mozilla
/
gecko-dev
зеркало из https://github.com/mozilla/gecko-dev.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность
gecko-dev/js/public/Stream.h

543 строки
21 KiB
C++
Исходник Ответственный История

/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*-
* vim: set ts=8 sts=2 et sw=2 tw=80:
* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
/*
* JSAPI functions and callbacks related to WHATWG Stream objects.
*
* Much of the API here mirrors the standard algorithms and standard JS methods
* of the objects defined in the Streams standard. One difference is that the
* functionality of the JS controller object is exposed to C++ as functions
* taking ReadableStream instances instead, for convenience.
*/
#ifndef js_Stream_h
#define js_Stream_h
#include <stddef.h>
#include "jstypes.h"
#include "js/RootingAPI.h"
#include "js/TypeDecls.h"
struct JSClass;
namespace JS {
/**
* Abstract base class for external underlying sources.
*
* The term "underlying source" is defined in the Streams spec:
* https://streams.spec.whatwg.org/#underlying-source
*
* A `ReadableStreamUnderlyingSource` is an underlying source that is
* implemented in C++ rather than JS. It can be passed to
* `JS::NewReadableExternalSourceStreamObject` to create a custom,
* embedding-defined ReadableStream.
*
* There are several API difference between this class and the standard API for
* underlying sources implemented in JS:
*
* - JS underlying sources can be either byte sources or non-byte sources.
* External underlying source are always byte sources.
*
* - The C++ API does not bother with controller objects. Instead of using
* controller methods, the underlying source directly calls API functions
* like JS::ReadableStream{UpdateDataAvailableFromSource,Close,Error}.
*
* - External readable streams are optimized to allow the embedding to
* interact with them with a minimum of overhead: chunks aren't enqueued as
* individual typed arrays; instead, the embedding only updates the amount
* of data available using
* JS::ReadableStreamUpdateDataAvailableFromSource. When JS requests data
* from a reader, writeIntoReadRequestBuffer is invoked, asking the
* embedding to write data directly into the buffer we're about to hand to
* JS.
*
* - The C++ API provides extra callbacks onClosed() and onErrored().
*
* - This class has a `finalize()` method, because C++ cares about lifetimes.
*
* Additionally, ReadableStreamGetExternalUnderlyingSource can be used to get
* the pointer to the underlying source. This locks the stream until it is
* released again using JS::ReadableStreamReleaseExternalUnderlyingSource.
*
* Embeddings can use this to optimize away the JS `ReadableStream` overhead
* when an embedding-defined C++ stream is passed to an embedding-defined C++
* consumer. For example, consider a ServiceWorker piping a `fetch` Response
* body to a TextDecoder. Instead of copying chunks of data into JS typed array
* buffers and creating a Promise per chunk, only to immediately resolve the
* Promises and read the data out again, the embedding can directly feed the
* incoming data to the TextDecoder.
*
* Compartment safety: All methods (except `finalize`) receive `cx` and
* `stream` arguments. SpiderMonkey enters the realm of the stream object
* before invoking these methods, so `stream` is never a wrapper. Other
* arguments may be wrappers.
*/
class JS_PUBLIC_API ReadableStreamUnderlyingSource {
public:
virtual ~ReadableStreamUnderlyingSource() = default;
/**
* Invoked whenever a reader desires more data from this source.
*
* The given `desiredSize` is the absolute size, not a delta from the
* previous desired size.
*/
virtual void requestData(JSContext* cx, HandleObject stream,
size_t desiredSize) = 0;
/**
* Invoked to cause the embedding to fill the given `buffer` with data from
* this underlying source.
*
* This is called only after the embedding has updated the amount of data
* available using JS::ReadableStreamUpdateDataAvailableFromSource. If at
* least one read request is pending when
* JS::ReadableStreamUpdateDataAvailableFromSource is called, this method
* is invoked immediately from under the call to
* JS::ReadableStreamUpdateDataAvailableFromSource. If not, it is invoked
* if and when a new read request is made.
*
* Note: This method *must not cause GC*, because that could potentially
* invalidate the `buffer` pointer.
*/
virtual void writeIntoReadRequestBuffer(JSContext* cx, HandleObject stream,
void* buffer, size_t length,
size_t* bytesWritten) = 0;
/**
* Invoked in reaction to the ReadableStream being canceled. This is
* equivalent to the `cancel` method on non-external underlying sources
* provided to the ReadableStream constructor in JavaScript.
*
* The underlying source may free up some resources in this method, but
* `*this` must not be destroyed until `finalize()` is called.
