Annotation of src/usr.bin/mandoc/term.c, Revision 1.136
1.136 ! schwarze 1: /* $OpenBSD: term.c,v 1.135 2018/08/16 13:49:40 schwarze Exp $ */
1.1 kristaps 2: /*
1.59 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.135 schwarze 4: * Copyright (c) 2010-2018 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>
1.1 kristaps 22: #include <stdio.h>
23: #include <stdlib.h>
24: #include <string.h>
25:
1.34 schwarze 26: #include "mandoc.h"
1.79 schwarze 27: #include "mandoc_aux.h"
1.16 schwarze 28: #include "out.h"
1.1 kristaps 29: #include "term.h"
1.16 schwarze 30: #include "main.h"
1.1 kristaps 31:
1.64 schwarze 32: static size_t cond_width(const struct termp *, int, int *);
1.126 schwarze 33: static void adjbuf(struct termp_col *, size_t);
1.59 schwarze 34: static void bufferc(struct termp *, char);
35: static void encode(struct termp *, const char *, size_t);
36: static void encode1(struct termp *, int);
1.124 schwarze 37: static void endline(struct termp *);
1.1 kristaps 38:
1.83 schwarze 39:
1.37 schwarze 40: void
1.129 schwarze 41: term_setcol(struct termp *p, size_t maxtcol)
42: {
43: if (maxtcol > p->maxtcol) {
44: p->tcols = mandoc_recallocarray(p->tcols,
45: p->maxtcol, maxtcol, sizeof(*p->tcols));
46: p->maxtcol = maxtcol;
47: }
48: p->lasttcol = maxtcol - 1;
49: p->tcol = p->tcols;
50: }
51:
52: void
1.37 schwarze 53: term_free(struct termp *p)
1.1 kristaps 54: {
1.126 schwarze 55: for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++)
56: free(p->tcol->buf);
57: free(p->tcols);
1.98 schwarze 58: free(p->fontq);
1.37 schwarze 59: free(p);
1.1 kristaps 60: }
61:
1.13 schwarze 62: void
1.83 schwarze 63: term_begin(struct termp *p, term_margin head,
1.106 schwarze 64: term_margin foot, const struct roff_meta *arg)
1.1 kristaps 65: {
66:
1.37 schwarze 67: p->headf = head;
68: p->footf = foot;
69: p->argf = arg;
70: (*p->begin)(p);
1.1 kristaps 71: }
72:
1.37 schwarze 73: void
74: term_end(struct termp *p)
1.1 kristaps 75: {
76:
1.37 schwarze 77: (*p->end)(p);
1.1 kristaps 78: }
79:
80: /*
1.82 schwarze 81: * Flush a chunk of text. By default, break the output line each time
82: * the right margin is reached, and continue output on the next line
83: * at the same offset as the chunk itself. By default, also break the
84: * output line at the end of the chunk.
1.27 schwarze 85: * The following flags may be specified:
1.1 kristaps 86: *
1.82 schwarze 87: * - TERMP_NOBREAK: Do not break the output line at the right margin,
88: * but only at the max right margin. Also, do not break the output
89: * line at the end of the chunk, such that the next call can pad to
90: * the next column. However, if less than p->trailspace blanks,
91: * which can be 0, 1, or 2, remain to the right margin, the line
92: * will be broken.
1.110 schwarze 93: * - TERMP_BRTRSP: Consider trailing whitespace significant
94: * when deciding whether the chunk fits or not.
1.82 schwarze 95: * - TERMP_BRIND: If the chunk does not fit and the output line has
96: * to be broken, start the next line at the right margin instead
97: * of at the offset. Used together with TERMP_NOBREAK for the tags
98: * in various kinds of tagged lists.
1.123 schwarze 99: * - TERMP_HANG: Do not break the output line at the right margin,
1.82 schwarze 100: * append the next chunk after it even if this one is too long.
101: * To be used together with TERMP_NOBREAK.
1.123 schwarze 102: * - TERMP_NOPAD: Start writing at the current position,
103: * do not pad with blank characters up to the offset.
1.1 kristaps 104: */
105: void
106: term_flushln(struct termp *p)
107: {
1.19 schwarze 108: size_t vis; /* current visual position on output */
109: size_t vbl; /* number of blanks to prepend to output */
1.33 schwarze 110: size_t vend; /* end of word visual position on output */
1.19 schwarze 111: size_t bp; /* visual right border position */
1.51 schwarze 112: size_t dv; /* temporary for visual pos calculations */
1.126 schwarze 113: size_t j; /* temporary loop index for p->tcol->buf */
1.71 schwarze 114: size_t jhy; /* last hyph before overflow w/r/t j */
1.42 schwarze 115: size_t maxvis; /* output position of visible boundary */
1.126 schwarze 116: int ntab; /* number of tabs to prepend */
1.130 schwarze 117: int breakline; /* after this word */
1.1 kristaps 118:
1.126 schwarze 119: vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ?
120: 0 : p->tcol->offset - p->viscol;
1.123 schwarze 121: if (p->minbl && vbl < p->minbl)
122: vbl = p->minbl;
1.126 schwarze 123: maxvis = p->tcol->rmargin > p->viscol + vbl ?
