tommy/libpostal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[normalize] Need to do a Latin-ASCII transliteration even if the string is entirely ASCII since it may contain HTML escapes

Browse Source
This commit is contained in:
Al
2015-08-11 23:36:08 -04:00
parent 87b275fcab
commit 66a71ab70d
1 changed files with 12 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
src/normalize.c
View File

@@ -116,11 +116,21 @@ string_tree_t *normalize_string(char *str, uint64_t options) {
if (options & NORMALIZE_STRING_LOWERCASE && is_ascii) {
utf8_normalized = normalize_string_utf8(str, NORMALIZE_STRING_LOWERCASE);
if (utf8_normalized != NULL) {
string_tree_add_string(tree, utf8_normalized);
if (options & NORMALIZE_STRING_LATIN_ASCII) {
transliterated = transliterate(LATIN_ASCII, utf8_normalized, len);
if (transliterated != NULL) {
string_tree_add_string(tree, transliterated);
free(transliterated);
transliterated = NULL;
}
} else {
string_tree_add_string(tree, utf8_normalized);
}
free(utf8_normalized);
utf8_normalized = NULL;
}
}
} else if (options & NORMALIZE_STRING_LATIN_ASCII && script == SCRIPT_LATIN && script_len > 0) {
add_latin_alternatives(tree, str, script_len, options);
} else if (options & NORMALIZE_STRING_TRANSLITERATE && script != SCRIPT_UNKNOWN && script_len > 0) {