Upgrade Core to e2f82c8a43a3a9b2ff53171fa3243eaa451f8979 (#851)

Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>

Author: jviotti

DEPENDENCIES

@@ -1,5 +1,5 @@
 vendorpull https://github.com/sourcemeta/vendorpull 1dcbac42809cf87cb5b045106b863e17ad84ba02
-core https://github.com/sourcemeta/core a75b850e323ad3713c6ec907424600107892028b
+core https://github.com/sourcemeta/core e2f82c8a43a3a9b2ff53171fa3243eaa451f8979
 jsonschema-test-suite https://github.com/json-schema-org/JSON-Schema-Test-Suite 60755c1097769e313fae3ec4d63bcc9d49b5d2d5
 jsonschema-2020-12 https://github.com/json-schema-org/json-schema-spec 769daad75a9553562333a8937a187741cb708c72
 jsonschema-2019-09 https://github.com/json-schema-org/json-schema-spec 41014ea723120ce70b314d72f863c6929d9f3cfd

src/configuration/parse.cc

@@ -58,21 +58,22 @@ auto Configuration::from_json(const sourcemeta::core::JSON &value,
       !value.defines("dependencies") || value.at("dependencies").is_object(),
       "The dependencies property must be an object", {"dependencies"});
 
+  const sourcemeta::core::JSON null_value{nullptr};
   result.title =
       sourcemeta::core::from_json<decltype(result.title)::value_type>(
-          value.at_or("title", sourcemeta::core::JSON{nullptr}));
+          value.at_or("title", null_value));
   result.description =
       sourcemeta::core::from_json<decltype(result.description)::value_type>(
-          value.at_or("description", sourcemeta::core::JSON{nullptr}));
+          value.at_or("description", null_value));
   result.email =
       sourcemeta::core::from_json<decltype(result.email)::value_type>(
-          value.at_or("email", sourcemeta::core::JSON{nullptr}));
+          value.at_or("email", null_value));
   result.github =
       sourcemeta::core::from_json<decltype(result.github)::value_type>(
-          value.at_or("github", sourcemeta::core::JSON{nullptr}));
+          value.at_or("github", null_value));
   result.website =
       sourcemeta::core::from_json<decltype(result.website)::value_type>(
-          value.at_or("website", sourcemeta::core::JSON{nullptr}));
+          value.at_or("website", null_value));
 
   if (value.defines("path")) {
     const std::filesystem::path path{value.at("path").to_string()};
@@ -113,7 +114,7 @@ auto Configuration::from_json(const sourcemeta::core::JSON &value,
 
   result.default_dialect =
       sourcemeta::core::from_json<decltype(result.default_dialect)::value_type>(
-          value.at_or("defaultDialect", sourcemeta::core::JSON{nullptr}));
+          value.at_or("defaultDialect", null_value));
 
   if (value.defines("extension")) {
     result.extension.clear();

src/frame/frame.cc

@@ -321,6 +321,24 @@ auto supports_id_anchors(
   }
 }
 
+// Generic URI normalisation only decodes unreserved characters (see RFC 3986,
+// section 6.2.2.2), so a reference destination may still spell its JSON
+// Pointer fragment with percent-encoded octets. Re-serialise such fragments
+// from their parsed form so that reference destinations match the URIs that
+// locations are framed under
+auto canonicalize_pointer_fragment(sourcemeta::core::URI &uri) -> void {
+  const auto fragment{uri.fragment()};
+  if (!fragment.has_value() ||
+      fragment.value().find('%') == std::string_view::npos) {
+    return;
+  }
+
+  const auto destination{sourcemeta::core::fragment_to_pointer(uri)};
+  if (destination.has_value()) {
+    uri.fragment(sourcemeta::core::to_string(destination.value()));
+  }
+}
+
 auto set_base_and_fragment(
     sourcemeta::blaze::SchemaFrame::ReferencesEntry &entry) -> void {
   const std::string_view destination_view{entry.destination};
@@ -1042,6 +1060,7 @@ auto SchemaFrame::analyse(const sourcemeta::core::JSON &root,
         }
 
         ref.canonicalize();
+        canonicalize_pointer_fragment(ref);
         auto ref_pointer{common_pointer_weak};
         ref_pointer.push_back(std::cref(KEYWORD_REF));
         const auto [it, inserted] = this->references_.insert_or_assign(
@@ -1127,6 +1146,7 @@ auto SchemaFrame::analyse(const sourcemeta::core::JSON &root,
         }
 
         ref.canonicalize();
+        canonicalize_pointer_fragment(ref);
         auto ref_string{ref.recompose()};
 
