#+TITLE: Emacs configuration file
#+AUTHOR: Lars Tveito
#+PROPERTY: header-args :tangle yes
#+STARTUP: overview

* About

  This is an Emacs configuration file written in [[http://orgmode.org][Org mode]]. It is an attempt to
  keep my =~/.emacs.d= tidy, but still be able to keep it all in one file. I
  aim to briefly explain all my configurations as I go along!

  I would not recommend using this configuration /as-is/, because it probably
  contains a lot you don't really need. I do, however, hope people find some
  golden nuggets that they can smuggle into their own configs.

  If you really do want to try this config out, this is how I'd go about it:

  Clone the repo.

  #+begin_src sh :tangle no

  git clone https://github.com/larstvei/dot-emacs

  #+end_src

  Backup your old =~/.emacs.d= (if necessary).

  #+begin_src sh :tangle no

  mv ~/.emacs.d ~/.emacs.d-bak

  #+end_src

  Backup your old =~/.emacs=-file (if necessary).

  #+begin_src sh :tangle no

  mv ~/.emacs ~/.emacs-bak

  #+end_src

  And finally

  #+begin_src sh :tangle no

  mv dot-emacs ~/.emacs.d

  #+end_src

  On first run it should install a bunch of packages (this might take a while),
  and you might have to restart your Emacs the first time. If you experience
  bugs, please let me know!

* Meta

  All changes to the configuration should be done in =init.org=, *not* in
  =init.el=. Any changes in the =init.el= will be overwritten by saving
  =init.org=. The =init.el= in this repo should not be tracked by git, and is
  replaced the first time Emacs is started (assuming it has been renamed to
  =~/.emacs.d=).

  Emacs can't load =.org=-files directly, but =org-mode= provides functions to
  extract the code blocks and write them to a file. There are multiple ways of
  handling this; like suggested by [[http://emacs.stackexchange.com/questions/3143/can-i-use-org-mode-to-structure-my-emacs-or-other-el-configuration-file][this StackOverflow post]], one could just use
  =org-babel-load-file=, but I had problems with byte-compilation. Previously I
  tracked both the =org.=- and =el.=-files, but the git commits got a little
  messy. So here is a new approach.

  When this configuration is loaded for the first time, the ~init.el~ is the
  file that is loaded. It looks like this:

  #+begin_src emacs-lisp :tangle no

  ;; This file replaces itself with the actual configuration at first run.

  ;; We can't tangle without org!
  (require 'org)
  ;; Open the configuration
  (find-file (concat user-emacs-directory "init.org"))
  ;; tangle it
  (org-babel-tangle)
  ;; load it
  (load-file (concat user-emacs-directory "init.el"))
  ;; finally byte-compile it
  (byte-compile-file (concat user-emacs-directory "init.el"))

  #+end_src

  It tangles the org-file, so that this file is overwritten with the actual
  configuration.

  There is no reason to track the =init.el= that is generated; by running the
  following command =git= will not bother tracking it:

  #+begin_src sh :tangle no

  git update-index --assume-unchanged init.el

  #+end_src

  If one wishes to make changes to the repo-version of =init.el= start tracking
  again with:

  #+begin_src sh :tangle no

  git update-index --no-assume-unchanged init.el

  #+end_src

  I want lexical scoping for the init-file, which can be specified in the
  header. The first line of the configuration is as follows:

  #+begin_src emacs-lisp

  ;;; -*- lexical-binding: t -*-

  #+end_src

  The =init.el= should (after the first run) mirror the source blocks in the
  =init.org=. We can use =C-c C-v t= to run =org-babel-tangle=, which extracts
  the code blocks from the current file into a source-specific file (in this
  case a =.el=-file).

  To avoid doing this each time a change is made we can add a function to the
  =after-save-hook= ensuring to always tangle and byte-compile the
  =org=-document after changes.

  #+begin_src emacs-lisp

  (defun tangle-init ()
    "If the current buffer is init.org the code-blocks are
  tangled, and the tangled file is compiled."
    (when (equal (buffer-file-name)
                 (expand-file-name (concat user-emacs-directory "init.org")))
      ;; Avoid running hooks when tangling.
      (let ((prog-mode-hook nil))
        (org-babel-tangle)
        (byte-compile-file (concat user-emacs-directory "init.el")))))

  (add-hook 'after-save-hook 'tangle-init)

  #+end_src

  I'd like to keep a few settings private, so we load a =private.el= if it
  exists after the init-file has loaded.

  #+begin_src emacs-lisp

  (add-hook
   'after-init-hook
   (lambda ()
     (let ((private-file (concat user-emacs-directory "private.el")))
       (when (file-exists-p private-file)
         (load-file private-file))
       (when custom-file
         (load-file custom-file))
       (server-start))))

  #+end_src

  A common optimization is to temporarily disable garbage collection during
  initialization. Here, we set the ~gc-cons-threshold~ to a ridiculously large
  number, and restore the default value after initialization.

  #+begin_src emacs-lisp

  (let ((old-gc-treshold gc-cons-threshold))
    (setq gc-cons-threshold most-positive-fixnum)
    (add-hook 'after-init-hook
              (lambda () (setq gc-cons-threshold old-gc-treshold))))

  #+end_src

* Packages
  
  John Wiegley's extremely popular [[https://github.com/jwiegley/use-package][use-package]] was included in [[https://lists.gnu.org/archive/html/emacs-devel/2022-12/msg00261.html][Emacs 29]]. It
  provides a powerful macro for isolating package configuration. After ignoring
  this for a decade, I'll budge and give it a whirl.

  #+begin_src emacs-lisp

  (require 'use-package)
  (setq use-package-always-ensure t)

  #+end_src

  Packages can be fetched from different mirrors, [[http://melpa.milkbox.net/#/][melpa]] is the largest archive
  and is well maintained.

  #+begin_src emacs-lisp

  (setq package-archives
        '(("GNU ELPA"     . "https://elpa.gnu.org/packages/")
          ("MELPA Stable" . "https://stable.melpa.org/packages/")
          ("MELPA"        . "https://melpa.org/packages/"))
        package-archive-priorities
        '(("GNU ELPA"     . 10)
          ("MELPA"        . 5)
          ("MELPA Stable" . 0)))

  #+end_src

* Sane defaults

  These are what /I/ consider to be saner defaults.

  We can set variables to whatever value we'd like using =setq=.

  #+begin_src emacs-lisp

  (setq auto-revert-interval 1            ; Refresh buffers fast
        default-input-method "TeX"        ; Use TeX when toggling input method
        echo-keystrokes 0.1               ; Show keystrokes asap
        frame-inhibit-implied-resize 1    ; Don't resize frame implicitly
        inhibit-startup-screen t          ; No splash screen please
        initial-scratch-message nil       ; Clean scratch buffer
        recentf-max-saved-items 10000     ; Show more recent files
        ring-bell-function 'ignore        ; Quiet
        scroll-margin 1                   ; Space between cursor and top/bottom
        sentence-end-double-space nil     ; No double space
        custom-file                       ; Customizations in a separate file
        (concat user-emacs-directory "custom.el"))
  ;; Some mac-bindings interfere with Emacs bindings.
  (when (boundp 'mac-pass-command-to-system)
    (setq mac-pass-command-to-system nil))

  #+end_src

  Some variables are buffer-local, so changing them using =setq= will only
  change them in a single buffer. Using =setq-default= we change the
  buffer-local variable's default value.

  #+begin_src emacs-lisp

  (setq-default tab-width 4                       ; Smaller tabs
                fill-column 79                    ; Maximum line width
                truncate-lines t                  ; Don't fold lines
                indent-tabs-mode nil              ; Use spaces instead of tabs
                split-width-threshold 160         ; Split verticly by default
                split-height-threshold nil        ; Split verticly by default
                frame-resize-pixelwise t          ; Fine-grained frame resize
                auto-fill-function 'do-auto-fill) ; Auto-fill-mode everywhere

  #+end_src

  The =load-path= specifies where Emacs should look for =.el=-files (or
  Emacs lisp files). I have a directory called =site-lisp= where I keep all
  extensions that have been installed manually (these are mostly my own
  projects).

  #+begin_src emacs-lisp

  (let ((default-directory (concat user-emacs-directory "site-lisp/")))
    (when (file-exists-p default-directory)
      (setq load-path
            (append
             (let ((load-path (copy-sequence load-path)))
               (normal-top-level-add-subdirs-to-load-path)) load-path))))

  #+end_src

  Answering /yes/ and /no/ to each question from Emacs can be tedious, a single
  /y/ or /n/ will suffice.

  #+begin_src emacs-lisp

  (fset 'yes-or-no-p 'y-or-n-p)

  #+end_src

  To avoid file system clutter we put all auto saved files in a single
  directory.

  #+begin_src emacs-lisp

  (defvar emacs-autosave-directory
    (concat user-emacs-directory "autosaves/")
    "This variable dictates where to put auto saves. It is set to a
    directory called autosaves located wherever your .emacs.d/ is
    located.")

  ;; Sets all files to be backed up and auto saved in a single directory.
  (setq backup-directory-alist
        `((".*" . ,emacs-autosave-directory))
        auto-save-file-name-transforms
        `((".*" ,emacs-autosave-directory t)))

  #+end_src

  Set =utf-8= as preferred coding system.

  #+begin_src emacs-lisp

  (set-language-environment "UTF-8")

  #+end_src

  By default the =narrow-to-region= command is disabled and issues a
  warning, because it might confuse new users. I find it useful sometimes,
  and don't want to be warned.

  #+begin_src emacs-lisp

  (put 'narrow-to-region 'disabled nil)

  #+end_src

  Automaticly revert =doc-view=-buffers when the file changes on disk.

  #+begin_src emacs-lisp

  (add-hook 'doc-view-mode-hook 'auto-revert-mode)

  #+end_src

* Visual

  First off, let's declutter. Remove clickies to give a nice and clean look.
  Also, the cursor can relax. We add this to the [[https://www.gnu.org/software/emacs/manual/html_node/emacs/Early-Init-File.html][early-init]], as it might be
  marginally faster, and look less wonky.

  #+begin_src emacs-lisp :tangle early-init.el

  (dolist (mode
           '(tool-bar-mode                ; No toolbars, more room for text
             scroll-bar-mode              ; No scroll bars either
             blink-cursor-mode))          ; The blinking cursor gets old
    (funcall mode 0))

  #+end_src

  Add a small border on the frame. This also goes in the early-init.

  #+begin_src emacs-lisp

  (add-to-list 'default-frame-alist '(internal-border-width . 24))

  #+end_src

  I am using a lot from [[https://github.com/rougier/nano-emacs][rougier's N Λ N O Emacs]], starting with the theme.

** Theme

   For the light theme, I keep the light background toned down a touch.

   #+begin_src emacs-lisp

   ;; N Λ N O theme
   (use-package nano-theme
     :init
     (setq nano-light-background "#fafafa"
           nano-light-highlight "#f5f7f8"))

   #+end_src

   The theme is set according to the system appearance (on macOS) if that is
   available, defaulting to a light theme.

   #+begin_src emacs-lisp

   (defun load-nano-theme (variant)
     (let ((theme (intern (concat "nano-" (symbol-name variant)))))
       (load-theme theme t)))

   (load-nano-theme (if (boundp 'ns-system-appearance) ns-system-appearance 'light))

   #+end_src

   Let's have Emacs change theme when the system appearance changes as well.

   #+begin_src emacs-lisp

   (when (boundp 'ns-system-appearance-change-functions)
     (add-hook 'ns-system-appearance-change-functions 'load-nano-theme))

   #+end_src    

   I want to be able to quickly switch between a light and a dark theme.

   #+begin_src emacs-lisp

   (defun cycle-themes ()
     "Returns a function that lets you cycle your themes."
     (let ((themes '(nano-light nano-dark)))
       (lambda ()
         (interactive)
         ;; Rotates the thme cycle and changes the current theme.
         (let ((rotated (nconc (cdr themes) (list (car themes)))))
           (load-theme (car (setq themes rotated)) t))
         (message (concat "Switched to " (symbol-name (car themes)))))))

   #+end_src

** Mode line

   This is my setup for [[https://github.com/rougier/nano-modeline][N Λ N O Modeline]] after version 1.0.0:

   #+begin_src emacs-lisp

   ;; N Λ N O modeline
   (use-package nano-modeline
     :hook ((prog-mode . nano-modeline-prog-mode)
            (text-mode . nano-modeline-text-mode)
            (org-mode . nano-modeline-org-mode)
            (pdf-view-mode . nano-modeline-pdf-mode)
            (mu4e-headers-mode . nano-modeline-mu4e-headers-mode)
            (mu4e-view-mode . nano-modeline-mu4e-message-mode)
            (elfeed-show-mode . nano-modeline-elfeed-entry-mode)
            (elfeed-search-mode . nano-modeline-elfeed-search-mode)
            (term-mode . nano-modeline-term-mode)
            (xwidget-webkit-mode . nano-modeline-xwidget-mode)
            (messages-buffer-mode . nano-modeline-message-mode)
            (org-capture-mode . nano-modeline-org-capture-mode)
            (org-agenda-mode . nano-modeline-org-agenda-mode))

     ;; Disable the default modeline and add a small margin around the frame
     :init
     (setq-default mode-line-format nil)   
     (add-to-list 'default-frame-alist '(internal-border-width . 24))
     ;; Set nano-modeline-text-mode as default
     :config
     (nano-modeline-text-mode 1))

   #+end_src

** Font

   Pick the first of the following fonts that is installed on the system.

   #+begin_src emacs-lisp

   (cond ((member "Source Code Pro" (font-family-list))
          (set-face-attribute 'default nil :font "Source Code Pro-15"))
         ((member "Roboto Mono" (font-family-list))
          (set-face-attribute 'default nil :font "Roboto Mono-14"))
         ((member "Fira Code" (font-family-list))
          (set-face-attribute 'default nil :font "Fira Code-15"))
         ((member "Inconsolata" (font-family-list))
          (set-face-attribute 'default nil :font "Inconsolata-14")))

   #+end_src

   New in Emacs 24.4 is the =prettify-symbols-mode=! It's neat.

   #+begin_src emacs-lisp

   (setq-default prettify-symbols-alist '(("lambda" . ?λ)
                                          ("delta" . ?Δ)
                                          ("gamma" . ?Γ)
                                          ("phi" . ?φ)
                                          ("psi" . ?ψ)))

   #+end_src

** Centering with Olivetti

   [[https://github.com/rnkn/olivetti][Olivetti]] is a package that simply centers the text of a buffer. It is very
   simple and beautiful. The default width is just a bit short.

   #+begin_src emacs-lisp

   ;; Minor mode for a nice writing environment
   (use-package olivetti
     :defer t
     :bind ("C-c o" . olivetti-mode)
     :config
     (setq-default olivetti-body-width (+ fill-column 3))
     (remove-hook 'olivetti-mode-on-hook 'visual-line-mode))

   #+end_src

** Focusing with focus

   [[https://github.com/larstvei/Focus][Focus]] is my own package. It looks pretty nice, especially in combination
   with Olivetti!

   #+begin_src emacs-lisp

   ;; Dim color of text in surrounding sections
   (use-package focus
     :defer t
     :bind ("C-c q" .
            (lambda ()
              (interactive)
              (focus-mode 1)
              (focus-read-only-mode 1))))

   #+end_src

** Dashboard

   Dashboard provides a nice welcome.

