Annotation of src/usr.bin/less/line.c, Revision 1.27
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: */
1.25 schwarze 115: static int
1.13 nicm 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];
1.24 schwarze 233: if (ctldisp == OPT_ONPLUS && c == ESC) {
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: {
1.27 ! schwarze 440: wchar_t ch, prev_ch;
! 441: int i, len, width;
! 442:
! 443: i = curr - 1;
! 444: if (utf_mode) {
! 445: while (i >= lmargin && IS_UTF8_TRAIL(linebuf[i]))
! 446: i--;
! 447: }
! 448: if (i < lmargin)
! 449: return (0);
! 450: if (utf_mode) {
! 451: len = mbtowc(&ch, linebuf + i, curr - i);
! 452: if (len == -1 || i + len < curr) {
! 453: (void)mbtowc(NULL, NULL, MB_CUR_MAX);
! 454: return (0);
! 455: }
! 456: } else
! 457: ch = linebuf[i];
1.10 shadchin 458:
459: /* This assumes that there is no '\b' in linebuf. */
1.13 nicm 460: while (curr > lmargin && column > lmargin &&
461: (!(attr[curr - 1] & (AT_ANSI|AT_BINARY)))) {
1.27 ! schwarze 462: curr = i--;
! 463: if (utf_mode) {
! 464: while (i >= lmargin && IS_UTF8_TRAIL(linebuf[i]))
! 465: i--;
! 466: }
! 467: if (i < lmargin)
! 468: prev_ch = L'\0';
! 469: else if (utf_mode) {
! 470: len = mbtowc(&prev_ch, linebuf + i, curr - i);
! 471: if (len == -1 || i + len < curr) {
! 472: (void)mbtowc(NULL, NULL, MB_CUR_MAX);
! 473: prev_ch = L'\0';
! 474: }
! 475: } else
! 476: prev_ch = linebuf[i];
1.10 shadchin 477: width = pwidth(ch, attr[curr], prev_ch);
478: column -= width;
479: if (width > 0)
1.13 nicm 480: return (1);
1.27 ! schwarze 481: if (prev_ch == L'\0')
! 482: return (0);
1.10 shadchin 483: ch = prev_ch;
484: }
1.13 nicm 485: return (0);
1.1 etheisen 486: }
487:
488: /*
1.25 schwarze 489: * Is a character the end of an ANSI escape sequence?
1.5 millert 490: */
1.13 nicm 491: static int
492: is_ansi_end(LWCHAR ch)
1.10 shadchin 493: {
494: if (!is_ascii_char(ch))
495: return (0);
1.13 nicm 496: return (strchr(end_ansi_chars, (char)ch) != NULL);
1.10 shadchin 497: }
498:
499: /*
500: *
501: */
1.13 nicm 502: int
503: is_ansi_middle(LWCHAR ch)
1.5 millert 504: {
1.10 shadchin 505: if (!is_ascii_char(ch))
506: return (0);
507: if (is_ansi_end(ch))
508: return (0);
1.13 nicm 509: return (strchr(mid_ansi_chars, (char)ch) != NULL);
1.5 millert 510: }
511:
512: /*
1.1 etheisen 513: * Append a character and attribute to the line buffer.
514: */
1.13 nicm 515: static int
516: store_char(LWCHAR ch, char a, char *rep, off_t pos)
1.1 etheisen 517: {
1.25 schwarze 518: int i;
1.10 shadchin 519: int w;
520: int replen;
521: char cs;
1.13 nicm 522: int matches;
1.10 shadchin 523:
1.13 nicm 524: if (is_hilited(pos, pos+1, 0, &matches)) {
525: /*
526: * This character should be highlighted.
527: * Override the attribute passed in.
528: */
529: if (a != AT_ANSI) {
530: if (highest_hilite != -1 && pos > highest_hilite)
531: highest_hilite = pos;
532: a |= AT_HILITE;
1.10 shadchin 533: }
1.5 millert 534: }
1.10 shadchin 535:
1.25 schwarze 536: w = -1;
537: if (ctldisp == OPT_ONPLUS) {
538: /*
539: * Set i to the beginning of an ANSI escape sequence
540: * that was begun and not yet ended, or to -1 otherwise.
541: */
542: for (i = curr - 1; i >= 0; i--) {
543: if (linebuf[i] == ESC)
544: break;
545: if (!is_ansi_middle(linebuf[i]))
546: i = 0;
547: }
548: if (i >= 0 && !is_ansi_end(ch) && !is_ansi_middle(ch)) {
1.10 shadchin 549: /* Remove whole unrecognized sequence. */
1.25 schwarze 550: curr = i;
1.13 nicm 551: return (0);
1.10 shadchin 552: }
1.25 schwarze 553: if (i >= 0 || ch == ESC) {
554: a = AT_ANSI; /* Will force re-AT_'ing around it. */
555: w = 0;
556: }
557: }
558: if (w == -1) {
559: wchar_t prev_ch;
560:
561: if (utf_mode) {
562: for (i = curr - 1; i >= 0; i--)
563: if (!IS_UTF8_TRAIL(linebuf[i]))
564: break;
565: if (i >= 0) {
566: w = mbtowc(&prev_ch, linebuf + i, curr - i);
1.27 ! schwarze 567: if (w == -1 || i + w < curr) {
! 568: (void)mbtowc(NULL, NULL, MB_CUR_MAX);
1.25 schwarze 569: prev_ch = L' ';
1.27 ! schwarze 570: }
1.25 schwarze 571: } else
572: prev_ch = L' ';
573: } else
574: prev_ch = curr > 0 ? linebuf[curr - 1] : L' ';
1.10 shadchin 575: w = pwidth(ch, a, prev_ch);
576: }
577:
1.5 millert 578: if (ctldisp != OPT_ON && column + w + attr_ewidth(a) > sc_width)
1.1 etheisen 579: /*
580: * Won't fit on screen.
581: */
582: return (1);
583:
1.13 nicm 584: if (rep == NULL) {
585: cs = (char)ch;
1.10 shadchin 586: rep = &cs;
587: replen = 1;
1.13 nicm 588: } else {
1.10 shadchin 589: replen = utf_len(rep[0]);
590: }
1.13 nicm 591: if (curr + replen >= size_linebuf-6) {
1.1 etheisen 592: /*
593: * Won't fit in line buffer.
1.5 millert 594: * Try to expand it.
1.1 etheisen 595: */
1.5 millert 596: if (expand_linebuf())
597: return (1);
598: }
1.1 etheisen 599:
1.13 nicm 600: while (replen-- > 0) {
1.10 shadchin 601: linebuf[curr] = *rep++;
602: attr[curr] = a;
603: curr++;
1.1 etheisen 604: }
605: column += w;
606: return (0);
607: }
608:
609: /*
1.5 millert 610: * Append a tab to the line buffer.
611: * Store spaces to represent the tab.
612: */
1.13 nicm 613: static int
614: store_tab(int attr, off_t pos)
1.5 millert 615: {
616: int to_tab = column + cshift - lmargin;
617: int i;
618:
619: if (ntabstops < 2 || to_tab >= tabstops[ntabstops-1])
620: to_tab = tabdefault -
1.13 nicm 621: ((to_tab - tabstops[ntabstops-1]) % tabdefault);
622: else {
1.21 deraadt 623: for (i = ntabstops - 2; i >= 0; i--)
1.5 millert 624: if (to_tab >= tabstops[i])
625: break;
626: to_tab = tabstops[i+1] - to_tab;
627: }
628:
1.13 nicm 629: if (column + to_tab - 1 + pwidth(' ', attr, 0) +
630: attr_ewidth(attr) > sc_width)
631: return (1);
1.10 shadchin 632:
1.5 millert 633: do {
1.26 schwarze 634: if (store_char(' ', attr, " ", pos))
635: return (1);
1.5 millert 636: } while (--to_tab > 0);
1.13 nicm 637: return (0);
1.5 millert 638: }
639:
1.13 nicm 640: static int
641: store_prchar(char c, off_t pos)
1.10 shadchin 642: {
643: char *s;
644:
645: /*
646: * Convert to printable representation.
647: */
648: s = prchar(c);
649:
650: /*
651: * Make sure we can get the entire representation
652: * of the character on this line.
653: */
1.13 nicm 654: if (column + (int)strlen(s) - 1 +
655: pwidth(' ', binattr, 0) + attr_ewidth(binattr) > sc_width)
656: return (1);
1.10 shadchin 657:
1.13 nicm 658: for (; *s != 0; s++) {
1.26 schwarze 659: if (store_char(*s, AT_BINARY, NULL, pos))
660: return (1);
1.13 nicm 661: }
662: return (0);
1.10 shadchin 663: }
664:
1.13 nicm 665: static int
666: flush_mbc_buf(off_t pos)
1.10 shadchin 667: {
668: int i;
669:
1.26 schwarze 670: for (i = 0; i < mbc_buf_index; i++) {
1.10 shadchin 671: if (store_prchar(mbc_buf[i], pos))
1.13 nicm 672: return (mbc_buf_index - i);
1.26 schwarze 673: }
1.13 nicm 674: return (0);
1.10 shadchin 675: }
676:
1.5 millert 677: /*
1.1 etheisen 678: * Append a character to the line buffer.
679: * Expand tabs into spaces, handle underlining, boldfacing, etc.
680: * Returns 0 if ok, 1 if couldn't fit in buffer.
681: */
1.13 nicm 682: int
683: pappend(char c, off_t pos)
1.1 etheisen 684: {
1.5 millert 685: int r;
686:
1.13 nicm 687: if (pendc) {
1.10 shadchin 688: if (do_append(pendc, NULL, pendpos))
1.1 etheisen 689: /*
690: * Oops. We've probably lost the char which
691: * was in pendc, since caller won't back up.
692: */
693: return (1);
694: pendc = '\0';
695: }
696:
1.13 nicm 697: if (c == '\r' && bs_mode == BS_SPECIAL) {
698: if (mbc_buf_len > 0) /* utf_mode must be on. */ {
1.10 shadchin 699: /* Flush incomplete (truncated) sequence. */
700: r = flush_mbc_buf(mbc_pos);
701: mbc_buf_index = r + 1;
702: mbc_buf_len = 0;
703: if (r)
704: return (mbc_buf_index);
705: }
706:
1.1 etheisen 707: /*
1.13 nicm 708: * Don't put the CR into the buffer until we see
1.1 etheisen 709: * the next char. If the next char is a newline,
710: * discard the CR.
711: */
712: pendc = c;
713: pendpos = pos;
714: return (0);
715: }
716:
1.13 nicm 717: if (!utf_mode) {
1.10 shadchin 718: r = do_append((LWCHAR) c, NULL, pos);
1.13 nicm 719: } else {
1.10 shadchin 720: /* Perform strict validation in all possible cases. */
1.13 nicm 721: if (mbc_buf_len == 0) {
722: retry:
1.10 shadchin 723: mbc_buf_index = 1;
724: *mbc_buf = c;
1.13 nicm 725: if (IS_ASCII_OCTET(c)) {
1.10 shadchin 726: r = do_append((LWCHAR) c, NULL, pos);
1.13 nicm 727: } else if (IS_UTF8_LEAD(c)) {
1.10 shadchin 728: mbc_buf_len = utf_len(c);
729: mbc_pos = pos;
730: return (0);
1.13 nicm 731: } else {
1.10 shadchin 732: /* UTF8_INVALID or stray UTF8_TRAIL */
733: r = flush_mbc_buf(pos);
1.13 nicm 734: }
735: } else if (IS_UTF8_TRAIL(c)) {
1.10 shadchin 736: mbc_buf[mbc_buf_index++] = c;
737: if (mbc_buf_index < mbc_buf_len)
738: return (0);
739: if (is_utf8_well_formed(mbc_buf))
1.13 nicm 740: r = do_append(get_wchar(mbc_buf), mbc_buf,
741: mbc_pos);
1.10 shadchin 742: else
743: /* Complete, but not shortest form, sequence. */
744: mbc_buf_index = r = flush_mbc_buf(mbc_pos);
745: mbc_buf_len = 0;
1.13 nicm 746: } else {
1.10 shadchin 747: /* Flush incomplete (truncated) sequence. */
748: r = flush_mbc_buf(mbc_pos);
749: mbc_buf_index = r + 1;
750: mbc_buf_len = 0;
751: /* Handle new char. */
752: if (!r)
753: goto retry;
1.13 nicm 754: }
1.10 shadchin 755: }
756:
1.5 millert 757: /*
758: * If we need to shift the line, do it.
759: * But wait until we get to at least the middle of the screen,
760: * so shifting it doesn't affect the chars we're currently
761: * pappending. (Bold & underline can get messed up otherwise.)
762: */
1.13 nicm 763: if (cshift < hshift && column > sc_width / 2) {
1.5 millert 764: linebuf[curr] = '\0';
765: pshift(hshift - cshift);
766: }
1.13 nicm 767: if (r) {
1.10 shadchin 768: /* How many chars should caller back up? */
769: r = (!utf_mode) ? 1 : mbc_buf_index;
770: }
1.5 millert 771: return (r);
1.1 etheisen 772: }
773:
1.13 nicm 774: static int
775: do_append(LWCHAR ch, char *rep, off_t pos)
1.1 etheisen 776: {
1.13 nicm 777: int a;
1.10 shadchin 778: LWCHAR prev_ch;
1.1 etheisen 779:
1.10 shadchin 780: a = AT_NORMAL;
1.1 etheisen 781:
1.13 nicm 782: if (ch == '\b') {
1.10 shadchin 783: if (bs_mode == BS_CONTROL)
1.5 millert 784: goto do_control_char;
1.10 shadchin 785:
786: /*
787: * A better test is needed here so we don't
788: * backspace over part of the printed
789: * representation of a binary character.
790: */
1.13 nicm 791: if (curr <= lmargin ||
792: column <= lmargin ||
793: (attr[curr - 1] & (AT_ANSI|AT_BINARY))) {
1.26 schwarze 794: if (store_prchar('\b', pos))
795: return (1);
1.13 nicm 796: } else if (bs_mode == BS_NORMAL) {
1.26 schwarze 797: if (store_char(ch, AT_NORMAL, NULL, pos))
798: return (1);
1.13 nicm 799: } else if (bs_mode == BS_SPECIAL) {
1.10 shadchin 800: overstrike = backc();
1.13 nicm 801: }
1.10 shadchin 802:
1.13 nicm 803: return (0);
1.10 shadchin 804: }
805:
1.13 nicm 806: if (overstrike > 0) {
1.1 etheisen 807: /*
808: * Overstrike the character at the current position
1.13 nicm 809: * in the line buffer. This will cause either
810: * underline (if a "_" is overstruck),
1.1 etheisen 811: * bold (if an identical character is overstruck),
812: * or just deletion of the character in the buffer.
813: */
1.10 shadchin 814: overstrike = utf_mode ? -1 : 0;
815: /* To be correct, this must be a base character. */
816: prev_ch = get_wchar(linebuf + curr);
817: a = attr[curr];
1.13 nicm 818: if (ch == prev_ch) {
1.5 millert 819: /*
820: * Overstriking a char with itself means make it bold.
821: * But overstriking an underscore with itself is
822: * ambiguous. It could mean make it bold, or
823: * it could mean make it underlined.
824: * Use the previous overstrike to resolve it.
825: */
1.13 nicm 826: if (ch == '_') {
1.10 shadchin 827: if ((a & (AT_BOLD|AT_UNDERLINE)) != AT_NORMAL)
828: a |= (AT_BOLD|AT_UNDERLINE);
1.20 natano 829: else if (curr > 0 && attr[curr - 1] & AT_UNDERLINE)
830: a |= AT_UNDERLINE;
831: else if (curr > 0 && attr[curr - 1] & AT_BOLD)
832: a |= AT_BOLD;
1.10 shadchin 833: else
1.20 natano 834: a |= AT_INDET;
1.13 nicm 835: } else {
1.10 shadchin 836: a |= AT_BOLD;
1.13 nicm 837: }
838: } else if (ch == '_') {
1.10 shadchin 839: a |= AT_UNDERLINE;
840: ch = prev_ch;
841: rep = linebuf + curr;
1.13 nicm 842: } else if (prev_ch == '_') {
1.10 shadchin 843: a |= AT_UNDERLINE;
844: }
845: /* Else we replace prev_ch, but we keep its attributes. */
1.13 nicm 846: } else if (overstrike < 0) {
847: if (is_composing_char(ch) ||
1.20 natano 848: is_combining_char(get_wchar(linebuf + curr), ch)) {
1.10 shadchin 849: /* Continuation of the same overstrike. */
1.20 natano 850: if (curr > 0)
851: a = attr[curr - 1] & (AT_UNDERLINE | AT_BOLD);
852: else
853: a = AT_NORMAL;
854: } else
1.10 shadchin 855: overstrike = 0;
856: }
857:
1.13 nicm 858: if (ch == '\t') {
1.5 millert 859: /*
860: * Expand a tab into spaces.
861: */
1.13 nicm 862: switch (bs_mode) {
1.1 etheisen 863: case BS_CONTROL:
864: goto do_control_char;
1.5 millert 865: case BS_NORMAL:
1.1 etheisen 866: case BS_SPECIAL:
1.26 schwarze 867: if (store_tab(a, pos))
868: return (1);
1.1 etheisen 869: break;
870: }
1.13 nicm 871: } else if ((!utf_mode || is_ascii_char(ch)) && control_char((char)ch)) {
872: do_control_char:
873: if (ctldisp == OPT_ON ||
1.24 schwarze 874: (ctldisp == OPT_ONPLUS && ch == ESC)) {
1.1 etheisen 875: /*
876: * Output as a normal character.
877: */
1.26 schwarze 878: if (store_char(ch, AT_NORMAL, rep, pos))
879: return (1);
1.13 nicm 880: } else {
1.26 schwarze 881: if (store_prchar(ch, pos))
882: return (1);
1.10 shadchin 883: }
1.13 nicm 884: } else if (utf_mode && ctldisp != OPT_ON && is_ubin_char(ch)) {
1.10 shadchin 885: char *s;
886:
887: s = prutfchar(ch);
1.1 etheisen 888:
1.13 nicm 889: if (column + (int)strlen(s) - 1 +
1.10 shadchin 890: pwidth(' ', binattr, 0) + attr_ewidth(binattr) > sc_width)
891: return (1);
1.1 etheisen 892:
1.26 schwarze 893: for (; *s != 0; s++) {
894: if (store_char(*s, AT_BINARY, NULL, pos))
895: return (1);
896: }
1.13 nicm 897: } else {
1.26 schwarze 898: if (store_char(ch, a, rep, pos))
899: return (1);
1.1 etheisen 900: }
1.13 nicm 901: return (0);
1.10 shadchin 902: }
903:
904: /*
905: *
906: */
1.13 nicm 907: int
908: pflushmbc(void)
1.10 shadchin 909: {
910: int r = 0;
1.1 etheisen 911:
1.13 nicm 912: if (mbc_buf_len > 0) {
1.10 shadchin 913: /* Flush incomplete (truncated) sequence. */
914: r = flush_mbc_buf(mbc_pos);
915: mbc_buf_len = 0;
916: }
1.13 nicm 917: return (r);
1.1 etheisen 918: }
919:
920: /*
921: * Terminate the line in the line buffer.
922: */
1.13 nicm 923: void
924: pdone(int endline, int forw)
1.1 etheisen 925: {
1.20 natano 926: int i;
927:
1.10 shadchin 928: (void) pflushmbc();
929:
1.1 etheisen 930: if (pendc && (pendc != '\r' || !endline))
931: /*
932: * If we had a pending character, put it in the buffer.
933: * But discard a pending CR if we are at end of line
934: * (that is, discard the CR in a CR/LF sequence).
935: */
1.10 shadchin 936: (void) do_append(pendc, NULL, pendpos);
1.20 natano 937:
938: for (i = curr - 1; i >= 0; i--) {
939: if (attr[i] & AT_INDET) {
940: attr[i] &= ~AT_INDET;
941: if (i < curr - 1 && attr[i + 1] & AT_BOLD)
942: attr[i] |= AT_BOLD;
943: else
944: attr[i] |= AT_UNDERLINE;
945: }
946: }
1.1 etheisen 947:
948: /*
1.5 millert 949: * Make sure we've shifted the line, if we need to.
950: */
951: if (cshift < hshift)
952: pshift(hshift - cshift);
953:
1.13 nicm 954: if (ctldisp == OPT_ONPLUS && is_ansi_end('m')) {
1.10 shadchin 955: /* Switch to normal attribute at end of line. */
956: char *p = "\033[m";
1.13 nicm 957: for (; *p != '\0'; p++) {
1.10 shadchin 958: linebuf[curr] = *p;
959: attr[curr++] = AT_ANSI;
960: }
961: }
962:
1.5 millert 963: /*
1.1 etheisen 964: * Add a newline if necessary,
965: * and append a '\0' to the end of the line.
1.10 shadchin 966: * We output a newline if we're not at the right edge of the screen,
967: * or if the terminal doesn't auto wrap,
968: * or if this is really the end of the line AND the terminal ignores
969: * a newline at the right edge.
1.13 nicm 970: * (In the last case we don't want to output a newline if the terminal
1.10 shadchin 971: * doesn't ignore it since that would produce an extra blank line.
972: * But we do want to output a newline if the terminal ignores it in case
973: * the next line is blank. In that case the single newline output for
974: * that blank line would be ignored!)
1.1 etheisen 975: */
1.13 nicm 976: if (column < sc_width || !auto_wrap || (endline && ignaw) ||
977: ctldisp == OPT_ON) {
1.1 etheisen 978: linebuf[curr] = '\n';
979: attr[curr] = AT_NORMAL;
980: curr++;
1.13 nicm 981: } else if (ignaw && column >= sc_width && forw) {
1.10 shadchin 982: /*
983: * Terminals with "ignaw" don't wrap until they *really* need
984: * to, i.e. when the character *after* the last one to fit on a
985: * line is output. But they are too hard to deal with when they
986: * get in the state where a full screen width of characters
987: * have been output but the cursor is sitting on the right edge
988: * instead of at the start of the next line.
1.13 nicm 989: * So we nudge them into wrapping by outputting a space
990: * character plus a backspace. But do this only if moving
1.10 shadchin 991: * forward; if we're moving backward and drawing this line at
992: * the top of the screen, the space would overwrite the first
1.13 nicm 993: * char on the next line. We don't need to do this "nudge"
1.10 shadchin 994: * at the top of the screen anyway.
995: */
996: linebuf[curr] = ' ';
997: attr[curr++] = AT_NORMAL;
1.13 nicm 998: linebuf[curr] = '\b';
1.10 shadchin 999: attr[curr++] = AT_NORMAL;
1.1 etheisen 1000: }
1001: linebuf[curr] = '\0';
1002: attr[curr] = AT_NORMAL;
1.10 shadchin 1003: }
1.5 millert 1004:
1.10 shadchin 1005: /*
1006: *
1007: */
1.13 nicm 1008: void
1009: set_status_col(char c)
1.10 shadchin 1010: {
1011: linebuf[0] = c;
1012: attr[0] = AT_NORMAL|AT_HILITE;
1.1 etheisen 1013: }
1014:
1015: /*
1016: * Get a character from the current line.
1017: * Return the character as the function return value,
1018: * and the character attribute in *ap.
1019: */
1.13 nicm 1020: int
1021: gline(int i, int *ap)
1.1 etheisen 1022: {
1.13 nicm 1023: if (is_null_line) {
1.1 etheisen 1024: /*
1025: * If there is no current line, we pretend the line is
1026: * either "~" or "", depending on the "twiddle" flag.
1027: */
1.13 nicm 1028: if (twiddle) {
1029: if (i == 0) {
1.10 shadchin 1030: *ap = AT_BOLD;
1.13 nicm 1031: return ('~');
1.10 shadchin 1032: }
1033: --i;
1034: }
1035: /* Make sure we're back to AT_NORMAL before the '\n'. */
1036: *ap = AT_NORMAL;
1.13 nicm 1037: return (i ? '\0' : '\n');
1.1 etheisen 1038: }
1039:
1040: *ap = attr[i];
1.10 shadchin 1041: return (linebuf[i] & 0xFF);
1.1 etheisen 1042: }
1043:
1044: /*
1045: * Indicate that there is no current line.
1046: */
1.13 nicm 1047: void
1048: null_line(void)
1.1 etheisen 1049: {
1050: is_null_line = 1;
1.5 millert 1051: cshift = 0;
1.1 etheisen 1052: }
1053:
1054: /*
1055: * Analogous to forw_line(), but deals with "raw lines":
1056: * lines which are not split for screen width.
1057: * {{ This is supposed to be more efficient than forw_line(). }}
1058: */
1.13 nicm 1059: off_t
1060: forw_raw_line(off_t curr_pos, char **linep, int *line_lenp)
1061: {
1062: int n;
1063: int c;
1064: off_t new_pos;
1065:
1066: if (curr_pos == -1 || ch_seek(curr_pos) ||
1067: (c = ch_forw_get()) == EOI)
1068: return (-1);
1.1 etheisen 1069:
1.5 millert 1070: n = 0;
1.13 nicm 1071: for (;;) {
1072: if (c == '\n' || c == EOI || ABORT_SIGS()) {
1.1 etheisen 1073: new_pos = ch_tell();
1074: break;
1075: }
1.13 nicm 1076: if (n >= size_linebuf-1) {
1077: if (expand_linebuf()) {
1.5 millert 1078: /*
1079: * Overflowed the input buffer.
1080: * Pretend the line ended here.
1081: */
1082: new_pos = ch_tell() - 1;
1083: break;
1084: }
1.1 etheisen 1085: }
1.13 nicm 1086: linebuf[n++] = (char)c;
1.1 etheisen 1087: c = ch_forw_get();
1088: }
1.5 millert 1089: linebuf[n] = '\0';
1.1 etheisen 1090: if (linep != NULL)
1091: *linep = linebuf;
1.10 shadchin 1092: if (line_lenp != NULL)
1093: *line_lenp = n;
1.1 etheisen 1094: return (new_pos);
1095: }
1096:
1097: /*
1098: * Analogous to back_line(), but deals with "raw lines".
1099: * {{ This is supposed to be more efficient than back_line(). }}
1100: */
1.13 nicm 1101: off_t
1102: back_raw_line(off_t curr_pos, char **linep, int *line_lenp)
1103: {
1104: int n;
1105: int c;
1106: off_t new_pos;
1107:
1108: if (curr_pos == -1 || curr_pos <= ch_zero() || ch_seek(curr_pos - 1))
1109: return (-1);
1.1 etheisen 1110:
1.5 millert 1111: n = size_linebuf;
1112: linebuf[--n] = '\0';
1.13 nicm 1113: for (;;) {
1.1 etheisen 1114: c = ch_back_get();
1.13 nicm 1115: if (c == '\n' || ABORT_SIGS()) {
1.1 etheisen 1116: /*
1117: * This is the newline ending the previous line.
1118: * We have hit the beginning of the line.
1119: */
1120: new_pos = ch_tell() + 1;
1121: break;
1122: }
1.13 nicm 1123: if (c == EOI) {
1.1 etheisen 1124: /*
1125: * We have hit the beginning of the file.
1126: * This must be the first line in the file.
1127: * This must, of course, be the beginning of the line.
1128: */
1129: new_pos = ch_zero();
1130: break;
1131: }
1.13 nicm 1132: if (n <= 0) {
1.5 millert 1133: int old_size_linebuf = size_linebuf;
1.13 nicm 1134: if (expand_linebuf()) {
1.5 millert 1135: /*
1136: * Overflowed the input buffer.
1137: * Pretend the line ended here.
1138: */
1139: new_pos = ch_tell() + 1;
1140: break;
1141: }
1.1 etheisen 1142: /*
1.5 millert 1143: * Shift the data to the end of the new linebuf.
1.1 etheisen 1144: */
1.5 millert 1145: n = size_linebuf - old_size_linebuf;
1.8 millert 1146: memmove(linebuf + n, linebuf, old_size_linebuf);
1.1 etheisen 1147: }
1.5 millert 1148: linebuf[--n] = c;
1.1 etheisen 1149: }
1150: if (linep != NULL)
1.5 millert 1151: *linep = &linebuf[n];
1.10 shadchin 1152: if (line_lenp != NULL)
1153: *line_lenp = size_linebuf - 1 - n;
1.1 etheisen 1154: return (new_pos);
1155: }