At 10/9/2007 02:53 PM, Ed Gould wrote:
Not speaking for Rick, but in agreement with him. As he mentioned
the viability of putting into production a $$ tool . I have seen
various programmers try to put various debugging tools into
production and have seen it slip by (into production once or twice)
don't go there. Being called (as a sysprog) at xx AM because the
tool doesn't work or its causing production stoppage is not fun and
it gets really nasty when it comes to politics at someone else on
the list says BTDTGTS . IF you let the tool run loose heaven help
you if the tool expires at xx AM and trying to get a hold of the
vendor that does 9-5 in another time zone or even worse in on the
other side of the pond (Atlantic or Pacific). NO THANKS and another BTDTGS.
Ed
Hi Ed,
Perhaps I've been reading things too hurriedly. Maybe I missed the
thrust of Rick's comment.
What you are concerned about is the possibility of letting a
distribution copy of a product get out the door with a debugging
interface still activated. That, of course, can lead to unhappy
situations at customer sites should the debugging interface get
executed. I agree. That sort of thing must never be allowed to happen.
But to my mind, avoiding such a situation is very easy: Just make the
debugging interface "fail-safe". Code it so as to require some
additional action of environmental characteristic without which the
interface simply does nothing, NOPs as if it did not exist. And there
are any number of ways to do that:
One way is to code a closed permanent branch around the interface
activation code. Then a manual zap by the developer would be
required, without which the code could never be executed.
Another might be to require the presence of a secret keyword ddname,
example //DEBUGME DD DUMMY. Then a simple TIOT scan would be all that
was needed for the debugging interface to know whether it should
allow debugging or just step aside.
Another might be to check the environment for your own computer's
local SYSPLEX name, SMF name, CPU id/serial#, TSO userid, RACF
ownerid, ... whatever. Absent the right value, the debugging
interface would not permit debugging.
I really don't see that there is a serious problem here. (Or am I
still missing the point?)
Dave Cole REPLY TO: [EMAIL PROTECTED]
Cole Software WEB PAGE: http://www.colesoft.com
736 Fox Hollow Road VOICE: 540-456-8536
Afton, VA 22920 FAX: 540-456-6658
----------------------------------------------------------------------
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
