Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47968
Updates the Inspector Proxy to report + log when a profiling build target (experimental) is registered. This notifies the developer that debugging is available for these app(s), which will not otherwise fetch development bundles from Metro.
Changelog: [Internal]
Reviewed By: rubennorte
Differential Revision: D66501771
fbshipit-source-id: e06dee279158094ad5c70bf8e6a90e7c983de48a
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47880
The previous diffs in this stack have aimed to make URL rewriting by inspector-proxy robust to any configuration of device->server, debugger->server and server->server connections.
Though rewriting was originally introduced to support Android emulator networking, we can now expand it to cover other use cases, like the device reaching the server over an internet address not reachable from the dev machine, or the debugger routing to the server through a tunnel on a different port, without needing CORS workarounds.
Changelog
[General][Fixed] dev-middleware: Rewrite URLs in the inspector proxy to cover all configurations, not just Android emulators.
Reviewed By: huntie
Differential Revision: D66247355
fbshipit-source-id: e9201ebc1f7f5fe2119c71cd4d7b4ca895645404
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47876
Currently, if a device is connected to the bundler via, say, `http://10.0.2.2:8081` (Android default), and a debugger is opened on `http://localhost:8081`, we rewrite hostnames `10.0.2.2.`<->`localhost` so that URLs are correct relative to each.
However, if the debugger is on a different port or protocol, this breaks down - because we only rewrite hostnames.
This fixes that by using the debugger's connecting `Host` header and `encrypted` state to derive the base URL of the server relative to the frontend. We then update the rewriting logic to use this actual full origin (protocol + host) in place of the device-relative origin.
Changelog:
[General][Fixed] dev-middleware: Fix URL rewriting where device and debugger reach the server on different ports/protocols.
Reviewed By: huntie
Differential Revision: D66077627
fbshipit-source-id: 01f6565149caa34b1e9e50dd58deb0122485657c
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47872
Largely a refactoring to the way we currently rewrite `url`/`urlRegex` in `Debugger.setBreakpointByURL` CDP requests (debugger->target).
Rewriting regexes is fragile, it only really works if we can assume `'localhost'` appears literally and that ports and protocols don't need changing. The intention here is to freeze the current behaviour so as not to break anyone relying on it (if anyone is), and decouple it from more robust rewriting we want to generalise.
Also adds simple regex escaping to host names (always IPv4 addresses) we inject into regex patterns, since the previous approach could've led to false matches in unlikely edge cases.
Changelog:
[General][Fixed] dev-middleware: Regex-escape IP addresses in urlRegex replacements
Reviewed By: huntie
Differential Revision: D66238782
fbshipit-source-id: 4cd0029081d68c193e36d3713057ffdc7ef0656f
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47685
Currently, we assume any URL with a hostname of `10.0.2.2` or `10.0.3.2` (device-relative) is eligible for rewriting to `localhost` (frontend-relative), because we assume the device is an Android emulator. We rewrite these URLs between device and dev machine so that the rewritten URLs are reachable from the dev machine.
This diff narrows this logic so that we'll only rewrite URLs where the hostname matches the pre-existing list *and* this matches the host the device is actually connected on, according to its headers from the original connection.
The main motivation for this change is to unblock removing assumptions about device-reachable vs server-reachable hosts. Later in the stack we'll drop the hardcoded listing of `10.0.2.2` etc in favour of identifying URLs that target the dev server, from whatever network.
There's also an edge case fix here that `10.0.2.2` etc might actually refer to a remote LAN server, and not be an Android emulator's alias for for an emulator host.
Changelog:
[General][Fixed] RN DevTools: Don't assume 10.0.2.2 is an alias for localhost unless it's used to establish a connection to the server
Reviewed By: huntie
Differential Revision: D66058704
fbshipit-source-id: bad28717b0c9b1ca43e2ea3391cef13f87892e6c
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47653
## Context
Currently, when `nativeSourceCodeFetching == false`, `inspector-proxy` attempts to pre-fetch source maps, given the URL from a `Debugger.scriptParsed` event, and embeds them into `Debugger.scriptParsed`'s `sourceMapURL` using a data URI.
This was originally to support frontends that did not perform HTTP requests or were blocked (eg by CORS), but we're retaining it for the moment because it's more performant than lazy loading the source map.
Similarly, we perform middleware->server fetches to respond to `Debugger.getScriptSource` events.
To make these fetches for URLs that target `10.0.2.2` (ie, addressable from within an Android emulator) (etc), we rewrite `10.0.2.2`->`localhost` and perform a `fetch` from the Node process running dev-middleware.
## The problem
Consider a setup where:
- Metro is running on a remote server, listening on `8081`.
- Dev machine tunnels `localhost:8082` -> remote `8081`.
- An app is running on an Android emulator on the dev machine, with bundle URL configured to `10.0.2.2:8082`.
In this case, we'll rewrite `10.0.2.2:8082` to `localhost:8082`, which *is* reachable and correct from the dev machine, but *not* from the machine where Metro is running, so the `fetch` of a source map from the inspector proxy will fail.
## Motivation
This might seem like a niche case, but it's part of fixing a series of unsafe assumptions that currently prevent us from running DevTools on an arbitrary port.
## This fix
Preserve the current behaviour (simple `10.0.2.2`<=>`localhost`) for URLs sent to the frontend, but construct a separate, server-relative URL, using the configured `serverBaseUrl`, for `fetch` calls within dev-middleware.
Changelog:
[General][Fixed] RN DevTools: Fix fetching sources and source maps when the dev-server is remote and not tunnelled via the same port+protocol.
Reviewed By: huntie
Differential Revision: D65993910
fbshipit-source-id: a0cdcf1644e97a2af3d8583f2da2aaa51276f68c
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47628
`serverBaseUrl` is currently documented as:
> The base URL to the dev server, as addressible from the local developer machine
This is problematic in general because `dev-middleware` on a server doesn't necessarily know about where clients might be reaching it from, how tunnels or port-forwards are set up, etc., and this can change over the lifetime of the server and vary between clients.
Indeed, our own use of `serverBaseUrl` from both `community-cli-plugin` and internally simply sets it to the host and port the dev server is listening on - ie it's the address of the dev server accessible *from the server*.
This PR changes the docs, redefining `serverBaseUrl`, to match the way we currently specify it.
One usage where we *do* want the previously documented behaviour is in responses to `/json/list` (`getPageDescriptions`) where the URLs in the response should be reachable by a browser requesting `/json/list`.
Here, we use the request (host header, etc.) to attempt to get working base URL.
History:
It should be mentioned that this is the latest in a series of changes like this:
- https://github.com/facebook/react-native/pull/39394
- https://github.com/facebook/react-native/pull/39456
Learning from those:
- This change does *not* break Android emulators, which routes `10.0.2.2` to localhost, or other routed devices, because `/open-debugger` still uses server-relative URLs, and now formally delegates to `BrowserLauncher` to decide what to do with those URLs (internally, VSCode / `xdg-open` handles port forwarding)
- Middleware configuration is no longer required to specify how it is reachable from clients.
This sets up some subsequent changes for more robust handling of tunnelled connections.
Changelog:
[General][Breaking] dev-middleware: Frameworks should specify `serverBaseUrl` relative to the middleware host.
Reviewed By: huntie
Differential Revision: D65974487
fbshipit-source-id: 1face8fc7715df387f75b329e80932d8543ee419
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47652
## Background
When the `nativeSourceCodeFetching` capability is disabled, `inspector-proxy` rewrites URLs exchanged over CDP between device and frontend so that URLs are addressable from CDT - in particular, when using an Android emulator `10.0.2.2` (host's address from within the emulator) is rewritten to and from `localhost` (the equivalent address reachable from the host).
Previously - before we implemented `Network.loadNetworkResource`, or on old frontends that don't attempt to use that method - this worked reasonably well. A `fetch` from CDT to Metro would succeed on the rewritten URL.
## Problem
Since we implemented `Network.loadNetworkResource`, but disabled the `nativeSourceCodeFetching` capability, source fetching is broken under Android emulators. We're rewriting URLs to be frontend-relative, but then attempting to fetch them through the device, because as far as CDT is aware, `Network.loadNetworkResource` should still be tried first.
When `Network.loadNetworkResource` responds with a CDP *error*, CDT falls back to a local fetch (which would work), but when it responds with a CDP *result* of `success: false`, there is no fallback.
## Fix
This diff adds an interception guarded behind `nativeSourceCodeFetching == false`, which rejects any calls to `Network.loadNetworkResource` with a CDP error. This restores the previous behaviour from before `Network.loadNetworkResource` was implemented at all.
NOTE: An alternative approach would be to rewrite URLs back to device-relative for `Network.loadNetworkResource`, but IMO it's more correct for the frontend to respect that the device is asserting that it doesn't have that capability, and not to try to use it.
Changelog:
[Android][Fixed] RN DevTools: Fix source loading when using an Android emulator connecting to a dev server on the host.
Reviewed By: huntie
Differential Revision: D66074731
fbshipit-source-id: f2050c014cd5cfa546bff5e9d0412413a5daff35
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47120
Fixes no-op behaviour of the "Open DevTools" Dev Menu item (bug on `main` introduced with D63329456).
This was caused by a change to the `description` field contents in our CDP `/json/list` response, when under Fusebox. In the `/open-debugger` call from the Dev Menu, we were still using the older `appId` param.
This did not affect `j` to debug, which uses the `target` param.
{F1937186832}
Changes:
In short: Matching against the `description` string is now fully eliminated for modern debugger targets.
- Update native Dev Menu implementation to omit `appId` parameter (`device` param alone is sufficient and fully precise on these platforms).
