That makes sense. Which-key doesn’t allow the timeout to take place after gc since it knows about gcc. While that functionality is useful when you don’t know which key to press next, in situations like these it would create problems. I think your best solution would still be to modify the mappings. Maybe remap gc to gcl (comment last) or something like that.
LSP maybe portable with it’s config if the LSP themselves are independent. Checkout Mason which seems to make it easier to bundle neovim and “portable” LSP. There was another project similar to Mason with some more features, but I forgot it’s name. So search around to see if that fits your requirements.
Like I said, I haven’t used mini.comment, but having both gc and gcc mapping may cause problems. If I remember correctly, in these situations, neovim waits after receiving gc command to figure out if you are trying to execute comment last or gcc with comment line. Depending on timelen (or timeout, forgot the exact name) setting the command you actually execute will defer based on the key pressed/not pressed after gc.
Also, if gc is mapped to comment last, and there was no last comment action performed, it might do nothing (maybe check the readme to be sure). So maybe try changing the mapping to see if that helps.
Try using :verbose nmap gc to find out what the shortcut is currently mapped to.
I don’t use mini.comment, but I assume it needs an operator after gc to for commenting. I have seen gcc used to comment out the line. Just wanted to make sure you are using the correct mapping.
I think it has to do with the way the Haskell packages are disturbed. So if you installed the pandoc from extra repository, and there was an update for any of the Haskell packages that pandoc depended on, then pandoc will have to be updated again. Meaning, instead of pandoc forcing Haskell updates, it is the other way around.
If you only use pandoc and don’t normally program in Haskell, check out pandoc-bin. This way you will only get updates for pandoc itself and its dependencies won’t force an update.
For me, it works as well on nightly without any modifications as it did when the archival announcement was made. But I only use formatters and one or two linters. So can’t confirm for other features.
That said, I am thinking about changing to conform.nivm for formatting.
The notice on packer suggested to use either lazy or pckr
NOTICE:
This repository is currently unmaintained. For the time being (as of August, 2023), it is recommended to use one of the following plugin managers instead:
lazy.nvim: Most stable and maintained plugin manager for Nvim.
pckr.nvim: Spiritual successor of packer.nvim. Functional but not as stable as lazy.nvim.
Considering null-ls still works on nightly, it should be fine.
Next time, try enabling display of whitespace characters through
set list. If I remember correctly, they can help distinguish between the Windows carriage return character and Linux end of line character.This may help for a quick overview.