]> git.cameronkatri.com Git - mandoc.git/blob - term.c
git.ckatri.com / mandoc.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
history | raw | HEAD
Purge duplicate error reporting from the .tr request parser:
[mandoc.git] / term.c
1 /* $Id: term.c,v 1.286 2022/04/27 13:41:13 schwarze Exp $ */
2 /*
3 * Copyright (c) 2010-2022 Ingo Schwarze <schwarze@openbsd.org>
4 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
5 *
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.
9 *
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS 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.
17 */
18 #include "config.h"
19
20 #include <sys/types.h>
21
22 #include <assert.h>
23 #include <ctype.h>
24 #include <stdint.h>
25 #include <stdio.h>
26 #include <stdlib.h>
27 #include <string.h>
28
29 #include "mandoc.h"
30 #include "mandoc_aux.h"
31 #include "out.h"
32 #include "term.h"
33 #include "main.h"
34
35 static size_t cond_width(const struct termp *, int, int *);
36 static void adjbuf(struct termp_col *, size_t);
37 static void bufferc(struct termp *, char);
38 static void encode(struct termp *, const char *, size_t);
39 static void encode1(struct termp *, int);
40 static void endline(struct termp *);
41 static void term_field(struct termp *, size_t, size_t);
42 static void term_fill(struct termp *, size_t *, size_t *,
43 size_t);
44
45
46 void
47 term_setcol(struct termp *p, size_t maxtcol)
48 {
49 if (maxtcol > p->maxtcol) {
50 p->tcols = mandoc_recallocarray(p->tcols,
51 p->maxtcol, maxtcol, sizeof(*p->tcols));
52 p->maxtcol = maxtcol;
53 }
54 p->lasttcol = maxtcol - 1;
55 p->tcol = p->tcols;
56 }
57
58 void
59 term_free(struct termp *p)
60 {
61 term_tab_free();
62 for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++)
63 free(p->tcol->buf);
64 free(p->tcols);
65 free(p->fontq);
66 free(p);
67 }
68
69 void
70 term_begin(struct termp *p, term_margin head,
71 term_margin foot, const struct roff_meta *arg)
72 {
73
74 p->headf = head;
75 p->footf = foot;
76 p->argf = arg;
77 (*p->begin)(p);
78 }
79
80 void
81 term_end(struct termp *p)
82 {
83
84 (*p->end)(p);
85 }
86
87 /*
88 * Flush a chunk of text. By default, break the output line each time
89 * the right margin is reached, and continue output on the next line
90 * at the same offset as the chunk itself. By default, also break the
91 * output line at the end of the chunk. There are many flags modifying
92 * this behaviour, see the comments in the body of the function.
93 */
94 void
95 term_flushln(struct termp *p)
96 {
97 size_t vbl; /* Number of blanks to prepend to the output. */
98 size_t vbr; /* Actual visual position of the end of field. */
99 size_t vfield; /* Desired visual field width. */
100 size_t vtarget; /* Desired visual position of the right margin. */
101 size_t ic; /* Character position in the input buffer. */
102 size_t nbr; /* Number of characters to print in this field. */
103
104 /*
105 * Normally, start writing at the left margin, but with the
106 * NOPAD flag, start writing at the current position instead.
107 */
108
109 vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ?
110 0 : p->tcol->offset - p->viscol;
111 if (p->minbl && vbl < p->minbl)
112 vbl = p->minbl;
113
114 if ((p->flags & TERMP_MULTICOL) == 0)
115 p->tcol->col = 0;
116
117 /* Loop over output lines. */
118
119 for (;;) {
120 vfield = p->tcol->rmargin > p->viscol + vbl ?
121 p->tcol->rmargin - p->viscol - vbl : 0;
122
123 /*
124 * Normally, break the line at the the right margin
125 * of the field, but with the NOBREAK flag, only
126 * break it at the max right margin of the screen,
127 * and with the BRNEVER flag, never break it at all.
128 */
129
130 vtarget = (p->flags & TERMP_NOBREAK) == 0 ? vfield :
131 p->maxrmargin > p->viscol + vbl ?
132 p->maxrmargin - p->viscol - vbl : 0;
133
134 /*
135 * Figure out how much text will fit in the field.
136 * If there is whitespace only, print nothing.
