Mailing List Archive

Re: Should document Sys::Hostname::Long requirement; ": default" = ?
On February 7, 2004, Meng Weng Wong wrote:
> On Sat, Feb 07, 2004 at 04:28:37PM -0800, Dan Harkless wrote:
> |
> | Hi. Thanks so much for SPF. Truly an idea whose time has come. I now have
> | SPF records published for harkless.org, have spf-milter in place, and link
> | to your site from my public Bookmarks page.
> |
> | A couple of things, though. One, the Mail::SPF::Query install process
> | should warn you if Sys::Hostname::Long isn't present (as it wasn't on my Red
> | Hat 9 box), and tell you that in this case the mail rejection URL will
> | contain a bare hostname with no domain, which may be confusing for the
> | bounce recipient. If it's hard to get the install process to note that
> | Sys::Hostname::Long isn't there, you should at least mention this issue in
> | the Mail::SPF::Query README.
>
> M:S:Q is supposed to work without it, did you get an error? The code
> falls back to Sys::Hostname when it's not present. At least, it's
> supposed to. If it doesn't, let me know what broke.

Yes, I saw how you used 'eval' to have it fail gracefully if
Sys::Hostname::Long isn't there. However, it's not desirable for the 5.7.1
message to contain a bare hostname with no domain (giving text on why.html
reading like "www rejected a message claiming..."). Therefore you should
document in README that this will occur if Sys::Hostname::Long is missing,
and/or have Makefile.PL give a prerequisite warning (not error). Otherwise
people have to dig through the script source to figure out why a FQDN isn't
being used.

> | Secondly, what is the meaning of the ": default" at the end of messages
> | like:
> |
> | Please see http://spf.pobox.com/why.html?sender=spf%40harkless.org&ip=64.165.228.246&receiver=www.harkless.org: default
> |
> | If this is documented anywhere, I was unable to find it, and this has the
> | potential to confuse bounce recipients. If the ": default" is meaningful
> | and you're going to keep it, you should at least have angle brackets around
> | the URL, so the user knows where to end their copy & paste:
> |
> | Please see <http://spf.pobox.com/why.html?sender=spf%40harkless.org&ip=64.165.228.246&receiver=www.harkless.org>: default
> |
>
> wow, I've never seen that before. That could be the milter. Where is
> it showing up?

No, it's not just the milter:

% perl -MMail::SPF::Query -le 'print for Mail::SPF::Query->new(ipv4=>shift, sender=>shift)->result' 1.2.3.4 spf-devel@harkless.org
Mail::SPF::Query: ->new() requires a "helo" argument.
Mail::SPF::Query::new('Mail::SPF::Query','ipv4',1.2.3.4,'sender','spf-devel@harkless.org') called at -e line 1
fail
Please see http://spf.pobox.com/why.html?sender=spf-devel%40harkless.org&ip=1.2.3.4&receiver=www.harkless.org: default
www.harkless.org: domain of spf-devel@harkless.org does not designate 1.2.3.4 as permitted sender
v=spf1 a mx -all

BTW, you might also want to update the example commandline in the
Mail::SPF::Query README to add ", helo=>undef" so people don't get that
warning above. I would also find it helpful to have that example
commandline documented in the Mail::SPF::Query man page, since the README
doesn't get installed anywhere.

> Please join spf-devel@v2.listbox.com and we can discuss this on the list
> where the maintainer for spf-milter is present.

There you go.

--
Dan Harkless
http://harkless.org/dan/

-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/?listname@Ë`Ì{5¤¨wâÇSÓ°)h
Re: Should document Sys::Hostname::Long requirement; ": default" = ? [ In reply to ]
On February 7, 2004, Dan Harkless <spf-devel@harkless.org> wrote:
> On February 7, 2004, Meng Weng Wong wrote:
> > On Sat, Feb 07, 2004 at 04:28:37PM -0800, Dan Harkless wrote:
> > | A couple of things, though. One, the Mail::SPF::Query install process
> > | should warn you if Sys::Hostname::Long isn't present (as it wasn't on my
> > | Red Hat 9 box), and tell you that in this case the mail rejection URL will
> > | contain a bare hostname with no domain, which may be confusing for the
> > | bounce recipient. If it's hard to get the install process to note that
> > | Sys::Hostname::Long isn't there, you should at least mention this issue in
> > | the Mail::SPF::Query README.
> >
> > M:S:Q is supposed to work without it, did you get an error? The code
> > falls back to Sys::Hostname when it's not present. At least, it's
> > supposed to. If it doesn't, let me know what broke.
>
> Yes, I saw how you used 'eval' to have it fail gracefully if
> Sys::Hostname::Long isn't there. However, it's not desirable for the 5.7.1
> message to contain a bare hostname with no domain (giving text on why.html
> reading like "www rejected a message claiming..."). Therefore you should
> document in README that this will occur if Sys::Hostname::Long is missing,
> and/or have Makefile.PL give a prerequisite warning (not error). Otherwise
> people have to dig through the script source to figure out why a FQDN isn't
> being used.
>
> > | Secondly, what is the meaning of the ": default" at the end of messages
> > | like:
> > |
> > | Please see http://spf.pobox.com/why.html?sender=spf-devel%40harkless.org&ip=64.165.228.246&receiver=www.harkless.org: default
> > |
> > | If this is documented anywhere, I was unable to find it, and this has the
> > | potential to confuse bounce recipients. If the ": default" is meaningful
> > | and you're going to keep it, you should at least have angle brackets around
> > | the URL, so the user knows where to end their copy & paste:
> > |
> > | Please see <http://spf.pobox.com/why.html?sender=spf-devel%40harkless.org&ip=64.165.228.246&receiver=www.harkless.org>: default
> > |
> >
> > wow, I've never seen that before. That could be the milter. Where is
> > it showing up?
>
> No, it's not just the milter:
>
> % perl -MMail::SPF::Query -le 'print for Mail::SPF::Query->new(ipv4=>shift, sender=>shift)->result' 1.2.3.4 spf-devel@harkless.org
> Mail::SPF::Query: ->new() requires a "helo" argument.
> Mail::SPF::Query::new('Mail::SPF::Query','ipv4',1.2.3.4,'sender','spf-devel@harkless.org') called at -e line 1
> fail
> Please see http://spf.pobox.com/why.html?sender=spf-devel%40harkless.org&ip=1.2.3.4&receiver=www.harkless.org: default
> www.harkless.org: domain of spf-devel@harkless.org does not designate 1.2.3.4 as permitted sender
> v=spf1 a mx -all

Hi, Meng. Never heard back from you on this, so I did some stepping in the
debugger to at least figure out where the ": default" was coming from. It's
line 335 of Query.pm (version 1.991):

my $smtp_comment = ($smtp_explanation && $smtp_why) ? "$smtp_explanation: $smtp_why" : ($smtp_explanation || $smtp_why);

$smtp_why is "default". The code there is pretty straightforward, so I'm
surprised by your comment that you've never seen this before...

The question is just whether it's beneficial for this ": default" string to
appear in the SMTP response string. I would say no...

> BTW, you might also want to update the example commandline in the
> Mail::SPF::Query README to add ", helo=>undef" so people don't get that
> warning above. I would also find it helpful to have that example
> commandline documented in the Mail::SPF::Query man page, since the README
> doesn't get installed anywhere.

--
Dan Harkless
http://harkless.org/dan/

-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/?listname@Ë`Ì{5¤¨wâÇSÓ°)h
Re: Re: Should document Sys::Hostname::Long requirement; ": default" = ? [ In reply to ]
On Tue, Feb 10, 2004 at 06:57:41PM -0800, Dan Harkless wrote:
|
| Hi, Meng. Never heard back from you on this, so I did some stepping in the
| debugger to at least figure out where the ": default" was coming from. It's
| line 335 of Query.pm (version 1.991):
|
| my $smtp_comment = ($smtp_explanation && $smtp_why) ? "$smtp_explanation: $smtp_why" : ($smtp_explanation || $smtp_why);
|
| $smtp_why is "default". The code there is pretty straightforward, so I'm
| surprised by your comment that you've never seen this before...
|
| The question is just whether it's beneficial for this ": default" string to
| appear in the SMTP response string. I would say no...

Thanks for the detective work. I'm planning to reqrite that section of
the code now that the spec calls for key-value-pairs.

I will fix everything in one swell foop.


-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/?listname@Ë`Ì{5¤¨wâÇSÓ°)h
Re: Re: Should document Sys::Hostname::Long requirement; ": default" = ? [ In reply to ]
Meng Weng Wong wrote:

>On Tue, Feb 10, 2004 at 06:57:41PM -0800, Dan Harkless wrote:
>|
>| Hi, Meng. Never heard back from you on this, so I did some stepping in the
>| debugger to at least figure out where the ": default" was coming from. It's
>| line 335 of Query.pm (version 1.991):
>|
>| my $smtp_comment = ($smtp_explanation && $smtp_why) ? "$smtp_explanation: $smtp_why" : ($smtp_explanation || $smtp_why);
>|
>| $smtp_why is "default". The code there is pretty straightforward, so I'm
>| surprised by your comment that you've never seen this before...
>|
>| The question is just whether it's beneficial for this ": default" string to
>| appear in the SMTP response string. I would say no...
>
>Thanks for the detective work. I'm planning to reqrite that section of
>the code now that the spec calls for key-value-pairs.
>
>I will fix everything in one swell foop.
>
>
The 'default' string comes from the 'all' mechanism. When the
smtp_comment is generated, the peice after the colon -- the smtp_why --
is 'why the mechanism matched'. For an IP4 mechanism it probably gives
the address that matched. For an exists: record it gives the domain that
matched or the TXT record that was returned. This latter feature makes
RBL checking work nicely. It also works nicely with my exists: record.
Try sending mail from baduser@gladstonefamily.net

Philip

--
Philip Gladstone
* Check out the live pondcam at http://pond.gladstonefamily.net

-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/?listname@Ë`Ì{5¤¨wâÇSÓ°)h