util

package
v0.0.0-...-2a696aa Latest Latest
Warning

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

Go to latest
Published: Oct 14, 2016 License: Apache-2.0, Apache-2.0 Imports: 38 Imported by: 0

Documentation

Overview

Package util implements various utility functions used in both testing and implementation of Kubernetes. Package util may not depend on any other package in the Kubernetes package tree.

Index

Constants

This section is empty.

Variables

This section is empty.

Functions

func AllPtrFieldsNil

func AllPtrFieldsNil(obj interface{}) bool

Tests whether all pointer fields in a struct are nil. This is useful when, for example, an API struct is handled by plugins which need to distinguish "no plugin accepted this spec" from "this spec is empty".

This function is only valid for structs and pointers to structs. Any other type will cause a panic. Passing a typed nil pointer will return true.

func ApplyRLimitForSelf

func ApplyRLimitForSelf(maxOpenFiles uint64)

func CertPoolFromFile

func CertPoolFromFile(filename string) (*x509.CertPool, error)

CertPoolFromFile returns an x509.CertPool containing the certificates in the given PEM-encoded file. Returns an error if the file could not be read, a certificate could not be parsed, or if the file does not contain any certificates

func CertsFromPEM

func CertsFromPEM(pemCerts []byte) ([]*x509.Certificate, error)

CertsFromPEM returns the x509.Certificates contained in the given PEM-encoded byte array Returns an error if a certificate could not be parsed, or if the data does not contain any certificates

func CompileRegexps

func CompileRegexps(regexpStrings []string) ([]*regexp.Regexp, error)

Takes a list of strings and compiles them into a list of regular expressions

func DeadlockWatchdog

func DeadlockWatchdog(lock sync.Locker, name string, maxLockPeriod time.Duration)

func DeadlockWatchdogReadLock

func DeadlockWatchdogReadLock(lock *sync.RWMutex, name string, maxLockPeriod time.Duration)

DeadlockWatchdogReadLock creates a watchdog on read/write mutex. If the mutex can not be acquired for read access within 'maxLockPeriod', the program exits via glog.Exitf() or os.Exit() if that fails 'name' is a semantic name that is useful for the user and is printed on exit.

func ExecuteTemplate

func ExecuteTemplate(w io.Writer, templateText string, data interface{}) error

ExecuteTemplate executes templateText with data and output written to w.

func ExecuteTemplateToString

func ExecuteTemplateToString(templateText string, data interface{}) (string, error)

func FileExists

func FileExists(filename string) (bool, error)

func FlushLogs

func FlushLogs()

FlushLogs flushes logs immediately.

func GenerateSelfSignedCert

func GenerateSelfSignedCert(host, certPath, keyPath string, alternateIPs []net.IP, alternateDNS []string) error

GenerateSelfSignedCert creates a self-signed certificate and key for the given host. Host may be an IP or a DNS name You may also specify additional subject alt names (either ip or dns names) for the certificate The certificate will be created with file mode 0644. The key will be created with file mode 0600. If the certificate or key files already exist, they will be overwritten. Any parent directories of the certPath or keyPath will be created as needed with file mode 0755.

func GetEnvAsFloat64OrFallback

func GetEnvAsFloat64OrFallback(key string, defaultValue float64) (float64, error)

func GetEnvAsIntOrFallback

func GetEnvAsIntOrFallback(key string, defaultValue int) (int, error)

func GetEnvAsStringOrFallback

func GetEnvAsStringOrFallback(key, defaultValue string) string

func InitFlags

func InitFlags()

InitFlags normalizes and parses the command line flags

func InitLogs

func InitLogs()

InitLogs initializes logs the way we want for kubernetes.

func IntPtr

func IntPtr(i int) *int

IntPtr returns a pointer to an int

func IntPtrDerefOr

func IntPtrDerefOr(ptr *int, def int) int

IntPtrDerefOr dereference the int ptr and returns it i not nil, else returns def.

func NewCRLFWriter

func NewCRLFWriter(w io.Writer) io.Writer

NewCRLFWriter implements a CR/LF line ending writer used for normalizing text for Windows platforms.

func NewLogger

func NewLogger(prefix string) *log.Logger

NewLogger creates a new log.Logger which sends logs to glog.Info.

func NewUUID

func NewUUID() types.UID

func ObjectDiff

func ObjectDiff(a, b interface{}) string

ObjectDiff writes the two objects out as JSON and prints out the identical part of the objects followed by the remaining part of 'a' and finally the remaining part of 'b'. For debugging tests.

func ObjectGoPrintDiff

func ObjectGoPrintDiff(a, b interface{}) string

ObjectGoPrintDiff is like ObjectDiff, but uses go-spew to print the objects, which shows absolutely everything by recursing into every single pointer (go's %#v formatters OTOH stop at a certain point). This is needed when you can't figure out why reflect.DeepEqual is returning false and nothing is showing you differences. This will.

func ObjectGoPrintSideBySide

func ObjectGoPrintSideBySide(a, b interface{}) string

ObjectGoPrintSideBySide prints a and b as textual dumps side by side, enabling easy visual scanning for mismatches.

func ReadDirNoExit

func ReadDirNoExit(dirname string) ([]os.FileInfo, []error, error)

borrowed from ioutil.ReadDir ReadDir reads the directory named by dirname and returns a list of directory entries, minus those with lstat errors

func RunInResourceContainer

func RunInResourceContainer(containerName string) error

Creates resource-only containerName if it does not already exist and moves the current process to it.

containerName must be an absolute container name.

func StringDiff

func StringDiff(a, b string) string

StringDiff diffs a and b and returns a human readable diff.

func Umask

func Umask(mask int) (old int, err error)

func UsingSystemdInitSystem

func UsingSystemdInitSystem() bool

Detects if using systemd as the init system Please note that simply reading /proc/1/cmdline can be misleading because some installation of various init programs can automatically make /sbin/init a symlink or even a renamed version of their main program. TODO(dchen1107): realiably detects the init system using on the system: systemd, upstart, initd, etc.

func WarnWordSepNormalizeFunc

func WarnWordSepNormalizeFunc(f *pflag.FlagSet, name string) pflag.NormalizedName

WarnWordSepNormalizeFunc changes and warns for flags that contain "_" separators

func WordSepNormalizeFunc

func WordSepNormalizeFunc(f *pflag.FlagSet, name string) pflag.NormalizedName

WordSepNormalizeFunc changes all flags that contain "_" separators

Types

type Backoff

type Backoff struct {
	sync.Mutex
	Clock Clock
	// contains filtered or unexported fields
}

func NewBackOff

func NewBackOff(initial, max time.Duration) *Backoff

func (*Backoff) GC

func (p *Backoff) GC()

Garbage collect records that have aged past maxDuration. Backoff users are expected to invoke this periodically.

func (*Backoff) Get

func (p *Backoff) Get(id string) time.Duration

Get the current backoff Duration

func (*Backoff) IsInBackOffSince

func (p *Backoff) IsInBackOffSince(id string, eventTime time.Time) bool

Returns True if the elapsed time since eventTime is smaller than the current backoff window

func (*Backoff) IsInBackOffSinceUpdate

func (p *Backoff) IsInBackOffSinceUpdate(id string, eventTime time.Time) bool

Returns True if time since lastupdate is less than the current backoff window.

func (*Backoff) Next

func (p *Backoff) Next(id string, eventTime time.Time)

move backoff to the next mark, capping at maxDuration

func (*Backoff) Reset

func (p *Backoff) Reset(id string)

Reset forces clearing of all backoff data for a given key.

