This is an old discussion. Why is storage allocated in doublewords on doubleword boundaries? This is done because the most exigent historical alignment requirements are met by doing so. (Very specialized quadword-alignment requirements do now sometimes arise.)
Storing current length in a signed halfword yields control-block overflow at 32768. On 2/5/13, Paul Gilmartin <[email protected]> wrote: > On Tue, 5 Feb 2013 11:11:04 -0500, John Gilmore wrote: > >>32760 = max[n | (n <= 32767) & (n = 0 mod(8))] >> > But why "n = 0 mod(8)"? The PARM length is specified in single > bytes in a halfword. As an experiment, I've passed a PARM of > 32767 bytes to HLASM, which processed it correctly while > misbehaving badly on 32768. > > -- gil > John Gilmore, Ashland, MA 01721 - USA ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
