Re: [whatwg/streams] Add support for ReadableStream "owning" type (PR #1271)

@saschanaz commented on this pull request.

BTW, should we care about [`Symbol.dispose` ECMA262 Stage 3 proposal](https://github.com/tc39/proposal-explicit-resource-management)? It looks reasonable for this API to support that, and if so it should be better sooner than later to prevent webcompat issue.

> @@ -651,8 +651,15 @@ enum ReadableStreamType { "bytes" };
   <p>For an example of how to set up a readable byte stream, including using the different
   controller interface, see [[#example-rbs-push]].
 
-  <p>Setting any value other than "{{ReadableStreamType/bytes}}" or undefined will cause the
-  {{ReadableStream()}} constructor to throw an exception.
+  <p>Can be set to "<dfn enum-value for="ReadableStreamType">owning</dfn>" to signal that the

```suggestion
  <p>If set to "<dfn enum-value for="ReadableStreamType">owning</dfn>", it signals that the
```

Can this and the above be "If set to"? "Can be set to" made sense when there was only one possibility, but repeating that makes less sense to me.

> @@ -651,8 +651,15 @@ enum ReadableStreamType { "bytes" };
   <p>For an example of how to set up a readable byte stream, including using the different
   controller interface, see [[#example-rbs-push]].
 
-  <p>Setting any value other than "{{ReadableStreamType/bytes}}" or undefined will cause the
-  {{ReadableStream()}} constructor to throw an exception.
+  <p>Can be set to "<dfn enum-value for="ReadableStreamType">owning</dfn>" to signal that the
+  constructed {{ReadableStream}} will own chunks (via transfer or serialization) before enqueuing them.
+  This ensures that enqueued chunks are not mutable by the source.

```suggestion
  This ensures that enqueued chunks are not mutated by the source.
```

"mutable by" sounds a bit unnatural.

> @@ -5188,7 +5217,7 @@ The following abstract operations support the implementation of the
  id="writable-stream-default-controller-close">WritableStreamDefaultControllerClose(|controller|)</dfn>
  performs the following steps:
 
- 1. Perform ! [$EnqueueValueWithSize$](|controller|, [=close sentinel=], 0).
+ 1. Perform ! [$EnqueueValueWithSize$](|controller|, [=close sentinel=], 0, undefined).

Should this `undefined` be `« »` for consistency, or vise versa?

> +<div algorithm>
+ <dfn abstract-op lt="StructuredTransferOrClone">StructuredTransferOrClone(|value|, |transferList|)</dfn>
+    performs the following steps:
+ 1. Let |serialized| be ? [$StructuredSerializeWithTransfer$](|value|, |transferList|).
+ 1. Let |deserialized| be ? [$StructuredDeserializeWithTransfer$](|serialized|, [=the current Realm=]).
+ 1. Return |deserialized|.\[[Deserialized]].
+</div>
+

(Please file an issue to not keep the duplication too long, or let me know so that I can file one)

> @@ -2950,13 +2969,21 @@ The following abstract operations support the implementation of the
 <div algorithm>
  <dfn abstract-op lt="ReadableStreamDefaultControllerEnqueue"
  id="readable-stream-default-controller-enqueue">ReadableStreamDefaultControllerEnqueue(|controller|,
- |chunk|)</dfn> performs the following steps:
+ |chunk|, |transferList|)</dfn> performs the following steps:

Not sure I understand your comment, I mean Streams have no existing example that receives empty list, while it has several optional parameters in section 9. But either way is fine I guess.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/streams/pull/1271#pullrequestreview-1468516963
You are receiving this because you are subscribed to this thread.

Message ID: <whatwg/streams/pull/1271/review/1468516963@github.com>