Define "usual constructor steps" #797
Conversation
This is for use in [HTMLConstructor] in whatwg/html#4915.
|
I'm not a big fan of this approach, personally; I think it'd be better to have something like
Any reason you're referring to the interface object, in particular? |
|
Sounds like my second option, right? Can do.
Nope, interface makes sense too. |
|
@Ms2ger I pushed a new approach, but I feel like there should be some warning text talking about how these steps are not general-purpose and are only used in special scenarios. Would you be able to help draft such text? I'm having some trouble doing so myself. |
| An interface may have <dfn export>overridden constructor steps</dfn>, which can | ||
| change the behavior of the [=interface object=] when called or constructed. By | ||
| default interfaces do not have such steps. | ||
|
|
There was a problem hiding this comment.
Maybe add a warning like we have a few already for legacy features, saying something like
In general, constructors are described by defining a [=constructor operation=] and its behavior.
The [=overridden constructor steps=] are used only for more complicated situations.
Editors who wish to use this feature are strongly advised to discuss this by filing an issue before proceeding.
| <div algorithm> | ||
|
|
||
| The [=interface object=] for a given [=interface=] |I| | ||
| with [=identifier=] |id| and in [=Realm=] |realm| | ||
| is <dfn lt="create an interface object">created</dfn> as follows: | ||
|
|
||
| 1. Let |steps| be the following steps: | ||
| 1. Let |steps| be |I|'s [=overriden constructor steps=] if they exist, or | ||
| the following steps otherwise: | ||
| 1. If |I| was not declared with a [=constructor operation=], | ||
| then [=ECMAScript/throw=] a {{ECMAScript/TypeError}}. | ||
| 1. If {{NewTarget}} is <emu-val>undefined</emu-val>, then |
There was a problem hiding this comment.
Should we keep the first two steps, even when overriding the rest of the algorithm?
Maybe also the last four steps?
There was a problem hiding this comment.
I guess that would work for [HTMLConstructor], although it feels cleaner to just have a general primitive that replaces the whole thing.
There was a problem hiding this comment.
I'll keep it like this for now, but we can refactor later if you think it's much better.
This is for use in [HTMLConstructor] in whatwg/html#4915.
There are alternative approaches and wordings that might be better, e.g.
But I thought I'd start here to get the ball rolling, as it's the first thing that occurred to me. Happy to tweak as desired.
Preview | Diff