Configuring Emacs - 7 months later - Act 3

Note: This is the final post in a series of 3 posts about my experience configuring Emacs over 7 months, starting from scratch.

• Post 1: Fundamentals
• Post 2: Keybindings and Package Integration
• Post 3: Writing Emacs Lisp Functions

Emacs Lisp is the language that is used to configure Emacs. It is a functional programming language and looks very similar to the prefix notation that is used when doing simple arithmetic:

(/ 2000 (+ 34 4.5))
51.94805194805195

;; Insert the current time string at point in the active buffer
(insert-string (current-time-string))

• Act 3: Writing Emacs Lisp Functions
  • Delete all comments from a buffer
  • Copy the contents of the current buffer

Delete all comments from a buffer

e501bb0a

My goal at this point was to write an Emacs function that would delete all the comments from a file. This goal was a fairly simple one (I thought) because I was using a Regular expression in Vim to do this: :g/^#/norm dd. I used the # because I deleted all comments only from shell files which were copies of the commands that I had run (My PS1 starts with a hash to ensure that when I ran this command, personal information like my host name or the path will be stripped).

In late October, 6 months and 9 days after I made my initial commit, I was able to write a function which actually deleted all comments from a buffer.

I faced some interesting challenges when I was writing this function. My original goal seemed extremely simple: Delete a line if it is a comment; run this for the whole buffer. Thinking like an editor, I decided to do this on a line-by-line basis.

One of the interesting challenges was figuring out when I was at the end of a buffer. Emacs does not seem to have any function to get the last line. There are functions to get the last byte position, but none to get the last line, that I was able to find. So, I had to manually keep track of whether I was deleting a line. This lead to a function which would call another function and depending on whether the line was deleted, subtract one from the original last-line-number value.

This process was full of other challenges that were related to my inability to find the name of the function that I wanted to use quickly. Emacs functions are not all named using some common pattern, which would make them easier to find. For e.g.:

• line-number-at-pos returns the line number at the current position
• (beginning-of-buffer) moves the cursor to the beginning of the buffer, (forward-line) moves one line ahead
• (kill-whole-line) deletes the current line.

As you can probably see, there’s no common convention here. All functions that deal with lines don’t start with line-, they simply have the line word in them!

Copy the contents of the current buffer

The goal for my second function was to copy the contents of the current buffer to the system clipboard. Writing this function took far less time, and I was able to find the functions I needed to use fairly easily. It was also much smaller; in fact, it had only 5 lines:

(defun copy-buffer ()
  "Copy the complete buffer to the system clipboard"
  (interactive)
  (kill-new (filter-buffer-substring (point-min) (point-max)))
  nil)

My main problem here was the terminology. The Emacs terminology for everything is different:

• yank – paste
• kill – cut, that is, delete and copy to the kill ring

What I wanted to do was copy, I could find anything close to this in the Emacs glossary. After reading some more, I realized that the kill-ring was similar to the clipboard and pushing things onto that ring is the same as copying something. And that is how I reached this function kill-new. This function pushes it’s first argument to the ring => essentially copying that string.