On the signup page there is the following code block:
{# we have to handle this error message here to have access to the request #}
{% if error == form.duplicate_email_error_label %}
{% trans login_url=url('account_login'), next_url=github_connect_url, provider_name=provider.name %}
A profile already exists with the selected email address.
<a href="{{ login_url }}" class="launch-login" data-next="{{ next_url }}">Connect your
{{ provider_name }} account to it.</a>
{% endtrans %}
{% else %}
From some testing(https://github.com/mdn/kuma/issues/5909#issuecomment-543634445) and speaking to @escattone it seems that the duplicate email constraint was never, or is no longer enforced.
@atopal Suggested that we look into the viability of(the amount of effort that will be required):
Sorry, there was a misunderstanding, my suggestion is that if we receive a validated email address that matches an already existing persona one we simply recover that account as part of the process, no emails being sent.
@atopal Ryan and I checked this out and found 160k Persona users.
Question 1: What do you prefer we do once we discover that someone attempting to log in was once a Persona user?
Question 2: what to do with a Persona user who has been banned? (stored in the MDN user acct)
Q1. Let鈥檚 go with option 1. They haven鈥檛 seen that data in at least 3 years.
Q2. Let鈥檚 ignore their banned status. This will most likely be extremely rare
Q2. Let鈥檚 ignore their banned status. This will most likely be extremely rare
Also, banning is usually about spam. If they're creating a membership account, they won't have edit access, so they won't be able to create spam. (However, we might want to prevent them from enabling edit access. But probably so rare a case as to be not worth spending any time on.)
@atopal @schalkneethling After looking at this in much more detail today (the code is not easy to grasp), I realized that this is more complicated than I originally thought (for example, deleting the user associated with an existing persona account may not be straightforward since that user may have made one or more document revisions). Now that I understand this user/social-account world much better, I think it would be really helpful if we could set-up some time to discuss this in a zoom meeting. Would that work for both of you?
Happy to meet, Ryan.
@escattone what else needs to be completed on this one?
And, is this estimate still accurate?
@escattone we should close this ticket and create a new one with acceptance criteria for the actual work. Can you do that?
@escattone we should close this ticket and create a new one with acceptance criteria for the actual work. Can you do that?
Yes, I think we should close this one, and I'll create a new one, thanks @tobinmori.
Closing this issue in favor of https://github.com/mdn/kuma/issues/6182
Most helpful comment
Q1. Let鈥檚 go with option 1. They haven鈥檛 seen that data in at least 3 years.
Q2. Let鈥檚 ignore their banned status. This will most likely be extremely rare