[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#1092497: mercurial: Please raise timeout for slower architectures and use --allow-slow-tests



Hi Julien,

On Wed, 2025-01-22 at 13:09 +0100, Julien Cristau wrote:
> On Wed, Jan 22, 2025 at 10:34:25 +0100, John Paul Adrian Glaubitz wrote:
> 
> > Hello Julien,
> > 
> > On Wed, 2025-01-22 at 10:15 +0100, Julien Cristau wrote:
> > > Hmm, so I don't think I want to do that:
> > > - the default timeout is 360, so 1800 is already way above that, an hour
> > >   per test is unreasonable
> > > - I don't want to be running the even slower tests on those already too
> > >   slow archs, which --allow-slow-tests would do
> > 
> > I understand your concern. But the problem with SPARC is that the single-core
> > performance is rather low and the architectures have a lot of cores to compensate
> > for that. Thus, if a test requires higher single-core performance, it will fail
> > here.
> 
> What --allow-slow-tests does is make some tests run that would otherwise
> be skipped, so if anything it's going to make things worse.
> 
> And I'd rather skip some tests that are problematic than let them run
> longer than the current timeout.

OK, then I misunderstood that option. I will retest with the option removed
and the timeout reduced to 1800 on sparc64 and report back.

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913


Reply to: