Annotation of src/usr.bin/mandoc/term.c, Revision 1.140
1.140 ! schwarze 1: /* $OpenBSD: term.c,v 1.139 2019/01/04 03:20:44 schwarze Exp $ */
1.1 kristaps 2: /*
1.59 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.138 schwarze 4: * Copyright (c) 2010-2019 Ingo Schwarze <schwarze@openbsd.org>
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 *);
1.138 schwarze 39: static void term_field(struct termp *, size_t, size_t,
40: size_t, size_t);
41: static void term_fill(struct termp *, size_t *, size_t *,
42: size_t);
1.1 kristaps 43:
1.83 schwarze 44:
1.37 schwarze 45: void
1.129 schwarze 46: term_setcol(struct termp *p, size_t maxtcol)
47: {
48: if (maxtcol > p->maxtcol) {
49: p->tcols = mandoc_recallocarray(p->tcols,
50: p->maxtcol, maxtcol, sizeof(*p->tcols));
51: p->maxtcol = maxtcol;
52: }
53: p->lasttcol = maxtcol - 1;
54: p->tcol = p->tcols;
55: }
56:
57: void
1.37 schwarze 58: term_free(struct termp *p)
1.1 kristaps 59: {
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:
1.138 schwarze 128: vtarget = p->flags & TERMP_BRNEVER ? SIZE_MAX :
129: (p->flags & TERMP_NOBREAK) == 0 ? vfield :
130: p->maxrmargin > p->viscol + vbl ?
131: p->maxrmargin - p->viscol - vbl : 0;
132:
133: /*
134: * Figure out how much text will fit in the field.
135: * If there is whitespace only, print nothing.
136: */
137:
138: term_fill(p, &nbr, &vbr, vtarget);
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:
157: term_field(p, vbl, nbr, vbr, vtarget);
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;
282: }
1.138 schwarze 283: /* Can break at the end of a word. */
284: if (breakline || vn > vtarget)
285: break;
286: if (graph) {
287: *nbr = ic;
288: *vbr = vis;
289: graph = 0;
1.33 schwarze 290: }
1.138 schwarze 291: vis = vn;
292: continue;
293:
294: case '\n': /* Escape \p (break at the end of the word). */
295: breakline = 1;
296: continue;
1.33 schwarze 297:
1.138 schwarze 298: case ASCII_HYPH: /* Breakable hyphen. */
299: graph = 1;
1.33 schwarze 300: /*
1.138 schwarze 301: * We are about to decide whether to break the
302: * line or not, so we no longer need this hyphen
303: * to be marked as breakable. Put back a real
304: * hyphen such that we get the correct width.
1.33 schwarze 305: */
1.138 schwarze 306: p->tcol->buf[ic] = '-';
307: vis += (*p->width)(p, '-');
308: if (vis > vtarget) {
309: ic++;
310: break;
1.61 schwarze 311: }
1.138 schwarze 312: *nbr = ic + 1;
313: *vbr = vis;
314: continue;
1.61 schwarze 315:
1.138 schwarze 316: case ASCII_NBRSP: /* Non-breakable space. */
317: p->tcol->buf[ic] = ' ';
318: /* FALLTHROUGH */
319: default: /* Printable character. */
320: graph = 1;
321: vis += (*p->width)(p, p->tcol->buf[ic]);
322: if (vis > vtarget && *nbr > 0)
323: return;
324: continue;
1.1 kristaps 325: }
1.138 schwarze 326: break;
327: }
1.130 schwarze 328:
1.138 schwarze 329: /*
330: * If the last word extends to the end of the field without any
331: * trailing whitespace, the loop could not check yet whether it
332: * can remain on this line. So do the check now.
333: */
1.130 schwarze 334:
1.138 schwarze 335: if (graph && (vis <= vtarget || *nbr == 0)) {
336: *nbr = ic;
337: *vbr = vis;
338: }
339: }
1.130 schwarze 340:
1.138 schwarze 341: /*
342: * Print the contents of one field
343: * with an indentation of vbl visual columns,
344: * an input string length of nbr characters,
345: * an output width of vbr visual columns,
346: * and a desired field width of vtarget visual columns.
347: */
348: static void
349: term_field(struct termp *p, size_t vbl, size_t nbr, size_t vbr, size_t vtarget)
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:
1.20 schwarze 591: term_fontrepl(p, TERMFONT_BOLD);
1.93 schwarze 592: continue;
1.83 schwarze 593: case ESCAPE_FONTITALIC:
1.20 schwarze 594: term_fontrepl(p, TERMFONT_UNDER);
1.93 schwarze 595: continue;
1.83 schwarze 596: case ESCAPE_FONTBI:
1.70 schwarze 597: term_fontrepl(p, TERMFONT_BI);
1.93 schwarze 598: continue;
1.83 schwarze 599: case ESCAPE_FONT:
1.136 schwarze 600: case ESCAPE_FONTCW:
1.83 schwarze 601: case ESCAPE_FONTROMAN:
1.20 schwarze 602: term_fontrepl(p, TERMFONT_NONE);
1.93 schwarze 603: continue;
1.83 schwarze 604: case ESCAPE_FONTPREV:
1.20 schwarze 605: term_fontlast(p);
1.130 schwarze 606: continue;
607: case ESCAPE_BREAK:
608: bufferc(p, '\n');
1.93 schwarze 609: continue;
1.83 schwarze 610: case ESCAPE_NOSPACE:
1.108 schwarze 611: if (p->flags & TERMP_BACKAFTER)
612: p->flags &= ~TERMP_BACKAFTER;
613: else if (*word == '\0')
1.97 schwarze 614: p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE);
1.121 schwarze 615: continue;
1.135 schwarze 616: case ESCAPE_DEVICE:
617: if (p->type == TERMTYPE_PDF)
618: encode(p, "pdf", 3);
619: else if (p->type == TERMTYPE_PS)
620: encode(p, "ps", 2);
621: else if (p->enc == TERMENC_ASCII)
622: encode(p, "ascii", 5);
623: else
624: encode(p, "utf8", 4);
625: continue;
1.121 schwarze 626: case ESCAPE_HORIZ:
1.133 schwarze 627: if (*seq == '|') {
628: seq++;
629: uc = -p->col;
630: } else
631: uc = 0;
1.128 schwarze 632: if (a2roffsu(seq, &su, SCALE_EM) == NULL)
1.121 schwarze 633: continue;
1.133 schwarze 634: uc += term_hen(p, &su);
1.121 schwarze 635: if (uc > 0)
636: while (uc-- > 0)
637: bufferc(p, ASCII_NBRSP);
638: else if (p->col > (size_t)(-uc))
639: p->col += uc;
640: else {
641: uc += p->col;
642: p->col = 0;
1.126 schwarze 643: if (p->tcol->offset > (size_t)(-uc)) {
1.121 schwarze 644: p->ti += uc;
1.126 schwarze 645: p->tcol->offset += uc;
1.121 schwarze 646: } else {
1.126 schwarze 647: p->ti -= p->tcol->offset;
648: p->tcol->offset = 0;
1.121 schwarze 649: }
1.122 schwarze 650: }
651: continue;
652: case ESCAPE_HLINE:
1.132 schwarze 653: if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL)
1.122 schwarze 654: continue;
1.131 schwarze 655: uc = term_hen(p, &su);
1.122 schwarze 656: if (uc <= 0) {
1.126 schwarze 657: if (p->tcol->rmargin <= p->tcol->offset)
1.122 schwarze 658: continue;
1.126 schwarze 659: lsz = p->tcol->rmargin - p->tcol->offset;
1.122 schwarze 660: } else
661: lsz = uc;
1.132 schwarze 662: if (*cp == seq[-1])
1.122 schwarze 663: uc = -1;
1.132 schwarze 664: else if (*cp == '\\') {
665: seq = cp + 1;
1.122 schwarze 666: esc = mandoc_escape(&seq, &cp, &sz);
667: switch (esc) {
668: case ESCAPE_UNICODE:
669: uc = mchars_num2uc(cp + 1, sz - 1);
670: break;
671: case ESCAPE_NUMBERED:
672: uc = mchars_num2char(cp, sz);
673: break;
674: case ESCAPE_SPECIAL:
675: uc = mchars_spec2cp(cp, sz);
676: break;
1.137 schwarze 677: case ESCAPE_UNDEF:
678: uc = *seq;
679: break;
1.122 schwarze 680: default:
681: uc = -1;
682: break;
683: }
684: } else
1.132 schwarze 685: uc = *cp;
1.122 schwarze 686: if (uc < 0x20 || (uc > 0x7E && uc < 0xA0))
687: uc = '_';
688: if (p->enc == TERMENC_ASCII) {
689: cp = ascii_uc2str(uc);
690: csz = term_strlen(p, cp);
691: ssz = strlen(cp);
692: } else
693: csz = (*p->width)(p, uc);
694: while (lsz >= csz) {
695: if (p->enc == TERMENC_ASCII)
696: encode(p, cp, ssz);
697: else
698: encode1(p, uc);
699: lsz -= csz;
1.121 schwarze 700: }
1.93 schwarze 701: continue;
1.83 schwarze 702: case ESCAPE_SKIPCHAR:
1.108 schwarze 703: p->flags |= TERMP_BACKAFTER;
1.93 schwarze 704: continue;
1.103 schwarze 705: case ESCAPE_OVERSTRIKE:
706: cp = seq + sz;
707: while (seq < cp) {
708: if (*seq == '\\') {
709: mandoc_escape(&seq, NULL, NULL);
710: continue;
711: }
712: encode1(p, *seq++);
1.108 schwarze 713: if (seq < cp) {
714: if (p->flags & TERMP_BACKBEFORE)
715: p->flags |= TERMP_BACKAFTER;
716: else
717: p->flags |= TERMP_BACKBEFORE;
718: }
1.103 schwarze 719: }
1.109 schwarze 720: /* Trim trailing backspace/blank pair. */
1.129 schwarze 721: if (p->tcol->lastcol > 2 &&
722: (p->tcol->buf[p->tcol->lastcol - 1] == ' ' ||
723: p->tcol->buf[p->tcol->lastcol - 1] == '\t'))
724: p->tcol->lastcol -= 2;
725: if (p->col > p->tcol->lastcol)
726: p->col = p->tcol->lastcol;
1.108 schwarze 727: continue;
1.20 schwarze 728: default:
1.93 schwarze 729: continue;
730: }
731:
732: /*
733: * Common handling for Unicode and numbered
734: * character escape sequences.
735: */
736:
737: if (p->enc == TERMENC_ASCII) {
738: cp = ascii_uc2str(uc);
739: encode(p, cp, strlen(cp));
740: } else {
741: if ((uc < 0x20 && uc != 0x09) ||
742: (uc > 0x7E && uc < 0xA0))
743: uc = 0xFFFD;
744: encode1(p, uc);
1.20 schwarze 745: }
746: }
1.75 schwarze 747: p->flags &= ~TERMP_NBRWORD;
1.1 kristaps 748: }
749:
750: static void
1.126 schwarze 751: adjbuf(struct termp_col *c, size_t sz)
1.1 kristaps 752: {
1.126 schwarze 753: if (c->maxcols == 0)
754: c->maxcols = 1024;
755: while (c->maxcols <= sz)
756: c->maxcols <<= 2;
757: c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf));
1.1 kristaps 758: }
759:
1.4 schwarze 760: static void
1.20 schwarze 761: bufferc(struct termp *p, char c)
762: {
1.124 schwarze 763: if (p->flags & TERMP_NOBUF) {
764: (*p->letter)(p, c);
765: return;
766: }
1.126 schwarze 767: if (p->col + 1 >= p->tcol->maxcols)
768: adjbuf(p->tcol, p->col + 1);
1.129 schwarze 769: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 770: p->tcol->buf[p->col] = c;
1.129 schwarze 771: if (p->tcol->lastcol < ++p->col)
772: p->tcol->lastcol = p->col;
1.20 schwarze 773: }
774:
1.59 schwarze 775: /*
776: * See encode().
777: * Do this for a single (probably unicode) value.
778: * Does not check for non-decorated glyphs.
779: */
780: static void
781: encode1(struct termp *p, int c)
782: {
783: enum termfont f;
784:
1.124 schwarze 785: if (p->flags & TERMP_NOBUF) {
786: (*p->letter)(p, c);
787: return;
788: }
789:
1.126 schwarze 790: if (p->col + 7 >= p->tcol->maxcols)
791: adjbuf(p->tcol, p->col + 7);
1.59 schwarze 792:
1.115 schwarze 793: f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ?
1.108 schwarze 794: p->fontq[p->fonti] : TERMFONT_NONE;
1.59 schwarze 795:
1.108 schwarze 796: if (p->flags & TERMP_BACKBEFORE) {
1.126 schwarze 797: if (p->tcol->buf[p->col - 1] == ' ' ||
798: p->tcol->buf[p->col - 1] == '\t')
1.109 schwarze 799: p->col--;
800: else
1.126 schwarze 801: p->tcol->buf[p->col++] = '\b';
1.108 schwarze 802: p->flags &= ~TERMP_BACKBEFORE;
803: }
1.126 schwarze 804: if (f == TERMFONT_UNDER || f == TERMFONT_BI) {
805: p->tcol->buf[p->col++] = '_';
806: p->tcol->buf[p->col++] = '\b';
807: }
808: if (f == TERMFONT_BOLD || f == TERMFONT_BI) {
809: if (c == ASCII_HYPH)
810: p->tcol->buf[p->col++] = '-';
1.70 schwarze 811: else
1.126 schwarze 812: p->tcol->buf[p->col++] = c;
813: p->tcol->buf[p->col++] = '\b';
1.70 schwarze 814: }
1.129 schwarze 815: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 816: p->tcol->buf[p->col] = c;
1.129 schwarze 817: if (p->tcol->lastcol < ++p->col)
818: p->tcol->lastcol = p->col;
1.108 schwarze 819: if (p->flags & TERMP_BACKAFTER) {
820: p->flags |= TERMP_BACKBEFORE;
821: p->flags &= ~TERMP_BACKAFTER;
822: }
1.59 schwarze 823: }
1.20 schwarze 824:
825: static void
826: encode(struct termp *p, const char *word, size_t sz)
1.4 schwarze 827: {
1.71 schwarze 828: size_t i;
1.124 schwarze 829:
830: if (p->flags & TERMP_NOBUF) {
831: for (i = 0; i < sz; i++)
832: (*p->letter)(p, word[i]);
833: return;
834: }
1.59 schwarze 835:
1.126 schwarze 836: if (p->col + 2 + (sz * 5) >= p->tcol->maxcols)
837: adjbuf(p->tcol, p->col + 2 + (sz * 5));
1.46 schwarze 838:
1.71 schwarze 839: for (i = 0; i < sz; i++) {
1.70 schwarze 840: if (ASCII_HYPH == word[i] ||
841: isgraph((unsigned char)word[i]))
842: encode1(p, word[i]);
1.119 schwarze 843: else {
1.129 schwarze 844: if (p->tcol->lastcol <= p->col ||
1.125 schwarze 845: (word[i] != ' ' && word[i] != ASCII_NBRSP))
1.126 schwarze 846: p->tcol->buf[p->col] = word[i];
1.125 schwarze 847: p->col++;
1.119 schwarze 848:
849: /*
850: * Postpone the effect of \z while handling
851: * an overstrike sequence from ascii_uc2str().
