1 /* $Id: term.c,v 1.201 2011/09/21 09:57:13 schwarze Exp $ */
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
6 * Permission to use, copy, modify, and distribute this software for any
7 * purpose with or without fee is hereby granted, provided that the above
8 * copyright notice and this permission notice appear in all copies.
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
22 #include <sys/types.h>
36 static void adjbuf(struct termp *p, int);
37 static void bufferc(struct termp *, char);
38 static void encode(struct termp *, const char *, size_t);
39 static void encode1(struct termp *, int);
42 term_free(struct termp *p)
48 mchars_free(p->symtab);
55 term_begin(struct termp *p, term_margin head,
56 term_margin foot, const void *arg)
67 term_end(struct termp *p)
74 * Flush a line of text. A "line" is loosely defined as being something
75 * that should be followed by a newline, regardless of whether it's
76 * broken apart by newlines getting there. A line can also be a
77 * fragment of a columnar list (`Bl -tag' or `Bl -column'), which does
78 * not have a trailing newline.
80 * The following flags may be specified:
82 * - TERMP_NOBREAK: this is the most important and is used when making
83 * columns. In short: don't print a newline and instead expect the
84 * next call to do the padding up to the start of the next column.
86 * - TERMP_TWOSPACE: make sure there is room for at least two space
87 * characters of padding. Otherwise, rather break the line.
89 * - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and
90 * the line is overrun, and don't pad-right if it's underrun.
92 * - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when
93 * overrunning, instead save the position and continue at that point
94 * when the next invocation.
96 * In-line line breaking:
98 * If TERMP_NOBREAK is specified and the line overruns the right
99 * margin, it will break and pad-right to the right margin after
100 * writing. If maxrmargin is violated, it will break and continue
101 * writing from the right-margin, which will lead to the above scenario
102 * upon exit. Otherwise, the line will break at the right margin.
105 term_flushln(struct termp *p)
107 int i; /* current input position in p->buf */
108 size_t vis; /* current visual position on output */
109 size_t vbl; /* number of blanks to prepend to output */
110 size_t vend; /* end of word visual position on output */
111 size_t bp; /* visual right border position */
112 size_t dv; /* temporary for visual pos calculations */
113 int j; /* temporary loop index for p->buf */
114 int jhy; /* last hyph before overflow w/r/t j */
115 size_t maxvis; /* output position of visible boundary */
116 size_t mmax; /* used in calculating bp */
119 * First, establish the maximum columns of "visible" content.
120 * This is usually the difference between the right-margin and
121 * an indentation, but can be, for tagged lists or columns, a
122 * small set of values.
124 assert (p->rmargin >= p->offset);
125 dv = p->rmargin - p->offset;
126 maxvis = (int)dv > p->overstep ? dv - (size_t)p->overstep : 0;
127 dv = p->maxrmargin - p->offset;
128 mmax = (int)dv > p->overstep ? dv - (size_t)p->overstep : 0;
130 bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
133 * Calculate the required amount of padding.
135 vbl = p->offset + p->overstep > p->viscol ?
136 p->offset + p->overstep - p->viscol : 0;
143 * Handle literal tab characters: collapse all
144 * subsequent tabs into a single huge set of spaces.
146 while (i < p->col && '\t' == p->buf[i]) {
147 vend = (vis / p->tabwidth + 1) * p->tabwidth;
154 * Count up visible word characters. Control sequences
155 * (starting with the CSI) aren't counted. A space
156 * generates a non-printing word, which is valid (the
157 * space is printed according to regular spacing rules).
160 for (j = i, jhy = 0; j < p->col; j++) {
161 if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j])
164 /* Back over the the last printed character. */
165 if (8 == p->buf[j]) {
167 vend -= (*p->width)(p, p->buf[j - 1]);
172 /* Break at the hyphen point if we overrun. */
173 if (vend > vis && vend < bp &&
174 ASCII_HYPH == p->buf[j])
177 vend += (*p->width)(p, p->buf[j]);
181 * Find out whether we would exceed the right margin.
182 * If so, break to the next line.
184 if (vend > bp && 0 == jhy && vis > 0) {
188 if (TERMP_NOBREAK & p->flags) {
190 vend += p->rmargin - p->offset;
194 /* Remove the p->overstep width. */
196 bp += (size_t)p->overstep;
200 /* Write out the [remaining] word. */
201 for ( ; i < p->col; i++) {
202 if (vend > bp && jhy > 0 && i > jhy)
204 if ('\t' == p->buf[i])
206 if (' ' == p->buf[i]) {
208 while (' ' == p->buf[i])
210 dv = (size_t)(i - j) * (*p->width)(p, ' ');
215 if (ASCII_NBRSP == p->buf[i]) {
216 vbl += (*p->width)(p, ' ');
221 * Now we definitely know there will be
222 * printable characters to output,
223 * so write preceding white space now.
226 (*p->advance)(p, vbl);
231 if (ASCII_HYPH == p->buf[i]) {
232 (*p->letter)(p, '-');
233 p->viscol += (*p->width)(p, '-');
237 (*p->letter)(p, p->buf[i]);
239 p->viscol -= (*p->width)(p, p->buf[i-1]);
241 p->viscol += (*p->width)(p, p->buf[i]);
247 * If there was trailing white space, it was not printed;
248 * so reset the cursor position accordingly.
256 if ( ! (TERMP_NOBREAK & p->flags)) {
262 if (TERMP_HANG & p->flags) {
263 /* We need one blank after the tag. */
264 p->overstep = (int)(vis - maxvis + (*p->width)(p, ' '));
267 * Behave exactly the same way as groff:
268 * If we have overstepped the margin, temporarily move
269 * it to the right and flag the rest of the line to be
271 * If we landed right at the margin, be happy.
272 * If we are one step before the margin, temporarily
273 * move it one step LEFT and flag the rest of the line
276 if (p->overstep < -1)
280 } else if (TERMP_DANGLE & p->flags)
283 /* If the column was overrun, break the line. */
285 ((TERMP_TWOSPACE & p->flags) ? (*p->width)(p, ' ') : 0)) {
293 * A newline only breaks an existing line; it won't assert vertical
294 * space. All data in the output buffer is flushed prior to the newline
298 term_newln(struct termp *p)
301 p->flags |= TERMP_NOSPACE;
302 if (p->col || p->viscol)
308 * Asserts a vertical space (a full, empty line-break between lines).
309 * Note that if used twice, this will cause two blank spaces and so on.
310 * All data in the output buffer is flushed prior to the newline
314 term_vspace(struct termp *p)
323 term_fontlast(struct termp *p)
328 p->fontl = p->fontq[p->fonti];
329 p->fontq[p->fonti] = f;
334 term_fontrepl(struct termp *p, enum termfont f)
337 p->fontl = p->fontq[p->fonti];
338 p->fontq[p->fonti] = f;
343 term_fontpush(struct termp *p, enum termfont f)
346 assert(p->fonti + 1 < 10);
347 p->fontl = p->fontq[p->fonti];
348 p->fontq[++p->fonti] = f;
353 term_fontq(struct termp *p)
356 return(&p->fontq[p->fonti]);
361 term_fonttop(struct termp *p)
364 return(p->fontq[p->fonti]);
369 term_fontpopq(struct termp *p, const void *key)
372 while (p->fonti >= 0 && key != &p->fontq[p->fonti])
374 assert(p->fonti >= 0);
379 term_fontpop(struct termp *p)
387 * Handle pwords, partial words, which may be either a single word or a
388 * phrase that cannot be broken down (such as a literal string). This
389 * handles word styling.
392 term_word(struct termp *p, const char *word)
394 const char *seq, *cp;
400 if ( ! (TERMP_NOSPACE & p->flags)) {
401 if ( ! (TERMP_KEEP & p->flags)) {
402 if (TERMP_PREKEEP & p->flags)
403 p->flags |= TERMP_KEEP;
405 if (TERMP_SENTENCE & p->flags)
408 bufferc(p, ASCII_NBRSP);
411 if ( ! (p->flags & TERMP_NONOSPACE))
412 p->flags &= ~TERMP_NOSPACE;
414 p->flags |= TERMP_NOSPACE;
416 p->flags &= ~(TERMP_SENTENCE | TERMP_IGNDELIM);
418 while ('\0' != *word) {
419 if ((ssz = strcspn(word, "\\")) > 0)
420 encode(p, word, ssz);
427 esc = mandoc_escape(&word, &seq, &sz);
428 if (ESCAPE_ERROR == esc)
431 if (TERMENC_ASCII != p->enc)
433 case (ESCAPE_UNICODE):
434 uc = mchars_num2uc(seq + 1, sz - 1);
439 case (ESCAPE_SPECIAL):
440 uc = mchars_spec2cp(p->symtab, seq, sz);
450 case (ESCAPE_UNICODE):
453 case (ESCAPE_NUMBERED):
454 c = mchars_num2char(seq, sz);
458 case (ESCAPE_SPECIAL):
459 cp = mchars_spec2str(p->symtab, seq, sz, &ssz);
465 case (ESCAPE_FONTBOLD):
466 term_fontrepl(p, TERMFONT_BOLD);
468 case (ESCAPE_FONTITALIC):
469 term_fontrepl(p, TERMFONT_UNDER);
473 case (ESCAPE_FONTROMAN):
474 term_fontrepl(p, TERMFONT_NONE);
476 case (ESCAPE_FONTPREV):
479 case (ESCAPE_NOSPACE):
481 p->flags |= TERMP_NOSPACE;
490 adjbuf(struct termp *p, int sz)
495 while (sz >= p->maxcols)
498 p->buf = mandoc_realloc
499 (p->buf, sizeof(int) * (size_t)p->maxcols);
503 bufferc(struct termp *p, char c)
506 if (p->col + 1 >= p->maxcols)
507 adjbuf(p, p->col + 1);
509 p->buf[p->col++] = c;
514 * Do this for a single (probably unicode) value.
