code review 5622053: test: test slice beyond len

test: test slice beyond len

When slicing a slice, the bounds may be > len as long as they
are <= cap.  Interestingly, gccgo got that wrong and still
passed the testsuite and all the library tests.

[iant, 2012-02-02 23:05:59]

Hello golang-dev@googlegroups.com,

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

[rsc, 2012-02-02 23:06:27]

LGTM

[rsc, 2012-02-02 23:07:02]

It must be the lower bound that was the problem?
Surely the upper bound being > len is all over the place.

[iant2, 2012-02-03 14:29:03]

Russ Cox <rsc@golang.org> writes:

> It must be the lower bound that was the problem?
> Surely the upper bound being > len is all over the place.

Yes, I was comparing the lower bound against len rather than cap.

Ian

[iant, 2012-02-03 14:29:38]

*** Submitted as http://code.google.com/p/go/source/detail?r=04ca0ed4614a ***

test: test slice beyond len

When slicing a slice, the bounds may be > len as long as they
are <= cap.  Interestingly, gccgo got that wrong and still
passed the testsuite and all the library tests.

R=golang-dev, rsc, iant
CC=golang-dev
http://codereview.appspot.com/5622053

[rsc, 2012-02-03 14:41:33]

On Fri, Feb 3, 2012 at 09:28, Ian Lance Taylor <iant@google.com> wrote:
> Yes, I was comparing the lower bound against len rather than cap.

FWIW, I believe that for [lo:hi] the simplest check is
lo <= hi && hi <= cap; lo doesn't have to be compared
against either len or cap.

Russ

[iant2, 2012-02-03 16:38:13]

Russ Cox <rsc@golang.org> writes:

> On Fri, Feb 3, 2012 at 09:28, Ian Lance Taylor <iant@google.com> wrote:
>> Yes, I was comparing the lower bound against len rather than cap.
>
> FWIW, I believe that for [lo:hi] the simplest check is
> lo <= hi && hi <= cap; lo doesn't have to be compared
> against either len or cap.

Yeah, I will do that when I disentangle ordinary index expressions from
slice expressions.

Ian