Christian Couder <christian.couder@xxxxxxxxx> writes:
> While at it, let's also use git_config_get_string_tmp() instead of
> git_config_get_string() to simplify memory management.
> ...
> - strvec_push(names, r->name);
> - strvec_push(urls, git_config_get_string(url_key, &url) ? NULL : url);
> + /* Only add remotes with a non empty URL */
> + if (!git_config_get_string_tmp(url_key, &url) && *url) {
> + strvec_push(names, r->name);
> + strvec_push(urls, url);
> + }
>
> - free(url);
Nice.
> +test_expect_success "clone with 'KnownName' and missing URL in the config" '
> + git -C server config promisor.advertise true &&
> +
> + # Clone from server to create a client
> + # Lazy fetching by the client from the LOP will fail because of the
> + # missing URL in the client config, so the server will have to lazy
> + # fetch from the LOP.
> + GIT_NO_LAZY_FETCH=0 git clone -c remote.lop.promisor=true \
> + -c promisor.acceptfromserver=KnownName \
> + --no-local --filter="blob:limit=5k" server client &&
> + test_when_finished "rm -rf client" &&
These are the other way around. When 'clone' fails, test_when_finished
is not run, so nobody arranges the new directory 'client' to be removed.
"git clone" does try to remove in such a case, but we are protecting
against a failing "clone", so swapping them around, i.e. arrange to
remove it and then try to create it, would make more sense.
> +test_expect_success "clone with 'KnownUrl' and url not configured on the server" '
> + git -C server config promisor.advertise true &&
> +
> + git -C server config unset remote.lop.url &&
> + test_when_finished "git -C server config set remote.lop.url \"file://$(pwd)/lop\"" &&
Probably the same principle applies here, but the case where "git
config" fails, it is likely that the file is not touched at all, or
it gets so corrupt beyond salvaging with another "config set", so it
matters much less than the previous one.
> + # Clone from server to create a client
> + # It should fail because the client will reject the LOP as URLs are
> + # different, and the server cannot lazy fetch as the LOP URL is
> + # missing, so the remote name will be used instead which will fail.
> + test_must_fail env GIT_NO_LAZY_FETCH=0 git clone -c remote.lop.promisor=true \
> + -c remote.lop.fetch="+refs/heads/*:refs/remotes/lop/*" \
> + -c remote.lop.url="file://$(pwd)/lop" \
> + -c promisor.acceptfromserver=KnownUrl \
> + --no-local --filter="blob:limit=5k" server client &&
> +
> + # Check that the largest object is still missing on the server
> + check_missing_objects server 1 "$oid"
> +'
> +
> +test_expect_success "clone with 'KnownUrl' and empty url, so not advertised" '
> + git -C server config promisor.advertise true &&
> +
> + git -C server config set remote.lop.url "" &&
> + test_when_finished "git -C server config set remote.lop.url \"file://$(pwd)/lop\"" &&
> +
> + # Clone from server to create a client
> + # It should fail because the client will reject the LOP as an empty URL is
> + # not advertised, and the server cannot lazy fetch as the LOP URL is empty,
> + # so the remote name will be used instead which will fail.
> + test_must_fail env GIT_NO_LAZY_FETCH=0 git clone -c remote.lop.promisor=true \
> + -c remote.lop.fetch="+refs/heads/*:refs/remotes/lop/*" \
> + -c remote.lop.url="file://$(pwd)/lop" \
> + -c promisor.acceptfromserver=KnownUrl \
> + --no-local --filter="blob:limit=5k" server client &&
> +
> + # Check that the largest object is still missing on the server
> + check_missing_objects server 1 "$oid"
> +'
The test clone is identical to the previous one except for the
four-line comment in the middle. The set-up on the other side is
different (the server has remote.lop.url set in the previous one to
empty, and unset in this one, which should amount to the same
thing). Makes sense.
> test_expect_success "clone with promisor.advertise set to 'true' but don't delete the client" '
> git -C server config promisor.advertise true &&
