Obsoleting hash-table-* functions in subr-x in favor of map-* functions

Obsoleting hash-table-* functions in subr-x in favor of map-* functions

[Nicolas Petton]

[-- Attachment #1: Type: text/plain, Size: 231 bytes --]

Hi,

I'm thinking about obsoleting `hash-table-empty-p', `hash-table-keys'
and `hash-table-values' defined in subr-x in favor of their map-*
equivalent.

WDYT?

Cheers,
Nico
-- 
Nicolas Petton
http://nicolas-petton.fr

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 512 bytes --]

Re: Obsoleting hash-table-* functions in subr-x in favor of map-* functions

[Bozhidar Batsov]

[-- Attachment #1: Type: text/plain, Size: 690 bytes --]

Sounds a bit strange to me. Those are somewhat basic hash-table-specific
stuff. If you're going to obsolete them, why not obsolete everything except
hash-table constructors? Somewhere a line has to be drawn.

Btw, I'm still not sure about the "map" naming, as this is a term that's
not used at all in Emacs Lisp terminology and will be the source of a lot
of confusing I guess.

On 21 June 2015 at 21:58, Nicolas Petton <nicolas@petton.fr> wrote:

> Hi,
>
> I'm thinking about obsoleting `hash-table-empty-p', `hash-table-keys'
> and `hash-table-values' defined in subr-x in favor of their map-*
> equivalent.
>
> WDYT?
>
> Cheers,
> Nico
> --
> Nicolas Petton
> http://nicolas-petton.fr
>

[-- Attachment #2: Type: text/html, Size: 1215 bytes --]

Re: Obsoleting hash-table-* functions in subr-x in favor of map-* functions

[Bozhidar Batsov]

[-- Attachment #1: Type: text/plain, Size: 1099 bytes --]

P.S. I'm aware this was borrowed from the likes of Clojure (as was seq),
but sometimes it's not a great idea to bring in external terminology 1:1.
Anyways, that's not major. When I added those hash-table specific function
the point was to extend a specific API, not to look for some generic
solution.

On 22 June 2015 at 08:28, Bozhidar Batsov <bozhidar@batsov.com> wrote:

> Sounds a bit strange to me. Those are somewhat basic hash-table-specific
> stuff. If you're going to obsolete them, why not obsolete everything except
> hash-table constructors? Somewhere a line has to be drawn.
>
> Btw, I'm still not sure about the "map" naming, as this is a term that's
> not used at all in Emacs Lisp terminology and will be the source of a lot
> of confusing I guess.
>
> On 21 June 2015 at 21:58, Nicolas Petton <nicolas@petton.fr> wrote:
>
>> Hi,
>>
>> I'm thinking about obsoleting `hash-table-empty-p', `hash-table-keys'
>> and `hash-table-values' defined in subr-x in favor of their map-*
>> equivalent.
>>
>> WDYT?
>>
>> Cheers,
>> Nico
>> --
>> Nicolas Petton
>> http://nicolas-petton.fr
>>
>
>

[-- Attachment #2: Type: text/html, Size: 1930 bytes --]

Re: Obsoleting hash-table-* functions in subr-x in favor of map-* functions

[Nicolas Petton]

[-- Attachment #1: Type: text/plain, Size: 288 bytes --]

Bozhidar Batsov <bozhidar@batsov.com> writes:

> Sounds a bit strange to me.

Might be. I was merely asking.  But now I think it should be the
opposite: map should use hash-table-* functions for hash-tables whenever
possible.

Nico
-- 
Nicolas Petton
http://nicolas-petton.fr

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 512 bytes --]

Re: Obsoleting hash-table-* functions in subr-x in favor of map-* functions

[Nicolas Petton]

[-- Attachment #1: Type: text/plain, Size: 465 bytes --]

Bozhidar Batsov <bozhidar@batsov.com> writes:

> P.S. I'm aware this was borrowed from the likes of Clojure (as was seq),
> but sometimes it's not a great idea to bring in external terminology 1:1.
> Anyways, that's not major.

We spent some time thinking about the name, and IIRC Stephan proposed
map, which I kind of like.

If you have a better name though, It's not too late to change :)

Cheers,
Nico
-- 
Nicolas Petton
http://nicolas-petton.fr

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 512 bytes --]

Re: Obsoleting hash-table-* functions in subr-x in favor of map-* functions

[Bozhidar Batsov]

[-- Attachment #1: Type: text/plain, Size: 399 bytes --]

Yep, this makes perfect sense.

On 22 June 2015 at 10:27, Nicolas Petton <nicolas@petton.fr> wrote:

> Bozhidar Batsov <bozhidar@batsov.com> writes:
>
> > Sounds a bit strange to me.
>
> Might be. I was merely asking.  But now I think it should be the
> opposite: map should use hash-table-* functions for hash-tables whenever
> possible.
>
> Nico
> --
> Nicolas Petton
> http://nicolas-petton.fr
>

[-- Attachment #2: Type: text/html, Size: 921 bytes --]