On Tue, Aug 18, 2020 at 11:03:27AM +0200, Marc Hartmayer wrote: > On Fri, Aug 14, 2020 at 03:29 PM +0200, Andrew Jones <drjones@xxxxxxxxxx> wrote: > > On Fri, Aug 14, 2020 at 03:06:36PM +0200, Marc Hartmayer wrote: > >> On Thu, Aug 13, 2020 at 10:30 AM +0200, Andrew Jones <drjones@xxxxxxxxxx> wrote: > >> > On Wed, Aug 12, 2020 at 11:27:04AM +0200, Marc Hartmayer wrote: > >> >> This allows us, for example, to auto generate a new test case based on > >> >> an existing test case. > >> >> > >> >> Signed-off-by: Marc Hartmayer <mhartmay@xxxxxxxxxxxxx> > >> >> --- > >> >> run_tests.sh | 2 +- > >> >> scripts/common.bash | 13 +++++++++++++ > >> >> scripts/mkstandalone.sh | 2 +- > >> >> 3 files changed, 15 insertions(+), 2 deletions(-) > >> >> > >> >> diff --git a/run_tests.sh b/run_tests.sh > >> >> index 24aba9cc3a98..23658392c488 100755 > >> >> --- a/run_tests.sh > >> >> +++ b/run_tests.sh > >> >> @@ -160,7 +160,7 @@ trap "wait; exit 130" SIGINT > >> >> # preserve stdout so that process_test_output output can write TAP to it > >> >> exec 3>&1 > >> >> test "$tap_output" == "yes" && exec > /dev/null > >> >> - for_each_unittest $config run_task > >> >> + for_each_unittest $config run_task arch_cmd > >> > > >> > Let's just require that arch cmd hook be specified by the "$arch_cmd" > >> > variable. Then we don't need to pass it to for_each_unittest. > >> > >> Where is it then specified? > > > > Just using it that way in the source is enough. We should probably call > > it $ARCH_CMD to indicate that it's a special variable. Also, we could > > return it from a $(arch_cmd) function, which is how $(migration_cmd) and > > $(timeout_cmd) work. > > My first approach was different… > > First we source the (common) functions that could be overridden by > architecture dependent code, and then source the architecture dependent > code. But I’m not sure which approach is cleaner - if you prefer your > proposed solution with the global variables I can change it. I prefer my proposed solution. It's not necessary to provide and source an arch-neutral function that will never do anything. And, it will never do anything, because the function is supposed to be arch-specific. If an arch doesn't implement the function, then we don't need to call anything at all. Thanks, drew > > Thanks for the feedback! > > […snip] > > -- > Kind regards / Beste Grüße > Marc Hartmayer > > IBM Deutschland Research & Development GmbH > Vorsitzender des Aufsichtsrats: Gregor Pillen > Geschäftsführung: Dirk Wittkopp > Sitz der Gesellschaft: Böblingen > Registergericht: Amtsgericht Stuttgart, HRB 243294 >
