[Fixes #3170] Skip ensure of repl available on xref functions

[ericdallo]

Fixes #3170 adding a extra arg to skip the ensure for xref functions.

---

Before submitting the PR make sure the following things have been done (and denote this
by checking the relevant checkboxes):

• The commits are consistent with our contribution guidelines
• You've added tests (if possible) to cover your change(s)
• All tests are passing (eldev test)
• All code passes the linter (eldev lint) which is based on elisp-lint and includes
  • byte-compilation, checkdoc, check-declare, packaging metadata, indentation, and trailing whitespace checks.
• You've updated the changelog (if adding/changing user-visible functionality)
• You've updated the user manual (if adding/changing user-visible functionality)

Thanks!

[vemv]

Looks like it would work!

Did you get to try it? (Of course, without any other workaround in place)

[ericdallo]

Yes @vemv, I called that function and it did return nil instead of throwing a error :)

[ericdallo]

Thank you!

[vemv]

🍻!

Feel free to propose bumping the commit in the Doom repo

[ericdallo]

that's what I just did! 😄

[vemv]

Just in case, I meant upstream e.g. doomemacs/doomemacs#6183

[ericdallo]

yes, I was talking with Henrik on Doom's discord, he said will bump cider soon with this fix :)

[vemv]

awesome!

[vemv]

This change had an oversight fixed in 3627780

[bbatsov]

I'm not sure this was the right approach to solving the problem, but you merged it before I saw the PR. Not that big of deal in general, but I was thinking that actually the predicate function should always return t/nil and there would be a cider-ensure-op-supported matching it for the cases where the other behavior is needed.

[bbatsov]

(likely the behavior you're trying to fix is somewhat accidental, given where it's coming from)

[ericdallo]

Yeah, I like precicate functions without side-effects, your idea sounds better indeed, but probably will require a little bit more refactor I think

[bbatsov]

We can always simply check if the test suite works fine if we make this the default behavior (suppressing the exception from cider-current-repl). I think this function didn't exists back when cider-nrepl-op-supported-p was written and its introduction altered the API. I assume that code that actually care about getting an exception is using cider-ensure-connected directly.

[vemv]

I'm not sure this was the right approach to solving the problem, but you merged it before I saw the PR.

Yeah that wasn't ideal, sorry, I guess I wanted to leave it bug-free for Doom/lsp users asap.

Given this 'api' has not been shared as a tagged release, I'd be happy for the solution to be changed if that's better.