On 02.05.2007 14:21, Niklas Edmundsson wrote: > > > It did, but it caused the freshness calculation to overflow so the end
Ouch. > result was bollocks. I hard-coded 100 years for the max-stale without > value case, not pretty but it works. I would suggest using *one* year here. Although 14.9.3 of RFC2616 states that max-stale without any value means that the client is willing to accept a stale response of *any* age I think this situation is somehow connected to the case where the server wants to mark a response as "never expires" via an Expires header (14.21). In this case the server sends an expires date one year from the time the response is sent. Comments? > > Updated patch attached. Looks fine to me except for the "100 years issues". Thanks. Regards RĂ¼diger