Re: [Emacs-diffs] master 8764544: Treat control characters in JSON strings as invalid

Re: [Emacs-diffs] master 8764544: Treat control characters in JSON strings as invalid

[Stefan Monnier]

> -  (or (char-after (point)) :json-eof))

I wonder why it used :json-eof instead of using nil ?
Any idea?


        Stefan

Re: [Emacs-diffs] master 8764544: Treat control characters in JSON strings as invalid

[Mark Oteiza]

On 18/08/17 at 03:51am, Stefan Monnier wrote:
>> -  (or (char-after (point)) :json-eof))
>
>I wonder why it used :json-eof instead of using nil ?
>Any idea?

Not sure--it has been there since json.el was added to Emacs.

Re: [Emacs-diffs] master 8764544: Treat control characters in JSON strings as invalid

[Stefan Monnier]

>>> -  (or (char-after (point)) :json-eof))
>> I wonder why it used :json-eof instead of using nil ?
>> Any idea?
> Not sure--it has been there since json.el was added to Emacs.

Then why use 0 now instead of nil?


        Stefan

Re: [Emacs-diffs] master 8764544: Treat control characters in JSON strings as invalid

[Philipp Stephani]

[-- Attachment #1: Type: text/plain, Size: 418 bytes --]

Stefan Monnier <monnier@iro.umontreal.ca> schrieb am Fr., 18. Aug. 2017 um
22:25 Uhr:

> >>> -  (or (char-after (point)) :json-eof))
> >> I wonder why it used :json-eof instead of using nil ?
> >> Any idea?
> > Not sure--it has been there since json.el was added to Emacs.
>
> Then why use 0 now instead of nil?
>
>
Because `following-char' returns 0 at EOB. "At the end of the buffer or
accessible region, return 0."

[-- Attachment #2: Type: text/html, Size: 756 bytes --]

Re: [Emacs-diffs] master 8764544: Treat control characters in JSON strings as invalid

[Stefan Monnier]

>> >>> -  (or (char-after (point)) :json-eof))
>> >> I wonder why it used :json-eof instead of using nil ?
>> >> Any idea?
>> > Not sure--it has been there since json.el was added to Emacs.
>> Then why use 0 now instead of nil?
> Because `following-char' returns 0 at EOB. "At the end of the buffer or
> accessible region, return 0."

But it also returns 0 in other cases.  E.g. if you had changed

    (or (char-after (point)) :json-eof)
to
    (char-after)

it would give you nil at EOB (instead of 0).


        Stefan

Re: [Emacs-diffs] master 8764544: Treat control characters in JSON strings as invalid

[Philipp Stephani]

[-- Attachment #1: Type: text/plain, Size: 830 bytes --]

Stefan Monnier <monnier@iro.umontreal.ca> schrieb am Do., 28. Sep. 2017 um
23:22 Uhr:

> >> >>> -  (or (char-after (point)) :json-eof))
> >> >> I wonder why it used :json-eof instead of using nil ?
> >> >> Any idea?
> >> > Not sure--it has been there since json.el was added to Emacs.
> >> Then why use 0 now instead of nil?
> > Because `following-char' returns 0 at EOB. "At the end of the buffer or
> > accessible region, return 0."
>
> But it also returns 0 in other cases.  E.g. if you had changed
>
>     (or (char-after (point)) :json-eof)
> to
>     (char-after)
>
> it would give you nil at EOB (instead of 0).
>
>
Agreed.
I'm wondering what the purpose of `following-char' is. Even in the rare
cases where you want to treat EOB and a following null character
identically it would be cleaner to write (or (char-after) 0).

[-- Attachment #2: Type: text/html, Size: 1254 bytes --]

Re: [Emacs-diffs] master 8764544: Treat control characters in JSON strings as invalid

[Stefan Monnier]

> I'm wondering what the purpose of `following-char' is.

I think it's a historical accident,


        Stefan