Implement top-level qpdf json parsing

37 files changed, 1053 insertions, 8 deletions

diff --git a/TODO b/TODO
index ef5d6d88..afc68a9c 100644
--- a/TODO
+++ b/TODO
@@ -54,6 +54,13 @@ Soon: Break ground on "Document-level work"
 Output JSON v2
 ==============
 
+Try to never flatten pages tree. Make sure we do something reasonable
+with pages tree repair. The problem is that if pages tree repair is
+done as a side effect of running --json, the qpdf part of the json may
+contain object numbers that aren't there. Maybe we need to indicate
+whether pages tree repair has been done in the json, but this would
+have to be known early in parsing, which is a problem.
+
 General things to remember:
 
 * Make sure all the information from --check and other informational
diff --git a/include/qpdf/QPDF.hh b/include/qpdf/QPDF.hh
index 61efa4ed..2e24b261 100644
--- a/include/qpdf/QPDF.hh
+++ b/include/qpdf/QPDF.hh
@@ -998,6 +998,7 @@ class QPDF
     class JSONReactor: public JSON::Reactor
     {
       public:
+        JSONReactor(QPDF&, bool must_be_complete);
         virtual ~JSONReactor() = default;
         virtual void dictionaryStart() override;
         virtual void arrayStart() override;
@@ -1006,6 +1007,32 @@ class QPDF
         virtual bool
         dictionaryItem(std::string const& key, JSON const& value) override;
         virtual bool arrayItem(JSON const& value) override;
+
+      private:
+        enum state_e {
+            st_initial,
+            st_top,
+            st_ignore,
+            st_qpdf,
+            st_objects_top,
+            st_trailer_top,
+            st_object_top,
+            st_stream,
+            st_object,
+        };
+
+        void containerStart();
+        void nestedState(std::string const& key, JSON const& value, state_e);
+
+        QPDF& pdf;
+        bool must_be_complete;
+        bool saw_qpdf;
+        bool saw_json_version;
+        bool saw_pdf_version;
+        bool saw_trailer;
+        state_e state;
+        state_e next_state;
+        std::vector<state_e> state_stack;
     };
     friend class JSONReactor;
 
diff --git a/libqpdf/QPDF_json.cc b/libqpdf/QPDF_json.cc
index 316c9935..2f74a673 100644
--- a/libqpdf/QPDF_json.cc
+++ b/libqpdf/QPDF_json.cc
@@ -1,42 +1,218 @@
 #include <qpdf/QPDF.hh>
 
 #include <qpdf/FileInputSource.hh>
+#include <qpdf/QTC.hh>
+#include <qpdf/QUtil.hh>
+#include <regex>
+
+namespace
+{
+    class JSONExc: public std::runtime_error
+    {
+      public:
+        JSONExc(JSON const& value, std::string const& msg) :
+            std::runtime_error(
+                "offset " + QUtil::uint_to_string(value.getStart()) + ": " +
+                msg)
+        {
+        }
+    };
+} // namespace
+
+static std::regex PDF_VERSION_RE("^\\d+\\.\\d+$");
+static std::regex OBJ_KEY_RE("^obj:(\\d+) (\\d+) R$");
+
+QPDF::JSONReactor::JSONReactor(QPDF& pdf, bool must_be_complete) :
+    pdf(pdf),
+    must_be_complete(must_be_complete),
+    saw_qpdf(false),
+    saw_json_version(false),
+    saw_pdf_version(false),
+    saw_trailer(false),
+    state(st_initial),
+    next_state(st_top)
+{
+    state_stack.push_back(st_initial);
+}
+
+void
+QPDF::JSONReactor::containerStart()
+{
+    state_stack.push_back(state);
+    state = next_state;
+}
 
 void
 QPDF::JSONReactor::dictionaryStart()
 {
-    // QXXXXQ
+    containerStart();
+    // QXXXQ
 }
 
 void
 QPDF::JSONReactor::arrayStart()
 {
-    // QXXXXQ
+    containerStart();
+    if (state == st_top) {
+        QTC::TC("qpdf", "QPDF_json top-level array");
+        throw std::runtime_error("QPDF JSON must be a dictionary");
+    }
+    // QXXXQ
 }
 
