In response to the mailing list message, it is the Free Software Foundation Steering Committee that needs to give the thumbs-up for this process to commence once Digital Mars is okay with the FSF copyright assignment and licensing policies.
The D programming language was conceived by Walter Bright himself at Digital Mars more than a decade ago and is slowly beginning to gain some traction. The design of the D programming language is largely influenced by C++. It was just recently with Fedora 14 that Red Hat pushed in a D compiler, but rather than using GDC, they used LDC. LDC is the LLVM D Compiler. The LLVM D Compiler targets version 1.0 of the D language (there's two versions) and is based upon the DMD front-end, but of course with the LLVM back-end, just as Clang is to C/C++ with LLVM. There's also Digital Mars' multi-platform DMD compiler for Windows / Linux / Mac OS X / BSD and a D.NET compiler too.
Pushing D language support into GCC via GDC will increase the language's presence and provide an alternative to those that may not want to use the LLVM-based D compiler. [It will also allow us to run more benchmarks in the future like we did for C/C++ with GCC vs. LLVM Clang, DragonEgg, LLVM-GCC performance, but for this alternative language.] Merging this D support is too late for GCC 4.6, so it would make it a GCC 4.7 target if everything pans out with the Free Software Foundation and Digital Mars. It was also just recently that Google's Go got the go-ahead to be pushed into GCC.