On Wed, 23 Oct 2013 21:47:34 -0700, Jon Perryman wrote:

>Sorry, I thought that the DD name was contained in the DATAID variable. It's 
>value was always in a format that ISPF uses for dynamic allocation. I always 
>assumed that it was the DD name.

When the dataset name is specified instead of DD name, then the dataset is 
allocated to the ISPnnnnn DD.
>
And I don't see that ISPF discloses that DDNAME to the programmer.  But
I guess the programmer isn't supposed to need it.

I don't know how I fell into the habit of LMINIT by DDNAME rather than by
DSNAME. Probably a mentor's code that I copied decades ago.

But what parameters are supported by ALLOCATE but not by LMINIT? 
I guess NEW, SPACE, RECFM, LRECL, ...  Others?  Having issued the
ALLOCATE it's simplest to LMINIT by DDNAME.  (I need to test whether
I can LMINIT a VIO data set.  Alas, not a DDNAME allocated to a UNIX
file.  That's *so* 20th Century!  It's a pain to copy a UNIX file to a
temporary data set merely so I can LMCOPY SHRW it to a library.
I suppose Kirk has a better idea.)

Hmmm.  Will 2.1 allow me to DYNALLOC a data set NEW, then immediately
downgrade the ENQ to SHR?  What command?  Assembler only?

Finally, I wonder why ISPF and SMP/E (perhaps others) generate DDNAMEs
with their component prefixes rather than allowing SVC 99 to generate
SYSnnnnn?  Perhaps more clarity reading dumps of the TIOT?

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to