Re: PATCH: Skip the test if the cross binary fails to run.

["H. J. Lu" <hjl at lucon dot org>]

On Mon, Sep 30, 2002 at 04:22:19PM -0700, Geoff Keating wrote:
> > Date: Mon, 30 Sep 2002 15:30:04 -0700
> > From: "H. J. Lu" <hjl@lucon.org>
> 
> > On Mon, Sep 30, 2002 at 03:22:55PM -0700, Geoff Keating wrote:
> > > 
> > > "H. J. Lu" <hjl@lucon.org> writes:
> > > 
> > > > If you have a cross compiler in PATH, but no target board, "make check"
> > > > will fail in binutils when it tries to run the cross binary on the build
> > > > machine. This patch skips the test.
> > > 
> > > Please don't do that, the test is important.  At least mark the test
> > > as ERROR rather than skipping it altogether.
> > > 
> > 
> > I must have missed something. Can you tell me how you are going to run
> > those tests if there is no hardware nor simulator to run on?
> 
> To run the test, you need hardware or simulator.  If you don't have
> hardware or simulator, the test can't be run.  If the test is not run,
> then objcopy hasn't been fully tested, and it's important to know
> that.
> 
> Why do you think that ERROR is wrong?  Doesn't it accurately describe
> what happened to this test?

Because it depends on if I have the cross compiler installed or not. Why should
it become an ERROR just because I have the cross compiler?


H.J.