Use a transformer to initialize custom elements and call initMethod. This is the

Use a transformer to initialize custom elements and call initMethod. This is the
first step to removes the dependence on mirrors in the deployable application.

R=jmesserly@google.com

Committed: https://code.google.com/p/dart/source/detail?r=32270

[Jennifer Messerly, 2014-02-03 19:09:32]

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
File pkg/polymer/lib/src/build/common.dart (right):

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/common.dart:127: AssetId resolve(AssetId source,
String url, TransformLogger logger, Span span,
instead of allowPackageColonUrls, is there a way we could build this as two
functions? e.g. resolveDartUri and resolveHtmlUri?

alternatively maybe just rename it "dartUris"

btw, great doc comment :)

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
File pkg/polymer/lib/src/build/script_compactor.dart (right):

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/script_compactor.dart:137: var url =
dartLibrary.path.startsWith('lib/')
refactor this into a function?

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/script_compactor.dart:153: // the show/hide
combinators.
hmm. This seems to be reimplementing top level name resolution?

Not saying this is wrong, but would be good to reuse an impl from dart2js or
dartanalyzer ... touching the show/hide stuff makes me a little nervous because
it seems like the kind of thing that could change and might have resolution
subtleties.

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/script_compactor.dart:155: var targetId =
resolve(dartLibrary, directive.uri.stringValue,
idea: pull this out into a separate method? e.g. _processExportDirective.
similar below for ClassDecl and FunctionDecl

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/script_compactor.dart:172: for (var node in
unit.declarations) {
is it possible for the library to conflict with an export or part? I think not
because would be an error in Dart too, so should be okay. Just checking.

[Siggi Cherem (dart-lang), 2014-02-04 00:20:34]

Thanks.

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
File pkg/polymer/lib/src/build/common.dart (right):

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/common.dart:127: AssetId resolve(AssetId source,
String url, TransformLogger logger, Span span,
On 2014/02/03 19:09:32, John Messerly wrote:
> instead of allowPackageColonUrls, is there a way we could build this as two
> functions? e.g. resolveDartUri and resolveHtmlUri?
> 
> alternatively maybe just rename it "dartUris"
> 
> btw, great doc comment :)

We could. This also gave me another idea. I can check whether
source.path.endsWith('.dart') then allow 'package:' automatically in that case.
Thoughts?

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
File pkg/polymer/lib/src/build/script_compactor.dart (right):

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/script_compactor.dart:137: var url =
dartLibrary.path.startsWith('lib/')
On 2014/02/03 19:09:32, John Messerly wrote:
> refactor this into a function?

Done.

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/script_compactor.dart:153: // the show/hide
combinators.
On 2014/02/03 19:09:32, John Messerly wrote:
> hmm. This seems to be reimplementing top level name resolution?
> 
> Not saying this is wrong, but would be good to reuse an impl from dart2js or
> dartanalyzer ... touching the show/hide stuff makes me a little nervous
because
> it seems like the kind of thing that could change and might have resolution
> subtleties. 

I hear you - I've added a TODO. I'm afraid of using that part of the analyzer
because it requires bringing in the full resolution.

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/script_compactor.dart:155: var targetId =
resolve(dartLibrary, directive.uri.stringValue,
On 2014/02/03 19:09:32, John Messerly wrote:
> idea: pull this out into a separate method? e.g. _processExportDirective.
> similar below for ClassDecl and FunctionDecl

Funny, I thought of it too, but wasn't sure it was worth it. It definitely makes
the code cleaner. Done.

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/script_compactor.dart:172: for (var node in
unit.declarations) {
On 2014/02/03 19:09:32, John Messerly wrote:
> is it possible for the library to conflict with an export or part? I think not
> because would be an error in Dart too, so should be okay. Just checking.

yeah, I think that would be caught as a Dart error too.

[Jennifer Messerly, 2014-02-04 00:37:40]

lgtm

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
File pkg/polymer/lib/src/build/common.dart (right):

https://codereview.chromium.org/151893003/diff/30001/pkg/polymer/lib/src/buil...
pkg/polymer/lib/src/build/common.dart:127: AssetId resolve(AssetId source,
String url, TransformLogger logger, Span span,
On 2014/02/04 00:20:35, Siggi Cherem (dart-lang) wrote:
> On 2014/02/03 19:09:32, John Messerly wrote:
> > instead of allowPackageColonUrls, is there a way we could build this as two
> > functions? e.g. resolveDartUri and resolveHtmlUri?
> > 
> > alternatively maybe just rename it "dartUris"
> > 
> > btw, great doc comment :)
> 
> We could. This also gave me another idea. I can check whether
> source.path.endsWith('.dart') then allow 'package:' automatically in that
case.
> Thoughts?

Love it!

[Siggi Cherem (dart-lang), 2014-02-04 01:08:57]

Committed patchset #2 manually as r32270 (presubmit successful).