ofz#18526 sw: WW8 import: don't insert control characters

Sanitize string before calling InsertString().

This segfaults since:

commit b522fc0646915d4da94df38dd249c88b28f25be7
Date:   Tue Sep 24 18:11:45 2019 +0200

sw: maintain fieldmarks in DeleteRange()/DeleteAndJoin()/ReplaceRange()

Change-Id: I9ef73d924420686f6838fa21900ec57b4d25c905
Reviewed-on: https://gerrit.libreoffice.org/81949
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
(cherry picked from commit 7ecda38cdaa2361e8510bf3e7206863c4936deab)
Reviewed-on: https://gerrit.libreoffice.org/82759
(cherry picked from commit d494a4c0ead7db481757d8d67fbce9e1b02e65df)

4 files changed, 48 insertions, 14 deletions

diff --git a/sw/qa/core/data/ww5/pass/ofz18526-1.doc b/sw/qa/core/data/ww5/pass/ofz18526-1.doc
new file mode 100644
index 000000000000..e651650f9a26
--- /dev/null
+++ b/sw/qa/core/data/ww5/pass/ofz18526-1.doc
diff --git a/sw/source/filter/ww8/ww8par.cxx b/sw/source/filter/ww8/ww8par.cxx
index 98569b059863..efb4b2942920 100644
--- a/sw/source/filter/ww8/ww8par.cxx
+++ b/sw/source/filter/ww8/ww8par.cxx
@@ -115,6 +115,8 @@
 #include <com/sun/star/document/XDocumentPropertiesSupplier.hpp>
 #include <com/sun/star/document/XViewDataSupplier.hpp>
 #include <com/sun/star/document/IndexedPropertyValues.hpp>
+
+#include <svl/lngmisc.hxx>
 #include <svl/itemiter.hxx>
 
 #include <comphelper/processfactory.hxx>
@@ -3391,13 +3393,37 @@ void SwWW8ImplReader::emulateMSWordAddTextToParagraph(const OUString& rAddString
     }
 }
 
+namespace sw {
+
+auto FilterControlChars(OUString const& rString) -> OUString
+{
+    OUStringBuffer buf(rString.getLength());
+    for (sal_Int32 i = 0; i < rString.getLength(); ++i)
+    {
+        sal_Unicode const ch(rString[i]);
+        if (!linguistic::IsControlChar(ch) || ch == '\r' || ch == '\n' || ch == '\t')
+        {
+            buf.append(ch);
+        }
+        else
+        {
+            SAL_INFO("sw.ww8", "filtering control character");
+        }
+    }
+    return buf.makeStringAndClear();
+}
+
+} // namespace sw
+
 void SwWW8ImplReader::simpleAddTextToParagraph(const OUString& rAddString)
 {
-    if (rAddString.isEmpty())
+    OUString const addString(sw::FilterControlChars(rAddString));
+
+    if (addString.isEmpty())
         return;
 
 #if OSL_DEBUG_LEVEL > 1
-    SAL_INFO("sw.ww8", "<addTextToParagraph>" << rAddString << "</addTextToParagraph>");
+    SAL_INFO("sw.ww8", "<addTextToParagraph>" << addString << "</addTextToParagraph>");
 #endif
 
     const SwContentNode *pCntNd = m_pPaM->GetContentNode();
@@ -3411,21 +3437,21 @@ void SwWW8ImplReader::simpleAddTextToParagraph(const OUString& rAddString)
     const sal_Int32 nCharsLeft = SAL_MAX_INT32 - pNd->GetText().getLength();
     if (nCharsLeft > 0)
     {
-        if (rAddString.getLength() <= nCharsLeft)
+        if (addString.getLength() <= nCharsLeft)
         {
-            m_rDoc.getIDocumentContentOperations().InsertString(*m_pPaM, rAddString);
+            m_rDoc.getIDocumentContentOperations().InsertString(*m_pPaM, addString);
         }
         else
         {
-            m_rDoc.getIDocumentContentOperations().InsertString(*m_pPaM, rAddString.copy(0, nCharsLeft));
+            m_rDoc.getIDocumentContentOperations().InsertString(*m_pPaM, addString.copy(0, nCharsLeft));
             AppendTextNode(*m_pPaM->GetPoint());
-            m_rDoc.getIDocumentContentOperations().InsertString(*m_pPaM, rAddString.copy(nCharsLeft));
+            m_rDoc.getIDocumentContentOperations().InsertString(*m_pPaM, addString.copy(nCharsLeft));
         }
     }
     else
     {
         AppendTextNode(*m_pPaM->GetPoint());
-        m_rDoc.getIDocumentContentOperations().InsertString(*m_pPaM, rAddString);
+        m_rDoc.getIDocumentContentOperations().InsertString(*m_pPaM, addString);
     }
 
     m_bReadTable = false;
@@ -3451,13 +3477,17 @@ bool SwWW8ImplReader::ReadChars(WW8_CP& rPos, WW8_CP nNextAttr, long nTextEnd,
                 nRequested = nMaxPossible;
             }
 
-            for (WW8_CP nCh = 0; nCh < nRequested; ++nCh)
+            if (!linguistic::IsControlChar(m_cSymbol)
+                || m_cSymbol == '\r' || m_cSymbol == '\n' || m_cSymbol == '\t')
             {
-                m_rDoc.getIDocumentContentOperations().InsertString( *m_pPaM, OUString(m_cSymbol) );
+                for (WW8_CP nCh = 0; nCh < nRequested; ++nCh)
+                {
+                    m_rDoc.getIDocumentContentOperations().InsertString(*m_pPaM, OUString(m_cSymbol));
+                }
+                m_xCtrlStck->SetAttr(*m_pPaM->GetPoint(), RES_CHRATR_FONT);
+                m_xCtrlStck->SetAttr(*m_pPaM->GetPoint(), RES_CHRATR_CJK_FONT);
+                m_xCtrlStck->SetAttr(*m_pPaM->GetPoint(), RES_CHRATR_CTL_FONT);
             }
-            m_xCtrlStck->SetAttr( *m_pPaM->GetPoint(), RES_CHRATR_FONT );
-            m_xCtrlStck->SetAttr( *m_pPaM->GetPoint(), RES_CHRATR_CJK_FONT );
-            m_xCtrlStck->SetAttr( *m_pPaM->GetPoint(), RES_CHRATR_CTL_FONT );
         }
         m_pStrm->SeekRel(nRequested);
         rPos = nEnd; // Ignore until attribute end
diff --git a/sw/source/filter/ww8/ww8par.hxx b/sw/source/filter/ww8/ww8par.hxx
index ee5a8b68b795..676745d07078 100644
--- a/sw/source/filter/ww8/ww8par.hxx
+++ b/sw/source/filter/ww8/ww8par.hxx
@@ -550,6 +550,8 @@ namespace sw
             sal_Int32 GetPtContent() { return mnPtContent; };
         };
     }
+
+    auto FilterControlChars(OUString const& rString) -> OUString;
 }
 
 class WW8FieldEntry
diff --git a/sw/source/filter/ww8/ww8par5.cxx b/sw/source/filter/ww8/ww8par5.cxx
index 1d2430db8d69..963364531d19 100644
--- a/sw/source/filter/ww8/ww8par5.cxx
+++ b/sw/source/filter/ww8/ww8par5.cxx
@@ -34,6 +34,7 @@
 #include <svl/urihelper.hxx>
 #include <svl/zforlist.hxx>
 #include <svl/zformat.hxx>
+#include <svl/lngmisc.hxx>
 #include <sfx2/linkmgr.hxx>
 #include <rtl/character.hxx>
 #include <unotools/charclass.hxx>
@@ -1906,7 +1907,8 @@ eF_ResT SwWW8ImplReader::Read_F_Symbol( WW8FieldDesc*, OUString& rStr )
     if( aQ.isEmpty() )
         return eF_ResT::TAGIGN;                      // -> no 0-char in text
 
-    if (sal_Unicode cChar = static_cast<sal_Unicode>(aQ.toInt32()))
+    sal_Unicode const cChar = static_cast<sal_Unicode>(aQ.toInt32());
+    if (!linguistic::IsControlChar(cChar) || cChar == '\r' || cChar == '\n' || cChar == '\t')
     {
         if (!aName.isEmpty())                           // Font Name set ?
         {
@@ -2686,11 +2688,11 @@ void SwWW8ImplReader::Read_SubF_Ruby( WW8ReadFieldParams& rReadParam)
                             if ((nBegin != -1) && (nEnd != -1) && (nBegin < nEnd))
                             {
                                 sText = sPart.copy(nBegin+1,nEnd-nBegin-1);
+                                sText = sw::FilterControlChars(sText);
                             }
                         }
                     }
                 }
-
             }
             break;
         }