As an experiment I added, to xr, a check for repetitions of expressions that can match the empty string, like

\(a*b*\)*

on the grounds that these are either mistakes or potentially risk being very slow on the backtracking regexp engine in Emacs. (A DFA would have no problem whatsoever; the regexp above would become a single state.)

Is this useful, or a pointless waste of time? Attached are the results from a relint run on the Emacs tree so that you can judge for yourselves. If the consensus is that it too often complains about perfectly fine regexps, then the check will be dropped unceremoniously.