New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Identity Providers, callback locking and the difference between being authenticated and being authorised #10
Comments
This issue extends so far, that I can log in with any wrong provider in my farm, supposedly unlocking a wiki. Only when trying to provide changes the yellow halo appears. |
When you login, with this security plugin, you are performing a login to the As a wiki farm host, you have to choose which authentication provider(s) you want to use. It makes sense to only configure a single provider, to avoid confusing wiki users. As a wiki user, when using a farm that provides more than one, you should pick one of the offered providers and stick to using that provider, while accessing that wiki farm. The 'lock' only indicated if you are authenticated, or not. So, the wiki farm knows who you are. It is not an indication of if you the owner of a particular wiki. This is much the same behaviour we have had with all the previous security implementations.
No. As part of the Google configuration you need to specify all the allowed callback URLs. If you don't use If you want to only login to a single wiki at a time, you must only use the Twitter integration as it does not require the oAuth callback url to be configured. |
This is still very confusing and hard to grasp. In continuation of #87 with a slightly different bias, I am receiving errors when logging in to subdomains within the farms denoted as
Exemplary debug output looks like:
How can I debug this further, i.e. why the PUT request is rejected?
Do you suggest the errors comes from I am not sure about how many scenarios we are looking here and would like to gain clarity about the dimensions at least. Am I missing something other than farmability (discrete; yes, no), number of strategies (discrete; N), callback restriction (discrete; yes, no)?
While it sounds rational to continue with the tradition, as a wiki user I would like to know before a page action (fork, move, ...) if it is going to be applied locally or remotely. This could potentially help to minimise confusion about the different halos. |
Oddly enough, some PUT actions went through, but I cannot reproduce the conditions. The last paragraphs in the following were added today, despite I can authenticate, but not authorise myself against the wiki: |
Almost certainly, if that is how you have claimed them. I expect that you have probably done this to create some separation of duty, but the reality is that just creating a trap for yourself and potentially others. |
As far as I know the PassportJS adapters can only provide access tokens for whole domains, without leaving subdomains to their respective owners, and as such locked.
https://github.com/fedwiki/wiki-security-passportjs/blob/master/docs/config-twitter.md#configure-wiki explains for example, how the Callback can be left open for the client to be changed.
How can I make sure that logging in to http://patterns.wiki.transformap.co/ does not unlock http://jon.patterns.wiki.transformap.co/ or the whole domain http://wiki.transformap.co/ ?
Would using the Google Adapter help out here?
Or is this closely tied to the
wikiDomains
configuration variable?Unfortunately I will not be able to change it every time a user creates a new subdomain.
I have the slight impression this may be just a display error:
Logging me into http://jon.fed.wiki/ provides an open lock on http://patternsofcommoning.fed.wiki/, while trying to edit it sends the changes to localstorage. Signing out here would also sign me out at the source.
In a private tab logging the respective user into http://patterns.wiki.transformap.co/ also unlocks http://jon.patterns.wiki.transformap.co/, while forking of a newer version of the workshop page also creates the yellow halo.
Thus my question ends up to be more about how we can turn the lock icon to be less ambiguous.
The text was updated successfully, but these errors were encountered: