The Complete Developers Guide (Golang)

Overview

info

My notes of the "Go: The Complete Developers Guide (Golang)" course on Udemy to learn Golang.

Notes

From Simple Start Chapter, questions:

1. how do we run the code in our project?
2. what does package main mean?
3. what does import "fmt" mean?
4. what's that func thing?
5. how is the main.go file organized

Simple Start

run code

go run main.go

• go run = compile & run
• go build = just compile
• go install, go get = installs pkg, download raw source code

package main

package == project == workspace

all files belonging to that "package" must declare so at the top with package main.

2 types of package = executable (runnable) and reusable (a 'helper', good for logic, libs)

package main is SACRED! only use it if you want to create a "runnable" file. a "non-main" package, compiles but does NOT run anything.

any executable package main must have a func main in it.

rename package to anything else and do go build does NOT build any executable file.

import fmt

gives us access to fmt library/package.

func

fun main() { - short for function.

main.go organized

• package declaration
• imports packages
• func main

Deeper into Go

variable declaration

this var Name string and name := "Ron" are EQUIVALENT in variable initialization and (initial) assignment the := will determine the type for you.

:= use only when declaring new vars, dont use for value "assignment".

functions

when writing up non-main functions, this is the format func funcName() typeToReturn e.g. func newCard() string { you'll get an error message if return type and return value type are mismatched.

slice and for Loops

Array = fixed list Slice = array that can grow/shrink

the TYPES in a slice must be of the same type.

declare a slice: cards := []string{"elements","inside","slice"}. the [] and string declares its a "slice" of type "string", and then {} to hold the elements of the slice.

adding elements to a slice: cards = append(cards, "newElement") <-- important to note here that the original "cards" slice is not modified with this new element, instead a NEW slice with the appended element is returned.

iterate over a slice, print every element: for i, card := range cards { fmt.Println(i, card)}

i = index, range = iterate over every element in slice.

oo vs go

Go is NOT OO language.

Think - go types (string, int, float, array, map), then extend type deck []string, then functions

• main.go - our main program that manipulates the deck
• deck.go - describe deck, how it works (spec?)
• deck_test.go - automatically test our deck

card deck go program

For our card app, what functions do we want?

1. newDeck - create return list of playing cards (array of strings)
2. print
3. shuffle
4. deal
5. saveToFile
6. newDeckFromFile

add the custom types and receivers as necessary.

custom types and receiver functions

we can do something like this to sort of simulate "extends" from OO approach

// in a separate .go file

// declare new type 
type deck []string

// create custom method for new type
func (d deck) print() {
  for i, card := range d {
      fmt.Println(i, card)
    }
}

We declared a custom type with type deck []string and can now use deck "type" anywhere in our main package code.

Printing the new type out now made easier with our print() function.

(d deck) is the bit that makes this func a "receiver function".

A receiver sets up methods on variables we create e.g. we create var card of type deck, and now the print func can be setup on ANY var of type deck e.g. card.print()

Think of d arg as this or self - in go, never use "this" or "self", and also always refer to the THING that you're setting method up on, by convention if your type is deck your arg name will be d, but you can do whatever as long as the references match i.e. (d deck) and range d match.

If you have a var that you don't care to use, and want to avoid the "you declared but haven't used this var" error message, replace them with _ underscore.

e.g.

func (d deck) print() {
  for i, card := range d {
    fmt.Println(i, card)
  }
}

slice range syntax

Dealing out a "hand", a slice of the 52 available cards.

slices indexed from 0, e.g. fruits[0]

for a subset, or range of the slice: fruits[0:2] will give you fruits of index 0,1 cos the first index 0 is "inclusive", but the 2nd index 2 is "up to, but NOT including" 2.

shorthand for this can be [:2] is the same as [0:2], another example is [2:] is the same as [2:n] i.e. "from index 2 until the end of the index"

when returning 2 x type deck values from the function, you assign them to two variables like this: hand, remainingDeck := deal(cards, 5) <-- because deal(cards, 5) gets (deck, deck) from the receiver function.

deck to string

trying to save to file, import io/ioutil we need to transfer our strings to a []byte "byte slice".

a way to do a "type conversion":

greeting = "Hi there!"
fmt.Println([]byte(greeting)) // changes the greeting string into a byte slice.

our process = start with deck --> []string --> string --> []byte

the string is ALL card strings smashed together to then convert to a []byte.

join slice of Strings

lookup and use strings library Join function.

save data to HDD

lookup and use io/iotuil library WriteFile function for writing a []byte byte slice to disk under "filename".

read data from HDD

use io/ioutil functions ReadFile to do the reverse and open a file from HDD, then use strings function Split to reverse what you did with Join, now you have a []string string slice AKA a deck type.

error handling

the err convention

...
if err != nil {
    fmt.Println("Error:", err)
    os.Exit(1)
  }

make sure to import os, and any non-zero number inside Exit() to signal a bad exit.

Summary

• create type deck
• tie methods to the new type deck using "receiver functions"
• receivers = can tack onto the type deck e.g. cards.print()
• no receiver for methods like .deal() because the "root" instance of the deck (??)

Assignment 1

write a go program that iterates over a range of numbers and evaluates even and odd and prints a statement to each.

my solution:

package main

import "fmt"

func main() {
    numbers := []int{1,2,3,4,5,6,7,8,9,10}

    for _, num := range numbers {
        if num % 2 == 0 {
            fmt.Println(num, " is even.")
        } else {
            fmt.Println(num, " is odd.")
          }
        }
      }
  }

Data Structures

structs

aka data structures are a "collection of properties that are related together".

we first define a "structure" e.g. of a person, then we create instances of people.

e.g.

type person struct {
    firstName string
    lastName string
}

few different ways to construct struct

1. alex := person{firstName: "Alex", lastName: "Anderson"}
2. var alex person then init with alex.firstName = "Alex" and alex.lastName = "Anderson"
3. var alex person then fmt.Printf("%+v", alex) with %+v printing out field names and values.

updating struct values

if you don't init the vars with values, if string they get assigned "" empty string value. if bool or int, they get 0.

so if you fmt.Println(alex) with no value assigned, you get { } printed out i.e. "empty strings"

using fmt.Printf() and syntax fmt.Printf("%+v", alex) as the 3rd way to show structs.

final form :

package main

import "fmt"

type person struct {
  firstName string
  lastName string
}

func main() {
  //alex := person{firstName: "Alex", lastName: "Anderson"} // version 1

  var alex person // version 2

  alex.firstName = "Alex"
  alex.lastName = "Anderson"

  fmt.Println(alex)
  fmt.Printf("%+v", alex)
}

embedded structs

embed on struct inside another struct.

type person
type contactInfo

you can use custom types inside our structs.

package main

import "fmt"

type contactInfo struct {
  email string
  zipCode int
}

type person struct {
  firstName string
  lastName string
  contact contactInfo
}

func main() {
  jim := person{
    firstName: "Jim",
    lastName: "Party",
    contact: contactInfo{
      email: "[email protected]",
      zipCode: 94000,
    },
  }
  fmt.Printf("%+v", jim)
}

another method of declaring our contactInfo is to remove the explicit variable name contact and just do the following:

type person struct {
  firstName string
  lastName string
  contactInfo
}

and then

function main() {
  jim := person{
    firstName: "Jim",
    lastName: "Party",
    contactInfo: contactInfo{
      email: "[email protected]",
      zipCode: 94000,
    },
  }
}

structs receiver functions

recap receiver functions

func (letter type) funcName(varName varType) { }

// e.g. func (p person) updateName(newFirstName string)

pointers

the reason the updateName function did not work.

Go is a "pass by values" means you work on a COPY of the value of the object/thing, not the original object/thing.

e.g. we contruct the person{firstName: "Jim"...} person, but when we pass the value to our function for a name update func (p person) updateName(newFirstName string), it creates a copy p of the person Jim:

RAM
Address	Value
0000
0001	person{firstName: "Jim"...}
0002
0003	person{firstName: "Jim"...}

original Jim object at 0001, "newFirstName" Jim object (i.e. p) at 0003.

to summarize, when updateName is called, go makes a COPY of that struct, and then makes the COPY (p) available to the function for processing.

why does go do this?

Pointer Operations

operators

& creates a memory ADDRESS pointer

* creates a memory VALUE pointer

&jim = "give me the memory address of the value this variable is pointing to"

so jimPointer := &jim is now pointing to the MEMORY ADDRESS of whatever &jim evaluated to.

*pointer = "give me the value this memory address is pointing to"

so func(pointerToPerson *person) says give me the VALUE at the pointer where this *pointer memory address is pointing.

understand difference between pointer as a TYPE and as a POINTER:

1. pointer in front of a type e.g. func (pointerToPerson *person) means this receiver can only accept a type of a "pointer to a person" i.e. something like jimPointer
2. pointer in front of a pointer e.g. (*pointerToPerson).firstName

Rules to Remember

use *address to turn an address into value.

use &value to turn a value into address.

Shortcuts

with this code:

  jimPointer := &jim
  jimPointer.updateName("jimmy")
  jim.print()

you can remove jimPointer := &jim, and the func (pointerToPerson *person) will automatically take your "type person" and turn it into a "pointer person (*person)" for you.

Pointer Gotchas

struct vs slice

with struct, you need pointers to update the actual values.

with stlices - values seem to update directly as you act upon the values.

reference vs value types

Arrays vs Slices - arrays are primitive, can't be resized. we use slices more.

A slice gives us BOTH a slice data structure (ptr, capacity, length) and array data structure (ptr points to the elements in our slice "array").

in memory, our SLICE is registered as the slice data structure at one memory address, and the actual array with our elements in it at another memory address.

when the "pass by value" happens when we pass our slice to a function, the COPY go makes of the slice is just the SLICE data structure, which goes into another address- BUT, this COPY still points to the SAME ARRAY values the original slice data structure points to.

so when we modify the SLICE, we are not modifying a COPY of the array values for the slice, there is no COPY, we are modifying the original array values from the slice.

slice is a "reference" type - it a data structure that refers to ANOTHER data structure in RAM.

Types

Reference types:

• maps
• channels
• pointers
• functions.

value types:

• int
• float
• string
• bool
• structs.

Maps

# MAP
key --> value
key --> value
key --> value

comparably: map(go) = hash(ruby) = object(javascript) = dict(python)

Maps = statically typed i.e. all keys = same type, all values = same type.

package main

import "fmt"

func main() {

  // 2 x ways of init a "null" map:
  // 1. var colors map[string]string
  // 2. colors := make(map[string]string)

  // example of using type=int
  // colors := make(map[int]string)
  // colors[10] = "#ffffff"
  // delete(colors, 10)
  // fmt.Println(color)

  colors := map[string]string{
    "red":   "#ff0000",
    "green": "#00ff00",
    "white": "#ffffff",
  }

  printMap(colors)
}

func printMap(c map[string]string) {
  // iterate over a map
  for color, hex := range c {
    fmt.Println("Hex code for", color, "is", hex)
  }
}

Always use [] braces with maps e.g. color["white"] = "#ffffff.

Let's break down the func, loop

// c = argument
// map[string]string = type of the argument
// color, hex = the variables that will receive the `key, value` during the loop
// range c = "iterate over the range 'c'"

func printMap(c map[string]string) {
  // iterate over a map
  for color, hex := range c {
    fmt.Println("Hex code for", color, "is", hex)
  }
}

Maps vs Structs

Maps = statically typed = all keys and values must be same type. Struct = values can be different type.

Maps = keys are indexed 0...12, can iterate over. Struct = keys dont support index, can't interate over.

Maps = original data structure of map directly modified Structs = original data structure is copied, and edited, original is unmodified.

When to use Maps?

• when representing a set of collection of closely related properties
• when your scenario doesn't know all keys, types at compile time (otherwise, look at structs if you know)

vast majority of golang = use structs.

Interfaces

Purpose of interfaces

reuse, generic code or code that has common factors - write an interface instead of duplicate code.

package main

import "fmt"

type bot interface {
  getGreeting() string
}

type englishBot struct{}
type spanishBot struct{}

func main() {
  eb := englishBot{}
  sb := spanishBot{}

  printGreeting(eb)
  printGreeting(sb)
}

func printGreeting(b bot) {
  fmt.Println(b.getGreeting())
}

func (eb englishBot) getGreeting() string {
  // very custom logic for generating english greeting
  return "Hi there!"
}

func (sb spanishBot) getGreeting() string {
  return "Hola!"
}

Create a new type bot that says, if you have a getGreeting() function, with return type string, you can be type cast as bot and use what bot can use i.e. printGreeting().

Rules of Interfaces

type bot interface {

  // input args = string, int
  // returns = string, error
  getGreeting(string, int) (string, error)

  // input args = none
  // returns = float
  getBotVersion() float

  // input args = user
  // returns = string
  respondToUser(user) string
}

Concrete Types = map, struct, int, englishBot, string -- can create values directly e.g. int a := 12 Interface Type = bot -- can't create value directly i.e. bot can't "equal 5"

Some key interface points:

1. Interfaces are NOT generic types.
2. Interfaces are implicit i.e. when you declare a type bot interface and then do a type englishBot struct{}, go will implicitly treat englishBot as type bot interface.
3. Interfaces are a contract to help us manages types.
4. Interfaces only help reuse code, doesn't check logic or test things for you, garbage in = garbage out.

The HTTP Package

A program that:

1. HTTP request --> google.com
2. print response to terminal

Get used to following the docs:

e.g. we are looking for where the "Body" of the HTTP request we made is, so we follow the code

https://pkg.go.dev/net/http#Get --> func Get(url string) (resp *Response, err error) --> https://pkg.go.dev/net/http#Response --> Body io.ReadCloser --> https://pkg.go.dev/io#ReadCloser --> type ReadCloser interface --> https://pkg.go.dev/io#Reader

tip

my thought: interfaces re-packages the data into a form that is generic enough, to be consumed by general code e.g. the Reader interface accepts all sources of Input and then "output" it to a byte[] slice, and that's a generic enough type to be handled by anything on the other side of the interface.

type Reader interface {
  Read(p []byte) (n int, err error)
}

this interface, has func Read, accepts input p of type []byte, and returns n int and err error.

Writer Interface

where the Reader is like: source of input --> Reader --> []byte

Writer is like: []byte --> Writer --> source of output

What, in the "standard library", implements the Writer interface? io.Copy()

So this:

  bs := make([]byte, 99999) //fixed byte size 99999 empty elements
  resp.Body.Read(bs)
  fmt.Println(string(bs))

does the same thing as this:

io.Copy(os.Stdout, resp.Body)

where resp is from resp, err := http.Get("http://google.com")

Copy interface, takes two types: func Copy(dst Writer, src Reader) (written int64, err error) a Writer or "something that implements the Writer interface" and a Reader or "something that implements the Reader interface" e.g. io.Copy(os.Stdout, resp.Body) i.e. io.Copy implmements the Writer interface, and resp.Body implements the Reader interface.