On Sat, Sep 12, 2009 at 12:00:49PM -0700, Russ Allbery wrote:
> Steve Langasek <vorlon@debi...> writes:
> > On Sat, Sep 12, 2009 at 10:07:13AM -0700, Don Armstrong wrote:
> >> No additional me too messages are needed; I just wanted to wait to see
> >> if there was some compelling objections before changing the default.
> >> Since there haven't been any, I'll be implementing the fast version
> >> (nnn@bdo and nnn-submitter@bdo always Cc: the submitter) and once I get
> >> a chance to delve into EoC (or whatever I end up using) fix it
> >> properly.
> > But I think making further changes to let submitters opt out via EoC is
> > *not* fixing it properly, and want to dissuade you from doing that :-)
> As a general principle I think it should always be possible for people to
> opt out of mail from any sort of automated or semi-automated system. I
> think supporting opt-out is a good idea. But I think that if the
> submitter opts out of receiving any mail about the bug, that should be
> clear to the package maintainer so that the package maintainer knows that
> follow-up questions will not receive a reply.
> Maybe the best opt-out mechanism would be to clear the submitter field?
I don't think this level of opt-out achieves anything. Perhaps owing to the
existing BTS handling, I'm very conscious of whether a given message I write
to the BTS should be seen by the submitter (and the answer is almost always
yes). A submitter opting out of receiving mail isn't going to get
substantially less mail from me, it's just going to make me work harder to
send it to them by Cc:or bounce when I find their email address through the
web interface. And if they make it impossible for me to send them follow-up
mails (such as by rejecting the mail), I'll close their bug.
We still have -maintonly for mail that /shouldn't/ go to the submitter; is
there any reason not to tune -maintonly behavior to also include
subscribers, if it doesn't today?