Emacs deal with it

After exporting my bookmarks from Firefox to an org-mode document I started to use Emacs as a bookmark manager. After a couple of days, I realized that I needed to automate some things. I didn't want to just insert new URLs to the org-mode document. I wanted to insert them as the org-mode links with informative descriptions (for example, the title of the web-page). Apparently, it is more complex than just CTRL+C and then CTRL+V the URL. I wanted the following:

  1. I copy a URL to the clipboard,
  2. switch to the Emacs window,
  3. press some magic keys,
  4. and Emacs inserts [[URL][title of the web-page]].

That would be awesome! So I implemented a function which can be bound to those magic keys:

(defun cliplink ()
  ;; Of course, this function is interactive. :)
  (interactive)
  (let (;; Remembering the current buffer, 'cause it is a destination
        ;; buffer we are inserting the org-link to.
        (dest-buffer (current-buffer))
        ;; Getting URL from the clipboard. Since it may contain
        ;; some text properties we are using substring-no-properties
        ;; function.
        (url (substring-no-properties (current-kill 0))))
    ;; Retrieving content by URL.
    (url-retrieve
     url
     ;; Performing an action on the retrieved content.
     `(lambda (s)
        (perform-cliplink ,dest-buffer ,url
                          (buffer-string))))))

As you can see, it's just a skeleton. The main work is done by perform-cliplink function:

(defun perform-cliplink (buffer url content)
  (let* (;; Decoding the content from UTF-8.
         (decoded-content (decode-coding-string content 'utf-8))
         ;; Extrating and preparing the title.
         (title (prepare-cliplink-title
                 (extract-title-from-html decoded-content))))
    ;; Inserting org-link.
    (with-current-buffer buffer
      (insert (format "[[%s][%s]]" url title)))))

Everything is clear except two functions: extract-title-from-html and prepare-cliplink-title. Let's have a look at the first one:

(defun extract-title-from-html (html)
  (let (;; Start index of the title.
        (start (string-match "<title>" html))
        ;; End index of the title.
        (end (string-match "</title>" html))
        ;; Amount of characters to skip the openning title tag.
        (chars-to-skip (length "<title>")))
    ;; If title is found ...
    (if (and start end (< start end))
        ;; ... extract it and return.
        (substring html (+ start chars-to-skip) end)
      nil)))

The second function is a bit more complex:

(defun prepare-cliplink-title (title)
  (let (;; Table of replacements which make this title usable for
        ;; org-link. Can be extended.
        (replace-table '(("\\[" . "{")
                         ("\\]" . "}")
                         ("&mdash;" . "—")))
        ;; Maximum length of the title.
        (max-length 77)
        ;; Removing redundant whitespaces from the title.
        (result (straight-string title)))
    ;; Applying every element of the replace-table.
    (dolist (x replace-table)
      (setq result (replace-regexp-in-string (car x) (cdr x) result)))
    ;; Cutting off the title according to its maximum length.
    (when (> (length result) max-length)
      (setq result (concat (substring result 0 max-length) "...")))
    ;; Returning result.
    result))

So, the last function is straight-string which helps us to remove redundant whitespace characters. Here is its implementation:

(defun straight-string (s)
  ;; Spliting the string and then concatenating it back.
  (mapconcat '(lambda (x) x) (split-string s) " "))

So much code! I hope it is not too hard to read. I just wanted to show my entire train of thoughts. I've put the complete code here.

UPD. Almost a year ago I decided to move this code out of my personal emacs config and create org-cliplink. Currently, it has much more features than the code in this post. Any bug reports and PRs are welcome.