Since our Upgrade to 3.5, the format of the registration tokens became more complex. It now includes a lot of special characters
e.g.:
Before: p3UHI9dCzOSrq9i1OQ1Jo4UV
Now: h<)vI1=:M*}*\xz2AKHb8t;+
I have a feeling that this is breaking the login for some users behind corporate proxies, maybe sometimes due to encoding issues?
I have tried setting registration.contents="-s" in the /token/init request, but it doesn’t seem to make difference.
Anywhere else I can set the format of these tokens?