The standard file with packaging instructions (setup.py with setuptools) for Python contains an author_email field. Such a package can then be published to PyPI, but the code is also available publicly on github.
Am I unnecessarily cautious if I want to obfuscate the email address in the setup.py file (e.g. by calling a base64.decodebytes())?
It seems superfluous to have the address as "john.doe[at]gmail.com" (with brackets) in a README file, but as plain text legit email in a python file, but I have seen no one obfuscating their setup.py address.
Nowadays, spam filters perform well, but is there some interest in doing what I suggest? Or does it just make my code look like virus?
[at]works is because they're usually in a long free text field, and the email is lost among other text, so it's really hard to write a parser that can just catch email address without getting lots of false positives. In this case, since this field is an email address field, it's much simpler to write a transformation function that can undo most of the transforms people would do while still making the email readable for humans. Especially if you're going to use common mail domains like gmail.com.author_emailfield and undoing all my obfuscation?email_address.replace("[at]", "@"), it wouldn't really be that difficult to write a scraper than can undo all the common obfuscations if you already know that the field must be an email address. It's a much easier problem than finding an obfuscated email address in a free text field.