Annotation of src/usr.bin/mandoc/term.c, Revision 1.144
1.144 ! schwarze 1: /* $OpenBSD: term.c,v 1.143 2021/08/10 12:36:42 schwarze Exp $ */
1.1 kristaps 2: /*
1.144 ! schwarze 3: * Copyright (c) 2010-2021 Ingo Schwarze <schwarze@openbsd.org>
1.59 schwarze 4: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.2 schwarze 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.106 schwarze 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.2 schwarze 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106 schwarze 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.2 schwarze 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.
1.1 kristaps 17: */
1.20 schwarze 18: #include <sys/types.h>
19:
1.1 kristaps 20: #include <assert.h>
1.20 schwarze 21: #include <ctype.h>
1.138 schwarze 22: #include <stdint.h>
1.1 kristaps 23: #include <stdio.h>
24: #include <stdlib.h>
25: #include <string.h>
26:
1.34 schwarze 27: #include "mandoc.h"
1.79 schwarze 28: #include "mandoc_aux.h"
1.16 schwarze 29: #include "out.h"
1.1 kristaps 30: #include "term.h"
1.16 schwarze 31: #include "main.h"
1.1 kristaps 32:
1.64 schwarze 33: static size_t cond_width(const struct termp *, int, int *);
1.126 schwarze 34: static void adjbuf(struct termp_col *, size_t);
1.59 schwarze 35: static void bufferc(struct termp *, char);
36: static void encode(struct termp *, const char *, size_t);
37: static void encode1(struct termp *, int);
1.124 schwarze 38: static void endline(struct termp *);
1.142 schwarze 39: static void term_field(struct termp *, size_t, size_t);
1.138 schwarze 40: static void term_fill(struct termp *, size_t *, size_t *,
41: size_t);
1.1 kristaps 42:
1.83 schwarze 43:
1.37 schwarze 44: void
1.129 schwarze 45: term_setcol(struct termp *p, size_t maxtcol)
46: {
47: if (maxtcol > p->maxtcol) {
48: p->tcols = mandoc_recallocarray(p->tcols,
49: p->maxtcol, maxtcol, sizeof(*p->tcols));
50: p->maxtcol = maxtcol;
51: }
52: p->lasttcol = maxtcol - 1;
53: p->tcol = p->tcols;
54: }
55:
56: void
1.37 schwarze 57: term_free(struct termp *p)
1.1 kristaps 58: {
1.144 ! schwarze 59: term_tab_free();
1.126 schwarze 60: for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++)
61: free(p->tcol->buf);
62: free(p->tcols);
1.98 schwarze 63: free(p->fontq);
1.37 schwarze 64: free(p);
1.1 kristaps 65: }
66:
1.13 schwarze 67: void
1.83 schwarze 68: term_begin(struct termp *p, term_margin head,
1.106 schwarze 69: term_margin foot, const struct roff_meta *arg)
1.1 kristaps 70: {
71:
1.37 schwarze 72: p->headf = head;
73: p->footf = foot;
74: p->argf = arg;
75: (*p->begin)(p);
1.1 kristaps 76: }
77:
1.37 schwarze 78: void
79: term_end(struct termp *p)
1.1 kristaps 80: {
81:
1.37 schwarze 82: (*p->end)(p);
1.1 kristaps 83: }
84:
85: /*
1.82 schwarze 86: * Flush a chunk of text. By default, break the output line each time
87: * the right margin is reached, and continue output on the next line
88: * at the same offset as the chunk itself. By default, also break the
1.138 schwarze 89: * output line at the end of the chunk. There are many flags modifying
90: * this behaviour, see the comments in the body of the function.
1.1 kristaps 91: */
92: void
93: term_flushln(struct termp *p)
94: {
1.138 schwarze 95: size_t vbl; /* Number of blanks to prepend to the output. */
96: size_t vbr; /* Actual visual position of the end of field. */
97: size_t vfield; /* Desired visual field width. */
98: size_t vtarget; /* Desired visual position of the right margin. */
99: size_t ic; /* Character position in the input buffer. */
100: size_t nbr; /* Number of characters to print in this field. */
101:
102: /*
103: * Normally, start writing at the left margin, but with the
104: * NOPAD flag, start writing at the current position instead.
105: */
1.1 kristaps 106:
1.126 schwarze 107: vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ?
108: 0 : p->tcol->offset - p->viscol;
1.123 schwarze 109: if (p->minbl && vbl < p->minbl)
110: vbl = p->minbl;
1.19 schwarze 111:
1.134 florian 112: if ((p->flags & TERMP_MULTICOL) == 0)
1.127 schwarze 113: p->tcol->col = 0;
1.138 schwarze 114:
115: /* Loop over output lines. */
116:
117: for (;;) {
118: vfield = p->tcol->rmargin > p->viscol + vbl ?
119: p->tcol->rmargin - p->viscol - vbl : 0;
1.127 schwarze 120:
1.22 schwarze 121: /*
1.138 schwarze 122: * Normally, break the line at the the right margin
123: * of the field, but with the NOBREAK flag, only
124: * break it at the max right margin of the screen,
125: * and with the BRNEVER flag, never break it at all.
1.30 schwarze 126: */
1.127 schwarze 127:
1.142 schwarze 128: vtarget = (p->flags & TERMP_NOBREAK) == 0 ? vfield :
1.138 schwarze 129: p->maxrmargin > p->viscol + vbl ?
130: p->maxrmargin - p->viscol - vbl : 0;
131:
132: /*
133: * Figure out how much text will fit in the field.
134: * If there is whitespace only, print nothing.
135: */
136:
1.142 schwarze 137: term_fill(p, &nbr, &vbr,
138: p->flags & TERMP_BRNEVER ? SIZE_MAX : vtarget);
1.138 schwarze 139: if (nbr == 0)
140: break;
1.139 schwarze 141:
142: /*
143: * With the CENTER or RIGHT flag, increase the indentation
144: * to center the text between the left and right margins
145: * or to adjust it to the right margin, respectively.
146: */
147:
148: if (vbr < vtarget) {
149: if (p->flags & TERMP_CENTER)
150: vbl += (vtarget - vbr) / 2;
151: else if (p->flags & TERMP_RIGHT)
152: vbl += vtarget - vbr;
153: }
154:
155: /* Finally, print the field content. */
1.138 schwarze 156:
1.142 schwarze 157: term_field(p, vbl, nbr);
1.22 schwarze 158:
1.1 kristaps 159: /*
1.138 schwarze 160: * If there is no text left in the field, exit the loop.
161: * If the BRTRSP flag is set, consider trailing
162: * whitespace significant when deciding whether
163: * the field fits or not.
1.1 kristaps 164: */
165:
1.138 schwarze 166: for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
167: switch (p->tcol->buf[ic]) {
168: case '\t':
169: if (p->flags & TERMP_BRTRSP)
170: vbr = term_tab_next(vbr);
171: continue;
172: case ' ':
173: if (p->flags & TERMP_BRTRSP)
174: vbr += (*p->width)(p, ' ');
1.130 schwarze 175: continue;
1.138 schwarze 176: case '\n':
177: case ASCII_BREAK:
178: continue;
179: default:
180: break;
1.130 schwarze 181: }
1.138 schwarze 182: break;
183: }
184: if (ic == p->tcol->lastcol)
185: break;
1.42 schwarze 186:
1.138 schwarze 187: /*
188: * At the location of an automtic line break, input
189: * space characters are consumed by the line break.
190: */
1.42 schwarze 191:
1.138 schwarze 192: while (p->tcol->col < p->tcol->lastcol &&
193: p->tcol->buf[p->tcol->col] == ' ')
194: p->tcol->col++;
1.42 schwarze 195:
1.138 schwarze 196: /*
197: * In multi-column mode, leave the rest of the text
198: * in the buffer to be handled by a subsequent
199: * invocation, such that the other columns of the
200: * table can be handled first.
201: * In single-column mode, simply break the line.
202: */
203:
204: if (p->flags & TERMP_MULTICOL)
205: return;
1.78 schwarze 206:
1.138 schwarze 207: endline(p);
208: p->viscol = 0;
1.1 kristaps 209:
210: /*
1.138 schwarze 211: * Normally, start the next line at the same indentation
212: * as this one, but with the BRIND flag, start it at the
213: * right margin instead. This is used together with
214: * NOBREAK for the tags in various kinds of tagged lists.
1.5 schwarze 215: */
1.127 schwarze 216:
1.138 schwarze 217: vbl = p->flags & TERMP_BRIND ?
218: p->tcol->rmargin : p->tcol->offset;
219: }
1.127 schwarze 220:
1.138 schwarze 221: /* Reset output state in preparation for the next field. */
1.66 schwarze 222:
1.138 schwarze 223: p->col = p->tcol->col = p->tcol->lastcol = 0;
224: p->minbl = p->trailspace;
225: p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE | TERMP_NOPAD);
1.120 schwarze 226:
1.138 schwarze 227: if (p->flags & TERMP_MULTICOL)
228: return;
1.120 schwarze 229:
1.138 schwarze 230: /*
231: * The HANG flag means that the next field
232: * always follows on the same line.
233: * The NOBREAK flag means that the next field
234: * follows on the same line unless the field was overrun.
235: * Normally, break the line at the end of each field.
236: */
1.66 schwarze 237:
1.138 schwarze 238: if ((p->flags & TERMP_HANG) == 0 &&
239: ((p->flags & TERMP_NOBREAK) == 0 ||
240: vbr + term_len(p, p->trailspace) > vfield))
241: endline(p);
242: }
1.30 schwarze 243:
1.138 schwarze 244: /*
245: * Store the number of input characters to print in this field in *nbr
246: * and their total visual width to print in *vbr.
247: * If there is only whitespace in the field, both remain zero.
248: * The desired visual width of the field is provided by vtarget.
249: * If the first word is longer, the field will be overrun.
250: */
251: static void
252: term_fill(struct termp *p, size_t *nbr, size_t *vbr, size_t vtarget)
253: {
254: size_t ic; /* Character position in the input buffer. */
255: size_t vis; /* Visual position of the current character. */
256: size_t vn; /* Visual position of the next character. */
257: int breakline; /* Break at the end of this word. */
258: int graph; /* Last character was non-blank. */
259:
260: *nbr = *vbr = vis = 0;
261: breakline = graph = 0;
262: for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
263: switch (p->tcol->buf[ic]) {
264: case '\b': /* Escape \o (overstrike) or backspace markup. */
265: assert(ic > 0);
266: vis -= (*p->width)(p, p->tcol->buf[ic - 1]);
267: continue;
1.127 schwarze 268:
1.138 schwarze 269: case '\t': /* Normal ASCII whitespace. */
270: case ' ':
271: case ASCII_BREAK: /* Escape \: (breakpoint). */
272: switch (p->tcol->buf[ic]) {
273: case '\t':
274: vn = term_tab_next(vis);
1.30 schwarze 275: break;
1.138 schwarze 276: case ' ':
277: vn = vis + (*p->width)(p, ' ');
1.1 kristaps 278: break;
1.138 schwarze 279: case ASCII_BREAK:
280: vn = vis;
1.22 schwarze 281: break;
1.141 schwarze 282: default:
283: abort();
1.22 schwarze 284: }
1.138 schwarze 285: /* Can break at the end of a word. */
286: if (breakline || vn > vtarget)
287: break;
288: if (graph) {
289: *nbr = ic;
290: *vbr = vis;
291: graph = 0;
1.33 schwarze 292: }
1.138 schwarze 293: vis = vn;
294: continue;
295:
296: case '\n': /* Escape \p (break at the end of the word). */
297: breakline = 1;
298: continue;
1.33 schwarze 299:
1.138 schwarze 300: case ASCII_HYPH: /* Breakable hyphen. */
301: graph = 1;
1.33 schwarze 302: /*
1.138 schwarze 303: * We are about to decide whether to break the
304: * line or not, so we no longer need this hyphen
305: * to be marked as breakable. Put back a real
306: * hyphen such that we get the correct width.
1.33 schwarze 307: */
1.138 schwarze 308: p->tcol->buf[ic] = '-';
309: vis += (*p->width)(p, '-');
310: if (vis > vtarget) {
311: ic++;
312: break;
1.61 schwarze 313: }
1.138 schwarze 314: *nbr = ic + 1;
315: *vbr = vis;
316: continue;
1.61 schwarze 317:
1.138 schwarze 318: case ASCII_NBRSP: /* Non-breakable space. */
319: p->tcol->buf[ic] = ' ';
320: /* FALLTHROUGH */
321: default: /* Printable character. */
322: graph = 1;
323: vis += (*p->width)(p, p->tcol->buf[ic]);
324: if (vis > vtarget && *nbr > 0)
325: return;
326: continue;
1.1 kristaps 327: }
1.138 schwarze 328: break;
329: }
1.130 schwarze 330:
1.138 schwarze 331: /*
332: * If the last word extends to the end of the field without any
333: * trailing whitespace, the loop could not check yet whether it
334: * can remain on this line. So do the check now.
335: */
1.130 schwarze 336:
1.138 schwarze 337: if (graph && (vis <= vtarget || *nbr == 0)) {
338: *nbr = ic;
339: *vbr = vis;
340: }
341: }
1.130 schwarze 342:
1.138 schwarze 343: /*
344: * Print the contents of one field
345: * with an indentation of vbl visual columns,
1.142 schwarze 346: * and an input string length of nbr characters.
1.138 schwarze 347: */
348: static void
1.142 schwarze 349: term_field(struct termp *p, size_t vbl, size_t nbr)
1.138 schwarze 350: {
351: size_t ic; /* Character position in the input buffer. */
352: size_t vis; /* Visual position of the current character. */
353: size_t dv; /* Visual width of the current character. */
354: size_t vn; /* Visual position of the next character. */
1.130 schwarze 355:
1.138 schwarze 356: vis = 0;
357: for (ic = p->tcol->col; ic < nbr; ic++) {
1.130 schwarze 358:
1.138 schwarze 359: /*
360: * To avoid the printing of trailing whitespace,
361: * do not print whitespace right away, only count it.
362: */
1.130 schwarze 363:
1.138 schwarze 364: switch (p->tcol->buf[ic]) {
365: case '\n':
366: case ASCII_BREAK:
367: continue;
368: case '\t':
369: vn = term_tab_next(vis);
370: vbl += vn - vis;
371: vis = vn;
372: continue;
373: case ' ':
374: case ASCII_NBRSP:
1.140 schwarze 375: dv = (*p->width)(p, ' ');
376: vbl += dv;
377: vis += dv;
1.138 schwarze 378: continue;
379: default:
380: break;
381: }
1.48 schwarze 382:
1.138 schwarze 383: /*
384: * We found a non-blank character to print,
385: * so write preceding white space now.
386: */
1.127 schwarze 387:
1.138 schwarze 388: if (vbl > 0) {
389: (*p->advance)(p, vbl);
390: p->viscol += vbl;
391: vbl = 0;
392: }
1.18 schwarze 393:
1.138 schwarze 394: /* Print the character and adjust the visual position. */
1.1 kristaps 395:
1.138 schwarze 396: (*p->letter)(p, p->tcol->buf[ic]);
397: if (p->tcol->buf[ic] == '\b') {
398: dv = (*p->width)(p, p->tcol->buf[ic - 1]);
399: p->viscol -= dv;
400: vis -= dv;
401: } else {
402: dv = (*p->width)(p, p->tcol->buf[ic]);
403: p->viscol += dv;
404: vis += dv;
405: }
406: }
407: p->tcol->col = nbr;
1.124 schwarze 408: }
409:
410: static void
411: endline(struct termp *p)
412: {
413: if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) {
414: p->mc = NULL;
415: p->flags &= ~TERMP_ENDMC;
416: }
417: if (p->mc != NULL) {
418: if (p->viscol && p->maxrmargin >= p->viscol)
419: (*p->advance)(p, p->maxrmargin - p->viscol + 1);
420: p->flags |= TERMP_NOBUF | TERMP_NOSPACE;
421: term_word(p, p->mc);
422: p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC);
423: }
424: p->viscol = 0;
425: p->minbl = 0;
426: (*p->endline)(p);
1.1 kristaps 427: }
428:
1.83 schwarze 429: /*
1.1 kristaps 430: * A newline only breaks an existing line; it won't assert vertical
431: * space. All data in the output buffer is flushed prior to the newline
432: * assertion.
