Annotation of mandoc/term.c, Revision 1.290
1.290 ! schwarze 1: /* $Id: term.c,v 1.289 2022/08/15 18:12:30 schwarze Exp $ */
1.1 kristaps 2: /*
1.285 schwarze 3: * Copyright (c) 2010-2022 Ingo Schwarze <schwarze@openbsd.org>
1.198 schwarze 4: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.74 kristaps 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: *
1.246 schwarze 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.74 kristaps 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.246 schwarze 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.74 kristaps 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: */
1.128 kristaps 18: #include "config.h"
19:
1.126 kristaps 20: #include <sys/types.h>
21:
1.1 kristaps 22: #include <assert.h>
1.122 kristaps 23: #include <ctype.h>
1.278 schwarze 24: #include <stdint.h>
1.22 kristaps 25: #include <stdio.h>
1.1 kristaps 26: #include <stdlib.h>
27: #include <string.h>
28:
1.137 kristaps 29: #include "mandoc.h"
1.218 schwarze 30: #include "mandoc_aux.h"
1.107 kristaps 31: #include "out.h"
1.71 kristaps 32: #include "term.h"
1.105 kristaps 33: #include "main.h"
1.1 kristaps 34:
1.203 schwarze 35: static size_t cond_width(const struct termp *, int, int *);
1.266 schwarze 36: static void adjbuf(struct termp_col *, size_t);
1.191 kristaps 37: static void bufferc(struct termp *, char);
38: static void encode(struct termp *, const char *, size_t);
1.194 kristaps 39: static void encode1(struct termp *, int);
1.264 schwarze 40: static void endline(struct termp *);
1.282 schwarze 41: static void term_field(struct termp *, size_t, size_t);
1.278 schwarze 42: static void term_fill(struct termp *, size_t *, size_t *,
43: size_t);
1.11 kristaps 44:
1.222 schwarze 45:
1.145 kristaps 46: void
1.269 schwarze 47: term_setcol(struct termp *p, size_t maxtcol)
48: {
49: if (maxtcol > p->maxtcol) {
50: p->tcols = mandoc_recallocarray(p->tcols,
51: p->maxtcol, maxtcol, sizeof(*p->tcols));
52: p->maxtcol = maxtcol;
53: }
54: p->lasttcol = maxtcol - 1;
55: p->tcol = p->tcols;
56: }
57:
58: void
1.71 kristaps 59: term_free(struct termp *p)
1.14 kristaps 60: {
1.284 schwarze 61: term_tab_free();
1.266 schwarze 62: for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++)
63: free(p->tcol->buf);
64: free(p->tcols);
1.238 schwarze 65: free(p->fontq);
1.142 kristaps 66: free(p);
67: }
68:
69: void
1.222 schwarze 70: term_begin(struct termp *p, term_margin head,
1.246 schwarze 71: term_margin foot, const struct roff_meta *arg)
1.142 kristaps 72: {
73:
74: p->headf = head;
75: p->footf = foot;
76: p->argf = arg;
1.146 kristaps 77: (*p->begin)(p);
1.142 kristaps 78: }
79:
80: void
81: term_end(struct termp *p)
82: {
83:
1.146 kristaps 84: (*p->end)(p);
1.14 kristaps 85: }
86:
1.71 kristaps 87: /*
1.221 schwarze 88: * Flush a chunk of text. By default, break the output line each time
89: * the right margin is reached, and continue output on the next line
90: * at the same offset as the chunk itself. By default, also break the
1.278 schwarze 91: * output line at the end of the chunk. There are many flags modifying
92: * this behaviour, see the comments in the body of the function.
1.71 kristaps 93: */
94: void
95: term_flushln(struct termp *p)
1.53 kristaps 96: {
1.278 schwarze 97: size_t vbl; /* Number of blanks to prepend to the output. */
98: size_t vbr; /* Actual visual position of the end of field. */
99: size_t vfield; /* Desired visual field width. */
100: size_t vtarget; /* Desired visual position of the right margin. */
101: size_t ic; /* Character position in the input buffer. */
102: size_t nbr; /* Number of characters to print in this field. */
103:
104: /*
105: * Normally, start writing at the left margin, but with the
106: * NOPAD flag, start writing at the current position instead.
107: */
1.53 kristaps 108:
1.266 schwarze 109: vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ?
110: 0 : p->tcol->offset - p->viscol;
1.263 schwarze 111: if (p->minbl && vbl < p->minbl)
112: vbl = p->minbl;
1.115 kristaps 113:
1.274 schwarze 114: if ((p->flags & TERMP_MULTICOL) == 0)
1.267 schwarze 115: p->tcol->col = 0;
1.278 schwarze 116:
117: /* Loop over output lines. */
118:
119: for (;;) {
120: vfield = p->tcol->rmargin > p->viscol + vbl ?
121: p->tcol->rmargin - p->viscol - vbl : 0;
1.267 schwarze 122:
1.71 kristaps 123: /*
1.278 schwarze 124: * Normally, break the line at the the right margin
125: * of the field, but with the NOBREAK flag, only
126: * break it at the max right margin of the screen,
127: * and with the BRNEVER flag, never break it at all.
1.138 schwarze 128: */
1.267 schwarze 129:
1.282 schwarze 130: vtarget = (p->flags & TERMP_NOBREAK) == 0 ? vfield :
1.278 schwarze 131: p->maxrmargin > p->viscol + vbl ?
132: p->maxrmargin - p->viscol - vbl : 0;
133:
134: /*
135: * Figure out how much text will fit in the field.
136: * If there is whitespace only, print nothing.
137: */
138:
1.282 schwarze 139: term_fill(p, &nbr, &vbr,
140: p->flags & TERMP_BRNEVER ? SIZE_MAX : vtarget);
1.278 schwarze 141: if (nbr == 0)
142: break;
1.279 schwarze 143:
144: /*
145: * With the CENTER or RIGHT flag, increase the indentation
146: * to center the text between the left and right margins
147: * or to adjust it to the right margin, respectively.
