The generate method modified in this change seems to be what is generating API (and session?) tokens. If this is the case, then I believe that using the v4() method instead of v1() would be more desirable as it would provide higher entropy and so be harder to reproduce. v4() is completely random while v1() is time-based.
Coverage remained the same at 97.697% when pulling 987a1e1236ec72d0466976f399d125ba5488a1a9 on ianwalter:patch-1 into c6468f068cdb0522f15f51ef3d05bf51e4a6c174 on adonisjs:develop.
The generate method modified in this change seems to be what is generating API (and session?) tokens. If this is the case, then I believe that using the v4() method instead of v1() would be more desirable as it would provide higher entropy and so be harder to reproduce. v4() is completely random while v1() is time-based.