- Update `/open-debugger` implementation to ignore the `appId` parameter for modern targets, and document this in the `dev-middleware` README.
Changelog: [Internal]
Reviewed By: robhogan
Differential Revision: D64597581
fbshipit-source-id: 46f536e7d0a4ececab0d52f4c0704e8698466cd0
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47098
Changelog: [Internal]
D63329456 updated the `description` field to be human-readable.
Unfortunately, InspectorProxy relies on this field to compare the incoming `/open-debugger` calls.
hoxyq discovered the symptom of Fusebox failing to launch with `No compatible apps connected. React Native DevTools can only be used with the Hermes engine.` in Metro
Reviewed By: hoxyq
Differential Revision: D64547367
fbshipit-source-id: deed6851f3ede2c74be2b492def1eba6e58c43e6
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/46780
This is primarily a debugger server change to better-align the `title` and `description` fields (visible in via the CDP `/json/list` endpoint), reporting them directly from the device.
For React Native users, the net effect of this change is to improve/align the display name for each debugger target in the CLI multi-select menu (`j` to debug). This change may also be useful for discovery in non-Fusebox frontends such as VS Code (`vscode-expo` extension).
Changes:
- Rename `title` prop on `InspectorPageDescription`/`IInspector::addPage` to `description` (no call site changes).
- Add `deviceName` param to `InspectorPackagerConnection`.
- Move the page `description` to the `description` JSON field.
- Update `InspectorPackagerConnection::Impl::pages()` to return new `title` and `description` fields.
- Align `OpenDebuggerKeyboardHandler` to display `title` field.
- Deprecate the nonstandard `deviceName` field.
**Before**
```
[
{
"id": "3c9f24bedab0e73fca6a1b295030e7af9346a8c0-1",
"title": "React Native Bridgeless [C++ connection]",
"description": "com.facebook.RNTester",
...
}
```
The `description` field was previously the closest thing we had to a target identifier. Today, this is not needed, since we have the stable `reactNative.logicalDeviceId` field.
**After**
```
[
{
"id": "3c9f24bedab0e73fca6a1b295030e7af9346a8c0-1",
"title": "com.facebook.RNTester (iPhone 16 Pro)",
"description": "React Native Bridgeless [C++ connection]",
...
}
```
The `title` field is now more human readable and aligned with what we render in the window title of React Native DevTools. The `description` field describes the type of debugger target (the specific React Native implementation).
Changelog: [Internal]
Reviewed By: vzaidman
Differential Revision: D63329456
fbshipit-source-id: cfe98f77e31c729431005925cfc66e2780ef8c72
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/45069
Currently, `/json/list` returns pages within each device in the iteration order of a C++ `unordered_map`, which doesn't tell us anything useful. Page IDs happen to be sequential, but only as an implementation detail.
Change this contract so that we guarantee ordering reflects addition order, allowing clients to consistently select e.g. most recently added page for a given device.
The implementation of this is as simple as switching from an `unordered_map` to a key-ordered`map`, because we already assign keys (page IDs) with an incrementing integer. Within the inspector proxy, devices already use an insertion (connection)-ordered JS `Map`, so we just document this guarantee.
Changelog:
[General][Changed] Debugger: Make `/json/list` return connection-addition-ordered targets.
Reviewed By: huntie
Differential Revision: D58735947
fbshipit-source-id: 7a132cc5e750475792a2b845afc9a42424690bf1
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/45035
Changelog: [General][Fixed] Avoid a zombie state when opening a second debugger frontend concurrently.
The problem here was that we were sending proxy-protocol messages to the device in the wrong order (`disconnect` *after* `connect`):
{F1701266597}
The root cause was that we were depending on the outgoing debugger socket's async `close` event to trigger sending the `disconnect` message to the device. This would happen after we'd already (synchronously) sent the `connect` message.
With this diff, we send the `disconnect` message synchronously with calling `close()` on the debugger socket, which fixes the ordering problem at the source. To avoid sending duplicate `disconnect` messages (e.g. one before calling `close()` and one from the `close` event handler), we store some extra state on `Device` (`#connectedPageIds`).
Reviewed By: robhogan, huntie
Differential Revision: D58730634
fbshipit-source-id: 0f54af2e4f8071a8f6d97cc9e3d8a4ea89a46f43
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/45027
Changelog: [Internal]
Changes the device ID collision handling logic to reuse `Device` instances instead of creating new ones. This enables further refactoring of `Device` to improve session state isolation.
Reviewed By: hoxyq
Differential Revision: D58724884
fbshipit-source-id: bc11ce45ce8c80c58c32dcd1b07b28f1d1753a62
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/45010
D46482492 added logic for handing off state across "device" connections that have the same ID. This logic currently has no test coverage. It also contains a bug whereby the new device's pages are removed from the target listing endpoint (`/json`) when the *old* device's socket is closed.
