[Libreoffice-commits] core.git: Branch 'libreoffice-6-0' - include/vcl vcl/source

[Libreoffice Gerrit user]

 include/vcl/filter/pdfdocument.hxx     |    4 ++++
 vcl/source/filter/ipdf/pdfdocument.cxx |   20 ++++++++++++++++----
 2 files changed, 20 insertions(+), 4 deletions(-)

New commits:
commit e6d5ef741c75895ed97801112823b8332557d79a
Author:     Caolán McNamara <caolanm at redhat.com>
AuthorDate: Wed Aug 1 12:04:30 2018 +0100
Commit:     Michael Stahl <Michael.Stahl at cib.de>
CommitDate: Fri Aug 3 11:03:18 2018 +0200

    forcepoint#65 pdf page visiting revisits itself
    
    Change-Id: I6d9eb75f0850a94814fb4d69ea1442b826674496
    Reviewed-on: https://gerrit.libreoffice.org/58418
    Tested-by: Jenkins
    Reviewed-by: Michael Stahl <Michael.Stahl at cib.de>

diff --git a/include/vcl/filter/pdfdocument.hxx b/include/vcl/filter/pdfdocument.hxx
index dfc13534bff8..697751b7f94d 100644
--- a/include/vcl/filter/pdfdocument.hxx
+++ b/include/vcl/filter/pdfdocument.hxx
@@ -50,6 +50,7 @@ class VCL_DLLPUBLIC PDFObjectElement : public PDFElement
     PDFDocument& m_rDoc;
     double m_fObjectValue;
     double m_fGenerationValue;
+    bool m_bVisiting;
     std::map<OString, PDFElement*> m_aDictionary;
     /// If set, the object contains this number element (outside any dictionary/array).
     PDFNumberElement* m_pNumberElement;
@@ -109,6 +110,9 @@ public:
     SvMemoryStream* GetStreamBuffer() const;
     void SetStreamBuffer(std::unique_ptr<SvMemoryStream>& pStreamBuffer);
     PDFDocument& GetDocument();
+
+    /// Visits the page tree recursively, looking for page objects.
+    void visitPages(std::vector<PDFObjectElement*>& rRet);
 };
 
 /// Array object: a list.
diff --git a/vcl/source/filter/ipdf/pdfdocument.cxx b/vcl/source/filter/ipdf/pdfdocument.cxx
index a9f78fbe7f8c..b55ad45096ef 100644
--- a/vcl/source/filter/ipdf/pdfdocument.cxx
+++ b/vcl/source/filter/ipdf/pdfdocument.cxx
@@ -1751,15 +1751,17 @@ const std::vector< std::unique_ptr<PDFElement> >& PDFDocument::GetElements()
 }
 
 /// Visits the page tree recursively, looking for page objects.
-static void visitPages(PDFObjectElement* pPages, std::vector<PDFObjectElement*>& rRet)
+void PDFObjectElement::visitPages(std::vector<PDFObjectElement*>& rRet)
 {
-    auto pKids = dynamic_cast<PDFArrayElement*>(pPages->Lookup("Kids"));
+    auto pKids = dynamic_cast<PDFArrayElement*>(Lookup("Kids"));
     if (!pKids)
     {
         SAL_WARN("vcl.filter", "visitPages: pages has no kids");
         return;
     }
 
+    m_bVisiting = true;
+
     for (const auto& pKid : pKids->GetElements())
     {
         auto pReference = dynamic_cast<PDFReferenceElement*>(pKid);
@@ -1770,14 +1772,23 @@ static void visitPages(PDFObjectElement* pPages, std::vector<PDFObjectElement*>&
         if (!pKidObject)
             continue;
 
+        // detect if visiting reenters itself
+        if (pKidObject->m_bVisiting)
+        {
+            SAL_WARN("vcl.filter", "visitPages: loop in hierarchy");
+            continue;
+        }
+
         auto pName = dynamic_cast<PDFNameElement*>(pKidObject->Lookup("Type"));
         if (pName && pName->GetValue() == "Pages")
             // Pages inside pages: recurse.
-            visitPages(pKidObject, rRet);
+            pKidObject->visitPages(rRet);
         else
             // Found an actual page.
             rRet.push_back(pKidObject);
     }
+
+    m_bVisiting = false;
 }
 
 std::vector<PDFObjectElement*> PDFDocument::GetPages()
@@ -1822,7 +1833,7 @@ std::vector<PDFObjectElement*> PDFDocument::GetPages()
         return aRet;
     }
 
-    visitPages(pPages, aRet);
+    pPages->visitPages(aRet);
 
     return aRet;
 }
@@ -2087,6 +2098,7 @@ PDFObjectElement::PDFObjectElement(PDFDocument& rDoc, double fObjectValue, doubl
     : m_rDoc(rDoc),
       m_fObjectValue(fObjectValue),
       m_fGenerationValue(fGenerationValue),
+      m_bVisiting(false),
       m_pNumberElement(nullptr),
       m_nDictionaryOffset(0),
       m_nDictionaryLength(0),