148: */
149:
150: if (vbr < vtarget) {
151: if (p->flags & TERMP_CENTER)
152: vbl += (vtarget - vbr) / 2;
153: else if (p->flags & TERMP_RIGHT)
154: vbl += vtarget - vbr;
155: }
156:
157: /* Finally, print the field content. */
1.278 schwarze 158:
1.282 schwarze 159: term_field(p, vbl, nbr);
1.290 ! schwarze 160: if (vbr < vtarget)
! 161: p->tcol->taboff += vbr;
! 162: else
! 163: p->tcol->taboff += vtarget;
! 164: p->tcol->taboff += (*p->width)(p, ' ');
1.138 schwarze 165:
166: /*
1.278 schwarze 167: * If there is no text left in the field, exit the loop.
168: * If the BRTRSP flag is set, consider trailing
169: * whitespace significant when deciding whether
170: * the field fits or not.
1.71 kristaps 171: */
172:
1.278 schwarze 173: for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
174: switch (p->tcol->buf[ic]) {
175: case '\t':
176: if (p->flags & TERMP_BRTRSP)
177: vbr = term_tab_next(vbr);
178: continue;
179: case ' ':
180: if (p->flags & TERMP_BRTRSP)
181: vbr += (*p->width)(p, ' ');
1.270 schwarze 182: continue;
1.278 schwarze 183: case '\n':
1.290 ! schwarze 184: case ASCII_NBRZW:
1.278 schwarze 185: case ASCII_BREAK:
1.290 ! schwarze 186: case ASCII_TABREF:
1.278 schwarze 187: continue;
188: default:
189: break;
1.270 schwarze 190: }
1.278 schwarze 191: break;
192: }
193: if (ic == p->tcol->lastcol)
194: break;
1.154 kristaps 195:
1.278 schwarze 196: /*
197: * At the location of an automtic line break, input
198: * space characters are consumed by the line break.
199: */
1.154 kristaps 200:
1.278 schwarze 201: while (p->tcol->col < p->tcol->lastcol &&
202: p->tcol->buf[p->tcol->col] == ' ')
203: p->tcol->col++;
1.154 kristaps 204:
1.278 schwarze 205: /*
206: * In multi-column mode, leave the rest of the text
207: * in the buffer to be handled by a subsequent
208: * invocation, such that the other columns of the
209: * table can be handled first.
210: * In single-column mode, simply break the line.
211: */
212:
213: if (p->flags & TERMP_MULTICOL)
214: return;
1.217 schwarze 215:
1.278 schwarze 216: endline(p);
1.53 kristaps 217:
1.71 kristaps 218: /*
1.278 schwarze 219: * Normally, start the next line at the same indentation
220: * as this one, but with the BRIND flag, start it at the
221: * right margin instead. This is used together with
222: * NOBREAK for the tags in various kinds of tagged lists.
1.81 kristaps 223: */
1.267 schwarze 224:
1.278 schwarze 225: vbl = p->flags & TERMP_BRIND ?
226: p->tcol->rmargin : p->tcol->offset;
227: }
1.267 schwarze 228:
1.278 schwarze 229: /* Reset output state in preparation for the next field. */
1.205 schwarze 230:
1.278 schwarze 231: p->col = p->tcol->col = p->tcol->lastcol = 0;
232: p->minbl = p->trailspace;
233: p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE | TERMP_NOPAD);
1.260 schwarze 234:
1.278 schwarze 235: if (p->flags & TERMP_MULTICOL)
236: return;
1.260 schwarze 237:
1.278 schwarze 238: /*
239: * The HANG flag means that the next field
240: * always follows on the same line.
241: * The NOBREAK flag means that the next field
242: * follows on the same line unless the field was overrun.
243: * Normally, break the line at the end of each field.
244: */
1.205 schwarze 245:
1.278 schwarze 246: if ((p->flags & TERMP_HANG) == 0 &&
247: ((p->flags & TERMP_NOBREAK) == 0 ||
248: vbr + term_len(p, p->trailspace) > vfield))
249: endline(p);
250: }
1.138 schwarze 251:
1.278 schwarze 252: /*
253: * Store the number of input characters to print in this field in *nbr
254: * and their total visual width to print in *vbr.
255: * If there is only whitespace in the field, both remain zero.
256: * The desired visual width of the field is provided by vtarget.
257: * If the first word is longer, the field will be overrun.
258: */
259: static void
260: term_fill(struct termp *p, size_t *nbr, size_t *vbr, size_t vtarget)
261: {
262: size_t ic; /* Character position in the input buffer. */
263: size_t vis; /* Visual position of the current character. */
264: size_t vn; /* Visual position of the next character. */
265: int breakline; /* Break at the end of this word. */
266: int graph; /* Last character was non-blank. */
1.290 ! schwarze 267: int taboff; /* Temporary offset for literal tabs. */
1.278 schwarze 268:
269: *nbr = *vbr = vis = 0;
270: breakline = graph = 0;
1.290 ! schwarze 271: taboff = p->tcol->taboff;
1.278 schwarze 272: for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
273: switch (p->tcol->buf[ic]) {
274: case '\b': /* Escape \o (overstrike) or backspace markup. */
275: assert(ic > 0);
276: vis -= (*p->width)(p, p->tcol->buf[ic - 1]);
277: continue;
1.267 schwarze 278:
1.278 schwarze 279: case ' ':
280: case ASCII_BREAK: /* Escape \: (breakpoint). */
1.287 schwarze 281: vn = vis;
282: if (p->tcol->buf[ic] == ' ')
283: vn += (*p->width)(p, ' ');
1.278 schwarze 284: /* Can break at the end of a word. */
285: if (breakline || vn > vtarget)
286: break;
287: if (graph) {
288: *nbr = ic;
289: *vbr = vis;
290: graph = 0;
1.136 schwarze 291: }
1.278 schwarze 292: vis = vn;
293: continue;
294:
295: case '\n': /* Escape \p (break at the end of the word). */
296: breakline = 1;
297: continue;
1.130 kristaps 298:
1.278 schwarze 299: case ASCII_HYPH: /* Breakable hyphen. */
300: graph = 1;
1.136 schwarze 301: /*
1.278 schwarze 302: * We are about to decide whether to break the
303: * line or not, so we no longer need this hyphen
304: * to be marked as breakable. Put back a real
305: * hyphen such that we get the correct width.
