opal-onigmo 0.1.0 → 0.1.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 42b7f0d79bf4c8caa956996c2c3e719e830e4b7d5cdc36a37f259bad0817d4be
-  data.tar.gz: 5d696a5c11e17b61aca28cefc0c275295dd9c5ffafd4e59cb173578e515d30db
+  metadata.gz: ec1caa38621cd815637b96284a30f5fb69f3049ccc759ff3e69a67ffc549c36a
+  data.tar.gz: 7d544017614820e67e952612a2fd5ef2d5496cfdcc01ed18b174630c2051f578
 SHA512:
-  metadata.gz: 6183bcc0febb8a96b3380a91ba1a1073f77ed304acc19619adf2ba570108fc68085cd3517e32abd4339c647564d8ce8cc203639fb3ff8ee80a6411ccf5d53ad3
-  data.tar.gz: e45bb5cbccbb07eb87ee9de044769aada834639012fe0fdfc8cfa0115308aaa6ad5c8548c9b0698eef34c963cd6da6be0c6e389b52381d9c1a718fc0c7cb339c
+  metadata.gz: e0d515da7c5d167ba85fee1cff7ffda106f6072568c0660c9e0997f0cff3499f621ae1472ac3c7175a2581a7b1b1c7f3ae33fea9868768bc3a250de7d419ce8c
+  data.tar.gz: f026ce0c670a78d70ff41f53c32e89f90c91e6ed68a65fb1cc2cd18d956aec49fbf5e1b3c1f82dba1a9947c79268f0388fc5ebb34c36901814841973a3f160fd

data/.gitmodules

@@ -1,3 +1,3 @@
 [submodule "Onigmo"]
 	path = Onigmo
-	url = https://github.com/interscript/onigmo
+	url = https://github.com/interscript/Onigmo

data/Gemfile

@@ -11,3 +11,4 @@ gem "pry"
 gem "opal", "~> 1.0"
 gem "opal-webassembly"
 gem "opal-rspec"
+gem "mini_racer"

data/Gemfile.lock

@@ -11,10 +11,10 @@ GEM
     coderay (1.1.3)
     concurrent-ruby (1.1.7)
     diff-lcs (1.4.4)
+    libv8 (8.4.255.0)
     method_source (1.0.0)
-    opal (1.0.3)
-      ast (>= 2.3.0)
-      parser (~> 2.6)
+    mini_racer (0.3.1)
+      libv8 (~> 8.4.255)
     opal-rspec (0.8.0.alpha2)
       opal (>= 0.11, < 1.1)
       opal-sprockets
@@ -54,7 +54,8 @@ PLATFORMS
   ruby
 
 DEPENDENCIES
-  opal (~> 1.0)
+  mini_racer
+  opal!
   opal-onigmo!
   opal-rspec
   opal-webassembly

data/Rakefile

@@ -9,7 +9,7 @@ Opal::RSpec::RakeTask.new("spec-opal") do |server, task|
   require 'opal/onigmo'
 end
 
-task :generate => ['opal/onigmo/constants.rb', 'opal/onigmo/onigmo.wasm'] do
+task :generate => ['opal/onigmo/constants.rb', 'opal/onigmo/onigmo-wasm.wasm'] do
 end
 
 def render_template(template, output, vars)
@@ -28,8 +28,11 @@ file 'opal/onigmo/constants.rb' => ['opal/onigmo/constants.rb.erb', 'Onigmo/onig
   )
 end
 
-file 'opal/onigmo/onigmo.wasm' => ['Onigmo/build-wasm.sh'] do |task|
-  sh "cd Onigmo; sh ./build-wasm.sh; mv onigmo.wasm #{task.name}"
+file 'opal/onigmo/onigmo-wasm.wasm' => ['Onigmo/build-wasm.sh'] do |task|
+  if File.exist? __dir__+"/../../wasi-sdk-11.0/share/wasi-sysroot/"
+    env = "BUILD=wasi WASI_PATH=#{__dir__}/../../wasi-sdk-11.0/share/wasi-sysroot/"
+  end
+  sh "cd Onigmo; #{env} sh ./build-wasm.sh; mv onigmo.wasm ../opal/onigmo/onigmo-wasm.wasm"
 end
 
 task :default => [:spec, "spec-opal"]

data/bin/opal-repl-onigmo

@@ -0,0 +1,4 @@
+#!/usr/bin/env ruby
+
+require 'bundler/setup'
+load "#{__dir__}/../exe/opal-repl-onigmo"

data/exe/opal-repl-onigmo

@@ -0,0 +1,8 @@
+#!/usr/bin/env ruby
+
+require 'opal/repl'
+require 'opal/onigmo'
+
+Opal.append_paths '.'
+
+repl = Opal::REPL.new.run ARGV.first

data/lib/opal/onigmo/version.rb

@@ -1,5 +1,5 @@
 module Opal
   module Onigmo
-    VERSION = "0.1.0"
+    VERSION = "0.1.2"
   end
 end

data/opal-onigmo.gemspec

@@ -27,5 +27,5 @@ Gem::Specification.new do |spec|
   spec.executables   = spec.files.grep(%r{^exe/}) { |f| File.basename(f) }
   spec.require_paths = ["lib"]
 
-  spec.add_dependency "opal-webassembly"
+  spec.add_dependency "opal-webassembly", "~> 0.1"
 end

data/opal/onigmo.rb

@@ -4,4 +4,17 @@ require 'onigmo/ffi'
 require 'onigmo/regexp'
 
 module Onigmo
+  def self.buffer(str)
+    Onigmo::FFI.context do
+      @buffer_size ||= 1024
+      @buffer ||= ::FFI::Pointer.new(:uint8, 640000) # 640 kB of RAM will be enough for everybody
+      current_size = @buffer_size
+      while str.length*2 > @buffer_size
+        @buffer_size *= 2
+      end
+      @buffer.resize(@buffer_size) if current_size != @buffer_size
+      @buffer.write_string(str)
+      @buffer
+    end
+  end
 end

data/opal/onigmo/core_ext.rb

@@ -18,4 +18,57 @@ class String
       match_before_onigmo?(re, pos, &block)
     end
   end
+
+  alias gsub_before_onigmo gsub
+  def gsub(from, to = undefined, &block)
+    if Onigmo::Regexp === from
+      out = []
+      index = 0
+      loop do
+        o = from.evaluate(self, index)
+        if o == nil # not found anymore
+          out << self[index..-1]
+          break
+        end
+        bgn, fin = from.ffi_region[0].map { |i| i/2 }
+        matches = from.matches
+
+        md = MatchData.new(from, from.js_matches(self))
+
+        out << self[index...bgn]
+
+        if block && `to === undefined`
+          out << block.(matches[0])
+        else
+          _to = to.JS.replace(%x{/([\\]+)([0-9+&`'\\])/g}) do |original, slashes, command|
+            if slashes.length % 2 == 0
+              original
+            else
+              slashes = slashes[0..-2]
+              case command
+              when "+", "&", "`", "'"
+                raise NotImplementedError
+              when "\\"
+                slashes + "\\"
+              else
+                slashes + ( matches[command.to_i] || "" )
+              end
+            end
+          end
+          out << _to
+        end
+
+        if bgn == fin
+          fin += 1
+          out << self[bgn]
+        end
+
+        index = fin
+      end
+
+      out.join
+    else
+      gsub_before_onigmo(from, to, &block)
+    end
+  end
 end

data/opal/onigmo/ffi.rb

@@ -10,6 +10,7 @@ module Onigmo
     ffi_lib "onigmo/onigmo-wasm"
 
     class Regexp < ::FFI::Struct
+      # This is defined mostly for debugging purposes
       layout :p, :pointer,
              :used, :uint,
              :alloc, :uint,
@@ -77,13 +78,15 @@ module Onigmo
 
       def self.default_compile_info
         @default ||= Onigmo::FFI.context do
+          casefold = Onigmo::FFI.library.exports[:OnigDefaultCaseFoldFlag]
+          casefold = casefold.value if casefold.respond_to? :value
           new.tap do |ci|
             ci[:num_of_elements] = 5
             ci[:pattern_enc] = Onigmo::FFI.library.exports[:OnigEncodingUTF_16LE]
             ci[:target_enc] = Onigmo::FFI.library.exports[:OnigEncodingUTF_16LE]
             ci[:syntax] = Onigmo::FFI.library.exports[:OnigSyntaxRuby]
             ci[:option] = Onigmo::FFI::ONIG_OPTION_NONE # To be overwritten
-            ci[:case_fold_flag] = Onigmo::FFI.library.exports[:OnigDefaultCaseFoldFlag].value
+            ci[:case_fold_flag] = casefold
           end
         end
       end
@@ -101,13 +104,19 @@ module Onigmo
       layout :enc, :pointer,
              :par, :pointer,
              :par_end, :pointer
+
+      def self.cached
+        @cached ||= new
+      end
     end
 
     attach_function :onig_new_deluxe, [RegexpPtr, :pointer, :pointer, CompileInfo, ErrorInfo], :int
-    attach_function :onig_region_new, [], Region
     attach_function :onig_match, [Regexp, :pointer, :pointer, :pointer, Region, :uint], :int
     attach_function :onig_search, [Regexp, :pointer, :pointer, :pointer, :pointer, Region, :uint], :int
+    attach_function :onig_free, [Regexp], :void
 
+    attach_function :onig_region_new, [], Region
+    attach_function :onig_region_free, [Region, :int], :void
   end
 
   Onigmo::FFI.library.memory.grow(128)

data/opal/onigmo/regexp.rb

@@ -24,28 +24,34 @@ module Onigmo
     end
 
     def initialize pattern, options = ''
+      if pattern.is_a? Onigmo::Regexp
+        pattern = pattern.pattern
+      end
+
       @lastIndex = 0
       @pattern = pattern.encode("UTF-16LE")
       @options = options
 
-      out = Onigmo::FFI.context do
-        @ffi_regexpptr = Onigmo::FFI::RegexpPtr.new
+      out = nil
+      Onigmo::FFI.context do
+        ffi_regexpptr = Onigmo::FFI::RegexpPtr.new
         ffi_options = Onigmo::FFI::CompileInfo.default_compile_info
-        ffi_errorinfo = Onigmo::FFI::ErrorInfo.new
+        ffi_errorinfo = Onigmo::FFI::ErrorInfo.cached
 
-        ffi_pattern = ::FFI::Pointer.from_string(@pattern)
+        ffi_pattern = Onigmo.buffer(@pattern)
         ffi_pattern_end = ffi_pattern + (@pattern.length * 2)
 
-        ffi_options.options = Onigmo::FFI::ONIG_OPTION_NONE
+        ffi_options.options  = Onigmo::FFI::ONIG_OPTION_NONE
         ffi_options.options |= Onigmo::FFI::ONIG_OPTION_IGNORECASE if options.include? "i"
         ffi_options.options |= Onigmo::FFI::ONIG_OPTION_MULTILINE if options.include? "m"
         ffi_options.options |= Onigmo::FFI::ONIG_OPTION_EXTEND if options.include? "x"
 
-        Onigmo::FFI.onig_new_deluxe(@ffi_regexpptr, ffi_pattern, ffi_pattern_end,
+        out = Onigmo::FFI.onig_new_deluxe(ffi_regexpptr, ffi_pattern, ffi_pattern_end,
                                     ffi_options, ffi_errorinfo)
-      end
 
-      @ffi_regexp = @ffi_regexpptr.value
+        @ffi_regexp = ffi_regexpptr.value
+        ffi_regexpptr.free
+      end
 
       @exec = proc { |re| js_exec(re) }
 
@@ -54,6 +60,8 @@ module Onigmo
       end
     end
 
+    attr_accessor :pattern, :options
+
     def ffi_evaluate string, offset = 0
       offset ||= 0
 
@@ -61,21 +69,37 @@ module Onigmo
 
       string = string.encode("UTF-16LE")
 
-      Onigmo::FFI.context do
-        ffi_string = ::FFI::Pointer.from_string(string)
+      out = Onigmo::FFI.context do
+        ffi_string = Onigmo.buffer(string)
         ffi_string_end = ffi_string + (string.length * 2)
         ffi_string_offset = ffi_string + offset * 2
 
         Onigmo::FFI.onig_search(@ffi_regexp, ffi_string, ffi_string_end, ffi_string_offset,
                                 ffi_string_end, @ffi_region, Onigmo::FFI::ONIG_OPTION_NONE)
       end
-    end
 
