Re: texteq/byteaeq: avoid detoast [REVIEW]

On Mon, Jan 17, 2011 at 2:56 AM, Peter Eisentraut <peter_e(at)gmx(dot)net> wrote:
> On mån, 2011-01-17 at 07:35 +0100, Magnus Hagander wrote:
>> For text, I think locales may make that impossible. Aren't there
>> locale rules where two different characters can "behave the same" when
>> comparing them? I know in Swedish at least w and v behave the same
>> when sorting (but not when comparing) in some variants of the locale.
>>
>> In fact, aren't there cases where the *length test* also fails? I
>> don't know this for sure, but unless we know for certain that two
>> different length strings can never be the same *independent of
>> locale*, this whole patch has a big problem...
>
> Currently, two text values are only equal of strcoll() considers them
> equal and the bits are the same.  So this patch is safe in that regard.
>
> There is, however, some desire to loosen this.  Possible applications
> are case-insensitive comparison and Unicode normalization.  It's not
> going to happen soon, but it may be worth considering not putting in an
> optimization that we'll end up having to rip out again in a year
> perhaps.

Hmm. I hate to give up on this - it's a nice optimization for the
cases to which it applies. Would it be possible to jigger things so
that we can still do it byte-for-byte when case-insensitive comparison
or Unicode normalization AREN'T in use?

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company