Mailing List Archive

On Wed, 24 Jul 2002, John Horne wrote:

> Exim 4.10 manual page 145 shows the redirect router has the 'allow_defer'
> and 'allow_fail' options, but then 'forbid_blackhole' (over the page). I
> tend to class these things together - defer/fail/blackhole - so in that
> respect shouldn't an 'allow_blackhole' be available? Or the other way -
> 'forbid_defer' and 'forbid_fail'?

I suppose it is a bit of a confusion. What I did was to decide which
would be "on" by default. For those I provided "forbid" options. For the
rest, which are "off" by default, I provided "allow" options. In other
words, which option there is depends on the default.

Providing both an "allow" and a "forbid" option for the same thing would
be confusing, because if both were given, which one should prevail?

Of course, I could turn them all into one or the other, with suitable
"true" or "false" defaults. In that case, it would have to be "allow", I
think, because "forbid_xxx = false" is a double negative, which is
always confusing.

--
Philip Hazel University of Cambridge Computing Service,
ph10@cus.cam.ac.uk Cambridge, England. Phone: +44 1223 334714.

On 24-Jul-2002 at 13:22:18 Philip Hazel wrote:
> On Wed, 24 Jul 2002, John Horne wrote:
>> Exim 4.10 manual page 145 shows the redirect router has the 'allow_defer'
>> and 'allow_fail' options, but then 'forbid_blackhole' (over the page). I
>> tend to class these things together - defer/fail/blackhole - so in that
>> respect shouldn't an 'allow_blackhole' be available? Or the other way -
>> 'forbid_defer' and 'forbid_fail'?
>
> Providing both an "allow" and a "forbid" option for the same thing would
> be confusing, because if both were given, which one should prevail?
>
Agreed.

> Of course, I could turn them all into one or the other, with suitable
> "true" or "false" defaults. In that case, it would have to be "allow", I
> think, because "forbid_xxx = false" is a double negative, which is
> always confusing.
>
I would say only change 'forbid_blackhole' to 'allow_blackhole'. This then
ties in with allow_defer and allow_fail. The 'forbid_pipe', 'forbid_file'
etc I likewise group together and since they are all 'forbid_' then that's
okay.

I've been trying to think why by default defer and fail are not allowed, but
blackhole is allowed in a redirect router. I can only think this stems from
Exim 3 when blackhole was introduced and potentially could be used in
/etc/aliases instead of /dev/null? I think would rather have all three
disabled by default :-)


John.

------------------------------------------------------------------------
John Horne, University of Plymouth, UK Tel: +44 (0)1752 233914
E-mail: jhorne@plymouth.ac.uk
PGP key available from public key servers

On Wed, 24 Jul 2002, John Horne wrote:

> I've been trying to think why by default defer and fail are not allowed, but
> blackhole is allowed in a redirect router. I can only think this stems from
> Exim 3 when blackhole was introduced and potentially could be used in
> /etc/aliases instead of /dev/null? I think would rather have all three
> disabled by default :-)

Defer isn't allowed because you don't want ordinary users to be able to
use it. At least, that's my judgement. Fail, I guess, is marginal. It
probably does no harm for ordinary users to be able to use it, as indeed
:blackhole: does no harm.

But it's too late to make incompatible changes now. I promised. So is it
worth actually making any change at all?

--
Philip Hazel University of Cambridge Computing Service,
ph10@cus.cam.ac.uk Cambridge, England. Phone: +44 1223 334714.

On 24-Jul-2002 at 15:30:45 Philip Hazel wrote:
> Defer isn't allowed because you don't want ordinary users to be able to
> use it. At least, that's my judgement. Fail, I guess, is marginal. It
> probably does no harm for ordinary users to be able to use it, as indeed
>:blackhole: does no harm.
>
> But it's too late to make incompatible changes now. I promised. So is it
> worth actually making any change at all?
>
No, not really.

Thanks,

John.

------------------------------------------------------------------------
John Horne, University of Plymouth, UK Tel: +44 (0)1752 233914
E-mail: jhorne@plymouth.ac.uk
PGP key available from public key servers