Annotation of mandoc/term.c, Revision 1.282
1.282 ! schwarze 1: /* $Id: term.c,v 1.281 2019/06/03 20:23:41 schwarze Exp $ */
1.1 kristaps 2: /*
1.198 schwarze 3: * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.282 ! schwarze 4: * Copyright (c) 2010-2020 Ingo Schwarze <schwarze@openbsd.org>
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: {
1.266 schwarze 61: for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++)
62: free(p->tcol->buf);
63: free(p->tcols);
1.238 schwarze 64: free(p->fontq);
1.142 kristaps 65: free(p);
66: }
67:
68: void
1.222 schwarze 69: term_begin(struct termp *p, term_margin head,
1.246 schwarze 70: term_margin foot, const struct roff_meta *arg)
1.142 kristaps 71: {
72:
73: p->headf = head;
74: p->footf = foot;
75: p->argf = arg;
1.146 kristaps 76: (*p->begin)(p);
1.142 kristaps 77: }
78:
79: void
80: term_end(struct termp *p)
81: {
82:
1.146 kristaps 83: (*p->end)(p);
1.14 kristaps 84: }
85:
1.71 kristaps 86: /*
1.221 schwarze 87: * Flush a chunk of text. By default, break the output line each time
88: * the right margin is reached, and continue output on the next line
89: * at the same offset as the chunk itself. By default, also break the
1.278 schwarze 90: * output line at the end of the chunk. There are many flags modifying
91: * this behaviour, see the comments in the body of the function.
1.71 kristaps 92: */
93: void
94: term_flushln(struct termp *p)
1.53 kristaps 95: {
1.278 schwarze 96: size_t vbl; /* Number of blanks to prepend to the output. */
97: size_t vbr; /* Actual visual position of the end of field. */
98: size_t vfield; /* Desired visual field width. */
99: size_t vtarget; /* Desired visual position of the right margin. */
100: size_t ic; /* Character position in the input buffer. */
101: size_t nbr; /* Number of characters to print in this field. */
102:
103: /*
104: * Normally, start writing at the left margin, but with the
105: * NOPAD flag, start writing at the current position instead.
106: */
1.53 kristaps 107:
1.266 schwarze 108: vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ?
109: 0 : p->tcol->offset - p->viscol;
1.263 schwarze 110: if (p->minbl && vbl < p->minbl)
111: vbl = p->minbl;
1.115 kristaps 112:
1.274 schwarze 113: if ((p->flags & TERMP_MULTICOL) == 0)
1.267 schwarze 114: p->tcol->col = 0;
1.278 schwarze 115:
116: /* Loop over output lines. */
117:
118: for (;;) {
119: vfield = p->tcol->rmargin > p->viscol + vbl ?
120: p->tcol->rmargin - p->viscol - vbl : 0;
1.267 schwarze 121:
1.71 kristaps 122: /*
1.278 schwarze 123: * Normally, break the line at the the right margin
124: * of the field, but with the NOBREAK flag, only
125: * break it at the max right margin of the screen,
126: * and with the BRNEVER flag, never break it at all.
1.138 schwarze 127: */
1.267 schwarze 128:
1.282 ! schwarze 129: vtarget = (p->flags & TERMP_NOBREAK) == 0 ? vfield :
1.278 schwarze 130: p->maxrmargin > p->viscol + vbl ?
131: p->maxrmargin - p->viscol - vbl : 0;
132:
133: /*
134: * Figure out how much text will fit in the field.
135: * If there is whitespace only, print nothing.
136: */
137:
1.282 ! schwarze 138: term_fill(p, &nbr, &vbr,
! 139: p->flags & TERMP_BRNEVER ? SIZE_MAX : vtarget);
1.278 schwarze 140: if (nbr == 0)
141: break;
1.279 schwarze 142:
143: /*
144: * With the CENTER or RIGHT flag, increase the indentation
145: * to center the text between the left and right margins
146: * or to adjust it to the right margin, respectively.
147: */
148:
149: if (vbr < vtarget) {
150: if (p->flags & TERMP_CENTER)
151: vbl += (vtarget - vbr) / 2;
152: else if (p->flags & TERMP_RIGHT)
153: vbl += vtarget - vbr;
154: }
155:
156: /* Finally, print the field content. */
1.278 schwarze 157:
1.282 ! schwarze 158: term_field(p, vbl, nbr);
1.138 schwarze 159:
160: /*
1.278 schwarze 161: * If there is no text left in the field, exit the loop.
162: * If the BRTRSP flag is set, consider trailing
163: * whitespace significant when deciding whether
164: * the field fits or not.
1.71 kristaps 165: */
166:
1.278 schwarze 167: for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
168: switch (p->tcol->buf[ic]) {
169: case '\t':
170: if (p->flags & TERMP_BRTRSP)
171: vbr = term_tab_next(vbr);
172: continue;
173: case ' ':
174: if (p->flags & TERMP_BRTRSP)
175: vbr += (*p->width)(p, ' ');
1.270 schwarze 176: continue;
1.278 schwarze 177: case '\n':
178: case ASCII_BREAK:
179: continue;
180: default:
181: break;
1.270 schwarze 182: }
1.278 schwarze 183: break;
184: }
185: if (ic == p->tcol->lastcol)
186: break;
1.154 kristaps 187:
1.278 schwarze 188: /*
189: * At the location of an automtic line break, input
190: * space characters are consumed by the line break.
191: */
1.154 kristaps 192:
1.278 schwarze 193: while (p->tcol->col < p->tcol->lastcol &&
194: p->tcol->buf[p->tcol->col] == ' ')
195: p->tcol->col++;
1.154 kristaps 196:
1.278 schwarze 197: /*
198: * In multi-column mode, leave the rest of the text
199: * in the buffer to be handled by a subsequent
200: * invocation, such that the other columns of the
201: * table can be handled first.
202: * In single-column mode, simply break the line.
203: */
204:
205: if (p->flags & TERMP_MULTICOL)
206: return;
1.217 schwarze 207:
1.278 schwarze 208: endline(p);
209: p->viscol = 0;
1.53 kristaps 210:
1.71 kristaps 211: /*
1.278 schwarze 212: * Normally, start the next line at the same indentation
213: * as this one, but with the BRIND flag, start it at the
214: * right margin instead. This is used together with
215: * NOBREAK for the tags in various kinds of tagged lists.
1.81 kristaps 216: */
1.267 schwarze 217:
1.278 schwarze 218: vbl = p->flags & TERMP_BRIND ?
219: p->tcol->rmargin : p->tcol->offset;
220: }
1.267 schwarze 221:
1.278 schwarze 222: /* Reset output state in preparation for the next field. */
1.205 schwarze 223:
1.278 schwarze 224: p->col = p->tcol->col = p->tcol->lastcol = 0;
225: p->minbl = p->trailspace;
226: p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE | TERMP_NOPAD);
1.260 schwarze 227:
1.278 schwarze 228: if (p->flags & TERMP_MULTICOL)
229: return;
1.260 schwarze 230:
1.278 schwarze 231: /*
232: * The HANG flag means that the next field
233: * always follows on the same line.
234: * The NOBREAK flag means that the next field
235: * follows on the same line unless the field was overrun.
236: * Normally, break the line at the end of each field.
237: */
1.205 schwarze 238:
1.278 schwarze 239: if ((p->flags & TERMP_HANG) == 0 &&
240: ((p->flags & TERMP_NOBREAK) == 0 ||
241: vbr + term_len(p, p->trailspace) > vfield))
242: endline(p);
243: }
1.138 schwarze 244:
1.278 schwarze 245: /*
246: * Store the number of input characters to print in this field in *nbr
247: * and their total visual width to print in *vbr.
248: * If there is only whitespace in the field, both remain zero.
249: * The desired visual width of the field is provided by vtarget.
250: * If the first word is longer, the field will be overrun.
251: */
252: static void
253: term_fill(struct termp *p, size_t *nbr, size_t *vbr, size_t vtarget)
254: {
255: size_t ic; /* Character position in the input buffer. */
256: size_t vis; /* Visual position of the current character. */
257: size_t vn; /* Visual position of the next character. */
258: int breakline; /* Break at the end of this word. */
259: int graph; /* Last character was non-blank. */
260:
261: *nbr = *vbr = vis = 0;
262: breakline = graph = 0;
263: for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) {
264: switch (p->tcol->buf[ic]) {
265: case '\b': /* Escape \o (overstrike) or backspace markup. */
266: assert(ic > 0);
267: vis -= (*p->width)(p, p->tcol->buf[ic - 1]);
268: continue;
1.267 schwarze 269:
1.278 schwarze 270: case '\t': /* Normal ASCII whitespace. */
271: case ' ':
272: case ASCII_BREAK: /* Escape \: (breakpoint). */
273: switch (p->tcol->buf[ic]) {
274: case '\t':
275: vn = term_tab_next(vis);
1.140 kristaps 276: break;
1.278 schwarze 277: case ' ':
278: vn = vis + (*p->width)(p, ' ');
1.138 schwarze 279: break;
1.278 schwarze 280: case ASCII_BREAK:
281: vn = vis;
1.71 kristaps 282: break;
1.281 schwarze 283: default:
284: abort();
1.136 schwarze 285: }
1.278 schwarze 286: /* Can break at the end of a word. */
287: if (breakline || vn > vtarget)
288: break;
289: if (graph) {
290: *nbr = ic;
291: *vbr = vis;
292: graph = 0;
1.136 schwarze 293: }
1.278 schwarze 294: vis = vn;
295: continue;
296:
297: case '\n': /* Escape \p (break at the end of the word). */
298: breakline = 1;
299: continue;
1.130 kristaps 300:
1.278 schwarze 301: case ASCII_HYPH: /* Breakable hyphen. */
302: graph = 1;
1.136 schwarze 303: /*
1.278 schwarze 304: * We are about to decide whether to break the
305: * line or not, so we no longer need this hyphen
306: * to be marked as breakable. Put back a real
307: * hyphen such that we get the correct width.
1.136 schwarze 308: */
1.278 schwarze 309: p->tcol->buf[ic] = '-';
310: vis += (*p->width)(p, '-');
311: if (vis > vtarget) {
312: ic++;
313: break;
1.200 schwarze 314: }
1.278 schwarze 315: *nbr = ic + 1;
316: *vbr = vis;
317: continue;
1.200 schwarze 318:
1.278 schwarze 319: case ASCII_NBRSP: /* Non-breakable space. */
320: p->tcol->buf[ic] = ' ';
321: /* FALLTHROUGH */
322: default: /* Printable character. */
323: graph = 1;
324: vis += (*p->width)(p, p->tcol->buf[ic]);
325: if (vis > vtarget && *nbr > 0)
326: return;
327: continue;
1.136 schwarze 328: }
1.278 schwarze 329: break;
330: }
1.270 schwarze 331:
1.278 schwarze 332: /*
333: * If the last word extends to the end of the field without any
334: * trailing whitespace, the loop could not check yet whether it
335: * can remain on this line. So do the check now.
336: */
1.270 schwarze 337:
1.278 schwarze 338: if (graph && (vis <= vtarget || *nbr == 0)) {
339: *nbr = ic;
340: *vbr = vis;
341: }
342: }
1.270 schwarze 343:
1.278 schwarze 344: /*
345: * Print the contents of one field
346: * with an indentation of vbl visual columns,
1.282 ! schwarze 347: * and an input string length of nbr characters.
1.278 schwarze 348: */
349: static void
1.282 ! schwarze 350: term_field(struct termp *p, size_t vbl, size_t nbr)
1.278 schwarze 351: {
352: size_t ic; /* Character position in the input buffer. */
353: size_t vis; /* Visual position of the current character. */
354: size_t dv; /* Visual width of the current character. */
355: size_t vn; /* Visual position of the next character. */
1.270 schwarze 356:
1.278 schwarze 357: vis = 0;
358: for (ic = p->tcol->col; ic < nbr; ic++) {
1.270 schwarze 359:
1.278 schwarze 360: /*
361: * To avoid the printing of trailing whitespace,
362: * do not print whitespace right away, only count it.
363: */
1.270 schwarze 364:
1.278 schwarze 365: switch (p->tcol->buf[ic]) {
366: case '\n':
367: case ASCII_BREAK:
368: continue;
369: case '\t':
370: vn = term_tab_next(vis);
371: vbl += vn - vis;
372: vis = vn;
373: continue;
374: case ' ':
375: case ASCII_NBRSP:
1.280 schwarze 376: dv = (*p->width)(p, ' ');
377: vbl += dv;
378: vis += dv;
1.278 schwarze 379: continue;
380: default:
381: break;
382: }
1.168 schwarze 383:
1.278 schwarze 384: /*
385: * We found a non-blank character to print,
386: * so write preceding white space now.
387: */
1.267 schwarze 388:
1.278 schwarze 389: if (vbl > 0) {
390: (*p->advance)(p, vbl);
391: p->viscol += vbl;
392: vbl = 0;
393: }
1.111 kristaps 394:
1.278 schwarze 395: /* Print the character and adjust the visual position. */
1.15 kristaps 396:
1.278 schwarze 397: (*p->letter)(p, p->tcol->buf[ic]);
398: if (p->tcol->buf[ic] == '\b') {
399: dv = (*p->width)(p, p->tcol->buf[ic - 1]);
400: p->viscol -= dv;
401: vis -= dv;
402: } else {
403: dv = (*p->width)(p, p->tcol->buf[ic]);
404: p->viscol += dv;
405: vis += dv;
406: }
407: }
408: p->tcol->col = nbr;
1.264 schwarze 409: }
410:
411: static void
412: endline(struct termp *p)
413: {
414: if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) {
415: p->mc = NULL;
416: p->flags &= ~TERMP_ENDMC;
417: }
418: if (p->mc != NULL) {
419: if (p->viscol && p->maxrmargin >= p->viscol)
420: (*p->advance)(p, p->maxrmargin - p->viscol + 1);
421: p->flags |= TERMP_NOBUF | TERMP_NOSPACE;
422: term_word(p, p->mc);
423: p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC);
424: }
425: p->viscol = 0;
426: p->minbl = 0;
427: (*p->endline)(p);
1.15 kristaps 428: }
429:
1.222 schwarze 430: /*
1.71 kristaps 431: * A newline only breaks an existing line; it won't assert vertical
432: * space. All data in the output buffer is flushed prior to the newline
433: * assertion.
434: */
435: void
436: term_newln(struct termp *p)
1.15 kristaps 437: {
438:
1.71 kristaps 439: p->flags |= TERMP_NOSPACE;
1.269 schwarze 440: if (p->tcol->lastcol || p->viscol)
1.200 schwarze 441: term_flushln(p);
1.16 kristaps 442: }
443:
1.71 kristaps 444: /*
445: * Asserts a vertical space (a full, empty line-break between lines).
446: * Note that if used twice, this will cause two blank spaces and so on.
447: * All data in the output buffer is flushed prior to the newline
448: * assertion.
