[perl #114008] Bleadperl around v5.16.0-150-g5f9f83b breaks LESPEA/Project-Euler-0.20.tar.gz

On Mon Feb 18 11:20:16 2013, jkeenan wrote:
> On Tue Aug 28 05:34:28 2012, rgs@consttype.org wrote:
> > On 28 August 2012 10:10, Jesse Luehrs <doy@tozt.net> wrote:
> > > On Tue, Aug 28, 2012 at 08:49:15AM +0200, Rafael Garcia-Suarez
> > wrote:
> > >> On 17 August 2012 23:27, Father Chrysostomos via RT
> > >> <perlbug-followup@perl.org> wrote:
> > >> > On Sat Jul 14 06:06:14 2012, rgs@consttype.org wrote:
> > >> >> Do you think the fix is incorrect ?
> > >> >
> > >> > Yes.
> > >> >
> > >> >> It actually cleans up a part of
> > >> >> the overload code that was not symmetrical and sanitizes the
> > semantics
> > >> >> of overload generation.
> > >> >
> > >> > But up until now, overloaded assignment operators have not been
> > >> > symmetrical.  Making them so will break any code that implements
> > them.
> > >> >
> > >> > From a user’s point of view, I wouldn’t expect to have to deal
> > with $x
> > >> > += $my_overloaded_object inside a += override.  Inside a +
> > override
> > >> > makes sense.  If fallback is 0, then $x += $pluseq_overloaded
> > should die.
> > >>
> > >> That's what my patch had fixed, IIRC. Previously, even with
> > >> fallback=>0 the overloaded operator += could have been generated
> > from
> > >> +.
> > >
> > > Right, but in fixing that, you broke something else (+= should not
> > be
> > > checking overloadedness of the right hand side).
> > 
> > This is right (while + needs to check overloadedness of both sides)
> > 
> > I'm a lot behind in my mail reading, do you know if we have already a
> > neat self-contained test case that I could use if I want to try to fix
> > my fix ?
> > 
> 
> This ticket is classified as a blocker to Perl 5.18.  I installed blead
> via perlbrew and ran 'cpan' with 'test Project::Euler' today.  I did not
> proceed when I noticed the long list of dependencies:
> 
> Creating new 'Build' script for 'Project-Euler' version '0.20'
> ---- Unsatisfied dependencies detected during ----
> ----     LESPEA/Project-Euler-0.20.tar.gz     ----
>     MooseX::Types [requires]
>     Test::Most [build_requires]
>     Modern::Perl [requires]
>     Exporter::Easy [requires]
>     Moose [requires]
>     Moose::Role [requires]
>     DateTime [requires]
>     String::Palindrome [requires]
>     Math::Big::Factors [requires]
>     Test::Exception [build_requires]
>     DateTime::Format::DateParse [requires]
>     namespace::autoclean [requires]
>     List::MoreUtils [requires]
>     MooseX::Types::Moose [requires]
>     Readonly [requires]
> 
> I then backscrolled in this RT and noticed a complaint about
> Math::Big::Factors.  I tested *that* module in the blead cpan shell and
> got this:
> 
> Running make test
> PERL_DL_NONLAZY=1
> /Users/jimk/perl5/perlbrew/perls/perl-blead/bin/perl5.17.9
> "-MExtUtils::Command::MM" "-e" "test_harness(0, 'inc', 'blib/lib',
> 'blib/arch')" t/*.t
> t/bigpm.t ...... overload arg '..' is invalid at
> /Users/jimk/.cpan/build/Math-BigInt-1.997-9JxLx4/blib/lib/Math/BigInt.pm
> line 153.
> t/bigpm.t ...... ok       
> t/factorspm.t .. overload arg '..' is invalid at
> /Users/jimk/.cpan/build/Math-BigInt-1.997-9JxLx4/blib/lib/Math/BigInt.pm
> line 153.
> t/factorspm.t .. ok     
> t/pod.t ........ ok   
> t/pod_cov.t .... ok   
> All tests successful.
> Files=4, Tests=182,  3 wallclock secs ( 0.26 usr  0.09 sys +  2.03 cusr
>  0.26 csys =  2.64 CPU)
> Result: PASS
>   TELS/math/Math-Big-1.12.tar.gz
>   /usr/bin/make test -- OK
> 
> The "overload arg '..' is invalid" warnings are the same as observed in
> https://rt.perl.org/rt3/Ticket/Display.html?id=113834.  My
> recommendation for this ticket is the same as for that ticket:  Keep the
> ticket open, but remove it from the 5.18-blockers list.
> 
> The ticket needs additional eyeballs.
> 
> Thank you very much.
> Jim Keenan


At http://www.cpantesters.org/distro/P/Project-Euler.html, Project-Euler v0.20 (a) is passing on all 5.19.* releases on Linux; and (b) is passing on 5.18.* on most OSes.

I'm taking this ticket for the purpose of closing it within 7 days unless someone objects.

Thank you very much.
Jim Keenan

---
via perlbug:  queue: perl5 status: open
https://rt.perl.org/Ticket/Display.html?id=114008

• [perl #114008] Bleadperl around v5.16.0-150-g5f9f83b breaks LESPEA/Project-Euler-0.20.tar.gz by James E Keenan via RT
• [perl #114008] Bleadperl around v5.16.0-150-g5f9f83b breaks LESPEA/Project-Euler-0.20.tar.gz by James E Keenan via RT