Function: lfunmf
Section: modular_forms
C-Name: lfunmf
Prototype: GD0,L,b
Help: lfunmf(mf,{flag=0}): If mf is an eigenform which is already
 embedded in C, outputs the corresponding L-function. If
 mf is a space, output the L-functions corresponding to all
 eigenforms, and in that case mf must be a split new space. The binary
 digits of flag mean 1: assume the form is real (resp. all eigenforms are
 real); 2: assume the form is cuspidal.
Doc: If \kbd{mf} is an eigenform which is already embedded in $\C$, outputs
 the corresponding L-function, ready for use with \kbd{lfunxxx} programs.
 The binary digits of \fl mean

 \item $1$: assume the form is real;

 \item $2$: assume the form is cuspidal.

 If \kbd{mf} is a space, it must be a split new space; output the
 vector of \kbd{lfuncreate}s attached to all eigenforms. The result
 is a vector whose length is the dimension of the space: each Galois orbit of
 dimension $d$ has $d$ corresponding \kbd{lfuncreate}s, one for each
 embedding.

 \bprog
 ? mf = mfsplit([35,2]); F = mfeigenbasis(mf)[2]; mffields(mf)
 %1 = [z, z^2 - z - 4]
 ? f = mfembed(F)[2];
 ? L = lfunmf(f);
 ? lfun(L,1)
 %4 = 0.46007635204895314548435893464149369804
 ? [ lfun(L,1) | L <- lfunmf(mf) ]
 %5 = [0.70291..., 0.81018..., 0.46007...]
 @eprog
