Reader: fix feed searching having incorrect following state

[samouri]

Barely a bug since it on an unreleased product.

In Search on Manage, search result items don't have the correct following status. This is due to the api returning url differently from following/mine.

feed-search --> ?
Following/mine --> feed_URL

The fix was just ignoring the url from feed-search and using the one found on a the feed/site.

To Test:

• head to the live branch, search for a feed you are already following, it should be green.
• clicking the urls below list items should open the website in a new tab (not the rss feed)

[matticbot]

Test live: https://calypso.live/?branch=fix/reader/feed-search-following-status

[matticbot]

@samouri This PR needs a rebase

[bluefuton]

I'd add "so we omit it and use the URL from the feed/site instead"

[bluefuton]

This hasn't done the trick for me @samouri. I'm following this feed, Awful Library Books:

http://calypso.localhost:3000/read/feeds/63937159

...but its following state isn't reflected:

[samouri]

@blowery: I didn't see subscribe_URL in the api response even though the reducer tests make me think it should be there 🤔

[samouri]

@bluefuton nice spot!
It was a case that I hadn't realized when initially writing the ReaderSubscriptionListItem. The siteUrl vs. feedUrl distinction.

[blowery]

if we're going to do this, it should be in a memoized selector, not here. This runs every time the redux store updates and would return a new instance of searchResults on each run. That would cause the connected component to always rerender...

[samouri]

good call. I'm still 🤞 on the api fix though :)

[samouri]

@blowery: i've updated this to use subscribe_url. The solution works (i think) but its not as clean as i think it could be.

What do you think about adding in feed_URL to the data-layer for both follows+feedSearches? (not the api, the normalization of the response)

[blowery]

What do you think about adding in feed_URL to the data-layer for both follows+feedSearches? (not the api, the normalization of the response)

Sounds fine. My first thought was that we'll need to update the schema for follows, but I'm not sure we have a schema for follows just yet?

[samouri]

done, its much nicer and explicit 👍

[blowery]

@samouri broke the tests. :)

[samouri]

Should be good now!

[blowery]

LGTM