Re: Emacs 28 on OSX: emacsclient.c:1415: warning: implicit declaration of function 'openat'

[Eli Zaretskii <eliz@gnu.org>]

> Date: Sun, 17 Apr 2022 13:36:58 -0700
> Cc: mattiase@acm.org, esq@lawlist.com, emacs-devel@gnu.org
> From: Paul Eggert <eggert@cs.ucla.edu>
> 
> On 4/17/22 12:12, Eli Zaretskii wrote:
> 
> > We
> > use 'open' and 'openat' in many places, and this change affects all of
> > them.
> 
> I don't see why. The patch affects none of the existing 'open' calls, 
> and just three of the 'openat' calls.

'openat' is implemented in src/w32.c for w32, and it calls 'sys_open',
which supports UTF-8 encoded file names on MS-Windows.  Your patch
replaced that with a call to 'open', which was basically not called
anywhere in Emacs before that, only internally in w32.c, via
'sys_open'.  And 'sys_open' is an important part of the Emacs ability
to support UTF-8 encoded file names on MS-Windows.  I hope it is now
more clear why such a change on a very stable branch cannot be
tolerated, unless the problem it solves is a critical one (which it
isn't).

I'm sorry, but users of old macOS versions will have to be more active
during the pretest to have their systems supported well by released
Emacs versions.  They can also install the patch locally if it solves
their problems.