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.

29 files changed, 275 insertions, 24 deletions

diff --git a/NEWS b/NEWS
index aa1087b6..ddc64095 100644
--- a/NEWS
+++ b/NEWS
@@ -3,6 +3,7 @@ serd (1.0.1) unstable;
   * Add SerdBuffer for mutable buffers to keep SerdChunk const-correct
   * Add SerdWorld for shared library state
   * Add option for writing terse output without newlines
+  * Add support for parsing variables
   * Add support for writing terse collections
   * Add support for xsd:float and xsd:double literals
   * Bring read/write interface closer to C standard
@@ -17,7 +18,7 @@ serd (1.0.1) unstable;
   * Use a fixed-size reader stack
   * Use char* for strings in public API
 
- -- David Robillard <d@drobilla.net>  Wed, 13 Jan 2021 13:29:44 +0000
+ -- David Robillard <d@drobilla.net>  Wed, 13 Jan 2021 13:37:55 +0000
 
 serd (0.30.11) unstable;
 
diff --git a/doc/serdi.1 b/doc/serdi.1
index f9c98492..495c3940 100644
--- a/doc/serdi.1
+++ b/doc/serdi.1
@@ -6,7 +6,7 @@
 .Nd read and write RDF syntax
 .Sh SYNOPSIS
 .Nm serdi
-.Op Fl abefhlqtv
+.Op Fl abefhlqtvx
 .Op Fl I Ar base
 .Op Fl c Ar prefix
 .Op Fl i Ar syntax
@@ -139,6 +139,13 @@ Display version information and exit.
 Write output to the given
 .Ar filename
 instead of stdout.
+.Pp
+.It Fl x
+Support parsing variable nodes.
+Variables can be written in SPARQL style, for example
+.Dq ?var
+or
+.Dq $var .
 .El
 .Sh EXIT STATUS
 .Nm
diff --git a/include/serd/serd.h b/include/serd/serd.h
index 82292389..a98b2b8a 100644
--- a/include/serd/serd.h
+++ b/include/serd/serd.h
@@ -535,7 +535,8 @@ typedef struct SerdNodeImpl SerdNode;
 
    An RDF node, in the abstract sense, can be either a resource, literal, or a
    blank.  This type is more precise, because syntactically there are two ways
-   to refer to a resource (by URI or CURIE).
+   to refer to a resource (by URI or CURIE).  Serd also has support for
+   variable nodes to support some features, which are not RDF nodes.
 
    There are also two ways to refer to a blank node in syntax (by ID or
    anonymously), but this is handled by statement flags rather than distinct
@@ -574,7 +575,16 @@ typedef enum {
      is meaningful only within this serialisation.  @see [RDF 1.1
      Turtle](http://www.w3.org/TR/turtle/#grammar-production-BLANK_NODE_LABEL)
   */
-  SERD_BLANK = 4
+  SERD_BLANK = 4,
+
+  /**
+     A variable node.
+
+     Value is a variable name without any syntactic prefix, like "name",
+     which is meaningful only within this serialisation.  @see [SPARQL 1.1
+     Query Language](https://www.w3.org/TR/sparql11-query/#rVar)
+  */
+  SERD_VARIABLE = 5
 } SerdNodeType;
 
 /**
@@ -1556,7 +1566,8 @@ typedef struct SerdReaderImpl SerdReader;
 
 /// Reader options
 typedef enum {
-  SERD_READ_LAX = 1u << 0u ///< Tolerate invalid input where possible
+  SERD_READ_LAX       = 1u << 0u, ///< Tolerate invalid input where possible
+  SERD_READ_VARIABLES = 1u << 1u  ///< Support variable nodes
 } SerdReaderFlag;
 
 /// Bitwise OR of SerdReaderFlag values
diff --git a/src/env.c b/src/env.c
index 2bae1f83..3ebcf646 100644
--- a/src/env.c
+++ b/src/env.c
@@ -322,6 +322,7 @@ serd_env_expand(const SerdEnv* env, const SerdNode* node)
   case SERD_CURIE:
     return expand_curie(env, node);
   case SERD_BLANK:
+  case SERD_VARIABLE:
     break;
   }
 
diff --git a/src/n3.c b/src/n3.c
index a6aa71ba..517dce8f 100644
--- a/src/n3.c
+++ b/src/n3.c
@@ -963,10 +963,53 @@ read_literal(SerdReader* const reader,
 }
 
