Re: Bug in XSHd5 futimens

[Michael Kerrisk <mtk.lists@xxxxxxxxx>]

Geoff Clare wrote:
> Michael Kerrisk <mtk.lists@gmail.com> wrote, on 15 May 2008:
> > Geoff Clare wrote:
> > > Michael Kerrisk <michael.kerrisk@gmail.com> wrote, on 14 May 2008:
> > > > @ page 969 line 32433 section futimens comment {utimensat-errs}
> > > I agree that this change is needed. Since draft 5.1 has already been
> > > produced, I suggest that we should take this as an aardvark against 5.1
> > > (the page and line numbers are the same), to be addressed in TC1.
> > Do I need to do anything to get that Aardvark moved to the TC1 queue?
> You don't need to do anything.  As per the minutes of yesterday's
> teleconference, Andrew is going to create a new set of aardvarks for
> the revision and yours will be the first entry.
>
> > In passing, does anyone know what the rationale was for having two 
> > different permission errors (EACCES and EPERM) for utime() and utimes() 
> > (and thus now utimesat() and futimens())?  Was this just standardization 
> > of existing practice, or is there another reason?
> I'm fairly sure it was just existing practice.
I see that the POSIX.1-2004 utimes() spec carries a similar mistake -- I 
guess that's where the working for utimenssat() came from.  Should I 
raise a separate Aardvark to get that fixed?

Cheers,

Michael