John Kalinich writes:
My understand of this in the OS/360 era was that if a module had the
"refreshable" attribute it could be replaced on the fly by the machine
check handler (MCH) in order to recover from memory errors.
and this is exactly right. In those days the official LE definition of a
reentrant module envisaged that it could perhaps modify itself (typically
once and typically early) iff it held "a global lock" while doing so, which
was/is more restrictive than "with serialization".
A refreshable module, which could not do so, was thus reentrant; but NOT
vice versa.
Everyone appears to have his own notion of what reentrant means, and in
consequence other terms, e.g., the CICS-jargon phrase "thread-safe", are
being introduced to disambiguate it. It may be, however, that "reentrant"
itself can be rehabilitated.
John Gilmore
Ashland, MA 01721-1817
USA
_________________________________________________________________
Try the next generation of search with Windows Live Search today!
http://imagine-windowslive.com/minisites/searchlaunch/?locale=en-us&source=hmtagline
----------------------------------------------------------------------
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