 void
 QPDF::JSONReactor::containerEnd(JSON const& value)
 {
-    // QXXXXQ
+    state = state_stack.back();
+    state_stack.pop_back();
+    if (state == st_initial) {
+        if (!this->saw_qpdf) {
+            QTC::TC("qpdf", "QPDF_json missing qpdf");
+            throw std::runtime_error("\"qpdf\" object was not seen");
+        }
+        if (!this->saw_json_version) {
+            QTC::TC("qpdf", "QPDF_json missing json version");
+            throw std::runtime_error("\"qpdf.jsonversion\" was not seen");
+        }
+        if (must_be_complete && !this->saw_pdf_version) {
+            QTC::TC("qpdf", "QPDF_json missing pdf version");
+            throw std::runtime_error("\"qpdf.pdfversion\" was not seen");
+        }
+        if (must_be_complete && !this->saw_trailer) {
+            /// QTC::TC("qpdf", "QPDF_json missing trailer");
+            throw std::runtime_error("\"qpdf.objects.trailer\" was not seen");
+        }
+    }
+
+    // QXXXQ
 }
 
 void
 QPDF::JSONReactor::topLevelScalar()
 {
-    // QXXXXQ
+    QTC::TC("qpdf", "QPDF_json top-level scalar");
+    throw std::runtime_error("QPDF JSON must be a dictionary");
+}
+
+void
+QPDF::JSONReactor::nestedState(
+    std::string const& key, JSON const& value, state_e next)
+{
+    // Use this method when the next state is for processing a nested
+    // dictionary.
+    if (!value.isDictionary()) {
+        throw JSONExc(value, "\"" + key + "\" must be a dictionary");
+    }
+    this->next_state = next;
 }
 
 bool
 QPDF::JSONReactor::dictionaryItem(std::string const& key, JSON const& value)
 {
-    // QXXXXQ
+    if (state == st_ignore) {
+        // ignore
+    } else if (state == st_top) {
+        if (key == "qpdf") {
+            this->saw_qpdf = true;
+            nestedState(key, value, st_qpdf);
+        } else {
+            // Ignore all other fields for forward compatibility.
+            // Don't use nestedState since this can be any type.
+            next_state = st_ignore;
+        }
+    } else if (state == st_qpdf) {
+        if (key == "jsonversion") {
+            this->saw_json_version = true;
+            std::string v;
+            if (!(value.getNumber(v) && (v == "2"))) {
+                QTC::TC("qpdf", "QPDF_json bad json version");
+                throw JSONExc(value, "only JSON version 2 is supported");
+            }
+        } else if (key == "pdfversion") {
+            this->saw_pdf_version = true;
+            bool version_okay = false;
+            std::string v;
+            if (value.getString(v)) {
+                std::smatch m;
+                if (std::regex_match(v, m, PDF_VERSION_RE)) {
+                    version_okay = true;
+                    this->pdf.m->pdf_version = v;
+                }
+            }
+            if (!version_okay) {
+                QTC::TC("qpdf", "QPDF_json bad pdf version");
+                throw JSONExc(value, "invalid PDF version (must be x.y)");
+            }
+        } else if (key == "objects") {
+            nestedState(key, value, st_objects_top);
+        } else {
+            // ignore unknown keys for forward compatibility
+        }
+    } else if (state == st_objects_top) {
+        std::smatch m;
+        if (key == "trailer") {
+            this->saw_trailer = true;
+            nestedState(key, value, st_trailer_top);
+            // QXXXQ
+        } else if (std::regex_match(key, m, OBJ_KEY_RE)) {
+            nestedState(key, value, st_object_top);
+            // QXXXQ
+        } else {
+            QTC::TC("qpdf", "QPDF_json bad object key");
+            throw JSONExc(
+                value, "object key should be \"trailer\" or \"obj:n n R\"");
+        }
+    } else if (state == st_object_top) {
+        if (key == "value") {
+            // Don't use nestedState since this can have any type.
+            next_state = st_object;
+            // QXXXQ
+        } else if (key == "stream") {
+            nestedState(key, value, st_stream);
+            // QXXXQ
+        } else {
+            // Ignore unknown keys for forward compatibility
+        }
+    } else if (state == st_trailer_top) {
+        if (key == "value") {
+            // The trailer must be a dictionary, so we can use nestedState.
+            nestedState("trailer.value", value, st_object);
+            // QXXXQ
+        } else if (key == "stream") {
+            QTC::TC("qpdf", "QPDF_json trailer stream");
+            throw JSONExc(value, "the trailer may not be a stream");
+        } else {
+            // Ignore unknown keys for forward compatibility
+        }
+    } else if (state == st_stream) {
+        if (key == "dict") {
+            // Since a stream dictionary must be a dictionary, we can
+            // use nestedState to transition to st_value.
+            nestedState("stream.dict", value, st_object);
+            // QXXXQ
+        } else if (key == "data") {
+            // QXXXQ
+        } else if (key == "datafile") {
+            // QXXXQ
+        } else {
+            // Ignore unknown keys for forward compatibility.
+            next_state = st_ignore;
+        }
+    } else if (state == st_object) {
+        // QXXXQ
+    } else {
+        throw std::logic_error(
+            "QPDF_json: unknown state " + QUtil::int_to_string(state));
+    }
+
+    // QXXXQ
     return true;
 }
 
 bool
 QPDF::JSONReactor::arrayItem(JSON const& value)
 {
-    // QXXXXQ
+    // QXXXQ
     return true;
 }
 
@@ -65,7 +241,12 @@ QPDF::updateFromJSON(std::shared_ptr<InputSource> is)
 }
 
 void
-QPDF::importJSON(std::shared_ptr<InputSource>, bool must_be_complete)
+QPDF::importJSON(std::shared_ptr<InputSource> is, bool must_be_complete)
 {
-    // QXXXQ
+    JSONReactor reactor(*this, must_be_complete);
+    try {
+        JSON::parse(*is, &reactor);
+    } catch (std::runtime_error& e) {
+        throw std::runtime_error(is->getName() + ": " + e.what());
+    }
 }
diff --git a/qpdf/qpdf.testcov b/qpdf/qpdf.testcov
index 4225fc64..69bd350b 100644
--- a/qpdf/qpdf.testcov
+++ b/qpdf/qpdf.testcov
@@ -650,3 +650,12 @@ QPDFJob json encrypt duplicate key length 0
 QPDFJob json encrypt missing password 0
 QPDFJob json pages no file 0
 qpdf-c called qpdf_empty_pdf 0
+QPDF_json missing qpdf 0
+QPDF_json missing json version 0
+QPDF_json missing pdf version 0
+QPDF_json top-level scalar 0
+QPDF_json bad json version 0
+QPDF_json bad pdf version 0
+QPDF_json top-level array 0
+QPDF_json bad object key 0
+QPDF_json trailer stream 0
diff --git a/qpdf/qtest/qpdf-json.test b/qpdf/qtest/qpdf-json.test
new file mode 100644
index 00000000..39078a87
--- /dev/null
+++ b/qpdf/qtest/qpdf-json.test
@@ -0,0 +1,50 @@
+#!/usr/bin/env perl
+require 5.008;
+use warnings;
+use strict;
+
+unshift(@INC, '.');
+require qpdf_test_helpers;
+
+chdir("qpdf") or die "chdir testdir failed: $!\n";
+
+require TestDriver;
+
+cleanup();
+
+my $td = new TestDriver('qpdf-json');
+
+my $n_tests = 0;
+
+my @badfiles = (
+    'no-qpdf-object',
+    'no-json-version',
+    'no-pdf-version',
+    'top-level-scalar',
+    'bad-json-version1',
+    'bad-json-version2',
+    'bad-pdf-version1',
+    'bad-pdf-version2',
+    'top-level-array',
+    'objects-not-dict',
+    'bad-object-key',
+    'object-not-dict',
+    'stream-not-dict',
+    'stream-dict-not-dict',
+    'trailer-not-dict',
+    'trailer-stream',
+    );
+
+$n_tests += scalar(@badfiles);
+
+foreach my $f (@badfiles)
+{
+    $td->runtest("bad: $f",
+                 {$td->COMMAND =>
+                      "qpdf --create-from-json=qjson-$f.json a.pdf"},
+                 {$td->FILE => "qjson-$f.out", $td->EXIT_STATUS => 2},
+                 $td->NORMALIZE_NEWLINES);
+}
+
+cleanup();
+$td->report($n_tests);
diff --git a/qpdf/qtest/qpdf/qjson-bad-json-version1.json b/qpdf/qtest/qpdf/qjson-bad-json-version1.json
new file mode 100644
index 00000000..78591a42
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-json-version1.json
@@ -0,0 +1,73 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "jsonversion": 16059,
+    "pdfversion": "1.3",
+    "maxobjectid": 6,
+    "objects": {
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-bad-json-version1.out b/qpdf/qtest/qpdf/qjson-bad-json-version1.out
new file mode 100644
index 00000000..1c921c2b
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-json-version1.out
@@ -0,0 +1 @@
+qpdf: qjson-bad-json-version1.json: offset 98: only JSON version 2 is supported
diff --git a/qpdf/qtest/qpdf/qjson-bad-json-version2.json b/qpdf/qtest/qpdf/qjson-bad-json-version2.json
new file mode 100644
index 00000000..20843caf
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-json-version2.json
@@ -0,0 +1,73 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "jsonversion": "potato",
+    "pdfversion": "1.3",
+    "maxobjectid": 6,
+    "objects": {
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-bad-json-version2.out b/qpdf/qtest/qpdf/qjson-bad-json-version2.out
new file mode 100644
index 00000000..756df7fe
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-json-version2.out
@@ -0,0 +1 @@
+qpdf: qjson-bad-json-version2.json: offset 98: only JSON version 2 is supported
diff --git a/qpdf/qtest/qpdf/qjson-bad-object-key.json b/qpdf/qtest/qpdf/qjson-bad-object-key.json
new file mode 100644
index 00000000..77441846
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-object-key.json
@@ -0,0 +1,75 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "jsonversion": 2,
+    "pdfversion": "1.3",
+    "maxobjectid": 6,
+    "objects": {
+      "potato": {
+      },
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-bad-object-key.out b/qpdf/qtest/qpdf/qjson-bad-object-key.out
new file mode 100644
index 00000000..91edfc8f
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-object-key.out
@@ -0,0 +1 @@
+qpdf: qjson-bad-object-key.json: offset 181: object key should be "trailer" or "obj:n n R"
diff --git a/qpdf/qtest/qpdf/qjson-bad-pdf-version1.json b/qpdf/qtest/qpdf/qjson-bad-pdf-version1.json
new file mode 100644
index 00000000..eccaeeca
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-pdf-version1.json
@@ -0,0 +1,73 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "jsonversion": 2,
+    "pdfversion": "potato",
+    "maxobjectid": 6,
+    "objects": {
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-bad-pdf-version1.out b/qpdf/qtest/qpdf/qjson-bad-pdf-version1.out
new file mode 100644
index 00000000..61331957
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-pdf-version1.out
@@ -0,0 +1 @@
+qpdf: qjson-bad-pdf-version1.json: offset 119: invalid PDF version (must be x.y)
diff --git a/qpdf/qtest/qpdf/qjson-bad-pdf-version2.json b/qpdf/qtest/qpdf/qjson-bad-pdf-version2.json
new file mode 100644
index 00000000..04443eea
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-pdf-version2.json
@@ -0,0 +1,73 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "jsonversion": 2,
+    "pdfversion": [],
+    "maxobjectid": 6,
+    "objects": {
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-bad-pdf-version2.out b/qpdf/qtest/qpdf/qjson-bad-pdf-version2.out
new file mode 100644
index 00000000..2ef92df4
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-bad-pdf-version2.out
@@ -0,0 +1 @@
+qpdf: qjson-bad-pdf-version2.json: offset 119: invalid PDF version (must be x.y)
diff --git a/qpdf/qtest/qpdf/qjson-no-json-version.json b/qpdf/qtest/qpdf/qjson-no-json-version.json
new file mode 100644
index 00000000..72f223ad
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-no-json-version.json
@@ -0,0 +1,72 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "pdfversion": "1.3",
+    "maxobjectid": 6,
+    "objects": {
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-no-json-version.out b/qpdf/qtest/qpdf/qjson-no-json-version.out
new file mode 100644
index 00000000..b745bc85
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-no-json-version.out
@@ -0,0 +1 @@
+qpdf: qjson-no-json-version.json: "qpdf.jsonversion" was not seen
diff --git a/qpdf/qtest/qpdf/qjson-no-pdf-version.json b/qpdf/qtest/qpdf/qjson-no-pdf-version.json
