Change algorithm interface to allow returning convergence info
This changes the way deconvolution algorithms are called. Now, the function returns a structure that will in the future also hold convergence information. This also eliminates the need for a separate reached_major_threshold parameter.
Edited by Andre Offringa