Refactor nag functionality in to NagTimer class.

Refactor nag functionality in to NagTimer class.

Add default 30 second nag timer to gclient subprocesses.

BUG=227537
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=198207

[szager1, 2013-05-02 06:49:10]

I just realized that the headline description is not the most important part of
this change.  It's this second line that's significant:

"Add default 30 second nag timer to gclient subprocesses."

[M-A Ruel, 2013-05-03 16:42:58]

https://codereview.chromium.org/14826003/diff/4001/subprocess2.py
File subprocess2.py (right):

https://codereview.chromium.org/14826003/diff/4001/subprocess2.py#newcode162
subprocess2.py:162: sleep_time = (self.last_output - now - 0.1) % self.interval
% with float? You should use an explicit cast in that case and also in
self.internal = int(interval)

[cmp_google, 2013-05-03 17:44:56]

lgtm

[szager1, 2013-05-03 18:16:58]

It occurs to me that committing this change as-is could cause builders to stall
during update.  If gclient gets stuck, but produces regular nag messages, then I
don't think buildbot will timeout.

I'm going to add "kill after three nag messages" before committing.

https://codereview.chromium.org/14826003/diff/4001/subprocess2.py
File subprocess2.py (right):

https://codereview.chromium.org/14826003/diff/4001/subprocess2.py#newcode162
subprocess2.py:162: sleep_time = (self.last_output - now - 0.1) % self.interval
On 2013/05/03 16:42:58, Marc-Antoine Ruel wrote:
> % with float? You should use an explicit cast in that case and also in
> self.internal = int(interval)

Can you explain why?  I can't think of a good reason to forbid floats here.  The
float type in python fully supports __mod__.

[M-A Ruel, 2013-05-03 18:32:34]

lgtm

https://codereview.chromium.org/14826003/diff/4001/subprocess2.py
File subprocess2.py (right):

https://codereview.chromium.org/14826003/diff/4001/subprocess2.py#newcode162
subprocess2.py:162: sleep_time = (self.last_output - now - 0.1) % self.interval
On 2013/05/03 18:16:58, szager1 wrote:
> On 2013/05/03 16:42:58, Marc-Antoine Ruel wrote:
> > % with float? You should use an explicit cast in that case and also in
> > self.internal = int(interval)
> 
> Can you explain why?  I can't think of a good reason to forbid floats here. 
The
> float type in python fully supports __mod__.

Ah ok I hadn't realized that.

[szager1, 2013-05-03 19:03:48]

Added nag_max, to timeout after a certain number of nags.

gclient uses nag_max=3.

PTAL

[cmp, 2013-05-03 21:51:27]

lgtm, +1 on nag_max, this is what I had in mind, too, no sense in waiting 10
mins for a hang whose cause will not have changed in the last 8m30s

[commit-bot: I haz the power, 2013-05-03 22:04:58]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/szager@chromium.org/14826003/14001

[commit-bot: I haz the power, 2013-05-03 22:06:38]

Change committed as 198207