[Xesam] Sortkeys: read this

[Mikkel Kamstrup Erlandsen]

2008/9/23 Philip Van Hoof <spam at pvanhoof.be>:
> On Tue, 2008-09-23 at 17:32 +0200, Philip Van Hoof wrote:
>> http://www.ietf.org/rfc/rfc4790.txt

I am particularly interested in the two sections on "sort keys"
because that is really what we want to generate .The "collation" here
is a comparison function.

I am too tired (haven't really slept since the hackfest ;-)) to fully
grasp the details. The way I read it there is both good and bad news.
Considering:

  "collation specification SHOULD describe the internal transformation
   algorithm to generate sort keys.  This algorithm can be applied to
   individual strings, and the result can be stored to potentially
   optimize future comparison operations.  A collation MAY specify that
   the sort key is generated by the identity function.  The sort key may
   have no meaning to a human.  The sort key may not be valid input to
   the collation."

So while the actual algorithm used to generate the sort keys should be
specced (which is what we want) there is no rule that the generated
sort keys are usable with memcpr().

If we can find a collation conforming to that spec that explicitly
states that memcpr() can be used as comparison function on the
generated sort keys then we should be home free (because according to
the spec the algorithm must be specified).

I believe Mikael Ottela found some ISO standard collation, but I don't
know if it had the desired behavior. Mikael do you have a link?
-- 
Cheers,
Mikkel