testing

package module
v0.0.0-...-c230cba Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Feb 24, 2015 License: LGPL-3.0 Imports: 28 Imported by: 0

README

juju/testing

This package provides additional base test suites to be used with gocheck.

Documentation

Index

Constants

View Source
const (
	// EchoQuotedArgs is a simple bash script that prints out the
	// basename of the command followed by the args as quoted strings.
	EchoQuotedArgsUnix = `#!/bin/bash --norc
name=` + "`basename $0`" + `
argfile="$name.out"
rm -f $argfile
printf "%s" $name | tee -a $argfile
for arg in "$@"; do
  printf " \"%s\""  "$arg" | tee -a $argfile
done
printf "\n" | tee -a $argfile
`
	EchoQuotedArgsWindows = `` /* 248-byte string literal not displayed */

)
View Source
const (

	// The default password to use when connecting to the mongo database.
	DefaultMongoPassword = "conn-from-name-secret"
)

Variables

View Source
var HookChannelSize = 10
View Source
var (
	// MgoServer is a shared mongo server used by tests.
	MgoServer = &MgoInstance{}
)
View Source
var Server = NewHTTPServer(5 * time.Second)

Functions

func AssertEchoArgs

func AssertEchoArgs(c *gc.C, execName string, args ...string)

AssertEchoArgs is used to check the args from an execution of a command that has been patchec using PatchExecutable containing EchoQuotedArgs.

func FindImports

func FindImports(packageName, prefix string) ([]string, error)

FindImports returns a sorted list of packages imported by the package with the given name that have the given prefix. The resulting list removes the common prefix, leaving just the short names.

func FindTCPPort

func FindTCPPort() int

FindTCPPort finds an unused TCP port and returns it. Use of this function has an inherent race condition - another process may claim the port before we try to use it. We hope that the probability is small enough during testing to be negligible.

func HomePath

func HomePath(names ...string) string

HomePath joins the specified path snippets and returns an absolute path under Juju home.

func HookCommandOutput

func HookCommandOutput(
	outputFunc *func(cmd *exec.Cmd) ([]byte, error), output []byte, err error) (<-chan *exec.Cmd, func())

HookCommandOutput intercepts CommandOutput to a function that passes the actual command and it's output back via a channel, and returns the error passed into this function. It also returns a cleanup function so you can restore the original function

func MgoDialInfo

func MgoDialInfo(certs *Certs, addrs ...string) *mgo.DialInfo

MgoDialInfo returns a DialInfo suitable for dialling an MgoInstance at any of the given addresses, optionally using TLS.

func MgoTestPackage

func MgoTestPackage(t *testing.T, certs *Certs)

MgoTestPackage should be called to register the tests for any package that requires a MongoDB server. If certs is non-nil, a secure SSL connection will be used from client to server.

func PatchExecutable

func PatchExecutable(c *gc.C, patcher EnvironmentPatcher, execName, script string)

PatchExecutable creates an executable called 'execName' in a new test directory and that directory is added to the path.

func PatchExecutableAsEchoArgs

func PatchExecutableAsEchoArgs(c *gc.C, patcher CleanupPatcher, execName string)

PatchExecutableAsEchoArgs creates an executable called 'execName' in a new test directory and that directory is added to the path. The content of the script is 'EchoQuotedArgs', and the args file is removed using a cleanup function.

func PatchExecutableThrowError

func PatchExecutableThrowError(c *gc.C, patcher CleanupPatcher, execName string, exitCode int)

PatchExecutableThrowError is needed to test cases in which we expect exit codes from executables called from the system path

Types

type Certs

type Certs struct {
	// CACert holds the CA certificate. This must certify the private key that
	// was used to sign the server certificate.
	CACert *x509.Certificate
	// ServerCert holds the certificate that certifies the server's
	// private key.
	ServerCert *x509.Certificate
	// ServerKey holds the server's private key.
	ServerKey *rsa.PrivateKey
}

Certs holds the certificates and keys required to make a secure SSL connection.

type CleanupFunc

type CleanupFunc func(*gc.C)

type CleanupPatcher

type CleanupPatcher interface {
	PatchEnvironment(name, value string)
	AddCleanup(cleanup CleanupFunc)
}

type CleanupSuite

type CleanupSuite struct {
	// contains filtered or unexported fields
}

CleanupSuite adds the ability to add cleanup functions that are called during either test tear down or suite tear down depending on the method called.

func (*CleanupSuite) AddCleanup

func (s *CleanupSuite) AddCleanup(cleanup CleanupFunc)

AddCleanup pushes the cleanup function onto the stack of functions to be called during TearDownTest.

func (*CleanupSuite) AddSuiteCleanup

func (s *CleanupSuite) AddSuiteCleanup(cleanup CleanupFunc)

AddSuiteCleanup pushes the cleanup function onto the stack of functions to be called during TearDownSuite.

func (*CleanupSuite) HookCommandOutput

func (s *CleanupSuite) HookCommandOutput(
	outputFunc *func(cmd *exec.Cmd) ([]byte, error),
	output []byte,
	err error,
) <-chan *exec.Cmd

HookCommandOutput calls the package function of the same name to mock out the result of a particular comand execution, and will call the restore function on test teardown.

func (*CleanupSuite) PatchEnvPathPrepend

func (s *CleanupSuite) PatchEnvPathPrepend(dir string)

PatchEnvPathPrepend prepends the given path to the environment $PATH and restores the original path on test teardown.

func (*CleanupSuite) PatchEnvironment

func (s *CleanupSuite) PatchEnvironment(name, value string)

PatchEnvironment sets the environment variable 'name' the the value passed in. The old value is saved and returned to the original value at test tear down time using a cleanup function.

func (*CleanupSuite) PatchValue

func (s *CleanupSuite) PatchValue(dest, value interface{})

PatchValue sets the 'dest' variable the the value passed in. The old value is saved and returned to the original value at test tear down time using a cleanup function. The value must be assignable to the element type of the destination.

func (*CleanupSuite) SetUpSuite

func (s *CleanupSuite) SetUpSuite(c *gc.C)

func (*CleanupSuite) SetUpTest

func (s *CleanupSuite) SetUpTest(c *gc.C)

func (*CleanupSuite) TearDownSuite

func (s *CleanupSuite) TearDownSuite(c *gc.C)

func (*CleanupSuite) TearDownTest

func (s *CleanupSuite) TearDownTest(c *gc.C)

type EnvironmentPatcher

type EnvironmentPatcher interface {
	PatchEnvironment(name, value string)
}

EnvironmentPatcher is an interface that requires just one method: PatchEnvironment.

type FakeHome

type FakeHome struct {
	// contains filtered or unexported fields
}

FakeHome stores information about the user's home environment so it can be cast aside for tests and restored afterwards.

func MakeFakeHome

func MakeFakeHome(c *gc.C) *FakeHome

func (*FakeHome) AddFiles

func (h *FakeHome) AddFiles(c *gc.C, files ...TestFile)

func (*FakeHome) FileContents

func (h *FakeHome) FileContents(c *gc.C, path string) string

FileContents returns the test file contents for the given specified path (which may be relative, so we compare with the base filename only).

func (*FakeHome) FileExists

func (h *FakeHome) FileExists(path string) bool

FileExists returns if the given relative file path exists in the fake home.

type FakeHomeSuite

type FakeHomeSuite struct {
	CleanupSuite
	LoggingSuite
	Home *FakeHome
}

FakeHomeSuite sets up a fake home directory before running tests.

func (*FakeHomeSuite) SetUpSuite

func (s *FakeHomeSuite) SetUpSuite(c *gc.C)

func (*FakeHomeSuite) SetUpTest

func (s *FakeHomeSuite) SetUpTest(c *gc.C)

func (*FakeHomeSuite) TearDownSuite

func (s *FakeHomeSuite) TearDownSuite(c *gc.C)

func (*FakeHomeSuite) TearDownTest

func (s *FakeHomeSuite) TearDownTest(c *gc.C)

type HTTPServer

type HTTPServer struct {
	URL     string
	Timeout time.Duration
	// contains filtered or unexported fields
}

func NewHTTPServer

func NewHTTPServer(timeout time.Duration) *HTTPServer

func (*HTTPServer) Flush

func (s *HTTPServer) Flush()

Flush discards all pending requests and responses.

func (*HTTPServer) Response

func (s *HTTPServer) Response(status int, headers map[string]string, body []byte)

Response prepares the test server to respond the following request using the provided response parameters.

func (*HTTPServer) ResponseFunc

func (s *HTTPServer) ResponseFunc(n int, f ResponseFunc)

ResponseFunc prepares the test server to respond the following n requests using f to build each response.

func (*HTTPServer) ResponseMap

func (s *HTTPServer) ResponseMap(n int, m ResponseMap)

ResponseMap prepares the test server to respond the following n requests using the m to obtain the responses.

func (*HTTPServer) Responses

func (s *HTTPServer) Responses(n int, status int, headers map[string]string, body []byte)

Responses prepares the test server to respond the following n requests using the provided response parameters.

func (*HTTPServer) ServeHTTP

func (s *HTTPServer) ServeHTTP(w http.ResponseWriter, req *http.Request)

func (*HTTPServer) Start

func (s *HTTPServer) Start()

func (*HTTPServer) WaitRequest

func (s *HTTPServer) WaitRequest() *http.Request

WaitRequest returns the next request made to the http server from the queue. If no requests were previously made, it waits until the timeout value for one to be made.

func (*HTTPServer) WaitRequests

func (s *HTTPServer) WaitRequests(n int) []*http.Request

WaitRequests returns the next n requests made to the http server from the queue. If not enough requests were previously made, it waits until the timeout value for them to be made.

type HTTPSuite

type HTTPSuite struct{}

func (*HTTPSuite) SetUpSuite

func (s *HTTPSuite) SetUpSuite(c *gc.C)

func (*HTTPSuite) SetUpTest

func (s *HTTPSuite) SetUpTest(c *gc.C)

func (*HTTPSuite) TearDownSuite

func (s *HTTPSuite) TearDownSuite(c *gc.C)

func (*HTTPSuite) TearDownTest

func (s *HTTPSuite) TearDownTest(c *gc.C)

func (*HTTPSuite) URL

func (s *HTTPSuite) URL(path string) string

type IsolationSuite

type IsolationSuite struct {
	OsEnvSuite
	CleanupSuite
	LoggingSuite
}

IsolationSuite isolates the tests from the underlaying system environment, sets up test logging and exposes cleanup facilities.

func (*IsolationSuite) SetUpSuite

func (s *IsolationSuite) SetUpSuite(c *gc.C)

func (*IsolationSuite) SetUpTest

func (s *IsolationSuite) SetUpTest(c *gc.C)

func (*IsolationSuite) TearDownSuite

func (s *IsolationSuite) TearDownSuite(c *gc.C)

func (*IsolationSuite) TearDownTest

func (s *IsolationSuite) TearDownTest(c *gc.C)

type LoggingCleanupSuite

type LoggingCleanupSuite struct {
	LoggingSuite
	CleanupSuite
}

LoggingCleanupSuite is defined for backward compatibility. Do not use this suite in new tests.

func (*LoggingCleanupSuite) SetUpSuite

func (s *LoggingCleanupSuite) SetUpSuite(c *gc.C)

func (*LoggingCleanupSuite) SetUpTest

func (s *LoggingCleanupSuite) SetUpTest(c *gc.C)

func (*LoggingCleanupSuite) TearDownSuite

func (s *LoggingCleanupSuite) TearDownSuite(c *gc.C)

func (*LoggingCleanupSuite) TearDownTest

func (s *LoggingCleanupSuite) TearDownTest(c *gc.C)

type LoggingSuite

type LoggingSuite struct{}

LoggingSuite redirects the juju logger to the test logger when embedded in a gocheck suite type.

func (*LoggingSuite) SetUpSuite

func (s *LoggingSuite) SetUpSuite(c *gc.C)

func (*LoggingSuite) SetUpTest

func (s *LoggingSuite) SetUpTest(c *gc.C)

func (*LoggingSuite) TearDownSuite

func (s *LoggingSuite) TearDownSuite(c *gc.C)

func (*LoggingSuite) TearDownTest

func (s *LoggingSuite) TearDownTest(c *gc.C)

type MgoInstance

type MgoInstance struct {

	// Params is a list of additional parameters that will be passed to
	// the mongod application
	Params []string

	// EnableJournal enables journaling.
	EnableJournal bool

	// EnableAuth enables authentication/authorization.
	EnableAuth bool

	// WithoutV8 is true if we believe this Mongo doesn't actually have the
	// V8 engine
	WithoutV8 bool
	// contains filtered or unexported fields
}

func (*MgoInstance) Addr

func (m *MgoInstance) Addr() string

Addr returns the address of the MongoDB server.

func (*MgoInstance) Destroy

func (inst *MgoInstance) Destroy()

Destroy kills mongod and cleans up its data directory.

func (*MgoInstance) DestroyWithLog

func (inst *MgoInstance) DestroyWithLog()

DestroyWithLog causes mongod to exit, cleans up its data directory, and captures the last N lines of mongod's log output.

func (*MgoInstance) Dial

func (inst *MgoInstance) Dial() (*mgo.Session, error)

