jsdoc/require-param produces syntactically invalid code

[RunDevelopment]

Expected behavior

When given a single-line doc-comment, the rule should add new lines when adding tags.

E.g.

/** Foo. */
function foo(a, b, c) {}

to

/**
 * Foo.
 * @param a
 * @param b
 * @param c
 */
function foo(a, b, c) {}

Actual behavior

* @param a
* @param b
* @param c
/** Foo. */
function foo(a, b, c) {}

Note: This happens with both ESLint and the VSCode ESLint extension.

ESLint Config

/** @type {import("eslint").Linter.Config} */
module.exports = {
	env: {
		browser: true,
		es6: true
	},
	root: true,
	plugins: [
		"jsdoc"
	],
	parserOptions: {
		ecmaVersion: 2018,
		sourceType: "module",
		ecmaFeatures: {
			node: true,
			spread: true
		}
	},
	rules: {
		"jsdoc/require-param": "error"
	},
}

ESLint sample

/** Foo. */
function foo(a, b, c) {}

Example project.

Environment

• Node version: v14.15.4
• ESLint version v7.26.0
• eslint-plugin-jsdoc version: 35.4.1

[gajus]

🎉 This issue has been resolved in version 35.4.6 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀

[brettz9]

I changed the behavior to create a normal multiline comment out of a single line comment per your example when missing tags were present.

However, I did not change handling for:

/** up here
*/

nor:

/**
 * down here */

...which will currently add the tags below and above the text, respectively.

[RunDevelopment]

...which will currently add the tags below and above the text, respectively.

So the rule still produces syntactically invalid code?

[brettz9]

...which will currently add the tags below and above the text, respectively.

So the rule still produces syntactically invalid code?

No.

It will produce:

/** up here
 * @param a
 * @param b
 * @param c
 */

/**
 * @param a
 * @param b
 * @param c
 * down here */

I don't want to get too far into policing where people want things. The latter is semantically invalid, but I'm not planning on adding time and complexity to address every possible bad case.

The single line comment is a reasonable one to support, and that has been fixed.

[RunDevelopment]

Thank you for the clarification!