124: p->tcol->rmargin - p->viscol - vbl : 0;
1.123 schwarze 125: bp = !(p->flags & TERMP_NOBREAK) ? maxvis :
126: p->maxrmargin > p->viscol + vbl ?
127: p->maxrmargin - p->viscol - vbl : 0;
1.53 schwarze 128: vis = vend = 0;
1.19 schwarze 129:
1.134 florian 130: if ((p->flags & TERMP_MULTICOL) == 0)
1.127 schwarze 131: p->tcol->col = 0;
1.129 schwarze 132: while (p->tcol->col < p->tcol->lastcol) {
1.127 schwarze 133:
1.22 schwarze 134: /*
1.42 schwarze 135: * Handle literal tab characters: collapse all
136: * subsequent tabs into a single huge set of spaces.
1.30 schwarze 137: */
1.127 schwarze 138:
1.66 schwarze 139: ntab = 0;
1.129 schwarze 140: while (p->tcol->col < p->tcol->lastcol &&
1.127 schwarze 141: p->tcol->buf[p->tcol->col] == '\t') {
1.120 schwarze 142: vend = term_tab_next(vis);
1.30 schwarze 143: vbl += vend - vis;
144: vis = vend;
1.66 schwarze 145: ntab++;
1.127 schwarze 146: p->tcol->col++;
1.30 schwarze 147: }
1.22 schwarze 148:
1.1 kristaps 149: /*
150: * Count up visible word characters. Control sequences
151: * (starting with the CSI) aren't counted. A space
152: * generates a non-printing word, which is valid (the
153: * space is printed according to regular spacing rules).
154: */
155:
1.127 schwarze 156: jhy = 0;
1.130 schwarze 157: breakline = 0;
1.129 schwarze 158: for (j = p->tcol->col; j < p->tcol->lastcol; j++) {
1.130 schwarze 159: if (p->tcol->buf[j] == '\n') {
160: if ((p->flags & TERMP_BRIND) == 0)
161: breakline = 1;
162: continue;
163: }
1.126 schwarze 164: if (p->tcol->buf[j] == ' ' || p->tcol->buf[j] == '\t')
1.1 kristaps 165: break;
1.42 schwarze 166:
1.117 krw 167: /* Back over the last printed character. */
1.126 schwarze 168: if (p->tcol->buf[j] == '\b') {
1.42 schwarze 169: assert(j);
1.126 schwarze 170: vend -= (*p->width)(p, p->tcol->buf[j - 1]);
1.42 schwarze 171: continue;
172: }
173:
174: /* Regular word. */
175: /* Break at the hyphen point if we overrun. */
1.83 schwarze 176: if (vend > vis && vend < bp &&
1.126 schwarze 177: (p->tcol->buf[j] == ASCII_HYPH||
178: p->tcol->buf[j] == ASCII_BREAK))
1.42 schwarze 179: jhy = j;
180:
1.78 schwarze 181: /*
182: * Hyphenation now decided, put back a real
183: * hyphen such that we get the correct width.
184: */
1.126 schwarze 185: if (p->tcol->buf[j] == ASCII_HYPH)
186: p->tcol->buf[j] = '-';
1.78 schwarze 187:
1.126 schwarze 188: vend += (*p->width)(p, p->tcol->buf[j]);
1.1 kristaps 189: }
190:
191: /*
1.5 schwarze 192: * Find out whether we would exceed the right margin.
1.33 schwarze 193: * If so, break to the next line.
1.5 schwarze 194: */
1.127 schwarze 195:
196: if (vend > bp && jhy == 0 && vis > 0 &&
1.124 schwarze 197: (p->flags & TERMP_BRNEVER) == 0) {
1.129 schwarze 198: if (p->flags & TERMP_MULTICOL)
1.127 schwarze 199: return;
200:
201: endline(p);
1.22 schwarze 202: vend -= vis;
1.66 schwarze 203:
1.120 schwarze 204: /* Use pending tabs on the new line. */
205:
206: vbl = 0;
207: while (ntab--)
208: vbl = term_tab_next(vbl);
209:
210: /* Re-establish indentation. */
1.66 schwarze 211:
1.123 schwarze 212: if (p->flags & TERMP_BRIND)
1.126 schwarze 213: vbl += p->tcol->rmargin;
1.123 schwarze 214: else
1.126 schwarze 215: vbl += p->tcol->offset;
216: maxvis = p->tcol->rmargin > vbl ?
217: p->tcol->rmargin - vbl : 0;
1.123 schwarze 218: bp = !(p->flags & TERMP_NOBREAK) ? maxvis :
219: p->maxrmargin > vbl ? p->maxrmargin - vbl : 0;
1.1 kristaps 220: }
1.30 schwarze 221:
1.127 schwarze 222: /*
223: * Write out the rest of the word.
