demo 980 resolution

test/test_980.py

@@ -0,0 +1,28 @@
+from rdflib import Graph
+
+
+def test_980():
+    """
+    The problem that this test ensures rdflib solves is that, previous to PR #1108, the
+    parsing of two triples with the same n-triples Blank Nodes IDs, here _:0, would
+    result in triples with the same rdflib internal BN IDs, e.g.
+    rdflib.term.BNode('Ne3fd8261b37741fca22d502483d88964'), see the Issue #980. They
+    should have different IDs.
+    """
+    graph1 = """
+    _:0 <http://purl.obolibrary.org/obo/RO_0002350> <http://www.gbif.org/species/0000001> .
+    """
+    graph2 = """
+    _:0 <http://purl.obolibrary.org/obo/RO_0002350> <http://www.gbif.org/species/0000002> .
+    """
+
+    g = Graph()
+    g.parse(data=graph1, format="nt")
+    g.parse(data=graph2, format="nt")
+
+    subs = 0
+    for s in g.subjects(None, None):
+        subs += 1
+
+    # we must see two different BN subjects
+    assert subs == 2