);
370 	if (hadbold) {
371 		putwchar('\r');
372 		for (cp=lbuf; *cp; cp++)
373 			putwchar(*cp=='_' ? ' ' : *cp);
374 		putwchar('\r');
375 		for (cp=lbuf; *cp; cp++)
376 			putwchar(*cp=='_' ? ' ' : *cp);
377 	}
378 }
379 
380 static void
381 iattr(void)
382 {
383 	int i;
384 	wchar_t lbuf[256];
385 	wchar_t *cp = lbuf;
386 
387 	for (i=0; i<maxcol; i++)
388 		switch (obuf[i].c_mode) {
389 		case NORMAL:	*cp++ = ' '; break;
390 		case ALTSET:	*cp++ = 'g'; break;
391 		case SUPERSC:	*cp++ = '^'; break;
392 		case SUBSC:	*cp++ = 'v'; break;
393 		case UNDERL:	*cp++ = '_'; break;
394 		case BOLD:	*cp++ = '!'; break;
395 		default:	*cp++ = 'X'; break;
396 		}
397 	for (*cp=' '; *cp==' '; cp--)
398 		*cp = 0;
399 	for (cp=lbuf; *cp; cp++)
400 		putwchar(*cp);
401 	putwchar('\n');
402 }
403 
404 static void
405 initbuf(void)
406 {
407 
408 	bzero((char *)obuf, sizeof (obuf));	/* depends on NORMAL == 0 */
409 	col = 0;
410 	maxcol = 0;
411 	mode &= ALTSET;
412 }
413 
414 static void
415 fwd(void)
416 {
417 	int oldcol, oldmax;
418 
419 	oldcol = col;
420 	oldmax = maxcol;
421 	flushln();
422 	col = oldcol;
423 	maxcol = oldmax;
424 }
425 
426 static void
427 reverse(void)
428 {
429 	upln++;
430 	fwd();
431 	PRINT(CURS_UP);
432 	PRINT(CURS_UP);
433 	upln++;
434 }
435 
436 static void
437 initcap(void)
438 {
439 	static char tcapbuf[512];
440 	char *bp = tcapbuf;
441 
442 	/* This nonsense attempts to work with both old and new termcap */
443 	CURS_UP =		tgetstr("up", &bp);
444 	CURS_RIGHT =		tgetstr("ri", &bp);
445 	if (CURS_RIGHT == NULL)
446 		CURS_RIGHT =	tgetstr("nd", &bp);
447 	CURS_LEFT =		tgetstr("le", &bp);
448 	if (CURS_LEFT == NULL)
449 		CURS_LEFT =	tgetstr("bc", &bp);
450 	if (CURS_LEFT == NULL && tgetflag("bs"))
451 		CURS_LEFT =	"\b";
452 
453 	ENTER_STANDOUT =	tgetstr("so", &bp);
454 	EXIT_STANDOUT =		tgetstr("se", &bp);
455 	ENTER_UNDERLINE =	tgetstr("us", &bp);
456 	EXIT_UNDERLINE =	tgetstr("ue", &bp);
457 	ENTER_DIM =		tgetstr("mh", &bp);
458 	ENTER_BOLD =		tgetstr("md", &bp);
459 	ENTER_REVERSE =		tgetstr("mr", &bp);
460 	EXIT_ATTRIBUTES =	tgetstr("me", &bp);
461 
462 	if (!ENTER_BOLD && ENTER_REVERSE)
463 		ENTER_BOLD = ENTER_REVERSE;
464 	if (!ENTER_BOLD && ENTER_STANDOUT)
465 		ENTER_BOLD = ENTER_STANDOUT;
466 	if (!ENTER_UNDERLINE && ENTER_STANDOUT) {
467 		ENTER_UNDERLINE = ENTER_STANDOUT;
468 		EXIT_UNDERLINE = EXIT_STANDOUT;
469 	}
470 	if (!ENTER_DIM && ENTER_STANDOUT)
471 		ENTER_DIM = ENTER_STANDOUT;
472 	if (!ENTER_REVERSE && ENTER_STANDOUT)
473 		ENTER_REVERSE = ENTER_STANDOUT;
474 	if (!EXIT_ATTRIBUTES && EXIT_STANDOUT)
475 		EXIT_ATTRIBUTES = EXIT_STANDOUT;
476 
477 	/*
478 	 * Note that we use REVERSE for the alternate character set,
479 	 * not the as/ae capabilities.  This is because we are modelling
480 	 * the model 37 teletype (since that's what nroff outputs) and
481 	 * the typical as/ae is more of a graphics set, not the greek
482 	 * letters the 37 has.
483 	 */
484 
485 	UNDER_CHAR =		tgetstr("uc", &bp);
486 	must_use_uc = (UNDER_CHAR && !ENTER_UNDERLINE);
487 }
488 
489 static int
490 outchar(int c)
491 {
492 	return (putwchar(c) != WEOF ? c : EOF);
493 }
494 
495 static int curmode = 0;
496 
497 static void
498 outc(wint_t c, int width)
499 {
500 	int i;
501 
502 	putwchar(c);
503 	if (must_use_uc && (curmode&UNDERL)) {
504 		for (i = 0; i < width; i++)
505 			PRINT(CURS_LEFT);
506 		for (i = 0; i < width; i++)
507 			PRINT(UNDER_CHAR);
508 	}
509 }
510 
511 static void
512 setnewmode(int newmode)
513 {
514 	if (!iflag) {
515 		if (curmode != NORMAL && newmode != NORMAL)
516 			setnewmode(NORMAL);
517 		switch (newmode) {
518 		case NORMAL:
519 			switch(curmode) {
520 			case NORMAL:
521 				break;
522 			case UNDERL:
523 				PRINT(EXIT_UNDERLINE);
524 				break;
525 			default:
526 				/* This includes standout */
527 				PRINT(EXIT_ATTRIBUTES);
528 				break;
529 			}
530 			break;
531 		case ALTSET:
532 			PRINT(ENTER_REVERSE);
533 			break;
534 		case SUPERSC:
535 			/*
536 			 * This only works on a few terminals.
537 			 * It should be fixed.
538 			 */
539 			PRINT(ENTER_UNDERLINE);
540 			PRINT(ENTER_DIM);
541 			break;
542 		case SUBSC:
543 			PRINT(ENTER_DIM);
544 			break;
545 		case UNDERL:
546 			PRINT(ENTER_UNDERLINE);
547 			break;
548 		case BOLD:
549 			PRINT(ENTER_BOLD);
550 			break;
551 		default:
552 			/*
553 			 * We should have some provision here for multiple modes
554 			 * on at once.  This will have to come later.
555 			 */
556 			PRINT(ENTER_STANDOUT);
557 			break;
558 		}
559 	}
560 	curmode = newmode;
561 }
562