Add support for parsing variables

This adds a reader flag and serdi option for extending a syntax with support
for SPARQL-like variables, for storing things like patterns or simple queries.

18 files changed, 150 insertions, 12 deletions

diff --git a/test/bad/bad-var.ttl b/test/bad/bad-var.ttl
new file mode 100644
index 00000000..29b5b008
--- /dev/null
+++ b/test/bad/bad-var.ttl
@@ -0,0 +1,2 @@
+?s
+	<http://example.org/p> <http://example.org/o> .
diff --git a/test/bad/manifest.ttl b/test/bad/manifest.ttl
index 96a0b593..466a4435 100644
--- a/test/bad/manifest.ttl
+++ b/test/bad/manifest.ttl
@@ -70,6 +70,7 @@
 		<#bad-string>
 		<#bad-subject>
 		<#bad-uri-escape>
+		<#bad-var>
 		<#bad-verb>
 		<#invalid-char-in-local>
 		<#invalid-char-in-prefix>
@@ -394,6 +395,11 @@
 	mf:name "bad-uri-escape" ;
 	mf:action <bad-uri-escape.ttl> .
 
+<#bad-var>
+	rdf:type rdft:TestTurtleNegativeSyntax ;
+	mf:name "bad-var" ;
+	mf:action <bad-var.ttl> .
+
 <#bad-verb>
 	rdf:type rdft:TestTurtleNegativeSyntax ;
 	mf:name "bad-verb" ;
diff --git a/test/meson.build b/test/meson.build
index 5063277b..222ec1bc 100644
--- a/test/meson.build
+++ b/test/meson.build
@@ -219,6 +219,14 @@ if get_option('utils')
        suite: ['rdf', 'serd'],
        timeout: 240)
 
+  manifest = files('pattern' / 'manifest.ttl')
+  base_uri = serd_base + 'pattern' + '/'
+  test('pattern', run_test_suite,
+       args: script_args + [manifest, base_uri, '--', '-x'],
+       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/pattern/bad-pattern-graph.nq b/test/pattern/bad-pattern-graph.nq
new file mode 100644
index 00000000..9c77ac9a
--- /dev/null
+++ b/test/pattern/bad-pattern-graph.nq
@@ -0,0 +1 @@
+<http://example.org/s> <http://example.org/p> <http://example.org/o> ?gµ .
diff --git a/test/pattern/bad-pattern-object.nq b/test/pattern/bad-pattern-object.nq
new file mode 100644
index 00000000..7f22b520
--- /dev/null
+++ b/test/pattern/bad-pattern-object.nq
@@ -0,0 +1 @@
+<http://example.org/s> <http://example.org/p> ?oµ <http://example.org/g> .
diff --git a/test/pattern/bad-pattern-object.nt b/test/pattern/bad-pattern-object.nt
new file mode 100644
index 00000000..aef057e4
--- /dev/null
+++ b/test/pattern/bad-pattern-object.nt
@@ -0,0 +1 @@
+<http://example.org/s> <http://example.org/p> ?oµ .
diff --git a/test/pattern/bad-pattern-object.ttl b/test/pattern/bad-pattern-object.ttl
new file mode 100644
index 00000000..cc2eee77
--- /dev/null
+++ b/test/pattern/bad-pattern-object.ttl
@@ -0,0 +1,2 @@
+<http://example.org/s>
+	<http://example.org/o> ?oµ .
diff --git a/test/pattern/bad-pattern-predicate.nq b/test/pattern/bad-pattern-predicate.nq
new file mode 100644
index 00000000..cfe466c9
--- /dev/null
+++ b/test/pattern/bad-pattern-predicate.nq
@@ -0,0 +1 @@
+<http://example.org/s> ?pµ <http://example.org/o> <http://example.org/g> .
diff --git a/test/pattern/bad-pattern-predicate.nt b/test/pattern/bad-pattern-predicate.nt
new file mode 100644
index 00000000..a7e0c994
--- /dev/null
+++ b/test/pattern/bad-pattern-predicate.nt
@@ -0,0 +1 @@
+<http://example.org/s> ?pµ <http://example.org/o> .
diff --git a/test/pattern/bad-pattern-predicate.ttl b/test/pattern/bad-pattern-predicate.ttl
new file mode 100644
index 00000000..cc2bd6b1
--- /dev/null
+++ b/test/pattern/bad-pattern-predicate.ttl
@@ -0,0 +1,2 @@
+<http://example.org/s>
+	?pµ <http://example.org/o> .
diff --git a/test/pattern/bad-pattern-subject.nq b/test/pattern/bad-pattern-subject.nq
new file mode 100644
index 00000000..341c437f
--- /dev/null
+++ b/test/pattern/bad-pattern-subject.nq
@@ -0,0 +1 @@
+?sµ <http://example.org/p> <http://example.org/o> <http://example.org/g> .
diff --git a/test/pattern/bad-pattern-subject.nt b/test/pattern/bad-pattern-subject.nt
new file mode 100644
index 00000000..5bbb29ff
--- /dev/null
+++ b/test/pattern/bad-pattern-subject.nt
@@ -0,0 +1 @@
+?sµ <http://example.org/p> <http://example.org/o> .
diff --git a/test/pattern/bad-pattern-subject.ttl b/test/pattern/bad-pattern-subject.ttl
new file mode 100644
index 00000000..5f3dbfdd
--- /dev/null
+++ b/test/pattern/bad-pattern-subject.ttl
@@ -0,0 +1,2 @@
+?sµ
+	<http://example.org/p> <http://example.org/o> .
diff --git a/test/pattern/manifest.ttl b/test/pattern/manifest.ttl
new file mode 100644
index 00000000..a7f238bc
--- /dev/null
+++ b/test/pattern/manifest.ttl
@@ -0,0 +1,89 @@
+@prefix mf: <http://www.w3.org/2001/sw/DataAccess/tests/test-manifest#> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
+@prefix rdft: <http://www.w3.org/ns/rdftest#> .
+
+<>
+	rdf:type mf:Manifest ;
+	rdfs:comment "Serd pattern syntax test cases" ;
+	mf:entries (
+		<#bad-pattern-graph-nq>
+		<#bad-pattern-object-nq>
+		<#bad-pattern-object-nt>
+		<#bad-pattern-object-ttl>
+		<#bad-pattern-predicate-nq>
+		<#bad-pattern-predicate-nt>
+		<#bad-pattern-predicate-ttl>
+		<#bad-pattern-subject-nq>
+		<#bad-pattern-subject-nt>
+		<#bad-pattern-subject-ttl>
+		<#test-graph-pattern>
+		<#test-pattern-nt>
+		<#test-pattern-ttl>
+	) .
+
+<#bad-pattern-graph-nq>
+	rdf:type rdft:TestNQuadsNegativeSyntax ;
+	mf:name "bad-pattern-graph-nq" ;
+	mf:action <bad-pattern-graph.nq> .
+
+<#bad-pattern-object-nq>
+	rdf:type rdft:TestNQuadsNegativeSyntax ;
+	mf:name "bad-pattern-object-nq" ;
+	mf:action <bad-pattern-object.nq> .
+
+<#bad-pattern-object-nt>
+	rdf:type rdft:TestNTriplesNegativeSyntax ;
+	mf:name "bad-pattern-object-nt" ;
+	mf:action <bad-pattern-object.nt> .
+
+<#bad-pattern-object-ttl>
+	rdf:type rdft:TestTurtleNegativeSyntax ;
+	mf:name "bad-pattern-object-ttl" ;
+	mf:action <bad-pattern-object.ttl> .
+
+<#bad-pattern-predicate-nq>
+	rdf:type rdft:TestNQuadsNegativeSyntax ;
+	mf:name "bad-pattern-predicate-nq" ;
+	mf:action <bad-pattern-predicate.nq> .
+
+<#bad-pattern-predicate-nt>
+	rdf:type rdft:TestNTriplesNegativeSyntax ;
+	mf:name "bad-pattern-predicate-nt" ;
+	mf:action <bad-pattern-predicate.nt> .
+
+<#bad-pattern-predicate-ttl>
+	rdf:type rdft:TestTurtleNegativeSyntax ;
+	mf:name "bad-pattern-predicate-ttl" ;
+	mf:action <bad-pattern-predicate.ttl> .
+
+<#bad-pattern-subject-nq>
+	rdf:type rdft:TestNQuadsNegativeSyntax ;
+	mf:name "bad-pattern-subject-nq" ;
+	mf:action <bad-pattern-subject.nq> .
+
+<#bad-pattern-subject-nt>
+	rdf:type rdft:TestNTriplesNegativeSyntax ;
+	mf:name "bad-pattern-subject-nt" ;
+	mf:action <bad-pattern-subject.nt> .
+
+<#bad-pattern-subject-ttl>
+	rdf:type rdft:TestTurtleNegativeSyntax ;
+	mf:name "bad-pattern-subject-ttl" ;
+	mf:action <bad-pattern-subject.ttl> .
+
+<#test-graph-pattern>
+	rdf:type rdft:TestNQuadsPositiveSyntax ;
+	mf:name "test-graph-pattern" ;
+	mf:action <test-graph-pattern.nq> .
+
+<#test-pattern-nt>
+	rdf:type rdft:TestNTriplesPositiveSyntax ;
+	mf:name "test-pattern-nt" ;
+	mf:action <test-pattern.nt> .
+
+<#test-pattern-ttl>
+	rdf:type rdft:TestTurtleEval ;
+	mf:name "test-pattern" ;
+	mf:action <test-pattern.ttl> ;
+	mf:result <test-pattern.nt> .
diff --git a/test/pattern/test-graph-pattern.nq b/test/pattern/test-graph-pattern.nq
new file mode 100644
index 00000000..453e9516
--- /dev/null
+++ b/test/pattern/test-graph-pattern.nq
@@ -0,0 +1 @@
+<http://a.example/s> <http://a.example/p> <http://a.example/o> ?g .
diff --git a/test/pattern/test-pattern.nt b/test/pattern/test-pattern.nt
new file mode 100644
index 00000000..ddfe6d3c
--- /dev/null
+++ b/test/pattern/test-pattern.nt
@@ -0,0 +1,6 @@
+?s <http://example.org/p1> <http://example.org/o1> .
+<http://example.org/s> ?p <http://example.org/o1> .
+<http://example.org/s> <http://example.org/p1> ?o .
+<http://example.org/s> <http://example.org/p2> _:b1 .
+_:b1 ?2p <http://example.org/o2> .
+_:b1 <http://example.org/p3> ?_o .
diff --git a/test/pattern/test-pattern.ttl b/test/pattern/test-pattern.ttl
new file mode 100644
index 00000000..3742e5ed
--- /dev/null
+++ b/test/pattern/test-pattern.ttl
@@ -0,0 +1,10 @@
+?s
+	<http://example.org/p1> <http://example.org/o1> .
+
+<http://example.org/s>
+	?p <http://example.org/o1> ;
+	<http://example.org/p1> ?o ;
+	<http://example.org/p2> [
+		?2p <http://example.org/o2> ;
+		<http://example.org/p3> ?_o
+	] .
diff --git a/test/test_overflow.c b/test/test_overflow.c
index 6d5c6d0c..335cd5c7 100644
--- a/test/test_overflow.c
+++ b/test/test_overflow.c
@@ -25,15 +25,16 @@ static const size_t min_stack_size = 4 * sizeof(size_t) + 256u;
 static const size_t max_stack_size = 1024u;
 
 static SerdStatus
-test_size(SerdWorld* const  world,
-          const char* const str,
-          const SerdSyntax  syntax,
-          const size_t      stack_size)
+test_size(SerdWorld* const      world,
+          const char* const     str,
+          const SerdSyntax      syntax,
+          const SerdReaderFlags flags,
+          const size_t          stack_size)
 {
   SerdSink*         sink        = serd_sink_new(NULL, NULL, NULL);
   SerdByteSource*   byte_source = serd_byte_source_new_string(str, NULL);
   SerdReader* const reader =
-    serd_reader_new(world, syntax, 0u, sink, stack_size);
+    serd_reader_new(world, syntax, flags, sink, stack_size);
 
   assert(reader);
 
@@ -47,17 +48,18 @@ test_size(SerdWorld* const  world,
 }
 
 static void
-test_all_sizes(SerdWorld* const  world,
-               const char* const str,
-               const SerdSyntax  syntax)
+test_all_sizes(SerdWorld* const      world,
+               const char* const     str,
+               const SerdSyntax      syntax,
+               const SerdReaderFlags flags)
 {
   // Ensure reading with the maximum stack size succeeds
-  SerdStatus st = test_size(world, str, syntax, max_stack_size);
+  SerdStatus st = test_size(world, str, syntax, flags, max_stack_size);
   assert(!st);
 
   // Test with an increasingly smaller stack
   for (size_t size = max_stack_size; size > min_stack_size; --size) {
-    if ((st = test_size(world, str, syntax, size))) {
+    if ((st = test_size(world, str, syntax, flags, size))) {
       assert(st == SERD_ERR_OVERFLOW);
     }
   }
@@ -76,7 +78,7 @@ test_ntriples_overflow(void)
   SerdWorld* const world = serd_world_new();
 
   for (const char* const* t = test_strings; *t; ++t) {
-    test_all_sizes(world, *t, SERD_NTRIPLES);
+    test_all_sizes(world, *t, SERD_NTRIPLES, 0u);
   }
 
   serd_world_free(world);
@@ -98,6 +100,7 @@ test_turtle_overflow(void)
     "<http://example.org/s> <http://example.org/p> _:blank .",
     "<http://example.org/s> <http://example.org/p> true .",
     "<http://example.org/s> <http://example.org/p> \"\"@en .",
+    "?subject ?predicate ?object .",
     "(((((((((42))))))))) <http://example.org/p> <http://example.org/o> .",
     "@prefix eg: <http://example.org/ns/test> .",
     "@base <http://example.org/base> .",
@@ -165,7 +168,7 @@ test_turtle_overflow(void)
   SerdWorld* const world = serd_world_new();
 
   for (const char* const* t = test_strings; *t; ++t) {
-    test_all_sizes(world, *t, SERD_TURTLE);
+    test_all_sizes(world, *t, SERD_TURTLE, SERD_READ_VARIABLES);
   }
 
   serd_world_free(world);