Greg is right.

While ECB=(DUMPECB,WRITE) would not really consider DUMPECB to be an 
A-type address (that being something more typical for list form where the 
expansion is expected to use an A-con), it is the case that the expansion 
for this is:
L     15,DUMPECB                   ADDRESS OF ECB         @L8A 
ST    15,16(0,1)             ADDR OF ECB INTO PARAM LIST  @L8A 

Which clearly expects DUMPECB to contain the address of the ECB.

It seems that "If an A-type address is specified" should simply be 
deleted. 

I don't defend this implementation, and it is certainly true that this 
sort of "what level of indirection is required" situation is all too often 
far too confusing in many IBM macros.
But it is what it is, and cannot compatibly be changed (and it is unlikely 
that someone would accept a requirement to add new analogous keywords such 
as ECB_NAME and ECB_ADDR by which it would be more clear on the invocation 
whether what was being provided was "the ECB" (expecting to take the 
address of it) or "the address of the ECB" (expecting, as in this case, to 
take the contents of it).

So the request as coded probably is treated as "no ECB at all" if the ECB 
is initialized to 0 (or the system attempts to post location 0).

Peter Relson
z/OS Core Technology Design

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

Reply via email to