regexp_parser 0.5.0 → 1.0.0

data/test/parser/test_groups.rb

@@ -82,9 +82,8 @@ class TestParserGroups < Test::Unit::TestCase
   def test_parse_option_switch_group
     t = RP.parse(/a(?i-m)b/m, 'ruby/1.8')
 
-    assert_equal Group::Options, t.expressions[1].class
-    assert_equal :options,       t.expressions[1].token
-    # TODO: change this ^ to :options_switch in v1.0.0
+    assert_equal Group::Options,  t.expressions[1].class
+    assert_equal :options_switch, t.expressions[1].token
 
     assert_equal false, t.expressions[1].m?
     assert_equal true,  t.expressions[1].i?
@@ -245,4 +244,24 @@ class TestParserGroups < Test::Unit::TestCase
       end
     end
   end
+
+  def test_parse_group_number
+    t = RP.parse('(a)(?=b)((?:c)(d|(e)))')
+    assert_equal 1,     t[0].number
+    assert_equal false, t[1].respond_to?(:number)
+    assert_equal 2,     t[2].number
+    assert_equal false, t[2][0].respond_to?(:number)
+    assert_equal 3,     t[2][1].number
+    assert_equal 4,     t[2][1][0][1][0].number
+  end
+
+  def test_parse_group_number_at_level
+    t = RP.parse('(a)(?=b)((?:c)(d|(e)))')
+    assert_equal 1,     t[0].number_at_level
+    assert_equal false, t[1].respond_to?(:number_at_level)
+    assert_equal 2,     t[2].number_at_level
+    assert_equal false, t[2][0].respond_to?(:number_at_level)
+    assert_equal 1,     t[2][1].number_at_level
+    assert_equal 1,     t[2][1][0][1][0].number_at_level
+  end
 end

data/test/parser/test_posix_classes.rb

@@ -0,0 +1,27 @@
+require File.expand_path("../../helpers", __FILE__)
+
+class TestParserPosixClasses < Test::Unit::TestCase
+  def test_parse_posix_class
+    root = RP.parse('[[:word:]]')
+    exp  = root[0][0]
+
+    assert_equal PosixClass, exp.class
+    assert_equal :posixclass, exp.type
+    assert_equal :word, exp.token
+    assert_equal 'word', exp.name
+    assert_equal '[:word:]', exp.text
+    refute       exp.negative?
+  end
+
+  def test_parse_negative_posix_class
+    root = RP.parse('[[:^word:]]')
+    exp  = root[0][0]
+
+    assert_equal PosixClass, exp.class
+    assert_equal :nonposixclass, exp.type
+    assert_equal :word, exp.token
+    assert_equal 'word', exp.name
+    assert_equal '[:^word:]', exp.text
+    assert       exp.negative?
+  end
+end

data/test/parser/test_properties.rb

@@ -1,314 +1,28 @@
 require File.expand_path("../../helpers", __FILE__)
 
