howard/hamacs
howard/hamacs
1
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Reformat with dash library to make code more readable

Browse source 
Using the anaphoric macros from the dash library allows the banner
code to be smaller and I feel more readable. Yes, I could have run the
`ls` command into another buffer and copied it, I have bigger plans
for displaying my directory listing.
This commit is contained in:
Howard Abrams 2022-09-29 09:18:59 -07:00
parent d5e30ec98f
commit b399c30bb6
1 changed files with 18 additions and 20 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

38
ha-eshell.org
View file

@ -563,7 +563,7 @@ The [[http://projects.ryuslash.org/eshell-fringe-status/][eshell-fringe-status]]
:hook (eshell-mode . eshell-fringe-status-mode)) :hook (eshell-mode . eshell-fringe-status-mode))
#+end_src #+end_src
** Opening Banner ** Opening Banner
Whenever I open a shell, I instinctively type =ls= … so why not do that automatically? Whenever I open a shell, I instinctively type =ls= … so why not do that automatically? The [[elisp:(describe-variable 'eshell-banner-message)][eshell-banner-message]] variable, while normally a string, can be a /form/ (an s-expression) that calls a function, so I made a customized =ls= that can be attractive:
#+begin_src emacs-lisp #+begin_src emacs-lisp
(defun ha-eshell-banner () (defun ha-eshell-banner ()
"Return a string containing the files in the current directory." "Return a string containing the files in the current directory."
@ -573,34 +573,32 @@ Whenever I open a shell, I instinctively type =ls= … so why not do that automa
(not "~") (not "~")
string-end)) string-end))
(files (directory-files default-directory nil non-hidden)) (files (directory-files default-directory nil non-hidden))
(longest (reduce (lambda (longest file) (max longest (length file))) (longest (--reduce-from (max acc (length it)) 1 files))
files :initial-value 1))
(padded (format "%%-%ds " longest)) (padded (format "%%-%ds " longest))
(width (window-total-width)) (width (window-total-width))
(columns (/ width (+ longest 3)))) (columns (/ width (+ longest 3))))
(cl-flet* ((process-file (cl-flet* ((process-file
(file) (file)
(let ((image-rx (rx "." (or "png" "jpg" "jpeg" "tif" "wav") string-end)) (let ((impor-rx (rx string-start "README"))
(image-rx (rx "." (or "png" "jpg" "jpeg" "tif" "wav") string-end))
(code-rx (rx "." (or "el" "py" "rb") string-end)) (code-rx (rx "." (or "el" "py" "rb") string-end))
(docs-rx (rx "." (or "org" "md") string-end))) (docs-rx (rx "." (or "org" "md") string-end)))
(format padded (cond (format padded (cond
((string-match image-rx file) (propertize file 'face '(:foreground "light pink"))) ((string-match impor-rx file)
((string-match code-rx file) (propertize file 'face '(:foreground "DarkSeaGreen1"))) (propertize file 'face '(:foreground "gold")))
((file-directory-p file) (propertize file 'face '(:foreground "SteelBlue"))) ((string-match image-rx file)
(t file))))) (propertize file 'face '(:foreground "light pink")))
;; This nasty little function was an attempt to make ((string-match code-rx file)
;; things readable, but who would have thunk that (propertize file 'face '(:foreground "DarkSeaGreen1")))
;; `mapconcat' couldn't access a symbol reference to a ((file-directory-p file)
;; function created by `cl-flet*'!? (propertize file 'face 'eshell-ls-directory))
(process-files (t
(table) file)))))
(mapconcat (lambda (line) (mapconcat
(lambda (file) (process-file file)) (process-lines (files) (s-join "• " (--map (process-file it) files)))
line
"• ")) (process-files (table) (s-join "\n" (--map (process-lines it) table))))
table
"\n")))
(concat (process-files (seq-partition files columns)) "\n\n")))) (concat (process-files (seq-partition files columns)) "\n\n"))))
#+end_src #+end_src