Annotation of src/usr.bin/mandoc/term.c, Revision 1.139
1.139 ! schwarze 1: /* $OpenBSD: term.c,v 1.138 2019/01/03 19:59:11 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:
375: vbl++;
376: vis++;
377: continue;
378: default:
379: break;
380: }
1.48 schwarze 381:
1.138 schwarze 382: /*
383: * We found a non-blank character to print,
384: * so write preceding white space now.
385: */
1.127 schwarze 386:
1.138 schwarze 387: if (vbl > 0) {
388: (*p->advance)(p, vbl);
389: p->viscol += vbl;
390: vbl = 0;
391: }
1.18 schwarze 392:
1.138 schwarze 393: /* Print the character and adjust the visual position. */
1.1 kristaps 394:
1.138 schwarze 395: (*p->letter)(p, p->tcol->buf[ic]);
396: if (p->tcol->buf[ic] == '\b') {
397: dv = (*p->width)(p, p->tcol->buf[ic - 1]);
398: p->viscol -= dv;
399: vis -= dv;
400: } else {
401: dv = (*p->width)(p, p->tcol->buf[ic]);
402: p->viscol += dv;
403: vis += dv;
404: }
405: }
406: p->tcol->col = nbr;
1.124 schwarze 407: }
408:
409: static void
410: endline(struct termp *p)
411: {
412: if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) {
413: p->mc = NULL;
414: p->flags &= ~TERMP_ENDMC;
415: }
416: if (p->mc != NULL) {
417: if (p->viscol && p->maxrmargin >= p->viscol)
418: (*p->advance)(p, p->maxrmargin - p->viscol + 1);
419: p->flags |= TERMP_NOBUF | TERMP_NOSPACE;
420: term_word(p, p->mc);
421: p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC);
422: }
423: p->viscol = 0;
424: p->minbl = 0;
425: (*p->endline)(p);
1.1 kristaps 426: }
427:
1.83 schwarze 428: /*
1.1 kristaps 429: * A newline only breaks an existing line; it won't assert vertical
430: * space. All data in the output buffer is flushed prior to the newline
431: * assertion.
432: */
433: void
434: term_newln(struct termp *p)
435: {
436:
437: p->flags |= TERMP_NOSPACE;
1.129 schwarze 438: if (p->tcol->lastcol || p->viscol)
1.61 schwarze 439: term_flushln(p);
1.1 kristaps 440: }
441:
442: /*
443: * Asserts a vertical space (a full, empty line-break between lines).
444: * Note that if used twice, this will cause two blank spaces and so on.
445: * All data in the output buffer is flushed prior to the newline
446: * assertion.
447: */
448: void
449: term_vspace(struct termp *p)
450: {
451:
452: term_newln(p);
1.29 schwarze 453: p->viscol = 0;
1.124 schwarze 454: p->minbl = 0;
1.63 schwarze 455: if (0 < p->skipvsp)
456: p->skipvsp--;
457: else
458: (*p->endline)(p);
1.1 kristaps 459: }
460:
1.98 schwarze 461: /* Swap current and previous font; for \fP and .ft P */
1.20 schwarze 462: void
463: term_fontlast(struct termp *p)
464: {
465: enum termfont f;
1.11 schwarze 466:
1.20 schwarze 467: f = p->fontl;
468: p->fontl = p->fontq[p->fonti];
469: p->fontq[p->fonti] = f;
470: }
471:
1.98 schwarze 472: /* Set font, save current, discard previous; for \f, .ft, .B etc. */
1.20 schwarze 473: void
474: term_fontrepl(struct termp *p, enum termfont f)
475: {
476:
477: p->fontl = p->fontq[p->fonti];
478: p->fontq[p->fonti] = f;
1.1 kristaps 479: }
480:
1.98 schwarze 481: /* Set font, save previous. */
1.20 schwarze 482: void
483: term_fontpush(struct termp *p, enum termfont f)
1.1 kristaps 484: {
1.7 schwarze 485:
1.20 schwarze 486: p->fontl = p->fontq[p->fonti];
1.98 schwarze 487: if (++p->fonti == p->fontsz) {
488: p->fontsz += 8;
489: p->fontq = mandoc_reallocarray(p->fontq,
1.116 schwarze 490: p->fontsz, sizeof(*p->fontq));
1.98 schwarze 491: }
492: p->fontq[p->fonti] = f;
1.20 schwarze 493: }
1.1 kristaps 494:
1.98 schwarze 495: /* Flush to make the saved pointer current again. */
1.20 schwarze 496: void
1.104 schwarze 497: term_fontpopq(struct termp *p, int i)
1.20 schwarze 498: {
1.1 kristaps 499:
1.104 schwarze 500: assert(i >= 0);
501: if (p->fonti > i)
502: p->fonti = i;
1.20 schwarze 503: }
1.1 kristaps 504:
1.98 schwarze 505: /* Pop one font off the stack. */
1.20 schwarze 506: void
507: term_fontpop(struct termp *p)
508: {
1.1 kristaps 509:
1.20 schwarze 510: assert(p->fonti);
511: p->fonti--;
1.1 kristaps 512: }
513:
514: /*
515: * Handle pwords, partial words, which may be either a single word or a
516: * phrase that cannot be broken down (such as a literal string). This
517: * handles word styling.
518: */
1.7 schwarze 519: void
520: term_word(struct termp *p, const char *word)
1.1 kristaps 521: {
1.121 schwarze 522: struct roffsu su;
1.75 schwarze 523: const char nbrsp[2] = { ASCII_NBRSP, 0 };
1.59 schwarze 524: const char *seq, *cp;
525: int sz, uc;
1.122 schwarze 526: size_t csz, lsz, ssz;
1.59 schwarze 527: enum mandoc_esc esc;
1.1 kristaps 528:
1.124 schwarze 529: if ((p->flags & TERMP_NOBUF) == 0) {
530: if ((p->flags & TERMP_NOSPACE) == 0) {
531: if ((p->flags & TERMP_KEEP) == 0) {
1.40 schwarze 532: bufferc(p, ' ');
1.124 schwarze 533: if (p->flags & TERMP_SENTENCE)
534: bufferc(p, ' ');
535: } else
536: bufferc(p, ASCII_NBRSP);
537: }
538: if (p->flags & TERMP_PREKEEP)
539: p->flags |= TERMP_KEEP;
540: if (p->flags & TERMP_NONOSPACE)
541: p->flags |= TERMP_NOSPACE;
542: else
543: p->flags &= ~TERMP_NOSPACE;
544: p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE);
545: p->skipvsp = 0;
1.31 schwarze 546: }
547:
1.59 schwarze 548: while ('\0' != *word) {
1.64 schwarze 549: if ('\\' != *word) {
1.75 schwarze 550: if (TERMP_NBRWORD & p->flags) {
551: if (' ' == *word) {
552: encode(p, nbrsp, 1);
553: word++;
554: continue;
555: }
556: ssz = strcspn(word, "\\ ");
557: } else
558: ssz = strcspn(word, "\\");
1.45 schwarze 559: encode(p, word, ssz);
1.64 schwarze 560: word += (int)ssz;
1.20 schwarze 561: continue;
1.64 schwarze 562: }
1.20 schwarze 563:
1.59 schwarze 564: word++;
565: esc = mandoc_escape(&word, &seq, &sz);
566: switch (esc) {
1.83 schwarze 567: case ESCAPE_UNICODE:
1.89 schwarze 568: uc = mchars_num2uc(seq + 1, sz - 1);
1.56 schwarze 569: break;
1.83 schwarze 570: case ESCAPE_NUMBERED:
1.93 schwarze 571: uc = mchars_num2char(seq, sz);
572: if (uc < 0)
573: continue;
1.20 schwarze 574: break;
1.83 schwarze 575: case ESCAPE_SPECIAL:
1.89 schwarze 576: if (p->enc == TERMENC_ASCII) {
1.114 schwarze 577: cp = mchars_spec2str(seq, sz, &ssz);
1.92 schwarze 578: if (cp != NULL)
1.89 schwarze 579: encode(p, cp, ssz);
580: } else {
1.114 schwarze 581: uc = mchars_spec2cp(seq, sz);
1.90 schwarze 582: if (uc > 0)
583: encode1(p, uc);
1.89 schwarze 584: }
1.93 schwarze 585: continue;
1.137 schwarze 586: case ESCAPE_UNDEF:
587: uc = *seq;
588: break;
1.83 schwarze 589: case ESCAPE_FONTBOLD:
1.20 schwarze 590: term_fontrepl(p, TERMFONT_BOLD);
1.93 schwarze 591: continue;
1.83 schwarze 592: case ESCAPE_FONTITALIC:
1.20 schwarze 593: term_fontrepl(p, TERMFONT_UNDER);
1.93 schwarze 594: continue;
1.83 schwarze 595: case ESCAPE_FONTBI:
1.70 schwarze 596: term_fontrepl(p, TERMFONT_BI);
1.93 schwarze 597: continue;
1.83 schwarze 598: case ESCAPE_FONT:
1.136 schwarze 599: case ESCAPE_FONTCW:
1.83 schwarze 600: case ESCAPE_FONTROMAN:
1.20 schwarze 601: term_fontrepl(p, TERMFONT_NONE);
1.93 schwarze 602: continue;
