Closed pagojo closed 10 years ago
Test case above actually may be due to #51. However the noscript
argument remains.
<div align="center">
<div class="article-img">
<img src="http://photos.appleinsider.com/v9/images/1x1-white.jpg" width="660" height="317" alt="iPhone Plus" class="lazy" data-original="http://cdn1.appleinsider.com/iphoneplus-130205.jpg"><noscript><img src="http://cdn1.appleinsider.com/iphoneplus-130205.jpg"></noscript>
</div>
<br><span class="minor2 small gray">Mockup of iPhone with 4.94-inch screen, <a href="http://www.marco.org/2013/01/31/iphone-plus-speculation">created by</a> Marco Arment.</span>
</div>
worse, there is double counting going on in clean_conditionally(node, candidates, selector)
as img
is counted both in the script/noscript cases inside a div. this skews the numbers.
The JS Readability seems to work fine with this kind of code, whereas the Ruby version doesn't.
The following snippet was lifted from AppleInsider
To me it seems that Ruby Readability should detect the
noscript
images and use them in its heuristics (modulo issue #51). Currently it doesn't