To me, the bug appears in use of the word octets, as it implies HTTPS without any Transfer-Encoding: header. I think it is easier known if length is determinate or indeterminate on the wire; otherwise, the numeric value is an assertion in HTTP. On Tuesday, December 4, 2012, Adrien W. de Croy wrote: > > ------ Original Message ------ > From: "Willy Tarreau" <w@1wt.eu> > >> Hi Martin, >> >> On Tue, Dec 04, 2012 at 03:08:01PM -0800, Martin Thomson wrote: >> >> >>> On 4 December 2012 15:05, Zhong Yu <zhong.j.yu@gmail.com> wrote: >>> >>>> On Tue, Dec 4, 2012 at 3:30 PM, Willy Tarreau <w@1wt.eu> wrote: >>>> >>>>> 6. If this is a request message and none of the above are true, >>>>> then >>>>> the message body length is zero (no message body is present). >>>>> >>>> >>>> I think it should simply state >>>> >>>> 6. If this is a request message and none of the above are true, >>>> then >>>> the message contains no body. >>>> >>> >>> Is it really useful to distinguish between no body and body with no >>> content? I can't imagine a use for such a distinction. >>> >>> >> >> I think the example with the POST that is rejected without a >> content-length >> is valid, I have already observed this one, though I don't remember on >> what server. >> >> maybe that's a bug in that server? > > Adrien > > >> >> Willy >> >> >> >> >> > >
Received on Wednesday, 5 December 2012 00:44:02 UTC