Re: [elpa] master 19862ff: Use font-lock-fontify-region instead of jit-lock-fontify-now

Re: [elpa] master 19862ff: Use font-lock-fontify-region instead of jit-lock-fontify-now

[Stefan Monnier]

> -            (jit-lock-fontify-now (min beg end) (max beg end))))
> +            (font-lock-fontify-region (min beg end) (max beg end))))
 
Maybe even better would be to use font-lock-ensure (when available)?


        Stefan

Re: [elpa] master 19862ff: Use font-lock-fontify-region instead of jit-lock-fontify-now

[Teemu Likonen]

[-- Attachment #1: Type: text/plain, Size: 906 bytes --]

Stefan Monnier [2016-01-30 19:45:23-05] wrote:

>> -            (jit-lock-fontify-now (min beg end) (max beg end))))
>> +            (font-lock-fontify-region (min beg end) (max beg end))))
>  
> Maybe even better would be to use font-lock-ensure (when available)?

It's unclear to me what's the difference between font-lock-ensure and
font-lock-fontify-region. It seems that font-lock-ensure is relatively
new, introduced in 24.4, so maybe not quite yet depend on it on an ELPA
package.

Are there established practices for writing a compatibility code in
situations like this? Perhaps calling a custom function
wcheck--fontify-region and having code like below?

    (eval-when-compile
      (if (fboundp 'font-lock-ensure)
          (defalias 'wcheck--fontify-region 'font-lock-ensure)
        (defun wcheck--fontify-region (beg end)
          (font-lock-fontify-region beg end))))

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]

Re: [elpa] master 19862ff: Use font-lock-fontify-region instead of jit-lock-fontify-now

[Teemu Likonen]

[-- Attachment #1: Type: text/plain, Size: 503 bytes --]

Teemu Likonen [2016-02-07 12:10:55+02] wrote:

> Are there established practices for writing a compatibility code in
> situations like this? Perhaps calling a custom function
> wcheck--fontify-region and having code like below?

There are some "alias wrappers" like below in Emacs code so maybe I'll
use similar with wcheck-mode.

    (eval-when-compile
      (defalias 'wcheck--fontify-region
        (if (fboundp 'font-lock-ensure)
            'font-lock-ensure
          'font-lock-fontify-region)))

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]

Re: [elpa] master 19862ff: Use font-lock-fontify-region instead of jit-lock-fontify-now

[Stefan Monnier]

>>> -            (jit-lock-fontify-now (min beg end) (max beg end))))
>>> +            (font-lock-fontify-region (min beg end) (max beg end))))
>> Maybe even better would be to use font-lock-ensure (when available)?
> It's unclear to me what's the difference between font-lock-ensure and
> font-lock-fontify-region.

font-lock-fontify-region will use font-lock rules and apply them to the
specified region, regardless of what happened before, and without
keeping track of the fact that those rules were applied there.
It's mostly an internal function in font-lock.

font-lock-ensure simply asks Emacs to make sure that the specified
region has had font-lock rules applied.  It's a function specifically
designed to be used by external packages.

> It seems that font-lock-ensure is relatively new, introduced in 24.4,

Indeed, it's brand new.

> so maybe not quite yet depend on it on an ELPA package.

(fboundp 'font-lock-ensure) will take care of that.

> There are some "alias wrappers" like below in Emacs code so maybe I'll
> use similar with wcheck-mode.
>     (eval-when-compile
>       (defalias 'wcheck--fontify-region
>         (if (fboundp 'font-lock-ensure)
>             'font-lock-ensure
>           'font-lock-fontify-region)))

You can use that  [ Tho, note that eval-when-compile is wrong here:
this function will be needed at run-time, not at compile time.  ]
But if there's only a single cal to that function, you might as well
inline it at its only call site.


        Stefan

Re: [elpa] master 19862ff: Use font-lock-fontify-region instead of jit-lock-fontify-now

[Teemu Likonen]

[-- Attachment #1: Type: text/plain, Size: 316 bytes --]

Stefan Monnier [2016-02-07 22:09:46-05] wrote:

>>         (if (fboundp 'font-lock-ensure)
>>             'font-lock-ensure
>>           'font-lock-fontify-region)))

> But if there's only a single cal to that function, you might as well
> inline it at its only call site.

Ok, thanks. I'll push the simple version.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]