RE: Nested Regexp's

[Doug Lewan]

> -----Original Message-----
> From: help-gnu-emacs-bounces+dougl=shubertticketing.com@gnu.org
> [mailto:help-gnu-emacs-bounces+dougl=shubertticketing.com@gnu.org] On
> Behalf Of Thorsten Jolitz
> Sent: Tuesday, 2012 September 18 12:41
> To: help-gnu-emacs@gnu.org
> Subject: Nested Regexp's
> 
> 
> Hi List,
> 
> I'm somehow stuck with a regexp I need for fontification purposes.

Thorsten,

It helps me a lot to think of regular expressions like programs and therefore 
to structure them like programs.

Here's how I define the REs that I think you want. (I assume that by "nested" 
you mean multiple.) I hope this helps.

,Doug

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(defconst my-regex-plain-text 
  (concat "\\([[:space:]]*[^}]+\\)[[:space:]]*" ; Matches "123$%^ Чебурашка 
&*(0-="
          )
  "Regular expression defining what 'plain text' is.")

(defconst  my-regex-bold-text
  (concat "\\(!{\\)"
          my-regex-plain-text
          "\\(}\\)")
  "Regular expression defining what 'bold text' is.")

(defconst my-regex-text
  (concat "\\("
          my-regex-plain-text
          "\\|"
          my-regex-bold-text
          "\\)")
  "Regular expression defining what 'text'.
Text is a mix of plain text and bold text.")

(defconst my-regex-list-item
  (concat "\\(-{\\)"
          my-regex-text "+"
          "\\(}\\)")
  "Regular expression defining what a 'list item' is.")

;; 
;; Sunny day test code
;; 
(defconst test-plain-text (list "foo"
                                "foo bar "
                                "       foo bar baz bat"
                                " --- 123$%^ Чебурашка &*(0-= --- "))
(defconst test-bold-text (mapcar (lambda (text)
                                   (concat "!{" text "}"))
                                 test-plain-text))
(defconst test-list-item (mapcar (lambda (list-text)
                                   (concat "-{" list-text "}"))
                                 (append test-plain-text test-bold-text)))

(mapc (lambda (test-spec)
        (let ((re (car test-spec))
              (test-data (cdr test-spec)))
          (mapc (lambda (item)
                  (if (string-match re item)
                      (message "PASS -- [[%s]] matches [[%s]]" re item)
                    (message "FAIL -- [[%s]] DIDN'T match [[%s]]" re item))
                  (sit-for 1))
                test-data)))
      (list (cons my-regex-plain-text test-plain-text)
            (cons my-regex-bold-text test-bold-text)
            (cons my-regex-list-item test-list-item)))
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

> 
> This is the syntax:
> 
> ,--------------------
> | -{bar}  ; list item
> | !{foo}  ; bold
> `--------------------
> 
> These are examples of nesting bold text into list items:
> 
> ,-------------------------------------------------------------------
> | -{foo bar foo bar}            ; fontified as list item
> | -{!{bold} foo bar}            ; bold fontified, list item text not
> | -{foo bar                     ; fontified as list item
> |    foo bar}
> | -{foo bar !{hello             ; 'hello bold' fontified, list item
> text not
> |    bold} foo bar}
> `-------------------------------------------------------------------
> 
> These are the regexps:
> 
> ,--------------------------------------------------
> | (defconst my-regex-list-item
> |   "\\(-{\\)\\([
> | ]*[^}]+\\)\\(}\\)"
> |   "Regular expression for matching a list item.")
> |
> | (defconst my-regex-bold
> |   "\\(!{\\)\\([
> | ]*[^}]+\\)\\(}\\)"
> |   "Regular expression for matching bold text.")
> `--------------------------------------------------
> 
> Now I would need to construct a 'my-regex-list-item' that allows for
> one
> or several nested 'bold' terms, with each the list item and the bold
> term
> possibly including line-breaks. The result should actually show the
> whole list item in the assigned color, and the bold terms in the same
> color, but bold.
> 
> I would appreciate any hint about how to construct the list-item regexp
> in a way that it can contain nested bold terms.
> 
> --
> cheers,
> Thorsten
> 
>