Where is part 2?

In D7987#187491, @deadalnix wrote:

Where is part 2?

I will work on adding the corresponding linter tomorrow, and take some to time to properly check that it does not degrade performances too much. Fabien told me that external linters are slower than a regex in PHP.

I already ran the command locally. There will be quite a few unused functions, variables and classes to be removed. I imagine I will do that in a separate diff too.

In D7987#187502, @PiRK wrote:

In D7987#187491, @deadalnix wrote:

Where is part 2?

I will work on adding the corresponding linter tomorrow, and take some to time to properly check that it does not degrade performances too much. Fabien told me that external linters are slower than a regex in PHP.

It turns out that this tool is not very reliable, for our codebase. vulture --min-confidence=60 finds many false positives, and vulture --min-confidence=70 does not detect obviously unused code added on purpose for testing.
It was disabled in Core: https://github.com/bitcoin/bitcoin/pull/16961

I think the problem is related to the way test_runner.py dynamically finds tests by doing an os.listdir and then runs them as external processes, instead of doing python imports and running functions. We basically use python as a scripting language.