summaryrefslogtreecommitdiff
path: root/unicode_constants.h
diff options
context:
space:
mode:
authorKarl Williamson <public@khwilliamson.com>2012-09-02 09:58:43 -0600
committerKarl Williamson <public@khwilliamson.com>2012-09-13 21:14:00 -0600
commit1b0f46bfdeb9373b604bc2b77173cd26c0d37ecc (patch)
treea206c86485b86272106c7dd890bff39dc6367233 /unicode_constants.h
parent525b6419a6e9037dd46cd37b578c11266e7cd2b2 (diff)
downloadperl-1b0f46bfdeb9373b604bc2b77173cd26c0d37ecc.tar.gz
Rename regen'd hdr to reflect expanded capabilities
The recently added utf8_strings.h has been expanded to include more than just strings. I'm renaming it to avoid confusion.
Diffstat (limited to 'unicode_constants.h')
-rw-r--r--unicode_constants.h48
1 files changed, 48 insertions, 0 deletions
diff --git a/unicode_constants.h b/unicode_constants.h
new file mode 100644
index 0000000000..f915d59b8e
--- /dev/null
+++ b/unicode_constants.h
@@ -0,0 +1,48 @@
+/* -*- buffer-read-only: t -*-
+ * !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
+ * This file is built by regen/unicode_constants.pl from Unicode data.
+ * Any changes made here will be lost!
+ */
+
+
+#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".
+ *
+ * The macros that have the suffix "_UTF8" may have further suffixes, as
+ * follows:
+ * "_FIRST_BYTE" if the value is just the first byte of the UTF-8
+ * representation; the value will be a numeric constant.
+ * "_TAIL" if instead it represents all but the first byte. This, and
+ * with no additional suffix are both string constants */
+
+#define COMBINING_GRAVE_ACCENT_UTF8 "\xCC\x80" /* U+0300 */
+#define COMBINING_ACUTE_ACCENT_UTF8 "\xCC\x81" /* U+0301 */
+#define COMBINING_DIAERESIS_UTF8 "\xCC\x88" /* U+0308 */
+
+#define GREEK_SMALL_LETTER_IOTA_UTF8_FIRST_BYTE 0xCE /* U+03B9 */
+#define GREEK_SMALL_LETTER_IOTA_UTF8_TAIL "\xB9" /* U+03B9 */
+
+#define GREEK_SMALL_LETTER_UPSILON_UTF8_FIRST_BYTE 0xCF /* U+03C5 */
+#define GREEK_SMALL_LETTER_UPSILON_UTF8_TAIL "\x85" /* U+03C5 */
+
+#define HANGUL_CHOSEONG_KIYEOK_UTF8 "\xE1\x84\x80" /* U+1100 */
+#define HANGUL_JUNGSEONG_FILLER_UTF8 "\xE1\x85\xA0" /* U+1160 */
+#define HANGUL_JONGSEONG_KIYEOK_UTF8 "\xE1\x86\xA8" /* U+11A8 */
+#define HYPHEN_UTF8 "\xE2\x80\x90" /* U+2010 */
+
+#define DELETE_NATIVE 0x007F /* U+007F */
+#define LATIN_SMALL_LETTER_SHARP_S_NATIVE 0x00DF /* U+00DF */
+#define LATIN_SMALL_LETTER_A_WITH_RING_ABOVE_NATIVE 0x00E5 /* U+00E5 */
+#define LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE_NATIVE 0x00C5 /* U+00C5 */
+#define LATIN_SMALL_LETTER_Y_WITH_DIAERESIS_NATIVE 0x00FF /* U+00FF */
+#define MICRO_SIGN_NATIVE 0x00B5 /* U+00B5 */
+#define NEXT_LINE_NATIVE 0x0085 /* U+0085 */
+
+#endif /* H_UNICODE_CONSTANTS */
+
+/* ex: set ro: */