Running that test case on the latest V3 CogVM (3427) should give you a 
debugger where the receiver got mixed up with some other variable.

Levente

On Fri, 23 Oct 2015, Eliot Miranda wrote:

> Nicolas,
>    what's (where's) the reproducible case?  
> 
> On Tue, Oct 20, 2015 at 8:52 AM, Nicolas Cellier <nicolas.cellier.aka.nice at gmail.com> wrote:
>        BTW, I had a fairly reproducible case, didn't I send it to this list?
>       Ah no, it was pharo...
> 
> ScaledDecimalTests>>testCoercion
>
>       https://www.mail-archive.com/pharo-dev@lists.pharo.org/msg31997.html
> 
> 2015-10-20 15:26 GMT+02:00 marcel.taeumel <Marcel.Taeumel at hpi.de>:
>       And if you update the VM, it will break.
>       And if you update the image, now, it will work again. ;-)
>
>       Well, here is the thing: I am not able to fix that VM/JIT bug although I
>       would rather like to. If you can fix it, please go ahead. If Eliot finds
>       time for it -- well, it works in Spur and Stack versions -- he will surely
>       do it.
>
>       Until then, we should fix it if possible to improve user experience. If you
>       find any other of those hundreds of places, please consider fixing them as
>       well. Not into trunk, but Squeak46 branch.
>
>       This VM bug has been known since mid of September at least:
>       http://forum.world.st/Non-Spur-CogVM-3427-and-3410-cannot-load-PNG-files-anymore-td4850308.html
>
>       Nothing has happened yet.
>
>       Best,
>       Marcel
> 
> 
>
>       --
>       View this message in context: http://forum.world.st/Squeak-4-6-Graphics-mt-316-mcz-tp4856712p4856747.html
>       Sent from the Squeak - Dev mailing list archive at Nabble.com.
> 
> 
> 
> 
> 
> 
> 
> 
> --
> _,,,^..^,,,_
> best, Eliot
> 
>