Hello @Jim-215-Fisher ,

it's an impressive undertaking what you have here!

I just gave the files a quick glance and noticed certain calls to real(val) might not be what was desired. To preserve the precision of the real part of a complex number you need to use the full syntax: real(cval,kind= ...). An alternative is to use cval%re. A discussion of this pitfall can be found here.

Hello @Jim-215-Fisher ,

it's an impressive undertaking what you have here!

I just gave the files a quick glance and noticed certain calls to real(val) might not be what was desired. To preserve the precision of the real part of a complex number you need to use the full syntax: real(cval,kind= ...). An alternative is to use cval%re. A discussion of this pitfall can be found here.

Thanks for the point. I will take a look of it. Originally, I was using cval%re and cval%im, but it did not pass the compilation on ubuntu 7,8,9.

Hello @Jim-215-Fisher ,

it's an impressive undertaking what you have here!

I just gave the files a quick glance and noticed certain calls to real(val) might not be what was desired. To preserve the precision of the real part of a complex number you need to use the full syntax: real(cval,kind= ...). An alternative is to use cval%re. A discussion of this pitfall can be found here.

I checked the Fortran 2018 Handbook, it states like this for real(A, [KIND]):
"if A is type complex and KIND is not present, the kind type parameter is the kind type parameter of A."
This is the result I desired in the code. Maybe I missed something.

I checked the Fortran 2018 Handbook, it states like this for real(A, [KIND]):
"if A is type complex and KIND is not present, the kind type parameter is the kind type parameter of A."
This is the result I desired in the code. Maybe I missed something.

In that case consider my comment void. I was basing my comment of the comments I read on Fortran Discourse.

@jvdp1 @awvwgk @gareth-nx , the files have been updated. Especially documentation for complex variable case was changed.

@jvdp1 @awvwgk @gareth-nx All precision for probability were added according to @gareth-nx . Once this PR is done, I will set up a new PR for uniform distribution to correct similar problems.

@Jim-215-Fisher @gareth-nx is this PR ready for merging?

@jvdp1 @Jim-215-Fisher I think it just needs some minor documentation fixes mentioned above (or, a comment about why they are not needed -- I haven't looked in detail since reviewing). I was intending to approve once those points were addressed.

@Jim-215-Fisher @gareth-nx is this PR ready for merging?

I think it is ready.

@jvdp1 @Jim-215-Fisher I think it just needs some minor documentation fixes mentioned above (or, a comment about why they are not needed -- I haven't looked in detail since reviewing). I was intending to approve once those points were addressed.

I don't think we mention precision in documentation though. By default, all precision are supported.

@jvdp1 @Jim-215-Fisher I think it just needs some minor documentation fixes mentioned above (or, a comment about why they are not needed -- I haven't looked in detail since reviewing). I was intending to approve once those points were addressed.

I don't think we mention precision in documentation though. By default, all precision are supported.

Ahh, I see. I am getting confused between type and kind in Fortran.

So at the moment the documentation indicates the output type, but not the kind, and it is assumed to be implied that all kinds are supported.

I think it would be nicer to be more explicit about this. The same point was made in this review comment by @jvdp1 on October 10. @Jim-215-Fisher would you mind to use phrasing like this when describing the output variables?

@jvdp1 @Jim-215-Fisher I think it just needs some minor documentation fixes mentioned above (or, a comment about why they are not needed -- I haven't looked in detail since reviewing). I was intending to approve once those points were addressed.

I don't think we mention precision in documentation though. By default, all precision are supported.

Ahh, I see. I am getting confused between type and kind in Fortran.

So at the moment the documentation indicates the output type, but not the kind, and it is assumed to be implied that all kinds are supported.

I think it would be nicer to be more explicit about this. The same point was made in this review comment by @jvdp1 on October 10. @Jim-215-Fisher would you mind to use phrasing like this when describing the output variables?

done.

@jvdp1 This is ready to merge from my view.