Re: ML-config changes
[=- Paul Walker wrote on Mon 7.Nov'05 at 0:39:41 +0000 -=]
> On Mon, Nov 07, 2005 at 01:05:00AM +0100, Vincent Lefevre wrote:
>
> > No, messages with attachments would not be affected by this
> > rule, as the message itself is *not* in base64.
>
> But the attachment is. How are you differentiating between the
> message section and the attachment section?
Hopefully it checks only the main header. I'll have a look at the
MJ2 code.
[=- Thomas Roessler wrote on Sun 6.Nov'05 at 19:18:17 +0100 -=]
> > { charset filtering }
>
> Doesn't help, I'd like to continue getting translations, even if
> possibly not compressed.
Together with the other reactions I understand not to use charset
filtering, fine with me.
So we (can) stick with the "/html" and "/alternative" types?
Just for mutt-dev, or other lists, too?
GLOBALly for all?
> Also, please note that mutt has configuration commands called
> "subscribe" and "unsubscribe", so filtering body content for
> these will be eminently useful if you want to keep people from
> posting configuration files or changes to certain parts of mutt.
a) they are not discarded, just diverted. Now that hopefully the
mod-queue stays small, mistakes might be fixed by moderators.
b) I quote myself:
------ QUOTE BEGIN ------
> For those interested, the current _global_ settings affecting
> _all_ mutt ML, probably incl. mutt-users (not sub'ed myself):
> http://wiki.mutt.org/?MuttChannel/Paste
>
> I'm sure some items should be removed, but I won't decide that all
> by myself. Comments/discussion welcome.
------- QUOTE END -------
Up to you which keywords/ patterns you want to keep and which to
remove. If you're not scared of any mistakes with majordomo usage
sent to the whole list(s), then we can drop all entries.
[=- Alain Bench wrote on Sat 5.Nov'05 at 15:52:01 +0100 -=]
> > size-limit of 20,000 bytes per post
>
> Good! But isn't the limit a *little* low for a developpers list?
> I mean patches, reports with traces, and such should be welcome
> once compressed and reasonable size, no?
Sure they are welcome. 20KB isn't reasonable when compressed? :)
> Let's see current practice: {...}
> {...}
> But as soon as at rank #10 by size, around 110Ko, there begins
> to be a majority of adequate posts, even if some people should
> be flamed for lack of zippage. And lack of thinking. And
> throwing dupes at BTS and MD!
So, you prefer what, ... 50KB compressed?
The flaming would then be taken over by the bounce (for the size
part).
I'll update the notice to include compressing advice.
> > I don't know whether it's better to silently ignore
> > non-members immediately not to cause useless bounces by faked
> > headers...
>
> Ah no: Silent dropping is not really acceptable, IMHO.
> Especially {...} from non-subscribers directed here by mutt -v
> "To contact the developers, please mail to <mutt-dev@xxxxxxxx>".
> I mean subscription to a dev list was not really requirable
> before. It is now forcibly required, but a notice *has* to be
> bounced. Hinting to subscription or Gnats. Pardonning
> inconvenience.
Alain, you're too polite. ;)
But I'll add your suggestion about bts.
> Humm... Would it be possible in the incoming flow to first
> despam most obvious with a high threshold (with a not too smart
> low power filter), then second bounce non-members and
> perhaps-spam? The specific difficulty with anti-spam here is
> that we receive sometimes legitimate spam, attached to a report
> of Mutt malfunction.
Current (soon to be) standing of filter-order:
1) Content-type bounce (so far only the 2 per above).
2) non-member bounce.
3) size bounce.
That's the best I can think of with my skills with MJ2.
> Anyway past filtering rules may have had a problem, as some
> legitimate mails were /dev/nulled. At least some mails by Tamo,
> Christian Ebert, myself... in the past monthes/year.
The filer-list is supposed to be thrown away unseen.
Everthing else will be revised.
(Once I've tracked down all the remaining mysteries with MJ2)
--
© Rado S. -- You must provide YOUR effort for your goal!
Even if it seems insignificant, in fact EVERY effort counts
for a shared task, at least to show your deserving attitude.