mediacloth 0.0.3 → 0.5

data/test/data/html14

@@ -0,0 +1,4 @@
+<h1><a name='Headline1'></a> Headline1 </h1>
+<pre>Something</pre><h2><a name='Headline2'></a> Headline2 </h2>
+<pre>Something</pre><h2><a name='Headline3'></a> Headline3 </h2>
+<pre>Something</pre>

data/test/data/html15

@@ -0,0 +1,29 @@
+<div class="wikitoc">
+<div class="wikitoctitle">Contents</div>
+<ul>
+<li><a href='#Headline1'><span class="wikitocnumber">1.</span><span class="wikitoctext">Headline1</span></a>
+<ul>
+<li><a href='#Sub_Headline_1a'><span class="wikitocnumber">1.1.</span><span class="wikitoctext">Sub Headline 1a</span></a>
+<li><a href='#Sub_Headline_1b'><span class="wikitocnumber">1.2.</span><span class="wikitoctext">Sub Headline 1b</span></a>
+</ul>
+<li><a href='#Headline2'><span class="wikitocnumber">2.</span><span class="wikitoctext">Headline2</span></a>
+<ul>
+<li><a href='#Sub_Headline_2a'><span class="wikitocnumber">2.1.</span><span class="wikitoctext">Sub Headline 2a</span></a>
+<ul>
+<li><a href='#Sub_Sub_Headline_2i'><span class="wikitocnumber">2.1.1.</span><span class="wikitoctext">Sub Sub Headline <i>2i</i></span></a>
+<li><a href='#Sub_Sub_Headline_2ii'><span class="wikitocnumber">2.1.2.</span><span class="wikitoctext">Sub Sub Headline <b>2ii</b></span></a>
+</ul>
+<li><a href='#Sub_Headline_2b'><span class="wikitocnumber">2.2.</span><span class="wikitoctext">Sub Headline 2b</span></a>
+</ul>
+</ul>
+</div>
+<p><br />
+
+</p><h2><a name='Headline1'></a> Headline1 </h2>
+<h3><a name='Sub_Headline_1a'></a> Sub Headline 1a </h3>
+<h3><a name='Sub_Headline_1b'></a> Sub Headline 1b </h3>
+<h2><a name='Headline2'></a> Headline2 </h2>
+<h3><a name='Sub_Headline_2a'></a> Sub Headline 2a </h3>
+<h4><a name='Sub_Sub_Headline_2i'></a> Sub Sub Headline <i>2i</i> </h4>
+<h4><a name='Sub_Sub_Headline_2ii'></a> Sub Sub Headline <b>2ii</b> </h4>
+<h3><a name='Sub_Headline_2b'></a> Sub Headline 2b </h3>

data/test/data/html16

@@ -0,0 +1,28 @@
+<h2><a name='Section'></a> Section </h2>
+<p>Some text.
+
+</p><pre class="indent">&lt;?xml version=&quot;1.0&quot; encoding=&quot;UTF-8&quot;?&gt;
+&lt;!DOCTYPE plist PUBLIC &quot;-//Apple Computer//DTD PLIST 1.0//EN&quot; &quot;<a href="http://www.apple.com/DTDs/PropertyList-1.0.dtd">http://www.apple.com/DTDs/PropertyList-1.0.dtd</a>&quot;&gt;
+&lt;plist version=&quot;1.0&quot;&gt;
+&lt;dict&gt;
+&lt;key&gt;Label&lt;/key&gt;
+&lt;string&gt;Update svn repo&lt;/string&gt;
+&lt;key&gt;ProgramArguments&lt;/key&gt;
+&lt;array&gt;
+&lt;string&gt;updatesvn.sh&lt;/string&gt;
+&lt;/array&gt;
+&lt;key&gt;StandardErrorPath&lt;/key&gt;
+&lt;string&gt;/dev/null&lt;/string&gt;
+&lt;key&gt;StandardOutPath&lt;/key&gt;
+&lt;string&gt;/dev/null&lt;/string&gt;
+&lt;key&gt;StartCalendarInterval&lt;/key&gt;
+&lt;dict&gt;
+&lt;key&gt;Hour&lt;/key&gt;
+&lt;integer&gt;16&lt;/integer&gt;
+&lt;key&gt;Minute&lt;/key&gt;
+&lt;integer&gt;40&lt;/integer&gt;
+&lt;/dict&gt;
+&lt;/dict&gt;
+&lt;/plist&gt;
+</pre><p>
+Some text.<br />Some other text.<br />Foo.<hr /><img src="foo" /></p>

data/test/data/html17

@@ -0,0 +1,10 @@
+<table cellpadding="5" border="1"><tr><td> foo
+bar
+zee
+</td><td> moo
+goo
+</td><td  attr="value" otherattr="othervalue" > nah
+mah
+bah
+</td></tr>
+</table>

data/test/data/html18

@@ -0,0 +1,8 @@
+<h2><a name='Foobar'></a> Foobar </h2>
+<table border="1px" cellpadding="5"><tr><th>test
+</th></tr>
+<tr><td>a
+</td></tr>
+</table>
+<h2><a name='Foobar'></a> Foobar </h2>
+<table border="1px"><thead><tr><th>test</th></tr></thead><tbody><tr><td>a</td></tr></tbody></table>

data/test/data/html19

@@ -0,0 +1,27 @@
+<pre>
+  foo bar
+</pre><p><code>
+  foo bar
+</code>
+
+</p><p> hi <br /> bye</p><p><br />
+
+</p><p>test &lt;br/&gt; test
+
+</p><p>There's something important in this example: <code>int a = <b>2</b>+2;</code> but I don't know what.
+
+</p><p>
+There's something important in this example: <code>int a = 2+2;<br />
+
+int b = a + 2;
+</code> but I don't know what.
+
+</p><pre>
+    Here's some code:
+
+    &lt;code&gt;
+    int a = 2+2;
+
+    int b = 3+4;
+    &lt;/code&gt;
+</pre>

data/test/data/html2

@@ -1,2 +1,2 @@
-<hr></hr><p>
+<hr/><p>
 This is a text</p>

data/test/data/html20

@@ -0,0 +1,7 @@
+<table border="1" cellpadding="2" cellspacing="0"><tr><td> xxx
+</td><td><pre class="indent">test
+test
+test
+</pre></td><td> yyy
+</td></tr>
+</table>

data/test/data/html21

@@ -0,0 +1,5 @@
+<div class="paste" style="white-space: pre-wrap;">This is '''some''' text pasted from anywhere<br />with<br />new<br />lines<br />and    multiple    spaces and	tabs<br /> and space<br />  and two spaces<br />with some<br />&lt;p attr=&quot;xxx&quot; moo=&quot;zzz&quot; bar=zzzzz&gt;HTML formatted &lt;a href=&quot;foo.bar&quot;&gt;text&lt;/a&gt; in &lt;/p&gt; it.<br /></div><p>
+
+
+</p><h2><a name='Section'></a> Section </h2>
+<div class="paste" style="white-space: pre-wrap;">Foo</div>

data/test/data/html22

@@ -0,0 +1,24 @@
+<div class="wikitoc">
+<div class="wikitoctitle">Contents</div>
+<ul>
+<li><a href='#Headline1'><span class="wikitocnumber">1.</span><span class="wikitoctext">Headline1</span></a>
+<ul>
+<li><a href='#Sub_Headline_1a'><span class="wikitocnumber">1.1.</span><span class="wikitoctext">Sub Headline 1a</span></a>
+<li><a href='#Sub_Headline_1b'><span class="wikitocnumber">1.2.</span><span class="wikitoctext">Sub Headline 1b</span></a>
+</ul>
+<li><a href='#Headline2'><span class="wikitocnumber">2.</span><span class="wikitoctext">Headline2</span></a>
+<ul>
+<li><a href='#Sub_Headline_2a'><span class="wikitocnumber">2.1.</span><span class="wikitoctext">Sub Headline 2a</span></a>
+<li><a href='#Sub_Headline_2b'><span class="wikitocnumber">2.2.</span><span class="wikitoctext">Sub Headline 2b</span></a>
+</ul>
+</ul>
+</div>
+<p><br />
+
+</p><p><div><h1><a name='Headline1'></a> Headline1 </h1>
+<span><h2><a name='Sub_Headline_1a'></a> Sub Headline 1a </h2>
+</span><h2><a name='Sub_Headline_1b'></a> Sub Headline 1b </h2>
+</div></p><h1><a name='Headline2'></a> Headline2 </h1>
+<table><tr><td><h2><a name='Sub_Headline_2a'></a> Sub Headline 2a </h2>
+</td><td><h2><a name='Sub_Headline_2b'></a> Sub Headline 2b </h2>
+</td></tr></table>

data/test/data/html23

@@ -0,0 +1,7 @@
+<pre>
+  This is
+    preformatted text
+
+</pre><pre class="indent">And this is
+  preformatted text
+</pre>

data/test/data/html3

@@ -1 +1 @@
-<p>--Sat Jan 01 01:01:01 UTC 2000CreatorCreator Sat Jan 01 01:01:01 UTC 2000</p>
+<p>--Sat Jan 01, 2000 01:01<a href="javascript:void(0)">Creator</a><a href="javascript:void(0)">Creator</a> Sat Jan 01, 2000 01:01</p>

data/test/data/html4

@@ -1,11 +1,60 @@
-<p>Test mixing of bold and italic formatting:</p><p><i>italic<b>bold</b>italic</i></p><p><b>bold<i>italic</i>bold</b></p><p><i><b>boldalic</b></i></p><p><i>italic<b>bold</b>italic</i><b>bold<i>italic</i>bold</b></p><p>
-Test bold and italic wrapped around inline links:</p><p><i><a href="http://example.com">http://example.com</a></i></p><p><b><a href="http://example.com">http://example.com</a></b></p><p><i><a href="http://example.com'">http://example.com'</a></i></p><p><b><a href="http://example.com'">http://example.com'</a> is good</b></p><p><i><a href="http://example.com'">http://example.com'</a> is good</i></p><p>
-Test unclosed bold and italic formatting:</p><p><i>Some italic and now </i>bold
-</p><h2>Heading</h2><p>
-Text</p><p><i>Some italic and now <b>bold
-</b></i></p><h2>Heading</h2><p>
-Text</p><p><i>Some italic and now <b>bold</b></i></p><h2>Heading</h2><p>
-Text</p><p><i>Some italic and now <b>bold
-</b></i></p><ul><li>one
-</li><li>two
-</li></ul><p><i>Some italic and now <b>bold</b></i></p><p>Text</p>
+<p>Test mixing of bold and italic formatting:
+
+</p><p><i>italic<b>bold</b>italic</i>
+
+</p><p><b>bold<i>italic</i>bold</b>
+
+</p><p><i><b>boldalic</b></i>
+
+</p><p><i>italic<b>bold</b>italic</i><b>bold<i>italic</i>bold</b>
+
+</p><p>
+Test bold and italic wrapped around inline links:
+
+</p><p><i><a href="http://example.com">http://example.com</a></i>
+
+</p><p><b><a href="http://example.com">http://example.com</a></b>
+
+</p><p><i><a href="http://example.com">http://example.com</a><b>
+
+</b></i></p><p><b><a href="http://example.com">http://example.com</a>' is good</b>
+
+</p><p><i><a href="http://example.com">http://example.com</a>' is good</i>
+
+</p><p>
+Test unclosed bold and italic formatting:
+
+</p><p><i>Some italic and now </i>bold
+</p><h2><a name='Heading'></a>Heading</h2>
+<p>Text
+
+</p><p><i>Some italic and now <b>bold
+</b></i></p><h2><a name='Heading'></a>Heading</h2>
+<p>
+Text
+
+</p><p><i>Some italic and now <b>bold
+
+</b></i></p><h2><a name='Heading'></a>Heading</h2>
+<p><br />
+
+</p><p>Text
+
+</p><p><i>Some italic and now <b>bold
+</b></i></p><ul><li> one
+</li><li> two
+</li></ul><p><br />
+
+</p><p><i>Some italic and now <b>bold
+
+</b></i></p><p>Text
+
+</p><p>Invalid and weird sections:
+</p><h6><a name='===='></a>====</h6>
+<p>
+=
+==
+</p><h1><a name='='></a>=</h1>
+<h1><a name='=='></a>==</h1>
+<h2><a name='='></a>=</h2>
+<h2><a name='=='></a>==</h2>

