Commit Graph

36 Commits

Author SHA1 Message Date
Austaras
bb02cdd26e
refactor(es/minifier): Simplify analyzer context (#8164) 2023-10-27 18:16:00 +00:00
Donny/강동윤
35601e4dcb
fix(es/minifier): Abort function inliner if keep_fnames is true (#8145)
**Related issue:**

 - https://github.com/vercel/next.js/issues/56408
2023-10-19 20:20:54 +00:00
Austaras
f8ca366cc1
fix(es/minifier): Abort seq inliner if var is not fn_local or reassigned (#7804)
**Description:**

It turns out that the original implementation of `infect`is incomplete
because it cannot cover function param, and the new implementation
introduced in #7772 is redundant because what
[terser](https://github.com/terser/terser/blob/master/lib/compress/tighten-body.js#L909C18-L909C28)
do is basically checking `fn_local`


**Related issue:**

 - Closes #7784
2023-08-17 05:15:59 +00:00
Austaras
ef8d12154d
fix(es/minifier): Abort seq inliner if a same var is defined in outer scope (#7772)
**Description:**

The algorithm here is directly copied from terser, I don't if it's correct or can be improved, but it does fix the issue.


**Related issue:**

 - Closes #7749
2023-08-09 20:01:59 +00:00
Austaras
ff1ad95b59
feat(es/minifier): Compress common sub expressions in sequences (#7587) 2023-07-04 05:47:37 +00:00
Donny/강동윤
47d2edd4dc
feat(es/minifier): Drop unused properties (#7534)
**Related issue:**

 - Closes #7472.
2023-07-04 05:11:33 +00:00
Austaras
5dbbbea2ef
fix(es/minifier): Mark all function params as potential property mutation (#7409)
**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.
2023-05-18 01:19:17 +00:00
Donny/강동윤
1dced17998
fix(es/minifier): Fix remapping of vars upon inlining (#7362)
**Related issue:**

 - Closes #7331.
2023-05-10 03:31:15 +00:00
Donny/강동윤
246300ae25
feat(es/minifier): Drop expressions using sequential inliner (#6936) 2023-04-29 22:19:00 +09:00
Anders Kaseorg
493a4f7042
fix(es/minifier): Remove wrong optimization of new RegExp(…) (#7091) 2023-03-17 03:22:02 +00:00
Donny/강동윤
86295ba8f2
fix(es/minifier): Preserve delete of unresolved variables (#7072)
**Related issue:**

 - Closes https://github.com/swc-project/swc/issues/7045.
2023-03-15 02:31:08 +00:00
Austaras
cfeb088c37
fix(es/minifier): Don't skip expressions with side effects from seq inliner (#7007) 2023-03-04 12:32:15 +09:00
Donny/강동윤
27ae59e77b
fix(es/minifier): Make sequential inliner respect resolution order (#6509)
**Related issue:**

 - Closes https://github.com/swc-project/swc/issues/6491.
2022-11-29 04:50:15 +00:00
Donny/강동윤
5fd7ab87b6
fix(es/minifier): Preserve op of the op-assignments in seq inliner (#6428)
**Description:**


**Related issue:**

 - Closes https://github.com/swc-project/swc/issues/6407.
2022-11-15 00:22:06 +00:00
Donny/강동윤
743a1aab4f
feat(es/minfiier): Compute more with sequential inliner (#6169) 2022-10-27 00:50:52 +00:00
Austaras
1cd7f617f9
feat(es/minifier): Inline a lazily initialized var if it's used once (#6237) 2022-10-26 06:02:22 +00:00
Donny/강동윤
86e265a024
feat(es/minifier): Support more statements in seqential inliner (#6248)
**Description:**

We now inline into the discriminant of a switch statement and into the initializer of for/for-in/for-of statements.
2022-10-26 13:37:10 +09:00
Austaras
b40d486253
fix(es/minifier): Consider function body cost while inlining function (#5342) 2022-10-15 05:47:56 +00:00
Donny/강동윤
d334c6e783
feat(es/minifier): Ignore return value if a param of IIFE is not used (#6115) 2022-10-14 02:58:00 +00:00
Donny/강동윤
66196a65be
feat(es/minifier): Inline pure array literal partially (#6099) 2022-10-12 02:29:40 +00:00
Austaras
657e5b3111
feat(es/minifier): Inline more lazily initialized vars (#6089) 2022-10-10 09:53:50 +09:00
Donny/강동윤
b0c57458c8
feat(es/minifier): Mark ref to fn as non-call in alias analyzer (#6088) 2022-10-09 09:21:54 +09:00
Donny/강동윤
9ce3df091f
feat(es/minifier): Distinguish calls in alias analyzer (#6080)
**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>
2022-10-08 03:15:18 +00:00
Austaras
c9427f1ec4
fix(es/minifier): Improve infection analysis (#6044) 2022-10-06 18:12:09 +09:00
Donny/강동윤
d65fba134d
fix(es/minifier): Fix infection analysis of sequential inliner (#6053)
**Description:**

This PR fixes the callee issue by fixing the infection analyzer.

**Related issue:**

 - Closes https://github.com/swc-project/swc/issues/6047
2022-10-06 02:46:22 +00:00
Donny/강동윤
197c4e269d
fix(es/minifier): Don't evaluate String.fromCharCode for non-ascii values (#6033) 2022-10-04 16:46:59 +09:00
Donny/강동윤
877936f626
feat(es/minifier): Clone trivial literals in sequential inliner (#6005)
**Description:**

As our minifier is two-pass by default, this is enough.

At the first pass, we inline all numeric literals using sequential inliner, and minifier can then inline constants correctly in the second pass.
2022-10-03 07:20:48 +00:00
Donny/강동윤
2b627524ac
feat(es/minifier): Ignore return values of ignored IIFEs (#6020)
**Description:**

As we are not using the return value of the function, we can call `ignore_return_value` on the argument of the return statements of IIFE.
2022-10-02 20:01:56 +09:00
Austaras
61807abf6e
feat(es/minifier): Inline function calls in more cases (#6010) 2022-10-01 18:18:45 +09:00
Donny/강동윤
aeb7f300a8
feat(es/minifier): Invoke sequential inliner for inlined IIFEs (#5991)
**Description:**

Inlined IIFEs have the exact pattern sequential inliner wants, so it's good to invoke it.
2022-09-29 11:03:06 +00:00
Donny/강동윤
f8358fbe2b
fix(es/minifier): Remove wrong variable joiner pass (#5992)
**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.
2022-09-29 19:11:18 +09:00
Donny/강동윤
8c4873f81e
test(es/minifier): Remove mangle-only snapshots (#5979)
**Description:**

We don't have to store snapshots of all fixtures.
2022-09-28 21:35:52 +09:00
Donny/강동윤
97647861bf
fix(es/minifier): Track aliasing via assignments (#5940) 2022-09-23 14:22:41 +09:00
Donny/강동윤
93964a82ea
test(es/minifier): Remove analysis snapshot (#5939)
**Description:**

This PR removes analysis snapshots as it disturbs code review process
2022-09-23 13:14:42 +09:00
Donny/강동윤
e9c704182b
feat(es/minifier): Improve analysis of reassignments (#5931) 2022-09-22 10:22:18 +00:00
Donny/강동윤
f2224132c1
fix(es/minifier): Prepend vars when dropping vars in a for loop initializer (#5930)
**Description:**

The minifier appends initializer of for statements instead of prepending it.

**Related issue:**

- https://github.com/vercel/next.js/discussions/30237#discussioncomment-3704795
2022-09-22 09:18:41 +00:00