tdf#124670: xml:space attribute may be specified for w:document top-level element

Treat xml:space specially in OOXMLFastContextHandler::startFastElement,
to allow this attribute to be handled for any element.

Change-Id: I81bd1e0642940ffdfc03d6c65d0ce9f421206c5e
Reviewed-on: https://gerrit.libreoffice.org/70723Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>

sw/qa/extras/ooxmlimport/ooxmlimport2.cxx

@@ -331,6 +331,16 @@ DECLARE_OOXMLIMPORT_TEST(testTdf121440, "tdf121440.docx")
         getProperty<sal_Int32>(getRun(getParagraph(1), 1), "CharEscapement"));
 }
 
+DECLARE_OOXMLIMPORT_TEST(testTdf124670, "tdf124670.docx")
+{
+    CPPUNIT_ASSERT_EQUAL(1, getParagraphs());
+    // We need to take xml:space attribute into account, even in w:document element
+    uno::Reference<text::XTextRange> paragraph = getParagraph(1);
+    CPPUNIT_ASSERT_EQUAL(
+        OUString("You won't believe, but that's how it was in markup of original      bugdoc!"),
+        paragraph->getString());
+}
+
 // tests should only be added to ooxmlIMPORT *if* they fail round-tripping in ooxmlEXPORT
 
 CPPUNIT_PLUGIN_IMPLEMENT();

writerfilter/source/ooxml/OOXMLFastContextHandler.cxx

@@ -146,6 +146,13 @@ void SAL_CALL OOXMLFastContextHandler::startFastElement
 (sal_Int32 Element,
  const uno::Reference< xml::sax::XFastAttributeList > & Attribs)
 {
+    // Set xml:space value early, to allow child contexts use it when dealing with strings.
+    if (Attribs && Attribs->hasAttribute(oox::NMSP_xml | oox::XML_space))
+    {
+        mbPreserveSpace = Attribs->getValue(oox::NMSP_xml | oox::XML_space) == "preserve";
+        mbPreserveSpaceSet = true;
+    }
+
     if (oox::getNamespace(Element) == NMSP_mce)
         m_bDiscardChildren = prepareMceContext(Element, Attribs);
 
@@ -883,6 +890,8 @@ bool OOXMLFastContextHandler::IsPreserveSpace() const
 {
     // xml:space attribute applies to all elements within the content of the element where it is specified,
     // unless overridden with another instance of the xml:space attribute
+    if (mbPreserveSpaceSet)
+        return mbPreserveSpace;
     if (mpParent)
         return mpParent->IsPreserveSpace();
     return false; // default value
@@ -895,9 +904,7 @@ bool OOXMLFastContextHandler::IsPreserveSpace() const
 OOXMLFastContextHandlerStream::OOXMLFastContextHandlerStream
 (OOXMLFastContextHandler * pContext)
 : OOXMLFastContextHandler(pContext),
-  mpPropertySetAttrs(new OOXMLPropertySet),
-  mbPreserveSpace(false),
-  mbPreserveSpaceSet(false)
+  mpPropertySetAttrs(new OOXMLPropertySet)
 {
 }
 
@@ -908,14 +915,7 @@ OOXMLFastContextHandlerStream::~OOXMLFastContextHandlerStream()
 void OOXMLFastContextHandlerStream::newProperty(Id nId,
                                                 const OOXMLValue::Pointer_t& pVal)
 {
-    if (nId == NS_ooxml::LN_CT_Text_space)
-    {
-        // Set <xml:space> value early, to allow
-        // child contexts use it when dealing with strings
-        mbPreserveSpace = pVal->getString() == "preserve";
-        mbPreserveSpaceSet = true;
-    }
-    else if (nId != 0x0)
+    if (nId != 0x0)
     {
         mpPropertySetAttrs->add(nId, pVal, OOXMLProperty::ATTRIBUTE);
     }
@@ -945,15 +945,6 @@ void OOXMLFastContextHandlerStream::handleHyperlink()
     aHyperlinkHandler.writetext();
 }
 
-bool OOXMLFastContextHandlerStream::IsPreserveSpace() const
-{
-    // xml:space attribute applies to all elements within the content of the element where it is specified,
-    // unless overridden with another instance of the xml:space attribute
-    if (mbPreserveSpaceSet)
-        return mbPreserveSpace;
-    return OOXMLFastContextHandler::IsPreserveSpace();
-}
-
 /*
   class OOXMLFastContextHandlerProperties
  */

writerfilter/source/ooxml/OOXMLFastContextHandler.hxx

@@ -51,7 +51,7 @@ public:
     virtual ~OOXMLFastContextHandler() override;
 
     // css::xml::sax::XFastContextHandler:
-    virtual void SAL_CALL startFastElement (sal_Int32 Element, const css::uno::Reference< css::xml::sax::XFastAttributeList >& Attribs) override;
+    virtual void SAL_CALL startFastElement (sal_Int32 Element, const css::uno::Reference< css::xml::sax::XFastAttributeList >& Attribs) override final;
 
     virtual void SAL_CALL startUnknownElement(const OUString & Namespace, const OUString & Name, const css::uno::Reference< css::xml::sax::XFastAttributeList > & Attribs) override;
 
@@ -224,9 +224,6 @@ protected:
     void startAction();
     void endAction();
 
-    // 2.10 of XML 1.0 specification
-    virtual bool IsPreserveSpace() const;
-
     const css::uno::Reference< css::uno::XComponentContext >& getComponentContext() { return m_xContext;}
 
     bool inPositionV;
@@ -237,9 +234,14 @@ private:
     /// Handles AlternateContent. Returns true, if children of the current element should be ignored.
     bool prepareMceContext(Token_t nElement, const css::uno::Reference<css::xml::sax::XFastAttributeList>& Attribs);
 
+    // 2.10 of XML 1.0 specification
+    bool IsPreserveSpace() const;
+
     css::uno::Reference< css::uno::XComponentContext > m_xContext;
     bool m_bDiscardChildren;
     bool m_bTookChoice; ///< Did we take the Choice or want Fallback instead?
+    bool mbPreserveSpace = false;
+    bool mbPreserveSpaceSet = false;
 
 };
 
@@ -259,13 +261,8 @@ public:
 
     void handleHyperlink();
 
-protected:
-    virtual bool IsPreserveSpace() const override;
-
 private:
     mutable OOXMLPropertySet::Pointer_t mpPropertySetAttrs;
-    bool mbPreserveSpace    : 1;
-    bool mbPreserveSpaceSet : 1;
 };
 
 class OOXMLFastContextHandlerProperties : public OOXMLFastContextHandler