On 12 February 2014 18:22, Gerhard Postpischil <gerha...@charter.net> wrote: > On 2/12/2014 6:06 PM, Tony Harminc wrote: >> >> I object far more to returning an address of 0 than to accepting a >> length of 0 on the request. To be sure, you are allowed to store no >> more than 0 bytes in your obtained area, so the 0 address sounds >> reasonable, but some instructions are allowed by the architecture to >> recognize access exceptions in the case where no data is stored, e.g. >> STCM with a zero mask.
> The way I get and free storage, I'd be happier with an abend (R form) or > non-zero return code. Sure, but as said, that's not going to happen for reasons of compatibility. > I generally save the length, and then do a FREEMAIN > (or STORAGE RELEASE) using the saved length. If the GETMAIN/OBTAIN was > accepted, freeing zero length is an instant disaster. You can't free 0 bytes at address 0? Now that is an inconsistency. Ah - the subpool thing on FREEMAIN. Is that also true for STORAGE RELEASE? >> So to be consistent, a non-zero address with appropriate access setup >> (primarily key) should probably be returned. > Consistent with what? It can't return an address because one wasn't assigned. Sure - it could assign one. It wouldn't have to be unique; just access-exception correct. Tony H. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN