Joining the parts of node identifiers is unnecessary and loses information by concealing the boundaries of rule names, label names, and label indices. The patch employs tuples instead of concatenated strings.
coverage: 85.904% (-0.3%) from 86.253%
when pulling edd51a40d92024e922a0b5b08149059658056baa on tuple-ids-v2
into 5e8d0c4f14f0943f2bde190148f33c66f105205e on master.
Joining the parts of node identifiers is unnecessary and loses information by concealing the boundaries of rule names, label names, and label indices. The patch employs tuples instead of concatenated strings.