From: Bill Klein (bill_at_[hidden])
Date: 1999-11-09 22:27:18
>You're assuming that comparing the pointed-to object identity
>is a less-likely or useful scenario than comparing the pointed-
>to object contents.
Nope, I was just suggesting that they might both
be useful, and hence letting us define the operator's
meaning might be useful.
I can see a lot of people feel strongly that == should
always be comparing pointers for shared_ptr... Like I
said before, I understand the desire for consistency in
the pointer-like semantics. I wouldn't be opposed to
doing as Darin Alder originally suggested and adding it
in like that.
-Bill Klein <bill_at_[hidden]>
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk