update SMS verification

[derhuerst]

This PR adapts the SMS verification modal to the changes in the server and adds a few more checks. Fixes #3514.

@chevdor Please check, I will keep the ropsten server running now.

22e1384 and c62c72d make the modal wait until the Puzzled event, posted by the server, has actually been mined. This prevents the follow-up confirm tx from failing (and therefore using up all gas).

As requested by @gavofyork, 6509252 makes the modal check if the code is valid using contract.confirm.call().

[chevdor]

I cannot check right now but I will within 2h max.

[derhuerst]

@jacogr @ngotchac

util/subscribe-to-event is a helper function that returns an event-based interface to listen for events.

Because the promise returned by Contract.prototype.subscribe resolves after the callback has been called for the first time, it is hard to unsubscribe if you already got what you want. You will have to store that you want to unsubscribe as soon as the promise resolves.

I think it now depends which interface is easier to consume. If you want many events, using the event-based interface is really nice. For one, it won't be worth the boilerplate.

[chevdor]

I merged jr-update-sms-verification and rebuilt.
I am still getting Failed to request a confirmation SMS: Failed to fetch

[derhuerst]

@chevdor Sorry, I forgot to mention that it only works on ropsten. For me and @jacogr, it just worked. Will ping you on Gitter with details.

[chevdor]

@derhuerst I also forgot to mention it I am also (now) on Ropsten. If it works for you both, I can propose you to connect with me and we check together or we leave it as it is, and I test again once the PRs is merged. Up to you.

[derhuerst]

Still to do:

• show the server errors in the UI – paritytech/sms-verification@7ab125a
• normalise number input – paritytech/sms-verification@5d135b2
• make number input friendlier – done in f1bc69a

[jacogr]

? Redux has isTest, don't want to maintain it in yet another place, we just simplified and got it down to 1. (And had to make changes.)

In addition, we are really starting to clutter /utils now with all these single function files. There is no reason why things can't be grouped, e.g. had to move nullable to protypes once we added another one. (General grumble)

[derhuerst]

Fixed as of 88cb1df.

[jacogr]

Would have really preferred a general solution that (a) doesn't add another way of doing things, (b) is available to everybody. But suppose it solves the immediate issue, so once we have a general solution in-place will have to come back.

[coveralls]

Changes Unknown when pulling f1bc69a on jr-update-sms-verification into ** on master**.

[coveralls]

Coverage increased (+0.003%) to 85.908% when pulling 88cb1df on jr-update-sms-verification into a7037f8 on master.