 static SerdStatus
-read_verb(SerdReader* const reader, SerdNode** const dest)
+read_VARNAME(SerdReader* const reader, SerdNode** const dest)
+{
+  // Simplified from SPARQL: VARNAME ::= (PN_CHARS_U | [0-9])+
+  SerdNode*  n  = *dest;
+  SerdStatus st = SERD_SUCCESS;
+  int        c  = 0;
+  peek_byte(reader);
+  while ((c = peek_byte(reader))) {
+    if (is_digit(c) || c == '_') {
+      st = push_byte(reader, n, eat_byte_safe(reader, c));
+    } else if ((st = read_PN_CHARS(reader, n))) {
+      st = st > SERD_FAILURE ? st : SERD_SUCCESS;
+      break;
+    }
+  }
+
+  return st;
+}
+
+static SerdStatus
+read_Var(SerdReader* const reader, SerdNode** const dest)
+{
+  if (!(reader->flags & SERD_READ_VARIABLES)) {
+    return r_err(
+      reader, SERD_ERR_BAD_SYNTAX, "syntax does not support variables\n");
+  }
+
+  if (!(*dest = push_node(reader, SERD_VARIABLE, "", 0))) {
+    return SERD_ERR_OVERFLOW;
+  }
+
+  assert(peek_byte(reader) == '$' || peek_byte(reader) == '?');
+  serd_byte_source_advance(reader->source);
+
+  return read_VARNAME(reader, dest);
+}
+
+static SerdStatus
+read_verb(SerdReader* reader, SerdNode** dest)
 {
   const size_t orig_stack_size = reader->stack.size;
-  if (peek_byte(reader) == '<') {
+
+  switch (peek_byte(reader)) {
+  case '$':
+  case '?':
+    return read_Var(reader, dest);
+  case '<':
     return read_IRIREF(reader, dest);
   }
 
@@ -1138,6 +1181,9 @@ read_object(SerdReader* const  reader,
     case '<':
     case '_':
       break;
+    case '$':
+    case '?':
+      break;
     default:
       return r_err(reader, SERD_ERR_BAD_SYNTAX, "expected: ':', '<', or '_'\n");
     }
@@ -1147,6 +1193,10 @@ read_object(SerdReader* const  reader,
   case EOF:
   case ')':
     return r_err(reader, SERD_ERR_BAD_SYNTAX, "expected object\n");
+  case '$':
+  case '?':
+    ret = read_Var(reader, &o);
+    break;
   case '[':
     simple = false;
     ret    = read_anon(reader, *ctx, false, &o);
@@ -1375,6 +1425,10 @@ read_subject(SerdReader* const reader,
   SerdStatus st      = SERD_SUCCESS;
   bool       ate_dot = false;
   switch ((*s_type = peek_byte(reader))) {
+  case '$':
+  case '?':
+    st = read_Var(reader, dest);
+    break;
   case '[':
     st = read_anon(reader, ctx, true, dest);
     break;
@@ -1643,6 +1697,7 @@ read_n3_statement(SerdReader* const reader)
       if (s_type == '(' || (s_type == '[' && !*ctx.flags)) {
         return r_err(reader, SERD_ERR_BAD_SYNTAX, "invalid graph name\n");
       }
+
       ctx.graph   = ctx.subject;
       ctx.subject = NULL;
       TRY(st, read_wrappedGraph(reader, &ctx));
@@ -1658,6 +1713,7 @@ read_n3_statement(SerdReader* const reader)
       }
 
       return st > SERD_FAILURE ? st : SERD_ERR_BAD_SYNTAX;
+
     } else if (!ate_dot) {
       read_ws_star(reader);
       st = eat_byte_check(reader, '.');
@@ -1717,10 +1773,22 @@ read_nquadsDoc(SerdReader* const reader)
       return SERD_ERR_BAD_SYNTAX;
     }
 
