On Mon, Feb 24, 2014 at 06:51:44PM +0100, Joerg Sonnenberger wrote:
> On Fri, Feb 21, 2014 at 03:17:18AM -0800, Renato Golin wrote:
> > This will create the notion that the IAS *will* validate the assembly
> > and be picky no matter what, which will push people back to using GNU
> > assembly for most cases. We want the exact opposite,
>
> (a) I think IAS should be doing this validation.
> (b) I don't think we are ready for that yet.
>
> I'm working on a nice standalone tool for addressing (b), but for now I
> think we shoul go back to not caring about the content of inline
> assembler for -S mode.
For this purpose, I think something like the attached change is minimal
intrusive.
Joerg
Index: lib/Driver/Tools.cpp
===================================================================
--- lib/Driver/Tools.cpp (revision 202049)
+++ lib/Driver/Tools.cpp (working copy)
@@ -2522,7 +2522,8 @@
CmdArgs.push_back("-masm-verbose");
if (!Args.hasFlag(options::OPT_fintegrated_as, options::OPT_fno_integrated_as,
- IsIntegratedAssemblerDefault))
+ IsIntegratedAssemblerDefault) ||
+ C.getArgs().hasArg(options::OPT_S))
CmdArgs.push_back("-no-integrated-as");
if (Args.hasArg(options::OPT_fdebug_pass_structure)) {
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
