Merge pull request #1 from zombocom/schneems/wip

Initial working implementation

Author: schneems

lib/syntax_error_search.rb

@@ -6,5 +6,34 @@
 
 module SyntaxErrorSearch
   class Error < StandardError; end
-  # Your code goes here...
+
+  # Used for counting spaces
+  module SpaceCount
+    def self.indent(string)
+      string.split(/\w/).first&.length || 0
+    end
+  end
+
+
+  def self.valid?(source)
+    source = source.join if source.is_a?(Array)
+    source = source.to_s
+
+    # Parser writes to stderr even if you catch the error
+    #
+    stderr = $stderr
+    $stderr = StringIO.new
+
+    Parser::CurrentRuby.parse(source)
+    true
+  rescue Parser::SyntaxError
+    false
+  ensure
+    $stderr = stderr if stderr
+  end
 end
+
+require_relative "syntax_error_search/code_line"
+require_relative "syntax_error_search/code_block"
+require_relative "syntax_error_search/code_frontier"
+require_relative "syntax_error_search/code_search"

lib/syntax_error_search/code_block.rb

@@ -0,0 +1,209 @@
+module SyntaxErrorSearch
+  # Multiple lines form a singular CodeBlock
+  #
+  # Source code is made of multiple CodeBlocks. A code block
+  # has a reference to the source code that created itself, this allows
+  # a code block to "expand" when needed
+  #
+  # The most important ability of a CodeBlock is this ability to expand:
+  #
+  # Example:
+  #
+  #   code_block.to_s # =>
+  #     #   def foo
+  #     #     puts "foo"
+  #     #   end
+  #
+  #   code_block.expand_until_next_boundry
+  #
+  #   code_block.to_s # =>
+  #     # class Foo
+  #     #   def foo
+  #     #     puts "foo"
+  #     #   end
+  #     # end
+  #
+  class CodeBlock
+    attr_reader :lines
+
+    def initialize(code_lines:, lines: [])
+      @lines = Array(lines)
+      @code_lines = code_lines
+    end
+
+    def is_end?
+      to_s.strip == "end"
+    end
+
+    def starts_at
+      @lines.first&.line_number
+    end
+
+    def code_lines
+      @code_lines
+    end
+
+    # This is used for frontier ordering, we are searching from
+    # the largest indentation to the smallest. This allows us to
+    # populate an array with multiple code blocks then call `sort!`
+    # on it without having to specify the sorting criteria
+    def <=>(other)
+      self.current_indent <=> other.current_indent
+    end
+
+    # Only the lines that are not empty and visible
+    def visible_lines
+      @lines
+        .select(&:not_empty?)
+        .select(&:visible?)
+    end
+
+    # This method is used to expand a code block to capture it's calling context
+    def expand_until_next_boundry
+      expand_to_indent(next_indent)
+      self
+    end
+
+    # This method expands the given code block until it captures
+    # its nearest neighbors. This is used to expand a single line of code
+    # to its smallest likely block.
+    #
+    #   code_block.to_s # =>
+    #     #     puts "foo"
+    #   code_block.expand_until_neighbors
+    #
+    #   code_block.to_s # =>
+    #     #     puts "foo"
+    #     #     puts "bar"
+    #     #     puts "baz"
+    #
+    def expand_until_neighbors
+      expand_to_indent(current_indent)
+
+      expand_hidden_parner_line if self.to_s.strip == "end"
+      self
+    end
+
+    def expand_hidden_parner_line
+      index = @lines.first.index
+      indent = current_indent
+      partner_line  = code_lines.select {|line| line.index < index && line.indent == indent }.last
+
+      if partner_line&.hidden?
+        partner_line.mark_visible
+        @lines.prepend(partner_line)
+      end
+    end
+
+    # This method expands the existing code block up (before)
+    # and down (after). It will break on change in indentation
+    # and empty lines.
+    #
+    #   code_block.to_s # =>
+    #     #   def foo
+    #     #     puts "foo"
+    #     #   end
+    #
+    #   code_block.expand_to_indent(0)
+    #   code_block.to_s # =>
+    #     # class Foo
+    #     #   def foo
+    #     #     puts "foo"
+    #     #   end
+    #     # end
+    #
+    private def expand_to_indent(indent)
+      array = []
+      before_lines(skip_empty: false).each do |line|
+        if line.empty?
+          array.prepend(line)
+          break
+        end
+
+        if line.indent == indent
+          array.prepend(line)
+        else
+          break
+        end
+      end
+
+      array << @lines
+
+      after_lines(skip_empty: false).each do |line|
+        if line.empty?
+          array << line
+          break
+        end
+
+        if line.indent == indent
+          array << line
+        else
+          break
+        end
+      end
+
+      @lines = array.flatten
+    end
+
+    def next_indent
+      [
+        before_line&.indent || 0,
+        after_line&.indent || 0
+      ].max
+    end
+
+    def current_indent
+      lines.detect(&:not_empty?)&.indent || 0
+    end
+
+    def before_line
+      before_lines.first
+    end
+
+    def after_line
+      after_lines.first
+    end
+
+    def before_lines(skip_empty: true)
+      index = @lines.first.index
+      lines = code_lines.select {|line| line.index < index }
+      lines.select!(&:not_empty?) if skip_empty
+      lines.select!(&:visible?)
+      lines.reverse!
+
+      lines
+    end
+
+    def after_lines(skip_empty: true)
+      index = @lines.last.index
+      lines = code_lines.select {|line| line.index > index }
+      lines.select!(&:not_empty?) if skip_empty
+      lines.select!(&:visible?)
+      lines
+    end
+
+    # Returns a code block of the source that does not include
+    # the current lines. This is useful for checking if a source
+    # with the given lines removed parses successfully. If so
+    #
+    # Then it's proof that the current block is invalid
+    def block_without
+      @block_without ||= CodeBlock.new(
+        source: @source,
+        lines: @source.code_lines - @lines
+      )
+    end
+
+    def document_valid_without?
+      block_without.valid?
+    end
+
+    def valid?
+      SyntaxErrorSearch.valid?(self.to_s)
+    end
+
+    def to_s
+      @lines.join
+    end
+  end
+end

