vcl/inc/pdf/objectcopier.hxx | 3 -
vcl/qa/cppunit/pdfexport/data/tdf160051.odt |binary
vcl/qa/cppunit/pdfexport/pdfexport.cxx | 44 ++++++++++++++++++++
vcl/qa/cppunit/pdfexport/pdfexport2.cxx | 6 +-
vcl/source/gdi/pdfobjectcopier.cxx | 60 +++++++++++++++++++++++++++-
vcl/source/gdi/pdfwriter_impl.cxx | 4 +
6 files changed, 110 insertions(+), 7 deletions(-)
New commits:
commit 05a075d23eb6003849e75582e12ef788e615a56d
Author: Tibor Nagy <tibor.nagy.ext...@allotropia.de>
AuthorDate: Thu Jan 9 23:02:55 2025 +0100
Commit: Nagy Tibor <tibor.nagy.ext...@allotropia.de>
CommitDate: Sun Jan 12 02:24:24 2025 +0100
tdf#160051 PDF export: Artifact present inside tagged content
If a PDF file containing artifacts is added to a document as an image,
and the document is then exported as a tagged PDF, these artifacts are
placed into a structure element (e.g., figure), which is not allowed.
This fix removes unnecessary artifact tags from the content stream.
Change-Id: I590ebec9a7aecdaa42520008824469bc8a9ff65b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/180041
Reviewed-by: Nagy Tibor <tibor.nagy.ext...@allotropia.de>
Tested-by: Jenkins
diff --git a/vcl/inc/pdf/objectcopier.hxx b/vcl/inc/pdf/objectcopier.hxx
index 0168f69717ae..6d4d8676e418 100644
--- a/vcl/inc/pdf/objectcopier.hxx
+++ b/vcl/inc/pdf/objectcopier.hxx
@@ -57,7 +57,8 @@ public:
/// Copies page one or more page streams from rContentStreams into rStream.
static sal_Int32 copyPageStreams(std::vector<filter::PDFObjectElement*>&
rContentStreams,
- SvMemoryStream& rStream, bool&
rCompressed);
+ SvMemoryStream& rStream, bool&
rCompressed,
+ bool bIsTaggedNonReferenceXObject =
false);
};
}
diff --git a/vcl/qa/cppunit/pdfexport/data/tdf160051.odt
b/vcl/qa/cppunit/pdfexport/data/tdf160051.odt
new file mode 100644
index 000000000000..39151e7e8d2c
Binary files /dev/null and b/vcl/qa/cppunit/pdfexport/data/tdf160051.odt differ
diff --git a/vcl/qa/cppunit/pdfexport/pdfexport.cxx
b/vcl/qa/cppunit/pdfexport/pdfexport.cxx
index c1f2ea43a6a0..c49c01aa7133 100644
--- a/vcl/qa/cppunit/pdfexport/pdfexport.cxx
+++ b/vcl/qa/cppunit/pdfexport/pdfexport.cxx
@@ -90,6 +90,50 @@ CPPUNIT_TEST_FIXTURE(PdfExportTest, testPopupRectangleSize)
}
}
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf160051)
+{
+ // A tagged PDF file which containing artifacts was added to the sample
file as an image.
+ // When the sample file exporting as a tagged PDF, these artifacts are
placed into a structure
+ // element (e.g.:figure) which is not allowed.
+
+ uno::Sequence<beans::PropertyValue> aFilterData(
+ comphelper::InitPropertySequence({ { "PDFUACompliance", uno::Any(true)
},
+ { "SelectPdfVersion",
uno::Any(sal_Int32(17)) } }));
+ aMediaDescriptor[u"FilterData"_ustr] <<= aFilterData;
+
+ vcl::filter::PDFDocument aDocument;
+ load(u"tdf160051.odt", aDocument);
+
+ std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+ CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+ // Directly go to the inner XObject Im7.
+ auto pInnerIm = aDocument.LookupObject(7);
+ CPPUNIT_ASSERT(pInnerIm);
+
+ vcl::filter::PDFStreamElement* pStream = pInnerIm->GetStream();
+ CPPUNIT_ASSERT(pStream);
+ SvMemoryStream& rObjectStream = pStream->GetMemory();
+
+ // Uncompress it.
+ SvMemoryStream aUncompressed;
+ ZCodec aZCodec;
+ aZCodec.BeginCompression();
+ rObjectStream.Seek(0);
+ aZCodec.Decompress(rObjectStream, aUncompressed);
+ CPPUNIT_ASSERT(aZCodec.EndCompression());
+
+ auto pStart = static_cast<const char*>(aUncompressed.GetData());
+ const char* pEnd = pStart + aUncompressed.GetSize();
+ OString aStr("/Artifact"_ostr);
+ auto pArtifact = std::search(pStart, pEnd, aStr.getStr(), aStr.getStr() +
aStr.getLength());
+
+ // Without the fix in place, this test would have failed with
+ // Expected: The content stream does not contain "/Artifact" element
+ // Actual: The content stream contains "/Artifact" element
+ CPPUNIT_ASSERT_EQUAL(pArtifact, pEnd);
+}
+
CPPUNIT_TEST_FIXTURE(PdfExportTest, testCommentAnnotation)
{
// Enable PDF/UA and Comment as PDF annotations
diff --git a/vcl/qa/cppunit/pdfexport/pdfexport2.cxx
b/vcl/qa/cppunit/pdfexport/pdfexport2.cxx
index f883a3c97540..25c707b5f0fc 100644
--- a/vcl/qa/cppunit/pdfexport/pdfexport2.cxx
+++ b/vcl/qa/cppunit/pdfexport/pdfexport2.cxx
@@ -830,7 +830,7 @@ CPPUNIT_TEST_FIXTURE(PdfExportTest2, testMultiPagePDF)
CPPUNIT_ASSERT(aZCodec.EndCompression());
// Just check that the size of the page stream is what is expected.
- CPPUNIT_ASSERT_EQUAL(sal_uInt64(1236), aUncompressed.Tell());
+ CPPUNIT_ASSERT_EQUAL(sal_uInt64(1218), aUncompressed.Tell());
}
{ // embedded PDF page 2
@@ -865,7 +865,7 @@ CPPUNIT_TEST_FIXTURE(PdfExportTest2, testMultiPagePDF)
CPPUNIT_ASSERT(aZCodec.EndCompression());
// Just check that the size of the page stream is what is expected.
- CPPUNIT_ASSERT_EQUAL(sal_uInt64(3911), aUncompressed.Tell());
+ CPPUNIT_ASSERT_EQUAL(sal_uInt64(3893), aUncompressed.Tell());
}
{ // embedded PDF page 3
@@ -900,7 +900,7 @@ CPPUNIT_TEST_FIXTURE(PdfExportTest2, testMultiPagePDF)
CPPUNIT_ASSERT(aZCodec.EndCompression());
// Just check that the size of the page stream is what is expected.
- CPPUNIT_ASSERT_EQUAL(sal_uInt64(373), aUncompressed.Tell());
+ CPPUNIT_ASSERT_EQUAL(sal_uInt64(355), aUncompressed.Tell());
}
#endif
}
diff --git a/vcl/source/gdi/pdfobjectcopier.cxx
b/vcl/source/gdi/pdfobjectcopier.cxx
index 56c3ba6e8138..3761520e3148 100644
--- a/vcl/source/gdi/pdfobjectcopier.cxx
+++ b/vcl/source/gdi/pdfobjectcopier.cxx
@@ -19,6 +19,8 @@
#include <pdf/objectcopier.hxx>
#include <pdf/pdfwriter_impl.hxx>
+#include <o3tl/string_view.hxx>
+
namespace vcl
{
PDFObjectCopier::PDFObjectCopier(PDFObjectContainer& rContainer)
@@ -304,7 +306,8 @@ void
PDFObjectCopier::copyPageResources(filter::PDFObjectElement* pPage, OString
}
sal_Int32
PDFObjectCopier::copyPageStreams(std::vector<filter::PDFObjectElement*>&
rContentStreams,
- SvMemoryStream& rStream, bool&
rCompressed)
+ SvMemoryStream& rStream, bool&
rCompressed,
+ bool bIsTaggedNonReferenceXObject)
{
for (auto pContent : rContentStreams)
{
@@ -344,7 +347,60 @@ sal_Int32
PDFObjectCopier::copyPageStreams(std::vector<filter::PDFObjectElement*
continue;
}
- rStream.WriteBytes(aMemoryStream.GetData(),
aMemoryStream.GetSize());
+ bool bHasArtifact = false;
+ if (bIsTaggedNonReferenceXObject)
+ {
+ auto pStart = static_cast<const
char*>(aMemoryStream.GetData());
+ const char* const pEnd = pStart + aMemoryStream.GetSize();
+ std::string_view aStreamView(pStart, pEnd - pStart);
+
+ std::string_view sArtifact = "/Artifact";
+ std::size_t nPosArtifact = aStreamView.find(sArtifact);
+ if (nPosArtifact != std::string_view::npos)
+ {
+ bHasArtifact = true;
+ SvMemoryStream aTmpStream;
+ std::string_view sBMC = "BMC";
+ std::string_view sBDC = "BDC";
+ std::string_view sEMC = "EMC";
+
+ while (!aStreamView.empty())
+ {
+ aTmpStream.WriteOString(aStreamView.substr(0,
nPosArtifact));
+ aStreamView.remove_prefix(nPosArtifact +
sArtifact.size());
+
+ std::size_t nPosBMC = aStreamView.find(sBMC);
+ std::size_t nPosBDC = aStreamView.find(sBDC);
+ std::size_t nPos = std::min(nPosBMC, nPosBDC);
+
+ if (nPos != std::string_view::npos)
+ {
+ if (nPos == nPosBMC)
+ aStreamView.remove_prefix(nPos + sBMC.size() +
1);
+ else
+ aStreamView.remove_prefix(nPos + sBDC.size() +
1);
+
+ std::size_t nPosEMC = aStreamView.find(sEMC);
+ if (nPosEMC != std::string_view::npos)
+ {
+ aTmpStream.WriteOString(aStreamView.substr(0,
nPosEMC));
+ aStreamView.remove_prefix(nPosEMC +
sEMC.size() + 1);
+ }
+ }
+
+ nPosArtifact = aStreamView.find(sArtifact);
+ if (nPosArtifact == std::string_view::npos)
+ {
+ aTmpStream.WriteOString(aStreamView);
+ break;
+ }
+ }
+ rStream.WriteBytes(aTmpStream.GetData(),
aTmpStream.GetSize());
+ }
+ }
+
+ if (!bHasArtifact)
+ rStream.WriteBytes(aMemoryStream.GetData(),
aMemoryStream.GetSize());
}
else
{
diff --git a/vcl/source/gdi/pdfwriter_impl.cxx
b/vcl/source/gdi/pdfwriter_impl.cxx
index 526dc7ec97b1..99a1ca024a84 100644
--- a/vcl/source/gdi/pdfwriter_impl.cxx
+++ b/vcl/source/gdi/pdfwriter_impl.cxx
@@ -9305,7 +9305,9 @@ void PDFWriterImpl::writeReferenceXObject(const
ReferenceXObjectEmit& rEmit)
SvMemoryStream aStream;
bool bCompressed = false;
- sal_Int32 nLength = PDFObjectCopier::copyPageStreams(aContentStreams,
aStream, bCompressed);
+ bool bIsTaggedNonReferenceXObject = m_aContext.Tagged &&
!m_aContext.UseReferenceXObject;
+ sal_Int32 nLength = PDFObjectCopier::copyPageStreams(aContentStreams,
aStream, bCompressed,
+
bIsTaggedNonReferenceXObject);
aLine.append(nLength);
aLine.append(">>
stream
");
