Re: [NonGNU Elpa] New package: ETT

[John Task]

Thank you very much for the comments.

> I would track the ignored files in a .elpaignore file that would look
> something like this

OK.

> -(defcustom ett-file (concat user-emacs-directory "tracks.ett")
> +(defcustom ett-file (locate-user-emacs-file "tracks.ett")
>    "File for storing ETT tracks."
>    :type '(file))

OK.
 
> -(defcustom ett-graph-icon "│"
> +(defcustom ett-graph-icon ?│
>    "Icon used for building ETT graphs.
>  It must be a single char string."
> -  :type '(string))
> +  :type 'character)
>  
>  (defface ett-heading
>    '((((class color) (min-colors 88) (background light))
> @@ -170,7 +170,7 @@ It must be a single char string."
>          (list "^\\(Less time .*\\): " '(1 'underline))
>          (list "^\\(Tags report\\):$" '(1 'underline))
>          (list "^\\(Pending goals\\):$" '(1 'underline))
> -        (list (concat ett-graph-icon "+") '(0 'ett-separator-face)))
> +        (list (concat (string ett-graph-icon) "+") '(0 'ett-separator-face)))
>    "Keywords for syntax highlighting on ETT report buffer.")

I should have thought about that myself, I forgot the string function
:)
 
>  (defvar ett-items-internal nil)
> @@ -269,41 +269,36 @@ Here are some translations (*starred* values are 
> current at the time):
>  3-2      ---> 03-02-*23*
>  03       ---> 03-*02*-*23*
>  3        ---> 03-*02*-*23*"
> -  (pcase (length date)
> -    ;; DD-MM-YY
> -    (8 date)
> -    ;; DD-MM
> -    (5 (setq date (concat date (format-time-string "-%y"))))
> -    ;; DD
> -    (2 (setq date (concat date (format-time-string "-%m-%y"))))
> -    ;; D
> -    (1 (setq date (concat "0" date (format-time-string "-%m-%y"))))
> -    ;; DD-M
> -    (4
> -     (setq date
> -           ;; Or D-MM?
> -           (pcase (string-search "0" date)
> -             (2 (concat "0" date (format-time-string "-%y")))
> -             (_ (concat
> +  (setq date
> +     (pcase (length date)
> +       ;; DD-MM-YY
> +       (8 date)
> +       ;; DD-MM
> +       (5 (concat date (format-time-string "-%y")))
> +       ;; DD
> +       (2 (concat date (format-time-string "-%m-%y")))
> +       ;; D
> +       (1 (concat "0" date (format-time-string "-%m-%y")))
> +       ;; DD-M
> +       (4
> +        ;; Or D-MM?
> +        (pcase (string-search "0" date)
> +          (2 (concat "0" date (format-time-string "-%y")))
> +          (_ (concat
>                   (replace-regexp-in-string "-\\(.*\\)" "-0\\1" date)
> -                 (format-time-string "-%y"))))))
> -    ;; D-M
> -    (3 (setq date
> -             (concat "0" (replace-regexp-in-string "-\\(.*\\)" "-0\\1" date)
> -                     (format-time-string "-%y"))))
> -    ;; D-MM-YY
> -    (7 (setq date
> -             ;; Or DD-M-YY?
> -             (pcase (string-search "0" date)
> -               (2 (concat "0" date))
> -               (_ (concat
> -                   (replace-regexp-in-string "-\\(.*\\)-" "-0\\1-" date))))))
> -    ;; D-M-YY (who writes like that?)
> -    (6 (setq date
> -             (concat "0"
> -                     (replace-regexp-in-string "-\\(.*\\)-" "-0\\1-" 
> date)))))
> -  ;; Return date
> -  date)
> +                 (format-time-string "-%y")))))
> +       ;; D-M
> +       (3 (concat "0" (replace-regexp-in-string "-\\(.*\\)" "-0\\1" date)
> +                  (format-time-string "-%y")))
> +       ;; D-MM-YY
> +       (7 ;; Or DD-M-YY?
> +        (pcase (string-search "0" date)
> +          (2 (concat "0" date))
> +          (_ (concat
> +              (replace-regexp-in-string "-\\(.*\\)-" "-0\\1-" date)))))
> +       ;; D-M-YY (who writes like that?)
> +       (6 (concat "0"
> +                  (replace-regexp-in-string "-\\(.*\\)-" "-0\\1-" date))))))

This is indeed better, thanks.

>  (defun ett-do-division (num1 num2)
> -  "Divide NUM1 between NUM2."
> +  "Divide NUM1 between NUM2."                ;can you explain why this is 
> needed?
>    (if (/= (% num1 num2) 0)
>        (/ (float num1) num2)
>      (/ num1 num2)))

I need the extra precision for various functions.  (/ 12 7) gives me
1, while (ett-do-division 12 7) gives me 1.7142857142857142.

>            (insert
>             (mapconcat #'concat (make-list (- col (current-column)) " ") "")))
>          (ett-insert-rectangle
> -         (make-list graph-f ett-graph-icon)))
> +         (make-list graph-f (string ett-graph-icon))))
>        (setq list (cdr list)

OK.

>  (defun ett-prettify-time (time)
>    "Return a prettified string for TIME."
>    (let ((hr 0))
> -    (while (>= time 60)
> +    (while (>= time 60)                      ;the loop here shouldn't be 
> necessary, you can calculate the same thing using remainder and floor
>        (setq hr (1+ hr)
>              time (- time 60)))

I'll look into that.  Is there any reason why I would need to avoid
that loop?  The code just works for now.

>      (save-excursion
>        ;; Sort items
>        ;; We need to do it in two steps because it's somehow complicated
> -      (sort-regexp-fields t "^.*$" "\\(:[0-9]+\\)" (point-min) (point-max))
> +      (sort-regexp-fields t "^.*$" "\\(:[0-9]+\\)" (point-min) (point-max)) 
> ;have you considered using rx?
>        (sort-regexp-fields t "^.*$" "\\([0-9]+:\\)" (point-min) (point-max))
>        ;; We now align
>        (align-regexp

I have considered it, but while my regexps are weird, they are
still readable for now.  Maybe I'll reconsider it.

> +(defvar ett-view-mode-map
> +  (let ((map (make-sparse-keymap)))
> +    (define-key map (kbd "C-c C-c") #'ett-add-track)
> +    (define-key map (kbd "SPC") #'ett-space-dwim)
> +    map))

I think you mean just ett-mode-map.  In that case, OK.

> (define-derived-mode ett-mode text-mode "ETT"
>    "Major mode for editing ETT based texts."
>    (setq font-lock-defaults '(ett-font-lock-keywords t))
> -  (setq-local outline-regexp "[0-9]+-[0-9]+-[0-9]+")
> +  (setq-local outline-regexp "[0-9]+-[0-9]+-[0-9]+") ;does this need to 
> start with a ^
>    ;; show-paren-mode can be somehow annoying here
>    (show-paren-local-mode -1))

Documentation of the variable says it isn't necessary, and it worked
as is on my tests, but I'm not an expert.

> +(defvar ett-view-mode-map
> +  (let ((map (make-sparse-keymap)))
> +    (define-key map (kbd "q") #'bury-buffer)
> +    (define-key map (kbd "SPC") #'ett-choose-view)
> +    (define-key map (kbd "<up>") #'scroll-down-command)
> +    (define-key map (kbd "<down>") #'scroll-up-command)
> +    map))

OK.

> -(define-key ett-view-mode-map (kbd "q") #'bury-buffer)
> -(define-key ett-view-mode-map (kbd "SPC") #'ett-choose-view)
> -(define-key ett-view-mode-map (kbd "<up>") #'scroll-down-command)
> -(define-key ett-view-mode-map (kbd "<down>") #'scroll-up-command)
> -(define-key ett-mode-map (kbd "C-c C-c") #'ett-add-track)
> -(define-key ett-mode-map (kbd "SPC") #'ett-space-dwim)
> -
>  ;;;###autoload
> -(add-to-list 'auto-mode-alist (cons (purecopy "\\.ett\\'") 'ett-mode))
> +(add-to-list 'auto-mode-alist '("\\.ett\\'" . ett-mode))

OK.

> Oh an by the way, it seems possible to reduce the minimum version of
> Emacs to at least 26 using Compat
> (https://elpa.gnu.org/packages/compat.html).  Would you be interested in
> that?

Yes, I plan to work on backward compatibility once the code is well
tested.  You know, better to see if it works on a version before
porting it to other ones :)

Thanks again.  Best regards.