Dial returns a new connection to the MongoDB server.

func (*MgoInstance) DialDirect

func (inst *MgoInstance) DialDirect() (*mgo.Session, error)

DialDirect returns a new direct connection to the shared MongoDB server. This must be used if you're connecting to a replicaset that hasn't been initiated yet.

func (*MgoInstance) DialInfo

func (inst *MgoInstance) DialInfo() *mgo.DialInfo

DialInfo returns information suitable for dialling the receiving MongoDB instance.

func (*MgoInstance) MustDial

func (inst *MgoInstance) MustDial() *mgo.Session

MustDial returns a new connection to the MongoDB server, and panics on errors.

func (*MgoInstance) MustDialDirect

func (inst *MgoInstance) MustDialDirect() *mgo.Session

MustDialDirect works like DialDirect, but panics on errors.

func (*MgoInstance) Port

func (m *MgoInstance) Port() int

Port returns the port of the MongoDB server.

func (*MgoInstance) Reset

func (inst *MgoInstance) Reset()

Reset deletes all content from the MongoDB server and panics if it encounters errors.

func (*MgoInstance) Restart

func (inst *MgoInstance) Restart()

Restart restarts the mongo server, useful for testing what happens when a state server goes down.

func (*MgoInstance) Start

func (inst *MgoInstance) Start(certs *Certs) error

Start starts a MongoDB server in a temporary directory.

type MgoSuite

type MgoSuite struct {
	Session *mgo.Session
}

MgoSuite is a suite that deletes all content from the shared MongoDB server at the end of every test and supplies a connection to the shared MongoDB server.

func (*MgoSuite) SetUpSuite

func (s *MgoSuite) SetUpSuite(c *gc.C)

func (*MgoSuite) SetUpTest

func (s *MgoSuite) SetUpTest(c *gc.C)

func (*MgoSuite) TearDownSuite

func (s *MgoSuite) TearDownSuite(c *gc.C)

func (*MgoSuite) TearDownTest

func (s *MgoSuite) TearDownTest(c *gc.C)

type OsEnvSuite

type OsEnvSuite struct {
	// contains filtered or unexported fields
}

OsEnvSuite isolates the tests from the underlaying system environment. Environment variables are reset in SetUpTest and restored in TearDownTest.

func (*OsEnvSuite) SetUpSuite

func (s *OsEnvSuite) SetUpSuite(c *gc.C)

func (*OsEnvSuite) SetUpTest

func (s *OsEnvSuite) SetUpTest(c *gc.C)

func (*OsEnvSuite) TearDownSuite

func (s *OsEnvSuite) TearDownSuite(c *gc.C)

func (*OsEnvSuite) TearDownTest

func (s *OsEnvSuite) TearDownTest(c *gc.C)

type Response

type Response struct {
	Status  int
	Headers map[string]string
	Body    []byte
}

type ResponseFunc

type ResponseFunc func(path string) Response

type ResponseMap

type ResponseMap map[string]Response

ResponseMap maps request paths to responses.

type Restorer

type Restorer func()

Restorer holds a function that can be used to restore some previous state.

func PatchEnvPathPrepend

func PatchEnvPathPrepend(dir string) Restorer

PatchEnvPathPrepend provides a simple way to prepend path to the start of the PATH environment variable. Returns a function that restores the environment to what it was before.

func PatchEnvironment

func PatchEnvironment(name, value string) Restorer

PatchEnvironment provides a test a simple way to override a single environment variable. A function is returned that will return the environment to what it was before.

func PatchValue

func PatchValue(dest, value interface{}) Restorer

PatchValue sets the value pointed to by the given destination to the given value, and returns a function to restore it to its original value. The value must be assignable to the element type of the destination.

func (Restorer) Add

func (f Restorer) Add(f1 Restorer) Restorer

Add returns a Restorer that restores first f1 and then f. It is valid to call this on a nil Restorer.

func (Restorer) Restore

func (r Restorer) Restore()

Restore restores some previous state.

type Stub

type Stub struct {
	// Calls is the list of calls that have been registered on the stub
	// (i.e. made on the stub's methods), in the order that they were
	// made.
	Calls []StubCall

	// Receivers is the list of receivers for all the recorded calls.
	// In the case of non-methods, the receiver is set to nil. The
	// receivers are tracked here rather than as a Receiver field on
	// StubCall because StubCall represents the common case for
	// testing. Typically the receiver does not need to be checked.
	Receivers []interface{}

	// Errors holds the list of error return values to use for
	// successive calls to methods that return an error. Each call
	// pops the next error off the list. An empty list (the default)
	// implies a nil error. nil may be precede actual errors in the
	// list, which means that the first calls will succeed, followed
	// by the failure. All this is facilitated through the Err method.
	Errors []error

	// DefaultError is the default error (when Errors is empty). The
	// typical Stub usage will leave this nil (i.e. no error).
	DefaultError error
}

Stub is used in testing to stand in for some other value, to record all calls to stubbed methods/functions, and to allow users to set the values that are returned from those calls. Stub is intended to be embedded in another struct that will define the methods to track:

type stubConn struct {
    *testing.Stub
    Response []byte
}

func newStubConn() *stubConn {
    return &stubConn{
        Stub: &testing.Stub{},
    }
}

// Send implements Connection.
func (fc *stubConn) Send(request string) []byte {
    fc.MethodCall(fc, "Send", request)
    return fc.Response, fc.NextErr()
}

As demonstrated in the example, embed a pointer to testing.Stub. This allows a single testing.Stub to be shared between multiple stubs.

Error return values are set through Stub.Errors. Set it to the errors you want returned (or use the convenience method `SetErrors`). The `NextErr` method returns the errors from Stub.Errors in sequence, falling back to `DefaultError` when the sequence is exhausted. Thus each stubbed method should call `NextErr` to get its error return value.

To validate calls made to the stub in a test, check Stub.Calls or call the CheckCalls (or CheckCall) method:

c.Check(s.stub.Calls, jc.DeepEquals, []StubCall{{
    FuncName: "Send",
    Args: []interface{}{
        expected,
    },
}})

s.stub.CheckCalls(c, []StubCall{{
    FuncName: "Send",
    Args: []interface{}{
        expected,
    },
}})

s.stub.CheckCall(c, 0, "Send", expected)

Not only is Stub useful for building a interface implementation to use in testing (e.g. a network API client), it is also useful in regular function patching situations:

type myStub struct {
    *testing.Stub
}

func (f *myStub) SomeFunc(arg interface{}) error {
    f.AddCall("SomeFunc", arg)
    return f.NextErr()
}

s.PatchValue(&somefunc, s.myStub.SomeFunc)

This allows for easily monitoring the args passed to the patched func, as well as controlling the return value from the func in a clean manner (by simply setting the correct field on the stub).

func (*Stub) AddCall

func (f *Stub) AddCall(funcName string, args ...interface{})

AddCall records a stubbed function call for later inspection using the CheckCalls method. A nil receiver is recorded. Thus for methods use MethodCall. All stubbed functions should call AddCall.

func (*Stub) CheckCall

func (f *Stub) CheckCall(c *gc.C, index int, funcName string, args ...interface{})

CheckCall checks the recorded call at the given index against the provided values. If the index is out of bounds then the check fails. The receiver is not checked. If it is significant for a test then it can be checked separately:

c.Check(mystub.Receivers[index], gc.Equals, expected)

func (*Stub) CheckCallNames

func (f *Stub) CheckCallNames(c *gc.C, expected ...string) bool

CheckCallNames verifies that the in-order list of called method names matches the expected calls.

func (*Stub) CheckCalls

func (f *Stub) CheckCalls(c *gc.C, expected []StubCall)

CheckCalls verifies that the history of calls on the stub's methods matches the expected calls. The receivers are not checked. If they are significant then check Stub.Receivers separately.

func (*Stub) MethodCall

func (f *Stub) MethodCall(receiver interface{}, funcName string, args ...interface{})

MethodCall records a stubbed method call for later inspection using the CheckCalls method. The receiver is added to Stub.Receivers.

func (*Stub) NextErr

func (f *Stub) NextErr() error

NextErr returns the error that should be returned on the nth call to any method on the stub. It should be called for the error return in all stubbed methods.

func (*Stub) SetErrors

func (f *Stub) SetErrors(errors ...error)

SetErrors sets the sequence of error returns for the stub. Each call to Err (thus each stub method call) pops an error off the front. So frontloading nil here will allow calls to pass, followed by a failure.

type StubCall

type StubCall struct {
	// Funcname is the name of the function that was called.
	FuncName string

	// Args is the set of arguments passed to the function. They are
	// in the same order as the function's parameters
	Args []interface{}
}

StubCall records the name of a called function and the passed args.

type TestFile

type TestFile struct {
	Name, Data string
}

Directories

Path Synopsis

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL