Not a major regression, but slightly inconvenient that the original change affected users without subrepos. Should the first patch have referenced the issue again in the summary (that was marked fixed and closed)? I didn't want to confuse the bot, though it might be useful for archeology.