In a recent note, Thomas H Puddicombe said: > Date: Tue, 16 Jan 2007 12:06:28 -0500 > > If SMS, and if DSORG=PS, then at allocation time, SMS will put a valid EOF > marker at the beginning of the newly allocated DASD extent. Without > knowing that the dataset is supposed to be physical sequential, SMS does > nothing to the newly allocated DASD extent - because the space may be > custom-formatted by the requestor using EXCP instead of a known access > method. > > RECFM, BLKSIZE, and LRECL might mean something to your dataset, but mean > nothing to the "semi-hysterical access method" I just dreamed up - so > there's no reason to populate those fields for my "SHAM" file. And I want > no EOF marker anywhere in my DASD extent (CA-DATACOM doesn't like them, > either IIRC). > Ummm... We've been through this one before. There's no guarantee that the first extent of a newly-allocated uninitialized data set doesn't begin with (or contain somewhere) an EOF. How does CA_DATACOM deal with this eventuality? Isn't your hypothetical "SHAM" subject to the same hazard? If allocation were unconditionally to write an EOF at the beginning of every newly allocated extent it would introduce no new hazard; it would merely make the behavior more predictable. Predictability is a good thing, except in cryptography.
However, I agree with not setting RECFM, LRECL, and BLKSIZE to non-zero values (although SMS is free to do so). o Such empty fields are an indicator to the DCB OPEN exit (or to SDB -- SDB is a good thing) that it must supply the information. o There is no guarantee that any BLKSIZE arbitrarily chosen would be compatible with the content of the uninitialized extent. Yet, unconditionally entering EODAD on any attempt to read beyond DS1LSTAR _before_ formatting the CCW or issuing the EXCP would avoid Charles's problem. -- gil -- StorageTek INFORMATION made POWERFUL ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
