RubyGems - tidy - Versions diffs - 1.0.0 - Mend

tidy 1.0.0

Files changed (13) hide show

data/CHANGES ADDED

@@ -0,0 +1,40 @@
+V 1.0.0
+  Oct/18/2004 - Declared stable, now packaged as a RubyGem
+VB.2.5
+  - Auto-detection and ENV variables removed, adds too much complexity
+    Module location is now specified by defining $TIDYLIB before require 'tidy'
+    $TIDYLIB is the system path to the library (ex: /usr/lib/tidylib.so)
+    Raises an error if $TIDYLIB is not defined
+V.B2.4
+  Now uses ENV['TIDY_LIB'] instead of $TIDY_LIB
+V B2.3
+  - Added library auto-detection
+    a) use $TIDY_LIB if defined
+    b) Search $: and ENV['PATH'] paths for (tidy|tidylib|libtidy|htmltidy).(dll|so)
+    c) Raise an error if not found, otherwise loads
+V.B2.2
+  - Tidy.to_b added
+V.B2.1
+  - load_config method added
+V.B2
+  - load(path) method added, library must be loaded explicity
+    this avoids hard coding the library name
+    require 'tidy'
+    Tidy.load('path/to/tidylib.so')
+V.B1
+  First release

data/MANIFEST ADDED

@@ -0,0 +1,12 @@
+lib/tidy.rb
+lib/tidy/tidybuf.rb
+lib/tidy/tidyerr.rb
+lib/tidy/tidylib.rb
+lib/tidy/tidyobj.rb
+lib/tidy/tidyopt.rb
+test/usage.rb
+CHANGES
+MANIFEST
+README.txt.en
+tidy.gemspec
+VERSION

data/README.txt.en ADDED

@@ -0,0 +1,35 @@
+Tidy README
+============
+  Ruby interface to HTML Tidy Library Project (http://tidy.sf.net)
+Requirements
+------------
+  * Recent version of Ruby
+  * RubyGems 1.8+ (http://rubygems.rubyforge.org)
+  * HTML Tidy Library (compiled)
+Install
+-------
+  - Download library from http://tidy.sf.net (pre-compiled versions available)
+  - Download the tidy .gem file to a directory. Then type:
+      $ gem install tidy .gem
+  - Open test/usage.rb, edit $TIDYLIB to point to your compiled tidy library, run.
+Usage
+-----
+  See API docs
+License
+-------
+  Distributes under the same terms as Ruby
+  http://www.ruby-lang.org/en/LICENSE.txt
+Kevin Howe <kh at newclear.ca>

data/VERSION ADDED

	@@ -0,0 +1 @@
1	+ 1.0.0

data/lib/tidy.rb ADDED

@@ -0,0 +1,65 @@
+# Ruby interface to HTML Tidy Library Project (http://tidy.sf.net)
+#
+# = Usage
+#
+#   $TIDYLIB = '/usr/lib/tidylib.so'
+#   require 'rubygems'
+#   require_gem 'tidy'
+#   html = '<html><title>title</title>Body</html>'
+#   xml = Tidy.open(:show_warnings=>true) do |tidy|
+#     tidy.options.output_xml = true
+#     puts tidy.options.show_warnings
+#     xml = tidy.clean(html)
+#     puts tidy.errors
+#     puts tidy.diagnostics
+#     xml
+#   end
+#   puts xml
+#
+# Author::  Kevin Howe
+# License:: Distributes under the same terms as Ruby
+#
+module Tidy
+  require 'dl/import'
+  require 'dl/struct'
+  require 'tidy/tidybuf'
+  require 'tidy/tidyerr'
+  require 'tidy/tidylib'
+  require 'tidy/tidyobj'
+  require 'tidy/tidyopt'
+  module_function
+  # Return a Tidyobj instance
+  #
+  def new(options=nil)
+    Tidyobj.new(options)
+  end
+  # With no block, open is a synonym for Tidy.new .
+  # If a block is present, it is passed aTidy as a parameter.
+  # aTidyObj.release is ensured at end of the block
+  #
+  def open(options=nil)
+    tidy = Tidy.new(options)
+    if block_given?
+      begin
+        yield tidy
+      ensure
+        tidy.release
+      end
+    else
+      tidy
+    end
+  end
+  # Convert to boolean.
+  # 0, false and nil return false, anything else true
+  #
+  def to_b(value)
+    return false if [0,false,nil].include?(value)
+    true
+  end
+end

data/lib/tidy/tidybuf.rb ADDED

@@ -0,0 +1,42 @@
+# Buffer structure
+#
+class Tidybuf
+  extend DL::Importable
+  # Access TidyBuffer instance
+  #
+  attr_reader(:struct)
+  # Mimic TidyBuffer
+  #
+  TidyBuffer = struct [
+    "byte* bp",
+    "uint size",
+    "uint allocated",
+    "uint next"
+  ]
+  def initialize()
+    @struct = TidyBuffer.malloc
+  end
+  # Free current contents and zero out
+  #
+  def free()
+    Tidylib.buf_free(@struct)
+  end
+  # Convert to array
+  #
+  def to_a
+    to_s.split("\r\n")
+  end
+  # Convert to string
+  #
+  def to_s
+    @struct.bp.to_s
+  end
+end

data/lib/tidy/tidyerr.rb ADDED

@@ -0,0 +1,30 @@
+# Parameterized error message
+#
+class Tidyerr < String
+  # Error parameter
+  #
+  attr_reader :severity, :line, :column, :message
+  # Create new instance
+  #
+  def initialize(error)
+    super(error.to_s)
+    parameterize
+  end
+  # Parse error message into parameters (where applicable)
+  #
+  def parameterize()
+    if to_str[0,4] == 'line'
+      tokens    = to_str.split(' ', 7)
+      @severity = tokens[5][0,1] # W or E
+      @line     = tokens[1].to_i
+      @column   = tokens[3].to_i
+      @message  = tokens[6]
+    end
+  end
+  protected :parameterize
+end

data/lib/tidy/tidylib.rb ADDED

@@ -0,0 +1,115 @@
+# Ruby wrapper for HTML Tidy Library Project (http://tidy.sf.net)
+#
+module Tidylib
+  extend DL::Importable
+  module_function
+  # Load library (requires that $TIDYLIB be defined)
+  #
+  def load()
+    raise LoadError, 'Tidy requires that $TIDYLIB be defined' if $TIDYLIB.nil?
+    begin
+      dlload($TIDYLIB)
+    rescue
+      raise LoadError, sprintf('Unable to load %s', $TIDYLIB)
+    end
+    extern "void *tidyCreate()"
+    extern "void tidyBufFree(void*)"
+    extern "int tidyCleanAndRepair(void*)"
+    extern "int tidyLoadConfig(void*, char*)"
+    extern "int tidyOptGetIdForName(char*)"
+    extern "char tidyOptGetValue(void*, unsigned int)"
+    extern "int tidyOptParseValue(void*, char*, char*)"
+    extern "int tidyParseString(void*, char*)"
+    extern "void tidyRelease(void*)"
+    extern "char* tidyReleaseDate()"
+    extern "int tidyRunDiagnostics(void*)"
+    extern "int tidySaveBuffer(void*, void*)"
+    extern "int tidySetErrorBuffer(void*, void*)"
+  end
+  # tidyBufFree
+  #
+  def buf_free(buf)
+    tidyBufFree(buf)
+  end
+  # tidyCreate
+  #
+  def create()
+    tidyCreate()
+  end
+  # tidyCleanAndRepair
+  #
+  def clean_and_repair(doc)
+    tidyCleanAndRepair(doc)
+  end
+  # tidyLoadConfig
+  #
+  def load_config(doc, file)
+    tidyLoadConfig(doc, file.to_s)
+  end
+  # tidyOptParseValue
+  #
+  def opt_parse_value(doc, name, value)
+    tidyOptParseValue(doc, translate_name(name), value.to_s)
+  end
+  # tidyOptGetValue (returns true/false instead of 1/0)
+  #
+  def opt_get_value(doc, name)
+    value = tidyOptGetValue(doc, tidyOptGetIdForName(translate_name(name)))
+    Tidy.to_b(value)
+  end
+  # tidyParseString
+  #
+  def parse_string(doc, str)
+    tidyParseString(doc, str.to_s)
+  end
+  # tidyRelease
+  #
+  def release(doc)
+    tidyRelease(doc)
+  end
+  # tidyReleaseDate
+  #
+  def release_date()
+    tidyReleaseDate()
+  end
+  # tidyRunDiagnostics
+  #
+  def run_diagnostics(doc)
+    tidyRunDiagnostics(doc)
+  end
+  # tidySaveBuffer
+  #
+  def save_buffer(doc, buf)
+    tidySaveBuffer(doc, buf)
+  end
+  # tidySetErrorBuffer
+  #
+  def set_error_buffer(doc, buf)
+    tidySetErrorBuffer(doc, buf)
+  end
+  # Convert to string replacing underscores with dashes.
+  # :output_xml becomes 'output-xml'
+  #
+  def translate_name(name)
+    name.to_s.sub('_', '-')
+  end
+end
+Tidylib.load

