[JSONSelection] Remove &mut errors: IndexSet<ApplyToError> from apply_to_path method signatures

[benjamn]

Following up on this suggestion from @nicholascioli: #5762 (comment)

The apply_to_path method used to have more than one mutable parameter (input_path as well as errors), but now that input_path is immutable, I agree it's cleaner to go all the way and avoid any mutable parameters, so apply_to_path can be a pure function.

One big benefit of this refactoring is that each apply_to_path method now returns just the errors resulting from the subtree that it handled, whereas the previous &mut errors: IndexSet<ApplyToError> strategy made it difficult to know which errors came from which apply_to_path method call. This will help if we ever decide to memoize the results of apply_to_path methods, or if we want to expose ApplyToErrors for processing by -> methods (imagine some sort of ->catch handler that captures/handles errors from further down its PathList chain).

As discussed in #5987 (comment), there might not be much point in deduplicating errors now, since their path and range information effectively makes them all unique, but if we decide to stop the deduplication that's now an easy change within JSONSelection::apply_with_vars.

These changes should be entirely an implementation detail. As evidence for this claim, I did not have to change any tests.

[router-perf]

CI performance tests

• connectors-const - Connectors stress test that runs with a constant number of users
• const - Basic stress test that runs with a constant number of users
• demand-control-instrumented - A copy of the step test, but with demand control monitoring and metrics enabled
• demand-control-uninstrumented - A copy of the step test, but with demand control monitoring enabled
• enhanced-signature - Enhanced signature enabled
• events - Stress test for events with a lot of users and deduplication ENABLED
• events_big_cap_high_rate - Stress test for events with a lot of users, deduplication enabled and high rate event with a big queue capacity
• events_big_cap_high_rate_callback - Stress test for events with a lot of users, deduplication enabled and high rate event with a big queue capacity using callback mode
• events_callback - Stress test for events with a lot of users and deduplication ENABLED in callback mode
• events_without_dedup - Stress test for events with a lot of users and deduplication DISABLED
• events_without_dedup_callback - Stress test for events with a lot of users and deduplication DISABLED using callback mode
• extended-reference-mode - Extended reference mode enabled
• large-request - Stress test with a 1 MB request payload
• no-tracing - Basic stress test, no tracing
• reload - Reload test over a long period of time at a constant rate of users
• step-jemalloc-tuning - Clone of the basic stress test for jemalloc tuning
• step-local-metrics - Field stats that are generated from the router rather than FTV1
• step-with-prometheus - A copy of the step test with the Prometheus metrics exporter enabled
• step - Basic stress test that steps up the number of users over time
• xlarge-request - Stress test with 10 MB request payload
• xxlarge-request - Stress test with 100 MB request payload

[dylan-apollo]

The more functional style feels much nicer to me, thanks! I think it will be a more efficient in the error case to switch to passing owned values instead of references which we reallocate on. Might shake up too many other things, though, so if it ends up being a mess, we can stick with this.

[dylan-apollo]

Nice trick, I always forget that I can hack in methods like this