Fix regex matching prefix

extend character classes to cover unicode letters
correctly escape all chars
add missing unicode regex modifiers
fixes #221

lib/elixir_sense.ex

@@ -535,7 +535,7 @@ defmodule ElixirSense do
 
     # Fix incomplete kw key, e.g. cursor after `option1: 1, opt`
     fix_incomplete_kw_key = fn text_before, text_after ->
-      if Regex.match?(~r/\,\s*([\p{L}_][\p{L}\p{N}_@]*[?!]?)?$/, text_before) do
+      if Regex.match?(~r/\,\s*([\p{L}_][\p{L}\p{N}_@]*[?!]?)?$/u, text_before) do
         text_before <> ": :__fake_value__" <> text_after
       end
     end

lib/elixir_sense/core/metadata_builder.ex

@@ -778,7 +778,7 @@ defmodule ElixirSense.Core.MetadataBuilder do
   defp pre({:@, meta_attr, [{name, meta, params}]}, state) when is_atom(name) do
     name_string = Atom.to_string(name)
 
-    if String.match?(name_string, ~r/^[_\p{Ll}\p{Lo}][\p{L}\p{N}_]*[?!]?$/) and
+    if String.match?(name_string, ~r/^[_\p{Ll}\p{Lo}][\p{L}\p{N}_]*[?!]?$/u) and
          not String.starts_with?(name_string, "__atom_elixir_marker_") do
       line = Keyword.fetch!(meta_attr, :line)
       column = Keyword.fetch!(meta_attr, :column)

lib/elixir_sense/core/parser.ex

@@ -246,7 +246,7 @@ defmodule ElixirSense.Core.Parser do
         |> Enum.join("\n")
 
       _ ->
-        if Regex.match?(~r/^[\p{L}_][\p{L}\p{N}_@]*[?!]?$/, token) do
+        if Regex.match?(~r/^[\p{L}_][\p{L}\p{N}_@]*[?!]?$/u, token) do
           remove_line(source, line)
         else
           replace_line_with_marker(source, line)

lib/elixir_sense/core/source.ex

@@ -111,7 +111,7 @@ defmodule ElixirSense.Core.Source do
 
     line_str = line |> String.slice(0, col - 1)
 
-    case Regex.run(~r/[\w0-9\._!\?\:@&\^~+<>=*\/|\\]+$/, line_str) do
+    case Regex.run(~r/[\p{L}\p{N}\.\_\!\?\:\@\&\^\~\+\-\<\>\=\*\/\|\\]+$/u, line_str) do
       nil -> ""
       [prefix] when is_binary(prefix) -> prefix
     end

lib/elixir_sense/plugins/ecto.ex

@@ -80,7 +80,7 @@ defmodule ElixirSense.Plugins.Ecto do
            end),
          assoc_code <- Source.text_after(text_before, line, col),
          [_, var] <-
-           Regex.run(~r/^assoc\(\s*([_\p{Ll}\p{Lo}][\p{L}\p{N}_]*[?!]?)\s*,/, assoc_code),
+           Regex.run(~r/^assoc\(\s*([_\p{Ll}\p{Lo}][\p{L}\p{N}_]*[?!]?)\s*,/u, assoc_code),
          %{^var => %{type: type}} <- Query.extract_bindings(text_before, from_info, env, meta),
          true <- function_exported?(type, :__schema__, 1) do
       {:override, Query.find_assoc_suggestions(type, hint)}

lib/elixir_sense/plugins/ecto/query.ex

@@ -215,15 +215,15 @@ defmodule ElixirSense.Plugins.Ecto.Query do
   def extract_bindings(prefix, %{pos: {{line, col}, _}} = func_info, env, buffer_metadata) do
     func_code = Source.text_after(prefix, line, col)
 
-    from_matches = Regex.scan(~r/^.+\(?\s*(#{@binding_r})/, func_code)
+    from_matches = Regex.scan(~r/^.+\(?\s*(#{@binding_r})/u, func_code)
 
     # TODO this code is broken
     # depends on join positions that we are unable to get from AST
     # line and col was previously assigned to each option in Source.which_func
     join_matches =
       for join when join in @joins <- func_info.options_so_far,
           code = Source.text_after(prefix, line, col),
-          match <- Regex.scan(~r/^#{join}\:\s*(#{@binding_r})/, code) do
+          match <- Regex.scan(~r/^#{join}\:\s*(#{@binding_r})/u, code) do
         match
       end
 

lib/elixir_sense/providers/suggestion/reducers/callbacks.ex

@@ -75,7 +75,7 @@ defmodule ElixirSense.Providers.Suggestion.Reducers.Callbacks do
     list = Enum.sort(list)
 
     cond do
-      Regex.match?(~r/\s(def|defmacro)\s+([_\p{Ll}\p{Lo}][\p{L}\p{N}_]*[?!]?)?$/, text_before) ->
+      Regex.match?(~r/\s(def|defmacro)\s+([_\p{Ll}\p{Lo}][\p{L}\p{N}_]*[?!]?)?$/u, text_before) ->
         {:halt, %{acc | result: list}}
 
       match?({_f, _a}, scope) ->

test/elixir_sense/suggestions_test.exs

@@ -213,6 +213,22 @@ defmodule ElixirSense.SuggestionsTest do
     assert list |> Enum.any?(&(&1.type == :function))
   end
 
+  test "functions from unicode module" do
+    buffer = """
+    defmodule :你好 do
+      def 运行 do
+        IO.puts("你好")
+      end
+    end
+
+    :你好.
+    """
+
+    list = ElixirSense.suggestions(buffer, 7, 5)
+
+    assert list |> Enum.any?(&(&1.type == :function && &1.name == "运行"))
+  end
+
   test "with an alias" do
     buffer = """
     defmodule MyModule do