[Intel-gfx] [PATCH i-g-t 2/5] lib/tests: don't use hard error status to indicate test failure

[Daniel Vetter]

On Mon, Nov 03, 2014 at 03:42:06PM +0000, Thomas Wood wrote:
> On 3 November 2014 15:02, Daniel Vetter <daniel at ffwll.ch> wrote:
> > On Mon, Nov 03, 2014 at 11:31:02AM +0000, Thomas Wood wrote:
> >> Signed-off-by: Thomas Wood <thomas.wood at intel.com>
> >> ---
> >>  lib/tests/igt_command_line.sh | 8 ++++----
> >>  1 file changed, 4 insertions(+), 4 deletions(-)
> >>
> >> diff --git a/lib/tests/igt_command_line.sh b/lib/tests/igt_command_line.sh
> >> index 7e6ca67..5cf2584 100755
> >> --- a/lib/tests/igt_command_line.sh
> >> +++ b/lib/tests/igt_command_line.sh
> >> @@ -48,20 +48,20 @@ for test in $TESTLIST; do
> >>
> >>       # check invalid option handling
> >>       echo "  Checking invalid option handling..."
> >> -     ./$test --invalid-option 2> /dev/null && exit 99
> >> +     ./$test --invalid-option 2> /dev/null && exit 1
> >
> > Just a curious question: What's better with hardcoding 1 than hardcoding
> > 99? Otherwise series lgtm.
> 
> From the automake manual:
> 
> "When no test protocol is in use, an exit status of 0 from a test
> script will denote a success, an exit status of 77 a skipped test, an
> exit status of 99 an hard error, and any other exit status will denote
> a failure."
> 
> So, an exit status of 99 is reported separately in the summary as an
> error, rather than as a test failure.

Ah, makes sense. Can you please add this bit of important information to
the commit message before pushing?

Thanks, Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch