[Webkit-unassigned] [Bug 201492] Keeps running obsolete WebProcess-es for too long

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=201492

--- Comment #11 from Carlos Garcia Campos <cgarcia at igalia.com> ---
(In reply to Milan Crha from comment #8)
> (In reply to Chris Dumez from comment #6)
> > Therefore, this indicates that Evolution is process-swapping and
> > that the cache is useful to it.
> 
> It's questionable whether it's useful. It's not from my point of view. As
> said above, the test-wk2.c only switches between two pages, but the
> processes are changing anyway, thus the cache doesn't work for it (might be
> due to some other constraint).
> 
> (In reply to Michael Catanzaro from comment #7)
> > That should never happen. Instead the page_id should change depending on
> > which web process is in use.
> 
> I'll surely check that out with the webkit-2.26 branch to see how it works.

Note that I'm going to disable PSON in 2.26, so you only need to care about this in trunk. Page id returned by the web view will never change in 2.26.

> > If you want to avoid the process swaps, changing URLs from :// to :/// like
> > we suggested earlier should *probably* suffice.
> 
> I'd like not to add workarounds for current behaviour. It led to the
> disaster with 2.26, where evo expects certain behaviour, which had been
> changed now. Thus I'd prefer to provide universal solution (if it's doable
> at all, aka in some extent universal).

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20190906/4b4f322c/attachment-0001.html>