dart2js: Add a header to the output of the startup emitter.

dart2js: Add a header to the output of the startup emitter.

R=sigmund@google.com

Committed: https://github.com/dart-lang/sdk/commit/7e9fcbf5d6942ef14a287984ee72207469f34943

[floitsch, 2015-08-04 15:06:37]

floitsch@google.com changed reviewers:
+ sigmund@google.com

[floitsch, 2015-08-04 15:06:38]

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
File pkg/compiler/lib/src/js_emitter/full_emitter/declarations.dart (left):

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
pkg/compiler/lib/src/js_emitter/full_emitter/declarations.dart:16: const String
GENERATED_BY = """
Moved.

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
File pkg/compiler/lib/src/js_emitter/startup_emitter/fragment_emitter.dart
(right):

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
pkg/compiler/lib/src/js_emitter/startup_emitter/fragment_emitter.dart:524: new
js.VariableDeclaration(global, allowRename: false));
It worked before, but since this variable is shared between two different
fragments, I prefer to make it explicit that it must not be renamed.

[Siggi Cherem (dart-lang), 2015-08-04 16:42:48]

lgtm

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
File pkg/compiler/lib/src/js_emitter/startup_emitter/model_emitter.dart (right):

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
pkg/compiler/lib/src/js_emitter/startup_emitter/model_emitter.dart:223: return
new jsAst.Comment(generatedBy(compiler));
should we add an argument to generatedBy so we can indicate which emitter we
used? (we used to have the "(fast startup)" as part of this header before).
Another idea is to print out all the flags/options in the header.

[floitsch, 2015-08-05 11:46:50]

I'm committing already (because of the merge-to-dev that is happening soon),
but, as usual, I can do a follow-up CL.

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
File pkg/compiler/lib/src/js_emitter/startup_emitter/model_emitter.dart (right):

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
pkg/compiler/lib/src/js_emitter/startup_emitter/model_emitter.dart:223: return
new jsAst.Comment(generatedBy(compiler));
On 2015/08/04 16:42:48, Siggi Cherem (dart-lang) wrote:
> should we add an argument to generatedBy so we can indicate which emitter we
> used? (we used to have the "(fast startup)" as part of this header before).
> Another idea is to print out all the flags/options in the header.

I added a 'flavor' argument to 'generatedBy'. That should bring us back to what
we had before. (Even more, since I take mirrors and CSP into account).
I think we don't want to print *all* options, but it is an option, and I
wouldn't be entirely against it. Up to you.

[floitsch, 2015-08-05 11:47:43]

Committed patchset #2 (id:20001) manually as
7e9fcbf5d6942ef14a287984ee72207469f34943 (presubmit successful).

[Siggi Cherem (dart-lang), 2015-08-05 15:51:02]

lgtm

[Siggi Cherem (dart-lang), 2015-08-05 15:53:36]

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
File pkg/compiler/lib/src/js_emitter/startup_emitter/model_emitter.dart (right):

https://codereview.chromium.org/1271953003/diff/1/pkg/compiler/lib/src/js_emi...
pkg/compiler/lib/src/js_emitter/startup_emitter/model_emitter.dart:223: return
new jsAst.Comment(generatedBy(compiler));
On 2015/08/05 11:46:50, floitsch wrote:
> On 2015/08/04 16:42:48, Siggi Cherem (dart-lang) wrote:
> > should we add an argument to generatedBy so we can indicate which emitter we
> > used? (we used to have the "(fast startup)" as part of this header before).
> > Another idea is to print out all the flags/options in the header.
> 
> I added a 'flavor' argument to 'generatedBy'. That should bring us back to
what
> we had before. (Even more, since I take mirrors and CSP into account).
> I think we don't want to print *all* options, but it is an option, and I
> wouldn't be entirely against it. Up to you.

I have a feeling that all flags might be too noisy too, but there might be some
others for which it may be useful (e.g. trust-type-annotations). We don't need
to add them now.