Mark for refontification

Mark for refontification

[Lennart Borgman]

Just one of my usual complaints: Could we please include a function
for marking a region for refontification? (Something like the function
mumamo-mark-for-refontification perhaps.)

Please also move the definition of the defmacro save-buffer-state
outside of eval-when-compile so that it could be used in other places
than font-lock.el. (Would it be enough to change to eval-and-compile?)

;;;###autoload
(defun mumamo-mark-for-refontification (min max)
  "Mark region between MIN and MAX for refontification."
    (save-restriction
      (widen)
      ;;(mumamo-with-buffer-prepared-for-jit-lock
      (mumamo-save-buffer-state nil
        (put-text-property min max 'fontified nil)
        )))

Re: Mark for refontification

[Stefan Monnier]

> Please also move the definition of the defmacro save-buffer-state
> outside of eval-when-compile so that it could be used in other places
> than font-lock.el.

We have with-silent-modifications.

> ;;;###autoload
> (defun mumamo-mark-for-refontification (min max)
>   "Mark region between MIN and MAX for refontification."
>     (save-restriction
>       (widen)
>       ;;(mumamo-with-buffer-prepared-for-jit-lock
>       (mumamo-save-buffer-state nil
>         (put-text-property min max 'fontified nil)
>         )))

Why widen?


        Stefan

Re: Mark for refontification

[Lennart Borgman]

On Fri, Apr 9, 2010 at 4:53 AM, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
>> Please also move the definition of the defmacro save-buffer-state
>> outside of eval-when-compile so that it could be used in other places
>> than font-lock.el.
>
> We have with-silent-modifications.


Thanks, I did not notice.


>> ;;;###autoload
>> (defun mumamo-mark-for-refontification (min max)
>>   "Mark region between MIN and MAX for refontification."
>>     (save-restriction
>>       (widen)
>>       ;;(mumamo-with-buffer-prepared-for-jit-lock
>>       (mumamo-save-buffer-state nil
>>         (put-text-property min max 'fontified nil)
>>         )))
>
> Why widen?


You mean it is not necessary for put-text-property?

Re: Mark for refontification

[Stefan Monnier]

>>> (defun mumamo-mark-for-refontification (min max)
>>>   "Mark region between MIN and MAX for refontification."
>>>     (save-restriction
>>>       (widen)
>>>       ;;(mumamo-with-buffer-prepared-for-jit-lock
>>>       (mumamo-save-buffer-state nil
>>>         (put-text-property min max 'fontified nil)
>>>         )))
>> 
>> Why widen?

> You mean it is not necessary for put-text-property?

No, my question doesn't have anything to do with put-text-property.
It's just that usually Elisp functions don't take buffer positions
outside of point-min...point-max and I see no reason why this one should
be an exception.


        Stefan

Re: Mark for refontification

[Lennart Borgman]

On Sat, Apr 10, 2010 at 2:57 AM, Stefan Monnier
<monnier@iro.umontreal.ca> wrote:
>
> >>> (defun mumamo-mark-for-refontification (min max)
> >>>   "Mark region between MIN and MAX for refontification."
> >>>     (save-restriction
> >>>       (widen)
> >>>       ;;(mumamo-with-buffer-prepared-for-jit-lock
> >>>       (mumamo-save-buffer-state nil
> >>>         (put-text-property min max 'fontified nil)
> >>>         )))
> >>
> >> Why widen?
>
> > You mean it is not necessary for put-text-property?
>
> No, my question doesn't have anything to do with put-text-property.
> It's just that usually Elisp functions don't take buffer positions
> outside of point-min...point-max and I see no reason why this one should
> be an exception.


I can't remember why I did it this way, but it is probably due to that
any change in the chunk dividing can result in that the rest of the
buffer should be refontified.

That can of course be optimized since it really may not need
refontification. I am trying to do that and reuse old chunk
fontification, but I don't remember how far I get with this at the
moment. (I have finally rewritten the chunk dividing routines so that
chunks always is part of a chain starting from the top of the buffer.
As usual the code can be improved.)

Still, when I find that an old chunk really is invalid I must mark (at
least part of) it for refontification.

Re: Mark for refontification

[Stefan Monnier]

> Still, when I find that an old chunk really is invalid I must mark (at
> least part of) it for refontification.

That's perfectly fine.  But if *you* need to refer to buffer positions
outside of point-min...point-max then *you* need to widen, not
"mark-for-refontification",


        Stefan