Thanks to Gregor for the test data. I can now duplicate the problem with a regression test using (small) synthetic data.
_______________________________________________ notmuch mailing list notmuch@notmuchmail.org https://notmuchmail.org/mailman/listinfo/notmuch
