Re: Remove message-valid-fqdn-regexp?

Re: Remove message-valid-fqdn-regexp?

[Lars Ingebrigtsen]

Lars-Johan Liman <liman@netnod.se> writes:

> Esteemed Sirs,
>
> Can I humbly suggest that the variable message-valid-fqdn-regexp and the
> use thereof be removed from GNUS? Or possibly set to a default value of
> ".*" (or similar) so that it's open but so that limits can be imposed by
> the user, if he/she so wishes?
>
> In my view it is simply overtaken by events. Back in the day, the root
> zone "moved" very slowly. New TLDs were introduced in singles per
> decade. Having a watchdog in your mail reader made a bit of sense.
> Today, we have several new TLDs every week. Any attempt at making
> distributed _code_ keep up with such a rapidly moving target is simply
> doomed.
>
> If I read things right, the current version of the variable in the GIT
> repository lists 277 TLDs, not counting the two-letter ccTLDs. The
> current root zone (serial no. 2015111500) has 859 ... and counting.

Yeah, that doesn't seem all that useful.  

> At first thought, it may seem reasonable to replace the static list with
> a dynamic DNS lookup when the data is needed, but I would not recommend
> that, as it's far from certain that you're connected to the Internet
> when your submitting your message for delivery*. Doing the lookup will
> in that case lead to long and bothersome timeouts.
>
> Removing or totally opening it up are the only viable alternatives to
> me. Possibly, one could envisage retaining a generic syntax checker, but
> that is only practical if done _after_ any IDN** translation from u-label
> to a-label.

Well, since Emacs is also a web browser, it has to have a way to keep
track of where it's valid to set cookies.  (foo.co.uk isn't allowed to
set cookies for co.uk.)

So Emacs updates its version of 

https://publicsuffix.org/list/public_suffix_list.dat

for every release.  So Emacs should already know about all these domains
without maintaining a separate variable in message.el.

I think I'll just adjust the code to query that data instead.  That
should work, right?

(Although it would always be somewhat outdated.)

> ** I don't know if GNUS supports IDNs, but if not, it hopefully will in
> the future.

It does now.  I think support was added a few months back.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: Remove message-valid-fqdn-regexp?

[Lars Ingebrigtsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> https://publicsuffix.org/list/public_suffix_list.dat

Hm...  that doesn't really have all the domains, perhaps?  Only those
that have special cookie rules?

Then I guess the best solution is to just remove
message-valid-fqdn-regexp.  Anybody have opinions here?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: Remove message-valid-fqdn-regexp?

[Lars-Johan Liman]

larsi@gnus.org:
>> https://publicsuffix.org/list/public_suffix_list.dat

larsi@gnus.org:
> Hm...  that doesn't really have all the domains, perhaps?  Only those
> that have special cookie rules?

> Then I guess the best solution is to just remove
> message-valid-fqdn-regexp.  Anybody have opinions here?

Thanks. That would be appreciated.

				Cheers,
				  /Liman

Re: Remove message-valid-fqdn-regexp?

[Lars Ingebrigtsen]

Lars-Johan Liman <liman@netnod.se> writes:

>> Then I guess the best solution is to just remove
>> message-valid-fqdn-regexp.  Anybody have opinions here?
>
> Thanks. That would be appreciated.

I've now removed it on the Emacs trunk.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: Remove message-valid-fqdn-regexp?

[Lars-Johan Liman]

larsi@gnus.org:
> Lars-Johan Liman <liman@netnod.se> writes:

>>> Then I guess the best solution is to just remove
>>> message-valid-fqdn-regexp.  Anybody have opinions here?
>> 
>> Thanks. That would be appreciated.

> I've now removed it on the Emacs trunk.

Thanks again! :-)

				Cheers,
				  /Liman