[PATCH 4/4] emacs: Simplify and fix `notmuch-mua-prompt-for-sender'

[Austin Clements]

Quoth Mark Walters on Feb 20 at 10:36 pm:
> 
> Patches 1-3 look fine +1.
> 
> This one I am less sure about. I agree with the principle but for my use
> case it is a little annoying:
> 
> I only use one name for all my addresses (Mark Walters), some addresses
> are mark at .. and some walters at ..
> 
> ido-completing-read is definitely less nice to use when all the addresses
> match mark and walters.

Ah, interesting.  I'd figured ido would make this change *less*
jarring because you could enter just the email substring.

> I wonder if we could get the old behaviour in a more robust fashion. Two
> possibilities we could consider are
> 
> 1) if getting the information from the config file (when there is
> necessarily a single name) then only complete the addresses.

This would be easy enough.  I'd chosen not to do this partly because I
thought the user may want to enter an address in "Name <email>" form
to override the name, especially if they only have one configured
identity (and since Emacs somehow lacks a address parser that can
extract the name without also transforming it to death, this is hard
to detect).  But I'm happy to ignore that situation.

> 2) make notmuch-identities a list of cons cells (name . email). Then
> there is no parsing and the old method could be robust.
> 
> OTOH I can get used to the change.
> 
> Best wishes
> 
> Mark
> 
> 
> 
> On Thu, 20 Feb 2014, Austin Clements <amdragon at MIT.EDU> wrote:
> > `notmuch-mua-prompt-for-sender' is over-engineered and often wrong.
> > It attempts to detect when all identities have the same name and
> > specialize the prompt to just the email address part, but this has
> > several problems.  First, it uses `mail-extract-address-components',
> > which is meant for displaying email addresses, not general-purpose
> > parsing, and hence performs many canonicalizations that can interfere
> > with this use.  For example, configuring notmuch-identities to
> > ("Austin <austin at example.com>"), will cause
> > `notmuch-mua-prompt-for-sender' to lose the name part entirely and
> > return " <austin at example.com>".  Second, though less serious, the
> > prompt specialization means the user can't enter a different name like
> > they can if their identities have different names.
> >
> > This patch rewrites `notmuch-mua-prompt-for-sender' to simply prompt
> > for a full identity, where the list of identities is derived from
> > either notmuch-identities or the user's Notmuch configuration.
> >
> > The original code also did several strange things, like using `eval'
> > and specifying that this function was interactive.  As a side-effect,
> > this patch fixes these problems.  And it adds a docstring.
> > ---
> >  emacs/notmuch-mua.el | 33 +++++++++------------------------
> >  1 file changed, 9 insertions(+), 24 deletions(-)
> >
> > diff --git a/emacs/notmuch-mua.el b/emacs/notmuch-mua.el
> > index f2df770..4a485a4 100644
> > --- a/emacs/notmuch-mua.el
> > +++ b/emacs/notmuch-mua.el
> > @@ -286,30 +286,15 @@ the From: header is already filled in by notmuch."
> >        (ad-activate 'ido-completing-read)))
> >  
> >  (defun notmuch-mua-prompt-for-sender ()
> > -  (interactive)
> > -  (let (name addresses one-name-only)
> > -    ;; If notmuch-identities is non-nil, check if there is a fixed user name.
> > -    (if notmuch-identities
> > -	(let ((components (mapcar 'mail-extract-address-components notmuch-identities)))
> > -	  (setq name          (caar components)
> > -		addresses     (mapcar 'cadr components)
> > -		one-name-only (eval
> > -			       (cons 'and
> > -				     (mapcar (lambda (identity)
> > -					       (string-equal name (car identity)))
> > -					     components)))))
> > -      ;; If notmuch-identities is nil, use values from the notmuch configuration file.
> > -      (setq name          (notmuch-user-name)
> > -	    addresses     (cons (notmuch-user-primary-email) (notmuch-user-other-email))
> > -	    one-name-only t))
> > -    ;; Now prompt the user, either for an email address only or for a full identity.
> > -    (if one-name-only
> > -	(let ((address
> > -	       (ido-completing-read (concat "Sender address for " name ": ") addresses
> > -				    nil nil nil 'notmuch-mua-sender-history (car addresses))))
> > -	  (concat name " <" address ">"))
> > -      (ido-completing-read "Send mail From: " notmuch-identities
> > -			   nil nil nil 'notmuch-mua-sender-history (car notmuch-identities)))))
> > +  "Prompt for a sender from the user's configured identities."
> > +  (let ((identities (or notmuch-identities
> > +			(let ((name (notmuch-user-name)))
> > +			  (mapcar (lambda (addr) (concat name " <" addr ">"))
> > +				  (cons (notmuch-user-primary-email)
> > +					(notmuch-user-other-email)))))))
> > +    (ido-completing-read "Send mail from: " identities
> > +			 nil nil nil 'notmuch-mua-sender-history
> > +			 (car identities))))
> >  
> >  (put 'notmuch-mua-new-mail 'notmuch-prefix-doc "... and prompt for sender")
> >  (defun notmuch-mua-new-mail (&optional prompt-for-sender)
> >
> > _______________________________________________
> > notmuch mailing list
> > notmuch at notmuchmail.org
> > http://notmuchmail.org/mailman/listinfo/notmuch

-- 
Austin Clements                                      MIT/'06/PhD/CSAIL
amdragon at mit.edu                           http://web.mit.edu/amdragon
       Somewhere in the dream we call reality you will find me,
              searching for the reality we call dreams.