The upper and lower limit of integration influence the results of covolution wrongly.
When the limit is [-5, 5], the results looks fine.
With other limit, the results is wrong.
Here is the demo
With other limit, the results is wrong.
Here is the demo
Is this a question or a bug report?