CUNY Academic Commons

Hi Ray,

The commit looks great. One niggle: for incorrect addresses, can we expunge the extra "from"?

...make sure to reply using the same email address at which you received the original notification from.

I pulled these latest changes to cdev, and reactivated the plugin. I'm tailing wp-content/bp-rbe-debug.log, and I can see when emails from incorrect addresses are coming in and getting blocked, but the error emails are being sent. Successful RBE emails are going out correctly, so email is working more generally. Could you have a look and see if I've overlooked something?

Love the "via email" bit :)

Hi Boone,

I'm tailing wp-content/bp-rbe-debug.log, and I can see when emails from incorrect addresses are coming in and getting blocked, but the error emails are being sent. Successful RBE emails are going out correctly, so email is working more generally. Could you have a look and see if I've overlooked something?

I made it a conscious choice to not log outgoing failure messages in bp-rbe-debug.log. Although, I believe I should now that you've talked me into it :)

Is this what you mean? If not, let me know!

Shoot. I meant "but the error emails are NOT being sent". Not sure how to debug that :)

What failure type did you try? (User doesn't exist, etc.)

I'll take a look on cdev.

Specifically, I tried user doesn't exist (responding with boone@gorg.es to a notification sent to boonebgorges@gmail.com).

Just tried the user doesn't exist failure type and I received the email notification at an alternate email address (Google Apps Mail FWIW).

Did you check your spam folder? If not, can you try again?

Update: Ah-ha!

I looked closer at the bp-rbe-debug.log and boone@gorg.es is already a valid user but is not a member of the group or banned.

My parsing logic doesn't send failure emails for members who are not a member of a group or a banned group user. That's why you weren't receiving the failure email!

I probably should send failure emails for users who are not a member of a group. Banned users shouldn't receive any failure emails as it's currently implemented.

What do you think?

Aha! Good find, Ray.

I guess the ideal situation would be that the reply-to hashes were user-specific. That way, the only address that can respond to a given notification is the user to whom it was originally sent.

If that's too difficult to implement, a group check would be more than enough. That'll mean that users can't join public groups by posting to them, as they can in the web interface, but that seems OK, since they won't be getting the RBE notification in the first place.

The capability to add user-specific hashes is already in RBE. See the hash used for members posting new group forum topics via email as an example.

However, for "Reply-To" emails, this isn't. Doing so will add an extra DB lookup to grab their user ID to add to the hash before sending each email. Might have to reconfigure parts of the codebase to make it work as well.

I think if we run into issues with spam, I can revisit this.

Did we want to send a failure email for users who are not a member of a group? Could be useful if a user decides to manually leave a group, but decided to reply to an older email?

I think if we run into issues with spam, I can revisit this.

Sounds good.

Did we want to send a failure email for users who are not a member of a group? Could be useful if a user decides to manually leave a group, but decided to reply to an older email?

Yes, if it's not too much trouble. We don't want to allow behavior over email (such as posting to a private group where one is not a member) that we don't allow on the site.

Part of commit 8ed8a2e adds failure emails for users who are not a member of a group.

If there aren't any other use cases, I think we can mark this issue as resolved!