   #+begin_src emacs-lisp

   ;; A startup screen extracted from Spacemacs
   (use-package dashboard
     :config
     (setq dashboard-banner-logo-title nil
           dashboard-center-content t
           dashboard-set-footer nil
           dashboard-page-separator "\n\n\n"
           dashboard-items '((projects . 15)
                             (recents  . 15)
                             (bookmarks . 5)))
     (dashboard-setup-startup-hook))

   #+end_src

* Mac OS X

  I run this configuration mostly on macOS, so we need a couple of settings to
  make things work smoothly. I use the =Command=-key as the =Meta=-key, Freeing
  up the =Option=-key, which I need for typing Norwegian characters on a US
  keyboard. In addition, it is more comfortable.

  I try to minimize the use of frames. The native compilation gives a lot of
  warnings, but they seem safe to ignore.

  #+begin_src emacs-lisp

  (when (memq window-system '(mac ns))
    (setq mac-option-modifier nil
          mac-command-modifier 'meta
          ns-pop-up-frames nil
          native-comp-async-report-warnings-errors nil))

  #+end_src

  The package [[https://github.com/purcell/exec-path-from-shell][exec-path-from-shell]] synchronizes environment variables from the
  shell to Emacs. This makes it a lot easier to deal with external programs on
  macOS.

  #+begin_src emacs-lisp

  (use-package exec-path-from-shell
    :if (memq window-system '(mac ns))
    :config
    (exec-path-from-shell-initialize))

  #+end_src

  I had some problems with Dired, and this seems to have solved it. I /think/
  the solutions was from [[https://stackoverflow.com/questions/4076360/error-in-dired-sorting-on-os-x][here]], and my problems were related, but not the same.

  #+begin_src emacs-lisp

  (use-package ls-lisp
    :ensure nil
    :if (memq window-system '(mac ns))
    :config
    (setq ls-lisp-use-insert-directory-program nil))
  
  #+end_src

  It is useful to be able to occasionally open the file associated with a
  buffer in macOS Finder.

  #+begin_src emacs-lisp

  (use-package reveal-in-osx-finder
    :if (memq window-system '(mac ns)))

  #+end_src

* Modes

  Here are a list of modes that I prefer enable by default.

  #+begin_src emacs-lisp

  (dolist (mode
           '(abbrev-mode                  ; E.g. sopl -> System.out.println
             column-number-mode           ; Show column number in mode line
             delete-selection-mode        ; Replace selected text
             dirtrack-mode                ; directory tracking in *shell*
             global-so-long-mode          ; Mitigate performance for long lines
             recentf-mode                 ; Recently opened files
             show-paren-mode))            ; Highlight matching parentheses
    (funcall mode 1))

  #+end_src

* Version control

  Magit is the best.

  #+begin_src emacs-lisp

  ;; A Git porcelain inside Emacs.
  (use-package magit
    :bind ("C-c m" . magit-status))

  #+end_src

  Have some visual indication where there are uncommitted changes.

  #+begin_src emacs-lisp

  ;; Highlight uncommitted changes using VC
  (use-package diff-hl
    :config
    (global-diff-hl-mode 1))

  #+end_src

* EditorConfig

  Using [[https://editorconfig.org/][EditorConfig]] is a must when collaborating with others. It is also a way
  of having multiple tools that want to format your buffer to agree (e.g. both
  the language's Emacs mode and some external formatter/prettifier).

  #+begin_src emacs-lisp

  ;; EditorConfig Emacs Plugin
  (use-package editorconfig
    :config
    (editorconfig-mode 1))

  #+end_src

* Projectile

  #+begin_src emacs-lisp

  ;; Manage and navigate projects in Emacs easily
  (use-package projectile
    :bind ("C-c p" . projectile-command-map))

  #+end_src

* Ivy/Counsel

  [[http://oremacs.com/swiper/][Ivy]] is a completion system, giving you completions and fuzzy search whenever
  you interact with the minibuffer. I transitioned to Ivy from [[https://emacs-helm.github.io/helm/][Helm]], mainly due
  to it being aesthetically noisy, and that I didn't fully take advantage of
  all its features (which are numerous). Here are some customization's that
  made the transition a bit easier.

  #+begin_src emacs-lisp

  ;; Incremental Vertical completion
  (use-package ivy
    :bind ("C-x b" . ivy-switch-buffer)
    :config
    (setq ivy-wrap t                          ; Easier access to the last candidate
          ivy-height 25                       ; Give me more candidates to look at
          ivy-use-virtual-buffers t           ; C-x b displays recents and bookmarks
          ivy-count-format "(%d/%d) "         ; Display both the index and the count
          ivy-on-del-error-function 'ignore   ; Lets me hold in backspace
          ivy-posframe-min-width 100          ; Keep ivy reasonably narrow
          ivy-posframe-height ivy-height      ; Maintain the height given by ivy
          ivy-virtual-abbreviate 'abbreviate) ; Disambiguate same file in different dirs
    (ivy-mode 1))

  #+end_src

  The completions are centered in a posframe (a frame at point).

  #+begin_src emacs-lisp

  ;; Using posframe to show Ivy
  (use-package ivy-posframe
    :config
    (ivy-posframe-mode 1))

  #+end_src

  Use counsel for =M-x=, yanking and finding files.

  #+begin_src emacs-lisp

  ;; Various completion functions using Ivy
  (use-package counsel
    :bind
    (("M-x" . counsel-M-x)
     ("M-y" . counsel-yank-pop)
     ("C-x C-f" . counsel-find-file)))

  #+end_src

  Use swiper for fancy search.

  #+begin_src emacs-lisp

  ;; Isearch with an overview.  Oh, man!
  (use-package swiper
    :bind ("C-c i" . swiper-isearch))

  #+end_src

  Have Ivy play with nice with Projectile.

  #+begin_src emacs-lisp

  ;; Ivy integration for Projectile
  (use-package counsel-projectile
    :config
    (counsel-projectile-mode 1))

  #+end_src

  Use smex to prioritize frequently used commands.

  #+begin_src emacs-lisp

  ;; M-x interface with Ido-style fuzzy matching
  (use-package smex)

  #+end_src

* PDF Tools

  [[https://github.com/vedang/pdf-tools][PDF Tools]] makes a huge improvement on the built-in [[http://www.gnu.org/software/emacs/manual/html_node/emacs/Document-View.html][doc-view-mode]]! Removing
  the =header-line-format= gives a very clean PDF-viewer; let's add that to a
  key.

  #+begin_src emacs-lisp

  ;; Emacs support library for PDF files
  (use-package pdf-tools
    :defer t
    :mode "\\.pdf\\'"
    :init (pdf-loader-install)
    :bind (:map pdf-view-mode-map
                ("c" . (lambda ()
                         (interactive)
                         (if header-line-format
                             (setq header-line-format nil)
                           (nano-modeline-pdf-mode))))
                ("j" . pdf-view-next-line-or-next-page)
                ("k" . pdf-view-previous-line-or-previous-page)))

  #+end_src

* Completion

  [[https://github.com/auto-complete/auto-complete][Auto-Complete]] has been a part of my config for years, but I want to try out
  [[http://company-mode.github.io/][company-mode]]. If I code in an environment with good completion, I've made an
  habit of trying to /guess/ function-names, and looking at the completions for
  the right one. So I want a pretty aggressive completion system, hence the no
  delay settings and short prefix length.

  #+begin_src emacs-lisp

  ;; Modular text completion framework
  (use-package company
    :bind (;; Note that a bug (probably) makes the quotes necessary here
           :map company-active-map
           ("C-d" . 'company-show-doc-buffer)
           ("C-n" . 'company-select-next)
           ("C-p" . 'company-select-previous))
    :config
    (setq company-idle-delay 0
          company-echo-delay 0
          company-dabbrev-downcase nil
          company-minimum-prefix-length 2
          company-selection-wrap-around t
          company-transformers '(company-sort-by-occurrence
                                 company-sort-by-backend-importance))
    (global-company-mode 1))

  #+end_src

* Spelling
** Flyspell

   Flyspell offers on-the-fly spell checking.

