Am 03.10.2008 um 19:37 schrieb Marc Mims:
* Matt S Trout <[EMAIL PROTECTED]> [081001 18:40]:
I'll happily give you svn space for it and we can figure out how to
deprecate
IC::File in favour of it later.
Just let me know when you're ready for it and get it into svn.
Consider $self->ident_values to get the PK values, IIRC.
Looks like it must be $self->_ident_values. ::PK::id returns:
(wantarray ? @pk : $pk[0]);
Although, many of the tables I work with have an 'id' column, so I end
up with an accessor that hides ::PK::id. It returns the same thing,
but
others might have an 'id' column that is not the PK.
In the interest of making it generic, I should probably not rely on
the
PK being an int, but the directory tree layout currently depends on
it,
and I think it is a reasonably sane default.
Rather than changing the int PK requirement, perhaps anyone wanting
something different should either use the ::SHA1 subclass or derive
their own. Thoughts?
The ::SHA1 subclass, should be able to accommodate any PK type.
However, I need an algorithm to join the segments of a multi-segment
PK
so it's always unique. A simple join isn't sufficient:
'A', 'B' => 'AB'
'AB', '' => 'AB' # clash
You can SHA each column of the primary key and then SHA the joined SHAs.
SHA(SHA(A).SHA(B)) != SHA(AB);
greets,
moritz
_______________________________________________
List: http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/dbix-class
IRC: irc.perl.org#dbix-class
SVN: http://dev.catalyst.perl.org/repos/bast/DBIx-Class/
Searchable Archive: http://www.grokbase.com/group/[EMAIL PROTECTED]
