From 6ca7dc69effddeb63a8fb8f759e29ff8649907ec Mon Sep 17 00:00:00 2001
From: Jean Boussier <jean.boussier@gmail.com>
Date: Mon, 22 Jul 2019 12:02:23 +0200
Subject: [ruby/psych] Deduplicate hash keys if they're strings

https://github.com/ruby/psych/commit/0414982ffd
---
 ext/psych/lib/psych/visitors/to_ruby.rb | 24 +++++++++++++++++++++++-
 test/psych/test_hash.rb                 | 14 ++++++++++++++
 2 files changed, 37 insertions(+), 1 deletion(-)

diff --git a/ext/psych/lib/psych/visitors/to_ruby.rb b/ext/psych/lib/psych/visitors/to_ruby.rb
index c265acb819..49447e124a 100644
--- a/ext/psych/lib/psych/visitors/to_ruby.rb
+++ b/ext/psych/lib/psych/visitors/to_ruby.rb
@@ -336,7 +336,7 @@ module Psych
       SHOVEL = '<<'
       def revive_hash hash, o
         o.children.each_slice(2) { |k,v|
-          key = accept(k)
+          key = deduplicate(accept(k))
           val = accept(v)
 
           if key == SHOVEL && k.tag != "tag:yaml.org,2002:str"
@@ -368,6 +368,28 @@ module Psych
         hash
       end
 
+      if String.method_defined?(:-@)
+        def deduplicate key
+          if key.is_a?(String)
+            # It is important to untaint the string, otherwise it won't
+            # be deduplicated into and fstring, but simply frozen.
+            -(key.untaint)
+          else
+            key
+          end
+        end
+      else
+        def deduplicate key
+          if key.is_a?(String)
+            # Deduplication is not supported by this implementation,
+            # but we emulate it's side effects
+            key.untaint.freeze
+          else
+            key
+          end
+        end
+      end
+
       def merge_key hash, key, val
       end
 
diff --git a/test/psych/test_hash.rb b/test/psych/test_hash.rb
index 2a563da1da..ba11b827da 100644
--- a/test/psych/test_hash.rb
+++ b/test/psych/test_hash.rb
@@ -111,5 +111,19 @@ bar:
 eoyml
       assert_equal({"foo"=>{"hello"=>"world"}, "bar"=>{"hello"=>"world"}}, hash)
     end
+
+    def test_key_deduplication
+      unless String.method_defined?(:-@) && (-("a" * 20)).equal?((-("a" * 20)))
+        skip "This Ruby implementation doesn't support string deduplication"
+      end
+
+      hashes = Psych.load(<<-eoyml)
+---
+- unique_identifier: 1
+- unique_identifier: 2
+eoyml
+
+      assert_same hashes[0].keys.first, hashes[1].keys.first
+    end
   end
 end
-- 
cgit v1.2.3