Adding upstream version 4:24.2.0.upstream/4%24.2.0

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

1 files changed, 2774 insertions, 0 deletions

diff --git a/vcl/qa/cppunit/pdfexport/pdfexport.cxx b/vcl/qa/cppunit/pdfexport/pdfexport.cxx
new file mode 100644
index 0000000000..d5515facb5
--- /dev/null
+++ b/vcl/qa/cppunit/pdfexport/pdfexport.cxx
@@ -0,0 +1,2774 @@
+/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
+/*
+ * This file is part of the LibreOffice project.
+ *
+ * This Source Code Form is subject to the terms of the Mozilla Public
+ * License, v. 2.0. If a copy of the MPL was not distributed with this
+ * file, You can obtain one at http://mozilla.org/MPL/2.0/.
+ */
+
+#include <sal/config.h>
+
+#include <memory>
+#include <string_view>
+
+#include <config_fonts.h>
+
+#include <com/sun/star/frame/XStorable.hpp>
+#include <com/sun/star/view/XPrintable.hpp>
+
+#include <comphelper/propertysequence.hxx>
+#include <test/unoapi_test.hxx>
+#include <unotools/mediadescriptor.hxx>
+#include <unotools/tempfile.hxx>
+#include <vcl/filter/pdfdocument.hxx>
+#include <tools/zcodec.hxx>
+#include <o3tl/string_view.hxx>
+
+#include <vcl/filter/PDFiumLibrary.hxx>
+
+using namespace ::com::sun::star;
+
+namespace
+{
+/// Tests the PDF export filter.
+class PdfExportTest : public UnoApiTest
+{
+protected:
+    utl::MediaDescriptor aMediaDescriptor;
+
+public:
+    PdfExportTest()
+        : UnoApiTest("/vcl/qa/cppunit/pdfexport/data/")
+    {
+    }
+
+    void saveAsPDF(std::u16string_view rFile);
+    void load(std::u16string_view rFile, vcl::filter::PDFDocument& rDocument,
+              bool bUseTaggedPDF = true);
+};
+
+void PdfExportTest::saveAsPDF(std::u16string_view rFile)
+{
+    // Import the bugdoc and export as PDF.
+    loadFromFile(rFile);
+    uno::Reference<frame::XStorable> xStorable(mxComponent, uno::UNO_QUERY);
+    xStorable->storeToURL(maTempFile.GetURL(), aMediaDescriptor.getAsConstPropertyValueList());
+}
+
+void PdfExportTest::load(std::u16string_view rFile, vcl::filter::PDFDocument& rDocument,
+                         bool bUseTaggedPDF)
+{
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    uno::Sequence<beans::PropertyValue> aFilterData(
+        comphelper::InitPropertySequence({ { "UseTaggedPDF", uno::Any(bUseTaggedPDF) } }));
+    aMediaDescriptor["FilterData"] <<= aFilterData;
+    saveAsPDF(rFile);
+
+    // Parse the export result.
+    SvFileStream aStream(maTempFile.GetURL(), StreamMode::READ);
+    CPPUNIT_ASSERT(rDocument.Read(aStream));
+}
+
+/// Tests that a pdf image is roundtripped back to PDF as a vector format.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf106059)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    // Explicitly enable the usage of the reference XObject markup.
+    uno::Sequence<beans::PropertyValue> aFilterData(
+        comphelper::InitPropertySequence({ { "UseReferenceXObject", uno::Any(true) } }));
+    aMediaDescriptor["FilterData"] <<= aFilterData;
+    saveAsPDF(u"tdf106059.odt");
+
+    // Parse the export result.
+    vcl::filter::PDFDocument aDocument;
+    SvFileStream aStream(maTempFile.GetURL(), StreamMode::READ);
+    CPPUNIT_ASSERT(aDocument.Read(aStream));
+
+    // Assert that the XObject in the page resources dictionary is a reference XObject.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+    vcl::filter::PDFObjectElement* pResources = aPages[0]->LookupObject("Resources"_ostr);
+    CPPUNIT_ASSERT(pResources);
+    auto pXObjects
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pResources->Lookup("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjects);
+    // The page has one image.
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pXObjects->GetItems().size());
+    vcl::filter::PDFObjectElement* pReferenceXObject
+        = pXObjects->LookupObject(pXObjects->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pReferenceXObject);
+    // The image is a reference XObject.
+    // This dictionary key was missing, so the XObject wasn't a reference one.
+    CPPUNIT_ASSERT(pReferenceXObject->Lookup("Ref"_ostr));
+}
+
+/// Tests export of PDF images without reference XObjects.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf106693)
+{
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf106693.odt", aDocument);
+
+    // Assert that the XObject in the page resources dictionary is a form XObject.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+    vcl::filter::PDFObjectElement* pResources = aPages[0]->LookupObject("Resources"_ostr);
+    CPPUNIT_ASSERT(pResources);
+    auto pXObjects
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pResources->Lookup("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjects);
+    // The page has one image.
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pXObjects->GetItems().size());
+    vcl::filter::PDFObjectElement* pXObject
+        = pXObjects->LookupObject(pXObjects->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pXObject);
+    // The image is a form XObject.
+    auto pSubtype = dynamic_cast<vcl::filter::PDFNameElement*>(pXObject->Lookup("Subtype"_ostr));
+    CPPUNIT_ASSERT(pSubtype);
+    CPPUNIT_ASSERT_EQUAL("Form"_ostr, pSubtype->GetValue());
+    // This failed: UseReferenceXObject was ignored and Ref was always created.
+    CPPUNIT_ASSERT(!pXObject->Lookup("Ref"_ostr));
+
+    // Assert that the form object refers to an inner form object, not a
+    // bitmap.
+    auto pInnerResources
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pXObject->Lookup("Resources"_ostr));
+    CPPUNIT_ASSERT(pInnerResources);
+    auto pInnerXObjects = dynamic_cast<vcl::filter::PDFDictionaryElement*>(
+        pInnerResources->LookupElement("XObject"_ostr));
+    CPPUNIT_ASSERT(pInnerXObjects);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pInnerXObjects->GetItems().size());
+    vcl::filter::PDFObjectElement* pInnerXObject
+        = pInnerXObjects->LookupObject(pInnerXObjects->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pInnerXObject);
+    auto pInnerSubtype
+        = dynamic_cast<vcl::filter::PDFNameElement*>(pInnerXObject->Lookup("Subtype"_ostr));
+    CPPUNIT_ASSERT(pInnerSubtype);
+    // This failed: this was Image (bitmap), not Form (vector).
+    CPPUNIT_ASSERT_EQUAL("Form"_ostr, pInnerSubtype->GetValue());
+}
+
+/// Tests that text highlight from Impress is not lost.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf105461)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf105461.odp");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // Make sure there is a filled rectangle inside.
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    int nYellowPathCount = 0;
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPdfPageObject = pPdfPage->getObject(i);
+        if (pPdfPageObject->getType() != vcl::pdf::PDFPageObjectType::Path)
+            continue;
+
+        if (pPdfPageObject->getFillColor() == COL_YELLOW)
+            ++nYellowPathCount;
+    }
+
+    // This was 0, the page contained no yellow paths.
+    CPPUNIT_ASSERT_EQUAL(1, nYellowPathCount);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf107868)
+{
+// No need to run it on Windows, since it would use GDI printing, and not trigger PDF export
+// which is the intent of the test.
+// FIXME: Why does this fail on macOS?
+#if !defined MACOSX && !defined _WIN32
+
+    // Import the bugdoc and print to PDF.
+    loadFromFile(u"tdf107868.odt");
+    uno::Reference<view::XPrintable> xPrintable(mxComponent, uno::UNO_QUERY);
+    CPPUNIT_ASSERT(xPrintable.is());
+    uno::Sequence<beans::PropertyValue> aOptions(comphelper::InitPropertySequence(
+        { { "FileName", uno::Any(maTempFile.GetURL()) }, { "Wait", uno::Any(true) } }));
+    xPrintable->print(aOptions);
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+    if (!pPdfDocument)
+        // Printing to PDF failed in a non-interesting way, e.g. CUPS is not
+        // running, there is no printer defined, etc.
+        return;
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // Make sure there is no filled rectangle inside.
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    int nWhitePathCount = 0;
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPdfPageObject = pPdfPage->getObject(i);
+        if (pPdfPageObject->getType() != vcl::pdf::PDFPageObjectType::Path)
+            continue;
+
+        if (pPdfPageObject->getFillColor() == COL_WHITE)
+            ++nWhitePathCount;
+    }
+
+    // This was 4, the page contained 4 white paths at problematic positions.
+    CPPUNIT_ASSERT_EQUAL(0, nWhitePathCount);
+#endif
+}
+
+/// Tests that embedded video from Impress is not exported as a linked one.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf105093)
+{
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf105093.odp", aDocument);
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    // Get page annotations.
+    auto pAnnots = dynamic_cast<vcl::filter::PDFArrayElement*>(aPages[0]->Lookup("Annots"_ostr));
+    CPPUNIT_ASSERT(pAnnots);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pAnnots->GetElements().size());
+    auto pAnnotReference
+        = dynamic_cast<vcl::filter::PDFReferenceElement*>(pAnnots->GetElements()[0]);
+    CPPUNIT_ASSERT(pAnnotReference);
+    vcl::filter::PDFObjectElement* pAnnot = pAnnotReference->LookupObject();
+    CPPUNIT_ASSERT(pAnnot);
+    CPPUNIT_ASSERT_EQUAL(
+        "Annot"_ostr,
+        static_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr))->GetValue());
+
+    // Get the Action -> Rendition -> MediaClip -> FileSpec.
+    auto pAction = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pAnnot->Lookup("A"_ostr));
+    CPPUNIT_ASSERT(pAction);
+    auto pRendition
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pAction->LookupElement("R"_ostr));
+    CPPUNIT_ASSERT(pRendition);
+    auto pMediaClip
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pRendition->LookupElement("C"_ostr));
+    CPPUNIT_ASSERT(pMediaClip);
+    auto pFileSpec
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pMediaClip->LookupElement("D"_ostr));
+    CPPUNIT_ASSERT(pFileSpec);
+    // Make sure the filespec refers to an embedded file.
+    // This key was missing, the embedded video was handled as a linked one.
+    CPPUNIT_ASSERT(pFileSpec->LookupElement("EF"_ostr));
+}
+
+/// Tests export of non-PDF images.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf106206)
+{
+    // Import the bugdoc and export as PDF.
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf106206.odt", aDocument);
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    // The page has a stream.
+    vcl::filter::PDFObjectElement* pContents = aPages[0]->LookupObject("Contents"_ostr);
+    CPPUNIT_ASSERT(pContents);
+    vcl::filter::PDFStreamElement* pStream = pContents->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());
+
+    // Make sure there is an image reference there.
+    OString aImage("/Im"_ostr);
+    auto pStart = static_cast<const char*>(aUncompressed.GetData());
+    const char* pEnd = pStart + aUncompressed.GetSize();
+    auto it = std::search(pStart, pEnd, aImage.getStr(), aImage.getStr() + aImage.getLength());
+    CPPUNIT_ASSERT(it != pEnd);
+
+    // And also that it's not an invalid one.
+    OString aInvalidImage("/Im0"_ostr);
+    it = std::search(pStart, pEnd, aInvalidImage.getStr(),
+                     aInvalidImage.getStr() + aInvalidImage.getLength());
+    // This failed, object #0 was referenced.
+    CPPUNIT_ASSERT(bool(it == pEnd));
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf127217)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf127217.odt");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // The page has one annotation.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfPage->getAnnotationCount());
+    std::unique_ptr<vcl::pdf::PDFiumAnnotation> pAnnot = pPdfPage->getAnnotation(0);
+
+    // Without the fix in place, this test would have failed here
+    CPPUNIT_ASSERT(!pAnnot->hasKey("DA"_ostr));
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf109143)
+{
+    // Import the bugdoc and export as PDF.
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf109143.odt", aDocument);
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    // Get access to the only image on the only page.
+    vcl::filter::PDFObjectElement* pResources = aPages[0]->LookupObject("Resources"_ostr);
+    CPPUNIT_ASSERT(pResources);
+    auto pXObjects
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pResources->Lookup("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjects);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pXObjects->GetItems().size());
+    vcl::filter::PDFObjectElement* pXObject
+        = pXObjects->LookupObject(pXObjects->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pXObject);
+
+    // Make sure it's re-compressed.
+    auto pLength = dynamic_cast<vcl::filter::PDFNumberElement*>(pXObject->Lookup("Length"_ostr));
+    CPPUNIT_ASSERT(pLength);
+    int nLength = pLength->GetValue();
+    // This failed: cropped TIFF-in-JPEG wasn't re-compressed, so crop was
+    // lost. Size was 59416, now is 11827.
+    CPPUNIT_ASSERT(nLength < 50000);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf106972)
+{
+    // Import the bugdoc and export as PDF.
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf106972.odt", aDocument);
+
+    // Get access to the only form object on the only page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+    vcl::filter::PDFObjectElement* pResources = aPages[0]->LookupObject("Resources"_ostr);
+    CPPUNIT_ASSERT(pResources);
+    auto pXObjects
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pResources->Lookup("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjects);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pXObjects->GetItems().size());
+    vcl::filter::PDFObjectElement* pXObject
+        = pXObjects->LookupObject(pXObjects->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pXObject);
+
+    // Get access to the only image inside the form object.