type BoolFlag

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

BoolFlag is a boolean flag compatible with flags and pflags that keeps track of whether it had a value supplied or not. Getting this flag to act like a normal bool, where true/false are not required needs a little bit of extra code, example:

f := cmd.Flags().VarPF(&BoolFlagVar, "flagname", "", "help about the flag")
f.NoOptDefVal = "true"

func (*BoolFlag) Default

func (f *BoolFlag) Default(value bool)

func (BoolFlag) Provided

func (f BoolFlag) Provided() bool

func (*BoolFlag) Set

func (f *BoolFlag) Set(value string) error

func (BoolFlag) String

func (f BoolFlag) String() string

func (*BoolFlag) Type

func (f *BoolFlag) Type() string

func (BoolFlag) Value

func (f BoolFlag) Value() bool

type Cache

type Cache []*cacheShard

func NewCache

func NewCache(maxSize int) Cache

func (*Cache) Add

func (c *Cache) Add(index uint64, obj interface{}) bool

Returns true if object already existed, false otherwise.

func (*Cache) Get

func (c *Cache) Get(index uint64) (obj interface{}, found bool)

type Clock

type Clock interface {
	Now() time.Time
	Since(time.Time) time.Duration
	After(d time.Duration) <-chan time.Time
}

Clock allows for injecting fake or real clocks into code that needs to do arbitrary things based on time.

type ConfigurationMap

type ConfigurationMap map[string]string

func (*ConfigurationMap) Set

func (m *ConfigurationMap) Set(value string) error

func (*ConfigurationMap) String

func (m *ConfigurationMap) String() string

func (*ConfigurationMap) Type

func (*ConfigurationMap) Type() string

type FakeClock

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

FakeClock implements Clock, but returns an arbitrary time.

func NewFakeClock

func NewFakeClock(t time.Time) *FakeClock

func (*FakeClock) After

func (f *FakeClock) After(d time.Duration) <-chan time.Time

Fake version of time.After(d).

func (*FakeClock) HasWaiters

func (f *FakeClock) HasWaiters() bool

Returns true if After has been called on f but not yet satisfied (so you can write race-free tests).

func (*FakeClock) Now

func (f *FakeClock) Now() time.Time

Now returns f's time.

func (*FakeClock) SetTime

func (f *FakeClock) SetTime(t time.Time)

Sets the time.

func (*FakeClock) Since

func (f *FakeClock) Since(ts time.Time) time.Duration

Since returns time since the time in f.

func (*FakeClock) Step

func (f *FakeClock) Step(d time.Duration)

Move clock by Duration, notify anyone that's called After

type GlogWriter

type GlogWriter struct{}

GlogWriter serves as a bridge between the standard log package and the glog package.

func (GlogWriter) Write

func (writer GlogWriter) Write(data []byte) (n int, err error)

Write implements the io.Writer interface.

type IntervalClock

type IntervalClock struct {
	Time     time.Time
	Duration time.Duration
}

IntervalClock implements Clock, but each invocation of Now steps the clock forward the specified duration

func (*IntervalClock) After

func (*IntervalClock) After(d time.Duration) <-chan time.Time

Unimplemented, will panic. TODO: make interval clock use FakeClock so this can be implemented.

func (*IntervalClock) Now

func (i *IntervalClock) Now() time.Time

Now returns i's time.

func (*IntervalClock) Since

func (i *IntervalClock) Since(ts time.Time) time.Duration

Since returns time since the time in i.

type LineDelimiter

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

A Line Delimiter is a filter that will

func NewLineDelimiter

func NewLineDelimiter(output io.Writer, delimiter string) *LineDelimiter

NewLineDelimiter allocates a new io.Writer that will split input on lines and bracket each line with the delimiter string. This can be useful in output tests where it is difficult to see and test trailing whitespace.

func (*LineDelimiter) Flush

func (ld *LineDelimiter) Flush() (err error)

Flush all lines up until now. This will assume insert a linebreak at the current point of the stream.

func (*LineDelimiter) Write

func (ld *LineDelimiter) Write(buf []byte) (n int, err error)

Write writes buf to the LineDelimiter ld. The only errors returned are ones encountered while writing to the underlying output stream.

