integrationtest

func AssertLeakDetector ¶

func AssertLeakDetector(t *testing.T, want int)

AssertLeakDetector creates a new subtest that ensures that the leak detector succeeded with exactly 'want' subprocesses. When the leak detector is not enabled, the subtest is skipped.

This function is untested.

func GetLogger ¶

func GetLogger() logging.Logger

GetLogger returns a pointer to a logger that is to be shared amongst all test cases.

This function can safely be called from different test cases, and in different goroutines at once. It will always return the same logger.

func NewServer ¶

func NewServer(options *proxyssh.Options, configurations ...proxyssh.Configuration) (testServer *ssh.Server, testLogger logging.Logger, cleanup func())

NewServer creates a new proxyssh.Server for an integration test. Returns the test server and a cleanup function to be defered.

When options is nil, sets up empty options.

When testServer.Handler remains unset, sets it to respond with "success" on stdout and exit immediatly. When testServer.HostSigners remains unset, generates a single reusable test server key.

This function should be used like:

testServer, cleanup := NewServer(testServer)
defer cleanup()

This function is untested.

func RunComposeTest ¶

func RunComposeTest(config string, files map[string]string, testcode ComposeTestCode) error

RunComposeTest runs a docker-compose based test. It first starts the docker-compose configuration contained in the string. Thenm it writes out addtional data files in the provided map and waits for it to become available. Finally, it runs the testcode function.

Calling this function is roughly equivalent to the following bash pseudo-code:

mktmp -d
echo $config > tmpdir/docker-compose.yml
docker-compose pull
docker-compose up -d tmpdir/docker-compose.yml
// ... call f() ...
docker-compose down -v
rm -rf tmpdir

For source code compatibility purposes, all occurences of '\t' in the compose file will be replace with four spaces.

This command depends on the 'docker-compose' executable to be available in path.

This function has two kinds of error conditions, those that occur during setting up the project, and those that occur in the testcode. If an error occurs during setup or teardown, panic() is called. If an error occurs during the testcase, and testcase does not call panic(), the error is returned by this function.

This function is itself untested.