Skip to content
Permalink
master
Switch branches/tags

presentations/go/oreilly/2019/slides.md.erb

Go to file
 
 
Cannot retrieve contributors at this time
3400 lines (2610 sloc) 62.7 KB
Raw Blame
Open in GitHub Desktop
<%
def exercise(title:, short_url:, start_code:, cheat_sheet:, solution:, long_url: nil)
<<-EOD.chomp
## Exercise: #{title}
`#{short_url}`
<!-- #{long_url} -->
## Exercise: #{title}
``` go
#{start_code}
```
## Exercise: #{title} cheat sheet
#{cheat_sheet}
`#{short_url}`
## Exercise: #{title} solution
``` go
#{solution}
```
EOD
end
%>
# Introduction to the Go Programming Language
## About me
* Author, _Head First Ruby_ and _Head First Go_
* 4 years experience as online software development instructor
* See my recent courses at `https://teamtreehouse.com`
## Where to Learn Go
`https://tour.golang.org`
![](images/go_tour.png)
(We'll repeat that link at the end.)
## Another humble recommendation
![](images/head_first_go_cover.png)
`https://headfirstgo.com`
(We'll repeat that link at the end too. Visit it for more exercises!)
# Why Go?
## Go at a glance
* C-like syntax
* Compiles to native code
* Type-safe
* Garbage collected
* Concurrency built into language
## OK, but what can you do with Go?
## Docker
![](images/docker.png)
## Docker
* "'go build' will embed everything you need. (No more 'install this in order to run my stuff'.)"
* "Extensive standard library and data types."
* "Strong duck typing."
—Jérôme Petazzoni, "Docker and Go: why did we decide to write Docker in Go?"
::: notes
https://www.slideshare.net/jpetazzo/docker-and-go-why-did-we-decide-to-write-docker-in-go
:::
## Kubernetes
![](images/kubernetes.png)
## Kubernetes
* "Code in Go isn't overly complex. People don't create FactoryFactory objects."
* "Something with the feel of C with more advanced features like anonymous functions is a great combo."
* "Garbage Collection: We all know how to clean up after our selves but it is so nice to not have to worry about it."
—Joe Beda, "Kubernetes + Go = Crazy Delicious"
## Poll: What do you want to make with Go?
1. A system utility
1. A web app or service
1. Something else entirely
1. I don't know yet
# Go Tools
## "go fmt"
* Automatically fixes code style
* Acts as community's style guide
* No more arguing tabs vs. spaces!
## "go fmt"
Before
``` go
package main
import "fmt"
func main() {
repeatLine("hello", 3 )
}
func repeatLine( line string ,times int) {
for i := 0; i < times; i++ {
fmt.Println(line)
}
}
```
## "go fmt"
``` go
$ go fmt repeat.go
```
## "go fmt"
After
``` go
package main
import "fmt"
func main() {
repeatLine("hello", 3)
}
func repeatLine(line string, times int) {
for i := 0; i < times; i++ {
fmt.Println(line)
}
}
```
## "go run"
* Compiles a Go source file and runs it
* No executable is saved
``` go
$ go run repeat.go
hello
hello
hello
```
## "go build"
* Compiles Go source file(s) into an executable
``` go
$ go build repeat.go
$ ls -l
total 2064
-rwxr-xr-x 1 jay staff 2106512 May 1 21:13 repeat
-rw-r--r-- 1 jay staff 166 May 1 21:13 repeat.go
$ ./repeat
hello
hello
hello
```
## Playground
* You don't even have to install Go to try it!
* Edit and run Go code in your browser
![](images/playground.png)
## Exercises for this training
* The Go Playground lets you save your code and share it at a URL
* Others can edit that code
* We'll use that ability for most of our exercises
* We'll post a link
* You visit it and follow the instructions there (add code, fill in blanks, etc.)
# Syntax
## Go file layout
* Package clause
* Imports
* Code
``` go
package main
import "fmt"
func main() {
fmt.Println("Hello, Go!") // => Hello, Go!
}
```
## Calling Functions
``` go
package main
import "fmt"
func main() {
// No arguments
fmt.Println() // Prints a blank line
// 1 argument
fmt.Println("argument 1") // => argument 1
// 2 arguments
fmt.Println("argument 1", "argument 2") // => argument 1 argument 2
}
```
## Calling Functions
* `fmt.Println` can take any number and type of arguments.
* Most functions require a _specific_ number and type of arguments.
``` go
package main
import "math"
func main() {
math.Floor(3.1415) // valid...
math.Floor() // not enough arguments
math.Floor(3.1415, 12.34) // too many arguments
math.Floor("a string") // wrong type
}
```
## Imports
``` go
package main
func main() {
fmt.Println(math.Floor(2.75))
fmt.Println(strings.Title("head first go"))
}
```
Compile errors:
``` go
prog.go:4:2: undefined: fmt
prog.go:4:14: undefined: math
prog.go:5:2: undefined: fmt
prog.go:5:14: undefined: strings
```
## Imports
Need to add `import` statement:
``` go
package main
import (
"fmt"
"math"
"strings"
)
func main() {
fmt.Println(math.Floor(2.75)) // => 2
fmt.Println(strings.Title("head first go")) // => Head First Go
}
```
## Unused imports not allowed
``` go
package main
import (
"fmt"
"os"
)
func main() {
fmt.Println("Hello, Go!")
}
```
Compile error:
```
temp.go:5:5: imported and not used: "os"
```
## "goimports"
* Wrapper for `go fmt`
* Automatically adds/removes imports
Install:
``` go
$ go get golang.org/x/tools/cmd/goimports
```
Do a web search for "goimports" for directions on integrating with your editor.
## "goimports"
Before saving
``` go
package main
func main() {
fmt.Println(math.Floor(2.75))
fmt.Println(strings.Title("head first go"))
}
```
## "goimports"
After saving
``` go
package main
import (
"fmt"
"math"
"strings"
)
func main() {
fmt.Println(math.Floor(2.75))
fmt.Println(strings.Title("head first go"))
}
```
## Variables
``` go
var myInteger int
myInteger = 1
var myFloat float64
myFloat = 3.1415
fmt.Println(myInteger) // => 1
fmt.Println(myFloat) // => 3.1415
fmt.Println(reflect.TypeOf(myInteger)) // => int
fmt.Println(reflect.TypeOf(myFloat)) // => float64
```
## Short Variable Declarations
``` go
myInteger := 1
myFloat := 3.1415
fmt.Println(myInteger) // => 1
fmt.Println(myFloat) // => 3.1415
fmt.Println(reflect.TypeOf(myInteger)) // => int
fmt.Println(reflect.TypeOf(myFloat)) // => float64
```
## Must use every variable you declare
``` go
subtotal := 24.70
tax := 1.89
fmt.Println(subtotal)
```
Compile error:
``` go
prog.go:9:2: tax declared and not used
```
## Type conversions
``` go
var length float64 = 1.2
var width int = 2
// Can't assign an `int` value
// to a `float64` variable:
length = width
fmt.Println(length)
```
Compile error:
``` go
cannot use width (type int) as type float64 in assignment
```
## Type conversions
``` go
var length float64 = 1.2
var width int = 2
// But you can if you do a type
// conversion!
length = float64(width)
fmt.Println(length) // => 2
```
## Type conversions
``` go
var length float64 = 1.2
var width int = 2
// Can't do a math operation with a float64 and an int:
fmt.Println("Area is", length*width)
// Or a comparison:
fmt.Println("length > width?", length > width)
```
Compile errors:
``` go
invalid operation: length * width (mismatched types float64 and int)
invalid operation: length > width (mismatched types float64 and int)
```
## Type conversions
``` go
var length float64 = 1.2
var width int = 2
// But you can if you do type conversions!
fmt.Println("Area is", length*float64(width))
fmt.Println("length > width?", length > float64(width))
```
Output:
``` go
Area is 2.4
length > width? false
```
## "if"
``` go
if 1 < 2 {
fmt.Println("It's true!")
}
if 1 > 2 {
fmt.Println("I won't be printed!")
}
```
Output:
``` go
It's true!
```
## "if"
Parentheses discouraged. `go fmt` will remove these:
``` go
if (1 < 2) {
fmt.Println("It's true!")
}
```
Opening curly brace _must_ be on same line as `if`. This is a syntax error:
``` go
if (1 < 2)
{
fmt.Println("It's true!")
