This is necessary to simplify the component stack handling to make way
for owner stacks. It also solves some hacks that we used to have but
don't quite make sense. It also solves the problem where things like key
warnings get silenced in RSC because they get deduped. It also surfaces
areas where we were missing key warnings to begin with.
Almost every type of warning is issued from the renderer. React Elements
are really not anything special themselves. They're just lazily invoked
functions and its really the renderer that determines there semantics.
We have three types of warnings that previously fired in
JSX/createElement:
- Fragment props validation.
- Type validation.
- Key warning.
It's nice to be able to do some validation in the JSX/createElement
because it has a more specific stack frame at the callsite. However,
that's the case for every type of component and validation. That's the
whole point of enableOwnerStacks. It's also not sufficient to do it in
JSX/createElement so we also have validation in the renderers too. So
this validation is really just an eager validation but also happens
again later.
The problem with these is that we don't really know what types are valid
until we get to the renderer. Additionally, by placing it in the
isomorphic code it becomes harder to do deduping of warnings in a way
that makes sense for that renderer. It also means we can't reuse logic
for managing stacks etc.
Fragment props validation really should just be part of the renderer
like any other component type. This also matters once we add Fragment
refs and other fragment features. So I moved this into Fiber. However,
since some Fragments don't have Fibers, I do the validation in
ChildFiber instead of beginWork where it would normally happen.
For `type` validation we already do validation when rendering. By
leaving it to the renderer we don't have to hard code an extra list.
This list also varies by context. E.g. class components aren't allowed
in RSC but client references are but we don't have an isomorphic way to
identify client references because they're defined by the host config so
the current logic is flawed anyway. I kept the early validation for now
without the `enableOwnerStacks` since it does provide a nicer stack
frame but with that flag on it'll be handled with nice stacks anyway. I
normalized some of the errors to ensure tests pass.
For `key` validation it's the same principle. The mechanism for the
heuristic is still the same - if it passes statically through a parent
JSX/createElement call then it's considered validated. We already did
print the error later from the renderer so this also disables the early
log in the `enableOwnerStacks` flag.
I also added logging to Fizz so that key warnings can print in SSR logs.
Flight is a bit more complex. For elements that end up on the client we
just pass the `validated` flag along to the client and let the client
renderer print the error once rendered. For server components we log the
error from Flight with the server component as the owner on the stack
which will allow us to print the right stack for context. The factoring
of this is a little tricky because we only want to warn if it's in an
array parent but we want to log the error later to get the right debug
info.
Fiber/Fizz has a similar factoring problem that causes us to create a
fake Fiber for the owner which means the logs won't be associated with
the right place in DevTools.
This is similar to #28771 but for isomorphic. We need a make over for
these dispatchers anyway so this is the first step. Also helps flush out
some internals usage that will break anyway.
It flattens the inner mutable objects onto the ReactSharedInternals.
This implements the concept of a DEV-only "owner" for Server Components.
The owner concept isn't really super useful. We barely use it anymore,
but we do have it as a concept in DevTools in a couple of cases so this
adds it for parity. However, this is mainly interesting because it could
be used to wire up future owner-based stacks.
I do this by outlining the DebugInfo for a Server Component
(ReactComponentInfo). Then I just rely on Flight deduping to refer to
that. I refer to the same thing by referential equality so that we can
associate a Server Component parent in DebugInfo with an owner.
If you suspend and replay a Server Component, we have to restore the
same owner. To do that, I did a little ugly hack and stashed it on the
thenable state object. Felt unnecessarily complicated to add a stateful
wrapper for this one dev-only case.
The owner could really be anything since it could be coming from a
different implementation. Because this is the first time we have an
owner other than Fiber, I have to fix up a bunch of places that assumes
Fiber. I mainly did the `typeof owner.tag === 'number'` to assume it's a
Fiber for now.
This also doesn't actually add it to DevTools / RN Inspector yet. I just
ignore them there for now.
Because Server Components can be async the owner isn't tracked after an
await. We need per-component AsyncLocalStorage for that. This can be
done in a follow up.
I'm a bit ambivalent about this one because it's not the main strategy
that I plan on pursuing. I plan on replacing most DEV-only specific
stacks like `console.error` stacks with a new take on owner stacks and
native stacks. The future owner stacks may or may not be exposed to
error boundaries in DEV but if they are they'd be a new errorInfo
property since they're owner based and not available in prod.
The use case in `console.error` mostly goes away in the future so this
PR is mainly for error boundaries. It doesn't hurt to have it in there
while I'm working on the better stacks though.
