On 1/28/22 00:41, IOhannes m zmölnig wrote:
i'm afraid i still do not fully understand which types of "users" there are.
afaics, there are three different types:
- django users this is what get's listed under <https://lists.mailman3.org/admin/auth/user/>
Correct.
- mailman accounts this is what is created when you visit <https://lists.mailman3.org/accounts/signup/>
No. That creates a Django user, not a Mailman user. What you think of as a Mailman account is actually a Django account.
- email addresses this is what is created when i subscribe to a list *without* having an account. e.g. via the "Subscribe" button on <https://lists.mailman3.org/mailman3/lists/mailman-users.mailman3.org/>
Mailman core has users and addresses. A user has one or more associated
addresses. When you subscribe to a list without having an account in
that way, or via email to list-join or list-subscribe or by being mass
subscribed by the list admin or by being imported by mailman import21,
assuming you are previously unknown, this creates a Mailman user with an
associated address and subscribes the user or address.
Each user has a primary address and zero or more additional linked addresses. The user can be subscribed to a list as a user in which case, delivery if any is to the user's primary address, or can be subscribed as an address.
afaiu, (multiple) "email addresses" can be associated with a single "mailman account" (but this is strictly optional).
Yes, see above.
if an "email address" is first used to subscribe to a list without a mailman account, and later a mailman account is created with that very email address, they get automatically associated. a single "email address" cannot be associated with more than one "mailman accounts".
Based on the above, your "mailman accounts" are actually Django accounts. With that in mind, the above is correct, but there is no real link between Mailman users and Django users. When you are logged in to the web UI you are logged in as a Django user, and Postorius and HyperKitty know that user's email address and can therefore know if it is a list member and know what address to subscribe or unsubscribe as requested.
i think i got this one right (as it seems pretty straightforward).
what i don't fully grasp yet is the relation between "django users" and "mailman accounts".
As I have said, what you are calling a mailman account is a Django user and not a Mailman user.
intuitively, i would have thought that they are identical.
They are, but they are not Mailman core users.
now mark wrote:
a Django user, which is distinct from a Mailman user. The users you added via the web API are Mailman users, but in order to access their lists via the web UI they have to
sign upto create a Django account.that seems to indicate, that a "django user" is a (strict) superset of a "mailman account".
No. A Django user is the same thing as what you call a mailman account. A Mailman core user is a separate thing entirely.
you (as admin) can create "mailman accounts" via the web API¹, but they are only expanded to "mailman accounts" if you (as user) go through the "Sign Up" process.
No. An admin can create Mailman users via the Postorius UI. A Django superuser can also create Django users through the Django admin UI.
otoh, if you create an initial "django user" by signing up, you automatically create a "mailman account".
Because they are the same thing. When you go to the Sign-up link on a Postorius or Hyperkitty page you go to https://example.com/accounts/signup/, and that's Django, not Mailman.
so are the two distinct concepts that are just associated via some common property (e.g. the email address), or is there some intrinsic relationship between the two (e.g. parent-child inheritance relationship)?
They are distinct. Django users have email addresses and Mailman users have email addresses and they can be associated in that way. There is no other link.
so my recent experiment (comparing the output of <https://lists.example.com/admin/auth/user/> with that of
[u for u in getUtility(IUserManager).users]) indicates that they are indeed distinct objects that *can* be associated with each other.could you please clarify this?
I hope I have.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan