From 2abcc25173ef40b29cdde28856a3f5c9234056d3 Mon Sep 17 00:00:00 2001 From: Hyeseong Kim Date: Thu, 24 Nov 2016 23:38:56 +0900 Subject: PLT-2077 Support CJK hashtags (#4555) * Add Korean character ranges to exist CJK pattern * Add constant for CJK hashtags Becuase most of keywords in CJK are two characters * Add CJK ranges to hashtag pattern to handle it. * Fixes hashtag pattern to apply numbers at last * Remove a wrong test case `test_` shouldn't be a hashtag * Fix hashtag regex to support standard unicodes * Remove wrong escapes from regex --- model/utils.go | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'model') diff --git a/model/utils.go b/model/utils.go index 457b64c09..0ce243fe7 100644 --- a/model/utils.go +++ b/model/utils.go @@ -304,7 +304,7 @@ func Etag(parts ...interface{}) string { return etag } -var validHashtag = regexp.MustCompile(`^(#[A-Za-zäöüÄÖÜß]+[A-Za-z0-9äöüÄÖÜß_\-]*[A-Za-z0-9äöüÄÖÜß])$`) +var validHashtag = regexp.MustCompile(`^(#\pL[\pL\d\-_.]*[\pL\d])$`) var puncStart = regexp.MustCompile(`^[^\pL\d\s#]+`) var hashtagStart = regexp.MustCompile(`^#{2,}`) var puncEnd = regexp.MustCompile(`[^\pL\d\s]+$`) -- cgit v1.2.3-1-g7c22