akket.com test case

[topotru]

Test-case for akket.com articles

[coveralls]

Coverage increased (+0.6%) to 96.818% when pulling 3dcb778 on topotru:akket.com-test into 992a112 on andreskrey:master.

[topotru]

sure, tests fails

[andreskrey]

Great, I'll debug it this week. Thanks :)

[andreskrey]

Wheew, this looks complex. Comparing the expected.html and the current results I'd say it's going to be quite tricky to reproduce your results.

I totally understand your desire to remove the buttons at the end of the text (class name is "tages") but the "follow us on twitter, bla bla bla" part I think it's not feasible with our current approach. That text is essentially part of the article. There's no way to programatically tell apart between actual content and that paragraph without actually analyzing the text and not just the semantic composition.

What I can do is target the link list to remove it. Unless you have a better idea for that chunk of text!

[topotru]

... but the "follow us on twitter, bla bla bla" part I think it's not feasible with our current approach. That text is essentially part of the article. There's no way to programatically tell apart between actual content and that paragraph without actually analyzing the text and not just the semantic composition...

Yeah! That's why I need original class-names )
Here is a paragraph without a class name, but in many cases I can rely on them for post-parsing processing.
Only where no way to programatically tell apart between actual content and the semantic composition...
But it's a chance to have more cleanest result... And I want to have this chance in some cases )

What I can do is target the link list to remove it. Unless you have a better idea for that chunk of text!

I was thinking about it too.
Logically - if the container contains only links, then it is more correct to get rid of it. Although, perhaps there are articles that contain links lists and they are directly related to the article itself. But imho this is a rare situation.
I think that the best solution is to introduce the option of cleanLinksLists (true by default) and everyone can decide use it or not )

[andreskrey]

Still fails in v2, link list still present, header image missing.