http://llvm.org/bugs/show_bug.cgi?id=20336

Hans Wennborg <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #9 from Hans Wennborg <[email protected]> ---
(In reply to comment #8)
> (In reply to comment #7)
> > (In reply to comment #6)
> > > Alternatively, could --show-includes use errs() instead of outs(), if it's
> > > something that doesn't disable the rest of compilation? This is how
> > > everything else I'm aware of in clang and other llvm tools work, which is
> > > why the raw_ostream code in question has been able to last almost 4 years 
> > > in
> > > the tree by now.
> > 
> > Unfortunately not. MSVC prints --show-includes to stdout, and build tools
> > rely on that, so clang-cl should do the same.
> 
> Ok. Does MSVC support "-o -" or something like it? If not, perhaps you could
> just issue an error if the user requests "-o -" and --show-includes at the
> same time.

I think it's really just /E (preprocess to stdout) that results in this
situation. I've committed r213589 which makes us ignore /showIncludes (with the
usual warning) when combined with /E or /EP.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
LLVMbugs mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs

Reply via email to