Create new form state for Enroll Account

[todda00]

Enroll account now has its own form state to allow more control over form labels and buttons, prompting user to set a password instead of changing a password.

Note: Ignore the reset and enroll hooks commit and revert, this change is in a separate PR.

[timbrandin]

This looks great, I will need to test this out so it works as expected, could you confirm these are the steps one need to take to validate this works? Taken from email flows in the guide:

Dependencies: accounts-password

1. Create a user without a password using Accounts.createUser
2. Trigger Accounts.sendEnrollmentEmail from the server for the user.
3. User clicks the link in their email
4. The LoginForm with formState: ENROLL_ACCOUNT shows up with the new label: setPassword which can is translated using softwarerero:accounts-t9n.

Also without accounts-password one should expect the form not to render a reset password or change password button. But that is another issue, but we could check that while testing this too.

[todda00]

That is my exact workflow I used to produce / test this.

Probably would be good to test for accounts-password, but would be a very rare situation since Accounts.sendEnrollmentEmail is part of the accounts-password package, so a user would only get in that situation if an app had it installed, sent out links, then removed the package before the user clicked the link.

[shoetten]

Hey @timbrandin,
i just saw you merged #58. Would you consider this PR for 1.2.10 as well?
This would finally allow me to get rid of my custom fork.

[todda00]

I resolved the conflicts, should be ready to merge.

[timbrandin]

I'm confused there's no changes in the files here, have I merged this already?

[shoetten]

Well, this is strange indeed. But the functionality is definitely not in the code base, yet.
For example helpers.js should include the ENROLL_ACCOUNT: Symbol('ENROLL_ACCOUNT') state and LoginForm.jsx should say

this.setState({
  formState: STATES.ENROLL_ACCOUNT
});
Session.set(KEY_PREFIX + 'state', null);

after line 59.

Maybe something got mangled, when resolving the conflicts @todda00?

[todda00]

Well, I probably messed it up somehow. I went ahead and manually re-applied the changes to the code in my branch, @timbrandin should be good to merge now.

[timbrandin]

LGTM!