Add `QPDFPageLabelDocumentHelper` integration

Integrate `QPDFPageLabelDocumentHelper` with `QPDF` for improved page label handling. Add methods for retrieving shared helper instances, validating page label structures, and streamline usage throughout the codebase.

Add `QPDFPageLabelDocumentHelper` integration
Integrate `QPDFPageLabelDocumentHelper` with `QPDF` for improved page label handling. Add methods for retrieving shared helper instances, validating page label structures, and streamline usage throughout the codebase.
m-holger
1 parent 551a9534
Showing 6 changed files with 55 additions and 15 deletions
include/qpdf/QPDF.hh
include/qpdf/QPDFPageLabelDocumentHelper.hh
libqpdf/QPDFJob.cc
libqpdf/QPDFPageLabelDocumentHelper.cc
libqpdf/qpdf/QPDF_private.hh
qpdf/test_driver.cc
@@ -63,6 +63,7 @@ class BufferInputSource;
 class QPDFLogger;
 class QPDFParser;
 class QPDFAcroFormDocumentHelper;
+class QPDFPageLabelDocumentHelper;
 class QPDF
 {
@@ -799,6 +800,7 @@ class QPDF
     inline bool reconstructed_xref() const;
     inline QPDFAcroFormDocumentHelper& acroform();
+    inline QPDFPageLabelDocumentHelper& page_labels();
     // For testing only -- do not add to DLL
     static bool test_json_validators();
@@ -22,11 +22,10 @@
 #include <qpdf/QPDFDocumentHelper.hh>
+#include <qpdf/DLL.h>
 #include <qpdf/QPDF.hh>
-#include <qpdf/QPDFNumberTreeObjectHelper.hh>
-#include <vector>
-#include <qpdf/DLL.h>
+#include <vector>
 // Page labels are discussed in the PDF spec (ISO-32000) in section 12.4.2.
 //
@@ -42,6 +41,21 @@
 class QPDFPageLabelDocumentHelper: public QPDFDocumentHelper
 {
   public:
+    // Get a shared document helper for a given QPDF object.
+    //
+    // Retrieving a document helper for a QPDF object rather than creating a new one avoids repeated
+    // validation of the PageLabels structure, which can be expensive.
+    QPDF_DLL
+    static QPDFPageLabelDocumentHelper& get(QPDF& qpdf);
+
+    // Re-validate the PageLabels structure. This is useful if you have modified the structure of
+    // the PageLabels dictionary in a way that could have invalidated the structure.
+    //
+    // If repair is true, the document will be repaired if possible if the validation encounters
+    // errors.
+    QPDF_DLL
+    void validate(bool repair = true);
+
     QPDF_DLL
     QPDFPageLabelDocumentHelper(QPDF&);
@@ -1053,7 +1053,7 @@ QPDFJob::doJSONPages(Pipeline* p, bool&amp; first, QPDF&amp; pdf)
     JSON::writeDictionaryKey(p, first, "pages", 1);
     bool first_page = true;
     JSON::writeArrayOpen(p, first_page, 2);
-    QPDFPageLabelDocumentHelper pldh(pdf);
+    auto& pldh = pdf.page_labels();
     QPDFOutlineDocumentHelper odh(pdf);
     int pageno = -1;
     for (auto& ph: QPDFPageDocumentHelper(pdf).getAllPages()) {
@@ -1116,7 +1116,7 @@ void
 QPDFJob::doJSONPageLabels(Pipeline* p, bool& first, QPDF& pdf)
 {
     JSON j_labels = JSON::makeArray();
-    QPDFPageLabelDocumentHelper pldh(pdf);
+    auto& pldh = pdf.page_labels();
     long long npages = QIntC::to_longlong(QPDFPageDocumentHelper(pdf).getAllPages().size());
     if (pldh.hasPageLabels()) {
         std::vector<QPDFObjectHandle> labels;
@@ -2548,7 +2548,7 @@ QPDFJob::handlePageSpecs(QPDF&amp; pdf, std::vector&lt;std::unique_ptr&lt;QPDF&gt;&gt;&amp; page_hea
             cis = page_spec_cfis[page_data.filename];
             cis->stayOpen(true);
         }
-        QPDFPageLabelDocumentHelper pldh(*page_data.qpdf);
+        auto& pldh = page_data.qpdf->page_labels();
         auto& other_afdh = page_data.qpdf->acroform();
         if (pldh.hasPageLabels()) {
             any_page_labels = true;
@@ -2992,7 +2992,7 @@ QPDFJob::doSplitPages(QPDF&amp; pdf)
         QPDFPageDocumentHelper dh(pdf);
         dh.removeUnreferencedResources();
     }
-    QPDFPageLabelDocumentHelper pldh(pdf);
+    auto& pldh = pdf.page_labels();
     auto& afdh = pdf.acroform();
     std::vector<QPDFObjectHandle> const& pages = pdf.getAllPages();
     size_t pageno_len = std::to_string(pages.size()).length();
 #include <qpdf/QPDFPageLabelDocumentHelper.hh>
+#include <qpdf/QPDFNumberTreeObjectHelper.hh>
 #include <qpdf/QPDFObjectHandle_private.hh>
+#include <qpdf/QPDF_private.hh>
+
+using namespace qpdf;
 class QPDFPageLabelDocumentHelper::Members
 {
@@ -16,14 +20,23 @@ QPDFPageLabelDocumentHelper::QPDFPageLabelDocumentHelper(QPDF&amp; qpdf) :
     QPDFDocumentHelper(qpdf),
     m(std::make_shared<Members>())
 {
-    QPDFObjectHandle root = qpdf.getRoot();
-    if (root.hasKey("/PageLabels")) {
+    validate();
+}
+
+QPDFPageLabelDocumentHelper&
+QPDFPageLabelDocumentHelper::get(QPDF& qpdf)
+{
+    return qpdf.page_labels();
+}
+
+void
+QPDFPageLabelDocumentHelper::validate(bool repair)
+{
+    m->labels = nullptr;
+    if (Dictionary labels = qpdf.getRoot()["/PageLabels"]) {
         m->labels = std::make_unique<QPDFNumberTreeObjectHelper>(
-            root.getKey("/PageLabels"),
-            this->qpdf,
-            [](QPDFObjectHandle const& o) -> bool { return o.isDictionary(); },
-            true);
-        m->labels->validate();
+            labels, qpdf, [](QPDFObjectHandle const& o) -> bool { return o.isDictionary(); }, true);
+        m->labels->validate(repair);
     }
 }
@@ -5,6 +5,7 @@
 #include <qpdf/QPDFAcroFormDocumentHelper.hh>
 #include <qpdf/QPDFObject_private.hh>
+#include <qpdf/QPDFPageLabelDocumentHelper.hh>
 #include <qpdf/QPDFTokenizer_private.hh>
 using namespace qpdf;
@@ -553,6 +554,7 @@ class QPDF::Members
     // Document Helpers;
     std::unique_ptr<QPDFAcroFormDocumentHelper> acroform;
+    std::unique_ptr<QPDFPageLabelDocumentHelper> page_labels;
 };
 // JobSetter class is restricted to QPDFJob.
@@ -584,4 +586,13 @@ QPDF::acroform()
     return *m->acroform;
 }
+inline QPDFPageLabelDocumentHelper&
+QPDF::page_labels()
+{
+    if (!m->page_labels) {
+        m->page_labels = std::make_unique<QPDFPageLabelDocumentHelper>(*this);
+    }
+    return *m->page_labels;
+}
+
 #endif // QPDF_PRIVATE_HH
@@ -1802,7 +1802,7 @@ static void
 test_47(QPDF& pdf, char const* arg2)
 {
     // Test page labels.
-    QPDFPageLabelDocumentHelper pldh(pdf);
+    auto& pldh = QPDFPageLabelDocumentHelper::get(pdf);
     long long npages = pdf.getRoot().getKey("/Pages").getKey("/Count").getIntValue();
     std::vector<QPDFObjectHandle> labels;
     pldh.getLabelsForPageRange(0, npages - 1, 1, labels);