chore: update deno_lint binary used in CI to v0.5.0

[magurotuna]

This PR updates third_paty submodule so we can use the latest deno_lint in CI.
The latest deno_lint brings no-unused-vars and no-deprecated-deno-api, which spot some of the existing code as errors. I dealt with them by fixing the code or adding ignore directives.

Additionally, I removed ESLint directives because we don't use ESLint now and deno_lint hasn't supported them since #8226.

[magurotuna]

Hmm, looks like this relative import doesn't work in CI... 🤔 I have no idea how to resolve it.

deno/tools/wpt/testharnessreport.js

Line 3 in 6d3ce37

import { writeAllSync } from "../../test_util/std/io/util.ts";

[magurotuna]

Shouldn't we make sure that all data have been written to stderr like the following?

Suggested change

	const bytesWritten = Deno.stderr.writeSync(data);
	if (bytesWritten !== data.byteLength) {
	throw new TypeError("failed to report test result");
	}
	let bytesWritten = 0;
	while (bytesWritten < data.byteLength) {
	bytesWritten += Deno.stderr.writeSync(data.subarray(byteWritten));
	}

[lucacasonato]

that works too :-)

[magurotuna]

so either is okay? then I'm fine with the current implementation of yours.
I feel a little bit skeptical though, because Deno.writeSync calls op_write_sync and then std::io::Write::write is used under the hood, which requires us to do a loop until all data are written if we wish to imitate something like write_all.
https://doc.rust-lang.org/std/io/trait.Write.html#examples

[lucacasonato]

I'm aware, I just think the messages are small enough to never fail. But if you want to add the loop, that is fine by me. It is a little more robust.

[magurotuna]

Makes sense. Indeed the tests run in CI succeeded, but I'd prefer a robust one so I'll add the loop :)

[lucacasonato]

LGTM! Thanks :-)