#309 Improved highlighting of Ruby inside diffs

7 files changed, 56 insertions, 33 deletions

diff --git a/lib/coderay/encoders/html.rb b/lib/coderay/encoders/html.rb
index 67d487c..724f3eb 100644
--- a/lib/coderay/encoders/html.rb
+++ b/lib/coderay/encoders/html.rb
@@ -96,16 +96,16 @@ module Encoders
     DEFAULT_OPTIONS = {
       :tab_width => 8,
       
-      :css => :class,
+      :css   => :class,
       :style => :alpha,
-      :wrap => nil,
+      :wrap  => nil,
       :title => 'CodeRay output',
       
-      :line_numbers => nil,
+      :line_numbers        => nil,
       :line_number_anchors => 'n',
-      :line_number_start => 1,
-      :bold_every => 10,
-      :highlight_lines => nil,
+      :line_number_start   => 1,
+      :bold_every          => 10,
+      :highlight_lines     => nil,
       
       :hint => false,
     }
diff --git a/lib/coderay/scanner.rb b/lib/coderay/scanner.rb
index 925acf7..ec2b3aa 100644
--- a/lib/coderay/scanner.rb
+++ b/lib/coderay/scanner.rb
@@ -61,6 +61,8 @@ module CodeRay
       
       KINDS_NOT_LOC = [:comment, :doctype, :docstring]
       
+      attr_accessor :state
+      
       class << self
         
         # Normalizes the given code into a string with UNIX newlines, in the
@@ -190,7 +192,14 @@ module CodeRay
         else
           raise ArgumentError, 'expected String, Array, or nil'
         end
-        scan_tokens @tokens, options
+        
+        begin
+          scan_tokens @tokens, options
+        rescue => e
+          message = "Error in %s#scan_tokens, initial state was: %p" % [self.class, defined?(state) && state]
+          raise_inspect e.message, @tokens, message, 30, e.backtrace
+        end
+        
         @cached_tokens = @tokens
         if source.is_a? Array
           @tokens.split_into_parts(*source.map { |part| part.size })
@@ -260,7 +269,7 @@ module CodeRay
       end
       
       # Scanner error with additional status information
-      def raise_inspect msg, tokens, state = 'No state given!', ambit = 30
+      def raise_inspect msg, tokens, state = self.state || 'No state given!', ambit = 30, backtrace = caller
         raise ScanError, <<-EOE % [
 
 
@@ -288,7 +297,7 @@ surrounding code:
           matched, state, bol?, eos?,
           string[pos - ambit, ambit],
           string[pos, ambit],
-        ]
+        ], backtrace
       end
       
       # Shorthand for scan_until(/\z/).
diff --git a/lib/coderay/scanners/diff.rb b/lib/coderay/scanners/diff.rb
index a8a63e9..fd12922 100644
--- a/lib/coderay/scanners/diff.rb
+++ b/lib/coderay/scanners/diff.rb
@@ -11,7 +11,7 @@ module Scanners
     
     DEFAULT_OPTIONS = {
       :highlight_code => true,
-      :inline_diff => true,
+      :inline_diff    => true,
     }
     
   protected
@@ -73,7 +73,7 @@ module Scanners
             next unless match = scan(/.+/)
             encoder.text_token match, :plain
           elsif match = scan(/@@(?>[^@\n]*)@@/)
-            content_scanner.instance_variable_set(:@state, :initial) unless match?(/\n\+/)
+            content_scanner.state = :initial unless match?(/\n\+/)
             content_scanner_entry_state = nil
             if check(/\n|$/)
               encoder.begin_line line_kind = :change
@@ -106,37 +106,39 @@ module Scanners
             encoder.begin_line line_kind = :delete
             encoder.text_token match, :delete
             if options[:inline_diff] && deleted_lines == 1 && check(/(?>.*)\n\+(?>.*)$(?!\n\+)/)
-              if content_scanner.instance_variable_defined?(:@state)
-                content_scanner_entry_state = content_scanner.instance_variable_get(:@state)
-              end
+              content_scanner_entry_state = content_scanner.state
               skip(/(.*)\n\+(.*)$/)
               head, deletion, insertion, tail = diff self[1], self[2]
               pre, deleted, post = content_scanner.tokenize [head, deletion, tail], :tokens => Tokens.new
               encoder.tokens pre
-              encoder.begin_group :eyecatcher
-              encoder.tokens deleted
-              encoder.end_group :eyecatcher
+              unless deleted.empty?
+                encoder.begin_group :eyecatcher
+                encoder.tokens deleted
+                encoder.end_group :eyecatcher
+              end
               encoder.tokens post
               encoder.end_line line_kind
               encoder.text_token "\n", :space
               encoder.begin_line line_kind = :insert
               encoder.text_token '+', :insert
-              content_scanner.instance_variable_set(:@state, content_scanner_entry_state || :initial)
+              content_scanner.state = content_scanner_entry_state || :initial
               pre, inserted, post = content_scanner.tokenize [head, insertion, tail], :tokens => Tokens.new
               encoder.tokens pre
-              encoder.begin_group :eyecatcher
-              encoder.tokens inserted
-              encoder.end_group :eyecatcher
+              unless inserted.empty?
+                encoder.begin_group :eyecatcher
+                encoder.tokens inserted
+                encoder.end_group :eyecatcher
+              end
               encoder.tokens post
             elsif match = scan(/.*/)
               if options[:highlight_code]
-                if deleted_lines == 1 && content_scanner.instance_variable_defined?(:@state)
-                  content_scanner_entry_state = content_scanner.instance_variable_get(:@state)
+                if deleted_lines == 1
+                  content_scanner_entry_state = content_scanner.state
                 end
                 content_scanner.tokenize match, :tokens => encoder unless match.empty?
                 if !match?(/\n-/)
                   if match?(/\n\+/)
-                    content_scanner.instance_variable_set(:@state, content_scanner_entry_state || :initial)
+                    content_scanner.state = content_scanner_entry_state || :initial
                   end
                   content_scanner_entry_state = nil
                 end
diff --git a/lib/coderay/scanners/ruby.rb b/lib/coderay/scanners/ruby.rb
index 299accc..b3c7de1 100644
--- a/lib/coderay/scanners/ruby.rb
+++ b/lib/coderay/scanners/ruby.rb
@@ -23,8 +23,9 @@ module Scanners
     end
     
     def scan_tokens encoder, options
+      state, heredocs = @state
+      heredocs = heredocs.dup if heredocs.is_a?(Array)
       
-      state = @state
       if state && state.instance_of?(self.class::StringState)
         encoder.begin_group state.type
       end
@@ -34,10 +35,15 @@ module Scanners
       method_call_expected = false
       value_expected = true
       
-      heredocs = nil
       inline_block_stack = nil
       inline_block_curly_depth = 0
       
+      if heredocs
+        state = heredocs.shift
+        encoder.begin_group state.type
+        heredocs = nil if heredocs.empty?
+      end
+      
       # def_object_stack = nil
       # def_object_paren_depth = 0
       
@@ -421,11 +427,14 @@ module Scanners
       
       # cleaning up
       if options[:keep_state]
-        @state = state
+        heredocs = nil if heredocs && heredocs.empty?
+        @state = state, heredocs
       end
+      
       if state.is_a? self.class::StringState
         encoder.end_group state.type
       end
+      
       if inline_block_stack
         until inline_block_stack.empty?
           state, = *inline_block_stack.pop
diff --git a/lib/coderay/scanners/ruby/string_state.rb b/lib/coderay/scanners/ruby/string_state.rb
index 14127c6..2f398d1 100644
--- a/lib/coderay/scanners/ruby/string_state.rb
+++ b/lib/coderay/scanners/ruby/string_state.rb
@@ -55,7 +55,7 @@ module Scanners
       def heredoc_pattern delim, interpreted, indented
         # delim = delim.dup  # workaround for old Ruby
         delim_pattern = Regexp.escape(delim)
-        delim_pattern = / \n #{ '(?>[ \t]*)' if indented } #{ Regexp.new delim_pattern } $ /x
+        delim_pattern = / (?:\A|\n) #{ '(?>[ \t]*)' if indented } #{ Regexp.new delim_pattern } $ /x
         if interpreted
           / (?= #{delim_pattern}() | \\ | \# [{$@] ) /mx  # $1 set == end of heredoc
         else
diff --git a/lib/coderay/styles/alpha.rb b/lib/coderay/styles/alpha.rb
index 5747abb..6399d7d 100644
--- a/lib/coderay/styles/alpha.rb
+++ b/lib/coderay/styles/alpha.rb
@@ -141,8 +141,8 @@ table.CodeRay td { padding: 2px 4px; vertical-align: top; }
 .head { color: #f8f; background: #505 }
 .head .filename { color: white; }
 
-.ins .eye { background-color: hsla(120,100%,50%,0.2) }
-.del .eye { background-color: hsla(0,100%,50%,0.2) }
+.del .eye { background-color: hsla(0,100%,50%,0.2); border: 1px solid hsla(0,100%,45%,0.5); margin: -1px; border-bottom: none; border-top-left-radius: 5px; border-top-right-radius: 5px; }
+.ins .eye { background-color: hsla(120,100%,50%,0.2); border: 1px solid hsla(120,100%,25%,0.5); margin: -1px; border-top: none; border-bottom-left-radius: 5px; border-bottom-right-radius: 5px; }
 
 .ins .ins { color: #080; background:transparent; font-weight:bold }
 .del .del { color: #800; background:transparent; font-weight:bold }
diff --git a/lib/coderay/tokens.rb b/lib/coderay/tokens.rb
index e7da7e8..f6f8845 100644
--- a/lib/coderay/tokens.rb
+++ b/lib/coderay/tokens.rb
@@ -223,12 +223,15 @@ module CodeRay
                 content_or_kind
               end
             end
-            parts << part.concat(closing)
-            part = Tokens.new
+            part.concat closing
+            begin
+              parts << part
+              part = Tokens.new
+              size = sizes[i += 1]
+            end until size.nil? || size > 0
             # ...and open them again.
             part.concat opened.flatten
             part_size = 0
-            size = sizes[i += 1]
             redo unless content.empty?
           else
             part << content << item