224: */
225:
1.129 schwarze 226: for ( ; p->tcol->col < p->tcol->lastcol; p->tcol->col++) {
1.127 schwarze 227: if (vend > bp && jhy > 0 && p->tcol->col > jhy)
1.30 schwarze 228: break;
1.130 schwarze 229: if (p->tcol->buf[p->tcol->col] == '\n')
230: continue;
1.127 schwarze 231: if (p->tcol->buf[p->tcol->col] == '\t')
1.1 kristaps 232: break;
1.127 schwarze 233: if (p->tcol->buf[p->tcol->col] == ' ') {
234: j = p->tcol->col;
1.129 schwarze 235: while (p->tcol->col < p->tcol->lastcol &&
1.127 schwarze 236: p->tcol->buf[p->tcol->col] == ' ')
237: p->tcol->col++;
238: dv = (p->tcol->col - j) * (*p->width)(p, ' ');
1.51 schwarze 239: vbl += dv;
240: vend += dv;
1.22 schwarze 241: break;
242: }
1.127 schwarze 243: if (p->tcol->buf[p->tcol->col] == ASCII_NBRSP) {
1.42 schwarze 244: vbl += (*p->width)(p, ' ');
1.33 schwarze 245: continue;
246: }
1.127 schwarze 247: if (p->tcol->buf[p->tcol->col] == ASCII_BREAK)
1.77 schwarze 248: continue;
1.33 schwarze 249:
250: /*
251: * Now we definitely know there will be
252: * printable characters to output,
253: * so write preceding white space now.
254: */
255: if (vbl) {
1.37 schwarze 256: (*p->advance)(p, vbl);
1.33 schwarze 257: p->viscol += vbl;
258: vbl = 0;
1.61 schwarze 259: }
260:
1.127 schwarze 261: (*p->letter)(p, p->tcol->buf[p->tcol->col]);
262: if (p->tcol->buf[p->tcol->col] == '\b')
263: p->viscol -= (*p->width)(p,
264: p->tcol->buf[p->tcol->col - 1]);
1.83 schwarze 265: else
1.127 schwarze 266: p->viscol += (*p->width)(p,
267: p->tcol->buf[p->tcol->col]);
1.1 kristaps 268: }
1.22 schwarze 269: vis = vend;
1.130 schwarze 270:
271: if (breakline == 0)
272: continue;
273:
274: /* Explicitly requested output line break. */
275:
276: if (p->flags & TERMP_MULTICOL)
277: return;
278:
279: endline(p);
280: breakline = 0;
281: vis = vend = 0;
282:
283: /* Re-establish indentation. */
284:
285: vbl = p->tcol->offset;
286: maxvis = p->tcol->rmargin > vbl ?
287: p->tcol->rmargin - vbl : 0;
288: bp = !(p->flags & TERMP_NOBREAK) ? maxvis :
289: p->maxrmargin > vbl ? p->maxrmargin - vbl : 0;
1.1 kristaps 290: }
1.48 schwarze 291:
292: /*
293: * If there was trailing white space, it was not printed;
294: * so reset the cursor position accordingly.
295: */
1.127 schwarze 296:
1.95 schwarze 297: if (vis > vbl)
1.61 schwarze 298: vis -= vbl;
1.95 schwarze 299: else
300: vis = 0;
1.18 schwarze 301:
1.129 schwarze 302: p->col = p->tcol->col = p->tcol->lastcol = 0;
1.124 schwarze 303: p->minbl = p->trailspace;
1.123 schwarze 304: p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE | TERMP_NOPAD);
1.1 kristaps 305:
1.129 schwarze 306: if (p->flags & TERMP_MULTICOL)
307: return;
308:
1.110 schwarze 309: /* Trailing whitespace is significant in some columns. */
1.127 schwarze 310:
1.110 schwarze 311: if (vis && vbl && (TERMP_BRTRSP & p->flags))
312: vis += vbl;
1.1 kristaps 313:
1.61 schwarze 314: /* If the column was overrun, break the line. */
1.124 schwarze 315: if ((p->flags & TERMP_NOBREAK) == 0 ||
316: ((p->flags & TERMP_HANG) == 0 &&
317: vis + p->trailspace * (*p->width)(p, ' ') > maxvis))
318: endline(p);
319: }
320:
321: static void
322: endline(struct termp *p)
323: {
324: if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) {
325: p->mc = NULL;
326: p->flags &= ~TERMP_ENDMC;
327: }
328: if (p->mc != NULL) {
329: if (p->viscol && p->maxrmargin >= p->viscol)
330: (*p->advance)(p, p->maxrmargin - p->viscol + 1);
331: p->flags |= TERMP_NOBUF | TERMP_NOSPACE;
332: term_word(p, p->mc);
333: p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC);
334: }
335: p->viscol = 0;
336: p->minbl = 0;
337: (*p->endline)(p);
1.1 kristaps 338: }
339:
1.83 schwarze 340: /*
1.1 kristaps 341: * A newline only breaks an existing line; it won't assert vertical
342: * space. All data in the output buffer is flushed prior to the newline
343: * assertion.
344: */
345: void
346: term_newln(struct termp *p)
347: {
348:
349: p->flags |= TERMP_NOSPACE;
1.129 schwarze 350: if (p->tcol->lastcol || p->viscol)
1.61 schwarze 351: term_flushln(p);
1.1 kristaps 352: }
353:
354: /*
355: * Asserts a vertical space (a full, empty line-break between lines).
356: * Note that if used twice, this will cause two blank spaces and so on.
357: * All data in the output buffer is flushed prior to the newline
358: * assertion.
359: */
360: void
361: term_vspace(struct termp *p)
362: {
363:
364: term_newln(p);
1.29 schwarze 365: p->viscol = 0;
1.124 schwarze 366: p->minbl = 0;
1.63 schwarze 367: if (0 < p->skipvsp)
368: p->skipvsp--;
369: else
370: (*p->endline)(p);
1.1 kristaps 371: }
372:
1.98 schwarze 373: /* Swap current and previous font; for \fP and .ft P */
1.20 schwarze 374: void
375: term_fontlast(struct termp *p)
376: {
377: enum termfont f;
1.11 schwarze 378:
1.20 schwarze 379: f = p->fontl;
380: p->fontl = p->fontq[p->fonti];
381: p->fontq[p->fonti] = f;
382: }
383:
1.98 schwarze 384: /* Set font, save current, discard previous; for \f, .ft, .B etc. */
1.20 schwarze 385: void
386: term_fontrepl(struct termp *p, enum termfont f)
387: {
388:
389: p->fontl = p->fontq[p->fonti];
390: p->fontq[p->fonti] = f;
1.1 kristaps 391: }
392:
1.98 schwarze 393: /* Set font, save previous. */
1.20 schwarze 394: void
395: term_fontpush(struct termp *p, enum termfont f)
1.1 kristaps 396: {
1.7 schwarze 397:
1.20 schwarze 398: p->fontl = p->fontq[p->fonti];
1.98 schwarze 399: if (++p->fonti == p->fontsz) {
400: p->fontsz += 8;
401: p->fontq = mandoc_reallocarray(p->fontq,
1.116 schwarze 402: p->fontsz, sizeof(*p->fontq));
1.98 schwarze 403: }
404: p->fontq[p->fonti] = f;
1.20 schwarze 405: }
1.1 kristaps 406:
1.98 schwarze 407: /* Flush to make the saved pointer current again. */
1.20 schwarze 408: void
1.104 schwarze 409: term_fontpopq(struct termp *p, int i)
1.20 schwarze 410: {
1.1 kristaps 411:
1.104 schwarze 412: assert(i >= 0);
413: if (p->fonti > i)
414: p->fonti = i;
1.20 schwarze 415: }
1.1 kristaps 416:
1.98 schwarze 417: /* Pop one font off the stack. */
1.20 schwarze 418: void
419: term_fontpop(struct termp *p)
420: {
1.1 kristaps 421:
1.20 schwarze 422: assert(p->fonti);
423: p->fonti--;
1.1 kristaps 424: }
425:
426: /*
427: * Handle pwords, partial words, which may be either a single word or a
428: * phrase that cannot be broken down (such as a literal string). This
429: * handles word styling.
430: */
1.7 schwarze 431: void
432: term_word(struct termp *p, const char *word)
1.1 kristaps 433: {
1.121 schwarze 434: struct roffsu su;
1.75 schwarze 435: const char nbrsp[2] = { ASCII_NBRSP, 0 };
1.59 schwarze 436: const char *seq, *cp;
437: int sz, uc;
1.122 schwarze 438: size_t csz, lsz, ssz;
1.59 schwarze 439: enum mandoc_esc esc;
1.1 kristaps 440:
1.124 schwarze 441: if ((p->flags & TERMP_NOBUF) == 0) {
442: if ((p->flags & TERMP_NOSPACE) == 0) {
443: if ((p->flags & TERMP_KEEP) == 0) {
1.40 schwarze 444: bufferc(p, ' ');
1.124 schwarze 445: if (p->flags & TERMP_SENTENCE)
446: bufferc(p, ' ');
447: } else
448: bufferc(p, ASCII_NBRSP);
449: }
450: if (p->flags & TERMP_PREKEEP)
451: p->flags |= TERMP_KEEP;
452: if (p->flags & TERMP_NONOSPACE)
453: p->flags |= TERMP_NOSPACE;
454: else
455: p->flags &= ~TERMP_NOSPACE;
456: p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE);
457: p->skipvsp = 0;
1.31 schwarze 458: }
459:
1.59 schwarze 460: while ('\0' != *word) {
1.64 schwarze 461: if ('\\' != *word) {
1.75 schwarze 462: if (TERMP_NBRWORD & p->flags) {
463: if (' ' == *word) {
464: encode(p, nbrsp, 1);
465: word++;
466: continue;
467: }
468: ssz = strcspn(word, "\\ ");
469: } else
470: ssz = strcspn(word, "\\");
1.45 schwarze 471: encode(p, word, ssz);
1.64 schwarze 472: word += (int)ssz;
1.20 schwarze 473: continue;
1.64 schwarze 474: }
1.20 schwarze 475:
1.59 schwarze 476: word++;
477: esc = mandoc_escape(&word, &seq, &sz);
478: if (ESCAPE_ERROR == esc)
1.85 schwarze 479: continue;
1.59 schwarze 480:
481: switch (esc) {
1.83 schwarze 482: case ESCAPE_UNICODE:
1.89 schwarze 483: uc = mchars_num2uc(seq + 1, sz - 1);
1.56 schwarze 484: break;
1.83 schwarze 485: case ESCAPE_NUMBERED:
1.93 schwarze 486: uc = mchars_num2char(seq, sz);
487: if (uc < 0)
488: continue;
1.20 schwarze 489: break;
1.83 schwarze 490: case ESCAPE_SPECIAL:
1.89 schwarze 491: if (p->enc == TERMENC_ASCII) {
1.114 schwarze 492: cp = mchars_spec2str(seq, sz, &ssz);
1.92 schwarze 493: if (cp != NULL)
1.89 schwarze 494: encode(p, cp, ssz);
495: } else {
1.114 schwarze 496: uc = mchars_spec2cp(seq, sz);
1.90 schwarze 497: if (uc > 0)
498: encode1(p, uc);
1.89 schwarze 499: }
1.93 schwarze 500: continue;
1.83 schwarze 501: case ESCAPE_FONTBOLD:
1.20 schwarze 502: term_fontrepl(p, TERMFONT_BOLD);
1.93 schwarze 503: continue;
1.83 schwarze 504: case ESCAPE_FONTITALIC:
1.20 schwarze 505: term_fontrepl(p, TERMFONT_UNDER);
1.93 schwarze 506: continue;
1.83 schwarze 507: case ESCAPE_FONTBI:
1.70 schwarze 508: term_fontrepl(p, TERMFONT_BI);
1.93 schwarze 509: continue;
1.83 schwarze 510: case ESCAPE_FONT:
1.136 ! schwarze 511: case ESCAPE_FONTCW:
1.83 schwarze 512: case ESCAPE_FONTROMAN:
1.20 schwarze 513: term_fontrepl(p, TERMFONT_NONE);
1.93 schwarze 514: continue;
1.83 schwarze 515: case ESCAPE_FONTPREV:
1.20 schwarze 516: term_fontlast(p);
1.130 schwarze 517: continue;
518: case ESCAPE_BREAK:
519: bufferc(p, '\n');
1.93 schwarze 520: continue;
1.83 schwarze 521: case ESCAPE_NOSPACE:
1.108 schwarze 522: if (p->flags & TERMP_BACKAFTER)
523: p->flags &= ~TERMP_BACKAFTER;
524: else if (*word == '\0')
1.97 schwarze 525: p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE);
1.121 schwarze 526: continue;
1.135 schwarze 527: case ESCAPE_DEVICE:
528: if (p->type == TERMTYPE_PDF)
529: encode(p, "pdf", 3);
530: else if (p->type == TERMTYPE_PS)
531: encode(p, "ps", 2);
532: else if (p->enc == TERMENC_ASCII)
533: encode(p, "ascii", 5);
534: else
535: encode(p, "utf8", 4);
536: continue;
1.121 schwarze 537: case ESCAPE_HORIZ:
1.133 schwarze 538: if (*seq == '|') {
539: seq++;
540: uc = -p->col;
541: } else
542: uc = 0;
1.128 schwarze 543: if (a2roffsu(seq, &su, SCALE_EM) == NULL)
1.121 schwarze 544: continue;
1.133 schwarze 545: uc += term_hen(p, &su);
1.121 schwarze 546: if (uc > 0)
547: while (uc-- > 0)
548: bufferc(p, ASCII_NBRSP);
549: else if (p->col > (size_t)(-uc))
550: p->col += uc;
551: else {
552: uc += p->col;
553: p->col = 0;
1.126 schwarze 554: if (p->tcol->offset > (size_t)(-uc)) {
1.121 schwarze 555: p->ti += uc;
1.126 schwarze 556: p->tcol->offset += uc;
1.121 schwarze 557: } else {
1.126 schwarze 558: p->ti -= p->tcol->offset;
559: p->tcol->offset = 0;
1.121 schwarze 560: }
1.122 schwarze 561: }
562: continue;
563: case ESCAPE_HLINE:
1.132 schwarze 564: if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL)
1.122 schwarze 565: continue;
1.131 schwarze 566: uc = term_hen(p, &su);
1.122 schwarze 567: if (uc <= 0) {
1.126 schwarze 568: if (p->tcol->rmargin <= p->tcol->offset)
1.122 schwarze 569: continue;
1.126 schwarze 570: lsz = p->tcol->rmargin - p->tcol->offset;
1.122 schwarze 571: } else
572: lsz = uc;
1.132 schwarze 573: if (*cp == seq[-1])
1.122 schwarze 574: uc = -1;
1.132 schwarze 575: else if (*cp == '\\') {
576: seq = cp + 1;
1.122 schwarze 577: esc = mandoc_escape(&seq, &cp, &sz);
578: switch (esc) {
579: case ESCAPE_UNICODE:
580: uc = mchars_num2uc(cp + 1, sz - 1);
581: break;
582: case ESCAPE_NUMBERED:
583: uc = mchars_num2char(cp, sz);
584: break;
585: case ESCAPE_SPECIAL:
586: uc = mchars_spec2cp(cp, sz);
587: break;
588: default:
589: uc = -1;
590: break;
591: }
592: } else
1.132 schwarze 593: uc = *cp;
1.122 schwarze 594: if (uc < 0x20 || (uc > 0x7E && uc < 0xA0))
595: uc = '_';
596: if (p->enc == TERMENC_ASCII) {
597: cp = ascii_uc2str(uc);
598: csz = term_strlen(p, cp);
599: ssz = strlen(cp);
600: } else
601: csz = (*p->width)(p, uc);
602: while (lsz >= csz) {
603: if (p->enc == TERMENC_ASCII)
604: encode(p, cp, ssz);
605: else
606: encode1(p, uc);
607: lsz -= csz;
1.121 schwarze 608: }
1.93 schwarze 609: continue;
1.83 schwarze 610: case ESCAPE_SKIPCHAR:
1.108 schwarze 611: p->flags |= TERMP_BACKAFTER;
1.93 schwarze 612: continue;
1.103 schwarze 613: case ESCAPE_OVERSTRIKE:
614: cp = seq + sz;
615: while (seq < cp) {
616: if (*seq == '\\') {
617: mandoc_escape(&seq, NULL, NULL);
618: continue;
619: }
620: encode1(p, *seq++);
1.108 schwarze 621: if (seq < cp) {
622: if (p->flags & TERMP_BACKBEFORE)
623: p->flags |= TERMP_BACKAFTER;
624: else
625: p->flags |= TERMP_BACKBEFORE;
626: }
1.103 schwarze 627: }
1.109 schwarze 628: /* Trim trailing backspace/blank pair. */
1.129 schwarze 629: if (p->tcol->lastcol > 2 &&
630: (p->tcol->buf[p->tcol->lastcol - 1] == ' ' ||
631: p->tcol->buf[p->tcol->lastcol - 1] == '\t'))
632: p->tcol->lastcol -= 2;
633: if (p->col > p->tcol->lastcol)
634: p->col = p->tcol->lastcol;
1.108 schwarze 635: continue;
1.20 schwarze 636: default:
1.93 schwarze 637: continue;
638: }
639:
640: /*
641: * Common handling for Unicode and numbered
642: * character escape sequences.
