X-Git-Url: https://git.distorted.org.uk/~mdw/disorder/blobdiff_plain/e5a5a1388e0236c4aa4084493e2f78ff3fbc8c5b..349b7b74df53fe1631a269f4cd4e87951cfe9397:/lib/unidata.h diff --git a/lib/unidata.h b/lib/unidata.h index 8595754..5f22127 100644 --- a/lib/unidata.h +++ b/lib/unidata.h @@ -33,6 +33,46 @@ enum unicode_gc_cat { unicode_gc_Zp, unicode_gc_Zs }; +enum unicode_Grapheme_Break { + unicode_Grapheme_Break_CR, + unicode_Grapheme_Break_Control, + unicode_Grapheme_Break_Extend, + unicode_Grapheme_Break_L, + unicode_Grapheme_Break_LF, + unicode_Grapheme_Break_LV, + unicode_Grapheme_Break_LVT, + unicode_Grapheme_Break_Other, + unicode_Grapheme_Break_T, + unicode_Grapheme_Break_V +}; +extern const char *const unicode_Grapheme_Break_names[]; +enum unicode_Word_Break { + unicode_Word_Break_ALetter, + unicode_Word_Break_Extend, + unicode_Word_Break_ExtendNumLet, + unicode_Word_Break_Format, + unicode_Word_Break_Katakana, + unicode_Word_Break_MidLetter, + unicode_Word_Break_MidNum, + unicode_Word_Break_Numeric, + unicode_Word_Break_Other +}; +extern const char *const unicode_Word_Break_names[]; +enum unicode_Sentence_Break { + unicode_Sentence_Break_ATerm, + unicode_Sentence_Break_Close, + unicode_Sentence_Break_Extend, + unicode_Sentence_Break_Format, + unicode_Sentence_Break_Lower, + unicode_Sentence_Break_Numeric, + unicode_Sentence_Break_OLetter, + unicode_Sentence_Break_Other, + unicode_Sentence_Break_STerm, + unicode_Sentence_Break_Sep, + unicode_Sentence_Break_Sp, + unicode_Sentence_Break_Upper +}; +extern const char *const unicode_Sentence_Break_names[]; enum unicode_flags { unicode_normalize_before_casefold = 1 }; @@ -46,6 +86,9 @@ struct unidata { unsigned char ccc; char gc; uint8_t flags; + char grapheme_break; + char word_break; + char sentence_break; }; extern const struct unidata *const unidata[]; #define UNICODE_NCHARS 195200