[PATCH] emacs: display tags in notmuch-show's header-line with links to search

[Damien Cassou]

Hi

On Wed, Nov 7, 2012 at 12:14 AM, Mark Walters <markwalters1009 at gmail.com> wrote:
> This is not a full review: just a couple of thoughts. It basically seems
> to work as expected. I am not quite sure what behaviour you would expect
> in a couple of corner cases:
>
> 1) what it the user toggles the display of matching messages (elide
> mode)? Do you still want all tags from all messages including those not
> visible?
> 2) What about any tags from excluded messages? Should they show up? What
> if they would be excluded but aren't because of the particular search?


to me, a thread's tags should be independent of what is visible and
thus should be stable. For example, when I want to star a thread, I
star the message I'm currently seeing and I expect the thread to be
starred for the rest of its life, even if the particular mail is not
currently visible. What is the opinion of others?

>>  emacs/notmuch-tager.el |   76 +++++++++++++++++++++++++++
>
> I would go for tagger rather than tager (but others can disagree).


I don't care and can change it without problem. I picked 'tager'
because it's shorter and I like to have a lot of information in my
function names. What is the opinion of others?

>> +(defun notmuch-query-thread-tags-from-id (thread-id)
>> +  "Return the tags of thread whose id is THREAD-ID.
>> +The thread tags are the union of the tags of emails in the
>> +thread."
>> +  (let ((tag-lists
>> +      (notmuch-query-map-forest
>> +       (lambda (msg) (plist-get msg :tags))
>> +       (car (notmuch-query-get-threads
>> +             (list (concat "thread:" thread-id)))))))
>> +    (case (length tag-lists)
>> +      (0 nil)
>> +      (1 (car tag-lists))
>> +      (otherwise (reduce (lambda (l1 l2)
>> +                        (union l1 l2 :test 'string=))
>> +                      tag-lists)))))
>
> Couldn't you do this with notmuch-show-mapc and avoid the extra call to
> notmuch? It also probably helps some cases of excluded tags and elide
> mode.
>


That would not work for my definition of a thread's tags. But if we
change the definition, I can change the implementation of this
function and avoid a call to notmuch.


> If for some reason the query is better than I think you work
> to remove the thread: from the thread-id below and then add it back in
> here?


I know and it is on purpose :-). When I first had a look at notmuch
sources, I was confused by all those thread-id and mail-id everywhere
that are sometimes ids (e.g., "000012") and sometimes queries (e.g.,
"thread:000012"). To me, the code should use ids everywhere and build
a query out of that when calling notmuch or when displaying a query.
This would also avoid the use of the optional `bare' parameter in some
existing functions.


Thank you for your review. So, I'm waiting for more opinions because
changing anything.

Best

--
Damien Cassou
http://damiencassou.seasidehosting.st

"Success is the ability to go from one failure to another without
losing enthusiasm."
Winston Churchill