[ruby/prism] Frozen strings in the AST

kddnewton · matzbot · commit 14b9098459b8 · 2025-01-12T18:41:42.000Z
ruby/prism@8d9d429155
diff --git a/lib/prism/translation/parser/compiler.rb b/lib/prism/translation/parser/compiler.rb
@@ -2049,7 +2049,7 @@ def visit_heredoc(node)
                 location = appendee.loc
                 location = location.with_expression(location.expression.join(child.loc.expression))
 
-                children[-1] = appendee.updated(:str, [appendee.children.first << child.children.first], location: location)
+                children[-1] = appendee.updated(:str, ["#{appendee.children.first}#{child.children.first}"], location: location)
               else
                 children << child
               end
diff --git a/lib/prism/translation/ruby_parser.rb b/lib/prism/translation/ruby_parser.rb
@@ -929,9 +929,9 @@ def visit_interpolated_x_string_node(node)
               if result == :space
                 # continue
               elsif result.is_a?(String)
-                results[0] << result
+                results[0] = "#{results[0]}#{result}"
               elsif result.is_a?(Array) && result[0] == :str
-                results[0] << result[1]
+                results[0] = "#{results[0]}#{result[1]}"
               else
                 results << result
                 state = :interpolated_content
@@ -940,7 +940,7 @@ def visit_interpolated_x_string_node(node)
               if result == :space
                 # continue
               elsif visited[index - 1] != :space && result.is_a?(Array) && result[0] == :str && results[-1][0] == :str && (results[-1].line_max == result.line)
-                results[-1][1] << result[1]
+                results[-1][1] = "#{results[-1][1]}#{result[1]}"
                 results[-1].line_max = result.line_max
               else
                 results << result
@@ -1440,6 +1440,7 @@ def visit_string_node(node)
           unescaped = node.unescaped
 
           if node.forced_binary_encoding?
+            unescaped = unescaped.dup
             unescaped.force_encoding(Encoding::BINARY)
           end
 
diff --git a/prism/templates/ext/prism/api_node.c.erb b/prism/templates/ext/prism/api_node.c.erb
@@ -34,7 +34,7 @@ pm_token_new(const pm_parser_t *parser, const pm_token_t *token, rb_encoding *en
 
 static VALUE
 pm_string_new(const pm_string_t *string, rb_encoding *encoding) {
-    return rb_enc_str_new((const char *) pm_string_source(string), pm_string_length(string), encoding);
+    return rb_obj_freeze(rb_enc_str_new((const char *) pm_string_source(string), pm_string_length(string), encoding));
 }
 
 VALUE
diff --git a/prism/templates/lib/prism/serialize.rb.erb b/prism/templates/lib/prism/serialize.rb.erb
@@ -241,14 +241,14 @@ module Prism
       end
 
       def load_embedded_string
-        io.read(load_varuint).force_encoding(encoding)
+        io.read(load_varuint).force_encoding(encoding).freeze
       end
 
       def load_string
         type = io.getbyte
         case type
         when 1
-          input.byteslice(load_varuint, load_varuint).force_encoding(encoding)
+          input.byteslice(load_varuint, load_varuint).force_encoding(encoding).freeze
         when 2
           load_embedded_string
         else
diff --git a/test/prism/encoding/regular_expression_encoding_test.rb b/test/prism/encoding/regular_expression_encoding_test.rb
@@ -119,8 +119,8 @@ def assert_regular_expression_encoding_flags(encoding, regexps)
         if expected.is_a?(Array) && actual.is_a?(Array)
           if expected.last.start_with?("/.../n has a non escaped non ASCII character in non ASCII-8BIT script:") &&
               actual.last.start_with?("/.../n has a non escaped non ASCII character in non ASCII-8BIT script:")
-              expected.last.clear
-              actual.last.clear
+            expected.pop
+            actual.pop
           end
         end
 
diff --git a/test/prism/result/string_test.rb b/test/prism/result/string_test.rb
@@ -0,0 +1,32 @@
+# frozen_string_literal: true
+
+require_relative "../test_helper"
+
+module Prism
+  class StringTest < TestCase
+    def test_regular_expression_node_unescaped_frozen
+      node = Prism.parse_statement("/foo/")
+      assert_predicate node.unescaped, :frozen?
+    end
+
+    def test_source_file_node_filepath_frozen
+      node = Prism.parse_statement("__FILE__")
+      assert_predicate node.filepath, :frozen?
+    end
+
+    def test_string_node_unescaped_frozen
+      node = Prism.parse_statement('"foo"')
+      assert_predicate node.unescaped, :frozen?
+    end
+
+    def test_symbol_node_unescaped_frozen
+      node = Prism.parse_statement(":foo")
+      assert_predicate node.unescaped, :frozen?
+    end
+
+    def test_xstring_node_unescaped_frozen
+      node = Prism.parse_statement("`foo`")
+      assert_predicate node.unescaped, :frozen?
+    end
+  end
+end

Original file line number	Diff line number	Diff line change
`@@ -34,7 +34,7 @@ pm_token_new(const pm_parser_t parser, const pm_token_t token, rb_encoding *en`
`34`	`34`
`35`	`35`	`static VALUE`
`36`	`36`	`pm_string_new(const pm_string_t string, rb_encoding encoding) {`
`37`		`- return rb_enc_str_new((const char *) pm_string_source(string), pm_string_length(string), encoding);`
	`37`	`+ return rb_obj_freeze(rb_enc_str_new((const char *) pm_string_source(string), pm_string_length(string), encoding));`
`38`	`38`	`}`
`39`	`39`
`40`	`40`	`VALUE`