From: Ann W. Harrison
Date: December 8 2008 8:15pm
Subject: Re: Please review fox for bug#34479
List-Archive: http://lists.mysql.com/falcon/276
Message-Id: <493D8059.8070801@mysql.com>
MIME-Version: 1.0
Content-Type: text/plain; format=flowed; charset=ISO-8859-1
Content-Transfer-Encoding: 7BIT

Vladislav Vaintroub wrote:
> 
> I believe there is nothing with Falcon internals. Prior solution seems to be
> a kludge, some workaround, possibly insufficient understanding of what one
> could do with strings library. IMO Falcon relies on MySQL collation for
> comparison and fail to see why on earth should we removing nowhere
> documented minChar.

The design of Falcon index keys calls for a representation of the key
value that produces the correct ordering from a byte-by-byte comparison.
For numeric keys with fewer than 19 digits of precision we use a
transformed double precision representation.  For string keys, we get
a byte-wise comparable transformation from the server that preserves
the semantics of the collation.  For strings we truncate trailing
spaces.  For numeric keys we truncate trailing zeros.

In the case of compound keys, there is additional processing that
allows us to truncate non-significant trailing characters (or bytes
when numeric) on leading segments of the key.

At least that's the way it was designed to work.

Cheers,

Ann