433: */
434: void
435: term_newln(struct termp *p)
436: {
437:
438: p->flags |= TERMP_NOSPACE;
1.129 schwarze 439: if (p->tcol->lastcol || p->viscol)
1.61 schwarze 440: term_flushln(p);
1.1 kristaps 441: }
442:
443: /*
444: * Asserts a vertical space (a full, empty line-break between lines).
445: * Note that if used twice, this will cause two blank spaces and so on.
446: * All data in the output buffer is flushed prior to the newline
447: * assertion.
448: */
449: void
450: term_vspace(struct termp *p)
451: {
452:
453: term_newln(p);
1.29 schwarze 454: p->viscol = 0;
1.124 schwarze 455: p->minbl = 0;
1.63 schwarze 456: if (0 < p->skipvsp)
457: p->skipvsp--;
458: else
459: (*p->endline)(p);
1.1 kristaps 460: }
461:
1.98 schwarze 462: /* Swap current and previous font; for \fP and .ft P */
1.20 schwarze 463: void
464: term_fontlast(struct termp *p)
465: {
466: enum termfont f;
1.11 schwarze 467:
1.20 schwarze 468: f = p->fontl;
469: p->fontl = p->fontq[p->fonti];
470: p->fontq[p->fonti] = f;
471: }
472:
1.98 schwarze 473: /* Set font, save current, discard previous; for \f, .ft, .B etc. */
1.20 schwarze 474: void
475: term_fontrepl(struct termp *p, enum termfont f)
476: {
477:
478: p->fontl = p->fontq[p->fonti];
479: p->fontq[p->fonti] = f;
1.1 kristaps 480: }
481:
1.98 schwarze 482: /* Set font, save previous. */
1.20 schwarze 483: void
484: term_fontpush(struct termp *p, enum termfont f)
1.1 kristaps 485: {
1.7 schwarze 486:
1.20 schwarze 487: p->fontl = p->fontq[p->fonti];
1.98 schwarze 488: if (++p->fonti == p->fontsz) {
489: p->fontsz += 8;
490: p->fontq = mandoc_reallocarray(p->fontq,
1.116 schwarze 491: p->fontsz, sizeof(*p->fontq));
1.98 schwarze 492: }
493: p->fontq[p->fonti] = f;
1.20 schwarze 494: }
1.1 kristaps 495:
1.98 schwarze 496: /* Flush to make the saved pointer current again. */
1.20 schwarze 497: void
1.104 schwarze 498: term_fontpopq(struct termp *p, int i)
1.20 schwarze 499: {
1.1 kristaps 500:
1.104 schwarze 501: assert(i >= 0);
502: if (p->fonti > i)
503: p->fonti = i;
1.20 schwarze 504: }
1.1 kristaps 505:
1.98 schwarze 506: /* Pop one font off the stack. */
1.20 schwarze 507: void
508: term_fontpop(struct termp *p)
509: {
1.1 kristaps 510:
1.20 schwarze 511: assert(p->fonti);
512: p->fonti--;
1.1 kristaps 513: }
514:
515: /*
516: * Handle pwords, partial words, which may be either a single word or a
517: * phrase that cannot be broken down (such as a literal string). This
518: * handles word styling.
519: */
1.7 schwarze 520: void
521: term_word(struct termp *p, const char *word)
1.1 kristaps 522: {
1.121 schwarze 523: struct roffsu su;
1.75 schwarze 524: const char nbrsp[2] = { ASCII_NBRSP, 0 };
1.59 schwarze 525: const char *seq, *cp;
526: int sz, uc;
1.122 schwarze 527: size_t csz, lsz, ssz;
1.59 schwarze 528: enum mandoc_esc esc;
1.1 kristaps 529:
1.124 schwarze 530: if ((p->flags & TERMP_NOBUF) == 0) {
531: if ((p->flags & TERMP_NOSPACE) == 0) {
532: if ((p->flags & TERMP_KEEP) == 0) {
1.40 schwarze 533: bufferc(p, ' ');
1.124 schwarze 534: if (p->flags & TERMP_SENTENCE)
535: bufferc(p, ' ');
536: } else
537: bufferc(p, ASCII_NBRSP);
538: }
539: if (p->flags & TERMP_PREKEEP)
540: p->flags |= TERMP_KEEP;
541: if (p->flags & TERMP_NONOSPACE)
542: p->flags |= TERMP_NOSPACE;
543: else
544: p->flags &= ~TERMP_NOSPACE;
545: p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE);
546: p->skipvsp = 0;
1.31 schwarze 547: }
548:
1.59 schwarze 549: while ('\0' != *word) {
1.64 schwarze 550: if ('\\' != *word) {
1.75 schwarze 551: if (TERMP_NBRWORD & p->flags) {
552: if (' ' == *word) {
553: encode(p, nbrsp, 1);
554: word++;
555: continue;
556: }
557: ssz = strcspn(word, "\\ ");
558: } else
559: ssz = strcspn(word, "\\");
1.45 schwarze 560: encode(p, word, ssz);
1.64 schwarze 561: word += (int)ssz;
1.20 schwarze 562: continue;
1.64 schwarze 563: }
1.20 schwarze 564:
1.59 schwarze 565: word++;
566: esc = mandoc_escape(&word, &seq, &sz);
567: switch (esc) {
1.83 schwarze 568: case ESCAPE_UNICODE:
1.89 schwarze 569: uc = mchars_num2uc(seq + 1, sz - 1);
1.56 schwarze 570: break;
1.83 schwarze 571: case ESCAPE_NUMBERED:
1.93 schwarze 572: uc = mchars_num2char(seq, sz);
573: if (uc < 0)
574: continue;
1.20 schwarze 575: break;
1.83 schwarze 576: case ESCAPE_SPECIAL:
1.89 schwarze 577: if (p->enc == TERMENC_ASCII) {
