jit-lock simplification?

jit-lock simplification?

[Kim F. Storm]

I was debugging a timer problem, and noticed some very odd lambda
forms in the timer-list.  It turns out to be the lambda generated by
the jit-lock code below.

Wouldn't it work just as well with the following patch?


*** jit-lock.el	25 Aug 2006 08:57:52 +0200	1.55
--- jit-lock.el	13 Sep 2006 14:12:18 +0200	
***************
*** 397,415 ****
             ;; eagerly extend the refontified region with
             ;; jit-lock-after-change-extend-region-functions.
             (when (< start orig-start)
!              (lexical-let ((start start)
!                            (orig-start orig-start)
!                            (buf (current-buffer)))
!                (run-with-timer
!                 0 nil (lambda ()
!                         (with-current-buffer buf
!                           (with-buffer-prepared-for-jit-lock
!                               (put-text-property start orig-start
!                                                  'fontified t)))))))
  
  	   ;; Find the start of the next chunk, if any.
  	   (setq start (text-property-any next end 'fontified nil))))))))
  
  \f
  ;;; Stealth fontification.
  
--- 397,415 ----
             ;; eagerly extend the refontified region with
             ;; jit-lock-after-change-extend-region-functions.
             (when (< start orig-start)
! 	     (run-with-timer 0 nil 'jit-lock-fontify-again
! 			     (current-buffer) start orig-start))
  
  	   ;; Find the start of the next chunk, if any.
  	   (setq start (text-property-any next end 'fontified nil))))))))
  
+ (defun jit-lock-fontify-again (buf start end)
+   "Fontify in buffer BUF from START to END."
+   (with-current-buffer buf
+     (with-buffer-prepared-for-jit-lock
+      (put-text-property start end 'fontified t))))
+ 
+ 
  \f
  ;;; Stealth fontification.
  

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: jit-lock simplification?

[Stefan Monnier]

> I was debugging a timer problem, and noticed some very odd lambda
> forms in the timer-list.  It turns out to be the lambda generated by
> the jit-lock code below.

> Wouldn't it work just as well with the following patch?

I see no reason why it shouldn't work as well.
I just find it less elegant ;-)


        Stefan


> *** jit-lock.el	25 Aug 2006 08:57:52 +0200	1.55
> --- jit-lock.el	13 Sep 2006 14:12:18 +0200	
> ***************
> *** 397,415 ****
>              ;; eagerly extend the refontified region with
>              ;; jit-lock-after-change-extend-region-functions.
>              (when (< start orig-start)
> !              (lexical-let ((start start)
> !                            (orig-start orig-start)
> !                            (buf (current-buffer)))
> !                (run-with-timer
> !                 0 nil (lambda ()
> !                         (with-current-buffer buf
> !                           (with-buffer-prepared-for-jit-lock
> !                               (put-text-property start orig-start
> !                                                  'fontified t)))))))
  
>   	   ;; Find the start of the next chunk, if any.
>   	   (setq start (text-property-any next end 'fontified nil))))))))
  
>   \f
>   ;;; Stealth fontification.
  
> --- 397,415 ----
>              ;; eagerly extend the refontified region with
>              ;; jit-lock-after-change-extend-region-functions.
>              (when (< start orig-start)
> ! 	     (run-with-timer 0 nil 'jit-lock-fontify-again
> ! 			     (current-buffer) start orig-start))
  
>   	   ;; Find the start of the next chunk, if any.
>   	   (setq start (text-property-any next end 'fontified nil))))))))
  
> + (defun jit-lock-fontify-again (buf start end)
> +   "Fontify in buffer BUF from START to END."
> +   (with-current-buffer buf
> +     (with-buffer-prepared-for-jit-lock
> +      (put-text-property start end 'fontified t))))
> + 
> + 
>   \f
>   ;;; Stealth fontification.
  

> -- 
> Kim F. Storm <storm@cua.dk> http://www.cua.dk



> _______________________________________________
> Emacs-devel mailing list
> Emacs-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: jit-lock simplification?

[David Kastrup]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> I was debugging a timer problem, and noticed some very odd lambda
>> forms in the timer-list.  It turns out to be the lambda generated by
>> the jit-lock code below.
>
>> Wouldn't it work just as well with the following patch?
>
> I see no reason why it shouldn't work as well.
> I just find it less elegant ;-)

Disagree.  If `run-with-timer' has the possibility of passing values,
that should be preferred over `lexical-let'.  Whether one wants to use
a named or an anonymous lambda function is, in contrast, more a matter
of taste.

If this were Scheme, using a closure would more or less be natural,
but in Emacs Lisp, this is just ugly.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: jit-lock simplification?

[Stefan Monnier]

>>>>> "David" == David Kastrup <dak@gnu.org> writes:

> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>>> I was debugging a timer problem, and noticed some very odd lambda
>>> forms in the timer-list.  It turns out to be the lambda generated by
>>> the jit-lock code below.
>> 
>>> Wouldn't it work just as well with the following patch?
>> 
>> I see no reason why it shouldn't work as well.
>> I just find it less elegant ;-)

> Disagree.  If `run-with-timer' has the possibility of passing values,
> that should be preferred over `lexical-let'.

My opinion obviously differs.  More specifically, I consider this ability to
pass extra values to be nothing but a hack to work around the lack of
lexical scoping in plain elisp.  Luckily, CL does provide us with lexical
scoping, and there's even hope that some future Emacs will support lexical
scoping natively.

But really, this is all a matter of taste.  If you guys want to change my
code, go for it.


        Stefan

Re: jit-lock simplification?

[Kim F. Storm]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> I was debugging a timer problem, and noticed some very odd lambda
>> forms in the timer-list.  It turns out to be the lambda generated by
>> the jit-lock code below.
>
>> Wouldn't it work just as well with the following patch?
>
> I see no reason why it shouldn't work as well.
> I just find it less elegant ;-)

I disagree, but YMMV.

Also, it's definitely not elegant when you look at the value of
timer-list :-)

Compare:

  ([nil 17671 61974 726488 nil (lambda (&rest --cl-rest--) (apply #[(G40900 G40901 G40902) "r\bJq\210\306 \x19\307\216\310\211\211\211\311\211\211\x1a^[\x1c\x1d\x1e\f\x1e\r\x1e\x0e\312\x0e\x0fJ\x0e\x10J\313\310$.
\207" [G40900 modified buffer-file-truename buffer-file-name deactivate-mark inhibit-modification-hooks buffer-modified-p ((byte-code "\b\204\b\0\301\302!\210\302\207" [modified restore-buffer-modified-p nil] 2)) t nil put-text-property fontified inhibit-point-motion-hooks inhibit-read-only buffer-undo-list G40902 G40901] 7] (quote --buf--) (quote --orig-start--) (quote --start--) --cl-rest--)) nil nil] [t 17671 61975 225097 0.5 blink-cursor-timer-function nil nil] [nil 17671 62024 0 60 display-time-event-handler nil nil])

to:

  ([nil 17671 61974 726488 nil jit-lock-fontify-again nil nil] [t 17671 61975 225097 0.5 blink-cursor-timer-function nil nil] [nil 17671 62024 0 60 display-time-event-handler nil nil])

>
>
>> *** jit-lock.el	25 Aug 2006 08:57:52 +0200	1.55
>> --- jit-lock.el	13 Sep 2006 14:12:18 +0200	
>> ***************
>> *** 397,415 ****
>>              ;; eagerly extend the refontified region with
>>              ;; jit-lock-after-change-extend-region-functions.
>>              (when (< start orig-start)
>> !              (lexical-let ((start start)
>> !                            (orig-start orig-start)
>> !                            (buf (current-buffer)))
>> !                (run-with-timer
>> !                 0 nil (lambda ()
>> !                         (with-current-buffer buf
>> !                           (with-buffer-prepared-for-jit-lock
>> !                               (put-text-property start orig-start
>> !                                                  'fontified t)))))))
>   
>>   	   ;; Find the start of the next chunk, if any.
>>   	   (setq start (text-property-any next end 'fontified nil))))))))
>   
>>   \f
>>   ;;; Stealth fontification.
>   
>> --- 397,415 ----
>>              ;; eagerly extend the refontified region with
>>              ;; jit-lock-after-change-extend-region-functions.
>>              (when (< start orig-start)
>> ! 	     (run-with-timer 0 nil 'jit-lock-fontify-again
>> ! 			     (current-buffer) start orig-start))
>   
>>   	   ;; Find the start of the next chunk, if any.
>>   	   (setq start (text-property-any next end 'fontified nil))))))))
>   
>> + (defun jit-lock-fontify-again (buf start end)
>> +   "Fontify in buffer BUF from START to END."
>> +   (with-current-buffer buf
>> +     (with-buffer-prepared-for-jit-lock
>> +      (put-text-property start end 'fontified t))))
>> + 
>> + 
>>   \f
>>   ;;; Stealth fontification.
>   
>
>> -- 
>> Kim F. Storm <storm@cua.dk> http://www.cua.dk
>
>
>
>> _______________________________________________
>> Emacs-devel mailing list
>> Emacs-devel@gnu.org
>> http://lists.gnu.org/mailman/listinfo/emacs-devel
>
>

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: jit-lock simplification?

[Richard Stallman]

I think it is clearer for maintenance if the timer function
has a name.

Re: jit-lock simplification?

[Kim F. Storm]

Richard Stallman <rms@gnu.org> writes:

> I think it is clearer for maintenance if the timer function
> has a name.

Ok. I installed my change.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk