Re: [Emacs-diffs] master a3b2101 1/2: New file with obarray functions.

Re: [Emacs-diffs] master a3b2101 1/2: New file with obarray functions.

[Stefan Monnier]

>     New file with obarray functions.
>     * lisp/obarray.el: basic obarray functions extracted from abbrev.el
>     * test/automated/obarray-tests.el: new file

Funnily enough, I think we should discourage the use of obarrays.
There are times/places where they are really indispensable (when we
really need to handle *symbols*), but in the majority of cases where
I could see this obarray.el being used, a hash-table would work better
(e.g. hash-table-p is more reliable, print+read actually works, access
is usually faster, ...).


        Stefan

Re: [Emacs-diffs] master a3b2101 1/2: New file with obarray functions.

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Date: Wed, 11 Nov 2015 15:26:53 -0500
> Cc: Przemysław Wojnowski <esperanto@cumego.com>
> 
> Funnily enough, I think we should discourage the use of obarrays.

At the very least we should fix these new warnings:

    ELC      obarray.elc

  In toplevel form:
  obarray.el:45:1:Warning: Lexical argument shadows the dynamic variable obarray
  obarray.el:50:1:Warning: Lexical argument shadows the dynamic variable obarray
  obarray.el:55:1:Warning: Lexical argument shadows the dynamic variable obarray
  obarray.el:60:1:Warning: Lexical argument shadows the dynamic variable obarray

Re: [Emacs-diffs] master a3b2101 1/2: New file with obarray functions.

[Przemysław Wojnowski]

W dniu 11.11.2015 o 21:26, Stefan Monnier pisze:
>>      New file with obarray functions.
>>      * lisp/obarray.el: basic obarray functions extracted from abbrev.el
>>      * test/automated/obarray-tests.el: new file
>
> Funnily enough, I think we should discourage the use of obarrays.
> There are times/places where they are really indispensable (when we
> really need to handle *symbols*), but in the majority of cases where
> I could see this obarray.el being used, a hash-table would work better
> (e.g. hash-table-p is more reliable, print+read actually works, access
> is usually faster, ...).
>
>
>          Stefan

Thanks for feedback.

I just extracted obarray functionality from abbrev.el. This separates 
abbrev responsibility from underlying collection type to make both 
easier to test, understand and change if needed. Replace of 
implementation was not purpose of this commit.

Re: [Emacs-diffs] master a3b2101 1/2: New file with obarray functions.

[Artur Malabarba]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> 
> Funnily enough, I think we should discourage the use of obarrays.

I do to, but I like this change. `abbrev.el' is easier to understand now
and we have more tests.


Eli Zaretskii <eliz@gnu.org> writes:

> At the very least we should fix these new warnings:
>
>     ELC      obarray.elc
>
>   In toplevel form:
>   obarray.el:45:1:Warning: Lexical argument shadows the dynamic variable obarray
>   obarray.el:50:1:Warning: Lexical argument shadows the dynamic variable obarray
>   obarray.el:55:1:Warning: Lexical argument shadows the dynamic variable obarray
>   obarray.el:60:1:Warning: Lexical argument shadows the dynamic variable obarray

Done.

Re: [Emacs-diffs] master a3b2101 1/2: New file with obarray 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. ]]]

  > Funnily enough, I think we should discourage the use of obarrays.
  > There are times/places where they are really indispensable (when we
  > really need to handle *symbols*), but in the majority of cases where
  > I could see this obarray.el being used, a hash-table would work better
  > (e.g. hash-table-p is more reliable, print+read actually works, access
  > is usually faster, ...).

Hash tables are a lot newer -- for a long time, obarrays were the only
Emacs data type that did hashing.  Perhaps that is why obarrays are used
in some places where hash tables would be better.

-- 
Dr Richard Stallman
President, Free Software Foundation (gnu.org, fsf.org)
Internet Hall-of-Famer (internethalloffame.org)
Skype: No way! See stallman.org/skype.html.

Re: [Emacs-diffs] master a3b2101 1/2: New file with obarray functions.

[Stefan Monnier]

> Hash tables are a lot newer -- for a long time, obarrays were the only
> Emacs data type that did hashing.

Of course, it's all relative: AFAIK they were introduced in Emacs-21
(i.e. in 2001), so they've already existed for almost half of Emacs's life.

> Perhaps that is why obarrays are used in some places where hash tables
> would be better.

I indeed also believe that's the reason.


        Stefan