1.114 schwarze 578: cp = mchars_spec2str(seq, sz, &ssz);
1.92 schwarze 579: if (cp != NULL)
1.89 schwarze 580: encode(p, cp, ssz);
581: } else {
1.114 schwarze 582: uc = mchars_spec2cp(seq, sz);
1.90 schwarze 583: if (uc > 0)
584: encode1(p, uc);
1.89 schwarze 585: }
1.93 schwarze 586: continue;
1.137 schwarze 587: case ESCAPE_UNDEF:
588: uc = *seq;
589: break;
1.83 schwarze 590: case ESCAPE_FONTBOLD:
1.143 schwarze 591: case ESCAPE_FONTCB:
1.20 schwarze 592: term_fontrepl(p, TERMFONT_BOLD);
1.93 schwarze 593: continue;
1.83 schwarze 594: case ESCAPE_FONTITALIC:
1.143 schwarze 595: case ESCAPE_FONTCI:
1.20 schwarze 596: term_fontrepl(p, TERMFONT_UNDER);
1.93 schwarze 597: continue;
1.83 schwarze 598: case ESCAPE_FONTBI:
1.70 schwarze 599: term_fontrepl(p, TERMFONT_BI);
1.93 schwarze 600: continue;
1.83 schwarze 601: case ESCAPE_FONT:
1.143 schwarze 602: case ESCAPE_FONTCR:
1.83 schwarze 603: case ESCAPE_FONTROMAN:
1.20 schwarze 604: term_fontrepl(p, TERMFONT_NONE);
1.93 schwarze 605: continue;
1.83 schwarze 606: case ESCAPE_FONTPREV:
1.20 schwarze 607: term_fontlast(p);
1.130 schwarze 608: continue;
609: case ESCAPE_BREAK:
610: bufferc(p, '\n');
1.93 schwarze 611: continue;
1.83 schwarze 612: case ESCAPE_NOSPACE:
1.108 schwarze 613: if (p->flags & TERMP_BACKAFTER)
614: p->flags &= ~TERMP_BACKAFTER;
615: else if (*word == '\0')
1.97 schwarze 616: p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE);
1.121 schwarze 617: continue;
1.135 schwarze 618: case ESCAPE_DEVICE:
619: if (p->type == TERMTYPE_PDF)
620: encode(p, "pdf", 3);
621: else if (p->type == TERMTYPE_PS)
622: encode(p, "ps", 2);
623: else if (p->enc == TERMENC_ASCII)
624: encode(p, "ascii", 5);
625: else
626: encode(p, "utf8", 4);
627: continue;
1.121 schwarze 628: case ESCAPE_HORIZ:
1.133 schwarze 629: if (*seq == '|') {
630: seq++;
631: uc = -p->col;
632: } else
633: uc = 0;
1.128 schwarze 634: if (a2roffsu(seq, &su, SCALE_EM) == NULL)
1.121 schwarze 635: continue;
1.133 schwarze 636: uc += term_hen(p, &su);
1.121 schwarze 637: if (uc > 0)
638: while (uc-- > 0)
639: bufferc(p, ASCII_NBRSP);
640: else if (p->col > (size_t)(-uc))
641: p->col += uc;
642: else {
643: uc += p->col;
644: p->col = 0;
1.126 schwarze 645: if (p->tcol->offset > (size_t)(-uc)) {
1.121 schwarze 646: p->ti += uc;
1.126 schwarze 647: p->tcol->offset += uc;
1.121 schwarze 648: } else {
1.126 schwarze 649: p->ti -= p->tcol->offset;
650: p->tcol->offset = 0;
1.121 schwarze 651: }
1.122 schwarze 652: }
653: continue;
654: case ESCAPE_HLINE:
1.132 schwarze 655: if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL)
1.122 schwarze 656: continue;
1.131 schwarze 657: uc = term_hen(p, &su);
1.122 schwarze 658: if (uc <= 0) {
1.126 schwarze 659: if (p->tcol->rmargin <= p->tcol->offset)
1.122 schwarze 660: continue;
1.126 schwarze 661: lsz = p->tcol->rmargin - p->tcol->offset;
1.122 schwarze 662: } else
663: lsz = uc;
1.132 schwarze 664: if (*cp == seq[-1])
1.122 schwarze 665: uc = -1;
1.132 schwarze 666: else if (*cp == '\\') {
667: seq = cp + 1;
1.122 schwarze 668: esc = mandoc_escape(&seq, &cp, &sz);
669: switch (esc) {
670: case ESCAPE_UNICODE:
671: uc = mchars_num2uc(cp + 1, sz - 1);
672: break;
673: case ESCAPE_NUMBERED:
674: uc = mchars_num2char(cp, sz);
675: break;
676: case ESCAPE_SPECIAL:
677: uc = mchars_spec2cp(cp, sz);
678: break;
1.137 schwarze 679: case ESCAPE_UNDEF:
680: uc = *seq;
681: break;
1.122 schwarze 682: default:
683: uc = -1;
684: break;
685: }
686: } else
1.132 schwarze 687: uc = *cp;
1.122 schwarze 688: if (uc < 0x20 || (uc > 0x7E && uc < 0xA0))
689: uc = '_';
690: if (p->enc == TERMENC_ASCII) {
691: cp = ascii_uc2str(uc);
692: csz = term_strlen(p, cp);
693: ssz = strlen(cp);
694: } else
695: csz = (*p->width)(p, uc);
696: while (lsz >= csz) {
697: if (p->enc == TERMENC_ASCII)
698: encode(p, cp, ssz);
699: else
700: encode1(p, uc);
701: lsz -= csz;
1.121 schwarze 702: }
1.93 schwarze 703: continue;
1.83 schwarze 704: case ESCAPE_SKIPCHAR:
1.108 schwarze 705: p->flags |= TERMP_BACKAFTER;
1.93 schwarze 706: continue;
1.103 schwarze 707: case ESCAPE_OVERSTRIKE:
708: cp = seq + sz;
709: while (seq < cp) {
710: if (*seq == '\\') {
711: mandoc_escape(&seq, NULL, NULL);
712: continue;
713: }
714: encode1(p, *seq++);
1.108 schwarze 715: if (seq < cp) {
716: if (p->flags & TERMP_BACKBEFORE)
717: p->flags |= TERMP_BACKAFTER;
718: else
719: p->flags |= TERMP_BACKBEFORE;
720: }
1.103 schwarze 721: }
1.109 schwarze 722: /* Trim trailing backspace/blank pair. */
1.129 schwarze 723: if (p->tcol->lastcol > 2 &&
724: (p->tcol->buf[p->tcol->lastcol - 1] == ' ' ||
725: p->tcol->buf[p->tcol->lastcol - 1] == '\t'))
726: p->tcol->lastcol -= 2;
727: if (p->col > p->tcol->lastcol)
728: p->col = p->tcol->lastcol;
1.108 schwarze 729: continue;
1.20 schwarze 730: default:
1.93 schwarze 731: continue;
732: }
733:
734: /*
735: * Common handling for Unicode and numbered
736: * character escape sequences.
737: */
738:
739: if (p->enc == TERMENC_ASCII) {
740: cp = ascii_uc2str(uc);
741: encode(p, cp, strlen(cp));
742: } else {
743: if ((uc < 0x20 && uc != 0x09) ||
744: (uc > 0x7E && uc < 0xA0))
745: uc = 0xFFFD;
746: encode1(p, uc);
1.20 schwarze 747: }
748: }
1.75 schwarze 749: p->flags &= ~TERMP_NBRWORD;
1.1 kristaps 750: }
751:
752: static void
1.126 schwarze 753: adjbuf(struct termp_col *c, size_t sz)
1.1 kristaps 754: {
1.126 schwarze 755: if (c->maxcols == 0)
756: c->maxcols = 1024;
757: while (c->maxcols <= sz)
758: c->maxcols <<= 2;
759: c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf));
1.1 kristaps 760: }
761:
1.4 schwarze 762: static void
1.20 schwarze 763: bufferc(struct termp *p, char c)
764: {
1.124 schwarze 765: if (p->flags & TERMP_NOBUF) {
766: (*p->letter)(p, c);
767: return;
768: }
1.126 schwarze 769: if (p->col + 1 >= p->tcol->maxcols)
770: adjbuf(p->tcol, p->col + 1);
1.129 schwarze 771: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 772: p->tcol->buf[p->col] = c;
1.129 schwarze 773: if (p->tcol->lastcol < ++p->col)
774: p->tcol->lastcol = p->col;
1.20 schwarze 775: }
776:
1.59 schwarze 777: /*
778: * See encode().
779: * Do this for a single (probably unicode) value.
780: * Does not check for non-decorated glyphs.
781: */
782: static void
783: encode1(struct termp *p, int c)
784: {
785: enum termfont f;
786:
1.124 schwarze 787: if (p->flags & TERMP_NOBUF) {
788: (*p->letter)(p, c);
789: return;
790: }
791:
1.126 schwarze 792: if (p->col + 7 >= p->tcol->maxcols)
793: adjbuf(p->tcol, p->col + 7);
1.59 schwarze 794:
1.115 schwarze 795: f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ?
1.108 schwarze 796: p->fontq[p->fonti] : TERMFONT_NONE;
1.59 schwarze 797:
1.108 schwarze 798: if (p->flags & TERMP_BACKBEFORE) {
1.126 schwarze 799: if (p->tcol->buf[p->col - 1] == ' ' ||
800: p->tcol->buf[p->col - 1] == '\t')
1.109 schwarze 801: p->col--;
802: else
1.126 schwarze 803: p->tcol->buf[p->col++] = '\b';
1.108 schwarze 804: p->flags &= ~TERMP_BACKBEFORE;
805: }
1.126 schwarze 806: if (f == TERMFONT_UNDER || f == TERMFONT_BI) {
807: p->tcol->buf[p->col++] = '_';
808: p->tcol->buf[p->col++] = '\b';
809: }
810: if (f == TERMFONT_BOLD || f == TERMFONT_BI) {
811: if (c == ASCII_HYPH)
812: p->tcol->buf[p->col++] = '-';
1.70 schwarze 813: else
1.126 schwarze 814: p->tcol->buf[p->col++] = c;
815: p->tcol->buf[p->col++] = '\b';
1.70 schwarze 816: }
1.129 schwarze 817: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 818: p->tcol->buf[p->col] = c;
1.129 schwarze 819: if (p->tcol->lastcol < ++p->col)
820: p->tcol->lastcol = p->col;
1.108 schwarze 821: if (p->flags & TERMP_BACKAFTER) {
822: p->flags |= TERMP_BACKBEFORE;
823: p->flags &= ~TERMP_BACKAFTER;
824: }
1.59 schwarze 825: }
1.20 schwarze 826:
827: static void
828: encode(struct termp *p, const char *word, size_t sz)
1.4 schwarze 829: {
1.71 schwarze 830: size_t i;
1.124 schwarze 831:
832: if (p->flags & TERMP_NOBUF) {
833: for (i = 0; i < sz; i++)
834: (*p->letter)(p, word[i]);
835: return;
836: }
1.59 schwarze 837:
1.126 schwarze 838: if (p->col + 2 + (sz * 5) >= p->tcol->maxcols)
839: adjbuf(p->tcol, p->col + 2 + (sz * 5));
1.46 schwarze 840:
1.71 schwarze 841: for (i = 0; i < sz; i++) {
1.70 schwarze 842: if (ASCII_HYPH == word[i] ||
843: isgraph((unsigned char)word[i]))
844: encode1(p, word[i]);
1.119 schwarze 845: else {
1.129 schwarze 846: if (p->tcol->lastcol <= p->col ||
1.125 schwarze 847: (word[i] != ' ' && word[i] != ASCII_NBRSP))
1.126 schwarze 848: p->tcol->buf[p->col] = word[i];
1.125 schwarze 849: p->col++;
1.119 schwarze 850:
851: /*
852: * Postpone the effect of \z while handling
853: * an overstrike sequence from ascii_uc2str().
854: */
855:
856: if (word[i] == '\b' &&
857: (p->flags & TERMP_BACKBEFORE)) {
858: p->flags &= ~TERMP_BACKBEFORE;
859: p->flags |= TERMP_BACKAFTER;
860: }
861: }
1.4 schwarze 862: }
1.129 schwarze 863: if (p->tcol->lastcol < p->col)
864: p->tcol->lastcol = p->col;
1.80 schwarze 865: }
866:
867: void
868: term_setwidth(struct termp *p, const char *wstr)
869: {
870: struct roffsu su;
1.107 schwarze 871: int iop, width;
1.80 schwarze 872:
1.81 schwarze 873: iop = 0;
874: width = 0;
1.80 schwarze 875: if (NULL != wstr) {
876: switch (*wstr) {
1.83 schwarze 877: case '+':
1.80 schwarze 878: iop = 1;
879: wstr++;
880: break;
1.83 schwarze 881: case '-':
1.80 schwarze 882: iop = -1;
883: wstr++;
884: break;
885: default:
886: break;
887: }
1.128 schwarze 888: if (a2roffsu(wstr, &su, SCALE_MAX) != NULL)
1.81 schwarze 889: width = term_hspan(p, &su);
890: else
1.80 schwarze 891: iop = 0;
892: }
893: (*p->setwidth)(p, iop, width);
1.4 schwarze 894: }
1.16 schwarze 895:
896: size_t
1.39 schwarze 897: term_len(const struct termp *p, size_t sz)
898: {
899:
1.112 schwarze 900: return (*p->width)(p, ' ') * sz;
1.39 schwarze 901: }
902:
1.64 schwarze 903: static size_t
904: cond_width(const struct termp *p, int c, int *skip)
905: {
906:
907: if (*skip) {
908: (*skip) = 0;
1.112 schwarze 909: return 0;
1.64 schwarze 910: } else
1.112 schwarze 911: return (*p->width)(p, c);
1.64 schwarze 912: }
1.39 schwarze 913:
914: size_t
915: term_strlen(const struct termp *p, const char *cp)
916: {
1.59 schwarze 917: size_t sz, rsz, i;
1.93 schwarze 918: int ssz, skip, uc;
1.50 schwarze 919: const char *seq, *rhs;
1.59 schwarze 920: enum mandoc_esc esc;
1.77 schwarze 921: static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH,
922: ASCII_BREAK, '\0' };
1.59 schwarze 923:
924: /*
925: * Account for escaped sequences within string length
926: * calculations. This follows the logic in term_word() as we
927: * must calculate the width of produced strings.
928: */
929:
930: sz = 0;
1.64 schwarze 931: skip = 0;
1.59 schwarze 932: while ('\0' != *cp) {
933: rsz = strcspn(cp, rej);
934: for (i = 0; i < rsz; i++)
1.64 schwarze 935: sz += cond_width(p, *cp++, &skip);
1.59 schwarze 936:
937: switch (*cp) {
1.83 schwarze 938: case '\\':
1.59 schwarze 939: cp++;
1.137 schwarze 940: rhs = NULL;
1.59 schwarze 941: esc = mandoc_escape(&cp, &seq, &ssz);
942: switch (esc) {
1.83 schwarze 943: case ESCAPE_UNICODE:
1.94 schwarze 944: uc = mchars_num2uc(seq + 1, ssz - 1);
1.59 schwarze 945: break;
1.83 schwarze 946: case ESCAPE_NUMBERED:
1.93 schwarze 947: uc = mchars_num2char(seq, ssz);
948: if (uc < 0)
949: continue;
1.50 schwarze 950: break;
1.83 schwarze 951: case ESCAPE_SPECIAL:
1.93 schwarze 952: if (p->enc == TERMENC_ASCII) {
1.114 schwarze 953: rhs = mchars_spec2str(seq, ssz, &rsz);
1.93 schwarze 954: if (rhs != NULL)
955: break;
956: } else {
1.114 schwarze 957: uc = mchars_spec2cp(seq, ssz);
1.93 schwarze 958: if (uc > 0)
959: sz += cond_width(p, uc, &skip);
1.89 schwarze 960: }
1.93 schwarze 961: continue;
1.137 schwarze 962: case ESCAPE_UNDEF:
963: uc = *seq;
964: break;
1.135 schwarze 965: case ESCAPE_DEVICE:
966: if (p->type == TERMTYPE_PDF) {
967: rhs = "pdf";
968: rsz = 3;
969: } else if (p->type == TERMTYPE_PS) {
970: rhs = "ps";
971: rsz = 2;
972: } else if (p->enc == TERMENC_ASCII) {
973: rhs = "ascii";
974: rsz = 5;
975: } else {
976: rhs = "utf8";
977: rsz = 4;
978: }
979: break;
1.83 schwarze 980: case ESCAPE_SKIPCHAR:
1.64 schwarze 981: skip = 1;
1.103 schwarze 982: continue;
983: case ESCAPE_OVERSTRIKE:
984: rsz = 0;
985: rhs = seq + ssz;
986: while (seq < rhs) {
987: if (*seq == '\\') {
988: mandoc_escape(&seq, NULL, NULL);
989: continue;
990: }
991: i = (*p->width)(p, *seq++);
992: if (rsz < i)
993: rsz = i;
994: }
995: sz += rsz;
1.93 schwarze 996: continue;
1.50 schwarze 997: default:
1.93 schwarze 998: continue;
1.50 schwarze 999: }
1.39 schwarze 1000:
1.93 schwarze 1001: /*
1002: * Common handling for Unicode and numbered
1003: * character escape sequences.
