1/*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1990, 1993, 1994
5 *	The Regents of the University of California.  All rights reserved.
6 *
7 * This code is derived from software contributed to Berkeley by
8 * Michael Rendell of the Memorial University of Newfoundland.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 *    notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 *    notice, this list of conditions and the following disclaimer in the
17 *    documentation and/or other materials provided with the distribution.
18 * 3. Neither the name of the University nor the names of its contributors
19 *    may be used to endorse or promote products derived from this software
20 *    without specific prior written permission.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * SUCH DAMAGE.
33 */
34
35#include <sys/capsicum.h>
36
37#include <capsicum_helpers.h>
38#include <err.h>
39#include <errno.h>
40#include <locale.h>
41#include <stdio.h>
42#include <stdlib.h>
43#include <string.h>
44#include <termios.h>
45#include <unistd.h>
46#include <wchar.h>
47#include <wctype.h>
48
49#define	BS	'\b'		/* backspace */
50#define	TAB	'\t'		/* tab */
51#define	SPACE	' '		/* space */
52#define	NL	'\n'		/* newline */
53#define	CR	'\r'		/* carriage return */
54#define	ESC	'\033'		/* escape */
55#define	SI	'\017'		/* shift in to normal character set */
56#define	SO	'\016'		/* shift out to alternate character set */
57#define	VT	'\013'		/* vertical tab (aka reverse line feed) */
58#define	RLF	'7'		/* ESC-7 reverse line feed */
59#define	RHLF	'8'		/* ESC-8 reverse half-line feed */
60#define	FHLF	'9'		/* ESC-9 forward half-line feed */
61
62/* build up at least this many lines before flushing them out */
63#define	BUFFER_MARGIN		32
64
65typedef char CSET;
66
67typedef struct char_str {
68#define	CS_NORMAL	1
69#define	CS_ALTERNATE	2
70	short		c_column;	/* column character is in */
71	CSET		c_set;		/* character set (currently only 2) */
72	wchar_t		c_char;		/* character in question */
73	int		c_width;	/* character width */
74} CHAR;
75
76typedef struct line_str LINE;
77struct line_str {
78	CHAR	*l_line;		/* characters on the line */
79	LINE	*l_prev;		/* previous line */
80	LINE	*l_next;		/* next line */
81	int	l_lsize;		/* allocated sizeof l_line */
82	int	l_line_len;		/* strlen(l_line) */
83	int	l_needs_sort;		/* set if chars went in out of order */
84	int	l_max_col;		/* max column in the line */
85};
86
87static void	addto_lineno(int *, int);
88static LINE	*alloc_line(void);
89static void	dowarn(int);
90static void	flush_line(LINE *);
91static void	flush_lines(int);
92static void	flush_blanks(void);
93static void	free_line(LINE *);
94static void	usage(void);
95
96static CSET	last_set;		/* char_set of last char printed */
97static LINE	*lines;
98static int	compress_spaces;	/* if doing space -> tab conversion */
99static int	fine;			/* if `fine' resolution (half lines) */
100static int	max_bufd_lines;		/* max # of half lines to keep in memory */
101static int	nblank_lines;		/* # blanks after last flushed line */
102static int	no_backspaces;		/* if not to output any backspaces */
103static int	pass_unknown_seqs;	/* pass unknown control sequences */
104
105#define	PUTC(ch) \
106	do {					\
107		if (putwchar(ch) == WEOF)	\
108			errx(1, "write error");	\
109	} while (0)
110
111int
112main(int argc, char **argv)
113{
114	wint_t ch;
115	CHAR *c;
116	CSET cur_set;			/* current character set */
117	LINE *l;			/* current line */
118	int extra_lines;		/* # of lines above first line */
119	int cur_col;			/* current column */
120	int cur_line;			/* line number of current position */
