r315126 - Driver: hoist the `wchar_t` handling to the driver

[Friedman, Eli via cfe-commits]

On 10/6/2017 4:09 PM, Saleem Abdulrasool via cfe-commits wrote:
> Modified: cfe/trunk/lib/Driver/ToolChains/Clang.cpp
> URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Driver/ToolChains/Clang.cpp?rev=315126&r1=315125&r2=315126&view=diff
> ==============================================================================
> --- cfe/trunk/lib/Driver/ToolChains/Clang.cpp (original)
> +++ cfe/trunk/lib/Driver/ToolChains/Clang.cpp Fri Oct  6 16:09:55 2017
> @@ -2601,6 +2601,33 @@ static void RenderModulesOptions(Compila
>     Args.AddLastArg(CmdArgs, options::OPT_fmodules_disable_diagnostic_validation);
>   }
>   
> +static void RenderCharacterOptions(const ArgList &Args, const llvm::Triple &T,
> +                                   ArgStringList &CmdArgs) {
> +  // -fsigned-char is default.
> +  if (const Arg *A = Args.getLastArg(options::OPT_fsigned_char,
> +                                     options::OPT_fno_signed_char,
> +                                     options::OPT_funsigned_char,
> +                                     options::OPT_fno_unsigned_char)) {
> +    if (A->getOption().matches(options::OPT_funsigned_char) ||
> +        A->getOption().matches(options::OPT_fno_signed_char)) {
> +      CmdArgs.push_back("-fno-signed-char");
> +    }
> +  } else if (!isSignedCharDefault(T)) {
> +    CmdArgs.push_back("-fno-signed-char");
> +  }
> +
> +  if (const Arg *A = Args.getLastArg(options::OPT_fshort_wchar,
> +                                     options::OPT_fno_short_wchar)) {
> +    if (A->getOption().matches(options::OPT_fshort_wchar)) {
> +      CmdArgs.push_back("-fwchar-type=short");
> +      CmdArgs.push_back("-fno-signed-wchar");
> +    } else {
> +      CmdArgs.push_back("-fwchar-type=int");
> +      CmdArgs.push_back("-fsigned-wchar");
> +    }
> +  }
> +}

It looks like this changes the behavior of "-fno-short-wchar". 
Specifically, on targets where the default wchar_t type is "unsigned 
int" (like ARM AAPCS), "-fno-short-wchar" used to be a no-op, but now it 
changes the sign of wchar_t.  Why are we overriding the default here?

-Eli

-- 
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project