Is the really a correct way to check if a filter is an FIR filter?

I found the following function in Matlab. It checks if a filter is has FIR. It basically checks that there is only one non-zero output term at at the first index. Is it really general? It is possible to design FIR filters with more than one non-zero output term.

function isfirflag = isfirisfir(b,a)
%ISFIR(B,A) True if FIR.
if nargin<2, a=[]; end
if isempty(a), a=1; end
if ~isvector(b) || ~isvector(a)
  error(message('signal:signalpolyutils:InvalidDimensions'));
end

if find(a ~= 0, 1, 'last') > 1,
  isfirflag = 0;
else
  isfirflag = 1;
end

[Daniel M]

It seems that this is what freqz() does internally to check for FIR filters too.