Bug #53611 for Moose: Attribute insertion order goes weird on reinsertion

Mon Jan 11 18:40:49 2010 david.tulloh [...] AirservicesAustralia.com - Ticket created

Subject:	Attribute insertion order goes weird on reinsertion
Date:	Tue, 12 Jan 2010 10:40:21 +1100
To:	<bug-Class-MOP [...] rt.cpan.org>
From:	"Tulloh, David" <david.tulloh [...] AirservicesAustralia.com>

If you insert an already existing attribute it's insertion order becomes equal to the last object inserted. I would expect it to either retain the original number or increment. Attached is a small test script that demonstrates the bug.

Message body is not shown because sender requested not to inline it.

Mon Sep 12 21:41:50 2011 DROLSKY [...] cpan.org - Queue changed from Class-MOP to Moose

Thu Feb 07 21:56:35 2013 cpan [...] prather.org - Correspondence added

On Mon Jan 11 18:40:49 2010, david.tulloh@AirservicesAustralia.com wrote: Show quoted text

> If you insert an already existing attribute it's insertion order becomes > equal to the last object inserted. > > I would expect it to either retain the original number or increment. > > > Attached is a small test script that demonstrates the bug.

This has been warnocked for two years but better late than never I suppose. This seems like it should throw an error rather than be something allowed. What would re- adding an existing method *mean*? Methods aren't just hash keys, they (potentially) carry a lot of behavior along with them. So I think this should just throw an error rather than either of the options presented here. A patch is welcome! -Chris

Thu Feb 07 21:56:38 2013 The RT System itself - Status changed from 'new' to 'open'