643: */
644:
645: if (p->enc == TERMENC_ASCII) {
646: cp = ascii_uc2str(uc);
647: encode(p, cp, strlen(cp));
648: } else {
649: if ((uc < 0x20 && uc != 0x09) ||
650: (uc > 0x7E && uc < 0xA0))
651: uc = 0xFFFD;
652: encode1(p, uc);
1.20 schwarze 653: }
654: }
1.75 schwarze 655: p->flags &= ~TERMP_NBRWORD;
1.1 kristaps 656: }
657:
658: static void
1.126 schwarze 659: adjbuf(struct termp_col *c, size_t sz)
1.1 kristaps 660: {
1.126 schwarze 661: if (c->maxcols == 0)
662: c->maxcols = 1024;
663: while (c->maxcols <= sz)
664: c->maxcols <<= 2;
665: c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf));
1.1 kristaps 666: }
667:
1.4 schwarze 668: static void
1.20 schwarze 669: bufferc(struct termp *p, char c)
670: {
1.124 schwarze 671: if (p->flags & TERMP_NOBUF) {
672: (*p->letter)(p, c);
673: return;
674: }
1.126 schwarze 675: if (p->col + 1 >= p->tcol->maxcols)
676: adjbuf(p->tcol, p->col + 1);
1.129 schwarze 677: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 678: p->tcol->buf[p->col] = c;
1.129 schwarze 679: if (p->tcol->lastcol < ++p->col)
680: p->tcol->lastcol = p->col;
1.20 schwarze 681: }
682:
1.59 schwarze 683: /*
684: * See encode().
685: * Do this for a single (probably unicode) value.
686: * Does not check for non-decorated glyphs.
687: */
688: static void
689: encode1(struct termp *p, int c)
690: {
691: enum termfont f;
692:
1.124 schwarze 693: if (p->flags & TERMP_NOBUF) {
694: (*p->letter)(p, c);
695: return;
696: }
697:
1.126 schwarze 698: if (p->col + 7 >= p->tcol->maxcols)
699: adjbuf(p->tcol, p->col + 7);
1.59 schwarze 700:
1.115 schwarze 701: f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ?
1.108 schwarze 702: p->fontq[p->fonti] : TERMFONT_NONE;
1.59 schwarze 703:
1.108 schwarze 704: if (p->flags & TERMP_BACKBEFORE) {
1.126 schwarze 705: if (p->tcol->buf[p->col - 1] == ' ' ||
706: p->tcol->buf[p->col - 1] == '\t')
1.109 schwarze 707: p->col--;
708: else
1.126 schwarze 709: p->tcol->buf[p->col++] = '\b';
1.108 schwarze 710: p->flags &= ~TERMP_BACKBEFORE;
711: }
1.126 schwarze 712: if (f == TERMFONT_UNDER || f == TERMFONT_BI) {
713: p->tcol->buf[p->col++] = '_';
714: p->tcol->buf[p->col++] = '\b';
715: }
716: if (f == TERMFONT_BOLD || f == TERMFONT_BI) {
717: if (c == ASCII_HYPH)
718: p->tcol->buf[p->col++] = '-';
1.70 schwarze 719: else
1.126 schwarze 720: p->tcol->buf[p->col++] = c;
721: p->tcol->buf[p->col++] = '\b';
1.70 schwarze 722: }
1.129 schwarze 723: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 724: p->tcol->buf[p->col] = c;
1.129 schwarze 725: if (p->tcol->lastcol < ++p->col)
726: p->tcol->lastcol = p->col;
1.108 schwarze 727: if (p->flags & TERMP_BACKAFTER) {
728: p->flags |= TERMP_BACKBEFORE;
729: p->flags &= ~TERMP_BACKAFTER;
730: }
1.59 schwarze 731: }
1.20 schwarze 732:
733: static void
734: encode(struct termp *p, const char *word, size_t sz)
1.4 schwarze 735: {
1.71 schwarze 736: size_t i;
1.124 schwarze 737:
738: if (p->flags & TERMP_NOBUF) {
739: for (i = 0; i < sz; i++)
740: (*p->letter)(p, word[i]);
741: return;
742: }
1.59 schwarze 743:
1.126 schwarze 744: if (p->col + 2 + (sz * 5) >= p->tcol->maxcols)
745: adjbuf(p->tcol, p->col + 2 + (sz * 5));
1.46 schwarze 746:
1.71 schwarze 747: for (i = 0; i < sz; i++) {
1.70 schwarze 748: if (ASCII_HYPH == word[i] ||
749: isgraph((unsigned char)word[i]))
750: encode1(p, word[i]);
1.119 schwarze 751: else {
1.129 schwarze 752: if (p->tcol->lastcol <= p->col ||
1.125 schwarze 753: (word[i] != ' ' && word[i] != ASCII_NBRSP))
1.126 schwarze 754: p->tcol->buf[p->col] = word[i];
1.125 schwarze 755: p->col++;
1.119 schwarze 756:
757: /*
758: * Postpone the effect of \z while handling
759: * an overstrike sequence from ascii_uc2str().