This diff adds tests and fixes the bug.
Changelog: [General][Fixed] inspector-proxy no longer accidentally detaches connected devices.
## Next steps
It seems that the device ID handoff logic exists to paper over a deeper problem with the inspector proxy protocol (or its implementation in React Native): The React Native runtime should not routinely be creating new "device" connections without tearing down previous ones.
In followup diffs, I'll explore changing this behaviour for Fusebox, based on the new test coverage.
Reviewed By: robhogan
Differential Revision: D51013056
fbshipit-source-id: e0c17678cc747366a3b75cef18ca2a722fc93acd
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/44811
Changelog:
[General][Fixed] - Debugger frontend socket-termination countdown now begins after the ping message is actually sent
The debugger is currently disconnected if a ping-pong message is missed.
This causes the debugger to be unusable if it happens to be lagging, e.g. when the initialisation is competing with the flood of log spam T191394188
There are a few ways to fix this as discused with motiz88 and robhogan:
1. Ensure the websocket has a chance to respond, e.g. in via web worker
1. Lengthen the time allowed for the pong resopnse
I've done some digging to find the root cause of the UI being blocked in CDT, However, profiling shows that most of the work is not simple to break up, i.e. the number of expensive re-layout calls. Diving into that rabbit hole could mean accidentally writing React.
Because we ping every 10 seconds, we could get un/lucky where CDT happens to be busy _at that exact moment_, making this a flaky symptom to fix, even if we lengthen the allowed time-to-respond.
# V2+
So upon further investigation, CDT websocket is actually responding to the pings in due time:
{F1679132204}
(CDT doesn't show the ping/pong API as frames, so a custom tick/tock message was used to visualise the timing)
Over here in dev-middleware, we currently start a timeout to terminate the socket after sending the ping:
https://www.internalfb.com/code/fbsource/[813870db697a8701f2512d25a7fed730f0ec6ed9]/xplat/js/react-native-github/packages/dev-middleware/src/inspector-proxy/InspectorProxy.js?lines=306-307
If CDT doesn't respond in time, websocket would be terminated.
But we saw CDT respond immediately above, even during the log spam, so the delay must be coming from somewhere else.
The intuition is that during the log-spam, the middleware takes a perf hit too when it's processing the spam from the device and forwarding it to the CDT websocket.
We can confirm this by passing a "sent" callback via `socket.ping(cb)`:
https://github.com/websockets/ws/blob/9bdb58070d64c33a9beeac7c732aac0f4e7e18b7/lib/websocket.js#L246-L254
This gives us the timing between calling `socket.ping()` and when the ping is actually sent.
Regular, stress-free operation without log-spam shows most pings are sent within the same millisecond:
{F1679223326}
With the pong response grace period at 5 seconds, there's plenty of time for CDT to `pong` back. That's why it has been working in most cases.
However, during the log-spam, we easily see this send-sent delay over 5 seconds. In extreme cases, almost 30 seconds would have passed before middleware sent a message to CDT, which then responded under 2 seconds:
{F1679163335}
This means while CDT is getting flooded and has observable lag in the UI, the smoking gun is actually the middleware.
Digging a little deeper, we know that incoming messages from the target goes into a Promise queue, including the console logs:
https://www.internalfb.com/code/fbsource/[d5d312082e9c]/xplat/js/react-native-github/packages/dev-middleware/src/inspector-proxy/Device.js?lines=155-157
This means during the flood of logs from the target, the Promise queue keeps getting chained rapidly for each message.
Meanhile, the `ws` lib uses the underlying NodeJS `Socket.write` method for `ping(…)` and `send(…)`:
https://github.com/websockets/ws/blob/9bdb58070d64c33a9beeac7c732aac0f4e7e18b7/lib/sender.js#L349
…which is guaranteed to fire the callback asynchronously:
https://github.com/nodejs/help/issues/1504#issuecomment-422879594
Promise queue is in the macro task queue, which gets priority before the micro task queue. So if the Promise queue is not cleared yet, the websocket queue will have a hard time getting executed in time – explaining the extreme send-sent durations during a log spam.
The fix is simple:
1. Start the terminate-socket-timer until the `ping` is actually sent
1. Treat any incoming message (along with `pong`s) as a terminate-socket-timer reset
1. This also applies if `pong` comes in between `send` and `sent`, which can happen sometimes due to the async nature of the callback:
{F1679288626}
# V1
~~In this diff, a more forgiving mechanism is introduced, i.e. CDT is allowed to miss a ping-pong roundtrip 3 times before the websocket connection is terminated.~~
~~This allows a bit more breathing room for CDT's initialisation during log spam while maintaining the same ping-pong interval for VS Code to keep the auto SSH tunnel alive.~~
Reviewed By: huntie
Differential Revision: D58220230
fbshipit-source-id: 7111c9878492d8755a6110a5cdf4ef622265001d
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/44835
As titled. The `vm` field is not part of the CDP spec and will not be used by the modern debugger frontend or proxy.
This change affects modern CDP targets only (using `InspectorPackagerConnection`). We aim to enable sharing of more detailed metadata over 1/ a new, dedicated CDP domain, and 2/ namespaced under the existing `reactNative` field (for the latter, strictly limited to metadata necessary for dev server functionality).
Changelog: [Internal]
(Note: `/json` endpoint behaviour is unchanged for legacy CDP targets)
Reviewed By: robhogan
Differential Revision: D58285587
fbshipit-source-id: dfef3a56b20486ba11891df9940f6c7bef59528e
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/44499
Changelog: [internal]
As [discussed](https://fb.workplace.com/groups/react.devx.team/permalink/930483712103526/), we'll begin segmenting Telemetry signals by Fusebox/non-Fusebox.
* Add new flag in event reporter for `debugger_command` (other events in subsequent diffs)
* Add new column to the Scuba destination
Reviewed By: blakef
Differential Revision: D57140479
fbshipit-source-id: 7ea813b1b4d53a282873fa95c8ee82e5d6f3d1d3
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/44498
Changelog: [internal]
Quick refactor to reduce commit noise in the following diff in the stack
Reviewed By: blakef
Differential Revision: D57140480
fbshipit-source-id: aa1fef83d5347b8a11651d3d5c4112b4adf7a7d5
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/44092
In a setup where a device retrieves a bundle from `http://127.0.0.1:8081`, but this is tunnelled to a remote host with only an IPv6 stack (eg, FB dev servers), the host running the inspector-proxy will fail to fetch source or source maps from 127.0.0.1 despite typically being on the same host (indeed, process) as Metro.
This causes a surprising inconsistency where using a bundler URL of `localhost` from the device results in source maps being inlined into `Debugger.scriptParsed`, but using a bundler URL of `127.0.0.1` causes inspector-proxy to fall back to preserving URLs, which are typically fetched lazily by CDT later.
This should be unnecessary once we've implemented CDP `Network.loadNetworkResource` and removed `Debugger.scriptParsed` rewriting, but for now it brings IPv6 tunnelled servers in line with local servers.
Changelog:
[General][Changed] Inspector proxy: Rewrite 127.0.0.1 to localhost in source map URLs for better IPv4->IPv6 tunnelling support.
Reviewed By: motiz88
Differential Revision: D56138742
fbshipit-source-id: b65c9cc8225a0ed54cf32171f640ef9e6408c762
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/44086
When a debugger frontend is connected to inspector-proxy via another proxy or tunnel that times out on idle (such as [VS Code's remote tunnel](https://github.com/microsoft/vscode/blob/main/src/vs/platform/tunnel/node/tunnelService.ts)), the connection between proxy and debugger may be dropped.
In addition, when the connection is dropped without a closing handshake, the proxy does *not* detect the disconnection - no disconnect is logged to the reporter and no notifications are sent to any connected devices.
This adds a mechanism using the WebSocket-standard `ping` and `pong` frames to:
1. Keep the connection alive
2. Detect when the debugger has gone away
Note that as all WebSocket clients already **must** reply to a ping with a pong, this is non-breaking for compliant implementations: https://datatracker.ietf.org/doc/html/rfc6455#section-5.5.2
Changelog:
[General][Added] Inspector proxy: Add ping/pong keepalive to debugger connections.
Reviewed By: hoxyq
Differential Revision: D56069185
fbshipit-source-id: e322de631c652a502f3d554c15ed5412a751ee04
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43697
Currently, messages from a device are handled by async handlers added to a promise chain.
If a handler rejects, the end of the chain becomes a rejected promise, picked up only asynchronously by Metro's global `unhandledRejection` handler.
This triggers a warning from Node.js, and worse, prevents any `then()` callback chained by subsequent messages from being invoked at all.
Handlers *should* attempt to gracefully deal with errors (as we do with source map fetching errors, for example), but this diff adds a catch-all fallback for anything we might've missed (in this case, a frontend socket disconnecting while we're busy fetching a source map). Errors are caught and logged to EventReporter.
**To follow**: Gracefully handle socket disconnections while an async handler is working or queued.
Changelog:
[General][Fixed] Inspector proxy: prevent errors proxying a device message from blocking the handler queue or spamming logs.
Reviewed By: EdmondChuiHW
Differential Revision: D55482735
fbshipit-source-id: bb726218495e105f9cb4f723a1d110c9815abdef
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43686
Under Node 20, the use of `new Buffer(string)` is deprecated and logs a warning. This replaces it with the recommended `Buffer.from(string)`.
Changelog:
[General][Fixed] FIx "Buffer() is deprecated" warning from debugger proxy.
