Previously, this test assumed that:
- a global watch would return only an event for the key in question
- only the delete event in question would be returned
Neither of these assumptions are correct for an etcd backend as long
as any other clients are interacting with the system. This commit
makes the watch more specific and extracts the correct event.
Signed-off-by: Steve Kuznetsov <skuznets@redhat.com>
Without these select statements, this test runs until the package-global
timeout and causes a panic. This change makes the test fail faster and
more legibly.
Signed-off-by: Steve Kuznetsov <skuznets@redhat.com>
If a staging repo has .gitattributes files containing the `export-subst`
attribute ([example](b6c06a95d7/staging/src/k8s.io/client-go/pkg/version/.gitattributes)), then git expands the specified placeholders
when git archive is used.
When a published repo is downloaded from GitHub, GitHub does a
"git archive" under the hood. This means that the placeholders get
replaced by their relevant values. This type of "git archive"
application sometimes leads to undesired values. See the example below.
- In client-go, [line 59 in `pkg/version/base.go`](b6c06a95d7/staging/src/k8s.io/client-go/pkg/version/base.go (L59))
is expanded on a git archive. This line is needed as a fallback to
inject k8s version info for client-go when this info is not provided
via ldflags during builds.
- However, when k/client-go is vendored, the line gets expanded to _the
commit of the project vendoring client-go_ -- which is not helpful at
all! This also means that the vendored client-go will now contain
different (expanded) commit shas for different projects.
- To ensure reproducibility of source, this commit helps remove
the .gitattributes files before publishing the staging repos.
Additionally, when client-go is used as a library, we don't care about
the line being expanded to inject version info so it is also safe to
remove these files.
When tests attempt to validate behavior in the case that a client asks
for a resource version that is "too large" for the underlying storage,
the previous implementation would simply add 1 to the latest revision
seen. This is only appropriate for storage backends that
a) provide a continuous monotonic logical clock
b) have no other events occurring while the test runs
For instance, when using a singe etcd backend as a shared fixture for
these tests, adding 1 to a previously-seen revision is not suffcient to
ensure that the resulting revision is "too large". By instead using the
largest possible integer value, we can be certain of this.
Signed-off-by: Steve Kuznetsov <skuznets@redhat.com>
