Add QDom internalSubset implementation

QDom's internalSubset() always returned empty because nothing
actually set the internal data member it returns. When parsing
the DECLTYPE, extract the internal subset and save it to the
doctype()'s member when present.

Pick-to: 5.15 6.2 6.3 6.4
Fixes: QTBUG-53661
Change-Id: I6e41ff8b914381168246073b3289d82205b1c255
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>

3 files changed, 100 insertions, 0 deletions

diff --git a/src/xml/dom/qdomhelpers.cpp b/src/xml/dom/qdomhelpers.cpp
index 39c7dcd360..b58d2f7ae9 100644
--- a/src/xml/dom/qdomhelpers.cpp
+++ b/src/xml/dom/qdomhelpers.cpp
@@ -52,6 +52,35 @@ bool QDomBuilder::startDTD(const QString &name, const QString &publicId, const Q
     return true;
 }
 
+QString QDomBuilder::dtdInternalSubset(const QString &dtd)
+{
+    // https://www.w3.org/TR/xml/#NT-intSubset
+    // doctypedecl: '<!DOCTYPE' S Name (S ExternalID)? S? ('[' intSubset ']' S?)? '>'
+    const QString &name = doc->doctype()->name;
+    QStringView tmp = QStringView(dtd).sliced(dtd.indexOf(name) + name.size());
+
+    const QString &publicId = doc->doctype()->publicId;
+    if (!publicId.isEmpty())
+        tmp = tmp.sliced(tmp.indexOf(publicId) + publicId.size());
+
+    const QString &systemId = doc->doctype()->systemId;
+    if (!systemId.isEmpty())
+        tmp = tmp.sliced(tmp.indexOf(systemId) + systemId.size());
+
+    const qsizetype obra = tmp.indexOf(u'[');
+    const qsizetype cbra = tmp.lastIndexOf(u']');
+    if (obra >= 0 && cbra >= 0)
+        return tmp.left(cbra).sliced(obra + 1).toString();
+
+    return QString();
+}
+
+bool QDomBuilder::parseDTD(const QString &dtd)
+{
+    doc->doctype()->internalSubset = dtdInternalSubset(dtd);
+    return true;
+}
+
 bool QDomBuilder::startElement(const QString &nsURI, const QString &qName,
                                const QXmlStreamAttributes &atts)
 {
@@ -272,6 +301,8 @@ bool QDomParser::parseProlog()
                         QDomParser::tr("Error occurred while processing document type declaration"));
                 return false;
             }
+            if (!domBuilder.parseDTD(reader->text().toString()))
+                return false;
             if (!parseMarkupDecl())
                 return false;
             break;
diff --git a/src/xml/dom/qdomhelpers_p.h b/src/xml/dom/qdomhelpers_p.h
index f2a0a79c3a..81de6e2e80 100644
--- a/src/xml/dom/qdomhelpers_p.h
+++ b/src/xml/dom/qdomhelpers_p.h
@@ -45,6 +45,7 @@ public:
     bool startEntity(const QString &name);
     bool endEntity();
     bool startDTD(const QString &name, const QString &publicId, const QString &systemId);
+    bool parseDTD(const QString &dtd);
     bool comment(const QString &characters);
     bool externalEntityDecl(const QString &name, const QString &publicId, const QString &systemId);
     bool notationDecl(const QString &name, const QString &publicId, const QString &systemId);
@@ -61,6 +62,8 @@ public:
     int errorColumn;
 
 private:
+    QString dtdInternalSubset(const QString &dtd);
+
     QDomDocumentPrivate *doc;
     QDomNodePrivate *node;
     QXmlStreamReader *reader;
diff --git a/tests/auto/xml/dom/qdom/tst_qdom.cpp b/tests/auto/xml/dom/qdom/tst_qdom.cpp
index f05020f61c..0736e158fc 100644
--- a/tests/auto/xml/dom/qdom/tst_qdom.cpp
+++ b/tests/auto/xml/dom/qdom/tst_qdom.cpp
@@ -97,6 +97,8 @@ private slots:
     void cloneDTD_QTBUG8398() const;
     void DTDNotationDecl();
     void DTDEntityDecl();
+    void DTDInternalSubset() const;
+    void DTDInternalSubset_data() const;
     void QTBUG49113_dontCrashWithNegativeIndex() const;
 
     void cleanupTestCase() const;
@@ -2044,5 +2046,69 @@ void tst_QDom::QTBUG49113_dontCrashWithNegativeIndex() const
     QVERIFY(node.isNull());
 }
 
+void tst_QDom::DTDInternalSubset() const
+{
+    QFETCH( QString, doc );
+    QFETCH( QString, internalSubset );
+    QXmlStreamReader reader(doc);
+    QDomDocument document;
+    QVERIFY(document.setContent(&reader, true));
+
+    QCOMPARE(document.doctype().internalSubset(), internalSubset);
+}
+
+void tst_QDom::DTDInternalSubset_data() const
+{
+    QTest::addColumn<QString>("doc");
+    QTest::addColumn<QString>("internalSubset");
+
+    QTest::newRow("data1") << "<?xml version='1.0'?>\n"
+                              "<!DOCTYPE note SYSTEM '/[abcd].dtd'>\n"
+                              "<note/>\n"
+                           << "" ;
+
+    QTest::newRow("data2") << "<?xml version='1.0'?>\n"
+                              "<!DOCTYPE note PUBLIC '-/freedesktop' 'https://[abcd].dtd'>\n"
+                              "<note/>\n"
+                           << "" ;
+
+    const QString internalSubset0(
+                "<!-- open brackets comment [ -->\n"
+                "<!-- colse brackets comment ] -->\n"
+                );
+    QTest::newRow("data3") << "<?xml version='1.0'?>\n"
+                              "<!DOCTYPE note ["
+                              + internalSubset0 +
+                              "]>\n"
+                              "<note/>\n"
+                           << internalSubset0;
+
+    const QString internalSubset1(
+                "<!ENTITY obra '['>\n"
+                "<!ENTITY cbra ']'>\n"
+                );
+    QTest::newRow("data4") << "<?xml version='1.0'?>\n"
+                              "<!DOCTYPE note ["
+                              + internalSubset1 +
+                              "]>\n"
+                              "<note/>\n"
+                           << internalSubset1;
+
+    QTest::newRow("data5") << "<?xml version='1.0'?>\n"
+          "<!DOCTYPE note  PUBLIC '-/freedesktop' 'https://[abcd].dtd' ["
+          + internalSubset0
+          + "]>\n"
+          "<note/>\n"
+       << internalSubset0;
+
+    QTest::newRow("data6") << "<?xml version='1.0'?>\n"
+          "<!DOCTYPE note  PUBLIC '-/freedesktop' "
+            "'2001:db8:130F:0000:0000:09C0:876A:130B://[abcd].dtd' ["
+          + internalSubset0
+          + "]>\n"
+          "<note/>\n"
+       << internalSubset0;
+}
+
 QTEST_MAIN(tst_QDom)
 #include "tst_qdom.moc"