Introduce anonymous structs

[niyumard]

No description provided.

[eliben]

Thanks for the contribution, it's useful and I'm inclined to accept it.

Let's work out the details: I feel this comment is inaccurate because dog is in fact an instance of a type, just a type that isn't named. It should say something more to the tune of unnamed struct type (but also the word anonymous should be there.

[niyumard]

@eliben Sorry for the late response.
Which one do you think is better?

Structs don't always have to be defined as instances of named types, you can define a struct anonymously.

or

Structs can be instances of unnamed types, therefore you can define structs anonymously.

or we could simply go with "You can define structs anonymously".

[eliben]

I'll approve this for now and will tweak the comment a bit

[niyumard]

I'll approve this for now and will tweak the comment a bit

@eliben
Thanks! But I'd argue against mentioning "used for a single value" because it could be misleading. My main use of anonymous structs is for table driven tests where I can do something like this.
Maybe table driven tests should go hand in hand with #421 and have their own section in the site though.
Here's a code example that could be used in case you're interested in introducing slices of anonymous structs (which could be more useful):

dog := []struct {
	name   string
	isGood bool
}{
	{"Rex", true},
	{"Bruiser", false},
}
fmt.Println(dog)

Its result:
[{Rex true} {Teddy false}]

[eliben]

We actually do have an example of table-driven tests already in https://gobyexample.com/testing-and-benchmarking -- I will add a link to it from this anonymous example.

"Value" is correct, even if it refers to a value of a slice type.

[niyumard]

Looks great, thank you.