137 */
138
139 term_fill(p, &nbr, &vbr,
140 p->flags & TERMP_BRNEVER ? SIZE_MAX : vtarget);
141 if (nbr == 0)
142 break;
143
144 /*
145 * With the CENTER or RIGHT flag, increase the indentation
146 * to center the text between the left and right margins
147 * or to adjust it to the right margin, respectively.
148 */
149
150 if (vbr < vtarget) {
151 if (p->flags & TERMP_CENTER)
152 vbl += (vtarget - vbr) / 2;
153 else if (p->flags & TERMP_RIGHT)
154 vbl += vtarget - vbr;
155 }
156
157 /* Finally, print the field content. */
158
159 term_field(p, vbl, nbr);
160
161 /*
162 * If there is no text left in the field, exit the loop.
163 * If the BRTRSP flag is set, consider trailing
164 * whitespace significant when deciding whether
165 * the field fits or not.
166 */
167
168 for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
169 switch (p->tcol->buf[ic]) {
170 case '\t':
171 if (p->flags & TERMP_BRTRSP)
172 vbr = term_tab_next(vbr);
173 continue;
174 case ' ':
175 if (p->flags & TERMP_BRTRSP)
176 vbr += (*p->width)(p, ' ');
177 continue;
178 case '\n':
179 case ASCII_BREAK:
180 continue;
181 default:
182 break;
183 }
184 break;
185 }
186 if (ic == p->tcol->lastcol)
187 break;
188
189 /*
190 * At the location of an automtic line break, input
191 * space characters are consumed by the line break.
192 */
193
194 while (p->tcol->col < p->tcol->lastcol &&
195 p->tcol->buf[p->tcol->col] == ' ')
196 p->tcol->col++;
197
198 /*
199 * In multi-column mode, leave the rest of the text
200 * in the buffer to be handled by a subsequent
201 * invocation, such that the other columns of the
202 * table can be handled first.
203 * In single-column mode, simply break the line.
204 */
205
206 if (p->flags & TERMP_MULTICOL)
207 return;
208
209 endline(p);
210 p->viscol = 0;
211
212 /*
213 * Normally, start the next line at the same indentation
214 * as this one, but with the BRIND flag, start it at the
215 * right margin instead. This is used together with
216 * NOBREAK for the tags in various kinds of tagged lists.
217 */
218
219 vbl = p->flags & TERMP_BRIND ?
220 p->tcol->rmargin : p->tcol->offset;
221 }
222
223 /* Reset output state in preparation for the next field. */
224
225 p->col = p->tcol->col = p->tcol->lastcol = 0;
226 p->minbl = p->trailspace;
227 p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE | TERMP_NOPAD);
228
229 if (p->flags & TERMP_MULTICOL)
230 return;
231
232 /*
233 * The HANG flag means that the next field
234 * always follows on the same line.
235 * The NOBREAK flag means that the next field
236 * follows on the same line unless the field was overrun.
237 * Normally, break the line at the end of each field.
238 */
239
240 if ((p->flags & TERMP_HANG) == 0 &&
241 ((p->flags & TERMP_NOBREAK) == 0 ||
242 vbr + term_len(p, p->trailspace) > vfield))
243 endline(p);
244 }
245
246 /*
247 * Store the number of input characters to print in this field in *nbr
248 * and their total visual width to print in *vbr.
249 * If there is only whitespace in the field, both remain zero.
250 * The desired visual width of the field is provided by vtarget.
251 * If the first word is longer, the field will be overrun.
252 */
253 static void
254 term_fill(struct termp *p, size_t *nbr, size_t *vbr, size_t vtarget)
255 {
256 size_t ic; /* Character position in the input buffer. */
257 size_t vis; /* Visual position of the current character. */
258 size_t vn; /* Visual position of the next character. */
259 int breakline; /* Break at the end of this word. */
260 int graph; /* Last character was non-blank. */
261
262 *nbr = *vbr = vis = 0;
263 breakline = graph = 0;
264 for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
265 switch (p->tcol->buf[ic]) {
266 case '\b': /* Escape \o (overstrike) or backspace markup. */
267 assert(ic > 0);
268 vis -= (*p->width)(p, p->tcol->buf[ic - 1]);
269 continue;
270
271 case '\t': /* Normal ASCII whitespace. */
272 case ' ':
273 case ASCII_BREAK: /* Escape \: (breakpoint). */
274 switch (p->tcol->buf[ic]) {
275 case '\t':
276 vn = term_tab_next(vis);
277 break;
278 case ' ':
279 vn = vis + (*p->width)(p, ' ');
280 break;
281 case ASCII_BREAK:
282 vn = vis;
283 break;
284 default:
285 abort();
286 }
287 /* Can break at the end of a word. */
288 if (breakline || vn > vtarget)
289 break;
290 if (graph) {
291 *nbr = ic;
292 *vbr = vis;
293 graph = 0;
294 }
295 vis = vn;
296 continue;
297
298 case '\n': /* Escape \p (break at the end of the word). */
299 breakline = 1;
300 continue;
301
302 case ASCII_HYPH: /* Breakable hyphen. */
303 graph = 1;
304 /*
305 * We are about to decide whether to break the
306 * line or not, so we no longer need this hyphen
307 * to be marked as breakable. Put back a real
308 * hyphen such that we get the correct width.
309 */
310 p->tcol->buf[ic] = '-';
311 vis += (*p->width)(p, '-');
312 if (vis > vtarget) {
313 ic++;
314 break;
315 }
316 *nbr = ic + 1;
317 *vbr = vis;
318 continue;
319
320 case ASCII_NBRSP: /* Non-breakable space. */
321 p->tcol->buf[ic] = ' ';
322 /* FALLTHROUGH */
323 default: /* Printable character. */
324 graph = 1;
325 vis += (*p->width)(p, p->tcol->buf[ic]);
326 if (vis > vtarget && *nbr > 0)
327 return;
328 continue;
329 }
330 break;
331 }
332
333 /*
334 * If the last word extends to the end of the field without any
335 * trailing whitespace, the loop could not check yet whether it
336 * can remain on this line. So do the check now.
337 */
338
339 if (graph && (vis <= vtarget || *nbr == 0)) {
340 *nbr = ic;
341 *vbr = vis;
342 }
343 }
344
345 /*
346 * Print the contents of one field
347 * with an indentation of vbl visual columns,
348 * and an input string length of nbr characters.
349 */
350 static void
351 term_field(struct termp *p, size_t vbl, size_t nbr)
352 {
353 size_t ic; /* Character position in the input buffer. */
354 size_t vis; /* Visual position of the current character. */
355 size_t dv; /* Visual width of the current character. */
356 size_t vn; /* Visual position of the next character. */
357
358 vis = 0;
359 for (ic = p->tcol->col; ic < nbr; ic++) {
360
361 /*
362 * To avoid the printing of trailing whitespace,
363 * do not print whitespace right away, only count it.
364 */
365
366 switch (p->tcol->buf[ic]) {
367 case '\n':
368 case ASCII_BREAK:
369 continue;
370 case '\t':
371 vn = term_tab_next(vis);
372 vbl += vn - vis;
373 vis = vn;
374 continue;
375 case ' ':
376 case ASCII_NBRSP:
377 dv = (*p->width)(p, ' ');
378 vbl += dv;
379 vis += dv;
380 continue;
381 default:
382 break;
383 }
384
385 /*
386 * We found a non-blank character to print,
387 * so write preceding white space now.
388 */
389
390 if (vbl > 0) {
391 (*p->advance)(p, vbl);
392 p->viscol += vbl;
393 vbl = 0;
394 }
395
396 /* Print the character and adjust the visual position. */
397
398 (*p->letter)(p, p->tcol->buf[ic]);
399 if (p->tcol->buf[ic] == '\b') {
400 dv = (*p->width)(p, p->tcol->buf[ic - 1]);
401 p->viscol -= dv;
402 vis -= dv;
403 } else {
404 dv = (*p->width)(p, p->tcol->buf[ic]);
405 p->viscol += dv;
406 vis += dv;
407 }
408 }
409 p->tcol->col = nbr;
410 }
411
412 static void
413 endline(struct termp *p)
414 {
415 if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) {
416 p->mc = NULL;
417 p->flags &= ~TERMP_ENDMC;
418 }
419 if (p->mc != NULL) {
420 if (p->viscol && p->maxrmargin >= p->viscol)
421 (*p->advance)(p, p->maxrmargin - p->viscol + 1);
422 p->flags |= TERMP_NOBUF | TERMP_NOSPACE;
423 term_word(p, p->mc);
424 p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC);
425 }
426 p->viscol = 0;
427 p->minbl = 0;
428 (*p->endline)(p);
429 }
430
431 /*
432 * A newline only breaks an existing line; it won't assert vertical
433 * space. All data in the output buffer is flushed prior to the newline
434 * assertion.
435 */
436 void
437 term_newln(struct termp *p)
438 {
439
440 p->flags |= TERMP_NOSPACE;
441 if (p->tcol->lastcol || p->viscol)
442 term_flushln(p);
443 }
444
445 /*
446 * Asserts a vertical space (a full, empty line-break between lines).
447 * Note that if used twice, this will cause two blank spaces and so on.
448 * All data in the output buffer is flushed prior to the newline
449 * assertion.
450 */
451 void
452 term_vspace(struct termp *p)
453 {
454
455 term_newln(p);
456 p->viscol = 0;
457 p->minbl = 0;
458 if (0 < p->skipvsp)
459 p->skipvsp--;
460 else
461 (*p->endline)(p);
462 }
463
464 /* Swap current and previous font; for \fP and .ft P */
465 void
466 term_fontlast(struct termp *p)
467 {
468 enum termfont f;
469
470 f = p->fontl;
471 p->fontl = p->fontq[p->fonti];
472 p->fontq[p->fonti] = f;
473 }
474
475 /* Set font, save current, discard previous; for \f, .ft, .B etc. */
476 void
477 term_fontrepl(struct termp *p, enum termfont f)
478 {
479
480 p->fontl = p->fontq[p->fonti];
481 p->fontq[p->fonti] = f;
482 }
483
484 /* Set font, save previous. */
485 void
486 term_fontpush(struct termp *p, enum termfont f)
487 {
488
489 p->fontl = p->fontq[p->fonti];
490 if (++p->fonti == p->fontsz) {
491 p->fontsz += 8;
492 p->fontq = mandoc_reallocarray(p->fontq,
493 p->fontsz, sizeof(*p->fontq));
494 }
495 p->fontq[p->fonti] = f;
496 }
497
498 /* Flush to make the saved pointer current again. */
499 void
500 term_fontpopq(struct termp *p, int i)
501 {
502
503 assert(i >= 0);
504 if (p->fonti > i)
505 p->fonti = i;
506 }
507
508 /* Pop one font off the stack. */
509 void
510 term_fontpop(struct termp *p)
511 {
512
513 assert(p->fonti);
514 p->fonti--;
515 }
516
517 /*
518 * Handle pwords, partial words, which may be either a single word or a
519 * phrase that cannot be broken down (such as a literal string). This
520 * handles word styling.
521 */
522 void
523 term_word(struct termp *p, const char *word)
524 {
525 struct roffsu su;
526 const char nbrsp[2] = { ASCII_NBRSP, 0 };
527 const char *seq, *cp;
528 int sz, uc;
529 size_t csz, lsz, ssz;
530 enum mandoc_esc esc;
531
532 if ((p->flags & TERMP_NOBUF) == 0) {
533 if ((p->flags & TERMP_NOSPACE) == 0) {
534 if ((p->flags & TERMP_KEEP) == 0) {
535 bufferc(p, ' ');
536 if (p->flags & TERMP_SENTENCE)
537 bufferc(p, ' ');
538 } else
539 bufferc(p, ASCII_NBRSP);
540 }
541 if (p->flags & TERMP_PREKEEP)
542 p->flags |= TERMP_KEEP;
543 if (p->flags & TERMP_NONOSPACE)
544 p->flags |= TERMP_NOSPACE;
545 else
546 p->flags &= ~TERMP_NOSPACE;
547 p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE);
548 p->skipvsp = 0;
549 }
550
551 while ('\0' != *word) {
552 if ('\\' != *word) {
553 if (TERMP_NBRWORD & p->flags) {
554 if (' ' == *word) {
555 encode(p, nbrsp, 1);
556 word++;
557 continue;
558 }
559 ssz = strcspn(word, "\\ ");
560 } else
561 ssz = strcspn(word, "\\");
562 encode(p, word, ssz);
563 word += (int)ssz;
564 continue;
565 }
566
567 word++;
568 esc = mandoc_escape(&word, &seq, &sz);
569 switch (esc) {
570 case ESCAPE_UNICODE:
571 uc = mchars_num2uc(seq + 1, sz - 1);
572 break;
573 case ESCAPE_NUMBERED:
574 uc = mchars_num2char(seq, sz);
575 if (uc < 0)
576 continue;
577 break;
578 case ESCAPE_SPECIAL:
579 if (p->enc == TERMENC_ASCII) {
580 cp = mchars_spec2str(seq, sz, &ssz);
581 if (cp != NULL)
582 encode(p, cp, ssz);
583 } else {
584 uc = mchars_spec2cp(seq, sz);
585 if (uc > 0)
586 encode1(p, uc);
587 }
588 continue;
589 case ESCAPE_UNDEF:
590 uc = *seq;
591 break;
592 case ESCAPE_FONTBOLD:
593 case ESCAPE_FONTCB:
594 term_fontrepl(p, TERMFONT_BOLD);
595 continue;
596 case ESCAPE_FONTITALIC:
597 case ESCAPE_FONTCI:
598 term_fontrepl(p, TERMFONT_UNDER);
599 continue;
600 case ESCAPE_FONTBI:
601 term_fontrepl(p, TERMFONT_BI);
602 continue;
603 case ESCAPE_FONT:
604 case ESCAPE_FONTCR:
605 case ESCAPE_FONTROMAN:
606 term_fontrepl(p, TERMFONT_NONE);
607 continue;
608 case ESCAPE_FONTPREV:
609 term_fontlast(p);
610 continue;
611 case ESCAPE_BREAK:
612 bufferc(p, '\n');
613 continue;
614 case ESCAPE_NOSPACE:
615 if (p->flags & TERMP_BACKAFTER)
616 p->flags &= ~TERMP_BACKAFTER;
617 else if (*word == '\0')
618 p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE);
619 continue;
620 case ESCAPE_DEVICE:
621 if (p->type == TERMTYPE_PDF)
622 encode(p, "pdf", 3);
623 else if (p->type == TERMTYPE_PS)
624 encode(p, "ps", 2);
625 else if (p->enc == TERMENC_ASCII)
626 encode(p, "ascii", 5);
627 else
628 encode(p, "utf8", 4);
629 continue;
630 case ESCAPE_HORIZ:
631 if (p->flags & TERMP_BACKAFTER) {
632 p->flags &= ~TERMP_BACKAFTER;
633 continue;
634 }
635 if (*seq == '|') {
636 seq++;
637 uc = -p->col;
638 } else
639 uc = 0;
640 if (a2roffsu(seq, &su, SCALE_EM) == NULL)
641 continue;
642 uc += term_hen(p, &su);
643 if (uc >= 0) {
644 while (uc > 0) {
645 uc -= term_len(p, 1);
646 if (p->flags & TERMP_BACKBEFORE)
647 p->flags &= ~TERMP_BACKBEFORE;
648 else
649 bufferc(p, ASCII_NBRSP);
650 }
651 continue;
652 }
653 if (p->flags & TERMP_BACKBEFORE) {
654 p->flags &= ~TERMP_BACKBEFORE;
655 assert(p->col > 0);
656 p->col--;
657 }
658 if (p->col >= (size_t)(-uc)) {
659 p->col += uc;
660 } else {
661 uc += p->col;
662 p->col = 0;
663 if (p->tcol->offset > (size_t)(-uc)) {
664 p->ti += uc;
665 p->tcol->offset += uc;
666 } else {
667 p->ti -= p->tcol->offset;
668 p->tcol->offset = 0;
669 }
670 }
671 continue;
672 case ESCAPE_HLINE:
673 if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL)
674 continue;
675 uc = term_hen(p, &su);
676 if (uc <= 0) {
677 if (p->tcol->rmargin <= p->tcol->offset)
678 continue;
679 lsz = p->tcol->rmargin - p->tcol->offset;
680 } else
681 lsz = uc;
682 if (*cp == seq[-1])
683 uc = -1;
684 else if (*cp == '\\') {
685 seq = cp + 1;
686 esc = mandoc_escape(&seq, &cp, &sz);
687 switch (esc) {
688 case ESCAPE_UNICODE:
689 uc = mchars_num2uc(cp + 1, sz - 1);
690 break;
691 case ESCAPE_NUMBERED:
692 uc = mchars_num2char(cp, sz);
693 break;
694 case ESCAPE_SPECIAL:
695 uc = mchars_spec2cp(cp, sz);
696 break;
697 case ESCAPE_UNDEF:
698 uc = *seq;
699 break;
700 default:
701 uc = -1;
702 break;
703 }
704 } else
705 uc = *cp;
706 if (uc < 0x20 || (uc > 0x7E && uc < 0xA0))
707 uc = '_';
708 if (p->enc == TERMENC_ASCII) {
709 cp = ascii_uc2str(uc);
710 csz = term_strlen(p, cp);
711 ssz = strlen(cp);
712 } else
713 csz = (*p->width)(p, uc);
714 while (lsz >= csz) {
715 if (p->enc == TERMENC_ASCII)
716 encode(p, cp, ssz);
717 else
718 encode1(p, uc);
719 lsz -= csz;
720 }
721 continue;
722 case ESCAPE_SKIPCHAR:
723 p->flags |= TERMP_BACKAFTER;
724 continue;
725 case ESCAPE_OVERSTRIKE:
726 cp = seq + sz;
727 while (seq < cp) {
728 if (*seq == '\\') {
729 mandoc_escape(&seq, NULL, NULL);
730 continue;
731 }
732 encode1(p, *seq++);
733 if (seq < cp) {
734 if (p->flags & TERMP_BACKBEFORE)
735 p->flags |= TERMP_BACKAFTER;
736 else
737 p->flags |= TERMP_BACKBEFORE;
738 }
739 }
740 /* Trim trailing backspace/blank pair. */
741 if (p->tcol->lastcol > 2 &&
742 (p->tcol->buf[p->tcol->lastcol - 1] == ' ' ||
743 p->tcol->buf[p->tcol->lastcol - 1] == '\t'))
744 p->tcol->lastcol -= 2;
745 if (p->col > p->tcol->lastcol)
746 p->col = p->tcol->lastcol;
747 continue;
748 default:
749 continue;
750 }
751
752 /*
753 * Common handling for Unicode and numbered
754 * character escape sequences.
755 */
756
757 if (p->enc == TERMENC_ASCII) {
758 cp = ascii_uc2str(uc);
759 encode(p, cp, strlen(cp));
760 } else {
761 if ((uc < 0x20 && uc != 0x09) ||
762 (uc > 0x7E && uc < 0xA0))
763 uc = 0xFFFD;
764 encode1(p, uc);
765 }
766 }
767 p->flags &= ~TERMP_NBRWORD;
768 }
769
770 static void
771 adjbuf(struct termp_col *c, size_t sz)
772 {
773 if (c->maxcols == 0)
774 c->maxcols = 1024;
775 while (c->maxcols <= sz)
776 c->maxcols <<= 2;
777 c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf));
778 }
779
780 static void
781 bufferc(struct termp *p, char c)
782 {
783 if (p->flags & TERMP_NOBUF) {
784 (*p->letter)(p, c);
785 return;
786 }
787 if (p->col + 1 >= p->tcol->maxcols)
788 adjbuf(p->tcol, p->col + 1);
789 if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
790 p->tcol->buf[p->col] = c;
791 if (p->tcol->lastcol < ++p->col)
792 p->tcol->lastcol = p->col;
793 }
794
795 /*
796 * See encode().
797 * Do this for a single (probably unicode) value.
798 * Does not check for non-decorated glyphs.
799 */
800 static void
801 encode1(struct termp *p, int c)
802 {
803 enum termfont f;
804
805 if (p->flags & TERMP_NOBUF) {
806 (*p->letter)(p, c);
807 return;
808 }
809
810 if (p->col + 7 >= p->tcol->maxcols)
811 adjbuf(p->tcol, p->col + 7);
812
813 f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ?
814 p->fontq[p->fonti] : TERMFONT_NONE;
815
816 if (p->flags & TERMP_BACKBEFORE) {
817 if (p->tcol->buf[p->col - 1] == ' ' ||
818 p->tcol->buf[p->col - 1] == '\t')
819 p->col--;
820 else
821 p->tcol->buf[p->col++] = '\b';
822 p->flags &= ~TERMP_BACKBEFORE;
823 }
824 if (f == TERMFONT_UNDER || f == TERMFONT_BI) {
825 p->tcol->buf[p->col++] = '_';
826 p->tcol->buf[p->col++] = '\b';
827 }
828 if (f == TERMFONT_BOLD || f == TERMFONT_BI) {
829 if (c == ASCII_HYPH)
830 p->tcol->buf[p->col++] = '-';
831 else
832 p->tcol->buf[p->col++] = c;
833 p->tcol->buf[p->col++] = '\b';
834 }
835 if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
836 p->tcol->buf[p->col] = c;
837 if (p->tcol->lastcol < ++p->col)
838 p->tcol->lastcol = p->col;
839 if (p->flags & TERMP_BACKAFTER) {
840 p->flags |= TERMP_BACKBEFORE;
841 p->flags &= ~TERMP_BACKAFTER;
842 }
843 }
844
845 static void
846 encode(struct termp *p, const char *word, size_t sz)
847 {
848 size_t i;
849
850 if (p->flags & TERMP_NOBUF) {
851 for (i = 0; i < sz; i++)
852 (*p->letter)(p, word[i]);
853 return;
854 }
855
856 if (p->col + 2 + (sz * 5) >= p->tcol->maxcols)
857 adjbuf(p->tcol, p->col + 2 + (sz * 5));
858
859 for (i = 0; i < sz; i++) {
860 if (ASCII_HYPH == word[i] ||
861 isgraph((unsigned char)word[i]))
862 encode1(p, word[i]);
863 else {
864 if (p->tcol->lastcol <= p->col ||
865 (word[i] != ' ' && word[i] != ASCII_NBRSP))
866 p->tcol->buf[p->col] = word[i];
867 p->col++;
868
869 /*
870 * Postpone the effect of \z while handling
871 * an overstrike sequence from ascii_uc2str().
872 */
873
874 if (word[i] == '\b' &&
875 (p->flags & TERMP_BACKBEFORE)) {
876 p->flags &= ~TERMP_BACKBEFORE;
877 p->flags |= TERMP_BACKAFTER;
878 }
879 }
880 }
881 if (p->tcol->lastcol < p->col)
882 p->tcol->lastcol = p->col;
883 }
884
885 void
886 term_setwidth(struct termp *p, const char *wstr)
887 {
888 struct roffsu su;
889 int iop, width;
890
891 iop = 0;
892 width = 0;
893 if (NULL != wstr) {
894 switch (*wstr) {
895 case '+':
896 iop = 1;
897 wstr++;
898 break;
899 case '-':
900 iop = -1;
901 wstr++;
902 break;
903 default:
904 break;
905 }
906 if (a2roffsu(wstr, &su, SCALE_MAX) != NULL)
907 width = term_hspan(p, &su);
908 else
909 iop = 0;
910 }
911 (*p->setwidth)(p, iop, width);
912 }
913
914 size_t
915 term_len(const struct termp *p, size_t sz)
916 {
917
918 return (*p->width)(p, ' ') * sz;
919 }
920
921 static size_t
922 cond_width(const struct termp *p, int c, int *skip)
923 {
924
925 if (*skip) {
926 (*skip) = 0;
927 return 0;
928 } else
929 return (*p->width)(p, c);
930 }
931
932 size_t
933 term_strlen(const struct termp *p, const char *cp)
934 {
935 size_t sz, rsz, i;
936 int ssz, skip, uc;
937 const char *seq, *rhs;
938 enum mandoc_esc esc;
939 static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH,
940 ASCII_BREAK, '\0' };
941
942 /*
943 * Account for escaped sequences within string length
944 * calculations. This follows the logic in term_word() as we
945 * must calculate the width of produced strings.
