sdktests

package
v2.0.0-alpha.12 Latest Latest
Warning

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

 Go to latest Published: Aug 29, 2022 License: Apache-2.0 Imports: 36 Imported by: 0

Documentation

Overview

Package sdktests contains the domain-specific SDK test logic.

Tests in this package use other packages as follows:

ldtest: the basic test scope framework

mockd: mock LaunchDarkly service components

servicedef: types used in communication with an SDK test service

data: test data file schemas and loader

Index

Constants

This section is empty.

Variables

This section is empty.

Functions

func Base64DecodedData 

func Base64DecodedData() m.MatcherTransform

func EqualReason 

func EqualReason(reason ldreason.EvaluationReason) m.Matcher

EqualReason is a type-safe replacement for m.Equal(EvaluationReason) that also provides better failure output, by treating it as a JSON object-- since the default implementation of String() for EvaluationReason returns a shorter string that doesn't include every property.

func EvalAllFlagsStateMap 

func EvalAllFlagsStateMap() m.MatcherTransform

func EvalAllFlagsValueForKeyShouldEqual 

func EvalAllFlagsValueForKeyShouldEqual(key string, value ldvalue.Value) m.Matcher

func EvalResponseReason 

func EvalResponseReason() m.MatcherTransform

func EvalResponseValue 

func EvalResponseValue() m.MatcherTransform

func EvalResponseVariation 

func EvalResponseVariation() m.MatcherTransform

func EventHasKind 

func EventHasKind(kind string) m.Matcher

func HasAnyCreationDate 

func HasAnyCreationDate() m.Matcher

func HasAuthorizationHeader 

func HasAuthorizationHeader(authKey string) m.Matcher

HasAuthorizationHeader is a matcher for an http.Header map that verifies that the Authorization header is present and contains the specified key. Some SDKs send just the raw key, while others prefix it with an "api_key" scheme identifier; the latter is more technically correct, but we need to allow both since LD allows both.

func HasContextKeys 

func HasContextKeys(context ldcontext.Context) m.Matcher

func HasContextObjectWithKey 

func HasContextObjectWithKey(key string) m.Matcher

func HasContextObjectWithMatchingKeys 

func HasContextObjectWithMatchingKeys(context ldcontext.Context) m.Matcher

func HasNoAuthorizationHeader 

func HasNoAuthorizationHeader() m.Matcher

func HasNoContextObject 

func HasNoContextObject() m.Matcher
func Header(name string) m.MatcherTransform

Header allows matchers to be applied to a specific named header from an http.Header map. It assumes that there is just one value for that name (i.e. it calls Header.Get()).

func IsCustomEvent 

func IsCustomEvent() m.Matcher

func IsCustomEventForEventKey 

func IsCustomEventForEventKey(key string) m.Matcher

func IsDebugEvent 

func IsDebugEvent() m.Matcher

func IsFeatureEvent 

func IsFeatureEvent() m.Matcher

func IsIdentifyEvent 

func IsIdentifyEvent() m.Matcher

func IsIdentifyEventForContext 

func IsIdentifyEventForContext(context ldcontext.Context) m.Matcher

func IsIndexEvent 

func IsIndexEvent() m.Matcher

func IsIndexEventForContext 

func IsIndexEventForContext(context ldcontext.Context) m.Matcher

func IsSummaryEvent 

func IsSummaryEvent() m.Matcher

func IsValidFeatureEventWithConditions 

func IsValidFeatureEventWithConditions(matchers ...m.Matcher) m.Matcher

func IsValidSummaryEventWithFlags 

func IsValidSummaryEventWithFlags(keyValueMatchers ...m.KeyValueMatcher) m.Matcher

func JSONMatchesContext 

func JSONMatchesContext(context ldcontext.Context) m.Matcher

JSONMatchesContext builds a Matcher to verify that the input JSON is a valid representation of the specified context. This is using the regular context schema (i.e. what would be sent to evaluation endpoints), not the event schema.

