trentforkert

Works for me using latest GDC binary release and Makefile generator. If you're using (your fork of) Ninja, d1715a83995190989b921070a50e510d069fd12c should fix it.

AFAICT, Visual D handles dependency tracking on its own. It might be nice for CMake to also know about the dependency info, but if VS/VD is handling it, I think...

I just confirmed that Ninja's .ninja_deps is being populated with D things. Ninja: mathstuf/ninja@763c6c65d501e03b6efa9577406ed66b6ce4341a CMake: 6733f3dbf72a9cba73a91957b80f38ab83e9f552 The dep files (and all other features I tested) work across several versions of...

This sort of approach might be for the best. One problem I encountered when doing Makefile generator support is that DMD (and the other D compilers) do full dependency analysis....

Thinking aloud again: If I were to implement a lexer for D we would lose: - Imports that show up via mixin() - Imports that show up via templates -...

Further thinking aloud: I've now written a simple test that uses PhilippeSigaud's Pegged (with a local fix for 64-bit) to produce a minimal D parser that only handles a single...

Fresh commit (ac1b5392372075231655f9769f84e828a793bfc9) implements DDeps, a dependency resolution parser for D, and the necessary code to have the Makefile and Ninja based generators use it. Things it handles: - Version/debug/unittest...

I just created a new default branch d_support3, which is freshly rebased on master. I've got it successfully building and generating projects on Linux, and a successful cross-compile targeting Windows...

> Not sure how easy it'd be for you, but could you get [Derelict](https://github.com/DerelictOrg) any my [bulletml](https://github.com/mathstuf/bulletml) library built and try [mu-cade](https://github.com/mathstuf/abagames-mu-cade)? I'm getting this build rule: I'll see about...

Hmm. It builds fine for me. No ddeps warnings, ddeps' files look fine, as do Ninja's. No extra/wrong/missing includes anywhere I can see.