852: */
853:
854: if (word[i] == '\b' &&
855: (p->flags & TERMP_BACKBEFORE)) {
856: p->flags &= ~TERMP_BACKBEFORE;
857: p->flags |= TERMP_BACKAFTER;
858: }
859: }
1.4 schwarze 860: }
1.129 schwarze 861: if (p->tcol->lastcol < p->col)
862: p->tcol->lastcol = p->col;
1.80 schwarze 863: }
864:
865: void
866: term_setwidth(struct termp *p, const char *wstr)
867: {
868: struct roffsu su;
1.107 schwarze 869: int iop, width;
1.80 schwarze 870:
1.81 schwarze 871: iop = 0;
872: width = 0;
1.80 schwarze 873: if (NULL != wstr) {
874: switch (*wstr) {
1.83 schwarze 875: case '+':
1.80 schwarze 876: iop = 1;
877: wstr++;
878: break;
1.83 schwarze 879: case '-':
1.80 schwarze 880: iop = -1;
881: wstr++;
882: break;
883: default:
884: break;
885: }
1.128 schwarze 886: if (a2roffsu(wstr, &su, SCALE_MAX) != NULL)
1.81 schwarze 887: width = term_hspan(p, &su);
888: else
1.80 schwarze 889: iop = 0;
890: }
891: (*p->setwidth)(p, iop, width);
1.4 schwarze 892: }
1.16 schwarze 893:
894: size_t
1.39 schwarze 895: term_len(const struct termp *p, size_t sz)
896: {
897:
1.112 schwarze 898: return (*p->width)(p, ' ') * sz;
1.39 schwarze 899: }
900:
1.64 schwarze 901: static size_t
902: cond_width(const struct termp *p, int c, int *skip)
903: {
904:
905: if (*skip) {
906: (*skip) = 0;
1.112 schwarze 907: return 0;
1.64 schwarze 908: } else
1.112 schwarze 909: return (*p->width)(p, c);
1.64 schwarze 910: }
1.39 schwarze 911:
912: size_t
913: term_strlen(const struct termp *p, const char *cp)
914: {
1.59 schwarze 915: size_t sz, rsz, i;
1.93 schwarze 916: int ssz, skip, uc;
1.50 schwarze 917: const char *seq, *rhs;
1.59 schwarze 918: enum mandoc_esc esc;
1.77 schwarze 919: static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH,
920: ASCII_BREAK, '\0' };
1.59 schwarze 921:
922: /*
923: * Account for escaped sequences within string length
924: * calculations. This follows the logic in term_word() as we
925: * must calculate the width of produced strings.
926: */
927:
928: sz = 0;
1.64 schwarze 929: skip = 0;
1.59 schwarze 930: while ('\0' != *cp) {
931: rsz = strcspn(cp, rej);
932: for (i = 0; i < rsz; i++)
1.64 schwarze 933: sz += cond_width(p, *cp++, &skip);
1.59 schwarze 934:
935: switch (*cp) {
1.83 schwarze 936: case '\\':
1.59 schwarze 937: cp++;
1.137 schwarze 938: rhs = NULL;
1.59 schwarze 939: esc = mandoc_escape(&cp, &seq, &ssz);
940: switch (esc) {
1.83 schwarze 941: case ESCAPE_UNICODE:
1.94 schwarze 942: uc = mchars_num2uc(seq + 1, ssz - 1);
1.59 schwarze 943: break;
1.83 schwarze 944: case ESCAPE_NUMBERED:
1.93 schwarze 945: uc = mchars_num2char(seq, ssz);
946: if (uc < 0)
947: continue;
1.50 schwarze 948: break;
1.83 schwarze 949: case ESCAPE_SPECIAL:
1.93 schwarze 950: if (p->enc == TERMENC_ASCII) {
1.114 schwarze 951: rhs = mchars_spec2str(seq, ssz, &rsz);
1.93 schwarze 952: if (rhs != NULL)
953: break;
954: } else {
1.114 schwarze 955: uc = mchars_spec2cp(seq, ssz);
1.93 schwarze 956: if (uc > 0)
957: sz += cond_width(p, uc, &skip);
1.89 schwarze 958: }
1.93 schwarze 959: continue;
1.137 schwarze 960: case ESCAPE_UNDEF:
961: uc = *seq;
962: break;
1.135 schwarze 963: case ESCAPE_DEVICE:
964: if (p->type == TERMTYPE_PDF) {
965: rhs = "pdf";
966: rsz = 3;
967: } else if (p->type == TERMTYPE_PS) {
968: rhs = "ps";
969: rsz = 2;
970: } else if (p->enc == TERMENC_ASCII) {
971: rhs = "ascii";
972: rsz = 5;
973: } else {
974: rhs = "utf8";
975: rsz = 4;
976: }
977: break;
1.83 schwarze 978: case ESCAPE_SKIPCHAR:
1.64 schwarze 979: skip = 1;
1.103 schwarze 980: continue;
981: case ESCAPE_OVERSTRIKE:
982: rsz = 0;
983: rhs = seq + ssz;
984: while (seq < rhs) {
985: if (*seq == '\\') {
986: mandoc_escape(&seq, NULL, NULL);
987: continue;
988: }
989: i = (*p->width)(p, *seq++);
990: if (rsz < i)
991: rsz = i;
992: }
993: sz += rsz;
1.93 schwarze 994: continue;
1.50 schwarze 995: default:
1.93 schwarze 996: continue;
1.50 schwarze 997: }
1.39 schwarze 998:
1.93 schwarze 999: /*
1000: * Common handling for Unicode and numbered
1001: * character escape sequences.
