[feature] anaphoric cl-callf(*)

[yyoncho]

here it is what you can do now:

(let ((ht (make-hash-table :test 'equal)))
  (cl-callf (lambda (s) (concat "FOO" s)) (gethash "BAR" ht))
  (cl-callf (lambda (s) (concat "-FOO2" s)) (gethash "BAR" ht))
  ht)

here it is what it can be with dash:

(let ((ht (make-hash-table :test 'equal)))
  (--callf (concat "FOO" it) (gethash "BAR" ht))
  (--callf (concat "-FOO2" it) (gethash "BAR" ht))
  ht)

[yyoncho]

FTR the lambda has to be more complex otherwise

(let ((ht (make-hash-table :test 'equal)))
  (cl-callf concat (gethash "BAR" ht) "FOO")
  (cl-callf concat (gethash "BAR" ht) "FOO")
  ht)

would work too.

[alphapapa]

I guess you mean "anaphoric." ;)

Yes, this would be a useful macro.

[yyoncho]

I guess you mean "anaphoric." ;)

Yes, fixed :)

[basil-conto]

FTR the lambda has to be more complex otherwise [...] would work too.

Yes, or you could use cl-callf2.

here it is what it can be with dash:

Sorry, I fail to see the relation between CL macros and a list processing library whose whole motto is "no CL needed".

The anaphoric macros that Dash provides are indeed handy and efficient shorthands, but that doesn't make Dash a library that provides anaphoras for everything under the sun.

[yyoncho]

Sorry, I fail to see the relation between CL

There is no relation, I used CL as an example. AFAICS macro can be implemented without CL dep.

list processing library

There are ->, ->>, -doto, etc which are not related to list processing but they are pretty much clojure ports. In this case, --callf is the equivalent to the swap!(disregarding the fact that elisp's places are not exactly equivalent to clojure's atoms). Having them will help with "no CL needed".

The anaphoric macros that Dash provides are indeed handy and efficient shorthands, but that doesn't make Dash a library that provides anaphoras for everything under the sun

I guess it is fine to reject the feature request.

[basil-conto]

AFAICS macro can be implemented without CL dep.

You mean by using gv.el? That would require Emacs >= 24.3.

There are ->, ->>, -doto, etc which are not related to list processing but they are pretty much clojure ports.

I never said I was a fan of those. ;)

I guess it is fine to reject the feature request.

I'm not going to because it's up to the maintainers.

It's just that to me, wrapping generalised variables in anaphora sounds both out of scope of efficient and pure list processing, and like it could potentially be a can of worms or slippery slope.