Renaming some functions

Renaming some functions

[Stefan Monnier]

Here's my first concrete renaming proposition:

    multibyte-string-p  =>  string-multibyte-p

Any objection?


        Stefan

Re: Renaming some functions

[Dmitry Gutov]

On 04.05.2020 00:21, Stefan Monnier wrote:
> Here's my first concrete renaming proposition:
> 
>      multibyte-string-p  =>  string-multibyte-p

Yes, please.

Re: Renaming some functions

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Date: Sun, 03 May 2020 17:21:31 -0400
> 
> Here's my first concrete renaming proposition:
> 
>     multibyte-string-p  =>  string-multibyte-p
> 
> Any objection?

We are not renaming anything, we are discussing aliases that could
make discoverability easier.

Re: Renaming some functions

[Stefan Monnier]

>> Here's my first concrete renaming proposition:
>>
>>     multibyte-string-p  =>  string-multibyte-p
>>
>> Any objection?
>
> We are not renaming anything, we are discussing aliases that could
> make discoverability easier.

Yes, and a general discussion is pointless (because I'm not interested
in a general renaming.  Only in a few specific ones).  Hence this specific
proposition which should be considered in isolation.

The purpose is not to make discoverability easier.  It's to make the
namespace more regular (which then brings various potential benefits such
as easier discoverability, better code completion, ...).

This function is used 51 times in Emacs's own source code, and once in
GNU ELPA, so changing it has a fairly low impact.  The only reason
I care about it is because I happen to use it fairly often in `M-:` and
after getting confused between the two names enough times, I added an
alias for myself and now I think it's time we admit that the original
name was simply an error.


        Stefan

Re: Renaming some functions

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: emacs-devel@gnu.org
> Date: Sun, 03 May 2020 22:59:39 -0400
> 
> >>     multibyte-string-p  =>  string-multibyte-p
> >>
> >> Any objection?
> >
> > We are not renaming anything, we are discussing aliases that could
> > make discoverability easier.
> 
> Yes, and a general discussion is pointless (because I'm not interested
> in a general renaming.  Only in a few specific ones).  Hence this specific
> proposition which should be considered in isolation.

You want to actually rename this?  Then you know the drill: leave the
old name as an obsolete alias, wait for 10 years, then remove the
alias.

However, based on what Richard wrote, maybe we should have both of
these functions ;-)

Re: Renaming some functions

[Stefan Monnier]

>> >>     multibyte-string-p  =>  string-multibyte-p
>> >>
>> >> Any objection?
>> >
>> > We are not renaming anything, we are discussing aliases that could
>> > make discoverability easier.
>> 
>> Yes, and a general discussion is pointless (because I'm not interested
>> in a general renaming.  Only in a few specific ones).  Hence this specific
>> proposition which should be considered in isolation.
>
> You want to actually rename this?

Yes.

> Then you know the drill: leave the old name as an obsolete alias, wait
> for 10 years, then remove the alias.

Yup.

> However, based on what Richard wrote, maybe we should have both of
> these functions ;-)

That would be worse than the status quo.


        Stefan

Re: Renaming some functions

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  >> >>     multibyte-string-p  =>  string-multibyte-p

  > > However, based on what Richard wrote, maybe we should have both of
  > > these functions ;-)

I did not suggest that we should define both names.
What I said was that each name goes naturally with a different
meaning.  I think implementing either one of them is sufficient.

-- 
Dr Richard Stallman
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Renaming some functions

[Philippe Vaucher]

>   >> >>     multibyte-string-p  =>  string-multibyte-p
>
>   > > However, based on what Richard wrote, maybe we should have both of
>   > > these functions ;-)
>
> I did not suggest that we should define both names.
> What I said was that each name goes naturally with a different
> meaning.  I think implementing either one of them is sufficient.

The current "multibyte-" prefix contains only a few functions. Maybe
it's too small to
warrant its own category?

Re: Renaming some functions

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > Here's my first concrete renaming proposition:

  >     multibyte-string-p  =>  string-multibyte-p

  > Any objection?

The actual behavior of the function is to return t if the argument is
a multibyte string, nil for any other type.  The clear name for that
is multibyte-string-p.

string-multibyte-p implies a function that accepts only strings as arguments
and returns t if the argument is multibyte.

ISTR that I thought about this when I defined that function, to choose
which of those two names to give it.

-- 
Dr Richard Stallman
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Renaming some functions

[Stefan Monnier]

>   > Here's my first concrete renaming proposition:
>
>   >     multibyte-string-p  =>  string-multibyte-p
>
>   > Any objection?
>
> The actual behavior of the function is to return t if the argument is
> a multibyte string, nil for any other type.  The clear name for that
> is multibyte-string-p.
>
> string-multibyte-p implies a function that accepts only strings as arguments
> and returns t if the argument is multibyte.

AFAICT, it's only ever called with a string as argument (IOW its
behavior on non-strings is basically of no consequence), so both names
work as well from that point of view.


        Stefan