fix(lsp): refactor escaping snippet text (#25611)

This commit is contained in:
Maria José Solano 2023-10-14 00:06:40 -07:00 committed by GitHub
parent 8ee8112b92
commit ee156ca60e
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 72 additions and 42 deletions

View File

@ -20,11 +20,15 @@ local format_capture = Cg(int / tonumber, 'capture')
local format_modifier = Cg(P('upcase') + P('downcase') + P('capitalize'), 'modifier') local format_modifier = Cg(P('upcase') + P('downcase') + P('capitalize'), 'modifier')
local tabstop = Cg(int / tonumber, 'tabstop') local tabstop = Cg(int / tonumber, 'tabstop')
-- These characters are always escapable in text nodes no matter the context.
local escapable = '$}\\'
--- Returns a function that unescapes occurrences of "special" characters. --- Returns a function that unescapes occurrences of "special" characters.
--- ---
--- @param special string --- @param special? string
--- @return fun(match: string): string --- @return fun(match: string): string
local function escape_text(special) local function escape_text(special)
special = special or escapable
return function(match) return function(match)
local escaped = match:gsub('\\(.)', function(c) local escaped = match:gsub('\\(.)', function(c)
return special:find(c) and c or '\\' .. c return special:find(c) and c or '\\' .. c
@ -33,25 +37,33 @@ local function escape_text(special)
end end
end end
-- Text nodes match "any character", but $, \, and } must be escaped. --- Returns a pattern for text nodes. Will match characters in `escape` when preceded by a backslash,
local escapable = '$}\\' --- and will stop with characters in `stop_with`.
local text = (backslash * S(escapable)) + (P(1) - S(escapable)) ---
local text_0, text_1 = (text ^ 0) / escape_text(escapable), text ^ 1 --- @param escape string
--- @param stop_with? string
--- @return vim.lpeg.Pattern
local function text(escape, stop_with)
stop_with = stop_with or escape
return (backslash * S(escape)) + (P(1) - S(stop_with))
end
-- For text nodes inside curly braces. It stops parsing when reaching an escapable character.
local braced_text = (text(escapable) ^ 0) / escape_text()
-- Within choice nodes, \ also escapes comma and pipe characters. -- Within choice nodes, \ also escapes comma and pipe characters.
local choice_text = C(((backslash * S(escapable .. ',|')) + (P(1) - S(escapable .. ',|'))) ^ 1) local choice_text = C(text(escapable .. ',|') ^ 1) / escape_text(escapable .. ',|')
/ escape_text(escapable .. ',|')
local if_text, else_text = Cg(text_0, 'if_text'), Cg(text_0, 'else_text')
-- Within format nodes, make sure we stop at / -- Within format nodes, make sure we stop at /
local format_text = C(((backslash * S(escapable)) + (P(1) - S(escapable .. '/'))) ^ 1) local format_text = C(text(escapable, escapable .. '/') ^ 1) / escape_text()
/ escape_text(escapable)
local if_text, else_text = Cg(braced_text, 'if_text'), Cg(braced_text, 'else_text')
-- Within ternary condition format nodes, make sure we stop at : -- Within ternary condition format nodes, make sure we stop at :
local if_till_colon_text = Cg( local if_till_colon_text = Cg(C(text(escapable, escapable .. ':') ^ 1) / escape_text(), 'if_text')
C(((backslash * S(escapable)) + (P(1) - S(escapable .. ':'))) ^ 1) / escape_text(escapable),
'if_text'
)
-- Matches the string inside //, allowing escaping of the closing slash. -- Matches the string inside //, allowing escaping of the closing slash.
local regex = Cg(((backslash * slash) + (P(1) - slash)) ^ 1, 'regex') local regex = Cg(text('/') ^ 1, 'regex')
-- Regex constructor flags (see https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/RegExp/RegExp#parameters). -- Regex constructor flags (see https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/RegExp/RegExp#parameters).
local options = Cg(S('dgimsuvy') ^ 0, 'options') local options = Cg(S('dgimsuvy') ^ 0, 'options')
@ -94,11 +106,11 @@ local G = P({
snippet = Ct(Cg( snippet = Ct(Cg(
Ct(( Ct((
V('any') + V('any') +
(Ct(Cg(text_1 / escape_text(escapable), 'text')) / node(Type.Text)) (Ct(Cg((text(escapable, '$') ^ 1) / escape_text(), 'text')) / node(Type.Text))
) ^ 1), 'children' ) ^ 1), 'children'
)) / node(Type.Snippet), ) * -P(1)) / node(Type.Snippet),
any_or_text = V('any') + (Ct(Cg(text_0 / escape_text(escapable), 'text')) / node(Type.Text)),
any = V('placeholder') + V('tabstop') + V('choice') + V('variable'), any = V('placeholder') + V('tabstop') + V('choice') + V('variable'),
any_or_text = V('any') + (Ct(Cg(braced_text, 'text')) / node(Type.Text)),
tabstop = Ct(dollar * (tabstop + (l_brace * tabstop * r_brace))) / node(Type.Tabstop), tabstop = Ct(dollar * (tabstop + (l_brace * tabstop * r_brace))) / node(Type.Tabstop),
placeholder = Ct(dollar * l_brace * tabstop * colon * Cg(V('any_or_text'), 'value') * r_brace) / node(Type.Placeholder), placeholder = Ct(dollar * l_brace * tabstop * colon * Cg(V('any_or_text'), 'value') * r_brace) / node(Type.Placeholder),
choice = Ct(dollar * choice = Ct(dollar *

