Changes the linear convolution domain used in pulsar likelihoods to be dynamically computed, rather than hard coded to support the pulsars in 47 Tuc and Ter 5. The new domain reaches up to either 5σ of the Gaussian error term or 250% larger than the peak of the cluster component, whichever is larger. This should be enough to probe all structure in the convolved distributions for both spin and orbital period likelihoods.
Changes the linear convolution domain used in pulsar likelihoods to be dynamically computed, rather than hard coded to support the pulsars in 47 Tuc and Ter 5. The new domain reaches up to either 5σ of the Gaussian error term or 250% larger than the peak of the cluster component, whichever is larger. This should be enough to probe all structure in the convolved distributions for both spin and orbital period likelihoods.