Il gio 17 set 2020, 18:53 Alex Bennée <alex.ben...@linaro.org> ha scritto:
> > Philippe Mathieu-Daudé <f4...@amsat.org> writes: > > > On 9/17/20 12:44 PM, Paolo Bonzini wrote: > >> Avoid that containers pile up. > >> > >> Signed-off-by: Paolo Bonzini <pbonz...@redhat.com> > >> --- > >> tests/docker/Makefile.include | 1 - > >> tests/docker/docker.py | 4 ++-- > >> 2 files changed, 2 insertions(+), 3 deletions(-) > >> > >> diff --git a/tests/docker/Makefile.include > b/tests/docker/Makefile.include > >> index 3daabaa2fd..75704268ff 100644 > >> --- a/tests/docker/Makefile.include > >> +++ b/tests/docker/Makefile.include > >> @@ -243,7 +243,6 @@ docker-run: docker-qemu-src > >> $(DOCKER_SCRIPT) run \ > >> $(if $(NOUSER),,--run-as-current-user) \ > >> --security-opt seccomp=unconfined \ > >> - $(if $V,,--rm) \ > >> $(if $(DEBUG),-ti,) \ > > > > Alternatively: > > > > - $(if $V,,--rm) > > - $(if $(DEBUG),-ti,) > > + $(if $(DEBUG),-ti,--rm) > > Surely that should b: > > $(if $(DEBUG),-ti --rm,) > > I think being able to keep the container around is useful for debug, I > have no problem with changing the behaviour for V=1. > You probably mean $(if $(DEBUG),-ti,--rm) but that would not work because the patch adds --rm unconditionally in docker.py. But $(DEBUG) gives you a shell to run the test from, so I don't think skipping --rm is useful even in the $(DEBUG) case. Paolo >