 class ParserProperties < Test::Unit::TestCase
-
   modes = ['p', 'P']
-  props = [
+  example_props = [
     'Alnum',
-    'Alpha',
-    'Ascii',
-    'Blank',
-    'Cntrl',
-    'Digit',
-    'Graph',
-    'Lower',
-    'Print',
-    'Punct',
-    'Space',
-    'Upper',
-    'Word',
-    'Xdigit',
-    'XPosixPunct',
-
     'Any',
-    'Assigned',
-
-    'L',
-    'Letter',
-
-    'Lu',
-    'Uppercase_Letter',
-
-    'Ll',
-    'Lowercase_Letter',
-
-    'Lt',
-    'Titlecase_Letter',
-
-    'Lm',
-    'Modifier_Letter',
-
-    'Lo',
-    'Other_Letter',
-
-    'M',
-    'Mark',
-
-    'Mn',
-    'Nonspacing_Mark',
-
-    'Mc',
-    'Spacing_Mark',
-
-    'Me',
-    'Enclosing_Mark',
-
-    'N',
-    'Number',
-
-    'Nd',
-    'Decimal_Number',
-
-    'Nl',
-    'Letter_Number',
-
-    'No',
-    'Other_Number',
-
-    'P',
-    'Punctuation',
-
-    'Pc',
-    'Connector_Punctuation',
-
-    'Pd',
-    'Dash_Punctuation',
-
-    'Ps',
-    'Open_Punctuation',
-
-    'Pe',
-    'Close_Punctuation',
-
-    'Pi',
-    'Initial_Punctuation',
-
-    'Pf',
-    'Final_Punctuation',
-
-    'Po',
-    'Other_Punctuation',
-
-    'S',
-    'Symbol',
-
-    'Sm',
-    'Math_Symbol',
-
-    'Sc',
-    'Currency_Symbol',
-
-    'Sk',
-    'Modifier_Symbol',
-
-    'So',
-    'Other_Symbol',
-
-    'Z',
-    'Separator',
-
-    'Zs',
-    'Space_Separator',
-
-    'Zl',
-    'Line_Separator',
-
-    'Zp',
-    'Paragraph_Separator',
-
-    'C',
-    'Other',
-
-    'Cc',
-    'Control',
-
-    'Cf',
-    'Format',
-
-    'Cs',
-    'Surrogate',
-
-    'Co',
-    'Private_Use',
-
-    'Cn',
-    'Unassigned',
-
     'Age=1.1',
-    'Age=2.0',
-    'Age=2.1',
-    'Age=3.0',
-    'Age=3.1',
-    'Age=3.2',
-    'Age=4.0',
-    'Age=4.1',
-    'Age=5.0',
-    'Age=5.1',
-    'Age=5.2',
-    'Age=6.0',
-    'Age=7.0',
-    'Age=8.0',
-    'Age=9.0',
-    'Age=10.0',
-
-    'ahex',
-    'ASCII_Hex_Digit',
-
-    'Alphabetic',
-
-    'Cased',
-
-    'cwcf',
-    'Changes_When_Casefolded',
-
-    'cwcm',
-    'Changes_When_Casemapped',
-
-    'cwl',
-    'Changes_When_Lowercased',
-
-    'cwt',
-    'Changes_When_Titlecased',
-
-    'cwu',
-    'Changes_When_Uppercased',
-
-    'ci',
-    'Case_Ignorable',
-
-    'bidic',
-    'Bidi_Control',
-
     'Dash',
-
-    'dep',
-    'Deprecated',
-
     'di',
     'Default_Ignorable_Code_Point',
-
-    'dia',
-    'Diacritic',
-
-    'ext',
-    'Extender',
-
-    'grbase',
-    'Grapheme_Base',
-
-    'grext',
-    'Grapheme_Extend',
-
-    'grlink',
-    'Grapheme_Link',
-
-    'hex',
-    'Hex_Digit',
-
-    'Hyphen',
-
-    'idc',
-    'ID_Continue',
-
-    'ideo',
-    'Ideographic',
-
-    'ids',
-    'ID_Start',
-
-    'idsb',
-    'IDS_Binary_Operator',
-
-    'idst',
-    'IDS_Trinary_Operator',
-
-    'joinc',
-    'Join_Control',
-
-    'loe',
-    'Logical_Order_Exception',
-
-    'Lowercase',
-
     'Math',
-
-    'nchar',
-    'Noncharacter_Code_Point',
-
-    'oalpha',
-    'Other_Alphabetic',
-
-    'odi',
-    'Other_Default_Ignorable_Code_Point',
-
-    'ogrext',
-    'Other_Grapheme_Extend',
-
-    'oidc',
-    'Other_ID_Continue',
-
-    'oids',
-    'Other_ID_Start',
-
-    'olower',
-    'Other_Lowercase',
-
-    'omath',
-    'Other_Math',
-
-    'oupper',
-    'Other_Uppercase',
-
-    'patsyn',
-    'Pattern_Syntax',
-
-    'patws',
-    'Pattern_Whitespace',
-
-    'qmark',
-    'quotationmark',
-
-    'radical',
-
-    'ri',
-    'Regional_Indicator',
-
+    'Noncharacter-Code_Point', # test dash
     'sd',
-    'Soft_Dotted',
-
+    'Soft Dotted', # test whitespace
     'sterm',
-
-    'term',
-    'Terminal_Punctuation',
-
-    'uideo',
-    'Unified_Ideograph',
-
-    'Uppercase',
-
-    'vs',
-    'Variation_Selector',
-
-    'wspace',
-    'whitespace',
-
-    'xids',
-    'XID_Start',
-
     'xidc',
     'XID_Continue',
-
     'Emoji',
-    'Emoji_Component',
-    'Emoji_Modifier',
-    'Emoji_Modifier_Base',
-    'Emoji_Presentation',
   ]
 
