code review 6494140: spec: unsafe.Alignof/Sizeof also accept non-variable ar...

spec: unsafe.Alignof/Sizeof also accept non-variable arguments

Both gc and gccgo permit calls such as unsafe.Sizeof(42). The
spec only permits variable arguments. This is a (backward-compatible)
spec change reflecting the status quo. Seems preferrable over
restricting the compilers.

[gri, 2012-09-14 22:06:01]

Hello r, rsc, iant, ken2 (cc: golang-dev@googlegroups.com),

I'd like you to review this change to
https://code.google.com/p/go

[r, 2012-09-14 22:20:11]

http://codereview.appspot.com/6494140/diff/1/doc/go_spec.html
File doc/go_spec.html (right):

http://codereview.appspot.com/6494140/diff/1/doc/go_spec.html#newcode5491
doc/go_spec.html:5491: of any type and return the alignment (or size,
respectively) of a hypothetical variable <code>v</code>
non-parallel construction; drop the parens

The functions Alignof and Size take an expression x ... and return the alignment
or size, respectively 

as for the merit of this: i suppose this is less work than locking down the
compilers.

[gri, 2012-09-14 22:25:47]

PTAL

http://codereview.appspot.com/6494140/diff/1/doc/go_spec.html
File doc/go_spec.html (right):

http://codereview.appspot.com/6494140/diff/1/doc/go_spec.html#newcode5491
doc/go_spec.html:5491: of any type and return the alignment (or size,
respectively) of a hypothetical variable <code>v</code>
On 2012/09/14 22:20:11, r wrote:
> non-parallel construction; drop the parens
> 
> The functions Alignof and Size take an expression x ... and return the
alignment
> or size, respectively 
> 
> as for the merit of this: i suppose this is less work than locking down the
> compilers.

Done.

http://codereview.appspot.com/6494140/diff/1/doc/go_spec.html#newcode5491
doc/go_spec.html:5491: of any type and return the alignment (or size,
respectively) of a hypothetical variable <code>v</code>
On 2012/09/14 22:20:11, r wrote:
> non-parallel construction; drop the parens
> 
> The functions Alignof and Size take an expression x ... and return the
alignment
> or size, respectively 
> 
> as for the merit of this: i suppose this is less work than locking down the
> compilers.

There appear uses for this in the library, e.g.:
http://tip.golang.org/src/pkg/encoding/gob/decode.go?h=unsafe.Sizeof#L367

[r, 2012-09-14 22:27:59]

LGTM

[iant, 2012-09-17 17:05:05]

LGTM

https://codereview.appspot.com/6494140/diff/3/doc/go_spec.html
File doc/go_spec.html (right):

https://codereview.appspot.com/6494140/diff/3/doc/go_spec.html#newcode3
doc/go_spec.html:3: "Subtitle": "Version of September 14, 2012",
s/14/17/

[rsc, 2012-09-17 18:22:32]

LGTM

[gri, 2012-09-17 19:23:44]

*** Submitted as http://code.google.com/p/go/source/detail?r=d27358d77d44 ***

spec: unsafe.Alignof/Sizeof also accept non-variable arguments

Both gc and gccgo permit calls such as unsafe.Sizeof(42). The
spec only permits variable arguments. This is a (backward-compatible)
spec change reflecting the status quo. Seems preferrable over
restricting the compilers.

R=r, rsc, iant, ken
CC=golang-dev
http://codereview.appspot.com/6494140