1.136 schwarze 306: */
1.278 schwarze 307: p->tcol->buf[ic] = '-';
308: vis += (*p->width)(p, '-');
309: if (vis > vtarget) {
310: ic++;
311: break;
1.200 schwarze 312: }
1.278 schwarze 313: *nbr = ic + 1;
314: *vbr = vis;
315: continue;
1.200 schwarze 316:
1.290 ! schwarze 317: case ASCII_TABREF:
! 318: taboff = -vis - (*p->width)(p, ' ');
! 319: continue;
! 320:
1.287 schwarze 321: default:
322: switch (p->tcol->buf[ic]) {
323: case '\t':
1.290 ! schwarze 324: if (taboff < 0 && (size_t)-taboff > vis)
! 325: vis = 0;
! 326: else
! 327: vis += taboff;
1.287 schwarze 328: vis = term_tab_next(vis);
1.290 ! schwarze 329: vis -= taboff;
1.287 schwarze 330: break;
1.289 schwarze 331: case ASCII_NBRZW: /* Non-breakable zero-width. */
332: break;
1.287 schwarze 333: case ASCII_NBRSP: /* Non-breakable space. */
334: p->tcol->buf[ic] = ' ';
335: /* FALLTHROUGH */
336: default: /* Printable character. */
337: vis += (*p->width)(p, p->tcol->buf[ic]);
338: break;
339: }
1.278 schwarze 340: graph = 1;
341: if (vis > vtarget && *nbr > 0)
342: return;
343: continue;
1.136 schwarze 344: }
1.278 schwarze 345: break;
346: }
1.270 schwarze 347:
1.278 schwarze 348: /*
349: * If the last word extends to the end of the field without any
350: * trailing whitespace, the loop could not check yet whether it
351: * can remain on this line. So do the check now.
352: */
1.270 schwarze 353:
1.278 schwarze 354: if (graph && (vis <= vtarget || *nbr == 0)) {
355: *nbr = ic;
356: *vbr = vis;
357: }
358: }
1.270 schwarze 359:
1.278 schwarze 360: /*
361: * Print the contents of one field
362: * with an indentation of vbl visual columns,
1.282 schwarze 363: * and an input string length of nbr characters.
1.278 schwarze 364: */
365: static void
1.282 schwarze 366: term_field(struct termp *p, size_t vbl, size_t nbr)
1.278 schwarze 367: {
368: size_t ic; /* Character position in the input buffer. */
369: size_t vis; /* Visual position of the current character. */
1.288 schwarze 370: size_t vt; /* Visual position including tab offset. */
1.278 schwarze 371: size_t dv; /* Visual width of the current character. */
1.290 ! schwarze 372: int taboff; /* Temporary offset for literal tabs. */
1.270 schwarze 373:
1.278 schwarze 374: vis = 0;
1.290 ! schwarze 375: taboff = p->tcol->taboff;
1.278 schwarze 376: for (ic = p->tcol->col; ic < nbr; ic++) {
1.270 schwarze 377:
1.278 schwarze 378: /*
379: * To avoid the printing of trailing whitespace,
380: * do not print whitespace right away, only count it.
381: */
1.270 schwarze 382:
1.278 schwarze 383: switch (p->tcol->buf[ic]) {
384: case '\n':
385: case ASCII_BREAK:
1.289 schwarze 386: case ASCII_NBRZW:
1.278 schwarze 387: continue;
1.290 ! schwarze 388: case ASCII_TABREF:
! 389: taboff = -vis - (*p->width)(p, ' ');
! 390: continue;
1.278 schwarze 391: case '\t':
392: case ' ':
393: case ASCII_NBRSP:
1.288 schwarze 394: if (p->tcol->buf[ic] == '\t') {
1.290 ! schwarze 395: if (taboff < 0 && (size_t)-taboff > vis)
! 396: vt = 0;
! 397: else
! 398: vt = vis + taboff;
1.288 schwarze 399: dv = term_tab_next(vt) - vt;
400: } else
401: dv = (*p->width)(p, ' ');
1.280 schwarze 402: vbl += dv;
403: vis += dv;
1.278 schwarze 404: continue;
405: default:
406: break;
407: }
1.168 schwarze 408:
1.278 schwarze 409: /*
410: * We found a non-blank character to print,
411: * so write preceding white space now.
412: */
1.267 schwarze 413:
1.278 schwarze 414: if (vbl > 0) {
415: (*p->advance)(p, vbl);
416: p->viscol += vbl;
417: vbl = 0;
418: }
1.111 kristaps 419:
1.278 schwarze 420: /* Print the character and adjust the visual position. */
1.15 kristaps 421:
1.278 schwarze 422: (*p->letter)(p, p->tcol->buf[ic]);
423: if (p->tcol->buf[ic] == '\b') {
424: dv = (*p->width)(p, p->tcol->buf[ic - 1]);
425: p->viscol -= dv;
426: vis -= dv;
427: } else {
428: dv = (*p->width)(p, p->tcol->buf[ic]);
429: p->viscol += dv;
430: vis += dv;
431: }
432: }
433: p->tcol->col = nbr;
1.264 schwarze 434: }
435:
436: static void
437: endline(struct termp *p)
438: {
439: if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) {
440: p->mc = NULL;
441: p->flags &= ~TERMP_ENDMC;
442: }
443: if (p->mc != NULL) {
444: if (p->viscol && p->maxrmargin >= p->viscol)
445: (*p->advance)(p, p->maxrmargin - p->viscol + 1);
446: p->flags |= TERMP_NOBUF | TERMP_NOSPACE;
447: term_word(p, p->mc);
448: p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC);
449: }
450: p->viscol = 0;
451: p->minbl = 0;
452: (*p->endline)(p);
1.15 kristaps 453: }
454:
1.222 schwarze 455: /*
1.71 kristaps 456: * A newline only breaks an existing line; it won't assert vertical
457: * space. All data in the output buffer is flushed prior to the newline
458: * assertion.
459: */
460: void
461: term_newln(struct termp *p)
1.15 kristaps 462: {
1.71 kristaps 463: p->flags |= TERMP_NOSPACE;
1.269 schwarze 464: if (p->tcol->lastcol || p->viscol)
1.200 schwarze 465: term_flushln(p);
1.290 ! schwarze 466: p->tcol->taboff = 0;
1.16 kristaps 467: }
468:
1.71 kristaps 469: /*
470: * Asserts a vertical space (a full, empty line-break between lines).
471: * Note that if used twice, this will cause two blank spaces and so on.
472: * All data in the output buffer is flushed prior to the newline
473: * assertion.
474: */
475: void
476: term_vspace(struct termp *p)
1.16 kristaps 477: {
478:
1.62 kristaps 479: term_newln(p);
1.139 schwarze 480: p->viscol = 0;
1.264 schwarze 481: p->minbl = 0;
1.202 schwarze 482: if (0 < p->skipvsp)
483: p->skipvsp--;
484: else
485: (*p->endline)(p);
1.16 kristaps 486: }
487:
1.238 schwarze 488: /* Swap current and previous font; for \fP and .ft P */
1.125 kristaps 489: void
490: term_fontlast(struct termp *p)
491: {
492: enum termfont f;
493:
494: f = p->fontl;
495: p->fontl = p->fontq[p->fonti];
496: p->fontq[p->fonti] = f;
497: }
498:
1.238 schwarze 499: /* Set font, save current, discard previous; for \f, .ft, .B etc. */
1.125 kristaps 500: void
501: term_fontrepl(struct termp *p, enum termfont f)
502: {
503:
504: p->fontl = p->fontq[p->fonti];
505: p->fontq[p->fonti] = f;
506: }
507:
1.238 schwarze 508: /* Set font, save previous. */
1.125 kristaps 509: void
510: term_fontpush(struct termp *p, enum termfont f)
511: {
512:
513: p->fontl = p->fontq[p->fonti];
1.238 schwarze 514: if (++p->fonti == p->fontsz) {
515: p->fontsz += 8;
516: p->fontq = mandoc_reallocarray(p->fontq,
1.256 schwarze 517: p->fontsz, sizeof(*p->fontq));
1.238 schwarze 518: }
519: p->fontq[p->fonti] = f;
1.125 kristaps 520: }
521:
1.238 schwarze 522: /* Flush to make the saved pointer current again. */
1.125 kristaps 523: void
1.244 schwarze 524: term_fontpopq(struct termp *p, int i)
1.125 kristaps 525: {
526:
1.244 schwarze 527: assert(i >= 0);
528: if (p->fonti > i)
529: p->fonti = i;
1.125 kristaps 530: }
1.94 kristaps 531:
1.238 schwarze 532: /* Pop one font off the stack. */
1.125 kristaps 533: void
534: term_fontpop(struct termp *p)
535: {
536:
537: assert(p->fonti);
538: p->fonti--;
1.17 kristaps 539: }
540:
1.71 kristaps 541: /*
542: * Handle pwords, partial words, which may be either a single word or a
543: * phrase that cannot be broken down (such as a literal string). This
544: * handles word styling.
545: */
1.86 kristaps 546: void
547: term_word(struct termp *p, const char *word)
1.65 kristaps 548: {
1.261 schwarze 549: struct roffsu su;
1.214 schwarze 550: const char nbrsp[2] = { ASCII_NBRSP, 0 };
1.191 kristaps 551: const char *seq, *cp;
1.194 kristaps 552: int sz, uc;
1.262 schwarze 553: size_t csz, lsz, ssz;
1.184 kristaps 554: enum mandoc_esc esc;
1.100 kristaps 555:
1.264 schwarze 556: if ((p->flags & TERMP_NOBUF) == 0) {
557: if ((p->flags & TERMP_NOSPACE) == 0) {
558: if ((p->flags & TERMP_KEEP) == 0) {
1.151 schwarze 559: bufferc(p, ' ');
1.264 schwarze 560: if (p->flags & TERMP_SENTENCE)
561: bufferc(p, ' ');
562: } else
563: bufferc(p, ASCII_NBRSP);
564: }
565: if (p->flags & TERMP_PREKEEP)
566: p->flags |= TERMP_KEEP;
567: if (p->flags & TERMP_NONOSPACE)
568: p->flags |= TERMP_NOSPACE;
569: else
570: p->flags &= ~TERMP_NOSPACE;
571: p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE);
572: p->skipvsp = 0;
1.133 kristaps 573: }
1.65 kristaps 574:
1.184 kristaps 575: while ('\0' != *word) {
1.203 schwarze 576: if ('\\' != *word) {
1.214 schwarze 577: if (TERMP_NBRWORD & p->flags) {
578: if (' ' == *word) {
579: encode(p, nbrsp, 1);
580: word++;
581: continue;
582: }
583: ssz = strcspn(word, "\\ ");
584: } else
585: ssz = strcspn(word, "\\");
1.162 kristaps 586: encode(p, word, ssz);
1.203 schwarze 587: word += (int)ssz;
1.124 kristaps 588: continue;
1.203 schwarze 589: }
1.124 kristaps 590:
1.184 kristaps 591: word++;
592: esc = mandoc_escape(&word, &seq, &sz);
