Unlinking algorithm always removes anchor element

We have a behavior difference in Firefox and Chrome where Firefox replaces an anchor with <span> when it has extra attributes while Chrome does not.

Repro: https://bug378931.bmoattachments.org/attachment.cgi?id=263746

Per the unlink command algorithm, it clears the value of the target anchor elements, where it removes href attribute in step 9. And then its step 10 has this note:

If we get past this step, we're something like <b class=foo> where we want to keep the extra attributes, so we stick them on a span.

Per the note the purpose of it is to keep the extra attributes of the element. But then the step says:

If element's specified command value for command is null, return the empty list.

Here, the "specified command value" becomes always null for "unlink" because the previous step just removed href:

If command is "createLink" or "unlink":
If element is an a element and has an href attribute, return the value of that attribute.
Return null.

Since the value is always null, the anchor elements always must disappear regardless of extra attributes.

The spec agrees with Chrome behavior but the note clearly suggests the original intention was for Firefox behavior.

What should be done here? Should Firefox be fixed or the spec?

Start with the linked execCommand unlink-command and clear-the-value sections, then reproduce attachment 263746 in Firefox and Chrome. Compare the observed behavior with steps 9 and 10 and the specified command value definition. Done means resolving whether the browser behavior or the specification should change.