bug#51320: 26.1; kill-region does not ignore its first two args if region is active

[Richard Lawrence <rwl@recursewithless.net>]

>> From: Richard Lawrence <rwl@recursewithless.net>
>> Date: Thu, 21 Oct 2021 14:29:11 +0200
>> 
>> In a function I'm writing, I tried to call kill-region like this:
>> 
>> (and (use-region-p) (kill-region nil nil t))
>> 
>> with its first two arguments being nil, because the docstring says:
>> 
>> "Supply two arguments, character positions BEG and END indicating the
>>  stretch of text to be killed.  If the optional argument REGION is
>>  non-nil, the function ignores BEG and END, and kills the current
>>  region instead."
>> 
>> But in fact, if both BEG and END are nil as in my function call,
>> kill-region doesn't ignore them but throws an error, since its first two
>> lines are:
>> 
>>   (unless (and beg end)
>>     (user-error "The mark is not set now, so there is no region"))
>> 
>> This was surprising to me, since I expected these values to be ignored.

Eli Zaretskii <eliz@gnu.org> writes:

> The error is thrown because there's no region in the buffer where you
> invoke this.  The fact that you pass BEG and END nil is not the reason
> for the error.

Forgive me, I'm no expert, but that doesn't sound right to me, for two
reasons:

1) In my example, kill-region is only called if use-region-p returns t.
(I also checked that (point) and (mark) return non-nil in this
context with the debugger.) 

2) Literally the first and only thing that kill-region does before
throwing the error is check whether BEG and END are non-nil. The error I
see is the one in the unless clause quoted above.

Am I missing something?    

-- 
Best,
Richard