Add JSON v2 support to C API

Jay Berkenbilt
1 parent 66f1fd2a
Showing 14 changed files with 535 additions and 11 deletions
ChangeLog
TODO
include/qpdf/qpdf-c.h
libqpdf/qpdf-c.cc
manual/release-notes.rst
qpdf/qpdf-ctest.c
qpdf/qtest/qpdf-json.test
qpdf/qtest/qpdf/minimal-update.json
qpdf/qtest/qpdf/minimal.json
qpdf/qtest/qpdf/qpdf-ctest-42-43.pdf
qpdf/qtest/qpdf/qpdf-ctest-44-45.pdf
qpdf/qtest/qpdf/qpdf-ctest-46.json
qpdf/qtest/qpdf/qpdf-ctest-47-4
qpdf/qtest/qpdf/qpdf-ctest-47.json
 2022-09-08  Jay Berkenbilt  <ejb@ql.org>
+        * Added new functions to the C API to support qpdf JSON:
+        qpdf_create_from_json_file, qpdf_create_from_json_data,
+        qpdf_update_from_json_file, qpdf_update_from_json_data, and
+        qpdf_write_json. Examples can be found in qpdf-ctest.c (in the
+        source tree), tests 42 through 47.
+
         * Add QPDFObjectHandle::isDestroyed() to test whether an indirect
         object was from a QPDF that has been destroyed.
@@ -8,12 +8,6 @@ Always
 Next
 ====
-Before Release:
-
-* Support json v2 in the C API. At a minimum, write_json,
-  create_from_json, and update_from_json need to be there and should
-  take the same kinds of functions as the C API for logger.
-
 Pending changes:
 * Consider also exposing a way to set a new logger and to get the
@@ -257,8 +257,6 @@ extern &quot;C&quot; {
     QPDF_DLL
     QPDF_ERROR_CODE qpdf_check_pdf(qpdf_data qpdf);
-    /* READ FUNCTIONS */
-
     /* READ PARAMETER FUNCTIONS -- must be called before qpdf_read */
     QPDF_DLL
@@ -267,6 +265,10 @@ extern &quot;C&quot; {
     QPDF_DLL
     void qpdf_set_attempt_recovery(qpdf_data qpdf, QPDF_BOOL value);
+    /* PROCESS FUNCTIONS */
+
+    /* This functions process a PDF or JSON input source. */
+
     /* Calling qpdf_read causes processFile to be called in the C++
      * API.  Basic parsing is performed, but data from the file is
      * only read as needed.  For files without passwords, pass a null
@@ -297,8 +299,40 @@ extern &quot;C&quot; {
     QPDF_DLL
     QPDF_ERROR_CODE qpdf_empty_pdf(qpdf_data qpdf);
-    /* Read functions below must be called after qpdf_read or
-     * qpdf_read_memory. */
+    /* Create a PDF from a JSON file. This calls createFromJSON in the
+     * C++ API.
+     */
+    QPDF_DLL
+    QPDF_ERROR_CODE
+    qpdf_create_from_json_file(qpdf_data qpdf, char const* filename);
+
+    /* Create a PDF from JSON data in a null-terminated string. This
+     * calls createFromJSON in the C++ API.
+     */
+    QPDF_DLL
+    QPDF_ERROR_CODE
+    qpdf_create_from_json_data(
+        qpdf_data qpdf, char const* buffer, unsigned long long size);
+
+    /* JSON UPDATE FUNCTIONS */
+
+    /* Update a QPDF object from a JSON file or buffer. These
+     * functions call updateFromJSON. One of the other processing
+     * functions has to be called first so that the QPDF object is
+     * initialized with PDF data.
+     */
+    QPDF_DLL
+    QPDF_ERROR_CODE
+    qpdf_update_from_json_file(qpdf_data qpdf, char const* filename);
+    QPDF_DLL
+    QPDF_ERROR_CODE
+    qpdf_update_from_json_data(
+        qpdf_data qpdf, char const* buffer, unsigned long long size);
+
+    /* READ FUNCTIONS */
+
+    /* Read functions below must be called after qpdf_read or any of
+     * the other functions that process a PDF. */
     /*
      * NOTE: Functions that return char* are returning a pointer to an
@@ -371,6 +405,39 @@ extern &quot;C&quot; {
     QPDF_DLL
     QPDF_BOOL qpdf_allow_modify_all(qpdf_data qpdf);
+    /* JSON WRITE FUNCTIONS */
+
+    /* This function serializes the PDF to JSON. This calls writeJSON
+     * from the C++ API.
+     *
+     * - version: the JSON version, currently must be 2
+     * - fn: a function that will be called with blocks of JSON data;
+     *   will be called with data, a length, and the value of the
+     *   udata parameter to this function
+     * - udata: will be passed as the third argument to fn with each
+     *   call; use this for your own tracking or pass a null pointer
+     *   if you don't need it
+     * - For decode_level, json_stream_data, file_prefix, and
+     *   wanted_objects, see comments in QPDF.hh. For this API,
+     *   wanted_objects should be a null-terminated array of
+     *   null-terminated strings. Pass a null pointer if you want all
+     *   objects.
+     */
+
+    /* Function should return 0 on success. */
+    typedef int (*qpdf_write_fn_t)(char const* data, size_t len, void* udata);
+
+    QPDF_DLL
+    QPDF_ERROR_CODE qpdf_write_json(
+        qpdf_data qpdf,
+        int version,
+        qpdf_write_fn_t fn,
+        void* udata,
+        enum qpdf_stream_decode_level_e decode_level,
+        enum qpdf_json_stream_data_e json_stream_data,
+        char const* file_prefix,
+        char const* const* wanted_objects);
+
     /* WRITE FUNCTIONS */
     /* Set up for writing.  No writing is actually performed until the
@@ -2,8 +2,10 @@
 #include <qpdf/QPDF.hh>
+#include <qpdf/BufferInputSource.hh>
 #include <qpdf/Pl_Buffer.hh>
 #include <qpdf/Pl_Discard.hh>
+#include <qpdf/Pl_Function.hh>
 #include <qpdf/QIntC.hh>
 #include <qpdf/QPDFExc.hh>
 #include <qpdf/QPDFLogger.hh>
@@ -2029,3 +2031,89 @@ qpdf_remove_page(qpdf_data qpdf, qpdf_oh page)
     auto p = qpdf_oh_item_internal(qpdf, page);
     return trap_errors(qpdf, [&p](qpdf_data q) { q->qpdf->removePage(p); });
 }
+
+QPDF_ERROR_CODE
+qpdf_create_from_json_file(qpdf_data qpdf, char const* filename)
+{
+    QPDF_ERROR_CODE status = QPDF_SUCCESS;
+    qpdf->filename = filename;
+    status = trap_errors(
+        qpdf, [](qpdf_data q) { q->qpdf->createFromJSON(q->filename); });
+    return status;
+}
+
+QPDF_ERROR_CODE
+qpdf_create_from_json_data(
+    qpdf_data qpdf, char const* buffer, unsigned long long size)
+{
+    QPDF_ERROR_CODE status = QPDF_SUCCESS;
+    qpdf->filename = "json buffer";
+    qpdf->buffer = buffer;
+    qpdf->size = size;
+    auto b =
+        new Buffer(QUtil::unsigned_char_pointer(buffer), QIntC::to_size(size));
+    auto is = std::make_shared<BufferInputSource>(qpdf->filename, b, true);
+    status =
+        trap_errors(qpdf, [&is](qpdf_data q) { q->qpdf->createFromJSON(is); });
+    return status;
+}
+
+QPDF_ERROR_CODE
+qpdf_update_from_json_file(qpdf_data qpdf, char const* filename)
+{
+    QPDF_ERROR_CODE status = QPDF_SUCCESS;
+    status = trap_errors(
+        qpdf, [filename](qpdf_data q) { q->qpdf->updateFromJSON(filename); });
+    return status;
+}
+
+QPDF_ERROR_CODE
+qpdf_update_from_json_data(
+    qpdf_data qpdf, char const* buffer, unsigned long long size)
+{
+    QPDF_ERROR_CODE status = QPDF_SUCCESS;
+    auto b =
+        new Buffer(QUtil::unsigned_char_pointer(buffer), QIntC::to_size(size));
+    auto is = std::make_shared<BufferInputSource>(qpdf->filename, b, true);
+    status =
+        trap_errors(qpdf, [&is](qpdf_data q) { q->qpdf->updateFromJSON(is); });
+    return status;
+}
+
+QPDF_ERROR_CODE
+qpdf_write_json(
+    qpdf_data qpdf,
+    int version,
+    qpdf_write_fn_t fn,
+    void* udata,
+    enum qpdf_stream_decode_level_e decode_level,
+    enum qpdf_json_stream_data_e json_stream_data,
+    char const* file_prefix,
+    char const* const* wanted_objects)
+{
+    QPDF_ERROR_CODE status = QPDF_SUCCESS;
+    auto p = std::make_shared<Pl_Function>("write_json", nullptr, fn, udata);
+    std::set<std::string> wanted_objects_set;
+    if (wanted_objects) {
+        for (auto i = wanted_objects; *i; ++i) {
+            wanted_objects_set.insert(*i);
+        }
+    }
+    status = trap_errors(
+        qpdf,
+        [version,
+         p,
+         decode_level,
+         json_stream_data,
+         file_prefix,
+         &wanted_objects_set](qpdf_data q) {
+            q->qpdf->writeJSON(
+                version,
+                p.get(),
+                decode_level,
+                json_stream_data,
+                file_prefix,
+                wanted_objects_set);
+        });
+    return status;
+}
@@ -40,6 +40,10 @@ For a detailed list of changes, please see the file
     - New C++ API calls: ``QPDF::writeJSON``,
       ``QPDF::createFromJSON``, ``QPDF::updateFromJSON``
+    - New C API calls: ``qpdf_create_from_json_file``,
+      ``qpdf_create_from_json_data``, ``qpdf_update_from_json_file``,
+      ``qpdf_update_from_json_data``, and ``qpdf_write_json``.
+
     - Complete documentation can be found at :ref:`json`. A
       comprehensive list of changes from version 1 to version 2 can be
       found at :ref:`json-v2-changes`.
@@ -133,6 +133,13 @@ count_progress(int percent, void* data)
     ++(*(int*)data);
 }
+static int
+write_to_file(char const* data, size_t size, void* udata)
+{
+    FILE* f = (FILE*)udata;
+    return fwrite(data, 1, size, f) != size;
+}
+
 static void
 test01(
     char const* infile,
@@ -1458,7 +1465,7 @@ test41(
     char const* outfile,
     char const* xarg)
 {
-    /* Empty PDF -- infile is ignored*/
+    /* Empty PDF -- infile is ignored */
     assert(qpdf_empty_pdf(qpdf) == 0);
     qpdf_init_write(qpdf, outfile);
     qpdf_set_static_ID(qpdf, QPDF_TRUE);
@@ -1466,6 +1473,110 @@ test41(
     report_errors();
 }
+static void
+test42(
+    char const* infile,
+    char const* password,
+    char const* outfile,
+    char const* xarg)
+{
+    assert(qpdf_create_from_json_file(qpdf, infile) == QPDF_SUCCESS);
+    qpdf_init_write(qpdf, outfile);
+    qpdf_set_static_ID(qpdf, QPDF_TRUE);
+    qpdf_write(qpdf);
+    report_errors();
+}
+
+static void
+test43(
+    char const* infile,
+    char const* password,
+    char const* outfile,
+    char const* xarg)
+{
+    char* buf = NULL;
+    unsigned long size = 0;
+    read_file_into_memory(infile, &buf, &size);
+    assert(qpdf_create_from_json_data(qpdf, buf, size) == QPDF_SUCCESS);
+    qpdf_init_write(qpdf, outfile);
+    qpdf_set_static_ID(qpdf, QPDF_TRUE);
+    qpdf_write(qpdf);
+    report_errors();
+    free(buf);
+}
+
+static void
+test44(
+    char const* infile,
+    char const* password,
+    char const* outfile,
+    char const* xarg)
+{
+    assert(qpdf_read(qpdf, infile, password) == 0);
+    assert(qpdf_update_from_json_file(qpdf, xarg) == QPDF_SUCCESS);
+    qpdf_init_write(qpdf, outfile);
+    qpdf_set_static_ID(qpdf, QPDF_TRUE);
+    qpdf_write(qpdf);
+    report_errors();
+}
+
+static void
+test45(
+    char const* infile,
+    char const* password,
+    char const* outfile,
+    char const* xarg)
+{
+    char* buf = NULL;
+    unsigned long size = 0;
+    read_file_into_memory(xarg, &buf, &size);
+    assert(qpdf_read(qpdf, infile, password) == 0);
+    assert(qpdf_update_from_json_data(qpdf, buf, size) == QPDF_SUCCESS);
+    qpdf_init_write(qpdf, outfile);
+    qpdf_set_static_ID(qpdf, QPDF_TRUE);
+    qpdf_write(qpdf);
+    report_errors();
+    free(buf);
+}
+
+static void
+test46(
+    char const* infile,
+    char const* password,
+    char const* outfile,
+    char const* xarg)
+{
+    FILE* f = safe_fopen(outfile, "wb");
+    assert(qpdf_read(qpdf, infile, password) == 0);
+    qpdf_write_json(
+        qpdf, 2, write_to_file, f, qpdf_dl_none, qpdf_sj_inline, "", NULL);
+    fclose(f);
+    report_errors();
+}
+
+static void
+test47(
+    char const* infile,
+    char const* password,
+    char const* outfile,
+    char const* xarg)
+{
+    FILE* f = safe_fopen(outfile, "wb");
+    assert(qpdf_read(qpdf, infile, password) == 0);
+    char const* wanted_objects[] = {"obj:4 0 R", "trailer", NULL};
+    qpdf_write_json(
+        qpdf,
+        2,
+        write_to_file,
+        f,
+        qpdf_dl_specialized,
+        qpdf_sj_file,
+        xarg,
+        wanted_objects);
+    fclose(f);
+    report_errors();
+}
+
 int
 main(int argc, char* argv[])
 {
@@ -1542,6 +1653,12 @@ main(int argc, char* argv[])
              : (n == 39) ? test39
              : (n == 40) ? test40
              : (n == 41) ? test41
+             : (n == 42) ? test42
+             : (n == 43) ? test43
+             : (n == 44) ? test44
+             : (n == 45) ? test45
+             : (n == 46) ? test46
+             : (n == 47) ? test47
                          : 0);
     if (fn == 0) {
@@ -288,5 +288,58 @@ $td-&gt;runtest(&quot;simple version of writeJSON&quot;,
              {$td->FILE => "minimal-write-json.json", $td->EXIT_STATUS => 0},
              $td->NORMALIZE_NEWLINES);
+$n_tests += 13;
+$td->runtest("C API create from json file",
+             {$td->COMMAND => "qpdf-ctest 42 minimal.json '' a.pdf"},
+             {$td->STRING => "C test 42 done\n", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("check C API create from file",
+             {$td->FILE => "a.pdf"},
+             {$td->FILE => "qpdf-ctest-42-43.pdf"});
+$td->runtest("C API create from json buffer",
+             {$td->COMMAND => "qpdf-ctest 43 minimal.json '' a.pdf"},
+             {$td->STRING => "C test 43 done\n", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("check C API create from buffer",
+             {$td->FILE => "a.pdf"},
+             {$td->FILE => "qpdf-ctest-42-43.pdf"});
+$td->runtest("C API update from json file",
+             {$td->COMMAND =>
+                  "qpdf-ctest 44 minimal.pdf '' a.pdf minimal-update.json"},
+             {$td->STRING => "C test 44 done\n", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("check C API update from file",
+             {$td->FILE => "a.pdf"},
+             {$td->FILE => "qpdf-ctest-44-45.pdf"});
+$td->runtest("C API update from json buffer",
+             {$td->COMMAND =>
+                  "qpdf-ctest 45 minimal.pdf '' a.pdf minimal-update.json"},
+             {$td->STRING => "C test 45 done\n", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("check C API update from buffer",
+             {$td->FILE => "a.pdf"},
+             {$td->FILE => "qpdf-ctest-44-45.pdf"});
+$td->runtest("C API write to JSON 1",
+             {$td->COMMAND =>
+                  "qpdf-ctest 46 minimal.pdf '' a.json"},
+             {$td->STRING => "C test 46 done\n", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("check C API write to JSON 1",
+             {$td->FILE => "a.json"},
+             {$td->FILE => "qpdf-ctest-46.json"},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("C API write to JSON 2",
+             {$td->COMMAND =>
+                  "qpdf-ctest 47 minimal.pdf '' a.json auto"},
+             {$td->STRING => "C test 47 done\n", $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("check C API write to JSON 2",
+             {$td->FILE => "a.json"},
+             {$td->FILE => "qpdf-ctest-47.json"},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("check C API write to JSON stream",
+             {$td->FILE => "auto-4"},
+             {$td->FILE => "qpdf-ctest-47-4"});
+
 cleanup();
 $td->report($n_tests);
+{
+  "qpdf": [
+    {
+      "jsonversion": 2,
+      "pushedinheritedpageresources": false,
+      "calledgetallpages": false
+    },
+    {
+      "obj:4 0 R": {
+        "stream": {
+          "data": "QlQKICAvRjEgMjQgVGYKICA3MiA3MjAgVGQKICAoU2FsYWQpIFRqCkVUCg==",
+          "dict": {}
+        }
+      }
+    }
+  ]
+}
+{
+  "qpdf": [
+    {
+      "jsonversion": 2,
+      "pdfversion": "1.3",
+      "pushedinheritedpageresources": false,
+      "calledgetallpages": false,
+      "maxobjectid": 6
+    },
+    {
+      "obj:1 0 R": {
+        "value": {
+          "/Pages": "2 0 R",
+          "/Type": "/Catalog"
+        }
+      },
+      "obj:2 0 R": {
+        "value": {
+          "/Count": 1,
+          "/Kids": [
+            "3 0 R"
+          ],
+          "/Type": "/Pages"
+        }
+      },
+      "obj:3 0 R": {
+        "value": {
+          "/Contents": "4 0 R",
+          "/MediaBox": [
+            0,
+            0,
+            612,
+            792
+          ],
+          "/Parent": "2 0 R",
+          "/Resources": {
+            "/Font": {
+              "/F1": "6 0 R"
+            },
+            "/ProcSet": "5 0 R"
+          },
+          "/Type": "/Page"
+        }
+      },
+      "obj:4 0 R": {
+        "stream": {
+          "data": "QlQKICAvRjEgMjQgVGYKICA3MiA3MjAgVGQKICAoUG90YXRvKSBUagpFVAo=",
+          "dict": {}
+        }
+      },
+      "obj:5 0 R": {
+        "value": [
+          "/PDF",
+          "/Text"
+        ]
+      },
+      "obj:6 0 R": {
+        "value": {
+          "/BaseFont": "/Helvetica",
+          "/Encoding": "/WinAnsiEncoding",
+          "/Name": "/F1",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  ]
+}
+{
+  "qpdf": [
+    {
+      "jsonversion": 2,
+      "pdfversion": "1.3",
+      "pushedinheritedpageresources": false,
+      "calledgetallpages": false,
+      "maxobjectid": 6
+    },
+    {
+      "obj:1 0 R": {
+        "value": {
+          "/Pages": "2 0 R",
+          "/Type": "/Catalog"
+        }
+      },
+      "obj:2 0 R": {
+        "value": {
+          "/Count": 1,
+          "/Kids": [
+            "3 0 R"
+          ],
+          "/Type": "/Pages"
+        }
+      },
+      "obj:3 0 R": {
+        "value": {
+          "/Contents": "4 0 R",
+          "/MediaBox": [
+            0,
+            0,
+            612,
+            792
+          ],
+          "/Parent": "2 0 R",
+          "/Resources": {
+            "/Font": {
+              "/F1": "6 0 R"
+            },
+            "/ProcSet": "5 0 R"
+          },
+          "/Type": "/Page"
+        }
+      },
+      "obj:4 0 R": {
+        "stream": {
+          "data": "QlQKICAvRjEgMjQgVGYKICA3MiA3MjAgVGQKICAoUG90YXRvKSBUagpFVAo=",
+          "dict": {}
+        }
+      },
+      "obj:5 0 R": {
+        "value": [
+          "/PDF",
+          "/Text"
+        ]
+      },
+      "obj:6 0 R": {
+        "value": {
+          "/BaseFont": "/Helvetica",
+          "/Encoding": "/WinAnsiEncoding",
+          "/Name": "/F1",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  ]
+}
+BT
+  /F1 24 Tf
+  72 720 Td
+  (Potato) Tj
+ET
+{
+  "qpdf": [
+    {
+      "jsonversion": 2,
+      "pdfversion": "1.3",
+      "pushedinheritedpageresources": false,
+      "calledgetallpages": false,
+      "maxobjectid": 6
+    },
+    {
+      "obj:4 0 R": {
+        "stream": {
+          "datafile": "auto-4",
+          "dict": {}
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  ]
+}