Re: #290: Motivate one-year limit for Expires

[Mark Nottingham <[e-mail]>]

On 24/07/2011, at 2:16 PM, Brian Pane wrote:

> I think the problem with ignoring on the client side is that the
> clients that would need to ignore large values the most are probably
> those least likely to detect the problem.  E.g, if you send this:
> 
>    Expires: Fri, 01 Jan 2100 00:00:01 GMT
> 
> to a client that uses 32-bit Unix time_t values to hold times, how
> will that client interpret it?  The implementation might be able to
> detect the overflow, but maybe not.


Right -- to me, that's a bigger concern, and we might amend the proposed text to note that specifically. I don't know that we can do much else; implementations are just going to have to change how they store and work with time as it approaches...

Cheers,

--
Mark Nottingham   http://www.mnot.net/