Pedro,
Thanks for catching this, I missed it when writing the quantile
module. As Rhys suggested, the fix is to make a new function
gsl_rstat_quantile_reset. I've made the changes on the git repository.
Thanks,
Patrick
On 07/25/2016 04:05 PM, Pedro Donato wrote:
Hi,
Looks like currently "gsl_rstat_reset" does not reset "median_workspace_p"
member of "gsl_rstat_workspace". This results in wrong median computations
after one reset. Minimum program attached shows the problem.
Actually "gsl_rstat_workspace" does not include an "reset" function which
would be useful.
I am willing to write a simple fix for that.
Thank you,
Pedro Donato
