NQuads parser doesn't pass test suite

src/JsonLD/Core/JsonLdProcessor.cs

@@ -66,8 +66,25 @@ public static JArray Expand(JToken input, JsonLdOptions opts)
         {
             // 1)
             // TODO: look into java futures/promises
-            // 2) TODO: better verification of DOMString IRI
-            if (input.Type == JTokenType.String && ((string)input).Contains(":"))
+
+            // 2) verification of DOMString IRI
+            bool isIriString = input.Type == JTokenType.String;
+            if (isIriString)
+            {
+                bool hasColon = false;
+                foreach (var c in ((string) input))
+                {
+                    if (c == ':')
+                        hasColon = true;
+                    if (!hasColon && (c == '{' || c == '['))
+                    {
+                        isIriString = false;
+                        break;
+                    }
+                }
+            }
+
+            if (isIriString)
             {
                 try
                 {

src/JsonLD/Core/JsonLdUtils.cs

@@ -199,7 +199,7 @@ internal static void MergeCompactedValue(JObject obj, string
         public static bool IsAbsoluteIri(string value)
         {
             // TODO: this is a bit simplistic!
-            return value.Contains(":");
+            return value != null && value.Contains(":");
         }
 
         /// <summary>Returns true if the given value is a subject with properties.</summary>

src/JsonLD/Core/RDFDatasetUtils.cs

@@ -544,14 +544,19 @@ public static string Escape(string str)
 
         private class Regex
         {
-            public static readonly Pattern Iri = Pattern.Compile("(?:<([^>]*)>)");
+            public static readonly Pattern HEX = Pattern.Compile("[0-9A-Fa-f]");
 
-            public static readonly Pattern Bnode = Pattern.Compile("(_:(?:[A-Za-z][A-Za-z0-9]*))"
-                );
+            public static readonly Pattern UCHAR = Pattern.Compile("\\\\u" + HEX + "{4}|\\\\U" + HEX + "{8}");
+
+            public static readonly Pattern Iri = Pattern.Compile("(?:<((?:[^\\x00-\\x20<>\"{}|^`\\\\]|" + UCHAR + ")*)>)");
 
-            public static readonly Pattern Plain = Pattern.Compile("\"([^\"\\\\]*(?:\\\\.[^\"\\\\]*)*)\""
+            public static readonly Pattern Bnode = Pattern.Compile("(_:(?:[A-Za-z0-9](?:[A-Za-z0-9\\-\\.]*[A-Za-z0-9])?))"
                 );
 
+            public static readonly Pattern ECHAR = Pattern.Compile("\\\\[tbnrf\"'\\\\]");
+
+            public static readonly Pattern Plain = Pattern.Compile("\"((?:[^\\x22\\x5C\\x0A\\x0D]|" + ECHAR + "|" + UCHAR + ")*)\"");
+
             public static readonly Pattern Datatype = Pattern.Compile("(?:\\^\\^" + Iri + ")"
                 );
 
@@ -561,19 +566,16 @@ private class Regex
             public static readonly Pattern Literal = Pattern.Compile("(?:" + Plain + "(?:" + 
                 Datatype + "|" + Language + ")?)");
 
-            public static readonly Pattern Ws = Pattern.Compile("[ \\t]+");
-
             public static readonly Pattern Wso = Pattern.Compile("[ \\t]*");
 
             public static readonly Pattern Eoln = Pattern.Compile("(?:\r\n)|(?:\n)|(?:\r)");
 
-            public static readonly Pattern Empty = Pattern.Compile("^" + Wso + "$");
+            public static readonly Pattern EmptyOrComment = Pattern.Compile("^" + Wso + "(#.*)?$");
 
             public static readonly Pattern Subject = Pattern.Compile("(?:" + Iri + "|" + Bnode
-                 + ")" + Ws);
+                 + ")" + Wso);
 
-            public static readonly Pattern Property = Pattern.Compile(Iri.GetPattern() + Ws.GetPattern
-                ());
+            public static readonly Pattern Property = Pattern.Compile(Iri.GetPattern() + Wso);
 
             public static readonly Pattern Object = Pattern.Compile("(?:" + Iri + "|" + Bnode
                  + "|" + Literal + ")" + Wso);
@@ -582,7 +584,7 @@ private class Regex
                  + Bnode + ")" + Wso + "\\.))");
 
             public static readonly Pattern Quad = Pattern.Compile("^" + Wso + Subject + Property
-                 + Object + Graph + Wso + "$");
+                 + Object + Graph + Wso + "(#.*)?$");
             // define partial regexes
             // final public static Pattern IRI =
             // Pattern.compile("(?:<([^:]+:[^>]*)>)");
@@ -606,7 +608,7 @@ public static RDFDataset ParseNQuads(string input)
             {
                 lineNumber++;
                 // skip empty lines
-                if (RDFDatasetUtils.Regex.Empty.Matcher(line).Matches())
+                if (RDFDatasetUtils.Regex.EmptyOrComment.Matcher(line).Matches())
                 {
                     continue;
                 }
@@ -621,19 +623,25 @@ public static RDFDataset ParseNQuads(string input)
                 RDFDataset.Node subject;
                 if (match.Group(1) != null)
                 {
-                    subject = new RDFDataset.IRI(Unescape(match.Group(1)));
+                    var subjectIri = Unescape(match.Group(1));
+                    AssertAbsoluteIri(subjectIri);
+                    subject = new RDFDataset.IRI(subjectIri);
                 }
                 else
                 {
                     subject = new RDFDataset.BlankNode(Unescape(match.Group(2)));
                 }
                 // get predicate
-                RDFDataset.Node predicate = new RDFDataset.IRI(Unescape(match.Group(3)));
+                var predicateIri = Unescape(match.Group(3));
+                AssertAbsoluteIri(predicateIri);
+                RDFDataset.Node predicate = new RDFDataset.IRI(predicateIri);
                 // get object
                 RDFDataset.Node @object;
                 if (match.Group(4) != null)
                 {
-                    @object = new RDFDataset.IRI(Unescape(match.Group(4)));
+                    var objectIri = Unescape(match.Group(4));
+                    AssertAbsoluteIri(objectIri);
+                    @object = new RDFDataset.IRI(objectIri);
                 }
                 else
                 {
@@ -646,6 +654,7 @@ public static RDFDataset ParseNQuads(string input)
                         string language = Unescape(match.Group(8));
                         string datatype = match.Group(7) != null ? Unescape(match.Group(7)) : match.Group
                             (8) != null ? JSONLDConsts.RdfLangstring : JSONLDConsts.XsdString;
+                        AssertAbsoluteIri(datatype);
                         string unescaped = Unescape(match.Group(6));
                         @object = new RDFDataset.Literal(unescaped, datatype, language);
                     }
@@ -655,6 +664,7 @@ public static RDFDataset ParseNQuads(string input)
                 if (match.Group(9) != null)
                 {
                     name = Unescape(match.Group(9));
+                    AssertAbsoluteIri(name);
                 }
                 else
                 {
@@ -683,5 +693,13 @@ public static RDFDataset ParseNQuads(string input)
             }
             return dataset;
         }
+
+        private static void AssertAbsoluteIri(string iri)
+        {
+            if (Uri.IsWellFormedUriString(Uri.EscapeUriString(iri), UriKind.Absolute) == false)
+            {
+                throw new JsonLdError(JsonLdError.Error.SyntaxError, "Invalid absolute URI <" + iri + ">");
+            }
+        }
     }
 }

src/JsonLD/Util/JavaCompat.cs

@@ -67,7 +67,7 @@ public static bool SafeCompare<T>(this JToken token, T val)
         {
             try
             {
-                return token.Value<T>().Equals(val);
+                return token == null ? val == null : token.Value<T>().Equals(val);
             }
             catch
             {

tests/JsonLD.Test/JsonLD.Test.csproj

@@ -58,8 +58,14 @@
   </ItemGroup>
   <ItemGroup>
     <Compile Include="ConformanceTests.cs" />
+    <Compile Include="NQuadsParserTests.cs" />
     <Compile Include="Properties\AssemblyInfo.cs" />
   </ItemGroup>
+  <ItemGroup>
+    <Content Include="NQuads\*">
+      <CopyToOutputDirectory>Always</CopyToOutputDirectory>
+    </Content>
+  </ItemGroup>
   <ItemGroup>
     <Content Include="W3C\compact-0001-context.jsonld">
       <CopyToOutputDirectory>PreserveNewest</CopyToOutputDirectory>
@@ -3005,6 +3011,7 @@
       <Name>JsonLD</Name>
     </ProjectReference>
   </ItemGroup>
+  <ItemGroup />
   <Import Project="$(MSBuildToolsPath)\Microsoft.CSharp.targets" />
   <Import Project="$(SolutionDir)\.nuget\NuGet.targets" Condition="Exists('$(SolutionDir)\.nuget\NuGet.targets')" />
   <Target Name="EnsureNuGetPackageBuildImports" BeforeTargets="PrepareForBuild">

tests/JsonLD.Test/NQuads/README

@@ -0,0 +1,25 @@
+This README is for the W3C RDF Working Group's N-Quads test suite.
+This test suite contains two kinds of tests:
+
+  Positive syntax (rdft:TestNQuadsPositiveSyntax) - an input N-Quads
+  file with no syntax errors.
+
+  Negative syntax (rdft:TestNQuadsNegativeSyntax) - an input N-Quads
+  file with at least one syntax error.
+
+The manifest.ttl file in this directory lists tests in the
+RDF WG's N-Quads test suite. All
+tests have a name (mf:name) and an input (mf:action).
+
+• An implementation passes a positive syntax test if it parses the
+  input.
+
+• An implementation passes a negative syntax test if it fails to parse
+  the input.
+
+The home of the test suite is <http://www.w3.org/2013/NQuadsTests/>.
+
+See http://www.w3.org/2011/rdf-wg/wiki/RDF_Test_Suites for more details.
+
+Eric Prud'hommeaux <[email protected]> - 11 June 2013.
+Gregg Kellogg <[email protected]> - 26 June 2013.

tests/JsonLD.Test/NQuads/comment_following_triple.nq

@@ -0,0 +1,5 @@
+<http://example/s> <http://example/p> <http://example/o> . # comment
+<http://example/s> <http://example/p> _:o . # comment
+<http://example/s> <http://example/p> "o" . # comment
+<http://example/s> <http://example/p> "o"^^<http://example/dt> . # comment
+<http://example/s> <http://example/p> "o"@en . # comment

tests/JsonLD.Test/NQuads/langtagged_string.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "chat"@en .

tests/JsonLD.Test/NQuads/lantag_with_subtag.nq

@@ -0,0 +1 @@
+<http://example.org/ex#a> <http://example.org/ex#b> "Cheers"@en-UK .

tests/JsonLD.Test/NQuads/literal.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "x" .

tests/JsonLD.Test/NQuads/literal_all_controls.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "\u0000\u0001\u0002\u0003\u0004\u0005\u0006\u0007\u0008\t\u000B\u000C\u000E\u000F\u0010\u0011\u0012\u0013\u0014\u0015\u0016\u0017\u0018\u0019\u001A\u001B\u001C\u001D\u001E\u001F" .

tests/JsonLD.Test/NQuads/literal_all_punctuation.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> " !\"#$%&():;<=>?@[]^_`{|}~" .

tests/JsonLD.Test/NQuads/literal_false.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "false"^^<http://www.w3.org/2001/XMLSchema#boolean> .

tests/JsonLD.Test/NQuads/literal_true.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "true"^^<http://www.w3.org/2001/XMLSchema#boolean> .

tests/JsonLD.Test/NQuads/literal_with_2_dquotes.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "x\"\"y" .

tests/JsonLD.Test/NQuads/literal_with_2_squotes.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "x''y" .

tests/JsonLD.Test/NQuads/literal_with_BACKSPACE.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "\b" .

tests/JsonLD.Test/NQuads/literal_with_CARRIAGE_RETURN.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "\r" .

tests/JsonLD.Test/NQuads/literal_with_CHARACTER_TABULATION.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "\t" .

tests/JsonLD.Test/NQuads/literal_with_FORM_FEED.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "\f" .

tests/JsonLD.Test/NQuads/literal_with_LINE_FEED.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "\n" .

tests/JsonLD.Test/NQuads/literal_with_REVERSE_SOLIDUS.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "\\" .

tests/JsonLD.Test/NQuads/literal_with_REVERSE_SOLIDUS2.nq

@@ -0,0 +1 @@
+<http://example.org/ns#s> <http://example.org/ns#p1> "test-\\" .

tests/JsonLD.Test/NQuads/literal_with_UTF8_boundaries.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "€߿ࠀ࿿က쿿퀀퟿�𐀀𿿽񀀀󿿽􀀀􏿽" .

tests/JsonLD.Test/NQuads/literal_with_dquote.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "x\"y" .

tests/JsonLD.Test/NQuads/literal_with_numeric_escape4.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "\u006F" .

tests/JsonLD.Test/NQuads/literal_with_numeric_escape8.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "\U0000006F" .

tests/JsonLD.Test/NQuads/literal_with_squote.nq

@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> "x'y" .