bug#24383: 25.1.50; Deprecation/internalization of `read-from-whole-string'

bug#24383: 25.1.50; Deprecation/internalization of `read-from-whole-string'

[Drew Adams]

What's the rationale behind the demotion of `read-from-whole-string'
and its replacement by "internal" function
`thing-at-point--read-from-whole-string'?

  "This is an internal thingatpt function and should not be used."

Should not be used?  Really?  Why?

This function has been around forever, and is used by Emacs users.
And there has been no change in the function definition.

Why should it suddenly be considered "internal" and something not
to be used?

And what does it have to do, in particular, with thingatpt.el?
(It should instead be promoted to a file such as subr.el.)

Based on this (seemingly gratuitous) change, existing user code now
has to fiddle, to be usable in more than one Emacs version (versions
0 to 24.5 vs version 25.1).

 (if (fboundp 'read-from-whole-string)
     (read-from-whole-string strg)
   (thing-at-point--read-from-whole-string strg))

In GNU Emacs 25.1.50.1 (i686-pc-mingw32)
 of 2015-12-10
Repository revision: 6148555ee5a3d0139ae517803718b3e0357933c7
Windowing system distributor 'Microsoft Corp.', version 6.1.7601
Configured using:
 'configure --prefix=/c/Devel/emacs/snapshot/trunk --enable-checking=yes
 --enable-check-lisp-object-type --without-compress-install 'CFLAGS=-Og
 -ggdb3' LDFLAGS=-Lc:/Devel/emacs/lib 'CPPFLAGS=-DGC_MCHECK=1
 -Ic:/Devel/emacs/include''

bug#24383: 25.1.50; Deprecation/internalization of `read-from-whole-string'

[Stefan Kangas]

Drew Adams <drew.adams@oracle.com> writes:

> What's the rationale behind the demotion of `read-from-whole-string'
> and its replacement by "internal" function
> `thing-at-point--read-from-whole-string'?
>
>   "This is an internal thingatpt function and should not be used."
>
> Should not be used?  Really?  Why?
>
> This function has been around forever, and is used by Emacs users.
> And there has been no change in the function definition.
>
> Why should it suddenly be considered "internal" and something not
> to be used?
>
> And what does it have to do, in particular, with thingatpt.el?
> (It should instead be promoted to a file such as subr.el.)
>
> Based on this (seemingly gratuitous) change, existing user code now
> has to fiddle, to be usable in more than one Emacs version (versions
> 0 to 24.5 vs version 25.1).

Glenn, you made the change mentioned above.  Do you have an opinion
here?

Best regards,
Stefan Kangas

bug#24383: 25.1.50; Deprecation/internalization of `read-from-whole-string'

[Lars Ingebrigtsen]

Drew Adams <drew.adams@oracle.com> writes:

> What's the rationale behind the demotion of `read-from-whole-string'
> and its replacement by "internal" function
> `thing-at-point--read-from-whole-string'?
>
>   "This is an internal thingatpt function and should not be used."
>
> Should not be used?  Really?  Why?

"Should not be used" is perhaps over-stating the case, but it's an
internal thingatpt function with a name that didn't reflect that, so
renaming it is appropriate.

> Based on this (seemingly gratuitous) change, existing user code now
> has to fiddle, to be usable in more than one Emacs version (versions
> 0 to 24.5 vs version 25.1).
>
>  (if (fboundp 'read-from-whole-string)
>      (read-from-whole-string strg)
>    (thing-at-point--read-from-whole-string strg))

It's up to you whether you want to use internal Emacs functions, of
course.

Closing.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#24383: [External] : Re: bug#24383: 25.1.50; Deprecation/internalization of `read-from-whole-string'

[Drew Adams]

> > What's the rationale behind the demotion of
> > `read-from-whole-string' and its replacement by
> > "internal" function
> > `thing-at-point--read-from-whole-string'?

?

> >   "This is an internal thingatpt function and should not be used."
> >
> > Should not be used?  Really?  Why?
> 
> "Should not be used" is perhaps over-stating the case, but it's an
> internal thingatpt function with a name that didn't reflect that, so
> renaming it is appropriate.

What makes it "internal"?  The name change is
based on it being internal, so the reason it's
considered internal can't be because of its name.

Does it have anything at all to do specifically
with thing-at-point?  What makes it a candidate
for being designated "internal"?

Does anything in the code or the doc string
suggest thing-at-point?

 "Read a Lisp expression from STR.
  Signal an error if the entire string was not used."

> It's up to you whether you want to use internal
> Emacs functions, of course.  Closing.

The alternative to using this function is to
reproduce its definition.  Is that better?

Wouldn't it be better to move this out of
thingatpt.el?  Isn't this function generally
useful?

I've copied the definition to two libraries,
unrelated to each other.  That says something
about how internal this is, no?