On Apr 26, 2010, at 2:55 AM, Eric Covener wrote: > On Mon, Apr 26, 2010 at 2:04 AM, Roy T. Fielding <[email protected]> wrote: >> I am confused. Why is this a good idea? Why is it unexpected to >> encounter a 413 response after a timeout due to a read of chunked >> body, > > The body is not too large, and the server is not rejecting the body > due to size. Isn't 413 a very misleading status?
Er, yeah, 413 is wrong -- I should have looked it up. It should be a 408 (Request Timeout). >> and how does changing it to a 400 somehow prevent a double >> errordoc? Why not just fix the double errordoc bug instead of >> the case that triggers it? > > This was the secondary part to me, but I'll open a bug with an example > (IIUC, filter that adds an error bucket but then also returns a > (non-http) error). Well, if it is fixed as Ruediger explained, then I am fine with it. I just couldn't see the fix in that patch. ....Roy
