Need better error message about incompatible constructor signature

[fictitious]

TypeScript Version: 2.7.0-dev.201xxxxx

Code

class Foo { constructor(a: string) { } }

let FooConstructor: { new(): Foo } = Foo;

Expected behavior:

Error message should tell the reason why constructor signatures are incompatible.

Actual behavior:

t.ts(3,5): error TS2322: Type 'typeof Foo' is not assignable to type 'new () => Foo'.

This is not particularly helpful, see comments below this answer, especially in comparison to the error message about incompatible methods:

let foo: { foo(a: string): void };
let foo2: { foo(): void };

foo2 = foo;

Type '{ foo(a: string): void; }' is not assignable to type '{ foo(): void; }'.
  Types of property 'foo' are incompatible.
    Type '(a: string) => void' is not assignable to type '() => void'.

So for constructor signature, ideally, the message could be

Type 'typeof Foo' is not assignable to type 'new () => Foo'.
  Types of constructor signature are incompatible.
    Type 'new(a: string) => Foo' is not assignable to type 'new() => Foo'.

[xiongmao86]

Can I take a stab of this, is anyone working on it?

[DanielRosenwasser]

Sure! Go for it. You'll need to modify diagnosticMessages.json and then run jake generate-diagnostics to create a new message. The logic is probably somewhere in signaturesRelatedTo.

To be more precise, it should be "Types of construct signature are incompatible.".

[xiongmao86]

Hi, @DanielRosenwasser. I'm looking into code and I have 2 question:

1. Which number should I pick when I am adding a new diagnostic message? The number after 2322, from 2323 to 2566 is occupied, should I choose 2567, it is a bit far from 2322 so I am asking here.

2. There are possibly 3 candidate place where I can place the call to report error message, the checker.ts is to big to provide a link so I am trying to put in here:

A:

if (getObjectFlags(source) & ObjectFlags.Instantiated && getObjectFlags(target) & ObjectFlags.Instantiated && source.symbol === target.symbol) {
    // We have instantiations of the same anonymous type (which typically will be the type of a
    // method). Simply do a pairwise comparison of the signatures in the two signature lists instead
    // of the much more expensive N * M comparison matrix we explore below. We erase type parameters
    // as they are known to always be the same.
    for (let i = 0; i < targetSignatures.length; i++) {
        const related = signatureRelatedTo(sourceSignatures[i], targetSignatures[i], /*erase*/ true, reportErrors);
        if (!related) {
            // A is here
            // ^^^^^^^^^
            return Ternary.False;
        }
        result &= related;
    }
}

B:

else {
    outer: for (const t of targetSignatures) {
        // Only elaborate errors from the first failure
        let shouldElaborateErrors = reportErrors;
        for (const s of sourceSignatures) {
            const related = signatureRelatedTo(s, t, /*erase*/ true, shouldElaborateErrors);
            if (related) {
                result &= related;
                errorInfo = saveErrorInfo;
                continue outer;
            }
            shouldElaborateErrors = false;
        }

        if (shouldElaborateErrors) {
            reportError(Diagnostics.Type_0_provides_no_match_for_the_signature_1,
                typeToString(source),
                signatureToString(t, /*enclosingDeclaration*/ undefined, /*flags*/ undefined, kind));
        }
        // B is here
        // ^^^^^^^^^
        return Ternary.False;
    }
}

And C is before the last return.

I can't figure out where to place calls by just looking at it, and would like put in console.log to figure it out, but found console is not recognized.

As a newbie I am always wondering how does an export reason through code, and any pointers would be appreciated. Thank you.

[xiongmao86]

I realized that I was coding the wrong way, and there is an alternative solution, which I will open a new pull request to.

[suhailsinghbains]

Hi @DanielRosenwasser
I think the PR has been closed and not merged, so does this mean that this issue is open? If it is open, would love to work on it!
Thanks

[amaksimovich2]

Hi! Can I try to do this issue?

[RyanCavanaugh]

See also #31046

[DanielRosenwasser]

Thanks @a-tarasyuk!