946 */
947
948 sz = 0;
949 skip = 0;
950 while ('\0' != *cp) {
951 rsz = strcspn(cp, rej);
952 for (i = 0; i < rsz; i++)
953 sz += cond_width(p, *cp++, &skip);
954
955 switch (*cp) {
956 case '\\':
957 cp++;
958 rhs = NULL;
959 esc = mandoc_escape(&cp, &seq, &ssz);
960 switch (esc) {
961 case ESCAPE_UNICODE:
962 uc = mchars_num2uc(seq + 1, ssz - 1);
963 break;
964 case ESCAPE_NUMBERED:
965 uc = mchars_num2char(seq, ssz);
966 if (uc < 0)
967 continue;
968 break;
969 case ESCAPE_SPECIAL:
970 if (p->enc == TERMENC_ASCII) {
971 rhs = mchars_spec2str(seq, ssz, &rsz);
972 if (rhs != NULL)
973 break;
974 } else {
975 uc = mchars_spec2cp(seq, ssz);
976 if (uc > 0)
977 sz += cond_width(p, uc, &skip);
978 }
979 continue;
980 case ESCAPE_UNDEF:
981 uc = *seq;
982 break;
983 case ESCAPE_DEVICE:
984 if (p->type == TERMTYPE_PDF) {
985 rhs = "pdf";
986 rsz = 3;
987 } else if (p->type == TERMTYPE_PS) {
988 rhs = "ps";
989 rsz = 2;
990 } else if (p->enc == TERMENC_ASCII) {
991 rhs = "ascii";
992 rsz = 5;
993 } else {
994 rhs = "utf8";
995 rsz = 4;
996 }
997 break;
998 case ESCAPE_SKIPCHAR:
999 skip = 1;
1000 continue;
1001 case ESCAPE_OVERSTRIKE:
1002 rsz = 0;
1003 rhs = seq + ssz;
1004 while (seq < rhs) {
1005 if (*seq == '\\') {
1006 mandoc_escape(&seq, NULL, NULL);
1007 continue;
1008 }
1009 i = (*p->width)(p, *seq++);
1010 if (rsz < i)
1011 rsz = i;
1012 }
1013 sz += rsz;
1014 continue;
1015 default:
1016 continue;
1017 }
1018
1019 /*
1020 * Common handling for Unicode and numbered
1021 * character escape sequences.
1022 */
1023
1024 if (rhs == NULL) {
1025 if (p->enc == TERMENC_ASCII) {
1026 rhs = ascii_uc2str(uc);
1027 rsz = strlen(rhs);
1028 } else {
1029 if ((uc < 0x20 && uc != 0x09) ||
1030 (uc > 0x7E && uc < 0xA0))
1031 uc = 0xFFFD;
1032 sz += cond_width(p, uc, &skip);
1033 continue;
1034 }
1035 }
1036
1037 if (skip) {
1038 skip = 0;
1039 break;
1040 }
1041
1042 /*
1043 * Common handling for all escape sequences
1044 * printing more than one character.
1045 */
1046
1047 for (i = 0; i < rsz; i++)
1048 sz += (*p->width)(p, *rhs++);
1049 break;
1050 case ASCII_NBRSP:
1051 sz += cond_width(p, ' ', &skip);
1052 cp++;
1053 break;
1054 case ASCII_HYPH:
1055 sz += cond_width(p, '-', &skip);
1056 cp++;
1057 break;
1058 default:
1059 break;
1060 }
1061 }
1062
1063 return sz;
1064 }
1065
1066 int
1067 term_vspan(const struct termp *p, const struct roffsu *su)
1068 {
1069 double r;
1070 int ri;
1071
1072 switch (su->unit) {
1073 case SCALE_BU:
1074 r = su->scale / 40.0;
1075 break;
1076 case SCALE_CM:
1077 r = su->scale * 6.0 / 2.54;
1078 break;
1079 case SCALE_FS:
1080 r = su->scale * 65536.0 / 40.0;
1081 break;
1082 case SCALE_IN:
1083 r = su->scale * 6.0;
1084 break;
1085 case SCALE_MM:
1086 r = su->scale * 0.006;
1087 break;
1088 case SCALE_PC:
1089 r = su->scale;
1090 break;
1091 case SCALE_PT:
1092 r = su->scale / 12.0;
1093 break;
1094 case SCALE_EN:
1095 case SCALE_EM:
1096 r = su->scale * 0.6;
1097 break;
1098 case SCALE_VS:
1099 r = su->scale;
1100 break;
1101 default:
1102 abort();
1103 }
1104 ri = r > 0.0 ? r + 0.4995 : r - 0.4995;
1105 return ri < 66 ? ri : 1;
1106 }
1107
1108 /*
1109 * Convert a scaling width to basic units, rounding towards 0.
1110 */
1111 int
1112 term_hspan(const struct termp *p, const struct roffsu *su)
1113 {
1114
1115 return (*p->hspan)(p, su);
1116 }
1117
1118 /*
1119 * Convert a scaling width to basic units, rounding to closest.
1120 */
1121 int
1122 term_hen(const struct termp *p, const struct roffsu *su)
1123 {
1124 int bu;
1125
1126 if ((bu = (*p->hspan)(p, su)) >= 0)
1127 return (bu + 11) / 24;
1128 else
1129 return -((-bu + 11) / 24);
1130 }
A git conversion mirror of mandoc.bsd.lv