aboutsummaryrefslogtreecommitdiff
path: root/runtime/lua/vim/lsp/util.lua
diff options
context:
space:
mode:
authorAshkan Kiani <ashkan.k.kiani@gmail.com>2019-11-13 12:55:26 -0800
committerBjörn Linse <bjorn.linse@gmail.com>2019-11-13 21:55:26 +0100
commit00dc12c5d8454a2d3c6806710f63bbb446076e96 (patch)
tree739a7b1f1343f2886868217ce0623717092304b2 /runtime/lua/vim/lsp/util.lua
parentdb436d5277a605606ac92cfabe9e58d86f48f64e (diff)
downloadrneovim-00dc12c5d8454a2d3c6806710f63bbb446076e96.tar.gz
rneovim-00dc12c5d8454a2d3c6806710f63bbb446076e96.tar.bz2
rneovim-00dc12c5d8454a2d3c6806710f63bbb446076e96.zip
lua LSP client: initial implementation (#11336)
Mainly configuration and RPC infrastructure can be considered "done". Specific requests and their callbacks will be improved later (and also served by plugins). There are also some TODO:s for the client itself, like incremental updates. Co-authored by at-tjdevries and at-h-michael, with many review/suggestion contributions.
Diffstat (limited to 'runtime/lua/vim/lsp/util.lua')
-rw-r--r--runtime/lua/vim/lsp/util.lua557
1 files changed, 557 insertions, 0 deletions
diff --git a/runtime/lua/vim/lsp/util.lua b/runtime/lua/vim/lsp/util.lua
new file mode 100644
index 0000000000..f96e0f01a8
--- /dev/null
+++ b/runtime/lua/vim/lsp/util.lua
@@ -0,0 +1,557 @@
+local protocol = require 'vim.lsp.protocol'
+local validate = vim.validate
+local api = vim.api
+
+local M = {}
+
+local split = vim.split
+local function split_lines(value)
+ return split(value, '\n', true)
+end
+
+local list_extend = vim.list_extend
+
+--- Find the longest shared prefix between prefix and word.
+-- e.g. remove_prefix("123tes", "testing") == "ting"
+local function remove_prefix(prefix, word)
+ local max_prefix_length = math.min(#prefix, #word)
+ local prefix_length = 0
+ for i = 1, max_prefix_length do
+ local current_line_suffix = prefix:sub(-i)
+ local word_prefix = word:sub(1, i)
+ if current_line_suffix == word_prefix then
+ prefix_length = i
+ end
+ end
+ return word:sub(prefix_length + 1)
+end
+
+local function resolve_bufnr(bufnr)
+ if bufnr == nil or bufnr == 0 then
+ return api.nvim_get_current_buf()
+ end
+ return bufnr
+end
+
+-- local valid_windows_path_characters = "[^<>:\"/\\|?*]"
+-- local valid_unix_path_characters = "[^/]"
+-- https://github.com/davidm/lua-glob-pattern
+-- https://stackoverflow.com/questions/1976007/what-characters-are-forbidden-in-windows-and-linux-directory-names
+-- function M.glob_to_regex(glob)
+-- end
+
+--- Apply the TextEdit response.
+-- @params TextEdit [table] see https://microsoft.github.io/language-server-protocol/specification
+function M.text_document_apply_text_edit(text_edit, bufnr)
+ bufnr = resolve_bufnr(bufnr)
+ local range = text_edit.range
+ local start = range.start
+ local finish = range['end']
+ local new_lines = split_lines(text_edit.newText)
+ if start.character == 0 and finish.character == 0 then
+ api.nvim_buf_set_lines(bufnr, start.line, finish.line, false, new_lines)
+ return
+ end
+ api.nvim_err_writeln('apply_text_edit currently only supports character ranges starting at 0')
+ error('apply_text_edit currently only supports character ranges starting at 0')
+ return
+ -- TODO test and finish this support for character ranges.
+-- local lines = api.nvim_buf_get_lines(0, start.line, finish.line + 1, false)
+-- local suffix = lines[#lines]:sub(finish.character+2)
+-- local prefix = lines[1]:sub(start.character+2)
+-- new_lines[#new_lines] = new_lines[#new_lines]..suffix
+-- new_lines[1] = prefix..new_lines[1]
+-- api.nvim_buf_set_lines(0, start.line, finish.line, false, new_lines)
+end
+
+-- textDocument/completion response returns one of CompletionItem[], CompletionList or null.
+-- https://microsoft.github.io/language-server-protocol/specification#textDocument_completion
+function M.extract_completion_items(result)
+ if type(result) == 'table' and result.items then
+ return result.items
+ elseif result ~= nil then
+ return result
+ else
+ return {}
+ end
+end
+
+--- Apply the TextDocumentEdit response.
+-- @params TextDocumentEdit [table] see https://microsoft.github.io/language-server-protocol/specification
+function M.text_document_apply_text_document_edit(text_document_edit, bufnr)
+ -- local text_document = text_document_edit.textDocument
+ -- TODO use text_document_version?
+ -- local text_document_version = text_document.version
+
+ -- TODO technically, you could do this without doing multiple buf_get/set
+ -- by getting the full region (smallest line and largest line) and doing
+ -- the edits on the buffer, and then applying the buffer at the end.
+ -- I'm not sure if that's better.
+ for _, text_edit in ipairs(text_document_edit.edits) do
+ M.text_document_apply_text_edit(text_edit, bufnr)
+ end
+end
+
+function M.get_current_line_to_cursor()
+ local pos = api.nvim_win_get_cursor(0)
+ local line = assert(api.nvim_buf_get_lines(0, pos[1]-1, pos[1], false)[1])
+ return line:sub(pos[2]+1)
+end
+
+--- Getting vim complete-items with incomplete flag.
+-- @params CompletionItem[], CompletionList or nil (https://microsoft.github.io/language-server-protocol/specification#textDocument_completion)
+-- @return { matches = complete-items table, incomplete = boolean }
+function M.text_document_completion_list_to_complete_items(result, line_prefix)
+ local items = M.extract_completion_items(result)
+ if vim.tbl_isempty(items) then
+ return {}
+ end
+ -- Only initialize if we have some items.
+ if not line_prefix then
+ line_prefix = M.get_current_line_to_cursor()
+ end
+
+ local matches = {}
+
+ for _, completion_item in ipairs(items) do
+ local info = ' '
+ local documentation = completion_item.documentation
+ if documentation then
+ if type(documentation) == 'string' and documentation ~= '' then
+ info = documentation
+ elseif type(documentation) == 'table' and type(documentation.value) == 'string' then
+ info = documentation.value
+ -- else
+ -- TODO(ashkan) Validation handling here?
+ end
+ end
+
+ local word = completion_item.insertText or completion_item.label
+
+ -- Ref: `:h complete-items`
+ table.insert(matches, {
+ word = remove_prefix(line_prefix, word),
+ abbr = completion_item.label,
+ kind = protocol.CompletionItemKind[completion_item.kind] or '',
+ menu = completion_item.detail or '',
+ info = info,
+ icase = 1,
+ dup = 0,
+ empty = 1,
+ })
+ end
+
+ return matches
+end
+
+-- @params WorkspaceEdit [table] see https://microsoft.github.io/language-server-protocol/specification
+function M.workspace_apply_workspace_edit(workspace_edit)
+ if workspace_edit.documentChanges then
+ for _, change in ipairs(workspace_edit.documentChanges) do
+ if change.kind then
+ -- TODO(ashkan) handle CreateFile/RenameFile/DeleteFile
+ error(string.format("Unsupported change: %q", vim.inspect(change)))
+ else
+ M.text_document_apply_text_document_edit(change)
+ end
+ end
+ return
+ end
+
+ if workspace_edit.changes == nil or #workspace_edit.changes == 0 then
+ return
+ end
+
+ for uri, changes in pairs(workspace_edit.changes) do
+ local fname = vim.uri_to_fname(uri)
+ -- TODO improve this approach. Try to edit open buffers without switching.
