planner

Documentation

Overview

Package planner contains logic to build and execute a test plan.

Index

• func RunTests(ctx context.Context, tests []*testing.TestInstance, out OutputStream, ...) error
• type Config
• type DownloadMode
• type OutputStream

Functions

func RunTests

func RunTests(ctx context.Context, tests []*testing.TestInstance, out OutputStream, pcfg *Config) error

RunTests runs a set of tests, writing outputs to out.

RunTests is responsible for building an efficient plan to run the given tests. Therefore the order of tests in the argument is ignored; it just specifies a set of tests to run.

RunTests runs tests on goroutines. If a test does not finish after reaching its timeout, this function returns with an error without waiting for its finish.

Types

type Config

type Config struct {
	// DataDir is the path to the base directory containing test data files.
	DataDir string
	// OutDir is the path to the base directory under which tests should write output files.
	OutDir string
	// Vars contains names and values of runtime variables used to pass out-of-band data to tests.
	Vars map[string]string
	// Features contains software/hardware features the DUT has.
	Features dep.Features
	// Devservers contains URLs of devservers that can be used to download files.
	Devservers []string
	// TLWServer is the address of TLW server
	TLWServer string
	// DUTName is the name given by the infra scheduler
	DUTName string
	// BuildArtifactsURL is the URL of Google Cloud Storage directory, ending with a slash,
	// containing build artifacts for the current Chrome OS image.
	BuildArtifactsURL string
	// RemoteData contains information relevant to remote tests.
	// It is nil for local tests.
	RemoteData *testing.RemoteData
	// TestHook is run before TestInstance.Func (and TestInstance.Pre.Prepare, when applicable) if non-nil.
	// The returned closure is executed after a test if not nil.
	TestHook func(context.Context, *testing.TestHookState) func(context.Context, *testing.TestHookState)
	// DownloadMode specifies a strategy to download external data files.
	DownloadMode DownloadMode
	// Fixtures is a map from a fixture name to its metadata.
	Fixtures map[string]*testing.Fixture
	// StartFixtureName is a name of a fixture to start test execution.
	// Tests requested to run should depend on the start fixture directly or
	// indirectly.
	// Since a start fixture is treated specially (e.g. no output directory is
	// created), metadata of a start fixture must not be contained in
	// Config.Fixtures. Instead, StartFixtureImpl gives an implementation of
	// a start fixture.
	StartFixtureName string
	// StartFixtureImpl gives an implementation of a start fixture.
	// If it is nil, a default stub implementation is used.
	StartFixtureImpl testing.FixtureImpl
}

Config contains details about how the planner should run tests.

type DownloadMode

type DownloadMode int

DownloadMode specifies a strategy to download external data files.

const (
	// DownloadBatch specifies that the planner downloads external data files
	// in batch before running tests.
	DownloadBatch DownloadMode = iota
	// DownloadLazy specifies that the planner download external data files
	// as needed between tests.
	DownloadLazy
)

type OutputStream

type OutputStream interface {
	// EntityStart reports that an entity has started.
	EntityStart(ei *testing.EntityInfo, outDir string) error
	// EntityLog reports an informational log message.
	EntityLog(ei *testing.EntityInfo, msg string) error
	// EntityError reports an error from an entity. An entity that reported one or more errors should be considered failure.
	EntityError(ei *testing.EntityInfo, e *testing.Error) error
	// EntityEnd reports that an entity has ended. If skipReasons is not empty it is considered skipped.
	EntityEnd(ei *testing.EntityInfo, skipReasons []string, timingLog *timing.Log) error
}

OutputStream is an interface to report streamed outputs of multiple entity runs. Note that testing.OutputStream is for a single entity in contrast.