---
We don't need to store whether a `PropertyLoad` happens within a conditional
(within its reactive scope). In fact, the PropertyLoad producing a rval often is
in a different ReactiveScope from where the rval is used.
We only need to add `#inConditionalWithinScope` when we actually visit a
reactive dependency.
Earlier PRs bailed out when the callee of an OptionalCallExpression was a
MemberExpression or OptionalMemberExpression (ie for optional method calls).
This PRs expands support for optional method calls, including when the receiver,
method, or both are optional. Even better, we don't need to add any additional
terminals or instruction variants for this case - the one new OptionalCall
terminal from earlier in the stack works for all these cases.
Tests, focusing on two key behaviors:
* Dependencies of the args are treated as conditional, since the call may not
happen
* Args cannot be memoized independently, even when that would be valid for a
non-optional call.
Implements HIR->ReactiveFunction conversion and Codegen for optional calls. We
add a new OptionalCall variant of ReactiveValue, which is a SequenceExpression
that describes the evaluation of the args and the call itself. This is then
straightforward to codgen.
Implements lowering for a subset of optional calls - specifically, we don't
(yet) support when the callee is a member expression or an optional member
expression. So `foo?.()` works but we bailout on `object?.foo()` and
`object.foo?.()`.
For `<calleee>?.(<args>)` we lower as roughly:
```
bb0:
t0 = <callee>
OptionalCall test=bb1 fallthrough=
bb1 (value):
Branch t0 consequent=bb2 alternate=bb3
bb2 (value):
...lower <args> here...
t1 = Call t0, args
StoreLocal res, t1
Goto bb4
bb3 (value):
t2 = undefined
StoreLocal res, t2
Goto bb4
bb4:
// result in `res` here
```
Adds a new `optional-call` terminal and sets up the appropriate handling in the
visitors, with lowering/reactivefunction/codegen as todos for now and
implemented in follow-ups.
---
Expand Hindley Milner type inference to infer dependent types.
Say `t` is a typevar and `t'` is some type (a built-in type, phi node, or
another typevar).
Our type equations are as follows (please edit/correct notation 😅)
- type substitution: `t = t'`,
- ~~dependent~~ polymorphic property load: `t = t'.prop`
- polymorphic function call `t = fnCall{returnType}`
- ~~dependent property call: `t = t'.prop` (only if t'.prop is a function
type)~~
- ~~dependent return type: `t = t'.[[returntype]]`~~
---
+10 −1,698 lines [[insert impacc macro]]
The ObjectShape stacks (#1350, #1358) used these tests to record changes in
inferred types (and associated ObjectShapes), reference effects, and mutable
ranges.
Now that those PRs have landed, we can delete these tests. They are somewhat
fragile (changing anytime HIR / printHIR is changed) and easily cause
rebase/merge conflicts.
---
This PR does not add inference for normal `CallExpression`s, since built-in
functions for `Array` and `Object` are usually only valid if called with a
correctly-typed `this`. If we want codegen to preserve source code semantics,
Forget should only add inferred types it is confident about.
This PR also adds `returnEffect` to FunctionSignature. `returnEffect = Store` if
this function is known to always return a captured value from `receiver` or
`args`.
---
Expand Hindley Milner type inference to infer dependent types.
Say `t` is a typevar and `t'` is some type (a built-in type, phi node, or
another typevar).
Our type equations are as follows (please edit/correct notation 😅)
- type substitution: `t = t'`,
- ~~dependent~~ polymorphic property load: `t = t'.prop`
- polymorphic function call `t = fnCall{returnType}`
- ~~dependent property call: `t = t'.prop` (only if t'.prop is a function
type)~~
- ~~dependent return type: `t = t'.[[returntype]]`~~
We limit the types of expressions allowed as switch case test values because we
our HIR doesn't yet preserve order-of-evaluation for switch test values (we
model them as being evaluated prior to entering the switch, as opposed to
lazily, when the case is reached). One common pattern internally is test case
values that are properties of a global, eg you have some bag of enum values and
are comparing against that:
```javascript
// at module scope, or imported from another module:
const OPTIONS = {FOO: 'foo'};
// in a component
switch (value) {
case OPTIONS.FOO: { ... }
}
```
This PR allows this specific case, ie member expressions where the innermost
object is a global identifier.
Now that we model the method resolution via a PropertyLoad or ComputedLoad, we
don't need to distinguish between PropertyCall and ComputedCall. These two call
variants are now combined into a single MethodCall variant.
This is the version of @mofeiZ's change for PropertyLoad, but made to work on
ComputedCall. We force the method to be evaluated in the same scope as the call
in InferReactiveScopeVariables.
---
(I'm not sure if these are already known issues. I found them while playing
around with lambda captures. They are also reproducible on main / stable)
I have some limited understanding of lambda captures after reading Sathya's
posts -- please correct if/where this is incorrect
```
function Component() {
// instr1
// instr2
const func3 = function(...) {
// func3instr1
}
}
```
We currently determine effects of captured references in `AnalyzeFunctions`,
before InferReferenceEffects.
- i.e. for some function
1. dependencies of all functions (func3.deps)
2. prefix traversal of all instructions (e.g. instr1, instr2, func3.deps,
func3instr1, ...)
- is this just an implementation decision? i.e. what is stopping us from postfix
traversal in InferReferenceEffects (e.g. instr1, instr2, func3instr1,
func3.deps)
As such, for each captured reference, `AnalyzeFunctions` needs to assign a
reference effect. We currently check `MutableRange`, which seems to miss a few
cases
- We do not model assignments to primitives correctly, since primitives do not
have a mutable range.
- We're not able to model captured (but not mutated) values correctly.
Would it be possible to consolidate `AnalyzeFunctions` into
InferReferenceEffects, using some post-order traversal (iterating over a
function's instructions to collect its dependencies + associated capture
effects)? I definitely don't understand lambdas completely, so please tell me
what I'm missing
---
This PR does not add inference for normal `CallExpression`s, since built-in
functions for `Array` and `Object` are usually only valid if called with a
correctly-typed `this`. If we want codegen to preserve source code semantics,
Forget should only add inferred types it is confident about.
This PR also adds `returnEffect` to FunctionSignature. `returnEffect = Store` if
this function is known to always return a captured value from `receiver` or
`args`.
---
I didn't properly understand Capture and Store effects previously, just
correcting those mistakes!
These functions are all synchronously mutative, so they should use Read /
Mutate, not Capture + Store
---
Expand Hindley Milner type inference to infer dependent types.
Say `t` is a typevar and `t'` is some type (a built-in type, phi node, or
another typevar).
Our type equations are as follows (please edit/correct notation 😅)
- type substitution: `t = t'`,
- ~~dependent~~ polymorphic property load: `t = t'.prop`
- polymorphic function call `t = fnCall{returnType}`
- ~~dependent property call: `t = t'.prop` (only if t'.prop is a function
type)~~
- ~~dependent return type: `t = t'.[[returntype]]`~~
How Forget currently lowers PropertyCall:
```js
// source: [[ calleeExpr ]].propertyName( [[ argExpr0 ]])
$0 = [[ calleeExpr ]]
$1 = [[ argExpr0 ]]
$2 = PropertyCall callee=$0 property="propertyName" args=[$1]
```
This PR changes the lowering:
```js
// source: [[ calleeExpr ]].propertyName( [[ argExpr0 ]])
$0 = [[ calleeExpr ]]
$1 = PropertyLoad $0 "propertyName"
$2 = [[ argExpr0 ]]
$3 = PropertyCall callee=$0 fn=$1 args=[$2]
```
From my understanding, `PropertyCall` needs the receiver to properly model JS
semantics which is something like `resolvedFn.apply(resolvedCallee, arg0, arg1,
...)`. This is additionally useful for:
- Fine-grained mutability / alias analysis. The property call is technically a
read of the resolved function, and a mutate of the callee.
- Dependency tracking. While we could special case PropertyCall, this
representation would correctly add both callee and callee.propertyName as
dependencies for PropertyCall.
e.g.
```js
let x = [];
mutate(x);
useFreeze(x);
let y = {};
y.a = x.bar();
return y;
```
Reverts #1199, which was added before we properly supported destructuring
assignment.
Next PR (changes to PropertyCall in #1384) will lower two references to the same
named identifier (the property call receiver)
The previous PR only updated simple assignment expressions (where the lvalue is
an identifier), this PR extends the same idea to all assignment variants. Note
that there is one case that doesn't work yet, which is complex destructuring
assignment as a value:
```javascript
let x = makeObject();
x.foo(([[x]] = makeObject()));
```
What happens here is that we lower the destructuring to a series of steps:
```
tmp1: Destructure Const [ tmp0 ] = makeObject();
tmp2: Destructure Reassign [ x ] = tmp0;
PropertyCall x, 'foo', [ tmp1 ]
```
Thankfully we can detect this case: if we have a const/let declaration with an
lvalue, that's invalid. See the new error test case which shows we correctly
detect & reject this case for now.
This PR subtly changes how we represent assignment expressions in order to
accurately model them _as expressions_. Specifically, the result of lowering an
assignment is now the temporary created for the assignment's lvalue. This allows
us to restore the assignment as a value (expression) during codegen. Note how
this fixes a bug and cleans up some output.
Updates ConstantPropagation so that each instruction is responsible for whether
to replace its `.value` with the resolved constant value (if found).
Specifically, for `StoreLocal` we don't want to replace the value — we want to
keep the assignment — but we do want to propagate the _result_ of the assignment
downstream. This more accurately models the semantics of assignment expressions,
and helps with subsequent PRs.
Rewrite ArrowFunctionExpression to FunctionDeclaration and compile it. This lets
us reuse all the export gating logic, rather than writing separate, specific
logic for ArrowFunctionExpression.
Cleans up duplicated code for processing call/constructor arguments. As a side
benefit, we now support spread elements for constructor arguments (and if we
want to change how we represent that, we can do it in one place).
We currently are not lowering property calls in evaluation order.
I wonder if the following lowering for a PropertyCall (with static or computed
property) is semantically equivalent to source:
1. eval + resolve receiver (store in t0)
2. eval computed property (if present)
3. resolve t0.property (binding the call to receiver and storing in t1)
4. eval args
5. eval t1(args)
Although codegen might then generate something like this (if args are named
temporaries)
```js
const tmp = receiver.property.bind(receiver);
// lower args to temporaries
tmp(arg1, arg2);
```
Adds a `DeclareLocal` instruction which represents declaring a named variable
without initializing it. Currently declarations without an initializer (`let x`)
are transformed into a declaration to undefined (`let x = undefined`) which
changes the semantics due to hoisting and TDZ (temporary dead zone). The correct
thing is to represent declaration without initialization.
These examples previously errored all the way in codegen, when we detected that
a value block (eg a `while` test expression) was declaring a new variable. We
now detect this in LeaveSSA and error. The actual fix is a bit tricky, we'd need
to add a new declaration in the nearest block scope (or selectively not DCE the
declaration if its reassigned in just this way).
If a logical or conditional expression is unused, then a phi node isn't created
for the identifier it assigns to. Then when we leave SSA form the two branches
will assign to separate values, and we aren't sure which identifier to use as
the lvalue of the resulting ReactiveInstruction (remember that
logicals/conditionals decompose into control flow in HIR, but are a single
compound instruction in ReactiveFunction). If the two sides don't assign to the
same location, it could be because of a bug in the compiler or because the value
wasn't used. Ideally we'd represent this explicitly, but for now i'm just making
this a TODO since most logicals/conditionals should have their value used.
Enables support for assignment expressions in value blocks (which includes in
loop init/test/update blocks). This was pretty straightforward, the main changes
are:
* During PropagateScopeDependencies, we currently record scope reassignments
based on `Identifier` object identity. In the case where a variable is
reassigned in multiple control-flow paths of a value block, however, there can
be multiple object identities. So we now de-dupe reassignments based on
identifier id.
* MergeOverlappingScopes now treats value blocks as regular blocks, allowing it
to correctly merge scopes from the value with other scopes from the outer block.
Otherwise this is mostly just lots of tests. Note that there is an outstanding
todo, which is that we currently error for ternaries and logicals whose value is
unused (eg `cond ? (x = 1) : null`). I'll address that in a follow-up.
This PR clarifies the logic for adjust mutable ranges of phis and their operands
during LeaveSSA. Previously we had logic in several places to determine
whether/how to extend the ranges of each phi and its operands: this occurred
while traversing reassignmentPhis (in 2+ places) and rewritePhis, as well as in
rewritePlace().
This was kind of a band-aid to make things work, but the logic was imprecise.
The actual rules are as follows:
If there is a back-edge, or the phi id is unnamed, then were extend the ranges
of the phi and its operands to min(starts) and max(ends). This ensures that the
operands are computed as one unit, ie put into a single reactive scope. For
loops this is necessary because...looping! For unnamed values this is necessary
because of the way we collapse logical and ternary expressions back to a
hierarchical ReactiveFunction — we need to make sure the final mutable range
extends from the start of the final instruction up to the end of the
logical/ternaries value blocks.
Otherwise this is a phi where operands come from predecessors and are named. If
the phi is mutated later, then we have to extend the end of each operand's range
to account for the fact that they can be mutated later. Else, we leave the
operands alone.
Behavior doesn't change, but we consolidate all of the mutable range logic in
one place.
Makes JSX memoized by default again, but adds an option to disable memoization
of JSX. Also adds a new test and fixtures directory to test the opt-in
no-jsx-memoization behavior.