mvregress fails to solve same problem as regress?

X = 1:10;  X=X';
Y = X;
% i.e. just the line y=x

regress(Y,X);  %  returns "1", as expected

mvregress(X,Y);  %  returns the following error:
%  Error using mvregress (line 452)
%  Covariance is not positive-definite.

Am I missing something here?

My definitions of X and Y for mvregress seem to correspond with the documentation...

[Roger Wohlwend]

My Matlab does not throw an error. Which release are you using?

[Matthew]

2014