Re: [PATCH v2 3/4] Add 'promisor-remote' capability to protocol v2

>
> I was looking at test breakages caused by this topic (in 'seen',
> t5710 fails leak checking).
>
> Then I noticed something strange.  Next to the "$(TRASH_DIRECTORY)",
> running this script leaves a few garbage files under the "t/"
> directory.
>
> I think the culprit is this helper function.
>
> > +initialize_server () {
> > +     # Repack everything first
> > +     git -C server -c repack.writebitmaps=false repack -a -d &&
> > +
> > +     # Remove promisor file in case they exist, useful when reinitializing
> > +     rm -rf server/objects/pack/*.promisor &&
> > +
> > +     # Repack without the largest object and create a promisor pack on server
> > +     git -C server -c repack.writebitmaps=false repack -a -d \
> > +         --filter=blob:limit=5k --filter-to="$(pwd)" &&
>
> This --filter-to="$(pwd)" expands to $(TRASH_DIRECTORY), which is
> "..../t/trash-directory.t5710-promisor-remote-capability".  I think
> that is the cause for two extra trash files that are _OUTSIDE_ the
> trash directory, which is an absolute no-no for tests to be safely
> runnable.  Next to the trash directory, this ends up creating three
> files
>
> trash directory.t5710-...-980d3ff591aae1651cdd52f7dfad4fb6319ee3c2.idx
> trash directory.t5710-...-980d3ff591aae1651cdd52f7dfad4fb6319ee3c2.pack
> trash directory.t5710-...-980d3ff591aae1651cdd52f7dfad4fb6319ee3c2.rev

> The first failure in leak check seems to be
>
> not ok 5 - fetch with promisor.advertise set to 'false'
> #
> #               git -C server config promisor.advertise false &&
> #
> #               # Clone from server to create a client
> #               GIT_NO_LAZY_FETCH=0 git clone -c remote.server2.promisor=true \
> #                       -c remote.server2.fetch="+refs/heads/*:refs/remotes/server2/*" \
> #                       -c remote.server2.url="file://$(pwd)/server2" \
> #                       -c promisor.acceptfromserver=All \
> #                       --no-local --filter="blob:limit=5k" server client &&
> #               test_when_finished "rm -rf client" &&
> #
> #               # Check that the largest object is not missing on the server
> #               check_missing_objects server 0 "" &&
> #
> #               # Reinitialize server so that the largest object is missing again
> #               initialize_server
>
> but I didn't dig further.  Can you take a look?  I'll eject the
> topic from 'seen' in the meantime to unblock the CI.