machete 0.1.0 → 0.2.0

data/CHANGELOG

@@ -0,0 +1,10 @@
+0.2.0 (2011-08-17)
+------------------
+
+* Extended method name grammar to cover more Ruby method names.
+* Extended symbol grammar to cover more Ruby symbols.
+
+0.1.0 (2011-07-27)
+------------------
+
+* Initial release.

data/VERSION

@@ -1 +1 @@
-0.1.0
+0.2.0

data/lib/machete/parser.y

@@ -1,6 +1,6 @@
 class Machete::Parser
 
-token VAR_NAME
+token METHOD_NAME
 token CLASS_NAME
 token SYMBOL
 token INTEGER
@@ -32,7 +32,14 @@ node : CLASS_NAME {
 attrs : attr
       | attrs "," attr { result = val[0].merge(val[2]) }
 
-attr : VAR_NAME "=" expression { result = { val[0].to_sym => val[2] } }
+attr : method_name "=" expression { result = { val[0].to_sym => val[2] } }
+
+# Hack to overcome the fact that "<", ">" and "|" will lex as simple tokens, not
+# METHOD_NAME tokens.
+method_name : METHOD_NAME
+            | "<"
+            | ">"
+            | "|"
 
 literal : SYMBOL  { result = LiteralMatcher.new(val[0][1..-1].to_sym) }
         | INTEGER {
@@ -94,9 +101,8 @@ private
 SIMPLE_TOKENS = ["|", "<", ">", ",", "="]
 
 COMPLEX_TOKENS = [
-  [:VAR_NAME,   /^[a-z_][a-zA-Z0-9_]*/],
-  [:CLASS_NAME, /^[A-Z][a-zA-Z0-9_]*/],
-  [:SYMBOL,     /^:[a-zA-Z_][a-zA-Z0-9_]*/],
+  # INTEGER needs to be before METHOD_NAME, otherwise e.g. "+1" would be
+  # recognized as two tokens.
   [
     :INTEGER,
     /^
@@ -144,6 +150,38 @@ COMPLEX_TOKENS = [
         "
       )
     /x
+  ],
+  # We exclude "<", ">" and "|" from method names since they are lexed as simple
+  # tokens. This is because they have also other meanings in Machette patterns
+  # beside Ruby method names.
+  [
+    :METHOD_NAME,
+    /^
+      (
+        # regular name
+        [a-z_][a-zA-Z0-9_]*[?!=]?
+        |
+        # operator (sorted by length, then alphabetically)
+        (<=>|===|\[\]=|\*\*|\+@|-@|<<|<=|==|=~|>=|>>|\[\]|[%&*+\-\/^`~])
+      )
+    /x
+  ],
+  [:CLASS_NAME,  /^[A-Z][a-zA-Z0-9_]*/],
+  [
+    :SYMBOL,
+    /^
+      :
+      (
+        # class name
+        [A-Z][a-zA-Z0-9_]*
+        |
+        # regular method name
+        [a-z_][a-zA-Z0-9_]*[?!=]?
+        |
+        # operator (sorted by length, then alphabetically)
+        (<=>|===|\[\]=|\*\*|\+@|-@|<<|<=|==|=~|>=|>>|\[\]|[%&*+\-\/<>^`|~])
+      )
+    /x
   ]
 ]
 
@@ -152,12 +190,8 @@ def next_token
 
   return false if remaining_input.empty?
 
-  SIMPLE_TOKENS.each do |token|
-    if remaining_input[0...token.length] == token
-      @pos += token.length
-      return [token, token]
-    end
-  end
+  # Complex tokens need to be before simple tokens, otherwise e.g. "<<" would be
+  # recognized as two tokens.
 
   COMPLEX_TOKENS.each do |type, regexp|
     if remaining_input =~ regexp
@@ -166,6 +200,13 @@ def next_token
     end
   end
 
+  SIMPLE_TOKENS.each do |token|
+    if remaining_input[0...token.length] == token
+      @pos += token.length
+      return [token, token]
+    end
+  end
+
   raise SyntaxError, "Unexpected character: #{remaining_input[0..0].inspect}."
 end
 

data/spec/machete/parser_spec.rb

@@ -65,6 +65,14 @@ module Machete
       'Foo<a = 42 | 43>'.should be_parsed_as(NodeMatcher.new(:Foo, :a => @ch4243))
     end
 
