upmark 0.9.0 → 0.10.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: ebfa7a03ebe86bf8d70709fe99002518ee9f7237
-  data.tar.gz: be21325fe8ab8cd94d9ba4a4df1fcb0025bfb64b
+  metadata.gz: 1e6c728e71452e9983f7fe6ef59948cca6fd691b
+  data.tar.gz: 87daa0b2e4e6105b43e88502b6f645a2e84e423a
 SHA512:
-  metadata.gz: d4eab42d3640a2f9011daf71244c97036f9e92bce80cf351fe9eeda00b1535f0ef5a0471440ec1c4dd9d5c0ac8c76d9c0263661f9c2c6f9d1425a3845296c901
-  data.tar.gz: 3ed9e855887dbab065c586517669657e7c4cb56311b1e5237161111e64853043c2e44097f2bff95b5841ecd5fe3424cd29811940fc07fe8e860472a4ad77b92c
+  metadata.gz: 25919f578df35140965011b7c2cb2ce57308f52adf0a05373ba26f4aff5973d3c810c6fb53b473efc9ce50744325b12a9331ee0a1934b715e61be6a8df830cf5
+  data.tar.gz: 018f24f9bf5db7e63e79c5301c3ee6035d7028ff9521951a4c03de346358b32a14fc716ead14ac8da41ad146d2e24edd0543a2ee26f42f5610fd0c3734102f01

data/lib/upmark/parser/xml.rb

@@ -79,14 +79,15 @@ module Upmark
       end
 
       rule(:double_quoted_attribute_value) do
-        (str('"').absent? >> (match(/[^<&]/) | entity_ref)).repeat
+        (str('"').absent? >> (match(/[^<]/) | string_entity | numeric_entity)).repeat
       end
 
       rule(:single_quoted_attribute_value) do
-        (str("'").absent? >> (match(/[^<&]/) | entity_ref)).repeat
+        (str("'").absent? >> (match(/[^<]/) | string_entity | numeric_entity)).repeat
       end
 
-      rule(:entity_ref) { match("&") >> name >> match(";") }
+      rule(:string_entity) { match("&") >> name >> match(";") }
+      rule(:numeric_entity) { match(/&#\d+;/) }
 
       rule(:space)  { match(/\s/).repeat(1) }
       rule(:space?) { space.maybe }

data/lib/upmark/transform/preprocess.rb

@@ -41,8 +41,10 @@ module Upmark
           .select { |c| Array === c }
           .map do |children|
             children.map do |child|
-              if child[:text]
-                child[:text].to_s.gsub!(/^\n */,'')
+              if child.is_a?(Hash) # if the td doesn't contain nested elements
+                if child[:text]
+                  child[:text].to_s.gsub!(/^\n */,'')
+                end
               end
               child
             end + ["\n"]

data/spec/acceptance/upmark_spec.rb

@@ -33,6 +33,28 @@ RSpec.describe Upmark, ".convert" do
     end
   end
 
+  context "<a> with numeric entity" do
+    specify 'converts as []()' do
+      expect(<<-HTML.strip
+        <p><a href=\"http://www.abc.net.au/news/2016-02-18/haylen-we-need-a-drug-summit-because-we&#39;re-losing-the-war/7177152\">blah</a></p>
+      HTML
+      ).to convert_to <<-MD.strip
+[blah](http://www.abc.net.au/news/2016-02-18/haylen-we-need-a-drug-summit-because-we&#39;re-losing-the-war/7177152 "")
+      MD
+    end
+  end
+
+  context "<a> with query string" do
+    specify 'converts as []()' do
+      expect(<<-HTML.strip
+        <p><a href=\"http://www.abc.net.au/news/2016-02-18/blah?blah=lol&lol=rofl\">blah</a></p>
+      HTML
+      ).to convert_to <<-MD.strip
+[blah](http://www.abc.net.au/news/2016-02-18/blah?blah=lol&lol=rofl "")
+      MD
+    end
+  end
+
   context "<img>" do
     specify 'converts as ![]()' do
       expect(<<-HTML.strip
@@ -289,4 +311,12 @@ messenger **bag** skateboard
       }.to raise_error(Upmark::ParseFailed)
     end
   end
+
+  context "nested table" do
+    let(:html) { "<table><tr><td><table><tr><td><p>Hi <br />there</p></td></tr></table></td></tr></table>"}
+
+    it "should strip both tables" do
+      expect(html).to convert_to("Hi\nthere")
+    end
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: upmark
 version: !ruby/object:Gem::Version
-  version: 0.9.0
+  version: 0.10.0
 platform: ruby
 authors:
 - Josh Bassett
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-11-08 00:00:00.000000000 Z
+date: 2016-06-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
@@ -111,12 +111,12 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: upmark
-rubygems_version: 2.4.5.1
+rubygems_version: 2.5.1
 signing_key: 
 specification_version: 4
 summary: A HTML to Markdown converter.
 test_files:
-- spec/unit/lib/upmark/transform/markdown_spec.rb
 - spec/unit/lib/upmark/parser/xml_spec.rb
+- spec/unit/lib/upmark/transform/markdown_spec.rb
 - spec/spec_helper.rb
 - spec/acceptance/upmark_spec.rb