Re: Regexp error scan (March 26)

[Andy Moreton <andrewjmoreton@gmail.com>]

On Wed 27 Mar 2019, Mattias Engdegård wrote:

> 27 mars 2019 kl. 04.43 skrev Basil L. Contovounesios <contovob@tcd.ie>:
>> 
>> Paul Eggert <eggert@cs.ucla.edu> writes:
>> 
>>>       (replace-regexp-in-string
>>> -       (format "[%s]"
>>> -               (mapconcat (lambda (char)
>>> -                            (regexp-quote (char-to-string char)))
>>> -                          erc-lurker-ignore-chars ""))
>>> +       (regexp-opt (cl-delete-duplicates
>>> +		    (mapcar #'char-to-string erc-lurker-ignore-chars)))
>>>        "" nick)
>>>     nick))
>> 
>> regexp-opt already deletes duplicates, so you can just:
>> 
>> (regexp-opt (mapcar #'char-to-string erc-lurker-ignore-chars))
>> 
>> Besides, cl-delete-duplicates defaults to eql for equality, so it's a
>> no-op here anyway.
>
> Or perhaps:
>
> (regexp-opt-charset (append erc-lurker-ignore-chars nil))

This *-charset naming is helpful, as it reminds users of the expected
argument type.

Given the previously noted problems with incorrect arguments for
`skip-chars-forward' and `skip-chars-backward', would it make sense to
rename these functions to `forward-charset' and `backward-charset' ?

    AndyM