X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=src%2Ficu_tokenizer.c;h=954ebc66fd957ee26adec0177db772624b75a03f;hb=03dccce903176e61f889b1558834aa5a09f9f8c0;hp=f9b4926c8ae410aacac32ab99fa519b3887bdd10;hpb=ded9dadedab82a379e0e52d78a61fe9632ac870c;p=yaz-moved-to-github.git diff --git a/src/icu_tokenizer.c b/src/icu_tokenizer.c index f9b4926..954ebc6 100644 --- a/src/icu_tokenizer.c +++ b/src/icu_tokenizer.c @@ -1,5 +1,5 @@ /* This file is part of the YAZ toolkit. - * Copyright (C) 1995-2010 Index Data + * Copyright (C) 1995-2012 Index Data * See the file LICENSE for details. */ @@ -19,6 +19,7 @@ #include +#include #include #include #include @@ -45,12 +46,9 @@ struct icu_tokenizer */ }; -struct icu_tokenizer *icu_tokenizer_create(const char *locale, char action, - UErrorCode *status) +static void icu_tokenizer_reset(struct icu_tokenizer *tokenizer, + char action) { - struct icu_tokenizer * tokenizer - = (struct icu_tokenizer *) xmalloc(sizeof(struct icu_tokenizer)); - tokenizer->action = action; tokenizer->bi = 0; tokenizer->buf16 = icu_buf_utf16_create(0); @@ -58,7 +56,32 @@ struct icu_tokenizer *icu_tokenizer_create(const char *locale, char action, tokenizer->token_id = 0; tokenizer->token_start = 0; tokenizer->token_end = 0; + tokenizer->bi = 0; +} + +struct icu_tokenizer *icu_tokenizer_clone(struct icu_tokenizer *old) +{ + int32_t bufferSize = U_BRK_SAFECLONE_BUFFERSIZE; + UErrorCode status = U_ZERO_ERROR; + struct icu_tokenizer * tokenizer + = (struct icu_tokenizer *) xmalloc(sizeof(struct icu_tokenizer)); + + assert(old); + icu_tokenizer_reset(tokenizer, old->action); + assert(old->bi); + tokenizer->bi = ubrk_safeClone(old->bi, NULL, &bufferSize, &status); + if (U_SUCCESS(status)) + return tokenizer; + return tokenizer; +} + +struct icu_tokenizer *icu_tokenizer_create(const char *locale, char action, + UErrorCode *status) +{ + struct icu_tokenizer * tokenizer + = (struct icu_tokenizer *) xmalloc(sizeof(struct icu_tokenizer)); + icu_tokenizer_reset(tokenizer, action); switch (tokenizer->action) { case 'l': @@ -172,14 +195,16 @@ int32_t icu_tokenizer_next_token(struct icu_tokenizer * tokenizer, { tokenizer->token_count++; tokenizer->token_id++; - } else { - tokenizer->token_id = 0; } + else + tokenizer->token_id = 0; + tokenizer->token_start = tkn_start; tokenizer->token_end = tkn_end; /* copying into token buffer if it exists */ - if (tkn16){ + if (tkn16) + { if (tkn16->utf16_cap < tkn_len) icu_buf_utf16_resize(tkn16, (size_t) tkn_len * 2);