elisp regular expression build problem

Question

I have the following line :

mod_name func_name /users/username/workspace/path_name/file.c 283

I wanted to capture the 3rd and 4th entry in this line, I am able to do that using regular expressions like so: ^(?:\S+\s){2}(\S+)\ ([0-9]+)

How do I translate this into a lisp regex ? so that elisp functions can understand. I couldn't use the above regex into re-builder within emacs. Says it is an invalid expression. What am I doing wrong ?

Answer 1

Emacs regexps do not use \S or \s for non-whitespace and whitepsace. See (info "(elisp) Syntax of Regexps") for details. You can evaluate that which will jump to the corresponding info page.

It's far easier and more readable to use the rx macro to create regexps in Emacs. Here is my translation to rx syntax of your example:

(rx line-start
    (= 2 (and (one-or-more any) " "))
    (group (one-or-more any))
    " "
    (group (one-or-more digit)))

And here is resulting regex string:

"^\\(?:.+ \\)\\{2\\}\\(.+\\) \\([0-9]+\\)"

Note that any which is expressed in regexps as "." does not match newlines in Emacs.