1002: */
1003:
1004: if (rhs == NULL) {
1005: if (p->enc == TERMENC_ASCII) {
1006: rhs = ascii_uc2str(uc);
1007: rsz = strlen(rhs);
1008: } else {
1009: if ((uc < 0x20 && uc != 0x09) ||
1010: (uc > 0x7E && uc < 0xA0))
1011: uc = 0xFFFD;
1012: sz += cond_width(p, uc, &skip);
1013: continue;
1014: }
1015: }
1.59 schwarze 1016:
1.64 schwarze 1017: if (skip) {
1018: skip = 0;
1019: break;
1020: }
1.93 schwarze 1021:
1022: /*
1023: * Common handling for all escape sequences
1024: * printing more than one character.
1025: */
1.64 schwarze 1026:
1.59 schwarze 1027: for (i = 0; i < rsz; i++)
1028: sz += (*p->width)(p, *rhs++);
1029: break;
1.83 schwarze 1030: case ASCII_NBRSP:
1.64 schwarze 1031: sz += cond_width(p, ' ', &skip);
1.55 schwarze 1032: cp++;
1.59 schwarze 1033: break;
1.83 schwarze 1034: case ASCII_HYPH:
1.64 schwarze 1035: sz += cond_width(p, '-', &skip);
1.55 schwarze 1036: cp++;
1.59 schwarze 1037: break;
1038: default:
1039: break;
1040: }
1041: }
1.39 schwarze 1042:
1.112 schwarze 1043: return sz;
1.39 schwarze 1044: }
1045:
1.100 schwarze 1046: int
1.39 schwarze 1047: term_vspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 1048: {
1049: double r;
1.101 schwarze 1050: int ri;
1.16 schwarze 1051:
1052: switch (su->unit) {
1.99 schwarze 1053: case SCALE_BU:
1054: r = su->scale / 40.0;
1055: break;
1.83 schwarze 1056: case SCALE_CM:
1.99 schwarze 1057: r = su->scale * 6.0 / 2.54;
1058: break;
1059: case SCALE_FS:
1060: r = su->scale * 65536.0 / 40.0;
1.16 schwarze 1061: break;
1.83 schwarze 1062: case SCALE_IN:
1.86 schwarze 1063: r = su->scale * 6.0;
1.16 schwarze 1064: break;
1.99 schwarze 1065: case SCALE_MM:
1066: r = su->scale * 0.006;
1067: break;
1.83 schwarze 1068: case SCALE_PC:
1.16 schwarze 1069: r = su->scale;
1070: break;
1.83 schwarze 1071: case SCALE_PT:
1.99 schwarze 1072: r = su->scale / 12.0;
1.16 schwarze 1073: break;
1.99 schwarze 1074: case SCALE_EN:
1075: case SCALE_EM:
1076: r = su->scale * 0.6;
1.16 schwarze 1077: break;
1.83 schwarze 1078: case SCALE_VS:
1.16 schwarze 1079: r = su->scale;
1080: break;
1081: default:
1.99 schwarze 1082: abort();
1.16 schwarze 1083: }
1.101 schwarze 1084: ri = r > 0.0 ? r + 0.4995 : r - 0.4995;
1.112 schwarze 1085: return ri < 66 ? ri : 1;
1.16 schwarze 1086: }
1087:
1.107 schwarze 1088: /*
1.131 schwarze 1089: * Convert a scaling width to basic units, rounding towards 0.
1.107 schwarze 1090: */
1.100 schwarze 1091: int
1.39 schwarze 1092: term_hspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 1093: {
1094:
1.112 schwarze 1095: return (*p->hspan)(p, su);
1.131 schwarze 1096: }
1097:
1098: /*
1099: * Convert a scaling width to basic units, rounding to closest.
1100: */
1101: int
1102: term_hen(const struct termp *p, const struct roffsu *su)
1103: {
1104: int bu;
1105:
1106: if ((bu = (*p->hspan)(p, su)) >= 0)
1107: return (bu + 11) / 24;
1108: else
1109: return -((-bu + 11) / 24);
1.16 schwarze 1110: }