   When working with several languages, we should be able to cycle through the
   languages we most frequently use. Every buffer should have a separate cycle
   of languages, so that cycling in one buffer does not change the state in a
   different buffer (this problem occurs if you only have one global cycle). We
   can implement this by using a [[http://www.gnu.org/software/emacs/manual/html_node/elisp/Closures.html][closure]].

   #+begin_src emacs-lisp

   (defun cycle-languages ()
     "Changes the ispell dictionary to the first element in
   ISPELL-LANGUAGES, and returns an interactive function that cycles
   the languages in ISPELL-LANGUAGES when invoked."
     (let ((ispell-languages (list "american" "norsk")))
       (lambda ()
         (interactive)
         ;; Rotates the languages cycle and changes the ispell dictionary.
         (let ((rotated (nconc (cdr ispell-languages) (list (car ispell-languages)))))
           (ispell-change-dictionary (car (setq ispell-languages rotated)))))))

   #+end_src

   We enable =flyspell-mode= for all text-modes, and use =flyspell-prog-mode=
   for spell checking comments and strings in all programming modes. We bind
   =C-c l= to a function returned from =cycle-languages=, giving a language
   switcher for every buffer where flyspell is enabled.

   #+begin_src emacs-lisp

   (use-package flyspell
     :defer t
     :if (executable-find "aspell")
     :hook ((text-mode . flyspell-mode)
            (prog-mode . flyspell-prog-mode)
            (flyspell-mode . (lambda ()
                               (local-set-key
                                (kbd "C-c l")
                                (cycle-languages)))))
     :config
     (ispell-change-dictionary "american" t))

   #+end_src

** Define word

   This super neat package looks up the word at point. I use it a lot!

   #+begin_src emacs-lisp

   ;; display the definition of word at point
   (use-package define-word
     :defer t
     :bind ("C-c D" . define-word-at-point))

   #+end_src

* Org

  I use Org mode extensively. Some of these configurations may be unfortunate,
  but it is a bit impractical to change, as I have years worth of org-files and
  want to avoid having to reformat a lot of files.

  One example is =org-adapt-indentation=, which changed default value in
  version 9.5 of Org mode. Another is that I for some unknown reason decided to
  content within source content not be indented by two spaces (which is the
  default).

  Note that I disable some safety features, so please don't copy and paste
  mindlessly (see the documentation for =org-confirm-babel-evaluate= and
  =org-export-allow-bind-keywords=).

  #+begin_src emacs-lisp

  ;; Outline-based notes management and organizer
  (use-package org
    :defer t
    :config
    (setq org-adapt-indentation t
          org-src-fontify-natively t
          org-edit-src-content-indentation 0))

  #+end_src

** LaTeX export

   For LaTeX export, I default to using XeLaTeX for compilation, and the
   [[https://github.com/tecosaur/engrave-faces][engrave-faces]] package for syntax highlighting source blocks after the Emacs
   color theme.

   #+begin_src emacs-lisp

   ;; Convert font-lock faces to other formats
   (use-package engrave-faces
     :defer t)

   #+end_src

   I have PDFs open directly in Emacs ([[PDF Tools]]). In addition, I have support
   for a couple of custom LaTeX classes.

   #+begin_src emacs-lisp

   ;; LaTeX Back-End for Org Export Engine
   (use-package ox-latex
     :ensure nil
     :after org
     :config
     (setq org-export-allow-bind-keywords t
           org-latex-src-block-backend 'engraved
           org-latex-pdf-process
           '("latexmk -pdflatex='xelatex -shell-escape -interaction nonstopmode' -pdf -f %f"))

     (add-to-list 'org-file-apps '("\\.pdf\\'" . emacs))

     (add-to-list 'org-latex-classes
                  '("ifimaster"
                    "\\documentclass{ifimaster}
   [DEFAULT-PACKAGES]
   [PACKAGES]
   [EXTRA]
   \\usepackage{babel,csquotes,ifimasterforside,url,varioref}"
                    ("\\chapter{%s}" . "\\chapter*{%s}")
                    ("\\section{%s}" . "\\section*{%s}")
                    ("\\subsection{%s}" . "\\subsection*{%s}")
                    ("\\subsubsection{%s}" . "\\subsubsection*{%s}")
                    ("\\paragraph{%s}" . "\\paragraph*{%s}")
                    ("\\subparagraph{%s}" . "\\subparagraph*{%s}")))

     (add-to-list 'org-latex-classes
                  '("easychair" "\\documentclass{easychair}"
                    ("\\section{%s}" . "\\section*{%s}")
                    ("\\subsection{%s}" . "\\subsection*{%s}")
                    ("\\subsubsection{%s}" . "\\subsubsection*{%s}")
                    ("\\paragraph{%s}" . "\\paragraph*{%s}")
                    ("\\subparagraph{%s}" . "\\subparagraph*{%s}"))))

   #+end_src

** Babel

   Add a few languages for Org babel. In addition, don't evaluate code on
   export by default.

   #+begin_src emacs-lisp

   ;; Working with Code Blocks in Org
   (use-package ob
     :ensure nil
     :after org
     :config
     (setq org-export-use-babel nil
           org-confirm-babel-evaluate nil)
     (org-babel-do-load-languages
      'org-babel-load-languages
      '((emacs-lisp . t)
        (python . t)
        (clojure . t)
        (chatgpt-shell . t))))

   #+end_src

   Default to use whatever interpreter is set by =python-shell-interpreter=.

   #+begin_src emacs-lisp

   ;; Babel Functions for Python
   (use-package ob-python
     :ensure nil
     :after (ob python)
     :config
     (setq org-babel-python-command python-shell-interpreter))

   #+end_src

** Tempo

   Since version 9.2 of Org mode, typing =<s= to get a source block (and
   similar variants) has been tucked away in the Org Tempo library, hoping that
   users rather use =C-c C-,=. Hopefully I'll stop typing =<s= at some point,
   and adapt the much saner =C-c C-,=.

   #+begin_src emacs-lisp

   ;; Template expansion for Org structures
   (use-package org-tempo
     :ensure nil
     :after org)

   #+end_src

** Bullets

   Touch up the headings a bit, with some fancy UTF-8 characters.

   #+begin_src emacs-lisp

   ;; Show bullets in org-mode as UTF-8 characters
   (use-package org-bullets
     :after org
     :hook (org-mode . (lambda () (org-bullets-mode 1))))

   #+end_src

** Email with org mode

   The package org-msg allows me to compose emails with Org mode. That means I
   easily can add headings, tables, source code, etc. It is really neat.

   #+begin_src emacs-lisp

   ;; Org mode to send and reply to email in HTML
   (use-package org-msg
     :after (org mu4e)
     :config
     (add-to-list 'mu4e-compose-pre-hook 'org-msg-mode)
     (setq org-msg-enforce-css (concat user-emacs-directory "email-style.css")
           org-msg-options "html-postamble:nil toc:nil num:nil author:nil email:nil"
           org-msg-default-alternatives '((new           . (text html))
                                          (reply-to-html . (text html))
                                          (reply-to-text . (text)))
           org-msg-signature "

     ,,#+begin_signature
     ,,#+begin_export html

     - Lars
     ,,#+end_export
     ,,#+end_signature\n"))

   #+end_src

** GitHub flavored markdown

   I guess I have to include my (semi-abandoned) mode [[https://github.com/larstvei/ox-gfm][ox-gfm]] for exporting org
   mode to GitHub Flavored Markdown.

   #+begin_src emacs-lisp

   ;; Export Github Flavored Markdown from Org
   (use-package ox-gfm
     :after (org))

   #+end_src

* Markdown

  #+begin_src emacs-lisp

  ;; Emacs Major mode for Markdown-formatted files
  (use-package markdown-mode
    :defer t)

  #+end_src

* Direnv

  I use [[https://direnv.net][direnv]] in combination with [[https://nixos.org][nix]] to allow for programs to only be
  available in certain directories. The [[https://github.com/wbolster/emacs-direnv][emacs-direnv]] makes Emacs play nice with
  direnv, so that it for instance can detect a language server that is only
  available within some project. The =direnv-always-show-summary= is set to
  =nil= to avoid having long messages pop up in the messages buffer whenever I
  enter a directory that interacts with direnv.