data/lib/tidy/tidyobj.rb ADDED

@@ -0,0 +1,103 @@
+# Ruby interface to Tidylib
+#
+class Tidyobj
+  # Diagnostics Buffer (Array of String)
+  #
+  attr_reader(:diagnostics)
+  # Access the tidy instance
+  #
+  attr_reader(:doc)
+  # Error Buffer (Array of Tidyerr)
+  #
+  attr_reader(:errors)
+  # Options interface (Tidyopt)
+  #
+  attr_reader(:options)
+  # Construct a new instance.
+  # Receives a hash of options to be set
+  #
+  def initialize(options=nil)
+    @diagnostics = Array.new
+    @doc = Tidylib.create
+    @errors = Array.new
+    @errbuf = Tidybuf.new
+    @outbuf = Tidybuf.new
+    @options = Tidyopt.new(@doc)
+    rc = Tidylib.set_error_buffer(@doc, @errbuf.struct)
+    verify_severe(rc)
+    unless options.nil?
+      options.each { |name, value| Tidylib.opt_parse_value(@doc, name, value) }
+    end
+  end
+  # Clean and Repair
+  #
+  def clean(str)
+    verify_doc
+    rc = -1
+    # Clean and repair the string
+    #
+    rc = Tidylib.parse_string(@doc, str)                                               # Parse the input
+    rc = Tidylib.clean_and_repair(@doc) if rc >= 0                                     # Tidy it up!
+    rc = (Tidylib.opt_parse_value(@doc, :force_output, true) == 1 ? rc : -1) if rc > 1 # If error, force output
+    rc = Tidylib.save_buffer(@doc, @outbuf.struct) if rc >= 0                          # Pretty Print
+    verify_severe(rc)
+    # Save and clear output/errors
+    #
+    output = @outbuf.to_s
+    @errors = @errbuf.to_a.collect { |e| Tidyerr.new(e) }
+    @outbuf.free
+    @errbuf.free
+    # Save diagnostics
+    #
+    rc = Tidylib.run_diagnostics(@doc)
+    verify_severe(rc)
+    @diagnostics = @errbuf.to_a
+    @errbuf.free
+    output
+  end
+  # Load a tidy config file
+  #
+  def load_config(file)
+    verify_doc
+    rc = Tidylib.load_config(@doc, file)
+    case rc
+      when -1 then raise LoadError, sprintf('%s does not exist', file)
+      when  1 then raise LoadError, sprintf('errors parsing %s', file)
+    end
+    rc
+  end
+  # Clear the tidy instance
+  #
+  def release()
+    verify_doc
+    Tidylib.release(@doc)
+    @doc = nil
+  end
+  # Raise an error if the tidy document is invalid
+  #
+  def verify_doc()
+    raise TypeError, 'Invalid Tidy document' unless @doc.class == DL::PtrData
+  end
+  # Raise severe error based on tidy status value
+  #
+  def verify_severe(rc)
+    raise sprintf("A severe error (%d) occurred.\n", rc) if rc < 0
+  end
+  protected :verify_doc, :verify_severe
+end

