Remove or obfuscate spam in the *-bugs mailing lists?

Remove or obfuscate spam in the *-bugs mailing lists?

[Christopher Faylor]

I would like to do something about the spam in the bugzilla backed
mailing lists.

I have two options:

1) Delete the messages (easy with mhonarc -rmm)
2) Munge the message body and subjects (already have an untested
script to do this)

I only discovered the mhonarc option after writing a script to munge the
spam.  I'd prefer to use the option since I just noticed that, in my
script, I had not taken mhonarc's database into consideration when
modifying index.html, msgNNNNN.html, and friends.

Is there a reason to keep bread crumbs in the archive or is it ok to
just eliminate any bug marked as spam?  I intend to do this
retroactively so that if a bug is marked as spam older messages will
also be munged/removed.

Everything will also be backed up, of course.

cgf

Re: Remove or obfuscate spam in the *-bugs mailing lists?

[Gerald Pfeifer]

On Tue, 9 Sep 2014, Christopher Faylor wrote:
> Is there a reason to keep bread crumbs in the archive or is it ok 
> to just eliminate any bug marked as spam?  I intend to do this
> retroactively so that if a bug is marked as spam older messages 
> will also be munged/removed.

In general "We have a strong policy of not editing the web archives"
( from https://gcc.gnu.org/lists.html ).

In a crazy situation like this, I believe it's the right thing for 
us to do, however, and you have my full support.

Thank you,
Gerald

Re: Remove or obfuscate spam in the *-bugs mailing lists?

[Christopher Faylor]

On Tue, Sep 09, 2014 at 07:21:16PM +0200, Gerald Pfeifer wrote:
>On Tue, 9 Sep 2014, Christopher Faylor wrote:
>> Is there a reason to keep bread crumbs in the archive or is it ok 
>> to just eliminate any bug marked as spam?  I intend to do this
>> retroactively so that if a bug is marked as spam older messages 
>> will also be munged/removed.
>
>In general "We have a strong policy of not editing the web archives"
>( from https://gcc.gnu.org/lists.html ).

Right.  There are similar words in sourceware.org/lists.html (which I
may have even written).  It's usually pointless to edit the archives
because the information eventually leaks out to the interwebs.  That +
it could become a full-time job to tweak the archives.

But, it really bugs me that some evil dudes could try to subvert our
system so I'd like to eliminate any remaining benefit they could see
from sending spam here.

>In a crazy situation like this, I believe it's the right thing for 
>us to do, however, and you have my full support.

Thanks.  Which option do you prefer: Deleting or keeping munged,
content-free copies around?  The only reason that I could see for the
latter is that existing url references to the archives would continue to
work.  But, it's difficult to see any valid reason for a reference to
bugzilla spam.

Nevertheless, I'm willing to do either.  If no one objects, I'll do
the deletion.  Otherwise, I'll just nuke content and subjects.

cgf

Re: Remove or obfuscate spam in the *-bugs mailing lists?

[Jeff Law]

On 09/09/14 12:53, Christopher Faylor wrote:
> Right.  There are similar words in sourceware.org/lists.html (which I
> may have even written).  It's usually pointless to edit the archives
> because the information eventually leaks out to the interwebs.  That +
> it could become a full-time job to tweak the archives.
>
> But, it really bugs me that some evil dudes could try to subvert our
> system so I'd like to eliminate any remaining benefit they could see
> from sending spam here.
Yea, and that's probably the biggest reason to delete the crap.

> Nevertheless, I'm willing to do either.  If no one objects, I'll do
> the deletion.  Otherwise, I'll just nuke content and subjects.
No strong preference, particularly since you're doing the work.  I'm 
happy with either choice.

jeff

Re: Remove or obfuscate spam in the *-bugs mailing lists?

[Gerald Pfeifer]

On Tue, 9 Sep 2014, Christopher Faylor wrote:
> But, it really bugs me that some evil dudes could try to subvert our
> system so I'd like to eliminate any remaining benefit they could see
> from sending spam here.

Agreed. :-)

> Thanks.  Which option do you prefer: Deleting or keeping munged,
> content-free copies around?  The only reason that I could see for the
> latter is that existing url references to the archives would continue to
> work.  But, it's difficult to see any valid reason for a reference to
> bugzilla spam.

Would other, non-spam mails change their URLs as well?  If that's
the case, munging might be better.  Otherwise I'd be in favor of
simply yanking the crap.  Which seems to be your preference as well.

Gerald

Re: Remove or obfuscate spam in the *-bugs mailing lists?

[Christopher Faylor]

On Tue, Sep 09, 2014 at 10:44:41PM +0200, Gerald Pfeifer wrote:
>On Tue, 9 Sep 2014, Christopher Faylor wrote:
>> But, it really bugs me that some evil dudes could try to subvert our
>> system so I'd like to eliminate any remaining benefit they could see
>> from sending spam here.
>
>Agreed. :-)
>
>> Thanks.  Which option do you prefer: Deleting or keeping munged,
>> content-free copies around?  The only reason that I could see for the
>> latter is that existing url references to the archives would continue to
>> work.  But, it's difficult to see any valid reason for a reference to
>> bugzilla spam.
>
>Would other, non-spam mails change their URLs as well?  If that's
>the case, munging might be better.  Otherwise I'd be in favor of
>simply yanking the crap.  Which seems to be your preference as well.

No, the mail would just disappear from the archive, leaving everything
else intact.  One other benefit with deletion is that since I'll be
using the mail archiver to delete it I won't have to worry about locking
the archives while I'm doing this.  That happens automatically.

So, it looks like I will be going with deletion.

If this seems to be reliable, I'll add a cron job to delete spam from
the archives automatically from now on.

cgf