On Tue, Jul 30, 2013 at 4:30 AM, Martin Jambor <mjam...@suse.cz> wrote:
>> I'm voting for references. References can be seen as yet another >> software structuring tool that instantly communicate some properties >> such as you mentioned above. In addition to that it's also a hint of >> ownership, i.e. if I get an object& from somewhere I know that I better >> not even think about whether to delete it or not. >> > > well, let me stress again that we should think about this in the > context of GCC. In GCC, we are used to C semantics of the dot > operator and have a lot of existing code that we will continue to use > and mix with new code with the same assumption. Putting a reference > where none has been before might result in silent and hard to spot > erroneous modifications. I cannot make sense of this. > At the same time, we are used to lookup whether the pointer we got can > be NULL or not if necessary. Moreover, in GCC, NULL-dereferences are > not particularly dangerous and are easy to debug. That should not be reason to avoid idiomatic constructions. > In the context of GCC, there will be no ownership hints of this kind > simply because new code will co-exist with tons of old code which uses > pointers and will not give any such hints and so you can't assume you > could free its pointers. Let me not even mention GC. References are not just about ownership (though they might imply it). -- Gaby