type RateLimiter

type RateLimiter interface {
	// TryAccept returns true if a token is taken immediately. Otherwise,
	// it returns false.
	TryAccept() bool
	// Accept returns once a token becomes available.
	Accept()
	// Stop stops the rate limiter, subsequent calls to CanAccept will return false
	Stop()
	// Saturation returns a percentage number which describes how saturated
	// this rate limiter is.
	// Usually we use token bucket rate limiter. In that case,
	// 1.0 means no tokens are available; 0.0 means we have a full bucket of tokens to use.
	Saturation() float64
}

func NewFakeAlwaysRateLimiter

func NewFakeAlwaysRateLimiter() RateLimiter

func NewFakeNeverRateLimiter

func NewFakeNeverRateLimiter() RateLimiter

func NewTokenBucketRateLimiter

func NewTokenBucketRateLimiter(qps float32, burst int) RateLimiter

NewTokenBucketRateLimiter creates a rate limiter which implements a token bucket approach. The rate limiter allows bursts of up to 'burst' to exceed the QPS, while still maintaining a smoothed qps rate of 'qps'. The bucket is initially filled with 'burst' tokens, and refills at a rate of 'qps'. The maximum number of tokens in the bucket is capped at 'burst'.

type RealClock

type RealClock struct{}

RealClock really calls time.Now()

func (RealClock) After

func (RealClock) After(d time.Duration) <-chan time.Time

Same as time.After(d).

func (RealClock) Now

func (RealClock) Now() time.Time

Now returns the current time.

func (RealClock) Since

func (RealClock) Since(ts time.Time) time.Duration

Since returns time since the specified timestamp.

type Runner

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

Runner is an abstraction to make it easy to start and stop groups of things that can be described by a single function which waits on a channel close to exit.

func NewRunner

func NewRunner(f ...func(stop chan struct{})) *Runner

NewRunner makes a runner for the given function(s). The function(s) should loop until the channel is closed.

func (*Runner) Start

func (r *Runner) Start()

Start begins running.

func (*Runner) Stop

func (r *Runner) Stop()

Stop stops running.

type StringFlag

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

StringFlag is a string flag compatible with flags and pflags that keeps track of whether it had a value supplied or not.

func NewStringFlag

func NewStringFlag(defaultVal string) StringFlag

func (*StringFlag) Default

func (f *StringFlag) Default(value string)

func (StringFlag) Provided

func (f StringFlag) Provided() bool

func (*StringFlag) Set

func (f *StringFlag) Set(value string) error

func (StringFlag) String

func (f StringFlag) String() string

func (*StringFlag) Type

func (f *StringFlag) Type() string

func (StringFlag) Value

func (f StringFlag) Value() string

type Trace

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

func NewTrace

func NewTrace(name string) *Trace

func (*Trace) Log

func (t *Trace) Log()

func (*Trace) LogIfLong

func (t *Trace) LogIfLong(threshold time.Duration)

func (*Trace) Step

func (t *Trace) Step(msg string)

func (*Trace) TotalTime

func (t *Trace) TotalTime() time.Duration

Directories

Path Synopsis
Package errors implements various utility functions and types around errors.
Package errors implements various utility functions and types around errors.
package jsonpath is a template engine using jsonpath syntax, which can be seen at http://goessner.net/articles/JsonPath/.
package jsonpath is a template engine using jsonpath syntax, which can be seen at http://goessner.net/articles/JsonPath/.
Package labels provides utilities to work with Kubernetes labels.
Package labels provides utilities to work with Kubernetes labels.
Package pod provides utilities to work with Kubernetes pod and pod templates.
Package pod provides utilities to work with Kubernetes pod and pod templates.
Package rand provides utilities related to randomization.
Package rand provides utilities related to randomization.
Package sets has auto-generated set types.
Package sets has auto-generated set types.
Package wait provides tools for polling or listening for changes to a condition.
Package wait provides tools for polling or listening for changes to a condition.
Package workqueue provides a simple queue that supports the following features: * Fair: items processed in the order in which they are added.
Package workqueue provides a simple queue that supports the following features: * Fair: items processed in the order in which they are added.

Jump to

Keyboard shortcuts

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