Bug 66020 - Load each instance of a cached resource if the cache policy says so
Summary: Load each instance of a cached resource if the cache policy says so
Status: RESOLVED DUPLICATE of bug 52153
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: James Simonsen
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-08-10 16:24 PDT by James Simonsen
Modified: 2011-08-10 18:37 PDT (History)
4 users (show)

See Also:


Attachments
Patch (8.25 KB, patch)
2011-08-10 16:24 PDT, James Simonsen
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description James Simonsen 2011-08-10 16:24:00 PDT
Load each instance of a cached resource if the cache policy says so
Comment 1 James Simonsen 2011-08-10 16:24:24 PDT
Created attachment 103553 [details]
Patch
Comment 2 James Simonsen 2011-08-10 16:33:17 PDT
There seems to be a new policy added in change 74807. It's causing problems with Google Docs as reported here:

http://code.google.com/p/chromium/issues/detail?id=86695

I couldn't find anything in either the HTML or HTTP spec that says we can skip reloading instances of the same resource on the same page in this case. So it seems we should probably take this out.
Comment 3 Alexey Proskuryakov 2011-08-10 18:27:04 PDT
The bug report seems rather vague. What exactly does cache policy say? For instance, what are the differences with bug 52044? Is this a duplicate of bug 52153?

I'm glad that cache-simulator.cgi was useful for this test, but it also hides some of the details, which should probably be spelled out in tests themselves. Just add a <p> explaining what exactly is being tested, and mention it in ChangeLog that the behavior matches IE and Firefox, if it does.
Comment 4 James Simonsen 2011-08-10 18:37:15 PDT
(In reply to comment #3)
> Is this a duplicate of bug 52153?

Indeed. Glad that you noticed that.

> 
> I'm glad that cache-simulator.cgi was useful for this test, but it also hides some of the details, which should probably be spelled out in tests themselves. Just add a <p> explaining what exactly is being tested, and mention it in ChangeLog that the behavior matches IE and Firefox, if it does.

Thanks. I'll fix that in a patch for the other bug. I'll also update my patch to only revalidate after the initial load completes, instead of all the time.

*** This bug has been marked as a duplicate of bug 52153 ***