On 11/1/11 4:55 PM, Mark Sapiro wrote:
Carl Zwanzig wrote:
In the case below, the first command attempted to subscribe an address which is already a member so it failed and the remaining commands were unprocessed. Why is that an error? I would expect it to ignored, just like setting a property to it's current value.
Because Mailman/commands/cmd_subscribe.py distinguishes it from success and calls it fatal. Granted, an argument could be made that it shouldn't be fatal and similarly that unsubscribe of a non-member address shouldn't be fatal either, but it isn't completely obvious that that should be the case.
One issue I see with this behavior, is that this breaks the privacy that mailman gives to lists. The web interface is very careful to NOT let you know if a given address is subscribed or not (unless the list publishes its subscriber list). This gives a method that you can get a message to one email address indicating that another, unrelated, email address is subscribed to the list.
-- Richard Damon