*
* The given `reason` is the JS::Value that was passed as an argument to
* ReadableStream#cancel().
*
* The returned JS::Value will be used to resolve the Promise returned by
* ReadableStream#cancel().
*/
virtual Value cancel(JSContext* cx, HandleObject stream,
HandleValue reason) = 0;
/**
* Invoked when the associated ReadableStream becomes closed.
*
* The underlying source may free up some resources in this method, but
* `*this` must not be destroyed until `finalize()` is called.
*/
virtual void onClosed(JSContext* cx, HandleObject stream) = 0;
/**
* Invoked when the associated ReadableStream becomes errored.
*
* The underlying source may free up some resources in this method, but
* `*this` must not be destroyed until `finalize()` is called.
*/
virtual void onErrored(JSContext* cx, HandleObject stream,
HandleValue reason) = 0;
/**
* Invoked when the associated ReadableStream object is finalized. The
* stream object is not passed as an argument, as it might not be in a
* valid state anymore.
*
* Note: Finalization can happen on a background thread, so the embedding
* must be prepared for `finalize()` to be invoked from any thread.
*/
virtual void finalize() = 0;
};
/**
* Returns a new instance of the ReadableStream builtin class in the current
* compartment, configured as a default stream.
* If a |proto| is passed, that gets set as the instance's [[Prototype]]
* instead of the original value of |ReadableStream.prototype|.
*/
extern JS_PUBLIC_API JSObject* NewReadableDefaultStreamObject(
JSContext* cx, HandleObject underlyingSource = nullptr,
HandleFunction size = nullptr, double highWaterMark = 1,
HandleObject proto = nullptr);
/**
* Returns a new instance of the ReadableStream builtin class in the current
* compartment.
*
* The instance is a byte stream backed by an embedding-provided underlying
* source, using the virtual methods of `underlyingSource` as callbacks. The
* embedding must ensure that `*underlyingSource` lives as long as the new
* stream object. The JS engine will call the finalize() method when the stream
* object is destroyed.
*
* `nsISupportsObject_alreadyAddreffed` is an optional pointer that can be used
* to make the new stream participate in Gecko's cycle collection. Here are the
* rules for using this parameter properly:
*
* - `*underlyingSource` must not be a cycle-collected object. (It would lead
* to memory leaks as the cycle collector would not be able to collect
* cycles containing that object.)
*
* - `*underlyingSource` must not contain nsCOMPtrs that point to cycle-
* collected objects. (Same reason.)
*
* - `*underlyingSource` may contain a pointer to a single cycle-collected
* object.
*
* - The pointer may be stored in `*underlyingSource` as a raw pointer.
*
* - The pointer to the nsISupports interface of the same object must be
* passed as the `nsISupportsObject_alreadyAddreffed` parameter to this
* function. (This is how the cycle collector knows about it, so omitting
* this would again cause leaks.)
*
* If `proto` is non-null, it is used as the instance's [[Prototype]] instead
* of the original value of `ReadableStream.prototype`.
*/
extern JS_PUBLIC_API JSObject* NewReadableExternalSourceStreamObject(
JSContext* cx, ReadableStreamUnderlyingSource* underlyingSource,
void* nsISupportsObject_alreadyAddreffed = nullptr,
HandleObject proto = nullptr);
/**
* Returns the embedding-provided underlying source of the given |stream|.
*
* Can be used to optimize operations if both the underlying source and the
* intended sink are embedding-provided. In that case it might be
* preferrable to pipe data directly from source to sink without interacting
* with the stream at all.
*
* Locks the stream until ReadableStreamReleaseExternalUnderlyingSource is
* called.
*
* Throws an exception if the stream is locked, i.e. if a reader has been
* acquired for the stream, or if ReadableStreamGetExternalUnderlyingSource
* has been used previously without releasing the external source again.
*
* Throws an exception if the stream isn't readable, i.e if it is errored or
* closed. This is different from ReadableStreamGetReader because we don't
* have a Promise to resolve/reject, which a reader provides.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*
* Asserts that the stream has an embedding-provided underlying source.
*/
extern JS_PUBLIC_API bool ReadableStreamGetExternalUnderlyingSource(
JSContext* cx, HandleObject stream,
ReadableStreamUnderlyingSource** source);
/**
* Releases the embedding-provided underlying source of the given |stream|,
* returning the stream into an unlocked state.
*
* Asserts that the stream was locked through
* ReadableStreamGetExternalUnderlyingSource.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*
* Asserts that the stream has an embedding-provided underlying source.
*/
extern JS_PUBLIC_API bool ReadableStreamReleaseExternalUnderlyingSource(
JSContext* cx, HandleObject stream);
/**
* Update the amount of data available at the underlying source of the given
* |stream|.
*
* Can only be used for streams with an embedding-provided underlying source.
* The JS engine will use the given value to satisfy read requests for the
* stream by invoking the writeIntoReadRequestBuffer method.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamUpdateDataAvailableFromSource(
JSContext* cx, HandleObject stream, uint32_t availableData);
/**
* Break the cycle between this object and the
* nsISupportsObject_alreadyAddreffed passed in
* NewReadableExternalSourceStreamObject().
*/
extern JS_PUBLIC_API void ReadableStreamReleaseCCObject(JSObject* stream);
/**
* Returns true if the given object is a ReadableStream object or an
* unwrappable wrapper for one, false otherwise.
*/
extern JS_PUBLIC_API bool IsReadableStream(JSObject* obj);
/**
* Returns true if the given object is a ReadableStreamDefaultReader or
* ReadableStreamBYOBReader object or an unwrappable wrapper for one, false
* otherwise.
*/
extern JS_PUBLIC_API bool IsReadableStreamReader(JSObject* obj);
/**
* Returns true if the given object is a ReadableStreamDefaultReader object
* or an unwrappable wrapper for one, false otherwise.
*/
extern JS_PUBLIC_API bool IsReadableStreamDefaultReader(JSObject* obj);
enum class ReadableStreamMode { Default, Byte, ExternalSource };
/**
* Returns the stream's ReadableStreamMode. If the mode is |Byte| or
* |ExternalSource|, it's possible to acquire a BYOB reader for more optimized
* operations.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamGetMode(JSContext* cx,
HandleObject stream,
ReadableStreamMode* mode);
enum class ReadableStreamReaderMode { Default };
/**
* Returns true if the given ReadableStream is readable, false if not.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamIsReadable(JSContext* cx,
HandleObject stream,
bool* result);
/**
* Returns true if the given ReadableStream is locked, false if not.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamIsLocked(JSContext* cx,
HandleObject stream,
bool* result);
/**
* Returns true if the given ReadableStream is disturbed, false if not.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamIsDisturbed(JSContext* cx,
HandleObject stream,
bool* result);
/**
* Cancels the given ReadableStream with the given reason and returns a
* Promise resolved according to the result.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API JSObject* ReadableStreamCancel(JSContext* cx,
HandleObject stream,
HandleValue reason);
/**
* Creates a reader of the type specified by the mode option and locks the
* stream to the new reader.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one. The returned object will always be created in the
* current cx compartment.
*/
extern JS_PUBLIC_API JSObject* ReadableStreamGetReader(
JSContext* cx, HandleObject stream, ReadableStreamReaderMode mode);
/**
* Tees the given ReadableStream and stores the two resulting streams in
* outparams. Returns false if the operation fails, e.g. because the stream is
* locked.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamTee(JSContext* cx, HandleObject stream,
MutableHandleObject branch1Stream,
MutableHandleObject branch2Stream);
/**
* Retrieves the desired combined size of additional chunks to fill the given
* ReadableStream's queue. Stores the result in |value| and sets |hasValue| to
* true on success, returns false on failure.
*
* If the stream is errored, the call will succeed but no value will be stored
* in |value| and |hasValue| will be set to false.
*
* Note: This is semantically equivalent to the |desiredSize| getter on
* the stream controller's prototype in JS. We expose it with the stream
* itself as a target for simplicity.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamGetDesiredSize(JSContext* cx,
JSObject* stream,
bool* hasValue,
double* value);
/**
* Close the given ReadableStream. This is equivalent to `controller.close()`
* in JS.
*
* This can fail with or without an exception pending under a variety of
* circumstances. On failure, the stream may or may not be closed, and
* downstream consumers may or may not have been notified.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamClose(JSContext* cx,
HandleObject stream);
/**
* Returns true if the given ReadableStream reader is locked, false otherwise.
*
* Asserts that |reader| is a ReadableStreamDefaultReader or
* ReadableStreamBYOBReader object or an unwrappable wrapper for one.
*/
extern JS_PUBLIC_API bool ReadableStreamReaderIsClosed(JSContext* cx,
HandleObject reader,
bool* result);
/**
* Enqueues the given chunk in the given ReadableStream.
*
* Throws a TypeError and returns false if the enqueing operation fails.
*
* Note: This is semantically equivalent to the |enqueue| method on
* the stream controller's prototype in JS. We expose it with the stream
* itself as a target for simplicity.
*
* If the ReadableStream has an underlying byte source, the given chunk must
* be a typed array or a DataView. Consider using
* ReadableByteStreamEnqueueBuffer.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamEnqueue(JSContext* cx,
HandleObject stream,
HandleValue chunk);
/**
* Errors the given ReadableStream, causing all future interactions to fail
* with the given error value.
*
* Throws a TypeError and returns false if the erroring operation fails.
*
* Note: This is semantically equivalent to the |error| method on
* the stream controller's prototype in JS. We expose it with the stream
* itself as a target for simplicity.
*
* Asserts that |stream| is a ReadableStream object or an unwrappable wrapper
* for one.
*/
extern JS_PUBLIC_API bool ReadableStreamError(JSContext* cx,
HandleObject stream,
HandleValue error);
/**
* C++ equivalent of `reader.cancel(reason)`
* (both <https://streams.spec.whatwg.org/#default-reader-cancel> and
* <https://streams.spec.whatwg.org/#byob-reader-cancel>).
*
* `reader` must be a stream reader created using `JS::ReadableStreamGetReader`
* or an unwrappable wrapper for one. (This function is meant to support using
* C++ to read from streams. It's not meant to allow C++ code to operate on
* readers created by scripts.)
*/
extern JS_PUBLIC_API bool ReadableStreamReaderCancel(JSContext* cx,
HandleObject reader,
HandleValue reason);
/**
* C++ equivalent of `reader.releaseLock()`
* (both <https://streams.spec.whatwg.org/#default-reader-release-lock> and
* <https://streams.spec.whatwg.org/#byob-reader-release-lock>).
*
* `reader` must be a stream reader created using `JS::ReadableStreamGetReader`
* or an unwrappable wrapper for one.
*/
extern JS_PUBLIC_API bool ReadableStreamReaderReleaseLock(JSContext* cx,
HandleObject reader);
/**
* C++ equivalent of the `reader.read()` method on default readers
* (<https://streams.spec.whatwg.org/#default-reader-read>).
*
* The result is a new Promise object, or null on OOM.
*
* `reader` must be the result of calling `JS::ReadableStreamGetReader` with
* `ReadableStreamReaderMode::Default` mode, or an unwrappable wrapper for such
* a reader.
*/
extern JS_PUBLIC_API JSObject* ReadableStreamDefaultReaderRead(
JSContext* cx, HandleObject reader);
class JS_PUBLIC_API WritableStreamUnderlyingSink {
public:
virtual ~WritableStreamUnderlyingSink() = default;
/**
* Invoked when the associated WritableStream object is finalized. The
* stream object is not passed as an argument, as it might not be in a
* valid state anymore.
*
* Note: Finalization can happen on a background thread, so the embedding
* must be prepared for `finalize()` to be invoked from any thread.
*/
virtual void finalize() = 0;
};
// ReadableStream.prototype.pipeTo SUPPORT
/**
* The signature of a function that, when passed an |AbortSignal| instance, will
* return the value of its "aborted" flag.
*
* This function will be called while |signal|'s realm has been entered.
*/
using AbortSignalIsAborted = bool (*)(JSObject* signal);
/**
* Dictate embedder-specific details necessary to implement certain aspects of
* the |ReadableStream.prototype.pipeTo| function. This should be performed
* exactly once, for a single context associated with a |JSRuntime|.
*
* The |ReadableStream.prototype.pipeTo| function accepts a |signal| argument
* that may be used to abort the piping operation. This argument must be either
* |undefined| (in other words, the piping operation can't be aborted) or an
* |AbortSignal| instance (that may be aborted using the signal's associated
* |AbortController|). |AbortSignal| is defined by WebIDL and the DOM in the
* web embedding. Therefore, embedders must use this function to specify how
* such objects can be recognized and how to perform various essential actions
* upon them.
*
* The provided |isAborted| function will be called with an unwrapped
* |AbortSignal| instance, while that instance's realm has been entered.
*
* If this function isn't called, and a situation arises where an "is this an
* |AbortSignal|?" question must be asked, that question will simply be answered
* "no".
*/
extern JS_PUBLIC_API void InitPipeToHandling(const JSClass* abortSignalClass,
AbortSignalIsAborted isAborted,
JSContext* cx);
} // namespace JS
#endif // js_Stream_h
Ссылка в новой задаче Показать git blame Копировать постоянную ссылку