Let's say I have a Go struct that looks something like this (please pardon syntax for the sake of explanation):
type Foo struct {
MyBar *Bar
}
type Bar struct {
value int
}
func (b* Bar) Bark() {
fmt.Printf("Hello I'm a bar with value: %v", b.value)
}
Now, let's say that this Foo struct is stored in memory on a process, and it takes some time for the MyBar field to be populated. As in, it is nil for sometime. Is there any way that I can check if its save to access the Bark() method at run time without manually doing nil checks?
For example, something like:
// Early on, this happens:
f := &Foo{}
.
.
.
.
// And later, I wanna see if its safe to do:
if isSafe(b.MyBar) {
b.MyBar.Bar()
}
The reason I ask this is for deeply nested pointers, this can become pretty painful. Any thoughts?
There is no shortcut for deep checking a struct for nil pointers. Your isSafe is essentially a nil check:
if b.MyBar!=nil {
b.MyBar.Bark()
}
Alternatively, you can also check for nil receiver in the method:
func (b* Bar) Bark() {
if b!=nil {
fmt.Printf("Hello I'm a bar with value: %v", b.value)
}
}
This last usage may hide bugs in your program, though, because it will run without panicking even when the receiver is not supposed to be nil.
Related
I have this problem which seems a bit weird to me. Take a look at this snippet of code:
package coreinterfaces
type FilterInterface interface {
Filter(s *string) bool
}
type FieldFilter struct {
Key string
Val string
}
func (ff *FieldFilter) Filter(s *string) bool {
// Some code
}
type FilterMapInterface interface {
AddFilter(f *FilterInterface) uuid.UUID
RemoveFilter(i uuid.UUID)
GetFilterByID(i uuid.UUID) *FilterInterface
}
type FilterMap struct {
mutex sync.Mutex
Filters map[uuid.UUID]FilterInterface
}
func (fp *FilterMap) AddFilter(f *FilterInterface) uuid.UUID {
// Some code
}
func (fp *FilterMap) RemoveFilter(i uuid.UUID) {
// Some code
}
func (fp *FilterMap) GetFilterByID(i uuid.UUID) *FilterInterface {
// Some code
}
On some other package, I have the following code:
func DoFilter() {
fieldfilter := &coreinterfaces.FieldFilter{Key: "app", Val: "152511"}
filtermap := &coreinterfaces.FilterMap{}
_ = filtermap.AddFilter(fieldfilter) // <--- Exception is raised here
}
The run-time won't accept the line mentioned because
"cannot use fieldfilter (type *coreinterfaces.FieldFilter) as type
*coreinterfaces.FilterInterface in argument to fieldint.AddFilter:
*coreinterfaces.FilterInterface is pointer to interface, not interface"
However, when changing the code to:
func DoBid() error {
bs := string(b)
var ifilterfield coreinterfaces.FilterInterface
fieldfilter := &coreinterfaces.FieldFilter{Key: "app", Val: "152511"}
ifilterfield = fieldfilter
filtermap := &coreinterfaces.FilterMap{}
_ = filtermap.AddFilter(&ifilterfield)
}
Everything is alright and when debugging the application it really seems to include
I'm a bit confused on this topic. When looking at other blog posts and stack overflow threads discussing this exact same issue (for example - This, or
This) the first snippet which raises this exception should work, because both fieldfilter and fieldmap are initialized as pointers to interfaces, rather than value of interfaces. I haven't been able to wrap my head around what actually happens here that I need to change in order for me not to declare a FieldInterface and assign the implementation for that interface. There must be an elegant way to do this.
So you're confusing two concepts here. A pointer to a struct and a pointer to an interface are not the same. An interface can store either a struct directly or a pointer to a struct. In the latter case, you still just use the interface directly, not a pointer to the interface. For example:
type Fooer interface {
Dummy()
}
type Foo struct{}
func (f Foo) Dummy() {}
func main() {
var f1 Foo
var f2 *Foo = &Foo{}
DoFoo(f1)
DoFoo(f2)
}
func DoFoo(f Fooer) {
fmt.Printf("[%T] %+v\n", f, f)
}
Output:
[main.Foo] {}
[*main.Foo] &{}
https://play.golang.org/p/I7H_pv5H3Xl
In both cases, the f variable in DoFoo is just an interface, not a pointer to an interface. However, when storing f2, the interface holds a pointer to a Foo structure.
Pointers to interfaces are almost never useful. In fact, the Go runtime was specifically changed a few versions back to no longer automatically dereference interface pointers (like it does for structure pointers), to discourage their use. In the overwhelming majority of cases, a pointer to an interface reflects a misunderstanding of how interfaces are supposed to work.
However, there is a limitation on interfaces. If you pass a structure directly into an interface, only value methods of that type (ie. func (f Foo) Dummy(), not func (f *Foo) Dummy()) can be used to fulfill the interface. This is because you're storing a copy of the original structure in the interface, so pointer methods would have unexpected effects (ie. unable to alter the original structure). Thus the default rule of thumb is to store pointers to structures in interfaces, unless there's a compelling reason not to.
Specifically with your code, if you change the AddFilter function signature to:
func (fp *FilterMap) AddFilter(f FilterInterface) uuid.UUID
And the GetFilterByID signature to:
func (fp *FilterMap) GetFilterByID(i uuid.UUID) FilterInterface
Your code will work as expected. fieldfilter is of type *FieldFilter, which fullfills the FilterInterface interface type, and thus AddFilter will accept it.
Here's a couple of good references for understanding how methods, types, and interfaces work and integrate with each other in Go:
https://medium.com/#agileseeker/go-interfaces-pointers-4d1d98d5c9c6
https://www.goinggo.net/2014/05/methods-interfaces-and-embedded-types.html
https://blog.golang.org/laws-of-reflection
GetFilterByID(i uuid.UUID) *FilterInterface
When I get this error, it's usually because I'm specifying a pointer to an interface instead of an interface ( that will actually be a pointer to my struct that fulfills the interface ).
There's a valid use for *interface{...} but more commonly I just am thinking 'this is a pointer' instead of 'this is an interface which happens to be a pointer in the code I'm writing'
This question already has answers here:
My object is not updated even if I use the pointer to a type to update it
(3 answers)
Closed 4 years ago.
here is a golang behavior that I am trying to understand and change: I wrote a method to populate a structure with slices in Golang. It works within the method itself, but the slice content gets lost outside of the method. I want however to keep the content. It probably comes from the fact that the pointers inside the slice where deleted at the end of the populateslice method, but how should I write it to prevent this to happen, ie. keep the content in mystruct.myslice after the function call ?
Here is how I wrote the code:
type BBDatacolumn struct {
Data []string
}
type Mystruct struc {
myslice []BBDatacolumn
}
//Method to populate the slice of the structure mystruct:
func (self mystruct) populateslice() {
for i:=0; i<imax; i++ {
bufferdatacolumn := NewBBDatacolumn()
//Here, code to populate bufferdatacolumns
self.myslice = append(self.myslice, bufferdatacolumn)
}
self.myslice.display() //Here, works fine: myslice contains the data of the BBDatacolumn correctly
}
//Later in the code (outside of the populateslice func):
mystructinstance.populateslice() //Populates slice OK at the end of the function
mystructinstance.display() //Problem: mystructinstance.myslice is empty: Instanciation of Mystruct does not contain the data in myslice anymore as it did inside the populateslice method
The method needs to be "on" a pointer of your struct, like so:
func (self *mystruct) Foo () {}
Otherwise the mystruct object that you call the method on is local only to the function.
In go the receiver of a method (the part between func and the method name) can either be a value receiver as you have or a pointer receiver. When you have a value receiver a copy of the object is passed to the method so any modifications made remain in that copy. If you want to modify the object then you must have a pointer receiver as so:
func (self *mystruct) populateslice() {
For a wider discussion of which is best in general see here:
Value receiver vs. Pointer receiver in Golang?
So let's say that we have a function of the following form:
func WorkMagic(obj interface{}) interface{} {
switch t := obj.(type) {
case string:
// Do string magic
default:
// Do slice magic
}
...
}
I am expecting obj to be either a string or a slice, which I can ascertain via the switch. In the case of a slice, I want to be able to do ordering work on any arbitrary slice, regardless of type. Seems like the best way to accomplish this is using the unsafe package in a similar fashion to that discussed in this article.
Here however, the function accepts a specific type of slice ([]string), whereas I would like to be able to work on any slice. So the question is, given that I am accepting an empty interface as input, how might I access the underlying slice / array using unsafe.Pointer so as to be able to loop through and modify which value is associate with which index?
You'll want to use reflection. It enables you to work generically without giving up type and memory safety like unsafe would. Read the Go blog's Laws of Reflection.
func actOnSlices(i interface{}) {
v := reflect.ValueOf(i)
for v.Kind() == reflect.Ptr { // dereference pointers
v = v.Elem()
}
if v.Kind() != reflect.Slice { // ensure you actually got a slice
panic("given argument is not a slice")
}
// do slice stuff
}
Edit to answer your second question:
Yes – this can be done: elements of a slice are adressable and hence settable. See the following working example:
package main
import (
"fmt"
"reflect"
)
func main() {
s := []string{"foo", "bar"}
fmt.Println(swapIndexes(s, 0, 1)) // prints [bar foo]
}
func swapIndexes(i interface{}, x, y int) interface{} {
v := reflect.ValueOf(i)
for v.Kind() == reflect.Ptr { // dereference pointers
v = v.Elem()
}
if v.Kind() != reflect.Slice { // ensure you actually got a slice
panic("given argument is not a slice")
}
t := v.Index(x).Interface()
v.Index(x).Set(v.Index(y))
v.Index(y).Set(reflect.ValueOf(t))
return v.Interface()
}
Edit to answer your third question:
The unsafe package is not something you'll encounter much in user-land code. It exists to implement certain features (e.g. reflection, C interaction) that need to circumvent Go's safety guarantees to work. Using unsafe is unsafe, as the name suggests, because you can mess up big time without even realizing. By using unsafe, you're incurring in a big trade-off, so it better be worth it. Quoting #twotwotwo:
The downside of unsafe is that if you mess up you're in the old days of segfaults, memory corruption, and buffer-overflow security holes.
Also, as #twotwotwo suggested; it's more "Go-like" to repeat code than using reflection to achieve genericity.
To Go's type-system, []string and []int are two completely separate and unrelated types. just as int and string would be. The relation (both are slices) is obvious only to the programmer. There is no way of expressing "a slice" without saying a slice of what.
I have an issue using this statement
m.Post(Model, binding.Form(Wish), func(wish Wish, r render.Render, db *mgo.Database) {
This worked fine if I use the struct define inside the prog like
m.Post(Model, binding.Form(Wish1{}) , func(wish Wish1, r render.Render, db *mgo.Database) {
but I need this to be an independent package.
I get "Wish is not a type" wish is the return of the binding function.
This worked with a primary Type struct. I am passing the strut as a interface{}
I am using GO with Martini.Classic() It is really complicated for me to change Martini or Binding package. Any suggestions.
This is the all code
package chlistpkg
import (
"github.com/codegangsta/martini"
"github.com/codegangsta/martini-contrib/binding"
"github.com/codegangsta/martini-contrib/render"
"labix.org/v2/mgo"
"time"
"fmt"
"html/template"
"reflect"
"adminStruct"
)
just to show the struct that I need to pass as to routine Doall
type Wish1 struct {
Name string `form:"name"`
Description string `form:"description"`
AnyDate time.Time `form:"anydate"`
Active bool `form:"active"`
Number int `form:"number"`
NumDec float32 `form:"numDec"`
}
DB Returns a martini.Handler
func DB() martini.Handler {
session, err := mgo.Dial("mongodb://localhost")
if err != nil {
panic(err)
}
return func(c martini.Context) {
s := session.Clone()
c.Map(s.DB("advent2"))
defer s.Close()
c.Next()
}
}
GetAll returns all Wishes in the database
func GetAll(db *mgo.Database, entList interface{}) interface{} {
db.C("wishes").Find(nil).All(entList)
fmt.Println("GettAll entList =", entList)
return entList
}
func Doall(Model string, Wish interface{}, Wish2 interface{}, Wishlist interface{} ) {
m := martini.Classic()
fmt.Println ("martini.Classic =", m)
m.Use(martini.Static("images")) // serve from the "images" directory as well
m.Use(render.Renderer(render.Options{
Directory: "templates",
Layout: "layout",
}))
m.Use(DB())
m.Get(Model, func(r render.Render, db *mgo.Database) {
r.HTML(200, "lista4", GetAll(db, Wishlist))
})
binding does not take a pointer. I have to pass the struct by reference on "Wish"
the issue is the return on "wish Wish" I got an error Wish is not a type
at compilation time
m.Post(Model, binding.Form(Wish), func(wish Wish, r render.Render, db *mgo.Database) {
fmt.Println("Input wish =", wish)
db.C("wishes").Insert(wish)
r.HTML(200, "lista4", GetAll(db, Wishlist))
})
m.Run()
Thanks in advance
Luis
The reason you are getting an error is that you have called your type Wish1 (with a numerical 1) but you are referring to the Wish type (which does not exist!) in your code.