}
```
## "for"
``` go
for x := 4; x <= 6; x++ {
fmt.Println("x is now", x)
}
```
Output:
``` go
x is now 4
x is now 5
x is now 6
```
<%
@title = <<-'EOD'.chomp
Go syntax
EOD
@short_url = "https://is.gd/goex_syntax"
@long_url = "https://play.golang.org/p/x9BXw0z5LXT"
@start_code = <<-'EOD'.chomp
// Replace the blanks ("____") in the below code so that it
// compiles, runs, and prints the message "Hello, O'Reilly!".
____ main
____ "fmt"
____ main() {
myString ____ "Hello, O'Reilly!"
fmt.Println(____)
}
EOD
@cheat_sheet = <<-'EOD'.chomp
* Every Go source file is part of a __package__.
* To use code from other packages, you have to __import__ them.
* The `main` function is called when a program first starts.
* Functions are declared using the `func` keyword.
* A function call needs parentheses following the function name: `mypackage.MyFunction("my argument")`
EOD
@solution = <<-'EOD'.chomp
package main
import "fmt"
func main() {
myString := "Hello, O'Reilly!"
fmt.Println(myString)
}
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: nil) %>
# Declaring Functions
## Declaring functions
``` go
func sayHi() {
fmt.Println("Hi!")
}
func main() {
sayHi() // => Hi!
}
```
## Function names
* Use `CamelCase`: capitalize each word after the first.
* If the first letter of a function name is `Capitalized`, it's considered __exported__: it can be used from other packages.
* If the first letter of a function name is `uncapitalized`, it's considered __unexported__: it can only be used _within_ its package.
* This is why all the names of standard library functions we've been calling are capitalized. (E.g. `fmt.Println`, `math.Floor`, etc.)
More on exported/unexported later.
## Parameters
``` go
// In parentheses, list parameter name(s)
// followed by type(s).
func say(phrase string, times int) {
for i := 0; i < times; i++ {
fmt.Print(phrase)
}
fmt.Print("\n")
}
func main() {
// Provide argument(s) when calling.
say("Hi", 4) // => HiHiHiHi
say("Bye", 2) // => ByeBye
}
```
## Variable scope
Variable scope limited to function where it's declared.
``` go
func myFunction() {
myVariable := 10
}
func main() {
myFunction()
fmt.Println(myVariable) // out of scope!
}
```
Compile error:
```
prog.go:11:14: undefined: myVariable
```
## Function return values
So how do we get a value from a function to its caller? Add a return value!
``` go
// Add return value type after parentheses
func myFunction() int {
// Use "return" keyword
return 10
}
func main() {
// Assign returned value to variable
myVariable := myFunction()
fmt.Println(myVariable) // => 10
}
```
## Multiple return values
``` go
func main() {
flag := strconv.ParseBool("true")
flag = strconv.ParseBool("foobar")
fmt.Println(flag)
}
```
Compile error:
``` go
prog.go:9:7: assignment mismatch: 1 variable but strconv.ParseBool returns 2 values
prog.go:10:7: assignment mismatch: 1 variable but strconv.ParseBool returns 2 values
```
## Multiple return values
``` go
func main() {
flag, err := strconv.ParseBool("true")
if err != nil {
log.Fatal(err)
}
fmt.Println(flag) // => true
flag, err = strconv.ParseBool("foobar")
if err != nil {
log.Fatal(err)
// => 2009/11/10 23:00:00 strconv.ParseBool:
// => parsing "foobar": invalid syntax
}
fmt.Println(flag)
}
```
## Error handling
"In Go, error handling is important. The language's design and conventions encourage you to explicitly check for errors where they occur (as distinct from the convention in other languages of throwing exceptions and **sometimes** catching them)." (Emphasis mine)
-Andrew Gerrand, https://blog.golang.org/error-handling-and-go
## Writing functions with multiple return values
``` go
func myParseBool(myString string) (bool, error) {
if myString == "true" {
return true, nil
} else if myString == "false" {
return false, nil
} else {
return false, fmt.Errorf("bad string %s", myString)
}
}
func main() {
trueFalse, err := myParseBool("false")
if err != nil {
log.Fatal(err)
}
fmt.Println(trueFalse) // => false
}
```
## Writing functions with multiple return values
``` go
func myParseBool(myString string) (bool, error) {
if myString == "true" {
return true, nil
} else if myString == "false" {
return false, nil
} else {
return false, fmt.Errorf("bad string %s", myString)
}
}
func main() {
trueFalse, err := myParseBool("foobar")
if err != nil {
log.Fatal(err) // => 2020-03-13 19:34:00 bad string foobar
}
fmt.Println(trueFalse)
}
```
<%
@title = <<-'EOD'.chomp
Declaring functions
EOD
@short_url = "https://is.gd/goex_define_functions"
@long_url = "https://play.golang.org/p/Jih76DwKh4_s"
@start_code = <<-'EOD'.chomp
package main
import (
"fmt"
)
// YOUR CODE HERE:
// Declare a "divide" function such that the call in the
// "main" function will compile and return 2.8.
// "divide" should accept two float64 values as parameters,
// and return a single float64 value that represents the
// first parameter divided by the second.
// EXTRA CREDIT:
// Have "divide" return TWO values, a float64 and an error.
// If the second parameter is 0, return an error value
// with the message "can't divide by 0". Otherwise, return
// nil for the error value. You can use the fmt.Errorf
// function to generate an error value. You'll also need
// to update the code in "main" to handle the error value.
func main() {
quotient := divide(5.6, 2)
fmt.Printf("%0.2f\n", quotient) // => 2.80
}
EOD
@cheat_sheet = <<-'EOD'.chomp
```
func oneReturnValue(param1 param1Type, param2 param2Type) returnType {
return valueToReturn
}
func twoReturnValues(param1 param1Type, param2 param2Type) (returnType1, returnType2) {
if thereIsAProblem {
return aMeaninglessValue, fmt.Errorf("an error message")
}
return valueToReturn, nil
}
```
EOD
@solution = <<-'EOD'.chomp
package main
import (
"fmt"
)
func divide(dividend float64, divisor float64) float64 {
return dividend / divisor
}
func main() {
quotient := divide(5.6, 2)
fmt.Printf("%0.2f\n", quotient)
}
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: @long_url) %>
## Exercise: Declaring functions extra credit
``` go
package main
import (
"fmt"
)
func divide(dividend float64, divisor float64) (float64, error) {
if divisor == 0.0 {
return 0, fmt.Errorf("can't divide by 0")
}
return dividend / divisor, nil
}
func main() {
quotient, err := divide(5.6, 0.0)
if err != nil {
fmt.Println(err)
} else {
fmt.Printf("%0.2f\n", quotient)
}
}
```
## Pass-by-value
* Go is a "pass-by-value" language (as opposed to "pass-by-reference").
* This means Go functions receive a copy of whatever values you pass to them.
* That's fine, until you want a function to alter a value...
## Pass-by-value
``` go
func main() {
amount := 6
// We want to set "amount" to 12
double(amount)
fmt.Println(amount) // But this prints "6"!
}
// double is SUPPOSED to take a value and double it
func double(number int) {
// But this doubles the COPY, not the original
number *= 2
}
```
## Pointers
The `&` ("address of") operator gets a pointer to a value.
``` go
amount := 6
fmt.Println(amount) // => 6
fmt.Println(&amount) // => 0x1040a124
```
## Pointers
We can get pointers to values of any type.
``` go
var myInt int
fmt.Println(&myInt) // => 0x1040a128
var myFloat float64
fmt.Println(&myFloat) // => 0x1040a140
var myBool bool
fmt.Println(&myBool) // => 0x1040a148
```
## Pointers
A pointer to an `int` is written `*int`, a pointer to a `bool` as `*bool`, and so on.
``` go
func main() {
var myInt int
fmt.Println(reflect.TypeOf(&myInt)) // => *int
var myFloat float64
fmt.Println(reflect.TypeOf(&myFloat)) // => *float64
var myBool bool
fmt.Println(reflect.TypeOf(&myBool)) // => *bool
}
```
## Pointers
You can declare variables that hold pointers:
``` go
var myInt int
var myIntPointer *int
myIntPointer = &myInt
fmt.Println(myIntPointer) // => 0x1040a128
var myFloat float64
var myFloatPointer *float64
myFloatPointer = &myFloat
fmt.Println(myFloatPointer) // => 0x1040a140
```
## Pointers
The `*` _operator_ gets the value a pointer refers to.
``` go
myInt := 4
myIntPointer := &myInt
fmt.Println(myIntPointer) // => 0x1040a124
fmt.Println(*myIntPointer) // => 4
myFloat := 98.6
myFloatPointer := &myFloat
fmt.Println(myFloatPointer) // => 0x1040a140
fmt.Println(*myFloatPointer) // => 98.6
```
## Pointers
The `*` operator can also be used to update the value at a pointer:
``` go
myInt := 4
fmt.Println(myInt) // => 4
myIntPointer := &myInt
// Update the value at the pointer.
*myIntPointer = 8
fmt.Println(*myIntPointer) // => 8
fmt.Println(myInt) // => 8
```
## Pointers
We can use pointers to fix our `double` function:
``` go
func main() {
amount := 6
// Pass pointer instead of value
double(&amount)
fmt.Println(amount) // => 12
}
// Accept pointer instead of value
func double(number *int) {
// Update value at pointer
*number *= 2
}
```
<!-- ******************EXERCISE****************** -->
<%
@title = <<-'EOD'.chomp
Passing pointers
EOD
@short_url = "https://is.gd/goex_pointers"
@long_url = "https://play.golang.org/p/wGKDcFH1Hr0"
@start_code = <<-'EOD'.chomp
// Update this program as described below.
package main
import "fmt"
// negate takes a boolean value and returns its
// opposite. E.g.: negate(false) returns true.
// But we WANT this function to accept a POINTER
// to a boolean value, and update the value at
// the pointer to its opposite. Once this change
// is made, the function doesn't need to return
// anything.
func negate(myBoolean bool) bool {
return !myBoolean
}
func main() {
truth := true
// Change this to pass a pointer.
negate(truth)
// Prints "true", but we want "false".
fmt.Println(truth)
lies := false
// Change this to pass a pointer.
negate(lies)
// Prints "false", but we want "true".
fmt.Println(lies)
}
EOD
@cheat_sheet = <<-'EOD'.chomp
* `!true` is `false`, `!false` is `true`.
* Pointer types are written as `*myType`.
* Get a pointer to a variable's value with `&myVariable`.
EOD
@solution = <<-'EOD'.chomp
package main
import "fmt"
func negate(myBoolean *bool) {
*myBoolean = !*myBoolean
}
func main() {
truth := true
negate(&truth)
fmt.Println(truth) // => false
lies := false
negate(&lies)
fmt.Println(lies) // => true
}
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: @long_url) %>
# Declaring Packages
## The "main" package
* Code intended for direct execution goes in the `main` package.
* Go looks for a `main` function and calls that first.
``` go
package main
import "fmt"
func Hello() {
fmt.Println("Hello!")
}
func Hi() {
fmt.Println("Hi!")
}
func main() {
Hello() // => Hello!
}
```
## The "main" package
But sticking everything in one package will only get you so far...
## The Go workspace
* A directory to hold package code.
* `~/go` by default.
* Or set `$GOPATH` environment variable to a different directory.
## Workspace subdirectories
* `bin`: holds binary executables.
* Add it to your `$PATH` and you can run them from anywhere.
* `pkg`: holds compiled package files.
* You generally don't need to touch this.
* `src`: holds source code.
* Including your code!
## Setting up a package
Let's move our functions to another package.
`~/go/src/greeting/greeting.go`
``` go
package greeting
import "fmt"
func Hello() {
fmt.Println("Hello!")
}
func Hi() {
fmt.Println("Hi!")
}
```
## Importing our package
`random_directory/hi.go`
``` go
package main
import "greeting"
func main() {
greeting.Hello() // => Hello!
greeting.Hi() // => Hi!
}
```
## "go run"
``` go
$ go run hi.go
Hello!
Hi!
```
## Moving "main" to the workspace
`~/go/src/hi/main.go`
``` go
package main
import "greeting"
func main() {
greeting.Hello() // => Hello!
greeting.Hi() // => Hi!
}
```
## "go install"
``` go
$ go install hi
$ tree ~/go
go
|-- bin
| `-- hi
`-- src
|-- greeting
| `-- greeting.go
`-- hi
`-- main.go
```
## "go install"
``` go
$ export PATH=$PATH:$HOME/go/bin
```
(Go installer does this for you.)
## "go install"
``` go
$ hi
Hello!
Hi!
```
## Exported
We ensured our function names were capitalized so they were exported:
`~/go/src/greeting/greeting.go`
``` go
package greeting
import "fmt"
func Hello() {
fmt.Println("Hello!")
}
func Hi() {
fmt.Println("Hi!")
}
```
## Unexported
What if we made them unexported?
`~/go/src/greeting/greeting.go`
``` go
package greeting
import "fmt"
func hello() {
fmt.Println("Hello!")
}
func hi() {
fmt.Println("Hi!")
}
```
## Unexported
Even if we update the function calls in `main` to match...
`~/go/src/hi/main.go`
``` go
package main
import "greeting"
func main() {
greeting.hello()
greeting.hi()
}
```
## Unexported
We're not allowed to call unexported functions from another package.
``` go
$ go install hi
# hi
go/src/hi/main.go:6:9: cannot refer to unexported name greeting.hello
go/src/hi/main.go:6:9: undefined: greeting.hello
go/src/hi/main.go:7:9: cannot refer to unexported name greeting.hi
go/src/hi/main.go:7:9: undefined: greeting.hi
```
## Unexported
So why would you ever make functions unexported?
* Unexported methods are Go's equivalent to Java's `private` methods.
* Use for helper functions that other packages shouldn't call.
* Once you export a function, you shouldn't change it any more.
* You can change how it works _internally_...
* But you shouldn't change its parameters, return value, etc.
* If you do, you risk breaking others' code!
* But you can change unexported functions all you want!
## Import paths
Suppose we want to add support for other languages...
We can nest them under the `greeting` directory.
``` go
$ tree ~/go/
go
`-- src
`-- greeting
|-- dansk
| `-- dansk.go
|-- deutsch
| `-- deutsch.go
`-- greeting.go
```
## Import paths
`~/go/src/greeting/deutsch/deutsch.go`
``` go
// Notice it's not "greeting/deutsch",
// it's just "deutsch".
package deutsch
import "fmt"
func Hallo() {
fmt.Println("Hallo!")
}
func GutenTag() {
fmt.Println("Guten Tag!")
}
```
## Import paths
`~/go/src/greeting/dansk/dansk.go`
``` go
// Notice it's not "greeting/dansk",
// it's just "dansk".
package dansk
import "fmt"
func Hej() {
fmt.Println("Hej!")
}
func GodMorgen() {
fmt.Println("God morgen!")
}
```
## Import paths
Now we can import and use these packages as well.
~/go/src/hi/main.go`
``` go
package main
import (
"greeting"
"greeting/dansk"
"greeting/deutsch"
)
func main() {
greeting.Hello() // => Hello!
greeting.Hi() // => Hi!
dansk.Hej() // => Hej!
dansk.GodMorgen() // => God morgen!
deutsch.Hallo() // => Hallo!
deutsch.GutenTag() // => Guten Tag!
}
```
## Import paths
* Notice the import paths are not the same as the package names!
* Package name is whatever is used in `package` clause in files: `package dansk`
* By convention, last segment of import path is used as package name.
| Import Path | Package Name |
| ----------- | ------------ |
| _greeting_ | _greeting_ |
| greeting/_dansk_ | _dansk_ |
| greeting/_deutsch_ | _deutsch_ |
## "go get"
I set up a repo at `https://github.com/headfirstgo/greeting`, and pushed the package code there...
![](images/github.png)
## "go get"
Now anyone can retrieve the package with `go get github.com/headfirstgo/greeting`:
``` go
$ go get github.com/headfirstgo/greeting
$ tree ~/go
go
`-- src
`-- github.com
`-- headfirstgo
`-- greeting
|-- dansk
| `-- dansk.go
|-- deutsch
| `-- deutsch.go
`-- greeting.go
```
## "go get"
`~/go/src/hi/main.go`
``` go
package main
import (
"github.com/headfirstgo/greeting"
"github.com/headfirstgo/greeting/dansk"
"github.com/headfirstgo/greeting/deutsch"
)
func main() {
greeting.Hello() // => Hello!
dansk.Hej() // => Hej!
deutsch.GutenTag() // => Guten Tag!
}
```
## "go doc"
* To document a package, just add an ordinary comment before its `package` clause.
* Comments can span as many lines as you need.
``` go
// Package greeting greets the user in English.
package greeting
import "fmt"
func Hello() {
fmt.Println("Hello!")
}
func Hi() {
fmt.Println("Hi!")
}
```
## "go doc"
* To document functions, add an ordinary comment before each.
``` go
// Package greeting greets the user in English.
package greeting
import "fmt"
// Hello prints the string "Hello!".
func Hello() {
fmt.Println("Hello!")
}
// Hi prints the string "Hi!".
func Hi() {
fmt.Println("Hi!")
}
```
## "go doc"
Get documentation on a package:
``` go
$ go doc github.com/headfirstgo/greeting
package greeting // import "github.com/headfirstgo/greeting"
Package greeting greets the user.
func Hello()
func Hi()
```
Get documentation on a function:
``` go
$ go doc github.com/headfirstgo/greeting Hello
func Hello()
Hello prints the string "Hello!".
```
## Web documentation
``` go
$ godoc -http=:6060
```
Then visit `http://localhost:6060/pkg/`...
## Web documentation
![](images/godoc.png)
## Web documentation
* Other servers like `godoc.org` make package documentation available on the web.
* For example, `https://godoc.org/github.com/headfirstgo/greeting` got created automatically for the `greeting` package.
![](images/godog.org.png)
## Web documentation
* Want to know more about the `fmt` package?
* Just Google "golang fmt"!
![](images/google.png)
## Web documentation
![](images/fmt_docs.png)
<!-- ******************EXERCISE****************** -->
<%
@title = <<-'EOD'.chomp
Using package documentation
EOD
@short_url = "https://is.gd/goex_documentation"
@long_url = "https://play.golang.org/p/0IoS8oGzrnw"
@start_code = <<-'EOD'.chomp
package main
import (
"fmt"
"log"
)
func main() {
string1 := "12.345"
string2 := "1.234"
// YOUR CODE HERE:
// Look up documentation for the "strconv" package's
// ParseFloat function. (You can use either "go doc"
// or a search engine.) Use ParseFloat to convert
// string1 to a float64 value. Assign the converted
// number to the variable number1, and any error value
// to the variable err. Use the integer 64 for
// ParseFloat's bitSize argument.
if err != nil {
log.Fatal("Could not parse string")
}
// YOUR CODE HERE:
// Use ParseFloat to convert string2 to a float64
// value. Assign the converted number to the variable
// number2, and any error value to the variable err.
if err != nil {
log.Fatal("Could not parse string")
}
fmt.Println(number1 - number2)
}
EOD
@cheat_sheet = <<-'EOD'.chomp
``` go
$ go doc strconv Parsefloat
func ParseFloat(s string, bitSize int) (float64, error)
ParseFloat converts the string s to a floating-point number...
```
* Don't forget to import the `"strconv"` package.
* Use the package name before the function name: `strconv.ParseFloat(...)`
* The first argument to `ParseFloat` is the string you want to convert to a `float64`.
* Use the integer `64` as the second argument.
* Provide variables for both the `float64` return value and the `error` return value: `number1, err := strconv.ParseFloat(string1, 64)`
EOD
@solution = <<-'EOD'.chomp
package main
import (
"fmt"
"log"
"strconv"
)
func main() {
string1 := "12.345"
string2 := "1.234"
number1, err := strconv.ParseFloat(string1, 64)
if err != nil {
log.Fatal("Could not parse string")
}
number2, err := strconv.ParseFloat(string2, 64)
if err != nil {
log.Fatal("Could not parse string")
}
fmt.Println(number1 - number2)
}
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: @long_url) %>
# Data Structures
# Arrays, Slices, and Maps
## What they are
* Arrays: a fixed-size collection of values, all of the same type.
* Slices: a collection of values just like arrays, except it's easy to add more values.
* Maps: a collection of keys, all of the same type. Each key has a corresponding value. All values are of the same type (though possibly different than keys).
## Array/Slice/Map type syntax
``` go
// Array type written as [size]ContainedType
var myArray [3]string
// Slice type written as []ContainedType
var mySlice []string
// Map type written as map[KeyType]ValueType
var myMap map[string]int
```
## Accessing arrays/slices/maps
``` go
var myArray [3]string
var mySlice []string
mySlice = make([]string, 2)
var myMap map[string]int
myMap = make(map[string]int)
myArray[0] = "Amy"
fmt.Println(myArray[0]) // => Amy
mySlice[1] = "Jose"
fmt.Println(mySlice[1]) // => Jose
myMap["Ben"] = 78
fmt.Println(myMap["Ben"]) // => 78
```
## Short declarations with slices/maps
``` go
var myArray [3]string
mySlice := make([]string, 2)
myMap := make(map[string]int)
myArray[0] = "Amy"
fmt.Println(myArray[0]) // => Amy
mySlice[1] = "Jose"
fmt.Println(mySlice[1]) // => Jose
myMap["Ben"] = 78
fmt.Println(myMap["Ben"]) // => 78
```
## Expanding slices is easy
``` go
primes := make([]int, 2)
primes[0] = 2
primes[1] = 3
primes = append(primes, 5)
primes = append(primes, 7)
fmt.Println(primes) // => [2 3 5 7]
```
* Want to do the same with an array? Have to throw it out and restart with a bigger one.
* In most cases you should use slices instead of arrays.
## Values can be of any type
```go
var flags [2]bool // Boolean array
flags[1] = true
fractions := make([]float64, 3) // Float slice
fractions[0] = 0.25
elements := make(map[string]int) // Integer map
elements["He"] = 2
fmt.Println(flags[1]) // => true
fmt.Println(fractions[0]) // => 0.25
fmt.Println(elements["He"]) // => 2
```
## Map keys can be of any type
``` go
binaryBooleans := make(map[bool]int)
binaryBooleans[true] = 1
binaryBooleans[false] = 0
fmt.Println(binaryBooleans[false]) // => 0
fmt.Println(binaryBooleans[true]) // => 1
```
## Array/slice/map literals
Create a collection and add data at the same time.
``` go
myArray := [3]string{"Amy", "Jose", "Ben"}
mySlice := []string{"Amy", "Jose", "Ben"}
myMap := map[string]int{"Amy": 84, "Jose": 96, "Ben": 78}
fmt.Println(myArray[1]) // => Jose
fmt.Println(mySlice[0]) // => Amy
fmt.Println(myMap["Ben"]) // => 78
```
## Collections and loops
Collection elements _can_ be accessed using a loop...
``` go
names := [3]string{"Amy", "Jose", "Ben"}
for i := 0; i < len(names); i++ {
fmt.Println(names[i])
}
```
Output:
``` go
Amy
Jose
Ben
```
## Collections and loops
Don't access/assign outside collection bounds; program will panic (a runtime error that crashes the program).
``` go
names := [3]string{"Amy", "Jose", "Ben"}
for i := 0; i <= len(names); i++ {
fmt.Println("index", i, names[i])
}
```
``` go
index 0 Amy
index 1 Jose
index 2 Ben
panic: runtime error: index out of range
goroutine 1 [running]:
main.main()
/tmp/sandbox741567581/prog.go:10 +0x180
```
## Arrays and "for ... range" loops
It's safer to use `for ... range` loops:
``` go
nameArray := [3]string{"Amy", "Jose", "Ben"}
for index, name := range nameArray {
fmt.Println(index, name)
}
```
Output:
``` go
0 Amy
1 Jose
2 Ben
```
## Slices and "for ... range" loops
``` go
nameSlice := []string{"Amy", "Jose", "Ben"}
for index, name := range nameSlice {
fmt.Println(index, name)
}
```
Output:
``` go
0 Amy
1 Jose
2 Ben
```
## Maps and "for ... range" loops
``` go
grades := map[string]int{"Amy": 84, "Jose": 96, "Ben": 78}
for name, grade := range grades {
fmt.Println(name, grade)
}
```
Output:
``` go
Amy 84
Jose 96
Ben 78
```
## "for ... range" loops and blank identifier
Don't want the index, or don't want the element? Assign it to the blank identifier.
``` go
names := [3]string{"Amy", "Jose", "Ben"}
for _, name := range names {
fmt.Println(name)
}
for index, _ := range names {
fmt.Println(index)
}
```
Output:
``` go
Amy
Jose
Ben
0
1
2
```
## "for ... range" loops and blank identifier
``` go
names := []string{"Amy", "Jose", "Ben"}
for _, name := range names {
fmt.Println(name)
}
for index, _ := range names {
fmt.Println(index)
}
```
Output:
``` go
Amy
Jose
Ben
0
1
2
```
## "for ... range" loops and blank identifier
``` go
grades := map[string]int{"Amy": 84, "Jose": 96, "Ben": 78}
for _, grade := range grades {
fmt.Println(grade)
}
for name, _ := range grades {
fmt.Println(name)
}
```
Output:
``` go
84
96
78
Amy
Jose
Ben
```
<!-- ******************EXERCISE****************** -->
<%
@title = <<-'EOD'.chomp
Slices and Maps
EOD
@short_url = "https://is.gd/goex_collections"
@long_url = "https://play.golang.org/p/LBwN8e1jMgb"
@start_code = <<-'EOD'.chomp
// Fill in the blanks so this program compiles and produces
// the output shown.
package main
import "fmt"
func main() {
// Create a variable to hold a slice of ints.
var primes ____int
// Create a slice with 2 elements.
primes = ____([]int, 2)
// Assign values to the first 2 elements.
primes[____] = 2
primes[____] = 3
// Add a third element to the end of the slice.
primes = ____(primes, 5)
fmt.Println(primes) // => [2 3 5]
// Write a map literal with int keys and string values.
elements := ____[int]string{1: "H", 2: "He", 3: "Li"}
// Loop over each key/value pair in the map.
for atomicNumber, symbol := ____ elements {
fmt.Println(atomicNumber, symbol)
}
// => 1 H
// => 2 He
// => 3 Li
}
EOD
@cheat_sheet = <<-'EOD'.chomp
* Slices:
* To declare a variable that holds a slice of `myType` values, write `var myVar []myType`.
* Slices can be created by calling the `make` function with 2 arguments: the type of the slice and the number of elements.
* Slice element indexes start at `0`, so `mySlice[0]` refers to the first element, `mySlice[1]` to the second, and so on.
* To append an element to `mySlice`, use `mySlice = append(mySlice, newElement)`.
* Maps:
``` go
myMap := map[string]int{"foo": 2, "bar": 1}
for key, value := range myMap {
fmt.Println(key, value)
}
```
EOD
@solution = <<-'EOD'.chomp
// Fill in the blanks so this program compiles and produces
// the output shown.
package main
import "fmt"
func main() {
// Create a variable to hold a slice of ints.
var primes []int
// Create a slice with 2 elements.
primes = make([]int, 2)
// Assign values to the first 2 elements.
primes[0] = 2
primes[1] = 3
// Add a third element to the end of the slice.
primes = append(primes, 5)
fmt.Println(primes) // => [2 3 5]
// Write a map literal with int keys and string values.
elements := map[int]string{1: "H", 2: "He", 3: "Li"}
// Loop over each key/value pair in the map.
for atomicNumber, symbol := range elements {
fmt.Println(atomicNumber, symbol)
}
// => 1 H
// => 2 He
// => 3 Li
}
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: @long_url) %>
# Structs
## Structs
Anonymous struct types...
``` go
var bucket struct {
number float64
word string
toggle bool
}
bucket.number = 3.14
bucket.word = "pie"
bucket.toggle = true
fmt.Println(bucket.number) // => 3.14
fmt.Println(bucket.word) // => pie
fmt.Println(bucket.toggle) // => true
```
## Custom types
`type myType` followed by an underlying type declares a new type.
``` go
type myType struct {
number float64
word string
toggle bool
}
```
## Custom types
``` go
func main() {
var bucket myType
bucket.number = 3.14
bucket.word = "pie"
bucket.toggle = true
fmt.Println(bucket.number) // => 3.14
fmt.Println(bucket.word) // => pie
fmt.Println(bucket.toggle) // => true
}
```
## Embedding structs is like inheriting fields
``` go
type Coordinates struct {
Latitude float64
Longitude float64
}
type Landmark struct {
Name string
// An "anonymous field"
// Has no name of its own, just a type
Coordinates
}
```
``` go
func main() {
var l Landmark
l.Name = "The Googleplex"
// Fields for "embedded struct" are "promoted"
l.Latitude = 37.42
l.Longitude = -122.08
fmt.Println(l.Name, l.Latitude, l.Longitude)
// => The Googleplex 37.42 -122.08
}
```
<!-- ******************EXERCISE****************** -->
<%
@title = <<-'EOD'.chomp
Struct types
EOD
@short_url = "https://is.gd/goex_structs"
@long_url = "https://play.golang.org/p/6TXfzxpTiCj"
@start_code = <<-'EOD'.chomp
package main
import (
"fmt"
)
type Subscriber struct {
Name string
Rate float64
Active bool
}
type Employee struct {
Name string
Salary float64
}
// YOUR CODE HERE:
// Define a struct type named Address that has Street, City, State,
// and PostalCode fields, each with a type of "string".
// Then embed the Address type within the Subscriber and Employee
// types using anonymous fields, so that the code in "main" will
// compile, run, and produce the output shown.
func main() {
var subscriber Subscriber
subscriber.Name = "Aman Singh"
subscriber.Street = "123 Oak St"
subscriber.City = "Omaha"
subscriber.State = "NE"
subscriber.PostalCode = "68111"
fmt.Println("Name:", subscriber.Name) // => Name: Aman Singh
fmt.Println("Street:", subscriber.Street) // => Street: 123 Oak St
fmt.Println("City:", subscriber.City) // => City: Omaha
fmt.Println("State:", subscriber.State) // => State: NE
fmt.Println("Postal Code:", subscriber.PostalCode) // => Postal Code: 68111
var employee Employee
employee.Name = "Joy Carr"
employee.Street = "456 Elm St"
employee.City = "Portland"
employee.State = "OR"
employee.PostalCode = "97222"
fmt.Println("Name:", employee.Name) // => Name: Joy Carr
fmt.Println("Street:", employee.Street) // => Street: 456 Elm St
fmt.Println("City:", employee.City) // => City: Portland
fmt.Println("State:", employee.State) // => State: OR
fmt.Println("Postal Code:", employee.PostalCode) // => Postal Code: 97222
}
EOD
@cheat_sheet = <<-'EOD'.chomp
``` go
type Profile struct {
Bio string
}
type User struct {
Profile
}
var user User
user.Bio = "I teach Go workshops."
```
EOD
@solution = <<-'EOD'.chomp
type Subscriber struct {
Name string
Rate float64
Active bool
Address
}
type Employee struct {
Name string
Salary float64
Address
}
type Address struct {
Street string
City string
State string
PostalCode string
}
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: @long_url) %>
# Defined Types
## Underlying basic types
A custom type can have an underlying basic type
``` go
type Liters float64
type Gallons float64
func main() {
var carFuel Gallons
var busFuel Liters
// Defining a type defines a conversion
// from the underlying type to the new type
carFuel = Gallons(10.0)
busFuel = Liters(240.0)
fmt.Println(carFuel) // => 10
fmt.Println(busFuel) // => 240
}
```
## Methods
``` go
type MyType string
// Specify a "receiver parameter" within a function
// definition to make it a method. The receiver
// parameter's type will be the type the method
// gets defined on.
func (m MyType) sayHi() {
fmt.Println("Hi")
}
```
## Methods
``` go
type MyType string
func (m MyType) sayHi() {
fmt.Println("Hi")
}
func main() {
value := MyType("a MyType value")
value.sayHi() // => Hi
anotherValue := MyType("another value")
anotherValue.sayHi() // => Hi
}
```
## Receiver parameter acts like just another parameter
``` go
type MyType string
func (m MyType) sayHi() {
fmt.Println("Hi from", m)
}
func main() {
value := MyType("a MyType value")
value.sayHi() // => Hi from a MyType value
anotherValue := MyType("another value")
anotherValue.sayHi() // => Hi from another value
}
```
## Underlying type is _not_ a superclass
The underlying type specifies how a type's data will be stored, so this is OK...
``` go
type MyType string
func (m MyType) sayHi() {
fmt.Println("Hi from", m)
}
type MyType2 MyType
func main() {
value2 := MyType2("a MyType2 value")
fmt.Println(value2)
}
```
## Underlying type is _not_ a superclass
But a type does _not_ inherit methods from its underlying type.
``` go
type MyType string
func (m MyType) sayHi() {
fmt.Println("Hi from", m)
}
type MyType2 MyType
func main() {
value2 := MyType2("a MyType2 value")
value2.sayHi()
}
```
Compile error:
``` go
prog.go:15:8: value2.sayHi undefined (type MyType2 has no field or method sayHi)
```
## Underlying type is _not_ a superclass
"Although Go has types and methods and allows an object-oriented style of programming, there is no type hierarchy."
—https://golang.org/doc/faq
* There is no method inheritance!
* But there's another way to get the same benefits...
## Embedding structs is like inheriting fields
Remember how fields for an embedded struct get promoted to the outer struct?
``` go
type Coordinates struct {
Latitude float64
Longitude float64
}
type Landmark struct {
Name string
// An "anonymous field"
// Has no name of its own, just a type
Coordinates
}
```
``` go
func main() {
var l Landmark
l.Name = "The Googleplex"
// Fields for "embedded struct" are "promoted"
l.Latitude = 37.42
l.Longitude = -122.08
fmt.Println(l.Name, l.Latitude, l.Longitude)
// => The Googleplex 37.42 -122.08
}
```
## Promotion of embedded types' methods
Methods for an embedded type get promoted too!
``` go
func (c Coordinates) Location() string {
return fmt.Sprintf("(%0.2f, %0.2f)",
c.Latitude, c.Longitude)
}
func main() {
var l Landmark
l.Name = "The Googleplex"
l.Latitude = 37.42
l.Longitude = -122.08
// Methods from embedded type are
// promoted to outer type
fmt.Println(l.Location())
// => (37.42, -122.08)
}
```
## Promotion of embedded types' methods
* Embed additional types to gain additional methods.
* You've heard "favor composition over inheritance"...
* Go implements that principle at the language level.
``` go
type Coordinates struct {
Latitude float64
Longitude float64
}
func (c Coordinates) Location() string {
return fmt.Sprintf("(%0.2f, %0.2f)",
c.Latitude, c.Longitude)
}
type Landmark struct {
Name string
Coordinates
}
```
<!-- ******************EXERCISE****************** -->
<%
@title = <<-'EOD'.chomp
Defined types
EOD
@short_url = "https://is.gd/goex_defined_types"
@long_url = "https://play.golang.org/p/TeMJ9D6bBOz"
@start_code = <<-'EOD'.chomp
package main
import "fmt"
// YOUR CODE HERE:
// Define a Rectangle struct type with Length and Width
// fields, each of which has a type of float64.
// YOUR CODE HERE:
// Define an Area method on the Rectangle type. It should
// accept no parameters (other than the receiver parameter).
// It should return a float64 value calculated by multiplying
// the receiver's Length by its Width.
// YOUR CODE HERE:
// Define a Perimeter method on the Rectangle type. It should
// accept no parameters. It should return a float64 value
// representing the receiver's perimeter (2 times its Length
// plus 2 times its Width).
func main() {
// Once you've defined the above code correctly,
// this code should compile and run.
var myRectangle Rectangle
myRectangle.Length = 2
myRectangle.Width = 3
fmt.Println("Area:", myRectangle.Area()) // => Area: 6
fmt.Println("Perimeter:", myRectangle.Perimeter()) // => Perimeter: 10
}
EOD
@cheat_sheet = <<-'EOD'.chomp
``` go
type User struct {
Name string
}
func (u User) ShoutName() string {
return strings.ToUpper(u.Name)
}
```
EOD
@solution = <<-'EOD'.chomp
package main
import (
"fmt"
)
type Rectangle struct {
Length float64
Width float64
}
func (r Rectangle) Area() float64 {
return r.Length * r.Width
}
func (r Rectangle) Perimeter() float64 {
return (2 * r.Length) + (2 * r.Width)
}
func main() {
var myRectangle Rectangle
myRectangle.Length = 2
myRectangle.Width = 3
fmt.Println("Area:", myRectangle.Area())
fmt.Println("Perimeter:", myRectangle.Perimeter())
}
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: @long_url) %>
# Interfaces
## Interfaces
A type with `Play` and `Stop` methods...
``` go
type TapePlayer struct {
Batteries string
}
func (t TapePlayer) Play(song string) {
fmt.Println("Playing", song)
}
func (t TapePlayer) Stop() {
fmt.Println("Stopped!")
}
```
## Interfaces
_Another_ type with `Play` and `Stop` methods...
``` go
type TapeRecorder struct {
Microphones int
}
func (t TapeRecorder) Play(song string) {
fmt.Println("Playing", song)
}
func (t TapeRecorder) Record() {
fmt.Println("Recording")
}
func (t TapeRecorder) Stop() {
fmt.Println("Stopped!")
}
```
## Interfaces
A function that accepts a `TapePlayer`...
``` go
func playList(device TapePlayer, songs []string) {
for _, song := range songs {
device.Play(song)
}
device.Stop()
}
```
## Interfaces
``` go
func main() {
mixtape := []string{"Jessie's Girl", "Whip It", "9 to 5"}
var player TapePlayer
playList(player, mixtape)
}
```
Output:
```
Playing Jessie's Girl
Playing Whip It
Playing 9 to 5
Stopped!
```
## Interfaces
But don't try to pass a `TapeRecorder` to `playList`!
``` go
func main() {
mixtape := []string{"Jessie's Girl", "Whip It", "9 to 5"}
var recorder TapeRecorder
playList(recorder, mixtape)
}
```
Compile error:
```
prog.go:40:10: cannot use recorder (type TapeRecorder) as type TapePlayer in argument to playList
```
## Interfaces
Define a `Player` interface with the methods you want:
``` go
type Player interface {
// Must have a Play method with
// a single string parameter
Play(string)
// Must have a Stop method with
// no parameters
Stop()
}
```
* Notice we don't have to modify the `TapePlayer` or `TapeRecorder` type definitions!
* Any type with `Play(string)` and `Stop()` methods automatically _satisfies_ this interface.
## Interfaces
Modify the `playList` function to accept a value of the `Player` (interface) type:
``` go
func playList(device Player, songs []string) {
for _, song := range songs {
device.Play(song)
}
device.Stop()
}
```
## Interfaces
Now, you can pass in a `TapePlayer` _or_ a `TapeRecorder` (or any other type with `Play` and `Stop` methods)!
``` go
func main() {
mixtape := []string{"Jessie's Girl", "Whip It", "9 to 5"}
var player TapePlayer
playList(player, mixtape)
var recorder TapeRecorder
playList(recorder, mixtape)
}
```
## Interfaces
Output:
``` go
Playing Jessie's Girl
Playing Whip It
Playing 9 to 5
Stopped!
Playing Jessie's Girl
Playing Whip It
Playing 9 to 5
Stopped!
```
## Type assertions
* If you have a value with an interface type, you can only call methods included in that interface.
* This is true even if the concrete value has that method!
## Type assertions
``` go
func main() {
// Even though we're passing in a TapeRecorder...
TryOut(TapeRecorder{})
}
func TryOut(player Player) {
player.Play("Test Track")
player.Stop()
// Player interface doesn't include this method!
player.Record()
}
```
Compile error:
``` go
player.Record undefined (type Player has no field or method Record)
```
## Type assertions
``` go
func main() {
TryOut(TapeRecorder{})
}
func TryOut(player Player) {
player.Play("Test Track")
player.Stop()
// Do a type assertion to get the concrete value back...
recorder := player.(TapeRecorder)
// Then you can call Record on that.
recorder.Record()
}
```
## Type assertions
Output:
``` go
Playing Test Track
Stopped!
Recording
```
<!-- ******************EXERCISE****************** -->
<%
@title = <<-'EOD'.chomp
Interfaces
EOD
@short_url = "https://is.gd/goex_interfaces"
@long_url = "https://play.golang.org/p/6g4_wz89Jes"
@start_code = <<-'EOD'.chomp
package main
import "fmt"
type Whistle string
func (w Whistle) MakeSound() {
fmt.Println("Tweet!")
}
type Horn string
func (h Horn) MakeSound() {
fmt.Println("Honk!")
}
type Robot string
func (r Robot) MakeSound() {
fmt.Println("Beep Boop")
}
func (r Robot) Walk() {
fmt.Println("Powering legs")
}
// YOUR CODE HERE:
// Define a NoiseMaker interface type, which the above
// Whistle, Horn, and Robot types will all satisfy.
// It should require one method, MakeSound, which has
// no parameters and no return values.
// YOUR CODE HERE:
// Define a Play function that accepts a parameter with
// the NoiseMaker interface. Play should call MakeSound
// on the parameter it receives.
func main() {
// When the above code has been implemented
// correctly, this code should run and produce
// the output shown.
Play(Whistle("Toyco Canary")) // => Tweet!
Play(Horn("Toyco Blaster")) // => Honk!
Play(Robot("Botco Ambler")) // => Beep Boop
}
EOD
@cheat_sheet = <<-'EOD'.chomp
``` go
type Pen string
func (p Pen) Write() {
fmt.Println("writing stuff")
}
type WritingInstrument interface {
Write()
}
func Test(w WritingInstrument) {
w.Write()
}
```
EOD
@solution = <<-'EOD'.chomp
package main
import "fmt"
type Whistle string
func (w Whistle) MakeSound() {
fmt.Println("Tweet!")
}
type Horn string
func (h Horn) MakeSound() {
fmt.Println("Honk!")
}
type Robot string
func (r Robot) MakeSound() {
fmt.Println("Beep Boop")
}
func (r Robot) Walk() {
fmt.Println("Powering legs")
}
type NoiseMaker interface {
MakeSound()
}
func play(n NoiseMaker) {
n.MakeSound()
}
func main() {
play(Whistle("Toyco Canary")) // => Tweet!
play(Horn("Toyco Blaster")) // => Honk!
play(Robot("Botco Ambler")) // => Beep Boop
}
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: @long_url) %>
# Handling Errors
## "defer"
It's usually polite to end conversations with "goodbye":
``` go
func Socialize() {
fmt.Println("Hello!")
fmt.Println("Nice weather, eh?")
fmt.Println("Goodbye!")
}
func main() {
Socialize()
}
```
Output:
``` go
Hello!
Nice weather, eh?
Goodbye!
```
## "defer"
Write `defer` before a function call, and it will be "deferred" until enclosing function ends.
``` go
func Socialize() {
// This call will be made when Socialize ends.
defer fmt.Println("Goodbye!")
fmt.Println("Hello!")
fmt.Println("Nice weather, eh?")
}
```
Output:
``` go
Hello!
Nice weather, eh?
Goodbye!
```
## "defer" calls made no matter what
``` go
func Socialize() error {
// Deferred call is made even if Socialize
// exits early (say, due to an error).
defer fmt.Println("Goodbye!")
fmt.Println("Hello!")
return fmt.Errorf("I don't want to talk.")
// The below code won't be run!
fmt.Println("Nice weather, eh?")
return nil
}
func main() {
err := Socialize()
if err != nil {
log.Fatal(err)
}
}
```
## "defer" calls made no matter what
Output:
``` go
Hello!
Goodbye!
2019/04/22 11:22:29 I don't want to talk.
exit status 1
```
## A (somewhat) more realistic example
* We need a function that prints the contents of a file, then closes it.
* If there's an error, it should be returned.
* But the file should be closed even if there's an error!
``` go
func main() {
err := PrintLines("lorem_ipsum.txt")
if err != nil {
log.Fatal(err)
}
}
```
## A (somewhat) more realistic example
``` go
func PrintLines(fileName string) error {
file, err := os.Open(fileName)
if err != nil {
return err
}
defer file.Close()
scanner := bufio.NewScanner(file)
for scanner.Scan() {
fmt.Println(scanner.Text())
}
if scanner.Err() != nil {
return scanner.Err()
}
return nil
}
```
## "panic"
* `panic` usually signals an _unanticipated_ error.
* This example is just to show its mechanics.
``` go
func Socialize() {
fmt.Println("Hello!")
panic("I need to get out of here!")
// The below code won't be run!
fmt.Println("Nice weather, eh?")
fmt.Println("Goodbye!")
}
```
## "panic"
Output:
``` go
Hello!
panic: I need to get out of here.
goroutine 1 [running]:
main.Socialize()
/Users/jay/socialize4_panic.go:9 +0x79
main.main()
/Users/jay/socialize4_panic.go:16 +0x20
exit status 2
```
## "panic" and "defer"
``` go
func Socialize() {
defer fmt.Println("Goodbye!")
fmt.Println("Hello!")
panic("I need to get out of here!")
// The below code won't be run!
fmt.Println("Nice weather, eh?")
}
```
## "panic" and "defer"
Output:
``` go
Hello!
Goodbye!
panic: I need to get out of here!
goroutine 1 [running]:
main.Socialize()
/Users/jay/socialize5_panic_defer.go:10 +0xd5
main.main()
/Users/jay/socialize5_panic_defer.go:16 +0x20
exit status 2
```
## "recover"
``` go
func CalmDown() {
// Halt the panic.
panicValue := recover()
// Print value passed to panic().
fmt.Println(panicValue)
}
func Socialize() {
defer fmt.Println("Goodbye!")
defer CalmDown()
fmt.Println("Hello!")
panic("I need to get out of here!")
// The below code won't be run!
fmt.Println("Nice weather, eh?")
}
```
## "recover"
Output:
``` go
Hello!
I need to get out of here!
Goodbye!
```
* Deferred `CalmDown` prints the `panic` value.
* Deferred `Println` prints "Goodbye!".
## "panic" should not be used like an exception
I know of one place in the standard library that `panic` is used in normal program flow: in a recursive parsing function that panics to unwind the call stack after a parsing error. (The function then recovers and handles the error normally.)
## "panic" should not be used like an exception
Generally, `panic` should be used only to indicate "impossible" situations:
``` go
func awardPrize() {
doorNumber := rand.Intn(3) + 1
if doorNumber == 1 {
fmt.Println("You win a cruise!")
} else if doorNumber == 2 {
fmt.Println("You win a car!")
} else if doorNumber == 3 {
fmt.Println("You win a goat!")
} else {
// This should never happen.
panic("invalid door number")
}
}
```
## "panic" should not be used like an exception
* If you know an error could happen, use normal control flow statements to handle it.
* Google "golang errors are values" (which should take you to `https://blog.golang.org/errors-are-values`) for some tips on making error handling more pleasant.
<!-- ******************EXERCISE****************** -->
<%
@title = <<-'EOD'.chomp
Handling errors
EOD
@short_url = "https://is.gd/goex_recovery"
@long_url = "https://play.golang.org/p/I16x3MSub57"
@start_code = <<-'EOD'.chomp
package main
import "fmt"
type Refrigerator struct {
Brand string
}
type Food string
func (r Refrigerator) Open() {
fmt.Println("Opening refrigerator")
}
func (r Refrigerator) Close() {
fmt.Println("Closing refrigerator")
}
func (r Refrigerator) FindFood(food string) (Food, error) {
// Food storage not implemented yet; always return error!
// Note: don't change FindFood as part of this exercise!
return Food(""), fmt.Errorf("%s not found", food)
}
// YOUR CODE HERE:
// Modify the code in the Eat function so that fridge.Close will
// always be called at the end, even if fridge.FindFood returns
// an error. Once you've figured the solution out, your changes
// will actually be quite small! Note: it wouldn't be appropriate
// to use either "panic" or "recover" in this exercise; we won't
// be using either one.
func Eat(fridge Refrigerator) error {
fridge.Open()
food, err := fridge.FindFood("bananas")
if err != nil {
return err
}
fmt.Println("Eating", food)
fridge.Close()
return nil
}
// CURRENT OUTPUT:
// Opening refrigerator
// bananas not found
// DESIRED OUTPUT:
// Opening refrigerator
// Closing refrigerator
// bananas not found
func main() {
var fridge Refrigerator
err := Eat(fridge)
if err != nil {
fmt.Println(err)
}
}
EOD
@cheat_sheet = <<-'EOD'.chomp
``` go
func Camp() error {
var fire Fire
fire.Light()
defer fire.Extinguish()
return fmt.Errorf("spotted a bear")
fmt.Println("Toasting marshmallows")
return nil
}
```
EOD
@solution = <<-'EOD'.chomp
package main
import "fmt"
type Refrigerator struct {
Brand string
}
type Food string
func (r Refrigerator) Open() {
fmt.Println("Opening refrigerator")
}
func (r Refrigerator) Close() {
fmt.Println("Closing refrigerator")
}
func (r Refrigerator) FindFood(food string) (Food, error) {
// Food storage not implemented yet; always return error!
