972406 - [e10s] transition after clicking a tile on the new tab page is weird

Status: RESOLVED WONTFIX

(Firefox :: New Tab Page, defect)

Description

[Tim Taubert [:ttaubert] (inactive)]

After clicking a tile on the new tab page, the remoteness of the <browser> is updated and set to remote="true". This leads to the page becoming a white canvas immediately and then loading the new page. Usually the old page stays visible for a while.

This will probably be fixed with bug 776477 but other chrome pages might suffer from the same problem.

Comment 1

[Matthew N. [:MattN]]

Mass-move to Firefox::New Tab Page.

Filter on new-tab-page-component.

Comment 2

[u428464]

Seems to happen when you navigate away from any in-content page (tested with add-ons and preferences).

Comment 3

[u428464]

Could this be put at an higher priority : it's what makes me disable e10s at the moment ?

Comment 4

[Tim Taubert [:ttaubert] (inactive)]

I agree this is slightly annoying but fixing this isn't too easy. We need to make about:newtab unprivileged, same goes for about:preferences, etc.

But... is that really the thing keeping e10s disabled? Come on, it's really not that bad and doesn't make the browser unusable in any way.

Comment 5

[u428464]

(In reply to Tim Taubert [:ttaubert] from comment #4)
> But... is that really the thing keeping e10s disabled? Come on, it's really
> not that bad and doesn't make the browser unusable in any way.

Yes I really hate those white frames : it makes the webpages loading feel sluggish.

Comment 6

[u428464]

Any progress on this ? I don't really see how e10s could land without this (meaning fixing the weird behavior with all about pages).

Comment 7

[:Gijs (he/him)]

(In reply to Guillaume C. [:ge3k0s] from comment #6)
> Any progress on this ? I don't really see how e10s could land without this
> (meaning fixing the weird behavior with all about pages).

I'm fairly confident that moving about:preferences and about:addons into the content process is not going to happen before we ship e10s. Doing so would be very complex and take a lot of time... Perhaps there is another way to work around this, but I also agree with Tim that this is just not that big of a deal.

about:newtab is more annoying, and I don't really understand why that's not in the content process yet - I believe it's being held up by changes as a result of the "remote new tab page" project, but I don't know why and/or if that's close to being resolved. Also not the subject of this bug though.

Comment 8

[u428464]

(In reply to :Gijs Kruitbosch from comment #7)
> (In reply to Guillaume C. [:ge3k0s] from comment #6)
> > Any progress on this ? I don't really see how e10s could land without this
> > (meaning fixing the weird behavior with all about pages).
> 
> I'm fairly confident that moving about:preferences and about:addons into the
> content process is not going to happen before we ship e10s. Doing so would
> be very complex and take a lot of time... Perhaps there is another way to
> work around this, but I also agree with Tim that this is just not that big
> of a deal.
> 
> about:newtab is more annoying, and I don't really understand why that's not
> in the content process yet - I believe it's being held up by changes as a
> result of the "remote new tab page" project, but I don't know why and/or if
> that's close to being resolved. Also not the subject of this bug though.
I agree that this is especially annoying about the new tab page and in a lesser extent about the home page. These two should really get fixed before e10s hit stable (Firefox 46 if I'm not mistaken).

Comment 9

[:Gijs (he/him)]

(In reply to Guillaume C. [:ge3k0s] from comment #8)
> (In reply to :Gijs Kruitbosch from comment #7)
> > about:newtab is more annoying, and I don't really understand why that's not
> > in the content process yet - I believe it's being held up by changes as a
> > result of the "remote new tab page" project, but I don't know why and/or if
> > that's close to being resolved. Also not the subject of this bug though.
> I agree that this is especially annoying about the new tab page and in a
> lesser extent about the home page. These two should really get fixed before
> e10s hit stable (Firefox 46 if I'm not mistaken).

It won't, at the current prioritization ("later"), and moving them into the content process and uplifting that change to beta isn't really going to happen, I don't think. I disagree with you about this bug's importance.

We're shooting for 46 but that might slip. Either way, if we want this prioritized we should actually ping Philipp and not remove the needinfo. Philipp, if you think this is important can you outline why/how and re-request e10s prioritization? Thanks.

Comment 10

[(Currently slow to respond) Philipp Sackl [:phlsa] (Firefox UX) please use needinfo]

The short answer is that I don't think that this should block us from shipping e10s.

The longer but much more fascinating answer is that this could actually be an opportunity. I recall reading about some paper(s) on perception of responsiveness, that mentioned exactly this behavior (making a screen blank after a user action, before it displays something). IIRC, the result was that the system was perceived as faster with that »blanking« than without it.

Comment 11

[Tim Taubert [:ttaubert] (inactive)]

If we really wanted to leave that in because it might be perceived as faster, I think in the long run we should make it faster (i.e. put about:newtab into the content process) *and* blank the page after the user clicked a link :)

Comment 12

[u428464]

I tend to disagree about the extra frame. I think it not really nice to have this flash in front of you every you leave the new tab page, but I do agree that an animation (Android style for example) would be nice (ie having the thumbnail animate to take the whole window).

Comment 13

[Jessilyn Davis]

Hello!

This bug has been closed due to inactivity and/or the potential for this bug to no longer be an issue with the new Discovery Stream-powered New Tab experience.

Please help us triage by reopening if this issue still persists and should be addressed.

Thanks!