Re: Dependent USING Does Not Work as Documented

[Paul Gilmartin]

On 2018-04-24, at 20:55:29, Steve Smith wrote:
>  
> ​Well, a bug should be reported via the regular channels.​  Since I don't
> recall what you're referring to, I can only say that overflow is ignored by
> design in addressing; from the birth of the architecture.
>  
"the birth of the architecture" meant 24-bit addressing.  We've outgrown
that.  But Jonathan's reply does not encourage using "the regular channels.​"

An overflow should not be ignored when it produces a mathematically
incorrect result.  IIRC, my example was something like:

         USING x'7FFFFF00',R2
         LA    R3,-256
-- gil