View File

@ -1,5 +1,6 @@
local helpers = require('test.functional.helpers')(after_each) local helpers = require('test.functional.helpers')(after_each)
local snippet = require('vim.lsp._snippet_grammar') local snippet = require('vim.lsp._snippet_grammar')
local type = snippet.NodeType
local eq = helpers.eq local eq = helpers.eq
local exec_lua = helpers.exec_lua local exec_lua = helpers.exec_lua
@ -15,28 +16,28 @@ describe('vim.lsp._snippet_grammar', function()
it('parses only text', function() it('parses only text', function()
eq({ eq({
{ type = snippet.NodeType.Text, data = { text = 'TE$}XT' } }, { type = type.Text, data = { text = 'TE$}XT' } },
}, parse('TE\\$\\}XT')) }, parse('TE\\$\\}XT'))
end) end)
it('parses tabstops', function() it('parses tabstops', function()
eq({ eq({
{ type = snippet.NodeType.Tabstop, data = { tabstop = 1 } }, { type = type.Tabstop, data = { tabstop = 1 } },
{ type = snippet.NodeType.Tabstop, data = { tabstop = 2 } }, { type = type.Tabstop, data = { tabstop = 2 } },
}, parse('$1${2}')) }, parse('$1${2}'))
end) end)
it('parses nested placeholders', function() it('parses nested placeholders', function()
eq({ eq({
{ {
type = snippet.NodeType.Placeholder, type = type.Placeholder,
data = { data = {
tabstop = 1, tabstop = 1,
value = { value = {
type = snippet.NodeType.Placeholder, type = type.Placeholder,
data = { data = {
tabstop = 2, tabstop = 2,
value = { type = snippet.NodeType.Tabstop, data = { tabstop = 3 } }, value = { type = type.Tabstop, data = { tabstop = 3 } },
}, },
}, },
}, },
@ -46,24 +47,24 @@ describe('vim.lsp._snippet_grammar', function()
it('parses variables', function() it('parses variables', function()
eq({ eq({
{ type = snippet.NodeType.Variable, data = { name = 'VAR' } }, { type = type.Variable, data = { name = 'VAR' } },
{ type = snippet.NodeType.Variable, data = { name = 'VAR' } }, { type = type.Variable, data = { name = 'VAR' } },
{ {
type = snippet.NodeType.Variable, type = type.Variable,
data = { data = {
name = 'VAR', name = 'VAR',
default = { type = snippet.NodeType.Tabstop, data = { tabstop = 1 } }, default = { type = type.Tabstop, data = { tabstop = 1 } },
}, },
}, },
{ {
type = snippet.NodeType.Variable, type = type.Variable,
data = { data = {
name = 'VAR', name = 'VAR',
regex = 'regex', regex = 'regex',
options = '', options = '',
format = { format = {
{ {
type = snippet.NodeType.Format, type = type.Format,
data = { capture = 1, modifier = 'upcase' }, data = { capture = 1, modifier = 'upcase' },
}, },
}, },
@ -75,7 +76,7 @@ describe('vim.lsp._snippet_grammar', function()
it('parses choice', function() it('parses choice', function()
eq({ eq({
{ {
