[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Questions on extensible matching rules

Okay, let's restate things:

A matching rule match function needs to know:
  a) the matching rule (functions can support multiple rules)
  b) the syntax of the asserted value
  c) the syntax of the stored value

[ b) is explicitly part of a).  c) might be missing. ]

Before calling the match function, the asserted value
should be normalized per the assertion syntax AND the
stored value should be normalized per its syntax.

The normalization of the asserted value may be specific
to the matching rule, but not specific to any particular
stored value syntax that the matching rule might support.

The normalization of the stored value could be specific
to the matching rule [but the current design doesn't
support this].

I believe the best thing to do is to remove the syntax
specifier from the call which normalizes the asserted
value.  The syntax is part of the matching rule (which
is provided).

Is this enough?