diff options
| author | Karl Williamson <khw@cpan.org> | 2015-07-28 22:17:01 -0600 |
|---|---|---|
| committer | Karl Williamson <khw@cpan.org> | 2015-07-28 22:17:01 -0600 |
| commit | a0b05c4bde4f97301b663b8de70677551eacff4c (patch) | |
| tree | e9329726efb7aa27f5fd08f1441943f95b68a594 /unicode_constants.h | |
| parent | 16d89be8495ff4e0a9b99e837b5444df8a4a6cc6 (diff) | |
| parent | 912fd71114648888551a4750534463f20ae16c7c (diff) | |
| download | perl-a0b05c4bde4f97301b663b8de70677551eacff4c.tar.gz | |
Allow perl to work again with all Unicode releases
Compiling perl to work with early Unicode releases is a goal, but has
been broken since 5.18. This sequence of commits gets this working
again, actually better than before.
Diffstat (limited to 'unicode_constants.h')
| -rw-r--r-- | unicode_constants.h | 21 |
1 files changed, 17 insertions, 4 deletions
diff --git a/unicode_constants.h b/unicode_constants.h index 7da4eb3a48..be7a38e92a 100644 --- a/unicode_constants.h +++ b/unicode_constants.h @@ -8,10 +8,11 @@ #ifndef H_UNICODE_CONSTANTS /* Guard against nested #includes */ #define H_UNICODE_CONSTANTS 1 -/* This file contains #defines for various Unicode code points. The values - * the macros expand to are the native Unicode code point, or all or portions - * of the UTF-8 encoding for the code point. In the former case, the macro - * name has the suffix "_NATIVE"; otherwise, the suffix "_UTF8". +/* This file contains #defines for the version of Unicode being used and + * various Unicode code points. The values the code point macros expand to + * are the native Unicode code point, or all or portions of the UTF-8 encoding + * for the code point. In the former case, the macro name has the suffix + * "_NATIVE"; otherwise, the suffix "_UTF8". * * The macros that have the suffix "_UTF8" may have further suffixes, as * follows: @@ -20,6 +21,10 @@ * "_TAIL" if instead it represents all but the first byte. This, and * with no additional suffix are both string constants */ +#define UNICODE_MAJOR_VERSION 8 +#define UNICODE_DOT_VERSION 0 +#define UNICODE_DOT_DOT_VERSION 0 + #if 'A' == 65 /* ASCII/Latin1 */ # define LATIN_SMALL_LETTER_LONG_S_UTF8 "\xC5\xBF" /* U+017F */ @@ -33,6 +38,8 @@ # define LATIN_SMALL_LIGATURE_LONG_S_T_UTF8 "\xEF\xAC\x85" /* U+FB05 */ # define LATIN_SMALL_LIGATURE_ST_UTF8 "\xEF\xAC\x86" /* U+FB06 */ +# define LATIN_CAPITAL_LETTER_I_WITH_DOT_ABOVE_UTF8 "\xC4\xB0" /* U+0130 */ +# define LATIN_SMALL_LETTER_DOTLESS_I_UTF8 "\xC4\xB1" /* U+0131 */ # define HYPHEN_UTF8 "\xE2\x80\x90" /* U+2010 */ # define FIRST_SURROGATE_UTF8_FIRST_BYTE 0xED /* U+D800 */ @@ -71,6 +78,8 @@ # define LATIN_SMALL_LIGATURE_LONG_S_T_UTF8 "\xDD\x72\x67\x46" /* U+FB05 */ # define LATIN_SMALL_LIGATURE_ST_UTF8 "\xDD\x72\x67\x47" /* U+FB06 */ +# define LATIN_CAPITAL_LETTER_I_WITH_DOT_ABOVE_UTF8 "\x8D\x57" /* U+0130 */ +# define LATIN_SMALL_LETTER_DOTLESS_I_UTF8 "\x8D\x58" /* U+0131 */ # define HYPHEN_UTF8 "\xCA\x41\x57" /* U+2010 */ # define FIRST_SURROGATE_UTF8_FIRST_BYTE 0xDD /* U+D800 */ @@ -109,6 +118,8 @@ # define LATIN_SMALL_LIGATURE_LONG_S_T_UTF8 "\xDD\x71\x66\x46" /* U+FB05 */ # define LATIN_SMALL_LIGATURE_ST_UTF8 "\xDD\x71\x66\x47" /* U+FB06 */ +# define LATIN_CAPITAL_LETTER_I_WITH_DOT_ABOVE_UTF8 "\x8C\x57" /* U+0130 */ +# define LATIN_SMALL_LETTER_DOTLESS_I_UTF8 "\x8C\x58" /* U+0131 */ # define HYPHEN_UTF8 "\xCA\x41\x57" /* U+2010 */ # define FIRST_SURROGATE_UTF8_FIRST_BYTE 0xDD /* U+D800 */ @@ -147,6 +158,8 @@ # define LATIN_SMALL_LIGATURE_LONG_S_T_UTF8 "\xDC\x74\x68\x46" /* U+FB05 */ # define LATIN_SMALL_LIGATURE_ST_UTF8 "\xDC\x74\x68\x47" /* U+FB06 */ +# define LATIN_CAPITAL_LETTER_I_WITH_DOT_ABOVE_UTF8 "\x8E\x58" /* U+0130 */ +# define LATIN_SMALL_LETTER_DOTLESS_I_UTF8 "\x8E\x59" /* U+0131 */ # define HYPHEN_UTF8 "\xCA\x41\x58" /* U+2010 */ # define FIRST_SURROGATE_UTF8_FIRST_BYTE 0xDC /* U+D800 */ |