test(keymap): react to Nightly changes regarding autoindent

[abeldekat]

• I have read CONTRIBUTING.md
• I have read CODE_OF_CONDUCT.md

Details:

• See PR 39961 in neovim/neovim

I wasn't able to fix this properly. The test now passes because I disabled autoindent, allowing the third line to be empty.

I wonder why the nvim-autopairs mock is more specific than the other mocks, which mostly add a line to the _G.log. Is the comment stating -- Mock exactly how 'nvim-autopairs' works here still applicable? Perhaps a simple _G.log addition is enough.

The failing test:

FAIL in tests/test_keymap.lua | map_multistep() | built-in steps | nvimautopairs_cr:
  Failed expectation for equality
  Cause: different values at key 3, left = ")", right = ""
  Left:  { "", "(", ")", "" }
  Right: { "", "(", "", ")" }
  Traceback:
    tests/test_keymap.lua:1088

EDIT:
I think this issue can also be closed...

EDIT2:
It's strange that the third empty line is now subject to truncation. Maybe the vim-patch is not selective enough?

[echasnovski]

Thanks for the PR!

I wonder why the nvim-autopairs mock is more specific than the other mocks, which mostly add a line to the _G.log. Is the comment stating -- Mock exactly how 'nvim-autopairs' works here still applicable?

Mostly because it requires extra key translation to be useful. And that sequence of keys is exactly what 'nvim-autopairs' uses.

As the matter of fact, the 'nvim-autopairs' plugin itself now acts on Nightly exactly as is detected by 'mini.keymap' test. It might be the problem in either vim-patch or 'nvim-autopairs'. Don't really have the bandwidth to investigate this (although it would probably be a net positive for Neovim ecosystem :) ).