[webkit-gtk] 2.14 slower to initialize webviews ?

[Carlos Garcia Campos]

El mié, 21-09-2016 a las 14:17 +0200, Jérémy Lal escribió:
> 2016-09-21 13:54 GMT+02:00 Carlos Garcia Campos <cgarcia at igalia.com>:
> > El mié, 21-09-2016 a las 12:05 +0200, Jérémy Lal escribió:
> > > I noticed version 2.14 is significantly slower to load several
> > web
> > > views one after another.
> > >
> > > By load time i mean the time between a call to
> > > `webkit_web_view_load_uri()` and the first
> > > WEBKIT_LOAD_STARTED change event.
> > >
> > > In version 2.12 it takes half a second for the first load and
> > less
> > > than 100ms
> > > when loading more views.
> > > in version 2.14 it always takes the same time.
> > 
> > So, you mean it takes more time the first load on every web view,
> > right? subsequents loads in the same web view are not slow, right?
> > 
> > If that's the case, yes, it's somehow expected, there's more work
> > to do
> > know when setting up web views, the web process has to create the
> > compositing thread and wait for it to start rendering.
> 
> Yes it is (much) faster to load again an url on an already "loaded"
> web view.
> Still, half a second seems to be an eternity.

Yes, we will try to improve that.

> I'm using webkitgtk quite a lot for server-side DOM "prerendering", a
> bit like
> what is chromium's headless branch for, and i wonder if it's
> something
> that is going to happen in webkitgtk as well (i.e. the ability to
> just disable
> any kind of graphic computations, let alone the ability to not use
> gtk).

Yes, I don't see why not. For now you could try disabling accelerated
compositing using WEBKIT_DISABLE_COMPOSITING_MODE env var.
 
> > > (I'm not sure it's related but for example epiphany 3.22 feels
> > much
> > > slower to open tabs).
> > >
> > > Reproductible on several machines with debian/sid, and different
> > > networks.
> > 
> > I don't think it has to do with networks. I assume you are trying
> > on
> > X11?
> 
> X11, Xvfb, and if i remember well on Wayland i had a similar result,
> but i'll check again and report if it's not.
> 
> Jérémy
> _______________________________________________
> webkit-gtk mailing list
> webkit-gtk at lists.webkit.org
> https://lists.webkit.org/mailman/listinfo/webkit-gtk
-- 
Carlos Garcia Campos
http://pgp.rediris.es:11371/pks/lookup?op=get&search=0xF3D322D0EC4582C3
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: This is a digitally signed message part
URL: <https://lists.webkit.org/pipermail/webkit-gtk/attachments/20160921/8da1ef88/attachment-0001.sig>