995651 - Cache-Control for Stale Content

Status: RESOLVED DUPLICATE of bug 1536511

(Core :: Networking: Cache, enhancement, P3)

Description

[Kevin Cox [@kevincox:matrix.org]]

RFC 5861 defines extensions for Cache-Control of stale content.  This allows the browser to serve stale content.  This can improve responsiveness of websites when caches become stale as well as allow a stale version of a site to be used if it is unavailable.

It defines two properties on the Cache-Control header.

"stale-while-revalidate" "=" delta-seconds
  Allows a cache to serve stale content while refreshing itself.

"stale-if-error" "=" delta-seconds
  Allows a cache to server stale content when the server experiences an error.

While these are not used by and browsers at the moment chromium is considering adding support.[0]  It would be nice to have support in firefox as well.

[0] http://code.google.com/p/chromium/issues/detail?id=348877

Comment 1

[Ben Kelly [:bkelly, not reviewing]]

Discussion of stale-while-revalidate on blink-dev is in this thread:

  https://groups.google.com/a/chromium.org/forum/#!msg/blink-dev/zchogDvIYrY/zqdQHn8fPkcJ

Comment 2

[Firefox Bug Husbandry Bot]

Bulk change to priority: https://bugzilla.mozilla.org/show_bug.cgi?id=1399258

Comment 3

[Firefox Bug Husbandry Bot]

Bulk change to priority: https://bugzilla.mozilla.org/show_bug.cgi?id=1399258

Comment 4

[herrernst]

Would love to see that implemented!

Comment 5

[Graham Perrin]

Is this implemented?

The developer of Boostaler https://addons.mozilla.org/addon/boostaler/ commented upon the age of this bug, and draws attention to https://dxr.mozilla.org/mozilla-central/search?q=stale. From amongst those results, I get (for example):

https://dxr.mozilla.org/mozilla-central/rev/7e40e33da3da2640e965a153254594a234231f76/netwerk/protocol/http/nsHttpResponseHead.cpp#935

Comment 6

[Kevin Cox [@kevincox:matrix.org]]

I'm not implementing this. I am not aware of any work on it.

One interesting update is that chromium has now implemented stale-while-revalidate https://bugs.chromium.org/p/chromium/issues/detail?id=348877

Comment 7

[Honza Bambas (:mayhemer)]

As we just did, except the stale-if-error handling which we don't intend to implement.

Comment 8

[brunoais]

I can't seem to be able to find the discussion about "stale-if-error".
Do you know where it is? At least, where may I find it?

Comment 9

[Honza Bambas (:mayhemer)]

(In reply to brunoais from comment #8)

I can't seem to be able to find the discussion about "stale-if-error".
Do you know where it is? At least, where may I find it?

Discussed only during the necko internal meeting, no public record exists.

Comment 10

[brunoais]

What is(are) the reason(s) for such decision?

Comment 11

[Honza Bambas (:mayhemer)]

(In reply to brunoais from comment #10)

What is(are) the reason(s) for such decision?

First, Chrome doesn't implement it either. Other reason is that that particular stale-if-error directive is not that beneficial as s-w-r while more complicated to implement, so the priority to implement it very low. We don't have data on actual use of that directive, tho. We could open a new bug just for s-i-e implementation, but it would be given a very low priority (P5), even for just discussing it.

Comment 12

[brunoais]

Noted.
I think making a bug for discussion is not out of question... Who knows? maybe someone may implement for you.
If someone implements it, would you keep it?

Comment 13

[Honza Bambas (:mayhemer)]

(In reply to brunoais from comment #12)

Noted.
I think making a bug for discussion is not out of question... Who knows? maybe someone may implement for you.
If someone implements it, would you keep it?

If it gets a production quality, meaning to have a test and pass reviews, then probably yes.

Feel free to file the bug under Core:Networking:HTTP.