760: */
761:
762: if (word[i] == '\b' &&
763: (p->flags & TERMP_BACKBEFORE)) {
764: p->flags &= ~TERMP_BACKBEFORE;
765: p->flags |= TERMP_BACKAFTER;
766: }
767: }
1.4 schwarze 768: }
1.129 schwarze 769: if (p->tcol->lastcol < p->col)
770: p->tcol->lastcol = p->col;
1.80 schwarze 771: }
772:
773: void
774: term_setwidth(struct termp *p, const char *wstr)
775: {
776: struct roffsu su;
1.107 schwarze 777: int iop, width;
1.80 schwarze 778:
1.81 schwarze 779: iop = 0;
780: width = 0;
1.80 schwarze 781: if (NULL != wstr) {
782: switch (*wstr) {
1.83 schwarze 783: case '+':
1.80 schwarze 784: iop = 1;
785: wstr++;
786: break;
1.83 schwarze 787: case '-':
1.80 schwarze 788: iop = -1;
789: wstr++;
790: break;
791: default:
792: break;
793: }
1.128 schwarze 794: if (a2roffsu(wstr, &su, SCALE_MAX) != NULL)
1.81 schwarze 795: width = term_hspan(p, &su);
796: else
1.80 schwarze 797: iop = 0;
798: }
799: (*p->setwidth)(p, iop, width);
1.4 schwarze 800: }
1.16 schwarze 801:
802: size_t
1.39 schwarze 803: term_len(const struct termp *p, size_t sz)
804: {
805:
1.112 schwarze 806: return (*p->width)(p, ' ') * sz;
1.39 schwarze 807: }
808:
1.64 schwarze 809: static size_t
810: cond_width(const struct termp *p, int c, int *skip)
811: {
812:
813: if (*skip) {
814: (*skip) = 0;
1.112 schwarze 815: return 0;
1.64 schwarze 816: } else
1.112 schwarze 817: return (*p->width)(p, c);
1.64 schwarze 818: }
1.39 schwarze 819:
820: size_t
821: term_strlen(const struct termp *p, const char *cp)
822: {
1.59 schwarze 823: size_t sz, rsz, i;
1.93 schwarze 824: int ssz, skip, uc;
1.50 schwarze 825: const char *seq, *rhs;
1.59 schwarze 826: enum mandoc_esc esc;
1.77 schwarze 827: static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH,
828: ASCII_BREAK, '\0' };
1.59 schwarze 829:
830: /*
831: * Account for escaped sequences within string length
832: * calculations. This follows the logic in term_word() as we
833: * must calculate the width of produced strings.
834: */
835:
836: sz = 0;
1.64 schwarze 837: skip = 0;
1.59 schwarze 838: while ('\0' != *cp) {
839: rsz = strcspn(cp, rej);
840: for (i = 0; i < rsz; i++)
1.64 schwarze 841: sz += cond_width(p, *cp++, &skip);
1.59 schwarze 842:
843: switch (*cp) {
1.83 schwarze 844: case '\\':
1.59 schwarze 845: cp++;
846: esc = mandoc_escape(&cp, &seq, &ssz);
847: if (ESCAPE_ERROR == esc)
1.85 schwarze 848: continue;
1.59 schwarze 849:
850: rhs = NULL;
1.50 schwarze 851:
1.59 schwarze 852: switch (esc) {
1.83 schwarze 853: case ESCAPE_UNICODE:
1.94 schwarze 854: uc = mchars_num2uc(seq + 1, ssz - 1);
1.59 schwarze 855: break;
1.83 schwarze 856: case ESCAPE_NUMBERED:
1.93 schwarze 857: uc = mchars_num2char(seq, ssz);
858: if (uc < 0)
859: continue;
1.50 schwarze 860: break;
1.83 schwarze 861: case ESCAPE_SPECIAL:
1.93 schwarze 862: if (p->enc == TERMENC_ASCII) {
1.114 schwarze 863: rhs = mchars_spec2str(seq, ssz, &rsz);
1.93 schwarze 864: if (rhs != NULL)
865: break;
866: } else {
1.114 schwarze 867: uc = mchars_spec2cp(seq, ssz);
1.93 schwarze 868: if (uc > 0)
869: sz += cond_width(p, uc, &skip);
1.89 schwarze 870: }
1.93 schwarze 871: continue;
1.135 schwarze 872: case ESCAPE_DEVICE:
873: if (p->type == TERMTYPE_PDF) {
874: rhs = "pdf";
875: rsz = 3;
876: } else if (p->type == TERMTYPE_PS) {
877: rhs = "ps";
878: rsz = 2;
879: } else if (p->enc == TERMENC_ASCII) {
880: rhs = "ascii";
881: rsz = 5;
882: } else {
883: rhs = "utf8";
884: rsz = 4;
885: }
886: break;
1.83 schwarze 887: case ESCAPE_SKIPCHAR:
1.64 schwarze 888: skip = 1;
1.103 schwarze 889: continue;
890: case ESCAPE_OVERSTRIKE:
891: rsz = 0;
892: rhs = seq + ssz;
893: while (seq < rhs) {
894: if (*seq == '\\') {
895: mandoc_escape(&seq, NULL, NULL);
896: continue;
897: }
898: i = (*p->width)(p, *seq++);
899: if (rsz < i)
900: rsz = i;
901: }
902: sz += rsz;
1.93 schwarze 903: continue;
1.50 schwarze 904: default:
1.93 schwarze 905: continue;
1.50 schwarze 906: }
1.39 schwarze 907:
1.93 schwarze 908: /*
909: * Common handling for Unicode and numbered
910: * character escape sequences.