The matcher should be tolerant of all allowable variants: for instance, it is legal to include `"anonymous": false` in the representation rather than omitting anonymous.

func JSONMatchesEventContext 

func JSONMatchesEventContext(context ldcontext.Context, redactedShouldBe []string) m.Matcher

JSONMatchesEventContext builds a Matcher to verify that the input JSON is a valid representation of the specified context within event data. The context should represent the attributes *after* any private attribute redaction; redactedShouldBe specifies what we expect to see in redactedAttributes.

The matcher should be tolerant of all allowable variants: for instance, it is legal to include `"anonymous": false` in the representation rather than omitting anonymous, and attribute names that appear in redactedAttributes could either use the literal syntax or the slash syntax.

func JSONPropertyKeysCanOnlyBe 

func JSONPropertyKeysCanOnlyBe(keys ...string) m.Matcher

func JSONPropertyNullOrAbsent 

func JSONPropertyNullOrAbsent(name string) m.Matcher

func JSONPropertyNullOrAbsentOrEqualTo 

func JSONPropertyNullOrAbsentOrEqualTo(name string, emptyValue interface{}) m.Matcher

func RedactedAttributesAre 

func RedactedAttributesAre(attrStrings ...string) m.Matcher

RedactedAttributesAre is a matcher for the value of an event context's redactedAttributes property, verifying that it has the specified attribute names/references and no others. This is not just a plain slice match, because 1. they can be in any order and 2. for simple attribute names, the SDK is allowed to send either "name" or "/name" (with any slashes or tildes escaped in the latter case).

func RunSDKTestSuite 

func RunSDKTestSuite(
	harness *harness.TestHarness,
	filter ldtest.Filter,
	testLogger ldtest.TestLogger,
) ldtest.Results

func SortedStrings 

func SortedStrings() m.MatcherTransform

func ValueIsPositiveNonZeroInteger 

func ValueIsPositiveNonZeroInteger() m.Matcher

Types

type BigSegmentStore 

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

BigSegmentStore is a test fixture that provides callback endpoints for SDK clients to connect to, behaving like a Big Segment store for a simulated database.

func NewBigSegmentStore 

func NewBigSegmentStore(t *ldtest.T, initialStatus ldreason.BigSegmentsStatus) *BigSegmentStore

NewBigSegmentStore creates a new BigSegmentStore with the specified initial status.

The object's lifecycle is tied to the test scope that created it; it will be automatically closed when this test scope exits. It can be reused by subtests until then. Debug output related to the data source will be attached to this test scope.

func (*BigSegmentStore) Configure 

func (b *BigSegmentStore) Configure(config *servicedef.SDKConfigParams) error

Configure updates the SDK client configuration for NewSDKClient, causing the SDK to connect to the appropriate base URI for the big segments test fixture.

func (*BigSegmentStore) ExpectMetadataQuery 

func (b *BigSegmentStore) ExpectMetadataQuery(t *ldtest.T, timeout time.Duration)

ExpectMetadataQuery blocks until the Big Segment store has received a metadata query.

func (*BigSegmentStore) ExpectNoMoreMetadataQueries 

func (b *BigSegmentStore) ExpectNoMoreMetadataQueries(t *ldtest.T, timeout time.Duration)

ExpectNoMoreMetadataQueries causes a test failure if the Big Segment store receives a metadata query.

func (*BigSegmentStore) GetMembershipQueries 

func (b *BigSegmentStore) GetMembershipQueries() []string

GetMembershipQueries returns the context hashes of all membership queries that have been received so far.

func (*BigSegmentStore) SetupGetMembership 

func (b *BigSegmentStore) SetupGetMembership(fn func(contextHash string) (map[string]bool, error))

SetupGetMembership causes the specified function to be called whenever the SDK calls the "get membership" method on the Big Segment store.

func (*BigSegmentStore) SetupGetMetadata 

func (b *BigSegmentStore) SetupGetMetadata(fn func() (ldtime.UnixMillisecondTime, error))

SetupGetMetadata causes the specified function to be called whenever the SDK calls the "get metadata" method on the Big Segment store.

func (*BigSegmentStore) SetupMemberships 

func (b *BigSegmentStore) SetupMemberships(t *ldtest.T, memberships map[string]map[string]bool)

SetupMemberships is a shortcut to call SetupGetMembership with appropriate logic for providing preconfigured results for each possible context hash. Any context hash whose key does not appear in the map will cause the test to fail.

func (*BigSegmentStore) SetupMetadataForStatus 

func (b *BigSegmentStore) SetupMetadataForStatus(status ldreason.BigSegmentsStatus)

SetupMetadataForStatus is a shortcut to call SetupGetMetadata with appropriate logic for making the Big Segment store return a current time ("healthy" status), an old time ("stale" status), or an error ("store error" status).

type CommonEvalParameterizedTestRunner 

type CommonEvalParameterizedTestRunner[SDKDataType mockld.SDKData] struct {
	SDKConfigurers       func(testmodel.EvalTestSuite[SDKDataType]) []SDKConfigurer
	FilterSDKData        func(SDKDataType) SDKDataType
	FilterExpectedReason func(ldreason.EvaluationReason) ldreason.EvaluationReason
}

func (CommonEvalParameterizedTestRunner[SDKDataType]) RunAll 

func (c CommonEvalParameterizedTestRunner[SDKDataType]) RunAll(t *ldtest.T, dirName string)

type CommonEventTests 

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

CommonEventTests groups together event-related test methods that are shared between server-side and client-side.

func NewCommonEventTests 

func NewCommonEventTests(t *ldtest.T, testName string, baseSDKConfigurers ...SDKConfigurer) CommonEventTests

func (CommonEventTests) BufferBehavior 

func (c CommonEventTests) BufferBehavior(t *ldtest.T)

func (CommonEventTests) CustomEvents 

func (c CommonEventTests) CustomEvents(t *ldtest.T)

func (CommonEventTests) DisablingEvents 

func (c CommonEventTests) DisablingEvents(t *ldtest.T)

func (CommonEventTests) EventContexts 

func (c CommonEventTests) EventContexts(t *ldtest.T)

func (CommonEventTests) IdentifyEvents 

func (c CommonEventTests) IdentifyEvents(t *ldtest.T)

func (CommonEventTests) RequestMethodAndHeaders 

func (c CommonEventTests) RequestMethodAndHeaders(t *ldtest.T, credential string)

func (CommonEventTests) RequestURLPath 

func (c CommonEventTests) RequestURLPath(t *ldtest.T, pathMatcher m.Matcher)

func (CommonEventTests) UniquePayloadIDs 

func (c CommonEventTests) UniquePayloadIDs(t *ldtest.T)

type CommonPollingTests 

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

CommonPollingTests groups together polling-related test methods that are shared between server-side and client-side.

Currently we do not have any tests that actually test *repeated* polling. This is because the SDKs enforce minimum polling intervals that would cause the tests to take a very long time. Therefore, the current tests only cover the behavior of the initial poll request.

func NewCommonPollingTests 

func NewCommonPollingTests(t *ldtest.T, testName string, baseSDKConfigurers ...SDKConfigurer) CommonPollingTests

func (CommonPollingTests) RequestContextProperties 

func (c CommonPollingTests) RequestContextProperties(t *ldtest.T, getPath string)

func (CommonPollingTests) RequestMethodAndHeaders 

func (c CommonPollingTests) RequestMethodAndHeaders(t *ldtest.T, credential string)

func (CommonPollingTests) RequestURLPath 

func (c CommonPollingTests) RequestURLPath(t *ldtest.T, pathMatcher func(flagRequestMethod) m.Matcher)

type CommonStreamingTests 

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

CommonStreamingTests groups together streaming-related test methods that are shared between server-side and client-side.

func NewCommonStreamingTests 

func NewCommonStreamingTests(t *ldtest.T, testName string, baseSDKConfigurers ...SDKConfigurer) CommonStreamingTests

func (CommonStreamingTests) RequestContextProperties 

func (c CommonStreamingTests) RequestContextProperties(t *ldtest.T, getPath string)

func (CommonStreamingTests) RequestMethodAndHeaders 

func (c CommonStreamingTests) RequestMethodAndHeaders(t *ldtest.T, credential string)

func (CommonStreamingTests) RequestURLPath 

func (c CommonStreamingTests) RequestURLPath(t *ldtest.T, pathMatcher func(flagRequestMethod) m.Matcher)

func (CommonStreamingTests) Updates 

func (c CommonStreamingTests) Updates(t *ldtest.T)

type CommonTagsTests 

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

CommonTagsTests groups together event-related test methods that are shared between server-side and client-side.

