1141476 - Fetch & parse page metadata in the background for items added via sync and missing metadata

Status: RESOLVED INCOMPLETE

(Firefox Graveyard :: Reading List, defect, P1)

Description

[Blair McBride [:Unfocused] (UNAVAILABLE)]

We can't assume all items were added to the server had complete metadata. For those items, we'll want to fetch the page in the background and extract the relevant metadata - then store and sync to the server.

Most notable case of this happening is when a URL is shared to Firefox on Android.

Comment 1

[Justin Dolske [:Dolske]]

Need bug 1149336 to be able to deal with syncing the updated item.

Comment 3

[Mark Hammond [:markh] [:mhammond]]

I guess there are 2 ways we could approach this:
1) Load the entire page in the background (just like we do for thumbnails) and fetch the data from that.
2) Fetch just the top-level page, parse that using JSDOMParser and extract it from that.

(1) should always be correct but slow. (2) should be faster but will not pick up things like JS mutating the DOM to setup the metadata we look for.  (2) *might* be fast enough to help with bug 1155014 (ie, PageMetaData.jsm could fallback to (2) if it finds the current DOM doesn't reflect reality)

Any thoughts on this?