bug tracker spam

bug tracker spam

[Glenn Morris]

On bug tracker spam:

1. I think Stefan's suggestion to reassign spam to a "spam" package
(as documented in admin/notes/bugtracker) is better than closing the
bugs. With the latter they get archived with all the rest, with the
former they are more easily separated from real reports.

2. I think somebody (or bodies) does a very good job of moderating
bug-gnu-emacs, so that I almost never see spam there (thank you!).

Yet the bug tracker collects a fair bit of spam that has to be dealt
with separately. Where is this coming from? Is there no way to combine
the moderation step with removing the spam from the tracker, to avoid
duplication of effort?

Re: bug tracker spam

[Andreas Schwab]

Glenn Morris <rgm@gnu.org> writes:

> 2. I think somebody (or bodies) does a very good job of moderating
> bug-gnu-emacs, so that I almost never see spam there (thank you!).

Take a look at the headers: like every mailing list at gnu.org it is
automatically filtered through a spam and virus checker.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: bug tracker spam

[Glenn Morris]

Andreas Schwab wrote:

> Take a look at the headers: like every mailing list at gnu.org it is
> automatically filtered through a spam and virus checker.

OK; though that doesn't rule out extra human moderation.

The question remains: how does the spam get into the tracker?

And if there's no spam on bug-gnu-emacs, and the tracker spam comes in
through some other route, would we lose much by making bug-gnu-emacs
the only way to open a new report in the tracker?

Re: bug tracker spam

[Eli Zaretskii]

> From: Glenn Morris <rgm@gnu.org>
> Date: Wed, 20 Aug 2008 16:11:48 -0400
> Cc: emacs-devel@gnu.org
> 
> Andreas Schwab wrote:
> 
> > Take a look at the headers: like every mailing list at gnu.org it is
> > automatically filtered through a spam and virus checker.
> 
> OK; though that doesn't rule out extra human moderation.

Most, if not all, gnu.org mailing lists _are_ human-moderated.  How do
you think we can have emacs-devel so spam-clean?

> And if there's no spam on bug-gnu-emacs, and the tracker spam comes in
> through some other route, would we lose much by making bug-gnu-emacs
> the only way to open a new report in the tracker?

Human moderation works mainly by holding mail from non-subcribers.
This is a good policy for a regular mailing list, where anyone who
wants to participate in discussions can simply subscribe to the list.
But for a bug-tracker gateway, this is IMO not a very good idea,
because someone who wants to report a bug won't normally consider
subscribing, as from her point of view reporting a bug is doing us a
favor.  Thus, implementing your suggestion might potentially become a
significant burden on the moderator due to a large volume of held
messages that the moderator would need to review.

Re: bug tracker spam

[Glenn Morris]

Eli Zaretskii wrote:

> Human moderation works mainly by holding mail from non-subcribers.
> This is a good policy for a regular mailing list, where anyone who
> wants to participate in discussions can simply subscribe to the list.
> But for a bug-tracker gateway, this is IMO not a very good idea,
> because someone who wants to report a bug won't normally consider
> subscribing, as from her point of view reporting a bug is doing us a
> favor.

Yes, and you can replace "bug-tracker" with "bug mailing list" above
and nothing changes.

> Thus, implementing your suggestion might potentially become a
> significant burden on the moderator due to a large volume of held
> messages that the moderator would need to review.

I'm not asking for the list to be moderated if it isn't already.

If it is being moderated, let's make more efficient use of the work
someone is so kindly doing; by using it at zero extra effort to keep
spam out of the tracker.

Re: bug tracker spam

[Eli Zaretskii]

> From: Glenn Morris <rgm@gnu.org>
> Cc: schwab@suse.de,  emacs-devel@gnu.org
> Date: Thu, 21 Aug 2008 12:02:59 -0400
> 
> Eli Zaretskii wrote:
> 
> > Human moderation works mainly by holding mail from non-subcribers.
> > This is a good policy for a regular mailing list, where anyone who
> > wants to participate in discussions can simply subscribe to the list.
> > But for a bug-tracker gateway, this is IMO not a very good idea,
> > because someone who wants to report a bug won't normally consider
> > subscribing, as from her point of view reporting a bug is doing us a
> > favor.
> 
> Yes, and you can replace "bug-tracker" with "bug mailing list" above
> and nothing changes.

Yes, it does: people tend to submit bugs to a tracker more than they
do to a list where they need to talk to a human.  I believe this was
one of the important reasons for starting the tracker in the first
place.

> I'm not asking for the list to be moderated if it isn't already.
> 
> If it is being moderated, let's make more efficient use of the work
> someone is so kindly doing; by using it at zero extra effort to keep
> spam out of the tracker.

I was trying to explain that it isn't zero extra effort, but I
obviously failed to explain that.

Re: bug tracker spam

[Glenn Morris]

Eli Zaretskii wrote:

> Yes, it does: people tend to submit bugs to a tracker more than they
> do to a list where they need to talk to a human.  I believe this was
> one of the important reasons for starting the tracker in the first
> place.

Really?

> I was trying to explain that it isn't zero extra effort, but I
> obviously failed to explain that.

Given your first paragraph, then your point obviously follows. It's
just both sentences of your first paragraph take me completely by
surprise.

Re: bug tracker spam

[Glenn Morris]

Actually, I still don't know what your point is.

Right now, everything (?) that goes to the bug tracker gets sent to
bug-gnu-emacs, even the interminable control messages, where it gets
moderated.

Maybe you're saying you think it's an unreasonable burden for someone
to moderate bug-gnu-emacs at all these days.

Re: bug tracker spam

[Eli Zaretskii]

> From: Glenn Morris <rgm@gnu.org>
> Cc: schwab@suse.de,  emacs-devel@gnu.org
> Date: Thu, 21 Aug 2008 13:52:33 -0400
> 
> Maybe you're saying you think it's an unreasonable burden for someone
> to moderate bug-gnu-emacs at all these days.

