Some (X)Emacs stuff

I am the author of two little developments for Emacs:

circul.el

Allows to:

Circulate Quickly between buffers (ex: ctrl-kpenter / ctrl-kpadd for backward / forward circulating). Customizable regular expressions allow to define which files may or may not appear when circulating.

Assign to a buffer or a place in a buffer some key of your keyboard keys (for example a key of the keypad, or an 'F' key) and come back to this buffer/place by just pressing this key again.

It is quite useful when one needs to quickly switch between different places in one or several buffers, typically when programming.

It is very light, as emacs provides something called "registers" which mainly does the work. Actually this package is only a wrapper mixing a list of buffers and registers.

The package does not assign any keystroke by itself, the user must assign them in its config files. The following code makes use of the keypad on the right of a standard pc french keyboard. Modify it as you want, a good idea is to use f(5-12) keys instead.

(setq circul-uninteresting-buffer-regexp
      "\\`\\*\\|\\` \\|.+\\.log\\'\\|bbdb\\|TAGS")
(setq circul-uninteresting-buffer-exception-regexp 
      "\\`\\*coq\\|\\`\\*cvs\\*")

; Navigation in interesting buffers

(global-set-key [(control kp-enter)]  'circul-bury-buffer)
(global-set-key [(control kp-add)]  'circul-unbury-buffer)
(global-set-key [(control meta button4)]  'circul-bury-buffer)
(global-set-key [(control meta button5)]  'circul-unbury-buffer)


; Assigning keys to buffers or registers

; In this example we assign keys of the numpad:
;  * C-kp-x assigns kpx to a buffer
;  * C-S-kp-x assigns une position in a buffer (i.e. a register) to k-px
;  * kp-x jumps to the position of buffer assigned to kp-x
; Note that ?0 .. ?9 mais be replaced by other characters, provided
; they are all different.

(circul-assign ?0 [(control kp-insert)] [(control shift kp-0)] [(kp-insert)])
(circul-assign ?1 [(control kp-end)] [(control shift kp-1)] [(kp-end)])
(circul-assign ?2 [(control kp-down)] [(control shift kp-2)] [(kp-down)])
(circul-assign ?3 [(control kp-next)] [(control shift kp-3)] [(kp-next)])
(circul-assign ?4 [(control kp-left)] [(control shift kp-4)] [(kp-left)])
(circul-assign ?5 [(control kp-begin)] [(control shift kp-5)] [(kp-begin)])
(circul-assign ?6 [(control kp-right)] [(control shift kp-6)] [(kp-right)])
(circul-assign ?7 [(control kp-home)] [(control shift kp-7)] [(kp-home)])
(circul-assign ?8 [(control kp-up)] [(control shift kp-8)] [(kp-up)])
(circul-assign ?9 [(control kp-prior)] [(control shift kp-9)] [(kp-prior)])

BUGS

tell me

holes.tgz
Updated on Mar 1th 2005 (corrected: bug on last buffer line).
This is inspired from other interactive programs ( Pcoq ). It allows to defines "holes" in your buffer. A hole is a piece of text (with nothing particular except it is highlighted) that may be replaced by another piece of text later. This feature is useful to build complicated expressions by copy pasting several peaces of text from different parts of a buffer (even from different buffers). Holes are just highlighted pieces of your buffer, nothing is written in your file and you can delete holes exactly like normal characters (there are other ways to delete a hole, see below).

USE

There is always one or zero active hole, highlighted with a different color.
To define a hole, two methods:
- Select a region with keyboard (ctrl-space) or mouse, then hit ctrl-c h
- Select text with mouse while pressing ctrl + meta.
If the region is empty, then a hole containing '#' is created.
To activate a hole, click on it with the button 1 of your mouse.

To forget a hole without deleting its text, click on it with the button 2(middle) of your mouse.

To destroy a hole and delete its text, click on it with the button 3 of your mouse.

To replace the active hole first make sure it is active, then two methods:
- Select text with keyboard (ctrl-space) or mouse and hit ctrl-c ctrl-y
- Select text with mouse while pressing ctrl + meta + shift. This is a generalization of the mouse-track-insert feature (ctrl + select text, if you don't know this, try it :-))
The following hole is automatically made active.
to jump to the active hole, just hit meta-return. You must be in the buffer containing the active hole. the point will move to the active hole, and the active hole will be destroyed so you can type something to put at its place. The following hole is automatically made active, so you can hit meta-return again.

It is also useful in conjonction with abbreviations, for example in c-mode I use "f" as an abbreviation for for(#;#;#) {#}, where each # is a hole, then hitting meta-return goes from one hole to the following and I can fill-in each hole very quickly. This last trick needs some very simple manipulation on your abbrev file. See some examples in my : abbrev file .

There are four files in this archive, three of them are used to deal with this f... Emacs/Xemacs incompatibility.