  #+begin_src emacs-lisp

  ;; direnv integration
  (use-package direnv
    :config
    (setq direnv-always-show-summary nil)
    (direnv-mode 1))

  #+end_src

* Email

  I've used Emacs for email in the past, where I've always had the need for a
  more standard email client in addition. I'm going to give it another go.

  #+begin_src emacs-lisp

  (defvar load-mail-setup
    (and (file-exists-p "~/Maildir")
         (executable-find "mbsync")
         (executable-find "msmtp")
         (executable-find "mu")))

  #+end_src

  I use [[http://www.djcbsoftware.nl/code/mu/mu4e.html][mu4e]] (which is a part of [[http://www.djcbsoftware.nl/code/mu/][mu]]) along with [[https://isync.sourceforge.io/][mbsync]].

  #+begin_src emacs-lisp
  (use-package mu4e
    :ensure nil
    :defer t
    :if (and (file-exists-p "~/Maildir")
             (executable-find "mbsync")
             (executable-find "msmtp")
             (executable-find "mu"))
    :bind ("C-x m" . mu4e)
    :config
    (setq
     mail-user-agent 'mu4e-user-agent
     user-full-name "Lars Tveito"            ; Your full name
     user-mail-address "larstvei@ifi.uio.no" ; And email-address

     sendmail-program (executable-find "msmtp")
     send-mail-function 'smtpmail-send-it

     message-sendmail-f-is-evil t
     message-sendmail-extra-arguments '("--read-envelope-from")
     message-send-mail-function 'message-send-mail-with-sendmail
     message-kill-buffer-on-exit t

     mu4e-get-mail-command (concat (executable-find "mbsync") " -a")
     mu4e-change-filenames-when-moving t
     mu4e-user-mail-address-list '("larstvei@ifi.uio.no")
     mu4e-maildir-shortcuts '(("/Inbox" . ?i) ("/Sent Items" . ?s))

     mu4e-sent-folder "/Sent Items"
     mu4e-trash-folder "/Deleted Items"
     mu4e-trash-folder "/Drafts"

     mu4e-use-fancy-chars t))
  #+end_src

* ChatGPT

  I have a line like this:

  #+begin_example

  machine api.openai.com password OPEN-AI-KEY

  #+end_example

  in my ~.authinfo~ file. Then the ~chatgpt-shell-openai-key~ can by set by:

  #+begin_src emacs-lisp

  ;; Interaction mode for ChatGPT
  (use-package chatgpt-shell
    :defer t
    :config
    (setq chatgpt-shell-openai-key
          (auth-source-pick-first-password
           :host "api.openai.com")))

  #+end_src

  #+begin_src emacs-lisp

  ;; Org babel functions for ChatGPT evaluation
  (use-package ob-chatgpt-shell
    :after ob)

  #+end_src

* Multiple cursors

  I use this /all the time/. Perhaps more than I should?

  #+begin_src emacs-lisp

  ;; Multiple cursors for Emacs
  (use-package multiple-cursors
    :defer t
    :bind (("C-c e" . mc/edit-lines)
           ("C-c a" . mc/mark-all-like-this)
           ("C-c n" . mc/mark-next-like-this)))

  #+end_src

* Expand region

  This is neat, and I use it way less than I should.

  #+begin_src emacs-lisp

  ;; Increase selected region by semantic units
  (use-package expand-region
    :defer t
    :bind ("C-=" . er/expand-region))

  #+end_src

* Try

  [[https://github.com/larstvei/Try][Try]] is my own package for trying out packages without installing them. It is
  the most useful of my packages (IMO).

  #+begin_src emacs-lisp

  ;; Try out Emacs packages
  (use-package try
    :defer t)

  #+end_src

* Interactive functions
  <<sec:defuns>>

  =just-one-space= removes all whitespace around a point - giving it a negative
  argument it removes newlines as well. We wrap a interactive function around
  it to be able to bind it to a key. In Emacs 24.4 =cycle-spacing= was
  introduced, and it works like =just-one-space=, but when run in succession it
  cycles between one, zero and the original number of spaces.

  #+begin_src emacs-lisp

  (defun cycle-spacing-delete-newlines ()
    "Removes whitespace before and after the point."
    (interactive)
    (if (version< emacs-version "24.4")
        (just-one-space -1)
      (cycle-spacing -1)))

  #+end_src

  Often I want to find other occurrences of a word I'm at, or more specifically
  the symbol (or tag) I'm at. The =isearch-forward-symbol-at-point= in Emacs
  24.4 works well for this, but I don't want to be bothered with the =isearch=
  interface. Rather jump quickly between occurrences of a symbol, or if non is
  found, don't do anything.

  #+begin_src emacs-lisp

  (defun jump-to-symbol-internal (&optional backwardp)
    "Jumps to the next symbol near the point if such a symbol
  exists. If BACKWARDP is non-nil it jumps backward."
    (let* ((point (point))
           (bounds (find-tag-default-bounds))
           (beg (car bounds)) (end (cdr bounds))
           (str (isearch-symbol-regexp (find-tag-default)))
           (search (if backwardp 'search-backward-regexp
                     'search-forward-regexp)))
      (goto-char (if backwardp beg end))
      (funcall search str nil t)
      (cond ((<= beg (point) end) (goto-char point))
            (backwardp (forward-char (- point beg)))
            (t  (backward-char (- end point))))))

  (defun jump-to-previous-like-this ()
    "Jumps to the previous occurrence of the symbol at point."
    (interactive)
    (jump-to-symbol-internal t))

  (defun jump-to-next-like-this ()
    "Jumps to the next occurrence of the symbol at point."
    (interactive)
    (jump-to-symbol-internal))

  #+end_src

  I sometimes regret killing the =*scratch*=-buffer, and have realized I never
  want to actually kill it. I just want to get it out of the way, and clean it
  up. The function below does just this for the =*scratch*=-buffer, and works
  like =kill-this-buffer= for any other buffer. It removes all buffer content
  and buries the buffer (this means making it the least likely candidate for
  =other-buffer=).

  #+begin_src emacs-lisp

  (defun kill-this-buffer-unless-scratch ()
    "Works like `kill-this-buffer' unless the current buffer is the
  ,*scratch* buffer. In witch case the buffer content is deleted and
  the buffer is buried."
    (interactive)
    (if (not (string= (buffer-name) "*scratch*"))
        (kill-this-buffer)
      (delete-region (point-min) (point-max))
      (switch-to-buffer (other-buffer))
      (bury-buffer "*scratch*")))

  #+end_src

  To duplicate either selected text or a line we define this interactive
  function.

  #+begin_src emacs-lisp

  (defun duplicate-thing (comment)
    "Duplicates the current line, or the region if active. If an argument is
  given, the duplicated region will be commented out."
    (interactive "P")
    (save-excursion
      (let ((start (if (region-active-p) (region-beginning) (line-beginning-position)))
            (end   (if (region-active-p) (region-end) (line-end-position)))
            (fill-column most-positive-fixnum))
        (goto-char end)
        (unless (region-active-p)
          (newline))
        (insert (buffer-substring start end))
        (when comment (comment-region start end)))))

  #+end_src

  To tidy up a buffer we define this function borrowed from [[https://github.com/simenheg][simenheg]].

  #+begin_src emacs-lisp

  (defun tidy ()
    "Ident, untabify and unwhitespacify current buffer, or region if active."
    (interactive)
    (let ((beg (if (region-active-p) (region-beginning) (point-min)))
          (end (if (region-active-p) (region-end) (point-max))))
      (indent-region beg end)
      (whitespace-cleanup)
      (untabify beg (if (< end (point-max)) end (point-max)))))

  #+end_src

  Org mode does currently not support synctex (which enables you to jump from a
  point in your TeX-file to the corresponding point in the pdf), and it [[http://comments.gmane.org/gmane.emacs.orgmode/69454][seems
  like a tricky problem]].

