From: Michal Sojka <sojkam1@fel.cvut.cz>
To: Mark Walters <markwalters1009@gmail.com>, notmuch@notmuchmail.org
Subject: Re: [PATCH v2] emacs: address completion,
 allow sender/recipient and filters
In-Reply-To: <1463153337-11381-1-git-send-email-markwalters1009@gmail.com>
References: <874mf2yeq7.fsf@steelpick.2x.cz>
 <1463153337-11381-1-git-send-email-markwalters1009@gmail.com>
User-Agent: Notmuch/0.22 (http://notmuchmail.org) Emacs/24.5.1
 (x86_64-pc-linux-gnu)
Date: Fri, 13 May 2016 23:41:58 +0200
Message-ID: <87h9e11vq1.fsf@steelpick.2x.cz>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: list

Hi Mark,

On Fri, May 13 2016, Mark Walters wrote:
> This commit lets the user customize the address completion.
>
> The first change controls whether to build the address completion list
> based on messages you have sent or you have received (the latter is
> much faster).
>
> The second change add a possible filter query to limit the messages
> used -- for example, setting this to date:1y..  would limit the
> address completions to addresses used in the last year. This speeds up
> the address harvest and may also make the search less cluttered as old
> addresses may well no longer be valid.
> ---
>
> I have fixed the bug pointed out by Michal in his review. It seems to
> work, but there are a lot of possible configurations,
>
> I also don't like my docstring for the notmuch-address-command
> defcustom, so any suggestions gratefully received.
>
> Also, I am not sure whether this is all to much complexity for this
> feature.

A big part of the patch is moving variable definitions and I'd say the
rest is OK (complexity-wise). Other comments below.

>  emacs/notmuch-address.el | 104 +++++++++++++++++++++++++++++++----------------
>  emacs/notmuch-company.el |   2 +-
>  2 files changed, 71 insertions(+), 35 deletions(-)
>
> diff --git a/emacs/notmuch-address.el b/emacs/notmuch-address.el
> index aafbe5f..8b84a4c 100644
> --- a/emacs/notmuch-address.el
> +++ b/emacs/notmuch-address.el
> @@ -28,15 +28,50 @@
>  ;;
>  (declare-function company-manual-begin "company")
>
> -(defcustom notmuch-address-command 'internal
> -  "The command which generates possible addresses. It must take a
> -single argument and output a list of possible matches, one per
> -line. The default value of `internal' uses built-in address
> -completion."
> +(defvar notmuch-address-last-harvest 0
> +  "Time of last address harvest")
> +
> +(defvar notmuch-address-completions (make-hash-table :test 'equal)
> +  "Hash of email addresses for completion during email composition.
> +  This variable is set by calling `notmuch-address-harvest'.")
> +
> +(defvar notmuch-address-full-harvest-finished nil
> +  "t indicates that full completion address harvesting has been
> +finished")
> +
> +(defcustom notmuch-address-command '(sent nil)
> +  "The command which generates possible addresses.
> +
> +It can be a (non-nil) list, in which case internal completion is
> +used; in this case the first list item 'sent/'received specifies
> +whether you match message sent by the user or received by the
> +user (note received by is much faster), and the second list item
> +should be nil or a filter-string, such as \"date:1y..\" to append
> +to the query.
> +
> +If this variable is nil then address completion is disabled.
> +
> +If it is a string then that string should be an external program
> +which must take a single argument and output a list of possible
> +matches, one per line."

What about the following? (I think that the string value should be
described first because it corresponds to the name of the variable)

     "Determines how to generate address completion candidates.

   If it is a string then that string should be an external program
   which must take a single argument (searched string) and output a list
   of completion candidates, one per line.

   Alternatively, it can be a (non-nil) list, in which case internal
   completion is used; in this case the list should have form
   '(DIRECTION FILTER), where DIRECTION is either sent or received and
   specifies whether the candidates are searched in messages sent by the
   user or received by the user (note received by is much faster), and
   FILTER is either nil or a filter-string, such as \"date:1y..\" to
   append to the query.

   If this variable is nil then address completion is disabled."

>    :type '(radio
> -	  (const :tag "Use internal address completion" internal)
> +	  (list :tag "Use internal address completion"
> +		(radio
> +		 :tag "Build list based on messages you have"

What about "Base completion on messages you have"?

> +		 :value sent
> +		 (const :tag "sent" sent)
> +		 (const :tag "received" received))
> +		(radio :tag "Filter messages used for completion"
> +		       (const :tag "Use all messages" nil)
> +		       (string :tag "Filter query")))
>  	  (const :tag "Disable address completion" nil)
> -	  (string :tag "Use external completion command" "notmuch-addresses"))
> +	  (string :tag "Use external completion command"))
> +  ;; We override set so that we can clear the cache when this changes
> +  :set (lambda (symbol value)
> +	 (set-default symbol value)
> +	 (setq notmuch-address-last-harvest 0)
> +	 (setq notmuch-address-completions (clrhash notmuch-address-completions))
> +	 (setq notmuch-address-full-harvest-finished nil))
>    :group 'notmuch-send
>    :group 'notmuch-external)
>
> @@ -51,17 +86,6 @@ to know how address selection is made by default."
>    :group 'notmuch-send
>    :group 'notmuch-external)
>
> -(defvar notmuch-address-last-harvest 0
> -  "Time of last address harvest")
> -
> -(defvar notmuch-address-completions (make-hash-table :test 'equal)
> -  "Hash of email addresses for completion during email composition.
> -  This variable is set by calling `notmuch-address-harvest'.")
> -
> -(defvar notmuch-address-full-harvest-finished nil
> -  "t indicates that full completion address harvesting has been
> -finished")
> -
>  (defun notmuch-address-selection-function (prompt collection initial-input)
>    "Call (`completing-read'
>        PROMPT COLLECTION nil nil INITIAL-INPUT 'notmuch-address-history)"
> @@ -83,7 +107,8 @@ finished")
>
>  (defun notmuch-address-setup ()
>    (let* ((use-company (and notmuch-address-use-company
> -			   (eq notmuch-address-command 'internal)
> +			   notmuch-address-command
> +			   (listp notmuch-address-command)
>  			   (require 'company nil t)))
>  	 (pair (cons notmuch-address-completion-headers-regexp
>  		     (if use-company
> @@ -111,11 +136,11 @@ The candidates are taken from `notmuch-address-completions'."
>  elisp-based implementation or older implementation requiring
>  external commands."
>    (cond
> -   ((eq notmuch-address-command 'internal)
> +   ((and notmuch-address-command (listp notmuch-address-command))
>      (when (not notmuch-address-full-harvest-finished)
>        ;; First, run quick synchronous harvest based on what the user
>        ;; entered so far
> -      (notmuch-address-harvest (format "to:%s*" original) t))
> +      (notmuch-address-harvest original t))
>      (prog1 (notmuch-address-matching original)
>        ;; Then start the (potentially long-running) full asynchronous harvest if necessary
>        (notmuch-address-harvest-trigger)))
> @@ -191,21 +216,32 @@ external commands."
>
>  The car is a partial harvest, and the cdr is a full harvest")
>
> -(defun notmuch-address-harvest (&optional filter-query synchronous callback)
> +(defun notmuch-address-harvest (&optional filter-string synchronous callback)
>    "Collect addresses completion candidates. It queries the
> -notmuch database for all messages sent by the user optionally
> -matching FILTER-QUERY (if not nil). It collects the destination
> -addresses from those messages and stores them in
> -`notmuch-address-completions'. Address harvesting may take some
> -time so the address collection runs asynchronously unless
> -SYNCHRONOUS is t. In case of asynchronous execution, CALLBACK is
> -called when harvesting finishes."
> -  (let* ((from-me-query (mapconcat (lambda (x) (concat "from:" x)) (notmuch-user-emails) " or "))
> -	 (query (if filter-query
> -		    (format "(%s) and (%s)" from-me-query filter-query)
> -		  from-me-query))
> +notmuch database for messages sent/received by the user
> +optionally with to/from matching FILTER-STRING (if not nil). It
> +collects the destination addresses from those messages and stores
> +them in `notmuch-address-completions'. Address harvesting may
> +take some time so the address collection runs asynchronously
> +unless SYNCHRONOUS is t. In case of asynchronous execution,
> +CALLBACK is called when harvesting finishes."

I'm a bit lost in the names of variables here so I suggest more
descriptive names below. Also the doc string could be IMHO better.

(defun notmuch-address-harvest (&optional addr-prefix synchronous callback)
      "Collect addresses completion candidates.

    It queries the notmuch database for messages sent/received (as
    configured with `notmuch-address-command`) by the user, collects
    destination/source addresses from those messages and stores them in
    `notmuch-address-completions'.

    If ADDR-PREFIX is not nil, only messages with to/from addresses
    matching ADDR-PREFIX*' are queried.

    Address harvesting may take some time so the address collection runs
    asynchronously unless SYNCHRONOUS is t. In case of asynchronous
    execution, CALLBACK is called when harvesting finishes."

> +  (let* ((sent (eq (car notmuch-address-command) 'sent))
> +	 (user-query (cadr notmuch-address-command))

config-query

> +	 (filter-query (when filter-string
> +			 (format "%s:%s*" (if sent "to" "from") filter-string)))

prefix-query

> +	 (from-or-to-me-query
> +	  (mapconcat (lambda (x)
> +		       (concat (if sent "from:" "to:") x))
> +		     (notmuch-user-emails) " or "))
> +	 (query (if (or filter-query user-query)
> +		    (concat (format "(%s)" from-or-to-me-query)
> +			    (when filter-query
> +			      (format " and (%s)" filter-query))
> +			    (when user-query
> +			      (format " and (%s)" user-query)))
> +		  from-or-to-me-query))
>  	 (args `("address" "--format=sexp" "--format-version=2"
> -		 "--output=recipients"
> +		 ,(if sent "--output=recipients" "--output=sender")
>  		 "--deduplicate=address"
>  		 ,query)))
>      (if synchronous
> diff --git a/emacs/notmuch-company.el b/emacs/notmuch-company.el
> index b881d6d..dcb59cd 100644
> --- a/emacs/notmuch-company.el
> +++ b/emacs/notmuch-company.el
> @@ -72,7 +72,7 @@
>  			  (lambda (callback)
>  			    ;; First run quick asynchronous harvest based on what the user entered so far
>  			    (notmuch-address-harvest
> -			     (format "to:%s*" arg) nil
> +			     arg nil
>  			     (lambda (_proc _event)
>  			       (funcall callback (notmuch-address-matching arg))
>  			       ;; Then start the (potentially long-running) full asynchronous harvest if necessary
> --
> 2.1.4

Cheers
-Michal