Hey, thanks for the PR!

I had a few suggestions:

• The doc mentions a "comma-separated list of exactly 8 characters." Maybe we could validate for that directly, instead of assuming?
• The split operation might be unnecessary sincevim.opt.winborder:get() already gives us a table.

Unrelated to your change, but while we're at it: for checking if the option exists, I think usingvim.fn.exists is a bit more direct and idiomatic than usingpcall.

The doc mentions a "comma-separated list of exactly 8 characters." Maybe we could validate for that directly, instead of assuming?

I don't think this is necessary since it should validate this when setting the option

The split operation might be unnecessary since vim.opt.winborder:get() already gives us a table.

Good point, I can try it out

Unrelated to your change, but while we're at it: for checking if the option exists, I think using vim.fn.exists is a bit more direct and idiomatic than using pcall.

I didn't know it worked for options🤔 will try this as well

I changed to usingvim.fn.exists, and also made use ofvim.opt.winborder:get(). We need to check the length of that.

Before custom borderchars, it returned a string, but now it returns a table. It should work for both scenarios.

Before custom border chars: None of the supported string values have a length of 1, and we will fallback tovim.opt.winborder:get() which just returns the string

After custom border chars: It will always return a table, and for the string values, we need to unpack it, and for the custom border chars, we just return the table

Thank you both!