Re: message-beginning-of-line behaves a bit strange

[Eric Abrahamsen <eric@ericabrahamsen.net>]

Emanuel Berg <moasenwood@zoho.eu> writes:

> Uwe Brauer wrote:
>
>>> Uwe Brauer <oub@mat.ucm.es> writes:
>>
>>> Try `message-beginning-of-header' instead. Alternately,
>>> fooling with the value of (variable)
>>> `message-beginning-of-line' while calling (function)
>>> `message-beginning-of-line'.
>>
>> Great, I did
>>
>> (defun my-add-subject-test ()
>>   (interactive)
>>   (message-goto-subject)
>> ;  (message-beginning-of-line)
>>   (message-beginning-of-header 'subject)
>>   (insert "[Test] "))
>>
>> And this work nicely for both, empty and non empty
>> subject lines!
>
> ?! Subject twice?
>
> What kind of code is that? I mean
>
>   (message-goto-subject)
>   (message-beginning-of-header 'subject)
>
> Much better:
>
> `message-goto-subject' could have an optional formal parameter
> BEG so when the argument is non-nil it does
> (message-beginning-of-header 'subject) last.
>
> But at the very least (message-beginning-of-header 'subject)
> should imply (message-goto-subject) first.
>
> Or one could make HANDLE-FOLDED optional and when nil, go to
> the current line's header beginning.
>
> Subject should appear one time!

As Robert points out, the 'subject argument is actually meaningless,
the argument's actual name is HANDLE-FOLDED.
`message-beginning-of-header' acts on whatever header point is on, it
doesn't differentiate between headers.