730101 - (prerendering) Implement prerendering in Firefox

Status: REOPENED

(Core :: DOM: Navigation, enhancement)

Description

[bug.zilla]

Can Firefox implement prerendering of pages like Chrome (http://code.google.com/chrome/whitepapers/prerender.html)?

This is one of the main reasons users report that Chrome is "faster" than Firefox even though Javascript performance is probably not noticeably different.

Comment 1

[bug.zilla]

Is this already included in Firefox? If so I can close this bug.

Comment 2

[bug.zilla]

If this has already been done, then I can close this?

Comment 3

[François Marier [:francois]]

I have tested this in Nightly 20.0a1 (2012-11-21) and prerender is not supported.

Comment 4

[doncodes]

This is a great feature in Chrome that definitely does _not_ exist in Firefox.  Firefox implements prefetching for single links into the cache.  Prerender on Chrome will render an entire page in the background and hot-swap it when the URL is actually requested, so the page load is instantaneous.

It strikes me that Firefox already has most of this functionality available in the bfcache (i.e. ability to have a hidden but active page that can be activated), so this should be relatively easy.

Comment 5

[doncodes]

This is a great feature in Chrome that definitely does _not_ exist in Firefox.  Firefox implements prefetching for single links into the cache.  Prerender on Chrome will render an entire page in the background and hot-swap it when the URL is actually requested, so the page load is instantaneous.

It strikes me that Firefox already has most of this functionality available in the bfcache (i.e. ability to have a hidden but active page that can be activated), so this should be relatively easy.

Comment 6

[Honza Bambas (:mayhemer)]

Why is this in Networking: Cache?  This more sounds like a meta bug to me.  Will be handled on more then one front.

Comment 7

[(no longer active)]

FWIW I will be mentoring an intern starting soon who will work on this.

Comment 8

[Honza Bambas (:mayhemer)]

I assume this will be based on data collected with the Network Seer?

Comment 9

[(no longer active)]

(In reply to comment #8)
> I assume this will be based on data collected with the Network Seer?

It could be, yes.  Once we have the support in the platform, we can look into various types of heuristics to determine when to prerender content, and the network seer could be one of them.

Comment 10

[Ashish]

We are a big tech company, and we have prerender/prefetch enabled on our products. Firefox, turns out to be the slowest for initial page loads, because of the absence of complete prerendering.

Comment 11

[yair even or]

Do it already Jesus Christ, sounds like a 1 hour task to me. at Least implement DND resolving if you hover a link for Christ sake.

Comment 13

[Caitlin Galimidi [:c8o]]

Francisco,
the FxOS 2.5 Architecture Evolution deliverables are dependent on this bug 730101 now that 1190805 is resolved as duplicate. is that correct? are we blocked in our delivery unless this is resolved in core?

Comment 14

[(no longer active)]

(In reply to Caitlin Galimidi [:c8o] from comment #13)
> Francisco,
> the FxOS 2.5 Architecture Evolution deliverables are dependent on this bug
> 730101 now that 1190805 is resolved as duplicate. is that correct? are we
> blocked in our delivery unless this is resolved in core?

This is not happening by November, as there is no active developer working on it, and in top of that, the implementation will probably take much longer.

Comment 17

[Jonathan Watt [:jwatt]]

Mass bug change to replace various 'parity' whiteboard flags with the new canonical keywords. (See bug 1443764 comment 13.)

Comment 18

[David Bruant]

Note that Chrome gave up on prerendering: https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/0nSxuuv9bBw
Interesting dicussion on the topic here: https://docs.google.com/document/d/16VCYGGWau483IMSxODpg5faZny1FJ6vNK2v-BuM5EhU/edit#heading=h.uajrcfabdbg5

Comment 19

[(no longer active)]

Sorry, changing the resolution was an accident...

Comment 20

[legowerewolf]

Update: Chrome hasn't given up on this yet, and there's a Mozilla positions ticket on it that hasn't seen much/any Mozillan response yet. Thoughts?

https://github.com/mozilla/standards-positions/issues/620

https://github.com/WICG/nav-speculation
https://developer.mozilla.org/en-US/docs/Web/API/Speculation_Rules_API