See comments in line below.
Lily
-----Original Message-----
From: Markus Hofmann [mailto:hofmann(_at_)bell-labs(_dot_)com]
Sent: Friday, April 06, 2001 10:59 AM
To: Yang, Lily L
Cc: 'Andre Beck'; ietf-openproxy(_at_)imc(_dot_)org
Subject: Re: IRML questions
Lily,
This again raises the question whether we really need this
"case-sensitive" attribute in the rule language. Why not go with the
following: If the protocol the rule refers to is case-senstive, the
matching will be case-sensitive and vice versa. Would this assumption
make sense?
No, because as Andre pointed out, the "case-sensitive" attribute in IRML is
about regular expression matching, not property (i.e. header) name. So
whether or not the property is case-sensitive depends on the protocol, but
the matching depends on IRML or rule authors.
I think typically people would not care, so insensitive is the default. But
if they do care, they can specify it. So I would say lets keep it there.
Would "url" be sufficient, or would it be helpful to also have
something like "server-name", "protocol-version", etc.? What is the
information we need to extract from the protocol header(s) in order to
allow flexible rule authoring?
I think potentially these may all be useful -- it enables more accurate
matching --
"url", "hostname". According to Rob, protocol-version may also be valuable
-- HTTP supports different set of header in different version, so version
checking might be necessary from time to time.
-Markus