Preserve set-only accessors in declaration emit

[DanielRosenwasser]

Playground Link

Related to #58112 (comment)

export let same = {
    set x(value: string) {
    },
    get x(): string {
        throw new Error("Not implemented.")
    },
};

export let divergent = {
    set x(value: string) {
    },
    get x(): number {
        throw new Error("Not implemented.")
    },
};

export let readonly = {
    get x(): string {
        throw new Error("Not implemented.")
    },
};

export let writeonly = {
    set x(value: string) {
    },
};

Current

Notice that in the declaration of writeonly, we don't preserve the accessor at all.

export declare let same: {
    x: string;
};
export declare let divergent: {
    get x(): number;
    set x(value: string);
};
export declare let readonly: {
    readonly x: string;
};
export declare let writeonly: {
    x: string;
};

Expected

If we had a writeonly modifier, I'd argue we could just use that as a parallel to the readonly declaration; however, we do support accessors in declaration emit, and we could emit that in the case of a write-only accessor.

It is technically a breaking change though for consumers of older TypeScript versions, and the only way to opt out would be to declare a no-op get accessor.

export declare let same: {
    x: string;
};
export declare let divergent: {
    get x(): number;
    set x(value: string);
};
export declare let readonly: {
    readonly x: string;
};
export declare let writeonly: {
    set x(): string;
};