**Description:**
- The option `hoist_props` now does what it's supposed to do.
- Dropping of unused properties now does not drop properties too aggressively.
- The initializer of a dropped variable declaration is now properly visited.
- Indexing with string literals is not marked as a dynamic index anymore. This is required to handle codes like c3f67ceb1e/crates/swc_ecma_minifier/tests/terser/compress/hoist_props/name_collision_1/input.js (L1-L7).
**Description:**
Note: Preserving native names is a hack, but it's used by `terser` and it's the only way to preserve the name of `class AbortSignal` while mangling without `keep_classnames: true`. We can special case `AbortSignal`, but let's just follow `terser`.
**Related issue:**
- https://github.com/vercel/next.js/pull/57904
**Description:**
This issue is more severe than I originally thought. It raises not in
array indexing, but in function calls and property mutation. We should
treat all function arguments as potentially be property mutated,
otherwise following example
```js
class A {
a = 1
toString() {
return this.a
}
}
const a = new A()
function foo(x) {
x.a++
}
const b = a + 1
foo(a)
console.log(b)
```
would be error(It should log 2, but logs 3 after compress).
As the result, massive regressions is unavoidable, since some of these
optimizations may indeed cause error. Part of them can be mitigated with
following optimization -- allow inline of ident even if its original
value is mutated. Consider
```js
export function foo(x) {
const y = x
x.a = 1
y.b = 2
}
```
If x is a primitive value, all mutations to its properties are ignored;
if x is a object, then y refers to the same object no matter what
mutation is performed.
And there's still room for more, currently following code
```js
export function foo(x) {
const y = Math.floor(x);
g(y);
}
```
But I'd rather do it in a separate PR.
**Related issue:**
- Closes#7402.
**Description:**
This PR improves the alias analyzer by distinguishing call and reference, thus reducing the number of identifiers.
---
Co-authored-by: Justin Ridgewell <justin@ridgewell.name>
**Description:**
We had two passes for joining variables. This PR removes one in the full optimizer, which is wrong.
**Related issue:**
- Closes https://github.com/swc-project/swc/issues/5989.
Safety:
For a function-local variable, an expression with side-effects would be a call, including an indirect one with a member expression.
- If the call is function-local, it will be analyzed by the analyzer and inliner will not work.
- If the call is not a function-local one, it cannot modify the local variable.
