PATCH for [ARM] subsequent use of plus and minus operators could be improved

[Richard Earnshaw]

> Ok. I moved my code into postreload phase. It works fine on ARM.
> I also tested on i386 with no problem, but no difference 
> occurs between normal and patched run.
> 
> I've never used other targets, so I don't know if the 
> problem occurs on them, or not. Could someone show me 
> another platform with an example where this problem occurs?

It's more likely to be a problem on RISC machines than CISC ones, since 
its a consequence of immediate constants having a limited range.

> Should I attach the patch for the generic solution of the 
> problem?

We need to see the patch.

R.

BTW.  Have you filed a copyright assignment with the FSF?