Closed Bug 481887 Opened 15 years ago Closed 15 years ago

How to cache and expire entries with bad date-headers

Categories

(Core :: Networking: Cache, defect)

defect
Not set
normal

Tracking

()

RESOLVED INVALID

People

(Reporter: bjarne, Unassigned)

Details

This is a spin-off from bug #203271. Ser https://bugzilla.mozilla.org/show_bug.cgi?id=203271#c27 .

RFC 2616 seems to assume that clocks on client and server are reasonably synchronized, but this may not always be true. The question is how a client should cache and expire entries from a server with a really different clock.
The algorithm in 2616 is pretty robust, in terms of clock skew; the only case it requires reasonable synchronisation in is if there's a HTTP/1.0 cache (i.e., one that doesn't generate Age) in the path. AFAIK almost all known caches have generated Age for at least a decade, so this shouldn't be a worry.
... or do you mean something else by "bad date headers" in the bug title? What circumstances in particular are you concerned about?
I was originally thinking about the example referred to in comment #0 where the clock on the server is one year before the clock on the client. This seemed to fail in FF, but it turned out the test-case was faulty and that FF actually handles this. See bug #203271#35 .

Marking this defect INVALID.
Status: NEW → RESOLVED
Closed: 15 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.