Remove support for the DeferredLibrary annotation and remove tests of it

Remove support for the DeferredLibrary annotation and remove tests of it

BUG=18562
R=floitsch@google.com, hausner@google.com

Committed: https://code.google.com/p/dart/source/detail?r=39797

[sigurdm, 2014-07-30 11:03:21]

https://codereview.chromium.org/429883002/diff/20001/tests/lib/async/deferred...
File tests/lib/async/deferred/deferred_fail_to_load_test.dart (left):

https://codereview.chromium.org/429883002/diff/20001/tests/lib/async/deferred...
tests/lib/async/deferred/deferred_fail_to_load_test.dart:10: const b = const
DeferredLibrary("NonExistingFile", uri: "wrong/wrong.js");
Unfortunately for this test the new syntax does not give us a way to specify the
uri of the part-file. So there is no way to test a bad uri.

[hausner, 2014-07-30 16:18:36]

https://codereview.chromium.org/429883002/diff/20001/tests/lib/async/deferred...
File tests/lib/async/deferred/deferred_fail_to_load_test.dart (left):

https://codereview.chromium.org/429883002/diff/20001/tests/lib/async/deferred...
tests/lib/async/deferred/deferred_fail_to_load_test.dart:10: const b = const
DeferredLibrary("NonExistingFile", uri: "wrong/wrong.js");
On 2014/07/30 11:03:21, sigurdm wrote:
> Unfortunately for this test the new syntax does not give us a way to specify
the
> uri of the part-file. So there is no way to test a bad uri.

I thought the new mechanism for catching load errors is that the future
completes with an error.

import "missing_file.dart" deferred as d;
...
d.loadLibrary().then((_) {
  ...
},
onError: (e) {
  print("load error: $e");
});

This is supposed to print the error when "missing_file.dart" can't be found.

I have a CL out that implements this behavior in the VM.

[sigurdm, 2014-07-31 11:35:22]

Yes, only this is hard to test (for the dart2js case) because dart2js needs the
deferred library to do the compilation.
It might make sense to have the test for the vm, and mark it as a
"CompileTimeError, Ok" for dart2js.

[floitsch, 2014-08-28 20:05:20]

Second patch set was diffing to a bad base.
I only reviewed patch set 1.

If possible I would remove the "DeferredLibrary" class directly. It doesn't work
anymore, and I expect users of the experimental approach to know how to update
to the new syntax.

LGTM.

https://chromiumcodereview.appspot.com/429883002/diff/20001/sdk/lib/_internal...
File sdk/lib/_internal/compiler/implementation/deferred_load.dart (right):

https://chromiumcodereview.appspot.com/429883002/diff/20001/sdk/lib/_internal...
sdk/lib/_internal/compiler/implementation/deferred_load.dart:701:
compiler.reportFatalError(import, MessageKind.DEFERRED_OLD_SYNTAX);
long line.

Maybe just a warning? Talk to Johnni/Peter.

https://chromiumcodereview.appspot.com/429883002/diff/20001/sdk/lib/async/def...
File sdk/lib/async/deferred_load.dart (right):

https://chromiumcodereview.appspot.com/429883002/diff/20001/sdk/lib/async/def...
sdk/lib/async/deferred_load.dart:14: class DeferredLibrary {
Any plan on when we want to remove the class?

[sigurdm, 2014-09-02 11:23:18]

https://codereview.chromium.org/429883002/diff/20001/sdk/lib/_internal/compil...
File sdk/lib/_internal/compiler/implementation/deferred_load.dart (right):

https://codereview.chromium.org/429883002/diff/20001/sdk/lib/_internal/compil...
sdk/lib/_internal/compiler/implementation/deferred_load.dart:701:
compiler.reportFatalError(import, MessageKind.DEFERRED_OLD_SYNTAX);
On 2014/08/28 20:05:20, floitsch wrote:
> long line.
> 
> Maybe just a warning? Talk to Johnni/Peter.

Giving a warning would entail keeping in all the functionality for handling the
old syntax.

https://codereview.chromium.org/429883002/diff/20001/sdk/lib/async/deferred_l...
File sdk/lib/async/deferred_load.dart (right):

https://codereview.chromium.org/429883002/diff/20001/sdk/lib/async/deferred_l...
sdk/lib/async/deferred_load.dart:14: class DeferredLibrary {
On 2014/08/28 20:05:20, floitsch wrote:
> Any plan on when we want to remove the class?

I think version 1.8 (after talking to seth)

[hausner, 2014-09-02 21:28:54]

tests/status lgtm. Can't speak to the compiler changes, though.

[sigurdm, 2014-09-03 08:34:55]

Committed patchset #3 (id:160001) manually as 39797 (presubmit successful).