RE: [XEN RFC for-4.14] Re: use of "stat -"

[Paul Durrant <xadimgnik@xxxxxxxxx>]

> -----Original Message-----
> From: Ian Jackson <ian.jackson@xxxxxxxxxx>
> Sent: 24 June 2020 17:20
> To: Jan Beulich <jbeulich@xxxxxxxx>
> Cc: Elliott Mitchell <ehem+xen@xxxxxxx>; Andrew Cooper 
> <Andrew.Cooper3@xxxxxxxxxx>; Jason Andryuk
> <jandryuk@xxxxxxxxx>; Paul Durrant <paul@xxxxxxx>; Wei Liu <wl@xxxxxxx>; xen-
> devel@xxxxxxxxxxxxxxxxxxxx
> Subject: [XEN RFC for-4.14] Re: use of "stat -"
> 
> Jan Beulich writes ("Re: use of "stat -""):
> > [CAUTION - EXTERNAL EMAIL] DO NOT reply, click links, or open attachments 
> > unless you have verified
> the sender and know the content is safe.
> > On 14.05.2020 13:02, Ian Jackson wrote:
> > > I've read this thread.  Jan, I'm sorry that this causes you
> > > inconvenience.  I'm hoping it won't come down to a choice between
> > > supporting people who want to ship a dom0 without perl, and people who
> > > want a dom0 using more-than-a-decade-old coreutils.
> > >
> > > Jan, can you tell me what the output is of this on your ancient
> > > system:
> > >
> > >   $ rm -f t
> > >   $ >t
> > >   $ exec 3<t
> > >   $ stat -L -c '%F %i' /dev/stdin <&3
> > >   regular empty file 393549
> > >   $ rm t
> > >   $ stat -L -c '%F %i' /dev/stdin <&3
> > >   regular empty file 393549
> > >   $ strace -ou stat -L -c '%F %i' /dev/stdin <&3
> > >   $
> >
> > $ rm -f t
> > $ >t
> > $ exec 3<t
> > $ stat -L -c '%F %i' /dev/stdin <&3
> > regular empty file 3380369
> > $ rm t
> > $ stat -L -c '%F %i' /dev/stdin <&3
> > regular empty file 3380369
> > $ strace -ou stat -L -c '%F %i' /dev/stdin <&3
> > regular empty file 3380369
> >
> > > Also, the contents of the file "u" afterwards, please.
> >
> > Attached.
> 
> Thanks.
> 
> I think this means that `stat -' can be replaced by `stat /dev/stdin'.
> 
> This script is only run on Linux where /dev/stdin has existed
> basically forever.  The strace output shows
>   stat("/dev/stdin", {st_mode=S_IFREG|0644, st_size=0, ...}) = 0
> and the transcript shows that your stat(1) behaves as I hope.
> 
> Jan, will you send a patch ?  It is best if someone else but me writes
> it and tests it because then I am a "clean" reviewer.
> 
> Paul, supposing I review such a patch and say it is low risk, and we
> commit it by Friday, can it have a release-ack ?
> 

In principle yes but, given Jason's response, it doesn't sound like we have 
agreement on what the final patch will look like yet.

  Paul

> Thanks,
> Ian.