Remove promises from "read all bytes"

This makes it easier for other specifications that use this algorithm, such as Fetch, to use streams while interfacing less directly with the JavaScript engine.

Helps with whatwg/fetch#1568. Helps with #1178.

Author: noamr

index.bs

@@ -6809,36 +6809,33 @@ a chunk</dfn> from a {{ReadableStreamDefaultReader}} |reader|, given a [=read re
 
 <div algorithm="read all bytes">
  <p>To <dfn export for="ReadableStreamDefaultReader" lt="read all bytes|reading all bytes">read all
- bytes</dfn> from a {{ReadableStreamDefaultReader}} |reader|, perform the following steps. The
- result will be a {{Promise}} for a [=byte sequence=].
-
- 1. Let |promise| be [=a new promise=].
- 1. Let |bytes| be an empty [=byte sequence=].
- 1. [=Read-loop=] given |reader|, |bytes|, and |promise|.
- 1. Return |promise|.
+ bytes</dfn> from a {{ReadableStreamDefaultReader}} |reader|, given |successSteps|,
+ which is an algorithm accepting a [=byte sequence=], and |failureSteps|, which is an algorithm
+ accepting a JavaScript value: [=read-loop=] given |reader|, a new [=byte sequence=],
+ |successSteps|, and |failureSteps|.
 
  <div algorithm="read-loop">
-  For the purposes of the above algorithm, to <dfn>read-loop</dfn> given |reader|, |bytes|, and
-  |promise|:
+  For the purposes of the above algorithm, to <dfn>read-loop</dfn> given |reader|, |bytes|,
+  |successSteps|, and |failureSteps|:
 
   1. Let |readRequest| be a new [=read request=] with the following [=struct/items=]:
    : [=read request/chunk steps=], given |chunk|
    ::
-    1. If |chunk| is not a {{Uint8Array}} object, [=reject=] |promise| with a {{TypeError}} and
+    1. If |chunk| is not a {{Uint8Array}} object, call |failureSteps| with a {{TypeError}} and
        abort these steps.
     1. Append the bytes represented by |chunk| to |bytes|.
-    1. [=Read-loop=] given |reader|, |bytes|, and |promise|.
+    1. [=Read-loop=] given |reader|, |bytes|, |successSteps|, and |failureSteps|.
        <p class="note">This recursion could potentially cause a stack overflow if implemented
        directly. Implementations will need to mitigate this, e.g. by using a non-recursive variant
        of this algorithm, or [=queue a microtask|queuing a microtask=], or using a more direct
        method of byte-reading as noted below.
 
    : [=read request/close steps=]
    ::
-    1. [=Resolve=] |promise| with |bytes|.
+    1. Call |successSteps| with |bytes|.
    : [=read request/error steps=], given |e|
    ::
-    1. [=Reject=] |promise| with |e|.
+    1. Call |failureSteps| with |e|.
   1. Perform ! [$ReadableStreamDefaultReaderRead$](|reader|, |readRequest|).
  </div>