1004: */
1005:
1006: if (rhs == NULL) {
1007: if (p->enc == TERMENC_ASCII) {
1008: rhs = ascii_uc2str(uc);
1009: rsz = strlen(rhs);
1010: } else {
1011: if ((uc < 0x20 && uc != 0x09) ||
1012: (uc > 0x7E && uc < 0xA0))
1013: uc = 0xFFFD;
1014: sz += cond_width(p, uc, &skip);
1015: continue;
1016: }
1017: }
1.59 schwarze 1018:
1.64 schwarze 1019: if (skip) {
1020: skip = 0;
1021: break;
1022: }
1.93 schwarze 1023:
1024: /*
1025: * Common handling for all escape sequences
1026: * printing more than one character.
1027: */
1.64 schwarze 1028:
1.59 schwarze 1029: for (i = 0; i < rsz; i++)
1030: sz += (*p->width)(p, *rhs++);
1031: break;
1.83 schwarze 1032: case ASCII_NBRSP:
1.64 schwarze 1033: sz += cond_width(p, ' ', &skip);
1.55 schwarze 1034: cp++;
1.59 schwarze 1035: break;
1.83 schwarze 1036: case ASCII_HYPH:
1.64 schwarze 1037: sz += cond_width(p, '-', &skip);
1.55 schwarze 1038: cp++;
1.59 schwarze 1039: break;
1040: default:
1041: break;
1042: }
1043: }
1.39 schwarze 1044:
1.112 schwarze 1045: return sz;
1.39 schwarze 1046: }
1047:
1.100 schwarze 1048: int
1.39 schwarze 1049: term_vspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 1050: {
1051: double r;
1.101 schwarze 1052: int ri;
1.16 schwarze 1053:
1054: switch (su->unit) {
1.99 schwarze 1055: case SCALE_BU:
1056: r = su->scale / 40.0;
1057: break;
1.83 schwarze 1058: case SCALE_CM:
1.99 schwarze 1059: r = su->scale * 6.0 / 2.54;
1060: break;
1061: case SCALE_FS:
1062: r = su->scale * 65536.0 / 40.0;
1.16 schwarze 1063: break;
1.83 schwarze 1064: case SCALE_IN:
1.86 schwarze 1065: r = su->scale * 6.0;
1.16 schwarze 1066: break;
1.99 schwarze 1067: case SCALE_MM:
1068: r = su->scale * 0.006;
1069: break;
1.83 schwarze 1070: case SCALE_PC:
1.16 schwarze 1071: r = su->scale;
1072: break;
1.83 schwarze 1073: case SCALE_PT:
1.99 schwarze 1074: r = su->scale / 12.0;
1.16 schwarze 1075: break;
1.99 schwarze 1076: case SCALE_EN:
1077: case SCALE_EM:
1078: r = su->scale * 0.6;
1.16 schwarze 1079: break;
1.83 schwarze 1080: case SCALE_VS:
1.16 schwarze 1081: r = su->scale;
1082: break;
1083: default:
1.99 schwarze 1084: abort();
1.16 schwarze 1085: }
1.101 schwarze 1086: ri = r > 0.0 ? r + 0.4995 : r - 0.4995;
1.112 schwarze 1087: return ri < 66 ? ri : 1;
1.16 schwarze 1088: }
1089:
1.107 schwarze 1090: /*
1.131 schwarze 1091: * Convert a scaling width to basic units, rounding towards 0.
1.107 schwarze 1092: */
1.100 schwarze 1093: int
1.39 schwarze 1094: term_hspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 1095: {
1096:
1.112 schwarze 1097: return (*p->hspan)(p, su);
1.131 schwarze 1098: }
1099:
1100: /*
1101: * Convert a scaling width to basic units, rounding to closest.
1102: */
1103: int
1104: term_hen(const struct termp *p, const struct roffsu *su)
1105: {
1106: int bu;
1107:
1108: if ((bu = (*p->hspan)(p, su)) >= 0)
1109: return (bu + 11) / 24;
1110: else
1111: return -((-bu + 11) / 24);
1.16 schwarze 1112: }