         // Note that here we cannot enforce the bookending requirement,

test/frame/frame_2020_12_test.cc

@@ -2932,6 +2932,76 @@ TEST(Frame_2020_12, ref_with_id) {
       "https://www.sourcemeta.com/schema#/$defs/string");
 }
 
+TEST(Frame_2020_12, ref_with_percent_encoded_colon_in_fragment) {
+  const sourcemeta::core::JSON document = sourcemeta::core::parse_json(R"JSON({
+    "$schema": "https://json-schema.org/draft/2020-12/schema",
+    "$ref": "#/$defs/https%3A~1~1example.com~1schema~1type",
+    "$defs": {
+      "https://example.com/schema/type": { "type": "string" }
+    }
+  })JSON");
+
+  sourcemeta::blaze::SchemaFrame frame{
+      sourcemeta::blaze::SchemaFrame::Mode::References};
+  frame.analyse(document, sourcemeta::blaze::schema_walker,
+                sourcemeta::blaze::schema_resolver);
+
+  EXPECT_EQ(frame.locations().size(), 6);
+
+  // JSON Pointers
+
+  EXPECT_ANONYMOUS_FRAME_STATIC_SUBSCHEMA(
+      frame, "", "", "https://json-schema.org/draft/2020-12/schema",
+      JSON_Schema_2020_12, std::nullopt, false, false);
+  EXPECT_ANONYMOUS_FRAME_STATIC_POINTER(
+      frame, "#/$schema", "/$schema",
+      "https://json-schema.org/draft/2020-12/schema", JSON_Schema_2020_12, "",
+      false, false);
+  EXPECT_ANONYMOUS_FRAME_STATIC_POINTER(
+      frame, "#/$ref", "/$ref", "https://json-schema.org/draft/2020-12/schema",
+      JSON_Schema_2020_12, "", false, false);
+  EXPECT_ANONYMOUS_FRAME_STATIC_POINTER(
+      frame, "#/$defs", "/$defs",
+      "https://json-schema.org/draft/2020-12/schema", JSON_Schema_2020_12, "",
+      false, false);
+  EXPECT_ANONYMOUS_FRAME_STATIC_SUBSCHEMA(
+      frame, "#/$defs/https:~1~1example.com~1schema~1type",
+      "/$defs/https:~1~1example.com~1schema~1type",
+      "https://json-schema.org/draft/2020-12/schema", JSON_Schema_2020_12, "",
+      false, true);
+  EXPECT_ANONYMOUS_FRAME_STATIC_POINTER(
+      frame, "#/$defs/https:~1~1example.com~1schema~1type/type",
+      "/$defs/https:~1~1example.com~1schema~1type/type",
+      "https://json-schema.org/draft/2020-12/schema", JSON_Schema_2020_12,
+      "/$defs/https:~1~1example.com~1schema~1type", false, true);
+
+  // References
+
+  EXPECT_EQ(frame.references().size(), 2);
+
+  EXPECT_STATIC_REFERENCE(
+      frame, "/$schema", "https://json-schema.org/draft/2020-12/schema",
+      "https://json-schema.org/draft/2020-12/schema", std::nullopt,
+      "https://json-schema.org/draft/2020-12/schema");
+  EXPECT_STATIC_REFERENCE(frame, "/$ref",
+                          "#/$defs/https:~1~1example.com~1schema~1type", "",
+                          "/$defs/https:~1~1example.com~1schema~1type",
+                          "#/$defs/https%3A~1~1example.com~1schema~1type");
+
+  // Reachability
+
+  EXPECT_FRAME_LOCATION_REACHABLE(frame, Static, "", frame.root());
+  EXPECT_FRAME_LOCATION_REACHABLE(frame, Static,
+                                  "#/$defs/https:~1~1example.com~1schema~1type",
+                                  frame.root());
+
+  EXPECT_FRAME_LOCATION_NON_REACHABLE(
+      frame, Static, "", "#/$defs/https:~1~1example.com~1schema~1type");
+  EXPECT_FRAME_LOCATION_REACHABLE(
+      frame, Static, "#/$defs/https:~1~1example.com~1schema~1type",
+      "#/$defs/https:~1~1example.com~1schema~1type");
+}
+
 TEST(Frame_2020_12, ref_from_definitions) {
   const sourcemeta::core::JSON document = sourcemeta::core::parse_json(R"JSON({
     "$id": "https://www.sourcemeta.com/schema",