+    # Canonical method_name is "a".
+    it "parses method_name" do
+      'Foo<a = 42>'.should be_parsed_as(node_matcher_with_attr(:a))
+      'Foo< < = 42>'.should be_parsed_as(node_matcher_with_attr(:<))
+      'Foo<> = 42>'.should be_parsed_as(node_matcher_with_attr(:>))
+      'Foo<| = 42>'.should be_parsed_as(node_matcher_with_attr(:|))
+    end
+
     # Canonical literal is "42".
     it "parses literal" do
       ':a'.should be_parsed_as(LiteralMatcher.new(:a))
@@ -72,8 +80,9 @@ module Machete
       '"abcd"'.should be_parsed_as(LiteralMatcher.new("abcd"))
     end
 
-    # Canonical VAR_NAME is "a".
-    it "parses VAR_NAME" do
+    # Canonical METHOD_NAME is "a".
+    it "parses METHOD_NAME" do
+      # Regular names
       'Foo<a = 42>'.should be_parsed_as(node_matcher_with_attr(:a))
       'Foo<z = 42>'.should be_parsed_as(node_matcher_with_attr(:z))
       'Foo<_ = 42>'.should be_parsed_as(node_matcher_with_attr(:_))
@@ -85,6 +94,33 @@ module Machete
       'Foo<a9 = 42>'.should be_parsed_as(node_matcher_with_attr(:a9))
       'Foo<a_ = 42>'.should be_parsed_as(node_matcher_with_attr(:a_))
       'Foo<abcd = 42>'.should be_parsed_as(node_matcher_with_attr(:abcd))
+      'Foo<a? = 42>'.should be_parsed_as(node_matcher_with_attr(:a?))
+      'Foo<a! = 42>'.should be_parsed_as(node_matcher_with_attr(:a!))
+      'Foo<a= = 42>'.should be_parsed_as(node_matcher_with_attr(:a=))
+
+      # Operators (sorted alphabetically)
+      'Foo<% = 42>'.should be_parsed_as(node_matcher_with_attr(:%))
+      'Foo<& = 42>'.should be_parsed_as(node_matcher_with_attr(:&))
+      'Foo<* = 42>'.should be_parsed_as(node_matcher_with_attr(:*))
+      'Foo<** = 42>'.should be_parsed_as(node_matcher_with_attr(:**))
+      'Foo<+ = 42>'.should be_parsed_as(node_matcher_with_attr(:+))
+      'Foo<+@ = 42>'.should be_parsed_as(node_matcher_with_attr(:+@))
+      'Foo<- = 42>'.should be_parsed_as(node_matcher_with_attr(:-))
+      'Foo<-@ = 42>'.should be_parsed_as(node_matcher_with_attr(:-@))
+      'Foo</ = 42>'.should be_parsed_as(node_matcher_with_attr(:/))
+      'Foo< << = 42>'.should be_parsed_as(node_matcher_with_attr(:<<))
+      'Foo< <= = 42>'.should be_parsed_as(node_matcher_with_attr(:<=))
+      'Foo< <=> = 42>'.should be_parsed_as(node_matcher_with_attr(:<=>))
+      'Foo< == = 42>'.should be_parsed_as(node_matcher_with_attr(:==))
+      'Foo< === = 42>'.should be_parsed_as(node_matcher_with_attr(:===))
+      'Foo< =~ = 42>'.should be_parsed_as(node_matcher_with_attr(:=~))
+      'Foo<>= = 42>'.should be_parsed_as(node_matcher_with_attr(:>=))
+      'Foo<>> = 42>'.should be_parsed_as(node_matcher_with_attr(:>>))
+      'Foo<[] = 42>'.should be_parsed_as(node_matcher_with_attr(:[]))
+      'Foo<[]= = 42>'.should be_parsed_as(node_matcher_with_attr(:[]=))
+      'Foo<^ = 42>'.should be_parsed_as(node_matcher_with_attr(:^))
+      'Foo<` = 42>'.should be_parsed_as(node_matcher_with_attr(:`))
+      'Foo<~ = 42>'.should be_parsed_as(node_matcher_with_attr(:~))
     end
 
