Custom character classes in Emacs' regexp

Question

Is it possible to add own character classes to emacs in order to use them in regular expressions?

Let's say, i want to add a class [[:consonant:]] which matches all letters that are not vowels in order to avoid writing [b-df-hj-np-tv-z] all the time (and yes, i am aware that my shortcut is almost as long as the term i want to avoid, take it as a simplification of my problem).

Is this possible at all or do i have to use format or concat, respectively? If it is possible, how do i do that?

An MWE could be like this:

(defun myfun ()
  "Finds clusters of three or more consonants"
  (interactive)
  (if (search-forward-regexp "[b-df-hj-np-tv-z]\\{3,\\}")
    (message "Yepp, here is a consonant cluster.")
))

(defun myfun-1 ()
  "Should also find clusters of three or more consonants."
  (interactive)
  (if (search-forward-regexp "[[:consonant:]]\\{3,\\}")
    (message "Yepp, here is a consonant cluster.")
))

Both functions myfun and myfun-1 should do the very same thing.

One step further i'd like to know if it is possible to put whole expressions in such "shortcuts", like

[[:ending:]] ==> "\\(?:en\\|st\\|t\\|e\\)"

Trey Jackson · Accepted Answer · 2021-10-08 23:20:21Z

3

Jordon-Biondo is correct, you cannot extend Emacs' character classes for regexp search. If you peek into Emacs' source code, you can see these defined in the routine re_wctype_parse on line 1510(ish) of regex-emacs.c. So adding to these natively would require modifying the .c file and rebuilding.

edited Oct 8, 2021 at 23:20

answered Oct 8, 2021 at 23:10

Trey Jackson

74.8k11 gold badges204 silver badges233 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Jordon Biondo · Accepted Answer · 2014-06-27 16:37:53Z

1

I do not believe you can do this. But there is something similar that was recently released in the ample-regexp package found HERE. This was taken from the readme as an example:

(define-arx h-w-rx
  '((h "Hello, ")
    (w "world"))) ;; -> hello-world-rx

(h-w-rx h w) ;; -> "Hello, world"

(h-w-rx (* h w)) ;; -> "\\(?:Hello, world\\)*"

You could use this to define a wide range of aliases in one big define-arx.

answered Jun 27, 2014 at 16:37

Jordon Biondo

4,0471 gold badge30 silver badges37 bronze badges

Collectives™ on Stack Overflow

Custom character classes in Emacs' regexp

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related