httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From (Rodent of Unusual Size)
Subject Re: regex syntax
Date Thu, 26 Jun 1997 03:32:45 GMT
>From the fingers of Randy Terbush flowed the following:
>Without deep knowledge of why we had to use ~ to invoke the regex 
>parsing, why do we need to have any kind of distinction. It would 
>be more intuitive to me to allow the regex without any special 
>directives or ~ switches.

    I'm not positive either, but I suspect it's to indicate whether
    things like "?" and "*" have their shell metacharacter definitions
    or canonical RE definitions.  Big difference there..

    As I mentioned, I like the idea of differentiating via the directive
    name rather than by overloading the syntax.  (We have too much of
    that already; witness the IP/name vhost confusions engendered by the
    overloading of <VirtualHost>.)

    I think we should decide upon a suffix that can be attached to
    virtually any directive that's appropriate that means "use RE
    processing, not shell metacharacters."  I submit "-Match" as a
    tentative suffix:


    (It's already there for BrowserMatch, after all..)  These would be
    in addition to the current syntaces, which would be downgraded to
    either display a deprecation message on detection of a "~" (but do
    the RE processing anyway) or else not do REs at all.

    The last part is a significant-enough departure from current
    behaviour that I don't think it's a candidate for anything before
    2.0.  The addition of the -Match suffixed directives could happen in
    anything following 1.3 (inclusive) though, IMHO.

    #ken    :-)}

View raw message