On Tue, Dec 23, 2014 at 10:14 AM, L. David Baron <dba...@dbaron.org> wrote:
> On Tuesday 2014-12-23 09:59 -0800, Martin Thomson wrote: > > On Tue, Dec 23, 2014 at 9:55 AM, L. David Baron <dba...@dbaron.org> > wrote: > > > > > > But that's an implicit constructor that's causing extra refcount > > > traffic, which is one of the things we didn't want here. > > > > > > > > I don't think that it's quite fair to complain about implicit conversion > if > > the whole point is to enable implicit conversion of another sort. And > > maybe something could be done about the refcount traffic. > > The problem is that something unexpected is happening. People > should expect the implicit conversions that are part of the > language. They generally don't expect implicit creation of smart > pointers to cause reference counting traffic when they make a > function call. > > Implicit conversions can be useful when they're carefully designed. > But C++ makes it too easy to create them by accident (by making > 1-parameter constructors do implicit conversion by default). And > this one doesn't look carefully designed, given that it does things > that users don't expect, and it's not clear how to fix that. Just to be clear, is your problem the implicit conversion itself or the reference count increment/decrement? On the specific topic of what users expect, what you expect may not be what I expect, but I would observe that both C++11 shared_ptr and Boost intrusive_ptr (the analog to our RefPtrs> automatically convert from SmartPtr<Derived> to SmartPtr<Base> and appear to increment the reference count in the same way that RefPtr does; this seems to suggest that at least some people do expect it to behave this way. -Ekr _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
