QMimeData: Prefer UTF-8 when multiple charsets are available

This was especially a problem on Wayland where both "text/plain" (us-ascii) and
"text/plain;charset=utf-8" are typically available. I.e. we would prefer ascii
over utf-8, losing special characters.

Fixes: QTBUG-54786
Change-Id: I985f66e16fcd5125e800c86c4d3949d210e939ba
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>

1 files changed, 5 insertions, 0 deletions

diff --git a/src/corelib/kernel/qmimedata.cpp b/src/corelib/kernel/qmimedata.cpp
index c8ad1bc43f..9a98db21d3 100644
--- a/src/corelib/kernel/qmimedata.cpp
+++ b/src/corelib/kernel/qmimedata.cpp
@@ -51,6 +51,7 @@ QT_BEGIN_NAMESPACE
 static inline QString textUriListLiteral() { return QStringLiteral("text/uri-list"); }
 static inline QString textHtmlLiteral() { return QStringLiteral("text/html"); }
 static inline QString textPlainLiteral() { return QStringLiteral("text/plain"); }
+static inline QString textPlainUtf8Literal() { return QStringLiteral("text/plain;charset=utf-8"); }
 static inline QString applicationXColorLiteral() { return QStringLiteral("application/x-color"); }
 static inline QString applicationXQtImageLiteral() { return QStringLiteral("application/x-qt-image"); }
 
@@ -399,6 +400,10 @@ bool QMimeData::hasUrls() const
 QString QMimeData::text() const
 {
     Q_D(const QMimeData);
+    QVariant utf8Text = d->retrieveTypedData(textPlainUtf8Literal(), QVariant::String);
+    if (!utf8Text.isNull())
+        return utf8Text.toString();
+
     QVariant data = d->retrieveTypedData(textPlainLiteral(), QVariant::String);
     return data.toString();
 }