bug#29489: Obsolete gnus-remove-if and gnus-remove-if-not

[Noam Postavsky <npostavs@users.sourceforge.net>]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> Here's another one of these.
>
> I've replaced these with `seq-remove' and `seq-filter', and I suppose
> someone might object to that -- the other option would be `cl-remove-if'
> and `cl-remove-if-not'. I don't believe there's much practical difference.

Neither the seq nor cl functions handle hash tables right?

> -             (gnus-remove-if-not 'gnus-valid-move-group-p gnus-active-hashtb t)
> +             (seq-filter 'gnus-valid-move-group-p gnus-active-hashtb)
>               nil prefix nil default))
>             ((= 1 (length split-name))
>              (gnus-group-completing-read
>               prom
> -	     (gnus-remove-if-not 'gnus-valid-move-group-p gnus-active-hashtb t)
> +	     (seq-filter 'gnus-valid-move-group-p gnus-active-hashtb)
>               nil prefix 'gnus-group-history (car split-name)))

> -(defun gnus-remove-if (predicate sequence &optional hash-table-p)
> -  "Return a copy of SEQUENCE with all items satisfying PREDICATE removed.
> -SEQUENCE should be a list, a vector, or a string.  Returns always a list.
> -If HASH-TABLE-P is non-nil, regards SEQUENCE as a hash table."

Although the docstring claims "hash table", while the code says...

> -    (if hash-table-p
> -	(mapatoms (lambda (symbol)
> -		    (unless (funcall predicate symbol)
> -		      (push symbol out)))
> -		  sequence)

obarray?