I mentioned autotools as part of the build system of some of our dependencies. We abandoned it ourselves in disgust many years ago (in favor of waf; meson may be in our future).
Boost is critical for us because we aim for compatibility (and for macOS that means buildability) on systems only reaching C++98, meaning that boost fills in large parts of the then-missing std:: namespace (such as shared_ptr<>). We will move to C++11 after our 6.0 release, but boost still provides several other very useful bits of functionality.
GTK+ remains useful even as we slowly move away from it, and there remains the big four (text entry, file browsing, treeviews and menus) that will be a huge effort to reimplement. Our goal is to eventually depend only upon GDK (i.e. a thin window system wrapper), but I doubt that we will ever reach that goal.
1998 was 22 years ago. At some point we've gotta let it go. And I'm not going to argue about Boost, but it'll suffice to say that I passionately hate C++ and Boost is a big part of why. That Ardour is a C++ project has a lot to do with its sprawling complexity, these are often correlated (which is why I guessed C++ from your earlier description). Also, I would definitely recommend looking into Meson instead of waf, having used both.
Anyway, I think we have reached something resembling an understanding. Your dependencies are fairly reasonable, but there's room to and reason to cull them down a bit.
I mentioned autotools as part of the build system of some of our dependencies. We abandoned it ourselves in disgust many years ago (in favor of waf; meson may be in our future).
Boost is critical for us because we aim for compatibility (and for macOS that means buildability) on systems only reaching C++98, meaning that boost fills in large parts of the then-missing std:: namespace (such as shared_ptr<>). We will move to C++11 after our 6.0 release, but boost still provides several other very useful bits of functionality.
GTK+ remains useful even as we slowly move away from it, and there remains the big four (text entry, file browsing, treeviews and menus) that will be a huge effort to reimplement. Our goal is to eventually depend only upon GDK (i.e. a thin window system wrapper), but I doubt that we will ever reach that goal.