utils

package module
v0.0.0-...-fc8c377 Latest Latest
Warning

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

Go to latest
Published: Jan 21, 2016 License: LGPL-3.0 Imports: 32 Imported by: 0

README

juju/utils

This package provides general utility packages and functions.

Documentation

Index

Examples

Constants

View Source
const (
	// VerifySSLHostnames ensures we verify the hostname on the certificate
	// matches the host we are connecting and is signed
	VerifySSLHostnames = SSLHostnameVerification(true)
	// NoVerifySSLHostnames informs us to skip verifying the hostname
	// matches a valid certificate
	NoVerifySSLHostnames = SSLHostnameVerification(false)
)
View Source
const (
	OSWindows   = "windows"
	OSDarwin    = "darwin"
	OSDragonfly = "dragonfly"
	OSFreebsd   = "freebsd"
	OSLinux     = "linux"
	OSNacl      = "nacl"
	OSNetbsd    = "netbsd"
	OSOpenbsd   = "openbsd"
	OSSolaris   = "solaris"
)

These are the names of the operating systems recognized by Go.

View Source
const (
	NoSuchUserErrRegexp = `user: unknown user [a-z0-9_-]*`
	NoSuchFileErrRegexp = `no such file or directory`
	MkdirFailErrRegexp  = `.* not a directory`
)

The following are strings/regex-es which match common Unix error messages that may be returned in case of failed calls to the system. Any extra leading/trailing regex-es are left to be added by the developer.

Variables

View Source
var (
	LowerAlpha = []rune("abcdefghijklmnopqrstuvwxyz")
	UpperAlpha = []rune("ABCDEFGHIJKLMNOPQRSTUVWXYZ")
	Digits     = []rune("0123456789")
)

Can be used as a sane default argument for RandomString

View Source
var CompatSalt = string([]byte{0x75, 0x82, 0x81, 0xca})

CompatSalt is because Juju 1.16 and older used a hard-coded salt to compute the password hash for all users and agents

View Source
var FastInsecureHash = false

FastInsecureHash specifies whether a fast, insecure version of the hash algorithm will be used. Changing this will cause PasswordHash to produce incompatible passwords. It should only be changed for testing purposes - to make tests run faster.

View Source
var MinAgentPasswordLength = base64.StdEncoding.EncodedLen(randomPasswordBytes)

MinAgentPasswordLength describes how long agent passwords should be. We require this length because we assume enough entropy in the Agent password that it is safe to not do extra rounds of iterated hashing.

OSUnix is the list of unix-like operating systems recognized by Go. See http://golang.org/src/path/filepath/path_unix.go.

View Source
var OutgoingAccessAllowed = true

OutgoingAccessAllowed determines whether connections other than localhost can be dialled.

View Source
var (
	UUIDSnippet = block1 + "-" + block2 + "-" + block3 + "-" + block4 + "-" + block5
)

regex for validating that the UUID matches RFC 4122. This package generates version 4 UUIDs but accepts any UUID version. http://www.ietf.org/rfc/rfc4122.txt

Functions

func AgentPasswordHash

func AgentPasswordHash(password string) string

AgentPasswordHash returns base64-encoded one-way hash of password. This is not suitable for User passwords because those will have limited entropy (see UserPasswordHash). However, since we generate long random passwords for agents, we can trust that there is sufficient entropy to prevent brute force search. And using a faster hash allows us to restart the state machines and have 1000s of agents log in in a reasonable amount of time.

func AtomicWriteFile

func AtomicWriteFile(filename string, contents []byte, perms os.FileMode) (err error)

AtomicWriteFile atomically writes the filename with the given contents and permissions, replacing any existing file at the same path.

func AtomicWriteFileAndChange

func AtomicWriteFileAndChange(filename string, contents []byte, change func(*os.File) error) (err error)

AtomicWriteFileAndChange atomically writes the filename with the given contents and calls the given function after the contents were written, but before the file is renamed.

func BasicAuthHeader

func BasicAuthHeader(username, password string) http.Header

BasicAuthHeader creates a header that contains just the "Authorization" entry. The implementation was originally taked from net/http but this is needed externally from the http request object in order to use this with our websockets. See 2 (end of page 4) http://www.ietf.org/rfc/rfc2617.txt "To receive authorization, the client sends the userid and password, separated by a single colon (":") character, within a base64 encoded string in the credentials."

func ChownPath

func ChownPath(path, username string) error

ChownPath sets the uid and gid of path to match that of the user specified.

func CommandString

func CommandString(args ...string) string

CommandString flattens a sequence of command arguments into a string suitable for executing in a shell, escaping slashes, variables and quotes as necessary; each argument is double-quoted if and only if necessary.

func CopyFile

func CopyFile(dest, source string) error

CopyFile writes the contents of the given source file to dest.

func EnsureBaseDir

func EnsureBaseDir(baseDir, path string) string

EnsureBaseDir ensures that path is always prefixed by baseDir, allowing for the fact that path might have a Window drive letter in it.

func EnvUsername

func EnvUsername() (string, error)

EnvUsername returns the username from the OS environment.

func GetAddressForInterface

func GetAddressForInterface(interfaceName string) (string, error)

GetAddressForInterface looks for the network interface and returns the IPv4 address from the possible addresses.

func GetHTTPClient

func GetHTTPClient(verify SSLHostnameVerification) *http.Client

GetHTTPClient returns either a standard http client or non validating client depending on the value of verify.

func GetIPv4Address

func GetIPv4Address(addresses []net.Addr) (string, error)

GetIPv4Address iterates through the addresses expecting the format from func (ifi *net.Interface) Addrs() ([]net.Addr, error)

func GetNonValidatingHTTPClient

func GetNonValidatingHTTPClient() *http.Client

GetNonValidatingHTTPClient returns a new http.Client that does not verify the server's certificate chain and hostname.

func GetValidatingHTTPClient

func GetValidatingHTTPClient() *http.Client

GetValidatingHTTPClient returns a new http.Client that verifies the server's certificate chain and hostname.

func Gunzip

func Gunzip(data []byte) ([]byte, error)

Gunzip uncompresses the given data.

func Gzip

func Gzip(data []byte) []byte

Gzip compresses the given data.

func Home

func Home() string

Home returns the os-specific home path as specified in the environment.

func IsUbuntu

func IsUbuntu() bool

IsUbuntu executes lxb_release to see if the host OS is Ubuntu.

func IsValidUUIDString

func IsValidUUIDString(s string) bool

IsValidUUIDString returns true, if the given string matches a valid UUID (version 4, variant 2).

func JoinServerPath

func JoinServerPath(elem ...string) string

JoinServerPath joins any number of path elements into a single path, adding a path separator (based on the current juju server OS) if necessary. The result is Cleaned; in particular, all empty strings are ignored.

func LocalUsername

func LocalUsername() (string, error)

LocalUsername determines the current username on the local host.

func MakeFileURL

func MakeFileURL(in string) string

MakeFileURL returns a file URL if a directory is passed in else it does nothing

func MoveFile

func MoveFile(source, destination string) (bool, error)

MoveFile atomically moves the source file to the destination, returning whether the file was moved successfully. If the destination already exists, it returns an error rather than overwrite it.

On unix systems, an error may occur with a successful move, if the source file location cannot be unlinked.

func NewHttpTLSTransport

func NewHttpTLSTransport(tlsConfig *tls.Config) *http.Transport

NewHttpTLSTransport returns a new http.Transport constructed with the TLS config and the necessary parameters for Juju.

func NormalizePath

func NormalizePath(dir string) (string, error)

NormalizePath expands a path containing ~ to its absolute form, and removes any .. or . path elements.

func OSIsUnix

func OSIsUnix(os string) bool

OSIsUnix determines whether or not the given OS name is one of the unix-like operating systems recognized by Go.

func OSUsername

func OSUsername() (string, error)

OSUsername returns the username of the current OS user (based on UID).

func ParseBasicAuthHeader

func ParseBasicAuthHeader(h http.Header) (userid, password string, err error)

ParseBasicAuth attempts to find an Authorization header in the supplied http.Header and if found parses it as a Basic header. See 2 (end of page 4) http://www.ietf.org/rfc/rfc2617.txt "To receive authorization, the client sends the userid and password, separated by a single colon (":") character, within a base64 encoded string in the credentials."

func ParseSize

func ParseSize(str string) (MB uint64, err error)

ParseSize parses the string as a size, in mebibytes.

The string must be a is a non-negative number with an optional multiplier suffix (M, G, T, P, E, Z, or Y). If the suffix is not specified, "M" is implied.

func RandomBytes

func RandomBytes(n int) ([]byte, error)

RandomBytes returns n random bytes.

func RandomPassword

func RandomPassword() (string, error)

RandomPassword generates a random base64-encoded password.

func RandomSalt

func RandomSalt() (string, error)

RandomSalt generates a random base64 data suitable for using as a password salt The pbkdf2 guideline is to use 8 bytes of salt, so we do 12 raw bytes into 16 base64 bytes. (The alternative is 6 raw into 8 base64).

func RandomString

func RandomString(n int, validRunes []rune) string

RandomString will return a string of length n that will only contain runes inside validRunes

func ReadFileSHA256

func ReadFileSHA256(filename string) (string, int64, error)

ReadFileSHA256 is like ReadSHA256 but reads the contents of the given file.

func ReadSHA256

func ReadSHA256(source io.Reader) (string, int64, error)

ReadSHA256 returns the SHA256 hash of the contents read from source (hex encoded) and the size of the source in bytes.

func ReadYaml

func ReadYaml(path string, obj interface{}) error

ReadYaml unmarshals the yaml contained in the file at path into obj. See goyaml.Unmarshal. If path is not found, the error returned will be compatible with os.IsNotExist.

func RelativeURLPath

func RelativeURLPath(basePath, targPath string) (string, error)

RelativeURLPath returns a relative URL path that is lexically equivalent to targpath when interpreted by url.URL.ResolveReference. On success, the returned path will always be non-empty and relative to basePath, even if basePath and targPath share no elements.

It is assumed that both basePath and targPath are normalized (have no . or .. elements).

An error is returned if basePath or targPath are not absolute paths.

func ReplaceFile

func ReplaceFile(source, destination string) error

ReplaceFile atomically replaces the destination file or directory with the source. The errors that are returned are identical to those returned by os.Rename.

func ResolveSudo

func ResolveSudo(username string) string

ResolveSudo returns the original username if sudo was used. The original username is extracted from the OS environment.

func ResolveUsername

func ResolveUsername(resolveSudo func(string) string, usernameFuncs ...func() (string, error)) (string, error)

ResolveUsername returns the username determined by the provided functions. The functions are tried in the same order in which they were passed in. An error returned from any of them is immediately returned. If an empty string is returned then that signals that the function did not find the username and the next function is tried. Once a username is found, the provided resolveSudo func (if any) is called with that username and the result is returned. If no username is found then errors.NotFound is returned.

func RunCommand

func RunCommand(command string, args ...string) (output string, err error)

RunCommand executes the command and return the combined output.

func SetHome

func SetHome(s string) error

SetHome sets the os-specific home path in the environment.

func ShQuote

func ShQuote(s string) string

ShQuote quotes s so that when read by bash, no metacharacters within s will be interpreted as such.

func Timeit

func Timeit(action string) func()

Start a timer, used for tracking time spent. Generally used with either defer, as in:

defer utils.Timeit("my func")()

Which will track how much time is spent in your function. Or if you want to track the time spent in a function you are calling then you would use:

toc := utils.Timeit("anotherFunc()")
anotherFunc()
toc()

This tracks nested calls by indenting the output, and will print out the full stack of timing when we reach the top of the stack.

func UniqueDirectory

func UniqueDirectory(path, name string) (string, error)

UniqueDirectory returns "path/name" if that directory doesn't exist. If it does, the method starts appending .1, .2, etc until a unique name is found.