1.83 schwarze 603: case ESCAPE_FONTPREV:
1.20 schwarze 604: term_fontlast(p);
1.130 schwarze 605: continue;
606: case ESCAPE_BREAK:
607: bufferc(p, '\n');
1.93 schwarze 608: continue;
1.83 schwarze 609: case ESCAPE_NOSPACE:
1.108 schwarze 610: if (p->flags & TERMP_BACKAFTER)
611: p->flags &= ~TERMP_BACKAFTER;
612: else if (*word == '\0')
1.97 schwarze 613: p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE);
1.121 schwarze 614: continue;
1.135 schwarze 615: case ESCAPE_DEVICE:
616: if (p->type == TERMTYPE_PDF)
617: encode(p, "pdf", 3);
618: else if (p->type == TERMTYPE_PS)
619: encode(p, "ps", 2);
620: else if (p->enc == TERMENC_ASCII)
621: encode(p, "ascii", 5);
622: else
623: encode(p, "utf8", 4);
624: continue;
1.121 schwarze 625: case ESCAPE_HORIZ:
1.133 schwarze 626: if (*seq == '|') {
627: seq++;
628: uc = -p->col;
629: } else
630: uc = 0;
1.128 schwarze 631: if (a2roffsu(seq, &su, SCALE_EM) == NULL)
1.121 schwarze 632: continue;
1.133 schwarze 633: uc += term_hen(p, &su);
1.121 schwarze 634: if (uc > 0)
635: while (uc-- > 0)
636: bufferc(p, ASCII_NBRSP);
637: else if (p->col > (size_t)(-uc))
638: p->col += uc;
639: else {
640: uc += p->col;
641: p->col = 0;
1.126 schwarze 642: if (p->tcol->offset > (size_t)(-uc)) {
1.121 schwarze 643: p->ti += uc;
1.126 schwarze 644: p->tcol->offset += uc;
1.121 schwarze 645: } else {
1.126 schwarze 646: p->ti -= p->tcol->offset;
647: p->tcol->offset = 0;
1.121 schwarze 648: }
1.122 schwarze 649: }
650: continue;
651: case ESCAPE_HLINE:
1.132 schwarze 652: if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL)
1.122 schwarze 653: continue;
1.131 schwarze 654: uc = term_hen(p, &su);
1.122 schwarze 655: if (uc <= 0) {
1.126 schwarze 656: if (p->tcol->rmargin <= p->tcol->offset)
1.122 schwarze 657: continue;
1.126 schwarze 658: lsz = p->tcol->rmargin - p->tcol->offset;
1.122 schwarze 659: } else
660: lsz = uc;
1.132 schwarze 661: if (*cp == seq[-1])
1.122 schwarze 662: uc = -1;
1.132 schwarze 663: else if (*cp == '\\') {
664: seq = cp + 1;
1.122 schwarze 665: esc = mandoc_escape(&seq, &cp, &sz);
666: switch (esc) {
667: case ESCAPE_UNICODE:
668: uc = mchars_num2uc(cp + 1, sz - 1);
669: break;
670: case ESCAPE_NUMBERED:
671: uc = mchars_num2char(cp, sz);
672: break;
673: case ESCAPE_SPECIAL:
674: uc = mchars_spec2cp(cp, sz);
675: break;
1.137 schwarze 676: case ESCAPE_UNDEF:
677: uc = *seq;
678: break;
1.122 schwarze 679: default:
680: uc = -1;
681: break;
682: }
683: } else
1.132 schwarze 684: uc = *cp;
1.122 schwarze 685: if (uc < 0x20 || (uc > 0x7E && uc < 0xA0))
686: uc = '_';
687: if (p->enc == TERMENC_ASCII) {
688: cp = ascii_uc2str(uc);
689: csz = term_strlen(p, cp);
690: ssz = strlen(cp);
691: } else
692: csz = (*p->width)(p, uc);
693: while (lsz >= csz) {
694: if (p->enc == TERMENC_ASCII)
695: encode(p, cp, ssz);
696: else
697: encode1(p, uc);
698: lsz -= csz;
1.121 schwarze 699: }
1.93 schwarze 700: continue;
1.83 schwarze 701: case ESCAPE_SKIPCHAR:
1.108 schwarze 702: p->flags |= TERMP_BACKAFTER;
1.93 schwarze 703: continue;
1.103 schwarze 704: case ESCAPE_OVERSTRIKE:
705: cp = seq + sz;
706: while (seq < cp) {
707: if (*seq == '\\') {
708: mandoc_escape(&seq, NULL, NULL);
709: continue;
710: }
711: encode1(p, *seq++);
1.108 schwarze 712: if (seq < cp) {
713: if (p->flags & TERMP_BACKBEFORE)
714: p->flags |= TERMP_BACKAFTER;
715: else
716: p->flags |= TERMP_BACKBEFORE;
717: }
1.103 schwarze 718: }
1.109 schwarze 719: /* Trim trailing backspace/blank pair. */
1.129 schwarze 720: if (p->tcol->lastcol > 2 &&
721: (p->tcol->buf[p->tcol->lastcol - 1] == ' ' ||
722: p->tcol->buf[p->tcol->lastcol - 1] == '\t'))
723: p->tcol->lastcol -= 2;
724: if (p->col > p->tcol->lastcol)
725: p->col = p->tcol->lastcol;
1.108 schwarze 726: continue;
1.20 schwarze 727: default:
1.93 schwarze 728: continue;
729: }
730:
731: /*
732: * Common handling for Unicode and numbered
733: * character escape sequences.
734: */
735:
736: if (p->enc == TERMENC_ASCII) {
737: cp = ascii_uc2str(uc);
738: encode(p, cp, strlen(cp));
739: } else {
740: if ((uc < 0x20 && uc != 0x09) ||
741: (uc > 0x7E && uc < 0xA0))
742: uc = 0xFFFD;
743: encode1(p, uc);
1.20 schwarze 744: }
745: }
1.75 schwarze 746: p->flags &= ~TERMP_NBRWORD;
1.1 kristaps 747: }
748:
749: static void
1.126 schwarze 750: adjbuf(struct termp_col *c, size_t sz)
1.1 kristaps 751: {
1.126 schwarze 752: if (c->maxcols == 0)
753: c->maxcols = 1024;
754: while (c->maxcols <= sz)
755: c->maxcols <<= 2;
756: c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf));
1.1 kristaps 757: }
758:
1.4 schwarze 759: static void
1.20 schwarze 760: bufferc(struct termp *p, char c)
761: {
1.124 schwarze 762: if (p->flags & TERMP_NOBUF) {
763: (*p->letter)(p, c);
764: return;
765: }
1.126 schwarze 766: if (p->col + 1 >= p->tcol->maxcols)
767: adjbuf(p->tcol, p->col + 1);
1.129 schwarze 768: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 769: p->tcol->buf[p->col] = c;
1.129 schwarze 770: if (p->tcol->lastcol < ++p->col)
771: p->tcol->lastcol = p->col;
1.20 schwarze 772: }
773:
1.59 schwarze 774: /*
775: * See encode().
776: * Do this for a single (probably unicode) value.
777: * Does not check for non-decorated glyphs.
778: */
779: static void
780: encode1(struct termp *p, int c)
781: {
782: enum termfont f;
783:
1.124 schwarze 784: if (p->flags & TERMP_NOBUF) {
785: (*p->letter)(p, c);
786: return;
787: }
788:
1.126 schwarze 789: if (p->col + 7 >= p->tcol->maxcols)
790: adjbuf(p->tcol, p->col + 7);
1.59 schwarze 791:
1.115 schwarze 792: f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ?
1.108 schwarze 793: p->fontq[p->fonti] : TERMFONT_NONE;
1.59 schwarze 794:
1.108 schwarze 795: if (p->flags & TERMP_BACKBEFORE) {
1.126 schwarze 796: if (p->tcol->buf[p->col - 1] == ' ' ||
797: p->tcol->buf[p->col - 1] == '\t')
1.109 schwarze 798: p->col--;
799: else
1.126 schwarze 800: p->tcol->buf[p->col++] = '\b';
1.108 schwarze 801: p->flags &= ~TERMP_BACKBEFORE;
802: }
1.126 schwarze 803: if (f == TERMFONT_UNDER || f == TERMFONT_BI) {
804: p->tcol->buf[p->col++] = '_';
805: p->tcol->buf[p->col++] = '\b';
806: }
807: if (f == TERMFONT_BOLD || f == TERMFONT_BI) {
808: if (c == ASCII_HYPH)
809: p->tcol->buf[p->col++] = '-';
1.70 schwarze 810: else
1.126 schwarze 811: p->tcol->buf[p->col++] = c;
812: p->tcol->buf[p->col++] = '\b';
1.70 schwarze 813: }
1.129 schwarze 814: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 815: p->tcol->buf[p->col] = c;
1.129 schwarze 816: if (p->tcol->lastcol < ++p->col)
817: p->tcol->lastcol = p->col;
1.108 schwarze 818: if (p->flags & TERMP_BACKAFTER) {
819: p->flags |= TERMP_BACKBEFORE;
820: p->flags &= ~TERMP_BACKAFTER;
821: }
1.59 schwarze 822: }
1.20 schwarze 823:
824: static void
825: encode(struct termp *p, const char *word, size_t sz)
1.4 schwarze 826: {
1.71 schwarze 827: size_t i;
1.124 schwarze 828:
829: if (p->flags & TERMP_NOBUF) {
830: for (i = 0; i < sz; i++)
831: (*p->letter)(p, word[i]);
832: return;
833: }
1.59 schwarze 834:
1.126 schwarze 835: if (p->col + 2 + (sz * 5) >= p->tcol->maxcols)
836: adjbuf(p->tcol, p->col + 2 + (sz * 5));
1.46 schwarze 837:
1.71 schwarze 838: for (i = 0; i < sz; i++) {
1.70 schwarze 839: if (ASCII_HYPH == word[i] ||
840: isgraph((unsigned char)word[i]))
841: encode1(p, word[i]);
1.119 schwarze 842: else {
1.129 schwarze 843: if (p->tcol->lastcol <= p->col ||
1.125 schwarze 844: (word[i] != ' ' && word[i] != ASCII_NBRSP))
1.126 schwarze 845: p->tcol->buf[p->col] = word[i];
1.125 schwarze 846: p->col++;
1.119 schwarze 847:
848: /*
849: * Postpone the effect of \z while handling
850: * an overstrike sequence from ascii_uc2str().