  Calling this function from an org-buffer jumps to the corresponding section
  in the exported pdf (given that the pdf-file exists), using pdf-tools.

  #+begin_src emacs-lisp

  (defun org-sync-pdf ()
    (interactive)
    (let ((headline (nth 4 (org-heading-components)))
          (pdf (concat (file-name-base (buffer-name)) ".pdf")))
      (when (file-exists-p pdf)
        (find-file-other-window pdf)
        (pdf-links-action-perform
         (cl-find headline (pdf-info-outline pdf)
                  :key (lambda (alist) (cdr (assoc 'title alist)))
                  :test 'string-equal)))))

  #+end_src

  The opposite of fill paragraph (from [[https://www.emacswiki.org/emacs/UnfillParagraph][EmacsWiki]]),

  #+begin_src emacs-lisp
  (defun unfill-paragraph ()
    (interactive)
    (let ((fill-column most-positive-fixnum))
      (fill-paragraph nil (region-active-p))))
  #+end_src

* Advice

  An advice can be given to a function to make it behave differently. This
  advice makes =eval-last-sexp= (bound to =C-x C-e=) replace the sexp with the
  value.

  #+begin_src emacs-lisp

  (defadvice eval-last-sexp (around replace-sexp (arg) activate)
    "Replace sexp when called with a prefix argument."
    (if arg
        (let ((pos (point)))
          ad-do-it
          (goto-char pos)
          (backward-kill-sexp)
          (forward-sexp))
      ad-do-it))

  #+end_src

  When interactively changing the theme (using =M-x load-theme=), the current
  custom theme is not disabled. This often gives weird-looking results; we can
  advice =load-theme= to always disable themes currently enabled themes.

  #+begin_src emacs-lisp

  (defadvice load-theme
      (before disable-before-load (theme &optional no-confirm no-enable) activate)
    (mapc 'disable-theme custom-enabled-themes))

  #+end_src

* global-scale-mode

  These functions provide something close to ~text-scale-mode~, but for every
  buffer, including the minibuffer and mode line.

  #+begin_src emacs-lisp

  (let* ((default (face-attribute 'default :height))
         (size default))

    (defun global-scale-default ()
      (interactive)
      (global-scale-internal (setq size default)))

    (defun global-scale-up ()
      (interactive)
      (global-scale-internal (setq size (+ size 20))))

    (defun global-scale-down ()
      (interactive)
      (global-scale-internal (setq size (- size 20))))

    (defun global-scale-internal (arg)
      (set-face-attribute 'default (selected-frame) :height arg)
      (set-temporary-overlay-map
       (let ((map (make-sparse-keymap)))
         (define-key map (kbd "C-=") 'global-scale-up)
         (define-key map (kbd "C-+") 'global-scale-up)
         (define-key map (kbd "C--") 'global-scale-down)
         (define-key map (kbd "C-0") 'global-scale-default) map))))

  #+end_src

* Mode specific
** Eglot

   I am using [[https://joaotavora.github.io/eglot/][eglot]], which is built in from [[https://git.savannah.gnu.org/cgit/emacs.git/tree/etc/NEWS?h=emacs-29#n3273][emacs 29.1]]. Some performance issues
   led me to set =eglot-events-buffer-size= to 0.

   #+begin_src emacs-lisp

   (setq eglot-events-buffer-size 0)
   (add-hook 'eglot-managed-mode-hook (lambda () (eglot-inlay-hints-mode -1)))

   #+end_src

** Compilation

   I often run ~latexmk -pdf -pvc~ in a compilation buffer, which recompiles
   the latex-file whenever it is changed. This often results in annoyingly
   large compilation buffers; the following snippet limits the buffer size in
   accordance with ~comint-buffer-maximum-size~, which defaults to 1024 lines.

   #+begin_src emacs-lisp

   (add-hook 'compilation-filter-hook 'comint-truncate-buffer)

   #+end_src

** vterm

   Inspired by [[https://github.com/torenord/.emacs.d][torenord]], I maintain quick access to shell buffers with bindings
   ~M-1~ to ~M-9~. In addition, the ~M-z~ toggles between the last visited
   shell, and the last visited non-shell buffer.

   #+begin_src emacs-lisp

   ;; A terminal via libvterm
   (use-package vterm
     :defer t
     :preface
     (let ((last-vterm ""))
       (defun toggle-vterm ()
         (interactive)
         (cond ((string-match-p "^\\vterm<[1-9][0-9]*>$" (buffer-name))
                (goto-non-vterm-buffer))
               ((get-buffer last-vterm) (switch-to-buffer last-vterm))
               (t (vterm (setq last-vterm "vterm<1>")))))

       (defun goto-non-vterm-buffer ()
         (let* ((r "^\\vterm<[1-9][0-9]*>$")
                (vterm-buffer-p (lambda (b) (string-match-p r (buffer-name b))))
                (non-vterms (cl-remove-if vterm-buffer-p (buffer-list))))
           (when non-vterms
             (switch-to-buffer (car non-vterms)))))

       (defun switch-vterm (n)
         (let ((buffer-name (format "vterm<%d>" n)))
           (setq last-vterm buffer-name)
           (cond ((get-buffer buffer-name)
                  (switch-to-buffer buffer-name))
                 (t (vterm buffer-name)
                    (rename-buffer buffer-name))))))

     :bind (("C-z" . toggle-vterm)
            ("M-1" . (lambda () (interactive) (switch-vterm 1)))
            ("M-2" . (lambda () (interactive) (switch-vterm 2)))
            ("M-3" . (lambda () (interactive) (switch-vterm 3)))
            ("M-4" . (lambda () (interactive) (switch-vterm 4)))
            ("M-5" . (lambda () (interactive) (switch-vterm 5)))
            ("M-6" . (lambda () (interactive) (switch-vterm 6)))
            ("M-7" . (lambda () (interactive) (switch-vterm 7)))
            ("M-8" . (lambda () (interactive) (switch-vterm 8)))
            ("M-9" . (lambda () (interactive) (switch-vterm 9))))

     :config
     ;; Don't query about killing vterm buffers, just kill it
     (defadvice vterm (after kill-with-no-query nil activate)
       (set-process-query-on-exit-flag (get-buffer-process ad-return-value) nil)))

   #+end_src

** Lisp

   I use [[https://paredit.org/][Paredit]] when editing lisp code, we enable this for all lisp-modes.
   Paredit version 25 [[https://paredit.org/cgit/paredit/plain/NEWS][seems to interfere]] with REPL-modes, and unbinding =RET=
   is the proposed fix.

   #+begin_src emacs-lisp

   ;; minor mode for editing parentheses
   (use-package paredit
     :defer t
     :bind (:map paredit-mode-map ("RET" . nil))
     :hook ((cider-repl-mode
             clojure-mode
             ielm-mode
             racket-mode
             racket-repl-mode
             slime-repl-mode
             lisp-mode
             emacs-lisp-mode
             lisp-interaction-mode
             scheme-mode) 
            . paredit-mode))

   #+end_src

*** Emacs Lisp

    In =emacs-lisp-mode= we can enable =eldoc-mode= to display information
    about a function or a variable in the echo area.

    #+begin_src emacs-lisp

    (add-hook 'emacs-lisp-mode-hook 'turn-on-eldoc-mode)
    (add-hook 'lisp-interaction-mode-hook 'turn-on-eldoc-mode)

    #+end_src

*** Clojure

    A very simple setup for Clojure. Cider works pretty much out of the box!

    #+begin_src emacs-lisp

    ;; Clojure Interactive Development Environment
    (use-package cider
      :defer t
      :bind (:map cider-repl-mode-map ("C-l" . cider-repl-clear-buffer)))

    #+end_src

    #+begin_src emacs-lisp

    ;; Commands for refactoring Clojure code
    (use-package clj-refactor
      :defer t)

    #+end_src

*** Racket

    A minimal setup for Racket.

