code review 101400043: runtime: add fake time support back.

runtime: add fake time support back.
Revived from CL 15690048.

Fixes issue 5356.

[minux, 2014-06-20 21:54:12]

Hello adg@golang.org (cc: golang-codereviews@googlegroups.com),

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

[dvyukov, 2014-06-21 00:57:49]

what about the deadlock scenario described in
https://codereview.appspot.com/73110043/
?

https://codereview.appspot.com/101400043/diff/40001/src/pkg/runtime/time.goc
File src/pkg/runtime/time.goc (right):

https://codereview.appspot.com/101400043/diff/40001/src/pkg/runtime/time.goc#...
src/pkg/runtime/time.goc:201: if(runtime·timens == 0) return nil;
return on the next line please

[rsc, 2014-10-27 23:42:22]

LGTM

The possible deadlock with external threads is not a problem. This code only
makes sense if there are no external threads. The fundamental idea is that the
clock jumps forward when everything is blocked until a given time; the clock
jumps to that time. If there are external threads, the runtime fundamentally
cannot know this. This code is not intended to be general. It is only for the
playground, where there are no external threads.

I will patch this in, fix the formatting, test that it works, and then submit
it.

[minux, 2014-10-27 23:58:30]

On Mon, Oct 27, 2014 at 7:42 PM,  <rsc@golang.org> wrote:
> I will patch this in, fix the formatting, test that it works, and then
> submit it.
Thanks Russ.

[rsc, 2014-10-28 00:35:19]

*** Submitted as https://code.google.com/p/go/source/detail?r=daf0179cc576 ***

runtime: add fake time support back.
Revived from CL 15690048.

Fixes issue 5356.

LGTM=rsc
R=adg, dvyukov, rsc
CC=golang-codereviews
https://codereview.appspot.com/101400043

Committer: Russ Cox <rsc@golang.org>

[adg1, 2014-10-28 00:45:01]

The submitted change also includes uncommenting these lines:

https://code.google.com/p/go/source/diff?spec=svndaf0179cc57600d8a960cf857ef2...

Is this right? Do we want fake time to be enabled by default for all NaCl
programs? I didn't think so.

[gobot, 2014-10-28 00:48:24]

This CL appears to have broken the nacl-amd64p32 builder.
See http://build.golang.org/log/a0d2cb09845055cc6043acb4f928359615894c27