new: py-lmfit

[jjstickel (Jonathan Stickel)]

New port for the python "lmfit" package. Portfile attached. I created a variant for the optional "unertainties" package, which can be found in ticket #39626.

[g5pw (Aljaž Srebrnič)]

Thanks for the submission!

Why is your portfile declaring a variant and then just adding a port dependency? I'ts better practice to force the setup script to actually use the port you list as dependency in a variant. Moreover, it looks like the uncertainties package is used just at runtime, so I think users can install the necessary package from MacPorts after installing lmfit. So, declaring a variant is pointless in my opinion, you can add a note about uncertainties if you want, but I wouldn't add a variant since the installed files are the same.

Also, your port is missing a platform keyword. Use the port lint command to check the port file in the future.

[g5pw (Aljaž Srebrnič)]

Cc Me!

[smithsp (Sterling Smith)]

The string comparison constructs should be ne instead of !=.

Regarding the uncertainties variant, if the uncertainties package is installed, does lmfit install things any differently?

[smithsp (Sterling Smith)]

Cc Me!

[jjstickel (Jonathan Stickel)]

Replying to smithsp@…:

Regarding the uncertainties variant, if the uncertainties package is installed, does lmfit install things any differently?

I don't think so. If uncertainties is installed, it can be used for additional calculations. I will remove the variant and add a note, per g5pw's suggestion.

[jjstickel (Jonathan Stickel)]

Please review the updated Portfile (ignore py-enable patch - mistaken upload that I can't remove).

[g5pw (Aljaž Srebrnič)]

Ok, I commited your patch in r116239, with the exception of adding a subport for python33, too. I tested the package with python33 and all the examples I could find run fine.