Orthogonal cleanups such as -ccc-clang-archs removal should be in a
separate patch generally speaking. I'll go ahead and apply that part to
save a round trip.

The test cases though look much like the original test cases in the patch I
cleaned up for you. I had to edit those test cases significantly to get
them to pass on Linux. Could you try to write the assertions in a manner
more similar to the ones I ended up checking in? Specifically there is no
reason to check every argument to the linker when you're really trying to
isolate and test a single feature. The way you've written in the tests are
simply too brittle.


On Tue, Oct 16, 2012 at 7:37 PM, David Hill <[email protected]> wrote:

> How is this?
>
> On Tue, Oct 16, 2012 at 05:24:59PM -0700, Eli Friedman wrote:
> > On Mon, Oct 15, 2012 at 9:00 AM, David Hill <[email protected]> wrote:
> > > Ping.  Can someone commit this?
> >
> > +    if (Args.hasArg(options::OPT_pthread)) {
> > +      if (!Args.hasArg(options::OPT_shared) &&
> > +          Args.hasArg(options::OPT_pg))
> >
> > I'd like to see a driver test where the check for -shared matters.
> >
> > -Eli
>
> _______________________________________________
> cfe-commits mailing list
> [email protected]
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
>
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to