   modes.each do |mode|
     token_type = mode == 'p' ? :property : :nonproperty
 
-    props.each do |property|
+    example_props.each do |property|
       define_method "test_parse_#{token_type}_#{property}" do
         t = RP.parse "ab\\#{mode}{#{property}}", 'ruby/2.5'
 
@@ -321,6 +35,13 @@ class ParserProperties < Test::Unit::TestCase
     end
   end
 
+  def test_parse_all_properties_of_current_ruby
+    unsupported = RegexpPropertyValues.all_for_current_ruby.reject do |prop|
+      begin RP.parse("\\p{#{prop}}"); rescue SyntaxError, StandardError; nil end
+    end
+    assert_empty unsupported
+  end
+
   def test_parse_property_negative
     t = RP.parse 'ab\p{L}cd', 'ruby/1.9'
 
@@ -345,6 +66,12 @@ class ParserProperties < Test::Unit::TestCase
     assert_equal false, t.expressions[1].negative?
   end
 
+  def test_parse_property_shortcut
+    assert_equal 'm',  RP.parse('\p{mark}').expressions[0].shortcut
+    assert_equal 'sc', RP.parse('\p{sc}').expressions[0].shortcut
+    assert_equal nil,  RP.parse('\p{in_bengali}').expressions[0].shortcut
+  end
+
   def test_parse_property_age
     t = RP.parse 'ab\p{age=5.2}cd', 'ruby/1.9'
 

data/test/parser/test_refcalls.rb

@@ -5,15 +5,15 @@ class TestParserRefcalls < Test::Unit::TestCase
     root = RP.parse('(abc)\1', 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::Number)
-    assert_equal '1',  exp.number
+    assert_instance_of Backreference::Number, exp
+    assert_equal 1,    exp.number
   end
 
   def test_parse_backref_named_ab
     root = RP.parse('(?<X>abc)\k<X>', 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::Name)
+    assert_instance_of Backreference::Name, exp
     assert_equal 'X',  exp.name
   end
 
@@ -21,7 +21,7 @@ class TestParserRefcalls < Test::Unit::TestCase
     root = RP.parse("(?<X>abc)\\k'X'", 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::Name)
+    assert_instance_of Backreference::Name, exp
     assert_equal 'X',  exp.name
   end
 
@@ -29,102 +29,142 @@ class TestParserRefcalls < Test::Unit::TestCase
     root = RP.parse('(abc)\k<1>', 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::Number)
-    assert_equal '1',  exp.number
+    assert_instance_of Backreference::Number, exp
+    assert_equal 1,    exp.number
   end
 
   def test_parse_backref_number_sq
     root = RP.parse("(abc)\\k'1'", 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::Number)
-    assert_equal '1',  exp.number
+    assert_instance_of Backreference::Number, exp
+    assert_equal 1,    exp.number
   end
 
   def test_parse_backref_number_relative_ab
     root = RP.parse('(abc)\k<-1>', 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NumberRelative)
-    assert_equal '-1', exp.number
+    assert_instance_of Backreference::NumberRelative, exp
+    assert_equal(-1,   exp.number)
   end
 
   def test_parse_backref_number_relative_sq
     root = RP.parse("(abc)\\k'-1'", 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NumberRelative)
-    assert_equal '-1', exp.number
+    assert_instance_of Backreference::NumberRelative, exp
+    assert_equal(-1,   exp.number)
   end
 
   def test_parse_backref_name_call_ab
     root = RP.parse('(?<X>abc)\g<X>', 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NameCall)
+    assert_instance_of Backreference::NameCall, exp
+    assert_equal 'X',  exp.name
   end
 
   def test_parse_backref_name_call_sq
     root = RP.parse("(?<X>abc)\\g'X'", 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NameCall)
+    assert_instance_of Backreference::NameCall, exp
+    assert_equal 'X',  exp.name
   end
 
   def test_parse_backref_number_call_ab
     root = RP.parse('(abc)\g<1>', 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NumberCall)
+    assert_instance_of Backreference::NumberCall, exp
+    assert_equal 1,    exp.number
   end
 
   def test_parse_backref_number_call_sq
     root = RP.parse("(abc)\\g'1'", 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NumberCall)
+    assert_instance_of Backreference::NumberCall, exp
+    assert_equal 1,    exp.number
   end
 
   def test_parse_backref_number_relative_call_ab
     root = RP.parse('(abc)\g<-1>', 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NumberCallRelative)
+    assert_instance_of Backreference::NumberCallRelative, exp
+    assert_equal(-1,   exp.number)
   end
 
   def test_parse_backref_number_relative_call_sq
     root = RP.parse("(abc)\\g'-1'", 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NumberCallRelative)
+    assert_instance_of Backreference::NumberCallRelative, exp
+    assert_equal(-1,   exp.number)
   end
 
-  def test_parse_backref_name_nest_level_ab
+  def test_parse_backref_name_recursion_level_ab
     root = RP.parse('(?<X>abc)\k<X-0>', 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NameNestLevel)
+    assert_instance_of Backreference::NameRecursionLevel, exp
+    assert_equal 'X',  exp.name
+    assert_equal 0,    exp.recursion_level
   end
 
-  def test_parse_backref_name_nest_level_sq
+  def test_parse_backref_name_recursion_level_sq
     root = RP.parse("(?<X>abc)\\k'X-0'", 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NameNestLevel)
+    assert_instance_of Backreference::NameRecursionLevel, exp
+    assert_equal 'X',  exp.name
+    assert_equal 0,    exp.recursion_level
   end
 
-  def test_parse_backref_number_nest_level_ab
+  def test_parse_backref_number_recursion_level_ab
     root = RP.parse('(abc)\k<1-0>', 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NumberNestLevel)
+    assert_instance_of Backreference::NumberRecursionLevel, exp
+    assert_equal 1,    exp.number
+    assert_equal 0,    exp.recursion_level
   end
 
-  def test_parse_backref_number_nest_level_sq
+  def test_parse_backref_number_recursion_level_sq
     root = RP.parse("(abc)\\k'1-0'", 'ruby/1.9')
     exp  = root.expressions.at(1)
 
-    assert_equal true, exp.is_a?(Backreference::NumberNestLevel)
+    assert_instance_of Backreference::NumberRecursionLevel, exp
+    assert_equal 1,    exp.number
+    assert_equal 0,    exp.recursion_level
+  end
+
+  def test_parse_backref_negative_number_recursion_level
+    root = RP.parse("(abc)\\k'-1+0'", 'ruby/1.9')
+    exp  = root.expressions.at(1)
+
+    assert_instance_of Backreference::NumberRecursionLevel, exp
+    assert_equal(-1,   exp.number)
+    assert_equal 0,    exp.recursion_level
   end
 
+  def test_parse_backref_number_positive_recursion_level
+    root = RP.parse("(abc)\\k'1+1'", 'ruby/1.9')
+    exp  = root.expressions.at(1)
+
+    assert_instance_of Backreference::NumberRecursionLevel, exp
+    assert_equal 1,    exp.number
+    assert_equal 1,    exp.recursion_level
+  end
+
+  def test_parse_backref_number_negative_recursion_level
+    root = RP.parse("(abc)\\k'1-1'", 'ruby/1.9')
+    exp  = root.expressions.at(1)
+
+    assert_instance_of Backreference::NumberRecursionLevel, exp
+    assert_equal 1,    exp.number
+    assert_equal(-1,   exp.recursion_level)
+  end
 end