Re: master f17bdee79b1: sgml-mode.el: Cosmetic fixes

Re: master f17bdee79b1: sgml-mode.el: Cosmetic fixes

[Robert Pluim]

>>>>> On Sat, 15 Jul 2023 11:41:39 -0400 (EDT), Stefan Monnier via Mailing list for Emacs changes <emacs-diffs@gnu.org> said:

    Stefan> diff --git a/lisp/textmodes/sgml-mode.el b/lisp/textmodes/sgml-mode.el
    Stefan> index b4f5ebf5cd0..d80cdc34775 100644
    Stefan> --- a/lisp/textmodes/sgml-mode.el
    Stefan> +++ b/lisp/textmodes/sgml-mode.el
    Stefan> @@ -66,7 +66,7 @@ When 2, attribute indentation looks like this:
    Stefan>    </element>"
    Stefan>    :version "25.1"
    Stefan>    :type 'integer
    Stefan> -  :safe 'integerp)
    Stefan> +  :safe #'integerp)

Isnʼt that already enforced by the ':typeʼ keyword? If so, we have
lots of redundant ':safe' in defcustoms.

Robert
-- 

Re: master f17bdee79b1: sgml-mode.el: Cosmetic fixes

[Stefan Monnier]

>     Stefan> diff --git a/lisp/textmodes/sgml-mode.el b/lisp/textmodes/sgml-mode.el
>     Stefan> index b4f5ebf5cd0..d80cdc34775 100644
>     Stefan> --- a/lisp/textmodes/sgml-mode.el
>     Stefan> +++ b/lisp/textmodes/sgml-mode.el
>     Stefan> @@ -66,7 +66,7 @@ When 2, attribute indentation looks like this:
>     Stefan>    </element>"
>     Stefan>    :version "25.1"
>     Stefan>    :type 'integer
>     Stefan> -  :safe 'integerp)
>     Stefan> +  :safe #'integerp)
>
> Isnʼt that already enforced by the ':typeʼ keyword?

No.

> If so, we have lots of redundant ':safe' in defcustoms.

It's not redundant: one describe the set of values that are meaningful
(so as to allow Custom to provide a more helpful UI), the other
specifies which values are believed to be safe when coming from
untrusted sources.

They're closely related and often identical, but not always.


        Stefan

Re: master f17bdee79b1: sgml-mode.el: Cosmetic fixes

[Robert Pluim]

>>>>> On Mon, 17 Jul 2023 08:57:03 -0400, Stefan Monnier <monnier@iro.umontreal.ca> said:

    Stefan> diff --git a/lisp/textmodes/sgml-mode.el b/lisp/textmodes/sgml-mode.el
    Stefan> index b4f5ebf5cd0..d80cdc34775 100644
    Stefan> --- a/lisp/textmodes/sgml-mode.el
    Stefan> +++ b/lisp/textmodes/sgml-mode.el
    Stefan> @@ -66,7 +66,7 @@ When 2, attribute indentation looks like this:
    Stefan> </element>"
    Stefan> :version "25.1"
    Stefan> :type 'integer
    Stefan> -  :safe 'integerp)
    Stefan> +  :safe #'integerp)
    >> 
    >> Isnʼt that already enforced by the ':typeʼ keyword?

    Stefan> No.

If I remove the ':safe' spec, and try to set the value to a string,
M-x customize refuses to apply it, so now I wonder when this failure
to enforce occurs.

    >> If so, we have lots of redundant ':safe' in defcustoms.

    Stefan> It's not redundant: one describe the set of values that are meaningful
    Stefan> (so as to allow Custom to provide a more helpful UI), the other
    Stefan> specifies which values are believed to be safe when coming from
    Stefan> untrusted sources.

    Stefan> They're closely related and often identical, but not always.

Iʼm having a hard time coming up with a situation where they wouldnʼt
be identical, but I guess they exist in the wild.

Robert
-- 

Re: master f17bdee79b1: sgml-mode.el: Cosmetic fixes

[Eli Zaretskii]

> From: Robert Pluim <rpluim@gmail.com>
> Cc: emacs-devel@gnu.org
> Date: Mon, 17 Jul 2023 15:17:36 +0200
> 
> >>>>> On Mon, 17 Jul 2023 08:57:03 -0400, Stefan Monnier <monnier@iro.umontreal.ca> said:
> 
>     Stefan> diff --git a/lisp/textmodes/sgml-mode.el b/lisp/textmodes/sgml-mode.el
>     Stefan> index b4f5ebf5cd0..d80cdc34775 100644
>     Stefan> --- a/lisp/textmodes/sgml-mode.el
>     Stefan> +++ b/lisp/textmodes/sgml-mode.el
>     Stefan> @@ -66,7 +66,7 @@ When 2, attribute indentation looks like this:
>     Stefan> </element>"
>     Stefan> :version "25.1"
>     Stefan> :type 'integer
>     Stefan> -  :safe 'integerp)
>     Stefan> +  :safe #'integerp)
>     >> 
>     >> Isnʼt that already enforced by the ':typeʼ keyword?
> 
>     Stefan> No.
> 
> If I remove the ':safe' spec, and try to set the value to a string,
> M-x customize refuses to apply it, so now I wonder when this failure
> to enforce occurs.

The :safe property is for using the option in file-local variables,
not for customizing it interactively.

Re: master f17bdee79b1: sgml-mode.el: Cosmetic fixes

[Basil Contovounesios]

Robert Pluim [2023-07-17 15:17 +0200] wrote:

>>>>>> On Mon, 17 Jul 2023 08:57:03 -0400, Stefan Monnier <monnier@iro.umontreal.ca> said:
>
>     Stefan> It's not redundant: one describe the set of values that are meaningful
>     Stefan> (so as to allow Custom to provide a more helpful UI), the other
>     Stefan> specifies which values are believed to be safe when coming from
>     Stefan> untrusted sources.
>
>     Stefan> They're closely related and often identical, but not always.
>
> Iʼm having a hard time coming up with a situation where they wouldnʼt
> be identical, but I guess they exist in the wild.

What about user options that can take on one of a set of predefined
atoms, or a custom function: then you may want to declare the defaults
as safe for use in a file-local variable, but not the arbitrary code.

-- 
Basil