Re: [Emacs-diffs] trunk r117640: * keyboard.c (safe_run_hook_funcall): Avoid consing around

Re: [Emacs-diffs] trunk r117640: * keyboard.c (safe_run_hook_funcall): Avoid consing around

[Stefan Monnier]

>  safe_run_hooks (Lisp_Object hook)
>  {
> -  /* FIXME: our `internal_condition_case' does not provide any way to pass data
> -     to its body or to its handlers other than via globals such as
> -     dynamically-bound variables ;-)  */
>    ptrdiff_t count = SPECPDL_INDEX ();
> +
>    specbind (Qinhibit_quit, hook);
> -
>    run_hook_with_args (1, &hook, safe_run_hook_funcall);
> -
>    unbind_to (count, Qnil);
>  }
 
You remove the comment but don't fix the code accordingly: if the FIXME
is not valid any more, than we should cleanly pass `hook' as argument
rather than let-binding it to inhibit-quit.


        Stefan

Re: [Emacs-diffs] trunk r117640: * keyboard.c (safe_run_hook_funcall): Avoid consing around

[Dmitry Antipov]

On 08/04/2014 11:28 AM, Stefan Monnier wrote:

> You remove the comment but don't fix the code accordingly: if the FIXME
> is not valid any more, than we should cleanly pass `hook' as argument
> rather than let-binding it to inhibit-quit.

Hm...isn't it required to bind inhibit-quit to non-nil value before running
hooks?

Dmitry

Re: [Emacs-diffs] trunk r117640: * keyboard.c (safe_run_hook_funcall): Avoid consing around

[Stefan Monnier]

>> You remove the comment but don't fix the code accordingly: if the FIXME
>> is not valid any more, than we should cleanly pass `hook' as argument
>> rather than let-binding it to inhibit-quit.
> Hm...isn't it required to bind inhibit-quit to non-nil value before running
> hooks?

Yes, binding it to a non-nil value is necessary.  But that doesn't mean
that hook needs to be passed via that dynamic var.


        Stefan