I thought #114 would be a quick fix, however, I was wrong. This is the start which adds a regression test to demonstrate the issue.
From my observation it would seem that with python3+ the actual request body will always be of type bytes. The cassette will always store utf-8, so the request/response body that is read in from the cassette should be converted to bytes. This is different than the code (some of which I previously added) does.
Unfortunately, I don't have more time tonight to look into it.
I thought #114 would be a quick fix, however, I was wrong. This is the start which adds a regression test to demonstrate the issue.
From my observation it would seem that with python3+ the actual request body will always be of type bytes. The cassette will always store utf-8, so the request/response body that is read in from the cassette should be converted to bytes. This is different than the code (some of which I previously added) does.
Unfortunately, I don't have more time tonight to look into it.