lib/syntax_error_search/code_frontier.rb

@@ -0,0 +1,110 @@
+module SyntaxErrorSearch
+  # This class is responsible for generating, storing, and sorting code blocks
+  class CodeFrontier
+    def initialize(code_lines: )
+      @code_lines = code_lines
+      @frontier = []
+      @indent_hash = {}
+      code_lines.each do |line|
+        next if line.empty?
+
+        @indent_hash[line.indent] ||= []
+        @indent_hash[line.indent] << line
+      end
+    end
+
+    # Returns true if the document is valid with all lines
+    # removed. By default it checks all blocks in present in
+    # the frontier array, but can be used for arbitrary arrays
+    # of codeblocks as well
+    def holds_all_syntax_errors?(block_array = @frontier)
+      lines = @code_lines
+      block_array.each do |block|
+        lines -= block.lines
+      end
+
+      return true if lines.empty?
+
+      CodeBlock.new(
+        code_lines: @code_lines,
+        lines: lines
+      ).valid?
+    end
+
+    # Returns a code block with the largest indentation possible
+    def pop
+      return nil if empty?
+
+      if generate_new_block?
+        self << next_block
+      end
+
+      return @frontier.pop
+    end
+
+    def next_block
+      indent = @indent_hash.keys.sort.last
+      lines = @indent_hash[indent].first
+
+      CodeBlock.new(
+        lines: lines,
+        code_lines: @code_lines
+      ).expand_until_neighbors
+    end
+
+    # This method is responsible for determining if a new code
+    # block should be generated instead of evaluating an already
+    # existing block in the frontier
+    def generate_new_block?
+      return false if @indent_hash.empty?
+      return true if @frontier.empty?
+
+      @frontier.last.current_indent <= @indent_hash.keys.sort.last
+    end
+
+    # Add a block to the frontier
+    #
+    # This method ensures the frontier always remains sorted (in indentation order)
+    # and that each code block's lines are removed from the indentation hash so we
+    # don't re-evaluate the same line multiple times.
+    def <<(block)
+      block.lines.each do |line|
+        @indent_hash[line.indent]&.delete(line)
+      end
+      @indent_hash.select! {|k, v| !v.empty?}
+
+      @frontier << block
+      @frontier.sort!
+
+      self
+    end
+
+    def any?
+      !empty?
+    end
+
+    def empty?
+      @frontier.empty? && @indent_hash.empty?
+    end
+
+    # Example:
+    #
+    #   combination([:a, :b, :c, :d])
+    #   # => [[:a], [:b], [:c], [:d], [:a, :b], [:a, :c], [:a, :d], [:b, :c], [:b, :d], [:c, :d], [:a, :b, :c], [:a, :b, :d], [:a, :c, :d], [:b, :c, :d], [:a, :b, :c, :d]]
+    def self.combination(array)
+      guesses = []
+      1.upto(array.length).each do |size|
+        guesses.concat(array.combination(size).to_a)
+      end
+      guesses
+    end
+
+    # Given that we know our syntax error exists somewhere in our frontier, we want to find
+    # the smallest possible set of blocks that contain all the syntax errors
+    def detect_invalid_blocks
+      self.class.combination(@frontier).detect do |block_array|
+        holds_all_syntax_errors?(block_array)
+      end || []
+    end
+  end
+end