1155151 - [e10s] about:performance should load in the child process

Status: RESOLVED WONTFIX

(Toolkit :: Performance Monitoring, defect)

Description

[Tim Taubert [:ttaubert] (inactive)]

about:performance was added without e10s support. We need to convert it to work and load in a child process.

Comment 1

[David Teller [:Yoric] - still alive but not very active]

Ah, I was planning to have about:performance run in the parent and fetch data from the child process, but we can of course do the opposite.

Comment 2

[David Teller [:Yoric] - still alive but not very active]

Tim, any idea how we can put a about:* in the child process? Do we already have such things on m-c?

Comment 3

[David Teller [:Yoric] - still alive but not very active]

mconley?

Comment 4

[Mike Conley (:mconley) (:⚙️)]

Yes, we do!

Mossop added flags in bug 1083281 that can be used to indicate which process the about: page is allowed to be run in.

What's better, is that he added a bunch of cool infrastructure to make converting these pages over simpler. He did about:plugins in only a few lines.

Here's the documentation for the infrastructure he added: https://developer.mozilla.org/en-US/Firefox/Multiprocess_Firefox/RemotePageManager

And here's where he made about:plugins load in the content process: https://reviewboard.mozilla.org/r/4419/diff/

Comment 5

[Tim Taubert [:ttaubert] (inactive)]

Sorry, I was hesitating to respond here because in bug 1155149 and on IRC Dave and I talked about the browser remoteness switching a tad more and it doesn't seem like we can realistically get rid of it. We should probably make it a first-class citizen of the sessionstore API sooner or later. What I'm saying is that you might not want to waste time converting about:performance as that page will be only one out of a few dozen that would block us from ever resolving bug 1155149.

Comment 6

[Tim Taubert [:ttaubert] (inactive)]

Let's wontfix this.

Comment 7

[David Teller [:Yoric] - still alive but not very active]

It would be useful for performance reasons, in particular since about:performance is relatively slow by itself.

Comment 8

[Florian Quèze [:florian]]

(In reply to David Teller [:Yoric] (away until end of July - please use "needinfo") from comment #7)
> It would be useful for performance reasons, in particular since
> about:performance is relatively slow by itself.

If about:performance is slow we should optimize it (and I intend to do it).