Re: A problem in using font-lock-add-keywords

[Dmitry Gutov <dgutov@yandex.ru>]

On 05.03.2013 13:33, source liu wrote:
> My value of `blink-matching-paren-distance' is 102400
> and when I type M-102400 C-f, it jump to the very place from where it
> can't match ")",

That's to be expected, see the description of the above variable.

it failed "M-x eval-buff" when the chars are beyond
> the range

This isn't. Works for me, I don't have any problems with evaluating your 
buffer.

> Come back to font-lock problem,  i can customize the font-lock
> keywords when the list is short, but it is not the case when i use the
> list defined in my attach.
> if you have time, please take a look at it.  many thanks
>
> in my .emacs file, i wrote
>   ================================================================
> (load-file  "~/.hyperworks.el")
> (add-hook 'tcl-mode-hook
> 	  (lambda ()
> 	    (font-lock-add-keywords nil
> 	     `((,(regexp-opt hypermesh-command-list2  t) . font-lock-constant-face)))))
>
> (display-time)
> =================================================================
>
> and hypermesh-command-list2 is defined in .hyperworks.el, in the attach file.
>
> The case is that, when i use `hypermesh-command-list2', it works ok,
> and i changed to `hypermesh-command-list1', it doesn't work, and it
> ruined the default font-lock-mode in tcl-mode.  the only deference
> between the two list is list length, i think.
>
> could you help? many thanks.

Here's how I debugged it:

1. Have the buggy buffer selected.
2. M-x toggle-debug-on-error
3. M-: (font-lock-fontify-region (point-min) (point-max))

At this point you should see the backtrace and that `re-search-forward' 
aborts complaining that the regexp is too big. It's not a restriction on 
the size of any Emacs primitive, the restriction is in the regexp engine.

I'm sure what to recommend here. Split the list into several, maybe.