I know that if I were the moderator, it would have been unbearable.

bug tracker spam

[Glenn Morris]

What's the recommended way to deal with spam in the bug tracker?
Just using "done" doesn't seem very good.

TIA.

Re: bug tracker spam

[Don Armstrong]

On Thu, 29 May 2008, Glenn Morris wrote:
> What's the recommended way to deal with spam in the bug tracker?
> Just using "done" doesn't seem very good.

If it's a new bug, I just kill them manually. If they're a message to
an existing bug, I kill off the message.


Don Armstrong

-- 
When I was a kid I used to pray every night for a new bicycle. Then I 
realised that the Lord doesn't work that way so I stole one and asked
Him to forgive me.
 -- Emo Philips.

http://www.donarmstrong.com              http://rzlab.ucr.edu

Re: bug tracker spam

[Glenn Morris]

Don Armstrong wrote:

> On Thu, 29 May 2008, Glenn Morris wrote:
>> What's the recommended way to deal with spam in the bug tracker?
>> Just using "done" doesn't seem very good.
>
> If it's a new bug, I just kill them manually. If they're a message to
> an existing bug, I kill off the message.

Right, but what do I actually _do_ when I'm sitting here looking at

#334, #333, #329, #325, #307, #304, #298, #290, #287, #275, #273,
#271, #257, #256, #255, #242, #241, #238, #235, #234, #228, #211,
#210

What commands do I email to where to remove them from the system
altogether?

Re: bug tracker spam

[Don Armstrong]

On Thu, 29 May 2008, Glenn Morris wrote:
> Right, but what do I actually _do_ when I'm sitting here looking at
> 
> #334, #333, #329, #325, #307, #304, #298, #290, #287, #275, #273,
> #271, #257, #256, #255, #242, #241, #238, #235, #234, #228, #211,
> #210
> 
> What commands do I email to where to remove them from the system
> altogether?

Just mail me; I use rm. ;-)


Don Armstrong

-- 
I really wanted to talk to her.
I just couldn't find an algorithm that fit.
 -- Peter Watts _Blindsight_ p294

http://www.donarmstrong.com              http://rzlab.ucr.edu

Re: bug tracker spam

[Glenn Morris]

Don Armstrong wrote:

> Just mail me; I use rm. ;-)

It would be nice if getting rid of spam was something maintainers
could do. Perhaps just tagging things as spam so that they appear in a
separate section of the summary until they are fully removed.

more spam: #304, #340

There are also things like #312, #313, #207 that seem to be mistakes.
I don't know what to do about those.

Couple of other questions:

How do I close a report in such a way that it is marked "this is not a bug"?
This isn't the same as "wontfix", IMO.

How do you (or will you) control who is allowed to manipulate bug reports?


PS. Thanks for being so involved with this.

Re: bug tracker spam

[Don Armstrong]

On Sat, 31 May 2008, Glenn Morris wrote:
> Don Armstrong wrote:
> > Just mail me; I use rm. ;-)
> 
> It would be nice if getting rid of spam was something maintainers
> could do. Perhaps just tagging things as spam so that they appear in
> a separate section of the summary until they are fully removed.

Usually we use a cgi script so people can mark bugs that contain or
are spam, and an administrator manually removes them later. The reason
why this shouldn't be handled by maintainers is that it's the only
action that cannot be reverted in the bts. [I suppose that eventually
I'll have to come up with a method that scales better than the current
ones, but it's scaled to > 450,000 bugs in Debian, and there are a
bunch of things that need fixing which are higher on my priority
list.]

> more spam: #304, #340

Deleted.

> There are also things like #312, #313, #207 that seem to be mistakes.
> I don't know what to do about those.

Anything that's not clearly spam, I suggest to keep. [They'll
disappear from the main page as they get archived, so it won't be a
problem long-term.]

> Couple of other questions:
> 
> How do I close a report in such a way that it is marked "this is not
> a bug"? This isn't the same as "wontfix", IMO.

You just close it, explaining that it's not a bug and why in the -done
message. [A lot of the -done messages I've seen so far have been far
too terse to extract any useful information from; it makes it really
useful if a done message includes at least the commit message that
caused a bug to be fixed and the revision.]

If this happens a lot, we can add a notabug tag.

> How do you (or will you) control who is allowed to manipulate bug
> reports?

We don't, generally speaking, because controlling it doesn't scale. If
there's a problem, we can blacklist people. [I have all of about 5
entries in Debian's blacklist out of tens of thousands who have
modified bugs, so it's rarely a problem.]


Don Armstrong

-- 
After the first battle of Sto Lat, I formulated a policy which has
stood me in good stead in other battles. It is this: if an enemy has
an impregnable stronghold, see he stays there.
 -- Terry Pratchett _Jingo_ p265

http://www.donarmstrong.com              http://rzlab.ucr.edu

Re: bug tracker spam

[Stefan Monnier]

> Don Armstrong wrote:
>> Just mail me; I use rm. ;-)

> It would be nice if getting rid of spam was something maintainers
> could do. Perhaps just tagging things as spam so that they appear in a
> separate section of the summary until they are fully removed.

Actually, now that I think about it, better than closing them, we can
just move them to a special "spam" package: "reassign <BN> spam".


        Stefan

Re: bug tracker spam

[David Hansen]

On Thu, 29 May 2008 19:26:00 -0400 Glenn Morris wrote:

> What's the recommended way to deal with spam in the bug tracker?
> Just using "done" doesn't seem very good.

Being bankrupt may be fixable.  AFAIK a small penis requires a complete
rewrite (but M-x doctor should offer a workaround).

David