new file mode 100644
index 00000000..39f21be2
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-no-pdf-version.json
@@ -0,0 +1,72 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "jsonversion": 2,
+    "maxobjectid": 6,
+    "objects": {
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-no-pdf-version.out b/qpdf/qtest/qpdf/qjson-no-pdf-version.out
new file mode 100644
index 00000000..cd752750
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-no-pdf-version.out
@@ -0,0 +1 @@
+qpdf: qjson-no-pdf-version.json: "qpdf.pdfversion" was not seen
diff --git a/qpdf/qtest/qpdf/qjson-no-qpdf-object.json b/qpdf/qtest/qpdf/qjson-no-qpdf-object.json
new file mode 100644
index 00000000..994a97e8
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-no-qpdf-object.json
@@ -0,0 +1,3 @@
+{
+  "potato": "salad"
+}
diff --git a/qpdf/qtest/qpdf/qjson-no-qpdf-object.out b/qpdf/qtest/qpdf/qjson-no-qpdf-object.out
new file mode 100644
index 00000000..765d1bc5
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-no-qpdf-object.out
@@ -0,0 +1 @@
+qpdf: qjson-no-qpdf-object.json: "qpdf" object was not seen
diff --git a/qpdf/qtest/qpdf/qjson-object-not-dict.json b/qpdf/qtest/qpdf/qjson-object-not-dict.json
new file mode 100644
index 00000000..c0831c74
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-object-not-dict.json
@@ -0,0 +1,68 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "jsonversion": 2,
+    "pdfversion": "1.3",
+    "maxobjectid": 6,
+    "objects": {
+      "obj:1 0 R": "potato",
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-object-not-dict.out b/qpdf/qtest/qpdf/qjson-object-not-dict.out
new file mode 100644
index 00000000..5ddeb7f2
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-object-not-dict.out
@@ -0,0 +1 @@
+qpdf: qjson-object-not-dict.json: offset 184: "obj:1 0 R" must be a dictionary
diff --git a/qpdf/qtest/qpdf/qjson-objects-not-dict.json b/qpdf/qtest/qpdf/qjson-objects-not-dict.json
new file mode 100644
index 00000000..0e16294a
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-objects-not-dict.json
@@ -0,0 +1,7 @@
+{
+  "qpdf": {
+    "jsonversion": 2,
+    "pdfversion": "1.7",
+    "objects": false
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-objects-not-dict.out b/qpdf/qtest/qpdf/qjson-objects-not-dict.out
new file mode 100644
index 00000000..7504428c
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-objects-not-dict.out
@@ -0,0 +1 @@
+qpdf: qjson-objects-not-dict.json: offset 77: "objects" must be a dictionary
diff --git a/qpdf/qtest/qpdf/qjson-stream-dict-not-dict.json b/qpdf/qtest/qpdf/qjson-stream-dict-not-dict.json
new file mode 100644
index 00000000..19bd08b1
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-stream-dict-not-dict.json
@@ -0,0 +1,13 @@
+{
+  "qpdf": {
+    "jsonversion": 2,
+    "pdfversion": "1.7",
+    "objects": {
+      "obj:1 0 R": {
+        "stream": {
+          "dict": "quack"
+        }
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-stream-dict-not-dict.out b/qpdf/qtest/qpdf/qjson-stream-dict-not-dict.out
new file mode 100644
index 00000000..0341766e
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-stream-dict-not-dict.out
@@ -0,0 +1 @@
+qpdf: qjson-stream-dict-not-dict.json: offset 137: "stream.dict" must be a dictionary
diff --git a/qpdf/qtest/qpdf/qjson-stream-not-dict.json b/qpdf/qtest/qpdf/qjson-stream-not-dict.json
new file mode 100644
index 00000000..cef86c95
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-stream-not-dict.json
@@ -0,0 +1,11 @@
+{
+  "qpdf": {
+    "jsonversion": 2,
+    "pdfversion": "1.7",
+    "objects": {
+      "obj:1 0 R": {
+        "stream": 3
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-stream-not-dict.out b/qpdf/qtest/qpdf/qjson-stream-not-dict.out
new file mode 100644
index 00000000..e1f85a29
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-stream-not-dict.out
@@ -0,0 +1 @@
+qpdf: qjson-stream-not-dict.json: offset 118: "stream" must be a dictionary
diff --git a/qpdf/qtest/qpdf/qjson-top-level-array.json b/qpdf/qtest/qpdf/qjson-top-level-array.json
new file mode 100644
index 00000000..2e2c15f8
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-top-level-array.json
@@ -0,0 +1 @@
+["potato"]
diff --git a/qpdf/qtest/qpdf/qjson-top-level-array.out b/qpdf/qtest/qpdf/qjson-top-level-array.out
new file mode 100644
index 00000000..b8a90532
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-top-level-array.out
@@ -0,0 +1 @@
+qpdf: qjson-top-level-array.json: QPDF JSON must be a dictionary
diff --git a/qpdf/qtest/qpdf/qjson-top-level-scalar.json b/qpdf/qtest/qpdf/qjson-top-level-scalar.json
new file mode 100644
index 00000000..a9d93e43
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-top-level-scalar.json
@@ -0,0 +1 @@
+"potato"
diff --git a/qpdf/qtest/qpdf/qjson-top-level-scalar.out b/qpdf/qtest/qpdf/qjson-top-level-scalar.out
new file mode 100644
index 00000000..75649cf1
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-top-level-scalar.out
@@ -0,0 +1 @@
+qpdf: qjson-top-level-scalar.json: QPDF JSON must be a dictionary
diff --git a/qpdf/qtest/qpdf/qjson-trailer-not-dict.json b/qpdf/qtest/qpdf/qjson-trailer-not-dict.json
new file mode 100644
index 00000000..69782074
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-trailer-not-dict.json
@@ -0,0 +1,70 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "jsonversion": 2,
+    "pdfversion": "1.3",
+    "maxobjectid": 6,
+    "objects": {
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": false,
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-trailer-not-dict.out b/qpdf/qtest/qpdf/qjson-trailer-not-dict.out
new file mode 100644
index 00000000..bec6c7ca
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-trailer-not-dict.out
@@ -0,0 +1 @@
+qpdf: qjson-trailer-not-dict.json: offset 1326: "trailer.value" must be a dictionary
diff --git a/qpdf/qtest/qpdf/qjson-trailer-stream.json b/qpdf/qtest/qpdf/qjson-trailer-stream.json
new file mode 100644
index 00000000..74891e5b
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-trailer-stream.json
@@ -0,0 +1,70 @@
+{
+  "version": 2,
+  "parameters": {
+    "decodelevel": "none"
+  },
+  "qpdf": {
+    "jsonversion": 2,
+    "pdfversion": "1.3",
+    "maxobjectid": 6,
+    "objects": {
+      "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",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "stream": {},
+      }
+    }
+  }
+}
diff --git a/qpdf/qtest/qpdf/qjson-trailer-stream.out b/qpdf/qtest/qpdf/qjson-trailer-stream.out
new file mode 100644
index 00000000..267ecf08
--- /dev/null
+++ b/qpdf/qtest/qpdf/qjson-trailer-stream.out
@@ -0,0 +1 @@
+qpdf: qjson-trailer-stream.json: offset 1327: the trailer may not be a stream