From: stt@houdini.camb.inmet.com (Tucker Taft)
Subject: Re: Finalization question
Date: 1997/09/23
Date: 1997-09-23T00:00:00+00:00 [thread overview]
Message-ID: <EGyznK.11I.0.-s@inmet.camb.inmet.com> (raw)
In-Reply-To: 3426E364.6036@bix.com
Tom Moran (tmoran@bix.com) wrote:
: 2 of 3 tested compilers do not call Finalize if Initialize did not
: complete normally. That's the way I'm inclined to read RM 7.6(4), but
: the wording could be interpreted otherwise. What's correct?
Finalize should only be called if Initialize (or an initial Adjust)
completes normally.
However, when there are multiple controlled subcomponents of a single
object, there is a running debate about how precisely should this rule
be enforced. If not enforced precisely, is it better to have extra
unAdjusted Finalizes, or extra unFinalized Adjusts?
Ideally, only those objects (including subcomponents) that successfully
Initialize/Adjust are subjected to finalization.
--
-Tucker Taft stt@inmet.com http://www.inmet.com/~stt/
Intermetrics, Inc. Burlington, MA USA
prev parent reply other threads:[~1997-09-23 0:00 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
1997-09-22 0:00 Finalization question Tom Moran
1997-09-23 0:00 ` Tucker Taft [this message]
replies disabled
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox