Add tests for SPARQL syntax codepoint escapes and invalid codepoint use

sparql11/data-sparql11/syntax-query/1val1STRING_LITERAL1_with_UTF8_boundaries.rq

@@ -0,0 +1,3 @@
+SELECT * WHERE {
+	?s <http://a.example/p1> '€߿ࠀ࿿က쿿퀀퟿�𐀀𿿽񀀀󿿽􀀀􏿽'
+}

sparql11/data-sparql11/syntax-query/1val1STRING_LITERAL1_with_UTF8_boundaries_escaped.rq

@@ -0,0 +1,3 @@
+SELECT * WHERE {
+	?s <http://a.example/p1> '\u0080\u07FF\u0800\u0FFF\u1000\uCFFF\uD000\uD7FF\uE000\uFFFD\U00010000\U0003FFFD\U00040000\U000FFFFD\U00100000\U0010FFFD'
+}

sparql11/data-sparql11/syntax-query/manifest.ttl

@@ -1,6 +1,6 @@
 @prefix :       <http://www.w3.org/2009/sparql/docs/tests/data-sparql11/syntax-query/manifest#> .
 @prefix rdf:    <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
-@prefix rdfs:	<http://www.w3.org/2000/01/rdf-schema#> .
+@prefix rdfs:   <http://www.w3.org/2000/01/rdf-schema#> .
 @prefix mf:     <http://www.w3.org/2001/sw/DataAccess/tests/test-manifest#> .
 @prefix mfx:    <http://jena.hpl.hp.com/2005/05/test-manifest-extra#> .
 @prefix qt:     <http://www.w3.org/2001/sw/DataAccess/tests/test-query#> .
@@ -103,6 +103,15 @@
 :test_pn_bad_13
 
 :test_pp_coll
+
+:test_codepoint_escape_01
+:test_codepoint_escape_02
+:test_codepoint_escape_03
+:test_codepoint_escape_bad_04
+:test_codepoint_escape_bad_05
+:test_codepoint_boundaries_01
+:test_codepoint_boundaries_01_escaped
+:test_invalid_codepoint_escaped_bad_01
 ) .
 
 :test_1 rdf:type   mf:PositiveSyntaxTest11 ;
@@ -622,3 +631,46 @@
    dawgt:approvedBy <http://www.w3.org/2009/sparql/meeting/2012-08-07#resolution_2> ;
    mf:name    "syn-pp-in-collection" ;
    mf:action  <syn-pp-in-collection.rq> ;.
+
+:test_codepoint_escape_01 rdf:type   mf:PositiveSyntaxTest11 ;
+   dawgt:approval dawgt:Proposed ;
+   mf:name    "\\U unicode codepoint escaping in literal" ;
+   mf:action  <syn-codepoint-escape-01.rq> ;.
+
+:test_codepoint_escape_02 rdf:type   mf:PositiveSyntaxTest11 ;
+   dawgt:approval dawgt:Proposed ;
+   mf:name    "\\U unicode codepoint escaping in query syntax" ;
+   mf:action  <syn-codepoint-escape-02.rq> ;.
+
+:test_codepoint_escape_03 rdf:type   mf:PositiveSyntaxTest11 ;
+   dawgt:approval dawgt:Proposed ;
+   mf:name    "\\U and \\u unicode codepoint escaping in variable name" ;
+   mf:action  <syn-codepoint-escape-03.rq> ;.
+
+:test_codepoint_escape_bad_04 rdf:type   mf:NegativeSyntaxTest11 ;
+   dawgt:approval dawgt:Proposed ;
+   mf:name    "Invalid multi-pass codepoint escaping (\\u then \\U)" ;
+   mf:description "Unescaping one escape sequence must not produce content that is used in another escape sequence" ;
+   mf:action  <syn-codepoint-escape-bad-04.rq> ;.
+
+:test_codepoint_escape_bad_05 rdf:type   mf:NegativeSyntaxTest11 ;
+   dawgt:approval dawgt:Proposed ;
+   mf:name    "Invalid multi-pass codepoint escaping (\\U then \\u)" ;
+   mf:description "Unescaping one escape sequence must not produce content that is used in another escape sequence" ;
+   mf:action  <syn-codepoint-escape-bad-05.rq> ;.
+
+:test_codepoint_boundaries_01 rdf:type mf:PositiveSyntaxTest11 ;
+   dawgt:approval dawgt:Proposed ;
+   mf:name    "utf8 literal using codepoints at notable unicode boundaries" ;
+   mf:action  <1val1STRING_LITERAL1_with_UTF8_boundaries.rq> ;.
+
+:test_codepoint_boundaries_01_escaped rdf:type mf:PositiveSyntaxTest11 ;
+   dawgt:approval dawgt:Proposed ;
+   mf:name    "\\U and \\u unicode codepoint escaping in literal using codepoints at notable unicode boundaries" ;
+   mf:action  <1val1STRING_LITERAL1_with_UTF8_boundaries_escaped.rq> ;.
+
+:test_invalid_codepoint_escaped_bad_01 rdf:type mf:NegativeSyntaxTest11 ;
+   dawgt:approval dawgt:Proposed ;
+   mf:name    "\\u unicode codepoint escaping in literal using partial surrogate pair" ;
+   mf:description "Using a codepoint that is half of a surrogate pair (in the range U+D800–U+DFFF) without a corresponding codepoint is illegal" ;
+   mf:action  <syn-invalid-codepoint-escaped-bad-01.rq> ;.

sparql11/data-sparql11/syntax-query/syn-codepoint-escape-01.rq

@@ -0,0 +1,3 @@
+SELECT * WHERE {
+    ?s ?p "\U0001f46a" .
+}

sparql11/data-sparql11/syntax-query/syn-codepoint-escape-02.rq

@@ -0,0 +1,3 @@
+SELECT \U0000002A WHERE \U0000007B # SELECT * WHERE {
+    ?s ?p true .
+}

sparql11/data-sparql11/syntax-query/syn-codepoint-escape-03.rq

@@ -0,0 +1,3 @@
+SELECT * WHERE {
+    ?s \u0024\U00000070 true . # ?s $p true
+}

sparql11/data-sparql11/syntax-query/syn-codepoint-escape-bad-04.rq

@@ -0,0 +1,7 @@
+SELECT * WHERE {
+    ?s ?p \u005cU00000031 .
+}
+
+# the first escape here produces '\' REVERSE SOLIDUS (U+5C)
+# its unescaping must not lead to a subsequent unescaping of "\U00000031" escape.
+# This should cause a syntax error as "?s ?p \U00000031" and NOT a valid query as "?s ?p 1"

sparql11/data-sparql11/syntax-query/syn-codepoint-escape-bad-05.rq

@@ -0,0 +1,7 @@
+SELECT * WHERE {
+    ?s ?p \U0000005cu0031 .
+}
+
+# the first escape here produces '\' REVERSE SOLIDUS (U+5C)
+# its unescaping must not lead to a subsequent unescaping of "\u0031" escape.
+# This should cause a syntax error as "?s ?p \u0031" and NOT a valid query as "?s ?p 1"

sparql11/data-sparql11/syntax-query/syn-invalid-codepoint-escaped-bad-01.rq

@@ -0,0 +1,3 @@
+SELECT * WHERE {
+	?s <http://a.example/p1> '\uD800'
+}