Avoid the _onSubscriptionStateChange being called twice in some cases.

Avoid the _onSubscriptionStateChange being called twice in some cases.

R=floitsch@google.com

Committed: https://code.google.com/p/dart/source/detail?r=17048

[floitsch, 2013-01-14 16:31:24]

LGTM.

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dart
File sdk/lib/async/stream_impl.dart (right):

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dar...
sdk/lib/async/stream_impl.dart:487: int timesPaused =
subscriber._setUnsubscribed();
Add comment: Unsubscribing a paused subscriber cancels the pauses. (Or similar).

s/timesPaused/subscriptionPauseCount

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dar...
sdk/lib/async/stream_impl.dart:642: if (!identical(wasPaused, _isPaused))
_onPauseStateChange();
why?

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dar...
sdk/lib/async/stream_impl.dart:670: // In the case where there only pending
events are generated ones.
Don't understand comment.

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_pipe.dart
File sdk/lib/async/stream_pipe.dart (right):

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_pipe.dar...
sdk/lib/async/stream_pipe.dart:43: * By default [:this:] subscribes to [_source]
and forwards all events to its own
80chars.

If 'this' is not available yet, we should probably use something else.
By default the stream subs...

[Lasse Reichstein Nielsen, 2013-01-15 08:52:54]

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dart
File sdk/lib/async/stream_impl.dart (right):

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dar...
sdk/lib/async/stream_impl.dart:487: int timesPaused =
subscriber._setUnsubscribed();
On 2013/01/14 16:31:24, floitsch wrote:
> Add comment: Unsubscribing a paused subscriber cancels the pauses. (Or
similar).
> 
> s/timesPaused/subscriptionPauseCount

Done.

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dar...
sdk/lib/async/stream_impl.dart:642: if (!identical(wasPaused, _isPaused))
_onPauseStateChange();
Probably overoptimizing :)
Reverted.

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dar...
sdk/lib/async/stream_impl.dart:670: // In the case where there only pending
events are generated ones.
It's commenting that in this particular class, it's possible to have pending
events even if there is no _pendingEvents object, which is why we can't use the
branch before.
Comment updated.

I think I'll change how generated events work, and instead generate them inside
the pending events object, so the rest of the object doesn't need changing.

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_pipe.dart
File sdk/lib/async/stream_pipe.dart (right):

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_pipe.dar...
sdk/lib/async/stream_pipe.dart:43: * By default [:this:] subscribes to [_source]
and forwards all events to its own
It's not that "this" is not available *yet*, it is that it's not an identifier
at all. You can't write generic expressions inside [...], only identifiers (and
qualified identifiers should also be allowed eventually, but it's not in the
spec yet).

[floitsch, 2013-01-15 14:40:19]

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dart
File sdk/lib/async/stream_impl.dart (right):

https://codereview.chromium.org/11880019/diff/1/sdk/lib/async/stream_impl.dar...
sdk/lib/async/stream_impl.dart:670: // In the case where there only pending
events are generated ones.
On 2013/01/15 08:52:54, Lasse Reichstein Nielsen wrote:
> It's commenting that in this particular class, it's possible to have pending
> events even if there is no _pendingEvents object, which is why we can't use
the
> branch before.
> Comment updated.
> 
> I think I'll change how generated events work, and instead generate them
inside
> the pending events object, so the rest of the object doesn't need changing.

Did you commit? The comment still looks the same.