Open joshuawilson opened 7 years ago
so I'm thinking we either:
(i) refactor the login stuff in fabric8-runtime-console into a service that we can opt out of in fabric8-ui and just use ngx-login-client directly or
(ii) refactor the core parts of fabric8-runtime-console out as an npm module (or separate repo) and let the login stuff sit in fabric8-ui
(iii) have 2 separate components that extend fabric8-runtime-console; one that uses vanilla OAuth in OpenShift; the other that uses ngx-login-client
As we need fabric8-runtime-console that runs on vanilla Kubernetes/OpenShift without using ngx-login-client so that we can just talk OAuth to OpenShift/Keycloak.
I'm guessing ngx-login-client can't handle the case where fabric8-runtime-console is running as a client directly on OpenShift with no planner/core API backend right?
Or we can update the login-client to handle generic oauth
yeah - current plan is I think to
lets separate out the "make ngx-login-client support vanilla OAuth" issue https://github.com/fabric8-ui/ngx-login-client/issues/20
then this issue becomes mostly 'migrate to ngx-login-client"
Setup the library to use the ngx-login-client and the auth that Platform is using.