code review 1081041: go spec: clarification of make arguments,

go spec: clarification of make arguments,
specification of runtime exceptions for
make and division by zero

[gri, 2010-05-03 20:24:01]

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

I'd like you to review this change.

[iant, 2010-05-03 23:45:52]

LGTM

[r, 2010-05-04 17:17:17]

http://codereview.appspot.com/1081041/diff/1/2
File doc/go_spec.html (right):

http://codereview.appspot.com/1081041/diff/1/2#newcode2926
doc/go_spec.html:2926: occurs or not is implementation-specific.
s/or not /

("whether" already includes that option)

http://codereview.appspot.com/1081041/diff/1/2#newcode4427
doc/go_spec.html:4427: Call             Type of T       Result
T is already a type; what does Type of T mean?

http://codereview.appspot.com/1081041/diff/1/2#newcode4429
doc/go_spec.html:4429: make(T, n)       slice type      slice T with length n
and capacity n
drop all the 'type' words in the middle column

the right hand column's T doesn't seem to be the same T as the left hand
column's.

[gri, 2010-05-04 19:51:02]

PTAL
- Robert

On Tue, May 4, 2010 at 10:17 AM, <r@golang.org> wrote:

>
> http://codereview.appspot.com/1081041/diff/1/2
> File doc/go_spec.html (right):
>
> http://codereview.appspot.com/1081041/diff/1/2#newcode2926
> doc/go_spec.html:2926: occurs or not is implementation-specific.
> s/or not /
>
> ("whether" already includes that option)
>
> http://codereview.appspot.com/1081041/diff/1/2#newcode4427
> doc/go_spec.html:4427: Call             Type of T       Result
> T is already a type; what does Type of T mean?
>
> http://codereview.appspot.com/1081041/diff/1/2#newcode4429
> doc/go_spec.html:4429: make(T, n)       slice type      slice T with
>
> length n and capacity n
> drop all the 'type' words in the middle column
>
> the right hand column's T doesn't seem to be the same T as the left hand
> column's.
>
>
> http://codereview.appspot.com/1081041/show
>

[r, 2010-05-04 19:53:38]

http://codereview.appspot.com/1081041/diff/6001/7001
File doc/go_spec.html (right):

http://codereview.appspot.com/1081041/diff/6001/7001#newcode4435
doc/go_spec.html:4435: make(T)          channel    synchronous channel of type T
with no buffer
synchronous and no buffer mean the same thing.
how about
synchronous channel of type T
asynchronous channel of type T, buffer size n

[rsc, 2010-05-04 19:56:25]

The restriction is actually 0 <= n <= m <= maxlen
where maxlen is the maximum positive value
representable by type int.

And then there's another restriction: for channel
and slice types, m <= addressspace/sizeof(element) -
you can't allocate more memory than there is
virtual address space.

The first restriction is probably worth stating,
the second probably not.

Russ

[gri, 2010-05-04 21:35:48]

PTAL

On Tue, May 4, 2010 at 12:56 PM, Russ Cox <rsc@golang.org> wrote:

> The restriction is actually 0 <= n <= m <= maxlen
> where maxlen is the maximum positive value
> representable by type int.
>
> And then there's another restriction: for channel
> and slice types, m <= addressspace/sizeof(element) -
> you can't allocate more memory than there is
> virtual address space.
>
> The first restriction is probably worth stating,
> the second probably not.
>
> Russ
>

[rsc1, 2010-05-04 21:41:24]

LGTM

[r, 2010-05-05 00:16:15]

LGTM

http://codereview.appspot.com/1081041/diff/13001/14001
File doc/go_spec.html (right):

http://codereview.appspot.com/1081041/diff/13001/14001#newcode4444
doc/go_spec.html:4444: <code>m</code> don't fit into an <code>int</code>.
s/don't fit into/cannot be represented by/ or maybe overflow?

[gri, 2010-05-05 00:31:44]

*** Submitted as http://code.google.com/p/go/source/detail?r=bb59d0168350 ***

go spec: clarification of make arguments,
specification of runtime exceptions for
make and division by zero

R=r, rsc, ken2, iant
CC=golang-dev
http://codereview.appspot.com/1081041