Add support for converting literals to canonical form

11 files changed, 239 insertions, 1 deletions

diff --git a/test/canon/bad-boolean.ttl b/test/canon/bad-boolean.ttl
new file mode 100644
index 00000000..c4fc3eb5
--- /dev/null
+++ b/test/canon/bad-boolean.ttl
@@ -0,0 +1,5 @@
+@base <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+[] <boolean> "      ja "^^xsd:boolean .
+
diff --git a/test/canon/bad-decimal-leading.ttl b/test/canon/bad-decimal-leading.ttl
new file mode 100644
index 00000000..0d18eac7
--- /dev/null
+++ b/test/canon/bad-decimal-leading.ttl
@@ -0,0 +1,4 @@
+@base <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+[] <decimal> " junk 1234.5678 "^^xsd:decimal .
diff --git a/test/canon/bad-decimal-trailing.ttl b/test/canon/bad-decimal-trailing.ttl
new file mode 100644
index 00000000..10882ef5
--- /dev/null
+++ b/test/canon/bad-decimal-trailing.ttl
@@ -0,0 +1,4 @@
+@base <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+[] <decimal> " 1234.5678 junk "^^xsd:decimal .
diff --git a/test/canon/bad-empty-boolean.ttl b/test/canon/bad-empty-boolean.ttl
new file mode 100644
index 00000000..9a390c46
--- /dev/null
+++ b/test/canon/bad-empty-boolean.ttl
@@ -0,0 +1,5 @@
+@base <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+[] <boolean> ""^^xsd:boolean .
+
diff --git a/test/canon/bad-integer-leading.ttl b/test/canon/bad-integer-leading.ttl
new file mode 100644
index 00000000..80c1a6af
--- /dev/null
+++ b/test/canon/bad-integer-leading.ttl
@@ -0,0 +1,4 @@
+@base <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+[] <integer> " junk 987654321 "^^xsd:integer .
diff --git a/test/canon/bad-integer-trailing.ttl b/test/canon/bad-integer-trailing.ttl
new file mode 100644
index 00000000..a94a9ec4
--- /dev/null
+++ b/test/canon/bad-integer-trailing.ttl
@@ -0,0 +1,4 @@
+@base <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+[] <integer> " 987654321 junk "^^xsd:integer .
diff --git a/test/canon/manifest.ttl b/test/canon/manifest.ttl
new file mode 100644
index 00000000..e2195212
--- /dev/null
+++ b/test/canon/manifest.ttl
@@ -0,0 +1,58 @@
+@prefix mf: <http://www.w3.org/2001/sw/DataAccess/tests/test-manifest#> .
+@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
+@prefix rdft: <http://www.w3.org/ns/rdftest#> .
+
+<>
+	a mf:Manifest ;
+	rdfs:comment "Serd canonical literal test cases" ;
+	mf:entries (
+		<#bad-boolean>
+		<#bad-decimal-leading>
+		<#bad-decimal-trailing>
+		<#bad-empty-boolean>
+		<#bad-integer-leading>
+		<#bad-integer-trailing>
+		<#test-canon>
+	) .
+
+<#bad-boolean>
+	a rdft:TestTurtleNegativeEval ;
+	mf:name "bad-boolean" ;
+	rdfs:comment "Invalid xsd::boolean syntax" ;
+	mf:action <bad-boolean.ttl> .
+
+<#bad-decimal-leading>
+	a rdft:TestTurtleNegativeEval ;
+	mf:name "bad-decimal-leading" ;
+	rdfs:comment "Invalid xsd::decimal syntax (leading garbage)" ;
+	mf:action <bad-decimal-leading.ttl> .
+
+<#bad-decimal-trailing>
+	a rdft:TestTurtleNegativeEval ;
+	mf:name "bad-decimal-trailing" ;
+	rdfs:comment "Invalid xsd::decimal syntax (trailing garbage)" ;
+	mf:action <bad-decimal-trailing.ttl> .
+
+<#bad-empty-boolean>
+	a rdft:TestTurtleNegativeEval ;
+	mf:name "bad-empty-boolean" ;
+	rdfs:comment "Invalid xsd::boolean syntax (no value)" ;
+	mf:action <bad-empty-boolean.ttl> .
+
+<#bad-integer-leading>
+	a rdft:TestTurtleNegativeEval ;
+	mf:name "bad-integer-leading" ;
+	rdfs:comment "Invalid xsd::integer syntax (leading garbage)" ;
+	mf:action <bad-integer-leading.ttl> .
+
+<#bad-integer-trailing>
+	a rdft:TestTurtleNegativeEval ;
+	mf:name "bad-integer-trailing" ;
+	rdfs:comment "Invalid xsd::integer syntax (trailing garbage)" ;
+	mf:action <bad-integer-trailing.ttl> .
+
+<#test-canon>
+	a rdft:TestTurtleEval ;
+	mf:name "test-canon" ;
+	mf:action <test-canon.ttl> ;
+	mf:result <test-canon.nt> .
diff --git a/test/canon/test-canon.nt b/test/canon/test-canon.nt
new file mode 100644
index 00000000..ff492890
--- /dev/null
+++ b/test/canon/test-canon.nt
@@ -0,0 +1,70 @@
+_:b1 <http://example.org/boolean> "false"^^<http://www.w3.org/2001/XMLSchema#boolean> .
+_:b1 <http://example.org/boolean> "false"^^<http://www.w3.org/2001/XMLSchema#boolean> .
+_:b1 <http://example.org/boolean> "true"^^<http://www.w3.org/2001/XMLSchema#boolean> .
+_:b1 <http://example.org/boolean> "true"^^<http://www.w3.org/2001/XMLSchema#boolean> .
+_:b1 <http://example.org/ieee754> "1.0E2"^^<http://www.w3.org/2001/XMLSchema#float> .
+_:b1 <http://example.org/ieee754> "-1.0E2"^^<http://www.w3.org/2001/XMLSchema#float> .
+_:b1 <http://example.org/ieee754> "1.0E3"^^<http://www.w3.org/2001/XMLSchema#double> .
+_:b1 <http://example.org/ieee754> "-1.0E3"^^<http://www.w3.org/2001/XMLSchema#double> .
+_:b1 <http://example.org/machine> "9223372036854775807"^^<http://www.w3.org/2001/XMLSchema#long> .
+_:b1 <http://example.org/machine> "-9223372036854775808"^^<http://www.w3.org/2001/XMLSchema#long> .
+_:b1 <http://example.org/machine> "2147483647"^^<http://www.w3.org/2001/XMLSchema#int> .
+_:b1 <http://example.org/machine> "-2147483648"^^<http://www.w3.org/2001/XMLSchema#int> .
+_:b1 <http://example.org/machine> "32767"^^<http://www.w3.org/2001/XMLSchema#short> .
+_:b1 <http://example.org/machine> "-32768"^^<http://www.w3.org/2001/XMLSchema#short> .
+_:b1 <http://example.org/machine> "127"^^<http://www.w3.org/2001/XMLSchema#byte> .
+_:b1 <http://example.org/machine> "-128"^^<http://www.w3.org/2001/XMLSchema#byte> .
+_:b1 <http://example.org/machine> "1"^^<http://www.w3.org/2001/XMLSchema#unsignedLong> .
+_:b1 <http://example.org/machine> "18446744073709551615"^^<http://www.w3.org/2001/XMLSchema#unsignedLong> .
+_:b1 <http://example.org/machine> "1"^^<http://www.w3.org/2001/XMLSchema#unsignedInt> .
+_:b1 <http://example.org/machine> "4294967295"^^<http://www.w3.org/2001/XMLSchema#unsignedInt> .
+_:b1 <http://example.org/machine> "1"^^<http://www.w3.org/2001/XMLSchema#unsignedShort> .
+_:b1 <http://example.org/machine> "65535"^^<http://www.w3.org/2001/XMLSchema#unsignedShort> .
+_:b1 <http://example.org/machine> "1"^^<http://www.w3.org/2001/XMLSchema#unsignedByte> .
+_:b1 <http://example.org/machine> "255"^^<http://www.w3.org/2001/XMLSchema#unsignedByte> .
+_:b1 <http://example.org/decimal> "0.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "0.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "-0.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "36893488147419103232.123"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "-36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "-36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "-36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "-36893488147419103232.0"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "-36893488147419103232.123"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "0.123"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "0.123"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "0.123"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "0.123"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "-0.123"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/decimal> "-0.123"^^<http://www.w3.org/2001/XMLSchema#decimal> .
+_:b1 <http://example.org/integer> "36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#integer> .
+_:b1 <http://example.org/integer> "36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#integer> .
+_:b1 <http://example.org/integer> "36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#integer> .
+_:b1 <http://example.org/integer> "36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#integer> .
+_:b1 <http://example.org/integer> "-36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#integer> .
+_:b1 <http://example.org/integer> "-36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#integer> .
+_:b1 <http://example.org/integer> "0"^^<http://www.w3.org/2001/XMLSchema#nonPositiveInteger> .
+_:b1 <http://example.org/integer> "-36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#nonPositiveInteger> .
+_:b1 <http://example.org/integer> "-1"^^<http://www.w3.org/2001/XMLSchema#negativeInteger> .
+_:b1 <http://example.org/integer> "-36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#negativeInteger> .
+_:b1 <http://example.org/integer> "0"^^<http://www.w3.org/2001/XMLSchema#nonNegativeInteger> .
+_:b1 <http://example.org/integer> "36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#nonNegativeInteger> .
+_:b1 <http://example.org/integer> "1"^^<http://www.w3.org/2001/XMLSchema#positiveInteger> .
+_:b1 <http://example.org/integer> "36893488147419103232"^^<http://www.w3.org/2001/XMLSchema#positiveInteger> .
+_:b1 <http://example.org/langString> "no language tag" .
+_:b1 <http://example.org/taggedString> "english"@en-ca .
+_:b1 <http://example.org/time> "P1Y6M"^^<http://www.w3.org/2001/XMLSchema#duration> .
+_:b1 <http://example.org/time> "12:15:01Z"^^<http://www.w3.org/2001/XMLSchema#time> .
+_:b1 <http://example.org/time> "2004-04-12Z"^^<http://www.w3.org/2001/XMLSchema#date> .
+_:b1 <http://example.org/binary> "A1B7F080"^^<http://www.w3.org/2001/XMLSchema#hexBinary> .
+_:b1 <http://example.org/binary> "Zm9vYmF="^^<http://www.w3.org/2001/XMLSchema#base64Binary> .
+_:b1 <http://example.org/other> "untyped" .
+_:b1 <http://example.org/other> <http://example.org/uri> .
+_:b1 <http://example.org/other> "notxsd"^^<http://example.org/sometype> .
+_:b1 <http://example.org/other> "unsupported"^^<http://www.w3.org/2001/XMLSchema#name> .
diff --git a/test/canon/test-canon.ttl b/test/canon/test-canon.ttl
new file mode 100644
index 00000000..0d0b4682
--- /dev/null
+++ b/test/canon/test-canon.ttl
@@ -0,0 +1,76 @@
+@base <http://example.org/> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+[
+	<boolean> " false "^^xsd:boolean ,
+		" 0 "^^xsd:boolean ,
+		" true "^^xsd:boolean ,