+    auto pFormResources
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pXObject->Lookup("Resources"_ostr));
+    CPPUNIT_ASSERT(pFormResources);
+    auto pImages = dynamic_cast<vcl::filter::PDFDictionaryElement*>(
+        pFormResources->LookupElement("XObject"_ostr));
+    CPPUNIT_ASSERT(pImages);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pImages->GetItems().size());
+    vcl::filter::PDFObjectElement* pImage
+        = pImages->LookupObject(pImages->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pImage);
+
+    // Assert resources of the image.
+    auto pImageResources
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pImage->Lookup("Resources"_ostr));
+    CPPUNIT_ASSERT(pImageResources);
+    // This failed: the PDF image had no Font resource.
+    CPPUNIT_ASSERT(pImageResources->LookupElement("Font"_ostr));
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf106972Pdf17)
+{
+    // Import the bugdoc and export as PDF.
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf106972-pdf17.odt", aDocument);
+
+    // Get access to the only image on the only page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+    vcl::filter::PDFObjectElement* pResources = aPages[0]->LookupObject("Resources"_ostr);
+    CPPUNIT_ASSERT(pResources);
+    auto pXObjects
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pResources->Lookup("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjects);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pXObjects->GetItems().size());
+    vcl::filter::PDFObjectElement* pXObject
+        = pXObjects->LookupObject(pXObjects->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pXObject);
+
+    // Assert that we now attempt to preserve the original PDF data, even if
+    // the original input was PDF >= 1.4.
+    CPPUNIT_ASSERT(pXObject->Lookup("Resources"_ostr));
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testSofthyphenPos)
+{
+// No need to run it on Windows, since it would use GDI printing, and not trigger PDF export
+// which is the intent of the test.
+// FIXME: Why does this fail on macOS?
+#if !defined MACOSX && !defined _WIN32
+
+    // Import the bugdoc and print to PDF.
+    loadFromFile(u"softhyphen_pdf.odt");
+    uno::Reference<view::XPrintable> xPrintable(mxComponent, uno::UNO_QUERY);
+    CPPUNIT_ASSERT(xPrintable.is());
+    uno::Sequence<beans::PropertyValue> aOptions(comphelper::InitPropertySequence(
+        { { "FileName", uno::Any(maTempFile.GetURL()) }, { "Wait", uno::Any(true) } }));
+    xPrintable->print(aOptions);
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+    if (!pPdfDocument)
+        // Printing to PDF failed in a non-interesting way, e.g. CUPS is not
+        // running, there is no printer defined, etc.
+        return;
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // tdf#96892 incorrect fractional part of font size caused soft-hyphen to
+    // be positioned inside preceding text (incorrect = 11.1, correct = 11.05)
+
+    // there are 3 texts currently, for line 1, soft-hyphen, line 2
+    bool haveText(false);
+
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPdfPageObject = pPdfPage->getObject(i);
+        CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFPageObjectType::Text, pPdfPageObject->getType());
+        haveText = true;
+        double const size = pPdfPageObject->getFontSize();
+        CPPUNIT_ASSERT_DOUBLES_EQUAL(11.05, size, 1E-06);
+    }
+
+    CPPUNIT_ASSERT(haveText);
+#endif
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf107013)
+{
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf107013.odt", aDocument);
+
+    // Get access to the only image on the only page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+    vcl::filter::PDFObjectElement* pResources = aPages[0]->LookupObject("Resources"_ostr);
+    CPPUNIT_ASSERT(pResources);
+    auto pXObjects
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pResources->Lookup("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjects);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pXObjects->GetItems().size());
+    vcl::filter::PDFObjectElement* pXObject
+        = pXObjects->LookupObject(pXObjects->GetItems().begin()->first);
+    // This failed, the reference to the image was created, but not the image.
+    CPPUNIT_ASSERT(pXObject);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf107018)
+{
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf107018.odt", aDocument);
+
+    // Get access to the only image on the only page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+    vcl::filter::PDFObjectElement* pResources = aPages[0]->LookupObject("Resources"_ostr);
+    CPPUNIT_ASSERT(pResources);
+    auto pXObjects
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pResources->Lookup("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjects);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pXObjects->GetItems().size());
+    vcl::filter::PDFObjectElement* pXObject
+        = pXObjects->LookupObject(pXObjects->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pXObject);
+
+    // Get access to the form object inside the image.
+    auto pXObjectResources
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pXObject->Lookup("Resources"_ostr));
+    CPPUNIT_ASSERT(pXObjectResources);
+    auto pXObjectForms = dynamic_cast<vcl::filter::PDFDictionaryElement*>(
+        pXObjectResources->LookupElement("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjectForms);
+    vcl::filter::PDFObjectElement* pForm
+        = pXObjectForms->LookupObject(pXObjectForms->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pForm);
+
+    // Get access to Resources -> Font -> F1 of the form.
+    auto pFormResources
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pForm->Lookup("Resources"_ostr));
+    CPPUNIT_ASSERT(pFormResources);
+    auto pFonts = dynamic_cast<vcl::filter::PDFDictionaryElement*>(
+        pFormResources->LookupElement("Font"_ostr));
+    CPPUNIT_ASSERT(pFonts);
+    auto pF1Ref = dynamic_cast<vcl::filter::PDFReferenceElement*>(pFonts->LookupElement("F1"_ostr));
+    CPPUNIT_ASSERT(pF1Ref);
+    vcl::filter::PDFObjectElement* pF1 = pF1Ref->LookupObject();
+    CPPUNIT_ASSERT(pF1);
+
+    // Check that Foo -> Bar of the font is of type Pages.
+    auto pFontFoo = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pF1->Lookup("Foo"_ostr));
+    CPPUNIT_ASSERT(pFontFoo);
+    auto pBar
+        = dynamic_cast<vcl::filter::PDFReferenceElement*>(pFontFoo->LookupElement("Bar"_ostr));
+    CPPUNIT_ASSERT(pBar);
+    vcl::filter::PDFObjectElement* pObject = pBar->LookupObject();
+    CPPUNIT_ASSERT(pObject);
+    auto pName = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT(pName);
+    // This was "XObject", reference in a nested dictionary wasn't updated when
+    // copying the page stream of a PDF image.
+    CPPUNIT_ASSERT_EQUAL("Pages"_ostr, pName->GetValue());
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf148706)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf148706.odt");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // The page has one annotation.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfPage->getAnnotationCount());
+    std::unique_ptr<vcl::pdf::PDFiumAnnotation> pAnnot = pPdfPage->getAnnotation(0);
+
+    CPPUNIT_ASSERT(pAnnot->hasKey("V"_ostr));
+    CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFObjectType::String, pAnnot->getValueType("V"_ostr));
+    OUString aV = pAnnot->getString("V"_ostr);
+
+    // Without the fix in place, this test would have failed with
+    // - Expected: 1821.84
+    // - Actual  :
+    CPPUNIT_ASSERT_EQUAL(OUString("1821.84"), aV);
+
+    CPPUNIT_ASSERT(pAnnot->hasKey("DV"_ostr));
+    CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFObjectType::String, pAnnot->getValueType("DV"_ostr));
+    OUString aDV = pAnnot->getString("DV"_ostr);
+
+    CPPUNIT_ASSERT_EQUAL(OUString("1821.84"), aDV);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf107089)
+{
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf107089.odt", aDocument);
+
+    // Get access to the only image on the only page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+    vcl::filter::PDFObjectElement* pResources = aPages[0]->LookupObject("Resources"_ostr);
+    CPPUNIT_ASSERT(pResources);
+    auto pXObjects
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pResources->Lookup("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjects);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), pXObjects->GetItems().size());
+    vcl::filter::PDFObjectElement* pXObject
+        = pXObjects->LookupObject(pXObjects->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pXObject);
+
+    // Get access to the form object inside the image.
+    auto pXObjectResources
+        = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pXObject->Lookup("Resources"_ostr));
+    CPPUNIT_ASSERT(pXObjectResources);
+    auto pXObjectForms = dynamic_cast<vcl::filter::PDFDictionaryElement*>(
+        pXObjectResources->LookupElement("XObject"_ostr));
+    CPPUNIT_ASSERT(pXObjectForms);
+    vcl::filter::PDFObjectElement* pForm
+        = pXObjectForms->LookupObject(pXObjectForms->GetItems().begin()->first);
+    CPPUNIT_ASSERT(pForm);
+
+    // Make sure 'Hello' is part of the form object's stream.
+    vcl::filter::PDFStreamElement* pStream = pForm->GetStream();
+    CPPUNIT_ASSERT(pStream);
+    SvMemoryStream aObjectStream;
+    ZCodec aZCodec;
+    aZCodec.BeginCompression();
+    pStream->GetMemory().Seek(0);
+    aZCodec.Decompress(pStream->GetMemory(), aObjectStream);
+    CPPUNIT_ASSERT(aZCodec.EndCompression());
+    aObjectStream.Seek(0);
+    OString aHello("Hello"_ostr);
+    auto pStart = static_cast<const char*>(aObjectStream.GetData());
+    const char* pEnd = pStart + aObjectStream.GetSize();
+    auto it = std::search(pStart, pEnd, aHello.getStr(), aHello.getStr() + aHello.getLength());
+    // This failed, 'Hello' was part only a mixed compressed/uncompressed stream, i.e. garbage.
+    CPPUNIT_ASSERT(it != pEnd);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf99680)
+{
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf99680.odt", aDocument);
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    // The page 1 has a stream.
+    vcl::filter::PDFObjectElement* pContents = aPages[0]->LookupObject("Contents"_ostr);
+    CPPUNIT_ASSERT(pContents);
+    vcl::filter::PDFStreamElement* pStream = pContents->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());
+
+    // tdf#130150 See infos in task - short: tdf#99680 was not the
+    // correct fix, so empty clip regions are valid - allow again in tests
+    //      Make sure there are no empty clipping regions.
+    //      OString aEmptyRegion("0 0 m h W* n");
+    //      auto it = std::search(pStart, pEnd, aEmptyRegion.getStr(), aEmptyRegion.getStr() + aEmptyRegion.getLength());
+    //      CPPUNIT_ASSERT_EQUAL_MESSAGE("Empty clipping region detected!", it, pEnd);
+
+    // Count save graphic state (q) and restore (Q) operators
+    // and ensure their amount is equal
+    auto pStart = static_cast<const char*>(aUncompressed.GetData());
+    const char* pEnd = pStart + aUncompressed.GetSize();
+    size_t nSaveCount = std::count(pStart, pEnd, 'q');
+    size_t nRestoreCount = std::count(pStart, pEnd, 'Q');
+    CPPUNIT_ASSERT_EQUAL_MESSAGE("Save/restore graphic state operators count mismatch!", nSaveCount,
+                                 nRestoreCount);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf99680_2)
+{
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf99680-2.odt", aDocument);
+
+    // For each document page
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(3), aPages.size());
+    for (size_t nPageNr = 0; nPageNr < aPages.size(); nPageNr++)
+    {
+        // Get page contents and stream.
+        vcl::filter::PDFObjectElement* pContents = aPages[nPageNr]->LookupObject("Contents"_ostr);
+        CPPUNIT_ASSERT(pContents);
+        vcl::filter::PDFStreamElement* pStream = pContents->GetStream();
+        CPPUNIT_ASSERT(pStream);
+        SvMemoryStream& rObjectStream = pStream->GetMemory();
+
+        // Uncompress the stream.
+        SvMemoryStream aUncompressed;
+        ZCodec aZCodec;
+        aZCodec.BeginCompression();
+        rObjectStream.Seek(0);
+        aZCodec.Decompress(rObjectStream, aUncompressed);
+        CPPUNIT_ASSERT(aZCodec.EndCompression());
+
+        // tdf#130150 See infos in task - short: tdf#99680 was not the
+        // correct fix, so empty clip regions are valid - allow again in tests
+        //      Make sure there are no empty clipping regions.
+        //      OString aEmptyRegion("0 0 m h W* n");
+        //      auto it = std::search(pStart, pEnd, aEmptyRegion.getStr(), aEmptyRegion.getStr() + aEmptyRegion.getLength());
+        //      CPPUNIT_ASSERT_EQUAL_MESSAGE("Empty clipping region detected!", it, pEnd);
+
+        // Count save graphic state (q) and restore (Q) operators
+        // and ensure their amount is equal
+        auto pStart = static_cast<const char*>(aUncompressed.GetData());
+        const char* pEnd = pStart + aUncompressed.GetSize();
+        size_t nSaveCount = std::count(pStart, pEnd, 'q');
+        size_t nRestoreCount = std::count(pStart, pEnd, 'Q');
+        CPPUNIT_ASSERT_EQUAL_MESSAGE("Save/restore graphic state operators count mismatch!",
+                                     nSaveCount, nRestoreCount);
+    }
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf108963)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf108963.odp");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // Test page size (28x15.75 cm, was 1/100th mm off, tdf#112690)
+    // bad: MediaBox[0 0 793.672440944882 446.428346456693]
+    // good: MediaBox[0 0 793.700787401575 446.456692913386]
+    const double aWidth = pPdfPage->getWidth();
+    CPPUNIT_ASSERT_DOUBLES_EQUAL(793.7, aWidth, 0.01);
+    const double aHeight = pPdfPage->getHeight();
+    CPPUNIT_ASSERT_DOUBLES_EQUAL(446.46, aHeight, 0.01);
+
+    // Make sure there is a filled rectangle inside.
