How to use string-to/as-unibyte correctly?

How to use string-to/as-unibyte correctly?

[Leo Liu]

Hi List,

As part of discussion on bug http://debbugs.gnu.org/23750, I am
wondering what is the deal with string-to/as-unibyte? How to use them
correctly or are they in their way out (obsolete)? Thanks.

,----[ Eli Zaretskii ]
| neither string-to-unibyte not string-as-unibyte should ever be used in
| applications, their use is more often than not a sign of some basic
| misunderstanding of text encoding.
`----

,----[ Eli Zaretskii ]
| Bottom line: like I said, there should be no reason to use
| string-*-unibyte in modern Emacs code on the url-http level or higher
| (maybe not at all). Its use is a sign of some basic misunderstanding,
| or a bug elsewhere, or remnant of old problems that no longer exist.
| So I think we should reconsider the solution on master as well.
`----

Leo

Re: How to use string-to/as-unibyte correctly?

[Stefan Monnier]

> As part of discussion on bug http://debbugs.gnu.org/23750, I am
> wondering what is the deal with string-to/as-unibyte? How to use them
> correctly or are they in their way out (obsolete)? Thanks.

The correct way to use them is to not use them and use
encode/decode-coding-string instead.


        Stefan

Re: How to use string-to/as-unibyte correctly?

[Eli Zaretskii]

> From: Leo Liu <sdl.web@gmail.com>
> Date: Tue, 21 Jun 2016 07:40:31 +0800
> 
> As part of discussion on bug http://debbugs.gnu.org/23750, I am
> wondering what is the deal with string-to/as-unibyte? How to use them
> correctly or are they in their way out (obsolete)?

Don't use them, period.  They are remnants from the era long gone,
before Emacs used Unicode and UTF-8 as the basis for its internal
representation of characters.  Nowadays the only place where such
conversions are sometimes needed are on the C level, deep within Emacs
internals.  Applications should never use them.

I'm okay with declaring them obsolete, if it helps.

Re: How to use string-to/as-unibyte correctly?

[Leo Liu]

On 2016-06-20 20:56 -0400, Stefan Monnier wrote:
[snipped 4 lines]
> The correct way to use them is to not use them and use
> encode/decode-coding-string instead.
>
>
>         Stefan

On 2016-06-21 05:36 +0300, Eli Zaretskii wrote:
[snipped 7 lines]
> Don't use them, period.  They are remnants from the era long gone,
> before Emacs used Unicode and UTF-8 as the basis for its internal
> representation of characters.  Nowadays the only place where such
> conversions are sometimes needed are on the C level, deep within Emacs
> internals.  Applications should never use them.
>
> I'm okay with declaring them obsolete, if it helps.

I had two cases of string-to-unibyte to mean I am expecting unibyte
string and will complain if given something else.

Looks like we'd better start the process of phasing out these functions
so that nobody will make the mistake of using them.

Leo

Re: How to use string-to/as-unibyte correctly?

[Andreas Schwab]

Leo Liu <sdl.web@gmail.com> writes:

> I had two cases of string-to-unibyte to mean I am expecting unibyte
> string and will complain if given something else.

There should be very few places that need a unibyte string.  Basically
only where we are communicating with outside of Emacs, immediately
before decoding or after encoding the string.

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

Re: How to use string-to/as-unibyte correctly?

[Eli Zaretskii]

> From: Leo Liu <sdl.web@gmail.com>
> Date: Wed, 22 Jun 2016 11:33:19 +0800
> 
> Looks like we'd better start the process of phasing out these functions
> so that nobody will make the mistake of using them.

You mean, deprecating them?  Patches to that effect are welcome.

Thanks.

Re: How to use string-to/as-unibyte correctly?

[Leo Liu]

On 2016-06-22 17:59 +0300, Eli Zaretskii wrote:
> You mean, deprecating them?  Patches to that effect are welcome.

Should we deprecate string-to/as-multibyte as well?

Leo

Re: How to use string-to/as-unibyte correctly?

[Eli Zaretskii]

> From: Leo Liu <sdl.web@gmail.com>
> Date: Sat, 25 Jun 2016 09:57:31 +0800
> 
> On 2016-06-22 17:59 +0300, Eli Zaretskii wrote:
> > You mean, deprecating them?  Patches to that effect are welcome.
> 
> Should we deprecate string-to/as-multibyte as well?

Yes.

Re: How to use string-to/as-unibyte correctly?

[Leo Liu]

Filed as http://debbugs.gnu.org/23850

Re: How to use string-to/as-unibyte correctly?

[Leo Liu]

On 2016-06-22 09:49 +0200, Andreas Schwab wrote:
> There should be very few places that need a unibyte string.  Basically
> only where we are communicating with outside of Emacs, immediately
> before decoding or after encoding the string.

Thanks for the input.

Leo