1218242 - Reader View eliminates article code's comments if they have "comment" in a class

Status: RESOLVED DUPLICATE of bug 1199651

(Toolkit :: Reader Mode, defect)

Description

[Franklin?]

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:41.0) Gecko/20100101 Firefox/41.0
Build ID: 20151014143721

Steps to reproduce:

Take a page on which Firefox allows "Reader View". Take one of the article's paragraph nodes and add the class "not-a-comment" to it. Click "Reader View".


Actual results:

That paragraph doesn't appear in Reader View.


Expected results:

It should.

----

As a real-world example, the following article uses the class "md-code-comment" to style comments within code examples:
https://ponyfoo.com/articles/es6-promises-in-depth
(archived: https://web.archive.org/web/20150928184517/http://ponyfoo.com/articles/es6-promises-in-depth)

Here's the code that strips out these nodes.
https://github.com/mozilla/readability/blob/a801846a456db91df31b6f3b150f2c15619bcb1a/Readability.js#L650

Here's the regex that decides it. ('unlikelyCandidates')
https://github.com/mozilla/readability/blob/a801846a456db91df31b6f3b150f2c15619bcb1a/Readability.js#L110

There are definitely uses for /.*comment.*/ classes. Maybe none of those unlikelyCandidate tags should be removed if they're the subnode of a "candidate" node with text, since at that point you already predicted that you're inside the article text.