code review 5702062: spec: apply method sets, embedding to all types, not ju...

spec: apply method sets, embedding to all types, not just named types

When we first wrote the method set definition, we had long
discussions about whether method sets applied to all types
or just named types, and we (or at least I) concluded that it
didn't matter: the two were equivalent points of view, because
the only way to introduce a new method was to write a method
function, which requires a named receiver type.

However, the addition of embedded types changed this.
Embedding can introduce a method without writing an explicit
method function, as in:

        var x struct {
                sync.Mutex
        }

        var px *struct {
                sync.Mutex
        }

        var _, _ sync.Locker = &x, px

The edits in this CL make clear that both &x and px satisfy
sync.Locker.  Today, gccgo already works this way; 6g does not.

[rsc, 2012-02-29 16:15:31]

Hello golang-dev@googlegroups.com,

I'd like you to review this change to
https://go.googlecode.com/hg/

[gri, 2012-02-29 16:24:52]

LGTM+++

gri

On Wed, Feb 29, 2012 at 8:15 AM,  <rsc@golang.org> wrote:
> Reviewers: golang-dev_googlegroups.com,
>
> Message:
> Hello golang-dev@googlegroups.com,
>
> I'd like you to review this change to
> https://go.googlecode.com/hg/
>
>
> Description:
> spec: apply method sets, embedding to all types, not just named types
>
> When we first wrote the method set definition, we had long
> discussions about whether method sets applied to all types
> or just named types, and we (or at least I) concluded that it
> didn't matter: the two were equivalent points of view, because
> the only way to introduce a new method was to write a method
> function, which requires a named receiver type.
>
> However, the addition of embedded types changed this.
> Embedding can introduce a method without writing an explicit
> method function, as in:
>
>        var x struct {
>                sync.Mutex
>        }
>
>        var px *struct {
>                sync.Mutex
>        }
>
>        var _, _ sync.Locker = &x, px
>
> The edits in this CL make clear that both &x and px satisfy
> sync.Locker.  Today, gccgo disagrees about px and
> 6g disagrees about both.  This is a language change,
> but one that resolves a disagreement between gccgo and 6g
> (by telling them that they're both wrong).
>
> Please review this at http://codereview.appspot.com/5702062/
>
> Affected files:
>  M doc/go_spec.html
>
>
> Index: doc/go_spec.html
> ===================================================================
> --- a/doc/go_spec.html
> +++ b/doc/go_spec.html
> @@ -678,7 +678,7 @@
>  A type may have a <i>method set</i> associated with it
>  (§<a href="#Interface_types">Interface types</a>, §<a
> href="#Method_declarations">Method declarations</a>).
>  The method set of an <a href="#Interface_types">interface type</a> is its
> interface.
> -The method set of any other named type <code>T</code>
> +The method set of any other type <code>T</code>
>  consists of all methods with receiver type <code>T</code>.
>  The method set of the corresponding pointer type <code>*T</code>
>  is the set of all methods with receiver <code>*T</code> or <code>T</code>
> @@ -954,7 +954,7 @@
>  <p>
>  Fields and methods (§<a href="#Method_declarations">Method
> declarations</a>) of an anonymous field are
>  promoted to be ordinary fields and methods of the struct (§<a
> href="#Selectors">Selectors</a>).
> -The following rules apply for a struct type named <code>S</code> and
> +The following rules apply for a struct type <code>S</code> and
>  a type named <code>T</code>:
>  </p>
>  <ul>
>
>

[rsc, 2012-02-29 16:27:50]

Ian corrected me about gccgo: this new spec is exactly what gccgo
already does.  I updated the CL description to reflect that.

[iant, 2012-02-29 16:38:53]

LGTM

[gri, 2012-02-29 17:49:32]

http://codereview.appspot.com/5702062/diff/4/doc/go_spec.html
File doc/go_spec.html (right):

http://codereview.appspot.com/5702062/diff/4/doc/go_spec.html#newcode681
doc/go_spec.html:681: The method set of any other type <code>T</code>
I think we should say something here about embedded types. As it stands it's
still confusing and one has to read both, this section and the section on
embedding in structs.

I am fine w/ doing this in a separate "cleanup" CL that doesn't make changes but
consolidates the method set rules.

[r, 2012-02-29 20:44:18]

LGTM
agree with gri, but in a second CL is fine and maybe better.

http://codereview.appspot.com/5702062/diff/4/doc/go_spec.html
File doc/go_spec.html (right):

http://codereview.appspot.com/5702062/diff/4/doc/go_spec.html#newcode3
doc/go_spec.html:3: "Subtitle": "Version of February 23, 2012"
this again

[rsc, 2012-02-29 20:54:12]

*** Submitted as http://code.google.com/p/go/source/detail?r=b1f0c3b7a935 ***

spec: apply method sets, embedding to all types, not just named types

When we first wrote the method set definition, we had long
discussions about whether method sets applied to all types
or just named types, and we (or at least I) concluded that it
didn't matter: the two were equivalent points of view, because
the only way to introduce a new method was to write a method
function, which requires a named receiver type.

However, the addition of embedded types changed this.
Embedding can introduce a method without writing an explicit
method function, as in:

        var x struct {
                sync.Mutex
        }

        var px *struct {
                sync.Mutex
        }

        var _, _ sync.Locker = &x, px

The edits in this CL make clear that both &x and px satisfy
sync.Locker.  Today, gccgo already works this way; 6g does not.

R=golang-dev, gri, iant, r
CC=golang-dev
http://codereview.appspot.com/5702062