1020419 - Extend nsHttpChannel to respect nsICacheEntry's forcedValidation

Status: RESOLVED FIXED

(Core :: Networking: HTTP, enhancement)

Comment 1

[Jeremy Poulin [:jaypoulz]]

Attachment: bug-1020419-fix.patch

Obviously this depends on #1020416, but I figured since I had some extra time, I'd start implementing this. I think I added the clause in the right place, but I'm wondering if there are any other major parts of the workflow in the file that are affected by cache validation.

Comment 2

[Honza Bambas (:mayhemer)]

Comment on attachment 8438081 [details] [diff] [review]
bug-1020419-fix.patch

Review of attachment 8438081 [details] [diff] [review]:
-----------------------------------------------------------------

yep, that's it :)

::: netwerk/protocol/http/nsHttpChannel.cpp
@@ -2144,5 @@
>          (entitySize != cachedContentLength)) {
>          LOG(("nsHttpChannel::ProcessPartialContent [this=%p] "
>               "206 has different total entity size than the content length "
>               "of the original partially cached entity.\n", this));
> -        

no ws only changes please.

Comment 3

[Jeremy Poulin [:jaypoulz]]

Attachment: Removed ws change.

Comment 4

[Jeremy Poulin [:jaypoulz]]

Attachment: bug-1020419-fix.patch

Actually removed the ws changes.

Comment 5

[Honza Bambas (:mayhemer)]

Comment on attachment 8458820 [details] [diff] [review]
bug-1020419-fix.patch

Review of attachment 8458820 [details] [diff] [review]:
-----------------------------------------------------------------

Push to try before landing.

::: netwerk/protocol/http/nsHttpChannel.cpp
@@ +2806,5 @@
>      bool canAddImsHeader = true;
>  
> +    bool isForcedValid;
> +    rv = entry->GetIsForcedValid(&isForcedValid);
> +    NS_ENSURE_SUCCESS(rv,rv);

hmm.. app cache entries don't implement this.  if the method fails, just set isForceValid = false and go on (no need to log, ns_warn etc...).

Comment 6

[Jeremy Poulin [:jaypoulz]]

Attachment: bug-1020419-fix.patch

Of course I'll push to try, but right now 1020416 is still in flux with Michals stuff so I'd rather wait for that to stabilize.

Comment 7

[Jeremy Poulin [:jaypoulz]]

Comment on attachment 8458857 [details] [diff] [review]
bug-1020419-fix.patch

Hi Pat - would you mind glancing over this since Honza is on PTO?

Comment 8

[Jeremy Poulin [:jaypoulz]]

https://tbpl.mozilla.org/?tree=Try&rev=b12c1d38a1da

Comment 9

[Patrick McManus [:mcmanus]]

Hey Jeremy - this seems sane to me, but I'm not real familiar with the new cache code yet. I'll delegate it to michal.

Comment 10

[Patrick McManus [:mcmanus]]

Comment on attachment 8458857 [details] [diff] [review]
bug-1020419-fix.patch

Review of attachment 8458857 [details] [diff] [review]:
-----------------------------------------------------------------

delegate to michal

Comment 11

[Jeremy Poulin [:jaypoulz]]

https://tbpl.mozilla.org/?tree=Try&rev=73f6656e05c2

Comment 12

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/27a066378f69

Comment 13

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/27a066378f69