func NewCommonTagsTests 

func NewCommonTagsTests(t *ldtest.T, testName string, baseSDKConfigurers ...SDKConfigurer) CommonTagsTests

func (CommonTagsTests) Run 

func (c CommonTagsTests) Run(t *ldtest.T)

type SDKClient 

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

SDKClient represents an SDK client instance in the test service which can be controlled by test logic.

func NewSDKClient 

func NewSDKClient(t *ldtest.T, configurers ...SDKConfigurer) *SDKClient

NewSDKClient tells the test service to create an SDK client instance.

The first parameter should be the current test scope. Any error in creating the client will cause the test to fail and terminate immediately. Debug output related to the client will be attached to this test scope.

You must always specify at least one SDKConfigurer to customize the SDK configuration, since a default SDK configuration would only connect to LaunchDarkly which is normally not what we want. Test fixture components such as SDKDataSource implement this interface so that they can insert the appropriate base URIs into the configuration, so a common pattern is: 

dataSource := NewSDKDataSource(t, ...)
eventSink := NewSDKEventSink(t, ...)
client := NewSDKClient(t, dataSource, eventSink)

Since the client will attempt to connect to its data source and possibly send events as soon as it starts up, the test fixtures must always be created first. You may reuse a previously created data source and event sink that was created in a parent test scope, if you do not need a new one for each client.

The object's lifecycle is tied to the test scope that created it; it will be automatically closed when this test scope exits. It can be reused by subtests until then.

func TryNewSDKClient 

func TryNewSDKClient(t *ldtest.T, configurers ...SDKConfigurer) (*SDKClient, error)

func (*SDKClient) Close 

func (c *SDKClient) Close() error

Close tells the test service to shut down the client instance. Normally this happens automatically at the end of a test.

func (*SDKClient) ContextBuild 

func (c *SDKClient) ContextBuild(t *ldtest.T, params servicedef.ContextBuildParams) servicedef.ContextBuildResponse

ContextBuild tells the test service to use the SDK's context builder to build a context and return it as JSON.

func (*SDKClient) ContextConvert 

func (c *SDKClient) ContextConvert(
	t *ldtest.T,
	params servicedef.ContextConvertParams,
) servicedef.ContextBuildResponse

ContextConvert tells the test service to use the SDK's JSON converters to unmarshal and remarshal a context.

func (*SDKClient) EvaluateAllFlags 

func (c *SDKClient) EvaluateAllFlags(
	t *ldtest.T,
	params servicedef.EvaluateAllFlagsParams,
) servicedef.EvaluateAllFlagsResponse

EvaluateAllFlags tells the SDK client to evaluate all feature flags. This corresponds to calling the SDK's AllFlags or AllFlagsState method.

Any error from the test service causes the test to terminate immediately.

func (*SDKClient) EvaluateFlag 

func (c *SDKClient) EvaluateFlag(t *ldtest.T, params servicedef.EvaluateFlagParams) servicedef.EvaluateFlagResponse

EvaluateFlag tells the SDK client to evaluate a feature flag. This corresponds to calling one of the SDK's Variation or VariationDetail methods, depending on the parameters.

Any error from the test service causes the test to terminate immediately.

func (*SDKClient) FlushEvents 

func (c *SDKClient) FlushEvents(t *ldtest.T)

FlushEvents tells the SDK client to initiate an event flush.

Any error from the test service causes the test to terminate immediately.

func (*SDKClient) GetBigSegmentStoreStatus 

func (c *SDKClient) GetBigSegmentStoreStatus(t *ldtest.T) servicedef.BigSegmentStoreStatusResponse

GetBigSegmentStoreStatus queries the big segment store status from the SDK client. The test harness will only call this method if the test service has the "big-segments" capability.

func (*SDKClient) GetSecureModeHash 

func (c *SDKClient) GetSecureModeHash(t *ldtest.T, context ldcontext.Context) string

GetSecureModeHash tells the SDK client to calculate a secure mode hash for a context. The test harness will only call this method if the test service has the "secure-mode-hash" capability.

func (*SDKClient) SendCustomEvent 

func (c *SDKClient) SendCustomEvent(t *ldtest.T, params servicedef.CustomEventParams)

SendCustomEvent tells the SDK client to send a custom event.

Any error from the test service causes the test to terminate immediately.

func (*SDKClient) SendIdentifyEvent 

func (c *SDKClient) SendIdentifyEvent(t *ldtest.T, context ldcontext.Context)

SendIdentifyEvent tells the SDK client to send an identify event.

Any error from the test service causes the test to terminate immediately.

type SDKConfigurer 

type SDKConfigurer helpers.ConfigOption[servicedef.SDKConfigParams]

SDKConfigurer is an interface for objects that can modify the configuration for StartSDKClient. It is implemented by types such as SDKDataSource.

func WithClientSideConfig 

func WithClientSideConfig(clientSideConfig servicedef.SDKConfigClientSideParams) SDKConfigurer

WithClientSideConfig is used with StartSDKClient to specify a non-default client-side SDK configuration.

func WithConfig 

func WithConfig(config servicedef.SDKConfigParams) SDKConfigurer

WithConfig is used with StartSDKClient to specify a non-default SDK configuration. Use this before any other SDKConfigurers or it will overwrite their effects.

func WithCredential 

func WithCredential(credential string) SDKConfigurer

WithCredential is used with StartSDKClient to set only the credential (SDK key, mobile key, or environment ID).

func WithEventsConfig 

func WithEventsConfig(eventsConfig servicedef.SDKConfigEventParams) SDKConfigurer

WithEventsConfig is used with StartSDKClient to specify a non-default events configuration.

func WithPollingConfig 

func WithPollingConfig(pollingConfig servicedef.SDKConfigPollingParams) SDKConfigurer

WithPollingConfig is used with StartSDKClient to specify a non-default polling configuration.

func WithServiceEndpointsConfig 

func WithServiceEndpointsConfig(endpointsConfig servicedef.SDKConfigServiceEndpointsParams) SDKConfigurer

WithServiceEndpointsConfig is used with StartSDKClient to specify non-default service endpoints. This will only work if the test service has the "service-endpoints" capability.

func WithStreamingConfig 

func WithStreamingConfig(streamingConfig servicedef.SDKConfigStreamingParams) SDKConfigurer

WithStreamingConfig is used with StartSDKClient to specify a non-default streaming configuration.

type SDKDataSource 

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

SDKDataSource is a test fixture that provides a callback endpoint for SDK clients to connect to, simulating the LaunchDarkly streaming or polling service.

func NewSDKDataSource 

func NewSDKDataSource(t *ldtest.T, data mockld.SDKData, options ...SDKDataSourceOption) *SDKDataSource

NewSDKDataSource creates a new SDKDataSource with the specified initial data set.

It can simulate either the streaming service or the polling service. If you don't explicitly specify DataSourceOptionPolling or DataSourceOptionStreaming, the default depends on what kind of SDK is being tested: server-side and mobile SDKs default to streaming, JS-based client-side SDKs default to polling.

It automatically detects (from the ldtest.T properties) whether we are testing a server-side, mobile, or JS-based client-side SDK, and configures the endpoint behavior as appropriate. The endpoints will enforce that the client only uses supported URL paths and HTTP methods; however, they do not do any validation of credentials (SDK key, mobile key, environment ID) since that would require this component to know more about the overall configuration than it knows. We have specific tests that do verify that the SDKs send appropriate credentials.

The object's lifecycle is tied to the test scope that created it; it will be automatically closed when this test scope exits. It can be reused by subtests until then. Debug output related to the data source will be attached to this test scope.

func NewSDKDataSourceWithoutEndpoint 

func NewSDKDataSourceWithoutEndpoint(t *ldtest.T, data mockld.SDKData, options ...SDKDataSourceOption) *SDKDataSource

NewSDKDataSourceWithoutEndpoint is the same as NewSDKDataSource, but it does not allocate an endpoint to accept incoming requests. Use this if you want to configure the endpoint separately, for instance if you want it to delegate some requests to the data source but return an error for some other requests.

func (*SDKDataSource) Configure 

func (d *SDKDataSource) Configure(config *servicedef.SDKConfigParams) error

Configure updates the SDK client configuration for NewSDKClient, causing the SDK to connect to the appropriate base URI for the data source test fixture. This only works if the data source was created along with its own endpoint, with NewSDKDataSource; if it was created as a handler to be used in a separately configured endpoint, you have to set the base URI in the test logic rather than using this shortcut.

func (*SDKDataSource) Endpoint 

func (d *SDKDataSource) Endpoint() *harness.MockEndpoint

Endpoint returns the low-level object that manages incoming requests.

func (*SDKDataSource) Handler 

func (d *SDKDataSource) Handler() http.Handler

Handler returns the HTTP handler for the service. Since StreamingService implements http.Handler already, this is the same as Service() but makes the purpose clearer.

func (*SDKDataSource) PollingService 

func (d *SDKDataSource) PollingService() *mockld.PollingService

PollingService returns the low-level object that manages the polling data, or nil if this is a streaming data source.

func (*SDKDataSource) StreamingService 

func (d *SDKDataSource) StreamingService() *mockld.StreamingService

StreamingService returns the low-level object that manages the stream data, or nil if this is a polling data source.

type SDKDataSourceOption 

type SDKDataSourceOption helpers.ConfigOption[sdkDataSourceConfig]

SDKDataSourceOption is the interface for options to NewSDKDataSource.

func DataSourceOptionPolling 

func DataSourceOptionPolling() SDKDataSourceOption

DataSourceOptionPolling makes an SDKDataSource simulate the polling service.

func DataSourceOptionStreaming 

func DataSourceOptionStreaming() SDKDataSourceOption

DataSourceOptionStreaming makes an SDKDataSource simulate the streaming service.

type SDKEventSink 

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

SDKEventSink is a test fixture that provides a callback endpoint for SDK clients to send event data to, simulating the LaunchDarkly event-recorder service.

func NewSDKEventSink 

func NewSDKEventSink(t *ldtest.T) *SDKEventSink

NewSDKEventSink creates a new SDKEventSink.

It automatically detects (from the ldtest.T properties) whether we are testing a server-side, mobile, or JS-based client-side SDK, and configures the endpoint behavior as appropriate. The endpoints will enforce that the client only uses supported URL paths and HTTP methods; however, they do not do any validation of credentials (SDK key, mobile key, environment ID) since that would require this component to know more about the overall configuration than it knows. We have specific tests that do verify that the SDKs send appropriate credentials.

The object's lifecycle is tied to the test scope that created it; it will be automatically closed when this test scope exits. It can be reused by subtests until then. Debug output related to the event sink will be attached to this test scope, and also to any of its subtests that are active when the output is generated.

func (*SDKEventSink) Configure 

func (e *SDKEventSink) Configure(config *servicedef.SDKConfigParams) error

Configure updates the SDK client configuration for NewSDKClient, causing the SDK to connect to the appropriate base URI for the test fixture.

func (*SDKEventSink) Endpoint 

func (e *SDKEventSink) Endpoint() *harness.MockEndpoint

Endpoint returns the low-level object that manages incoming requests.

func (*SDKEventSink) ExpectAnalyticsEvents 

func (e *SDKEventSink) ExpectAnalyticsEvents(t require.TestingT, timeout time.Duration) mockld.Events

ExpectAnalyticsEvents waits for event data to be posted to the endpoint, and then calls matchers.ItemsInAnyOrder with the specified eventMatchers, verifying that the payload contains one event matching each of the matchers regardless of ordering.

If no new events arrive before the timeout, the test immediately fails and terminates.

The number of events posted must be the same as the number of matchers.

func (*SDKEventSink) ExpectNoAnalyticsEvents 

func (e *SDKEventSink) ExpectNoAnalyticsEvents(t require.TestingT, timeout time.Duration)

ExpectNoAnalyticsEvents waits for the specified timeout and fails if any events are posted before then.

func (*SDKEventSink) Service 

func (e *SDKEventSink) Service() *mockld.EventsService

Service returns the underlying mock events service component, for access to special options.

type SDKTestContext 

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

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.