Re: master 40db29f: * lisp/mail/sendmail.el (mail-from-style):

Re: master 40db29f: * lisp/mail/sendmail.el (mail-from-style):

[Robert Pluim]

rgm@gnu.org (Glenn Morris) writes:

>  (make-obsolete-variable
>   'mail-from-style
> - "Only the `angles' value is valid according to RFC2822" "27.1")
> + "only the `angles' value is valid according to RFC2822." "27.1" 'set)

This raises two questions. From 'emacs -Q', C-h v mail-from-style
gives:

    mail-from-style is a variable defined in ‘sendmail.el’.
    Its value is ‘default’
    Original value was angles

Which implies that the value has been changed even though it hasnʼt.

Also, why does describe-variable only surround the
original value of the variable in quotes?

Robert

Re: master 40db29f: * lisp/mail/sendmail.el (mail-from-style):

[Glenn Morris]

Robert Pluim wrote:

> This raises two questions. 

Neither question seems related to my change, AFAICS?

> From 'emacs -Q', C-h v mail-from-style gives:
>
>     mail-from-style is a variable defined in 'sendmail.el'.
>     Its value is 'default'
>     Original value was angles

I can't reproduce this.

> Also, why does describe-variable only surround the
> original value of the variable in quotes?

(I guess you mean, why does it not quote the original value?)
Perhaps an oversight in 36311ec?

Re: master 40db29f: * lisp/mail/sendmail.el (mail-from-style):

[Paul Eggert]

On 07/03/2018 08:49 AM, Glenn Morris wrote:
> Perhaps an oversight in 36311ec?

Hmm, well, even before 36311ec the old value was printed differently 
from the new value. That is, in 36311ec I was hesitant to redo quoting 
of old-values since they weren't quoted correctly via the old rules 
either. It sounds like this should be rethought, as I don't see why 
describe-variable should display old values in a different style from 
new ones.

Re: master 40db29f: * lisp/mail/sendmail.el (mail-from-style):

[Robert Pluim]

Glenn Morris <rgm@gnu.org> writes:

> Robert Pluim wrote:
>
>> This raises two questions. 
>
> Neither question seems related to my change, AFAICS?
>
>> From 'emacs -Q', C-h v mail-from-style gives:
>>
>>     mail-from-style is a variable defined in 'sendmail.el'.
>>     Its value is 'default'
>>     Original value was angles
>
> I can't reproduce this.

Neither can I, after a 'make bootstrap'. Does the build cache custom
values somewhere?

>> Also, why does describe-variable only surround the
>> original value of the variable in quotes?
>
> (I guess you mean, why does it not quote the original value?)

Yes.

> Perhaps an oversight in 36311ec?

It was like that before 36311ec.

Regards

Robert

Re: master 40db29f: * lisp/mail/sendmail.el (mail-from-style):

[Andy Moreton]

On Wed 04 Jul 2018, Robert Pluim wrote:

> Glenn Morris <rgm@gnu.org> writes:
>
>> Robert Pluim wrote:
>>
>>> This raises two questions. 
>>
>> Neither question seems related to my change, AFAICS?
>>
>>> From 'emacs -Q', C-h v mail-from-style gives:
>>>
>>>     mail-from-style is a variable defined in 'sendmail.el'.
>>>     Its value is 'default'
>>>     Original value was angles
>>
>> I can't reproduce this.
>
> Neither can I, after a 'make bootstrap'. Does the build cache custom
> values somewhere?

mail-from-style is an autoload, so ends up being defined in loaddefs.el.
It is possible that the default value in sendmail.el was different until
loaddefs.el was updated.

    AndyM

Re: master 40db29f: * lisp/mail/sendmail.el (mail-from-style):

[Robert Pluim]

Andy Moreton <andrewjmoreton@gmail.com> writes:

> On Wed 04 Jul 2018, Robert Pluim wrote:
>
>> Glenn Morris <rgm@gnu.org> writes:
>>
>>> Robert Pluim wrote:
>>>
>>>> This raises two questions. 
>>>
>>> Neither question seems related to my change, AFAICS?
>>>
>>>> From 'emacs -Q', C-h v mail-from-style gives:
>>>>
>>>>     mail-from-style is a variable defined in 'sendmail.el'.
>>>>     Its value is 'default'
>>>>     Original value was angles
>>>
>>> I can't reproduce this.
>>
>> Neither can I, after a 'make bootstrap'. Does the build cache custom
>> values somewhere?
>
> mail-from-style is an autoload, so ends up being defined in loaddefs.el.
> It is possible that the default value in sendmail.el was different until
> loaddefs.el was updated.

Yes, the default value changed a while ago. Looks like loaddefs.el
doesnʼt get updated in that case.

Robert