More problems with general power of a matrix
Even though in version 8.2 the code for the general power of a matrix has been improved (c.f. question 41622), it still doesn't work in some cases, as i.e. this singular, diagonalizable matrix
A=matrix(QQbar,3,3,[[-2,-8,-12],[1,4,4],[0,0,1]])
k=var('k')
A**k
shows.
Concerning the remark in trac ticket 25520: Why not defining $0^x=1$ for $x\in {\bf N}$, which seems reasonable, since the number of functions $\emptyset \to \emptyset$ is 1?