develooper Front page | perl.perl5.porters | Postings from May 2013

[perl #116831] Benchmark tests produce a lot of warnings in 5.17.9

Thread Next
James E Keenan via RT
May 20, 2013 00:36
[perl #116831] Benchmark tests produce a lot of warnings in 5.17.9
Message ID:
On Fri Mar 01 08:39:28 2013, davem wrote:
> On Sun, Feb 24, 2013 at 09:53:03PM -0500, Ricardo Signes wrote:
> > * demerphq <> [2013-02-24T00:53:55]
> > > Good catch. I actually wonder at the wisdom of this at all. Having
> the
> > > warnings be on a different channel than the data means you dont
> know
> > > which item being benchmarked is responsible for the warning as the
> two
> > > handles can get out of sync due to buffering.
> >
> > I think if that's the case, we might want to improve the warning
> message to
> > make it clearer, rather than put the warnings on the standard output
> stream.
> >
> > I also wonder whether some of the tests that are emitting these
> warnings are
> > emitting them with good reason.  That is: are they actually testing
> timings
> > badly?  I have not yet investigated.
> I've reverted the change for now with commit
> a2656c4ab81b216a427d1e0db6a3aa25b4350ba8.
> Presumably we can re-address this issue post-5.18.

Now that we're post-5.18.0, can someone sum up where we stand with the
issues discussed in this ticket?

If the issues are unresolved, it would be good to know whether we should
continue discussing them in this ticket, versus opening up a new ticket
with a clean statement of the problems.

Thank you very much.
Jim Keenan

via perlbug:  queue: perl5 status: open

Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About