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

3 files changed, 3 insertions, 9 deletions

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]+$`)
diff --git a/webapp/tests/formatting_hashtags.test.jsx b/webapp/tests/formatting_hashtags.test.jsx
index 37b84a4a8..1c7de1541 100644
--- a/webapp/tests/formatting_hashtags.test.jsx
+++ b/webapp/tests/formatting_hashtags.test.jsx
@@ -73,12 +73,6 @@ describe('TextFormatting.Hashtags', function() {
             "<p><a class='mention-link' href='#' data-hashtag='#test'>#test</a>.</p>"
         );
 
-        // Known issue, trailing underscore is captured by the clientside regex but not the serverside one
-        assert.equal(
-            TextFormatting.formatText('#test_').trim(),
-            "<p><a class='mention-link' href='#' data-hashtag='#test_'>#test_</a></p>"
-        );
-
         assert.equal(
             TextFormatting.formatText('This is a sentence #test containing a hashtag').trim(),
             "<p>This is a sentence <a class='mention-link' href='#' data-hashtag='#test'>#test</a> containing a hashtag</p>"
diff --git a/webapp/utils/text_formatting.jsx b/webapp/utils/text_formatting.jsx
index 9c2edf954..9f983a1ee 100644
--- a/webapp/utils/text_formatting.jsx
+++ b/webapp/utils/text_formatting.jsx
@@ -11,7 +11,7 @@ import XRegExp from 'xregexp';
 
 // pattern to detect the existance of a Chinese, Japanese, or Korean character in a string
 // http://stackoverflow.com/questions/15033196/using-javascript-to-check-whether-a-string-contains-japanese-characters-includi
-const cjkPattern = /[\u3000-\u303f\u3040-\u309f\u30a0-\u30ff\uff00-\uff9f\u4e00-\u9faf\u3400-\u4dbf]/;
+const cjkPattern = /[\u3000-\u303f\u3040-\u309f\u30a0-\u30ff\uff00-\uff9f\u4e00-\u9faf\u3400-\u4dbf\uac00-\ud7a3]/;
 
 // Performs formatting of user posts including highlighting mentions and search terms and converting urls, hashtags,
 // @mentions and ~channels to links by taking a user's message and returning a string of formatted html. Also takes
@@ -342,7 +342,7 @@ function autolinkHashtags(text, tokens) {
         return prefix + alias;
     }
 
-    return output.replace(/(^|\W)(#[a-zA-ZäöüÄÖÜß][a-zA-Z0-9äöüÄÖÜß.\-_]*)\b/g, replaceHashtagWithToken);
+    return output.replace(XRegExp.cache('(^|\\W)(#\\pL[\\pL\\d\\-_.]*[\\pL\\d])', 'g'), replaceHashtagWithToken);
 }
 
 const puncStart = XRegExp.cache('^[^\\pL\\d\\s#]+');