Re: AI 2000-05-010: proposed interface

[Antoine Leca <yyyyyyyyyyyy@xxxxxxxxxx>]

Ulrich Drepper wrote:
> 
> David Korn <yyy@xxxxxxxxxxxxxxxx> writes:
> 
> > If it is true, that no two characters can have the same collation
> > order, (which I assume is required), then wouldn't == mean that the
> > strings are equal?
> 
> Yes.

I am a bit lost here.

Are you both saying that L"\u00E0" should be differentiated from
L"a\u0300"? (U+00E0 is à, and U+0300 is the combining grave; so these
two sequences are the same as per a Unicode/iso10646 point of view).

I can understand that they should have different collation *sequence*.
But Mr Korn is talking about collation *order* (so it refers to
strcoll rather than the proposed strseq, I presume).

I understand that the difference between coll.sequence and coll.order
is the very point that have driven us so far.

Or is it just a typo?

(and yes, I know Mr Korn overlooked the problem with multicharacter
 collating element, and L"a\u0300" certainly is such an example)


Antoine