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/plain_test.go | 13 +++++++++++++
 internal/text/util.go       | 30 ++++++++++++++++++++++++++++--
 2 files changed, 41 insertions(+), 2 deletions(-)

(limited to 'internal/text')

diff --git a/internal/text/plain_test.go b/internal/text/plain_test.go
index 48280bb44..fac54a38e 100644
--- a/internal/text/plain_test.go
+++ b/internal/text/plain_test.go
@@ -36,6 +36,8 @@ const (
 	moreComplexExpected        = "<p>Another test <span class=\"h-card\"><a href=\"http://fossbros-anonymous.io/@foss_satan\" class=\"u-url mention\" rel=\"nofollow noreferrer noopener\" target=\"_blank\">@<span>foss_satan</span></a></span><br><br><a href=\"http://localhost:8080/tags/hashtag\" class=\"mention hashtag\" rel=\"tag nofollow noreferrer noopener\" target=\"_blank\">#<span>Hashtag</span></a><br><br>Text<br><br>:rainbow:</p>"
 	withUTF8Link               = "here's a link with utf-8 characters in it: https://example.org/söme_url"
 	withUTF8LinkExpected       = "<p>here's a link with utf-8 characters in it: <a href=\"https://example.org/s%C3%B6me_url\" rel=\"nofollow noreferrer noopener\" target=\"_blank\">https://example.org/söme_url</a></p>"
+	withFunkyTags              = "#hashtag1 pee #hashtag2\u200Bpee #hashtag3|poo #hashtag4\uFEFFpoo"
+	withFunkyTagsExpected      = "<p><a href=\"http://localhost:8080/tags/hashtag1\" class=\"mention hashtag\" rel=\"tag nofollow noreferrer noopener\" target=\"_blank\">#<span>hashtag1</span></a> pee <a href=\"http://localhost:8080/tags/hashtag2\" class=\"mention hashtag\" rel=\"tag nofollow noreferrer noopener\" target=\"_blank\">#<span>hashtag2</span></a>\u200bpee <a href=\"http://localhost:8080/tags/hashtag3\" class=\"mention hashtag\" rel=\"tag nofollow noreferrer noopener\" target=\"_blank\">#<span>hashtag3</span></a>|poo <a href=\"http://localhost:8080/tags/hashtag4\" class=\"mention hashtag\" rel=\"tag nofollow noreferrer noopener\" target=\"_blank\">#<span>hashtag4</span></a>\ufeffpoo</p>"
 )
 
 type PlainTestSuite struct {
@@ -136,6 +138,17 @@ func (suite *PlainTestSuite) TestDeriveHashtagsOK() {
 	suite.Equal("올빼미", tags[0].Name)
 }
 
+func (suite *PlainTestSuite) TestFunkyTags() {
+	formatted := suite.FromPlain(withFunkyTags)
+	suite.Equal(withFunkyTagsExpected, formatted.HTML)
+
+	tags := formatted.Tags
+	suite.Equal("hashtag1", tags[0].Name)
+	suite.Equal("hashtag2", tags[1].Name)
+	suite.Equal("hashtag3", tags[2].Name)
+	suite.Equal("hashtag4", tags[3].Name)
+}
+
 func (suite *PlainTestSuite) TestDeriveMultiple() {
 	statusText := `Another test @foss_satan@fossbros-anonymous.io
 
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 v1.3