+		" 1 "^^xsd:boolean ;
+	<ieee754> " +0100.0 "^^xsd:float ,
+		" -0100.0 "^^xsd:float ,
+		" +01000.0 "^^xsd:double ,
+		" -01000.0 "^^xsd:double ;
+	<machine> " +09223372036854775807 "^^xsd:long ,
+		" -09223372036854775808 "^^xsd:long ,
+		" +02147483647 "^^xsd:int ,
+		" -02147483648 "^^xsd:int ,
+		" +032767 "^^xsd:short ,
+		" -032768 "^^xsd:short ,
+		" +0127 "^^xsd:byte ,
+		" -0128 "^^xsd:byte ,
+		" 01 "^^xsd:unsignedLong ,
+		" 018446744073709551615 "^^xsd:unsignedLong ,
+		" 01 "^^xsd:unsignedInt ,
+		" 04294967295 "^^xsd:unsignedInt ,
+		" 01 "^^xsd:unsignedShort ,
+		" 065535 "^^xsd:unsignedShort ,
+		" 01 "^^xsd:unsignedByte ,
+		" 0255 "^^xsd:unsignedByte ;
+	<decimal> " 00 "^^xsd:decimal ,
+		" +0 "^^xsd:decimal ,
+		" -0 "^^xsd:decimal ,
+		" 36893488147419103232 "^^xsd:decimal ,
+		" 0036893488147419103232 "^^xsd:decimal ,
+		" +36893488147419103232 "^^xsd:decimal ,
+		" +0036893488147419103232 "^^xsd:decimal ,
+		" +0036893488147419103232. "^^xsd:decimal ,
+		" +0036893488147419103232.00 "^^xsd:decimal ,
+		" +0036893488147419103232.12300 "^^xsd:decimal ,
+		" -36893488147419103232 "^^xsd:decimal ,
+		" -0036893488147419103232 "^^xsd:decimal ,
+		" -0036893488147419103232. "^^xsd:decimal ,
+		" -0036893488147419103232.00 "^^xsd:decimal ,
+		" -0036893488147419103232.12300 "^^xsd:decimal ,
+		" 00.12300 "^^xsd:decimal ,
+		" .12300 "^^xsd:decimal ,
+		" +.12300 "^^xsd:decimal ,
+		" +00.12300 "^^xsd:decimal ,
+		" -.12300 "^^xsd:decimal ,
+		" -00.12300 "^^xsd:decimal ;
+	<integer> " 36893488147419103232 "^^xsd:integer ,
+		" 0036893488147419103232 "^^xsd:integer ,
+		" +36893488147419103232 "^^xsd:integer ,
+		" +0036893488147419103232 "^^xsd:integer ,
+		" -36893488147419103232 "^^xsd:integer ,
+		" -0036893488147419103232 "^^xsd:integer ,
+		" 00 "^^xsd:nonPositiveInteger ,
+		" -036893488147419103232 "^^xsd:nonPositiveInteger ,
+		" -01 "^^xsd:negativeInteger ,
+		" -036893488147419103232 "^^xsd:negativeInteger ,
+		" 00 "^^xsd:nonNegativeInteger ,
+		" 036893488147419103232 "^^xsd:nonNegativeInteger ,
+		" +01 "^^xsd:positiveInteger ,
+		" 036893488147419103232 "^^xsd:positiveInteger ;
+	<langString> "no language tag"^^rdf:langString ;
+	<taggedString> "english"@EN-CA ;
+	<time> " P1Y6M0D "^^xsd:duration ,
+		" 12:15:01+00:00 "^^xsd:time ,
+		" 02004-04-12+00:00 "^^xsd:date ;
+	<binary> "A 1 B7 F080"^^xsd:hexBinary ,
+		" Zm 9v  Y m F="^^xsd:base64Binary ;
+	<other> "untyped" ,
+		<uri> ,
+		"notxsd"^^<sometype> ,
+		"unsupported"^^xsd:name
+] .
diff --git a/test/meson.build b/test/meson.build
index c4dc4b3b..e8392559 100644
--- a/test/meson.build
+++ b/test/meson.build
@@ -249,6 +249,14 @@ if get_option('utils')
        suite: ['rdf', 'serd'],
        timeout: 240)
 
+  manifest = files('canon' / 'manifest.ttl')
+  base_uri = serd_base + 'canon' + '/'
+  test('canon', run_test_suite,
+       args: script_args + [manifest, base_uri, '--', '-C'],
+       env: test_env,
+       suite: ['rdf', 'serd'],
+       timeout: 240)
+
   ### The lax suite is special because it is run twice...
   lax_manifest = files('lax/manifest.ttl')
   lax_base_uri = serd_base + name + '/'
diff --git a/test/test_string.c b/test/test_string.c
index 2bd47680..f8534001 100644
--- a/test/test_string.c
+++ b/test/test_string.c
@@ -32,7 +32,7 @@ test_strerror(void)
 {
   const char* msg = serd_strerror(SERD_SUCCESS);
   assert(!strcmp(msg, "Success"));
-  for (int i = SERD_FAILURE; i <= SERD_ERR_BAD_INDEX; ++i) {
+  for (int i = SERD_FAILURE; i <= SERD_ERR_INVALID; ++i) {
     msg = serd_strerror((SerdStatus)i);
     assert(strcmp(msg, "Success"));
   }