fix(keys): refactor retrigger mechanism

[MurdeRM3L0DY]

should fix #286.

[folke]

I did already fix that, no?

I use mini.ai myself with those keybiondings and it works as expected

[MurdeRM3L0DY]

yes, you fixed the main problem. However consider this case:

1. you have the string "hello"
2. if you try to ciw with the cursor on the character h, you'll actually end up performing the action on the previous character " which happens to be a word in vim

I suspect this happens because c operator enters insert mode so when you retrigger to re-enter normal mode, the cursor shifts backwards by one position.

[folke]

Thank you for the clarification. Will need to check this PR with some issues from the past related to keymaps.

I'm also not that sure of the vim.schedule in there

[MurdeRM3L0DY]

Yeah. I agree with the sentiment about using vim.schedule.

I've got another solution via this branch which uses expr and remap to retrigger the original keymap.

Edit: I've figured out a better solution for the retrigger logic

[folke]

I implemented the solution you propose a while back, but it failed some issues people had in the past. Although, I did not add that <ignore>.

What I'll do first, I'll add a bunch of unit tests for those issues I'm talking about, so we have something to test this with. Will do so on Monday.

[MurdeRM3L0DY]

Understood. From my various "tests" it works pretty well:

• loading multiple plugins with one keymap. e.g gzaiw[char] loads mini.surround then mini.ai for me
• count works as expected

[mathjiajia]

yes, you fixed the main problem. However consider this case:

1. you have the string "hello"
2. if you try to ciw with the cursor on the character h, you'll actually end up performing the action on the previous character " which happens to be a word in vim

I suspect this happens because c operator enters insert mode so when you retrigger to re-enter normal mode, the cursor shifts backwards by one position.

When will this pr be merged?
I also encountered this issue.

[folke]

I still need to check existing issues related to keymaps and make sure they still work as expected. If you want to help, add a comment to this PR linking to existing issues I should re-check with this PR

[MurdeRM3L0DY]

AFAICT, these are the more relevant issues regarding the keys handler:

• bug: Introduced regression with key handling #308
• bug: key was broken for <M-t> #307
• bug: operator-mode mappings don't get retriggered correctly the first time. #286
• bug: when a count is given to a key through which a module is triggered, the module gets loaded, but the count is forgotten #252
• key lazyloading with which-key.nvim needs to be used twice #71
• Plugins lazily loaded on keys with an identical mapping in config won't show on first request #28

All of these seems to be fixed using neovim master but not on neovim stable (0.8.x). Particularly:

• keys need to be pressed twice(sometimes)
  • Upon opening neovim, When I perform any(literally) action, subsequent registered keys need to be pressed twice to trigger
• keys are pasted when triggering in visual mode.

[folke]

ok, thanks. I'll try to look into it later today. I do think the pending keys probably need to be added to your PR, but I'll see if it's needed when I encounter issues

[folke]

I checked all existing issues and these changes indeed seem to work for all of them. Thanks for fixing this!

[MurdeRM3L0DY]

Happy to help! Though I did mention that there were some issues with neovim stable. Were you able to verify this?

[folke]

It all seemed to work for me on stable. What issue in particular didnt work on stable for you?

[folke]

This one seems to work on nightly, but not on stable. #511

Will look into it