On 9 November 2013 15:37, Klaim - Joël Lamotte wrote:
I think I would have have the advice for most non-standard containers, but I strongly disagree in the case of flat_set, flat_map and stable_vector. The standard containers don't provide enough basic alternatives and, at least in my experience and experiences I've read online and discussed with other C++ devs, we end-up often getting back to std::vector for almost anything.
Good, it's the right most for most situations :-)
Boost's flat_set and flat_map are preferable as default container than std::set/map in almost all cases I've been working on in the last 5 years (which include time I couldn't use these containers in practice).
So you almost never need iterator and reference stability? Because if you don't need them, then yes, a sorted std::vector might be a better choice, or another container such as boost::flat_set, which is great and readily available from Boost - maybe you've heard of it ;-)
My understanding of the standard is that it should provide useful and widely used tools, like containers, which are then standard because everybody should use them by default.
According to Stepanov (and IIRC Stroustrup) everyone should use vector by default :-)
It looks like I'm not the only one praising these kind of containers: http://www.slashslash.info/2013/10/ode-to-a-flat-set/
I'm not saying they're not useful. I specifically said they're useful, you even quoted me saying that. Stop trying to convince me they're useful. I'm objecting to the (IMHO too common) view that everything useful should be put in the standard. There are things missing from the standard (networking, databases, XML/XSLT processing, graphics) that would be better for the committee and implementors to spend time on than different containers with small-ish variations on the interfaces and properties of the existing standard containers. The flat_map and flat_set containers exist and are perfectly usable for those who want them.