regextest 0.1.6 → 0.1.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: dee3ca4abbe70edabbe42f690c1a3c586fe689e0
-  data.tar.gz: 7bfe0081c0432cb7e9f9da3e4918042a36770bc1
+  metadata.gz: 865e62c608efb9bb2cf95f3de6e426eef8a00a2a
+  data.tar.gz: ca284223e0405480e8361f00fdbb59f71129ee21
 SHA512:
-  metadata.gz: 4b2bf4c54259660e4f4509ee944676388530e6203266820b22a4cc32bb82d1adb849e8c535b8d4044057517fa8169ad9c39ec2ff23391ff43672042a492515f4
-  data.tar.gz: 8f9faee644cd13ade02be3cdcc1ab51eb8e4f14985402eb6ea732bfb988ac2eb363c82fa2da595bd39c2f3e2bbc514ead114a33603275000fd717167eae77fc3
+  metadata.gz: 29c2cb6d4bb6a678bb978fb760341f4bf4eee750ff5deaea9f631df6df0270f7d18914d920ec155c6403ca78816f009eed7f98518511290abf83fca9d466e46b
+  data.tar.gz: 467200d8e3ba06a936ac58b7a746f6d4cdd3cc0e0c4fc8826c4e4b51c2e1b5a2a23213a89c83e4ada76366151f265579b7ed5eeef49193f5e0cf0a0f42554a6d

data/.gitignore

@@ -8,4 +8,8 @@
 /spec/reports/
 /tmp/
 /work/
+/lib/regextest/unicode.rb
+/lib/regextest/front/case-folding.rb
+/lib/regextest/front/parser.rb
+/lib/regextest/front/bracket-parser.rb
 /*.gem

data/Rakefile

@@ -5,7 +5,7 @@ RSpec::Core::RakeTask.new(:spec)
 
 # task :default => :spec
 
-task :default => [:make, :spec]
+task :default => [:make, :doc, :spec]
 
 # Generating parser
 file 'lib/regextest/front/parser.rb' => 'lib/regextest/front/parser.y' do
@@ -59,4 +59,16 @@ task :test => :make do
   sh 'ruby test.rb'
 end
 
+# Generating documents
+file 'doc/index.html' => [ 'lib/regextest/regexp.rb', 'README.md' ] do
+  puts "making document for regextest"
+  sh 'yardoc lib/regextest/regexp.rb'
+end
+
+task :doc =>
+        [
+          'doc/index.html',
+        ] do 
+  puts "Generate documents"
+end
 

data/lib/regextest.rb

@@ -178,7 +178,7 @@ class Regextest
   def verify(result_string)
     md = nil
     begin
-      timeout(TstConstTimeout){
+      Timeout.timeout(TstConstTimeout){
         md = @reg_exp.match(result_string)
       }
     rescue Timeout::Error => ex

data/lib/regextest/back/main.rb

@@ -280,7 +280,7 @@ class Regextest::Back::Main
     end
     
     # puts "repeat=#{repeat} quit=#{@quit_mode} nest=#{@nest}"
-    repeat.times do
+    repeat.times do | current_repeat |
       if( elem = generate_candidates(target["value"], param))
         result.push elem
       else
@@ -288,9 +288,13 @@ class Regextest::Back::Main
         break
       end
       
-      # quit to repeat if the first element is begin anchor
       elem.flatten! if Array === elem  # flatten considering duplicated repeat
-      if elem.size > 0 && elem[0].respond_to?(:command) && elem[-1].respond_to?(:command)
+      
+      # quit to repeat if the first element is begin anchor
+      if current_repeat >= min_repeat &&
+         elem.size > 0 &&
+         elem[0].respond_to?(:command) &&
+         elem[-1].respond_to?(:command)
         break if elem[0].command == :CMD_ANC_LINE_BEGIN && !elem[-1].new_line?
         break if elem[0].command == :CMD_ANC_STRING_BEGIN
       end
@@ -327,11 +331,9 @@ class Regextest::Back::Main
       # puts "relative: #{relative_num}, nest=#{@nest}, :#{target}"
       result = @parens_hash[target["refer_name"]][:generated][relative_num]
       
-      # Somehow /(^a)\1/ must match with "aa"
-      if result.size > 0 &&
-         (result[0].command == :CMD_ANC_LINE_BEGIN ||
-          result[0].command == :CMD_ANC_STRING_BEGIN)
-        result = result[1..-1]  # ignore first anchor
+      # Somehow /(^a)\1/ or /(\ba)\1/ must match with "aa"
+      if result.size > 0
+        result = result.select{|elem| elem.command == :CMD_SELECT}
       end
     else
       result = nil

data/lib/regextest/back/result.rb

@@ -382,16 +382,24 @@ class Regextest::Back::Result
             @results.push (Regextest::Back::Element.any_char)
             bound_process(@results[0], @results[1])
           elsif offset == @results.size
-            @results.push (Regextest::Back::Element.any_char)
-            if !bound_process(@results[-2], @results[-1])
+            if !is_end_anchor?
+              @results.push (Regextest::Back::Element.any_char)
+              if !bound_process(@results[-2], @results[-1])
+                return false
+              end
+            elsif !@results[-1].word_elements?
               return false
             end
           elsif offset == 0
-            if !unshift_params(1)
-              return false
-            end
-            @results.unshift (Regextest::Back::Element.any_char)
-            if !bound_process(@results[0], @results[1])
+            if !is_begin_anchor?
+              if !unshift_params(1)
+                return false
+              end
+              @results.unshift (Regextest::Back::Element.any_char)
+              if !bound_process(@results[0], @results[1])
+                return false
+              end
+            elsif !@results[0].word_elements?
               return false
             end
           end
@@ -409,16 +417,24 @@ class Regextest::Back::Result
             @results.push (Regextest::Back::Element.any_char)
             unbound_process(@results[0], @results[1])
           elsif offset == @results.size
-            @results.push (Regextest::Back::Element.any_char)
-            if !unbound_process(@results[-2], @results[-1])
+            if !is_end_anchor?
+              @results.push (Regextest::Back::Element.any_char)
+              if !unbound_process(@results[-2], @results[-1])
+                return false
+              end
+            elsif @results[-1].word_elements?
               return false
             end
           elsif offset == 0
-            if !unshift_params(1)
-              return false
-            end
-            @results.unshift (Regextest::Back::Element.any_char)
-            if !unbound_process(@results[0], @results[1])
+            if !is_begin_anchor?
+              if !unshift_params(1)
+                return false
+              end
+              @results.unshift (Regextest::Back::Element.any_char)
+              if !unbound_process(@results[0], @results[1])
+                return false
+              end
+            elsif @results[0].word_elements?
               return false
             end
           end
@@ -481,6 +497,24 @@ class Regextest::Back::Result
     end
     true
   end
+  
+  # return true if begin anchor exists
+  def is_begin_anchor?
+    (@positional_anchors[:CMD_ANC_STRING_BEGIN] &&
+     @positional_anchors[:CMD_ANC_STRING_BEGIN][0] == 0) ||
+    (@positional_anchors[:CMD_ANC_LINE_BEGIN] &&
+     @positional_anchors[:CMD_ANC_LINE_BEGIN][0] == 0)
+  end
+
+  # return true if end anchor exists
+  def is_end_anchor?
+    (@positional_anchors[:CMD_ANC_STRING_END] &&
+     @positional_anchors[:CMD_ANC_STRING_END][-1] == @results.size) ||
+    (@positional_anchors[:CMD_ANC_STRING_END2] &&
+     @positional_anchors[:CMD_ANC_STRING_END2][-1] == @results.size) ||
+    (@positional_anchors[:CMD_ANC_LINE_END] &&
+     @positional_anchors[:CMD_ANC_LINE_END][-1] == @results.size)
+  end
 
   # Fixes results
   def fix

data/lib/regextest/front/char-class.rb

@@ -16,6 +16,7 @@ module Regextest::Front::CharClass
     @@id = 0   # a class variable for generating unique name of element
     @@ascii_whole_set   = nil
     @@unicode_whole_set = nil
+    @@any_code_whole_set = nil
     
     attr_reader :candidates, :offset, :length
 
@@ -80,20 +81,28 @@ module Regextest::Front::CharClass
       @is_reverse = true
     end
     
-    def set_reverse(options)
-      TstLog("CharClass set_reverse"); 
+    def set_reverse(options, default_whole_set = nil)
+      TstLog("CharClass set_reverse")
+      
+      elems = @candidates.inject([]){|result, elem| result += elem.enumerate}
       
       # Calc whole set of letters (depends on language environment)
-      @whole_set = get_whole_set(options)
+      @whole_set = default_whole_set || get_whole_set(options)
 
       # delete characters from whole set
-      whole = @whole_set.dup
-      @candidates.each do | candidate |
-        whole -= candidate.enumerate
+      whole = @whole_set.sort
+      if !default_whole_set && elems.find{| elem | !whole.bsearch {|x| elem <=> x }}
+        if options[:reg_options].is_unicode?
+          set_reverse(options, get_any_whole_set)
+        else  # ascii or default mode
+          set_reverse(options, @@unicode_whole_set)
+        end
+      else 
+        whole -= elems
+        
+        # reconstructing valid character set using TRange objects
+        @candidates = reconstruct_candidates(whole)
       end
-      
-      # reconstructing valid character set using TRange objects
-      @candidates = reconstruct_candidates(whole)
       self
     end
 
@@ -158,7 +167,7 @@ module Regextest::Front::CharClass
       ascii_set & print_set
     end
     
-    # Get whole code set of unicode
+    # Get code set of specified unicode classes( by environment variable)
     def get_unicode_whole_set
       char_set = []
       TstConstUnicodeCharSet.split("|").each do | elem |
@@ -167,6 +176,15 @@ module Regextest::Front::CharClass
       char_set
     end
     
+    # Get whole (assigned) unicode set
+    def get_any_whole_set
+      if @@any_code_whole_set
+        @@any_code_whole_set
+      else
+        @@any_code_whole_set = Regextest::Unicode.enumerate("assigned")
+      end
+    end
+    
     # enumerate nomimated letters
     def enumerate
       TstLog("CharClass enumerate")

data/lib/regextest/version.rb

@@ -1,5 +1,5 @@
 # Copyright (C) 2016 Mikio Ikoma
 
 class Regextest
-  VERSION = "0.1.6"
+  VERSION = "0.1.8"
 end

data/lib/tst-reg-test.rb

@@ -89,7 +89,7 @@ class Regextest::Test
       reg = line.gsub(/^\s*(?:x1|x2|n)\(\"|\"\)\s*$/, "")
       begin
         rc = nil
-        timeout(timeout_seconds){
+        Timeout.timeout(timeout_seconds){
           rc = eval(line)
         }
         if(rc[:result] == :ok)
@@ -144,6 +144,7 @@ class Regextest::Test
         if !line.match(/ONIG_SYNTAX_PERL/)
           if(md = line.match(/^\s*(?:x|x2|n)\s*\(.+?$/u) rescue nil)
             line.sub!(/,\s*\".+?$/, ")") rescue nil
+            line.gsub!(/\\\\x\{(\h+)\}/) { "\\u{#{$1}}" }
             lines.push line if line
           end
         else

data/regextest.gemspec

@@ -2,6 +2,12 @@
 lib = File.expand_path('../lib', __FILE__)
 $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
 require 'regextest/version'
+additional_files = [
+  "lib/regextest/unicode.rb",
+  "lib/regextest/front/case-folding.rb",
+  "lib/regextest/front/parser.rb",
+  "lib/regextest/front/bracket-parser.rb"
+]
 
 Gem::Specification.new do |spec|
   spec.name          = "regextest"
@@ -14,7 +20,8 @@ Gem::Specification.new do |spec|
   spec.homepage      = "https://bitbucket.org/ikomamik/regextest"
   spec.license       = "2-clause BSD license (see the file LICENSE.txt)"
 
-  spec.files         = `git ls-files -z`.split("\x0").reject { |f| f.match(%r{^(test|spec|features)/}) }
+  spec.files         = `git ls-files -z`.split("\x0").reject { |f| f.match(%r{^(test|spec|features)/}) } +
+                       additional_files
   spec.bindir        = "exe"
   spec.executables   = spec.files.grep(%r{^exe/}) { |f| File.basename(f) }
   spec.require_paths = ["lib"]

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: regextest
 version: !ruby/object:Gem::Version
-  version: 0.1.6
+  version: 0.1.8
 platform: ruby
 authors:
 - IKOMA, Mikio
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2016-09-17 00:00:00.000000000 Z
+date: 2016-10-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ruby
@@ -143,18 +143,15 @@ files:
 - lib/regextest/front.rb
 - lib/regextest/front/anchor.rb
 - lib/regextest/front/back-refer.rb
-- lib/regextest/front/bracket-parser.rb
 - lib/regextest/front/bracket-parser.y
 - lib/regextest/front/bracket-scanner.rb
 - lib/regextest/front/bracket.rb
 - lib/regextest/front/builtin-functions.rb
-- lib/regextest/front/case-folding.rb
 - lib/regextest/front/char-class.rb
 - lib/regextest/front/empty.rb
 - lib/regextest/front/letter.rb
 - lib/regextest/front/manage-parentheses.rb
 - lib/regextest/front/parenthesis.rb
-- lib/regextest/front/parser.rb
 - lib/regextest/front/parser.y
 - lib/regextest/front/range.rb
 - lib/regextest/front/repeat.rb
@@ -165,10 +162,13 @@ files:
 - lib/regextest/front/special-letter.rb
 - lib/regextest/regex-option.rb
 - lib/regextest/regexp.rb
-- lib/regextest/unicode.rb
 - lib/regextest/version.rb
 - lib/tst-reg-test.rb
 - regextest.gemspec
+- lib/regextest/unicode.rb
+- lib/regextest/front/case-folding.rb
+- lib/regextest/front/parser.rb
+- lib/regextest/front/bracket-parser.rb
 homepage: https://bitbucket.org/ikomamik/regextest
 licenses:
 - 2-clause BSD license (see the file LICENSE.txt)