930598 - Generic HAWK client

Status: VERIFIED DUPLICATE of bug 935232

(Cloud Services :: Firefox: Common, defect)

Description

[Zachary Carter [:zaach]]

Services such as Sync and Firefox Accounts on desktop and FxOS require HAWK for certain API requests. We'll need a simple HAWK client and a helper for deriving credentials from a token.

Comment 1

[Jed Parsons (use needinfo, please) [:jedp, :jparsons]]

Can we give this a higher-level name, like FXAClient?

Comment 2

[Tim Taubert [:ttaubert] (inactive)]

Yes, I was already told that HAWK is not a good name. FXAClient sounds better to me for a start.

Comment 3

[Tim Taubert [:ttaubert] (inactive)]

Attachment: 0001-Bug-930598-Add-generic-client-for-the-Firefox-Accoun.patch

Here's the first version mostly taken from the existing HAWK.jsm in the fxaccounts working branch. Do we need to support all request specified for the FxAccounts API or are the FxOS requirements similar to the one on Desktop? Can we use services-common/utils and services-crypto/utils on FxOS?

Comment 4

[Tim Taubert [:ttaubert] (inactive)]

Oh and we should of course add some tests. I will look into using SJS files and the test vectors defined in the API document.

Comment 5

[Tim Taubert [:ttaubert] (inactive)]

No, SJS won't work. Let's just use an HTTP Server instance. I will additionally take a look at toolkit/modules/Http.jsm. Seems like we should probably use that.

Comment 6

[Richard Newman [:rnewman]]

(In reply to Tim Taubert [:ttaubert] from comment #3)
> Can we use services-common/utils and services-crypto/utils on FxOS?

According to my reading of services/moz.build, yes. (They're already used by FHR on FxOS.)

Comment 7

[Tim Taubert [:ttaubert] (inactive)]

Attachment: Add some more helper functions to services-common/crypto utils

Most of the stuff stolen from what Brian checked into markh's feature branch.

Comment 8

[Chris Karlof [:ckarlof]]

Much of the attachment 824639 [details] [diff] [review] is covered by the work in bug 935232. There are some hkdf tests in attachment 824639 [details] [diff] [review] that are not represented in the attachments in 935232. We should probably move these over. I'm resolving as duplicate.

Comment 9

[James Bonacci [:jbonacci]]

OK.