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    int nYellowPathCount = 0;
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPdfPageObject = pPdfPage->getObject(i);
+        if (pPdfPageObject->getType() != vcl::pdf::PDFPageObjectType::Path)
+            continue;
+
+        if (pPdfPageObject->getFillColor() == COL_YELLOW)
+        {
+            ++nYellowPathCount;
+            // The path described a yellow rectangle, but it was not rotated.
+            int nSegments = pPdfPageObject->getPathSegmentCount();
+            CPPUNIT_ASSERT_EQUAL(5, nSegments);
+            std::unique_ptr<vcl::pdf::PDFiumPathSegment> pSegment
+                = pPdfPageObject->getPathSegment(0);
+            CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFSegmentType::Moveto, pSegment->getType());
+            basegfx::B2DPoint aPoint = pSegment->getPoint();
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(245, aPoint.getX(), 0.999);
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(244, aPoint.getY(), 0.999);
+            CPPUNIT_ASSERT(!pSegment->isClosed());
+
+            pSegment = pPdfPageObject->getPathSegment(1);
+            CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFSegmentType::Lineto, pSegment->getType());
+            aPoint = pSegment->getPoint();
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(275, aPoint.getX(), 0.999);
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(267, aPoint.getY(), 0.999);
+            CPPUNIT_ASSERT(!pSegment->isClosed());
+
+            pSegment = pPdfPageObject->getPathSegment(2);
+            CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFSegmentType::Lineto, pSegment->getType());
+            aPoint = pSegment->getPoint();
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(287, aPoint.getX(), 0.999);
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(251, aPoint.getY(), 0.999);
+            CPPUNIT_ASSERT(!pSegment->isClosed());
+
+            pSegment = pPdfPageObject->getPathSegment(3);
+            CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFSegmentType::Lineto, pSegment->getType());
+            aPoint = pSegment->getPoint();
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(257, aPoint.getX(), 0.999);
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(228, aPoint.getY(), 0.999);
+            CPPUNIT_ASSERT(!pSegment->isClosed());
+
+            pSegment = pPdfPageObject->getPathSegment(4);
+            CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFSegmentType::Lineto, pSegment->getType());
+            aPoint = pSegment->getPoint();
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(245, aPoint.getX(), 0.999);
+            CPPUNIT_ASSERT_DOUBLES_EQUAL(244, aPoint.getY(), 0.999);
+            CPPUNIT_ASSERT(pSegment->isClosed());
+        }
+    }
+
+    CPPUNIT_ASSERT_EQUAL(1, nYellowPathCount);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testAlternativeText)
+{
+    aMediaDescriptor["FilterName"] <<= OUString("impress_pdf_Export");
+
+    uno::Sequence<beans::PropertyValue> aFilterData(
+        comphelper::InitPropertySequence({ { "UseTaggedPDF", uno::Any(true) } }));
+    aMediaDescriptor["FilterData"] <<= aFilterData;
+    saveAsPDF(u"alternativeText.fodp");
+
+    // Parse the export result.
+    vcl::filter::PDFDocument aDocument;
+    SvFileStream aStream(maTempFile.GetURL(), StreamMode::READ);
+    CPPUNIT_ASSERT(aDocument.Read(aStream));
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    for (const auto& aElement : aDocument.GetElements())
+    {
+        auto pObject = dynamic_cast<vcl::filter::PDFObjectElement*>(aElement.get());
+        if (!pObject)
+            continue;
+        auto pType = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("Type"_ostr));
+        if (pType && pType->GetValue() == "StructElem")
+        {
+            auto pS = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("S"_ostr));
+            if (pS && pS->GetValue() == "Figure")
+            {
+                CPPUNIT_ASSERT_EQUAL(u"This is the text alternative - This is the description"_ustr,
+                                     ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(
+                                         *dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                                             pObject->Lookup("Alt"_ostr))));
+            }
+        }
+    }
+
+    // tdf#67866 check that Catalog contains Lang
+    auto* pCatalog = aDocument.GetCatalog();
+    CPPUNIT_ASSERT(pCatalog);
+    auto* pCatalogDictionary = pCatalog->GetDictionary();
+    CPPUNIT_ASSERT(pCatalogDictionary);
+    auto pLang = dynamic_cast<vcl::filter::PDFLiteralStringElement*>(
+        pCatalogDictionary->LookupElement("Lang"_ostr));
+    CPPUNIT_ASSERT(pLang);
+    CPPUNIT_ASSERT_EQUAL("en-US"_ostr, pLang->GetValue());
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf105972)
+{
+    vcl::filter::PDFDocument aDocument;
+    // Loading fails with tagged PDF enabled
+    load(u"tdf105972.fodt", aDocument, false);
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    auto pAnnots = dynamic_cast<vcl::filter::PDFArrayElement*>(aPages[0]->Lookup("Annots"_ostr));
+    CPPUNIT_ASSERT(pAnnots);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(3), pAnnots->GetElements().size());
+
+    sal_uInt32 nTextFieldCount = 0;
+    for (const auto& aElement : aDocument.GetElements())
+    {
+        auto pObject = dynamic_cast<vcl::filter::PDFObjectElement*>(aElement.get());
+        if (!pObject)
+            continue;
+        auto pType = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("FT"_ostr));
+        if (pType && pType->GetValue() == "Tx")
+        {
+            ++nTextFieldCount;
+
+            auto pT
+                = dynamic_cast<vcl::filter::PDFLiteralStringElement*>(pObject->Lookup("T"_ostr));
+            CPPUNIT_ASSERT(pT);
+            auto pAA = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pObject->Lookup("AA"_ostr));
+            CPPUNIT_ASSERT(pAA);
+            CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(2), pAA->GetItems().size());
+            auto pF
+                = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pAA->LookupElement("F"_ostr));
+            CPPUNIT_ASSERT(pF);
+            CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(2), pF->GetItems().size());
+
+            if (nTextFieldCount == 1)
+            {
+                CPPUNIT_ASSERT_EQUAL("CurrencyField"_ostr, pT->GetValue());
+
+                auto pJS = dynamic_cast<vcl::filter::PDFLiteralStringElement*>(
+                    pF->LookupElement("JS"_ostr));
+                CPPUNIT_ASSERT_EQUAL("AFNumber_Format\\(4, 0, 0, 0, \"\\\\u20ac\",true\\);"_ostr,
+                                     pJS->GetValue());
+            }
+            else if (nTextFieldCount == 2)
+            {
+                CPPUNIT_ASSERT_EQUAL("TimeField"_ostr, pT->GetValue());
+
+                auto pJS = dynamic_cast<vcl::filter::PDFLiteralStringElement*>(
+                    pF->LookupElement("JS"_ostr));
+                CPPUNIT_ASSERT_EQUAL("AFTime_FormatEx\\(\"h:MM:sstt\"\\);"_ostr, pJS->GetValue());
+            }
+            else
+            {
+                CPPUNIT_ASSERT_EQUAL("DateField"_ostr, pT->GetValue());
+
+                auto pJS = dynamic_cast<vcl::filter::PDFLiteralStringElement*>(
+                    pF->LookupElement("JS"_ostr));
+                CPPUNIT_ASSERT_EQUAL("AFDate_FormatEx\\(\"yy-mm-dd\"\\);"_ostr, pJS->GetValue());
+            }
+        }
+    }
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf148442)
+{
+    vcl::filter::PDFDocument aDocument;
+    // Loading fails with tagged PDF enabled
+    load(u"tdf148442.odt", aDocument, false);
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    auto pAnnots = dynamic_cast<vcl::filter::PDFArrayElement*>(aPages[0]->Lookup("Annots"_ostr));
+    CPPUNIT_ASSERT(pAnnots);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(3), pAnnots->GetElements().size());
+
+    sal_uInt32 nBtnCount = 0;
+    for (const auto& aElement : aDocument.GetElements())
+    {
+        auto pObject = dynamic_cast<vcl::filter::PDFObjectElement*>(aElement.get());
+        if (!pObject)
+            continue;
+        auto pType = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("FT"_ostr));
+        if (pType && pType->GetValue() == "Btn")
+        {
+            ++nBtnCount;
+            auto pT
+                = dynamic_cast<vcl::filter::PDFLiteralStringElement*>(pObject->Lookup("T"_ostr));
+            CPPUNIT_ASSERT(pT);
+            auto pAS = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("AS"_ostr));
+            CPPUNIT_ASSERT(pAS);
+
+            auto pAP = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pObject->Lookup("AP"_ostr));
+            CPPUNIT_ASSERT(pAP);
+            auto pN
+                = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pAP->LookupElement("N"_ostr));
+            CPPUNIT_ASSERT(pN);
+            CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(2), pN->GetItems().size());
+
+            if (nBtnCount == 1)
+            {
+                CPPUNIT_ASSERT_EQUAL("Checkbox1"_ostr, pT->GetValue());
+                CPPUNIT_ASSERT_EQUAL("Yes"_ostr, pAS->GetValue());
+                CPPUNIT_ASSERT(!pN->GetItems().count("ref"_ostr));
+                CPPUNIT_ASSERT(pN->GetItems().count("Yes"_ostr));
+                CPPUNIT_ASSERT(pN->GetItems().count("Off"_ostr));
+            }
+            else if (nBtnCount == 2)
+            {
+                CPPUNIT_ASSERT_EQUAL("Checkbox2"_ostr, pT->GetValue());
+                CPPUNIT_ASSERT_EQUAL("Yes"_ostr, pAS->GetValue());
+
+                // Without the fix in place, this test would have failed here
+                CPPUNIT_ASSERT(pN->GetItems().count("ref"_ostr));
+                CPPUNIT_ASSERT(!pN->GetItems().count("Yes"_ostr));
+                CPPUNIT_ASSERT(pN->GetItems().count("Off"_ostr));
+            }
+            else
+            {
+                CPPUNIT_ASSERT_EQUAL("Checkbox3"_ostr, pT->GetValue());
+                CPPUNIT_ASSERT_EQUAL("Off"_ostr, pAS->GetValue());
+                CPPUNIT_ASSERT(pN->GetItems().count("ref"_ostr));
+                CPPUNIT_ASSERT(!pN->GetItems().count("Yes"_ostr));
+
+                // tdf#143612: Without the fix in place, this test would have failed here
+                CPPUNIT_ASSERT(!pN->GetItems().count("Off"_ostr));
+                CPPUNIT_ASSERT(pN->GetItems().count("refOff"_ostr));
+            }
+        }
+    }
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf118244_radioButtonGroup)
+{
+    vcl::filter::PDFDocument aDocument;
+    // Loading fails with tagged PDF enabled
+    load(u"tdf118244_radioButtonGroup.odt", aDocument, false);
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    // There are eight radio buttons.
+    auto pAnnots = dynamic_cast<vcl::filter::PDFArrayElement*>(aPages[0]->Lookup("Annots"_ostr));
+    CPPUNIT_ASSERT(pAnnots);
+    CPPUNIT_ASSERT_EQUAL_MESSAGE("# of radio buttons", static_cast<size_t>(8),
+                                 pAnnots->GetElements().size());
+
+    sal_uInt32 nRadioGroups = 0;
+    for (const auto& aElement : aDocument.GetElements())
+    {
+        auto pObject = dynamic_cast<vcl::filter::PDFObjectElement*>(aElement.get());
+        if (!pObject)
+            continue;
+        auto pType = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("FT"_ostr));
+        if (pType && pType->GetValue() == "Btn")
+        {
+            auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObject->Lookup("Kids"_ostr));
+            if (pKids)
+            {
+                size_t expectedSize = 2;
+                ++nRadioGroups;
+                if (nRadioGroups == 3)
+                    expectedSize = 3;
+                CPPUNIT_ASSERT_EQUAL(expectedSize, pKids->GetElements().size());
+            }
+        }
+    }
+    CPPUNIT_ASSERT_EQUAL_MESSAGE("# of radio groups", sal_uInt32(3), nRadioGroups);
+}
+
+/// Test writing ToUnicode CMAP for LTR ligatures.
+// This requires Carlito font, if it is missing the test will most likely
+// fail.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf115117_1)
+{
+#if HAVE_MORE_FONTS
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf115117-1.odt", aDocument);
+
+    vcl::filter::PDFObjectElement* pToUnicode = nullptr;
+
+    // Get access to ToUnicode of the first font
+    for (const auto& aElement : aDocument.GetElements())
+    {
+        auto pObject = dynamic_cast<vcl::filter::PDFObjectElement*>(aElement.get());
+        if (!pObject)
+            continue;
+        auto pType = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("Type"_ostr));
+        if (pType && pType->GetValue() == "Font")
+        {
+            auto pToUnicodeRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                pObject->Lookup("ToUnicode"_ostr));
+            CPPUNIT_ASSERT(pToUnicodeRef);
+            pToUnicode = pToUnicodeRef->LookupObject();
+            break;
+        }
+    }
+
+    CPPUNIT_ASSERT(pToUnicode);
+    auto pStream = pToUnicode->GetStream();
+    CPPUNIT_ASSERT(pStream);
+    SvMemoryStream aObjectStream;
+    ZCodec aZCodec;
+    aZCodec.BeginCompression();
+    pStream->GetMemory().Seek(0);
+    aZCodec.Decompress(pStream->GetMemory(), aObjectStream);
+    CPPUNIT_ASSERT(aZCodec.EndCompression());
+    aObjectStream.Seek(0);
+    // The first values, <01> <02> etc., are glyph ids, they might change order
+    // if we changed how font subsets are created.
+    // The second values, <00740069> etc., are Unicode code points in hex,
+    // <00740069> is U+0074 and U+0069 i.e. "ti" which is a ligature in
+    // Carlito/Calibri. This test is failing if any of the second values
+    // changed which means we are not detecting ligatures and writing CMAP
+    // entries for them correctly. If glyph order in the subset changes then
+    // the order here will changes and the PDF has to be carefully inspected to
+    // ensure that the new values are correct before updating the string below.
+    OString aCmap("9 beginbfchar\n"
+                  "<01> <00740069>\n"
+                  "<02> <0020>\n"
+                  "<03> <0074>\n"
+                  "<04> <0065>\n"
+                  "<05> <0073>\n"
+                  "<06> <00660069>\n"
+                  "<07> <0066006C>\n"
+                  "<08> <006600660069>\n"
+                  "<09> <00660066006C>\n"
+                  "endbfchar"_ostr);
+    auto pStart = static_cast<const char*>(aObjectStream.GetData());
+    const char* pEnd = pStart + aObjectStream.GetSize();
+    auto it = std::search(pStart, pEnd, aCmap.getStr(), aCmap.getStr() + aCmap.getLength());
+    CPPUNIT_ASSERT(it != pEnd);
+#endif
+}
+
+/// Test writing ToUnicode CMAP for RTL ligatures.
+// This requires DejaVu Sans font, if it is missing the test will most likely
+// fail.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf115117_2)
+{
+#if HAVE_MORE_FONTS
+    // See the comments in testTdf115117_1() for explanation.
+
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf115117-2.odt", aDocument);
+
+    vcl::filter::PDFObjectElement* pToUnicode = nullptr;
+
+    for (const auto& aElement : aDocument.GetElements())
+    {
+        auto pObject = dynamic_cast<vcl::filter::PDFObjectElement*>(aElement.get());
+        if (!pObject)
+            continue;
+        auto pType = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("Type"_ostr));
+        if (pType && pType->GetValue() == "Font")
+        {
+            auto pToUnicodeRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                pObject->Lookup("ToUnicode"_ostr));
+            CPPUNIT_ASSERT(pToUnicodeRef);
+            pToUnicode = pToUnicodeRef->LookupObject();
+            break;
+        }
+    }
+
+    CPPUNIT_ASSERT(pToUnicode);
+    auto pStream = pToUnicode->GetStream();
+    CPPUNIT_ASSERT(pStream);
+    SvMemoryStream aObjectStream;
+    ZCodec aZCodec;
+    aZCodec.BeginCompression();
+    pStream->GetMemory().Seek(0);
+    aZCodec.Decompress(pStream->GetMemory(), aObjectStream);
+    CPPUNIT_ASSERT(aZCodec.EndCompression());
+    aObjectStream.Seek(0);
+    OString aCmap("7 beginbfchar\n"
+                  "<01> <06440627>\n"
+                  "<02> <0020>\n"
+                  "<03> <0641>\n"
+                  "<04> <0642>\n"
+                  "<05> <0648>\n"
+                  "<06> <06440627>\n"
+                  "<07> <0628>\n"
+                  "endbfchar"_ostr);
+    auto pStart = static_cast<const char*>(aObjectStream.GetData());
+    const char* pEnd = pStart + aObjectStream.GetSize();
+    auto it = std::search(pStart, pEnd, aCmap.getStr(), aCmap.getStr() + aCmap.getLength());
+    CPPUNIT_ASSERT(it != pEnd);
+#endif
+}
+
+/// Text extracting LTR text with ligatures.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf115117_1a)
+{
+#if HAVE_MORE_FONTS
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf115117-1.odt");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pPdfTextPage = pPdfPage->getTextPage();
+    CPPUNIT_ASSERT(pPdfTextPage);
+
+    // Extract the text from the page. This pdfium API is a bit higher level
+    // than we want and might apply heuristic that give false positive, but it
+    // is a good approximation in addition to the check in testTdf115117_1().
+    int nChars = pPdfTextPage->countChars();
+    CPPUNIT_ASSERT_EQUAL(44, nChars);
+
+    std::vector<sal_uInt32> aChars(nChars);
+    for (int i = 0; i < nChars; i++)
+        aChars[i] = pPdfTextPage->getUnicode(i);
+    OUString aActualText(aChars.data(), aChars.size());
+    CPPUNIT_ASSERT_EQUAL(OUString("ti ti test ti\r\nti test fi fl ffi ffl test fi"), aActualText);
+#endif
+}
+
+/// Test extracting RTL text with ligatures.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf115117_2a)
+{
+#if HAVE_MORE_FONTS
+    // See the comments in testTdf115117_1a() for explanation.
+
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf115117-2.odt");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pPdfTextPage = pPdfPage->getTextPage();
+    CPPUNIT_ASSERT(pPdfTextPage);
+
+    int nChars = pPdfTextPage->countChars();
+    CPPUNIT_ASSERT_EQUAL(13, nChars);
+
+    std::vector<sal_uInt32> aChars(nChars);
+    for (int i = 0; i < nChars; i++)
+        aChars[i] = pPdfTextPage->getUnicode(i);
+    OUString aActualText(aChars.data(), aChars.size());
+    CPPUNIT_ASSERT_EQUAL(u"\u0627\u0644 \u0628\u0627\u0644 \u0648\u0642\u0641 \u0627\u0644"_ustr,
+                         aActualText);
+#endif
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf154549)
+{
+// FIXME: On Windows, the number of chars is 4 instead of 3
+#ifndef _WIN32
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf154549.odt");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pPdfTextPage = pPdfPage->getTextPage();
+    CPPUNIT_ASSERT(pPdfTextPage);
+
+    int nChars = pPdfTextPage->countChars();
+
+    CPPUNIT_ASSERT_EQUAL(3, nChars);
+
+    std::vector<sal_uInt32> aChars(nChars);
+    for (int i = 0; i < nChars; i++)
+        aChars[i] = pPdfTextPage->getUnicode(i);
+    OUString aActualText(aChars.data(), aChars.size());
+
+    // Without the fix in place, this test would have failed with
+    // - Expected: ִبي
+    // - Actual  : بִي
+    CPPUNIT_ASSERT_EQUAL(u"\u05B4\u0628\u064A"_ustr, aActualText);
+#endif
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf150846)
+{
+    // Without the fix in place, this test would have failed with
+    // An uncaught exception of type com.sun.star.io.IOException
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf150846.txt");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pPdfTextPage = pPdfPage->getTextPage();
+    CPPUNIT_ASSERT(pPdfTextPage);
+
+    int nChars = pPdfTextPage->countChars();
+
+    CPPUNIT_ASSERT_EQUAL(5, nChars);
+
+    std::vector<sal_uInt32> aChars(nChars);
+    for (int i = 0; i < nChars; i++)
+        aChars[i] = pPdfTextPage->getUnicode(i);
+    OUString aActualText(aChars.data(), aChars.size());
+    CPPUNIT_ASSERT_EQUAL(u"hello"_ustr, aActualText);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf103492)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf103492.odt");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has two page.
+    CPPUNIT_ASSERT_EQUAL(2, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage1 = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage1);
+
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pPdfTextPage1 = pPdfPage1->getTextPage();
+    CPPUNIT_ASSERT(pPdfTextPage1);
+
+    int nChars1 = pPdfTextPage1->countChars();
+
+    // Without the fix in place, this test would have failed with
+    // - Expected: 15
+    // - Actual  : 18
+    CPPUNIT_ASSERT_EQUAL(15, nChars1);
+
+    std::vector<sal_uInt32> aChars1(nChars1);
+    for (int i = 0; i < nChars1; i++)
+        aChars1[i] = pPdfTextPage1->getUnicode(i);
+    OUString aActualText1(aChars1.data(), aChars1.size());
+    CPPUNIT_ASSERT_EQUAL(u"يوسف My name is"_ustr, aActualText1);
+
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage2 = pPdfDocument->openPage(/*nIndex=*/1);
+    CPPUNIT_ASSERT(pPdfPage2);
+
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pPdfTextPage2 = pPdfPage2->getTextPage();
+    CPPUNIT_ASSERT(pPdfTextPage2);
+
+    int nChars2 = pPdfTextPage2->countChars();
+
+    CPPUNIT_ASSERT_EQUAL(15, nChars2);
+
+    std::vector<sal_uInt32> aChars2(nChars2);
+    for (int i = 0; i < nChars2; i++)
+        aChars2[i] = pPdfTextPage2->getUnicode(i);
+    OUString aActualText2(aChars2.data(), aChars2.size());
+    CPPUNIT_ASSERT_EQUAL(u"My name is يوسف"_ustr, aActualText2);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf145274)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf145274.docx");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+
+    auto pPage = pPdfDocument->openPage(0);
+    CPPUNIT_ASSERT(pPage);
+
+    int nPageObjectCount = pPage->getObjectCount();
+
+    // Without the fix in place, this test would have failed with
+    // - Expected: 6
+    // - Actual  : 4
+    CPPUNIT_ASSERT_EQUAL(6, nPageObjectCount);
+
+    auto pTextPage = pPage->getTextPage();
+
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPage->getObject(i);
+        if (pPageObject->getType() != vcl::pdf::PDFPageObjectType::Text)
+            continue;
+
+        CPPUNIT_ASSERT_EQUAL(11.0, pPageObject->getFontSize());
+        CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFTextRenderMode::Fill, pPageObject->getTextRenderMode());
+        CPPUNIT_ASSERT_EQUAL(COL_RED, pPageObject->getFillColor());
+    }
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf156685)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf156685.docx");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+
+    auto pPage = pPdfDocument->openPage(0);
+    CPPUNIT_ASSERT(pPage);
+
+    int nPageObjectCount = pPage->getObjectCount();
+
+    CPPUNIT_ASSERT_EQUAL(9, nPageObjectCount);
+
+    auto pTextPage = pPage->getTextPage();
+
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPage->getObject(i);
+        if (pPageObject->getType() != vcl::pdf::PDFPageObjectType::Text)
+            continue;
+
+        CPPUNIT_ASSERT_EQUAL(11.0, pPageObject->getFontSize());
+        CPPUNIT_ASSERT_EQUAL(vcl::pdf::PDFTextRenderMode::Fill, pPageObject->getTextRenderMode());
+
+        // Without the fix in place, this test would have failed with
+        // - Expected: rgba[000000ff]
+        // - Actual  : rgba[ffffffff]
+        CPPUNIT_ASSERT_EQUAL(COL_BLACK, pPageObject->getFillColor());
+    }
+}
+
+/// Test writing ToUnicode CMAP for doubly encoded glyphs.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf66597_1)
+{
+#if HAVE_MORE_FONTS
+    // This requires Amiri font, if it is missing the test will fail.
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf66597-1.odt", aDocument);
+
+    {
+        // Get access to ToUnicode of the first font
+        vcl::filter::PDFObjectElement* pToUnicode = nullptr;
+        for (const auto& aElement : aDocument.GetElements())
+        {
+            auto pObject = dynamic_cast<vcl::filter::PDFObjectElement*>(aElement.get());
+            if (!pObject)
+                continue;
+            auto pType = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("Type"_ostr));
+            if (pType && pType->GetValue() == "Font")
+            {
+                auto pName
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("BaseFont"_ostr));
+                auto aName = pName->GetValue().copy(7); // skip the subset id
+                CPPUNIT_ASSERT_EQUAL_MESSAGE("Unexpected font name", "Amiri-Regular"_ostr, aName);
+
+                auto pToUnicodeRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObject->Lookup("ToUnicode"_ostr));
+                CPPUNIT_ASSERT(pToUnicodeRef);
+                pToUnicode = pToUnicodeRef->LookupObject();
+                break;
+            }
+        }
+
+        CPPUNIT_ASSERT(pToUnicode);
+        auto pStream = pToUnicode->GetStream();
+        CPPUNIT_ASSERT(pStream);
+        SvMemoryStream aObjectStream;
+        ZCodec aZCodec;
+        aZCodec.BeginCompression();
+        pStream->GetMemory().Seek(0);
+        aZCodec.Decompress(pStream->GetMemory(), aObjectStream);
+        CPPUNIT_ASSERT(aZCodec.EndCompression());
+        aObjectStream.Seek(0);
+        // The <01> is glyph id, <2044> is code point.
+        // The document has two characters <2044><2215><2044>, but the font
+        // reuses the same glyph for U+2044 and U+2215 so we should have a single
+        // CMAP entry for the U+2044, and U+2215 will be handled with ActualText
+        // (tested below).