121	int max_line;			/* max value of cur_line */
122	int this_line;			/* line l points to */
123	int nflushd_lines;		/* number of lines that were flushed */
124	int adjust, opt, warned, width;
125	const char *errstr;
126
127	(void)setlocale(LC_CTYPE, "");
128
129	if (caph_limit_stdio() == -1)
130		err(1, "unable to limit stdio");
131
132	if (caph_enter() < 0)
133		err(1, "unable to enter capability mode");
134
135	max_bufd_lines = 256;
136	compress_spaces = 1;		/* compress spaces into tabs */
137	while ((opt = getopt(argc, argv, "bfhl:px")) != -1)
138		switch (opt) {
139		case 'b':		/* do not output backspaces */
140			no_backspaces = 1;
141			break;
142		case 'f':		/* allow half forward line feeds */
143			fine = 1;
144			break;
145		case 'h':		/* compress spaces into tabs */
146			compress_spaces = 1;
147			break;
148		case 'l':		/* buffered line count */
149			max_bufd_lines = strtonum(optarg, 1,
150			    (INT_MAX - BUFFER_MARGIN) / 2, &errstr) * 2;
151			if (errstr != NULL)
152				errx(1, "bad -l argument, %s: %s", errstr,
153					optarg);
154			break;
155		case 'p':		/* pass unknown control sequences */
156			pass_unknown_seqs = 1;
157			break;
158		case 'x':		/* do not compress spaces into tabs */
159			compress_spaces = 0;
160			break;
161		case '?':
162		default:
163			usage();
164		}
165
166	if (optind != argc)
167		usage();
168
169	adjust = cur_col = extra_lines = warned = 0;
170	cur_line = max_line = nflushd_lines = this_line = 0;
171	cur_set = last_set = CS_NORMAL;
172	lines = l = alloc_line();
173
174	while ((ch = getwchar()) != WEOF) {
175		if (!iswgraph(ch)) {
176			switch (ch) {
177			case BS:		/* can't go back further */
178				if (cur_col == 0)
179					continue;
180				--cur_col;
181				continue;
182			case CR:
183				cur_col = 0;
184				continue;
185			case ESC:		/* just ignore EOF */
186				switch(getwchar()) {
187				/*
188				 * In the input stream, accept both the
189				 * XPG5 sequences ESC-digit and the
190				 * traditional BSD sequences ESC-ctrl.
191				 */
192				case '\007':
193					/* FALLTHROUGH */
194				case RLF:
195					addto_lineno(&cur_line, -2);
196					break;
197				case '\010':
198					/* FALLTHROUGH */
199				case RHLF:
200					addto_lineno(&cur_line, -1);
201					break;
202				case '\011':
203					/* FALLTHROUGH */
204				case FHLF:
205					addto_lineno(&cur_line, 1);
206					if (cur_line > max_line)
207						max_line = cur_line;
208				}
209				continue;
210			case NL:
211				addto_lineno(&cur_line, 2);
212				if (cur_line > max_line)
213					max_line = cur_line;
214				cur_col = 0;
215				continue;
216			case SPACE:
217				++cur_col;
218				continue;
219			case SI:
220				cur_set = CS_NORMAL;
221				continue;
222			case SO:
223				cur_set = CS_ALTERNATE;
224				continue;
225			case TAB:		/* adjust column */
226				cur_col |= 7;
227				++cur_col;
228				continue;
229			case VT:
230				addto_lineno(&cur_line, -2);
231				continue;
232			}
233			if (iswspace(ch)) {
234				if ((width = wcwidth(ch)) > 0)
235					cur_col += width;
236				continue;
237			}
238			if (!pass_unknown_seqs)
239				continue;
240		}
241
242		/* Must stuff ch in a line - are we at the right one? */
243		if (cur_line + adjust != this_line) {
244			LINE *lnew;
245
246			/* round up to next line */
247			adjust = !fine && (cur_line & 1);
248
249			if (cur_line + adjust < this_line) {
250				while (cur_line + adjust < this_line &&
251				    l->l_prev != NULL) {
252					l = l->l_prev;
253					this_line--;
254				}
255				if (cur_line + adjust < this_line) {
256					if (nflushd_lines == 0) {
257						/*
258						 * Allow backup past first
259						 * line if nothing has been
260						 * flushed yet.
