Make DAGCircuit's topological iterators infallible
#15434
Conversation
Both `topological_nodes` and `topological_op_nodes` are already guaranteed to be infallible by the data-consistency requirements of the `DAGCircuit`. This guarantees it in the interface as well. The return type of `DAGCircuit::topological_nodes` was needlessly being hidden, where users could make use of it.
jakelishman
added
Changelog: None
|
One or more of the following people are relevant to this code:
|
| rustworkx_core::dag_algo::TopologicalSortError::CycleOrBadInitialState => { | ||
| PyValueError::new_err(format!("{e}")) | ||
| } | ||
| rustworkx_core::dag_algo::TopologicalSortError::KeyError(_) => { |
There was a problem hiding this comment.
Hi @jakelishman , thank you so much for your contribution! This PR is very cleanly and simplistically written; significantly cleaning up call sites and removing a great deal of error handling boilerplate. This implementation is flawless. The PR LGTM. I had but one doubt however, for some clarity.
In this line, you have removed the KeyError branch in the match statement, and I would like to know your thought process behind doing so, considering the branch was already unreachable, and was put in place just to make the match exhaustive, and in turn make it robust in compilation; is there a reason that making this change does not affect compilation?
3 participants
Both
topological_nodesandtopological_op_nodesare already guaranteed to be infallible by the data-consistency requirements of theDAGCircuit. This guarantees it in the interface as well.The return type of
DAGCircuit::topological_nodeswas needlessly being hidden, where users could make use of it.Summary
Details and comments
This is just fixing a small annoyance I had when looking at other things to do with the topological sorters. It's not particularly important.