On Fri, Jan 20, 2023 at 10:01:33PM +0100, Christian Weisgerber wrote:
> Tracey Emery:
> 
> > After setting up regress, it was found that 32 was too low of a ulimit
> > for sparc64. The following is an idea to fix this for sparc64 and keep
> > other arches untouched.
> 
> Why would sparc64 be different in this respect?  That doesn't make any
> sense.
> 
> > --- regress/cmdline/checkout.sh
> > +++ regress/cmdline/checkout.sh
> > @@ -901,7 +901,17 @@ test_checkout_ulimit_n() {
> >  	# Checkout should still work; if it does not, then either there is
> >  	# a bug or the fixed limit used by this test case is no longer valid
> >  	# and must be raised. Use a subshell to avoid changing global ulimit.
> > -	(ulimit -n 32; got checkout -q $testroot/repo $testroot/wt \
> 
> There is absolutely no reason that limit should vary by architecture.
> 
> -- 
> Christian "naddy" Weisgerber                          naddy@mips.inka.de
> 

32 fails on sparc. That's the only reason.

-- 

Tracey Emery