Various small fixes

[anka-213]

Three fixes that prevented code I was working on from being unpacked

• fix: Make isIIFE match what it claims to match
  More specifically:

  !function(){...}()
  

  wasn't matched, only

  !function(){...}
  

  which isn't an IIFE

• Fix: Remove buggy and redundant transform

  No need to remove import foo as foo. The desired behaviour happens automatically and other code expects the local field (as foo) to exist to not crash.

  import { foo as foo } to import { foo }
  

  will automatically be printed as

  import { foo } to import { foo }
  

  without this transform

• feat: Allow re-exports from modules in Webpack5
  If a module re-exports values from another module,
  Webpack5 would export them as

    var r = n(645); n.d(t, { foo: () => r.foo })
  

  so this adds member-expressions to the expected possible values

• feat: Allow Arrow functions in module definitions

  Example:

  (self.webpackChunk_N_E=self.webpackChunk_N_E ||
       []).push([[888],{2189: (e,t,n) => {...} }
  

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
wakaru	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 1, 2024 5:26pm

[pionxzh]

Can you help to update the test cases accordingly?

[anka-213]

Will do! I was apparently wrong in my assumption regarding the foo as foo. I thought I had tested it, but apparently my tests didn't cover those cases.

[pionxzh]

I want to know more about the "buggy" part of foo as foo transformation. What would be the real-world issue? Semantic changes or wakaru throwing error?

[anka-213]

wakaru throwing error. More specifically it was when running wakaru all somefile.js. I got crashes from two of the transforms. Let me try to reproduce it.

I believe it's a bug in the scopes library from ast-types

[anka-213]

These are the lines that the crash originated from:
https://github.com/benjamn/ast-types/blob/v0.16.1/src/scope.ts#L256-L258

As you can see it assumes that one of local, name or id exists on the binding, but if you call j.importSpecifier with only one argument none of them is set. When I tried using astexplorer, it does set local even if it's only a single value with no as.

[pionxzh]

Do you have a short reproduction? We can also patch ast-types by ourself if this is indeed a bug from ast-types.

[anka-213]

I unfortunately haven't extracted a short reproduction yet, but I have found a simple fix. Constructing a new identifier and using it for both arguments does produce the desired behavior.

I'll see if I can make a short reproduction as well.

[anka-213]

I also removed another change that caused test failures. Sorry for not running the tests before sending a PR

[pionxzh]

Don't worry.

I also updated another iife matching that should be corrected.

[pionxzh]

one more thing. Please help to update the commit message from "Fix" to fix. The Changelog tool will be mad about it 🤣

[codecov]

Codecov Report

Attention: Patch coverage is 65.62500% with 11 lines in your changes are missing coverage. Please review.

Project coverage is 89.31%. Comparing base (e099371) to head (cc4a10a).

❗ Current head cc4a10a differs from pull request most recent head 14c4377. Consider uploading reports for the commit 14c4377 to get more accurate results

Files	Patch %	Lines
packages/unpacker/src/extractors/webpack/jsonp.ts	11.11%	8 Missing ⚠️
packages/ast-utils/src/matchers/isIIFE.ts	83.33%	2 Missing ⚠️
...ckages/unpacker/src/extractors/webpack/webpack5.ts	87.50%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #128      +/-   ##
==========================================
- Coverage   89.33%   89.31%   -0.02%     
==========================================
  Files         100      100              
  Lines       12308    12316       +8     
  Branches     1632     1635       +3     
==========================================
+ Hits        10995    11000       +5     
  Misses       1259     1259              
- Partials       54       57       +3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[anka-213]

Haha, done! I also restored the isIIFE fix now that your commit fixed the test failure.

[anka-213]

Here's a minimal reproduction of the bug. It's a combination of the first test case for un-import-rename and the first test case for un-optional-chaining, but it works with any combination of un-import-rename and something that uses @wakaru/ast-utils/scope.

import { foo as a, bar as b, code } from '_';

console.log(a, b, code);

var _a;
(_a = a) === null || _a === void 0 ? void 0 : _a.b;

edit: I would add it as a test, but I don't know how to add tests that requires multiple different transforms

[pionxzh]

You can inline combined rules if there is only one test case, or declare it as a variable.

wakaru/packages/unminify/src/transformations/__tests__/un-default-parameter.spec.ts

Line 5 in e099371

const inlineTest = defineInlineTest([unFlipComparisons, transform])

wakaru/packages/unminify/src/transformations/__tests__/un-nullish-coalescing.spec.ts

Lines 122 to 125 in e099371

	defineInlineTest([transform, unParameters])('SWC - ES5',
	`
	foo !== null && foo !== void 0 ? foo : "bar";

I would recommend you put it at the test of un-import-rename 🤔

[pionxzh]

Thanks for fixing some many issues. 👍