data/test/data/html5

@@ -1,18 +1,57 @@
-<ul><li>A
+<p>Unordered lists
+
+</p><ul><li>A
 <ul><li>A
 <ul><li>a
 </li><li>b
 <ul><li>a
 </li></ul></li></ul></li></ul></li><li>B
 <ul><li>b
-</li></ul></li></ul><ol><li>a
+</li></ul></li></ul><ul><li>A
+</li><li>B
+</li><li></li></ul><p>
+Ordered lists
+
+</p><ol><li>a
 <ol><li>a
 <ol><li>a
 <ol><li>a
 </li></ol></li></ol></li></ol></li><li>a
 <ol><li>aa
-</li></ol></li></ol><ul><li>A
-</li><li>B
-</li><li></li></ul><ol><li>A
+</li></ol></li></ol><ol><li>A
 </li><li>B
-</li><li></li></ol>
+</li><li></li></ol><p>
+Mixed lists
+
+</p><ul><li>a
+<ol><li>1
+</li><li>2
+</li></ol></li><li>b
+</li></ul><ol><li>a
+<ul><li>1
+</li><li>2
+<ul><li>2
+</li></ul></li></ul></li><li>b
+</li></ol><p>
+Dictionary lists
+
+</p><dl><dt>d
+</dt></dl><p><br />
+
+</p><dl><dd>t
+</dd></dl><p><br />
+
+</p><dl><dt>d
+</dt><dd>t
+</dd></dl><p><br />
+
+</p><dl><dt>d
+</dt><dd>t
+</dd><dd>t
+</dd></dl><p><br />
+
+</p><dl><dt>d
+</dt><dd>t
+</dd></dl><dl><dt>d
+</dt><dd>t
+</dd></dl><dl><dt>d </dt><dd>t <a href="image:foo.jpg">image:foo.jpg</a></dd></dl>

data/test/data/html6

@@ -1,8 +1,8 @@
 <p>This is some text with <b>bold</b> and <i>italic</i> formating.
 The list is also here:
-</p><ul><li>List Item 1
-<ul><li>Sub list item 1
-</li><li>Sub list item 2
-</li></ul></li><li>List Item 2
+</p><ul><li> List Item 1
+<ul><li> Sub list item 1
+</li><li> Sub list item 2
+</li></ul></li><li> List Item 2
 </li></ul><p>The end
-</p><hr></hr>
+</p><hr/>

data/test/data/html7

@@ -1 +1,59 @@
-<p><a href="http://sun.com">http://sun.com</a></p><p><a href="http://sun.com">http://sun.com</a></p><p><a href="mailto:joe@sun.com">mailto:joe@sun.com</a></p><p><a href="http://sun.com">stars</a></p><p><a href="http://sun.com">stars and moon</a></p><p><a href="http://sun.com">stars and <i>moon</i>and <b>trees</b>and birds</a></p><p><a href="javascript:void(0)">sun</a></p><p><a href="javascript:void(0)">All about Sun</a></p><p><a href="javascript:void(0)">image:sun(All about Sun)</a></p><p><a href="javascript:void(0)">nofollow|All about Sun</a></p><p><a href="javascript:void(0)">image:sun(nofollow, All about Sun)</a></p><p><a href="javascript:void(0)">image:sun(All about <a href="javascript:void(0)">Sun</a>)</a></p><p><a href="javascript:void(0)">image:sun(All about <a href="javascript:void(0)">More about</a>)</a></p><p><a href="javascript:void(0)">image:sun(one, two, three)</a></p><p>[]</p><p>[ ]</p><p>[[]]</p><p>[[ ]]</p>
+<p><a href="http://sun.com">http://sun.com</a>
+
+</p><p><a href="http://sun.com:80">http://sun.com:80</a>
+
+</p><p><a href="http://sun.com">http://sun.com</a>
+
+</p><p><a href="mailto:joe@sun.com">mailto:joe@sun.com</a>
+
+</p><p><a href="http://sun.com">stars</a>
+
+</p><p><a href="http://sun.com">stars and moon</a>
+
+</p><p><a href="http://sun.com">stars and <i>moon</i> and <b>trees</b> and birds</a>
+
+</p><p><a href="http://sun.com"><tt>sun</tt> and <tt>moon</tt></a>
+
+</p><p><a href="http://sun.com">sun &gt; moon</a>
+
+</p><p><a href="javascript:void(0)">sun</a>
+
+</p><p><a href="javascript:void(0)">All about Sun</a>
+
+</p><p><a href="image:sun">image:sun</a>
+
+</p><p><a href="javascript:void(0)">nofollow|All about Sun</a>
+
+</p><p><a href="javascript:void(0)">nofollow|Sun &gt; Moon</a>
+
+</p><p><a href="javascript:void(0)">nofollow|<tt>sun</tt> and <tt>moon</tt></a>
+
+</p><p><a href="image:sun">image:sun</a>
+
+</p><p><a href="image:sun">image:sun</a>
+
+</p><p><a href="image:sun">image:sun</a>
+
+</p><p><a href="image:sun">image:sun</a>
+
+</p><p><a href="image:sun">image:sun</a>
+
+</p><p>[]
+
+</p><p>[ ]
+
+</p><p>[[]]
+
+</p><p><a href="javascript:void(0)"> </a>
+
+</p><p>
+
+</p><p>
+
+</p><p>
+
+</p><p><a href="javascript:void(0)">Category</a>
+
+</p><p><a href="javascript:void(0)">Category name</a>
+
+</p><p><a href="javascript:void(0)">Foo</a></p>

data/test/data/html8

@@ -1 +1 @@
-<h3> foo </h3>
+<h3><a name='foo'></a> foo </h3>

data/test/data/html9

@@ -1,6 +1,14 @@
-<p>Some examples of wiki escaping.</p><p>A paragraph with some wiki characters: ''italic'' '''bold''' [[Link]]</p><p>An escaped list:
+<p>Some examples of wiki escaping.
+
+</p><p>A paragraph with some wiki characters: ''italic'' '''bold''' [[Link]]
+
+</p><p>An escaped list:
 
 * one
 * two
 * three
-</p><p>Text escaped in a heading:</p><h2>Heading with '''</h2>
+
+
+</p><p>Text escaped in a heading:
+
+</p><h2><a name='Heading_with____'></a>Heading with '''</h2>

data/test/data/input1

@@ -8,6 +8,8 @@ Another one.
 
 This is text with [[Internal Link]] and [http://www.example.com external link].
 
+http://inline.link/with?at@character
+
 We can have headlines:
 
 
@@ -32,3 +34,5 @@ This is a text after the line.
 
 *foo
 * foo2
+
+== There aren't any closing quotes inside this headline's name ==

data/test/data/input11

@@ -0,0 +1,19 @@
+==Examples of XHTML markup and escaping of XML characters==
+
+Here is a <tt>code sample</tt>.
+
+Another <tt class='code'>code sample</tt>.
+Another <tt class="code">code sample</tt>.
+Another <tt class=code>code sample</tt>.
+
+An example of a <nowiki><tt>code sample</tt></nowiki>.
+
+Some math:
+
+  1 < 2 & 1 > 0
+
+And more text.
+
+* one
+* two <span>a b
+c d e f</span>

data/test/data/input12

@@ -0,0 +1,32 @@
+one & two
+1 < 2
+<lt; is the "lt" symbol
+1 < 2 & 3 > 2
+
+this &lt;&gt; that
+
+{{ref}}
+
+{{}}
+
+{{ }}
+
+{{ref|attr1}}
+
+{{ref|attr1|attr2|attr3}}
+
+{{ref with spaces|attr1 with spaces}}
+
+{{NS:ref}}
+
+{{NS:ref|attr}}
+
+{{NS: ref}}
+
+{{NS: ref|attr}}
+
+{{ref|{{inner}}}}
+
+{{ref|{{inner}}|{{inner2}}}}
+
+{{ref|{{inner|{{subinner}}}}}}

data/test/data/input13

@@ -0,0 +1,10 @@
+__TOC__
+
+= Headline1 =
+== Sub Headline 1a ==
+== Sub Headline 1b ==
+= Headline2 =
+== Sub Headline 2a ==
+=== Sub Sub Headline <i>2i</i> ===
+=== Sub Sub Headline '''2ii''' ===
+== Sub Headline 2b ==

data/test/data/input14

@@ -0,0 +1,8 @@
+= Headline1 =
+<pre>Something</pre>
+
+== Headline2 ==
+<pre>Something</pre>
+
+== Headline3 ==
+<pre>Something</pre>

data/test/data/input15

@@ -0,0 +1,10 @@
+__TOC__
+
+== Headline1 ==
+=== Sub Headline 1a ===
+=== Sub Headline 1b ===
+== Headline2 ==
+=== Sub Headline 2a ===
+==== Sub Sub Headline <i>2i</i> ====
+==== Sub Sub Headline '''2ii''' ====
+=== Sub Headline 2b ===

data/test/data/input16

@@ -0,0 +1,28 @@
+== Section ==
+Some text.
+
+ <?xml version="1.0" encoding="UTF-8"?>
+ <!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+ <plist version="1.0">
+ <dict>
+ <key>Label</key>
+ <string>Update svn repo</string>
+ <key>ProgramArguments</key>
+ <array>
+ <string>updatesvn.sh</string>
+ </array>
+ <key>StandardErrorPath</key>
+ <string>/dev/null</string>
+ <key>StandardOutPath</key>
+ <string>/dev/null</string>
+ <key>StartCalendarInterval</key>
+ <dict>
+ <key>Hour</key>
+ <integer>16</integer>
+ <key>Minute</key>
+ <integer>40</integer>
+ </dict>
+ </dict>
+ </plist>
+
+Some text.<br>Some other text.<br/>Foo.<hr><img src="foo">

data/test/data/input17

@@ -0,0 +1,10 @@
+{|
+| foo
+bar
+zee
+| moo
+goo
+| attr="value" otherattr="othervalue" | nah
+mah
+bah
+|}

data/test/data/input18

@@ -0,0 +1,16 @@
+== Foobar ==
+{|border="1px"
+!test
+|-
+|a
+|}
+
+== Foobar ==
+<table border="1px">
+<thead>
+<tr><th>test</th></tr>
+</thead>
+<tbody>
+<tr><td>a</td></tr>
+</tbody>
+</table>

data/test/data/input19

@@ -0,0 +1,29 @@
+<pre>
+  foo bar
+</pre>
+
+<code>
+  foo bar
+</code>
+
+<p> hi <br/> bye</p>
+
+<nowiki>test <br/> test</nowiki>
+
+There's something important in this example: <code>int a = '''2'''+2;</code> but I don't know what.
+
+
+There's something important in this example: <code>int a = 2+2;<br/>
+
+int b = a + 2;
+</code> but I don't know what.
+
+<pre>
+    Here's some code:
+
+    <code>
+    int a = 2+2;
+
+    int b = 3+4;
+    </code>
+</pre>

data/test/data/input20

@@ -0,0 +1,8 @@
+{| border="1" cellpadding="2" cellspacing="0"
+| xxx
+|
+  test
+  test
+  test
+| yyy
+|}

data/test/data/input21

@@ -0,0 +1,18 @@
+<paste>
+This is '''some''' text pasted from anywhere
+with
+new
+lines
+and    multiple    spaces and	tabs
+ and space
+  and two spaces
+with some
+<p attr="xxx" moo="zzz" bar=zzzzz>HTML formatted <a href="foo.bar">text</a> in </p> it.
+</paste>
+<paste></paste>
+<paste/>
+
+
+== Section ==
+
+<paste>Foo</paste>