Re: ord(), chr(), and &url.asciitbale.com;
| From: | Christoph M. Becker | Date: | Fri, 22 Jun 2018 10:48:15 +0000 |
| Subject: | Re: ord(), chr(), and &url.asciitbale.com; | ||
| References: | 1 | Groups: | php.doc |
| Request: | Send a blank email to phpdoc+get-969386978@lists.php.net to get a copy of this message | ||
On 10.06.2018 at 19:59, Rowan Collins wrote:
> I'd like to raise a discussion of the patch I've just proposed at
> https://edit.php.net/?patchID=2939&project=PHP because
> it's a relatively
> major change.
>
> Currently, the ord() and chr() manual pages describe their input and
> output in terms of "ASCII" and "characters", which is misleading for a
> number of reasons:
>
> 1) ASCII is a 7-bit encoding, but these functions work with 8-bit
> numbers (0 to 255)
> 2) The term "character" is often used in the context of multibyte
> encodings, but these functions *always* work with a single byte
> 3) Neither function is encoding aware in any way, and both can in fact
> be used on arbitrary binary strings
>
> In the patch linked above, I've had a go at completely rewording these
> manual pages to make this clearer. Unfortunately, the "preview" function
> in the online editor doesn't work for me, so I may have made some
> formatting errors, but I'd also like feedback on the proposed wording.
Thanks. In my opinion, this is an overall improvement! However, I'm
not happy with the <refpurpose>s which appear to be far too general.
Regarding the changed parameter name: we should actually use the names
which are returned by Reflection, which is $codepoint for chr() (cf.
<https://3v4l.org/CrcS2>).
> Finally, both pages currently link to http://www.asciitable.com
> which
> has a reasonable table for ASCII values, and then one labelled "extended
> ASCII", which appears to be the top half of DOS Code Page 437, of all
> things. I think this should be replaced with a more suitable link, but
> I'm not sure where to; Wikipedia has good charts of encodings, for
> instance, but what page would we link to? ASCII, ISO-8859-1, Windows
> 1252, something else?
Good question! Maybe we should remove the link altogether?
--
Christoph M. Becker