Change your struct to be:
// Note: "Wish", not "Wish1"
type Wish struct {
Name string `form:"name"`
Description string `form:"description"`
AnyDate time.Time `form:"anydate"`
Active bool `form:"active"`
Number int `form:"number"`
NumDec float32 `form:"numDec"`
}
If you want to put your type into another package (tip: don't overdo the sub-packages), then it will need to become a pkgname.Wish as names are fully qualified.
Added
After a second look, you're also messing things up here:
func Doall(Model string, Wish interface{}, Wish2 interface{}, Wishlist interface{} ) {
m := martini.Classic()
fmt.Println ("martini.Classic =", m)
m.Use(martini.Static("images")) // serve from the "images" directory as well
Your parameter list needs to provide a name for each type; you can't pass Wish interface{} as a parameter as Wish is a type, not a variable name.
You should either:
func DoAll(model string, wish interface{}, wish2 interface{}, wishList interface{}) { ... }
Or, better still, stop using interface{} like this and write:
func DoAll(model string, wishList []Wish, wishes... Wish) { ... }
However, your DoAll function does not seem to be referenced elsewhere, and is creating its own Martini instance. I highly suggest thinking about why things are "split out" like this if you're just starting out. Keep it simple - e.g.
func main() {
m := martini.Classic()
m.Use(martini.Static("images"))
m.Use(DB())
m.Use(render.Renderer(render.Options{...}))
// No need for an anonymous function, which just adds clutter
m.Get("/wishes/all", GetAllWishes)
// Same goes for here
m.Post("/wishes/new", PostWish)
m.Run()
}
PS: I've fixed the formatting of your code, as it has a lot of unnecessary spacing before/after parenthesis. Make sure to use gofmt, which is included with the Go install and can be hooked into most popular editors.
Correction:
Link #1 http://play.golang.org/p/CKRNyWYF8X
Link #2 http://play.golang.org/p/oT2yKzFwep
From the first link,
I am sure that the panic error comes from this
func (A *DoublyLinkedList) AddHead(input_value interface{}) {
temp_node := &Node{value: input_value, prev: nil, next: A.head}
original_head_node := A.head
original_head_node.prev = temp_node
A.length++
}
But when I use this for doubly linked list, it panics little later. And still fails because this one below does not connect the original head with previous pointer.
func (A *DoublyLinkedList) AddHead(input_value interface{}) {
A.head = NewNode(input_value, nil, A.head)
A.length++
}
This is the one. This one has the similar problem.
cannot assign to target_node.GetPrevNode().GetNextNode()
Does go not support pointer reference this way? I did fix this just assigning a new variable every time I need to get the pointer. But my first question on the top still does not compile.
In short, how do I connect the doubly linked list when adding a new element in Go?
You need to initialize the properties inside the DoublyLinkedList. It seems to me, you are currently creating a reference to it in NewDoublyLinkedList() with 2 nil properties.
type DoublyLinkedList struct {
head *Node // nil
tail *Node // nil
length int
}
And when doing this
original_head_node := A.head // A.head == nil
original_head_node.prev = temp_node // You are trying to access a property in nil