"R.S." <[EMAIL PROTECTED]>
wrote in message news:<[EMAIL PROTECTED]>...
> Ted MacNEIL wrote:
> 
<snip>
> Regarding bodies for testing: do you have any for testing new
> OS/database/CICS/others releases ? Maybe it's not obvious, but all the
> compilation work is done by computer itself (no human attendance), the
> result is the list of compilation errors. Quite short one.
> I can recompile any of my application programs and I'm quite sure that
> it will work unchanged - this is due to code management tool which
> assures that compiled version on Prod system comes from the source in
> "master" library.
<snip> 

Are you really saying that just because the same code compiles "cleanly"
with the old and new compilers, that you think this is all the testing you
need to do?

If so, I think I have a canoe to sell you that is just as was tested to be
the same thing as the Brooklyn Bridge<G>.

IBM certainly "attempts" to insure that (assuming the SAME compiler and
run-time options) that "old and new" compiled code will:
  A) yield the same results
  B) "perform comparably"

However, I don't know of many shops that ASSUME this will always be true (or
any place that IBM guarantees that this will be the case; only that they
will accept defect reports when it is not true.).  Certainly if you don't
spend a LOT of time researching "new" compiler options (for example) - much
less changes in default compiler and run-time options, I would HOPE that you
would do SOME "systems" testing to insure acceptable production performance
before migrating (en mass) all the newly compiled code into production.

Bottom-Line: 
  (to me - and YMMV) if you do a mass recompile, then you really, REALLY
need to do extensive performance AND "same result" testing - at least for a
fairly large "sample" of programs/applications.

----------------------------------------------------------------------
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

Reply via email to