C API: expose functions for indirect objects (fixes #588)

8 files changed, 213 insertions, 1 deletions

diff --git a/ChangeLog b/ChangeLog
index 9efbdc15..f100817f 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,9 @@
 2021-12-10  Jay Berkenbilt  <ejb@ql.org>
 
+	* C API: add qpdf_get_object_by_id, qpdf_make_indirect_object, and
+	qpdf_replace_object, exposing the corresponding methods in QPDF
+	and QPDFObjectHandle. Fixes #588.
+
 	* Add missing QPDF_DLL to QPDFObjectHandle::addTokenFilter so that
 	it is actually accessible as part of the public interface as
 	intended. Fixes #580.
diff --git a/include/qpdf/qpdf-c.h b/include/qpdf/qpdf-c.h
index 90219123..fc393152 100644
--- a/include/qpdf/qpdf-c.h
+++ b/include/qpdf/qpdf-c.h
@@ -618,6 +618,15 @@ extern "C" {
     QPDF_DLL
     qpdf_oh qpdf_get_root(qpdf_data data);
 
+    /* Retrieve and replace indirect objects */
+    QPDF_DLL
+    qpdf_oh qpdf_get_object_by_id(qpdf_data qpdf, int objid, int generation);
+    QPDF_DLL
+    qpdf_oh qpdf_make_indirect_object(qpdf_data qpdf, qpdf_oh oh);
+    QPDF_DLL
+    void qpdf_replace_object(
+        qpdf_data qpdf, int objid, int generation, qpdf_oh oh);
+
     /* Wrappers around QPDFObjectHandle methods. Be sure to read
      * corresponding comments in QPDFObjectHandle.hh to understand
      * what each function does and what kinds of objects it applies
diff --git a/libqpdf/qpdf-c.cc b/libqpdf/qpdf-c.cc
index b7e67e85..1183ffd2 100644
--- a/libqpdf/qpdf-c.cc
+++ b/libqpdf/qpdf-c.cc
@@ -972,6 +972,12 @@ qpdf_oh qpdf_get_root(qpdf_data qpdf)
         });
 }
 
+qpdf_oh qpdf_get_object_by_id(qpdf_data qpdf, int objid, int generation)
+{
+    QTC::TC("qpdf", "qpdf-c called qpdf_get_object_by_id");
+    return new_object(qpdf, qpdf->qpdf->getObjectByID(objid, generation));
+}
+
 template<class RET>
 static RET do_with_oh(
     qpdf_data qpdf, qpdf_oh oh,
@@ -1008,6 +1014,15 @@ static void do_with_oh_void(
         });
 }
 
+void qpdf_replace_object(qpdf_data qpdf, int objid, int generation, qpdf_oh oh)
+{
+    do_with_oh_void(
+        qpdf, oh, [&qpdf, &objid, &generation](QPDFObjectHandle& o) {
+            QTC::TC("qpdf", "qpdf-c called qpdf_replace_object");
+            qpdf->qpdf->replaceObject(objid, generation, o);
+        });
+}
+
 QPDF_BOOL qpdf_oh_is_initialized(qpdf_data qpdf, qpdf_oh oh)
 {
     QTC::TC("qpdf", "qpdf-c called qpdf_oh_is_initialized");
@@ -1421,6 +1436,16 @@ void qpdf_oh_make_direct(qpdf_data qpdf, qpdf_oh oh)
         });
 }
 
+qpdf_oh qpdf_make_indirect_object(qpdf_data qpdf, qpdf_oh oh)
+{
+    return do_with_oh<qpdf_oh>(
+        qpdf, oh,
+        return_uninitialized(qpdf),
+        [&qpdf](QPDFObjectHandle& o) {
+            return new_object(qpdf, qpdf->qpdf->makeIndirectObject(o));
+        });
+}
+
 static QPDFObjectHandle
 qpdf_oh_item_internal(qpdf_data qpdf, qpdf_oh item)
 {
diff --git a/manual/qpdf-manual.xml b/manual/qpdf-manual.xml
index f7d224e5..42b1d25c 100644
--- a/manual/qpdf-manual.xml
+++ b/manual/qpdf-manual.xml
@@ -5263,6 +5263,15 @@ print "\n";
           C API. This allows you to clone an object handle.
          </para>
         </listitem>
+        <listitem>
+         <para>
+	  Add <function>qpdf_get_object_by_id</function>,
+	  <function>qpdf_make_indirect_object</function>, and
+	  <function>qpdf_replace_object</function>, exposing the
+	  corresponding methods in <classname>QPDF</classname> and
+	  <classname>QPDFObjectHandle</classname>.
+         </para>
+        </listitem>
        </itemizedlist>
       </listitem>
      </itemizedlist>
diff --git a/qpdf/qpdf-ctest.c b/qpdf/qpdf-ctest.c
index 6ee19004..d789aa61 100644
--- a/qpdf/qpdf-ctest.c
+++ b/qpdf/qpdf-ctest.c
@@ -881,6 +881,56 @@ static void test31(char const* infile,
     report_errors();
 }
 
+static void test32(char const* infile,
+		   char const* password,
+		   char const* outfile,
+		   char const* outfile2)
+{
+    /* This test case is designed for minimal.pdf. */
+    assert(qpdf_read(qpdf, infile, password) == 0);
+    qpdf_oh page = qpdf_get_object_by_id(qpdf, 3, 0);
+    assert(qpdf_oh_is_dictionary(qpdf, page));
+    assert(qpdf_oh_has_key(qpdf, page, "/MediaBox"));
+    report_errors();
+}
+
+static void test33(char const* infile,
+		   char const* password,
+		   char const* outfile,
+		   char const* outfile2)
+{
+    /* This test case is designed for minimal.pdf. */
+
+    /* Convert a direct object to an indirect object and replace it. */
+    assert(qpdf_read(qpdf, infile, password) == 0);
+    qpdf_oh root = qpdf_get_root(qpdf);
+    qpdf_oh pages = qpdf_oh_get_key(qpdf, root, "/Pages");
+    qpdf_oh kids = qpdf_oh_get_key(qpdf, pages, "/Kids");
+    qpdf_oh page1 = qpdf_oh_get_array_item(qpdf, kids, 0);
+    qpdf_oh mediabox = qpdf_oh_get_key(qpdf, page1, "/MediaBox");
+    assert(! qpdf_oh_is_indirect(qpdf, mediabox));
+    qpdf_oh i_mediabox = qpdf_make_indirect_object(qpdf, mediabox);
+    assert(qpdf_oh_is_indirect(qpdf, i_mediabox));
+    qpdf_oh_replace_key(qpdf, page1, "/MediaBox", i_mediabox);
+
+    /* Replace a different indirect object */
+    qpdf_oh resources = qpdf_oh_get_key(qpdf, page1, "/Resources");
+    qpdf_oh procset = qpdf_oh_get_key(qpdf, resources, "/ProcSet");
+    assert(qpdf_oh_is_indirect(qpdf, procset));
+    qpdf_replace_object(
+        qpdf,
+        qpdf_oh_get_object_id(qpdf, procset),
+        qpdf_oh_get_generation(qpdf, procset),
+        qpdf_oh_parse(qpdf, "[/PDF]"));
+
+    qpdf_init_write(qpdf, outfile);
+    qpdf_set_static_ID(qpdf, QPDF_TRUE);
+    qpdf_set_qdf_mode(qpdf, QPDF_TRUE);
+    qpdf_set_suppress_original_object_IDs(qpdf, QPDF_TRUE);
+    qpdf_write(qpdf);
+    report_errors();
+}
+
 int main(int argc, char* argv[])
 {
     char* p = 0;
@@ -952,6 +1002,8 @@ int main(int argc, char* argv[])
 	  (n == 29) ? test29 :
 	  (n == 30) ? test30 :
 	  (n == 31) ? test31 :
+	  (n == 32) ? test32 :
+	  (n == 33) ? test33 :
 	  0);
 
     if (fn == 0)
diff --git a/qpdf/qpdf.testcov b/qpdf/qpdf.testcov
index 67311b01..9b4e3cea 100644
--- a/qpdf/qpdf.testcov
+++ b/qpdf/qpdf.testcov
@@ -606,3 +606,5 @@ qpdf-c called qpdf_oh_new_uninitialized 0
 qpdf-c warn about oh error 1
 qpdf-c registered oh error handler 0
 qpdf-c cleanup warned about unhandled error 0
+qpdf-c called qpdf_get_object_by_id 0
+qpdf-c called qpdf_replace_object 0
diff --git a/qpdf/qtest/qpdf.test b/qpdf/qtest/qpdf.test
index 384d9dca..c7881e02 100644
--- a/qpdf/qtest/qpdf.test
+++ b/qpdf/qtest/qpdf.test
@@ -4812,7 +4812,7 @@ foreach my $i (@c_check_types)
 show_ntests();
 # ----------
 $td->notify("--- C API Object Handle ---");
-$n_tests += 10;
+$n_tests += 13;
 
