launchpad-dev team mailing list archive

["Francis J. Lacoste" <francis.lacoste@xxxxxxxxxxxxx>]

On 11-10-17 12:22 AM, Jeroen Vermeulen wrote:
> On 2011-10-13 15:06, Robert Collins wrote:
> 
>> Curtis script expunges dangling team join requests; expiring them
>> would involve notifications ('sorry your request was not replied to'),
>> and that would be bogus if e.g. the person had been deleted (or
>> alternatively merged into someone in the team). Neither case can
>> happen if the person merge bug is fixed.
> 
> It was stupid of me to ignore notifications.  But is this really a case
> of "membership request expiry is harder with the broken data still in
> place," or is it actually "notifying users of anything related to
> membership requests is harder with the broken data still in place"?
> 
> ISTM the notification code needs to be robust against this sort of thing
> as a matter of necessity.  Wasn't similar robustness built in back when
> we realized that users don't always have preferred email addresses?


This is an affirmation that I've seen mentioned a couple of time
recently, but I cannot assert his truth value. In my understanding, this
is still a myth. Launchpad users always have a preferred email address.
Teams might not have one, as person records that we imported but nobody
"activated". But as soon as a user "logs in" Launchpad, and thus become
a user, they have a preferred address.

If that's not the case, please explain how this invariant can now be broken.

-- 
Francis J. Lacoste
francis.lacoste@xxxxxxxxxxxxx

Attachment: signature.asc
Description: OpenPGP digital signature