Add generic addNode and addEdge overloads

[sbaldu]

I was thinking that it might be nice to add overloads for the addNode and addEdge methods that take an arbitrary number of nodes or edges.
This could be done with something like:

template <typename T1, typename... Tn>
void Graph<T>::addNode(T1 node, Tn... nodes) {

and in addition it would be nice to define a type trait that makes sure that the template parameter is in fact a node/edge, like:

template <typename T>
struct is_node : std::false_type {};

template <typename T>
struct is_node<const Node<T>*> : std::true_type {};

so that the complete function declaration becomes:

template <typename T1, typename... Tn>
requires is_node<T1>::value && (is_node<Tn>::value && ...)
void Graph<T>::addNode(T1 node, Tn... nodes) {

This might be an overkill but I think that it could be a useful addition;
What do you think @ZigRazor @nrkramer?

P.S. the use of requires needs C++20, but considering issue #266 I suppose that there was already the intention of updating the standard version.

[nrkramer]

I don't think forcing users to upgrade their version of the C++ standard is necessarily a good thing.

Supporting new features from the standard is fine as long as it's hidden in the library - but if it's in the headers it becomes a far more troublesome problem.

Having the maximum amount of backwards compatibility is a good thing. Most library implementers should strive for that.

This being said, the template variadic list is a neat idea.

[nrkramer]

I think we should discuss and consider the possibility of distributing a standard hpp/cpp library, as the drawbacks of header-only are starting to manifest in this library.

[nrkramer]

This SO article goes over some of the basic points: https://softwareengineering.stackexchange.com/questions/305618/are-header-only-libraries-more-efficient

I do think the low administrative costs of a header-only library make it attractive to keep.

[ZigRazor]

I think we should discuss and consider the possibility of distributing a standard hpp/cpp library, as the drawbacks of header-only are starting to manifest in this library.

I think we can open a discussion in discussion section.

[ZigRazor]

I do think the low administrative costs of a header-only library make it attractive to keep.

I Agree with @nrkramer

[sbaldu]

This being said, the template variadic list is a neat idea.

If you think that it would be a useful addition we could still implement the new method without using the C++20 requires.
I think that we can implement the check on the template parameters by simply using std::enable_if, which is from C++11 so we don't need to update the standard.
What do you think?

[ZigRazor]

In this way, I think is good!