type = snippet.NodeType.Choice, type = type.Choice,
data = { tabstop = 1, values = { ',', '|' } }, data = { tabstop = 1, values = { ',', '|' } },
}, },
}, parse('${1|\\,,\\||}')) }, parse('${1|\\,,\\||}'))
@ -85,30 +86,30 @@ describe('vim.lsp._snippet_grammar', function()
eq( eq(
{ {
{ {
type = snippet.NodeType.Variable, type = type.Variable,
data = { data = {
name = 'VAR', name = 'VAR',
regex = 'regex', regex = 'regex',
options = '', options = '',
format = { format = {
{ {
type = snippet.NodeType.Format, type = type.Format,
data = { capture = 1, modifier = 'upcase' }, data = { capture = 1, modifier = 'upcase' },
}, },
{ {
type = snippet.NodeType.Format, type = type.Format,
data = { capture = 1, if_text = 'if_text' }, data = { capture = 1, if_text = 'if_text' },
}, },
{ {
type = snippet.NodeType.Format, type = type.Format,
data = { capture = 1, else_text = 'else_text' }, data = { capture = 1, else_text = 'else_text' },
}, },
{ {
type = snippet.NodeType.Format, type = type.Format,
data = { capture = 1, if_text = 'if_text', else_text = 'else_text' }, data = { capture = 1, if_text = 'if_text', else_text = 'else_text' },
}, },
{ {
type = snippet.NodeType.Format, type = type.Format,
data = { capture = 1, else_text = 'else_text' }, data = { capture = 1, else_text = 'else_text' },
}, },
}, },
@ -124,24 +125,24 @@ describe('vim.lsp._snippet_grammar', function()
it('parses empty strings', function() it('parses empty strings', function()
eq({ eq({
{ {
type = snippet.NodeType.Placeholder, type = type.Placeholder,
data = { data = {
tabstop = 1, tabstop = 1,
value = { type = snippet.NodeType.Text, data = { text = '' } }, value = { type = type.Text, data = { text = '' } },
}, },
}, },
{ {
type = snippet.NodeType.Text, type = type.Text,
data = { text = ' ' }, data = { text = ' ' },
}, },
{ {
type = snippet.NodeType.Variable, type = type.Variable,
data = { data = {
name = 'VAR', name = 'VAR',
regex = 'erg', regex = 'erg',
format = { format = {
{ {
type = snippet.NodeType.Format, type = type.Format,
data = { capture = 1, if_text = '' }, data = { capture = 1, if_text = '' },
}, },
}, },
@ -150,4 +151,21 @@ describe('vim.lsp._snippet_grammar', function()
}, },
}, parse('${1:} ${VAR/erg/${1:+}/g}')) }, parse('${1:} ${VAR/erg/${1:+}/g}'))
end) end)
it('parses closing curly brace as text', function()
eq(
{
{ type = type.Text, data = { text = 'function ' } },
{ type = type.Tabstop, data = { tabstop = 1 } },
{ type = type.Text, data = { text = '() {\n ' } },
{ type = type.Tabstop, data = { tabstop = 0 } },
{ type = type.Text, data = { text = '\n}' } },
},
parse(table.concat({
'function $1() {',
' $0',
'}',
}, '\n'))
)
end)
end) end)