hgfred([a,b],[0],x) and hgfred([a,0],[0],x)

While playing around with hypergeometrics again, I noticed that
hgfred([a,b],[0],x) returns the noun form.  Wolfram Alpha says it's complex
infinity.  This seems right.  Maybe we should return infinity too?

For fun, I wanted to see what maxima gives for hgfred([a,0],[0],x).  Maxima
says 1/(1-x)^a.  When I ask Wolfram Alpha for 2f1(a,0,;0,x), it says it's
identically 1.  But it also gives a series expansion for 2f1(a,0;0;x) which
is obviously not 1.  In fact, it looks like 1/(1-x)^a.

The moral?  I guess we should always check our work and check our
references too. :-)

Ray