refactor: move zero width replacement to a function (#5594)

src/util/Helpers.cpp

@@ -7,6 +7,18 @@
 #include <QRegularExpression>
 #include <QUuid>
 
+namespace {
+
+const QString ZERO_WIDTH_JOINER = QStringLiteral("\u200D");
+
+// Note: \U requires /utf-8 for MSVC
+// See https://mm2pl.github.io/emoji_rfc.pdf
+const QRegularExpression ESCAPE_TAG_REGEX(
+    QStringLiteral("(?<!\U000E0002)\U000E0002"),
+    QRegularExpression::UseUnicodePropertiesOption);
+
+}  // namespace
+
 namespace chatterino {
 
 namespace helpers::detail {
@@ -283,4 +295,10 @@ bool compareEmoteStrings(const QString &a, const QString &b)
     return k < 0;
 }
 
+QString unescapeZeroWidthJoiner(QString escaped)
+{
+    escaped.replace(ESCAPE_TAG_REGEX, ZERO_WIDTH_JOINER);
+    return escaped;
+}
+
 }  // namespace chatterino

src/util/Helpers.hpp

@@ -182,4 +182,11 @@ constexpr std::optional<std::decay_t<T>> makeConditionedOptional(bool condition,
     return std::nullopt;
 }
 
+/// @brief Unescapes zero width joiners (ZWJ; U+200D) from Twitch messages
+///
+/// Older Chatterino versions escape ZWJ with an ESCAPE TAG (U+E0002), following
+/// https://mm2pl.github.io/emoji_rfc.pdf. This function unescapes all tags with
+/// a ZWJ. See also: https://github.com/Chatterino/chatterino2/issues/3384.
+QString unescapeZeroWidthJoiner(QString escaped);
+
 }  // namespace chatterino