Re: trunk r117128: Allow any non-nil value to count as true in bool-vector.

Re: trunk r117128: Allow any non-nil value to count as true in bool-vector.

[Michael Albinus]

Paul Eggert <eggert@cs.ucla.edu> writes:

> === modified file 'src/dbusbind.c'
> --- a/src/dbusbind.c	2014-05-19 07:49:09 +0000
> +++ b/src/dbusbind.c	2014-05-19 19:19:05 +0000
> @@ -387,7 +387,8 @@
>        break;
>
>      case DBUS_TYPE_BOOLEAN:
> -      CHECK_BOOLEAN (object);
> +      /* Every Emacs Lisp object serves as a boolean, so there's nothing
> +	 to check.  */
>        sprintf (signature, "%c", dtype);
>        break;
>

That's wrong. dbusbind.c maps Lisp types to D-Bus types, and it expects
exactly t and nil as boolean values, nothing else. See also
(info "(dbus)Type Conversion")

I've reverted your patch (r117128) and the one from Dmitry (r117125).

Best regards, Michael.

Re: trunk r117128: Allow any non-nil value to count as true in bool-vector.

[Davis Herring]

>>        sprintf (signature, "%c", dtype);

sprintf catches the eye, of course.  Why not just write this?

  *signature = dtype;

Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.

Re: trunk r117128: Allow any non-nil value to count as true in bool-vector.

[Michael Albinus]

Davis Herring <herring@lanl.gov> writes:

>>>        sprintf (signature, "%c", dtype);
>
> sprintf catches the eye, of course.  Why not just write this?
>
>   *signature = dtype;

Maybe. But the "%c" format says explicitely "one byte", so I find it
more readable.

Of course, there are different tastes ...

> Davis

Best regards, Michael.