Conver APIListDataSource to use Future.Then() rather than Get().

Convert APIListDataSource to use Future.Then() rather than Get().
I needed to change the behaviour of Future.Then() to implement this.

BUG=306341
R=yoz@chromium.org
NOTRY=true

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=287633

[not at google - send to devlin, 2014-08-05 18:47:18]

does this look sensible?

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
File chrome/common/extensions/docs/server2/api_list_data_source.py (right):

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
chrome/common/extensions/docs/server2/api_list_data_source.py:62:
platform).IterModels():
IterModels() might be a problem, it's an implicit Get() operation, albeit an
optimised one which tries to run in parallel. Something to note.

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
File chrome/common/extensions/docs/server2/future_test.py (right):

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
chrome/common/extensions/docs/server2/future_test.py:164: self.assertEqual(val,
42)
assertEquals isn't a function, assertEqual is. how did this pass?

[ahernandez, 2014-08-05 18:58:23]

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
File chrome/common/extensions/docs/server2/future_test.py (right):

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
chrome/common/extensions/docs/server2/future_test.py:164: self.assertEqual(val,
42)
On 2014/08/05 18:47:18, kalman wrote:
> assertEquals isn't a function, assertEqual is. how did this pass?

assertEquals is deprecated according to
https://docs.python.org/2/library/unittest.html#deprecated-aliases. 

I'd be more than happy to submit a patch later converting all uses of
assertEquals to assertEqual in docserver tests, if that's desirable.

[not at google - send to devlin, 2014-08-05 20:01:26]

Btw this isn't related to this CL at all, but rather your other FutureLoop CL -
Yoyo brought up a good point, why the big speed regression? It should basically
have been a no-op right?

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
File chrome/common/extensions/docs/server2/future_test.py (right):

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
chrome/common/extensions/docs/server2/future_test.py:164: self.assertEqual(val,
42)
On 2014/08/05 18:58:23, ahernandez wrote:
> On 2014/08/05 18:47:18, kalman wrote:
> > assertEquals isn't a function, assertEqual is. how did this pass?
> 
> assertEquals is deprecated according to
> https://docs.python.org/2/library/unittest.html#deprecated-aliases. 
> 
> I'd be more than happy to submit a patch later converting all uses of
> assertEquals to assertEqual in docserver tests, if that's desirable.

not fussed. I'm constantly confused because gtest is ASSERT_EQ, python is
assertEqual, and the JS test framework we use is assertEquals.

[ahernandez, 2014-08-05 20:35:24]

On 2014/08/05 20:01:26, kalman wrote:
> Btw this isn't related to this CL at all, but rather your other FutureLoop CL
-
> Yoyo brought up a good point, why the big speed regression? It should
basically
> have been a no-op right?

I don't know why; the modifications I made to Get()/Raise() should not have
incurred much extra overhead. Then again, there are a lot of Futures created, so
even the slightest additional work is magnified? I once counted the number of
Futures created with callbacks during an integration test run, and IIRC it was
over 300,000.

[Yoyo Zhou, 2014-08-05 21:52:37]

LGTM, I think it's reasonable.

https://chromiumcodereview.appspot.com/438403002/diff/1/chrome/common/extensi...
File chrome/common/extensions/docs/server2/future.py (right):

https://chromiumcodereview.appspot.com/438403002/diff/1/chrome/common/extensi...
chrome/common/extensions/docs/server2/future.py:73: If |callback| returns a
Future then Then() will resolve to the same value
These 2 sentences are confusing next to each other.

[not at google - send to devlin, 2014-08-05 22:51:42]

On 2014/08/05 20:35:24, ahernandez wrote:
> On 2014/08/05 20:01:26, kalman wrote:
> > Btw this isn't related to this CL at all, but rather your other FutureLoop
CL
> -
> > Yoyo brought up a good point, why the big speed regression? It should
> basically
> > have been a no-op right?
> 
> I don't know why; the modifications I made to Get()/Raise() should not have
> incurred much extra overhead. Then again, there are a lot of Futures created,
so
> even the slightest additional work is magnified? I once counted the number of
> Futures created with callbacks during an integration test run, and IIRC it was
> over 300,000.

Hm, maybe. The pure operations required to construct 300,000 Future instances
should be fine, but maybe you were doing some kind of expensive threading
operations/checks as part of that?

[not at google - send to devlin, 2014-08-05 23:00:54]

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
File chrome/common/extensions/docs/server2/future.py (right):

https://codereview.chromium.org/438403002/diff/1/chrome/common/extensions/doc...
chrome/common/extensions/docs/server2/future.py:73: If |callback| returns a
Future then Then() will resolve to the same value
On 2014/08/05 21:52:36, Yoyo Zhou wrote:
> These 2 sentences are confusing next to each other.

This is tricky to explain. I had another go.

[not at google - send to devlin, 2014-08-05 23:01:08]

The CQ bit was checked by kalman@chromium.org

[commit-bot: I haz the power, 2014-08-05 23:03:53]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/kalman@chromium.org/438403002/20001

[commit-bot: I haz the power, 2014-08-05 23:06:20]

Change committed as 287633