    #+begin_src emacs-lisp

    ;; Major mode for Racket language
    (use-package racket-mode
      :defer t)

    #+end_src

*** Common lisp

    #+begin_quote
    Note that I haven't used Common Lisp for a very long time, and this setup
    might be broken. I keep it around for reference.
    #+end_quote

    I use [[http://www.common-lisp.net/project/slime/][Slime]] along with =lisp-mode= to edit Common Lisp code. Slime provides
    code evaluation and other great features, a must have for a Common Lisp
    developer. You can install the Common Lisp slime counterpart using
    [[http://www.quicklisp.org/beta/][Quicklisp]], creating a helper that can be loaded.

    We can specify what Common Lisp program Slime should use (I use SBCL). More
    sensible =loop= indentation is borrowed from [[https://github.com/simenheg][simenheg]].

    #+begin_src emacs-lisp

    ;; Superior Lisp Interaction Mode for Emacs
    (use-package slime
      :disabled
      :defer t
      :bind (:map slime-repl-mode-map ("C-l" . slime-repl-clear-buffer))
      :hook (common-lisp-mode . activate-slime-helper)
      :config
      (when (file-exists-p "~/.quicklisp/slime-helper.el")
        (load (expand-file-name "~/.quicklisp/slime-helper.el")))

      (setq inferior-lisp-program "sbcl")

      (setq lisp-loop-forms-indentation   6
            lisp-simple-loop-indentation  2
            lisp-loop-keyword-indentation 6))

    #+end_src

** Python

   #+begin_src emacs-lisp

   (setq python-shell-interpreter "python3.10")
   (add-hook 'python-mode-hook
             (lambda () (setq forward-sexp-function nil)))

   #+end_src

** C

   The =c-mode-common-hook= is a general hook that work on all C-like languages
   (C, C++, Java, etc...). I like being able to quickly compile using =C-c C-c=
   (instead of =M-x compile=), a habit from =latex-mode=.

   #+begin_src emacs-lisp

   (defun c-setup ()
     (local-set-key (kbd "C-c C-c") 'compile))

   (add-hook 'c-mode-hook 'c-setup)

   #+end_src

** Java

   Some statements in Java appear often, and become tedious to write out. We
   can use abbrevs to speed this up.

   #+begin_src emacs-lisp

   (define-abbrev-table 'java-mode-abbrev-table
     '(("psv" "public static void main(String[] args) {" nil 0)
       ("sopl" "System.out.println" nil 0)
       ("sop" "System.out.printf" nil 0)))

   #+end_src

   To be able to use the abbrev table defined above, =abbrev-mode= must be
   activated.

   #+begin_src emacs-lisp

   (add-hook 'java-mode-hook 'eglot-ensure)

   #+end_src

** Assembler

   When writing assembler code I use =#= for comments. By defining
   =comment-start= we can add comments using =M-;= like in other programming
   modes. Also in assembler should one be able to compile using =C-c C-c=.

   #+begin_src emacs-lisp

   (defun asm-setup ()
     (setq comment-start "#")
     (local-set-key (kbd "C-c C-c") 'compile))

   (add-hook 'asm-mode-hook 'asm-setup)

   #+end_src

** LaTeX

   #+begin_src emacs-lisp

   ;; Integrated environment for *TeX*
   (use-package auctex
     :defer t)

   #+end_src

** Erlang

   Erlang mode works out of the box.

   #+begin_src emacs-lisp

   ;; Erlang major mode
   (use-package erlang
     :defer t)

   #+end_src

** Nix

   #+begin_src emacs-lisp

   ;; Major mode for editing .nix files
   (use-package nix-mode
     :defer t
     :hook (nix-mode . eglot-ensure))

   #+end_src

** Haskell

   =haskell-doc-mode= is similar to =eldoc=, it displays documentation in the
   echo area. Haskell has several indentation modes - I prefer using
   =haskell-indent=.

   #+begin_src emacs-lisp

   ;; A Haskell editing mode
   (use-package haskell-mode
     :defer t
     :hook ((haskell-mode . interactive-haskell-mode)
            (haskell-mode . turn-on-haskell-doc-mode)
            (haskell-mode . turn-on-haskell-indent)))

   #+end_src

** Maude

   Use =---= for comments in Maude.

   #+begin_src emacs-lisp

   ;; Emacs mode for the programming language Maude
   (use-package maude-mode
     :defer t
     :hook (maude-mode . (lambda () (setq-local comment-start "---")))
     :config
     (add-to-list 'maude-command-options "-no-wrap"))

   #+end_src

** Minizinc

   Provide a default =compile-command=.

   #+begin_src emacs-lisp

   (defun minizinc-setup-compile-command ()
     (let ((command (concat "minizinc " (buffer-file-name) " "))
           (f (concat (file-name-base (buffer-file-name)) ".dzn")))
       (local-set-key (kbd "C-c C-c") 'recompile)
       (setq-local compile-command (concat command (if (file-exists-p f) f "")))))

   #+end_src

   Use =minizinc-mode=, and hook up the =minizinc-setup-compile-command= above.

   #+begin_src emacs-lisp

   ;; Major mode for MiniZinc code
   (use-package minizinc-mode
     :disabled
     :defer t
     :mode "\\.mzn\\'"
     :hook (minizinc-mode . minizinc-setup-compile-command))

   #+end_src

** Coq

   [[https://proofgeneral.github.io/][Proof General]] is really great for working with proof assistants. I have only
   tried it with Coq.

   #+begin_src emacs-lisp

   ;; A generic Emacs interface for proof assistants
   (use-package proof-general
     :disabled
     :defer t)

   #+end_src

   For completions, I use [[https://github.com/cpitclaudel/company-coq][company-coq]].

   #+begin_src emacs-lisp

   ;; A collection of extensions PG's Coq mode
   (use-package company-coq
     :disabled
     :defer t
     :hook (coq-mode . company-coq-mode))

   #+end_src

** Rust

   #+begin_src emacs-lisp

   ;; Rust development environment
   (use-package rustic
     :defer t
     :config
     (setq rustic-lsp-client 'eglot))

   #+end_src

** Go

   #+begin_src emacs-lisp

   ;; Major mode for the Go programming language
   (use-package go-mode
     :defer t
     :mode "\\.go\\'"
     :hook (go-mode . eglot-ensure))

   #+end_src

** Lua

   #+begin_src emacs-lisp

   ;; a major-mode for editing Lua scripts
   (use-package lua-mode
     :defer t)

   #+end_src

** Webdev

   My webdev setup isn't much, but with eglot, I don't find myself missing
   much. It depends on [[https://tree-sitter.github.io/tree-sitter/][Tree-sitter]], which was added in [[https://git.savannah.gnu.org/cgit/emacs.git/tree/etc/NEWS?h=emacs-29#n36][emacs 29.1]].

   The following is what I use for plain Javascript:

   #+begin_src emacs-lisp

   (add-to-list 'auto-mode-alist '("\\.jsx?\\'" . js-ts-mode))
   (add-hook 'js-ts-mode-hook 'eglot-ensure)

   #+end_src

   Similarly for Typescript:

   #+begin_src emacs-lisp

   (add-to-list 'auto-mode-alist '("\\.tsx?\\'" . tsx-ts-mode))
   (add-hook 'tsx-ts-mode-hook 'eglot-ensure)
   (setq typescript-ts-mode-indent-offset 4)

   #+end_src

   I am using [[https://svelte.dev][Svelte]] for some projects, where I find [[https://web-mode.org][web-mode]] along with the
   [[https://github.com/sveltejs/language-tools][Svelte Language Server]] to work well.

   #+begin_src emacs-lisp

   (add-to-list 'auto-mode-alist '("\\.svelte\\'" . web-mode))
   (add-hook 'web-mode-hook 'eglot-ensure)
   (with-eval-after-load "web-mode"
     (add-to-list 'web-mode-engines-alist '("svelte" . "\\.svelte\\'")))
   (with-eval-after-load "eglot"
     (add-to-list 'eglot-server-programs
                  '(web-mode . ("svelteserver" "--stdio"))))

   #+end_src

** Z3

   I mostly use [[https://github.com/Z3Prover/z3][Z3]] as a Python library, but occasionally I'll run some SMT-LIB
   code directly.