449: */
450: void
451: term_vspace(struct termp *p)
1.16 kristaps 452: {
453:
1.62 kristaps 454: term_newln(p);
1.139 schwarze 455: p->viscol = 0;
1.264 schwarze 456: p->minbl = 0;
1.202 schwarze 457: if (0 < p->skipvsp)
458: p->skipvsp--;
459: else
460: (*p->endline)(p);
1.16 kristaps 461: }
462:
1.238 schwarze 463: /* Swap current and previous font; for \fP and .ft P */
1.125 kristaps 464: void
465: term_fontlast(struct termp *p)
466: {
467: enum termfont f;
468:
469: f = p->fontl;
470: p->fontl = p->fontq[p->fonti];
471: p->fontq[p->fonti] = f;
472: }
473:
1.238 schwarze 474: /* Set font, save current, discard previous; for \f, .ft, .B etc. */
1.125 kristaps 475: void
476: term_fontrepl(struct termp *p, enum termfont f)
477: {
478:
479: p->fontl = p->fontq[p->fonti];
480: p->fontq[p->fonti] = f;
481: }
482:
1.238 schwarze 483: /* Set font, save previous. */
1.125 kristaps 484: void
485: term_fontpush(struct termp *p, enum termfont f)
486: {
487:
488: p->fontl = p->fontq[p->fonti];
1.238 schwarze 489: if (++p->fonti == p->fontsz) {
490: p->fontsz += 8;
491: p->fontq = mandoc_reallocarray(p->fontq,
1.256 schwarze 492: p->fontsz, sizeof(*p->fontq));
1.238 schwarze 493: }
494: p->fontq[p->fonti] = f;
1.125 kristaps 495: }
496:
1.238 schwarze 497: /* Flush to make the saved pointer current again. */
1.125 kristaps 498: void
1.244 schwarze 499: term_fontpopq(struct termp *p, int i)
1.125 kristaps 500: {
501:
1.244 schwarze 502: assert(i >= 0);
503: if (p->fonti > i)
504: p->fonti = i;
1.125 kristaps 505: }
1.94 kristaps 506:
1.238 schwarze 507: /* Pop one font off the stack. */
1.125 kristaps 508: void
509: term_fontpop(struct termp *p)
510: {
511:
512: assert(p->fonti);
513: p->fonti--;
1.17 kristaps 514: }
515:
1.71 kristaps 516: /*
517: * Handle pwords, partial words, which may be either a single word or a
518: * phrase that cannot be broken down (such as a literal string). This
519: * handles word styling.
520: */
1.86 kristaps 521: void
522: term_word(struct termp *p, const char *word)
1.65 kristaps 523: {
1.261 schwarze 524: struct roffsu su;
1.214 schwarze 525: const char nbrsp[2] = { ASCII_NBRSP, 0 };
1.191 kristaps 526: const char *seq, *cp;
1.194 kristaps 527: int sz, uc;
1.262 schwarze 528: size_t csz, lsz, ssz;
1.184 kristaps 529: enum mandoc_esc esc;
1.100 kristaps 530:
1.264 schwarze 531: if ((p->flags & TERMP_NOBUF) == 0) {
532: if ((p->flags & TERMP_NOSPACE) == 0) {
533: if ((p->flags & TERMP_KEEP) == 0) {
1.151 schwarze 534: bufferc(p, ' ');
1.264 schwarze 535: if (p->flags & TERMP_SENTENCE)
536: bufferc(p, ' ');
537: } else
538: bufferc(p, ASCII_NBRSP);
539: }
540: if (p->flags & TERMP_PREKEEP)
541: p->flags |= TERMP_KEEP;
542: if (p->flags & TERMP_NONOSPACE)
543: p->flags |= TERMP_NOSPACE;
544: else
545: p->flags &= ~TERMP_NOSPACE;
546: p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE);
547: p->skipvsp = 0;
1.133 kristaps 548: }
1.65 kristaps 549:
1.184 kristaps 550: while ('\0' != *word) {
1.203 schwarze 551: if ('\\' != *word) {
1.214 schwarze 552: if (TERMP_NBRWORD & p->flags) {
553: if (' ' == *word) {
554: encode(p, nbrsp, 1);
555: word++;
556: continue;
557: }
558: ssz = strcspn(word, "\\ ");
559: } else
560: ssz = strcspn(word, "\\");
1.162 kristaps 561: encode(p, word, ssz);
1.203 schwarze 562: word += (int)ssz;
1.124 kristaps 563: continue;
1.203 schwarze 564: }
1.124 kristaps 565:
1.184 kristaps 566: word++;
567: esc = mandoc_escape(&word, &seq, &sz);
568: switch (esc) {
1.222 schwarze 569: case ESCAPE_UNICODE:
1.229 schwarze 570: uc = mchars_num2uc(seq + 1, sz - 1);
1.192 kristaps 571: break;
1.222 schwarze 572: case ESCAPE_NUMBERED:
1.233 schwarze 573: uc = mchars_num2char(seq, sz);
574: if (uc < 0)
575: continue;
1.184 kristaps 576: break;
1.222 schwarze 577: case ESCAPE_SPECIAL:
1.229 schwarze 578: if (p->enc == TERMENC_ASCII) {
1.254 schwarze 579: cp = mchars_spec2str(seq, sz, &ssz);
1.232 schwarze 580: if (cp != NULL)
1.229 schwarze 581: encode(p, cp, ssz);
582: } else {
1.254 schwarze 583: uc = mchars_spec2cp(seq, sz);
1.230 schwarze 584: if (uc > 0)
585: encode1(p, uc);
1.229 schwarze 586: }
1.233 schwarze 587: continue;
1.277 schwarze 588: case ESCAPE_UNDEF:
589: uc = *seq;
590: break;
1.222 schwarze 591: case ESCAPE_FONTBOLD:
1.125 kristaps 592: term_fontrepl(p, TERMFONT_BOLD);
1.233 schwarze 593: continue;
1.222 schwarze 594: case ESCAPE_FONTITALIC:
1.125 kristaps 595: term_fontrepl(p, TERMFONT_UNDER);
1.233 schwarze 596: continue;
1.222 schwarze 597: case ESCAPE_FONTBI:
1.209 schwarze 598: term_fontrepl(p, TERMFONT_BI);
1.233 schwarze 599: continue;
1.222 schwarze 600: case ESCAPE_FONT:
1.276 schwarze 601: case ESCAPE_FONTCW:
1.222 schwarze 602: case ESCAPE_FONTROMAN:
1.125 kristaps 603: term_fontrepl(p, TERMFONT_NONE);
1.233 schwarze 604: continue;
1.222 schwarze 605: case ESCAPE_FONTPREV:
1.125 kristaps 606: term_fontlast(p);
1.270 schwarze 607: continue;
608: case ESCAPE_BREAK:
609: bufferc(p, '\n');
1.233 schwarze 610: continue;
1.222 schwarze 611: case ESCAPE_NOSPACE:
1.248 schwarze 612: if (p->flags & TERMP_BACKAFTER)
613: p->flags &= ~TERMP_BACKAFTER;
614: else if (*word == '\0')
1.237 schwarze 615: p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE);
1.261 schwarze 616: continue;
1.275 schwarze 617: case ESCAPE_DEVICE:
618: if (p->type == TERMTYPE_PDF)
619: encode(p, "pdf", 3);
620: else if (p->type == TERMTYPE_PS)
621: encode(p, "ps", 2);
622: else if (p->enc == TERMENC_ASCII)
623: encode(p, "ascii", 5);
624: else
625: encode(p, "utf8", 4);
626: continue;
1.261 schwarze 627: case ESCAPE_HORIZ:
1.273 schwarze 628: if (*seq == '|') {
629: seq++;
630: uc = -p->col;
631: } else
632: uc = 0;
1.268 schwarze 633: if (a2roffsu(seq, &su, SCALE_EM) == NULL)
1.261 schwarze 634: continue;
1.273 schwarze 635: uc += term_hen(p, &su);
1.261 schwarze 636: if (uc > 0)
637: while (uc-- > 0)
638: bufferc(p, ASCII_NBRSP);
639: else if (p->col > (size_t)(-uc))
640: p->col += uc;
641: else {
642: uc += p->col;
643: p->col = 0;
1.266 schwarze 644: if (p->tcol->offset > (size_t)(-uc)) {
1.261 schwarze 645: p->ti += uc;
1.266 schwarze 646: p->tcol->offset += uc;
1.261 schwarze 647: } else {
1.266 schwarze 648: p->ti -= p->tcol->offset;
649: p->tcol->offset = 0;
1.261 schwarze 650: }
1.262 schwarze 651: }
652: continue;
653: case ESCAPE_HLINE:
1.272 schwarze 654: if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL)
1.262 schwarze 655: continue;
1.271 schwarze 656: uc = term_hen(p, &su);
1.262 schwarze 657: if (uc <= 0) {
1.266 schwarze 658: if (p->tcol->rmargin <= p->tcol->offset)
1.262 schwarze 659: continue;
1.266 schwarze 660: lsz = p->tcol->rmargin - p->tcol->offset;
1.262 schwarze 661: } else
662: lsz = uc;
1.272 schwarze 663: if (*cp == seq[-1])
1.262 schwarze 664: uc = -1;
1.272 schwarze 665: else if (*cp == '\\') {
666: seq = cp + 1;
1.262 schwarze 667: esc = mandoc_escape(&seq, &cp, &sz);
668: switch (esc) {
669: case ESCAPE_UNICODE:
670: uc = mchars_num2uc(cp + 1, sz - 1);
671: break;
672: case ESCAPE_NUMBERED:
673: uc = mchars_num2char(cp, sz);
674: break;
675: case ESCAPE_SPECIAL:
676: uc = mchars_spec2cp(cp, sz);
677: break;
1.277 schwarze 678: case ESCAPE_UNDEF:
679: uc = *seq;
680: break;
1.262 schwarze 681: default:
682: uc = -1;
683: break;
684: }
685: } else
1.272 schwarze 686: uc = *cp;
1.262 schwarze 687: if (uc < 0x20 || (uc > 0x7E && uc < 0xA0))
688: uc = '_';
689: if (p->enc == TERMENC_ASCII) {
690: cp = ascii_uc2str(uc);
691: csz = term_strlen(p, cp);
692: ssz = strlen(cp);
693: } else
694: csz = (*p->width)(p, uc);
695: while (lsz >= csz) {
696: if (p->enc == TERMENC_ASCII)
697: encode(p, cp, ssz);
698: else
699: encode1(p, uc);
700: lsz -= csz;
1.261 schwarze 701: }
1.233 schwarze 702: continue;
1.222 schwarze 703: case ESCAPE_SKIPCHAR:
1.248 schwarze 704: p->flags |= TERMP_BACKAFTER;
1.233 schwarze 705: continue;
1.243 schwarze 706: case ESCAPE_OVERSTRIKE:
707: cp = seq + sz;
708: while (seq < cp) {
709: if (*seq == '\\') {
710: mandoc_escape(&seq, NULL, NULL);
711: continue;
712: }
713: encode1(p, *seq++);
1.248 schwarze 714: if (seq < cp) {
715: if (p->flags & TERMP_BACKBEFORE)
716: p->flags |= TERMP_BACKAFTER;
717: else
718: p->flags |= TERMP_BACKBEFORE;
719: }
1.243 schwarze 720: }
1.249 schwarze 721: /* Trim trailing backspace/blank pair. */
1.269 schwarze 722: if (p->tcol->lastcol > 2 &&
723: (p->tcol->buf[p->tcol->lastcol - 1] == ' ' ||
724: p->tcol->buf[p->tcol->lastcol - 1] == '\t'))
725: p->tcol->lastcol -= 2;
726: if (p->col > p->tcol->lastcol)
727: p->col = p->tcol->lastcol;
1.248 schwarze 728: continue;
1.124 kristaps 729: default:
1.233 schwarze 730: continue;
731: }
732:
733: /*
734: * Common handling for Unicode and numbered
735: * character escape sequences.
736: */
737:
738: if (p->enc == TERMENC_ASCII) {
739: cp = ascii_uc2str(uc);
740: encode(p, cp, strlen(cp));
741: } else {
742: if ((uc < 0x20 && uc != 0x09) ||
743: (uc > 0x7E && uc < 0xA0))
744: uc = 0xFFFD;
745: encode1(p, uc);
1.124 kristaps 746: }
747: }
1.214 schwarze 748: p->flags &= ~TERMP_NBRWORD;
1.65 kristaps 749: }
750:
1.71 kristaps 751: static void
1.266 schwarze 752: adjbuf(struct termp_col *c, size_t sz)
1.51 kristaps 753: {
1.266 schwarze 754: if (c->maxcols == 0)
755: c->maxcols = 1024;
756: while (c->maxcols <= sz)
757: c->maxcols <<= 2;
758: c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf));
1.51 kristaps 759: }
760:
1.79 kristaps 761: static void
1.125 kristaps 762: bufferc(struct termp *p, char c)
763: {
1.264 schwarze 764: if (p->flags & TERMP_NOBUF) {
765: (*p->letter)(p, c);
766: return;
767: }
1.266 schwarze 768: if (p->col + 1 >= p->tcol->maxcols)
769: adjbuf(p->tcol, p->col + 1);
1.269 schwarze 770: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.266 schwarze 771: p->tcol->buf[p->col] = c;
1.269 schwarze 772: if (p->tcol->lastcol < ++p->col)
773: p->tcol->lastcol = p->col;
1.125 kristaps 774: }
775:
1.194 kristaps 776: /*
777: * See encode().
778: * Do this for a single (probably unicode) value.
779: * Does not check for non-decorated glyphs.
780: */
781: static void
782: encode1(struct termp *p, int c)
783: {
784: enum termfont f;
785:
1.264 schwarze 786: if (p->flags & TERMP_NOBUF) {
787: (*p->letter)(p, c);
788: return;
789: }
790:
1.266 schwarze 791: if (p->col + 7 >= p->tcol->maxcols)
792: adjbuf(p->tcol, p->col + 7);
1.194 kristaps 793:
1.255 schwarze 794: f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ?
1.248 schwarze 795: p->fontq[p->fonti] : TERMFONT_NONE;
1.194 kristaps 796:
1.248 schwarze 797: if (p->flags & TERMP_BACKBEFORE) {
1.266 schwarze 798: if (p->tcol->buf[p->col - 1] == ' ' ||
799: p->tcol->buf[p->col - 1] == '\t')
1.249 schwarze 800: p->col--;
801: else
1.266 schwarze 802: p->tcol->buf[p->col++] = '\b';
1.248 schwarze 803: p->flags &= ~TERMP_BACKBEFORE;
804: }
1.266 schwarze 805: if (f == TERMFONT_UNDER || f == TERMFONT_BI) {
806: p->tcol->buf[p->col++] = '_';
807: p->tcol->buf[p->col++] = '\b';
808: }
809: if (f == TERMFONT_BOLD || f == TERMFONT_BI) {
810: if (c == ASCII_HYPH)
811: p->tcol->buf[p->col++] = '-';
1.209 schwarze 812: else
1.266 schwarze 813: p->tcol->buf[p->col++] = c;
814: p->tcol->buf[p->col++] = '\b';
1.209 schwarze 815: }
1.269 schwarze 816: if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP))
1.266 schwarze 817: p->tcol->buf[p->col] = c;
1.269 schwarze 818: if (p->tcol->lastcol < ++p->col)
819: p->tcol->lastcol = p->col;
1.248 schwarze 820: if (p->flags & TERMP_BACKAFTER) {
821: p->flags |= TERMP_BACKBEFORE;
822: p->flags &= ~TERMP_BACKAFTER;
823: }
1.194 kristaps 824: }
825:
1.125 kristaps 826: static void
827: encode(struct termp *p, const char *word, size_t sz)
828: {
1.210 schwarze 829: size_t i;
1.264 schwarze 830:
831: if (p->flags & TERMP_NOBUF) {
832: for (i = 0; i < sz; i++)
833: (*p->letter)(p, word[i]);
834: return;
835: }
1.188 kristaps 836:
1.266 schwarze 837: if (p->col + 2 + (sz * 5) >= p->tcol->maxcols)
838: adjbuf(p->tcol, p->col + 2 + (sz * 5));
1.165 kristaps 839:
1.210 schwarze 840: for (i = 0; i < sz; i++) {
1.209 schwarze 841: if (ASCII_HYPH == word[i] ||
842: isgraph((unsigned char)word[i]))
843: encode1(p, word[i]);
1.259 schwarze 844: else {
1.269 schwarze 845: if (p->tcol->lastcol <= p->col ||
1.265 schwarze 846: (word[i] != ' ' && word[i] != ASCII_NBRSP))
1.266 schwarze 847: p->tcol->buf[p->col] = word[i];
1.265 schwarze 848: p->col++;
1.259 schwarze 849:
850: /*
851: * Postpone the effect of \z while handling
852: * an overstrike sequence from ascii_uc2str().
853: */
854:
855: if (word[i] == '\b' &&
856: (p->flags & TERMP_BACKBEFORE)) {
857: p->flags &= ~TERMP_BACKBEFORE;
858: p->flags |= TERMP_BACKAFTER;
859: }
860: }
1.79 kristaps 861: }
1.269 schwarze 862: if (p->tcol->lastcol < p->col)
863: p->tcol->lastcol = p->col;
1.219 schwarze 864: }
865:
866: void
867: term_setwidth(struct termp *p, const char *wstr)
868: {
869: struct roffsu su;
1.247 schwarze 870: int iop, width;
1.219 schwarze 871:
1.220 schwarze 872: iop = 0;
873: width = 0;
1.219 schwarze 874: if (NULL != wstr) {
875: switch (*wstr) {
1.222 schwarze 876: case '+':
1.219 schwarze 877: iop = 1;
878: wstr++;
879: break;
1.222 schwarze 880: case '-':
1.219 schwarze 881: iop = -1;
882: wstr++;
883: break;
884: default:
885: break;
886: }
1.268 schwarze 887: if (a2roffsu(wstr, &su, SCALE_MAX) != NULL)
1.220 schwarze 888: width = term_hspan(p, &su);
889: else
1.219 schwarze 890: iop = 0;
891: }
892: (*p->setwidth)(p, iop, width);
1.79 kristaps 893: }
1.106 kristaps 894:
1.107 kristaps 895: size_t
1.149 kristaps 896: term_len(const struct termp *p, size_t sz)
897: {
898:
1.252 schwarze 899: return (*p->width)(p, ' ') * sz;
1.149 kristaps 900: }
901:
1.203 schwarze 902: static size_t
903: cond_width(const struct termp *p, int c, int *skip)
904: {
905:
906: if (*skip) {
907: (*skip) = 0;
1.252 schwarze 908: return 0;
1.203 schwarze 909: } else
1.252 schwarze 910: return (*p->width)(p, c);
1.203 schwarze 911: }
1.149 kristaps 912:
913: size_t
914: term_strlen(const struct termp *p, const char *cp)
915: {
1.184 kristaps 916: size_t sz, rsz, i;
1.233 schwarze 917: int ssz, skip, uc;
1.171 kristaps 918: const char *seq, *rhs;
1.196 kristaps 919: enum mandoc_esc esc;
1.216 schwarze 920: static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH,
921: ASCII_BREAK, '\0' };
1.171 kristaps 922:
1.184 kristaps 923: /*
924: * Account for escaped sequences within string length
925: * calculations. This follows the logic in term_word() as we
926: * must calculate the width of produced strings.
927: */
928:
929: sz = 0;
1.203 schwarze 930: skip = 0;
1.189 kristaps 931: while ('\0' != *cp) {
932: rsz = strcspn(cp, rej);
933: for (i = 0; i < rsz; i++)
1.203 schwarze 934: sz += cond_width(p, *cp++, &skip);
1.189 kristaps 935:
1.184 kristaps 936: switch (*cp) {
1.222 schwarze 937: case '\\':
1.189 kristaps 938: cp++;
1.277 schwarze 939: rhs = NULL;
1.196 kristaps 940: esc = mandoc_escape(&cp, &seq, &ssz);
941: switch (esc) {
1.222 schwarze 942: case ESCAPE_UNICODE:
1.234 schwarze 943: uc = mchars_num2uc(seq + 1, ssz - 1);
1.194 kristaps 944: break;
1.222 schwarze 945: case ESCAPE_NUMBERED:
1.233 schwarze 946: uc = mchars_num2char(seq, ssz);
947: if (uc < 0)
948: continue;
1.171 kristaps 949: break;
1.222 schwarze 950: case ESCAPE_SPECIAL:
1.233 schwarze 951: if (p->enc == TERMENC_ASCII) {
1.254 schwarze 952: rhs = mchars_spec2str(seq, ssz, &rsz);
1.233 schwarze 953: if (rhs != NULL)
954: break;
955: } else {
1.254 schwarze 956: uc = mchars_spec2cp(seq, ssz);
1.233 schwarze 957: if (uc > 0)
958: sz += cond_width(p, uc, &skip);
1.229 schwarze 959: }
1.233 schwarze 960: continue;
1.277 schwarze 961: case ESCAPE_UNDEF:
962: uc = *seq;
963: break;
1.275 schwarze 964: case ESCAPE_DEVICE:
965: if (p->type == TERMTYPE_PDF) {
966: rhs = "pdf";
967: rsz = 3;
968: } else if (p->type == TERMTYPE_PS) {
969: rhs = "ps";
970: rsz = 2;
971: } else if (p->enc == TERMENC_ASCII) {
972: rhs = "ascii";
973: rsz = 5;
974: } else {
975: rhs = "utf8";
976: rsz = 4;
977: }
978: break;
1.222 schwarze 979: case ESCAPE_SKIPCHAR:
1.203 schwarze 980: skip = 1;
1.243 schwarze 981: continue;
982: case ESCAPE_OVERSTRIKE:
983: rsz = 0;
984: rhs = seq + ssz;
985: while (seq < rhs) {
986: if (*seq == '\\') {
987: mandoc_escape(&seq, NULL, NULL);
988: continue;
989: }
990: i = (*p->width)(p, *seq++);
991: if (rsz < i)
992: rsz = i;
993: }
994: sz += rsz;
1.233 schwarze 995: continue;
1.171 kristaps 996: default:
1.233 schwarze 997: continue;
1.171 kristaps 998: }
1.149 kristaps 999:
1.233 schwarze 1000: /*
1001: * Common handling for Unicode and numbered
1002: * character escape sequences.
1003: */
1004:
1005: if (rhs == NULL) {
1006: if (p->enc == TERMENC_ASCII) {
1007: rhs = ascii_uc2str(uc);
1008: rsz = strlen(rhs);
1009: } else {
1010: if ((uc < 0x20 && uc != 0x09) ||
1011: (uc > 0x7E && uc < 0xA0))
1012: uc = 0xFFFD;
1013: sz += cond_width(p, uc, &skip);
1014: continue;
1015: }
1016: }
1.184 kristaps 1017:
1.203 schwarze 1018: if (skip) {
1019: skip = 0;
1020: break;
1021: }
1.233 schwarze 1022:
1023: /*
1024: * Common handling for all escape sequences
1025: * printing more than one character.
1026: */
1.203 schwarze 1027:
1.184 kristaps 1028: for (i = 0; i < rsz; i++)
1029: sz += (*p->width)(p, *rhs++);
1030: break;
1.222 schwarze 1031: case ASCII_NBRSP:
1.203 schwarze 1032: sz += cond_width(p, ' ', &skip);
1.176 kristaps 1033: cp++;
1.184 kristaps 1034: break;
1.222 schwarze 1035: case ASCII_HYPH:
1.203 schwarze 1036: sz += cond_width(p, '-', &skip);
1.176 kristaps 1037: cp++;
1.184 kristaps 1038: break;
1039: default:
1040: break;
1041: }
1.189 kristaps 1042: }
1.149 kristaps 1043:
1.252 schwarze 1044: return sz;
1.149 kristaps 1045: }
1046:
1.240 schwarze 1047: int
1.149 kristaps 1048: term_vspan(const struct termp *p, const struct roffsu *su)
1.106 kristaps 1049: {
1050: double r;
1.241 schwarze 1051: int ri;
1.106 kristaps 1052:
1.107 kristaps 1053: switch (su->unit) {
1.239 schwarze 1054: case SCALE_BU:
1055: r = su->scale / 40.0;
1056: break;
1.222 schwarze 1057: case SCALE_CM:
1.239 schwarze 1058: r = su->scale * 6.0 / 2.54;
1059: break;
1060: case SCALE_FS:
1061: r = su->scale * 65536.0 / 40.0;
1.106 kristaps 1062: break;
1.222 schwarze 1063: case SCALE_IN:
1.225 schwarze 1064: r = su->scale * 6.0;
1.106 kristaps 1065: break;
1.239 schwarze 1066: case SCALE_MM:
1067: r = su->scale * 0.006;
1068: break;
1.222 schwarze 1069: case SCALE_PC:
1.107 kristaps 1070: r = su->scale;
1.106 kristaps 1071: break;
1.222 schwarze 1072: case SCALE_PT:
1.239 schwarze 1073: r = su->scale / 12.0;
1.106 kristaps 1074: break;
1.239 schwarze 1075: case SCALE_EN:
1076: case SCALE_EM:
1077: r = su->scale * 0.6;
1.106 kristaps 1078: break;
1.222 schwarze 1079: case SCALE_VS:
1.107 kristaps 1080: r = su->scale;
1.106 kristaps 1081: break;
1082: default:
1.239 schwarze 1083: abort();
1.106 kristaps 1084: }
1.241 schwarze 1085: ri = r > 0.0 ? r + 0.4995 : r - 0.4995;
1.252 schwarze 1086: return ri < 66 ? ri : 1;
1.106 kristaps 1087: }
1088:
1.247 schwarze 1089: /*
1.271 schwarze 1090: * Convert a scaling width to basic units, rounding towards 0.
1.247 schwarze 1091: */
1.240 schwarze 1092: int
1.149 kristaps 1093: term_hspan(const struct termp *p, const struct roffsu *su)
1.106 kristaps 1094: {
1.108 kristaps 1095:
1.252 schwarze 1096: return (*p->hspan)(p, su);
1.271 schwarze 1097: }
1098:
1099: /*
1100: * Convert a scaling width to basic units, rounding to closest.
1101: */
1102: int
1103: term_hen(const struct termp *p, const struct roffsu *su)
1104: {
1105: int bu;
1106:
1107: if ((bu = (*p->hspan)(p, su)) >= 0)
1108: return (bu + 11) / 24;
1109: else
1110: return -((-bu + 11) / 24);
1.106 kristaps 1111: }
CVSweb