911: */
912:
913: if (rhs == NULL) {
914: if (p->enc == TERMENC_ASCII) {
915: rhs = ascii_uc2str(uc);
916: rsz = strlen(rhs);
917: } else {
918: if ((uc < 0x20 && uc != 0x09) ||
919: (uc > 0x7E && uc < 0xA0))
920: uc = 0xFFFD;
921: sz += cond_width(p, uc, &skip);
922: continue;
923: }
924: }
1.59 schwarze 925:
1.64 schwarze 926: if (skip) {
927: skip = 0;
928: break;
929: }
1.93 schwarze 930:
931: /*
932: * Common handling for all escape sequences
933: * printing more than one character.
934: */
1.64 schwarze 935:
1.59 schwarze 936: for (i = 0; i < rsz; i++)
937: sz += (*p->width)(p, *rhs++);
938: break;
1.83 schwarze 939: case ASCII_NBRSP:
1.64 schwarze 940: sz += cond_width(p, ' ', &skip);
1.55 schwarze 941: cp++;
1.59 schwarze 942: break;
1.83 schwarze 943: case ASCII_HYPH:
1.64 schwarze 944: sz += cond_width(p, '-', &skip);
1.55 schwarze 945: cp++;
1.59 schwarze 946: break;
947: default:
948: break;
949: }
950: }
1.39 schwarze 951:
1.112 schwarze 952: return sz;
1.39 schwarze 953: }
954:
1.100 schwarze 955: int
1.39 schwarze 956: term_vspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 957: {
958: double r;
1.101 schwarze 959: int ri;
1.16 schwarze 960:
961: switch (su->unit) {
1.99 schwarze 962: case SCALE_BU:
963: r = su->scale / 40.0;
964: break;
1.83 schwarze 965: case SCALE_CM:
1.99 schwarze 966: r = su->scale * 6.0 / 2.54;
967: break;
968: case SCALE_FS:
969: r = su->scale * 65536.0 / 40.0;
1.16 schwarze 970: break;
1.83 schwarze 971: case SCALE_IN:
1.86 schwarze 972: r = su->scale * 6.0;
1.16 schwarze 973: break;
1.99 schwarze 974: case SCALE_MM:
975: r = su->scale * 0.006;
976: break;
1.83 schwarze 977: case SCALE_PC:
1.16 schwarze 978: r = su->scale;
979: break;
1.83 schwarze 980: case SCALE_PT:
1.99 schwarze 981: r = su->scale / 12.0;
1.16 schwarze 982: break;
1.99 schwarze 983: case SCALE_EN:
984: case SCALE_EM:
985: r = su->scale * 0.6;
1.16 schwarze 986: break;
1.83 schwarze 987: case SCALE_VS:
1.16 schwarze 988: r = su->scale;
989: break;
990: default:
1.99 schwarze 991: abort();
1.16 schwarze 992: }
1.101 schwarze 993: ri = r > 0.0 ? r + 0.4995 : r - 0.4995;
1.112 schwarze 994: return ri < 66 ? ri : 1;
1.16 schwarze 995: }
996:
1.107 schwarze 997: /*
1.131 schwarze 998: * Convert a scaling width to basic units, rounding towards 0.
1.107 schwarze 999: */
1.100 schwarze 1000: int
1.39 schwarze 1001: term_hspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 1002: {
1003:
1.112 schwarze 1004: return (*p->hspan)(p, su);
1.131 schwarze 1005: }
1006:
1007: /*
1008: * Convert a scaling width to basic units, rounding to closest.
1009: */
1010: int
1011: term_hen(const struct termp *p, const struct roffsu *su)
1012: {
1013: int bu;
1014:
1015: if ((bu = (*p->hspan)(p, su)) >= 0)
1016: return (bu + 11) / 24;
1017: else
1018: return -((-bu + 11) / 24);
1.16 schwarze 1019: }