
Niall Douglas <s_sourceforge <at> nedprod.com> writes:
It's very simple: If you weren't in GSoC this summer, I'd assume you'd spend your summer working on your TMP library anyway. So simply list what you hope to finish this summer, and propose the first half of that for GSoC as concrete items and the second half as optional "would like if have time" items. Post that here when you're done.
First, thank you for taking your time to help me. Here is the list of MPL11-related tasks I plan to clear during the summer. Note that full unit testing and documentation is implied for each task. - Write a suite of benchmarks to compare the compilation time and memory usage of different TMP techniques. The benchmarks should run automatically on different input datasets and produce charts for easy comparison. This can be done with e.g. Ruby and a Ruby binding for Gnuplot. - Implement a compile-time map and a compile-time set with variadic templates. Several implementation techniques should be tested and benchmarked to pick the best one. For the interface, I'm thinking about an something similar to Haskell's Data.Map.Strict for the map and to Data.Set for the set. Note that I will have to decide what kind of type equality to support: type identity or "deep" type equivalence? - Write numeric metafunctions that can perform efficiently on homogeneous sequences of integral constants (using constexpr). - Write a tutorial that does not only present the core concepts of the library, but really shows how to use it for concrete tasks. - Setup a Boost-compatible build system; the library currently uses CMake. - Design a cross-tag dispatching system for binary metafunctions (probably following N3351). This is a bit like the mixed-type numeric metafunctions of the MPL, except it should work with arbitrary (non-numeric) types and metafunctions. - Design a small DSL to write inline metafunction classes. This is pretty much like lambda expressions in the MPL, but: - It should be possible to refer to arguments in an outer scope. This issue can be addressed either by let-expressions or De Bruijn indices. - It should be possible and easy to define custom placeholders. - It should be possible to have variadic placeholders, i.e. placeholders that are replaced by more than one argument. Note that this is tricky because the template in which the placeholder appears might not have the arity required to perform the substitution. These are optional tasks I would like to complete if I have the time. Again, each task implies full unit testing and documentation. - Make std::tuple, MPL sequences and Fusion sequences valid MPL11 sequences, so they can be used in MPL11 algorithms. - Implement pattern matching on template specializations, much like Proto's match<> but for arbitrary template specializations. Pattern modifiers should be provided. At the minimum: - or_<Patterns...> matches if any of the Patterns match. - and_<Patterns...> matches if all of the Patterns match. - not_<Pattern> matches if the Pattern does not match. - when<F> matches if apply<F, T>::type::value is true, where T is the expression being matched against. - Implement a compile-time graph data structure. The graph should support an interface similar to Boost.Graph's adjacency_list. Some basic graph algorithms should also be provided. At the minimum: - depth-first search - breadth-first search - topological sort - transitive closure This is useful e.g. to order the execution of computations at compile-time. By creating a compile-time graph where nodes are computations and edges are dependencies between them, one can determine which computations can be parallelized and which ones must be linearized. Comments welcome! Regards, Louis