bug#29309: 25.2; Gnus: please make message-from-style 'angles the only option

bug#29309: 25.2; Gnus: please make message-from-style 'angles the only option

[Michael Stapelberg]

Package: emacs,gnus

Currently, the custom option message-from-style can be one of:

'parens, resulting in e.g. “stapelberg@debian.org (Michael Stapelberg)”

'angles, resulting in e.g. “Michael Stapelberg <stapelberg@debian.org>”

However, https://tools.ietf.org/html/rfc2822#section-3.4 (from 2001)
states:

>    Note: Some legacy implementations used the simple form where the
>    addr-spec appears without the angle brackets, but included the name
>    of the recipient in parentheses as a comment following the addr-spec.
>    Since the meaning of the information in a comment is unspecified,
>    implementations SHOULD use the full name-addr form of the mailbox,
>    instead of the legacy form, to specify the display name associated
>    with a mailbox.  Also, because some legacy implementations interpret
>    the comment, comments generally SHOULD NOT be used in address fields
>    to avoid confusing such implementations.

Let us please make Gnus a better citizen by following the RFC’s
recommendation and removing the legacy form.

Thanks,

-- 
Best regards,
Michael

bug#29309: 25.2; Gnus: please make message-from-style 'angles the only option

[Robert Pluim]

Michael Stapelberg <stapelberg@debian.org> writes:

> Package: emacs,gnus
>
> Currently, the custom option message-from-style can be one of:
>
> 'parens, resulting in e.g. “stapelberg@debian.org (Michael Stapelberg)”
>
> 'angles, resulting in e.g. “Michael Stapelberg <stapelberg@debian.org>”
>
> However, https://tools.ietf.org/html/rfc2822#section-3.4 (from 2001)
> states:
>
>>    Note: Some legacy implementations used the simple form where the
>>    addr-spec appears without the angle brackets, but included the name
>>    of the recipient in parentheses as a comment following the addr-spec.
>>    Since the meaning of the information in a comment is unspecified,
>>    implementations SHOULD use the full name-addr form of the mailbox,
>>    instead of the legacy form, to specify the display name associated
>>    with a mailbox.  Also, because some legacy implementations interpret
>>    the comment, comments generally SHOULD NOT be used in address fields
>>    to avoid confusing such implementations.
>
> Let us please make Gnus a better citizen by following the RFC’s
> recommendation and removing the legacy form.

I see no reason to remove the legacy form, unless you want to go to
the trouble of writing code to nag people to change their setup. You
might have an argument for changing the default to be 'angles though.

Robert

bug#29309: 25.2; Gnus: please make message-from-style 'angles the only option

[Michael Stapelberg]

On Wed, Nov 15, 2017 at 8:03 PM, Robert Pluim <rpluim@gmail.com> wrote:
> Michael Stapelberg <stapelberg@debian.org> writes:
>
>> Package: emacs,gnus
>>
>> Currently, the custom option message-from-style can be one of:
>>
>> 'parens, resulting in e.g. “stapelberg@debian.org (Michael Stapelberg)”
>>
>> 'angles, resulting in e.g. “Michael Stapelberg <stapelberg@debian.org>”
>>
>> However, https://tools.ietf.org/html/rfc2822#section-3.4 (from 2001)
>> states:
>>
>>>    Note: Some legacy implementations used the simple form where the
>>>    addr-spec appears without the angle brackets, but included the name
>>>    of the recipient in parentheses as a comment following the addr-spec.
>>>    Since the meaning of the information in a comment is unspecified,
>>>    implementations SHOULD use the full name-addr form of the mailbox,
>>>    instead of the legacy form, to specify the display name associated
>>>    with a mailbox.  Also, because some legacy implementations interpret
>>>    the comment, comments generally SHOULD NOT be used in address fields
>>>    to avoid confusing such implementations.
>>
>> Let us please make Gnus a better citizen by following the RFC’s
>> recommendation and removing the legacy form.
>
> I see no reason to remove the legacy form, unless you want to go to
> the trouble of writing code to nag people to change their setup. You
> might have an argument for changing the default to be 'angles though.

