Date: Tue, 16 Feb 1999 17:13:09 -0500
From: Gregory Sereda <gsereda(_at_)maillennium(_dot_)att(_dot_)com>
The spec should make a statement concerning the "case" of the
sieve language. Since there is no statement, I assume that
sieve language is case-sensitive and must be lower case as
shown in the examples.
Good point. I intended case insensitivity but am indifferent. I'd
better make that clear. I'll add this to 2.1:
| Tokens in the ASCII range are considered case-insensitive.
If I edit section 2 again and do any significant rewrite I'll add more
discussion of tokens.
Syntax: [ < ":is" / ":contains" / ":matches" > <key: string> ]
^^^^^^^^^^^^^
This is incorrect, or at least misleading.
It is incorrect, and I've removed the key bit. I don't think that this
is what we want.
In general, moving the tagged arguments before the positional arguments
results in some confusion because the tagged arguments and the positional
arguments which they affect may be interspersed with other arguments.
This, I think, makes sieve less intuitive. Nevertheless, my major need
is that we clearly specify the current sieve and resist the temptation
to tweak the language syntax, except for very good reasons.
I agree with you, and I think several other people do as well (this
opinion was stated in Orlando: we don't like it, but we don't feel like
changing it). I have no problem with allowing tagged arguments anywhere
on the command line.
The kind of comparison is done, such as whether or not the comparison
done is case-insensitive, is specified as a comparator argument to
the test.
add referrence to ADDRESS-PART, for example....
For convienence, the ADDRESS-PART syntax element is defined here as
follows:
Oh, @#^&*. Fixed.
Thanks for the rest of the corrections, they've all been made in the source.
--
Tim Showalter <tjs+(_at_)andrew(_dot_)cmu(_dot_)edu>