Annotation of src/usr.bin/less/line.c, Revision 1.23
1.1 etheisen 1: /*
1.12 shadchin 2: * Copyright (C) 1984-2012 Mark Nudelman
1.14 nicm 3: * Modified for use with illumos by Garrett D'Amore.
4: * Copyright 2014 Garrett D'Amore <garrett@damore.org>
1.1 etheisen 5: *
1.5 millert 6: * You may distribute under the terms of either the GNU General Public
7: * License or the Less License, as specified in the README file.
1.1 etheisen 8: *
1.12 shadchin 9: * For more information, see the README file.
1.13 nicm 10: */
1.1 etheisen 11:
12: /*
13: * Routines to manipulate the "line buffer".
14: * The line buffer holds a line of output as it is being built
15: * in preparation for output to the screen.
16: */
17:
1.23 ! schwarze 18: #include <wchar.h>
! 19:
1.19 mmcc 20: #include "charset.h"
1.1 etheisen 21: #include "less.h"
22:
1.10 shadchin 23: static char *linebuf = NULL; /* Buffer which holds the current output line */
1.5 millert 24: static char *attr = NULL; /* Extension of linebuf to hold attributes */
1.13 nicm 25: int size_linebuf = 0; /* Size of line buffer (and attr buffer) */
1.5 millert 26:
1.10 shadchin 27: static int cshift; /* Current left-shift of output line buffer */
1.13 nicm 28: int hshift; /* Desired left-shift of output line buffer */
29: int tabstops[TABSTOP_MAX] = { 0 }; /* Custom tabstops */
30: int ntabstops = 1; /* Number of tabstops */
31: int tabdefault = 8; /* Default repeated tabstops */
32: off_t highest_hilite; /* Pos of last hilite in file found so far */
1.1 etheisen 33:
34: static int curr; /* Index into linebuf */
1.13 nicm 35: static int column; /* Printable length, accounting for backspaces, etc. */
1.1 etheisen 36: static int overstrike; /* Next char should overstrike previous char */
37: static int is_null_line; /* There is no current line */
1.5 millert 38: static int lmargin; /* Left margin */
1.1 etheisen 39: static char pendc;
1.13 nicm 40: static off_t pendpos;
1.5 millert 41: static char *end_ansi_chars;
1.10 shadchin 42: static char *mid_ansi_chars;
1.1 etheisen 43:
1.13 nicm 44: static int attr_swidth(int);
45: static int attr_ewidth(int);
46: static int do_append(LWCHAR, char *, off_t);
1.1 etheisen 47:
1.11 millert 48: extern volatile sig_atomic_t sigs;
1.1 etheisen 49: extern int bs_mode;
50: extern int linenums;
51: extern int ctldisp;
52: extern int twiddle;
53: extern int binattr;
1.5 millert 54: extern int status_col;
1.1 etheisen 55: extern int auto_wrap, ignaw;
56: extern int bo_s_width, bo_e_width;
57: extern int ul_s_width, ul_e_width;
58: extern int bl_s_width, bl_e_width;
59: extern int so_s_width, so_e_width;
60: extern int sc_width, sc_height;
1.5 millert 61: extern int utf_mode;
1.13 nicm 62: extern off_t start_attnpos;
63: extern off_t end_attnpos;
1.5 millert 64:
1.10 shadchin 65: static char mbc_buf[MAX_UTF_CHAR_LEN];
66: static int mbc_buf_len = 0;
67: static int mbc_buf_index = 0;
1.13 nicm 68: static off_t mbc_pos;
1.10 shadchin 69:
1.5 millert 70: /*
71: * Initialize from environment variables.
72: */
1.13 nicm 73: void
74: init_line(void)
1.5 millert 75: {
76: end_ansi_chars = lgetenv("LESSANSIENDCHARS");
77: if (end_ansi_chars == NULL || *end_ansi_chars == '\0')
78: end_ansi_chars = "m";
1.10 shadchin 79:
80: mid_ansi_chars = lgetenv("LESSANSIMIDCHARS");
81: if (mid_ansi_chars == NULL || *mid_ansi_chars == '\0')
82: mid_ansi_chars = "0123456789;[?!\"'#%()*+ ";
83:
1.13 nicm 84: linebuf = ecalloc(LINEBUF_SIZE, sizeof (char));
85: attr = ecalloc(LINEBUF_SIZE, sizeof (char));
1.5 millert 86: size_linebuf = LINEBUF_SIZE;
87: }
88:
89: /*
90: * Expand the line buffer.
91: */
1.13 nicm 92: static int
93: expand_linebuf(void)
1.5 millert 94: {
1.10 shadchin 95: /* Double the size of the line buffer. */
1.7 millert 96: int new_size = size_linebuf * 2;
1.10 shadchin 97:
98: /* Just realloc to expand the buffer, if we can. */
1.22 deraadt 99: char *new_buf = recallocarray(linebuf, size_linebuf, new_size, 1);
100: char *new_attr = recallocarray(attr, size_linebuf, new_size, 1);
1.13 nicm 101: if (new_buf == NULL || new_attr == NULL) {
1.15 mmcc 102: free(new_attr);
103: free(new_buf);
1.13 nicm 104: return (1);
1.5 millert 105: }
106: linebuf = new_buf;
107: attr = new_attr;
108: size_linebuf = new_size;
1.13 nicm 109: return (0);
1.5 millert 110: }
1.1 etheisen 111:
112: /*
1.10 shadchin 113: * Is a character ASCII?
114: */
1.13 nicm 115: int
116: is_ascii_char(LWCHAR ch)
1.10 shadchin 117: {
118: return (ch <= 0x7F);
119: }
120:
121: /*
1.1 etheisen 122: * Rewind the line buffer.
123: */
1.13 nicm 124: void
125: prewind(void)
1.1 etheisen 126: {
127: curr = 0;
128: column = 0;
1.10 shadchin 129: cshift = 0;
1.1 etheisen 130: overstrike = 0;
1.10 shadchin 131: mbc_buf_len = 0;
1.1 etheisen 132: is_null_line = 0;
133: pendc = '\0';
1.5 millert 134: lmargin = 0;
135: if (status_col)
136: lmargin += 1;
1.1 etheisen 137: }
138:
139: /*
140: * Insert the line number (of the given position) into the line buffer.
141: */
1.13 nicm 142: void
143: plinenum(off_t pos)
1.1 etheisen 144: {
1.16 mmcc 145: off_t linenum = 0;
1.13 nicm 146: int i;
1.5 millert 147:
1.13 nicm 148: if (linenums == OPT_ONPLUS) {
1.5 millert 149: /*
150: * Get the line number and put it in the current line.
151: * {{ Note: since find_linenum calls forw_raw_line,
1.13 nicm 152: * it may seek in the input file, requiring the caller
1.5 millert 153: * of plinenum to re-seek if necessary. }}
154: * {{ Since forw_raw_line modifies linebuf, we must
155: * do this first, before storing anything in linebuf. }}
156: */
157: linenum = find_linenum(pos);
158: }
1.1 etheisen 159:
160: /*
1.5 millert 161: * Display a status column if the -J option is set.
1.1 etheisen 162: */
1.13 nicm 163: if (status_col) {
1.5 millert 164: linebuf[curr] = ' ';
1.13 nicm 165: if (start_attnpos != -1 &&
1.5 millert 166: pos >= start_attnpos && pos < end_attnpos)
1.10 shadchin 167: attr[curr] = AT_NORMAL|AT_HILITE;
1.5 millert 168: else
1.10 shadchin 169: attr[curr] = AT_NORMAL;
1.5 millert 170: curr++;
171: column++;
172: }
1.1 etheisen 173: /*
1.5 millert 174: * Display the line number at the start of each line
175: * if the -N option is set.
1.1 etheisen 176: */
1.13 nicm 177: if (linenums == OPT_ONPLUS) {
1.18 mmcc 178: char buf[23];
1.5 millert 179: int n;
1.1 etheisen 180:
1.17 mmcc 181: postoa(linenum, buf, sizeof(buf));
1.5 millert 182: n = strlen(buf);
183: if (n < MIN_LINENUM_WIDTH)
184: n = MIN_LINENUM_WIDTH;
1.6 millert 185: snprintf(linebuf+curr, size_linebuf-curr, "%*s ", n, buf);
1.21 deraadt 186: n++; /* One space after the line number. */
1.5 millert 187: for (i = 0; i < n; i++)
188: attr[curr+i] = AT_NORMAL;
189: curr += n;
190: column += n;
191: lmargin += n;
192: }
1.1 etheisen 193:
194: /*
1.5 millert 195: * Append enough spaces to bring us to the lmargin.
1.1 etheisen 196: */
1.13 nicm 197: while (column < lmargin) {
1.1 etheisen 198: linebuf[curr] = ' ';
199: attr[curr++] = AT_NORMAL;
200: column++;
1.5 millert 201: }
202: }
203:
204: /*
1.10 shadchin 205: * Shift the input line left.
206: * This means discarding N printable chars at the start of the buffer.
1.5 millert 207: */
1.13 nicm 208: static void
209: pshift(int shift)
1.10 shadchin 210: {
211: LWCHAR prev_ch = 0;
212: unsigned char c;
213: int shifted = 0;
214: int to;
215: int from;
1.5 millert 216: int len;
1.10 shadchin 217: int width;
218: int prev_attr;
219: int next_attr;
220:
221: if (shift > column - lmargin)
222: shift = column - lmargin;
223: if (shift > curr - lmargin)
224: shift = curr - lmargin;
1.5 millert 225:
1.10 shadchin 226: to = from = lmargin;
1.5 millert 227: /*
1.10 shadchin 228: * We keep on going when shifted == shift
229: * to get all combining chars.
1.5 millert 230: */
1.13 nicm 231: while (shifted <= shift && from < curr) {
1.10 shadchin 232: c = linebuf[from];
1.13 nicm 233: if (ctldisp == OPT_ONPLUS && IS_CSI_START(c)) {
1.10 shadchin 234: /* Keep cumulative effect. */
235: linebuf[to] = c;
236: attr[to++] = attr[from++];
1.13 nicm 237: while (from < curr && linebuf[from]) {
1.10 shadchin 238: linebuf[to] = linebuf[from];
239: attr[to++] = attr[from];
240: if (!is_ansi_middle(linebuf[from++]))
241: break;
1.13 nicm 242: }
1.10 shadchin 243: continue;
244: }
245:
246: width = 0;
247:
1.13 nicm 248: if (!IS_ASCII_OCTET(c) && utf_mode) {
1.10 shadchin 249: /* Assumes well-formedness validation already done. */
250: LWCHAR ch;
251:
252: len = utf_len(c);
253: if (from + len > curr)
254: break;
255: ch = get_wchar(linebuf + from);
1.13 nicm 256: if (!is_composing_char(ch) &&
257: !is_combining_char(prev_ch, ch))
1.10 shadchin 258: width = is_wide_char(ch) ? 2 : 1;
259: prev_ch = ch;
1.13 nicm 260: } else {
1.10 shadchin 261: len = 1;
262: if (c == '\b')
263: /* XXX - Incorrect if several '\b' in a row. */
1.13 nicm 264: width = (utf_mode && is_wide_char(prev_ch)) ?
265: -2 : -1;
1.10 shadchin 266: else if (!control_char(c))
267: width = 1;
268: prev_ch = 0;
269: }
270:
271: if (width == 2 && shift - shifted == 1) {
272: /* Should never happen when called by pshift_all(). */
273: attr[to] = attr[from];
274: /*
275: * Assume a wide_char will never be the first half of a
276: * combining_char pair, so reset prev_ch in case we're
277: * followed by a '\b'.
278: */
279: prev_ch = linebuf[to++] = ' ';
280: from += len;
281: shifted++;
282: continue;
283: }
284:
285: /* Adjust width for magic cookies. */
286: prev_attr = (to > 0) ? attr[to-1] : AT_NORMAL;
287: next_attr = (from + len < curr) ? attr[from + len] : prev_attr;
1.13 nicm 288: if (!is_at_equiv(attr[from], prev_attr) &&
289: !is_at_equiv(attr[from], next_attr)) {
1.10 shadchin 290: width += attr_swidth(attr[from]);
291: if (from + len < curr)
292: width += attr_ewidth(attr[from]);
1.13 nicm 293: if (is_at_equiv(prev_attr, next_attr)) {
1.10 shadchin 294: width += attr_ewidth(prev_attr);
295: if (from + len < curr)
296: width += attr_swidth(next_attr);
1.5 millert 297: }
298: }
299:
1.10 shadchin 300: if (shift - shifted < width)
301: break;
302: from += len;
303: shifted += width;
304: if (shifted < 0)
305: shifted = 0;
306: }
1.13 nicm 307: while (from < curr) {
1.10 shadchin 308: linebuf[to] = linebuf[from];
309: attr[to++] = attr[from++];
310: }
311: curr = to;
312: column -= shifted;
313: cshift += shifted;
1.5 millert 314: }
315:
316: /*
1.10 shadchin 317: *
1.5 millert 318: */
1.13 nicm 319: void
320: pshift_all(void)
1.5 millert 321: {
1.10 shadchin 322: pshift(column);
1.1 etheisen 323: }
324:
325: /*
326: * Return the printing width of the start (enter) sequence
327: * for a given character attribute.
328: */
1.13 nicm 329: static int
330: attr_swidth(int a)
1.1 etheisen 331: {
1.10 shadchin 332: int w = 0;
333:
334: a = apply_at_specials(a);
335:
336: if (a & AT_UNDERLINE)
337: w += ul_s_width;
338: if (a & AT_BOLD)
339: w += bo_s_width;
340: if (a & AT_BLINK)
341: w += bl_s_width;
342: if (a & AT_STANDOUT)
343: w += so_s_width;
344:
1.13 nicm 345: return (w);
1.1 etheisen 346: }
347:
348: /*
349: * Return the printing width of the end (exit) sequence
350: * for a given character attribute.
351: */
1.13 nicm 352: static int
353: attr_ewidth(int a)
1.1 etheisen 354: {
1.10 shadchin 355: int w = 0;
356:
357: a = apply_at_specials(a);
358:
359: if (a & AT_UNDERLINE)
360: w += ul_e_width;
361: if (a & AT_BOLD)
362: w += bo_e_width;
363: if (a & AT_BLINK)
364: w += bl_e_width;
365: if (a & AT_STANDOUT)
366: w += so_e_width;
367:
1.13 nicm 368: return (w);
1.1 etheisen 369: }
370:
371: /*
372: * Return the printing width of a given character and attribute,
373: * if the character were added to the current position in the line buffer.
374: * Adding a character with a given attribute may cause an enter or exit
375: * attribute sequence to be inserted, so this must be taken into account.
376: */
1.13 nicm 377: static int
1.23 ! schwarze 378: pwidth(wchar_t ch, int a, wchar_t prev_ch)
1.1 etheisen 379: {
1.10 shadchin 380: int w;
1.1 etheisen 381:
1.23 ! schwarze 382: /*
! 383: * In case of a backspace, back up by the width of the previous
! 384: * character. If that is non-printable (for example another
! 385: * backspace) or zero width (for example a combining accent),
! 386: * the terminal may actually back up to a character even further
! 387: * back, but we no longer know how wide that may have been.
! 388: * The best guess possible at this point is that it was
! 389: * hopefully width one.
! 390: */
! 391: if (ch == L'\b') {
! 392: w = wcwidth(prev_ch);
! 393: if (w <= 0)
! 394: w = 1;
! 395: return (-w);
! 396: }
! 397:
! 398: w = wcwidth(ch);
! 399:
! 400: /*
! 401: * Non-printable characters can get here if the -r flag is in
! 402: * effect, and possibly in some other situations (XXX check that!).
! 403: * Treat them as zero width.
! 404: * That may not always match their actual behaviour,
! 405: * but there is no reasonable way to be more exact.
! 406: */
! 407: if (w == -1)
! 408: w = 0;
1.13 nicm 409:
1.23 ! schwarze 410: /*
! 411: * Combining accents take up no space.
! 412: * Some terminals, upon failure to compose them with the
! 413: * characters that precede them, will actually take up one column
! 414: * for the combining accent; there isn't much we could do short
! 415: * of testing the (complex) composition process ourselves and
! 416: * printing a binary representation when it fails.
! 417: */
! 418: if (w == 0)
! 419: return (0);
1.1 etheisen 420:
421: /*
1.10 shadchin 422: * Other characters take one or two columns,
1.1 etheisen 423: * plus the width of any attribute enter/exit sequence.
424: */
1.10 shadchin 425: if (curr > 0 && !is_at_equiv(attr[curr-1], a))
1.1 etheisen 426: w += attr_ewidth(attr[curr-1]);
1.10 shadchin 427: if ((apply_at_specials(a) != AT_NORMAL) &&
428: (curr == 0 || !is_at_equiv(attr[curr-1], a)))
1.1 etheisen 429: w += attr_swidth(a);
430: return (w);
431: }
432:
433: /*
1.10 shadchin 434: * Delete to the previous base character in the line buffer.
435: * Return 1 if one is found.
1.1 etheisen 436: */
1.13 nicm 437: static int
438: backc(void)
1.1 etheisen 439: {
1.10 shadchin 440: LWCHAR prev_ch;
441: char *p = linebuf + curr;
442: LWCHAR ch = step_char(&p, -1, linebuf + lmargin);
443: int width;
444:
445: /* This assumes that there is no '\b' in linebuf. */
1.13 nicm 446: while (curr > lmargin && column > lmargin &&
447: (!(attr[curr - 1] & (AT_ANSI|AT_BINARY)))) {
1.10 shadchin 448: curr = p - linebuf;
449: prev_ch = step_char(&p, -1, linebuf + lmargin);
450: width = pwidth(ch, attr[curr], prev_ch);
451: column -= width;
452: if (width > 0)
1.13 nicm 453: return (1);
1.10 shadchin 454: ch = prev_ch;
455: }
456:
1.13 nicm 457: return (0);
1.1 etheisen 458: }
459:
460: /*
1.5 millert 461: * Are we currently within a recognized ANSI escape sequence?
462: */
1.13 nicm 463: static int
464: in_ansi_esc_seq(void)
1.5 millert 465: {
1.10 shadchin 466: char *p;
1.5 millert 467:
468: /*
469: * Search backwards for either an ESC (which means we ARE in a seq);
470: * or an end char (which means we're NOT in a seq).
471: */
1.21 deraadt 472: for (p = &linebuf[curr]; p > linebuf; ) {
1.10 shadchin 473: LWCHAR ch = step_char(&p, -1, linebuf);
474: if (IS_CSI_START(ch))
1.5 millert 475: return (1);
1.10 shadchin 476: if (!is_ansi_middle(ch))
1.5 millert 477: return (0);
478: }
479: return (0);
480: }
481:
482: /*
483: * Is a character the end of an ANSI escape sequence?
484: */
1.13 nicm 485: int
486: is_ansi_end(LWCHAR ch)
1.10 shadchin 487: {
488: if (!is_ascii_char(ch))
489: return (0);
1.13 nicm 490: return (strchr(end_ansi_chars, (char)ch) != NULL);
1.10 shadchin 491: }
492:
493: /*
494: *
495: */
1.13 nicm 496: int
497: is_ansi_middle(LWCHAR ch)
1.5 millert 498: {
1.10 shadchin 499: if (!is_ascii_char(ch))
500: return (0);
501: if (is_ansi_end(ch))
502: return (0);
1.13 nicm 503: return (strchr(mid_ansi_chars, (char)ch) != NULL);
1.5 millert 504: }
505:
506: /*
1.1 etheisen 507: * Append a character and attribute to the line buffer.
508: */
1.13 nicm 509: #define STORE_CHAR(ch, a, rep, pos) \
510: if (store_char((ch), (a), (rep), (pos))) \
511: return (1)
512:
513: static int
514: store_char(LWCHAR ch, char a, char *rep, off_t pos)
1.1 etheisen 515: {
1.10 shadchin 516: int w;
517: int replen;
518: char cs;
1.13 nicm 519: int matches;
1.10 shadchin 520:
1.13 nicm 521: if (is_hilited(pos, pos+1, 0, &matches)) {
522: /*
523: * This character should be highlighted.
524: * Override the attribute passed in.
525: */
526: if (a != AT_ANSI) {
527: if (highest_hilite != -1 && pos > highest_hilite)
528: highest_hilite = pos;
529: a |= AT_HILITE;
1.10 shadchin 530: }
1.5 millert 531: }
1.10 shadchin 532:
1.13 nicm 533: if (ctldisp == OPT_ONPLUS && in_ansi_esc_seq()) {
1.10 shadchin 534: if (!is_ansi_end(ch) && !is_ansi_middle(ch)) {
535: /* Remove whole unrecognized sequence. */
536: char *p = &linebuf[curr];
537: LWCHAR bch;
538: do {
539: bch = step_char(&p, -1, linebuf);
540: } while (p > linebuf && !IS_CSI_START(bch));
541: curr = p - linebuf;
1.13 nicm 542: return (0);
1.10 shadchin 543: }
544: a = AT_ANSI; /* Will force re-AT_'ing around it. */
1.5 millert 545: w = 0;
1.13 nicm 546: } else if (ctldisp == OPT_ONPLUS && IS_CSI_START(ch)) {
1.10 shadchin 547: a = AT_ANSI; /* Will force re-AT_'ing around it. */
548: w = 0;
1.13 nicm 549: } else {
1.10 shadchin 550: char *p = &linebuf[curr];
551: LWCHAR prev_ch = step_char(&p, -1, linebuf);
552: w = pwidth(ch, a, prev_ch);
553: }
554:
1.5 millert 555: if (ctldisp != OPT_ON && column + w + attr_ewidth(a) > sc_width)
1.1 etheisen 556: /*
557: * Won't fit on screen.
558: */
559: return (1);
560:
1.13 nicm 561: if (rep == NULL) {
562: cs = (char)ch;
1.10 shadchin 563: rep = &cs;
564: replen = 1;
1.13 nicm 565: } else {
1.10 shadchin 566: replen = utf_len(rep[0]);
567: }
1.13 nicm 568: if (curr + replen >= size_linebuf-6) {
1.1 etheisen 569: /*
570: * Won't fit in line buffer.
1.5 millert 571: * Try to expand it.
1.1 etheisen 572: */
1.5 millert 573: if (expand_linebuf())
574: return (1);
575: }
1.1 etheisen 576:
1.13 nicm 577: while (replen-- > 0) {
1.10 shadchin 578: linebuf[curr] = *rep++;
579: attr[curr] = a;
580: curr++;
1.1 etheisen 581: }
582: column += w;
583: return (0);
584: }
585:
586: /*
1.5 millert 587: * Append a tab to the line buffer.
588: * Store spaces to represent the tab.
589: */
1.13 nicm 590: #define STORE_TAB(a, pos) \
591: if (store_tab((a), (pos))) \
592: return (1)
1.5 millert 593:
1.13 nicm 594: static int
595: store_tab(int attr, off_t pos)
1.5 millert 596: {
597: int to_tab = column + cshift - lmargin;
598: int i;
599:
600: if (ntabstops < 2 || to_tab >= tabstops[ntabstops-1])
601: to_tab = tabdefault -
1.13 nicm 602: ((to_tab - tabstops[ntabstops-1]) % tabdefault);
603: else {
1.21 deraadt 604: for (i = ntabstops - 2; i >= 0; i--)
1.5 millert 605: if (to_tab >= tabstops[i])
606: break;
607: to_tab = tabstops[i+1] - to_tab;
608: }
609:
1.13 nicm 610: if (column + to_tab - 1 + pwidth(' ', attr, 0) +
611: attr_ewidth(attr) > sc_width)
612: return (1);
1.10 shadchin 613:
1.5 millert 614: do {
1.10 shadchin 615: STORE_CHAR(' ', attr, " ", pos);
1.5 millert 616: } while (--to_tab > 0);
1.13 nicm 617: return (0);
1.5 millert 618: }
619:
1.13 nicm 620: #define STORE_PRCHAR(c, pos) \
621: if (store_prchar((c), (pos))) \
622: return (1)
1.10 shadchin 623:
1.13 nicm 624: static int
625: store_prchar(char c, off_t pos)
1.10 shadchin 626: {
627: char *s;
628:
629: /*
630: * Convert to printable representation.
631: */
632: s = prchar(c);
633:
634: /*
635: * Make sure we can get the entire representation
636: * of the character on this line.
637: */
1.13 nicm 638: if (column + (int)strlen(s) - 1 +
639: pwidth(' ', binattr, 0) + attr_ewidth(binattr) > sc_width)
640: return (1);
1.10 shadchin 641:
1.13 nicm 642: for (; *s != 0; s++) {
1.10 shadchin 643: STORE_CHAR(*s, AT_BINARY, NULL, pos);
1.13 nicm 644: }
645: return (0);
1.10 shadchin 646: }
647:
1.13 nicm 648: static int
649: flush_mbc_buf(off_t pos)
1.10 shadchin 650: {
651: int i;
652:
653: for (i = 0; i < mbc_buf_index; i++)
654: if (store_prchar(mbc_buf[i], pos))
1.13 nicm 655: return (mbc_buf_index - i);
1.10 shadchin 656:
1.13 nicm 657: return (0);
1.10 shadchin 658: }
659:
1.5 millert 660: /*
1.1 etheisen 661: * Append a character to the line buffer.
662: * Expand tabs into spaces, handle underlining, boldfacing, etc.
663: * Returns 0 if ok, 1 if couldn't fit in buffer.
664: */
1.13 nicm 665: int
666: pappend(char c, off_t pos)
1.1 etheisen 667: {
1.5 millert 668: int r;
669:
1.13 nicm 670: if (pendc) {
1.10 shadchin 671: if (do_append(pendc, NULL, pendpos))
1.1 etheisen 672: /*
673: * Oops. We've probably lost the char which
674: * was in pendc, since caller won't back up.
675: */
676: return (1);
677: pendc = '\0';
678: }
679:
1.13 nicm 680: if (c == '\r' && bs_mode == BS_SPECIAL) {
681: if (mbc_buf_len > 0) /* utf_mode must be on. */ {
1.10 shadchin 682: /* Flush incomplete (truncated) sequence. */
683: r = flush_mbc_buf(mbc_pos);
684: mbc_buf_index = r + 1;
685: mbc_buf_len = 0;
686: if (r)
687: return (mbc_buf_index);
688: }
689:
1.1 etheisen 690: /*
1.13 nicm 691: * Don't put the CR into the buffer until we see
1.1 etheisen 692: * the next char. If the next char is a newline,
693: * discard the CR.
694: */
695: pendc = c;
696: pendpos = pos;
697: return (0);
698: }
699:
1.13 nicm 700: if (!utf_mode) {
1.10 shadchin 701: r = do_append((LWCHAR) c, NULL, pos);
1.13 nicm 702: } else {
1.10 shadchin 703: /* Perform strict validation in all possible cases. */
1.13 nicm 704: if (mbc_buf_len == 0) {
705: retry:
1.10 shadchin 706: mbc_buf_index = 1;
707: *mbc_buf = c;
1.13 nicm 708: if (IS_ASCII_OCTET(c)) {
1.10 shadchin 709: r = do_append((LWCHAR) c, NULL, pos);
1.13 nicm 710: } else if (IS_UTF8_LEAD(c)) {
1.10 shadchin 711: mbc_buf_len = utf_len(c);
712: mbc_pos = pos;
713: return (0);
1.13 nicm 714: } else {
1.10 shadchin 715: /* UTF8_INVALID or stray UTF8_TRAIL */
716: r = flush_mbc_buf(pos);
1.13 nicm 717: }
718: } else if (IS_UTF8_TRAIL(c)) {
1.10 shadchin 719: mbc_buf[mbc_buf_index++] = c;
720: if (mbc_buf_index < mbc_buf_len)
721: return (0);
722: if (is_utf8_well_formed(mbc_buf))
1.13 nicm 723: r = do_append(get_wchar(mbc_buf), mbc_buf,
724: mbc_pos);
1.10 shadchin 725: else
726: /* Complete, but not shortest form, sequence. */
727: mbc_buf_index = r = flush_mbc_buf(mbc_pos);
728: mbc_buf_len = 0;
1.13 nicm 729: } else {
1.10 shadchin 730: /* Flush incomplete (truncated) sequence. */
731: r = flush_mbc_buf(mbc_pos);
732: mbc_buf_index = r + 1;
733: mbc_buf_len = 0;
734: /* Handle new char. */
735: if (!r)
736: goto retry;
1.13 nicm 737: }
1.10 shadchin 738: }
739:
1.5 millert 740: /*
741: * If we need to shift the line, do it.
742: * But wait until we get to at least the middle of the screen,
743: * so shifting it doesn't affect the chars we're currently
744: * pappending. (Bold & underline can get messed up otherwise.)
745: */
1.13 nicm 746: if (cshift < hshift && column > sc_width / 2) {
1.5 millert 747: linebuf[curr] = '\0';
748: pshift(hshift - cshift);
749: }
1.13 nicm 750: if (r) {
1.10 shadchin 751: /* How many chars should caller back up? */
752: r = (!utf_mode) ? 1 : mbc_buf_index;
753: }
1.5 millert 754: return (r);
1.1 etheisen 755: }
756:
1.13 nicm 757: static int
758: do_append(LWCHAR ch, char *rep, off_t pos)
1.1 etheisen 759: {
1.13 nicm 760: int a;
1.10 shadchin 761: LWCHAR prev_ch;
1.1 etheisen 762:
1.10 shadchin 763: a = AT_NORMAL;
1.1 etheisen 764:
1.13 nicm 765: if (ch == '\b') {
1.10 shadchin 766: if (bs_mode == BS_CONTROL)
1.5 millert 767: goto do_control_char;
1.10 shadchin 768:
769: /*
770: * A better test is needed here so we don't
771: * backspace over part of the printed
772: * representation of a binary character.
773: */
1.13 nicm 774: if (curr <= lmargin ||
775: column <= lmargin ||
776: (attr[curr - 1] & (AT_ANSI|AT_BINARY))) {
1.10 shadchin 777: STORE_PRCHAR('\b', pos);
1.13 nicm 778: } else if (bs_mode == BS_NORMAL) {
1.10 shadchin 779: STORE_CHAR(ch, AT_NORMAL, NULL, pos);
1.13 nicm 780: } else if (bs_mode == BS_SPECIAL) {
1.10 shadchin 781: overstrike = backc();
1.13 nicm 782: }
1.10 shadchin 783:
1.13 nicm 784: return (0);
1.10 shadchin 785: }
786:
1.13 nicm 787: if (overstrike > 0) {
1.1 etheisen 788: /*
789: * Overstrike the character at the current position
1.13 nicm 790: * in the line buffer. This will cause either
791: * underline (if a "_" is overstruck),
1.1 etheisen 792: * bold (if an identical character is overstruck),
793: * or just deletion of the character in the buffer.
794: */
1.10 shadchin 795: overstrike = utf_mode ? -1 : 0;
796: /* To be correct, this must be a base character. */
797: prev_ch = get_wchar(linebuf + curr);
798: a = attr[curr];
1.13 nicm 799: if (ch == prev_ch) {
1.5 millert 800: /*
801: * Overstriking a char with itself means make it bold.
802: * But overstriking an underscore with itself is
803: * ambiguous. It could mean make it bold, or
804: * it could mean make it underlined.
805: * Use the previous overstrike to resolve it.
806: */
1.13 nicm 807: if (ch == '_') {
1.10 shadchin 808: if ((a & (AT_BOLD|AT_UNDERLINE)) != AT_NORMAL)
809: a |= (AT_BOLD|AT_UNDERLINE);
1.20 natano 810: else if (curr > 0 && attr[curr - 1] & AT_UNDERLINE)
811: a |= AT_UNDERLINE;
812: else if (curr > 0 && attr[curr - 1] & AT_BOLD)
813: a |= AT_BOLD;
1.10 shadchin 814: else
1.20 natano 815: a |= AT_INDET;
1.13 nicm 816: } else {
1.10 shadchin 817: a |= AT_BOLD;
1.13 nicm 818: }
819: } else if (ch == '_') {
1.10 shadchin 820: a |= AT_UNDERLINE;
821: ch = prev_ch;
822: rep = linebuf + curr;
1.13 nicm 823: } else if (prev_ch == '_') {
1.10 shadchin 824: a |= AT_UNDERLINE;
825: }
826: /* Else we replace prev_ch, but we keep its attributes. */
1.13 nicm 827: } else if (overstrike < 0) {
828: if (is_composing_char(ch) ||
1.20 natano 829: is_combining_char(get_wchar(linebuf + curr), ch)) {
1.10 shadchin 830: /* Continuation of the same overstrike. */
1.20 natano 831: if (curr > 0)
832: a = attr[curr - 1] & (AT_UNDERLINE | AT_BOLD);
833: else
834: a = AT_NORMAL;
835: } else
1.10 shadchin 836: overstrike = 0;
837: }
838:
1.13 nicm 839: if (ch == '\t') {
1.5 millert 840: /*
841: * Expand a tab into spaces.
842: */
1.13 nicm 843: switch (bs_mode) {
1.1 etheisen 844: case BS_CONTROL:
845: goto do_control_char;
1.5 millert 846: case BS_NORMAL:
1.1 etheisen 847: case BS_SPECIAL:
1.10 shadchin 848: STORE_TAB(a, pos);
1.1 etheisen 849: break;
850: }
1.13 nicm 851: } else if ((!utf_mode || is_ascii_char(ch)) && control_char((char)ch)) {
852: do_control_char:
853: if (ctldisp == OPT_ON ||
854: (ctldisp == OPT_ONPLUS && IS_CSI_START(ch))) {
1.1 etheisen 855: /*
856: * Output as a normal character.
857: */
1.10 shadchin 858: STORE_CHAR(ch, AT_NORMAL, rep, pos);
1.13 nicm 859: } else {
860: STORE_PRCHAR((char)ch, pos);
1.10 shadchin 861: }
1.13 nicm 862: } else if (utf_mode && ctldisp != OPT_ON && is_ubin_char(ch)) {
1.10 shadchin 863: char *s;
864:
865: s = prutfchar(ch);
1.1 etheisen 866:
1.13 nicm 867: if (column + (int)strlen(s) - 1 +
1.10 shadchin 868: pwidth(' ', binattr, 0) + attr_ewidth(binattr) > sc_width)
869: return (1);
1.1 etheisen 870:
1.13 nicm 871: for (; *s != 0; s++)
1.10 shadchin 872: STORE_CHAR(*s, AT_BINARY, NULL, pos);
1.13 nicm 873: } else {
1.10 shadchin 874: STORE_CHAR(ch, a, rep, pos);
1.1 etheisen 875: }
1.13 nicm 876: return (0);
1.10 shadchin 877: }
878:
879: /*
880: *
881: */
1.13 nicm 882: int
883: pflushmbc(void)
1.10 shadchin 884: {
885: int r = 0;
1.1 etheisen 886:
1.13 nicm 887: if (mbc_buf_len > 0) {
1.10 shadchin 888: /* Flush incomplete (truncated) sequence. */
889: r = flush_mbc_buf(mbc_pos);
890: mbc_buf_len = 0;
891: }
1.13 nicm 892: return (r);
1.1 etheisen 893: }
894:
895: /*
896: * Terminate the line in the line buffer.
897: */
1.13 nicm 898: void
899: pdone(int endline, int forw)
1.1 etheisen 900: {
1.20 natano 901: int i;
902:
1.10 shadchin 903: (void) pflushmbc();
904:
1.1 etheisen 905: if (pendc && (pendc != '\r' || !endline))
906: /*
907: * If we had a pending character, put it in the buffer.
908: * But discard a pending CR if we are at end of line
909: * (that is, discard the CR in a CR/LF sequence).
910: */
1.10 shadchin 911: (void) do_append(pendc, NULL, pendpos);
1.20 natano 912:
913: for (i = curr - 1; i >= 0; i--) {
914: if (attr[i] & AT_INDET) {
915: attr[i] &= ~AT_INDET;
916: if (i < curr - 1 && attr[i + 1] & AT_BOLD)
917: attr[i] |= AT_BOLD;
918: else
919: attr[i] |= AT_UNDERLINE;
920: }
921: }
1.1 etheisen 922:
923: /*
1.5 millert 924: * Make sure we've shifted the line, if we need to.
925: */
926: if (cshift < hshift)
927: pshift(hshift - cshift);
928:
1.13 nicm 929: if (ctldisp == OPT_ONPLUS && is_ansi_end('m')) {
1.10 shadchin 930: /* Switch to normal attribute at end of line. */
931: char *p = "\033[m";
1.13 nicm 932: for (; *p != '\0'; p++) {
1.10 shadchin 933: linebuf[curr] = *p;
934: attr[curr++] = AT_ANSI;
935: }
936: }
937:
1.5 millert 938: /*
1.1 etheisen 939: * Add a newline if necessary,
940: * and append a '\0' to the end of the line.
1.10 shadchin 941: * We output a newline if we're not at the right edge of the screen,
942: * or if the terminal doesn't auto wrap,
943: * or if this is really the end of the line AND the terminal ignores
944: * a newline at the right edge.
1.13 nicm 945: * (In the last case we don't want to output a newline if the terminal
1.10 shadchin 946: * doesn't ignore it since that would produce an extra blank line.
947: * But we do want to output a newline if the terminal ignores it in case
948: * the next line is blank. In that case the single newline output for
949: * that blank line would be ignored!)
1.1 etheisen 950: */
1.13 nicm 951: if (column < sc_width || !auto_wrap || (endline && ignaw) ||
952: ctldisp == OPT_ON) {
1.1 etheisen 953: linebuf[curr] = '\n';
954: attr[curr] = AT_NORMAL;
955: curr++;
1.13 nicm 956: } else if (ignaw && column >= sc_width && forw) {
1.10 shadchin 957: /*
958: * Terminals with "ignaw" don't wrap until they *really* need
959: * to, i.e. when the character *after* the last one to fit on a
960: * line is output. But they are too hard to deal with when they
961: * get in the state where a full screen width of characters
962: * have been output but the cursor is sitting on the right edge
963: * instead of at the start of the next line.
1.13 nicm 964: * So we nudge them into wrapping by outputting a space
965: * character plus a backspace. But do this only if moving
1.10 shadchin 966: * forward; if we're moving backward and drawing this line at
967: * the top of the screen, the space would overwrite the first
1.13 nicm 968: * char on the next line. We don't need to do this "nudge"
1.10 shadchin 969: * at the top of the screen anyway.
970: */
971: linebuf[curr] = ' ';
972: attr[curr++] = AT_NORMAL;
1.13 nicm 973: linebuf[curr] = '\b';
1.10 shadchin 974: attr[curr++] = AT_NORMAL;
1.1 etheisen 975: }
976: linebuf[curr] = '\0';
977: attr[curr] = AT_NORMAL;
1.10 shadchin 978: }
1.5 millert 979:
1.10 shadchin 980: /*
981: *
982: */
1.13 nicm 983: void
984: set_status_col(char c)
1.10 shadchin 985: {
986: linebuf[0] = c;
987: attr[0] = AT_NORMAL|AT_HILITE;
1.1 etheisen 988: }
989:
990: /*
991: * Get a character from the current line.
992: * Return the character as the function return value,
993: * and the character attribute in *ap.
994: */
1.13 nicm 995: int
996: gline(int i, int *ap)
1.1 etheisen 997: {
1.13 nicm 998: if (is_null_line) {
1.1 etheisen 999: /*
1000: * If there is no current line, we pretend the line is
1001: * either "~" or "", depending on the "twiddle" flag.
1002: */
1.13 nicm 1003: if (twiddle) {
1004: if (i == 0) {
1.10 shadchin 1005: *ap = AT_BOLD;
1.13 nicm 1006: return ('~');
1.10 shadchin 1007: }
1008: --i;
1009: }
1010: /* Make sure we're back to AT_NORMAL before the '\n'. */
1011: *ap = AT_NORMAL;
1.13 nicm 1012: return (i ? '\0' : '\n');
1.1 etheisen 1013: }
1014:
1015: *ap = attr[i];
1.10 shadchin 1016: return (linebuf[i] & 0xFF);
1.1 etheisen 1017: }
1018:
1019: /*
1020: * Indicate that there is no current line.
1021: */
1.13 nicm 1022: void
1023: null_line(void)
1.1 etheisen 1024: {
1025: is_null_line = 1;
1.5 millert 1026: cshift = 0;
1.1 etheisen 1027: }
1028:
1029: /*
1030: * Analogous to forw_line(), but deals with "raw lines":
1031: * lines which are not split for screen width.
1032: * {{ This is supposed to be more efficient than forw_line(). }}
1033: */
1.13 nicm 1034: off_t
1035: forw_raw_line(off_t curr_pos, char **linep, int *line_lenp)
1036: {
1037: int n;
1038: int c;
1039: off_t new_pos;
1040:
1041: if (curr_pos == -1 || ch_seek(curr_pos) ||
1042: (c = ch_forw_get()) == EOI)
1043: return (-1);
1.1 etheisen 1044:
1.5 millert 1045: n = 0;
1.13 nicm 1046: for (;;) {
1047: if (c == '\n' || c == EOI || ABORT_SIGS()) {
1.1 etheisen 1048: new_pos = ch_tell();
1049: break;
1050: }
1.13 nicm 1051: if (n >= size_linebuf-1) {
1052: if (expand_linebuf()) {
1.5 millert 1053: /*
1054: * Overflowed the input buffer.
1055: * Pretend the line ended here.
1056: */
1057: new_pos = ch_tell() - 1;
1058: break;
1059: }
1.1 etheisen 1060: }
1.13 nicm 1061: linebuf[n++] = (char)c;
1.1 etheisen 1062: c = ch_forw_get();
1063: }
1.5 millert 1064: linebuf[n] = '\0';
1.1 etheisen 1065: if (linep != NULL)
1066: *linep = linebuf;
1.10 shadchin 1067: if (line_lenp != NULL)
1068: *line_lenp = n;
1.1 etheisen 1069: return (new_pos);
1070: }
1071:
1072: /*
1073: * Analogous to back_line(), but deals with "raw lines".
1074: * {{ This is supposed to be more efficient than back_line(). }}
1075: */
1.13 nicm 1076: off_t
1077: back_raw_line(off_t curr_pos, char **linep, int *line_lenp)
1078: {
1079: int n;
1080: int c;
1081: off_t new_pos;
1082:
1083: if (curr_pos == -1 || curr_pos <= ch_zero() || ch_seek(curr_pos - 1))
1084: return (-1);
1.1 etheisen 1085:
1.5 millert 1086: n = size_linebuf;
1087: linebuf[--n] = '\0';
1.13 nicm 1088: for (;;) {
1.1 etheisen 1089: c = ch_back_get();
1.13 nicm 1090: if (c == '\n' || ABORT_SIGS()) {
1.1 etheisen 1091: /*
1092: * This is the newline ending the previous line.
1093: * We have hit the beginning of the line.
1094: */
1095: new_pos = ch_tell() + 1;
1096: break;
1097: }
1.13 nicm 1098: if (c == EOI) {
1.1 etheisen 1099: /*
1100: * We have hit the beginning of the file.
1101: * This must be the first line in the file.
1102: * This must, of course, be the beginning of the line.
1103: */
1104: new_pos = ch_zero();
1105: break;
1106: }
1.13 nicm 1107: if (n <= 0) {
1.5 millert 1108: int old_size_linebuf = size_linebuf;
1.13 nicm 1109: if (expand_linebuf()) {
1.5 millert 1110: /*
1111: * Overflowed the input buffer.
1112: * Pretend the line ended here.
1113: */
1114: new_pos = ch_tell() + 1;
1115: break;
1116: }
1.1 etheisen 1117: /*
1.5 millert 1118: * Shift the data to the end of the new linebuf.
1.1 etheisen 1119: */
1.5 millert 1120: n = size_linebuf - old_size_linebuf;
1.8 millert 1121: memmove(linebuf + n, linebuf, old_size_linebuf);
1.1 etheisen 1122: }
1.5 millert 1123: linebuf[--n] = c;
1.1 etheisen 1124: }
1125: if (linep != NULL)
1.5 millert 1126: *linep = &linebuf[n];
1.10 shadchin 1127: if (line_lenp != NULL)
1128: *line_lenp = size_linebuf - 1 - n;
1.1 etheisen 1129: return (new_pos);
1130: }