Intent to Deprecate and Remove: legacy caller for HTMLEmbedElement and HTMLObjectElement

91 visualizzazioni
Passa al primo messaggio da leggere

Philip Jägenstedt

da leggere,
15 nov 2016, 16:29:0515/11/16
a blink-dev

Summary

Drop [Custom=LegacyCallAsFunction] for HTMLEmbedElement and HTMLObjectElement, so that document.createElement("embed")() throws with "x is not a function".


Motivation

The base motivation is getting rid of more custom bindings and reducing unimportant differences between our IDL and the spec.


A legacy caller in WebIDL allows an instance to be invoked as a function. HTML used to have this for embed and object, but no more:

https://github.com/whatwg/html/pull/1979 https://github.com/w3c/web-platform-tests/pull/4084


Per the tests for embed and object, the legacy caller is supported by Chrome and Firefox, but not Edge or Safari. It is being removed from Gecko.


It's not entirely clear to me when this code did anything useful, probably it was for NPAPI. Testing with http://www.sjgames.com/ourgames/flash.html and http://wwwns.akamai.com/hdnetwork/demo/flash/default.html it looks like if a (Flash) plugin is actually initialized, calling as a function throws TypeError. Starting at V8HTMLPlugInElementCustom.cpp I haven't been able to figure out if the call is ever intercepted to do anything.

Compatibility Risk

Any content that invokes embeds or objects as functions will throw exceptions. There's not much hope of identifying this in httparchive, so use counters are all we have to estimate risk, and they are rounding to zero.


Alternative implementation suggestion for web developers

Just remove these calls. If there is in fact some plugin out there where the legacy caller does something useful, please comment on the issue.


Usage information from UseCounter

https://www.chromestatus.com/metrics/feature/timeline/popularity/152 https://www.chromestatus.com/metrics/feature/timeline/popularity/153


OWP launch tracking bug

https://crbug.com/663662


Entry on the feature dashboard


Requesting approval to remove too?

Yes, but I'll add a deprecation message for one release cycle before removing.


TAMURA, Kent

da leggere,
15 nov 2016, 18:02:1215/11/16
a Philip Jägenstedt, blink-dev
LGTM1

--
TAMURA Kent
Software Engineer, Google


Rick Byers

da leggere,
17 nov 2016, 10:59:2417/11/16
a TAMURA, Kent, Philip Jägenstedt, blink-dev
LGTM2

Chris Harrelson

da leggere,
17 nov 2016, 12:20:4917/11/16
a Rick Byers, TAMURA, Kent, Philip Jägenstedt, blink-dev
LGTM3

LGTM2
--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.

Rispondi a tutti
Rispondi all'autore
Inoltra
0 nuovi messaggi