clang is only a fast, memory efficient C/C++ frontend.  To become a full compiler clang requires the llvm backend (a distinct and - on my Ubuntu box - separately installed set of shared libraries).

gcc has no equivalent to clang's externally exposed AST access and navigation API.  Therefore supporting clang as a semantic source does not constitute encouraging use of clang over gcc.

It saddens me that FSF doctrine would deny the emacs community a big CEDET step forward in functionality and performance.  Especially when at least some competing IDEs are exploiting clang to provide features and performance that currently CEDET cannot match.

All projects on which I work (both professionally and in my free time) use the gnu toolchain.  I have clang installed only to be able to use rtags.  Does that make me an imperfect supporter of the FSF's goals and agenda?


On Thu, Nov 7, 2013 at 11:04 AM, David Engster <> wrote:
John Yates writes:
> As you can imagine I would be delighted if semantic could become a client of
> the rtags server.

Note that code that encourages usage of clang over gcc cannot be merged
to Emacs proper (which is why I am not inclined to work on that).