 $td->runtest("C check object handles",
              {$td->COMMAND => "qpdf-ctest 24 minimal.pdf '' a.pdf"},
@@ -4831,6 +4831,14 @@ $td->runtest("check output",
 	     {$td->FILE => 'a.pdf'},
 	     {$td->FILE => 'c-object-handle-creation-out.pdf'});
 
+$td->runtest("C indirect objects",
+             {$td->COMMAND => "qpdf-ctest 33 minimal.pdf '' a.pdf"},
+             {$td->STRING => "", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("check output",
+	     {$td->FILE => 'a.pdf'},
+	     {$td->FILE => 'c-indirect-objects-out.pdf'});
+
 $td->runtest("C uninitialized objects",
              {$td->COMMAND => "qpdf-ctest 26 '' '' ''"},
              {$td->STRING => "", $td->EXIT_STATUS => 0},
@@ -4855,6 +4863,10 @@ $td->runtest("C type mismatch warning",
              {$td->COMMAND => "qpdf-ctest 31 minimal.pdf '' ''"},
              {$td->FILE => "c-type-warning.out", $td->EXIT_STATUS => 0},
              $td->NORMALIZE_NEWLINES);
+$td->runtest("C get object by ID",
+             {$td->COMMAND => "qpdf-ctest 32 minimal.pdf '' ''"},
+             {$td->STRING => "", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
 
 show_ntests();
 # ----------
diff --git a/qpdf/qtest/qpdf/c-indirect-objects-out.pdf b/qpdf/qtest/qpdf/c-indirect-objects-out.pdf
new file mode 100644
index 00000000..fe6dc178
--- /dev/null
+++ b/qpdf/qtest/qpdf/c-indirect-objects-out.pdf
@@ -0,0 +1,99 @@
+%PDF-1.3
+%¿÷¢þ
+%QDF-1.0
+
+1 0 obj
+<<
+  /Pages 2 0 R
+  /Type /Catalog
+>>
+endobj
+
+2 0 obj
+<<
+  /Count 1
+  /Kids [
+    3 0 R
+  ]
+  /Type /Pages
+>>
+endobj
+
+%% Page 1
+3 0 obj
+<<
+  /Contents 4 0 R
+  /MediaBox 6 0 R
+  /Parent 2 0 R
+  /Resources <<
+    /Font <<
+      /F1 7 0 R
+    >>
+    /ProcSet 8 0 R
+  >>
+  /Type /Page
+>>
+endobj
+
+%% Contents for page 1
+4 0 obj
+<<
+  /Length 5 0 R
+>>
+stream
+BT
+  /F1 24 Tf
+  72 720 Td
+  (Potato) Tj
+ET
+endstream
+endobj
+
+5 0 obj
+44
+endobj
+
+6 0 obj
+[
+  0
+  0
+  612
+  792
+]
+endobj
+
+7 0 obj
+<<
+  /BaseFont /Helvetica
+  /Encoding /WinAnsiEncoding
+  /Name /F1
+  /Subtype /Type1
+  /Type /Font
+>>
+endobj
+
+8 0 obj
+[
+  /PDF
+]
+endobj
+
+xref
+0 9
+0000000000 65535 f 
+0000000025 00000 n 
+0000000079 00000 n 
+0000000161 00000 n 
+0000000348 00000 n 
+0000000447 00000 n 
+0000000466 00000 n 
+0000000506 00000 n 
+0000000624 00000 n 
+trailer <<
+  /Root 1 0 R
+  /Size 9
+  /ID [<31415926535897932384626433832795><31415926535897932384626433832795>]
+>>
+startxref
+651
+%%EOF