func UseMultipleCPUs

func UseMultipleCPUs()

UseMultipleCPUs sets GOMAXPROCS to the number of CPU cores unless it has already been overridden by the GOMAXPROCS environment variable.

func UserHomeDir

func UserHomeDir(userName string) (hDir string, err error)

UserHomeDir returns the home directory for the specified user, or the home directory for the current user if the specified user is empty.

func UserPasswordHash

func UserPasswordHash(password string, salt string) string

UserPasswordHash returns base64-encoded one-way hash password that is computationally hard to crack by iterating through possible passwords.

func WinCmdQuote

func WinCmdQuote(s string) string

WinCmdQuote quotes s so that when read by cmd.exe, no metacharacters within s will be interpreted as such.

func WinPSQuote

func WinPSQuote(s string) string

WinPSQuote quotes s so that when read by powershell, no metacharacters within s will be interpreted as such.

func WriteYaml

func WriteYaml(path string, obj interface{}) error

WriteYaml marshals obj as yaml to a temporary file in the same directory as path, than atomically replaces path with the temporary file.

Types

type Attempt

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

func (*Attempt) HasNext

func (a *Attempt) HasNext() bool

HasNext returns whether another attempt will be made if the current one fails. If it returns true, the following call to Next is guaranteed to return true.

Example
package main

import (
	"time"

	gc "gopkg.in/check.v1"

	"github.com/juju/utils"
)

func doSomething() (int, error) { return 0, nil }

func shouldRetry(error) bool { return false }

func doSomethingWith(int) {}

func main() {
	// This example shows how Attempt.HasNext can be used to help
	// structure an attempt loop. If the godoc example code allowed
	// us to make the example return an error, we would uncomment
	// the commented return statements.
	attempts := utils.AttemptStrategy{
		Total: 1 * time.Second,
		Delay: 250 * time.Millisecond,
	}
	for attempt := attempts.Start(); attempt.Next(); {
		x, err := doSomething()
		if shouldRetry(err) && attempt.HasNext() {
			continue
		}
		if err != nil {
			// return err
			return
		}
		doSomethingWith(x)
	}
	// return ErrTimedOut
	return
}

func (*utilsSuite) TestAttemptTiming(c *gc.C) {
	testAttempt := utils.AttemptStrategy{
		Total: 0.25e9,
		Delay: 0.1e9,
	}
	want := []time.Duration{0, 0.1e9, 0.2e9, 0.2e9}
	got := make([]time.Duration, 0, len(want)) // avoid allocation when testing timing
	t0 := time.Now()
	for a := testAttempt.Start(); a.Next(); {
		got = append(got, time.Now().Sub(t0))
	}
	got = append(got, time.Now().Sub(t0))
	c.Assert(got, gc.HasLen, len(want))
	const margin = 0.01e9
	for i, got := range want {
		lo := want[i] - margin
		hi := want[i] + margin
		if got < lo || got > hi {
			c.Errorf("attempt %d want %g got %g", i, want[i].Seconds(), got.Seconds())
		}
	}
}

func (*utilsSuite) TestAttemptNextHasNext(c *gc.C) {
	a := utils.AttemptStrategy{}.Start()
	c.Assert(a.Next(), gc.Equals, true)
	c.Assert(a.Next(), gc.Equals, false)

	a = utils.AttemptStrategy{}.Start()
	c.Assert(a.Next(), gc.Equals, true)
	c.Assert(a.HasNext(), gc.Equals, false)
	c.Assert(a.Next(), gc.Equals, false)

	a = utils.AttemptStrategy{Total: 2e8}.Start()
	c.Assert(a.Next(), gc.Equals, true)
	c.Assert(a.HasNext(), gc.Equals, true)
	time.Sleep(2e8)
	c.Assert(a.HasNext(), gc.Equals, true)
	c.Assert(a.Next(), gc.Equals, true)
	c.Assert(a.Next(), gc.Equals, false)

	a = utils.AttemptStrategy{Total: 1e8, Min: 2}.Start()
	time.Sleep(1e8)
	c.Assert(a.Next(), gc.Equals, true)
	c.Assert(a.HasNext(), gc.Equals, true)
	c.Assert(a.Next(), gc.Equals, true)
	c.Assert(a.HasNext(), gc.Equals, false)
	c.Assert(a.Next(), gc.Equals, false)
}
Output:

func (*Attempt) Next

func (a *Attempt) Next() bool

Next waits until it is time to perform the next attempt or returns false if it is time to stop trying. It always returns true the first time it is called - we are guaranteed to make at least one attempt.

type AttemptStrategy

type AttemptStrategy struct {
	Total time.Duration // total duration of attempt.
	Delay time.Duration // interval between each try in the burst.
	Min   int           // minimum number of retries; overrides Total
}

AttemptStrategy represents a strategy for waiting for an action to complete successfully.

func (AttemptStrategy) Start

func (s AttemptStrategy) Start() *Attempt

Start begins a new sequence of attempts for the given strategy.

type BackoffTimer

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

BackoffTimer implements Countdown. A backoff timer starts at min and gets multiplied by factor until it reaches max. Jitter determines whether a small randomization is added to the duration.

func NewBackoffTimer

func NewBackoffTimer(config BackoffTimerConfig) *BackoffTimer

NewBackoffTimer creates and initializes a new BackoffTimer A backoff timer starts at min and gets multiplied by factor until it reaches max. Jitter determines whether a small randomization is added to the duration.

func (*BackoffTimer) Reset

func (t *BackoffTimer) Reset()

Reset implements the Timer interface.

func (*BackoffTimer) Start

func (t *BackoffTimer) Start()

Start implements the Timer interface. Any existing timer execution is stopped before a new one is created.

type BackoffTimerConfig

type BackoffTimerConfig struct {
	// The minimum duration after which Func is called.
	Min time.Duration

	// The maximum duration after which Func is called.
	Max time.Duration

	// Determines whether a small randomization is applied to
	// the duration.
	Jitter bool

	// The factor by which you want the duration to increase
	// every time.
	Factor int64

	// Func is the function that will be called when the countdown reaches 0.
	Func func()

	// Clock provides the AfterFunc function used to call func.
	// It is exposed here so it's easier to mock it in tests.
	Clock clock.Clock
}

BackoffTimerConfig is a helper struct for backoff timer that encapsulates config information.

type Countdown

type Countdown interface {
	// Reset stops the timer and resets its duration to the minimum one.
	// Start must be called to start the timer again.
	Reset()

	// Start starts the internal timer.
	// At the end of the timer, if Reset hasn't been called in the mean time
	// Func will be called and the duration is increased for the next call.
	Start()
}

Countdown implements a timer that will call a provided function. after a internally stored duration. The steps as well as min and max durations are declared upon initialization and depend on the particular implementation.

type Limiter

type Limiter interface {
	// Acquire another unit of the resource.
	// Acquire returns false to indicate there is no more availability,
	// until another entity calls Release.
	Acquire() bool
	// AcquireWait requests a unit of resource, but blocks until one is
	// available.
	AcquireWait()
	// Release returns a unit of the resource. Calling Release when there
	// are no units Acquired is an error.
	Release() error
}

Limiter represents a limited resource (eg a semaphore).

func NewLimiter

func NewLimiter(max int) Limiter

type SSLHostnameVerification

type SSLHostnameVerification bool

SSLHostnameVerification is used as a switch for when a given provider might use self-signed credentials and we should not try to verify the hostname on the TLS/SSL certificates

type SizeTracker

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

SizeTracker tracks the number of bytes passing through its Write method (which is otherwise a no-op).

Use SizeTracker with io.MultiWriter() to track number of bytes written. Use with io.TeeReader() to track number of bytes read.

func (SizeTracker) Size

func (st SizeTracker) Size() int64

Size returns the number of bytes written so far.

func (*SizeTracker) Write

func (st *SizeTracker) Write(data []byte) (n int, err error)

Write implements io.Writer.

type UUID

type UUID [16]byte

UUID represent a universal identifier with 16 octets.

func MustNewUUID

func MustNewUUID() UUID

MustNewUUID returns a new uuid, if an error occurs it panics.

func NewUUID

func NewUUID() (UUID, error)

NewUUID generates a new version 4 UUID relying only on random numbers.

func UUIDFromString

func UUIDFromString(s string) (UUID, error)

func (UUID) Copy

func (uuid UUID) Copy() UUID

Copy returns a copy of the UUID.

func (UUID) Raw

func (uuid UUID) Raw() [16]byte

Raw returns a copy of the UUID bytes.

func (UUID) String

func (uuid UUID) String() string

String returns a hexadecimal string representation with standardized separators.

Directories

Path Synopsis
Package bzr offers an interface to manage branches of the Bazaar VCS.
Package bzr offers an interface to manage branches of the Bazaar VCS.
Package cache provides a simple caching mechanism that limits the age of cache entries and tries to avoid large repopulation events by staggering refresh times.
Package cache provides a simple caching mechanism that limits the age of cache entries and tries to avoid large repopulation events by staggering refresh times.
Package debugstatus provides facilities for inspecting information about a running HTTP service.
Package debugstatus provides facilities for inspecting information about a running HTTP service.
The deque package implements an efficient double-ended queue data structure called Deque.
The deque package implements an efficient double-ended queue data structure called Deque.
The featureflag package gives other parts of Juju the ability to easily check to see if a feature flag has been defined.
The featureflag package gives other parts of Juju the ability to easily check to see if a feature flag has been defined.
utils/filestorage provides types for abstracting and implementing a system that stores files, including their metadata.
utils/filestorage provides types for abstracting and implementing a system that stores files, including their metadata.
Package fslock provides an on-disk mutex protecting a resource
Package fslock provides an on-disk mutex protecting a resource
The hash package provides utilities that support use of the stdlib hash.Hash.
The hash package provides utilities that support use of the stdlib hash.Hash.
The keyvalues package implements a set of functions for parsing key=value data, usually passed in as command-line parameters to juju subcommands, e.g.
The keyvalues package implements a set of functions for parsing key=value data, usually passed in as command-line parameters to juju subcommands, e.g.
Package os provides access to operating system related configuration.
Package os provides access to operating system related configuration.
Package packaging abstracts away differences between package managers like apt and yum and allows for easy extension for other package managers/distros.
Package packaging abstracts away differences between package managers like apt and yum and allows for easy extension for other package managers/distros.
commands
Package commands contains an interface which returns common package-manager related commands and the reference implementation for apt and yum-based systems.
Package commands contains an interface which returns common package-manager related commands and the reference implementation for apt and yum-based systems.
config
The config package defines an interface which returns packaging-related configuration options and operations depending on the desired package-management system.
The config package defines an interface which returns packaging-related configuration options and operations depending on the desired package-management system.
manager
The manager package defines an interface which can carry out numerous package-management related operations on the local system and the respective implementations on apt and yum-based systems.
The manager package defines an interface which can carry out numerous package-management related operations on the local system and the respective implementations on apt and yum-based systems.
manager/testing
This package contains a mock implementation of the manager.PackageManager interface which always returns positive outcomes and a nil error.
This package contains a mock implementation of the manager.PackageManager interface which always returns positive outcomes and a nil error.
The parallel package provides utilities for running tasks concurrently.
The parallel package provides utilities for running tasks concurrently.
series provides helpers for determining the series of a host, and translating from os to series.
series provides helpers for determining the series of a host, and translating from os to series.
ssh
Package ssh contains utilities for dealing with SSH connections, key management, and so on.
Package ssh contains utilities for dealing with SSH connections, key management, and so on.
This package provides convenience helpers on top of archive/tar to be able to tar/untar files with a functionality closer to gnu tar command.
This package provides convenience helpers on top of archive/tar to be able to tar/untar files with a functionality closer to gnu tar command.
Package voyeur implements a concurrency-safe value that can be watched for changes.
Package voyeur implements a concurrency-safe value that can be watched for changes.

Jump to

Keyboard shortcuts

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