Redirect pure isolate print to helper isolate.

Redirect pure isolate print to helper isolate.

R=podivilov@chromium.org,vsm@google.com
BUG=4686
Committed: https://code.google.com/p/dart/source/detail?r=20668

[Anton Muhin, 2013-03-29 09:47:24]

https://codereview.chromium.org/13185006/diff/3001/sdk/lib/html/dart2js/html_...
File sdk/lib/html/dart2js/html_dart2js.dart (left):

https://codereview.chromium.org/13185006/diff/3001/sdk/lib/html/dart2js/html_...
sdk/lib/html/dart2js/html_dart2js.dart:32784: -const _NEW_TIMER = 'NEW_TIMER';
that looks wrong, mind if I move the whole Timer into native_DOMImplementation?

[vsm, 2013-03-29 14:14:54]

lgtm

https://codereview.chromium.org/13185006/diff/3001/sdk/lib/html/dart2js/html_...
File sdk/lib/html/dart2js/html_dart2js.dart (left):

https://codereview.chromium.org/13185006/diff/3001/sdk/lib/html/dart2js/html_...
sdk/lib/html/dart2js/html_dart2js.dart:32784: -const _NEW_TIMER = 'NEW_TIMER';
How about an IsolateTimer (or similar) file?  I could imagine dart2js using this
code in the future.

On 2013/03/29 09:47:24, Anton Muhin wrote:
> that looks wrong, mind if I move the whole Timer into
native_DOMImplementation?

https://codereview.chromium.org/13185006/diff/3001/tools/dom/src/native_DOMIm...
File tools/dom/src/native_DOMImplementation.dart (right):

https://codereview.chromium.org/13185006/diff/3001/tools/dom/src/native_DOMIm...
tools/dom/src/native_DOMImplementation.dart:159: print('[From isolate]
$message');
We have the id we're using for interop (_Utils._getNewIsolateId) ... if that
works on background isolates.

Do you know if workers print any ID?  That is what dart2js isolates should be
mapping to.

https://codereview.chromium.org/13185006/diff/3001/tools/dom/src/native_DOMIm...
tools/dom/src/native_DOMImplementation.dart:164: final _printClosure =
window.console.log;
Is there a way at the Dart-level to programmatically tell whether I'm on a DOM
isolate (without a try-catch)?  Could use a pattern like:

final _printClosure = _hasDomAccess ? window.console.log :
_pureIsolatePrintClosure;

instead of wiring up in c++.

[Anton Muhin, 2013-03-29 16:14:32]

https://codereview.chromium.org/13185006/diff/3001/sdk/lib/html/dart2js/html_...
File sdk/lib/html/dart2js/html_dart2js.dart (left):

https://codereview.chromium.org/13185006/diff/3001/sdk/lib/html/dart2js/html_...
sdk/lib/html/dart2js/html_dart2js.dart:32784: -const _NEW_TIMER = 'NEW_TIMER';
I'd prefer to move it to Dartium specific code for now.

If later you'd like to share bits of code, let's recreate Timer.dart.

But I am not insisting on this particular approach.

On 2013/03/29 14:14:54, vsm wrote:
> How about an IsolateTimer (or similar) file?  I could imagine dart2js using
this
> code in the future.
> 
> On 2013/03/29 09:47:24, Anton Muhin wrote:
> > that looks wrong, mind if I move the whole Timer into
> native_DOMImplementation?
>

https://codereview.chromium.org/13185006/diff/3001/tools/dom/src/native_DOMIm...
File tools/dom/src/native_DOMImplementation.dart (right):

https://codereview.chromium.org/13185006/diff/3001/tools/dom/src/native_DOMIm...
tools/dom/src/native_DOMImplementation.dart:159: print('[From isolate]
$message');
Sorry, no, I do not immediately now.  Ideally we should probably print something
like root library url and some unique id.

On 2013/03/29 14:14:54, vsm wrote:
> We have the id we're using for interop (_Utils._getNewIsolateId) ... if that
> works on background isolates.
> 
> Do you know if workers print any ID?  That is what dart2js isolates should be
> mapping to.

https://codereview.chromium.org/13185006/diff/3001/tools/dom/src/native_DOMIm...
tools/dom/src/native_DOMImplementation.dart:164: final _printClosure =
window.console.log;
Not right now.  We can expose it via Utils, but I am not sure if it's worth it
or not.

On 2013/03/29 14:14:54, vsm wrote:
> Is there a way at the Dart-level to programmatically tell whether I'm on a DOM
> isolate (without a try-catch)?  Could use a pattern like:
> 
> final _printClosure = _hasDomAccess ? window.console.log :
> _pureIsolatePrintClosure;
> 
> instead of wiring up in c++.

[vsm, 2013-03-29 16:41:12]

still lgtm

[Anton Muhin, 2013-03-29 16:45:35]

Committed patchset #2 manually as r20668 (presubmit successful).