+        std::string aCmap("1 beginbfchar\n"
+                          "<01> <2044>\n"
+                          "endbfchar");
+        std::string aData(static_cast<const char*>(aObjectStream.GetData()),
+                          aObjectStream.GetSize());
+        auto nPos = aData.find(aCmap);
+        CPPUNIT_ASSERT(nPos != std::string::npos);
+    }
+
+    {
+        auto aPages = aDocument.GetPages();
+        CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+        // Get page contents and stream.
+        auto pContents = aPages[0]->LookupObject("Contents"_ostr);
+        CPPUNIT_ASSERT(pContents);
+        auto pStream = pContents->GetStream();
+        CPPUNIT_ASSERT(pStream);
+        auto& rObjectStream = pStream->GetMemory();
+
+        // Uncompress the stream.
+        SvMemoryStream aUncompressed;
+        ZCodec aZCodec;
+        aZCodec.BeginCompression();
+        rObjectStream.Seek(0);
+        aZCodec.Decompress(rObjectStream, aUncompressed);
+        CPPUNIT_ASSERT(aZCodec.EndCompression());
+
+        // Make sure the expected ActualText is present.
+        std::string aData(static_cast<const char*>(aUncompressed.GetData()),
+                          aUncompressed.GetSize());
+
+        std::string aActualText("/Span<</ActualText<");
+        size_t nCount = 0;
+        size_t nPos = 0;
+        while ((nPos = aData.find(aActualText, nPos)) != std::string::npos)
+        {
+            nCount++;
+            nPos += aActualText.length();
+        }
+        CPPUNIT_ASSERT_EQUAL_MESSAGE("The should be one ActualText entry!", static_cast<size_t>(1),
+                                     nCount);
+
+        aActualText = "/Span<</ActualText<FEFF2215>>>";
+        nPos = aData.find(aActualText);
+        CPPUNIT_ASSERT_MESSAGE("ActualText not found!", nPos != std::string::npos);
+    }
+#endif
+}
+
+/// Test writing ActualText for RTL many to one glyph to Unicode mapping.
+// This requires Reem Kufi font, if it is missing the test will fail.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf66597_2)
+{
+#if HAVE_MORE_FONTS
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf66597-2.odt", aDocument);
+
+    {
+        // Get access to ToUnicode of the first font
+        vcl::filter::PDFObjectElement* pToUnicode = nullptr;
+        for (const auto& aElement : aDocument.GetElements())
+        {
+            auto pObject = dynamic_cast<vcl::filter::PDFObjectElement*>(aElement.get());
+            if (!pObject)
+                continue;
+            auto pType = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("Type"_ostr));
+            if (pType && pType->GetValue() == "Font")
+            {
+                auto pName
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("BaseFont"_ostr));
+                auto aName = pName->GetValue().copy(7); // skip the subset id
+                CPPUNIT_ASSERT_EQUAL_MESSAGE("Unexpected font name", "ReemKufi-Regular"_ostr,
+                                             aName);
+
+                auto pToUnicodeRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObject->Lookup("ToUnicode"_ostr));
+                CPPUNIT_ASSERT(pToUnicodeRef);
+                pToUnicode = pToUnicodeRef->LookupObject();
+                break;
+            }
+        }
+
+        CPPUNIT_ASSERT(pToUnicode);
+        auto pStream = pToUnicode->GetStream();
+        CPPUNIT_ASSERT(pStream);
+        SvMemoryStream aObjectStream;
+        ZCodec aZCodec;
+        aZCodec.BeginCompression();
+        pStream->GetMemory().Seek(0);
+        aZCodec.Decompress(pStream->GetMemory(), aObjectStream);
+        CPPUNIT_ASSERT(aZCodec.EndCompression());
+        aObjectStream.Seek(0);
+        std::string aCmap("8 beginbfchar\n"
+                          "<02> <0632>\n"
+                          "<03> <0020>\n"
+                          "<04> <0648>\n"
+                          "<05> <0647>\n"
+                          "<06> <062F>\n"
+                          "<08> <062C>\n"
+                          "<0A> <0628>\n"
+                          "<0C> <0623>\n"
+                          "endbfchar");
+        std::string aData(static_cast<const char*>(aObjectStream.GetData()),
+                          aObjectStream.GetSize());
+        auto nPos = aData.find(aCmap);
+        CPPUNIT_ASSERT(nPos != std::string::npos);
+    }
+
+    {
+        auto aPages = aDocument.GetPages();
+        CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+        // Get page contents and stream.
+        auto pContents = aPages[0]->LookupObject("Contents"_ostr);
+        CPPUNIT_ASSERT(pContents);
+        auto pStream = pContents->GetStream();
+        CPPUNIT_ASSERT(pStream);
+        auto& rObjectStream = pStream->GetMemory();
+
+        // Uncompress the stream.
+        SvMemoryStream aUncompressed;
+        ZCodec aZCodec;
+        aZCodec.BeginCompression();
+        rObjectStream.Seek(0);
+        aZCodec.Decompress(rObjectStream, aUncompressed);
+        CPPUNIT_ASSERT(aZCodec.EndCompression());
+
+        // Make sure the expected ActualText is present.
+        std::string aData(static_cast<const char*>(aUncompressed.GetData()),
+                          aUncompressed.GetSize());
+
+        std::vector<std::string> aCodes({ "0632", "062C", "0628", "0623" });
+        std::string aActualText("/Span<</ActualText<");
+        size_t nCount = 0;
+        size_t nPos = 0;
+        while ((nPos = aData.find(aActualText, nPos)) != std::string::npos)
+        {
+            nCount++;
+            nPos += aActualText.length();
+        }
+        CPPUNIT_ASSERT_EQUAL_MESSAGE("Number of ActualText entries does not match!", aCodes.size(),
+                                     nCount);
+
+        for (const auto& aCode : aCodes)
+        {
+            aActualText = "/Span<</ActualText<FEFF" + aCode + ">>>";
+            nPos = aData.find(aActualText);
+            CPPUNIT_ASSERT_MESSAGE("ActualText not found for " + aCode, nPos != std::string::npos);
+        }
+    }
+#endif
+}
+
+/// Test writing ActualText for LTR many to one glyph to Unicode mapping.
+// This requires Gentium Basic font, if it is missing the test will fail.
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf66597_3)
+{
+#if HAVE_MORE_FONTS
+    vcl::filter::PDFDocument aDocument;
+    load(u"tdf66597-3.odt", aDocument);
+
+    {
+        // Get access to ToUnicode of the first font
+        vcl::filter::PDFObjectElement* pToUnicode = nullptr;
+        for (const auto& aElement : aDocument.GetElements())
+        {
+            auto pObject = dynamic_cast<vcl::filter::PDFObjectElement*>(aElement.get());
+            if (!pObject)
+                continue;
+            auto pType = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("Type"_ostr));
+            if (pType && pType->GetValue() == "Font")
+            {
+                auto pName
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObject->Lookup("BaseFont"_ostr));
+                auto aName = pName->GetValue().copy(7); // skip the subset id
+                CPPUNIT_ASSERT_EQUAL_MESSAGE("Unexpected font name", "GentiumBasic"_ostr, aName);
+
+                auto pToUnicodeRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObject->Lookup("ToUnicode"_ostr));
+                CPPUNIT_ASSERT(pToUnicodeRef);
+                pToUnicode = pToUnicodeRef->LookupObject();
+                break;
+            }
+        }
+
+        CPPUNIT_ASSERT(pToUnicode);
+        auto pStream = pToUnicode->GetStream();
+        CPPUNIT_ASSERT(pStream);
+        SvMemoryStream aObjectStream;
+        ZCodec aZCodec;
+        aZCodec.BeginCompression();
+        pStream->GetMemory().Seek(0);
+        aZCodec.Decompress(pStream->GetMemory(), aObjectStream);
+        CPPUNIT_ASSERT(aZCodec.EndCompression());
+        aObjectStream.Seek(0);
+        std::string aCmap("2 beginbfchar\n"
+                          "<01> <1ECB0331030B>\n"
+                          "<05> <0020>\n"
+                          "endbfchar");
+        std::string aData(static_cast<const char*>(aObjectStream.GetData()),
+                          aObjectStream.GetSize());
+        auto nPos = aData.find(aCmap);
+        CPPUNIT_ASSERT(nPos != std::string::npos);
+    }
+
+    {
+        auto aPages = aDocument.GetPages();
+        CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+        // Get page contents and stream.
+        auto pContents = aPages[0]->LookupObject("Contents"_ostr);
+        CPPUNIT_ASSERT(pContents);
+        auto pStream = pContents->GetStream();
+        CPPUNIT_ASSERT(pStream);
+        auto& rObjectStream = pStream->GetMemory();
+
+        // Uncompress the stream.
+        SvMemoryStream aUncompressed;
+        ZCodec aZCodec;
+        aZCodec.BeginCompression();
+        rObjectStream.Seek(0);
+        aZCodec.Decompress(rObjectStream, aUncompressed);
+        CPPUNIT_ASSERT(aZCodec.EndCompression());
+
+        // Make sure the expected ActualText is present.
+        std::string aData(static_cast<const char*>(aUncompressed.GetData()),
+                          aUncompressed.GetSize());
+
+        std::string aActualText("/Span<</ActualText<FEFF1ECB0331030B>>>");
+        size_t nCount = 0;
+        size_t nPos = 0;
+        while ((nPos = aData.find(aActualText, nPos)) != std::string::npos)
+        {
+            nCount++;
+            nPos += aActualText.length();
+        }
+        CPPUNIT_ASSERT_EQUAL_MESSAGE("Number of ActualText entries does not match!",
+                                     static_cast<size_t>(4), nCount);
+    }
+#endif
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf105954)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    uno::Sequence<beans::PropertyValue> aFilterData(comphelper::InitPropertySequence(
+        { { "ReduceImageResolution", uno::Any(true) },
+          { "MaxImageResolution", uno::Any(static_cast<sal_Int32>(300)) } }));
+    aMediaDescriptor["FilterData"] <<= aFilterData;
+    saveAsPDF(u"tdf105954.odt");
+
+    // Parse the export result with pdfium.
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // There is a single image on the page.
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    CPPUNIT_ASSERT_EQUAL(1, nPageObjectCount);
+
+    // Check width of the image.
+    std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(/*index=*/0);
+    Size aMeta = pPageObject->getImageSize(*pPdfPage);
+    // This was 2000, i.e. the 'reduce to 300 DPI' request was ignored.
+    // This is now around 238 (228 on macOS).
+    CPPUNIT_ASSERT_LESS(static_cast<tools::Long>(250), aMeta.getWidth());
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf157679)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("impress_pdf_Export");
+    saveAsPDF(u"tdf157679.pptx");
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // Without the fix in place, this test would have failed with
+    // - Expected: 3
+    // - Actual  : 5
+    CPPUNIT_ASSERT_EQUAL(3, pPdfPage->getObjectCount());
+
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pTextPage = pPdfPage->getTextPage();
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        // Check there are not Text objects
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(i);
+        CPPUNIT_ASSERT(pPageObject->getType() != vcl::pdf::PDFPageObjectType::Text);
+    }
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf128445)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("impress_pdf_Export");
+    saveAsPDF(u"tdf128445.odp");
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // Without the fix in place, this test would have failed with
+    // - Expected: 7
+    // - Actual  : 6
+    CPPUNIT_ASSERT_EQUAL(7, pPdfPage->getObjectCount());
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf128630)
+{
+    // FIXME: the DPI check should be removed when either (1) the test is fixed to work with
+    // non-default DPI; or (2) unit tests on Windows are made to use svp VCL plugin.
+    if (!IsDefaultDPI())
+        return;
+
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("impress_pdf_Export");
+    saveAsPDF(u"tdf128630.odp");
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has one page.
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+
+    // Assert the size of the only bitmap on the page.
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(i);
+        if (pPageObject->getType() != vcl::pdf::PDFPageObjectType::Image)
+            continue;
+
+        std::unique_ptr<vcl::pdf::PDFiumBitmap> pBitmap = pPageObject->getImageBitmap();
+        CPPUNIT_ASSERT(pBitmap);
+        int nWidth = pBitmap->getWidth();
+        // Without the accompanying fix in place, this test would have failed with:
+        // - Expected: 466
+        // - Actual  : 289
+        // i.e. the rotated + scaled arrow was more thin than it should be.
+        CPPUNIT_ASSERT_DOUBLES_EQUAL(466, nWidth, 1);
+        int nHeight = pBitmap->getHeight();
+        CPPUNIT_ASSERT_EQUAL(nWidth, nHeight);
+    }
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf106702)
+{
+    // Import the bugdoc and export as PDF.
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf106702.odt");
+
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has two pages.
+    CPPUNIT_ASSERT_EQUAL(2, pPdfDocument->getPageCount());
+
+    // First page already has the correct image position.
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+    int nExpected = 0;
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(i);
+        if (pPageObject->getType() != vcl::pdf::PDFPageObjectType::Image)
+            continue;
+
+        // Top, but upside down.
+        nExpected = pPageObject->getBounds().getMaxY();
+        break;
+    }
+
+    // Second page had an incorrect image position.
+    pPdfPage = pPdfDocument->openPage(/*nIndex=*/1);
+    CPPUNIT_ASSERT(pPdfPage);
+    int nActual = 0;
+    nPageObjectCount = pPdfPage->getObjectCount();
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(i);
+        if (pPageObject->getType() != vcl::pdf::PDFPageObjectType::Image)
+            continue;
+
+        // Top, but upside down.
+        nActual = pPageObject->getBounds().getMaxY();
+        break;
+    }
+
+    // This failed, vertical pos is 818 points, was 1674 (outside visible page
+    // bounds).
+    CPPUNIT_ASSERT_EQUAL(nExpected, nActual);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf113143)
+{
+    aMediaDescriptor["FilterName"] <<= OUString("impress_pdf_Export");
+    uno::Sequence<beans::PropertyValue> aFilterData(comphelper::InitPropertySequence({
+        { "ExportNotesPages", uno::Any(true) },
+        // ReduceImageResolution is on by default and that hides the bug we
+        // want to test.
+        { "ReduceImageResolution", uno::Any(false) },
+        // Set a custom PDF version.
+        { "SelectPdfVersion", uno::Any(static_cast<sal_Int32>(16)) },
+    }));
+    aMediaDescriptor["FilterData"] <<= aFilterData;
+    saveAsPDF(u"tdf113143.odp");
+
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+
+    // The document has two pages.
+    CPPUNIT_ASSERT_EQUAL(2, pPdfDocument->getPageCount());
+
+    // First has the original (larger) image.
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+    int nLarger = 0;
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(i);
+        if (pPageObject->getType() != vcl::pdf::PDFPageObjectType::Image)
+            continue;
+
+        nLarger = pPageObject->getBounds().getWidth();
+        break;
+    }
+
+    // Second page has the scaled (smaller) image.
+    pPdfPage = pPdfDocument->openPage(/*nIndex=*/1);
+    CPPUNIT_ASSERT(pPdfPage);
+    int nSmaller = 0;
+    nPageObjectCount = pPdfPage->getObjectCount();
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(i);
+        if (pPageObject->getType() != vcl::pdf::PDFPageObjectType::Image)
+            continue;
+
+        nSmaller = pPageObject->getBounds().getWidth();
+        break;
+    }
+
+    // This failed, both were 319, now nSmaller is 169.
+    CPPUNIT_ASSERT_LESS(nLarger, nSmaller);
+
+    // The following check used to fail in the past, header was "%PDF-1.5":
+    maMemory.Seek(0);
+    OString aExpectedHeader("%PDF-1.6"_ostr);
+    OString aHeader(read_uInt8s_ToOString(maMemory, aExpectedHeader.getLength()));
+    CPPUNIT_ASSERT_EQUAL(aExpectedHeader, aHeader);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testForcePoint71)
+{
+    // I just care it doesn't crash
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"forcepoint71.key");
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testForcePoint80)
+{
+    // printing asserted in SwCellFrame::FindStartEndOfRowSpanCell
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"forcepoint80-1.rtf");
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testForcePoint3)
+{
+    // printing asserted in SwFrame::GetNextSctLeaf()
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"flowframe_null_ptr_deref.sample");
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf84283)
+{
+    // Without the fix in place, this test would have crashed
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf84283.doc");
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf115262)
+{
+    aMediaDescriptor["FilterName"] <<= OUString("calc_pdf_Export");
+    saveAsPDF(u"tdf115262.ods");
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+    CPPUNIT_ASSERT_EQUAL(8, pPdfDocument->getPageCount());
+
+    // Get the 6th page.
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/5);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // Look up the position of the first image and the 400th row.
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pTextPage = pPdfPage->getTextPage();
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    int nFirstImageTop = 0;
+    int nRowTop = 0;
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(i);
+        // Top, but upside down.
+        float fTop = pPageObject->getBounds().getMaxY();
+
+        if (pPageObject->getType() == vcl::pdf::PDFPageObjectType::Image)
+        {
+            nFirstImageTop = fTop;
+        }
+        else if (pPageObject->getType() == vcl::pdf::PDFPageObjectType::Text)
+        {
+            OUString sText = pPageObject->getText(pTextPage);
+            if (sText == "400")
+                nRowTop = fTop;
+        }
+    }
+    // Make sure that the top of the "400" is below the top of the image (in
+    // bottom-right-corner-based PDF coordinates).
+    // This was: expected less than 144, actual is 199.
+    CPPUNIT_ASSERT_LESS(nFirstImageTop, nRowTop);
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf121962)
+{
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf121962.odt");
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+
+    // Get the first page
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pTextPage = pPdfPage->getTextPage();
+
+    // Make sure the table sum is displayed as "0", not faulty expression.
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(i);
+        if (pPageObject->getType() != vcl::pdf::PDFPageObjectType::Text)
+            continue;
+        OUString sText = pPageObject->getText(pTextPage);
+        CPPUNIT_ASSERT(sText != "** Expression is faulty **");
+    }
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf139065)
+{
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf139065.odt");
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+
+    // Without the fix in place, this test would have failed with
+    // - Expected: 15
+    // - Actual  : 6
+    CPPUNIT_ASSERT_EQUAL(15, pPdfPage->getObjectCount());
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf157816)
+{
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+
+    // Enable PDF/UA
+    uno::Sequence<beans::PropertyValue> aFilterData(
+        comphelper::InitPropertySequence({ { "PDFUACompliance", uno::Any(true) } }));
+    aMediaDescriptor["FilterData"] <<= aFilterData;
+    saveAsPDF(u"tdf157816.fodt");
+
+    vcl::filter::PDFDocument aDocument;
+    SvFileStream aStream(maTempFile.GetURL(), StreamMode::READ);
+    CPPUNIT_ASSERT(aDocument.Read(aStream));
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    vcl::filter::PDFObjectElement* pDocument(nullptr);
+    for (const auto& rDocElement : aDocument.GetElements())
+    {
+        auto pObject1 = dynamic_cast<vcl::filter::PDFObjectElement*>(rDocElement.get());
+        if (!pObject1)
+            continue;
+        auto pType1 = dynamic_cast<vcl::filter::PDFNameElement*>(pObject1->Lookup("Type"_ostr));
+        if (pType1 && pType1->GetValue() == "StructElem")
+        {
+            auto pS1 = dynamic_cast<vcl::filter::PDFNameElement*>(pObject1->Lookup("S"_ostr));
+            if (pS1 && pS1->GetValue() == "Document")
+            {
+                pDocument = pObject1;
+            }
+        }
+    }
+    CPPUNIT_ASSERT(pDocument);
+
+    auto pKidsD = dynamic_cast<vcl::filter::PDFArrayElement*>(pDocument->Lookup("K"_ostr));
+    CPPUNIT_ASSERT(pKidsD);
+    // assume there are no MCID ref at this level
+    auto pKidsDv = pKidsD->GetElements();
+    auto pRefKidD2 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsDv[2]);
+    CPPUNIT_ASSERT(pRefKidD2);
+    auto pObjectD2 = pRefKidD2->LookupObject();
+    CPPUNIT_ASSERT(pObjectD2);
+    auto pTypeD2 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD2->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD2->GetValue());
+    auto pSD2 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD2->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Text#20body"_ostr, pSD2->GetValue());
+
+    auto pKidsD2 = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD2->Lookup("K"_ostr));
+    CPPUNIT_ASSERT(pKidsD2);
+    auto pKidsD2v = pKidsD2->GetElements();
+    auto pRefKidD20 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD2v[0]);
+    // MCID for text
+    CPPUNIT_ASSERT(!pRefKidD20);
+    auto pRefKidD21 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD2v[1]);
+    // MCID for text
+    CPPUNIT_ASSERT(!pRefKidD21);
+
+    auto pRefKidD22 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD2v[2]);
+    CPPUNIT_ASSERT(pRefKidD22);
+    auto pObjectD22 = pRefKidD22->LookupObject();
+    CPPUNIT_ASSERT(pObjectD22);
+    auto pTypeD22 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD22->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD22->GetValue());
+    auto pSD22 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD22->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Link"_ostr, pSD22->GetValue());
+    {
+        auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD22->Lookup("K"_ostr));
+        auto nMCID(0);
+        auto nRef(0);
+        for (size_t i = 0; i < pKids->GetElements().size(); ++i)
+        {
+            auto pNum = dynamic_cast<vcl::filter::PDFNumberElement*>(pKids->GetElement(i));
+            auto pObjR = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pKids->GetElement(i));
+            if (pNum)
+            {
+                ++nMCID;
+            }
+            if (pObjR)
+            {
+                ++nRef;
+                auto pOType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObjR->LookupElement("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("OBJR"_ostr, pOType->GetValue());
+                auto pAnnotRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObjR->LookupElement("Obj"_ostr));
+                auto pAnnot = pAnnotRef->LookupObject();
+                auto pAType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Annot"_ostr, pAType->GetValue());
+                auto pASubtype
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Subtype"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Link"_ostr, pASubtype->GetValue());
+                auto pAContents = dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                    pAnnot->Lookup("Contents"_ostr));
+                CPPUNIT_ASSERT_EQUAL(
+                    u"Error: Reference source not found"_ustr,
+                    ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(*pAContents));
+                auto pStructParent = dynamic_cast<vcl::filter::PDFNumberElement*>(
+                    pAnnot->Lookup("StructParent"_ostr));
+                CPPUNIT_ASSERT(pStructParent); // every link must have it!
+                auto pARect
+                    = dynamic_cast<vcl::filter::PDFArrayElement*>(pAnnot->Lookup("Rect"_ostr));
+                CPPUNIT_ASSERT(pARect);
+                const auto& rElements = pARect->GetElements();
+                CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), rElements.size());
+                const auto* pNumL = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[0]);
+                CPPUNIT_ASSERT(pNumL);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(95.143, pNumL->GetValue(), 1e-3);
+                const auto* pNumT = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[1]);
+                CPPUNIT_ASSERT(pNumT);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(674.589, pNumT->GetValue(), 1e-3);
+                const auto* pNumR = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[2]);
+                CPPUNIT_ASSERT(pNumR);
+                // this changed to the end of the text, not the start of the fly
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(187.207, pNumR->GetValue(), 1e-3);
+                const auto* pNumB = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[3]);
+                CPPUNIT_ASSERT(pNumB);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(688.389, pNumB->GetValue(), 1e-3);
+            }
+        }
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nMCID)>(1), nMCID);
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nRef)>(1), nRef);
+    }
+
+    auto pRefKidD23 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD2v[3]);
+    CPPUNIT_ASSERT(pRefKidD23);
+    auto pObjectD23 = pRefKidD23->LookupObject();
+    CPPUNIT_ASSERT(pObjectD23);
+    auto pTypeD23 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD23->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD23->GetValue());
+    auto pSD23 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD23->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Link"_ostr, pSD23->GetValue());
+    {
+        auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD23->Lookup("K"_ostr));
+        auto nMCID(0);
+        auto nRef(0);
+        for (size_t i = 0; i < pKids->GetElements().size(); ++i)
+        {
+            auto pNum = dynamic_cast<vcl::filter::PDFNumberElement*>(pKids->GetElement(i));
+            auto pObjR = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pKids->GetElement(i));
+            if (pNum)
+            {
+                ++nMCID;
+            }
+            if (pObjR)
+            {
+                ++nRef;
+                auto pOType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObjR->LookupElement("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("OBJR"_ostr, pOType->GetValue());
+                auto pAnnotRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObjR->LookupElement("Obj"_ostr));
+                auto pAnnot = pAnnotRef->LookupObject();
+                auto pAType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Annot"_ostr, pAType->GetValue());
+                auto pASubtype
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Subtype"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Link"_ostr, pASubtype->GetValue());
+                auto pAContents = dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                    pAnnot->Lookup("Contents"_ostr));
+                CPPUNIT_ASSERT_EQUAL(
+                    u"Error: Reference source not found"_ustr,
+                    ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(*pAContents));
+                auto pStructParent = dynamic_cast<vcl::filter::PDFNumberElement*>(
+                    pAnnot->Lookup("StructParent"_ostr));
+                CPPUNIT_ASSERT(pStructParent); // every link must have it!
