Annotation of src/usr.bin/mandoc/term.c, Revision 1.151
1.151 ! jmc 1: /* $OpenBSD: term.c,v 1.150 2022/08/16 17:44:53 schwarze Exp $ */
1.1 kristaps 2: /*
1.145 schwarze 3: * Copyright (c) 2010-2022 Ingo Schwarze <schwarze@openbsd.org>
1.59 schwarze 4: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.2 schwarze 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.106 schwarze 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.2 schwarze 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106 schwarze 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.2 schwarze 13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 17: */
1.20 schwarze 18: #include <sys/types.h>
19:
1.1 kristaps 20: #include <assert.h>
1.20 schwarze 21: #include <ctype.h>
1.138 schwarze 22: #include <stdint.h>
1.1 kristaps 23: #include <stdio.h>
24: #include <stdlib.h>
25: #include <string.h>
26:
1.34 schwarze 27: #include "mandoc.h"
1.79 schwarze 28: #include "mandoc_aux.h"
1.16 schwarze 29: #include "out.h"
1.1 kristaps 30: #include "term.h"
1.16 schwarze 31: #include "main.h"
1.1 kristaps 32:
1.64 schwarze 33: static size_t cond_width(const struct termp *, int, int *);
1.126 schwarze 34: static void adjbuf(struct termp_col *, size_t);
1.59 schwarze 35: static void bufferc(struct termp *, char);
36: static void encode(struct termp *, const char *, size_t);
37: static void encode1(struct termp *, int);
1.124 schwarze 38: static void endline(struct termp *);
1.142 schwarze 39: static void term_field(struct termp *, size_t, size_t);
1.138 schwarze 40: static void term_fill(struct termp *, size_t *, size_t *,
41: size_t);
1.1 kristaps 42:
1.83 schwarze 43:
1.37 schwarze 44: void
1.129 schwarze 45: term_setcol(struct termp *p, size_t maxtcol)
46: {
47: if (maxtcol > p->maxtcol) {
48: p->tcols = mandoc_recallocarray(p->tcols,
49: p->maxtcol, maxtcol, sizeof(*p->tcols));
50: p->maxtcol = maxtcol;
51: }
52: p->lasttcol = maxtcol - 1;
53: p->tcol = p->tcols;
54: }
55:
56: void
1.37 schwarze 57: term_free(struct termp *p)
1.1 kristaps 58: {
1.144 schwarze 59: term_tab_free();
1.126 schwarze 60: for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++)
61: free(p->tcol->buf);
62: free(p->tcols);
1.98 schwarze 63: free(p->fontq);
1.37 schwarze 64: free(p);
1.1 kristaps 65: }
66:
1.13 schwarze 67: void
1.83 schwarze 68: term_begin(struct termp *p, term_margin head,
1.106 schwarze 69: term_margin foot, const struct roff_meta *arg)
1.1 kristaps 70: {
71:
1.37 schwarze 72: p->headf = head;
73: p->footf = foot;
74: p->argf = arg;
75: (*p->begin)(p);
1.1 kristaps 76: }
77:
1.37 schwarze 78: void
79: term_end(struct termp *p)
1.1 kristaps 80: {
81:
1.37 schwarze 82: (*p->end)(p);
1.1 kristaps 83: }
84:
85: /*
1.82 schwarze 86: * Flush a chunk of text. By default, break the output line each time
87: * the right margin is reached, and continue output on the next line
88: * at the same offset as the chunk itself. By default, also break the
1.138 schwarze 89: * output line at the end of the chunk. There are many flags modifying
90: * this behaviour, see the comments in the body of the function.
1.1 kristaps 91: */
92: void
93: term_flushln(struct termp *p)
94: {
1.138 schwarze 95: size_t vbl; /* Number of blanks to prepend to the output. */
96: size_t vbr; /* Actual visual position of the end of field. */
97: size_t vfield; /* Desired visual field width. */
98: size_t vtarget; /* Desired visual position of the right margin. */
99: size_t ic; /* Character position in the input buffer. */
100: size_t nbr; /* Number of characters to print in this field. */
101:
102: /*
103: * Normally, start writing at the left margin, but with the
104: * NOPAD flag, start writing at the current position instead.
105: */
1.1 kristaps 106:
1.126 schwarze 107: vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ?
108: 0 : p->tcol->offset - p->viscol;
1.123 schwarze 109: if (p->minbl && vbl < p->minbl)
110: vbl = p->minbl;
1.19 schwarze 111:
1.134 florian 112: if ((p->flags & TERMP_MULTICOL) == 0)
1.127 schwarze 113: p->tcol->col = 0;
1.138 schwarze 114:
115: /* Loop over output lines. */
116:
117: for (;;) {
118: vfield = p->tcol->rmargin > p->viscol + vbl ?
119: p->tcol->rmargin - p->viscol - vbl : 0;
1.127 schwarze 120:
1.22 schwarze 121: /*
1.138 schwarze 122: * Normally, break the line at the the right margin
123: * of the field, but with the NOBREAK flag, only
124: * break it at the max right margin of the screen,
125: * and with the BRNEVER flag, never break it at all.
1.30 schwarze 126: */
1.127 schwarze 127:
1.142 schwarze 128: vtarget = (p->flags & TERMP_NOBREAK) == 0 ? vfield :
1.138 schwarze 129: p->maxrmargin > p->viscol + vbl ?
130: p->maxrmargin - p->viscol - vbl : 0;
131:
132: /*
133: * Figure out how much text will fit in the field.
134: * If there is whitespace only, print nothing.
135: */
136:
1.142 schwarze 137: term_fill(p, &nbr, &vbr,
138: p->flags & TERMP_BRNEVER ? SIZE_MAX : vtarget);
1.138 schwarze 139: if (nbr == 0)
140: break;
1.139 schwarze 141:
142: /*
143: * With the CENTER or RIGHT flag, increase the indentation
144: * to center the text between the left and right margins
145: * or to adjust it to the right margin, respectively.
146: */
147:
148: if (vbr < vtarget) {
149: if (p->flags & TERMP_CENTER)
150: vbl += (vtarget - vbr) / 2;
151: else if (p->flags & TERMP_RIGHT)
152: vbl += vtarget - vbr;
153: }
154:
155: /* Finally, print the field content. */
1.138 schwarze 156:
1.142 schwarze 157: term_field(p, vbl, nbr);
1.150 schwarze 158: if (vbr < vtarget)
159: p->tcol->taboff += vbr;
160: else
161: p->tcol->taboff += vtarget;
162: p->tcol->taboff += (*p->width)(p, ' ');
1.22 schwarze 163:
1.1 kristaps 164: /*
1.138 schwarze 165: * If there is no text left in the field, exit the loop.
166: * If the BRTRSP flag is set, consider trailing
167: * whitespace significant when deciding whether
168: * the field fits or not.
1.1 kristaps 169: */
170:
1.138 schwarze 171: for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
172: switch (p->tcol->buf[ic]) {
173: case '\t':
174: if (p->flags & TERMP_BRTRSP)
175: vbr = term_tab_next(vbr);
176: continue;
177: case ' ':
178: if (p->flags & TERMP_BRTRSP)
179: vbr += (*p->width)(p, ' ');
1.130 schwarze 180: continue;
1.138 schwarze 181: case '\n':
1.150 schwarze 182: case ASCII_NBRZW:
1.138 schwarze 183: case ASCII_BREAK:
1.150 schwarze 184: case ASCII_TABREF:
1.138 schwarze 185: continue;
186: default:
187: break;
1.130 schwarze 188: }
1.138 schwarze 189: break;
190: }
191: if (ic == p->tcol->lastcol)
192: break;
1.42 schwarze 193:
1.138 schwarze 194: /*
1.151 ! jmc 195: * At the location of an automatic line break, input
1.138 schwarze 196: * space characters are consumed by the line break.
197: */
1.42 schwarze 198:
1.138 schwarze 199: while (p->tcol->col < p->tcol->lastcol &&
200: p->tcol->buf[p->tcol->col] == ' ')
201: p->tcol->col++;
1.42 schwarze 202:
1.138 schwarze 203: /*
204: * In multi-column mode, leave the rest of the text
205: * in the buffer to be handled by a subsequent
206: * invocation, such that the other columns of the
207: * table can be handled first.
208: * In single-column mode, simply break the line.
209: */
210:
211: if (p->flags & TERMP_MULTICOL)
212: return;
1.78 schwarze 213:
1.138 schwarze 214: endline(p);
1.1 kristaps 215:
216: /*
1.138 schwarze 217: * Normally, start the next line at the same indentation
218: * as this one, but with the BRIND flag, start it at the
219: * right margin instead. This is used together with
220: * NOBREAK for the tags in various kinds of tagged lists.
1.5 schwarze 221: */
1.127 schwarze 222:
1.138 schwarze 223: vbl = p->flags & TERMP_BRIND ?
224: p->tcol->rmargin : p->tcol->offset;
225: }
1.127 schwarze 226:
1.138 schwarze 227: /* Reset output state in preparation for the next field. */
1.66 schwarze 228:
1.138 schwarze 229: p->col = p->tcol->col = p->tcol->lastcol = 0;
230: p->minbl = p->trailspace;
231: p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE | TERMP_NOPAD);
1.120 schwarze 232:
1.138 schwarze 233: if (p->flags & TERMP_MULTICOL)
234: return;
1.120 schwarze 235:
1.138 schwarze 236: /*
237: * The HANG flag means that the next field
238: * always follows on the same line.
239: * The NOBREAK flag means that the next field
240: * follows on the same line unless the field was overrun.
241: * Normally, break the line at the end of each field.
242: */
1.66 schwarze 243:
1.138 schwarze 244: if ((p->flags & TERMP_HANG) == 0 &&
245: ((p->flags & TERMP_NOBREAK) == 0 ||
246: vbr + term_len(p, p->trailspace) > vfield))
247: endline(p);
248: }
1.30 schwarze 249:
1.138 schwarze 250: /*
251: * Store the number of input characters to print in this field in *nbr
252: * and their total visual width to print in *vbr.
253: * If there is only whitespace in the field, both remain zero.
254: * The desired visual width of the field is provided by vtarget.
255: * If the first word is longer, the field will be overrun.
256: */
257: static void
258: term_fill(struct termp *p, size_t *nbr, size_t *vbr, size_t vtarget)
259: {
260: size_t ic; /* Character position in the input buffer. */
261: size_t vis; /* Visual position of the current character. */
262: size_t vn; /* Visual position of the next character. */
263: int breakline; /* Break at the end of this word. */
264: int graph; /* Last character was non-blank. */
1.150 schwarze 265: int taboff; /* Temporary offset for literal tabs. */
1.138 schwarze 266:
267: *nbr = *vbr = vis = 0;
268: breakline = graph = 0;
1.150 schwarze 269: taboff = p->tcol->taboff;
1.138 schwarze 270: for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
271: switch (p->tcol->buf[ic]) {
272: case '\b': /* Escape \o (overstrike) or backspace markup. */
273: assert(ic > 0);
274: vis -= (*p->width)(p, p->tcol->buf[ic - 1]);
275: continue;
1.127 schwarze 276:
1.138 schwarze 277: case ' ':
278: case ASCII_BREAK: /* Escape \: (breakpoint). */
1.147 schwarze 279: vn = vis;
280: if (p->tcol->buf[ic] == ' ')
281: vn += (*p->width)(p, ' ');
1.138 schwarze 282: /* Can break at the end of a word. */
283: if (breakline || vn > vtarget)
284: break;
285: if (graph) {
286: *nbr = ic;
287: *vbr = vis;
288: graph = 0;
1.33 schwarze 289: }
1.138 schwarze 290: vis = vn;
291: continue;
292:
293: case '\n': /* Escape \p (break at the end of the word). */
294: breakline = 1;
295: continue;
1.33 schwarze 296:
1.138 schwarze 297: case ASCII_HYPH: /* Breakable hyphen. */
298: graph = 1;
1.33 schwarze 299: /*
1.138 schwarze 300: * We are about to decide whether to break the
301: * line or not, so we no longer need this hyphen
302: * to be marked as breakable. Put back a real
303: * hyphen such that we get the correct width.
1.33 schwarze 304: */
1.138 schwarze 305: p->tcol->buf[ic] = '-';
306: vis += (*p->width)(p, '-');
307: if (vis > vtarget) {
308: ic++;
309: break;
1.61 schwarze 310: }
1.138 schwarze 311: *nbr = ic + 1;
312: *vbr = vis;
313: continue;
1.61 schwarze 314:
1.150 schwarze 315: case ASCII_TABREF:
316: taboff = -vis - (*p->width)(p, ' ');
317: continue;
318:
1.147 schwarze 319: default:
320: switch (p->tcol->buf[ic]) {
321: case '\t':
1.150 schwarze 322: if (taboff < 0 && (size_t)-taboff > vis)
323: vis = 0;
324: else
325: vis += taboff;
1.147 schwarze 326: vis = term_tab_next(vis);
1.150 schwarze 327: vis -= taboff;
1.147 schwarze 328: break;
1.149 schwarze 329: case ASCII_NBRZW: /* Non-breakable zero-width. */
330: break;
1.147 schwarze 331: case ASCII_NBRSP: /* Non-breakable space. */
332: p->tcol->buf[ic] = ' ';
333: /* FALLTHROUGH */
334: default: /* Printable character. */
335: vis += (*p->width)(p, p->tcol->buf[ic]);
336: break;
337: }
1.138 schwarze 338: graph = 1;
339: if (vis > vtarget && *nbr > 0)
340: return;
341: continue;
1.1 kristaps 342: }
1.138 schwarze 343: break;
344: }
1.130 schwarze 345:
1.138 schwarze 346: /*
347: * If the last word extends to the end of the field without any
348: * trailing whitespace, the loop could not check yet whether it
349: * can remain on this line. So do the check now.
350: */
1.130 schwarze 351:
1.138 schwarze 352: if (graph && (vis <= vtarget || *nbr == 0)) {
353: *nbr = ic;
354: *vbr = vis;
355: }
356: }
1.130 schwarze 357:
1.138 schwarze 358: /*
359: * Print the contents of one field
360: * with an indentation of vbl visual columns,
1.142 schwarze 361: * and an input string length of nbr characters.
1.138 schwarze 362: */
363: static void
1.142 schwarze 364: term_field(struct termp *p, size_t vbl, size_t nbr)
1.138 schwarze 365: {
366: size_t ic; /* Character position in the input buffer. */
367: size_t vis; /* Visual position of the current character. */
1.148 schwarze 368: size_t vt; /* Visual position including tab offset. */
1.138 schwarze 369: size_t dv; /* Visual width of the current character. */
1.150 schwarze 370: int taboff; /* Temporary offset for literal tabs. */
1.130 schwarze 371:
1.138 schwarze 372: vis = 0;
1.150 schwarze 373: taboff = p->tcol->taboff;
1.138 schwarze 374: for (ic = p->tcol->col; ic < nbr; ic++) {
1.130 schwarze 375:
1.138 schwarze 376: /*
377: * To avoid the printing of trailing whitespace,
378: * do not print whitespace right away, only count it.
379: */
1.130 schwarze 380:
1.138 schwarze 381: switch (p->tcol->buf[ic]) {
382: case '\n':
383: case ASCII_BREAK:
1.149 schwarze 384: case ASCII_NBRZW:
1.138 schwarze 385: continue;
1.150 schwarze 386: case ASCII_TABREF:
387: taboff = -vis - (*p->width)(p, ' ');
388: continue;
1.138 schwarze 389: case '\t':
390: case ' ':
391: case ASCII_NBRSP:
1.148 schwarze 392: if (p->tcol->buf[ic] == '\t') {
1.150 schwarze 393: if (taboff < 0 && (size_t)-taboff > vis)
394: vt = 0;
395: else
396: vt = vis + taboff;
1.148 schwarze 397: dv = term_tab_next(vt) - vt;
398: } else
399: dv = (*p->width)(p, ' ');
1.140 schwarze 400: vbl += dv;
401: vis += dv;
1.138 schwarze 402: continue;
403: default:
404: break;
405: }
1.48 schwarze 406:
1.138 schwarze 407: /*
408: * We found a non-blank character to print,
409: * so write preceding white space now.
410: */
1.127 schwarze 411:
1.138 schwarze 412: if (vbl > 0) {
413: (*p->advance)(p, vbl);
414: p->viscol += vbl;
415: vbl = 0;
416: }
1.18 schwarze 417:
1.138 schwarze 418: /* Print the character and adjust the visual position. */
1.1 kristaps 419:
1.138 schwarze 420: (*p->letter)(p, p->tcol->buf[ic]);
421: if (p->tcol->buf[ic] == '\b') {
422: dv = (*p->width)(p, p->tcol->buf[ic - 1]);
423: p->viscol -= dv;
424: vis -= dv;
425: } else {
426: dv = (*p->width)(p, p->tcol->buf[ic]);
427: p->viscol += dv;
428: vis += dv;
429: }
430: }
431: p->tcol->col = nbr;
1.124 schwarze 432: }
433:
434: static void
435: endline(struct termp *p)
436: {
437: if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) {
438: p->mc = NULL;
439: p->flags &= ~TERMP_ENDMC;
440: }
441: if (p->mc != NULL) {
442: if (p->viscol && p->maxrmargin >= p->viscol)
443: (*p->advance)(p, p->maxrmargin - p->viscol + 1);
444: p->flags |= TERMP_NOBUF | TERMP_NOSPACE;
445: term_word(p, p->mc);
446: p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC);
447: }
448: p->viscol = 0;
449: p->minbl = 0;
450: (*p->endline)(p);
1.1 kristaps 451: }
452:
1.83 schwarze 453: /*
1.1 kristaps 454: * A newline only breaks an existing line; it won't assert vertical
455: * space. All data in the output buffer is flushed prior to the newline
456: * assertion.
457: */
458: void
459: term_newln(struct termp *p)
460: {
461: p->flags |= TERMP_NOSPACE;
1.129 schwarze 462: if (p->tcol->lastcol || p->viscol)
1.61 schwarze 463: term_flushln(p);
1.150 schwarze 464: p->tcol->taboff = 0;
1.1 kristaps 465: }
466:
467: /*
468: * Asserts a vertical space (a full, empty line-break between lines).
469: * Note that if used twice, this will cause two blank spaces and so on.
470: * All data in the output buffer is flushed prior to the newline
471: * assertion.
472: */
473: void
474: term_vspace(struct termp *p)
475: {
476:
477: term_newln(p);
1.29 schwarze 478: p->viscol = 0;
1.124 schwarze 479: p->minbl = 0;
1.63 schwarze 480: if (0 < p->skipvsp)
481: p->skipvsp--;
482: else
483: (*p->endline)(p);
1.1 kristaps 484: }
485:
1.98 schwarze 486: /* Swap current and previous font; for \fP and .ft P */
1.20 schwarze 487: void
488: term_fontlast(struct termp *p)
489: {
490: enum termfont f;
1.11 schwarze 491:
1.20 schwarze 492: f = p->fontl;
493: p->fontl = p->fontq[p->fonti];
494: p->fontq[p->fonti] = f;
495: }
496:
1.98 schwarze 497: /* Set font, save current, discard previous; for \f, .ft, .B etc. */
1.20 schwarze 498: void
499: term_fontrepl(struct termp *p, enum termfont f)
500: {
501:
502: p->fontl = p->fontq[p->fonti];
503: p->fontq[p->fonti] = f;
1.1 kristaps 504: }
505:
1.98 schwarze 506: /* Set font, save previous. */
1.20 schwarze 507: void
508: term_fontpush(struct termp *p, enum termfont f)
1.1 kristaps 509: {
1.7 schwarze 510:
1.20 schwarze 511: p->fontl = p->fontq[p->fonti];
1.98 schwarze 512: if (++p->fonti == p->fontsz) {
513: p->fontsz += 8;
514: p->fontq = mandoc_reallocarray(p->fontq,
1.116 schwarze 515: p->fontsz, sizeof(*p->fontq));
1.98 schwarze 516: }
517: p->fontq[p->fonti] = f;
1.20 schwarze 518: }
1.1 kristaps 519:
1.98 schwarze 520: /* Flush to make the saved pointer current again. */
1.20 schwarze 521: void
1.104 schwarze 522: term_fontpopq(struct termp *p, int i)
1.20 schwarze 523: {
1.1 kristaps 524:
1.104 schwarze 525: assert(i >= 0);
526: if (p->fonti > i)
527: p->fonti = i;
1.20 schwarze 528: }
1.1 kristaps 529:
1.98 schwarze 530: /* Pop one font off the stack. */
1.20 schwarze 531: void
532: term_fontpop(struct termp *p)
533: {
1.1 kristaps 534:
1.20 schwarze 535: assert(p->fonti);
536: p->fonti--;
1.1 kristaps 537: }
538:
539: /*
540: * Handle pwords, partial words, which may be either a single word or a
541: * phrase that cannot be broken down (such as a literal string). This
542: * handles word styling.
543: */
1.7 schwarze 544: void
545: term_word(struct termp *p, const char *word)
1.1 kristaps 546: {
1.121 schwarze 547: struct roffsu su;
1.75 schwarze 548: const char nbrsp[2] = { ASCII_NBRSP, 0 };
1.59 schwarze 549: const char *seq, *cp;
550: int sz, uc;
1.122 schwarze 551: size_t csz, lsz, ssz;
1.59 schwarze 552: enum mandoc_esc esc;
1.1 kristaps 553:
1.124 schwarze 554: if ((p->flags & TERMP_NOBUF) == 0) {
555: if ((p->flags & TERMP_NOSPACE) == 0) {
556: if ((p->flags & TERMP_KEEP) == 0) {
1.40 schwarze 557: bufferc(p, ' ');
1.124 schwarze 558: if (p->flags & TERMP_SENTENCE)
559: bufferc(p, ' ');
560: } else
561: bufferc(p, ASCII_NBRSP);
562: }
563: if (p->flags & TERMP_PREKEEP)
564: p->flags |= TERMP_KEEP;
565: if (p->flags & TERMP_NONOSPACE)
566: p->flags |= TERMP_NOSPACE;
567: else
568: p->flags &= ~TERMP_NOSPACE;
569: p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE);
570: p->skipvsp = 0;
1.31 schwarze 571: }
572:
1.59 schwarze 573: while ('\0' != *word) {
1.64 schwarze 574: if ('\\' != *word) {
1.75 schwarze 575: if (TERMP_NBRWORD & p->flags) {
576: if (' ' == *word) {
577: encode(p, nbrsp, 1);
578: word++;
579: continue;
580: }
581: ssz = strcspn(word, "\\ ");
582: } else
583: ssz = strcspn(word, "\\");
1.45 schwarze 584: encode(p, word, ssz);
1.64 schwarze 585: word += (int)ssz;
1.20 schwarze 586: continue;
1.64 schwarze 587: }
1.20 schwarze 588:
1.59 schwarze 589: word++;
590: esc = mandoc_escape(&word, &seq, &sz);
