1037721 - Figure out whether we can get rid of the forceGet argument to location.reload

Status: NEW

(Core :: DOM: Core & HTML, defect, P5)

Description

[Boris Zbarsky [:bzbarsky]]

It's not there in the spec.

Comment 1

[Boris Zbarsky [:bzbarsky]]

So both W3Schools and MDN claim the arg exists.  However location.reload.length == 0 in every single non-Gecko-based browser I can find.  And at least source inspection suggests webkit/blink don't have that arg.

In our tree, we have a dozen or two uses of reload(true) at first glance, mostly in tests.  I don't know about non-literals being passed in, or extensions.  :(

Comment 2

[Jake Archibald [:jakea]]

Interested to know Ilya Grigorik's take on this. I've seen him use location.reload(true) in examples.

Comment 3

[Anne (:annevk)]

The bug on the specification for adding this turned out to be INVALID. I didn't know it was only Gecko that supported this. Do please reopen that bug if the circumstances here change:

  https://www.w3.org/Bugs/Public/show_bug.cgi?id=27764

Comment 4

[Ilya Grigorik]

FWIW, I do believe "forceGet" is being used in the wild by some popular sites. The general pattern is:

- http://www.stevesouders.com/blog/2012/05/22/self-updating-scripts/
- http://www.stevesouders.com/blog/2013/02/26/reloading-post-onload-resources/
- http://www.phpied.com/update-a-far-future-expiring-component/

Comment 5

[Jake Archibald [:jakea]]

"forceGet" always worked in Chrome because we make conditional requests on `location.reload()`, so we were able to cache-bust resources with a future max-age even without `location.reload(true)`. Appears that Firefox also makes conditional requests on `location.reload()`, so the boolean can be dropped and work as Chrome does now.

Maybe Firefox used to fetch resources with future max-age from the cache on `location.reload()`? That might explain why the boolean used to be useful.

Comment 6

[Boris Zbarsky [:bzbarsky]]

No, having the boolean false still forced cache revalidation via conditional request.

The boolean being true does a lot more than do conditional requests.  It completely bypasses the cache, it forces creation of a new HTTP connection instead of reuse of an existing one, it disables pipelining on that connection, and if we're doing an HTTP 1.1 request sends a "Cache-Control: no-cache" on the outgoing request (and a "Pragma: no-cache" if doing HTTP 1.0), to bypass proxy caches.

The boolean being false will do "Cache-Control: max-age=0" on the request if HTTP 1.1, else also "Pragma: no-cache".  And of course it'll force the use of a conditional request no matter what max-age our cache has.  But it won't do those other bits.

Comment 7

[Jake Archibald [:jakea]]

Sorry, I wasn't clear. I was referring to the posts Ilya posted. I mean their techniques still work without boolean-true, because of the conditional requests, which is why they work in Chrome too.

Comment 8

[Firefox Bug Husbandry Bot]

https://bugzilla.mozilla.org/show_bug.cgi?id=1472046

Move all DOM bugs that haven’t been updated in more than 3 years and has no one currently assigned to P5.

If you have questions, please contact :mdaly.