<<< Chronological >>> Author Index    Subject Index <<< Threads >>>

Re: thoughts on updates for obj with guarded attribute


Daniel Karrenberg <Daniel.Karrenberg@localhost writes
 *   > if
 *   >    a guarded attribute does match exactly or is omitted altogether.
 *   >    Warning(s) should be generted only if there is a mismatch.
 * 
 * You are perfectly right. This is how it should work.
 * 
 * Goes on the to-be-fixed list.

To be honest, this should already work. In the last check before updating we
do [updatecheck.pl]:


sub addguard {

	local(*new, *cur) = @_;
	local($stat) = $OK;

        foreach $i (split(/\s+/, $GRD{$type})) {
		if ($new{$i} ne $cur{$i}) {
                	$new{$i} = $cur{$i};
                	&addwarning(*new, "guarded field $ATTL{$i} reset to current value");
		}
        }
	return;
}

which checks that the new guarded values are the same as the old values, and
only adds a warning if there is a difference. The only time when you can get
a warning even without changing the fields is when the values are all the
same but the ordering has changed. This needs fixing. 

Wilfried, if you still have the example you sent in, please forward it to us
to see if we can reproduce and fix.

Cheers from Interop,

-Marten



<<< Chronological >>> Author    Subject <<< Threads >>>