Define multi-character parentheses in Emacs

Question

In Verilog you enclose code blocks between the words 'begin' and 'end' like this:

if(foo) begin
   x <= 1'b0;
   y <= 1'b0;
end else begin
   x <= x_d;
   y <= y_d;
end

Is there any way to set begin and end as parentheses in Emacs, so that you can use check-parens or similar to find any that are mismatched?

I've tried adding this (and variations of) to my ~/.emacs file, but it doesn't like it...

(modify-syntax-entry ?begin "(end" )
(modify-syntax-entry ?end ")begin" )

Thanks.

Answer 1

Sadly, Emacs's parenthesis matching infrastructure doesn't understand much of multi-character tokens. The new library SMIE added in Emacs-23.4 is partly intended to address this issue. It lets major modes describe the syntax of the language (in a very limited kind of grammar) after which things like C-M-f and C-M-b will know how to jump over logical elements, e.g. skip from a begin to its matching end. As LindyDancer notes, such things are usually needed for indentation, and indeed the main motivation behind SMIE was to provide a generic indentation engine.

Now Verilog doesn't use SMIE but it does implement similar navigation commands. So you could probably try something like

(defun sm-verilog-check-parens ()
  (save-excursion
    (goto-char (point-min))
    (while (not (eobp))
      (verilog-forward-sexp))))

Tho I do not know if verilog-forward-sexp will give you appropriate warnings/errors if it bumps into the end of buffer at an unexpected moment.