R help - Aug 2013 - Problem when running an SVAR-AB model in vars package

Is "B" a reserve word in vars package?

I tried to run an SVAR-AB model by SVAR function and find the IRF in vars
package. The problem is that when B matrix is named by "B", an error
message occurs. However, if the same matrix is named by "Bm", then
things
run smoothly.  What's wrong? Is "B" a reserve word in vars
package?
> Amatexo
     [,1] [,2]
[1,]    1    0
[2,]   NA    1
> Bm
     [,1] [,2]
[1,]   NA    0
[2,]    0   NA
> B<-Bm
> B
     [,1] [,2]
[1,]   NA    0
[2,]    0   NA
> svar.Aexo<-SVAR(var_dl, estmethod ="direct", Amat=Amatexo,
Bmat=Bm,hessian=TRUE)
Warning message:
In SVAR(var_dl, estmethod = "direct", Amat = Amatexo, Bmat = Bm,  :
  The AB-model is just identified. No test possible.
> irf.svaraexo<-irf(svar.Aexo,  boot=TRUE, n.ahead=12)
There were 50 or more warnings (use warnings() to see the first
50)
> svar.Aexo<-SVAR(var_dl, estmethod ="direct", Amat=Amatexo,
Bmat=B,hessian=TRUE)
Warning message:
In SVAR(var_dl, estmethod = "direct", Amat = Amatexo, Bmat = B,  :
  The AB-model is just identified. No test possible.
> irf.svaraexo<-irf(svar.Aexo,  boot=TRUE, n.ahead=12)
Error in determinant.matrix(x, logarithm = TRUE, ...) :
  'x' must be a square matrix
In addition: Warning message:
In optim(start, logLc, ...) :
  one-dimensional optimization by Nelder-Mead is unreliable:
use "Brent" or optimize() directly

	[[alternative HTML version deleted]]