851: */
852:
853: if (word[i] == '\b' &&
854: (p->flags & TERMP_BACKBEFORE)) {
855: p->flags &= ~TERMP_BACKBEFORE;
856: p->flags |= TERMP_BACKAFTER;
857: }
858: }
1.4 schwarze 859: }
1.129 schwarze 860: if (p->tcol->lastcol < p->col)
861: p->tcol->lastcol = p->col;
1.80 schwarze 862: }
863:
864: void
865: term_setwidth(struct termp *p, const char *wstr)
866: {
867: struct roffsu su;
1.107 schwarze 868: int iop, width;
1.80 schwarze 869:
1.81 schwarze 870: iop = 0;
871: width = 0;
1.80 schwarze 872: if (NULL != wstr) {
873: switch (*wstr) {
1.83 schwarze 874: case '+':
1.80 schwarze 875: iop = 1;
876: wstr++;
877: break;
1.83 schwarze 878: case '-':
1.80 schwarze 879: iop = -1;
880: wstr++;
881: break;
882: default:
883: break;
884: }
1.128 schwarze 885: if (a2roffsu(wstr, &su, SCALE_MAX) != NULL)
1.81 schwarze 886: width = term_hspan(p, &su);
887: else
1.80 schwarze 888: iop = 0;
889: }
890: (*p->setwidth)(p, iop, width);
1.4 schwarze 891: }
1.16 schwarze 892:
893: size_t
1.39 schwarze 894: term_len(const struct termp *p, size_t sz)
895: {
896:
1.112 schwarze 897: return (*p->width)(p, ' ') * sz;
1.39 schwarze 898: }
899:
1.64 schwarze 900: static size_t
901: cond_width(const struct termp *p, int c, int *skip)
902: {
903:
904: if (*skip) {
905: (*skip) = 0;
1.112 schwarze 906: return 0;
1.64 schwarze 907: } else
1.112 schwarze 908: return (*p->width)(p, c);
1.64 schwarze 909: }
1.39 schwarze 910:
911: size_t
912: term_strlen(const struct termp *p, const char *cp)
913: {
1.59 schwarze 914: size_t sz, rsz, i;
1.93 schwarze 915: int ssz, skip, uc;
1.50 schwarze 916: const char *seq, *rhs;
1.59 schwarze 917: enum mandoc_esc esc;
1.77 schwarze 918: static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH,
919: ASCII_BREAK, '\0' };
1.59 schwarze 920:
921: /*
922: * Account for escaped sequences within string length
923: * calculations. This follows the logic in term_word() as we
924: * must calculate the width of produced strings.
925: */
926:
927: sz = 0;
1.64 schwarze 928: skip = 0;
1.59 schwarze 929: while ('\0' != *cp) {
930: rsz = strcspn(cp, rej);
931: for (i = 0; i < rsz; i++)
1.64 schwarze 932: sz += cond_width(p, *cp++, &skip);
1.59 schwarze 933:
934: switch (*cp) {
1.83 schwarze 935: case '\\':
1.59 schwarze 936: cp++;
1.137 schwarze 937: rhs = NULL;
1.59 schwarze 938: esc = mandoc_escape(&cp, &seq, &ssz);
939: switch (esc) {
1.83 schwarze 940: case ESCAPE_UNICODE:
1.94 schwarze 941: uc = mchars_num2uc(seq + 1, ssz - 1);
1.59 schwarze 942: break;
1.83 schwarze 943: case ESCAPE_NUMBERED:
1.93 schwarze 944: uc = mchars_num2char(seq, ssz);
945: if (uc < 0)
946: continue;
1.50 schwarze 947: break;
1.83 schwarze 948: case ESCAPE_SPECIAL:
1.93 schwarze 949: if (p->enc == TERMENC_ASCII) {
1.114 schwarze 950: rhs = mchars_spec2str(seq, ssz, &rsz);
1.93 schwarze 951: if (rhs != NULL)
952: break;
953: } else {
1.114 schwarze 954: uc = mchars_spec2cp(seq, ssz);
1.93 schwarze 955: if (uc > 0)
956: sz += cond_width(p, uc, &skip);
1.89 schwarze 957: }
1.93 schwarze 958: continue;
1.137 schwarze 959: case ESCAPE_UNDEF:
960: uc = *seq;
961: break;
1.135 schwarze 962: case ESCAPE_DEVICE:
963: if (p->type == TERMTYPE_PDF) {
964: rhs = "pdf";
965: rsz = 3;
966: } else if (p->type == TERMTYPE_PS) {
967: rhs = "ps";
968: rsz = 2;
969: } else if (p->enc == TERMENC_ASCII) {
970: rhs = "ascii";
971: rsz = 5;
972: } else {
973: rhs = "utf8";
974: rsz = 4;
975: }
976: break;
1.83 schwarze 977: case ESCAPE_SKIPCHAR:
1.64 schwarze 978: skip = 1;
1.103 schwarze 979: continue;
980: case ESCAPE_OVERSTRIKE:
981: rsz = 0;
982: rhs = seq + ssz;
983: while (seq < rhs) {
984: if (*seq == '\\') {
985: mandoc_escape(&seq, NULL, NULL);
986: continue;
987: }
988: i = (*p->width)(p, *seq++);
989: if (rsz < i)
990: rsz = i;
991: }
992: sz += rsz;
1.93 schwarze 993: continue;
1.50 schwarze 994: default:
1.93 schwarze 995: continue;
1.50 schwarze 996: }
1.39 schwarze 997:
1.93 schwarze 998: /*
999: * Common handling for Unicode and numbered
1000: * character escape sequences.
