Re: [PATCH 1/1] Use notmuch-extract-patch if available

[Kyle Meyer <kyle@kyleam.com>]

sourcehut@relevant-information.com writes (in cover letter):

> Instead of getting all the messages in the thread like my last patch I thought
> using notmuch-extract-patch would be cleaner. The code is adapted from
> elpa-mailscripts that is linked from the piem's documentation. It might be a
> good alternative to b4.

sourcehut@relevant-information.com writes (in commit message):

> notmuch-extract-patch is a command line tool from the elpa-mailscripts
> package that extracts patches from a thread. It's a useful way to
> extract the latest patch series from an email thread and filter out the
> replies and reviews.

It's been a little while since I've looked over mailscripts, but I think
it's got a lot of neat functionality (and in general am excited to see
work in this space).  I obviously decided to focus piem's patch
extraction functionality around b4, but I'm happy to consider some
support for notmuch-extract-patch if there are people that 1) want to
use notmuch-extract-patch and 2) for whatever reason, prefer to use piem
rather than mailscripts.el.

> diff --git a/piem-notmuch.el b/piem-notmuch.el
> index 8b2a353..05c03ab 100644
> --- a/piem-notmuch.el
> +++ b/piem-notmuch.el
> @@ -81,7 +81,9 @@ (defun piem-notmuch-am-ready-mbox ()
>    "Return a function that inserts an am-ready mbox.
>  If the buffer has any MIME parts that look like a patch, use
>  those parts' contents (in order) as the mbox.  Otherwise, use the
> -message itself if it looks like a patch."
> +message itself if it looks like a patch. If the executable
> +notmuch-extract-patch exists on the path, use that to get the
> +latest patch series from the notmuch thread."
>    (when (derived-mode-p 'notmuch-show-mode)
>      (let* ((handle (piem-notmuch--with-current-message
>                       (mm-dissect-buffer)))
> @@ -90,10 +92,25 @@ (defun piem-notmuch-am-ready-mbox ()
>        (if (= n-attachments 0)
>            (when (string-match-p piem-patch-subject-re
>                                  (notmuch-show-get-subject))
> -            (let ((id (notmuch-show-get-message-id)))
> +            (let ((id (notmuch-show-get-message-id))
> +                  (thread-id notmuch-show-thread-id))
>                (lambda ()
> -                (call-process notmuch-command nil t nil
> -                              "show" "--format=mbox" id))))
> +                (if-let ((cmd (executable-find "notmuch-extract-patch"))
> +                         (tid
> +                          ;; If `notmuch-show' was called with a notmuch query rather
> +                          ;; than a thread ID, as `org-notmuch-follow-link' in
> +                          ;; org-notmuch.el does, then `notmuch-show-thread-id' might
> +                          ;; be an arbitrary notmuch query instead of a thread ID.  We
> +                          ;; need to wrap such a query in thread:{} before passing it
> +                          ;; to notmuch-extract-patch(1), or we might not get a whole
> +                          ;; thread extracted (e.g. if the query is just id:foo)
> +                          (if (string= (substring thread-id 0 7) "thread:")
> +                              thread-id
> +                            (concat "thread:{" thread-id "}"))))

I believe this comment and (if ...) is coming straight from
mailscripts.el.  You note in the cover letter that some of this is
copied from mailscripts.el, but I think there should be a code comment
as well.

However, I don't really understand why the condition is necessary.
Wouldn't always using

  thread:{notmuch-show-get-message-id return value}

work?

> +                    (call-process cmd nil t nil
> +                                  tid)
> +                  (call-process notmuch-command nil t nil
> +                                "show" "--format=mbox" id)))))

Rather than add this functionality to piem-notmuch-am-ready-mbox, I'd
prefer to add a dedicated function for notmuch-extract-patch.  Then this
function can be added instead of (or even just ahead of)
piem-notmuch-am-ready-mbox in piem-am-ready-mbox-functions.  Whether
that function is added to piem-am-ready-mbox-functions could be
controlled by a user option.