Hmmm. Seems to me that the advantages of the new instructions would have to be pretty dramatic to justify the additional testing time such a strategy would mandate.
i ------ Original Message ------ Received: 05:48 PM COT, 10/17/2010 From: john gilmore <[email protected]> To: [email protected] Subject: Re: YOPs to be avoided? > There seem to me to be some separable issues here. > > The notion that long-displacement instructions should be avoided where they are available seems to me to be a very dubious one. That said, it is clear beyond cavil that they cannot be used where they are not available. > > Resort to the macro language---I have made my view that those who cannot use it are in effect crippled as assembly-language programmers clear on many occasions---provides simple solutions. > > There is an OPTABLE assembler option, and there is a system variable symbol, &SYSOPT_OPTABLE, that makes the value of OPTABLE that is in effect for a particular assembly available at assembly time. It is thus easy to use the usual conditional-assembly machinery of the HLASM to assemble one set of instructions for modern hardware and and another, different set of instructions for obsolescent hardware where there is a need to support such obsolescent hardware. I have done this, twice; and on both occasions I found it straightforward to do, although it would be disingenuous not to concede that it is also tedious to do. > > To support USING ranges in assembling long-displacement instructions would be, I think, to deprive these instructions of their raison d'être. > > It may well be possible to equip the HLASM with machinery for issuing suppressible warning messages when such USING-range excursions are detected. Whether doing so would constitute a good use of the very limited resources that appear to be available for maintaining the assembler is an again separable question that is worth examining. > > John Gilmore > Ashland, MA 01721, USA >
