Object desctructuring to remove object properties does not work after minification

[koljada]

Version 1.21.4

Describe the bug I'm using object desctructuring to remove object properties. For example: const { avatar, timeline, backup, ...copy } = { ...user }; After this line copy object does not have avatar, timeline and backup properties. But after minification not used variables avatar, timeline and backup are removed. And as a result my copy object is identical to original user object. Which is a big problem in my scenario. I tried to set RemoveUnneededCode settings to false but it didn't work.

To Reproduce

string src = """
    const func = () => {
        const user = {
            avatar: Date(),
            timeline: Date(),
            backup: Date(),
            id: 123,
            name: 'Roman',
        };
        const { avatar, timeline, backup, ...copy } = user;
        return copy;
    };
    console.log(func());
    """;
var result = Uglify.Js(src);

Minified output or stack trace const func=()=>{const n={avatar:Date(),timeline:Date(),backup:Date(),id:123,name:"Roman"},{...copy}=n;return copy};console.log(func())

Excepted output code const func=()=>{const n={avatar:Date(),timeline:Date(),backup:Date(),id:123,name:"Roman"},{avatar, timeline, backup,...copy}=n;return copy};console.log(func())

[trullock]

This is a new one.

Its because it doesn't think theyre used as it doesn't properly understand the object destructuring with the spread operator, i.e. the preceeding arguments/destructured variables change what gets ... spread into copy.

The change needs to be in one of the Analysers, can't recall off the top of my head, pull requests welcome

[failwyn]

I found an issue with destructuring objects with default values, haven’t had a chance to enter it yet, but if anyone can give ideas of where to look, I’ll try to take a pass at all if them. I’ve set aside tomorrow (5/9) to try and fix them.

[trullock]

That would be great, thanks

Without going deep into it i cant tell you exactly where, but from memory its to do with how it parses object literals. I cant remember exactly what the problem was, but the parser gets confused between two different notations/types of objects and it needed more than a quick bugfix to deal with it, that part of the parser needed reconsidering/rewriting

[trullock]

Probably in ParseObjectLiteralProperty

[failwyn]

This one is a bit more than I have time for this week, I'll try to come back to it when I have another free day.

[Matey-Tsilov]

Hello @failwyn and @trullock Is the bug mentioned by @koljada fixed now? If yes from which version on?

[trullock]

I haven't fixed it. PRs welcome