No, you can't preserve attribute order, period. Doing so would constrain the implementation and add performance overhead. Rewrite the other application to not be sensitive to this.
It could be worse. We could have used a hashtable rather than an insertion-sort to drive our by-name lookup of attributes, and the resulting sort order wouldn't make even this much sense... and it'd _still_ be correct behavior. ______________________________________ Joe Kesselman / IBM Research --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
