Add support for :all protocols

.vscode/settings.json

@@ -3,4 +3,7 @@
   "[ruby]": {
     "editor.defaultFormatter": "Shopify.ruby-lsp"
   },
+  "[markdown]": {
+    "editor.defaultFormatter": "esbenp.prettier-vscode"
+  },
 }

README.md

@@ -76,7 +76,7 @@ attributes: {
 
 # URL handling protocols to allow in specific attributes. By default, no
 # protocols are allowed. Use :relative in place of a protocol if you want
-# to allow relative URLs sans protocol.
+# to allow relative URLs sans protocol. Set to `:all` to allow any protocol.
 protocols: {
     "a" => { "href" => ["http", "https", "mailto", :relative] },
     "img" => { "href" => ["http", "https"] },

ext/selma/src/sanitizer.rs

@@ -211,20 +211,23 @@ impl SelmaSanitizer {
                     }
                     Some(protocol_list) => protocol_list.push(allowed_protocol.to_string()),
                 }
-            } else if allowed_protocol.is_kind_of(class::symbol())
-                && allowed_protocol.inspect() == ":relative"
-            {
-                match protocol_list {
-                    None => {
-                        protocol_sanitizers.insert(
-                            attr_name.to_string(),
-                            vec!["#".to_string(), "/".to_string()],
-                        );
-                    }
-                    Some(protocol_list) => {
-                        protocol_list.push("#".to_string());
-                        protocol_list.push("/".to_string());
+            } else if allowed_protocol.is_kind_of(class::symbol()) {
+                let protocol_config = allowed_protocol.inspect();
+                if protocol_config == ":relative" {
+                    match protocol_list {
+                        None => {
+                            protocol_sanitizers.insert(
+                                attr_name.to_string(),
+                                vec!["#".to_string(), "/".to_string()],
+                            );
+                        }
+                        Some(protocol_list) => {
+                            protocol_list.push("#".to_string());
+                            protocol_list.push("/".to_string());
+                        }
                     }
+                } else if protocol_config == ":all" {
+                    protocol_sanitizers.insert(attr_name.to_string(), vec!["all".to_string()]);
                 }
             }
         }
@@ -388,6 +391,10 @@ impl SelmaSanitizer {
     }
 
     fn has_allowed_protocol(protocols_allowed: &[String], attr_val: &String) -> bool {
+        if protocols_allowed.contains(&"all".to_string()) {
+            return true;
+        }
+
         // FIXME: is there a more idiomatic way to do this?
         let mut pos: usize = 0;
         let mut chars = attr_val.chars();

lib/selma/sanitizer.rb

@@ -66,7 +66,12 @@ def allow_class(element, *klass)
     end
 
     def allow_protocol(element, attr, protos)
-      protos = [protos] unless protos.is_a?(Array)
+      if protos.is_a?(Array)
+        raise ArgumentError, "`:all` must be passed outside of an array" if protos.include?(:all)
+      else
+        protos = [protos]
+      end
+
       set_allowed_protocols(element, attr, protos)
     end
 

lib/selma/sanitizer/config/default.rb

@@ -28,7 +28,7 @@ module Config
 
         # URL handling protocols to allow in specific attributes. By default, no
         # protocols are allowed. Use :relative in place of a protocol if you want
-        # to allow relative URLs sans protocol.
+        # to allow relative URLs sans protocol. Set to `:all` to allow any protocol.
         protocols: {},
 
         # An Array of element names whose contents will be removed. The contents

lib/selma/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Selma
-  VERSION = "0.2.2"
+  VERSION = "0.3.0"
 end

test/selma_maliciousness_test.rb

@@ -154,4 +154,17 @@ def test_that_it_raises_on_handle_text_returning_non_string
       Selma::Rewriter.new(sanitizer: nil, handlers: [GarbageTextOptions.new]).rewrite(frag)
     end
   end
+
+  def test_sanitizer_expects_all_as_symbol
+    html = "<a href='https://google.com'>wow!</a>"
+    sanitizer = Selma::Sanitizer.new({
+      elements: ["a"],
+      attributes: { "a" => ["href"] },
+      protocols: { "a" => { "href" => [:all] } },
+    })
+
+    assert_raises(ArgumentError) do
+      Selma::Rewriter.new(sanitizer: sanitizer).rewrite(html)
+    end
+  end
 end

test/selma_sanitizer_elements_test.rb

@@ -252,6 +252,24 @@ def test_should_allow_relative_urls_containing_colons_when_the_colon_is_part_of_
           assert_equal("<a>Footnote 1</a>", Selma::Rewriter.new(sanitizer: sanitizer).rewrite(input))
         end
 
+        def test_should_allow_all_protocols_if_asked
+          input = <<~HTML
+            <a href="/foo/bar">Link</a>
+            <a href="http://wow.com/foo/bar">Link</a>
+            <a href="https://wow.com/foo/bar">Link</a>
+            <a href="ftp://wow.com/foo/bar">Link</a>
+            <a href="ssh://127.0.0.1">Link</a>
+          HTML
+
+          sanitizer = Selma::Sanitizer.new({
+            elements: ["a"],
+            attributes: { "a" => ["href"] },
+            protocols: { "a" => { "href" => :all } },
+          })
+
+          assert_equal(input, Selma::Rewriter.new(sanitizer: sanitizer).rewrite(input))
+        end
+
         def test_should_remove_the_contents_of_filtered_nodes_when_remove_contents_is_true
           sanitizer = Selma::Sanitizer.new({ remove_contents: true })