Andres Freund <andres(at)anarazel(dot)de> writes:
> Is there a special reason for not using the normal function calling
> mechanisms? It looks to me as it was just done to have an easy way to store it
> in pg_trigger.tgargs.
Well, this is all very historical, dating from Berkeley days AFAIK.
If we had it to do over, I bet we'd do it differently --- but the pain
of changing it seems to exceed any likely benefit.
regards, tom lane