Integrate `QPDFEmbeddedFileDocumentHelper` with `QPDF` to streamline embedded fi…

…le handling. Add shared helper retrieval, validation methods, and update usages across the codebase.

Integrate `QPDFEmbeddedFileDocumentHelper` with `QPDF` to streamline embedded fi…
…le handling. Add shared helper retrieval, validation methods, and update usages across the codebase.
m-holger
1 parent 8d762145
Showing 7 changed files with 54 additions and 13 deletions
examples/pdf-attach-file.cc
include/qpdf/QPDF.hh
include/qpdf/QPDFEmbeddedFileDocumentHelper.hh
libqpdf/QPDFEmbeddedFileDocumentHelper.cc
libqpdf/QPDFJob.cc
libqpdf/qpdf/QPDF_private.hh
qpdf/test_driver.cc
@@ -79,7 +79,7 @@ process(
     }
     // Add the embedded file at the document level as an attachment.
-    auto efdh = QPDFEmbeddedFileDocumentHelper(q);
+    auto& efdh = QPDFEmbeddedFileDocumentHelper::get(q);
     efdh.replaceEmbeddedFile(key, fs);
     // Create a file attachment annotation.
@@ -63,6 +63,7 @@ class BufferInputSource;
 class QPDFLogger;
 class QPDFParser;
 class QPDFAcroFormDocumentHelper;
+class QPDFEmbeddedFileDocumentHelper;
 class QPDFPageLabelDocumentHelper;
 class QPDF
@@ -800,6 +801,7 @@ class QPDF
     inline bool reconstructed_xref() const;
     inline QPDFAcroFormDocumentHelper& acroform();
+    inline QPDFEmbeddedFileDocumentHelper& embedded_files();
     inline QPDFPageLabelDocumentHelper& page_labels();
     // For testing only -- do not add to DLL
@@ -36,6 +36,21 @@
 class QPDFEmbeddedFileDocumentHelper: 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 EmbeddedFiles structure, which can be expensive.
+    QPDF_DLL
+    static QPDFEmbeddedFileDocumentHelper& get(QPDF& qpdf);
+
+    // Re-validate the EmbeddedFiles structure. This is useful if you have modified the structure of
+    // the EmbeddedFiles dictionary in a way that would invalidate the cache.
+    //
+    // 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
     QPDFEmbeddedFileDocumentHelper(QPDF&);
 #include <qpdf/QPDFEmbeddedFileDocumentHelper.hh>
+#include <qpdf/QPDFNameTreeObjectHelper.hh>
+#include <qpdf/QPDF_private.hh>
+
 // File attachments are stored in the /EmbeddedFiles (name tree) key of the /Names dictionary from
 // the document catalog. Each entry points to a /FileSpec, which in turn points to one more Embedded
 // File Streams. Note that file specs can appear in other places as well, such as file attachment
@@ -44,6 +47,19 @@ QPDFEmbeddedFileDocumentHelper::QPDFEmbeddedFileDocumentHelper(QPDF&amp; qpdf) :
     QPDFDocumentHelper(qpdf),
     m(std::make_shared<Members>())
 {
+    validate();
+}
+
+QPDFEmbeddedFileDocumentHelper&
+QPDFEmbeddedFileDocumentHelper::get(QPDF& qpdf)
+{
+    return qpdf.embedded_files();
+}
+
+void
+QPDFEmbeddedFileDocumentHelper::validate(bool repair)
+{
+    m->embedded_files.reset();
     auto names = qpdf.getRoot().getKey("/Names");
     if (names.isDictionary()) {
         auto embedded_files = names.getKey("/EmbeddedFiles");
@@ -53,7 +69,7 @@ QPDFEmbeddedFileDocumentHelper::QPDFEmbeddedFileDocumentHelper(QPDF&amp; qpdf) :
                 qpdf,
                 [](QPDFObjectHandle const& o) -> bool { return o.isDictionary(); },
                 true);
-            m->embedded_files->validate();
+            m->embedded_files->validate(repair);
         }
     }
 }
@@ -13,15 +13,12 @@
 #include <qpdf/Pl_StdioFile.hh>
 #include <qpdf/Pl_String.hh>
 #include <qpdf/QIntC.hh>
-#include <qpdf/QPDFAcroFormDocumentHelper.hh>
 #include <qpdf/QPDFCryptoProvider.hh>
-#include <qpdf/QPDFEmbeddedFileDocumentHelper.hh>
 #include <qpdf/QPDFExc.hh>
 #include <qpdf/QPDFLogger.hh>
 #include <qpdf/QPDFObjectHandle_private.hh>
 #include <qpdf/QPDFOutlineDocumentHelper.hh>
 #include <qpdf/QPDFPageDocumentHelper.hh>
-#include <qpdf/QPDFPageLabelDocumentHelper.hh>
 #include <qpdf/QPDFPageObjectHelper.hh>
 #include <qpdf/QPDFSystemError.hh>
 #include <qpdf/QPDFUsage.hh>
@@ -894,7 +891,7 @@ QPDFJob::doShowPages(QPDF&amp; pdf)
 void
 QPDFJob::doListAttachments(QPDF& pdf)
 {