Reviewed By: huntie
Differential Revision: D55472025
fbshipit-source-id: 8b5af9e2d7e026cbdf6aa68f71ff0f856fb164db
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43639
Currently, the `react-native/dev-middleware` inspector proxy intercepts `Debugger.scriptParsed` notifications from the target and replaces `sourceMapURL` with a data uri, via an async fetch from Metro. During this async fetch, other notifications from the debugger may pass through the proxy, which results in the frontend receiving them before `Debugger.scriptParsed`.
This reordering causes problems in breakpoint resolution and pausing, because `Debugger.breakpointResolved` and `Debugger.paused` events may reference `scriptId`s unknown to the frontend while the corresponding `Debugger.scriptParsed` is delayed.
In particular, breakpoint UI state and backend state can fall out of sync, and breakpoints hit may open to the incorrect source location.
This diff modifies the proxy to use a simple per-target promise queue to ensure messages are handled in the order they were received from the target.
Changelog:
[General][Fixed] Fix breakpoints opening to incorrect location or disappearing from debugger frontend UI.
Reviewed By: motiz88
Differential Revision: D55200617
fbshipit-source-id: 27c95f822266875ed668d0bf8a525da49554cafd
Summary:
This is a proposal for the `react-native/dev-middleware` package, to allow implementers to extend the CDP capabilities of the `InspectorProxy`. It's unfortunately needed until we can move to the native Hermes CDP layer.
At Expo, we extend the CDP capabilities of this `InspectorProxy` by injecting functionality on the device level. This proposed API does the same, but without having to overwrite internal functions of both the `InspectorProxy` and `InspectorDevice`.
A good example of this is the network inspector's capabilities. This currently works through the inspection proxy, and roughly like:
- Handle any incoming `Expo(Network.receivedResponseBody)` from the _**device**_, store it, and stop event from propagating
- Handle the incoming `Network.getResponseBody` from the _**debugger**_, return the data, and stop event from propagating.
This API brings back that capability in a more structured way.
## API:
```ts
import { createDevMiddleware } from 'react-native/dev-middleware';
const { middleware, websocketEndpoints } = createDevMiddleware({
unstable_customInspectorMessageHandler: ({ page, deviceInfo, debuggerInfo }) => {
// Do not enable handler for page other than "SOMETHING", or for vscode debugging
// Can also include `page.capabilities` to determine if handler is required
if (page.title !== 'SOMETHING' || debuggerInfo.userAgent?.includes('vscode')) {
return null;
}
return {
handleDeviceMessage(message) {
if (message.type === 'CDP_MESSAGE') {
// Do something and stop message from propagating with return `true`
return true;
}
},
handleDebuggerMessage(message) {
if (message.type === 'CDP_MESSAGE') {
// Do something and stop message from propagating with return `true`
return true;
}
},
};
},
});
```
## Changelog:
<!-- Help reviewers and the release process by writing your own changelog entry.
Pick one each for the category and type tags:
For more details, see:
https://reactnative.dev/contributing/changelogs-in-pull-requests
-->
[GENERAL] [ADDED] - Add inspector proxy device message middleware API
Pull Request resolved: https://github.com/facebook/react-native/pull/43291
Test Plan: See added tests and code above
Reviewed By: huntie
Differential Revision: D54804503
Pulled By: motiz88
fbshipit-source-id: ae918dcd5b7e76d3fb31db4c84717567ae60fa96
Summary:
At Expo, we use [Expo Tools](https://github.com/expo/vscode-expo/blob/main/src/expoDebuggers.ts) to connect the [built-in vscode-js-debug](https://github.com/microsoft/vscode-js-debug) to Hermes.
Since there are a few differences in vscode vs chrome devtools, we need to enable a couple of modifications through the [`customMessageHandler` API](https://github.com/facebook/react-native/pull/43291). Unfortunately, vscode itself doesn't set the `user-agent` header when connecting to the inspector proxy. Becuase of that, we'd need a fallback to "manually" mark the debugger as being vscode ([we use this query parameter here](https://github.com/expo/vscode-expo/blob/main/src/expoDebuggers.ts#L208)).
This PR supports setting the `user-agent` through `?userAgent=` when the header is not set.
## Changelog:
<!-- Help reviewers and the release process by writing your own changelog entry.
Pick one each for the category and type tags:
[ANDROID|GENERAL|IOS|INTERNAL] [BREAKING|ADDED|CHANGED|DEPRECATED|REMOVED|FIXED|SECURITY] - Message
For more details, see:
https://reactnative.dev/contributing/changelogs-in-pull-requests
-->
[GENERAL] [ADDED] - Fallback to query parameter based `user-agent` when header is unset
Pull Request resolved: https://github.com/facebook/react-native/pull/43364
Test Plan:
- Install [Expo Tools](https://marketplace.visualstudio.com/items?itemName=expo.vscode-expo-tools)
- Start Metro with this change.
- Connect a device.
- Run the vscode command `"Expo: Debug Expo app ..."`
- Debugger should connect, and have it's user-agent marked as:
`vscode/1.87.0 vscode-expo-tools/1.3.0`
Reviewed By: huntie
Differential Revision: D54804556
Pulled By: motiz88
fbshipit-source-id: 1ff558ba5350811ad042d08a713438e046759feb
Summary:
The inspector proxy is inlining source maps on `Debugger.scriptParsed` CDP events. The inlining prevents Chrome DevTools from downloading this remotely, as that's not supported in newer versions anymore.
The current implementation locks this inlining mechanism to just `localhost` and/or `127.0.0.1` addresses, making it incompatible with LAN or tunnel device connections.
This PR removes that limitation to allow source map inlining on these LAN and tunnel connections.
## Changelog:
<!-- Help reviewers and the release process by writing your own changelog entry.
Pick one each for the category and type tags:
[ANDROID|GENERAL|IOS|INTERNAL] [BREAKING|ADDED|CHANGED|DEPRECATED|REMOVED|FIXED|SECURITY] - Message
For more details, see:
https://reactnative.dev/contributing/changelogs-in-pull-requests
-->
[GENERAL][FIXED] Allow Inspector proxy to inline source maps on LAN connections
Pull Request resolved: https://github.com/facebook/react-native/pull/43307
Test Plan:
- See added test
- Start Metro and connect a device over LAN, open the chrome devtools
Reviewed By: huntie
Differential Revision: D54485247
Pulled By: robhogan
fbshipit-source-id: 6fcb0c6dd762d2f0a013497ba0a1126095b9130b
Summary:
This adds the `nativeNetworkInspection` target capability flag, to enable/disable the proxy-side network inspection handling.
## Changelog:
<!-- Help reviewers and the release process by writing your own changelog entry.
Pick one each for the category and type tags:
[ANDROID|GENERAL|IOS|INTERNAL] [BREAKING|ADDED|CHANGED|DEPRECATED|REMOVED|FIXED|SECURITY] - Message
For more details, see:
https://reactnative.dev/contributing/changelogs-in-pull-requests
-->
[GENERAL][ADDED] Add inspector proxy `nativeNetworkInspection` target capability flag
Pull Request resolved: https://github.com/facebook/react-native/pull/43310
Test Plan:
Once this lands, and is published through `react-native/dev-middleware`, we (Expo) can disable the proxy-side network inspection handling.
See https://github.com/expo/expo/pull/27425/commits/1a1b601a29fbc5766628238db7259121689f6cd6 on PR expo/expo#27425
Reviewed By: christophpurrer, motiz88
Differential Revision: D54486516
Pulled By: huntie
fbshipit-source-id: cc151349c816fb3866d3ec07af1a29a5f4ff9b00
Summary:
Changelog: [General][Fixed] Re-enable listing Hermes debugger targets in chrome://inspect, broken in 0.74 RC
Fixes https://github.com/facebook/react-native/issues/43259.
Reverts D52958725 and fixes the original `Content-Length` Unicode bug using a different approach.
Reviewed By: fabriziocucci
Differential Revision: D54409847
fbshipit-source-id: ed5bb464ab67f37535947646b124814d8bbf797c
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/42885
## Context
We're introducing the concept of **capability flags** to provide granular control of behaviours in the Inspector Proxy, to replace the recently added `type: 'Legacy' | 'Modern'` target switch.
A capability flag disables a specific feature/hack in the Inspector Proxy layer by indicating that the target supports one or more modern CDP features.
## This diff
Following D53355413, we're now able to remove the previous `type: 'Legacy' | 'Modern'` page concept, implemented in this diff.
Changelog: [Internal]
Reviewed By: robhogan
Differential Revision: D53358480
fbshipit-source-id: 62e53a1bd60760291ada3479121dfca9e1f6edbc
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/42818
## Context
We're introducing the concept of **capability flags** to provide granular control of behaviours in the Inspector Proxy, to replace the recently added `type: 'Legacy' | 'Modern'` target switch.
A capability flag disables a specific feature/hack in the Inspector Proxy layer by indicating that the target supports one or more modern CDP features.
## This diff
Implements a second granular flag, `nativeSourceCodeFetching`, and adds tests for this.
Changelog: [Internal]
Reviewed By: robhogan
Differential Revision: D53352242
fbshipit-source-id: 94b62d84c731c903c5f99f8206d5c91bc501d030
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/42817
## Context
We're introducing the concept of **capability flags** to provide granular control of behaviours in the Inspector Proxy, to replace the recently added `type: 'Legacy' | 'Modern'` target switch.
A capability flag disables a specific feature/hack in the Inspector Proxy layer by indicating that the target supports one or more modern CDP features.
## This diff
- Implements capability flags in `InspectorProxy`, via an optional `"capabilities"` key returned by a device's CDP server.
- Wires up an initial flag, `nativePageReloads`, to disable the legacy "React Native Experimental (Improved Chrome Reloads)" page and emulated page reload behaviour.
Changelog: [Internal]
Reviewed By: robhogan
Differential Revision: D53352244
fbshipit-source-id: 622fc6028174919b9bf776e3ac52724d97ca2734
Summary:
This change removes Content-Length header from proxy inspector response.
The presence of this header was resulting in the response being cropped under some circumstances because of erroneously calculated length.
The `Content-Length` header value represents the number of bytes in the response. In the code, `string.length` was used to calculate that value, but in JavaScript it gives the number of characters in a string instead of its size in bytes. Specifically, if there are some UTF characters in the string that occupy more than byte, there would be a mismatch in this size. This mismatch resulted in the response being cropped.
The easiest way to reproduce this problem is to set the simulator name to contain a two-byte UTF character.
This change works according to the HTTP spec, which states that when Content-Length is not present, the end of the response stream indicates the end of the response. Since in the code `response.end(data)` is use, it terminates the stream and hence there is no need to provide the length in the header.
## Changelog:
[GENERAL] [FIXED] - fix issue with debugger not working when device name contain two-byte UTF characters
Pull Request resolved: https://github.com/facebook/react-native/pull/42590
Test Plan:
1. Change your iOS simulator name to contain some two-byte UTF character (for example this one: "–")
2. Run metro and connect your app with it
3. Go to http://localhost:8081/json/list in your browser – see the response being marked invalid as it is cropped
4. Apply the change and see that the resulting JSON in the response is now correct
5. Open debugger workflow to confirm it sees the connected device
Reviewed By: robhogan
Differential Revision: D52958725
Pulled By: motiz88
fbshipit-source-id: 92c32893cbbf8552237585d824e4a44737fa3968
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/42303
Changelog: [Internal]
Adds a coarse-grained mechanism to `inspector-proxy` for distinguishing between legacy and modern debug targets. The guiding principles are:
1. `inspector-proxy` does not interfere in the CDP message stream between the debugger frontend and a modern target, or in the lifecycle of a target.
2. Legacy runtimes (current React Native, React Native Desktop, etc) that rely on `inspector-proxy`'s existing invasive semantics must continue to work seamlessly for now. We'll decide on the right time to deprecate/remove this legacy code in the future.
NOTE: This is an experimental addition to the proxy protocol that may be replaced at any time.
Reviewed By: hoxyq
Differential Revision: D50967795
fbshipit-source-id: bb9c39a8fe755ef3661e2c61507dd324d8dc8894
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/42282
Changelog: [Internal] `inspector-proxy` now assumes each app will report pages with locally unique IDs.
In order to simplify some upcoming logic changes in `inspector-proxy`, in this diff we begin to enforce the assumption that each app ( = platform-specific implementation of `InspectorPackagerConnection`) assigns a locally unique ID to each inspector page. The inspector proxy will silently drop page descriptors that have conflicting IDs, and log a message to `debug()`.
NOTE: As an implementation detail, integrators may use `DEBUG=Metro:InspectorProxy` to see debug messages from `inspector-proxy`.
Reviewed By: huntie
Differential Revision: D50969752
fbshipit-source-id: a4e6faa91d97594fc5343ce4bee66233523cd175
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/41342
While rewriting `Debugger.getScriptSource` messages to fetch code and source map over HTTP, we weren't checking the status code of the fetch calls. This diff fixes that and adds corresponding tests (as well as for the filesystem error case).
Changelog: [Internal]
Reviewed By: robhogan
Differential Revision: D51013054
fbshipit-source-id: 58e7bb9fcd6a3cf92329b43fb8a139093c80d305
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/41341
* Extends `dev-middleware`'s test utilities to enable testing against an HTTPS server with a self-signed cert.
* Runs the CDP transport integration tests (D51002261) using both HTTP and HTTPS.
* Adds a test to explicitly cover the `ws=...` / `wss=...` variation in `devtoolsFrontendUrl` first introduced in D49158227.
Changelog: [Internal]
Reviewed By: blakef
Differential Revision: D51006835
fbshipit-source-id: df3db8cd865898248cd0d8f307f75949a7f313fd
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/41343
Changelog: [Internal]
Adds basic tests for two-way communication between a debugger (frontend) and a target (backend) using CDP over `inspector-proxy`.
Reviewed By: blakef
Differential Revision: D51002261
fbshipit-source-id: 44e571f89437c26e76ef6e6192b2bf6244665cf0
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/41331
`inspector-proxy`'s `Device` class currently leaks a `setInterval` handle. This is mostly harmless in current usage. In the test suite added up the stack, it shows up as a leak that prevents Jest from exiting cleanly, so let's clean it up properly.
Changelog: [Internal]
Reviewed By: blakef
Differential Revision: D51002263
fbshipit-source-id: ca36797ce1196aa049ceb3a8e96ee53d34893fdc