[oxlint] result inconsistencies for rule no-cycle

[folkvir]

There are inconsistencies in the result of oxlint using --threads N and in the results compared to eslint.

There are two types of problems:

• inconsistencies with the --threads option. When I use --threads 1 I get 2 results and when I use --threads 10 I get 5 results. The expected is 7 cycles; what eslint is reporting.
• Eslint reports 7 cycles while oxlint reports only 5 of them. The 7 cycles are explained in the test repository.

Reproductible repo: https://github.com/folkvir/oxc_no-cycle_test
The cycles are explained in the README

[Boshen]

Regarding different number of cycles comparing to eslint - I think both results are correct because a cycle is formed, the difference seems to come from a depth first or a breath first search.

I'll look into --threads next.

[Boshen]

I can't fix the mismatch for the time being, it is a bit difficult figure out what's wrong with --threads=1.

Given that no-cycles are still being detected for all the entry files, I'm going to close this as won't fix for now.