-    QPDFEmbeddedFileDocumentHelper efdh(pdf);
+    auto& efdh = pdf.embedded_files();
     if (efdh.hasEmbeddedFiles()) {
         for (auto const& i: efdh.getEmbeddedFiles()) {
             std::string const& key = i.first;
@@ -934,7 +931,7 @@ QPDFJob::doListAttachments(QPDF&amp; pdf)
 void
 QPDFJob::doShowAttachment(QPDF& pdf)
 {
-    QPDFEmbeddedFileDocumentHelper efdh(pdf);
+    auto& efdh = pdf.embedded_files();
     auto fs = efdh.getEmbeddedFile(m->attachment_to_show);
     if (!fs) {
         throw std::runtime_error("attachment " + m->attachment_to_show + " not found");
@@ -1321,7 +1318,7 @@ QPDFJob::doJSONAttachments(Pipeline* p, bool&amp; first, QPDF&amp; pdf)
     };
     JSON j_attachments = JSON::makeDictionary();
-    QPDFEmbeddedFileDocumentHelper efdh(pdf);
+    auto& efdh = pdf.embedded_files();
     for (auto const& iter: efdh.getEmbeddedFiles()) {
         std::string const& key = iter.first;
         auto fsoh = iter.second;
@@ -2073,7 +2070,7 @@ void
 QPDFJob::addAttachments(QPDF& pdf)
 {
     maybe_set_pagemode(pdf, "/UseAttachments");
-    QPDFEmbeddedFileDocumentHelper efdh(pdf);
+    auto& efdh = pdf.embedded_files();
     std::vector<std::string> duplicated_keys;
     for (auto const& to_add: m->attachments_to_add) {
         if ((!to_add.replace) && efdh.getEmbeddedFile(to_add.key)) {
@@ -2117,7 +2114,7 @@ void
 QPDFJob::copyAttachments(QPDF& pdf)
 {
     maybe_set_pagemode(pdf, "/UseAttachments");
-    QPDFEmbeddedFileDocumentHelper efdh(pdf);
+    auto& efdh = pdf.embedded_files();
     std::vector<std::string> duplicates;
     for (auto const& to_copy: m->attachments_to_copy) {
         doIfVerbose([&](Pipeline& v, std::string const& prefix) {
@@ -2125,7 +2122,7 @@ QPDFJob::copyAttachments(QPDF&amp; pdf)
         });
         std::unique_ptr<QPDF> other;
         processFile(other, to_copy.path.c_str(), to_copy.password.c_str(), false, false);
-        QPDFEmbeddedFileDocumentHelper other_efdh(*other);
+        auto& other_efdh = other->embedded_files();
         auto other_attachments = other_efdh.getEmbeddedFiles();
         for (auto const& iter: other_attachments) {
             std::string new_key = to_copy.prefix + iter.first;
@@ -2259,7 +2256,7 @@ QPDFJob::handleTransformations(QPDF&amp; pdf)
         pdf.getRoot().replaceKey("/PageLabels", page_labels);
     }
     if (!m->attachments_to_remove.empty()) {
-        QPDFEmbeddedFileDocumentHelper efdh(pdf);
+        auto& efdh = pdf.embedded_files();
         for (auto const& key: m->attachments_to_remove) {
             if (efdh.removeEmbeddedFile(key)) {
                 doIfVerbose([&](Pipeline& v, std::string const& prefix) {
@@ -4,6 +4,7 @@
 #include <qpdf/QPDF.hh>
 #include <qpdf/QPDFAcroFormDocumentHelper.hh>
+#include <qpdf/QPDFEmbeddedFileDocumentHelper.hh>
 #include <qpdf/QPDFObject_private.hh>
 #include <qpdf/QPDFPageLabelDocumentHelper.hh>
 #include <qpdf/QPDFTokenizer_private.hh>
@@ -554,6 +555,7 @@ class QPDF::Members
     // Document Helpers;
     std::unique_ptr<QPDFAcroFormDocumentHelper> acroform;
+    std::unique_ptr<QPDFEmbeddedFileDocumentHelper> embedded_files;
     std::unique_ptr<QPDFPageLabelDocumentHelper> page_labels;
 };
@@ -586,6 +588,15 @@ QPDF::acroform()
     return *m->acroform;
 }
+inline QPDFEmbeddedFileDocumentHelper&
+QPDF::embedded_files()
+{
+    if (!m->embedded_files) {
+        m->embedded_files = std::make_unique<QPDFEmbeddedFileDocumentHelper>(*this);
+    }
+    return *m->embedded_files;
+}
+
 inline QPDFPageLabelDocumentHelper&
 QPDF::page_labels()
 {
@@ -2624,7 +2624,7 @@ test_76(QPDF&amp; pdf, char const* arg2)
 {
     // Embedded files. arg2 is a file to attach. Hard-code the
     // mime type and file name for test purposes.
-    QPDFEmbeddedFileDocumentHelper efdh(pdf);
+    auto &efdh = QPDFEmbeddedFileDocumentHelper::get(pdf);
     auto fs1 = QPDFFileSpecObjectHelper::createFileSpec(pdf, "att1.txt", arg2);
     fs1.setDescription("some text");
     auto efs1 = QPDFEFStreamObjectHelper(fs1.getEmbeddedFileStream());