Re:Re: [patch] * lisp/international/mule-cmds.el (deactivate-input-method): Tiny improve.

[tumashu <tumashu@163.com>]

[-- Attachment #1: Type: text/plain, Size: 1177 bytes --]

maybe, but test a variable value is really a function before funcall it, seem to be better way, I think :-)





--
发自我的网易邮箱手机智能版
<br/><br/><br/>


----- Original Message -----
From: "Eli Zaretskii" <eliz@gnu.org>
To: tumashu <tumashu@163.com>
Cc: emacs-devel@gnu.org
Sent: Wed, 28 Apr 2021 18:03:10 +0300
Subject: Re: [patch] * lisp/international/mule-cmds.el (deactivate-input-method): Tiny improve.

> Date: Wed, 28 Apr 2021 20:33:23 +0800 (CST)
> From: tumashu  <tumashu@163.com>
> Cc: "emacs-devel@gnu.org" <emacs-devel@gnu.org>
> 
> >> -   (funcall deactivate-current-input-method-function))
> >> +          (when (functionp deactivate-current-input-method-function)
> >> +     (funcall deactivate-current-input-method-function)))
> >
> >Thanks, but could you please explain why this is needed?  IOW, in
> >which use case did you see this function called when
> >deactivate-current-input-method-function wasn't a function?
> 
> Yes, I have faced some bugs of pyim in evil environment, sometimes, this variable maybe is nil.
> I can not find why.

I see traces of using defadvise there.  Could it be that the advice
calls deactivate-input-method incorrectly?

[-- Attachment #2: Type: text/html, Size: 2099 bytes --]