2 * Copyright (C) 1984-2014 Mark Nudelman
4 * You may distribute under the terms of either the GNU General Public
5 * License or the Less License, as specified in the README file.
7 * For more information, see the README file.
12 * Functions to define the character set
13 * and do things specific to the character set.
25 public int utf_mode = 0;
28 * Predefined character sets,
29 * selected by the LESSCHARSET environment variable.
36 { "ascii", NULL, "8bcccbcc18b95.b" },
37 { "utf-8", &utf_mode, "8bcccbcc18b95.b126.bb" },
38 { "iso8859", NULL, "8bcccbcc18b95.33b." },
39 { "latin3", NULL, "8bcccbcc18b95.33b5.b8.b15.b4.b12.b18.b12.b." },
40 { "arabic", NULL, "8bcccbcc18b95.33b.3b.7b2.13b.3b.b26.5b19.b" },
41 { "greek", NULL, "8bcccbcc18b95.33b4.2b4.b3.b35.b44.b" },
42 { "greek2005", NULL, "8bcccbcc18b95.33b14.b35.b44.b" },
43 { "hebrew", NULL, "8bcccbcc18b95.33b.b29.32b28.2b2.b" },
44 { "koi8-r", NULL, "8bcccbcc18b95.b." },
45 { "KOI8-T", NULL, "8bcccbcc18b95.b8.b6.b8.b.b.5b7.3b4.b4.b3.b.b.3b." },
46 { "georgianps", NULL, "8bcccbcc18b95.3b11.4b12.2b." },
47 { "tcvn", NULL, "b..b...bcccbccbbb7.8b95.b48.5b." },
48 { "TIS-620", NULL, "8bcccbcc18b95.b.4b.11b7.8b." },
49 { "next", NULL, "8bcccbcc18b95.bb125.bb" },
50 { "dos", NULL, "8bcccbcc12bc5b95.b." },
51 { "windows-1251", NULL, "8bcccbcc12bc5b95.b24.b." },
52 { "windows-1252", NULL, "8bcccbcc12bc5b95.b.b11.b.2b12.b." },
53 { "windows-1255", NULL, "8bcccbcc12bc5b95.b.b8.b.5b9.b.4b." },
54 { "ebcdic", NULL, "5bc6bcc7bcc41b.9b7.9b5.b..8b6.10b6.b9.7b9.8b8.17b3.3b9.7b9.8b8.6b10.b.b.b." },
55 { "IBM-1047", NULL, "4cbcbc3b9cbccbccbb4c6bcc5b3cbbc4bc4bccbc191.b" },
60 * Support "locale charmap"/nl_langinfo(CODESET) values, as well as others.
67 { "ANSI_X3.4-1968", "ascii" },
68 { "US-ASCII", "ascii" },
69 { "latin1", "iso8859" },
70 { "ISO-8859-1", "iso8859" },
71 { "latin9", "iso8859" },
72 { "ISO-8859-15", "iso8859" },
73 { "latin2", "iso8859" },
74 { "ISO-8859-2", "iso8859" },
75 { "ISO-8859-3", "latin3" },
76 { "latin4", "iso8859" },
77 { "ISO-8859-4", "iso8859" },
78 { "cyrillic", "iso8859" },
79 { "ISO-8859-5", "iso8859" },
80 { "ISO-8859-6", "arabic" },
81 { "ISO-8859-7", "greek" },
82 { "IBM9005", "greek2005" },
83 { "ISO-8859-8", "hebrew" },
84 { "latin5", "iso8859" },
85 { "ISO-8859-9", "iso8859" },
86 { "latin6", "iso8859" },
87 { "ISO-8859-10", "iso8859" },
88 { "latin7", "iso8859" },
89 { "ISO-8859-13", "iso8859" },
90 { "latin8", "iso8859" },
91 { "ISO-8859-14", "iso8859" },
92 { "latin10", "iso8859" },
93 { "ISO-8859-16", "iso8859" },
95 { "EBCDIC-US", "ebcdic" },
96 { "IBM1047", "IBM-1047" },
97 { "KOI8-R", "koi8-r" },
98 { "KOI8-U", "koi8-r" },
99 { "GEORGIAN-PS", "georgianps" },
100 { "TCVN5712-1", "tcvn" },
101 { "NEXTSTEP", "next" },
102 { "windows", "windows-1252" }, /* backward compatibility */
103 { "CP1251", "windows-1251" },
104 { "CP1252", "windows-1252" },
105 { "CP1255", "windows-1255" },
109 #define IS_BINARY_CHAR 01
110 #define IS_CONTROL_CHAR 02
112 static char chardef[256];
113 static char *binfmt = NULL;
114 static char *utfbinfmt = NULL;
115 public int binattr = AT_STANDOUT;
119 * Define a charset, given a description string.
120 * The string consists of 256 letters,
121 * one for each character in the charset.
122 * If the string is shorter than 256 letters, missing letters
123 * are taken to be identical to the last one.
124 * A decimal number followed by a letter is taken to be a
125 * repetition of the letter.
127 * Each letter is one of:
130 * c control character
154 v = IS_BINARY_CHAR|IS_CONTROL_CHAR;
157 case '0': case '1': case '2': case '3': case '4':
158 case '5': case '6': case '7': case '8': case '9':
159 n = (10 * n) + (s[-1] - '0');
163 error("invalid chardef", NULL_PARG);
170 if (cp >= chardef + sizeof(chardef))
172 error("chardef longer than 256", NULL_PARG);
181 while (cp < chardef + sizeof(chardef))
186 * Define a charset, given a charset name.
187 * The valid charset names are listed in the "charsets" array.
190 icharset(name, no_error)
194 register struct charset *p;
195 register struct cs_alias *a;
197 if (name == NULL || *name == '\0')
200 /* First see if the name is an alias. */
201 for (a = cs_aliases; a->name != NULL; a++)
203 if (strcmp(name, a->name) == 0)
210 for (p = charsets; p->name != NULL; p++)
212 if (strcmp(name, p->name) == 0)
215 if (p->p_flag != NULL)
222 error("invalid charset name", NULL_PARG);
230 * Define a charset, given a locale name.
237 for (c = 0; c < (int) sizeof(chardef); c++)
242 chardef[c] = IS_CONTROL_CHAR;
244 chardef[c] = IS_BINARY_CHAR|IS_CONTROL_CHAR;
250 * Define the printing format for control (or binary utf) chars.
253 setbinfmt(s, fmtvarptr, default_fmt)
260 /* It would be too hard to account for width otherwise. */
264 if (*t < ' ' || *t > '~')
274 if (s == NULL || *s == '\0' ||
275 (*s == '*' && (s[1] == '\0' || s[2] == '\0' || strchr(s + 2, 'n'))) ||
276 (*s != '*' && strchr(s, 'n')))
280 * Select the attributes if it starts with "*".
287 case 'd': binattr = AT_BOLD; break;
288 case 'k': binattr = AT_BLINK; break;
289 case 's': binattr = AT_STANDOUT; break;
290 case 'u': binattr = AT_UNDERLINE; break;
291 default: binattr = AT_NORMAL; break;
307 * See if environment variable LESSCHARSET is defined.
309 s = lgetenv("LESSCHARSET");
314 * LESSCHARSET is not defined: try LESSCHARDEF.
316 s = lgetenv("LESSCHARDEF");
317 if (s != NULL && *s != '\0')
326 * Try using the codeset name as the charset name.
328 s = nl_langinfo(CODESET);
336 * Check whether LC_ALL, LC_CTYPE or LANG look like UTF-8 is used.
338 if ((s = lgetenv("LC_ALL")) != NULL ||
339 (s = lgetenv("LC_CTYPE")) != NULL ||
340 (s = lgetenv("LANG")) != NULL)
342 if ( strstr(s, "UTF-8") != NULL || strstr(s, "utf-8") != NULL
343 || strstr(s, "UTF8") != NULL || strstr(s, "utf8") != NULL)
344 if (icharset("utf-8", 1))
351 * Get character definitions from locale functions,
352 * rather than from predefined charset entry.
359 (void) icharset("dos", 1);
362 * Default to "latin1".
364 (void) icharset("latin1", 1);
370 * Initialize charset data structures.
378 setlocale(LC_ALL, "");
383 s = lgetenv("LESSBINFMT");
384 setbinfmt(s, &binfmt, "*s<%02X>");
386 s = lgetenv("LESSUTFBINFMT");
387 setbinfmt(s, &utfbinfmt, "<U+%04lX>");
391 * Is a given character a "binary" character?
398 return (is_ubin_char(c));
400 return (chardef[c] & IS_BINARY_CHAR);
404 * Is a given character a "control" character?
411 return (chardef[c] & IS_CONTROL_CHAR);
415 * Return the printable form of a character.
416 * For example, in the "ascii" charset '\3' is printed as "^C".
422 /* {{ This buffer can be overrun if LESSBINFMT is a long string. }} */
426 if ((c < 128 || !utf_mode) && !control_char(c))
427 SNPRINTF1(buf, sizeof(buf), "%c", (int) c);
431 else if (!binary_char(c) && c < 64)
432 SNPRINTF1(buf, sizeof(buf), "^%c",
434 * This array roughly inverts CONTROL() #defined in less.h,
435 * and should be kept in sync with CONTROL() and IBM-1047.
441 "..V....D....TU.Z"[c]);
443 else if (c < 128 && !control_char(c ^ 0100))
444 SNPRINTF1(buf, sizeof(buf), "^%c", (int) (c ^ 0100));
447 SNPRINTF1(buf, sizeof(buf), binfmt, c);
452 * Return the printable form of a UTF-8 character.
462 else if (ch < 128 && control_char(ch))
464 if (!control_char(ch ^ 0100))
465 SNPRINTF1(buf, sizeof(buf), "^%c", ((char) ch) ^ 0100);
467 SNPRINTF1(buf, sizeof(buf), binfmt, (char) ch);
468 } else if (is_ubin_char(ch))
470 SNPRINTF1(buf, sizeof(buf), utfbinfmt, ch);
474 if (ch >= 0x80000000)
475 ch = 0xFFFD; /* REPLACEMENT CHARACTER */
483 * Get the length of a UTF-8 character in bytes.
489 if ((ch & 0x80) == 0)
491 if ((ch & 0xE0) == 0xC0)
493 if ((ch & 0xF0) == 0xE0)
495 if ((ch & 0xF8) == 0xF0)
497 if ((ch & 0xFC) == 0xF8)
499 if ((ch & 0xFE) == 0xFC)
501 /* Invalid UTF-8 encoding. */
506 * Does the parameter point to the lead byte of a well-formed UTF-8 character?
509 is_utf8_well_formed(s)
515 if (IS_UTF8_INVALID(s[0]))
518 len = utf_len((char) s[0]);
528 mask = (~((1 << (8-len)) - 1)) & 0xFF;
529 if (s[0] == mask && (s[1] & mask) == 0x80)
533 for (i = 1; i < len; i++)
534 if (!IS_UTF8_TRAIL(s[i]))
540 * Return number of invalid UTF-8 sequences found in a buffer.
543 utf_bin_count(data, len)
550 if (is_utf8_well_formed(data))
552 int clen = utf_len(*data);
557 /* Skip to next lead byte. */
562 } while (len > 0 && !IS_UTF8_LEAD(*data));
569 * Get the value of a UTF-8 character.
575 switch (utf_len(p[0]))
583 /* 110xxxxx 10xxxxxx */
585 ((p[0] & 0x1F) << 6) |
588 /* 1110xxxx 10xxxxxx 10xxxxxx */
590 ((p[0] & 0x0F) << 12) |
591 ((p[1] & 0x3F) << 6) |
594 /* 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */
596 ((p[0] & 0x07) << 18) |
597 ((p[1] & 0x3F) << 12) |
598 ((p[2] & 0x3F) << 6) |
601 /* 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx */
603 ((p[0] & 0x03) << 24) |
604 ((p[1] & 0x3F) << 18) |
605 ((p[2] & 0x3F) << 12) |
606 ((p[3] & 0x3F) << 6) |
609 /* 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx */
611 ((p[0] & 0x01) << 30) |
612 ((p[1] & 0x3F) << 24) |
613 ((p[2] & 0x3F) << 18) |
614 ((p[3] & 0x3F) << 12) |
615 ((p[4] & 0x3F) << 6) |
621 * Store a character into a UTF-8 string.
628 if (!utf_mode || ch < 0x80)
631 *(*pp)++ = (char) ch;
632 } else if (ch < 0x800)
634 /* 110xxxxx 10xxxxxx */
635 *(*pp)++ = (char) (0xC0 | ((ch >> 6) & 0x1F));
636 *(*pp)++ = (char) (0x80 | (ch & 0x3F));
637 } else if (ch < 0x10000)
639 /* 1110xxxx 10xxxxxx 10xxxxxx */
640 *(*pp)++ = (char) (0xE0 | ((ch >> 12) & 0x0F));
641 *(*pp)++ = (char) (0x80 | ((ch >> 6) & 0x3F));
642 *(*pp)++ = (char) (0x80 | (ch & 0x3F));
643 } else if (ch < 0x200000)
645 /* 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */
646 *(*pp)++ = (char) (0xF0 | ((ch >> 18) & 0x07));
647 *(*pp)++ = (char) (0x80 | ((ch >> 12) & 0x3F));
648 *(*pp)++ = (char) (0x80 | ((ch >> 6) & 0x3F));
649 *(*pp)++ = (char) (0x80 | (ch & 0x3F));
650 } else if (ch < 0x4000000)
652 /* 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx */
653 *(*pp)++ = (char) (0xF0 | ((ch >> 24) & 0x03));
654 *(*pp)++ = (char) (0x80 | ((ch >> 18) & 0x3F));
655 *(*pp)++ = (char) (0x80 | ((ch >> 12) & 0x3F));
656 *(*pp)++ = (char) (0x80 | ((ch >> 6) & 0x3F));
657 *(*pp)++ = (char) (0x80 | (ch & 0x3F));
660 /* 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx */
661 *(*pp)++ = (char) (0xF0 | ((ch >> 30) & 0x01));
662 *(*pp)++ = (char) (0x80 | ((ch >> 24) & 0x3F));
663 *(*pp)++ = (char) (0x80 | ((ch >> 18) & 0x3F));
664 *(*pp)++ = (char) (0x80 | ((ch >> 12) & 0x3F));
665 *(*pp)++ = (char) (0x80 | ((ch >> 6) & 0x3F));
666 *(*pp)++ = (char) (0x80 | (ch & 0x3F));
671 * Step forward or backward one character in a string.
674 step_char(pp, dir, limit)
685 /* It's easy if chars are one byte. */
687 ch = (LWCHAR) ((p < limit) ? *p++ : 0);
689 ch = (LWCHAR) ((p > limit) ? *--p : 0);
704 while (p > limit && IS_UTF8_TRAIL(p[-1]))
716 * Unicode characters data
717 * Actual data is in the generated *.uni files.
720 #define DECLARE_RANGE_TABLE_START(name) \
721 static struct wchar_range name##_array[] = {
722 #define DECLARE_RANGE_TABLE_END(name) \
723 }; struct wchar_range_table name##_table = { name##_array, sizeof(name##_array)/sizeof(*name##_array) };
725 DECLARE_RANGE_TABLE_START(compose)
726 #include "compose.uni"
727 DECLARE_RANGE_TABLE_END(compose)
729 DECLARE_RANGE_TABLE_START(ubin)
731 DECLARE_RANGE_TABLE_END(ubin)
733 DECLARE_RANGE_TABLE_START(wide)
735 DECLARE_RANGE_TABLE_END(wide)
737 /* comb_table is special pairs, not ranges. */
738 static struct wchar_range comb_table[] = {
739 {0x0644,0x0622}, {0x0644,0x0623}, {0x0644,0x0625}, {0x0644,0x0627},
744 is_in_table(ch, table)
746 struct wchar_range_table *table;
751 /* Binary search in the table. */
752 if (ch < table->table[0].first)
755 hi = table->count - 1;
758 int mid = (lo + hi) / 2;
759 if (ch > table->table[mid].last)
761 else if (ch < table->table[mid].first)
770 * Is a character a UTF-8 composing character?
771 * If a composing character follows any char, the two combine into one glyph.
774 is_composing_char(ch)
777 return is_in_table(ch, &compose_table);
781 * Should this UTF-8 character be treated as binary?
787 return is_in_table(ch, &ubin_table);
791 * Is this a double width UTF-8 character?
797 return is_in_table(ch, &wide_table);
801 * Is a character a UTF-8 combining character?
802 * A combining char acts like an ordinary char, but if it follows
803 * a specific char (not any char), the two combine into one glyph.
806 is_combining_char(ch1, ch2)
810 /* The table is small; use linear search. */
812 for (i = 0; i < sizeof(comb_table)/sizeof(*comb_table); i++)
814 if (ch1 == comb_table[i].first &&
815 ch2 == comb_table[i].last)