We have a commit that adds a configuration flag so that anything that can't be interpreted as UTF-8 will be a binary string instead. I'll submit it as a PR. Aren
while using the python3 module with 3.2.1 I got the following errors:
mod_populate_vptuple:402, vp->da->name: User-Password python_error_log:209, Exception type: <class 'UnicodeDecodeError'>, Exception value: 'utf-8' codec can't decode byte 0xb3 in position 0: invalid start byte do_python_single:496, authorize - mod_populate_vps failed do_python_single:675, authorize - RLM_MODULE_FAIL
The cause is a switch whichs software mangles the User-Password to binary garbage.
Is this handling of binary values in the python module favorable? To my feeling it would be cleaner to convert to bytes (instead of a string). Then python can handle binary data well.
Or make the conversion using replacement character/ignoring encoding errors.
The switch will be fixed, so this is only a design question.
Thanks for all your great work, greetings Hermann