LGTM
Little comment in the code might be nice too.
 On Jun 10, 2014 6:00 PM, <rsc@golang.org> wrote:
> Reviewers: bradfitz,
>
> Message:
> Hello bradfitz, bradfitz@golang.org (cc:
> golang-codereviews@googlegroups.com),
>
> I'd like you to review this change to
> https://code.google.com/p/go
>
>
> Description:
> net: try to deflake TestVariousDeadlines4Proc
>
> Timeouts are only checked when the non-blocking
> network read returns EAGAIN. If the writer is racing
> far enough ahead of the reader, it is possible that the
> reader may never catch up, and therefore never
> reach the timeout check.
>
> This theory is supported by the fact that it is always
> the longer timeouts that fail in the build failures.
> They give the writer more time to get ahead.
>
> Sleeping in the writer should give the reader time
> to catch up.
>
> Please review this at https://codereview.appspot.com/106950044/
>
> Affected files (+1, -0 lines):
> M src/pkg/net/timeout_test.go
>
>
> Index: src/pkg/net/timeout_test.go
> ===================================================================
> --- a/src/pkg/net/timeout_test.go
> +++ b/src/pkg/net/timeout_test.go
> @@ -409,6 +409,7 @@
> for i := range p {
> p[i] = byte(b)
> }
> + time.Sleep(1 * time.Millisecond)
> return len(p), nil
> }
>
>
>
>