code review 14415043: spec: clarify rules for blank identifiers

spec: clarify rules for blank identifiers

This documents the status quo more precisely.
Not a language change.

Fixes issue 6006.

[gri, 2013-10-05 01:16:38]

Hello r@golang.org (cc: golang-dev@googlegroups.com),

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

[r, 2013-10-05 16:02:36]

https://codereview.appspot.com/14415043/diff/5001/doc/go_spec.html
File doc/go_spec.html (right):

https://codereview.appspot.com/14415043/diff/5001/doc/go_spec.html#newcode1525
doc/go_spec.html:1525: like any other identifier, but it does not introduce a
binding and thus is not
grammar depends on order.

may be used like any other identifier in a declaration, ...

https://codereview.appspot.com/14415043/diff/5001/doc/go_spec.html#newcode2105
doc/go_spec.html:2105: </p>
or short declaration? i know they're not operands on the left of := but this
seems to contradict reality. also the negation with exception makes this hard to
parse. how about:

The <>blank identifier<> may appear as an operand only as the operand on the
left-hand side of an <>assignment<>.

https://codereview.appspot.com/14415043/diff/5001/doc/go_spec.html#newcode4337
doc/go_spec.html:4337: x, _ = f()  // evaluate f() but ignore second result
value
a, _, c = f(), g(), h() // evaluate all three functions but discard return value
of call to g

[gri, 2013-10-07 19:58:11]

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

Please take another look.

[gri, 2013-10-07 19:58:29]

PTAL

https://codereview.appspot.com/14415043/diff/5001/doc/go_spec.html
File doc/go_spec.html (right):

https://codereview.appspot.com/14415043/diff/5001/doc/go_spec.html#newcode1525
doc/go_spec.html:1525: like any other identifier, but it does not introduce a
binding and thus is not
On 2013/10/05 16:02:36, r wrote:
> grammar depends on order.
> 
> may be used like any other identifier in a declaration, ...

Done.

https://codereview.appspot.com/14415043/diff/5001/doc/go_spec.html#newcode2105
doc/go_spec.html:2105: </p>
On 2013/10/05 16:02:36, r wrote:
> or short declaration? i know they're not operands on the left of := but this
> seems to contradict reality. also the negation with exception makes this hard
to
> parse. how about:
> 
> The <>blank identifier<> may appear as an operand only as the operand on the
> left-hand side of an <>assignment<>.

Done.

[r, 2013-10-07 22:08:07]

LGTM but wait for others

https://codereview.appspot.com/14415043/diff/18001/doc/go_spec.html
File doc/go_spec.html (right):

https://codereview.appspot.com/14415043/diff/18001/doc/go_spec.html#newcode1597
doc/go_spec.html:1597: The <i>blank identifier</i> is the underscore character
<code>_</code>.
s/is/is represented by/

https://codereview.appspot.com/14415043/diff/18001/doc/go_spec.html#newcode4287
doc/go_spec.html:4287: <code>y</code> where <i>op</i> is a binary arithmetic
operation is equivalent
s2/is//

[gri, 2013-10-07 22:26:14]

PTAL

I also added the word "first" on line 4383 since untyped constants, once
converted to their default type must still be assignable after the conversion
(they are not automatically assignable to say, a non-empty interface).

https://codereview.appspot.com/14415043/diff/18001/doc/go_spec.html
File doc/go_spec.html (right):

https://codereview.appspot.com/14415043/diff/18001/doc/go_spec.html#newcode1597
doc/go_spec.html:1597: The <i>blank identifier</i> is the underscore character
<code>_</code>.
On 2013/10/07 22:08:07, r wrote:
> s/is/is represented by/

Done.

https://codereview.appspot.com/14415043/diff/18001/doc/go_spec.html#newcode4287
doc/go_spec.html:4287: <code>y</code> where <i>op</i> is a binary arithmetic
operation is equivalent
On 2013/10/07 22:08:07, r wrote:
> s2/is//

Done.

[iant, 2013-10-08 00:15:27]

LGTM

https://codereview.appspot.com/14415043/diff/24001/doc/go_spec.html
File doc/go_spec.html (right):

https://codereview.appspot.com/14415043/diff/24001/doc/go_spec.html#newcode4392
doc/go_spec.html:4392: If the left-hand side is the blank identifier, any typed
or non-constant
Perhaps s/the left-hand side/a left-hand operand/, because there may be multiple
operands on the left-hand side.

[gri, 2013-10-08 00:21:58]

https://codereview.appspot.com/14415043/diff/24001/doc/go_spec.html
File doc/go_spec.html (right):

https://codereview.appspot.com/14415043/diff/24001/doc/go_spec.html#newcode4392
doc/go_spec.html:4392: If the left-hand side is the blank identifier, any typed
or non-constant
On 2013/10/08 00:15:28, iant wrote:
> Perhaps s/the left-hand side/a left-hand operand/, because there may be
multiple
> operands on the left-hand side.

Done.

[rsc, 2013-11-13 01:43:21]

LGTM

[rsc, 2013-11-13 02:07:01]

*** Submitted as https://code.google.com/p/go/source/detail?r=4daa80747394 ***

spec: clarify rules for blank identifiers

This documents the status quo more precisely.
Not a language change.

Fixes issue 6006.

R=r, rsc, iant, ken
CC=golang-dev
https://codereview.appspot.com/14415043

Committer: Russ Cox <rsc@golang.org>