-    // subject predicate object
     if ((st = read_subject(reader, ctx, &ctx.subject, &s_type)) ||
-        !read_ws_star(reader) || (st = read_IRIREF(reader, &ctx.predicate)) ||
-        !read_ws_star(reader) ||
+        !read_ws_star(reader)) {
+      return st;
+    }
+
+    switch (peek_byte(reader)) {
+    case '$':
+    case '?':
+      st = read_Var(reader, &ctx.predicate);
+      break;
+    case '<':
+      st = read_IRIREF(reader, &ctx.predicate);
+      break;
+    }
+
+    if (st || !read_ws_star(reader) ||
         (st = read_object(reader, &ctx, false, &ate_dot))) {
       return st;
     }
@@ -1730,6 +1798,9 @@ read_nquadsDoc(SerdReader* const reader)
       switch (peek_byte(reader)) {
       case '.':
         break;
+      case '?':
+        TRY(st, read_Var(reader, &ctx.graph));
+        break;
       case '_':
         TRY(st, read_BLANK_NODE_LABEL(reader, &ctx.graph, &ate_dot));
         break;
diff --git a/src/node.c b/src/node.c
index bd306875..70f53033 100644
--- a/src/node.c
+++ b/src/node.c
@@ -170,7 +170,8 @@ result(const SerdStatus status, const size_t count)
 SerdNode*
 serd_new_simple_node(const SerdNodeType type, const SerdStringView str)
 {
-  if (type != SERD_BLANK && type != SERD_CURIE && type != SERD_URI) {
+  if (type != SERD_BLANK && type != SERD_CURIE && type != SERD_URI &&
+      type != SERD_VARIABLE) {
     return NULL;
   }
 
diff --git a/src/reader.c b/src/reader.c
index fe88ee14..d70cbb53 100644
--- a/src/reader.c
+++ b/src/reader.c
@@ -196,6 +196,7 @@ serd_reader_new(SerdWorld* const      world,
   me->sink    = sink;
   me->stack   = serd_stack_new(stack_size, serd_node_align);
   me->syntax  = syntax;
+  me->flags   = flags;
   me->next_id = 1;
   me->strict  = !(flags & SERD_READ_LAX);
 
diff --git a/src/reader.h b/src/reader.h
index 76f46506..b08c3fd0 100644
--- a/src/reader.h
+++ b/src/reader.h
@@ -53,6 +53,7 @@ struct SerdReaderImpl {
   SerdByteSource* source;
   SerdStack       stack;
   SerdSyntax      syntax;
+  SerdReaderFlags flags;
   unsigned        next_id;
   uint8_t*        buf;
   char*           bprefix;
diff --git a/src/serdi.c b/src/serdi.c
index 3025b494..9a5495af 100644
--- a/src/serdi.c
+++ b/src/serdi.c
@@ -74,6 +74,7 @@ print_usage(const char* const name, const bool error)
   fprintf(os, "  -t           Write terser output without newlines.\n");
   fprintf(os, "  -v           Display version information and exit.\n");
   fprintf(os, "  -w FILENAME  Write output to FILENAME instead of stdout.\n");
+  fprintf(os, "  -x           Support parsing variable nodes like `?x'.\n");
   return error ? 1 : 0;
 }
 
@@ -190,6 +191,8 @@ main(int argc, char** argv)
         writer_flags |= SERD_WRITE_TERSE;
       } else if (opt == 'v') {
         return print_version();
+      } else if (opt == 'x') {
+        reader_flags |= SERD_READ_VARIABLES;
       } else if (argv[a][1] == 'I') {
         if (++a == argc) {
           return missing_arg(prog, 'I');
diff --git a/src/statement.c b/src/statement.c
index 51788e05..0ec6520a 100644
--- a/src/statement.c
+++ b/src/statement.c
@@ -28,7 +28,8 @@ is_resource(const SerdNode* const node)
 {
   const SerdNodeType type = serd_node_type(node);
 
-  return type == SERD_URI || type == SERD_CURIE || type == SERD_BLANK;
+  return type == SERD_URI || type == SERD_CURIE || type == SERD_BLANK ||
+         type == SERD_VARIABLE;
 }
 
 bool
diff --git a/src/writer.c b/src/writer.c
index 2cee09c5..1ad62a18 100644
--- a/src/writer.c
+++ b/src/writer.c
@@ -899,6 +899,18 @@ write_blank(SerdWriter* const        writer,
 }
 
 SERD_WARN_UNUSED_RESULT static SerdStatus
+write_variable(SerdWriter* const writer, const SerdNode* const node)
+{
+  SerdStatus st = SERD_SUCCESS;
+
+  TRY(st, esink("?", 1, writer));
+  TRY(st, esink(serd_node_string(node), node->length, writer));
+
+  writer->last_sep = SEP_NONE;
+  return st;
+}
+
+SERD_WARN_UNUSED_RESULT static SerdStatus
 write_node(SerdWriter* const        writer,
            const SerdNode* const    node,
            const SerdField          field,
@@ -919,6 +931,9 @@ write_node(SerdWriter* const        writer,
   case SERD_BLANK:
     st = write_blank(writer, node, field, flags);
     break;
+  case SERD_VARIABLE:
+    st = write_variable(writer, node);
+    break;
   }
 
   return st;
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);