On 2017-03-24, at 09:17, Charles Mills wrote: > > Warnings are not much better or different than errors. It sounds great to > say "issue a warning" but the fact is that then the programmer has to either > (a.) live with and ignore RC=4 for the life of the project; of (b.) change > the code. If the former then why issue the warning at all; if the latter > then it is no better or different than an error. I think warnings in this > situation are kind of wishy-washy: either it's wrong or it's not. I get the > difficulty in deciding whether LHI 0,X'FFFF' is wrong or not, but saying it > is half wrong (4 rather than 8) is just bailing on the decision. > Agreed. It appears that some instructions report a range violation by a Logical Immediate operand with a warning; others as an error. I see no rationale for this. (I'd call them errors.)
And no alert is noted for Logical Immediate Fullword violations because 2**32-1 can not be represented as an expression value. -- gil
