Speedup Literal.__hash__ and Literal.__eq__ by accessing directly _da…

[rchateauneu]

Literal.hash and Literal.eq use self.datatype and self. language properties. It is faster to directly use the private members Literal._dataype and Literal.__language.

Test importing orkg.nt

Before:

After:

[coveralls]

Coverage decreased (-0.08%) to 75.627% when pulling 282a8b7 on rchateauneu:speedup_term_hash into 58b23fd on RDFLib:master.

[coveralls]

Coverage decreased (-0.08%) to 75.627% when pulling 282a8b7 on rchateauneu:speedup_term_hash into 58b23fd on RDFLib:master.

[coveralls]

Coverage decreased (-0.08%) to 75.627% when pulling 282a8b7 on rchateauneu:speedup_term_hash into 58b23fd on RDFLib:master.

[coveralls]

Coverage decreased (-0.08%) to 75.627% when pulling 282a8b7 on rchateauneu:speedup_term_hash into 58b23fd on RDFLib:master.

[coveralls]

Coverage decreased (-0.08%) to 75.627% when pulling 282a8b7 on rchateauneu:speedup_term_hash into 58b23fd on RDFLib:master.

[coveralls]

Coverage remained the same at 75.705% when pulling 42c90d9 on rchateauneu:speedup_term_hash into 58b23fd on RDFLib:master.

[rchateauneu]

It failed only Python 3.7, but 3.6 and 3.8 worked OK. This is strange because the changes are very simple.

The failed test is:

ERROR: test_query_with_added_rdf_prolog (test.test_sparqlstore.SPARQLStoreDBPediaTestCase)
...
  File "/usr/local/lib/python3.7/socket.py", line 716, in create_connection
    sock.connect(sa)
OSError: [Errno 99] Cannot assign requested address

I tend to think that it is a transient error related to a temporary network problem.

Is it possible to rerun a build on the same PR ?

[nicholascar]

Don't worry about the "cannot assign address" test failure you're noticing: that's spurious and won't prevent this PR going through.