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/44672
Swaps out and simplifies the internals of the debugger launch flow.
We observed that we could achieve better launch/windowing behaviour by passing the `--app` argument directly to the detected Chrome path.
This shares the user's default Chrome profile:
- Fixes unwanted behaviour such as a separate dock icon on macOS (which, when clicked, would launch an unwanted empty window).
- Enables settings persistence.
This change also removes the `LaunchedBrowser.kill` API.
Changelog: [Internal]
Reviewed By: hoxyq
Differential Revision: D57726649
fbshipit-source-id: fc3a715dc852a50559048d1d1c378f64aeb2013f
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/44638
Reverts the debugger launch flow to use the default `ChromeLauncher` profile. This is the approach used in the current `--experimental-debugger` experiment and by Expo.
This is motivated after a review of the tradeoffs of a guest profile — which allow us to programatically quit the browser process, however takes over system URL handling.
Changelog: [Internal]
Reviewed By: hoxyq
Differential Revision: D57619542
fbshipit-source-id: 3713e1cf8eed61e7a70ed1e4eb58f02da845155f
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/44362
Packages that are built and directly run in the monorepo no longer need to worry about
conditionally registering themselves to transpile Flow -> JS at runtime. Our build step
strips this file now.
Changelog: [Internal] changes in published packages no longer require conditional calls to Babel register.
Reviewed By: huntie
Differential Revision: D56839521
fbshipit-source-id: 6bec706c639f1ab4138e0b790be8a07654333046
Summary: Changelog: [General][Removed] `launchId` query param for `/debugger-frontend` is no longer generated automatically for each `/open-debugger` call. Caller of `/open-debugger` is now responsible for generating the `launchId`, which will be passed along to `/debugger-frontend`.
Reviewed By: robhogan
Differential Revision: D55164645
fbshipit-source-id: b83303eda77b6fb86ebf50f699d9f308676533c6
Summary:
Supports the removal of Flipper from the template in 0.74, paried with additional blog post messaging: https://reactnative.dev/blog/2024/04/22/release-0.74#removal-of-flipper-react-native-plugin.
Changelog:
[General][Changed] - Update "Open Debugger" action to print extended Flipper guidance
Reviewed By: cipolleschi
Differential Revision: D56705236
fbshipit-source-id: d7e869625262ebb02bc2454c924f832cccfbcd31
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/44180
Changelog: [General][Changed] Update Chrome launch flags for `--experimental-debugger` launch flow
Internally at Meta, we've been testing the experimental debugger launch flow with a different set of Chrome flags than are currently shipped in open source. This diff fixes those differences:
* Removes `--disable-backgrounding-occluded-windows`
* Adds `--guest`
Reviewed By: EdmondChuiHW
Differential Revision: D56418271
fbshipit-source-id: 884c5746e93cad89f17e4ef9e3ef193a2a454eb5
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
