romanbsd-curb-openuri 0.1.0

data/VERSION.yml

@@ -0,0 +1,4 @@
+--- 
+:patch: 0
+:minor: 1
+:major: 0

data/lib/curb_openuri.rb

@@ -0,0 +1,32 @@
+require 'curl_agent'
+
+module Kernel
+  private
+  alias open_uri_original_open open # :nodoc:
+
+  # makes possible to open various resources including URIs.
+  # If the first argument respond to `open' method,
+  # the method is called with the rest arguments.
+  #
+  # If the first argument is a string which begins with xxx://,
+  # it is parsed by URI.parse.  If the parsed object respond to `open' method,
+  # the method is called with the rest arguments.
+  #
+  # Otherwise original open is called.
+  #
+  # Since open-uri.rb provides URI::HTTP#open, URI::HTTPS#open and
+  # URI::FTP#open,
+  # Kernel[#.]open can accepts such URIs and strings which begins with
+  # http://, https:// and ftp://.
+  # In these case, the opened file object is extended by OpenURI::Meta.
+  def open(name, *rest, &block) # :doc:
+    if name.respond_to?(:open)
+      name.open(*rest, &block)
+    elsif name.respond_to?(:to_s) && %r{\A(ftp|https?)://} =~ name
+      CurlAgent.open(name, *rest, &block)
+    else
+      open_uri_original_open(name, *rest, &block)
+    end
+  end
+  module_function :open
+end

data/lib/curl_agent.rb

@@ -0,0 +1,94 @@
+require 'rubygems'
+require 'stringio'
+require 'curb'
+
+class CurlAgent
+  def initialize(url)
+    @curl = Curl::Easy.new(url)
+    # Defaults
+    @curl.headers['User-Agent'] = 'Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.6) Gecko/2009011913 Firefox/3.0.6'
+    @curl.follow_location = true
+    @curl.max_redirects = 2
+    @curl.enable_cookies = true
+    @curl.connect_timeout = 5
+    @curl.timeout = 30
+    @performed = false
+  end
+
+  # Do the actual fetch, after which it's possible to call body_str method
+  def perform!
+    @curl.perform
+    @performed = true
+  end
+
+  # Returns the charset of the page
+  def charset
+    perform! unless @performed
+    content_type = @curl.content_type || ''
+    charset = if content_type.match(/charset\s*=\s*([a-zA-Z0-9-]+)/ni)
+        $1
+      elsif ! body_str.nil? and (m = body_str.slice(0,1000).match(%r{<meta.*http-equiv\s*=\s*['"]?Content-Type['"]?.*?>}mi)) and
+        m[0].match(%r{content=['"]text/html.*?charset=(.*?)['"]}mi)
+        $1
+      else
+        ''
+      end.downcase
+  end
+
+  # Proxies all calls to Curl::Easy instance
+  def respond_to?(symbol)
+    @curl.respond_to?(symbol)
+  end
+
+  # Proxies all calls to Curl::Easy instance
+  def method_missing(symbol, *args)
+    @curl.send(symbol, *args)
+  end
+
+  # This method opens the URL and returns an IO object.
+  # If a block is provided, it's called with that object.
+  # You can override defaults and provide configuration directives
+  # to Curl::Easy with symbol hash keys, for example:
+  # open('http://www.example.com/', :timeout => 10)
+  # all the rest keys will be passed as headers, for example:
+  # open('http://www.example.com/', :timeout => 10, 'User-Agent'=>'curl')
+  def self.open(name, *rest, &block)
+    mode, perm, rest = scan_open_optional_arguments(*rest)
+    options = rest.shift if !rest.empty? && Hash === rest.first
+    raise ArgumentError.new("extra arguments") if !rest.empty?
+
+    unless mode == nil || mode == 'r' || mode == 'rb' || mode == File::RDONLY
+      raise ArgumentError.new("invalid access mode #{mode} (resource is read only.)")
+    end
+
+    agent = CurlAgent.new(name)
+
+    options ||= {}
+    options.each {|k, v|
+      # Strings will be passed as headers, as in original open-uri
+      next unless k.is_a? Symbol
+      agent.send("#{k}=".intern, v)
+      options.delete(k)
+    }
+    # All that's left should be considered headers
+    agent.headers.merge!(options)
+
+    agent.perform!
+    io = StringIO.new(agent.body_str)
+    if block
+      block.call(io)
+    else
+      io
+    end
+  end
+
+  def self.scan_open_optional_arguments(*rest) # :nodoc:
+    if !rest.empty? && (String === rest.first || Integer === rest.first)
+      mode = rest.shift
+      if !rest.empty? && Integer === rest.first
+        perm = rest.shift
+      end
+    end
+    return mode, perm, rest
+  end
+end

data/spec/curb_openuri_spec.rb

@@ -0,0 +1,22 @@
+require File.dirname(__FILE__) + '/spec_helper'
+
+describe "CurbOpenuri" do
+  describe "shall override Kernel::open" do
+    ['http','https','ftp'].each do |p|
+      it "shall use curl for #{p}" do
+        CurlAgent.should_receive(:open).and_return('')
+        open("#{p}://www.example.com/")
+      end
+    end
+
+    it 'shall not use curl for other protocols' do
+      CurlAgent.should_not_receive(:open)
+      lambda {open('file:///dev/null')}.should raise_error(Errno::ENOENT)
+    end
+
+    it 'shall not use curl for files' do
+      CurlAgent.should_not_receive(:open)
+      open('/dev/null') {|f| }
+    end
+  end
+end

data/spec/curl_agent_spec.rb

@@ -0,0 +1,84 @@
+require File.dirname(__FILE__) + '/spec_helper'
+
+describe "CurlAgent" do
+
+  describe 'new method' do
+    it 'shall permit to override user-agent' do
+      curl = CurlAgent.new('http://www.example.com/')
+      curl.headers['User-Agent'].should_not be_nil
+      curl.headers['User-Agent'] = 'curl'
+      curl.headers['User-Agent'].should == 'curl'
+    end
+  end
+
+  describe 'when used alone' do
+    before(:each) do
+      @mock = mock('curl_easy')
+      @headers = {'User-Agent' => 'foo'}
+      @mock.stub!(:headers).and_return(@headers)
+      @mock.stub!(:'follow_location=')
+      @mock.stub!(:'max_redirects=')
+      @mock.stub!(:'enable_cookies=')
+      @mock.stub!(:'connect_timeout=')
+      @mock.stub!(:'timeout=')
+      @mock.should_receive(:perform)
+      Curl::Easy.should_receive(:new).and_return(@mock)
+    end
+
+    it 'should recognize charset' do
+      @mock.stub!(:content_type).and_return('Content-Type: text/html;charset=utf-8')
+      curl = CurlAgent.new('http://www.example.com/')
+      curl.charset.should == 'utf-8'
+    end
+
+    it 'should recognize upper case charset' do
+      @mock.stub!(:content_type).and_return('Content-Type: text/html;charset=Windows-1251')
+      curl = CurlAgent.new('http://www.example.com/')
+      curl.charset.should == 'windows-1251'
+    end
+
+    it 'should return empty str for empty charset' do
+      @mock.stub!(:content_type).and_return('Content-Type: text/html')
+      @mock.should_receive(:body_str).once
+      curl = CurlAgent.new('http://www.example.com/')
+      curl.charset.should == ''
+    end
+
+    it 'should attempt to find charset in html' do
+      @mock.stub!(:content_type).and_return('Content-Type: text/html')
+      @mock.stub!(:body_str).and_return(<<EOF)
+      <html>
+      <head>
+      <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"/>
+      </head>
+      <body></body>
+      </html>
+EOF
+      curl = CurlAgent.new('http://www.example.com/')
+      curl.charset.should == 'iso-8859-1'
+    end
+  end
+
+  describe 'when used with open' do
+    before(:each) do
+      @headers = {'User-Agent'=>'foo'}
+      @curl = mock('curl')
+      @curl.stub!(:headers).and_return(@headers)
+      @curl.stub!(:perform!)
+      @curl.stub!(:body_str).and_return('')
+      CurlAgent.should_receive(:new).and_return(@curl)
+    end
+
+    it 'shall permit to specify user-agent' do
+      @curl.headers['User-Agent'].should_not == 'curl'
+      CurlAgent.open('http://www.example.com/', 'User-Agent'=>'curl')
+      @curl.headers['User-Agent'].should == 'curl'
+    end
+
+    it 'shall permit to override timeout' do
+      @curl.should_receive(:'timeout=').once.with(10)
+      CurlAgent.open('http://www.example.com/', :timeout => 10)
+    end
+
+  end
+end

data/spec/spec_helper.rb

@@ -0,0 +1,9 @@
+require 'spec'
+
+$LOAD_PATH.unshift(File.dirname(__FILE__))
+
+require 'curb_openuri'
+
+Spec::Runner.configure do |config|
+  
+end

metadata

@@ -0,0 +1,68 @@
+--- !ruby/object:Gem::Specification 
+name: romanbsd-curb-openuri
+version: !ruby/object:Gem::Version 
+  version: 0.1.0
+platform: ruby
+authors: 
+- Roman Shterenzon
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2009-02-18 00:00:00 -08:00
+default_executable: 
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: curb
+  type: :runtime
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: 0.1.4
+    version: 
+description: open-uri drop-in replacement that uses curb
+email: romanbsd@yahoo.com
+executables: []
+
+extensions: []
+
+extra_rdoc_files: []
+
+files: 
+- VERSION.yml
+- lib/curb_openuri.rb
+- lib/curl_agent.rb
+- spec/curb_openuri_spec.rb
+- spec/spec_helper.rb
+- spec/curl_agent_spec.rb
+has_rdoc: true
+homepage: http://github.com/romanbsd/curb-openuri
+post_install_message: 
+rdoc_options: 
+- --inline-source
+- --charset=UTF-8
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.2.0
+signing_key: 
+specification_version: 2
+summary: open-uri drop-in replacement that uses curb
+test_files: []
+