Amos Jeffries wrote:
> Tsantilas Christos wrote:
>> Hi all.
>> I am sending the second version.
>> Amos Jeffries wrote:
>>>
> <snip>
>>>
>>> src/cf.data.pre:
>>>
>>> The use of [http::] is a bit congruous here. Since the body data is
>>> not necessarily HTTP protocol from the receiver.
>>> Given that there is no other obvious scope, can you change the docs
>>> to say "HTTP-equivalent body data" from the upstream source? That
>>> should allow then the comment about FTP listings.
>> OK. I modified a little the comment
>>
>>> As a corollary what about gopher/whois listings, and other internal
>>> loaded/generated objects such as errors and icons?
>> Unfortunately the patch currently does not handle gopher and whois
>> protocols. I am not sure if it is useful, but if required, i think we
>> can include these protocols.
>
> Gopher certainly is. It generates download bodies and directory listings
> similar to FTP.
>
> I believe if FTP is caught by the addVirginReplyBody() position then
> gopher and whois shopuld already be caught in the counter as well
> without you noticing.
>
> For the documentation you can write "FTP/Gopher listings" or "FTP and
> Gopher listings".
Actually, after reading Alex reply I don't think its worth mentioning
these as special anyway. The term "HTTP-equivalent body" covers
de-capsulating from FTP/gopher/etc to HTTP.
What they send can be considered the same as a compressed encoding which
gets inflated to HTTP body then accounted.
The special mention is only worth saying to imply on-wire accounting
which doesn't count their particular on-wire.
Amos
-- Please be using Current Stable Squid 2.7.STABLE9 or 3.1.6 Beta testers wanted for 3.2.0.1Received on Wed Aug 04 2010 - 10:02:42 MDT
This archive was generated by hypermail 2.2.0 : Thu Aug 05 2010 - 12:00:03 MDT