Remove tolerance for minimum/maximum of truncated distributions?

[vankesteren]

From @MWeiper:

I have data that starts at 0, but the truncated normal distribution shows a minimum of -1e-08. So there is a chance (however small) of generating negative values. This is strange to me.

Should we remove this tolerance?