On Tue, Dec 22, 2015 at 04:54:06PM +0100, Igor Mammedov wrote:
> On Tue, 22 Dec 2015 17:28:42 +0200
> "Michael S. Tsirkin" <m...@redhat.com> wrote:
>
> > On Thu, Dec 10, 2015 at 12:40:55AM +0100, Igor Mammedov wrote:
> > > Signed-off-by: Igor Mammedov <imamm...@redhat.com>
> > > ---
> > > tests/bios-tables-test.c | 7 +++++++
> > > 1 file changed, 7 insertions(+)
> > >
> > > diff --git a/tests/bios-tables-test.c b/tests/bios-tables-test.c
> > > index 6d37332..50678b5 100644
> > > --- a/tests/bios-tables-test.c
> > > +++ b/tests/bios-tables-test.c
> > > @@ -580,6 +580,13 @@ static void test_acpi_asl(test_data *data)
> > > (gchar *)&signature,
> > > sdt->asl_file, sdt->aml_file,
> > > exp_sdt->asl_file, exp_sdt->aml_file);
> > > + if (getenv("V")) {
> >
> > How do you test? make check V=1?
> yep
> make V=1 check
>
> >
> > > + int ret G_GNUC_UNUSED;
> > > + char *diff = g_strdup_printf("git diff --exit-code
> > > %s %s",
> > > + exp_sdt->asl_file, sdt->asl_file);
> >
> > This assumes running under git which is generally wrong e.g.
> > people build from tarballs.
> > git is also not a build dependency.
> usually only developers use V=1 as they care about verbose output
> and they probably use/have git so it doesn't have to builddep,
> if they don't they won't get any useful output here.
>
> my choice of 'git diff' was dictated by the fact that
> it produces nice colored diff of ASL while usual diff doesn't.
> And that helps a lot (to me) as that diff is intended
> for human consumption.
Well for one, I want make check to pass before
I commit a new expected, not afterwards.
Add another variable to specify which diff to use,
install colordiff and use that.
> >
> >
> > > + ret = system(diff) ;
> > > + g_free(diff);
> > > + }
> > > }
> > > }
> > > g_string_free(asl, true);
> > > --
> > > 1.8.3.1
> > >
