Re: [PATCH] torture: Correctly fetch CPUs for kvm-build.sh with all native language

[Frederic Weisbecker <frederic@kernel.org>]

On Thu, Apr 01, 2021 at 02:02:53PM -0700, Paul E. McKenney wrote:
> On Thu, Apr 01, 2021 at 10:41:13PM +0200, Frederic Weisbecker wrote:
> > On Thu, Apr 01, 2021 at 01:40:22PM -0700, Paul E. McKenney wrote:
> > > On Thu, Apr 01, 2021 at 10:31:12PM +0200, Frederic Weisbecker wrote:
> > > > On Thu, Apr 01, 2021 at 11:51:16AM -0700, Paul E. McKenney wrote:
> > > > > On Thu, Apr 01, 2021 at 03:26:02PM +0200, Frederic Weisbecker wrote:
> > > > > > Grepping for "CPU" on lscpu output isn't always successful, depending
> > > > > > on the local language setting. As a result, the build can be aborted
> > > > > > early with:
> > > > > > 
> > > > > > 	"make: the '-j' option requires a positive integer argument"
> > > > > > 
> > > > > > Prefer a more generic solution.
> > > > > > 
> > > > > > Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
> > > > > 
> > > > > Good catch, applied, thank you!
> > > > > 
> > > > > There is a similar construct in kvm-remote.sh, so I added a similar
> > > > > fix to your patch.
> > > > > 
> > > > > But what about this in functions.sh?
> > > > > 
> > > > > nt="`lscpu | grep '^NUMA node0' | sed -e 's/^[^,]*,\([0-9]*\),.*$/\1/'`"
> > > > > 
> > > > > I am guessing that "node0" is human-language-independent, but is "NUMA"?
> > > > 
> > > > I thought they wouldn't bother translating that, but they did...
> > > > 
> > > >     NUMA node0 CPU(s):               0-7
> > > > 
> > > > becomes:
> > > > 
> > > >     Nœud NUMA 0 de processeur(s) : 0-7
> > > > 
> > > > Not sure about the best way to fix it.
> > > 
> > > The rude and crude fix is for the scripts to force the local language
> > > to English.  ;-)
> > 
> > I don't have a better answer :o)
> 
> If you set the environment variable LANG to en_US.UTF-8, does that
> make things work for you?  Huh.  Setting it to fr_FR.UTF-8 does not
> shift lscpu out of English for me, so I am guessing "no".

Maybe that language isn't installed in your system. I would expect
en_US.UTF-8 to be supported pretty much everywhere though. At least it
works for me with: "LANG=en_US.UTF-8 lscpu".

Thanks.

> 
> Help?
> 
> 							Thanx, Paul