1001: */
1002:
1003: if (rhs == NULL) {
1004: if (p->enc == TERMENC_ASCII) {
1005: rhs = ascii_uc2str(uc);
1006: rsz = strlen(rhs);
1007: } else {
1008: if ((uc < 0x20 && uc != 0x09) ||
1009: (uc > 0x7E && uc < 0xA0))
1010: uc = 0xFFFD;
1011: sz += cond_width(p, uc, &skip);
1012: continue;
1013: }
1014: }
1.59 schwarze 1015:
1.64 schwarze 1016: if (skip) {
1017: skip = 0;
1018: break;
1019: }
1.93 schwarze 1020:
1021: /*
1022: * Common handling for all escape sequences
1023: * printing more than one character.
1024: */
1.64 schwarze 1025:
1.59 schwarze 1026: for (i = 0; i < rsz; i++)
1027: sz += (*p->width)(p, *rhs++);
1028: break;
1.83 schwarze 1029: case ASCII_NBRSP:
1.64 schwarze 1030: sz += cond_width(p, ' ', &skip);
1.55 schwarze 1031: cp++;
1.59 schwarze 1032: break;
1.83 schwarze 1033: case ASCII_HYPH:
1.64 schwarze 1034: sz += cond_width(p, '-', &skip);
1.55 schwarze 1035: cp++;
1.59 schwarze 1036: break;
1037: default:
1038: break;
1039: }
1040: }
1.39 schwarze 1041:
1.112 schwarze 1042: return sz;
1.39 schwarze 1043: }
1044:
1.100 schwarze 1045: int
1.39 schwarze 1046: term_vspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 1047: {
1048: double r;
1.101 schwarze 1049: int ri;
1.16 schwarze 1050:
1051: switch (su->unit) {
1.99 schwarze 1052: case SCALE_BU:
1053: r = su->scale / 40.0;
1054: break;
1.83 schwarze 1055: case SCALE_CM:
1.99 schwarze 1056: r = su->scale * 6.0 / 2.54;
1057: break;
1058: case SCALE_FS:
1059: r = su->scale * 65536.0 / 40.0;
1.16 schwarze 1060: break;
1.83 schwarze 1061: case SCALE_IN:
1.86 schwarze 1062: r = su->scale * 6.0;
1.16 schwarze 1063: break;
1.99 schwarze 1064: case SCALE_MM:
1065: r = su->scale * 0.006;
1066: break;
1.83 schwarze 1067: case SCALE_PC:
1.16 schwarze 1068: r = su->scale;
1069: break;
1.83 schwarze 1070: case SCALE_PT:
1.99 schwarze 1071: r = su->scale / 12.0;
1.16 schwarze 1072: break;
1.99 schwarze 1073: case SCALE_EN:
1074: case SCALE_EM:
1075: r = su->scale * 0.6;
1.16 schwarze 1076: break;
1.83 schwarze 1077: case SCALE_VS:
1.16 schwarze 1078: r = su->scale;
1079: break;
1080: default:
1.99 schwarze 1081: abort();
1.16 schwarze 1082: }
1.101 schwarze 1083: ri = r > 0.0 ? r + 0.4995 : r - 0.4995;
1.112 schwarze 1084: return ri < 66 ? ri : 1;
1.16 schwarze 1085: }
1086:
1.107 schwarze 1087: /*
1.131 schwarze 1088: * Convert a scaling width to basic units, rounding towards 0.
1.107 schwarze 1089: */
1.100 schwarze 1090: int
1.39 schwarze 1091: term_hspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 1092: {
1093:
1.112 schwarze 1094: return (*p->hspan)(p, su);
1.131 schwarze 1095: }
1096:
1097: /*
1098: * Convert a scaling width to basic units, rounding to closest.
1099: */
1100: int
1101: term_hen(const struct termp *p, const struct roffsu *su)
1102: {
1103: int bu;
1104:
1105: if ((bu = (*p->hspan)(p, su)) >= 0)
1106: return (bu + 11) / 24;
1107: else
1108: return -((-bu + 11) / 24);
1.16 schwarze 1109: }