Is there a general consensus amongst top chess players, composers, analyzists, and organizations, such as FIDE, for what exactly defines an en passant checkmate?
Is there any history for it?
As far as I know, there are two possibilities for what defines an en passant checkmate:
When the pawn does the passant, moving as to allow another piece's attack range deliver the checkmate.
When the pawn does the passant, and the pawn itself makes the checkmate.
After some research I could not find any "official" definition by a chess association/federation. This doesn't really surprise me, as there really is no need for a definition of "e.p. checkmate" to ensure correct play.
However the general internet consensus on the usage of the term seems to agree on the following requirements:
An exemplary notated move would be "29... cxd e.p. #" (although notating e.p. is optional under the FIDE laws of chess (C.9)) - regardless if the pawn capturing delivers check or not. The Wikipedia examples follow this definition.
As there is no opinion by a regulatory instance about this topic, it seems to be up to you to make up your mind on this topic or bringing it to the attention of said regulatory instance. I am taking the "democratic" approach with the above definition.
I don't think it is defined at all. But in your first case, the pawn is delivering the mate. So I think that would be the 'en passant mate'. The 2nd case is 'just' a discovered mate.
