FlightReplyServer are for client->server and ReactFlightClient is for
server->client. They're not 100% symmetrical.
We did a number of refactors to ReactFlightClient in PRs like #29823 and
#33664 to change the structure of the resolution. This PR brings those
changes to synchronize the two approaches. Which addresses deep
resolution of cycles and deferred error handling.
This also fixes a critical security vulnerability.
For Edge Flight servers, that use Web Streams, we're defining the
`debugChannel` option as:
```
debugChannel?: {readable?: ReadableStream, writable?: WritableStream, ...}
```
Whereas for Node.js Flight servers, that use Node.js Streams, we're
defining it as:
```
debugChannel?: Readable | Writable | Duplex | WebSocket
```
For the Edge Flight clients, there is currently only one direction of
the debug channel supported, so we define the option as:
```
debugChannel?: {readable?: ReadableStream, ...}
```
Consequently, for the Node.js Flight clients, we define the option as:
```
debugChannel?: Readable
```
The presence of a readable debug channel is passed to the Flight client
internally via the `hasReadable` flag on the internal `debugChannel`
option. For the Node.js clients, that flag was accidentally derived from
the public option `debugChannel.readable`, which is conceptually
incorrect, because `debugChannel` is a `Readable` stream, not an options
object with a `readable` property. However, a `Readable` also has a
`readable` property, which is a boolean that indicates whether the
stream is in a readable state. This meant that the `hasReadable` flag
was incidentally still set correctly. Regardless, this was confusing and
unintentional, so we're now fixing it to always set `hasReadable` to
`true` when a `debugChannel` is provided to the Node.js clients. We'll
revisit this in case we ever add support for writable debug channels in
Node.js (and Edge) clients.
When you use the `createFromFetch` API we assume that the start time of
the request is the same time as when you call `createFromFetch` but in
principle you could use it with a Promise that starts earlier and just
happens to resolve to a `Response`.
When you use `createFromReadableStream` that is almost definitely the
case. E.g. you might have started it way earlier and you don't call
`createFromReadableStream` until you get the headers back (the fetch
promise resolves).
This adds an option to pass in the start time for debug purposes if you
started the request before starting to parse it.
One thing that can suspend is the downloading of the RSC stream itself.
This tracks an I/O entry for each Promise (`SomeChunk<T>`) that
represents the request to the RSC stream. As the value we use the
`Response` for `createFromFetch` (or the `ReadableStream` for
`createFromReadableStream`). The start time is when you called those.
Since we're not awaiting the whole stream, each I/O entry represents the
part of the stream up until it got unblocked. However, in a production
environment with TLS packets and buffering in practice the chunks
received by the client isn't exactly at the boundary of each row. It's a
bit longer into larger chunks. From testing, it seems like multiples of
16kb or 64kb uncompressed are common. To simulate a production
environment we group into roughly 64kb chunks if they happen in rapid
sequence. Note that this might be too small to give a good idea because
of the throttle many boundaries might be skipped anyway so this might
show too many.
The React DevTools will see each I/O entry as separate but dedupe if an
outer boundary already depends on the same chunk. This deduping makes it
so that small boundaries that are blocked on the same chunk, don't get
treated as having unique suspenders. If you have a boundary with large
content, then that content will likely be in a separate chunk which is
not in the parent and then it gets marked as.
This is all just an approximation. The goal of this is just to highlight
that very large boundaries will very likely suspend even if they don't
suspend on any I/O on the server. In practice, these boundaries can
float around a lot and it's really any Suspense boundary that might
suspend but some are more likely than others which this is meant to
highlight.
It also just lets you inspect how many bytes needs to be transferred
before you can show a particular part of the content, to give you an
idea that it's not just I/O on the server that might suspend.
If you don't use the debug channel it can be misleading since the data
in development mode stream will have a lot more data in it which leads
to more chunking.
Similarly to "client references" these I/O infos don't have an "env"
since it's the client that has the I/O and so those are excluded from
flushing in the Server performance tracks.
Note that currently the same Response can appear many times in the same
Instance of SuspenseNode in DevTools when there are multiple chunks. In
a follow up I'll show only the last one per Response at any given level.
Note that when a separate debugChannel is used it has its own I/O entry
that's on the `_debugInfo` for the debug chunks in that channel.
However, if everything works correctly these should never leak into the
DevTools UI since they should never be propagated from a debug chunk to
the values waited by the runtime. This is easy to break though.
When the Flight Client is waiting for pending debug chunks, it drops the
debug info if there is no writable side of the debug channel defined.
However, it should instead check if there's no readable side defined.
Fixing this is not only important for browser clients that don't want or
need a return channel, but it's also crucial for server-side rendering,
because the Node and Edge clients only accept a readable side of the
debug channel. So they can't even define a noop writable side as a
workaround.
When a debug channel is defined, we must ensure that we don't close the
Flight Client's response when the debug channel's readable is done, but
the RSC stream is still flowing. Now, we wait for both streams to end
before closing the response.
When a debug channel is used between the Flight server and a browser
Flight client, we want to allow the same RSC stream to be used for
server-side rendering. To support this, the Edge and Node Flight clients
also need to accept a `debugChannel` option. Without it, debug
information would be missing (e.g. for SSR error stacks), and in some
cases this could result in `Connection closed` errors.
This PR adds support for the `debugChannel` option in the Edge and Node
clients for ESM, Parcel, Turbopack, and Webpack. Unlike the browser
clients, these clients only support a one-way channel, since the Flight
server’s return protocol is not designed for multiple clients.
The implementation follows the approach used in the browser clients, but
excludes the writable parts.
This adds a "suspended by" row for each chunk that is referenced from a
client reference. So when you select a client component, you can see
what bundles will block that client component when loading on the
client.
This is only done in the browser build since if we added it on the
server, it would show up as a blocking resource and while it's possible
we expect that a typical server request won't block on loading JS.
<img width="664" height="486" alt="Screenshot 2025-08-17 at 3 45 14 PM"
src="https://github.com/user-attachments/assets/b1f83445-2a4e-4470-9a20-7cd215ab0482"
/>
<img width="745" height="678" alt="Screenshot 2025-08-17 at 3 46 58 PM"
src="https://github.com/user-attachments/assets/3558eae1-cf34-4e11-9d0e-02ec076356a4"
/>
Currently this is only included if it ends up wrapped in a lazy like in
the typical type position of a Client Component, but there's a general
issue that maybe hard references need to transfer their debug info to
the parent which can transfer it to the Fiber.
This lets us pass a writable on the server side and readable on the
client side to send debug info through a separate channel so that it
doesn't interfere with the main payload as much. The main payload refers
to chunks defined in the debug info which means it's still blocked on it
though. This ensures that the debug data has loaded by the time the
value is rendered so that the next step can forward the data.
This will be a bit fragile to race conditions until #33665 lands.
Another follow up needed is the ability to skip the debug channel on the
receiving side. Right now it'll block forever if you don't provide one
since we're blocking on the debug data.
This adds plumbing for opening a stream from the Flight Client to the
Flight Server so it can ask for more data on-demand. In this mode, the
Flight Server keeps the connection open as long as the client is still
alive and there's more objects to load. It retains any depth limited
objects so that they can be asked for later. In this first PR it just
releases the object when it's discovered on the server and doesn't
actually lazy load it yet. That's coming in a follow up.
This strategy is built on the model that each request has its own
channel for this. Instead of some global registry. That ensures that
referential identity is preserved within a Request and the Request can
refer to previously written objects by reference.
The fixture implements a WebSocket per request but it doesn't have to be
done that way. It can be multiplexed through an existing WebSocket for
example. The current protocol is just a Readable(Stream) on the server
and WritableStream on the client. It could even be sent through a HTTP
request body if browsers implemented full duplex (which they don't).
This PR only implements the direction of messages from Client to Server.
However, I also plan on adding Debug Channel in the other direction to
allow debug info (optionally) be sent from Server to Client through this
channel instead of through the main RSC request. So the `debugChannel`
option will be able to take writable or readable or both.
---------
Co-authored-by: Hendrik Liebau <mail@hendrik-liebau.de>
This effectively lets us consume Web Streams in a Node build. In fact
the Node entry point is now just adding Node stream APIs.
For the client, this is simple because the configs are not actually
stream type specific. The server is a little trickier.
Reverts #33457, #33456 and #33442.
There are too many issues with wrappers, lazy init, stateful modules,
duplicate instantiation of async_hooks and duplication of code.
Instead, we'll just do a wrapper polyfill that uses Node Streams
internally.
I kept the client indirection files that I added for consistency with
the server though.
This is used to register Server References that exist in the current
environment but also exists in the server it might call into. Such as a
remote server.
If the value comes from the remote server in the first place then this
is called automatically to ensure that you can pass a reference back to
where it came from - even if the `serverModuleMap` option is used. This
was already the case when `serverModuleMap` wasn't passed. This is how
you can pass server references back to the server. However, when we
added `serverModuleMap` that pass was skipped because we were getting
real functions instead of proxies.
For functions that wasn't yet passed from the remote server to the
current server, we can register them eagerly just like we do for
`import('/server').registerServerReference()`. You can now also do this
with `import('/client').registerServerReference()`. We could make them
shared so you only have to do this once but it might not be possible to
pass to the remote server and the remote server might not even be the
same RSC renderer. Therefore I split them. It's up to the compiler
whether it should do that or not. It has to know that any function you
might call might be able to receive it. This is currently global to a
specific RSC renderer.
This implements `findSourceMapURL` in react-server-dom-parcel, enabling
source maps for replayed server errors on the client. It utilizes a new
endpoint in the Parcel dev server that returns the source map for a
given bundle/file. The error overlay UI has also been updated to handle
these stacks. See https://github.com/parcel-bundler/parcel/pull/10082
Also updated the fixture to the latest Parcel canary. A few APIs have
changed. We do have a higher level library wrapper now (`@parcel/rsc`
added in https://github.com/parcel-bundler/parcel/pull/10074) but I left
the fixture using the lower level APIs directly here since it is easier
to see how react-server-dom-parcel is used.
Corresponding Parcel PR:
https://github.com/parcel-bundler/parcel/pull/10073
Parcel avoids [cascading cache
invalidation](https://philipwalton.com/articles/cascading-cache-invalidation/)
by injecting a bundle manifest containing a mapping of stable bundle ids
to hashed URLs. When using an HTML entry point, this is done (as of the
above PR) via a native import map. This means that if a bundle's hash
changes, only that bundle will be invalidated (plus the HTML itself
which typically has a short caching policy), not any other bundles that
reference it.
For RSCs, we cannot currently use native import maps because of client
side navigations, where a new HTML file is not requested. Eventually,
multiple `<script type="importmap">` elements will be supported
(https://github.com/whatwg/html/pull/10528) ([coming Chrome
133](https://chromestatus.com/feature/5121916248260608)), at which point
React could potentially inject them. In the meantime, I've added some
APIs to Parcel to polyfill this. With this change, an import map can be
sent along with a client reference, containing a mapping for any dynamic
imports and URL dependencies (e.g. images) that are referenced by the JS
bundles. On the client, the import map is extended with these new
mappings prior to executing the referenced bundles. This preserves the
caching advantages described above while supporting client navigations.
Followup to #31725
This implements `prepareDestinationForModule` in the Parcel Flight
client. On the Parcel side, the `<Resources>` component now only inserts
`<link>` elements for stylesheets (along with a bootstrap script when
needed), and React is responsible for inserting scripts. This ensures
that components that are conditionally dynamic imported during render
are also preloaded.
CSS must be added to the RSC tree using `<Resources>` to avoid FOUC.
This must be manually rendered in both the top-level page, and in any
component that is dynamic imported. It would be nice if there was a way
for React to automatically insert CSS as well, but unfortunately
`prepareDestinationForModule` only knows about client components and not
CSS for server components. Perhaps there could be a way we could
annotate components at code splitting boundaries with the resources they
need? More thoughts in this thread:
https://github.com/facebook/react/pull/31725#discussion_r1884867607
We support streaming `multipart/form-data` in Node.js using Busboy since
that's kind of the idiomatic ecosystem way for handling these stream
there. There's not really anything idiomatic like that for Edge that's
universal yet.
This adds a version that's basically just
`AsyncIterable.from(formData)`. It could also be a `ReadableStream` of
those entries since those are also `AsyncIterable`.
I imagine that in the future we might add one from a binary
`ReadableStream` that does the parsing built-in.
Follow up to #31725.
I diffed against the Turbopack one to find any unexpected discrepancies.
Some parts are forked enough that it's hard to diff but I think I got
most of it.
This adds a new `react-server-dom-parcel-package`, which is an RSC
integration for the Parcel bundler. It is mostly copied from the
existing webpack/turbopack integrations, with some changes to utilize
Parcel runtime APIs for loading and executing bundles/modules.
See https://github.com/parcel-bundler/parcel/pull/10043 for the Parcel
side of this, which includes the plugin needed to generate client and
server references. https://github.com/parcel-bundler/rsc-examples also
includes examples of various ways to use RSCs with Parcel.
Differences from other integrations:
* Client and server modules are all part of the same graph, and we use
Parcel's
[environments](https://parceljs.org/plugin-system/transformer/#the-environment)
to distinguish them. The server is the Parcel build entry point, and it
imports and renders server components in route handlers. When a `"use
client"` directive is seen, the environment changes and Parcel creates a
new client bundle for the page, combining all client modules together.
CSS from both client and server components are also combined
automatically.
* There is no separate manifest file that needs to be passed around by
the user. A [Runtime](https://parceljs.org/plugin-system/runtime/)
plugin injects client and server references as needed into the relevant
bundles, and registers server action ids using `react-server-dom-parcel`
automatically.
* A special `<Resources>` component is also generated by Parcel to
render the `<script>` and `<link rel="stylesheet">` elements needed for
a page, using the relevant info from the bundle graph.
Note: I've already published a 0.0.x version of this package to npm for
testing purposes but happy to add whoever needs access to it as well.
### Questions
* How to test this in the React repo. I'll have integration tests in
Parcel, but setting up all the different mocks and environments to
simulate that here seems challenging. I could try to copy how
Webpack/Turbopack do it but it's a bit different.
* Where to put TypeScript types. Right now I have some ambient types in
my [example
repo](https://github.com/parcel-bundler/rsc-examples/blob/main/types.d.ts)
but it would be nice for users not to copy and paste these. Can I
include them in the package or do they need to maintained separately in
definitelytyped? I would really prefer not to have to maintain code in
three different repos ideally.
---------
Co-authored-by: Sebastian Markbage <sebastian@calyptus.eu>
