Allow optional fields in json "schema" checks

6 files changed, 113 insertions, 29 deletions

diff --git a/ChangeLog b/ChangeLog
index b70d6699..2433a76a 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,9 @@
 2022-01-22  Jay Berkenbilt  <ejb@ql.org>
 
+	* JSON: for (qpdf-specific, not official) "schema" checking, add
+	the ability to treat missing fields as optional. Also ensure that
+	values in the schema are dictionary, array, or string.
+
 	* Add convenience methods isNameAndEquals and isDictionaryOfType
 	to QPDFObjectHandle with corresponding functions added to the C
 	API. Thanks to m-holger for the contribution.
diff --git a/include/qpdf/JSON.hh b/include/qpdf/JSON.hh
index 3b13b4fe..d45fb3e1 100644
--- a/include/qpdf/JSON.hh
+++ b/include/qpdf/JSON.hh
@@ -107,21 +107,39 @@ class JSON
     //     single-element arrays, and strings only.
     //   * Recursively walk the schema
     //   * If the current value is a dictionary, this object must have
-    //     a dictionary in the same place with the same keys
+    //     a dictionary in the same place with the same keys. If flags
+    //     contains f_optional, a key in the schema does not have to
+    //     be present in the object. Otherwise, all keys have to be
+    //     present. Any key in the object must be present in the
+    //     schema.
     //   * If the current value is an array, this object must have an
     //     array in the same place. The schema's array must contain a
     //     single element, which is used as a schema to validate each
     //     element of this object's corresponding array.
-    //   * Otherwise, the value is ignored.
+    //   * Otherwise, the value must be a string whose value is a
+    //     description of the object's corresponding value, which may
+    //     have any type.
     //
     // QPDF's JSON output conforms to certain strict compatibility
     // rules as discussed in the manual. The idea is that a JSON
     // structure created manually in qpdf.cc doubles as both JSON help
     // information and a schema for validating the JSON that qpdf
     // generates. Any discrepancies are a bug in qpdf.
+    //
+    // Flags is a bitwise or of values from check_flags_e.
+    enum check_flags_e {
+        f_none = 0,
+        f_optional = 1 << 0,
+    };
+    QPDF_DLL
+    bool checkSchema(JSON schema, unsigned long flags,
+                     std::list<std::string>& errors);
+
+    // Same as passing 0 for flags
     QPDF_DLL
     bool checkSchema(JSON schema, std::list<std::string>& errors);
 
+
     // Create a JSON object from a string.
     QPDF_DLL
     static JSON parse(std::string const&);
@@ -180,6 +198,7 @@ class JSON
 
     static bool
     checkSchemaInternal(JSON_value* this_v, JSON_value* sch_v,
+                        unsigned long flags,
                         std::list<std::string>& errors,
                         std::string prefix);
 
diff --git a/libqpdf/JSON.cc b/libqpdf/JSON.cc
index af98553e..f70145a7 100644
--- a/libqpdf/JSON.cc
+++ b/libqpdf/JSON.cc
@@ -394,12 +394,21 @@ JSON::checkSchema(JSON schema, std::list<std::string>& errors)
 {
     return checkSchemaInternal(this->m->value.getPointer(),
                                schema.m->value.getPointer(),
-                               errors, "");
+                               0, errors, "");
 }
 
+bool
+JSON::checkSchema(JSON schema, unsigned long flags,
+                  std::list<std::string>& errors)
+{
+    return checkSchemaInternal(this->m->value.getPointer(),
+                               schema.m->value.getPointer(),
+                               flags, errors, "");
+}
 
 bool
 JSON::checkSchemaInternal(JSON_value* this_v, JSON_value* sch_v,
+                          unsigned long flags,
                           std::list<std::string>& errors,
                           std::string prefix)
 {
@@ -409,6 +418,8 @@ JSON::checkSchemaInternal(JSON_value* this_v, JSON_value* sch_v,
     JSON_array* sch_arr = dynamic_cast<JSON_array*>(sch_v);
     JSON_dictionary* sch_dict = dynamic_cast<JSON_dictionary*>(sch_v);
 
+    JSON_string* sch_str = dynamic_cast<JSON_string*>(sch_v);
+
     std::string err_prefix;
     if (prefix.empty())
     {
@@ -446,34 +457,38 @@ JSON::checkSchemaInternal(JSON_value* this_v, JSON_value* sch_v,
         {
             std::string const& key = iter.first;
             checkSchemaInternal(
-                this_dict->members[key].getPointer(),
-                pattern_schema,
-                errors, prefix + "." + key);
+                this_dict->members[key].getPointer(), pattern_schema,
+                flags, errors, prefix + "." + key);
         }
     }
     else if (sch_dict)
     {
-        for (std::map<std::string, PointerHolder<JSON_value> >::iterator iter =
-                 sch_dict->members.begin();
-             iter != sch_dict->members.end(); ++iter)
+        for (auto& iter: sch_dict->members)
         {
-            std::string const& key = (*iter).first;
+            std::string const& key = iter.first;
             if (this_dict->members.count(key))
             {
                 checkSchemaInternal(
                     this_dict->members[key].getPointer(),
-                    (*iter).second.getPointer(),
-                    errors, prefix + "." + key);
+                    iter.second.getPointer(),
+                    flags, errors, prefix + "." + key);
             }
             else
             {
-                QTC::TC("libtests", "JSON key missing in object");
-                errors.push_back(
-                    err_prefix + ": key \"" + key +
-                    "\" is present in schema but missing in object");
+                if (flags & f_optional)
+                {
+                    QTC::TC("libtests", "JSON optional key");
+                }
+                else
+                {
+                    QTC::TC("libtests", "JSON key missing in object");
+                    errors.push_back(
+                        err_prefix + ": key \"" + key +
+                        "\" is present in schema but missing in object");
+                }
             }
         }
-        for (std::map<std::string, PointerHolder<JSON_value> >::iterator iter =
+        for (std::map<std::string, PointerHolder<JSON_value>>::iterator iter =
                  this_dict->members.begin();
              iter != this_dict->members.end(); ++iter)
         {
@@ -510,9 +525,16 @@ JSON::checkSchemaInternal(JSON_value* this_v, JSON_value* sch_v,
             checkSchemaInternal(
                 (*iter).getPointer(),
                 sch_arr->elements.at(0).getPointer(),
-                errors, prefix + "." + QUtil::int_to_string(i));
+                flags, errors, prefix + "." + QUtil::int_to_string(i));
         }
     }
+    else if (! sch_str)
+    {
+        QTC::TC("libtests", "JSON schema other type");
+        errors.push_back(err_prefix +
+                         " schema value is not dictionary, array, or string");
+        return false;
+    }
 
     return errors.empty();
 }
diff --git a/libtests/json.cc b/libtests/json.cc
index 006f00cf..754c2b3e 100644
--- a/libtests/json.cc
+++ b/libtests/json.cc
@@ -112,12 +112,12 @@ static void test_main()
     assert(dvalue == xdvalue);
 }
 
-static void check_schema(JSON& obj, JSON& schema, bool exp,
-                         std::string const& description)
+static void check_schema(JSON& obj, JSON& schema, unsigned long flags,
+                         bool exp, std::string const& description)
 {
     std::list<std::string> errors;
     std::cout << "--- " << description << std::endl;
-    assert(exp == obj.checkSchema(schema, errors));
+    assert(exp == obj.checkSchema(schema, flags, errors));
     for (std::list<std::string>::iterator iter = errors.begin();
          iter != errors.end(); ++iter)
     {
@@ -134,8 +134,7 @@ static void test_schema()
     "a": {
       "q": "queue",
       "r": {
-        "x": "ecks",
-        "y": "(bool) why"
+        "x": "ecks"
       },
       "s": [
         "esses"
@@ -151,14 +150,14 @@ static void test_schema()
   "three": {
     "<objid>": {
       "z": "ebra",
-      "o": "(optional, string) optional"
+      "o": "ptional"
     }
   }
 }
 )");
 
     JSON a = JSON::parse(R"(["not a", "dictionary"])");
-    check_schema(a, schema, false, "top-level type mismatch");
+    check_schema(a, schema, 0, false, "top-level type mismatch");
     JSON b = JSON::parse(R"(
 {
   "one": {
@@ -205,10 +204,42 @@ static void test_schema()
 }
 )");
 
-    check_schema(b, schema, false, "missing items");
-    check_schema(a, a, false, "top-level schema array error");
-    check_schema(b, b, false, "lower-level schema array error");
-    check_schema(schema, schema, true, "pass");
+    check_schema(b, schema, 0, false, "missing items");
+    check_schema(a, a, 0, false, "top-level schema array error");
+    check_schema(b, b, 0, false, "lower-level schema array error");
+
+    JSON bad_schema = JSON::parse(R"({"a": true, "b": "potato?"})");
+    check_schema(bad_schema, bad_schema, 0, false, "bad schema field type");
+
+    JSON good = JSON::parse(R"(
+{
+  "one": {
+    "a": {
+      "q": "potato",
+      "r": {
+        "x": [1, null]
+      },
+      "s": [
+        null,
+        "anything"
+      ]
+    }
+  },
+  "two": [
+    {
+      "glarp": "enspliel",
+      "goose": 3.14
+    }
+  ],
+  "three": {
+    "<objid>": {
+      "z": "ebra"
+    }
+  }
+}
+)");
+    check_schema(good, schema, 0, false, "not optional");
+    check_schema(good, schema, JSON::f_optional, true, "pass");
 }
 
 int main()
diff --git a/libtests/libtests.testcov b/libtests/libtests.testcov
index 6284c0e8..907e6579 100644
--- a/libtests/libtests.testcov
+++ b/libtests/libtests.testcov
@@ -89,3 +89,5 @@ JSON parse premature end of u 0
 JSON parse bad hex after u 0
 JSONHandler unhandled value 0
 JSONHandler unexpected key 0
+JSON schema other type 0
+JSON optional key 0
diff --git a/libtests/qtest/json/json.out b/libtests/qtest/json/json.out
index 1a320dcf..c7cb85a7 100644
--- a/libtests/qtest/json/json.out
+++ b/libtests/qtest/json/json.out
@@ -21,6 +21,12 @@ top-level object schema array contains other than one item
 json key ".one.a.r" schema array contains other than one item
 json key ".two" schema array contains other than one item
 ---
+--- bad schema field type
+json key ".a" schema value is not dictionary, array, or string
+---
+--- not optional
+json key ".three.<objid>": key "o" is present in schema but missing in object
+---
 --- pass
 ---
 end of json tests