I am not sure if this test (and tests like this) in rendered_markdown.test.js needs to be changed since all the node-tests are passing.

run_test("code playground none", ({override, mock_template}) => {
    override(realm_playground, "get_playground_info_for_languages", (language) => {
        assert.equal(language, "javascript");
        return undefined;
    });

    override(copied_tooltip, "show_copied_confirmation", noop);

    const {prepends, $button_container, $view_code} = test_code_playground(mock_template, false);
    assert.deepEqual(prepends, [$button_container]);
    assert_clipboard_setup();

    assert.equal($view_code.attr("data-tippy-content"), undefined);
    assert.equal($view_code.attr("aria-label"), undefined);
});

@amanagr are you up for reviewing this one? I haven't tested.

LGTM

Merged, thanks @nimishmedatwal!

I guess @karlstolley FYI; this is a case where brittle JS made refactoring the HTML structure break stuff. I don't think there's an easy way to catch this sort of thing except experimentally, so I wouldn't spend undue additional time on refactoring. I don't think this type of bad pattern is common.

I guess we didn't fix the brittleness in this PR. @nimishmedatwal would you put up a follow-up that instead uses .closest() and .find() to find the right element without making assumptions about how many containing elements there are?

I guess we didn't fix the brittleness in this PR. @nimishmedatwal would you put up a follow-up that instead uses .closest() and .find() to find the right element without making assumptions about how many containing elements there are?

Sure! I'll do that.