Merge pull request #1284 from bettio/fix-elixir-inspect

Fix Elixir `inspect()`

Closer to Elixir one (and more reliable), but not yet perfect.
e.g. structs are still printed as maps.

These changes are made under both the "Apache 2.0" and the "GNU Lesser General
Public License 2.1 or later" license terms (dual license).

SPDX-License-Identifier: Apache-2.0 OR LGPL-2.1-or-later

Author: bettio

CHANGELOG.md

@@ -38,6 +38,8 @@ instead
 - Fix memory corruption in `unicode:characters_to_binary`
 - Fix handling of large literal indexes
 - `unicode:characters_to_list`: fixed bogus out_of_memory error on some platforms such as ESP32
+- Fix crash in Elixir library when doing `inspect(:atom)`
+- General inspect() compliance with Elixir behavior (but there are still some minor differences)
 
 ## [0.6.4] - 2024-08-18
 

libs/exavmlib/lib/Kernel.ex

@@ -42,14 +42,19 @@ defmodule Kernel do
   def inspect(term, opts \\ []) when is_list(opts) do
     case term do
       t when is_atom(t) ->
-        [?:, atom_to_string(t)]
+        atom_to_string(t, ":")
 
       t when is_integer(t) ->
         :erlang.integer_to_binary(t)
 
       t when is_list(t) ->
-        # TODO: escape unprintable lists
-        :erlang.list_to_binary(t)
+        if is_printable_list(t) do
+          str = :erlang.list_to_binary(t)
+          <<?'::utf8, str::binary, ?'::utf8>>
+        else
+          [?[ | t |> inspect_join(?])]
+          |> :erlang.list_to_binary()
+        end
 
       t when is_pid(t) ->
         :erlang.pid_to_list(t)
@@ -64,15 +69,19 @@ defmodule Kernel do
         |> :erlang.list_to_binary()
 
       t when is_binary(t) ->
-        # TODO: escape unprintable binaries
-        t
+        if is_printable_binary(t) do
+          <<?"::utf8, t::binary, ?"::utf8>>
+        else
+          ["<<" | t |> :erlang.binary_to_list() |> inspect_join(">>")]
+          |> :erlang.list_to_binary()
+        end
 
       t when is_reference(t) ->
         :erlang.ref_to_list(t)
         |> :erlang.list_to_binary()
 
       t when is_float(t) ->
-        :erlang.float_to_binary(t)
+        :erlang.float_to_binary(term, [{:decimals, 17}, :compact])
 
       t when is_map(t) ->
         [?%, ?{ | t |> inspect_kv() |> join(?})]
@@ -88,6 +97,10 @@ defmodule Kernel do
     [inspect(e), last]
   end
 
+  defp inspect_join([h | e], last) when not is_list(e) do
+    [inspect(h), " | ", inspect(e), last]
+  end
+
   defp inspect_join([h | t], last) do
     [inspect(h), ?,, ?\s | inspect_join(t, last)]
   end
@@ -118,12 +131,56 @@ defmodule Kernel do
     )
   end
 
-  defp atom_to_string(atom) do
-    # TODO: use unicode rather than plain latin1
-    # handle spaces and special characters
-    :erlang.atom_to_binary(atom, :latin1)
+  defp atom_to_string(atom, prefix \\ "") do
+    case atom do
+      true ->
+        "true"
+
+      false ->
+        "false"
+
+      nil ->
+        "nil"
+
+      any_atom ->
+        case :erlang.atom_to_binary(any_atom) do
+          <<"Elixir.", displayable::binary>> ->
+            displayable
+
+          other ->
+            <<prefix::binary, other::binary>>
+        end
+    end
   end
 
+  defp is_printable_list([]), do: false
+
+  defp is_printable_list([char]) do
+    is_printable_ascii(char)
+  end
+
+  defp is_printable_list([char | t]) do
+    if is_printable_ascii(char) do
+      is_printable_list(t)
+    else
+      false
+    end
+  end
+
+  defp is_printable_list(_any), do: false
+
+  defp is_printable_ascii(char) do
+    is_integer(char) and char >= 32 and char < 127 and char != ?'
+  end
+
+  defp is_printable_binary(<<>>), do: true
+
+  defp is_printable_binary(<<char::utf8, rest::binary>>) when char >= 32 do
+    is_printable_binary(rest)
+  end
+
+  defp is_printable_binary(_any), do: false
+
   @doc """
   Returns the biggest of the two given terms according to
   Erlang's term ordering.

tests/libs/exavmlib/Tests.ex

@@ -19,13 +19,19 @@
 #
 
 defmodule Tests do
+  #  defstruct [
+  #    :field1,
+  #    field2: 42
+  #  ]
+
   @compile {:no_warn_undefined, :undef}
 
   def start() do
     :ok = IO.puts("Running Elixir tests")
     :ok = test_enum()
     :ok = test_exception()
     :ok = test_chars_protocol()
+    :ok = test_inspect()
     :ok = IO.puts("Finished Elixir tests")
   end
 
@@ -234,7 +240,62 @@ defmodule Tests do
     :ok
   end
 
+  def test_inspect() do
+    "true" = inspect(true)
+    "false" = inspect(false)
+    "nil" = inspect(nil)
+
+    ":test" = inspect(:test)
+    ":アトム" = inspect(:アトム)
+    "Test" = inspect(Test)
+
+    "5" = inspect(5)
+    "5.0" = inspect(5.0)
+
+    ~s[""] = inspect("")
+    ~s["hello"] = inspect("hello")
+    ~s["アトム"] = inspect("アトム")
+
+    "<<10>>" = inspect("\n")
+    "<<0, 1, 2, 3>>" = inspect(<<0, 1, 2, 3>>)
+    "<<195, 168, 0>>" = inspect(<<195, 168, 0>>)
+
+    "[]" = inspect([])
+    "[0]" = inspect([0])
+    "[9, 10]" = inspect([9, 10])
+    ~s'["test"]' = inspect(["test"])
+    "'hello'" = inspect('hello')
+    "[127]" = inspect([127])
+    "[104, 101, 108, 108, 248]" = inspect('hellø')
+
+    ~s([5 | "hello"]) = inspect([5 | "hello"])
+
+    "{}" = inspect({})
+    "{1, 2}" = inspect({1, 2})
+    "{:test, 1}" = inspect({:test, 1})
+
+    "%{}" = inspect(%{})
+    either("%{a: 1, b: 2}", "%{b: 2, a: 1}", inspect(%{a: 1, b: 2}))
+    either(~s[%{"a" => 1, "b" => 2}], ~s[%{"b" => 2, "a" => 1}], inspect(%{"a" => 1, "b" => 2}))
+
+    # TODO: structs are not yet supported
+    # either(
+    #   ~s[%#{__MODULE__}{field1: nil, field2: 42}],
+    #   ~s[%#{__MODULE__}{field2: 42, field1: nil}],
+    #   inspect(%__MODULE__{})
+    # )
+
+    :ok
+  end
+
   defp fact(n) when n < 0, do: :test
   defp fact(0), do: 1
   defp fact(n), do: fact(n - 1) * n
+
+  def either(a, b, value) do
+    case value do
+      ^a -> a
+      ^b -> b
+    end
+  end
 end