515 * Does not check for non-decorated glyphs.
518 encode1(struct termp *p, int c)
522 if (p->col + 4 >= p->maxcols)
523 adjbuf(p, p->col + 4);
527 if (TERMFONT_NONE == f) {
528 p->buf[p->col++] = c;
530 } else if (TERMFONT_UNDER == f) {
531 p->buf[p->col++] = '_';
533 p->buf[p->col++] = c;
535 p->buf[p->col++] = 8;
536 p->buf[p->col++] = c;
540 encode(struct termp *p, const char *word, size_t sz)
549 * Encode and buffer a string of characters. If the current
550 * font mode is unset, buffer directly, else encode then buffer
551 * character by character.
554 if (TERMFONT_NONE == (f = term_fonttop(p))) {
555 if (p->col + len >= p->maxcols)
556 adjbuf(p, p->col + len);
557 for (i = 0; i < len; i++)
558 p->buf[p->col++] = word[i];
562 /* Pre-buffer, assuming worst-case. */
564 if (p->col + 1 + (len * 3) >= p->maxcols)
565 adjbuf(p, p->col + 1 + (len * 3));
567 for (i = 0; i < len; i++) {
568 if (ASCII_HYPH != word[i] &&
569 ! isgraph((unsigned char)word[i])) {
570 p->buf[p->col++] = word[i];
574 if (TERMFONT_UNDER == f)
575 p->buf[p->col++] = '_';
576 else if (ASCII_HYPH == word[i])
577 p->buf[p->col++] = '-';
579 p->buf[p->col++] = word[i];
581 p->buf[p->col++] = 8;
582 p->buf[p->col++] = word[i];
587 term_len(const struct termp *p, size_t sz)
590 return((*p->width)(p, ' ') * sz);
595 term_strlen(const struct termp *p, const char *cp)
599 const char *seq, *rhs;
601 static const char rej[] = { '\\', ASCII_HYPH, ASCII_NBRSP, '\0' };
604 * Account for escaped sequences within string length
605 * calculations. This follows the logic in term_word() as we
606 * must calculate the width of produced strings.
610 while ('\0' != *cp) {
611 rsz = strcspn(cp, rej);
612 for (i = 0; i < rsz; i++)
613 sz += (*p->width)(p, *cp++);
619 esc = mandoc_escape(&cp, &seq, &ssz);
620 if (ESCAPE_ERROR == esc)
623 if (TERMENC_ASCII != p->enc)
625 case (ESCAPE_UNICODE):
630 sz += (*p->width)(p, c);
632 case (ESCAPE_SPECIAL):
634 (p->symtab, seq, ssz);
637 sz += (*p->width)(p, c);
646 case (ESCAPE_UNICODE):
647 sz += (*p->width)(p, '?');
649 case (ESCAPE_NUMBERED):
650 c = mchars_num2char(seq, ssz);
652 sz += (*p->width)(p, c);
654 case (ESCAPE_SPECIAL):
655 rhs = mchars_spec2str
656 (p->symtab, seq, ssz, &rsz);
671 for (i = 0; i < rsz; i++)
672 sz += (*p->width)(p, *rhs++);
675 sz += (*p->width)(p, ' ');
679 sz += (*p->width)(p, '-');
692 term_vspan(const struct termp *p, const struct roffsu *su)
710 r = su->scale / 1000;
722 return(/* LINTED */(size_t)
727 term_hspan(const struct termp *p, const struct roffsu *su)
731 v = ((*p->hspan)(p, su));
734 return((size_t) /* LINTED */