261						 */
262						while (cur_line + adjust
263						    < this_line) {
264							lnew = alloc_line();
265							l->l_prev = lnew;
266							lnew->l_next = l;
267							l = lines = lnew;
268							extra_lines++;
269							this_line--;
270						}
271					} else {
272						if (!warned++)
273							dowarn(cur_line);
274						cur_line = this_line - adjust;
275					}
276				}
277			} else {
278				/* may need to allocate here */
279				while (cur_line + adjust > this_line) {
280					if (l->l_next == NULL) {
281						l->l_next = alloc_line();
282						l->l_next->l_prev = l;
283					}
284					l = l->l_next;
285					this_line++;
286				}
287			}
288			if (this_line > nflushd_lines &&
289			    this_line - nflushd_lines >=
290			    max_bufd_lines + BUFFER_MARGIN) {
291				if (extra_lines) {
292					flush_lines(extra_lines);
293					extra_lines = 0;
294				}
295				flush_lines(this_line - nflushd_lines -
296				    max_bufd_lines);
297				nflushd_lines = this_line - max_bufd_lines;
298			}
299		}
300		/* grow line's buffer? */
301		if (l->l_line_len + 1 >= l->l_lsize) {
302			int need;
303
304			need = l->l_lsize ? l->l_lsize * 2 : 90;
305			if ((l->l_line = realloc(l->l_line,
306			    (unsigned)need * sizeof(CHAR))) == NULL)
307				err(1, NULL);
308			l->l_lsize = need;
309		}
310		c = &l->l_line[l->l_line_len++];
311		c->c_char = ch;
312		c->c_set = cur_set;
313		c->c_column = cur_col;
314		c->c_width = wcwidth(ch);
315		/*
316		 * If things are put in out of order, they will need sorting
317		 * when it is flushed.
318		 */
319		if (cur_col < l->l_max_col)
320			l->l_needs_sort = 1;
321		else
322			l->l_max_col = cur_col;
323		if (c->c_width > 0)
324			cur_col += c->c_width;
325	}
326	if (ferror(stdin))
327		err(1, NULL);
328	if (extra_lines) {
329		/*
330		 * Extra lines only exist if no lines have been flushed
331		 * yet. This means that 'lines' must point to line zero
332		 * after we flush the extra lines.
333		 */
334		flush_lines(extra_lines);
335		l = lines;
336		this_line = 0;
337	}
338
339	/* goto the last line that had a character on it */
340	for (; l->l_next; l = l->l_next)
341		this_line++;
342	flush_lines(this_line - nflushd_lines + 1);
343
344	/* make sure we leave things in a sane state */
345	if (last_set != CS_NORMAL)
346		PUTC(SI);
347
348	/* flush out the last few blank lines */
349	if (max_line >= this_line)
350		nblank_lines = max_line - this_line + (max_line & 1);
351	if (nblank_lines == 0)
352		/* end with a newline even if the source doesn't */
353		nblank_lines = 2;
354	flush_blanks();
355	exit(0);
356}
357
358/*
359 * Prints the first 'nflush' lines. Printed lines are freed.
360 * After this function returns, 'lines' points to the first
361 * of the remaining lines, and 'nblank_lines' will have the
362 * number of half line feeds between the final flushed line
363 * and the first remaining line.
364 */
365static void
366flush_lines(int nflush)
367{
368	LINE *l;
369
370	while (--nflush >= 0) {
371		l = lines;
372		lines = l->l_next;
373		if (l->l_line) {
374			flush_blanks();
375			flush_line(l);
376			free(l->l_line);
377		}
378		if (l->l_next)
379			nblank_lines++;
380		free_line(l);
381	}
382	if (lines)
383		lines->l_prev = NULL;
384}
385
386/*
387 * Print a number of newline/half newlines.
388 * nblank_lines is the number of half line feeds.
389 */
390static void
391flush_blanks(void)
392{
393	int half, i, nb;
394
395	half = 0;
396	nb = nblank_lines;
397	if (nb & 1) {
398		if (fine)
399			half = 1;
400		else
401			nb++;
402	}
403	nb /= 2;
404	for (i = nb; --i >= 0;)
405		PUTC('\n');
406	if (half) {
407		PUTC(ESC);
408		PUTC(FHLF);
409		if (!nb)
410			PUTC('\r');
411	}
412	nblank_lines = 0;
413}
414
415/*
416 * Write a line to stdout taking care of space to tab conversion (-h flag)
417 * and character set shifts.
418 */
419static void
420flush_line(LINE *l)
421{
422	CHAR *c, *endc;
423	int i, j, nchars, last_col, save, this_col, tot;
424
425	last_col = 0;
426	nchars = l->l_line_len;
427
428	if (l->l_needs_sort) {
429		static CHAR *sorted;
430		static int count_size, *count, sorted_size;
431
432		/*
433		 * Do an O(n) sort on l->l_line by column being careful to
434		 * preserve the order of characters in the same column.
