Beginnings of user accounts, sqlalchemy etc. #85
Conversation
|
Travis being green is a lie, and that has been fixed in the other branches. When we're working on the same story that will help... |
deploystream/templates/index.html
Outdated
There was a problem hiding this comment.
Isn't this what base-angular is bringing in?
There was a problem hiding this comment.
My mistake. Didn't take it out after the refactor.
I meant to ask you, is this bit of styling needed?
There was a problem hiding this comment.
Maybe not. AFAIR it is supposed to prevent some visualisation glitches when rendering angular views.
|
I made some changes re: styling in a PR #86 to this branch |
deploystream/apps/oauth/views.py
Outdated
There was a problem hiding this comment.
I am a bit unclear about how we really want to handle user registration and linking external accounts / oauth tokens. In particular how much user management we want in the core "open-source" project (in particular, self-registration).
E.g. now, any non-registered user can log in via github Oauth, and have a user account created with no email. Whereas for self-registered users, email is a required field.
Up for a discussion elsewhere.
There was a problem hiding this comment.
Thinking about it some more I'd be happy with username+password for non oauth accounts as opposed to email+password - although I'm not sure that's your concern.
Happy to have a chat about this.
|
I forgot to mention in the PR itself (it's in the commit log) that a lot of the user stuff was taken from https://github.com/mitsuhiko/flask/wiki/Large-app-how-to |
…opover Alternative login form as a popover
|
There's one failing test in this that will need some more thought. |
|
Oops, all the fontawesome files are not meant to be under version control, they are bower dependencies... |
There was a problem hiding this comment.
This is some architectural departure from the API backend / Angular front-end approach.
Is this just meant as a temporary arrangement or intentional?
There was a problem hiding this comment.
This could either be temporary or not. I wasn't sure how much we wanted Angular to "own". I had recalled an early conversation where one of the advantages of angular was the fact it forces us to create a nice API from the start - and I wasn't sure whether registering or logging in would fit into that model.
Either way profile makes sense to move into Angular. I'll add a task to asana for it.
|
It'd be nice to consolidate what we've got so far and merge in to To do this, #83 will require updating the PR. I'm going to work at least in the morning on Friday on this, but after that it may be some time before I get the chance to do some - it'd be really nice to leave on a clean slate in terms of PRs - even if there are some outstanding tasks to be picked up by either of us in Asana for user signup. |
Beginnings of user accounts, sqlalchemy etc.
To create the DB run the
create_db.pyscript.There's work to be done on getting the login, register and front page to all look good.
I was thinking that we could go for the github approach that assumes you'll mostly stay logged in. ie. the front page has a registration field on it, with a button at the top for signing in.
However, I think I'm missing a bunch of less files from bootstrap so wanted to talk to you about that - feel free on taking that task though!
Unfortunately asana.com doesn't provide decent issue numbers (they're super long like 5641387051775) so we may need to move away from that. Someone on twitter recommended binfire.com to me. Thoughts?
[I've added some failing tests so that I know what I need to finish up when I get back to this]