In Corollary 4.2, there should be an assumption that L is not Galois over Q. Reported by Rakvi.

In the proof of Lemma 8.14 (in the published version) there is a gap in the C_2xC_20 case. We claim that the Q-rational 10-torsion point Q has to become divisible by 2 in the quartic field K, which is not true. Hence the case when Q is not divisible by 2 in K needs to be eliminated. Reported by Andreas Schweizer. The proof that this is not possible can be found in https://web.math.pmf.unizg.hr/~fnajman/8_14_fix.txt which is a Magma computation together with explanations.