bug#28302: 26.0.50; [PATCH] Make ucs-names a hash table

[Thien-Thi Nguyen <ttn@gnu.org>]

[-- Attachment #1: Type: text/plain, Size: 1722 bytes --]


() Eli Zaretskii <eliz@gnu.org>
() Fri, 01 Sep 2017 09:23:05 +0300

   I suggest to update them to support a hash table as well as an
   alist.

Tangent: Maybe we could extend ‘ucs-names’ (the function) to
take args and DTRT accordingly.  Something like:

 ;; backward compatible
 (ucs-names) => ALIST

 ;; lookup
 (ucs-names 'forward-lookup CHAR-NAME) => CHAR-CODE
 (ucs-names 'reverse-lookup CHAR-CODE) => CHAR-NAME

 ;; bonus: reflection
 (ucs-names 'as-alist) => ALIST
 (ucs-names 'as-hash-table) => HASH-TABLE

Admittedly, this is not very idiomatic Emacs Lisp.  OTOH, i
would guess the vast majority of callers use it for lookup, so
centralizing that functionality would be a net win, despite the
style drift.

Really, given the rise of lexical binding and w/ niceties like
‘apply-partially’ already in the mix, i expect that sooner or
later, someone will put into place something like:

 (defun callable-hash-table (source)
   (let ((ht (ELABORATE source)))
     ;; rv
     (lambda (&optional cmd)
       (case cmd
         (as-hash-table ht)
         (as-alist ...)
         (forward-lookup ...)
         (reverse-lookup ...)
         ...))))

 (fset ucs-names (callable-hash-table ucs-names))

IOW:  Style drift be damned!  Up the idioms!  HOP things now!
(Insert more FP slogans here. :-D)

Or maybe this is already done?  What am i missing?  More coffee!

-- 
Thien-Thi Nguyen -----------------------------------------------
 (defun responsep (query)
   (pcase (context query)
     (`(technical ,ml) (correctp ml))
     ...))                              748E A0E8 1CB8 A748 9BFA
--------------------------------------- 6CE4 6703 2224 4C80 7502


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 197 bytes --]