diff options
Diffstat (limited to 'src/modules/filters/utf8nfkd.cpp')
-rw-r--r-- | src/modules/filters/utf8nfkd.cpp | 46 |
1 files changed, 46 insertions, 0 deletions
diff --git a/src/modules/filters/utf8nfkd.cpp b/src/modules/filters/utf8nfkd.cpp new file mode 100644 index 0000000..450cbbf --- /dev/null +++ b/src/modules/filters/utf8nfkd.cpp @@ -0,0 +1,46 @@ +/****************************************************************************** +* +* utf8nfkd - SWFilter decendant to perform NFKD (compatability decomposition +* normalization) on UTF-8 text +*/ + +#ifdef _ICU_ + +#include <stdlib.h> +#include <string.h> + +#ifdef __GNUC__ +#include <unixstr.h> +#endif + +#include <utf8nfkd.h> + +UTF8NFKD::UTF8NFKD() { + conv = ucnv_open("UTF-8", &err); +} + +UTF8NFKD::~UTF8NFKD() { + ucnv_close(conv); +} + +char UTF8NFKD::ProcessText(char *text, int maxlen, const SWKey *key, const SWModule *module) +{ + int32_t len = strlen(text) * 2; + source = new UChar[len + 1]; //each char could become a surrogate pair + + // Convert UTF-8 string to UTF-16 (UChars) + len = ucnv_toUChars(conv, source, len, text, -1, &err); + target = new UChar[len + 1]; + + //compatability decomposition + unorm_normalize(source, len, UNORM_NFKD, 0, target, len, &err); + + ucnv_fromUChars(conv, text, maxlen, target, -1, &err); + + delete [] source; + delete [] target; + + return 0; +} + +#endif |