593: switch (esc) {
1.222 schwarze 594: case ESCAPE_UNICODE:
1.229 schwarze 595: uc = mchars_num2uc(seq + 1, sz - 1);
1.192 kristaps 596: break;
1.222 schwarze 597: case ESCAPE_NUMBERED:
1.233 schwarze 598: uc = mchars_num2char(seq, sz);
1.289 schwarze 599: if (uc >= 0)
600: break;
601: bufferc(p, ASCII_NBRZW);
602: continue;
1.222 schwarze 603: case ESCAPE_SPECIAL:
1.229 schwarze 604: if (p->enc == TERMENC_ASCII) {
1.254 schwarze 605: cp = mchars_spec2str(seq, sz, &ssz);
1.232 schwarze 606: if (cp != NULL)
1.229 schwarze 607: encode(p, cp, ssz);
1.289 schwarze 608: else
609: bufferc(p, ASCII_NBRZW);
1.229 schwarze 610: } else {
1.254 schwarze 611: uc = mchars_spec2cp(seq, sz);
1.230 schwarze 612: if (uc > 0)
613: encode1(p, uc);
1.289 schwarze 614: else
615: bufferc(p, ASCII_NBRZW);
1.229 schwarze 616: }
1.233 schwarze 617: continue;
1.277 schwarze 618: case ESCAPE_UNDEF:
619: uc = *seq;
620: break;
1.222 schwarze 621: case ESCAPE_FONTBOLD:
1.283 schwarze 622: case ESCAPE_FONTCB:
1.125 kristaps 623: term_fontrepl(p, TERMFONT_BOLD);
1.233 schwarze 624: continue;
1.222 schwarze 625: case ESCAPE_FONTITALIC:
1.283 schwarze 626: case ESCAPE_FONTCI:
1.125 kristaps 627: term_fontrepl(p, TERMFONT_UNDER);
1.233 schwarze 628: continue;
1.222 schwarze 629: case ESCAPE_FONTBI:
1.209 schwarze 630: term_fontrepl(p, TERMFONT_BI);
1.233 schwarze 631: continue;
1.222 schwarze 632: case ESCAPE_FONT:
1.283 schwarze 633: case ESCAPE_FONTCR:
1.222 schwarze 634: case ESCAPE_FONTROMAN:
1.125 kristaps 635: term_fontrepl(p, TERMFONT_NONE);
1.233 schwarze 636: continue;
1.222 schwarze 637: case ESCAPE_FONTPREV:
1.125 kristaps 638: term_fontlast(p);
1.270 schwarze 639: continue;
640: case ESCAPE_BREAK:
641: bufferc(p, '\n');
1.233 schwarze 642: continue;
1.222 schwarze 643: case ESCAPE_NOSPACE:
1.248 schwarze 644: if (p->flags & TERMP_BACKAFTER)
645: p->flags &= ~TERMP_BACKAFTER;
646: else if (*word == '\0')
1.237 schwarze 647: p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE);
1.261 schwarze 648: continue;
1.275 schwarze 649: case ESCAPE_DEVICE:
650: if (p->type == TERMTYPE_PDF)
651: encode(p, "pdf", 3);
652: else if (p->type == TERMTYPE_PS)
653: encode(p, "ps", 2);
654: else if (p->enc == TERMENC_ASCII)
655: encode(p, "ascii", 5);
656: else
657: encode(p, "utf8", 4);
658: continue;
1.261 schwarze 659: case ESCAPE_HORIZ:
1.286 schwarze 660: if (p->flags & TERMP_BACKAFTER) {
661: p->flags &= ~TERMP_BACKAFTER;
662: continue;
663: }
1.273 schwarze 664: if (*seq == '|') {
665: seq++;
666: uc = -p->col;
667: } else
668: uc = 0;
1.268 schwarze 669: if (a2roffsu(seq, &su, SCALE_EM) == NULL)
1.261 schwarze 670: continue;
1.273 schwarze 671: uc += term_hen(p, &su);
1.286 schwarze 672: if (uc >= 0) {
1.285 schwarze 673: while (uc > 0) {
674: uc -= term_len(p, 1);
1.286 schwarze 675: if (p->flags & TERMP_BACKBEFORE)
676: p->flags &= ~TERMP_BACKBEFORE;
677: else
678: bufferc(p, ASCII_NBRSP);
1.285 schwarze 679: }
1.286 schwarze 680: continue;
681: }
682: if (p->flags & TERMP_BACKBEFORE) {
683: p->flags &= ~TERMP_BACKBEFORE;
684: assert(p->col > 0);
685: p->col--;
686: }
687: if (p->col >= (size_t)(-uc)) {
1.261 schwarze 688: p->col += uc;
1.285 schwarze 689: } else {
1.261 schwarze 690: uc += p->col;
691: p->col = 0;
1.266 schwarze 692: if (p->tcol->offset > (size_t)(-uc)) {
1.261 schwarze 693: p->ti += uc;
1.266 schwarze 694: p->tcol->offset += uc;
1.261 schwarze 695: } else {
1.266 schwarze 696: p->ti -= p->tcol->offset;
697: p->tcol->offset = 0;
1.261 schwarze 698: }
1.262 schwarze 699: }
700: continue;
701: case ESCAPE_HLINE:
1.272 schwarze 702: if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL)
1.262 schwarze 703: continue;
1.271 schwarze 704: uc = term_hen(p, &su);
1.262 schwarze 705: if (uc <= 0) {
1.266 schwarze 706: if (p->tcol->rmargin <= p->tcol->offset)
1.262 schwarze 707: continue;
1.266 schwarze 708: lsz = p->tcol->rmargin - p->tcol->offset;
1.262 schwarze 709: } else
710: lsz = uc;
1.272 schwarze 711: if (*cp == seq[-1])
