pagespeed-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Joshua Marantz <jmara...@google.com.INVALID>
Subject Re: Emitting Cache-Control: immutable
Date Wed, 29 Nov 2017 15:35:14 GMT
OK anyway declaring it 'immutable' clarifies intent in a manner that's
consistent with the wording quoted, so I'm in favor of it.


On Wed, Nov 29, 2017 at 10:18 AM, Otto van der Schaaf <oschaaf@we-amp.com>
wrote:

> Quoting one of the comments from Patrick McManus:
>
> *"The core concept is that without immutable the concepts of fresh and
> current are*
> *not clearly separated. http caching allows for fresh (i.e. cache
> replayable) data *
> *that isn’t the most current representation of a resource – and of course
> firefox *
> *normally uses that. But when you press refresh we in the past have
> revalidated *
> *even fresh things (i.e. with a high max age) to check to see if they were
> still the *
> *most current. immutable lets the cache know that there is never more than
> 1 *
> *version of that resource, so we only need to worry about freshness and
> never *
> *re-validate under those circumstances."*
>
> From what I understand the behavior of browsers should change when pages
> are
> reloaded, for any resources linked that have cache-control: immutable.
>
> But.. I just tested, and I can not confirm that Chrome actually uses
> revalidation
> when I reload pages with cache-extended resources over at modpagespeed.com
> .
> It pulls them from its cache instead.
> So it looks like this needs some more testing with other browsers, and
> perhaps
> proxies, to see if there are clear advantages.
>
> Otto
>
>
> On Wed, Nov 29, 2017 at 3:31 PM Joshua Marantz <jmarantz@google.com.invalid
> >
> wrote:
>
> > This sounds like a reasonable idea to me.  I agree that .pagespeed. URLs
> > are a good candidate for this, as long as the hash is matching and
> > therefore we are sending 1-year caching directives.
> >
> > However I'm confused why about why revalidation would be needed in such
> > cases, even without 'immutable', unless the item has been in cache for >1
> > year.
> >
> >
> > On Thu, Nov 23, 2017 at 3:33 PM, Otto van der Schaaf <oschaaf@we-amp.com
> >
> > wrote:
> >
> > > Hi all,
> > >
> > > I wanted to discuss emitting the "immutable" cache-control directive in
> > > responses for requests to .pagespeed. urls.
> > >
> > > After reading up on this fairly new cache-control directive, think we
> may
> > > be able to avoid revalidations by doing so.
> > > The RFC mentions versioned urls as a candidate for doing this, which I
> > > think also includes our fingerprinted .pagespeed. urls
> > > (these have a hash that changes when any of the underlying resources
> > > change)
> > >
> > > Rough implementation:
> > > https://github.com/pagespeed/mod_pagespeed/compare/oschaaf-
> > > cc-immutable?expand=1
> > >
> > > Context:
> > > https://hacks.mozilla.org/2017/01/using-immutable-
> > > caching-to-speed-up-the-web/
> > > https://tools.ietf.org/html/rfc8246
> > >
> > > Would love to hear thoughts on this!
> > >
> > > Otto
> > >
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message