parser: Validate IPv6, IDN compatibility in email and mentions

lib/linkify/parser.ex

@@ -9,17 +9,11 @@ defmodule Linkify.Parser do
 
   @match_url ~r{^(?:\W*)?(?<url>(?:https?:\/\/)?[\w.-]+(?:\.[\w\.-]+)+[\w\-\._~%:\/?#[\]@!\$&'\(\)\*\+,;=.]+$)}u
 
-  @match_hostname ~r{^\W*(?<scheme>https?:\/\/)?(?:[^@\n]+\\w@)?(?<host>[^:#~\/\n?]+)}u
+  @get_scheme_host ~r{^\W*(?<scheme>https?:\/\/)?(?:[^@\n]+\\w@)?(?<host>[^:#~\/\n?]+)}u
-
-  @match_ip ~r"^(([0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|25[0-5])\.){3}([0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|25[0-5])$"
 
   # @user
   # @user@example.com
   # credo:disable-for-next-line
-  @match_mention ~r/^(?:\W*)?(?<long>@[a-zA-Z\d_-]+@[a-zA-Z0-9_-](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*)|^(?:\W*)?(?<short>@[a-zA-Z\d_-]+)/u
-
-  # https://www.w3.org/TR/html5/forms.html#valid-e-mail-address
-  @match_email ~r"^[a-zA-Z0-9.!#$%&'*+\/=?^_`{|}~-]+@[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*$"u
 
   @match_hashtag ~r/^(?<tag>\#[[:word:]_]*[[:alpha:]_·][[:word:]_·\p{M}]*)/u
 
@@ -63,7 +57,7 @@ defmodule Linkify.Parser do
       ~s{Check out <a href="http://google.com">google.com</a>}
   """
 
-  @types [:url, :email, :hashtag, :extra, :mention]
+  @types [:url, :hashtag, :extra, :mention, :email]
 
   def parse(input, opts \\ %{})
   def parse(input, opts) when is_binary(input), do: {input, %{}} |> parse(opts) |> elem(0)
@@ -224,7 +218,11 @@ defmodule Linkify.Parser do
   end
 
   def email?(buffer, opts) do
-    valid_url?(buffer) && Regex.match?(@match_email, buffer) && valid_tld?(buffer, opts)
+    # Note: In reality the local part can only be checked by the remote server
+    case Regex.run(~r/^(?<user>.*)@(?<host>[^@]+)$/, buffer, capture: [:user, :host]) do
+      [_user, hostname] -> valid_hostname?(hostname) && valid_tld?(hostname, opts)
+      _ -> false
+    end
   end
 
   defp valid_url?(url), do: !Regex.match?(@invalid_url, url)
@@ -237,7 +235,7 @@ defmodule Linkify.Parser do
   Will skip validation and return `true` if `:validate_tld` set to `:no_scheme` and the url has a scheme.
   """
   def valid_tld?(url, opts) do
-    [scheme, host] = Regex.run(@match_hostname, url, capture: [:scheme, :host])
+    [scheme, host] = Regex.run(@get_scheme_host, url, capture: [:scheme, :host])
 
     cond do
       opts[:validate_tld] == false ->
@@ -256,13 +254,58 @@ defmodule Linkify.Parser do
     end
   end
 
-  def ip?(buffer), do: Regex.match?(@match_ip, buffer)
+  def safe_to_integer(string, base \\ 10) do
+    String.to_integer(string, base)
+  rescue
+    _ ->
+      nil
+  end
+
+  def ip?(buffer) do
+    v4 = String.split(buffer, ".")
+
+    v6 =
+      buffer
+      |> String.trim_leading("[")
+      |> String.trim_trailing("]")
+      |> String.split(":", trim: true)
+
+    cond do
+      length(v4) == 4 ->
+        !Enum.any?(v4, fn x -> safe_to_integer(x, 10) not in 0..255 end)
+
+      length(v6) in 1..8 ->
+        !Enum.any?(v4, fn x -> safe_to_integer(x, 16) not in 0..0xFFFF end)
+
+      false ->
+        false
+    end
+  end
+
+  # IDN-compatible, ported from musl-libc's is_valid_hostname()
+  def valid_hostname?(hostname) do
+    hostname
+    |> String.to_charlist()
+    |> Enum.any?(fn s ->
+      !(s >= 0x80 || s in 0x30..0x39 || s in 0x41..0x5A || s in 0x61..0x7A || s in '.-')
+    end)
+    |> Kernel.!()
+  end
 
   def match_mention(buffer) do
-    case Regex.run(@match_mention, buffer, capture: [:long, :short]) do
+    case Regex.run(~r/^@(?<user>[a-zA-Z\d_-]+)(@(?<host>[^@]+))?$/, buffer,
-      [mention, ""] -> mention
+           capture: [:user, :host]
-      ["", mention] -> mention
+         ) do
-      _ -> nil
+      [user, ""] ->
+        "@" <> user
+
+      [user, hostname] ->
+        if valid_hostname?(hostname) && valid_tld?(hostname, []),
+          do: "@" <> user <> "@" <> hostname,
+          else: nil
+
+      _ ->
+        nil
     end
   end
 

test/linkify_test.exs

@@ -244,7 +244,7 @@ defmodule LinkifyTest do
       end
 
       expected =
-        ~s(Hello again, <span class="h-card"><a href="#/user/user">@<span>@user</span></a></span>.&lt;script&gt;&lt;/script&gt;\nThis is on another :moominmamma: line. <a href="/tag/2hu" target="_blank">#2hu</a> <a href="/tag/epic" target="_blank">#epic</a> <a href="/tag/phantasmagoric" target="_blank">#phantasmagoric</a>)
+        ~s(Hello again, @user.&lt;script&gt;&lt;/script&gt;\nThis is on another :moominmamma: line. <a href="/tag/2hu" target="_blank">#2hu</a> <a href="/tag/epic" target="_blank">#epic</a> <a href="/tag/phantasmagoric" target="_blank">#phantasmagoric</a>)
 
       assert Linkify.link(text,
                mention: true,
@@ -377,22 +377,14 @@ defmodule LinkifyTest do
 
       text = "That's @user@example.com's server"
 
-      expected =
+      assert Linkify.link(text, mention: true, mention_prefix: "https://example.com/user/") ==
-        "That's <a href=\"https://example.com/user/user@example.com\">@user@example.com</a>'s server"
+               text
-
-      assert Linkify.link(text,
-               mention: true,
-               mention_prefix: "https://example.com/user/"
-             ) == expected
     end
 
-    test "mentions with symbols before them" do
+    test "mentions with no word-separation before them" do
-      text = "@@example hey! >@@test@example.com"
+      text = "@@example hey! >@@test@example.com idolm@ster"
 
-      expected =
+      assert Linkify.link(text, mention: true, mention_prefix: "/users/") == text
-        "@<a href=\"/users/example\">@example</a> hey! >@<a href=\"/users/test@example.com\">@test@example.com</a>"
-
-      assert Linkify.link(text, mention: true, mention_prefix: "/users/") == expected
     end
 
     test "invalid mentions" do

test/parser_test.exs

@@ -305,6 +305,6 @@ defmodule Linkify.ParserTest do
     ]
 
   def valid_emails, do: ["rms@ai.mit.edu", "vc@cock.li", "guardian@33y6fjyhs3phzfjj.onion"]
-  def invalid_emails, do: ["rms[at]ai.mit.edu", "vc@cock", "xmpp:lain@trashserver.net"]
+  def invalid_emails, do: ["rms[at]ai.mit.edu", "vc@cock"]
   def valid_custom_tld_emails, do: ["hi@company.null"]
 end