When using route-maps + prefix-list combination for bgp peer, if you have both ipv4 and ipv6 prefix lists in the same route-map rule, neither is matched thus no advertisement occurs or perhaps an outage occurs.

set policy route-map EXPORT-xxxxxx rule 11 match ip prefix-list EXPORT-blahh
set policy route-map EXPORT-xxxxxx rule 11 match ipv6 prefix-list EXPORT-blahh

which converts to (in FRR):

!
route-map EXPORT-xxxxxx permit 11
 match ip address prefix-list EXPORT-xxxxxx
 match ipv6 address prefix-list EXPORT-xxxxxx
exit

We either need some validation logic to prevent multiple address-classes in the route-map rule definitions from being committed to FRR or for FRR to not break :)