[PATCH] clang-cl: implement /fallback mode

[Hans Wennborg]

================
Comment at: lib/Driver/Tools.cpp:6721
@@ +6720,3 @@
+
+  // FIXME: If we've put clang-cl as cl.exe on the path, then we might have a problem now :/
+  const char *Exec =
----------------
Reid Kleckner wrote:
> Hans Wennborg wrote:
> > Reid Kleckner wrote:
> > > That seems problematic.  You could use sys::fs::equivalent() to skip a path if it's the current executable, but we'd have to sink that into GetProgramPath().
> > On the good side, we're not passing /fallback through to the second invocation, so we don't get stuck forever :)
> > 
> > On the bad side, I don't think checking sys::fs::equivalent() would work, because the cl.exe on the PATH could be a copy of clang-cl.exe, as in the case of the current VS integration :/
> > 
> > It seems that automatically looking for the "true" cl.exe could get pretty messy.
> > 
> > What do you think about having an option, like /fallbackcommand <foo>? Ideally, the VS integration could set this automagically, and it would also be useful for testing (we could set this to "echo" or something and test the actual execution of the fallback).
> I agree it could get messy, but I think there's an important common case, which is where the currently running compiler is the cl.exe on PATH, and the next cl.exe on PATH is MSVC.  That's the set up we use for VS project files, so it's pretty important for users.  I'm fine handling that as a followup, though.
Ah, yes. I was thinking we'd need to have a general "is cl.exe really clang-cl.exe" check, but that's not what we need for this case. I'll address this in a follow-up.


http://llvm-reviews.chandlerc.com/D1711