jzmaddock

Need to add last-mod-time.cpp to this list as well (msvc-12.0).

>astar_maze works for me (osx) Works for me on Ubuntu - it's just msvc where it's crashing. Here's what I've found: * MSVC has an "uninitialised variable guard" in debug...

>last-mod-time has problems with the stat call on osx (non fatal, but incorrect results). This would be easiest to fix by using the c++17 filesystem header, is that permissible? Ooops...

OK, there is a bug somewhere, but it's masked by some very clever named-return-value optimisations with gcc and clang. Here's what I have so far: * During: ``` maze random_maze(std::size_t...

Making `filtered_graph_base` non-copyable breaks a lot of code - there are a lot of factory functions that create filtered_graph's. filter_graph is actually documented to store a reference and mentions that...

Pushed fixes for astar_maze.cpp: https://github.com/boostorg/graph/commit/32bc0e1dff9554e179f54cb287af9a5541359a25

This looks like a low-risk merge to me - anyone else care to comment?

@jeremy-murphy, @Belcourt, any thoughts on this one?

>Quick question: Given a floating point type T, how to I form the associated complex type in the multiprecision case? To wit, std::complex will fail for float128, cpp_bin_float_100, etc. Ugh,...

> #include We really shouldn't need that, and it might get rid of the error? And of course I should have said `decltype` since `typeof` is rather last century (like...