bug#60620: [PATCH] copy.c: replace set_acl() with chmod_or_fchmod()

[Ondrej Valousek]

Well, true, I do not know for sure what will happen on different platforms, yes 
- I can't test.
On Linux both work the same way.
Either case, the purpose of qset_acl() is to set ACLs the way so that it 
corresponds with the mode argument right?
That's what I would expect how it's supposed to work.
If yes, then isn't it better to handle this task to the filesystem driver in 
kernel and just call chmod?

Zasláno z Outlooku pro Android<https://aka.ms/AAb9ysg>
________________________________
From: Pádraig Brady <pixelbeat@gmail.com> on behalf of Pádraig Brady 
<P@draigBrady.com>
Sent: Sunday, January 8, 2023 10:04:19 PM
To: Ondrej Valousek <ondrej.valousek.xm@renesas.com>; 60620@debbugs.gnu.org 
<60620@debbugs.gnu.org>
Subject: Re: bug#60620: [PATCH] copy.c: replace set_acl() with chmod_or_fchmod()

On 06/01/2023 15:23, Ondrej Valousek wrote:
> This patch replaces set_acl() funclion call with chmod_or_fchmod()
> Both functions works (AFAIK) the same way (at least in Linux) so should be 
> possible.
> Using chmod_or_fchmod would also help us to reduce dependency on libacl
> (see the forthcoming patch to qcopy-acl.c from Gnulib).

The "at least in Linux" qualification worries me.
Having a very quick look at the qset_acl() code suggests
it clears ACLs on some platforms at least,
which chmod_or_fchmod() does not.
Am I reading that wrong?

thanks,
Pádraig