BUGS
- Dont try to make overlapping holes, it doesn't work. (what would it mean anyway?)
- With FSF emacs, cutting or pasting a hole wil not produce new holes, and undoing on holes cannot make reappear holes. With Xemacs it will, but if you copy paste the active hole, you will get several holes highlighted as the active one, which is annoying.
- tell me

To use these two files, download them and put them (untar holes.tgz with the command tar xvfz holes.tgz, you get a directory with four files) somewhere Xemacs can find them. Then put this in you initialization file:


	(require 'holes-load)
	

	(require 'circul)

To toggle holes-mode: M-x holes-mode. You can put (holes-mode 1) in the hooks of your modes in order to have holes-mode automatically enabled. All shortcuts described above will only work on buffers where holes-mode is on. Of course you can edit holes.el to change the shortcuts.

The files span.el, span-extent.el and span.el are from the ProofGeneral mode. Everything is under GPL. It should work on both FSF Emacs(21) and Xemacs. Ideas for improvement and bug reports are welcome.

Emacs and git (magit)

magit is quite easy to use for nomianl use. I was unhappy with magit's way of splitting/deleting windows, so I put this in my initialization file (notice that for quitting I only redefine the "q" keybinding, the menu "quit" still kills the window):

(defvar pc:magit-from-buffer
  "Buffer from where magit-status were last called."
  nil)


(defun pc:magit-status-buffer-switch (buf)
  "My replacement for `magit-status-buffer-switch-function'.
`magit-status' does not split windows (switch to magit buffer
instead). Also store the current buffer to switch back to it when
quitting.
TODO: store the whole frame config instead?"
  (setq pc:magit-from-buffer (current-buffer))
  (switch-to-buffer buf))

(setq magit-status-buffer-switch-function 'pc:magit-status-buffer-switch)

(defun pc:magit-quit-window (&optional kill-buffer)
  "My replacement for \"q\" keybinding in magit.
Bury the current (magit) buffer and switch to original buffer.
With a prefix argument, kill the magit buffer instead."
  (interactive "P")
  (if kill-buffer (kill-buffer) (bury-buffer))
  (switch-to-buffer pc:magit-from-buffer))

(add-hook 'magit-mode-hook
	  '(lambda () (define-key magit-mode-map (kbd "q") 'pc:magit-quit-window)))

Emacs and LaTeX

Here is a minimum config you should add to your .emacs (replace (control fxx) by the keystroke you want). RefTeX should be installed by default if you have a standard installation of (x)emacs:

 (add-hook 'LaTeX-mode-hook '(lambda () 
    (turn-on-reftex); Turn on RefTeX Minor Mode for all LaTeX files
    (define-key LaTeX-mode-map [(control f12)] 'TeX-next-error)
    (define-key LaTeX-mode-map [(control f11)]
      '(lambda () (interactive) 
         (TeX-command-menu "LaTeX")))

Now hit (control f11), it compiles, if there are some errors, type (control f12) and it goes straight to the error and shows the message. See also the ref menu to navigate through the table of content of your document.

A hack to have xdvi updated when latex compile is finished:

This one is a bit outdated since everybody uses pdf output nowadays. It allows to have a keystroke which runs latex on your file, then waits for the process to finish, and then sends a signal to xdvi telling him to watch the file (change "xdvi.bin" by the exact name of the process if necessary). That way you don't have to switch to xdvi's window each time. Thanks to Eric Pop helping make this emacs compatible.


    
(defvar latex-process-timer nil
  "the timer used to watch latex and dvips processes.")

(defun send-sig (process-string sig)
  "sends sig signal to every process having name process-string"
  (interactive)
  (call-process "killall" nil nil nil sig process-string))
(defun wait-latex-xdvi ()
  "wait for the latex process to stop, and send the USR1 signal to all
  running xdvi.bin processes"
  (if (TeX-active-process)
      ()
    (let ((x (current-message)))
      (send-sig "xdvi.bin" "-USR1")
      (if (featurep 'xemacs) (delete-itimer latex-process-timer) 
	(cancel-timer latex-process-timer))
      (message x))
    )
  )

(defun latex-xdvi () 
  (interactive) 
  (TeX-command-menu "LaTeX")
  (setq latex-process-timer (run-with-timer 1 1 'wait-latex-xdvi)) ;; latex-process-timer is a global variable
  )


    
    Then replace the control f11 shortcut above by this one:
    (add-hook 'LaTeX-mode-hook '(lambda () 
  ...
  (define-key LaTeX-mode-map [(control f11)] 'latex-xdvi))
)

Last modified: Wed Sep 6 17:56:47 CEST 2006

Some (X)Emacs stuff

circul.el

BUGS

holes.tgz

USE

BUGS

Emacs and git (magit)

Emacs and LaTeX

A hack to have xdvi updated when latex compile is finished: