How can url-show-status be void when (featurep 'url-vars) is t?

How can url-show-status be void when (featurep 'url-vars) is t?

[Lennart Borgman]

It looks impossible to me that url-show-status is void as a variable
when url-vars.elc has been loaded, but that is what I see here. Also
url-show-status seems to have been initialized by defcustom since it
has (standard-value (t)) in the property list.

url-show-status is a boolean defcustom without any :set function. So
custom-initialize-reset should have created the variable by calling
set-default.

Any ideas how then url-show-status can be void after loading url-vars?

Re: How can url-show-status be void when (featurep 'url-vars) is t?

[Andreas Schwab]

Lennart Borgman <lennart.borgman@gmail.com> writes:

> Any ideas how then url-show-status can be void after loading url-vars?

ELISP> (let (url-show-status) (load "url-vars"))
t
ELISP> (boundp 'url-show-status)
nil

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: How can url-show-status be void when (featurep 'url-vars) is t?

[Chong Yidong]

Lennart Borgman <lennart.borgman@gmail.com> writes:

> It looks impossible to me that url-show-status is void as a variable
> when url-vars.elc has been loaded, but that is what I see here. Also
> url-show-status seems to have been initialized by defcustom since it
> has (standard-value (t)) in the property list.
>
> url-show-status is a boolean defcustom without any :set function. So
> custom-initialize-reset should have created the variable by calling
> set-default.
>
> Any ideas how then url-show-status can be void after loading url-vars?

I can't reproduce this.  Maybe you have a rogue `provide' in your code.

Re: How can url-show-status be void when (featurep 'url-vars) is t?

[Lennart Borgman]

On Sun, Jun 27, 2010 at 8:08 PM, Andreas Schwab <schwab@linux-m68k.org> wrote:
> Lennart Borgman <lennart.borgman@gmail.com> writes:
>
>> Any ideas how then url-show-status can be void after loading url-vars?
>
> ELISP> (let (url-show-status) (load "url-vars"))
> t
> ELISP> (boundp 'url-show-status)
> nil

Ah, thanks. I managed to move (require 'url-vars) inside the let-clause.

That is pretty easy to do when you are restructuring something. I did
not notice until many months later so I had forgotten about the change
of course.

Wouldn't it be possible for defvar, defcustom etc to detect that situation?

Re: How can url-show-status be void when (featurep 'url-vars) is t?

[Kevin Rodgers]

Lennart Borgman wrote:
> On Sun, Jun 27, 2010 at 8:08 PM, Andreas Schwab <schwab@linux-m68k.org> wrote:
>> Lennart Borgman <lennart.borgman@gmail.com> writes:
>>
>>> Any ideas how then url-show-status can be void after loading url-vars?
>> ELISP> (let (url-show-status) (load "url-vars"))
>> t
>> ELISP> (boundp 'url-show-status)
>> nil
> 
> Ah, thanks. I managed to move (require 'url-vars) inside the let-clause.
> 
> That is pretty easy to do when you are restructuring something. I did
> not notice until many months later so I had forgotten about the change
> of course.

`C-h f defvar'

If SYMBOL has a local binding, then this form affects the local
binding.  This is usually not what you want.  Thus, if you need to
load a file defining variables, with this form or with `defconst' or
`defcustom', you should always load that file _outside_ any bindings
for these variables.  (`defconst' and `defcustom' behave similarly in
this respect.)

> Wouldn't it be possible for defvar, defcustom etc to detect that situation?

defvar does this:

	{ /* Check if there is really a global binding rather than just a let
	     binding that shadows the global unboundness of the var.  */
	  volatile struct specbinding *pdl = specpdl_ptr;
	  while (--pdl >= specpdl)
	    {
	      if (EQ (pdl->symbol, sym) && !pdl->func
		  && EQ (pdl->old_value, Qunbound))
		{
		  message_with_string ("Warning: defvar ignored because %s is let-bound",
				       SYMBOL_NAME (sym), 1);
		  break;
		}
	    }
	}

-- 
Kevin Rodgers
Denver, Colorado, USA

Re: How can url-show-status be void when (featurep 'url-vars) is t?

[Lennart Borgman]

On Tue, Jun 29, 2010 at 7:45 AM, Kevin Rodgers
<kevin.d.rodgers@gmail.com> wrote:
> Lennart Borgman wrote:
>>
>> On Sun, Jun 27, 2010 at 8:08 PM, Andreas Schwab <schwab@linux-m68k.org>
>> wrote:
>>>
>>> Lennart Borgman <lennart.borgman@gmail.com> writes:
>>>
>>>> Any ideas how then url-show-status can be void after loading url-vars?
>>>
>>> ELISP> (let (url-show-status) (load "url-vars"))
>>> t
>>> ELISP> (boundp 'url-show-status)
>>> nil
>>
>> Ah, thanks. I managed to move (require 'url-vars) inside the let-clause.
>>
>> That is pretty easy to do when you are restructuring something. I did
>> not notice until many months later so I had forgotten about the change
>> of course.
>
>> Wouldn't it be possible for defvar, defcustom etc to detect that
>> situation?
>
> defvar does this:
>
>        { /* Check if there is really a global binding rather than just a let
>             binding that shadows the global unboundness of the var.  */
>          volatile struct specbinding *pdl = specpdl_ptr;
>          while (--pdl >= specpdl)
>            {
>              if (EQ (pdl->symbol, sym) && !pdl->func
>                  && EQ (pdl->old_value, Qunbound))
>                {
>                  message_with_string ("Warning: defvar ignored because %s is
> let-bound",
>                                       SYMBOL_NAME (sym), 1);
>                  break;
>                }
>            }
>        }


Thanks, but should not that be an error?

Re: How can url-show-status be void when (featurep 'url-vars) is t?

[Stefan Monnier]

>>                  message_with_string ("Warning: defvar ignored because %s is
>> let-bound",
>>                                       SYMBOL_NAME (sym), 1);

> Thanks, but should not that be an error?

Feel free to turn it into an error: I think that's what it *should* do,
but backward compatibility means we should accept it.


        Stefan

Re: How can url-show-status be void when (featurep 'url-vars) is t?

[David Kastrup]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>>>                  message_with_string ("Warning: defvar ignored because %s is
>>> let-bound",
>>>                                       SYMBOL_NAME (sym), 1);
>
>> Thanks, but should not that be an error?
>
> Feel free to turn it into an error: I think that's what it *should* do,
> but backward compatibility means we should accept it.

I don't think we should bother all too much about preserving
compatibility to what never was anything but an error.

-- 
David Kastrup