return Food(""), fmt.Errorf("%s not found", food)
}
func Eat(fridge Refrigerator) error {
fridge.Open()
defer fridge.Close()
food, err := fridge.FindFood("bananas")
if err != nil {
return err
}
fmt.Println("Eating", food)
return nil
}
func main() {
var fridge Refrigerator
err := Eat(fridge)
if err != nil {
fmt.Println(err)
}
}
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: @long_url) %>
# Concurrency
## A non-concurrent program
``` go
// responseSize retrieves "url" and prints
// the response length in bytes.
func responseSize(url string) {
fmt.Println("Getting", url)
// Note: errors ignored with _!
response, _ := http.Get(url)
defer response.Body.Close()
body, _ := ioutil.ReadAll(response.Body)
fmt.Println(len(body))
}
```
## A non-concurrent program
``` go
func main() {
// Note the time we started.
start := time.Now()
responseSize("https://example.com/")
responseSize("https://golang.org/")
responseSize("https://golang.org/doc")
// Print how long everything took.
fmt.Println(time.Since(start).Seconds(), "seconds")
}
```
## A non-concurrent program
Output:
``` go
Getting https://example.com/
1270
Getting https://golang.org/
8158
Getting https://golang.org/doc
12558
1.5341211000000001 seconds
```
If only we could make additional calls to `responseSize` while we were waiting for HTTP responses...
## Goroutines
* `responseSize` function unchanged.
* Just add `go` keyword before each call to it.
``` go
func main() {
start := time.Now()
go responseSize("https://example.com/")
go responseSize("https://golang.org/")
go responseSize("https://golang.org/doc")
fmt.Println(time.Since(start).Seconds())
}
```
## Goroutines
Output:
``` go
3.1e-06
```
* Run time so brief the duration is printed in scientific notation.
* None of the `responseSize` goroutines get to even request their URL.
* Problem is, `main` goroutine exits, ending the program, without waiting for `responseSize` goroutines.
## Channels
* Modify `responseSize` to accept a "channel" as a parameter.
``` go
// A channel type is written as "chan" followed by data type.
func responseSize(url string, channel chan int) {
fmt.Println("Getting", url) // Unchanged
response, _ := http.Get(url) // Unchanged
defer response.Body.Close() // Unchanged
body, _ := ioutil.ReadAll(response.Body) // Unchanged
// Send body length value via channel.
channel <- len(body)
}
```
## Channels
``` go
func main() {
start := time.Now() // Unchanged
// Make a channel to carry ints.
sizes := make(chan int)
// Pass channel to each call to responseSize.
go responseSize("https://example.com/", sizes)
go responseSize("https://golang.org/", sizes)
go responseSize("https://golang.org/doc", sizes)
// Read and print values from channel.
fmt.Println(<-sizes)
fmt.Println(<-sizes)
fmt.Println(<-sizes)
fmt.Println(time.Since(start).Seconds()) // Unchanged
}
```
## Channels
``` go
Getting https://golang.org/doc
Getting https://golang.org/
Getting https://example.com/
1270
8158
12558
0.695384291
```
* Finishes in half the time of the original! (YMMV.)
* The channel accomplishes two things:
* Channel reads cause `main` goroutine to block until `responseSize` goroutines send, so they have time to finish before program ends.
* The channel transmits data from the `responseSize` goroutines back to the `main` goroutine.
<!-- ******************EXERCISE****************** -->
<%
@title = <<-'EOD'.chomp
Goroutines and channels
EOD
@short_url = "https://is.gd/goex_goroutines"
@long_url = "https://play.golang.org/p/mtfvNLts6Vm"
@start_code = <<-'EOD'.chomp
// This program should call the "repeat" function twice, using two
// separate goroutines. The first goroutine should print the string
// "x" repeatedly, and the second goroutine should print "y"
// repeatedly. You'll also need to create a channel that carries
// boolean values to pass to "repeat", so the goroutine can signal
// when it's done.
//
// Output will vary, but here's one possible result:
// yyyyyyyyyyyyyyyyyyyyxxxxxxxxxxxxxxxxxxxxxyyyyyyyyyxxxxxxxxxy
//
// Replace the blanks ("____") in the code so the program will
// compile and run.
package main
import (
"fmt"
)
// repeat prints a string multiple times, then writes "true" to the
// provided channel to signal it's done.
func repeat(s string, channel ____ bool) {
for i := 0; i < 30; i++ {
fmt.Print(s)
}
channel ____ true
}
func main() {
channel := ____(chan bool)
____ repeat("x", channel)
____ repeat("y", channel)
<-channel
<-channel
}
EOD
@cheat_sheet = <<-'EOD'.chomp
``` go
func myFunc(channel chan int) {
channel <- 42
}
func main() {
channel := make(chan int)
go myFunc(channel)
go myFunc(channel)
fmt.Println(<-channel)
fmt.Println(<-channel)
}
```
EOD
@solution = <<-'EOD'.chomp
package main
import "fmt"
func repeat(s string, channel chan bool) {
for i := 0; i < 30; i++ {
fmt.Print(s)
}
channel <- true
}
func main() {
channel := make(chan bool)
go repeat("x", channel)
go repeat("y", channel)
<-channel
<-channel
}
// => yyyyyyyyyyyyyyyyyyyyxxxxxxxxxxxxxxxxxxxxxyyyyyyyyyxxxxxxxxxy
EOD
%>
<%= exercise(title: @title, short_url: @short_url, start_code: @start_code, cheat_sheet: @cheat_sheet, solution: @solution, long_url: @long_url) %>
# Where to Go Next
## "go test"
Suppose we're not certain this function works correctly...
``` go
func JoinWithCommas(phrases []string) string {
if len(phrases) == 2 {
return phrases[0] + " and " + phrases[1]
} else {
result := strings.Join(phrases[:len(phrases)-1], ", ")
result += ", and "
result += phrases[len(phrases)-1]
return result
}
}
```
## "go test"
We can use the `testing` package to write a test function...
``` go
import (
"fmt"
"testing"
)
func TestTwoElements(t *testing.T) {
list := []string{"apple", "orange"}
want := "apple and orange"
got := JoinWithCommas(list)
if got != want {
t.Error(errorString(list, got, want))
}
}
```
## "go test"
Add tests for all the functionality we want...
``` go
import (
"fmt"
"testing"
)
func TestOneElement(t *testing.T) {
list := []string{"apple"}
want := "apple"
got := JoinWithCommas(list)
if got != want {
t.Error(errorString(list, got, want))
}
}
func TestTwoElements(t *testing.T) {
list := []string{"apple", "orange"}
want := "apple and orange"
got := JoinWithCommas(list)
if got != want {
t.Error(errorString(list, got, want))
}
}
func TestThreeElements(t *testing.T) {
list := []string{"apple", "orange", "pear"}
want := "apple, orange, and pear"
got := JoinWithCommas(list)
if got != want {
t.Error(errorString(list, got, want))
}
}
func errorString(list []string, got string, want string) string {
return fmt.Sprintf("JoinWithCommas(%#v) = \"%s\", want \"%s\"", list, got, want)
}
```
## "go test"
Then just run `go test` and get a summary of the tests that failed!
``` go
$ go test github.com/headfirstgo/prose
--- FAIL: TestOneElement (0.00s)
lists_test.go:13: JoinWithCommas([]string{"apple"}) = ", and apple", want "apple"
FAIL
FAIL github.com/headfirstgo/prose 0.006s
```
## Web development
The standard library includes the `net/http` package:
``` go
package main
import "net/http"
func helloHandler(writer http.ResponseWriter, request *http.Request) {
// Note: unhandled error!
writer.Write([]byte("<h1>Hello, web!</h1>"))
}
func main() {
http.HandleFunc("/hello", helloHandler)
// Note: unhandled error!
http.ListenAndServe("localhost:8080", nil)
}
```
## Web development
![](images/web_app.png)
## Go Gopher
By Renee French, used under a CC-Attribution-3.0 license.
![](images/gopher_color.png)
## Other resources
* Go Tour: `https://tour.golang.org`
* Go Playground: `https://play.golang.org`
Visit `https://headfirstgo.com` for info on the book, and for more exercises!