Bug: no-invalid-label-refs reports escaped label references

[KumJungMin]

Environment

ESLint version: 10.8.0
@eslint/markdown version: 8.0.3
Node version: 22.18.0
npm version: 10.9.3

Which language are you using?

commonmark

What did you do?

Configuration

import { defineConfig } from "eslint/config";
import markdown from "@eslint/markdown";

export default defineConfig([
    {
        files: ["**/*.md"],
        plugins: {
            markdown,
        },
        language: "markdown/commonmark",
        rules: {
            "markdown/no-invalid-label-refs": "error",
        },
    },
]);

<!-- eslint markdown/no-invalid-label-refs: "error" -->

<!-- OK -->
\[foo][ ]
\\\[foo][ ]

<!-- NOT OK -->
[foo][ ]
\\[foo][ ]
\\\\[foo][ ]

What did you expect to happen?

he rule should not report the cases with an odd number of backslashes before the opening bracket:

\[foo][ ]
\\\[foo][ ]

In these cases, the opening bracket is escaped, so the text does not form a label reference.

The rule should report only the cases where the opening bracket is not escaped:

[foo][ ]
\\[foo][ ]
\\\\[foo][ ]

An even number of backslashes does not escape the opening bracket,
so these cases still form invalid label references.

What actually happened?

no-invalid-label-refs reports the cases where the opening bracket is escaped:

   5:7   error  Label reference 'foo' is invalid due to white space between [ and ]  markdown/no-invalid-la...

The rule appears to detect the ][ ] portion and search backward for an opening bracket without checking whether that bracket is escaped. As a result, escaped text is incorrectly treated as an invalid label reference.

Link to Minimal Reproducible Example

https://stackblitz.com/edit/vitejs-vite-ddfg51us?file=eslint.config.js,test.md

Participation

• I am willing to submit a pull request for this issue.

AI acknowledgment

• I did not use AI to generate this issue report.
• (If the above is not checked) I have reviewed the AI-generated content before submitting.

Additional comments

Disclosure: I'm a participant of open source contribution program OSSCA

A related case was addressed in #490, which updated several rules to distinguish escaped syntax based on whether it is preceded by an odd or even number of consecutive backslashes.

That change covered no-missing-label-refs, no-reversed-media-syntax, and no-missing-atx-heading-space. However, no-invalid-label-refs appears to perform a separate label-reference check that does not account for whether the opening bracket is escaped.

[lumirlumir]

Disclosure: I'm a participant of open source contribution program OSSCA: confirmed.

I can reproduce it and confirm that it’s a bug.

Actually, from a historical perspective, I noticed this bug while working on #490 and several other similar fixes. However, this one involved some additional edge cases related to handling consecutive backslashes, which made it quite tricky at the time. I didn’t have enough time to investigate it further, so I left it as is.

Another edge case to consider for this bug would be:

[foo\][]

[foo]: hi

Looking at the Code Explorer, cases with an odd number of backslashes should not be reported. Only cases with an even number of backslashes should be reported.

• Recognized as LinkReference and Definition:

[foo][]

[foo]: hi

• Recognized as Text and Definition:

\[foo][]

[foo]: hi

• Recognized as LinkReference and Definition

\\[foo][]

[foo]: hi

• Recognized as Text and Definition

\\\[foo][]

[foo]: hi

---

Using the same regular expression pattern as the one below might be helpful in this case:

markdown/src/rules/no-reversed-media-syntax.js

Lines 23 to 25 in ac31775

	/** Matches reversed link/image syntax like `(text)[url]`, ignoring escaped characters like `\(text\)[url]`. */
	const reversedPattern =
	/(?<=(?<!\\)(?:\\{2})*)\((?<label>(?:\\.|[^()\\]|\([\s\S]*\))*)\)\[(?<url>(?:\\.|[^\]\\\r\n])*)\](?!\()/gu;

markdown/src/rules/no-missing-label-refs.js

Lines 46 to 47 in ac31775

	const labelPattern =
	/(?<=(?<!\\)(?:\\{2})*)\[(?<left>(?:\\.|[^[\]\\])*)\](?:\[(?<right>(?:\\.|[^\]\\])*)\])?/dgu;