[ruby/prism] Add a custom builder class for the parser translator

I want to add new node types to the parser translator, for example `itblock`. The bulk of the work is already done by prism itself. In the `parser`
builder, this would be a 5-line change at most but we don't control that here.

Instead, we can add our own builder and either overwrite the few methods we need,
or just inline the complete builder. I'm not sure yet which would be better.

`rubocop-ast` uses its own builder for `parser`. For this to correctly work, it must explicitly choose to extend the
prism builder and use it, same as it currently chooses to use a different parser when prism is used.

I'd like to enforce that the builder for prism extends its custom one since it will lead to
some pretty weird issues otherwise. But first, I'd like to change `rubocop-ast` to make use of this.

ruby/prism@b080e608a8

lib/prism/prism.gemspec

@@ -97,6 +97,7 @@ Gem::Specification.new do |spec|
     "lib/prism/translation/parser33.rb",
     "lib/prism/translation/parser34.rb",
     "lib/prism/translation/parser35.rb",
+    "lib/prism/translation/parser/builder.rb",
     "lib/prism/translation/parser/compiler.rb",
     "lib/prism/translation/parser/lexer.rb",
     "lib/prism/translation/ripper.rb",

lib/prism/translation/parser.rb

@@ -31,6 +31,11 @@ def initialize(message, level, reason, location)
         end
       end
 
+      # Create the parser with our custom builder class
+      def initialize(builder = Parser::Builder.new)
+        super
+      end
+
       Racc_debug_parser = false # :nodoc:
 
       # By using the `:parser` keyword argument, you can translate in a way that is compatible with
@@ -342,6 +347,7 @@ def convert_for_prism(version)
         end
       end
 
+      require_relative "parser/builder"
       require_relative "parser/compiler"
       require_relative "parser/lexer"
 

lib/prism/translation/parser/builder.rb

@@ -0,0 +1,13 @@
+# frozen_string_literal: true
+
+module Prism
+  module Translation
+    class Parser
+      # A builder that knows how to convert more modern Ruby syntax
+      # into whitequark/parser gem's syntax tree.
+      class Builder < ::Parser::Builders::Default
+
+      end
+    end
+  end
+end

test/prism/ruby/parser_test.rb

@@ -16,6 +16,7 @@
 
 # First, opt in to every AST feature.
 Parser::Builders::Default.modernize
+Prism::Translation::Parser::Builder.modernize
 
 # The parser gem rejects some strings that would most likely lead to errors
 # in consumers due to encoding problems. RuboCop however monkey-patches this