Fix jump regression in LSP references action handler

[adriangoransson]

Description

This change restores the possibility to exclude the current line when invoking the lsp_references picker.

• opts.include_current_line is by default unset, so the previous equality check would fail unless the option was set explicitly.
• vim.api.nvim_win_get_cursor() returns both line and column, so it can't be directly compared with the returned line number.
• The actual comparison was expecting quickfix-like items, when in actuality, we were dealing with raw LSP location objects.

Type of change

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

• Triggering lsp_references on a symbol which is only referenced in one other location jumps directly to that location.
• Triggering lsp_references on a symbol which is referenced in multiple other locations lists reference locations. References on the cursor line where the picker was triggered are excluded.
• Triggering lsp_references with opts = { include_current_line = true } yields the same behavior as it does today.

Configuration:

• Neovim version (nvim --version):

  NVIM v0.9.5
  Build type: Release
  LuaJIT 2.1.1713773202

• Operating system and version:

  macOs 14.4.1 (23E224)

Checklist:

• My code follows the style guidelines of this project (stylua)
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation (lua annotations) N/A, I think?

[adriangoransson]

Everything has been working for a couple of days, but I got warning messages:

locations_to_items must be called with valid offset encoding

I added offset_encoding as a parameter to action_handlers to resolve it. It's not very pretty, and currently not annotated correctly either, but the warnings are gone. Any assistance to make it nicer is kindly accepted!

[jamestrew]

Thanks!
For the offset_encoding part, I think it's probably better to call vim.lsp.util.locations_to_items earlier, before calling apply_action_handler, then pass the return value of it to apply_action_handler.

This is what we used to do before my (terrible 😅) refactor. See here:

telescope.nvim/lua/telescope/builtin/__lsp.lua

Lines 25 to 67 in b744cf5

	local locations = {}
	if result then
	local results = vim.lsp.util.locations_to_items(result, vim.lsp.get_client_by_id(ctx.client_id).offset_encoding)
	if not include_current_line then
	locations = vim.tbl_filter(function(v)
	-- Remove current line from result
	return not (v.filename == filepath and v.lnum == lnum)
	end, vim.F.if_nil(results, {}))
	else
	locations = vim.F.if_nil(results, {})
	end
	end
	if vim.tbl_isempty(locations) then
	return
	end
	if #locations == 1 and opts.jump_type ~= "never" then
	if filepath ~= locations[1].filename then
	if opts.jump_type == "tab" then
	vim.cmd "tabedit"
	elseif opts.jump_type == "split" then
	vim.cmd "new"
	elseif opts.jump_type == "vsplit" then
	vim.cmd "vnew"
	elseif opts.jump_type == "tab drop" then
	vim.cmd("tab drop " .. locations[1].filename)
	end
	end
	-- jump to location
	local location = locations[1]
	local bufnr = opts.bufnr
	if location.filename then
	local uri = location.filename
	if not utils.is_uri(uri) then
	uri = vim.uri_from_fname(uri)
	end
	bufnr = vim.uri_to_bufnr(uri)
	end
	vim.api.nvim_win_set_buf(0, bufnr)
	vim.api.nvim_win_set_cursor(0, { location.lnum, location.col - 1 })
	return

[adriangoransson]

For the offset_encoding part, I think it's probably better to call vim.lsp.util.locations_to_items earlier, before calling apply_action_handler, then pass the return value of it to apply_action_handler.

Thanks for the help (and for your work on the plugin! 🌟). I think I managed to solve it in an okay way, by letting action handlers receive and process both the lsp locations as well as the qf items. It requires some discipline to keep the processing "synced" between the two, but it keeps the calling function (that doesn't yet know which representation it is going to use) simpler.

I added a couple of refactoring commits to the PR as well, but can drop those if you wish.

[jamestrew]

thanks lgtm!