1697117 - oAuth2.0 for Yahoo Plus

Status: NEW

(Thunderbird :: Account Manager, defect)

Description

[Matt]

+++ Yahoo oAth was implemented in #1293958 +++

That implementation used the general free server names, but yahoo plus apparently uses

plus.pop.mail.yahoo.com
and
plus.smtp.mail.yahoo.com

Implementing Bug 1591782 would remove the need for these type of bugs on a regular basis

Comment 1

[Ben Bucksch (:BenB)]

If you have a Yahoo Plus account, and you use the account creation wizard to set up your email address, and you simply follow the dialog, without changing anything, what happens? I know we would configure imap.mail.yahoo.com, and (if all goes as expected) we would use OAuth2 to authenticate the user. But: Is there anything going wrong? I presume that imap.mail.yahoo.com would also work for plus users, no?

Comment 2

[Matt]

I filled this bug based on the discussion here https://support.mozilla.org/en-US/questions/1327586 although https://support.mozilla.org/en-US/questions/1310273 covers the same ground. Perhaps sfhowes can off you an insight.

This is not really a new account wizard issue, it is an issue that those doing as the provider says can't get access to oAuth is my understanding.

Comment 3

[sfhowes]

I don't have a Plus account to test, but I suspect that if the Plus servers support OAuth2 (probably), it won't be possible to use it in TB, even if the authentication is set to OAuth2 manually in Config.editor (setting mail.server.serverN.authMethod to 10), since it's not hardcoded into TB for Plus, as it is e.g. with gmail, AOL, mail.ru and Yahoo free (mail.ru example https://bugzilla.mozilla.org/show_bug.cgi?id=1231642).

Comment 4

[Ben Bucksch (:BenB)]

The key question is: Does imap.mail.yahoo.com also work for Plus users?
If that's the case, then the account creation wizard creates the correct config out of the box for these users as well, including OAuth2.

Comment 5

[Matt]

No the key question in my opinion is what can we do to make the application work correctly using the settings the provider offers. If we discount Yahoo plus for a moment, the wizard does not work for ATT, Frontier and a host of others that use Yahoo infrastructure and their own server names. Hence my opening comment that Bug 1591782 would be the better way to address the issue.

The current arrangement of hardcoded provider server names to enable oAuth is not sustainable in the long term.

Comment 6

[Ben Bucksch (:BenB)]

Does imap.mail.yahoo.com also work for Plus users?

because that tells me how the application behaves. This needs to be tested by somebody who has a plus account. I cannot proceed without somebody helping testing.

ATT, Frontier and a host of others that use Yahoo infrastructure and their own server names

If that doesn't work, I would like to fix that. Same here, we need users with accounts on these servers, to test the actual behavior, otherwise we cannot know what works and what doesn't. If the default config fails to fetch mail, and we have testers, could you please make a list and file a new bug against me? We need to pre-configure these in the ISP DB to use the right settings, so that it works by default for end users.

Comment 7

[Matt]

I have filed bug 1698316 for oauth on ATT and associated domains. I have no testers, but support has been full of those complaining for years. The issue is clearly documented in support and on the ATT web site. We also know using the yahoo server works for ATT customers. But that is not a suitable result as ATT support will have the user change that as soon as they connect them, even if it is a web search where that advises the "correct" servers to use. We need to offer yahoo for the ATT.com servers.

Comment 8

[Ben Bucksch (:BenB)]

I have no testers

If there are so many issues in support that you are in contact with, there should be a user that can help us.

None of these bugs can progress without testers that have accounts on these servers. I need a number of questions answered. In any case, we will not roll any changes out without having tested it.

Please make sure that we have a user that is willing to actively cooperate in testing.