tomtaylor-geo-spider 0.2.1 → 0.2.2

data/lib/geo-spider.rb

@@ -5,6 +5,7 @@ require 'rubygems'
 require 'hpricot'
 require 'open-uri'
 
+require 'geo-spider/exceptions'
 require 'geo-spider/page'
 require 'geo-spider/site'
 

data/lib/geo-spider/extractors/base.rb

@@ -5,6 +5,7 @@ module GeoSpider
     class Base
       
       def initialize(element)
+        raise InvalidElement if element.class != Hpricot::Elem || element.nil?
         @element = element
       end
   

data/lib/geo-spider/site.rb

@@ -28,7 +28,7 @@ module GeoSpider
           seen << url
           next_links = (page.internal_links - seen - queue) # only add internal links that we've not seen or already have queued.
           queue.concat(next_links)
-        rescue Timeout::Error, OpenURI::HTTPError
+        rescue Timeout::Error, OpenURI::HTTPError, InvalidElement
           next
         end
       end

data/lib/geo-spider/version.rb

@@ -2,7 +2,7 @@ module GeoSpider
   module VERSION #:nodoc:
     MAJOR = 0
     MINOR = 2
-    TINY  = 1
+    TINY  = 2
 
     STRING = [MAJOR, MINOR, TINY].join('.')
   end

data/spec/geo-spider/page_spec.rb

@@ -135,4 +135,17 @@ describe Page, "which is parsing a page with a string in a URL that happens to m
     @page.locations.should be_empty
   end
   
+end
+
+describe Page, "which is a parsing a page which doesn't contain the specific content_css_selector" do
+  
+  before(:each) do
+    OpenURI.should_receive(:open_uri).and_return(page_as_string('page_with_links.html'))
+    @page = Page.new("http://www.example.com", :content_css_selector => "notreal")
+  end
+  
+  it "should raise InvalidElement" do
+    lambda { @page.locations }.should raise_error(InvalidElement)
+  end
+  
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: tomtaylor-geo-spider
 version: !ruby/object:Gem::Version 
-  version: 0.2.1
+  version: 0.2.2
 platform: ruby
 authors: 
 - Tom Taylor