From 9477fd7eba9bda6813b65c6c54380904892ca35e Mon Sep 17 00:00:00 2001
From: tobi <31960611+tsmethurst@users.noreply.github.com>
Date: Sun, 8 Dec 2024 16:03:00 +0100
Subject: [feature] Allow partial-word hashtags using non-breaking spaces
 (#3606)

* [feature] Allow partial-word hashtags using non-breaking spaces

* update docs
---
 internal/text/util.go | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)

(limited to 'internal/text/util.go')

diff --git a/internal/text/util.go b/internal/text/util.go
index 204c64838..af45cfaf0 100644
--- a/internal/text/util.go
+++ b/internal/text/util.go
@@ -38,8 +38,34 @@ func isPermittedInHashtag(r rune) bool {
 // is a recognized break character for before
 // or after a #hashtag.
 func isHashtagBoundary(r rune) bool {
-	return unicode.IsSpace(r) ||
-		(unicode.IsPunct(r) && r != '_')
+	switch {
+
+	// Zero width space.
+	case r == '\u200B':
+		return true
+
+	// Zero width no-break space.
+	case r == '\uFEFF':
+		return true
+
+	// Pipe character sometimes
+	// used as workaround.
+	case r == '|':
+		return true
+
+	// Standard Unicode white space.
+	case unicode.IsSpace(r):
+		return true
+
+	// Non-underscore punctuation.
+	case unicode.IsPunct(r) && r != '_':
+		return true
+
+	// Not recognized
+	// hashtag boundary.
+	default:
+		return false
+	}
 }
 
 // isMentionBoundary returns true if rune r
-- 
cgit v1.2.3