435		 */
436		if (l->l_lsize > sorted_size) {
437			sorted_size = l->l_lsize;
438			if ((sorted = realloc(sorted,
439			    (unsigned)sizeof(CHAR) * sorted_size)) == NULL)
440				err(1, NULL);
441		}
442		if (l->l_max_col >= count_size) {
443			count_size = l->l_max_col + 1;
444			if ((count = realloc(count,
445			    (unsigned)sizeof(int) * count_size)) == NULL)
446				err(1, NULL);
447		}
448		memset(count, 0, sizeof(int) * l->l_max_col + 1);
449		for (i = nchars, c = l->l_line; --i >= 0; c++)
450			count[c->c_column]++;
451
452		/*
453		 * calculate running total (shifted down by 1) to use as
454		 * indices into new line.
455		 */
456		for (tot = 0, i = 0; i <= l->l_max_col; i++) {
457			save = count[i];
458			count[i] = tot;
459			tot += save;
460		}
461
462		for (i = nchars, c = l->l_line; --i >= 0; c++)
463			sorted[count[c->c_column]++] = *c;
464		c = sorted;
465	} else
466		c = l->l_line;
467	while (nchars > 0) {
468		this_col = c->c_column;
469		endc = c;
470		do {
471			++endc;
472		} while (--nchars > 0 && this_col == endc->c_column);
473
474		/* if -b only print last character */
475		if (no_backspaces) {
476			c = endc - 1;
477			if (nchars > 0 &&
478			    this_col + c->c_width > endc->c_column)
479				continue;
480		}
481
482		if (this_col > last_col) {
483			int nspace = this_col - last_col;
484
485			if (compress_spaces && nspace > 1) {
486				while (1) {
487					int tab_col, tab_size;
488
489					tab_col = (last_col + 8) & ~7;
490					if (tab_col > this_col)
491						break;
492					tab_size = tab_col - last_col;
493					if (tab_size == 1)
494						PUTC(' ');
495					else
496						PUTC('\t');
497					nspace -= tab_size;
498					last_col = tab_col;
499				}
500			}
501			while (--nspace >= 0)
502				PUTC(' ');
503			last_col = this_col;
504		}
505
506		for (;;) {
507			if (c->c_set != last_set) {
508				switch (c->c_set) {
509				case CS_NORMAL:
510					PUTC(SI);
511					break;
512				case CS_ALTERNATE:
513					PUTC(SO);
514				}
515				last_set = c->c_set;
516			}
517			PUTC(c->c_char);
518			if ((c + 1) < endc)
519				for (j = 0; j < c->c_width; j++)
520					PUTC('\b');
521			if (++c >= endc)
522				break;
523		}
524		last_col += (c - 1)->c_width;
525	}
526}
527
528/*
529 * Increment or decrement a line number, checking for overflow.
530 * Stop one below INT_MAX such that the adjust variable is safe.
531 */
532void
533addto_lineno(int *lno, int offset)
534{
535	if (offset > 0) {
536		if (*lno >= INT_MAX - offset)
537			errx(1, "too many lines");
538	} else {
539		if (*lno < INT_MIN - offset)
540			errx(1, "too many reverse line feeds");
541	}
542	*lno += offset;
543}
544
545#define	NALLOC 64
546
547static LINE *line_freelist;
548
549static LINE *
550alloc_line(void)
551{
552	LINE *l;
553	int i;
554
555	if (!line_freelist) {
556		if ((l = realloc(NULL, sizeof(LINE) * NALLOC)) == NULL)
557			err(1, NULL);
558		line_freelist = l;
559		for (i = 1; i < NALLOC; i++, l++)
560			l->l_next = l + 1;
561		l->l_next = NULL;
562	}
563	l = line_freelist;
564	line_freelist = l->l_next;
565
566	memset(l, 0, sizeof(LINE));
567	return (l);
568}
569
570static void
571free_line(LINE *l)
572{
573
574	l->l_next = line_freelist;
575	line_freelist = l;
576}
577
578static void
579usage(void)
580{
581
582	(void)fprintf(stderr, "usage: col [-bfhpx] [-l nline]\n");
583	exit(1);
584}
585
586static void
587dowarn(int line)
588{
589
590	warnx("warning: can't back up %s",
591		line < 0 ? "past first line" : "-- line already flushed");
592}
593