data/lib/tidy/tidyopt.rb ADDED

@@ -0,0 +1,31 @@
+# Ruby interface to Tidylib options
+#
+class Tidyopt
+  # Construct a new instance
+  #
+  def initialize(doc)
+    @doc = doc
+  end
+  # Reader for options (Hash syntax)
+  #
+  def [](name)
+    Tidylib.opt_get_value(@doc, name)
+  end
+  # Writer for options (Hash syntax)
+  #
+  def []=(name, value)
+    Tidylib.opt_parse_value(@doc, name, value)
+  end
+  # Reader/Writer for options (Object syntax)
+  #
+  def method_missing(name, value=:none, *args)
+    name = name.to_s.sub('=', '')
+    return self[name] if value == :none
+    self[name] = value
+  end
+end

data/test/usage.rb ADDED

@@ -0,0 +1,13 @@
+$TIDYLIB = '/usr/lib/tidylib.so'
+require 'rubygems'
+require_gem 'tidy'
+html = '<html><title>title</title>Body</html>'
+xml = Tidy.open(:show_warnings=>true) do |tidy|
+  tidy.options.output_xml = true
+  puts tidy.options.show_warnings
+  xml = tidy.clean(html)
+  puts tidy.errors
+  puts tidy.diagnostics
+  xml
+end
+puts xml

data/tidy.gemspec ADDED

@@ -0,0 +1,13 @@
+require 'rubygems'
+spec = Gem::Specification.new do |s|
+  s.name = "tidy"
+  s.version = File.read("VERSION").strip
+  s.summary = "Ruby interface to HTML Tidy Library Project"
+  s.description = s.summary+" (http://tidy.sf.net)"
+  s.files = Dir.glob("**/*").delete_if { |item| item.include?("CVS") }
+  s.require_path = 'lib'
+  s.autorequire = 'tidy'
+  s.has_rdoc = true
+  s.author = "Kevin Howe"
+  s.homepage = "http://www.newclear.ca/ruby/tidy/"
+end

metadata ADDED

@@ -0,0 +1,50 @@
+--- !ruby/object:Gem::Specification
+rubygems_version: 0.8.1
+specification_version: 1
+name: tidy
+version: !ruby/object:Gem::Version
+  version: 1.0.0
+date: 2004-11-04
+summary: Ruby interface to HTML Tidy Library Project
+require_paths:
+  - lib
+author: Kevin Howe
+email:
+homepage: http://www.newclear.ca/ruby/tidy/
+rubyforge_project:
+description: Ruby interface to HTML Tidy Library Project (http://tidy.sf.net)
+autorequire: tidy
+default_executable:
+bindir: bin
+has_rdoc: true
+required_ruby_version: !ruby/object:Gem::Version::Requirement
+  requirements:
+    -
+      - ">"
+      - !ruby/object:Gem::Version
+        version: 0.0.0
+  version:
+platform: ruby
+files:
+  - lib
+  - test
+  - CHANGES
+  - MANIFEST
+  - README.txt.en
+  - VERSION
+  - tidy.gemspec
+  - lib/tidy
+  - lib/tidy.rb
+  - lib/tidy/tidyopt.rb
+  - lib/tidy/tidyobj.rb
+  - lib/tidy/tidylib.rb
+  - lib/tidy/tidyerr.rb
+  - lib/tidy/tidybuf.rb
+  - test/usage.rb
+test_files: []
+rdoc_options: []
+extra_rdoc_files: []
+executables: []
+extensions: []
+requirements: []
+dependencies: []