1.262 schwarze 712: uc = -1;
1.272 schwarze 713: else if (*cp == '\\') {
714: seq = cp + 1;
1.262 schwarze 715: esc = mandoc_escape(&seq, &cp, &sz);
716: switch (esc) {
717: case ESCAPE_UNICODE:
718: uc = mchars_num2uc(cp + 1, sz - 1);
719: break;
720: case ESCAPE_NUMBERED:
721: uc = mchars_num2char(cp, sz);
722: break;
723: case ESCAPE_SPECIAL:
724: uc = mchars_spec2cp(cp, sz);
725: break;
1.277 schwarze 726: case ESCAPE_UNDEF:
727: uc = *seq;
728: break;
1.262 schwarze 729: default:
730: uc = -1;
731: break;
732: }
733: } else
1.272 schwarze 734: uc = *cp;
1.262 schwarze 735: if (uc < 0x20 || (uc > 0x7E && uc < 0xA0))
736: uc = '_';
737: if (p->enc == TERMENC_ASCII) {
738: cp = ascii_uc2str(uc);
739: csz = term_strlen(p, cp);
740: ssz = strlen(cp);
741: } else
742: csz = (*p->width)(p, uc);
743: while (lsz >= csz) {
744: if (p->enc == TERMENC_ASCII)
745: encode(p, cp, ssz);
746: else
747: encode1(p, uc);
748: lsz -= csz;
1.261 schwarze 749: }
1.233 schwarze 750: continue;
1.222 schwarze 751: case ESCAPE_SKIPCHAR:
1.248 schwarze 752: p->flags |= TERMP_BACKAFTER;
1.233 schwarze 753: continue;
1.243 schwarze 754: case ESCAPE_OVERSTRIKE:
755: cp = seq + sz;
756: while (seq < cp) {
757: if (*seq == '\\') {
758: mandoc_escape(&seq, NULL, NULL);
759: continue;
760: }
761: encode1(p, *seq++);
1.248 schwarze 762: if (seq < cp) {
763: if (p->flags & TERMP_BACKBEFORE)
764: p->flags |= TERMP_BACKAFTER;
765: else
766: p->flags |= TERMP_BACKBEFORE;
767: }
1.243 schwarze 768: }
1.249 schwarze 769: /* Trim trailing backspace/blank pair. */
1.269 schwarze 770: if (p->tcol->lastcol > 2 &&
771: (p->tcol->buf[p->tcol->lastcol - 1] == ' ' ||
772: p->tcol->buf[p->tcol->lastcol - 1] == '\t'))
773: p->tcol->lastcol -= 2;
774: if (p->col > p->tcol->lastcol)
775: p->col = p->tcol->lastcol;
1.248 schwarze 776: continue;
1.289 schwarze 777: case ESCAPE_IGNORE:
778: bufferc(p, ASCII_NBRZW);
779: continue;
1.124 kristaps 780: default:
1.233 schwarze 781: continue;
782: }
783:
784: /*
785: * Common handling for Unicode and numbered
786: * character escape sequences.
787: */
788:
789: if (p->enc == TERMENC_ASCII) {
790: cp = ascii_uc2str(uc);
791: encode(p, cp, strlen(cp));
792: } else {
793: if ((uc < 0x20 && uc != 0x09) ||
794: (uc > 0x7E && uc < 0xA0))
795: uc = 0xFFFD;
796: encode1(p, uc);
1.124 kristaps 797: }
798: }
1.214 schwarze 799: p->flags &= ~TERMP_NBRWORD;
1.65 kristaps 800: }
801:
1.71 kristaps 802: static void
1.266 schwarze 803: adjbuf(struct termp_col *c, size_t sz)
1.51 kristaps 804: {
1.266 schwarze 805: if (c->maxcols == 0)
806: c->maxcols = 1024;
807: while (c->maxcols <= sz)
808: c->maxcols <<= 2;
809: c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf));
1.51 kristaps 810: }
811:
1.79 kristaps 812: static void
1.125 kristaps 813: bufferc(struct termp *p, char c)
814: {
1.264 schwarze 815: if (p->flags & TERMP_NOBUF) {
816: (*p->letter)(p, c);
817: return;
818: }
1.266 schwarze 819: if (p->col + 1 >= p->tcol->maxcols)
820: adjbuf(p->tcol, p->col + 1);
1.269 schwarze 821: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.266 schwarze 822: p->tcol->buf[p->col] = c;
1.269 schwarze 823: if (p->tcol->lastcol < ++p->col)
824: p->tcol->lastcol = p->col;
1.125 kristaps 825: }
826:
1.290 ! schwarze 827: void
! 828: term_tab_ref(struct termp *p)
! 829: {
! 830: if (p->tcol->lastcol && p->tcol->lastcol <= p->col &&
! 831: (p->flags & TERMP_NOBUF) == 0)
! 832: bufferc(p, ASCII_TABREF);
! 833: }
! 834:
1.194 kristaps 835: /*
836: * See encode().
837: * Do this for a single (probably unicode) value.
838: * Does not check for non-decorated glyphs.
839: */
840: static void
841: encode1(struct termp *p, int c)
842: {
843: enum termfont f;
844:
1.264 schwarze 845: if (p->flags & TERMP_NOBUF) {
846: (*p->letter)(p, c);
847: return;
848: }
849:
1.266 schwarze 850: if (p->col + 7 >= p->tcol->maxcols)
851: adjbuf(p->tcol, p->col + 7);
1.194 kristaps 852:
1.255 schwarze 853: f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ?
