fix: HTML detection for documents starting with <html>

Bionus · Nov 5, 2022 · 319b523 · 319b523
1 parent df00f39
commit 319b523
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 1 deletion.
diff --git a/src/lib/src/downloader/file-downloader.cpp b/src/lib/src/downloader/file-downloader.cpp
@@ -60,7 +60,7 @@ void FileDownloader::replyFinished()
 	const auto msg = m_reply->errorString();
 	const QUrl redirectUrl = m_reply->attribute(QNetworkRequest::RedirectionTargetAttribute).toUrl();
 	const bool failedLastWrite = data.length() > 0 && written < 0;
-	const bool invalidHtml = !m_allowHtmlResponses && QString(data.left(100)).trimmed().startsWith("<!DOCTYPE", Qt::CaseInsensitive);
+	const bool invalidHtml = !m_allowHtmlResponses && isHtml(data);
 	const bool emptyFile = m_readSize == 0 && redirectUrl.isEmpty();
 
 	if (error != NetworkReply::NetworkError::NoError || failedLastWrite || invalidHtml || emptyFile) {

diff --git a/src/lib/src/functions.cpp b/src/lib/src/functions.cpp
@@ -1210,3 +1210,15 @@ void renameSettingsKey(QSettings *settings, const QString &before, const QString
 	settings->setValue(after, settings->value(before));
 	settings->remove(before);
 }
+
+
+/**
+ * Helper function to know if a byte array contains HTML.
+ * Only performs a basic check and might fail for XML or some differently formatted documents.
+ */
+bool isHtml(const QByteArray &data)
+{
+	const QString left = QString(data.left(100)).trimmed();
+	return left.startsWith("<!DOCTYPE", Qt::CaseInsensitive)
+		|| left.startsWith("<html>", Qt::CaseInsensitive);
+}
diff --git a/src/lib/src/functions.h b/src/lib/src/functions.h
@@ -103,6 +103,8 @@ QRect stringToRect(const QString &str);
 void renameSettingsGroup(QSettings *settings, const QString &before, const QString &after);
 void renameSettingsKey(QSettings *settings, const QString &before, const QString &after);
 
+bool isHtml(const QByteArray &data);
+
 
 
 template <typename T>