Subject: mutable
From: Mike Nordell (tamlin@algonet.se)
Date: Sat Jul 07 2001 - 14:15:18 CDT
Sorry for pointing out the obvious.
With the current discussion regarding this keyword, please respect its
purpose. If an object changes its _logical_ state due to changing a mutable
data member, that data member has no business being mutable.
It is however useful for e.g. caching purposes where the object logically
doesn't change state, and that is AFAIK what has been proposed (e.g. the
docPos of a frag is a cached).
/Mike
This archive was generated by hypermail 2b25 : Sat Jul 07 2001 - 14:14:28 CDT