Remove more multi-thread support

Remove more multi-thread support. Sorry about how hard it is to review this; diff doesn't handle this kind of change very well.

The assumption is that tasks cannot be interrupted in the single-threaded world of Dart so we no longer need to guard against the case where the state of analysis has changed while executing a task. Please think carefully about the assumption and also check carefully that I have not made any mistakes while making this change (the tests all run, but I'm not sure how well tested this portion of the code base is).

R=scheglov@google.com

Committed: https://code.google.com/p/dart/source/detail?r=41051

[Brian Wilkerson, 2014-10-10 18:06:02]

brianwilkerson@google.com changed reviewers:
+ scheglov@google.com

[scheglov, 2014-10-10 19:52:56]

LGTM

https://codereview.chromium.org/647033002/diff/1/pkg/analyzer/lib/src/generat...
File pkg/analyzer/lib/src/generated/engine.dart (left):

https://codereview.chromium.org/647033002/diff/1/pkg/analyzer/lib/src/generat...
pkg/analyzer/lib/src/generated/engine.dart:3983: int sourceTime =
getModificationStamp(source);
The source cannot be modified in _contentCache.
But it may be modified on disk.
So, I wonder if we still need to check it.

OTOH, the context will be notified about these file changes eventually, and
invalidate related information.
It seems to be more effective, so that we don't have to talk to the file system.

[Brian Wilkerson, 2014-10-10 20:24:56]

https://codereview.chromium.org/647033002/diff/1/pkg/analyzer/lib/src/generat...
File pkg/analyzer/lib/src/generated/engine.dart (left):

https://codereview.chromium.org/647033002/diff/1/pkg/analyzer/lib/src/generat...
pkg/analyzer/lib/src/generated/engine.dart:3983: int sourceTime =
getModificationStamp(source);
> The source cannot be modified in _contentCache.
> But it may be modified on disk.
> So, I wonder if we still need to check it.
> 
> OTOH, the context will be notified about these file changes eventually, and
> invalidate related information.
> It seems to be more effective, so that we don't have to talk to the file
system.

Yes, that was my conclusion as well.

The original code was guarding against in-memory updates on separate threads. We
don't need to guard against changes on disk because those will be dealt with by
other tasks.

[Brian Wilkerson, 2014-10-10 20:25:20]

Committed patchset #2 (id:50001) manually as 41051 (presubmit successful).