IPT tutorial: issue with multiplying G_tau

[tayral]

Hi,
On branch "unstable", the ipynb 05-IPT... returns an error when calling the solve method of the IPT solver. It is caused by the line Sigma_t << U/2 * Gt * Gt * Gt:

TypeError                                 Traceback (most recent call last)
<ipython-input-3-9ca55848cf11> in <module>()
     17 for i in range(n_loops):
     18     S.g0 << inverse( iOmega_n - t**2 * S.g )
---> 19     S.solve(U = U)
     20 
     21     # Get real axis function with Pade approximation

<ipython-input-2-f1c49a130509> in solve(self, U)
     16     def solve(self, U):
     17         self.g0t << InverseFourier(self.g0)
---> 18         self.sigmat << (U**2) * self.g0t * self.g0t * self.g0t
     19         self.sigma << Fourier(self.sigmat)
     20 

/home/tayral/Work/TRIQS_1.6/install_gcc7/lib/python2.7/dist-packages/pytriqs/gf/gf.pyc in __mul__(self, y)
    447     def __mul__(self,y):
    448         c = self.copy()
--> 449         c *= y
    450         return c
    451 

/home/tayral/Work/TRIQS_1.6/install_gcc7/lib/python2.7/dist-packages/pytriqs/gf/gf.pyc in __imul__(self, arg)
    441         if descriptor_base.is_lazy(arg): return lazy_expressions.make_lazy(self) * arg
    442         # Multiply by a GF !?
--> 443         self._data[:] *= arg
    444         if self._singularity : self._singularity *= arg
    445         return self

TypeError: ufunc 'multiply' output (typecode 'O') could not be coerced to provided output parameter (typecode 'D') according to the casting rule ''same_kind''

I posted a boiled-down version of the problem on the TRIQS issue list (issue 404)

Thomas