Re: Improve `replace-regexp-in-string' ergonomics?

[Stefan Monnier]

>                 (replace-regexp-in-string
>                  "\\[" "(" (replace-regexp-in-string
>                             "\\]" ")" (replace-regexp-in-string
>                                        ",[[:space:]]" " "
>                                      (replace-regexp-in-string
>                                       "'" "\"" results))))))

To me "the right way" is to pass a function as the replacement.
I'm not sure the way REP currently works in `replace-regexp-in-string`
is the best we can do, but it works:

    (replace-regexp-in-string
     "[][']\\|,[[:space:]]\\(\\)"
     (lambda (s)
       (if (match-end 1) " "
         (pcase-exhaustive (aref s 0)
           (?\[ "(")
           (?\] ")")
           (?\' "\""))))
     results)

It'd be nice to have a front-end that lets you write a kind of lex-like
set of rules, but I think it'd be important for that front-end to allow
*computing* the replacement rather than only selecting it based on the
matched regexp.  Something like

    (foo-replace results
      ("\\[" "(")
      ("\\]" ")")
      ("[[:lower:]]+" (upcase (match-string 0)))
      ("siglo \\([0-9]+\\)"
       (integer-to-roman (string-to-number (match-string 1))))
      ...)


-- Stefan