Error during cleanup about php-readability

[techexo]

Here is some more information about the issue:

1. It occurs only when a header element is empty. Even a space between the tags "fixes" the issue.
2. It occurs with every header element >= h2 (i.e. h3, h4 etc.). (I suppose h1 are interpreted by default in graby as containing the title of the article, thus are stripped.)
3. If the element has no class attribute, the following elements of same level (with or without class) are stripped from the code! (i.e. a <h2></h2> instruction will strip all following <h2> markups, even not empty!)

<h2></h2>
<p>Lorem ipsum.</p>
<h2 class="truc">Truc</h2>
<p>Lorem ipsum 2.</p>

<!-- will give -->

<p>Lorem ipsum.</p>
<p>Lorem ipsum 2.</p>

4. It is not "class" dependant (i.e. <h2 id="truc"></h2> will give the same issue)

<h2 id="truc"></h2>
<p>Lorem ipsum.</p>
<h2 id="truc">Truc</h2>
<p>Lorem ipsum 2.</p>

<!-- will give -->

<h2>Lorem ipsum.
Truc</h2>
<p>Lorem ipsum 2.</p>

5. The issue only appears up to the next occurrence of a header of same level:

<h2 class="truc"></h2>
<p>Lorem ipsum.</p>
<h2 class="truc">Truc</h2>
<h2 class="truc">Truc 2</h2>
<p>Lorem ipsum 3.</p>

<!-- will give -->

<h2>Lorem ipsum.
Truc</h2>
<h2 class="truc">Truc2</h2>
<p>Lorem ipsum 2.</p>

6. Interestingly however, the problem disappears when the next header is not at the same level

<h2 class="truc"></h2>
<p>Lorem ipsum.</p>
<h3 class="truc">Truc</h3>
<p>Lorem ipsum 2.</p>

<!-- will give -->

<p>Lorem ipsum.</p>
<h3>Truc</h3>
<p>Lorem ipsum 2.</p>

It seems to be an issue finding the closing HTML markup when the markups are empty.

[techexo]

New test: Using Tidy to clean the examples above with the options specified in php-readability doesn't fail.

[Kdecherf]

@j0k3r this issue can be closed as per j0k3r/graby#178