-    def ffi_region
-      @ffi_region[:num_regs].times.map do |i|
+      @region = @ffi_region[:num_regs].times.map do |i|
         [@ffi_region[:beg].get(:long, i*4),
          @ffi_region[:end].get(:long, i*4) ]
       end
+
+      Onigmo::FFI.onig_region_free(@ffi_region, 1)
+
+      @matches = @region.map { |b,e| string[b/2...e/2] }
+
+      out
+    end
+
+    def ffi_region
+      @region
+    end
+
+    def matches
+      @matches
+    end
+
+    def ffi_free
+      Onigmo::FFI.onig_free(@ffi_regexp)
     end
 
     def evaluate string, offset = 0
@@ -87,21 +111,17 @@ module Onigmo
       end
     end
 
-    def matches(str)
-      ffi_region.map { |b,e| str[b/2...e/2] }
-    end
-
     def js_matches(str)
-      matches = matches(str)
+      ms = matches
       region = ffi_region
       out = {
         input: str,
         index: region[0][0]/2,
-        length: matches.length,
+        length: ms.length,
         slice: proc { |i| matches[i..-1] }
       }
       out = out.to_n
-      matches.each_with_index { |m,i| `out[#{i}] = #{m}` }
+      ms.each_with_index { |m,i| `out[#{i}] = #{m}` }
       @lastIndex = region[0][1]/2
       out
     end
@@ -114,5 +134,10 @@ module Onigmo
         nil.to_n
       end
     end
+
+    def reset
+      @lastIndex = 0
+      self
+    end
   end
 end

data/spec-opal/regexp_spec.rb

@@ -20,10 +20,24 @@ RSpec.describe Onigmo::Regexp do
   end
 
   it "should support String#gsub" do
+    require 'onigmo/core_ext'
     re = Onigmo::Regexp.new('aa')
     expect('aababaaaabbaaaba'.gsub(re, 'ZZ')).to eq('ZZbabZZZZbbZZaba')
   end
 
+  it "should support String#gsub with $ correctly" do
+    require 'onigmo/core_ext'
+    re = Onigmo::Regexp.new('$')
+    expect('lol'.gsub(re, '2')).to eq('lol2')
+  end
+
+  it "should support String#gsub with backparameters correctly" do
+    require 'onigmo/core_ext'
+    re = Onigmo::Regexp.new('(?<=te)(..)(?=st)')
+    expect('texxst'.gsub(re, '\1\1')).to eq('texxxxst')
+    expect('teyyst'.gsub(re) { "123" }).to eq('te123st')
+  end
+
   it "should support String#sub" do
     re = Onigmo::Regexp.new('aa')
     expect('aababaaaabbaaaba'.sub(re, 'ZZ')).to eq('ZZbabaaaabbaaaba')

metadata

@@ -1,11 +1,11 @@
 --- !ruby/object:Gem::Specification
 name: opal-onigmo
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.2
 platform: ruby
 authors:
 - Ribose Inc.
-autorequire:
+autorequire: 
 bindir: exe
 cert_chain: []
 date: 2020-10-05 00:00:00.000000000 Z
@@ -14,21 +14,22 @@ dependencies:
   name: opal-webassembly
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '0.1'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '0.1'
 description: Execute regexps in Opal with a native Ruby regexp engine - Onigmo - compiled
   using clang to WebAssembly
 email:
 - open.source@ribose.com
-executables: []
+executables:
+- opal-repl-onigmo
 extensions: []
 extra_rdoc_files: []
 files:
@@ -44,7 +45,9 @@ files:
 - README.adoc
 - Rakefile
 - bin/console
+- bin/opal-repl-onigmo
 - bin/setup
+- exe/opal-repl-onigmo
 - lib/opal/onigmo.rb
 - lib/opal/onigmo/version.rb
 - opal-onigmo.gemspec
@@ -64,7 +67,7 @@ metadata:
   homepage_uri: https://github.com/interscript/opal-onigmo
   source_code_uri: https://github.com/interscript/opal-onigmo
   changelog_uri: https://github.com/interscript/opal-onigmo/commits/master
-post_install_message:
+post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
@@ -79,8 +82,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.3
-signing_key:
+rubygems_version: 3.1.2
+signing_key: 
 specification_version: 4
 summary: Execute regexps with Opal using Onigmo
 test_files: []