[squid-users] Re: squid doubts

Provided the client has properly escaped the multibyte characters into
ascii according to the current standards Squid should have no problem
processing them.

Unicode/multibyte characters in URL paths (the part after the hostname)
should be escaped using the standard URL escapes. The HTTP specification
currently does not accept non-ascii characters in the protocol.

UTF-8 characters in hostnames should according to the last
internet-drafts I have seen on how to manage UTF-8 in DNS be escaped
into ascii representation by the user agent before sent to
itermediaries/proxies (at least this how I read the draft).

What is sure is that you quite likely won't see any unicode support in
Squid using methods which is not likely to become internet standards. If
there is a standards tracks RFC or even IETF internet draft which is
seen as likely to become a standards track RFC then I would be happy to
get them implemented in Squid if any changes to Squid is required.

Implementing non-standard methods which is not likely to become standard
is somewhat counter-productive as it is only a temporary solution, takes
development resources, increases code maintenance cost and also quite
likely further delays the IETF standardization process.

Regards
Henrik Nordström
Squid Hacker
MARA Systems AB, Sweden

Rohit Vatsa wrote:
>
> hi ,
> i wanted to know whether squid supports multi-byte requests sent
> to it. i was making a client that actually supports some languages
> that require unicode support.
>
> so the request that i give (GET somepage) is also in a wide string
> (unicode). now if i send that to squid, will squid be able to
> recognize and service the request ???
> that is, does squid have unicode support ??
> so far in my tests, i have not been able to have any luck, could u
> please help ?
>
> thanks and regards,
> rohit.
Received on Mon Oct 01 2001 - 04:34:26 MDT