+ ShEx 2.1: new ShExC parser, ShExJ-shaped AST, spec validator

Replace the ShEx1-era .ssx parser and RulePattern AST with a ShEx 2.1
implementation that passes all 1729 applicable shexSpec/shexTest cases
(schemas, negativeSyntax, negativeStructure and validation manifests,
including the EXTENDS/ABSTRACT hierarchy tests).

Parser: lib/ShExCParser.ypp ports the productions of shex.js's
ShExJison.jison (the reference ShExC grammar) to this repo's bison
conventions; the scanner gains REGEXP, REPEAT_RANGE, CODE, ATPNAME and
LANG_STRING tokens plus /* */ comments. Conflict-free LALR.

AST + validation: lib/ShExSchema.{hpp,cpp} mirror ShExJ (shape
expressions, triple expressions, node constraint components, value
sets with stems/ranges/exclusions). The validator follows jena-shex:
predicate-based pre-matching, eager recursive valueExpr filtering,
iteration over matchings accepted by interval computation on the SORBE
form of each triple expression, CLOSED/EXTRA handling, cyclic-schema
assumptions, and EXTENDS splits with supertype constraint checks.
Structure checks enforce defined references, triple-constraint-crossing
cycles, no negation (or EXTRA'd constraint) in reference cycles, and
shape/tripleExpr label separation.

Semantic actions: the http://shex.io/extensions/Test/ language is
interpreted (its fail(...) aborts); other languages are no-ops.

tests/test_ShEx.cpp walks the shexTest manifests (fetched by
tests/fetch-test-suites.sh) and registers each entry as a test case,
mapping the suite's web URLs to the local checkout and resolving
focus/shape blank nodes by source label. test_ShEx moves into CI's
blocking bucket.

AtomFactory now preserves the lexical form of xsd:boolean literals and
interns them by it ("TRUE"^^xsd:boolean is not "true"^^xsd:boolean),
which the ShEx ValidLexicalForm tests require.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Author: Eric Prud'hommeaux

.github/workflows/ci.yml

@@ -41,17 +41,17 @@ jobs:
       - name: Run test suites
         run: >
           ctest --test-dir build --output-on-failure --timeout 600
-          -E 'test_(SPARQL|SPARQL11|QueryMap|ShEx)$'
+          -E 'test_(SPARQL|SPARQL11|QueryMap)$'
 
       # These suites carry known pre-existing failures:
       #  - test_SPARQL11: md5/sha* builtins are unimplemented (8 cases)
       #  - test_SPARQL: function_library expects CSV <foo> parsed as an IRI,
       #    a csv-regex branch deliberately disabled upstream (~2013)
-      #  - test_QueryMap, test_ShEx: heap-layout-dependent nondeterminism in
-      #    the query rewriter / ShEx normalizer makes single cases flaky
+      #  - test_QueryMap: heap-layout-dependent nondeterminism in the query
+      #    rewriter makes single cases flaky
       # Run them for visibility without gating the build.
       - name: Run suites with known pre-existing failures (non-blocking)
         continue-on-error: true
         run: >
           ctest --test-dir build --output-on-failure --timeout 600
-          -R 'test_(SPARQL|SPARQL11|QueryMap|ShEx)$'
+          -R 'test_(SPARQL|SPARQL11|QueryMap)$'

.gitignore

@@ -33,3 +33,4 @@ tests/RDFXMLTests/
 tests/sparql11-test-suite/
 bin/parseAndFree
 bin/dm-materialize
+tests/shexTest/

README

@@ -38,6 +38,15 @@ you'll need these packages:
 On macOS: brew install cmake boost bison flex expat
 (cmake prefers Homebrew's bison/flex over the older /usr/bin ones.)
 
+# ShEx
+
+lib/ShExCParser.ypp and lib/ShExSchema.{hpp,cpp} implement ShEx 2.1
+(ShExC syntax including IMPORT, ABSTRACT and EXTENDS; validation by
+interval-based matching over SORBE-normalized triple expressions, after
+jena-shex). tests/test_ShEx.cpp runs the shexSpec/shexTest suite
+(schemas, negativeSyntax, negativeStructure and validation manifests);
+tests/fetch-test-suites.sh clones the test data.
+
 # More info
 
 To learn more about SWObjects, checkout some tutorials given at SWAT4LS in 2010 and 2011:

lib/SWObjects.cpp

@@ -2240,18 +2240,9 @@ void RecursiveExpressor::valuesClause (const ValuesClause* const, const ResultSe
 	} else if (p_URI == TTerm::URI_xsd_dateTime) {
 	    return getDateTimeRDFLiteral(p_String.c_str());
 	} else if (p_URI == TTerm::URI_xsd_boolean) {
-	    if (p_String == "0" || p_String == "false")
-		return getBooleanRDFLiteral("false", 0);
-	    bool b;
-#ifdef LIBC_PARSES_BOOL
-	    is >> b;
-#else
-	    b = true;
-#endif
-// 	    std::stringstream canonical;
-// 	    canonical << std::boolalpha << b;
-// 	    return getNumericRDFLiteral(canonical.str().c_str(), b);
-	    return getBooleanRDFLiteral(p_String.c_str(), b);
+	    // Preserve the lexical form; the value is true for the two true
+	    // lexical representations.
+	    return getBooleanRDFLiteral(p_String, p_String == "1" || p_String == "true");
 	}
 
 	std::stringstream buf;
@@ -2496,7 +2487,9 @@ void RecursiveExpressor::valuesClause (const ValuesClause* const, const ResultSe
 
     const BooleanRDFLiteral* AtomFactory::getBooleanRDFLiteral (std::string p_String, bool p_value) {
 	std::stringstream buf;
-	buf << "\"" << (p_value ? "true" : "false") << "\"^^<http://www.w3.org/2001/XMLSchema#boolean>";
+	// key on the lexical form: "TRUE"^^xsd:boolean and "true"^^xsd:boolean
+	// are distinct RDF terms
+	buf << "\"" << p_String << "\"^^<http://www.w3.org/2001/XMLSchema#boolean>";
 	std::string key(buf.str());
 	RDFLiteralMap::const_iterator vi = rdfLiterals_static.find(key);
 	if (vi == rdfLiterals_static.end()) {