From d0d597e2d93035c35b6b653d181f550d9dd522fd Mon Sep 17 00:00:00 2001
From: Yorick Peterse <yorickpeterse@gmail.com>
Date: Sat, 23 May 2015 10:46:49 +0200
Subject: [PATCH] Allow script/template in various table elements

Fixes #105
---
 lib/oga/xml/lexer.rb                          | 14 +++--
 .../html/lexer/closing_rules/table_spec.rb    | 62 ++++++++++++-------
 .../html/lexer/closing_rules/tbody_spec.rb    | 20 ++++++
 .../html/lexer/closing_rules/tfoot_spec.rb    | 20 ++++++
 .../html/lexer/closing_rules/thead_spec.rb    | 20 ++++++
 spec/oga/html/lexer/closing_rules/tr_spec.rb  | 20 ++++++
 6 files changed, 128 insertions(+), 28 deletions(-)

diff --git a/lib/oga/xml/lexer.rb b/lib/oga/xml/lexer.rb
index e2c724d..465fd2a 100644
--- a/lib/oga/xml/lexer.rb
+++ b/lib/oga/xml/lexer.rb
@@ -50,6 +50,10 @@ module Oga
         %w{thead tbody tfoot tr caption colgroup col}
       )
 
+      HTML_SCRIPT_ELEMENTS = Whitelist.new(%w{script template})
+
+      HTML_TABLE_ROW_ELEMENTS = Whitelist.new(%w{tr}) + HTML_SCRIPT_ELEMENTS
+
       # Elements that should be closed automatically before a new opening tag is
       # processed.
       HTML_CLOSE_SELF = {
@@ -71,11 +75,11 @@ module Oga
         'option'   => Blacklist.new(%w{optgroup option}),
         'colgroup' => Whitelist.new(%w{col template}),
         'caption'  => HTML_TABLE_ALLOWED.to_blacklist,
-        'table'    => HTML_TABLE_ALLOWED,
-        'thead'    => Whitelist.new(%w{tr}),
-        'tbody'    => Whitelist.new(%w{tr}),
-        'tfoot'    => Whitelist.new(%w{tr}),
-        'tr'       => Whitelist.new(%w{td th}),
+        'table'    => HTML_TABLE_ALLOWED + HTML_SCRIPT_ELEMENTS,
+        'thead'    => HTML_TABLE_ROW_ELEMENTS,
+        'tbody'    => HTML_TABLE_ROW_ELEMENTS,
+        'tfoot'    => HTML_TABLE_ROW_ELEMENTS,
+        'tr'       => Whitelist.new(%w{td th}) + HTML_SCRIPT_ELEMENTS,
         'td'       => Blacklist.new(%w{td th}) + HTML_TABLE_ALLOWED,
         'th'       => Blacklist.new(%w{td th}) + HTML_TABLE_ALLOWED
       }
diff --git a/spec/oga/html/lexer/closing_rules/table_spec.rb b/spec/oga/html/lexer/closing_rules/table_spec.rb
index 5a7f437..160f80c 100644
--- a/spec/oga/html/lexer/closing_rules/table_spec.rb
+++ b/spec/oga/html/lexer/closing_rules/table_spec.rb
@@ -1,30 +1,46 @@
 require 'spec_helper'
 
 describe Oga::XML::Lexer do
-  describe 'HTML tables' do
-    describe 'with unclosed <tr> tags' do
-      it 'lexes a <tr> tag followed by a <tbody> tag' do
-        lex_html('<tr>foo<tbody></tbody>').should == [
-          [:T_ELEM_NAME, 'tr', 1],
-          [:T_TEXT, 'foo', 1],
-          [:T_ELEM_END, nil, 1],
-          [:T_ELEM_NAME, 'tbody', 1],
-          [:T_ELEM_END, nil, 1]
-        ]
-      end
+  describe 'using HTML <table> elements' do
+    it 'lexes two unclosed <table> elements following each other as separate elements' do
+      lex_html('<table>foo<table>bar').should == [
+        [:T_ELEM_NAME, 'table', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_NAME, 'table', 1],
+        [:T_TEXT, 'bar', 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
 
-      it 'lexes an unclosed <th> tag followed by a <tbody> tag' do
-        lex_html('<tr><th>foo<tbody>bar</tbody>').should == [
-          [:T_ELEM_NAME, 'tr', 1],
-          [:T_ELEM_NAME, 'th', 1],
-          [:T_TEXT, 'foo', 1],
-          [:T_ELEM_END, nil, 1],
-          [:T_ELEM_END, nil, 1],
-          [:T_ELEM_NAME, 'tbody', 1],
-          [:T_TEXT, 'bar', 1],
-          [:T_ELEM_END, nil, 1]
-        ]
-      end
+    it 'lexes a <table> element containing a <thead> element' do
+      lex_html('<table><thead>foo</thead></table>').should == [
+        [:T_ELEM_NAME, 'table', 1],
+        [:T_ELEM_NAME, 'thead', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
+
+    it 'lexes a <table> element containing a <script> element' do
+      lex_html('<table><script>foo</script></table>').should == [
+        [:T_ELEM_NAME, 'table', 1],
+        [:T_ELEM_NAME, 'script', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
+
+    it 'lexes a <table> element containing a <template> element' do
+      lex_html('<table><template>foo</template></table>').should == [
+        [:T_ELEM_NAME, 'table', 1],
+        [:T_ELEM_NAME, 'template', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
     end
   end
 end
diff --git a/spec/oga/html/lexer/closing_rules/tbody_spec.rb b/spec/oga/html/lexer/closing_rules/tbody_spec.rb
index f024f1c..4864dae 100644
--- a/spec/oga/html/lexer/closing_rules/tbody_spec.rb
+++ b/spec/oga/html/lexer/closing_rules/tbody_spec.rb
@@ -33,5 +33,25 @@ describe Oga::XML::Lexer do
         [:T_ELEM_END, nil, 1]
       ]
     end
+
+    it 'lexes a <tbody> element containing a <script> element' do
+      lex_html('<tbody><script>foo</script></tbody>').should == [
+        [:T_ELEM_NAME, 'tbody', 1],
+        [:T_ELEM_NAME, 'script', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
+
+    it 'lexes a <tbody> element containing a <template> element' do
+      lex_html('<tbody><template>foo</template></tbody>').should == [
+        [:T_ELEM_NAME, 'tbody', 1],
+        [:T_ELEM_NAME, 'template', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
   end
 end
diff --git a/spec/oga/html/lexer/closing_rules/tfoot_spec.rb b/spec/oga/html/lexer/closing_rules/tfoot_spec.rb
index 4acb170..2bf2c90 100644
--- a/spec/oga/html/lexer/closing_rules/tfoot_spec.rb
+++ b/spec/oga/html/lexer/closing_rules/tfoot_spec.rb
@@ -33,5 +33,25 @@ describe Oga::XML::Lexer do
         [:T_ELEM_END, nil, 1]
       ]
     end
+
+    it 'lexes a <tfoot> element containing a <script> element' do
+      lex_html('<tfoot><script>foo</script></tfoot>').should == [
+        [:T_ELEM_NAME, 'tfoot', 1],
+        [:T_ELEM_NAME, 'script', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
+
+    it 'lexes a <tfoot> element containing a <template> element' do
+      lex_html('<tfoot><template>foo</template></tfoot>').should == [
+        [:T_ELEM_NAME, 'tfoot', 1],
+        [:T_ELEM_NAME, 'template', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
   end
 end
diff --git a/spec/oga/html/lexer/closing_rules/thead_spec.rb b/spec/oga/html/lexer/closing_rules/thead_spec.rb
index efcaf27..f9b1d71 100644
--- a/spec/oga/html/lexer/closing_rules/thead_spec.rb
+++ b/spec/oga/html/lexer/closing_rules/thead_spec.rb
@@ -33,5 +33,25 @@ describe Oga::XML::Lexer do
         [:T_ELEM_END, nil, 1]
       ]
     end
+
+    it 'lexes a <thead> element containing a <script> element' do
+      lex_html('<thead><script>foo</script></thead>').should == [
+        [:T_ELEM_NAME, 'thead', 1],
+        [:T_ELEM_NAME, 'script', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
+
+    it 'lexes a <thead> element containing a <template> element' do
+      lex_html('<thead><template>foo</template></thead>').should == [
+        [:T_ELEM_NAME, 'thead', 1],
+        [:T_ELEM_NAME, 'template', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
   end
 end
diff --git a/spec/oga/html/lexer/closing_rules/tr_spec.rb b/spec/oga/html/lexer/closing_rules/tr_spec.rb
index 5dd1f56..4952b17 100644
--- a/spec/oga/html/lexer/closing_rules/tr_spec.rb
+++ b/spec/oga/html/lexer/closing_rules/tr_spec.rb
@@ -32,5 +32,25 @@ describe Oga::XML::Lexer do
         [:T_ELEM_END, nil, 1]
       ]
     end
+
+    it 'lexes a <tr> element containing a <script> element' do
+      lex_html('<tr><script>foo</script></tr>').should == [
+        [:T_ELEM_NAME, 'tr', 1],
+        [:T_ELEM_NAME, 'script', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
+
+    it 'lexes a <tr> element containing a <template> element' do
+      lex_html('<tr><template>foo</template></tr>').should == [
+        [:T_ELEM_NAME, 'tr', 1],
+        [:T_ELEM_NAME, 'template', 1],
+        [:T_TEXT, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_ELEM_END, nil, 1]
+      ]
+    end
   end
 end