Kjetil Torgrim Homme wrote:
On Thu, 2005-09-22 at 13:26 -0400, Mark E. Mallett wrote:
On Wed, Sep 21, 2005 at 04:41:58PM +0100, Alexey Melnikov wrote:
Hi folks,
One of the questions raised during the WGLC of
draft-ietf-sieve-3431bis-01.txt was concerning the following text in the
draft:
3.1. Match Type Value
[...]
Leading and trailing white space MUST be removed from the value of
the message for the comparison. White space is defined as
SP / HTAB / CRLF
I would like to poll people who've implemented the relational draft/RFC
if they've been following this rule.
Yes but probably not because of that draft, but because that's what the
associated tests do, relational or not. As I mentioned, if I had [yet]
implemented the variables draft, I probably would question the
whitespace skipping on the first argument in the "string" test.
well, this stripping only happens for relational tests, not ":is" and
friends. although it's a trap for the unwary, it's not a big problem,
IMO. you can work around it when needed by using explicit begin and end
markers in both strings:
if string :gt "!${var1}!" "!${var2}!" ...
an example of this technique may be useful.
This would work but it is a hack!
I would rather say that for headers, relational strips leading and
trailing spaces, but for other things, like addresses and variables - it
does not.