+                auto pARect
+                    = dynamic_cast<vcl::filter::PDFArrayElement*>(pAnnot->Lookup("Rect"_ostr));
+                CPPUNIT_ASSERT(pARect);
+                const auto& rElements = pARect->GetElements();
+                CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), rElements.size());
+                const auto* pNumL = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[0]);
+                CPPUNIT_ASSERT(pNumL);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(56.693, pNumL->GetValue(), 1e-3);
+                const auto* pNumT = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[1]);
+                CPPUNIT_ASSERT(pNumT);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(660.789, pNumT->GetValue(), 1e-3);
+                const auto* pNumR = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[2]);
+                CPPUNIT_ASSERT(pNumR);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(146.157, pNumR->GetValue(), 1e-3);
+                const auto* pNumB = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[3]);
+                CPPUNIT_ASSERT(pNumB);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(674.589, pNumB->GetValue(), 1e-3);
+            }
+        }
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nMCID)>(1), nMCID);
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nRef)>(1), nRef);
+    }
+
+    auto pRefKidD24 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD2v[4]);
+    CPPUNIT_ASSERT(pRefKidD24);
+    auto pObjectD24 = pRefKidD24->LookupObject();
+    CPPUNIT_ASSERT(pObjectD24);
+    auto pTypeD24 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD24->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD24->GetValue());
+    auto pSD24 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD24->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Link"_ostr, pSD24->GetValue());
+    {
+        auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD24->Lookup("K"_ostr));
+        auto nMCID(0);
+        auto nRef(0);
+        for (size_t i = 0; i < pKids->GetElements().size(); ++i)
+        {
+            auto pNum = dynamic_cast<vcl::filter::PDFNumberElement*>(pKids->GetElement(i));
+            auto pObjR = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pKids->GetElement(i));
+            if (pNum)
+            {
+                ++nMCID;
+            }
+            if (pObjR)
+            {
+                ++nRef;
+                auto pOType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObjR->LookupElement("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("OBJR"_ostr, pOType->GetValue());
+                auto pAnnotRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObjR->LookupElement("Obj"_ostr));
+                auto pAnnot = pAnnotRef->LookupObject();
+                auto pAType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Annot"_ostr, pAType->GetValue());
+                auto pASubtype
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Subtype"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Link"_ostr, pASubtype->GetValue());
+                auto pAContents = dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                    pAnnot->Lookup("Contents"_ostr));
+                CPPUNIT_ASSERT_EQUAL(
+                    u"Error: Reference source not found"_ustr,
+                    ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(*pAContents));
+                auto pStructParent = dynamic_cast<vcl::filter::PDFNumberElement*>(
+                    pAnnot->Lookup("StructParent"_ostr));
+                CPPUNIT_ASSERT(pStructParent); // every link must have it!
+                auto pARect
+                    = dynamic_cast<vcl::filter::PDFArrayElement*>(pAnnot->Lookup("Rect"_ostr));
+                CPPUNIT_ASSERT(pARect);
+                const auto& rElements = pARect->GetElements();
+                CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), rElements.size());
+                const auto* pNumL = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[0]);
+                CPPUNIT_ASSERT(pNumL);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(146.093, pNumL->GetValue(), 1e-3);
+                const auto* pNumT = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[1]);
+                CPPUNIT_ASSERT(pNumT);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(660.789, pNumT->GetValue(), 1e-3);
+                const auto* pNumR = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[2]);
+                CPPUNIT_ASSERT(pNumR);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(179.457, pNumR->GetValue(), 1e-3);
+                const auto* pNumB = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[3]);
+                CPPUNIT_ASSERT(pNumB);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(674.589, pNumB->GetValue(), 1e-3);
+            }
+        }
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nMCID)>(1), nMCID);
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nRef)>(1), nRef);
+    }
+
+    auto pRefKidD25 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD2v[5]);
+    CPPUNIT_ASSERT(pRefKidD25);
+    auto pObjectD25 = pRefKidD25->LookupObject();
+    CPPUNIT_ASSERT(pObjectD25);
+    auto pTypeD25 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD25->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD25->GetValue());
+    auto pSD25 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD25->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Link"_ostr, pSD25->GetValue());
+    {
+        auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD25->Lookup("K"_ostr));
+        auto nMCID(0);
+        auto nRef(0);
+        for (size_t i = 0; i < pKids->GetElements().size(); ++i)
+        {
+            auto pNum = dynamic_cast<vcl::filter::PDFNumberElement*>(pKids->GetElement(i));
+            auto pObjR = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pKids->GetElement(i));
+            if (pNum)
+            {
+                ++nMCID;
+            }
+            if (pObjR)
+            {
+                ++nRef;
+                auto pOType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObjR->LookupElement("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("OBJR"_ostr, pOType->GetValue());
+                auto pAnnotRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObjR->LookupElement("Obj"_ostr));
+                auto pAnnot = pAnnotRef->LookupObject();
+                auto pAType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Annot"_ostr, pAType->GetValue());
+                auto pASubtype
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Subtype"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Link"_ostr, pASubtype->GetValue());
+                auto pAContents = dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                    pAnnot->Lookup("Contents"_ostr));
+                CPPUNIT_ASSERT_EQUAL(
+                    u"Error: Reference source not found"_ustr,
+                    ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(*pAContents));
+                auto pStructParent = dynamic_cast<vcl::filter::PDFNumberElement*>(
+                    pAnnot->Lookup("StructParent"_ostr));
+                CPPUNIT_ASSERT(pStructParent); // every link must have it!
+                auto pARect
+                    = dynamic_cast<vcl::filter::PDFArrayElement*>(pAnnot->Lookup("Rect"_ostr));
+                CPPUNIT_ASSERT(pARect);
+                const auto& rElements = pARect->GetElements();
+                CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), rElements.size());
+                const auto* pNumL = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[0]);
+                CPPUNIT_ASSERT(pNumL);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(56.693, pNumL->GetValue(), 1e-3);
+                const auto* pNumT = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[1]);
+                CPPUNIT_ASSERT(pNumT);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(646.989, pNumT->GetValue(), 1e-3);
+                const auto* pNumR = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[2]);
+                CPPUNIT_ASSERT(pNumR);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(174.757, pNumR->GetValue(), 1e-3);
+                const auto* pNumB = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[3]);
+                CPPUNIT_ASSERT(pNumB);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(660.789, pNumB->GetValue(), 1e-3);
+            }
+        }
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nMCID)>(1), nMCID);
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nRef)>(1), nRef);
+    }
+
+    auto pRefKidD26 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD2v[6]);
+    CPPUNIT_ASSERT(pRefKidD26);
+    auto pObjectD26 = pRefKidD26->LookupObject();
+    CPPUNIT_ASSERT(pObjectD26);
+    auto pTypeD26 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD26->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD26->GetValue());
+    auto pSD26 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD26->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Link"_ostr, pSD26->GetValue());
+    {
+        auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD26->Lookup("K"_ostr));
+        auto nMCID(0);
+        auto nRef(0);
+        for (size_t i = 0; i < pKids->GetElements().size(); ++i)
+        {
+            auto pNum = dynamic_cast<vcl::filter::PDFNumberElement*>(pKids->GetElement(i));
+            auto pObjR = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pKids->GetElement(i));
+            if (pNum)
+            {
+                ++nMCID;
+            }
+            if (pObjR)
+            {
+                ++nRef;
+                auto pOType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObjR->LookupElement("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("OBJR"_ostr, pOType->GetValue());
+                auto pAnnotRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObjR->LookupElement("Obj"_ostr));
+                auto pAnnot = pAnnotRef->LookupObject();
+                auto pAType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Annot"_ostr, pAType->GetValue());
+                auto pASubtype
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Subtype"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Link"_ostr, pASubtype->GetValue());
+                auto pAContents = dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                    pAnnot->Lookup("Contents"_ostr));
+                CPPUNIT_ASSERT_EQUAL(
+                    u"Error: Reference source not found"_ustr,
+                    ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(*pAContents));
+                auto pStructParent = dynamic_cast<vcl::filter::PDFNumberElement*>(
+                    pAnnot->Lookup("StructParent"_ostr));
+                CPPUNIT_ASSERT(pStructParent); // every link must have it!
+                auto pARect
+                    = dynamic_cast<vcl::filter::PDFArrayElement*>(pAnnot->Lookup("Rect"_ostr));
+                CPPUNIT_ASSERT(pARect);
+                const auto& rElements = pARect->GetElements();
+                CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), rElements.size());
+                const auto* pNumL = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[0]);
+                CPPUNIT_ASSERT(pNumL);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(56.693, pNumL->GetValue(), 1e-3);
+                const auto* pNumT = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[1]);
+                CPPUNIT_ASSERT(pNumT);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(633.189, pNumT->GetValue(), 1e-3);
+                const auto* pNumR = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[2]);
+                CPPUNIT_ASSERT(pNumR);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(86.807, pNumR->GetValue(), 1e-3);
+                const auto* pNumB = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[3]);
+                CPPUNIT_ASSERT(pNumB);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(646.989, pNumB->GetValue(), 1e-3);
+            }
+        }
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nMCID)>(1), nMCID);
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nRef)>(1), nRef);
+    }
+
+    auto pRefKidD27 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD2v[7]);
+    // MCID for text
+    CPPUNIT_ASSERT(!pRefKidD27);
+
+    // the problem was that in addition to the 5 links with SE there were 3 more
+    auto pAnnots = dynamic_cast<vcl::filter::PDFArrayElement*>(aPages[0]->Lookup("Annots"_ostr));
+    CPPUNIT_ASSERT(pAnnots);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(5), pAnnots->GetElements().size());
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf157816Link)
+{
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+
+    // Enable PDF/UA
+    uno::Sequence<beans::PropertyValue> aFilterData(
+        comphelper::InitPropertySequence({ { "PDFUACompliance", uno::Any(true) } }));
+    aMediaDescriptor["FilterData"] <<= aFilterData;
+    saveAsPDF(u"LinkWithFly.fodt");
+
+    vcl::filter::PDFDocument aDocument;
+    SvFileStream aStream(maTempFile.GetURL(), StreamMode::READ);
+    CPPUNIT_ASSERT(aDocument.Read(aStream));
+
+    // The document has one page.
+    std::vector<vcl::filter::PDFObjectElement*> aPages = aDocument.GetPages();
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(1), aPages.size());
+
+    vcl::filter::PDFObjectElement* pDocument(nullptr);
+    for (const auto& rDocElement : aDocument.GetElements())
+    {
+        auto pObject1 = dynamic_cast<vcl::filter::PDFObjectElement*>(rDocElement.get());
+        if (!pObject1)
+            continue;
+        auto pType1 = dynamic_cast<vcl::filter::PDFNameElement*>(pObject1->Lookup("Type"_ostr));
+        if (pType1 && pType1->GetValue() == "StructElem")
+        {
+            auto pS1 = dynamic_cast<vcl::filter::PDFNameElement*>(pObject1->Lookup("S"_ostr));
+            if (pS1 && pS1->GetValue() == "Document")
+            {
+                pDocument = pObject1;
+            }
+        }
+    }
+    CPPUNIT_ASSERT(pDocument);
+
+    auto pKidsD = dynamic_cast<vcl::filter::PDFArrayElement*>(pDocument->Lookup("K"_ostr));
+    CPPUNIT_ASSERT(pKidsD);
+    // assume there are no MCID ref at this level
+    auto pKidsDv = pKidsD->GetElements();
+    auto pRefKidD0 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsDv[0]);
+    CPPUNIT_ASSERT(pRefKidD0);
+    auto pObjectD0 = pRefKidD0->LookupObject();
+    CPPUNIT_ASSERT(pObjectD0);
+    auto pTypeD0 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD0->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD0->GetValue());
+    auto pSD0 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD0->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Standard"_ostr, pSD0->GetValue());
+
+    auto pKidsD0 = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD0->Lookup("K"_ostr));
+    CPPUNIT_ASSERT(pKidsD0);
+    auto pKidsD0v = pKidsD0->GetElements();
+
+    auto pRefKidD00 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD0v[0]);
+    CPPUNIT_ASSERT(pRefKidD00);
+    auto pObjectD00 = pRefKidD00->LookupObject();
+    CPPUNIT_ASSERT(pObjectD00);
+    auto pTypeD00 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD00->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD00->GetValue());
+    auto pSD00 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD00->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Link"_ostr, pSD00->GetValue());
+    {
+        auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD00->Lookup("K"_ostr));
+        auto nMCID(0);
+        auto nRef(0);
+        for (size_t i = 0; i < pKids->GetElements().size(); ++i)
+        {
+            auto pNum = dynamic_cast<vcl::filter::PDFNumberElement*>(pKids->GetElement(i));
+            auto pObjR = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pKids->GetElement(i));
+            if (pNum)
+            {
+                ++nMCID;
+            }
+            if (pObjR)
+            {
+                ++nRef;
+                auto pOType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObjR->LookupElement("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("OBJR"_ostr, pOType->GetValue());
+                auto pAnnotRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObjR->LookupElement("Obj"_ostr));
+                auto pAnnot = pAnnotRef->LookupObject();
+                auto pAType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Annot"_ostr, pAType->GetValue());
+                auto pASubtype
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Subtype"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Link"_ostr, pASubtype->GetValue());
+                auto pAContents = dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                    pAnnot->Lookup("Contents"_ostr));
+                CPPUNIT_ASSERT_EQUAL(
+                    u"https://www.mozilla.org/en-US/firefox/119.0/releasenotes/"_ustr,
+                    ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(*pAContents));
+                auto pStructParent = dynamic_cast<vcl::filter::PDFNumberElement*>(
+                    pAnnot->Lookup("StructParent"_ostr));
+                CPPUNIT_ASSERT(pStructParent); // every link must have it!
+                auto pARect
+                    = dynamic_cast<vcl::filter::PDFArrayElement*>(pAnnot->Lookup("Rect"_ostr));
+                CPPUNIT_ASSERT(pARect);
+                const auto& rElements = pARect->GetElements();
+                CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), rElements.size());
+                const auto* pNumL = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[0]);
+                CPPUNIT_ASSERT(pNumL);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(56.693, pNumL->GetValue(), 1e-3);
+                const auto* pNumT = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[1]);
+                CPPUNIT_ASSERT(pNumT);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(771.389, pNumT->GetValue(), 1e-3);
+                const auto* pNumR = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[2]);
+                CPPUNIT_ASSERT(pNumR);
+                // this changed to the end of the text, not the start of the fly
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(191.707, pNumR->GetValue(), 1e-3);
+                const auto* pNumB = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[3]);
+                CPPUNIT_ASSERT(pNumB);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(785.189, pNumB->GetValue(), 1e-3);
+            }
+        }
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nMCID)>(1), nMCID);
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nRef)>(1), nRef);
+    }
+
+    auto pRefKidD01 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD0v[1]);
+    CPPUNIT_ASSERT(pRefKidD01);
+    auto pObjectD01 = pRefKidD01->LookupObject();
+    CPPUNIT_ASSERT(pObjectD01);
+    auto pTypeD01 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD01->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD01->GetValue());
+    auto pSD01 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD01->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Link"_ostr, pSD01->GetValue());
+    {
+        auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD01->Lookup("K"_ostr));
+        auto nMCID(0);
+        auto nRef(0);
+        for (size_t i = 0; i < pKids->GetElements().size(); ++i)
+        {
+            auto pNum = dynamic_cast<vcl::filter::PDFNumberElement*>(pKids->GetElement(i));
+            auto pObjR = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pKids->GetElement(i));
+            if (pNum)
+            {
+                ++nMCID;
+            }
+            if (pObjR)
+            {
+                ++nRef;
+                auto pOType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObjR->LookupElement("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("OBJR"_ostr, pOType->GetValue());
+                auto pAnnotRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObjR->LookupElement("Obj"_ostr));
+                auto pAnnot = pAnnotRef->LookupObject();
+                auto pAType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Annot"_ostr, pAType->GetValue());
+                auto pASubtype
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Subtype"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Link"_ostr, pASubtype->GetValue());
+                auto pAContents = dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                    pAnnot->Lookup("Contents"_ostr));
+                CPPUNIT_ASSERT_EQUAL(
+                    u"https://www.mozilla.org/en-US/firefox/119.0/releasenotes/"_ustr,
+                    ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(*pAContents));
+                auto pStructParent = dynamic_cast<vcl::filter::PDFNumberElement*>(
+                    pAnnot->Lookup("StructParent"_ostr));
+                CPPUNIT_ASSERT(pStructParent); // every link must have it!
+                auto pARect
+                    = dynamic_cast<vcl::filter::PDFArrayElement*>(pAnnot->Lookup("Rect"_ostr));
+                CPPUNIT_ASSERT(pARect);
+                const auto& rElements = pARect->GetElements();
+                CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), rElements.size());
+                const auto* pNumL = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[0]);
+                CPPUNIT_ASSERT(pNumL);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(387.843, pNumL->GetValue(), 1e-3);
+                const auto* pNumT = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[1]);
+                CPPUNIT_ASSERT(pNumT);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(771.389, pNumT->GetValue(), 1e-3);
+                const auto* pNumR = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[2]);
+                CPPUNIT_ASSERT(pNumR);
+                // this changed to the end of the text, not the start of the fly
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(534.407, pNumR->GetValue(), 1e-3);
+                const auto* pNumB = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[3]);
+                CPPUNIT_ASSERT(pNumB);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(785.189, pNumB->GetValue(), 1e-3);
+            }
+        }
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nMCID)>(1), nMCID);
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nRef)>(1), nRef);
+    }
+
+    auto pRefKidD02 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD0v[2]);
+    CPPUNIT_ASSERT(pRefKidD02);
+    auto pObjectD02 = pRefKidD02->LookupObject();
+    CPPUNIT_ASSERT(pObjectD02);
+    auto pTypeD02 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD02->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD02->GetValue());
+    auto pSD02 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD02->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Figure"_ostr, pSD02->GetValue());
+
+    auto pRefKidD1 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsDv[1]);
+    CPPUNIT_ASSERT(pRefKidD1);
+    auto pObjectD1 = pRefKidD1->LookupObject();
+    CPPUNIT_ASSERT(pObjectD1);
+    auto pTypeD1 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD1->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD1->GetValue());
+    auto pSD1 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD1->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Standard"_ostr, pSD1->GetValue());
+
+    auto pKidsD1 = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD1->Lookup("K"_ostr));
+    CPPUNIT_ASSERT(pKidsD1);
+    auto pKidsD1v = pKidsD1->GetElements();
+
+    auto pRefKidD10 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD1v[0]);
+    CPPUNIT_ASSERT(pRefKidD10);
+    auto pObjectD10 = pRefKidD10->LookupObject();
+    CPPUNIT_ASSERT(pObjectD10);
+    auto pTypeD10 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD10->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD10->GetValue());
+    auto pSD10 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD10->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Link"_ostr, pSD10->GetValue());
+    {
+        auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD10->Lookup("K"_ostr));
+        auto nMCID(0);
+        auto nRef(0);
+        for (size_t i = 0; i < pKids->GetElements().size(); ++i)
+        {
+            auto pNum = dynamic_cast<vcl::filter::PDFNumberElement*>(pKids->GetElement(i));
+            auto pObjR = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pKids->GetElement(i));
+            if (pNum)
+            {
+                ++nMCID;
+            }
+            if (pObjR)
+            {
+                ++nRef;
+                auto pOType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObjR->LookupElement("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("OBJR"_ostr, pOType->GetValue());
+                auto pAnnotRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObjR->LookupElement("Obj"_ostr));
+                auto pAnnot = pAnnotRef->LookupObject();
+                auto pAType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Annot"_ostr, pAType->GetValue());
+                auto pASubtype
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Subtype"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Link"_ostr, pASubtype->GetValue());
+                auto pAContents = dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                    pAnnot->Lookup("Contents"_ostr));
+                CPPUNIT_ASSERT_EQUAL(
+                    u"https://www.mozilla.org/en-US/firefox/118.0/releasenotes/"_ustr,
+                    ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(*pAContents));
+                auto pStructParent = dynamic_cast<vcl::filter::PDFNumberElement*>(
+                    pAnnot->Lookup("StructParent"_ostr));
+                CPPUNIT_ASSERT(pStructParent); // every link must have it!
+                auto pARect
+                    = dynamic_cast<vcl::filter::PDFArrayElement*>(pAnnot->Lookup("Rect"_ostr));
+                CPPUNIT_ASSERT(pARect);
+                const auto& rElements = pARect->GetElements();
+                CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), rElements.size());
+                const auto* pNumL = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[0]);
+                CPPUNIT_ASSERT(pNumL);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(56.693, pNumL->GetValue(), 1e-3);
+                const auto* pNumT = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[1]);
+                CPPUNIT_ASSERT(pNumT);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(757.589, pNumT->GetValue(), 1e-3);
+                const auto* pNumR = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[2]);
+                CPPUNIT_ASSERT(pNumR);
+                // this changed to the end of the text, not the start of the fly
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(191.707, pNumR->GetValue(), 1e-3);
+                const auto* pNumB = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[3]);
+                CPPUNIT_ASSERT(pNumB);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(771.389, pNumB->GetValue(), 1e-3);
+            }
+        }
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nMCID)>(1), nMCID);
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nRef)>(1), nRef);
+    }
+
+    auto pRefKidD11 = dynamic_cast<vcl::filter::PDFReferenceElement*>(pKidsD1v[1]);
+    CPPUNIT_ASSERT(pRefKidD11);
+    auto pObjectD11 = pRefKidD11->LookupObject();
+    CPPUNIT_ASSERT(pObjectD11);
+    auto pTypeD11 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD11->Lookup("Type"_ostr));
+    CPPUNIT_ASSERT_EQUAL("StructElem"_ostr, pTypeD11->GetValue());
+    auto pSD11 = dynamic_cast<vcl::filter::PDFNameElement*>(pObjectD11->Lookup("S"_ostr));
+    CPPUNIT_ASSERT_EQUAL("Link"_ostr, pSD11->GetValue());
+    {
+        auto pKids = dynamic_cast<vcl::filter::PDFArrayElement*>(pObjectD11->Lookup("K"_ostr));
+        auto nMCID(0);
+        auto nRef(0);
+        for (size_t i = 0; i < pKids->GetElements().size(); ++i)
+        {
+            auto pNum = dynamic_cast<vcl::filter::PDFNumberElement*>(pKids->GetElement(i));
+            auto pObjR = dynamic_cast<vcl::filter::PDFDictionaryElement*>(pKids->GetElement(i));
+            if (pNum)
+            {
+                ++nMCID;
+            }
+            if (pObjR)
+            {
+                ++nRef;
+                auto pOType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pObjR->LookupElement("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("OBJR"_ostr, pOType->GetValue());
+                auto pAnnotRef = dynamic_cast<vcl::filter::PDFReferenceElement*>(
+                    pObjR->LookupElement("Obj"_ostr));
+                auto pAnnot = pAnnotRef->LookupObject();
+                auto pAType
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Type"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Annot"_ostr, pAType->GetValue());
+                auto pASubtype
+                    = dynamic_cast<vcl::filter::PDFNameElement*>(pAnnot->Lookup("Subtype"_ostr));
+                CPPUNIT_ASSERT_EQUAL("Link"_ostr, pASubtype->GetValue());
+                auto pAContents = dynamic_cast<vcl::filter::PDFHexStringElement*>(
+                    pAnnot->Lookup("Contents"_ostr));
+                CPPUNIT_ASSERT_EQUAL(
+                    u"https://www.mozilla.org/en-US/firefox/118.0/releasenotes/"_ustr,
+                    ::vcl::filter::PDFDocument::DecodeHexStringUTF16BE(*pAContents));
+                auto pStructParent = dynamic_cast<vcl::filter::PDFNumberElement*>(
+                    pAnnot->Lookup("StructParent"_ostr));
+                CPPUNIT_ASSERT(pStructParent); // every link must have it!
+                auto pARect
+                    = dynamic_cast<vcl::filter::PDFArrayElement*>(pAnnot->Lookup("Rect"_ostr));
+                CPPUNIT_ASSERT(pARect);
+                const auto& rElements = pARect->GetElements();
+                CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), rElements.size());
+                const auto* pNumL = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[0]);
+                CPPUNIT_ASSERT(pNumL);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(387.843, pNumL->GetValue(), 1e-3);
+                const auto* pNumT = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[1]);
+                CPPUNIT_ASSERT(pNumT);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(757.589, pNumT->GetValue(), 1e-3);
+                const auto* pNumR = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[2]);
+                CPPUNIT_ASSERT(pNumR);
+                // this changed to the end of the text, not the start of the fly
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(534.407, pNumR->GetValue(), 1e-3);
+                const auto* pNumB = dynamic_cast<vcl::filter::PDFNumberElement*>(rElements[3]);
+                CPPUNIT_ASSERT(pNumB);
+                CPPUNIT_ASSERT_DOUBLES_EQUAL(771.389, pNumB->GetValue(), 1e-3);
+            }
+        }
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nMCID)>(1), nMCID);
+        CPPUNIT_ASSERT_EQUAL(static_cast<decltype(nRef)>(1), nRef);
+    }
+
+    // the problem was that in addition to the 4 links with SE there was 1 more
+    auto pAnnots = dynamic_cast<vcl::filter::PDFArrayElement*>(aPages[0]->Lookup("Annots"_ostr));
+    CPPUNIT_ASSERT(pAnnots);
+    CPPUNIT_ASSERT_EQUAL(static_cast<size_t>(4), pAnnots->GetElements().size());
+}
+
+CPPUNIT_TEST_FIXTURE(PdfExportTest, testTdf115967)
+{
+    aMediaDescriptor["FilterName"] <<= OUString("writer_pdf_Export");
+    saveAsPDF(u"tdf115967.odt");
+    std::unique_ptr<vcl::pdf::PDFiumDocument> pPdfDocument = parsePDFExport();
+    CPPUNIT_ASSERT_EQUAL(1, pPdfDocument->getPageCount());
+
+    // Get the first page
+    std::unique_ptr<vcl::pdf::PDFiumPage> pPdfPage = pPdfDocument->openPage(/*nIndex=*/0);
+    CPPUNIT_ASSERT(pPdfPage);
+    std::unique_ptr<vcl::pdf::PDFiumTextPage> pTextPage = pPdfPage->getTextPage();
+
+    // Make sure the elements inside a formula in a RTL document are exported
+    // LTR ( m=750abc ) and not RTL ( m=057cba )
+    int nPageObjectCount = pPdfPage->getObjectCount();
+    OUString sText;
+    for (int i = 0; i < nPageObjectCount; ++i)
+    {
+        std::unique_ptr<vcl::pdf::PDFiumPageObject> pPageObject = pPdfPage->getObject(i);
+        if (pPageObject->getType() != vcl::pdf::PDFPageObjectType::Text)
+            continue;
+        OUString sChar = pPageObject->getText(pTextPage);
+        sText += o3tl::trim(sChar);
+    }
+    CPPUNIT_ASSERT_EQUAL(OUString("m=750abc"), sText);
+}
+
+} // end anonymous namespace
+
+CPPUNIT_PLUGIN_IMPLEMENT();
+
+/* vim:set shiftwidth=4 softtabstop=4 expandtab: */