From 8b98f92c7f1883989f9e54739c551872a503dde3 Mon Sep 17 00:00:00 2001 From: Gaurav Chande Date: Wed, 3 Jun 2015 19:40:49 +0000 Subject: [PATCH] Extract tokenize logic from Template to a RubyTokenizer --- lib/liquid.rb | 1 + lib/liquid/template.rb | 22 +----------------- lib/liquid/tokenizer.rb | 38 ++++++++++++++++++++++++++++++++ test/unit/tokenizer_unit_test.rb | 11 ++++----- 4 files changed, 46 insertions(+), 26 deletions(-) create mode 100644 lib/liquid/tokenizer.rb diff --git a/lib/liquid.rb b/lib/liquid.rb index 9724aac9..14afa757 100644 --- a/lib/liquid.rb +++ b/lib/liquid.rb @@ -68,6 +68,7 @@ require 'liquid/template' require 'liquid/standardfilters' require 'liquid/condition' require 'liquid/utils' +require 'liquid/tokenizer' require 'liquid/token' # Load all the tags of the standard library diff --git a/lib/liquid/template.rb b/lib/liquid/template.rb index a6e36b1b..05b4114d 100644 --- a/lib/liquid/template.rb +++ b/lib/liquid/template.rb @@ -228,28 +228,8 @@ module Liquid private - # Uses the Liquid::TemplateParser regexp to tokenize the passed source def tokenize(source) - source = source.source if source.respond_to?(:source) - return [] if source.to_s.empty? - - tokens = calculate_line_numbers(source.split(TemplateParser)) - - # removes the rogue empty element at the beginning of the array - tokens.shift if tokens[0] && tokens[0].empty? - - tokens - end - - def calculate_line_numbers(raw_tokens) - return raw_tokens unless @line_numbers - - current_line = 1 - raw_tokens.map do |token| - Token.new(token, current_line).tap do - current_line += token.count("\n") - end - end + Tokenizer.new(source, @line_numbers) end def with_profiling diff --git a/lib/liquid/tokenizer.rb b/lib/liquid/tokenizer.rb new file mode 100644 index 00000000..9196d2ae --- /dev/null +++ b/lib/liquid/tokenizer.rb @@ -0,0 +1,38 @@ +module Liquid + class Tokenizer + attr_reader :tokens + + def initialize(source, line_numbers = false) + @source, @line_numbers = source, line_numbers + @tokens = tokenize + end + + def shift + @tokens.shift + end + + private + + def tokenize + @source = @source.source if @source.respond_to?(:source) + return [] if @source.to_s.empty? + + tokens = @source.split(TemplateParser) + tokens = @line_numbers ? calculate_line_numbers(tokens) : tokens + + # removes the rogue empty element at the beginning of the array + tokens.shift if tokens[0] && tokens[0].empty? + + tokens + end + + def calculate_line_numbers(tokens) + current_line = 1 + tokens.map do |token| + Token.new(token, current_line).tap do + current_line += token.count("\n") + end + end + end + end +end diff --git a/test/unit/tokenizer_unit_test.rb b/test/unit/tokenizer_unit_test.rb index f0ae7c99..71b6205e 100644 --- a/test/unit/tokenizer_unit_test.rb +++ b/test/unit/tokenizer_unit_test.rb @@ -24,15 +24,16 @@ class TokenizerTest < Minitest::Test def test_calculate_line_numbers_per_token_with_profiling template = Liquid::Template.parse("", :profile => true) - assert_equal [1], template.send(:tokenize, "{{funk}}").map(&:line_number) - assert_equal [1, 1, 1], template.send(:tokenize, " {{funk}} ").map(&:line_number) - assert_equal [1, 2, 2], template.send(:tokenize, "\n{{funk}}\n").map(&:line_number) - assert_equal [1, 1, 3], template.send(:tokenize, " {{\n funk \n}} ").map(&:line_number) + assert_equal [1], template.send(:tokenize, "{{funk}}").tokens.map(&:line_number) + assert_equal [1, 1, 1], template.send(:tokenize, " {{funk}} ").tokens.map(&:line_number) + assert_equal [1, 2, 2], template.send(:tokenize, "\n{{funk}}\n").tokens.map(&:line_number) + assert_equal [1, 1, 3], template.send(:tokenize, " {{\n funk \n}} ").tokens.map(&:line_number) end private def tokenize(source) - Liquid::Template.new.send(:tokenize, source) + tokenizer = Liquid::Tokenizer.new(source) + tokenizer.tokens end end