   #+begin_src emacs-lisp

   ;; z3/SMTLIBv2 interactive development
   (use-package z3-mode
     :disabled
     :defer t)

   #+end_src

* Modal meow

  I have been wanting to try out modal editing. [[https://github.com/meow-edit/meow][meow]] seems like a nice package,
  where I can still use a lot of the ten years of Emacs that are already in my
  fingers. These are the default settings for qwerty.

  #+begin_src emacs-lisp

  (use-package meow
    :config
    (setq meow-cheatsheet-layout meow-cheatsheet-layout-qwerty)
    (meow-motion-overwrite-define-key
     '("j" . meow-next)
     '("k" . meow-prev)
     '("<escape>" . ignore))
    (meow-leader-define-key
     ;; SPC j/k will run the original command in MOTION state.
     '("j" . "H-j")
     '("k" . "H-k")
     ;; Use SPC (0-9) for digit arguments.
     '("1" . meow-digit-argument)
     '("2" . meow-digit-argument)
     '("3" . meow-digit-argument)
     '("4" . meow-digit-argument)
     '("5" . meow-digit-argument)
     '("6" . meow-digit-argument)
     '("7" . meow-digit-argument)
     '("8" . meow-digit-argument)
     '("9" . meow-digit-argument)
     '("0" . meow-digit-argument)
     '("/" . meow-keypad-describe-key)
     '("?" . meow-cheatsheet))
    (meow-normal-define-key
     '("0" . meow-expand-0)
     '("9" . meow-expand-9)
     '("8" . meow-expand-8)
     '("7" . meow-expand-7)
     '("6" . meow-expand-6)
     '("5" . meow-expand-5)
     '("4" . meow-expand-4)
     '("3" . meow-expand-3)
     '("2" . meow-expand-2)
     '("1" . meow-expand-1)
     '("-" . negative-argument)
     '(";" . meow-reverse)
     '("," . meow-inner-of-thing)
     '("." . meow-bounds-of-thing)
     '("[" . meow-beginning-of-thing)
     '("]" . meow-end-of-thing)
     '("a" . meow-append)
     '("A" . meow-open-below)
     '("b" . meow-back-word)
     '("B" . meow-back-symbol)
     '("c" . meow-change)
     '("d" . meow-delete)
     '("D" . meow-backward-delete)
     '("e" . meow-next-word)
     '("E" . meow-next-symbol)
     '("f" . meow-find)
     '("g" . meow-cancel-selection)
     '("G" . meow-grab)
     '("h" . meow-left)
     '("H" . meow-left-expand)
     '("i" . meow-insert)
     '("I" . meow-open-above)
     '("j" . meow-next)
     '("J" . meow-next-expand)
     '("k" . meow-prev)
     '("K" . meow-prev-expand)
     '("l" . meow-right)
     '("L" . meow-right-expand)
     '("m" . meow-join)
     '("n" . meow-search)
     '("o" . meow-block)
     '("O" . meow-to-block)
     '("p" . meow-yank)
     '("q" . meow-quit)
     '("Q" . meow-goto-line)
     '("r" . meow-replace)
     '("R" . meow-swap-grab)
     '("s" . meow-kill)
     '("t" . meow-till)
     '("u" . meow-undo)
     '("U" . meow-undo-in-selection)
     '("v" . meow-visit)
     '("w" . meow-mark-word)
     '("W" . meow-mark-symbol)
     '("x" . meow-line)
     '("X" . meow-goto-line)
     '("y" . meow-save)
     '("Y" . meow-sync-grab)
     '("z" . meow-pop-selection)
     '("'" . repeat)
     '("<escape>" . ignore))
    (meow-global-mode 1))

  #+end_src

* Key bindings

  Inspired by [[http://stackoverflow.com/questions/683425/globally-override-key-binding-in-emacs][this StackOverflow post]] I keep a =custom-bindings-map= that holds
  all my custom bindings. This map can be activated by toggling a simple
  =minor-mode= that does nothing more than activating the map. This inhibits
  other =major-modes= to override these bindings. I keep this at the end of the
  init-file to make sure that all functions are actually defined.

  #+begin_src emacs-lisp

  (defvar custom-bindings-map (make-keymap)
    "A keymap for custom bindings.")

  #+end_src

** Which key

   [[https://github.com/justbur/emacs-which-key][Which key]] is nice for discoverability.

   #+begin_src emacs-lisp

   ;; Display available keybindings in popup
   (use-package which-key
     :config
     (which-key-mode 1))

   #+end_src

** Bindings for [[https://github.com/magnars/multiple-cursors.el][multiple-cursors]]

  #+begin_src emacs-lisp

  (define-key custom-bindings-map (kbd "C-c e")  'mc/edit-lines)
  (define-key custom-bindings-map (kbd "C-c a")  'mc/mark-all-like-this)
  (define-key custom-bindings-map (kbd "C-c n")  'mc/mark-next-like-this)

  #+end_src

** Bindings for built-ins

  #+begin_src emacs-lisp

  (define-key custom-bindings-map (kbd "M-u")         'upcase-dwim)
  (define-key custom-bindings-map (kbd "M-c")         'capitalize-dwim)
  (define-key custom-bindings-map (kbd "M-l")         'downcase-dwim)
  (define-key custom-bindings-map (kbd "M-]")         'other-frame)
  (define-key custom-bindings-map (kbd "C-j")         'newline-and-indent)
  (define-key custom-bindings-map (kbd "C-c s")       'ispell-word)
  (define-key comint-mode-map     (kbd "C-l")         'comint-clear-buffer)

  #+end_src

** Bindings for functions defined [[sec:defuns][above]].

  #+begin_src emacs-lisp

  (define-key global-map          (kbd "M-p")     'jump-to-previous-like-this)
  (define-key global-map          (kbd "M-n")     'jump-to-next-like-this)
  (define-key custom-bindings-map (kbd "M-,")     'jump-to-previous-like-this)
  (define-key custom-bindings-map (kbd "M-.")     'jump-to-next-like-this)
  (define-key custom-bindings-map (kbd "C-c .")   (cycle-themes))
  (define-key custom-bindings-map (kbd "C-x k")   'kill-this-buffer-unless-scratch)
  (define-key custom-bindings-map (kbd "C-c C-0") 'global-scale-default)
  (define-key custom-bindings-map (kbd "C-c C-=") 'global-scale-up)
  (define-key custom-bindings-map (kbd "C-c C-+") 'global-scale-up)
  (define-key custom-bindings-map (kbd "C-c C--") 'global-scale-down)
  (define-key custom-bindings-map (kbd "C-c j")   'cycle-spacing-delete-newlines)
  (define-key custom-bindings-map (kbd "C-c d")   'duplicate-thing)
  (define-key custom-bindings-map (kbd "<C-tab>") 'tidy)

  (with-eval-after-load 'org
    (define-key org-mode-map (kbd "C-'") 'org-sync-pdf))

  #+end_src

  Lastly we need to activate the map by creating and activating the
  =minor-mode=.

  #+begin_src emacs-lisp

  (define-minor-mode custom-bindings-mode
    "A mode that activates custom-bindings."
    t nil custom-bindings-map)

  #+end_src

* License

  My Emacs configurations written in Org mode.

  Copyright (c) 2013 - 2023 Lars Tveito

  This program is free software: you can redistribute it and/or modify it under
  the terms of the GNU General Public License as published by the Free Software
  Foundation, either version 3 of the License, or (at your option) any later
  version.

  This program is distributed in the hope that it will be useful, but WITHOUT
  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
  FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
  details.

  You should have received a copy of the GNU General Public License along with
  this program. If not, see <http://www.gnu.org/licenses/>.