diff options
| author | vng <viktor.govako@gmail.com> | 2015-07-01 16:50:49 +0300 |
|---|---|---|
| committer | Alex Zolotarev <alex@maps.me> | 2015-09-23 02:54:47 +0300 |
| commit | 671dcce02a7a73079c66023b6713f5660aa5379f (patch) | |
| tree | cf9b93386240352bf6c14b217b08272df8700bcc /indexer/search_string_utils.hpp | |
| parent | 90457ec1e4e5d8307df931ba75a3a84b1baa544c (diff) | |
[search] Treat “Đ,đ” as “d” letter in normalization.
Diffstat (limited to 'indexer/search_string_utils.hpp')
| -rw-r--r-- | indexer/search_string_utils.hpp | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/indexer/search_string_utils.hpp b/indexer/search_string_utils.hpp index d27e43991a..86c9b10092 100644 --- a/indexer/search_string_utils.hpp +++ b/indexer/search_string_utils.hpp @@ -17,6 +17,10 @@ inline strings::UniString NormalizeAndSimplifyString(string const & s) UniChar & c = uniString[i]; switch (c) { + // Replace "d with stroke" to simple d letter. Used in Vietnamese. + // (unicode-compliant implementation leaves it unchanged) + case 0x0110: + case 0x0111: c = 'd'; break; // Replace small turkish dotless 'ı' with dotted 'i'. // Our own invented hack to avoid well-known Turkish I-letter bug. case 0x0131: c = 'i'; break; |