Replacing Boost data structures with STL

[GYuvanShankar]

Many files still use boost::tuple, even though std::tuple provides similar features. Would it be worthwhile to change it?

[mglisse]

In places where it is an implementation detail, it would make sense to change it. Some implementations of std::tuple store objects in the reverse order of boost::tuple, so there is a slight risk that it impacts memory or performance, but it doesn't seem too likely.
In places where this is part of the public interface, changing it will break users' code. It may still make sense in the long run, but I don't know what the right strategy would be to minimize the trouble for users.

[GYuvanShankar]

Thanks. Ill try changing in a few instances.

[sloriot]

In the Kernel it won't be easy as boost::tuple does not need to know the complete type of objects stored while std::tuple does. I think I already did replace all the easy one.