The `componentStack` property exposed to error boundaries is more like
production behavior similar to `new Error().stack` (which even in DEV
won't ever expose owner stacks because `console.createTask` doesn't
affect these). I'm not sure it's worth adding server components in DEV
(this PR) because then you have forked behavior between dev and prod.
However, since even in the future there won't be any other place to get
the *parent* stack, maybe this can be useful information even if it's
only dev. We could expose a third property on errorInfo that's DEV only
and parent stack but including server components. That doesn't seem
worth it over just having the stack differ in dev and prod.
I don't plan on adding line/column number to these particular stacks.
A follow up could be to add this to Fizz prerender too but only in DEV.
Along with all the places using it like the `_debugSource` on Fiber.
This still lets them be passed into `createElement` (and JSX dev
runtime) since those can still be used in existing already compiled code
and we don't want that to start spreading to DOM attributes.
We used to have a DEV mode that compiles the source location of JSX into
the compiled output. This was nice because we could get the actual call
site of the JSX (instead of just somewhere in the component). It had a
bunch of issues though:
- It only works with JSX.
- The way this source location is compiled is different in all the
pipelines along the way. It relies on this transform being first and the
source location we want to extract but it doesn't get preserved along
source maps and don't have a way to be connected to the source hosted by
the source maps. Ideally it should just use the mechanism other source
maps use.
- Since it's expensive it only works in DEV so if it's used for
component stacks it would vary between dev and prod.
- It only captures the callsite of the JSX and not the stack between the
component and that callsite. In the happy case it's in the component but
not always.
Instead, we have another zero-cost trick to extract the call site of
each component lazily only if it's needed. This ensures that component
stacks are the same in DEV and PROD. At the cost of worse line number
information.
The better way to get the JSX call site would be to get it from `new
Error()` or `console.createTask()` inside the JSX runtime which can
capture the whole stack in a consistent way with other source mappings.
We might explore that in the future.
This removes source location info from React DevTools and React Native
Inspector. The "jump to source code" feature or inspection can be made
lazy instead by invoking the lazy component stack frame generation. That
way it can be made to work in prod too. The filtering based on file path
is a bit trickier.
When redesigned this UI should ideally also account for more than one
stack frame.
With this change the DEV only Babel transforms are effectively
deprecated since they're not necessary for anything.
## Summary
There's a bug with the existing stack comparison algorithm in
`describeNativeComponentFrame` — specifically how it attempts to find a
common root frame between the control and sample stacks. This PR
attempts to fix that bug by injecting a frame that can have a guaranteed
string in it for us to search for in both stacks to find a common root.
## Brief Background/How it works now
Right now `describeNativeComponentFrame` does the following to leverage
native browser/VM stack frames to get details (e.g. script path, row and
col #s) for a single component:
1. Throwing and catching a control error in the function
2. Calling the component which should eventually throw an error (most of
the time), that we'll catch as our sample error.
3. Diffing the stacks in the control and sample errors to find the line
which should represent our component call.
## What's broken
To account for potential stack trace truncation, the stack diffing
algorithm first attempts to find a common "root" frame by inspecting the
earliest frame of the sample stack and searching for an identical frame
in the control stack starting from the bottom. However, there are a
couple of scenarios which I've hit that cause the above approach to not
work correctly.
First, it's possible that for render passes of extremely large component
trees to have a lot of repeating internal react function calls, which
can result in an incorrect common or "root" frame found. Here's a small
example from a stack trace using React Fizz for SSR.
Our control frame can look like this:
```
Error:
at Fake (...)
at construct (native)
at describeNativeComponentFrame (...)
at describeClassComponentFrame (...)
at getStackByComponentStackNode (...)
at getCurrentStackInDEV (...)
at renderNodeDestructive (...)
at renderElement (...)
at renderNodeDestructiveImpl (...) // <-- Actual common root frame with the sample stack
at renderNodeDestructive (...)
at renderElement (...)
at renderNodeDestructiveImpl (...) // <-- Incorrectly chosen common root frame
at renderNodeDestructive (...)
```
And our sample stack can look like this:
```
Error:
at set (...)
at PureComponent (...)
at call (native)
at apply (native)
at ErrorBoundary (...)
at construct (native)
at describeNativeComponentFrame (...)
at describeClassComponentFrame (...)
at getStackByComponentStackNode (...)
at getCurrentStackInDEV (...)
at renderNodeDestructive (...)
at renderElement (...)
at renderNodeDestructiveImpl (...) // <-- Root frame that's common in the control stack
```
Here you can see that the earliest trace in the sample stack, the
`renderNodeDestructiveImpl` call, can exactly match with multiple
`renderNodeDestructiveImpl` calls in the control stack (including file
path and line + col #s). Currently the algorithm will chose the
earliest/last frame with the `renderNodeDestructiveImpl` call (which is
the second last frame in our control stack), which is incorrect. The
actual matching frame in the control stack is the latest or first frame
(when traversing from the top) with the `renderNodeDestructiveImpl`
call. This leads to the rest of the stack diffing associating an
incorrect frame (`at getStackByComponentStackNode (...)`) for the
component.
Another issue with this approach is that it assumes all VMs will
truncate stack traces at the *bottom*, [which isn't the case for the
Hermes
VM](df07cf713a/lib/VM/JSError.cpp (L688-L699))
which **truncates stack traces in the middle**, placing a
```
at renderNodeDestructiveImpl (...)
... skipping {n} frames
at renderNodeDestructive (...)
```
line in the middle of the stack trace for all stacks that contain more
than 100 traces. This causes stack traces for React Native apps using
the Hermes VM to potentially break for large component trees. Although
for this specific case with Hermes, it's possible to account for this by
either manually grepping and removing the `... skipping` line and
everything below it (see draft PR: #26999), or by implementing the
non-standard `prepareStackTrace` API which Hermes also supports to
manually generate a stack trace that truncates from the bottom ([example
implementation](https://github.com/facebook/react/compare/main...KarimP:react:component-stack-hermes-fix)).
## The Fix
I found different ways to go about fixing this. The first was to search
for a common stack frame starting from the top/latest frame. It's a
relatively small change ([see
implementation](https://github.com/facebook/react/compare/main...KarimP:react:component-stack-fix-2)),
although it is less performant by being n^2 (albeit with `n`
realistically being <= 5 here). It's also a bit more buggy for class
components given that different VMs insert a different amount of
additional lines for new/construct calls...
Another fix would be to actually implement a [longest common
substring](https://en.wikipedia.org/wiki/Longest_common_substring)
algorithm, which can also be roughly n^2 time (assuming the longest
common substring between control and sample will be most of the sample
frame).
The fix I ended up going with was have the lines that throw the control
error and the lines that call/instantiate the component be inside a
distinct method under an object property
(`"DescribeNativeComponentFrameRoot"`). All major VMs (Safari's
JavaScriptCore, Firefox's SpiderMonkey, V8, Hermes, and Bun) should
display the object property name their stack trace. I've also set the
`name` and `displayName` properties for method as well to account for
minification, any advanced optimizations (e.g. key crushing), and VM
inconsistencies (both Bun and Safari seem to exclusively use the value
under `displayName` and not `name` in traces for methods defined under
an object's own property...).
We can then find this "common" frame by simply finding the line that has
our special method name (`"DescribeNativeComponentFrameRoot"`), and the
rest of the code to determine the actual component line works as
expected. If by any chance we don't find a frame with our special method
name in either control or sample stack traces, we then revert back to
the existing approach mentioned above by searching for the last line of
the sample frame in the control frame.
## How did you test this change?
1. There are bunch of existing tests that ensure a properly formatted
component trace is logged for certain scenarios, so I ensured the
existing full test suite passed
2. I threw an error in a component that's deep in the component
hierarchy of a large React app (facebook) to ensure there's stack trace
truncation, and ensured the correct component stack trace was logged for
Chrome, Safari, and Firefox, and with and without minification.
3. Ran a large React app (facebook) on the Hermes VM, threw an error in
a component that's deep in the component hierarchy, and ensured that
component frames are generated despite stack traces being truncated in
the middle.
Adds a development warning to complement the error introduced by
https://github.com/facebook/react/pull/27019.
We can detect and warn about async client components by checking the
prototype of the function. This won't work for environments where async
functions are transpiled, but for native async functions, it allows us
to log an earlier warning during development, including in cases that
don't trigger the infinite loop guard added in
https://github.com/facebook/react/pull/27019. It does not supersede the
infinite loop guard, though, because that mechanism also prevents the
app from crashing.
I also added a warning for calling a hook inside an async function. This
one fires even during a transition. We could add a corresponding warning
to Flight, since hooks are not allowed in async Server Components,
either. (Though in both environments, this is better handled by a lint
rule.)
Added an explicit type to all $FlowFixMe suppressions to reduce
over-suppressions of new errors that might be caused on the same lines.
Also removes suppressions that aren't used (e.g. in a `@noflow` file as
they're purely misleading)
Test Plan:
yarn flow-ci
The old version of prettier we were using didn't support the Flow syntax
to access properties in a type using `SomeType['prop']`. This updates
`prettier` and `rollup-plugin-prettier` to the latest versions.
I added the prettier config `arrowParens: "avoid"` to reduce the diff
size as the default has changed in Prettier 2.0. The largest amount of
changes comes from function expressions now having a space. This doesn't
have an option to preserve the old behavior, so we have to update this.
* Facebook -> Meta in copyright
rg --files | xargs sed -i 's#Copyright (c) Facebook, Inc. and its affiliates.#Copyright (c) Meta Platforms, Inc. and affiliates.#g'
* Manual tweaks
This PR adds a component stack field to the `schedule-state-update` event. The algorithm is as follows:
* During profiling, whenever a state update happens collect the parents of the fiber that caused the state update and store it in a map
* After profiling finishes, post process the `schedule-state-update` event and using the parent fibers, generate the component stack by using`describeFiber`, a function that uses error throwing to get the location of the component by calling the component without props.
---
Co-authored-by: Blake Friedman <blake.friedman@gmail.com>
## Summary
We're experiencing some issues internally where the component stack is
getting into our way of fixing them as it causes the page to become
unresponsive. This adds a flag so that we can disable this feature as a
temporary workaround.
More internal context: https://fburl.com/go9yoklm
## Test Plan
I tried to default this flag to `__VARIANT__` but the variant tests
(`yarn test-www --variant`) started to fail across the board since a lot
of tests depend on the component tree, things like this:
https://user-images.githubusercontent.com/458591/100771192-6a1e1c00-33fe-11eb-9ab0-8ff46ba378a2.png
So, it seems to work :-)
Given that it's unhandy to update the hundreds of tests that are failing
I decided to hard code this to `false` like we already do for some other
options.
* Remove Blocks
* Remove Flight Server Runtime
There's no need for this now that the JSResource is part of the bundler
protocol. Might need something for Webpack plugin specifically later.
* Devtools
This could be used to do custom formatting of the stack trace in a way
that isn't compatible with how we use it. So we disable it while we use
it.
In theory we could call this ourselves with the result of our stack.
It would be a lot of extra production code though. My personal opinion
is that this should always be done server side instead of on the client.
We could expose a custom parser that converts it and passes it through
prepareStackTrace as structured data. That way it's external and doesn't
have to be built-in to React.
* Add more edge cases to fixture
Also adjust some expectations. I think the column should ideally be 1 but varies.
The Example row is one line off because it throws on the hook but should ideally be the component.
Similarly class components with constructors may have the line in the constructor.
* Account for the construct call taking a stack frame
We do this by first searching for the first different frame, then find
the same frames and then find the first different frame again.
* Throw controls
Otherwise they don't get a stack frame associated with them in IE.
* Protect against generating stacks failing
Errors while generating stacks will bubble to the root. Since this technique
is a bit sketchy, we should probably protect against it.
* Don't construct the thing that throws
Instead, we pass the prototype as the "this". It's new every time anyway.
* Implement component stack extraction hack
* Normalize errors in tests
This drops the requirement to include owner to pass the test.
* Special case tests
* Add destructuring to force toObject which throws before the side-effects
This ensures that we don't double call yieldValue or advanceTime in tests.
Ideally we could use empty destructuring but ES lint doesn't like it.
* Cache the result in DEV
In DEV it's somewhat likely that we'll see many logs that add component
stacks. This could be slow so we cache the results of previous components.
* Fixture
* Add Reflect to lint
* Log if out of range.
* Fix special case when the function call throws in V8
In V8 we need to ignore the first line. Normally we would never get there
because the stacks would differ before that, but the stacks are the same if
we end up throwing at the same place as the control.
We can't patch the row. We could give these their own "built-in" stack
frame since they're conceptually HoCs. However, from a debugging
perspective this is not very useful meta data and quite noisy. So I'm
just going to exclude them.
* Add feature flag
* Split stack from current fiber
You can get stack from any fiber, not just current.
* Refactor description of component frames
These should use fiber tags for switching. This also puts the relevant code
behind DEV flags.
* We no longer expose StrictMode in component stacks
They're not super useful and will go away later anyway.
* Update tests
Context is no longer part of SSR stacks. This was already the case on the
client.
forwardRef no longer is wrapped on the stack. It's still in getComponentName
but it's probably just noise in stacks. Eventually we'll remove the wrapper
so it'll go away anyway. If we use native stack frames they won't have this
extra wrapper.
It also doesn't pick up displayName from the outer wrapper. We could maybe
transfer it but this will also be fixed by removing the wrapper.
* Forward displayName onto the inner function for forwardRef and memo in DEV
This allows them to show up in stack traces.
I'm not doing this for lazy because lazy is supposed to be called on the
consuming side so you shouldn't assign it a name on that end. Especially
not one that mutates the inner.
* Use multiple instances of the fake component
We mutate the inner component for its name so we need multiple copies.
