troy d.straszheim wrote:
Would seem reasonable to me to allow both and have the docs reflect that a cast is preferred, but if the consensus is to turn off the constructor-from-raw-pointer, if not now, when?
There is no consensus to turn off the raw pointer constructor. While replacements are frequently being proposed - some of them predating boost::shared_ptr - none of them have proven useful in practice; or if they have, I am not aware of them. For now I'm inclined to think of this recurring idea as something that looks good on paper but doesn't offer significant benefits. Of course I may be wrong. ;-) Another alternative is to tackle the problem at runtime. In a "safe mode" with some help from the heap manager, it's possible for the pointer constructor to detect invalid arguments (the current implementation has a proof of concept). However, interestingly enough, none of the "safe mode" STLs offer such safeguards for std::auto_ptr (to the best of my knowledge); there's probably no demand.