+ -- Not sure how to handle files which aren't open. This is deprecated
+ -- anyway, so I guess it could be left as is.
+ api.nvim_command('edit '..fname)
+ for _, change in ipairs(changes) do
+ M.text_document_apply_text_edit(change)
+ end
+ end
+end
+
+--- Convert any of MarkedString | MarkedString[] | MarkupContent into markdown text lines
+-- see https://microsoft.github.io/language-server-protocol/specifications/specification-3-14/#textDocument_hover
+-- Useful for textDocument/hover, textDocument/signatureHelp, and potentially others.
+function M.convert_input_to_markdown_lines(input, contents)
+ contents = contents or {}
+ -- MarkedString variation 1
+ if type(input) == 'string' then
+ list_extend(contents, split_lines(input))
+ else
+ assert(type(input) == 'table', "Expected a table for Hover.contents")
+ -- MarkupContent
+ if input.kind then
+ -- The kind can be either plaintext or markdown. However, either way we
+ -- will just be rendering markdown, so we handle them both the same way.
+ -- TODO these can have escaped/sanitized html codes in markdown. We
+ -- should make sure we handle this correctly.
+
+ -- Some servers send input.value as empty, so let's ignore this :(
+ -- assert(type(input.value) == 'string')
+ list_extend(contents, split_lines(input.value or ''))
+ -- MarkupString variation 2
+ elseif input.language then
+ -- Some servers send input.value as empty, so let's ignore this :(
+ -- assert(type(input.value) == 'string')
+ table.insert(contents, "```"..input.language)
+ list_extend(contents, split_lines(input.value or ''))
+ table.insert(contents, "```")
+ -- By deduction, this must be MarkedString[]
+ else
+ -- Use our existing logic to handle MarkedString
+ for _, marked_string in ipairs(input) do
+ M.convert_input_to_markdown_lines(marked_string, contents)
+ end
+ end
+ end
+ if contents[1] == '' or contents[1] == nil then
+ return {}
+ end
+ return contents
+end
+
+function M.make_floating_popup_options(width, height, opts)
+ validate {
+ opts = { opts, 't', true };
+ }
+ opts = opts or {}
+ validate {
+ ["opts.offset_x"] = { opts.offset_x, 'n', true };
+ ["opts.offset_y"] = { opts.offset_y, 'n', true };
+ }
+
+ local anchor = ''
+ local row, col
+
+ if vim.fn.winline() <= height then
+ anchor = anchor..'N'
+ row = 1
+ else
+ anchor = anchor..'S'
+ row = 0
+ end
+
+ if vim.fn.wincol() + width <= api.nvim_get_option('columns') then
+ anchor = anchor..'W'
+ col = 0
+ else
+ anchor = anchor..'E'
+ col = 1
+ end
+
+ return {
+ anchor = anchor,
+ col = col + (opts.offset_x or 0),
+ height = height,
+ relative = 'cursor',
+ row = row + (opts.offset_y or 0),
+ style = 'minimal',
+ width = width,
+ }
+end
+
+function M.open_floating_preview(contents, filetype, opts)
+ validate {
+ contents = { contents, 't' };
+ filetype = { filetype, 's', true };
+ opts = { opts, 't', true };
+ }
+
+ -- Trim empty lines from the end.
+ for i = #contents, 1, -1 do
+ if #contents[i] == 0 then
+ table.remove(contents)
+ else
+ break
+ end
+ end
+
+ local width = 0
+ local height = #contents
+ for i, line in ipairs(contents) do
+ -- Clean up the input and add left pad.
+ line = " "..line:gsub("\r", "")
+ -- TODO(ashkan) use nvim_strdisplaywidth if/when that is introduced.
+ local line_width = vim.fn.strdisplaywidth(line)
+ width = math.max(line_width, width)
+ contents[i] = line
+ end
+ -- Add right padding of 1 each.
+ width = width + 1
+
+ local floating_bufnr = api.nvim_create_buf(false, true)
+ if filetype then
+ api.nvim_buf_set_option(floating_bufnr, 'filetype', filetype)
+ end
+ local float_option = M.make_floating_popup_options(width, height, opts)
+ local floating_winnr = api.nvim_open_win(floating_bufnr, false, float_option)
+ if filetype == 'markdown' then
+ api.nvim_win_set_option(floating_winnr, 'conceallevel', 2)
+ end
+ api.nvim_buf_set_lines(floating_bufnr, 0, -1, true, contents)
+ api.nvim_buf_set_option(floating_bufnr, 'modifiable', false)
+ api.nvim_command("autocmd CursorMoved <buffer> ++once lua pcall(vim.api.nvim_win_close, "..floating_winnr..", true)")
+ return floating_bufnr, floating_winnr
+end
+
+local function validate_lsp_position(pos)
+ validate { pos = {pos, 't'} }
+ validate {
+ line = {pos.line, 'n'};
+ character = {pos.character, 'n'};
+ }
+ return true
+end
+
+function M.open_floating_peek_preview(bufnr, start, finish, opts)
+ validate {
+ bufnr = {bufnr, 'n'};
+ start = {start, validate_lsp_position, 'valid start Position'};
+ finish = {finish, validate_lsp_position, 'valid finish Position'};
+ opts = { opts, 't', true };
+ }
+ local width = math.max(finish.character - start.character + 1, 1)
+ local height = math.max(finish.line - start.line + 1, 1)
+ local floating_winnr = api.nvim_open_win(bufnr, false, M.make_floating_popup_options(width, height, opts))
+ api.nvim_win_set_cursor(floating_winnr, {start.line+1, start.character})
+ api.nvim_command("autocmd CursorMoved * ++once lua pcall(vim.api.nvim_win_close, "..floating_winnr..", true)")
+ return floating_winnr
+end
+
+
+local function highlight_range(bufnr, ns, hiname, start, finish)
+ if start[1] == finish[1] then
+ -- TODO care about encoding here since this is in byte index?
+ api.nvim_buf_add_highlight(bufnr, ns, hiname, start[1], start[2], finish[2])
+ else
+ api.nvim_buf_add_highlight(bufnr, ns, hiname, start[1], start[2], -1)
+ for line = start[1] + 1, finish[1] - 1 do
+ api.nvim_buf_add_highlight(bufnr, ns, hiname, line, 0, -1)
+ end
+ api.nvim_buf_add_highlight(bufnr, ns, hiname, finish[1], 0, finish[2])
+ end
+end
+
+do
+ local all_buffer_diagnostics = {}
+
+ local diagnostic_ns = api.nvim_create_namespace("vim_lsp_diagnostics")
+
+ local default_severity_highlight = {
+ [protocol.DiagnosticSeverity.Error] = { guifg = "Red" };
+ [protocol.DiagnosticSeverity.Warning] = { guifg = "Orange" };
+ [protocol.DiagnosticSeverity.Information] = { guifg = "LightBlue" };
+ [protocol.DiagnosticSeverity.Hint] = { guifg = "LightGrey" };
+ }
+
+ local underline_highlight_name = "LspDiagnosticsUnderline"
+ api.nvim_command(string.format("highlight %s gui=underline cterm=underline", underline_highlight_name))
+
+ local function find_color_rgb(color)
+ local rgb_hex = api.nvim_get_color_by_name(color)
+ validate { color = {color, function() return rgb_hex ~= -1 end, "valid color name"} }
+ return rgb_hex
+ end
+
+ --- Determine whether to use black or white text
+ -- Ref: https://stackoverflow.com/a/1855903/837964
+ -- https://stackoverflow.com/questions/596216/formula-to-determine-brightness-of-rgb-color
+ local function color_is_bright(r, g, b)
+ -- Counting the perceptive luminance - human eye favors green color
+ local luminance = (0.299*r + 0.587*g + 0.114*b)/255
+ if luminance > 0.5 then
+ return true -- Bright colors, black font
+ else
+ return false -- Dark colors, white font
+ end
+ end
+
+ local severity_highlights = {}
+
+ function M.set_severity_highlights(highlights)
+ validate {highlights = {highlights, 't'}}
+ for severity, default_color in pairs(default_severity_highlight) do
+ local severity_name = protocol.DiagnosticSeverity[severity]
+ local highlight_name = "LspDiagnostics"..severity_name
+ local hi_info = highlights[severity] or default_color
+ -- Try to fill in the foreground color with a sane default.
+ if not hi_info.guifg and hi_info.guibg then
+ -- TODO(ashkan) move this out when bitop is guaranteed to be included.
+ local bit = require 'bit'
+ local band, rshift = bit.band, bit.rshift
+ local rgb = find_color_rgb(hi_info.guibg)
+ local is_bright = color_is_bright(rshift(rgb, 16), band(rshift(rgb, 8), 0xFF), band(rgb, 0xFF))
+ hi_info.guifg = is_bright and "Black" or "White"
+ end
+ if not hi_info.ctermfg and hi_info.ctermbg then
+ -- TODO(ashkan) move this out when bitop is guaranteed to be included.
+ local bit = require 'bit'
+ local band, rshift = bit.band, bit.rshift
+ local rgb = find_color_rgb(hi_info.ctermbg)
+ local is_bright = color_is_bright(rshift(rgb, 16), band(rshift(rgb, 8), 0xFF), band(rgb, 0xFF))
+ hi_info.ctermfg = is_bright and "Black" or "White"
+ end
+ local cmd_parts = {"highlight", highlight_name}
+ for k, v in pairs(hi_info) do
+ table.insert(cmd_parts, k.."="..v)
+ end
+ api.nvim_command(table.concat(cmd_parts, ' '))
+ severity_highlights[severity] = highlight_name
+ end
+ end
+
+ function M.buf_clear_diagnostics(bufnr)
+ validate { bufnr = {bufnr, 'n', true} }
+ bufnr = bufnr == 0 and api.nvim_get_current_buf() or bufnr
+ api.nvim_buf_clear_namespace(bufnr, diagnostic_ns, 0, -1)
+ end
+
+ -- Initialize with the defaults.
+ M.set_severity_highlights(default_severity_highlight)
+
+ function M.get_severity_highlight_name(severity)
+ return severity_highlights[severity]
+ end
+
+ function M.show_line_diagnostics()
+ local bufnr = api.nvim_get_current_buf()
+ local line = api.nvim_win_get_cursor(0)[1] - 1
+ -- local marks = api.nvim_buf_get_extmarks(bufnr, diagnostic_ns, {line, 0}, {line, -1}, {})
+ -- if #marks == 0 then
+ -- return
+ -- end
+ -- local buffer_diagnostics = all_buffer_diagnostics[bufnr]
+ local lines = {"Diagnostics:"}
+ local highlights = {{0, "Bold"}}
+
+ local buffer_diagnostics = all_buffer_diagnostics[bufnr]
+ if not buffer_diagnostics then return end
+ local line_diagnostics = buffer_diagnostics[line]
+ if not line_diagnostics then return end
+
+ for i, diagnostic in ipairs(line_diagnostics) do
+ -- for i, mark in ipairs(marks) do
+ -- local mark_id = mark[1]
+ -- local diagnostic = buffer_diagnostics[mark_id]
+
+ -- TODO(ashkan) make format configurable?
+ local prefix = string.format("%d. ", i)
+ local hiname = severity_highlights[diagnostic.severity]
+ local message_lines = split_lines(diagnostic.message)
+ table.insert(lines, prefix..message_lines[1])
+ table.insert(highlights, {#prefix + 1, hiname})
+ for j = 2, #message_lines do
+ table.insert(lines, message_lines[j])
+ table.insert(highlights, {0, hiname})
+ end
+ end
+ local popup_bufnr, winnr = M.open_floating_preview(lines, 'plaintext')
+ for i, hi in ipairs(highlights) do
+ local prefixlen, hiname = unpack(hi)
+ -- Start highlight after the prefix
+ api.nvim_buf_add_highlight(popup_bufnr, -1, hiname, i-1, prefixlen, -1)
+ end
+ return popup_bufnr, winnr
+ end
+
+ function M.buf_diagnostics_save_positions(bufnr, diagnostics)
+ validate {
+ bufnr = {bufnr, 'n', true};
+ diagnostics = {diagnostics, 't', true};
+ }
+ if not diagnostics then return end
+ bufnr = bufnr == 0 and api.nvim_get_current_buf() or bufnr
+
+ if not all_buffer_diagnostics[bufnr] then
+ -- Clean up our data when the buffer unloads.
+ api.nvim_buf_attach(bufnr, false, {
+ on_detach = function(b)
+ all_buffer_diagnostics[b] = nil
+ end
+ })
+ end
+ all_buffer_diagnostics[bufnr] = {}
+ local buffer_diagnostics = all_buffer_diagnostics[bufnr]
+
+ for _, diagnostic in ipairs(diagnostics) do
+ local start = diagnostic.range.start
+ -- local mark_id = api.nvim_buf_set_extmark(bufnr, diagnostic_ns, 0, start.line, 0, {})
+ -- buffer_diagnostics[mark_id] = diagnostic
+ local line_diagnostics = buffer_diagnostics[start.line]
+ if not line_diagnostics then
+ line_diagnostics = {}
+ buffer_diagnostics[start.line] = line_diagnostics
+ end
+ table.insert(line_diagnostics, diagnostic)
+ end
+ end
+
+
+ function M.buf_diagnostics_underline(bufnr, diagnostics)
+ for _, diagnostic in ipairs(diagnostics) do
+ local start = diagnostic.range.start
+ local finish = diagnostic.range["end"]
+
+ -- TODO care about encoding here since this is in byte index?
+ highlight_range(bufnr, diagnostic_ns, underline_highlight_name,
+ {start.line, start.character},
+ {finish.line, finish.character}
+ )
+ end
+ end
+
+ function M.buf_diagnostics_virtual_text(bufnr, diagnostics)
+ local buffer_line_diagnostics = all_buffer_diagnostics[bufnr]
+ if not buffer_line_diagnostics then
+ M.buf_diagnostics_save_positions(bufnr, diagnostics)
+ end
+ buffer_line_diagnostics = all_buffer_diagnostics[bufnr]
+ if not buffer_line_diagnostics then
+ return
+ end
+ for line, line_diags in pairs(buffer_line_diagnostics) do
+ local virt_texts = {}
+ for i = 1, #line_diags - 1 do
+ table.insert(virt_texts, {"■", severity_highlights[line_diags[i].severity]})
+ end
+ local last = line_diags[#line_diags]
+ -- TODO(ashkan) use first line instead of subbing 2 spaces?
+ table.insert(virt_texts, {"■ "..last.message:gsub("\r", ""):gsub("\n", " "), severity_highlights[last.severity]})
+ api.nvim_buf_set_virtual_text(bufnr, diagnostic_ns, line, virt_texts, {})
+ end
+ end
+end
+
+function M.buf_loclist(bufnr, locations)
+ local targetwin
+ for _, winnr in ipairs(api.nvim_list_wins()) do
+ local winbuf = api.nvim_win_get_buf(winnr)
+ if winbuf == bufnr then
+ targetwin = winnr
+ break
+ end
+ end
+ if not targetwin then return end
+
+ local items = {}
+ local path = api.nvim_buf_get_name(bufnr)
+ for _, d in ipairs(locations) do
+ -- TODO: URL parsing here?
+ local start = d.range.start
+ table.insert(items, {
+ filename = path,
+ lnum = start.line + 1,
+ col = start.character + 1,
+ text = d.message,
+ })
+ end
+ vim.fn.setloclist(targetwin, items, ' ', 'Language Server')
+end
+
+return M
+-- vim:sw=2 ts=2 et