[PATCH 0/5] Linux: lchmod and AT_SYMLINK_NOFOLLOW support for fchmodat

[Andreas Schwab]

On Aug 25 2020, Florian Weimer wrote:

> * Andreas Schwab:
>
>> On Aug 25 2020, Florian Weimer wrote:
>>
>>> * Andreas Schwab:
>>>
>>>> On Aug 25 2020, Florian Weimer wrote:
>>>>
>>>>> * Andreas Schwab:
>>>>>
>>>>>>> I have not been able to convince the kernel developers to add the
>>>>>>> required system call.  Maybe you will have more luck with that?
>>>>>>
>>>>>> What's the point of having lchmod if it always fails anyway?
>>>>>
>>>>> It used to be stub that always failed, but it does not anymore.
>>>>
>>>> Yes, it does.  The only point of lchmod is to call it on a symlink, and
>>>> that always fails.
>>>
>>> I'm not sure your assumption is correct.  lchown-type
>>
>> We are talking about lchmod, not lchown.
>
> The rationale for using lchmod in coreutils is the same, I think.

You cannot compare lchmod with lchown.  The latter has a useful, working
implementation, whereas the former *always* fails, so implementing it
doesn't make any sense.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."