Strip HEREDOC

Heredocs in ruby are common sources of indentation even though they cannot cause either an "unmatched" or "unexpected" end syntax errors. To help avoid these false positives when parsing source, we first find and parse all heredocs. If the lines that contain the heredocs are valid, they'll be commented-out/hidden early on. This drastically decreases the chance that heredoc based indentation will lead to false positives.

CHANGELOG.md

@@ -1,5 +1,7 @@
 ## HEAD (unreleased)
 
+- Strip out heredocs in documents first (https://github.com/zombocom/syntax_search/pull/19)
+
 ## 0.1.4
 
 - Parser gem replaced with Ripper (https://github.com/zombocom/syntax_search/pull/17)

lib/syntax_search.rb

@@ -141,3 +141,4 @@ def self.invalid_type(source)
 
 require_relative "syntax_search/code_search"
 require_relative "syntax_search/who_dis_syntax_error"
+require_relative "syntax_search/heredoc_block_parse"

lib/syntax_search/code_block.rb

@@ -31,6 +31,10 @@ def is_end?
       to_s.strip == "end"
     end
 
+    def hidden?
+      @lines.all?(&:hidden?)
+    end
+
     def starts_at
       @starts_at ||= @lines.first&.line_number
     end

lib/syntax_search/code_search.rb

@@ -28,13 +28,14 @@ class CodeSearch
     private; attr_reader :frontier; public
     public; attr_reader :invalid_blocks, :record_dir, :code_lines
 
-    def initialize(string, record_dir: ENV["SYNTAX_SEARCH_RECORD_DIR"])
+    def initialize(source, record_dir: ENV["SYNTAX_SEARCH_RECORD_DIR"])
+      @source = source
       if record_dir
         @time = Time.now.strftime('%Y-%m-%d-%H-%M-%s-%N')
         @record_dir = Pathname(record_dir).join(@time).tap {|p| p.mkpath }
         @write_count = 0
       end
-      @code_lines = string.lines.map.with_index do |line, i|
+      @code_lines = source.lines.map.with_index do |line, i|
         CodeLine.new(line: line, index: i)
       end
       @frontier = CodeFrontier.new(code_lines: @code_lines)
@@ -107,8 +108,19 @@ def expand_invalid_block
       push(block, name: "expand")
     end
 
+
+    def sweep_heredocs
+      HeredocBlockParse.new(
+        source: @source,
+        code_lines: @code_lines
+      ).call.each do |block|
+        push(block, name: "heredoc")
+      end
+    end
+
     # Main search loop
     def call
+      sweep_heredocs
       until frontier.holds_all_syntax_errors?
         @tick += 1
 

lib/syntax_search/display_invalid_blocks.rb

@@ -20,7 +20,7 @@ def initialize(code_lines: ,blocks:, io: $stderr, filename: nil, terminal: false
     end
 
     def call
-      if @blocks.any?
+      if @blocks.any? { |b| !b.hidden? }
         found_invalid_blocks
       else
         @io.puts "Syntax OK"

lib/syntax_search/heredoc_block_parse.rb

@@ -0,0 +1,30 @@
+# frozen_string_literal: true
+
+module SyntaxErrorSearch
+  # Takes in a source, and returns blocks containing each heredoc
+  class HeredocBlockParse
+    private; attr_reader :code_lines, :lex; public
+
+    def initialize(source:, code_lines: )
+      @code_lines = code_lines
+      @lex = Ripper.lex(source)
+    end
+
+    def call
+      blocks = []
+      beginning = []
+      @lex.each do |(line, col), event, *_|
+        case event
+        when :on_heredoc_beg
+          beginning << line
+        when :on_heredoc_end
+          start_index = beginning.pop - 1
+          end_index = line - 1
+          blocks << CodeBlock.new(lines: code_lines[start_index..end_index])
+        end
+      end
+
+      blocks
+    end
+  end
+end