I don’t follow. Why would people need to change their setup? We could
just discard the option altogether and always use 'angles, couldn’t
we?


-- 
Best regards,
Michael

bug#29309: 25.2; Gnus: please make message-from-style 'angles the only option

[Robert Pluim]

Michael Stapelberg <stapelberg@debian.org> writes:

>>> Let us please make Gnus a better citizen by following the RFC’s
>>> recommendation and removing the legacy form.
>>
>> I see no reason to remove the legacy form, unless you want to go to
>> the trouble of writing code to nag people to change their setup. You
>> might have an argument for changing the default to be 'angles though.
>
> I don’t follow. Why would people need to change their setup? We could
> just discard the option altogether and always use 'angles, couldn’t
> we?

Ah, you want to remove the configuration variable entirely. I think
we'd have to obsolete it first, wait for a release cycle, and then
maybe remove it. Removing variables tends not to be fast for Emacs.

Robert

bug#29309: 25.2; Gnus: please make message-from-style 'angles the only option

[Michael Stapelberg]

On Wed, Nov 15, 2017 at 9:20 PM, Robert Pluim <rpluim@gmail.com> wrote:
> Michael Stapelberg <stapelberg@debian.org> writes:
>
>>>> Let us please make Gnus a better citizen by following the RFC’s
>>>> recommendation and removing the legacy form.
>>>
>>> I see no reason to remove the legacy form, unless you want to go to
>>> the trouble of writing code to nag people to change their setup. You
>>> might have an argument for changing the default to be 'angles though.
>>
>> I don’t follow. Why would people need to change their setup? We could
>> just discard the option altogether and always use 'angles, couldn’t
>> we?
>
> Ah, you want to remove the configuration variable entirely. I think
> we'd have to obsolete it first, wait for a release cycle, and then
> maybe remove it. Removing variables tends not to be fast for Emacs.

Could we make it ineffective before removing it entirely? I think
making it ineffective could be done right away.

-- 
Best regards,
Michael

bug#29309: 25.2; Gnus: please make message-from-style 'angles the only option

[Robert Pluim]

Michael Stapelberg <stapelberg@debian.org> writes:

> On Wed, Nov 15, 2017 at 9:20 PM, Robert Pluim <rpluim@gmail.com> wrote:
>> Ah, you want to remove the configuration variable entirely. I think
>> we'd have to obsolete it first, wait for a release cycle, and then
>> maybe remove it. Removing variables tends not to be fast for Emacs.
>
> Could we make it ineffective before removing it entirely? I think
> making it ineffective could be done right away.

We could, but then people might get annoyed at us changing behaviour
without warning, hence the obsolescence phase.

Robert

bug#29309: 25.2; Gnus: please make message-from-style 'angles the only option

[Lars Ingebrigtsen]

Michael Stapelberg <stapelberg@debian.org> writes:

>>    Note: Some legacy implementations used the simple form where the
>>    addr-spec appears without the angle brackets, but included the name
>>    of the recipient in parentheses as a comment following the addr-spec.
>>    Since the meaning of the information in a comment is unspecified,
>>    implementations SHOULD use the full name-addr form of the mailbox,
>>    instead of the legacy form, to specify the display name associated
>>    with a mailbox.

I agree.  The RFC says SHOULD, so Message shouldn't have an easy way to
generate From headers that deviate from the `angles' style.

So I propose to remove the code that reacts to `message-from-style'
(i.e., always generate `angles' From headers), and mark
`message-from-style' as obsolete and note in the doc string that it's no
longer used.

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

bug#29309: 25.2; Gnus: please make message-from-style 'angles the only option

[Lars Ingebrigtsen]

Eli didn't like the idea of removing the possibility of generating
invalid headers, but I flipped the default to `angles' and marked the
variable as obsolete.

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