591: switch (esc) {
1.83 schwarze 592: case ESCAPE_UNICODE:
1.89 schwarze 593: uc = mchars_num2uc(seq + 1, sz - 1);
1.56 schwarze 594: break;
1.83 schwarze 595: case ESCAPE_NUMBERED:
1.93 schwarze 596: uc = mchars_num2char(seq, sz);
1.149 schwarze 597: if (uc >= 0)
598: break;
599: bufferc(p, ASCII_NBRZW);
600: continue;
1.83 schwarze 601: case ESCAPE_SPECIAL:
1.89 schwarze 602: if (p->enc == TERMENC_ASCII) {
1.114 schwarze 603: cp = mchars_spec2str(seq, sz, &ssz);
1.92 schwarze 604: if (cp != NULL)
1.89 schwarze 605: encode(p, cp, ssz);
1.149 schwarze 606: else
607: bufferc(p, ASCII_NBRZW);
1.89 schwarze 608: } else {
1.114 schwarze 609: uc = mchars_spec2cp(seq, sz);
1.90 schwarze 610: if (uc > 0)
611: encode1(p, uc);
1.149 schwarze 612: else
613: bufferc(p, ASCII_NBRZW);
1.89 schwarze 614: }
1.93 schwarze 615: continue;
1.137 schwarze 616: case ESCAPE_UNDEF:
617: uc = *seq;
618: break;
1.83 schwarze 619: case ESCAPE_FONTBOLD:
1.143 schwarze 620: case ESCAPE_FONTCB:
1.20 schwarze 621: term_fontrepl(p, TERMFONT_BOLD);
1.93 schwarze 622: continue;
1.83 schwarze 623: case ESCAPE_FONTITALIC:
1.143 schwarze 624: case ESCAPE_FONTCI:
1.20 schwarze 625: term_fontrepl(p, TERMFONT_UNDER);
1.93 schwarze 626: continue;
1.83 schwarze 627: case ESCAPE_FONTBI:
1.70 schwarze 628: term_fontrepl(p, TERMFONT_BI);
1.93 schwarze 629: continue;
1.83 schwarze 630: case ESCAPE_FONT:
1.143 schwarze 631: case ESCAPE_FONTCR:
1.83 schwarze 632: case ESCAPE_FONTROMAN:
1.20 schwarze 633: term_fontrepl(p, TERMFONT_NONE);
1.93 schwarze 634: continue;
1.83 schwarze 635: case ESCAPE_FONTPREV:
1.20 schwarze 636: term_fontlast(p);
1.130 schwarze 637: continue;
638: case ESCAPE_BREAK:
639: bufferc(p, '\n');
1.93 schwarze 640: continue;
1.83 schwarze 641: case ESCAPE_NOSPACE:
1.108 schwarze 642: if (p->flags & TERMP_BACKAFTER)
643: p->flags &= ~TERMP_BACKAFTER;
644: else if (*word == '\0')
1.97 schwarze 645: p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE);
1.121 schwarze 646: continue;
1.135 schwarze 647: case ESCAPE_DEVICE:
648: if (p->type == TERMTYPE_PDF)
649: encode(p, "pdf", 3);
650: else if (p->type == TERMTYPE_PS)
651: encode(p, "ps", 2);
652: else if (p->enc == TERMENC_ASCII)
653: encode(p, "ascii", 5);
654: else
655: encode(p, "utf8", 4);
656: continue;
1.121 schwarze 657: case ESCAPE_HORIZ:
1.146 schwarze 658: if (p->flags & TERMP_BACKAFTER) {
659: p->flags &= ~TERMP_BACKAFTER;
660: continue;
661: }
1.133 schwarze 662: if (*seq == '|') {
663: seq++;
664: uc = -p->col;
665: } else
666: uc = 0;
1.128 schwarze 667: if (a2roffsu(seq, &su, SCALE_EM) == NULL)
1.121 schwarze 668: continue;
1.133 schwarze 669: uc += term_hen(p, &su);
1.146 schwarze 670: if (uc >= 0) {
1.145 schwarze 671: while (uc > 0) {
672: uc -= term_len(p, 1);
1.146 schwarze 673: if (p->flags & TERMP_BACKBEFORE)
674: p->flags &= ~TERMP_BACKBEFORE;
675: else
676: bufferc(p, ASCII_NBRSP);
1.145 schwarze 677: }
1.146 schwarze 678: continue;
679: }
680: if (p->flags & TERMP_BACKBEFORE) {
681: p->flags &= ~TERMP_BACKBEFORE;
682: assert(p->col > 0);
683: p->col--;
684: }
685: if (p->col >= (size_t)(-uc)) {
1.121 schwarze 686: p->col += uc;
1.145 schwarze 687: } else {
1.121 schwarze 688: uc += p->col;
689: p->col = 0;
1.126 schwarze 690: if (p->tcol->offset > (size_t)(-uc)) {
1.121 schwarze 691: p->ti += uc;
1.126 schwarze 692: p->tcol->offset += uc;
1.121 schwarze 693: } else {
1.126 schwarze 694: p->ti -= p->tcol->offset;
695: p->tcol->offset = 0;
1.121 schwarze 696: }
1.122 schwarze 697: }
698: continue;
699: case ESCAPE_HLINE:
1.132 schwarze 700: if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL)
1.122 schwarze 701: continue;
1.131 schwarze 702: uc = term_hen(p, &su);
1.122 schwarze 703: if (uc <= 0) {
1.126 schwarze 704: if (p->tcol->rmargin <= p->tcol->offset)
1.122 schwarze 705: continue;
1.126 schwarze 706: lsz = p->tcol->rmargin - p->tcol->offset;
1.122 schwarze 707: } else
708: lsz = uc;
1.132 schwarze 709: if (*cp == seq[-1])
1.122 schwarze 710: uc = -1;
1.132 schwarze 711: else if (*cp == '\\') {
712: seq = cp + 1;
1.122 schwarze 713: esc = mandoc_escape(&seq, &cp, &sz);
714: switch (esc) {
715: case ESCAPE_UNICODE:
716: uc = mchars_num2uc(cp + 1, sz - 1);
717: break;
718: case ESCAPE_NUMBERED:
719: uc = mchars_num2char(cp, sz);
720: break;
721: case ESCAPE_SPECIAL:
722: uc = mchars_spec2cp(cp, sz);
723: break;
1.137 schwarze 724: case ESCAPE_UNDEF:
725: uc = *seq;
726: break;
1.122 schwarze 727: default:
728: uc = -1;
729: break;
730: }
731: } else
1.132 schwarze 732: uc = *cp;
1.122 schwarze 733: if (uc < 0x20 || (uc > 0x7E && uc < 0xA0))
734: uc = '_';
735: if (p->enc == TERMENC_ASCII) {
736: cp = ascii_uc2str(uc);
737: csz = term_strlen(p, cp);
738: ssz = strlen(cp);
739: } else
740: csz = (*p->width)(p, uc);
741: while (lsz >= csz) {
742: if (p->enc == TERMENC_ASCII)
743: encode(p, cp, ssz);
744: else
745: encode1(p, uc);
746: lsz -= csz;
1.121 schwarze 747: }
1.93 schwarze 748: continue;
1.83 schwarze 749: case ESCAPE_SKIPCHAR:
1.108 schwarze 750: p->flags |= TERMP_BACKAFTER;
1.93 schwarze 751: continue;
1.103 schwarze 752: case ESCAPE_OVERSTRIKE:
753: cp = seq + sz;
754: while (seq < cp) {
755: if (*seq == '\\') {
756: mandoc_escape(&seq, NULL, NULL);
757: continue;
758: }
759: encode1(p, *seq++);
1.108 schwarze 760: if (seq < cp) {
761: if (p->flags & TERMP_BACKBEFORE)
762: p->flags |= TERMP_BACKAFTER;
763: else
764: p->flags |= TERMP_BACKBEFORE;
765: }
1.103 schwarze 766: }
1.109 schwarze 767: /* Trim trailing backspace/blank pair. */
1.129 schwarze 768: if (p->tcol->lastcol > 2 &&
769: (p->tcol->buf[p->tcol->lastcol - 1] == ' ' ||
770: p->tcol->buf[p->tcol->lastcol - 1] == '\t'))
771: p->tcol->lastcol -= 2;
772: if (p->col > p->tcol->lastcol)
773: p->col = p->tcol->lastcol;
1.108 schwarze 774: continue;
1.149 schwarze 775: case ESCAPE_IGNORE:
776: bufferc(p, ASCII_NBRZW);
777: continue;
1.20 schwarze 778: default:
1.93 schwarze 779: continue;
780: }
781:
782: /*
783: * Common handling for Unicode and numbered
784: * character escape sequences.
785: */
786:
787: if (p->enc == TERMENC_ASCII) {
788: cp = ascii_uc2str(uc);
789: encode(p, cp, strlen(cp));
790: } else {
791: if ((uc < 0x20 && uc != 0x09) ||
792: (uc > 0x7E && uc < 0xA0))
793: uc = 0xFFFD;
794: encode1(p, uc);
1.20 schwarze 795: }
796: }
1.75 schwarze 797: p->flags &= ~TERMP_NBRWORD;
1.1 kristaps 798: }
799:
800: static void
1.126 schwarze 801: adjbuf(struct termp_col *c, size_t sz)
1.1 kristaps 802: {
1.126 schwarze 803: if (c->maxcols == 0)
804: c->maxcols = 1024;
805: while (c->maxcols <= sz)
806: c->maxcols <<= 2;
807: c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf));
1.1 kristaps 808: }
809:
1.4 schwarze 810: static void
1.20 schwarze 811: bufferc(struct termp *p, char c)
812: {
1.124 schwarze 813: if (p->flags & TERMP_NOBUF) {
814: (*p->letter)(p, c);
815: return;
816: }
1.126 schwarze 817: if (p->col + 1 >= p->tcol->maxcols)
818: adjbuf(p->tcol, p->col + 1);
1.129 schwarze 819: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 820: p->tcol->buf[p->col] = c;
1.129 schwarze 821: if (p->tcol->lastcol < ++p->col)
822: p->tcol->lastcol = p->col;
1.20 schwarze 823: }
824:
1.150 schwarze 825: void
826: term_tab_ref(struct termp *p)
827: {
828: if (p->tcol->lastcol && p->tcol->lastcol <= p->col &&
829: (p->flags & TERMP_NOBUF) == 0)
830: bufferc(p, ASCII_TABREF);
831: }
832:
1.59 schwarze 833: /*
834: * See encode().
835: * Do this for a single (probably unicode) value.
836: * Does not check for non-decorated glyphs.
837: */
838: static void
839: encode1(struct termp *p, int c)
840: {
841: enum termfont f;
842:
1.124 schwarze 843: if (p->flags & TERMP_NOBUF) {
844: (*p->letter)(p, c);
845: return;
846: }
847:
1.126 schwarze 848: if (p->col + 7 >= p->tcol->maxcols)
849: adjbuf(p->tcol, p->col + 7);
1.59 schwarze 850:
1.115 schwarze 851: f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ?
1.108 schwarze 852: p->fontq[p->fonti] : TERMFONT_NONE;
1.59 schwarze 853:
1.108 schwarze 854: if (p->flags & TERMP_BACKBEFORE) {
1.126 schwarze 855: if (p->tcol->buf[p->col - 1] == ' ' ||
856: p->tcol->buf[p->col - 1] == '\t')
1.109 schwarze 857: p->col--;
858: else
1.126 schwarze 859: p->tcol->buf[p->col++] = '\b';
1.108 schwarze 860: p->flags &= ~TERMP_BACKBEFORE;
861: }
1.126 schwarze 862: if (f == TERMFONT_UNDER || f == TERMFONT_BI) {
863: p->tcol->buf[p->col++] = '_';
864: p->tcol->buf[p->col++] = '\b';
865: }
866: if (f == TERMFONT_BOLD || f == TERMFONT_BI) {
867: if (c == ASCII_HYPH)
868: p->tcol->buf[p->col++] = '-';
1.70 schwarze 869: else
1.126 schwarze 870: p->tcol->buf[p->col++] = c;
871: p->tcol->buf[p->col++] = '\b';
1.70 schwarze 872: }
1.129 schwarze 873: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.126 schwarze 874: p->tcol->buf[p->col] = c;
1.129 schwarze 875: if (p->tcol->lastcol < ++p->col)
876: p->tcol->lastcol = p->col;
1.108 schwarze 877: if (p->flags & TERMP_BACKAFTER) {
878: p->flags |= TERMP_BACKBEFORE;
879: p->flags &= ~TERMP_BACKAFTER;
880: }
1.59 schwarze 881: }
1.20 schwarze 882:
883: static void
884: encode(struct termp *p, const char *word, size_t sz)
1.4 schwarze 885: {
1.71 schwarze 886: size_t i;
1.124 schwarze 887:
888: if (p->flags & TERMP_NOBUF) {
889: for (i = 0; i < sz; i++)
890: (*p->letter)(p, word[i]);
891: return;
892: }
1.59 schwarze 893:
1.126 schwarze 894: if (p->col + 2 + (sz * 5) >= p->tcol->maxcols)
895: adjbuf(p->tcol, p->col + 2 + (sz * 5));
1.46 schwarze 896:
1.71 schwarze 897: for (i = 0; i < sz; i++) {
1.70 schwarze 898: if (ASCII_HYPH == word[i] ||
899: isgraph((unsigned char)word[i]))
900: encode1(p, word[i]);
1.119 schwarze 901: else {
1.129 schwarze 902: if (p->tcol->lastcol <= p->col ||
1.125 schwarze 903: (word[i] != ' ' && word[i] != ASCII_NBRSP))
1.126 schwarze 904: p->tcol->buf[p->col] = word[i];
1.125 schwarze 905: p->col++;
1.119 schwarze 906:
907: /*
908: * Postpone the effect of \z while handling
909: * an overstrike sequence from ascii_uc2str().
910: */
911:
912: if (word[i] == '\b' &&
913: (p->flags & TERMP_BACKBEFORE)) {
914: p->flags &= ~TERMP_BACKBEFORE;
915: p->flags |= TERMP_BACKAFTER;
916: }
917: }
1.4 schwarze 918: }
1.129 schwarze 919: if (p->tcol->lastcol < p->col)
920: p->tcol->lastcol = p->col;
1.80 schwarze 921: }
922:
923: void
924: term_setwidth(struct termp *p, const char *wstr)
925: {
926: struct roffsu su;
1.107 schwarze 927: int iop, width;
1.80 schwarze 928:
1.81 schwarze 929: iop = 0;
930: width = 0;
1.80 schwarze 931: if (NULL != wstr) {
932: switch (*wstr) {
1.83 schwarze 933: case '+':
1.80 schwarze 934: iop = 1;
935: wstr++;
936: break;
1.83 schwarze 937: case '-':
1.80 schwarze 938: iop = -1;
939: wstr++;
940: break;
941: default:
942: break;
943: }
1.128 schwarze 944: if (a2roffsu(wstr, &su, SCALE_MAX) != NULL)
1.81 schwarze 945: width = term_hspan(p, &su);
946: else
1.80 schwarze 947: iop = 0;
948: }
949: (*p->setwidth)(p, iop, width);
1.4 schwarze 950: }
1.16 schwarze 951:
952: size_t
1.39 schwarze 953: term_len(const struct termp *p, size_t sz)
954: {
955:
1.112 schwarze 956: return (*p->width)(p, ' ') * sz;
1.39 schwarze 957: }
958:
1.64 schwarze 959: static size_t
960: cond_width(const struct termp *p, int c, int *skip)
961: {
962:
963: if (*skip) {
964: (*skip) = 0;
1.112 schwarze 965: return 0;
1.64 schwarze 966: } else
1.112 schwarze 967: return (*p->width)(p, c);
1.64 schwarze 968: }
1.39 schwarze 969:
970: size_t
971: term_strlen(const struct termp *p, const char *cp)
972: {
1.59 schwarze 973: size_t sz, rsz, i;
1.93 schwarze 974: int ssz, skip, uc;
1.50 schwarze 975: const char *seq, *rhs;
1.59 schwarze 976: enum mandoc_esc esc;
1.149 schwarze 977: static const char rej[] = { '\\', ASCII_NBRSP, ASCII_NBRZW,
1.150 schwarze 978: ASCII_BREAK, ASCII_HYPH, ASCII_TABREF, '\0' };
1.59 schwarze 979:
980: /*
981: * Account for escaped sequences within string length
982: * calculations. This follows the logic in term_word() as we
983: * must calculate the width of produced strings.
984: */
985:
986: sz = 0;
1.64 schwarze 987: skip = 0;
1.59 schwarze 988: while ('\0' != *cp) {
989: rsz = strcspn(cp, rej);
990: for (i = 0; i < rsz; i++)
1.64 schwarze 991: sz += cond_width(p, *cp++, &skip);
1.59 schwarze 992:
993: switch (*cp) {
1.83 schwarze 994: case '\\':
1.59 schwarze 995: cp++;
1.137 schwarze 996: rhs = NULL;
1.59 schwarze 997: esc = mandoc_escape(&cp, &seq, &ssz);
998: switch (esc) {
1.83 schwarze 999: case ESCAPE_UNICODE:
1.94 schwarze 1000: uc = mchars_num2uc(seq + 1, ssz - 1);
1.59 schwarze 1001: break;
1.83 schwarze 1002: case ESCAPE_NUMBERED:
1.93 schwarze 1003: uc = mchars_num2char(seq, ssz);
1004: if (uc < 0)
1005: continue;
1.50 schwarze 1006: break;
1.83 schwarze 1007: case ESCAPE_SPECIAL:
1.93 schwarze 1008: if (p->enc == TERMENC_ASCII) {
1.114 schwarze 1009: rhs = mchars_spec2str(seq, ssz, &rsz);
1.93 schwarze 1010: if (rhs != NULL)
1011: break;
1012: } else {
1.114 schwarze 1013: uc = mchars_spec2cp(seq, ssz);
1.93 schwarze 1014: if (uc > 0)
1015: sz += cond_width(p, uc, &skip);
1.89 schwarze 1016: }
1.93 schwarze 1017: continue;
1.137 schwarze 1018: case ESCAPE_UNDEF:
1019: uc = *seq;
1020: break;
1.135 schwarze 1021: case ESCAPE_DEVICE:
1022: if (p->type == TERMTYPE_PDF) {
1023: rhs = "pdf";
1024: rsz = 3;
1025: } else if (p->type == TERMTYPE_PS) {
1026: rhs = "ps";
1027: rsz = 2;
1028: } else if (p->enc == TERMENC_ASCII) {
1029: rhs = "ascii";
1030: rsz = 5;
1031: } else {
1032: rhs = "utf8";
1033: rsz = 4;
1034: }
1035: break;
1.83 schwarze 1036: case ESCAPE_SKIPCHAR:
1.64 schwarze 1037: skip = 1;
1.103 schwarze 1038: continue;
1039: case ESCAPE_OVERSTRIKE:
1040: rsz = 0;
1041: rhs = seq + ssz;
1042: while (seq < rhs) {
1043: if (*seq == '\\') {
1044: mandoc_escape(&seq, NULL, NULL);
1045: continue;
1046: }
1047: i = (*p->width)(p, *seq++);
1048: if (rsz < i)
1049: rsz = i;
1050: }
1051: sz += rsz;
1.93 schwarze 1052: continue;
1.50 schwarze 1053: default:
1.93 schwarze 1054: continue;
1.50 schwarze 1055: }
1.39 schwarze 1056:
1.93 schwarze 1057: /*
1058: * Common handling for Unicode and numbered
1059: * character escape sequences.
1060: */
1061:
1062: if (rhs == NULL) {
1063: if (p->enc == TERMENC_ASCII) {
1064: rhs = ascii_uc2str(uc);
1065: rsz = strlen(rhs);
1066: } else {
1067: if ((uc < 0x20 && uc != 0x09) ||
1068: (uc > 0x7E && uc < 0xA0))
1069: uc = 0xFFFD;
1070: sz += cond_width(p, uc, &skip);
1071: continue;
1072: }
1073: }
1.59 schwarze 1074:
1.64 schwarze 1075: if (skip) {
1076: skip = 0;
1077: break;
1078: }
1.93 schwarze 1079:
1080: /*
1081: * Common handling for all escape sequences
1082: * printing more than one character.
1083: */
1.64 schwarze 1084:
1.59 schwarze 1085: for (i = 0; i < rsz; i++)
1086: sz += (*p->width)(p, *rhs++);
1087: break;
1.83 schwarze 1088: case ASCII_NBRSP:
1.64 schwarze 1089: sz += cond_width(p, ' ', &skip);
1.55 schwarze 1090: cp++;
1.59 schwarze 1091: break;
1.83 schwarze 1092: case ASCII_HYPH:
1.64 schwarze 1093: sz += cond_width(p, '-', &skip);
1.55 schwarze 1094: cp++;
1.59 schwarze 1095: break;
1096: default:
1097: break;
1098: }
1099: }
1.39 schwarze 1100:
1.112 schwarze 1101: return sz;
1.39 schwarze 1102: }
1103:
1.100 schwarze 1104: int
1.39 schwarze 1105: term_vspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 1106: {
1107: double r;
1.101 schwarze 1108: int ri;
1.16 schwarze 1109:
1110: switch (su->unit) {
1.99 schwarze 1111: case SCALE_BU:
1112: r = su->scale / 40.0;
1113: break;
1.83 schwarze 1114: case SCALE_CM:
1.99 schwarze 1115: r = su->scale * 6.0 / 2.54;
1116: break;
1117: case SCALE_FS:
1118: r = su->scale * 65536.0 / 40.0;
1.16 schwarze 1119: break;
1.83 schwarze 1120: case SCALE_IN:
1.86 schwarze 1121: r = su->scale * 6.0;
1.16 schwarze 1122: break;
1.99 schwarze 1123: case SCALE_MM:
1124: r = su->scale * 0.006;
1125: break;
1.83 schwarze 1126: case SCALE_PC:
1.16 schwarze 1127: r = su->scale;
1128: break;
1.83 schwarze 1129: case SCALE_PT:
1.99 schwarze 1130: r = su->scale / 12.0;
1.16 schwarze 1131: break;
1.99 schwarze 1132: case SCALE_EN:
1133: case SCALE_EM:
1134: r = su->scale * 0.6;
1.16 schwarze 1135: break;
1.83 schwarze 1136: case SCALE_VS:
1.16 schwarze 1137: r = su->scale;
1138: break;
1139: default:
1.99 schwarze 1140: abort();
1.16 schwarze 1141: }
1.101 schwarze 1142: ri = r > 0.0 ? r + 0.4995 : r - 0.4995;
1.112 schwarze 1143: return ri < 66 ? ri : 1;
1.16 schwarze 1144: }
1145:
1.107 schwarze 1146: /*
1.131 schwarze 1147: * Convert a scaling width to basic units, rounding towards 0.
1.107 schwarze 1148: */
1.100 schwarze 1149: int
1.39 schwarze 1150: term_hspan(const struct termp *p, const struct roffsu *su)
1.16 schwarze 1151: {
1152:
1.112 schwarze 1153: return (*p->hspan)(p, su);
1.131 schwarze 1154: }
1155:
1156: /*
1157: * Convert a scaling width to basic units, rounding to closest.
1158: */
1159: int
1160: term_hen(const struct termp *p, const struct roffsu *su)
1161: {
1162: int bu;
1163:
1164: if ((bu = (*p->hspan)(p, su)) >= 0)
1165: return (bu + 11) / 24;
1166: else
1167: return -((-bu + 11) / 24);
1.16 schwarze 1168: }
![[BACK]](/icons/back.gif){kind=icon}
Return to term.c CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [local] / src / usr.bin / mandoc