     # Canonical CLASS_NAME is "A".
@@ -103,10 +139,21 @@ module Machete
 
     # Canonical SYMBOL is ":a".
     it "parses SYMBOL" do
-      ':a'.should be_parsed_as(LiteralMatcher.new(:a))
-      ':z'.should be_parsed_as(LiteralMatcher.new(:z))
+      # Class names
       ':A'.should be_parsed_as(LiteralMatcher.new(:A))
       ':Z'.should be_parsed_as(LiteralMatcher.new(:Z))
+      ':Aa'.should be_parsed_as(LiteralMatcher.new(:Aa))
+      ':Az'.should be_parsed_as(LiteralMatcher.new(:Az))
+      ':AA'.should be_parsed_as(LiteralMatcher.new(:AA))
+      ':AZ'.should be_parsed_as(LiteralMatcher.new(:AZ))
+      ':A0'.should be_parsed_as(LiteralMatcher.new(:A0))
+      ':A9'.should be_parsed_as(LiteralMatcher.new(:A9))
+      ':A_'.should be_parsed_as(LiteralMatcher.new(:A_))
+      ':Abcd'.should be_parsed_as(LiteralMatcher.new(:Abcd))
+
+      # Regular method names
+      ':a'.should be_parsed_as(LiteralMatcher.new(:a))
+      ':z'.should be_parsed_as(LiteralMatcher.new(:z))
       ':_'.should be_parsed_as(LiteralMatcher.new(:_))
       ':aa'.should be_parsed_as(LiteralMatcher.new(:aa))
       ':az'.should be_parsed_as(LiteralMatcher.new(:az))
@@ -116,6 +163,36 @@ module Machete
       ':a9'.should be_parsed_as(LiteralMatcher.new(:a9))
       ':a_'.should be_parsed_as(LiteralMatcher.new(:a_))
       ':abcd'.should be_parsed_as(LiteralMatcher.new(:abcd))
+      ':a?'.should be_parsed_as(LiteralMatcher.new(:a?))
+      ':a!'.should be_parsed_as(LiteralMatcher.new(:a!))
+      ':a='.should be_parsed_as(LiteralMatcher.new(:a=))
+
+      # Operators (sorted alphabetically)
+      ':%'.should be_parsed_as(LiteralMatcher.new(:%))
+      ':&'.should be_parsed_as(LiteralMatcher.new(:&))
+      ':*'.should be_parsed_as(LiteralMatcher.new(:*))
+      ':**'.should be_parsed_as(LiteralMatcher.new(:**))
+      ':+'.should be_parsed_as(LiteralMatcher.new(:+))
+      ':+@'.should be_parsed_as(LiteralMatcher.new(:+@))
+      ':-'.should be_parsed_as(LiteralMatcher.new(:-))
+      ':-@'.should be_parsed_as(LiteralMatcher.new(:-@))
+      ':/'.should be_parsed_as(LiteralMatcher.new(:/))
+      ':<'.should be_parsed_as(LiteralMatcher.new(:<))
+      ':<<'.should be_parsed_as(LiteralMatcher.new(:<<))
+      ':<='.should be_parsed_as(LiteralMatcher.new(:<=))
+      ':<=>'.should be_parsed_as(LiteralMatcher.new(:<=>))
+      ':=='.should be_parsed_as(LiteralMatcher.new(:==))
+      ':==='.should be_parsed_as(LiteralMatcher.new(:===))
+      ':=~'.should be_parsed_as(LiteralMatcher.new(:=~))
+      ':>'.should be_parsed_as(LiteralMatcher.new(:>))
+      ':>='.should be_parsed_as(LiteralMatcher.new(:>=))
+      ':>>'.should be_parsed_as(LiteralMatcher.new(:>>))
+      ':[]'.should be_parsed_as(LiteralMatcher.new(:[]))
+      ':[]='.should be_parsed_as(LiteralMatcher.new(:[]=))
+      ':^'.should be_parsed_as(LiteralMatcher.new(:^))
+      ':`'.should be_parsed_as(LiteralMatcher.new(:`))
+      ':|'.should be_parsed_as(LiteralMatcher.new(:|))
+      ':~'.should be_parsed_as(LiteralMatcher.new(:~))
     end
 
     # Canonical INTEGER is "42".

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: machete
 version: !ruby/object:Gem::Version 
-  hash: 27
+  hash: 23
   prerelease: 
   segments: 
   - 0
-  - 1
+  - 2
   - 0
-  version: 0.1.0
+  version: 0.2.0
 platform: ruby
 authors: 
 - David Majda
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-07-27 00:00:00 +02:00
+date: 2011-08-17 00:00:00 +02:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -99,6 +99,7 @@ extra_rdoc_files: []
 files: 
 - .gitignore
 - .yardopts
+- CHANGELOG
 - LICENSE
 - README.md
 - Rakefile