1.248 schwarze 854: p->fontq[p->fonti] : TERMFONT_NONE;
1.194 kristaps 855:
1.248 schwarze 856: if (p->flags & TERMP_BACKBEFORE) {
1.266 schwarze 857: if (p->tcol->buf[p->col - 1] == ' ' ||
858: p->tcol->buf[p->col - 1] == '\t')
1.249 schwarze 859: p->col--;
860: else
1.266 schwarze 861: p->tcol->buf[p->col++] = '\b';
1.248 schwarze 862: p->flags &= ~TERMP_BACKBEFORE;
863: }
1.266 schwarze 864: if (f == TERMFONT_UNDER || f == TERMFONT_BI) {
865: p->tcol->buf[p->col++] = '_';
866: p->tcol->buf[p->col++] = '\b';
867: }
868: if (f == TERMFONT_BOLD || f == TERMFONT_BI) {
869: if (c == ASCII_HYPH)
870: p->tcol->buf[p->col++] = '-';
1.209 schwarze 871: else
1.266 schwarze 872: p->tcol->buf[p->col++] = c;
873: p->tcol->buf[p->col++] = '\b';
1.209 schwarze 874: }
1.269 schwarze 875: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.266 schwarze 876: p->tcol->buf[p->col] = c;
1.269 schwarze 877: if (p->tcol->lastcol < ++p->col)
878: p->tcol->lastcol = p->col;
1.248 schwarze 879: if (p->flags & TERMP_BACKAFTER) {
880: p->flags |= TERMP_BACKBEFORE;
881: p->flags &= ~TERMP_BACKAFTER;
882: }
1.194 kristaps 883: }
884:
1.125 kristaps 885: static void
886: encode(struct termp *p, const char *word, size_t sz)
887: {
1.210 schwarze 888: size_t i;
1.264 schwarze 889:
890: if (p->flags & TERMP_NOBUF) {
891: for (i = 0; i < sz; i++)
892: (*p->letter)(p, word[i]);
893: return;
894: }
1.188 kristaps 895:
1.266 schwarze 896: if (p->col + 2 + (sz * 5) >= p->tcol->maxcols)
897: adjbuf(p->tcol, p->col + 2 + (sz * 5));
1.165 kristaps 898:
1.210 schwarze 899: for (i = 0; i < sz; i++) {
1.209 schwarze 900: if (ASCII_HYPH == word[i] ||
901: isgraph((unsigned char)word[i]))
902: encode1(p, word[i]);
1.259 schwarze 903: else {
1.269 schwarze 904: if (p->tcol->lastcol <= p->col ||
1.265 schwarze 905: (word[i] != ' ' && word[i] != ASCII_NBRSP))
1.266 schwarze 906: p->tcol->buf[p->col] = word[i];
1.265 schwarze 907: p->col++;
1.259 schwarze 908:
909: /*
910: * Postpone the effect of \z while handling
911: * an overstrike sequence from ascii_uc2str().
912: */
913:
914: if (word[i] == '\b' &&
915: (p->flags & TERMP_BACKBEFORE)) {
916: p->flags &= ~TERMP_BACKBEFORE;
917: p->flags |= TERMP_BACKAFTER;
918: }
919: }
1.79 kristaps 920: }
1.269 schwarze 921: if (p->tcol->lastcol < p->col)
922: p->tcol->lastcol = p->col;
1.219 schwarze 923: }
924:
925: void
926: term_setwidth(struct termp *p, const char *wstr)
927: {
928: struct roffsu su;
1.247 schwarze 929: int iop, width;
1.219 schwarze 930:
1.220 schwarze 931: iop = 0;
932: width = 0;
1.219 schwarze 933: if (NULL != wstr) {
934: switch (*wstr) {
1.222 schwarze 935: case '+':
1.219 schwarze 936: iop = 1;
937: wstr++;
938: break;
1.222 schwarze 939: case '-':
1.219 schwarze 940: iop = -1;
941: wstr++;
942: break;
943: default:
944: break;
945: }
1.268 schwarze 946: if (a2roffsu(wstr, &su, SCALE_MAX) != NULL)
1.220 schwarze 947: width = term_hspan(p, &su);
948: else
1.219 schwarze 949: iop = 0;
950: }
951: (*p->setwidth)(p, iop, width);
1.79 kristaps 952: }
1.106 kristaps 953:
1.107 kristaps 954: size_t
1.149 kristaps 955: term_len(const struct termp *p, size_t sz)
956: {
957:
1.252 schwarze 958: return (*p->width)(p, ' ') * sz;
1.149 kristaps 959: }
960:
1.203 schwarze 961: static size_t
962: cond_width(const struct termp *p, int c, int *skip)
963: {
964:
965: if (*skip) {
966: (*skip) = 0;
1.252 schwarze 967: return 0;
1.203 schwarze 968: } else
1.252 schwarze 969: return (*p->width)(p, c);
1.203 schwarze 970: }
1.149 kristaps 971:
972: size_t
973: term_strlen(const struct termp *p, const char *cp)
974: {
1.184 kristaps 975: size_t sz, rsz, i;
1.233 schwarze 976: int ssz, skip, uc;
1.171 kristaps 977: const char *seq, *rhs;
1.196 kristaps 978: enum mandoc_esc esc;
1.289 schwarze 979: static const char rej[] = { '\\', ASCII_NBRSP, ASCII_NBRZW,
1.290 ! schwarze 980: ASCII_BREAK, ASCII_HYPH, ASCII_TABREF, '\0' };
1.171 kristaps 981:
1.184 kristaps 982: /*
983: * Account for escaped sequences within string length
984: * calculations. This follows the logic in term_word() as we
985: * must calculate the width of produced strings.
986: */
987:
988: sz = 0;
1.203 schwarze 989: skip = 0;
1.189 kristaps 990: while ('\0' != *cp) {
991: rsz = strcspn(cp, rej);
992: for (i = 0; i < rsz; i++)
1.203 schwarze 993: sz += cond_width(p, *cp++, &skip);
1.189 kristaps 994:
1.184 kristaps 995: switch (*cp) {
1.222 schwarze 996: case '\\':
1.189 kristaps 997: cp++;
1.277 schwarze 998: rhs = NULL;
1.196 kristaps 999: esc = mandoc_escape(&cp, &seq, &ssz);
1000: switch (esc) {
1.222 schwarze 1001: case ESCAPE_UNICODE:
1.234 schwarze 1002: uc = mchars_num2uc(seq + 1, ssz - 1);
1.194 kristaps 1003: break;
1.222 schwarze 1004: case ESCAPE_NUMBERED:
1.233 schwarze 1005: uc = mchars_num2char(seq, ssz);
1006: if (uc < 0)
1007: continue;
1.171 kristaps 1008: break;
1.222 schwarze 1009: case ESCAPE_SPECIAL:
1.233 schwarze 1010: if (p->enc == TERMENC_ASCII) {
1.254 schwarze 1011: rhs = mchars_spec2str(seq, ssz, &rsz);
1.233 schwarze 1012: if (rhs != NULL)
1013: break;
1014: } else {
1.254 schwarze 1015: uc = mchars_spec2cp(seq, ssz);
1.233 schwarze 1016: if (uc > 0)
1017: sz += cond_width(p, uc, &skip);
1.229 schwarze 1018: }
1.233 schwarze 1019: continue;
1.277 schwarze 1020: case ESCAPE_UNDEF:
1021: uc = *seq;
1022: break;
1.275 schwarze 1023: case ESCAPE_DEVICE:
1024: if (p->type == TERMTYPE_PDF) {
1025: rhs = "pdf";
1026: rsz = 3;
1027: } else if (p->type == TERMTYPE_PS) {
1028: rhs = "ps";
1029: rsz = 2;
1030: } else if (p->enc == TERMENC_ASCII) {
1031: rhs = "ascii";
1032: rsz = 5;
1033: } else {
1034: rhs = "utf8";
1035: rsz = 4;
1036: }
1037: break;
1.222 schwarze 1038: case ESCAPE_SKIPCHAR:
1.203 schwarze 1039: skip = 1;
1.243 schwarze 1040: continue;
1041: case ESCAPE_OVERSTRIKE:
1042: rsz = 0;
1043: rhs = seq + ssz;
1044: while (seq < rhs) {
1045: if (*seq == '\\') {
1046: mandoc_escape(&seq, NULL, NULL);
1047: continue;
1048: }
1049: i = (*p->width)(p, *seq++);
1050: if (rsz < i)
1051: rsz = i;
1052: }
1053: sz += rsz;
1.233 schwarze 1054: continue;
1.171 kristaps 1055: default:
1.233 schwarze 1056: continue;
1.171 kristaps 1057: }
1.149 kristaps 1058:
1.233 schwarze 1059: /*
1060: * Common handling for Unicode and numbered
1061: * character escape sequences.
1062: */
1063:
1064: if (rhs == NULL) {
1065: if (p->enc == TERMENC_ASCII) {
1066: rhs = ascii_uc2str(uc);
1067: rsz = strlen(rhs);
1068: } else {
1069: if ((uc < 0x20 && uc != 0x09) ||
1070: (uc > 0x7E && uc < 0xA0))
1071: uc = 0xFFFD;
1072: sz += cond_width(p, uc, &skip);
1073: continue;
1074: }
1075: }
1.184 kristaps 1076:
1.203 schwarze 1077: if (skip) {
1078: skip = 0;
1079: break;
1080: }
1.233 schwarze 1081:
1082: /*
1083: * Common handling for all escape sequences
1084: * printing more than one character.
1085: */
1.203 schwarze 1086:
1.184 kristaps 1087: for (i = 0; i < rsz; i++)
1088: sz += (*p->width)(p, *rhs++);
1089: break;
1.222 schwarze 1090: case ASCII_NBRSP:
1.203 schwarze 1091: sz += cond_width(p, ' ', &skip);
1.176 kristaps 1092: cp++;
1.184 kristaps 1093: break;
1.222 schwarze 1094: case ASCII_HYPH:
1.203 schwarze 1095: sz += cond_width(p, '-', &skip);
1.176 kristaps 1096: cp++;
1.184 kristaps 1097: break;
1098: default:
1099: break;
1100: }
1.189 kristaps 1101: }
1.149 kristaps 1102:
1.252 schwarze 1103: return sz;
1.149 kristaps 1104: }
1105:
1.240 schwarze 1106: int
1.149 kristaps 1107: term_vspan(const struct termp *p, const struct roffsu *su)
1.106 kristaps 1108: {
1109: double r;
1.241 schwarze 1110: int ri;
1.106 kristaps 1111:
1.107 kristaps 1112: switch (su->unit) {
1.239 schwarze 1113: case SCALE_BU:
1114: r = su->scale / 40.0;
1115: break;
1.222 schwarze 1116: case SCALE_CM:
1.239 schwarze 1117: r = su->scale * 6.0 / 2.54;
1118: break;
1119: case SCALE_FS:
1120: r = su->scale * 65536.0 / 40.0;
1.106 kristaps 1121: break;
1.222 schwarze 1122: case SCALE_IN:
1.225 schwarze 1123: r = su->scale * 6.0;
1.106 kristaps 1124: break;
1.239 schwarze 1125: case SCALE_MM:
1126: r = su->scale * 0.006;
1127: break;
1.222 schwarze 1128: case SCALE_PC:
1.107 kristaps 1129: r = su->scale;
1.106 kristaps 1130: break;
1.222 schwarze 1131: case SCALE_PT:
1.239 schwarze 1132: r = su->scale / 12.0;
1.106 kristaps 1133: break;
1.239 schwarze 1134: case SCALE_EN:
1135: case SCALE_EM:
1136: r = su->scale * 0.6;
1.106 kristaps 1137: break;
1.222 schwarze 1138: case SCALE_VS:
1.107 kristaps 1139: r = su->scale;
1.106 kristaps 1140: break;
1141: default:
1.239 schwarze 1142: abort();
1.106 kristaps 1143: }
1.241 schwarze 1144: ri = r > 0.0 ? r + 0.4995 : r - 0.4995;
1.252 schwarze 1145: return ri < 66 ? ri : 1;
1.106 kristaps 1146: }
1147:
1.247 schwarze 1148: /*
1.271 schwarze 1149: * Convert a scaling width to basic units, rounding towards 0.
1.247 schwarze 1150: */
1.240 schwarze 1151: int
1.149 kristaps 1152: term_hspan(const struct termp *p, const struct roffsu *su)
1.106 kristaps 1153: {
1.108 kristaps 1154:
1.252 schwarze 1155: return (*p->hspan)(p, su);
1.271 schwarze 1156: }
1157:
1158: /*
1159: * Convert a scaling width to basic units, rounding to closest.
1160: */
1161: int
1162: term_hen(const struct termp *p, const struct roffsu *su)
1163: {
1164: int bu;
1165:
1166: if ((bu = (*p->hspan)(p, su)) >= 0)
1167: return (bu + 11) / 24;
1168: else
1169: return -((-bu + 11) / 24);
1.106 kristaps 1170: }
CVSweb