hypothesis / product-backlog

Where new feature ideas and current bugs for the Hypothesis product live
118 stars 7 forks source link

bogus rel-canonical results in inappropriate coalescence #263

Open judell opened 7 years ago

judell commented 7 years ago

Steps to reproduce

Annotate (in a fresh group) the three URLs mentioned in https://hypothes-is.slack.com/archives/C03QZM0K4/p1491481119704670, namely

http://digitallearning.middcreate.net/summit/reflections-on-the-digital/five-ways-of-looking-at-the-digital-reflections-on-false-binaries/

http://digitallearning.middcreate.net/middcreate/the-digital-is-about-doing/

http://digitallearning.middcreate.net/critical-digital-pedagogy/exploring-digital-sanctuary/

Expected behaviour

An AP card for each of the 3 URLs

Actual behaviour

Two URLs share a card:

image

Additional details

Slack: https://hypothes-is.slack.com/archives/C03QZM0K4/p1491481119704670

Drive: https://docs.google.com/document/d/1R094mP9j5sQSLDQasoFBl1JqWk6Jkdh_DxJLDhGSkT8/edit

judell commented 7 years ago

I suspect we could safely ignore these:

image

Hmm Are we already ignoring application/rss+xml? It is present in the page but It doesn't show up in that query result.

Here is what the client sends:

link:Array(6)

0:Object href:"http://digitallearning.middcreate.net/middcreate/the-digital-is-about-doing/"

1:Object href:"http://digitallearning.middcreate.net/events/?ical=1" rel:"alternate" type:"text/calendar"

2:Object href:"http://digitallearning.middcreate.net/middcreate/the-digital-is-about-doing/" rel:"canonical" type:""

3:Object href:"http://wp.me/p7rMZ8-eZ" rel:"shortlink" type:""

4:Object href:"http://digitallearning.middcreate.net/wp-json/oembed/1.0/embed?url=http%3A%2F%2Fdigitallearning.middcreate.net%2Fmiddcreate%2Fthe-digital-is-about-doing%2F" rel:"alternate" type:"application/json+oembed"

5:Object href:"http://digitallearning.middcreate.net/wp-json/oembed/1.0/embed?url=http%3A%2F%2Fdigitallearning.middcreate.net%2Fmiddcreate%2Fthe-digital-is-about-doing%2F&format=xml" rel:"alternate" type:"text/xml+oembed"

Questions:

  1. Why isn't a link of type application/rss+xml showing up here?
  2. Are these other (likely unnecessary) alternates triggering the problem?