The new "device registration" feature of Firefox Accounts requires that Firefox instances register themselves with Firefox Accounts. The new endpoints are outlined in https://github.com/mozilla/fxa/blob/phil/fxa-45-requirements/features/FxA-45-device-registration-api/README.md services/fxaccounts/FxAccountsClient.jsm contains methods used to make requests to the FxA auth server and may be a natural place for this work.

Add new methods to FxAccountsClient.jsm to interface `devices` API methods on the auth server. * `deviceRegister` * `deviceUpdate` * `deviceDestroy` * `deviceList`

:markh I do not think the above patch is ready for final review because I still have questions about how change password, reset password, and force auth will work where an existing deviceID must be associated with a new sessionToken. It's possible `updateDevice` will need to be modified or a new method added to handle these cases. Still, early feedback is very much appreciated!

Comment on attachment 8683622 [details] [diff] [review] Add FxA "devices" API methods to FxAccountsClient.jsm Review of attachment 8683622 [details] [diff] [review]: ----------------------------------------------------------------- This LGTM, although note that FxAccountsClient is really only abstracting the request/response handling away, so we are also going to want a more "public" interface to this - currently this is really a "private" module with no external callers. I'm not sure if that should be on FxAccounts.jsm or on a new, say, FxDevice.jsm - I'd need to think more about how tightly bound the account itself is with the device.

(In reply to Mark Hammond [:markh] from comment #3) > or on a new, say, FxDevice.jsm - I'd need to think more about how tightly > bound the account itself is with the device. Thinking more, and given we are talking about using FxA storage for device stuff (which sounds fine), it sounds like it should just be on FxAccounts.jsm

Thanks for the feedback Mark, and w.r.t. the public interface in FxAccounts.jsm, yup. I was ostrich approaching that work until [1] is a bit more fleshed out. Do you think it is reasonable to finalize/merge this patch and then do a second patch w/ the public interface, or would it be easier to roll those into a single patch? [1] - https://github.com/mozilla/fxa/blob/fxa-45-device-registration-api-firefox/features/FxA-45-device-registration-api/FX-DESKTOP.md (In reply to Mark Hammond [:markh] from comment #3) > Comment on attachment 8683622 [details] [diff] [review] > Add FxA "devices" API methods to FxAccountsClient.jsm > > Review of attachment 8683622 [details] [diff] [review]: > ----------------------------------------------------------------- > > This LGTM, although note that FxAccountsClient is really only abstracting > the request/response handling away, so we are also going to want a more > "public" interface to this - currently this is really a "private" module > with no external callers. I'm not sure if that should be on FxAccounts.jsm > or on a new, say, FxDevice.jsm - I'd need to think more about how tightly > bound the account itself is with the device.

(In reply to Shane Tomlinson [:stomlinson] from comment #5) > Do you think it is reasonable to finalize/merge this patch and then do a > second patch w/ the public interface, or would it be easier to roll those > into a single patch? I'm not a big fan of "speculative" code, so if there's no other compelling reason I'd rather see the final integrated code land.

Closing this in favour of the greater scope of #1227527, the patch and review comments from have been noted though.