On Wed, Sep 28, 2022 at 09:32:57AM +0200, Thomas Huth wrote: > On 27/09/2022 23.26, Michael S. Tsirkin wrote: > > On Tue, Sep 06, 2022 at 02:10:56PM +0100, Daniel P. Berrangé wrote: > > > On Tue, Jul 12, 2022 at 12:22:10PM +0530, Ani Sinha wrote: > > > > > > > > > > > > On Mon, 11 Jul 2022, John Snow wrote: > > > > > > > > > On Sun, Jul 10, 2022 at 1:01 PM Ani Sinha <a...@anisinha.ca> wrote: > > > > > > > > > > > > Integrate the pytest framework with the meson build system. This > > > > > > will make meson > > > > > > run all the pytests under the pytest directory. > > > > > > > > > > > > Signed-off-by: Ani Sinha <a...@anisinha.ca> > > > > > > --- > > > > > > tests/Makefile.include | 4 +++- > > > > > > tests/meson.build | 1 + > > > > > > tests/pytest/meson.build | 49 > > > > > > ++++++++++++++++++++++++++++++++++++++++ > > > > > > 3 files changed, 53 insertions(+), 1 deletion(-) > > > > > > create mode 100644 tests/pytest/meson.build > > > > > > > > > > > > diff --git a/tests/Makefile.include b/tests/Makefile.include > > > > > > index 3accb83b13..40755a6bd1 100644 > > > > > > --- a/tests/Makefile.include > > > > > > +++ b/tests/Makefile.include > > > > > > @@ -3,12 +3,14 @@ > > > > > > .PHONY: check-help > > > > > > check-help: > > > > > > @echo "Regression testing targets:" > > > > > > - @echo " $(MAKE) check Run block, > > > > > > qapi-schema, unit, softfloat, qtest and decodetree tests" > > > > > > + @echo " $(MAKE) check Run block, > > > > > > qapi-schema, unit, softfloat, qtest, pytest and decodetree tests" > > > > > > > > > > Does this mean that "make check" *requires* an internet connection? > > > > > > > > No. My test will be skipped if it is unable to download the artifacts it > > > > requires due to lack of Internet connectivity. > > > > > > That's not the only concern, there are also people who have metered > > > internet connections, or whose connections are slow and thus have > > > long download times. Any test that downloads should be opt-in only. > > > > > > > > > With regards, > > > Daniel > > > > <rant> > > This is why I wanted git submodules. A well understood decentralized > > model. Now we are reinventing them badly. > > I don't see much of a difference here with submodules with regards to > Daniel's concerns: If you clone the QEMU repo with a good internet > connection, you also don't get the submodules by default. If you then move > to a bad internet connection or offline and the build system tries to pull > in the submodule, you've also lost. > OTOH, with downloads, you can also make sure to have all downloads cached > while you're still on the good internet connection. The cache should > continue working if you're then moving offline.
submodules as a cache include checksumming, a way to make local changes without too much pain, documentation .... > > I asked on the maintainers summit what issues people have with > > submodules, no one volunteered any information. > > You know my position already > (https://lore.kernel.org/qemu-devel/d7a7b28f-a665-2567-0fb6-e31e7ecbb...@redhat.com/), > so at least I did not feel the need to repeat that again (and we were pretty > short in time anyway). Pity, but ok, I will revive that thread. > > Thomas