Revert "fix math export/import in docx/rtf"

This reverts commit 2b5953a19e36a02040f2ff08bc87efe4785f80bd.  Whatever that
"gcc4.4 (and 4.3 and possibly older) have a problem with dynamic_cast directly
to the target class" issue actually was:  For one, our GCC 7 baseline presumably
would no longer have such an issue.  And for another, the added asserts that the
results of the dynamic_casts must be non-null were presumably all bogus (and
have in part been reverted again in the meantime), as all the sources are UNO
interface types that can presumably point at implementation objects of other
than the expected C++ class types.  (Those dynamic_casts from UNO interface
types will be addressed in a follow-up commit.  See the upcoming commit
introducing loplugin:unocast on why such dynamic_casts are dangerous.)

Conflicts:
sw/qa/extras/ooxmlexport/ooxmlexport.cxx
sw/qa/extras/rtfexport/rtfexport.cxx
sw/source/filter/ww8/docxattributeoutput.cxx
sw/source/filter/ww8/rtfattributeoutput.cxx
writerfilter/Library_writerfilter.mk
writerfilter/source/ooxml/OOXMLFastContextHandler.cxx
writerfilter/source/rtftok/rtfdocumentimpl.cxx

Change-Id: I0c330a3541e64ce08bfe30ff15d51a2fd8a243b8
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/144336
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

2 files changed, 5 insertions, 11 deletions

diff --git a/writerfilter/source/ooxml/OOXMLFastContextHandler.cxx b/writerfilter/source/ooxml/OOXMLFastContextHandler.cxx
index b479e587ebc0..5fcd1dd518cb 100644
--- a/writerfilter/source/ooxml/OOXMLFastContextHandler.cxx
+++ b/writerfilter/source/ooxml/OOXMLFastContextHandler.cxx
@@ -2282,11 +2282,8 @@ void OOXMLFastContextHandlerMath::process()
     if (!ref.is())
         return;
     uno::Reference< uno::XInterface > component(ref->getComponent(), uno::UNO_QUERY_THROW);
-// gcc4.4 (and 4.3 and possibly older) have a problem with dynamic_cast directly to the target class,
-// so help it with an intermediate cast. I'm not sure what exactly the problem is, seems to be unrelated
-// to RTLD_GLOBAL, so most probably a gcc bug.
-    oox::FormulaImportBase& import = dynamic_cast<oox::FormulaImportBase&>(dynamic_cast<SfxBaseModel&>(*component));
-    import.readFormulaOoxml(buffer);
+    if( oox::FormulaImportBase* import = dynamic_cast< oox::FormulaImportBase* >( component.get()))
+        import->readFormulaOoxml( buffer );
     if (!isForwardEvents())
         return;
 
diff --git a/writerfilter/source/rtftok/rtfdocumentimpl.cxx b/writerfilter/source/rtftok/rtfdocumentimpl.cxx
index 8a888d8348f6..921bb61efceb 100644
--- a/writerfilter/source/rtftok/rtfdocumentimpl.cxx
+++ b/writerfilter/source/rtftok/rtfdocumentimpl.cxx
@@ -2967,12 +2967,9 @@ RTFError RTFDocumentImpl::beforePopState(RTFParserState& rState)
             {
                 uno::Reference<util::XCloseable> xComponent(xObject->getComponent(),
                                                             uno::UNO_SET_THROW);
-                // gcc4.4 (and 4.3 and possibly older) have a problem with dynamic_cast directly to the target class,
-                // so help it with an intermediate cast. I'm not sure what exactly the problem is, seems to be unrelated
-                // to RTLD_GLOBAL, so most probably a gcc bug.
-                auto& rImport = dynamic_cast<oox::FormulaImportBase&>(
-                    dynamic_cast<SfxBaseModel&>(*xComponent));
-                rImport.readFormulaOoxml(m_aMathBuffer);
+                if (oox::FormulaImportBase* pImport
+                    = dynamic_cast<oox::FormulaImportBase*>(xComponent.get()))
+                    pImport->readFormulaOoxml(m_aMathBuffer);
 
                 auto pValue = new RTFValue(xObject);
                 RTFSprms aMathAttributes;