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

[perl #119413] Reconsider e6c4c33 to ext/B/ in light of test failures

Thread Previous
From:
James E Keenan
Date:
August 21, 2013 23:29
Subject:
[perl #119413] Reconsider e6c4c33 to ext/B/ in light of test failures
Message ID:
rt-3.6.HEAD-1873-1377127732-143.119413-75-0@perl.org
# New Ticket Created by  James E Keenan 
# Please include the string:  [perl #119413]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=119413 >


In https://rt.perl.org/rt3/Ticket/Display.html?id=119351, I reported 
that Devel::Cover, then at version 1.06, was failing its tests when I 
and others attempted to install it on Perl 5.18.1.  Recent versions of 
Devel::Cover had successfully installed on 5.18.0 in May of this year, 
so these test failures were surprising.  In fact, those other recent 
versions of Devel::Cover *also* had test failures when tried on 5.18.1.

We identified Perl 5 commit e6c4c33 as the likely cause of the problem 
from the Perl 5 end.  That diagnosis can be found in this post:
https://rt.perl.org/rt3/Ticket/Display.html?id=119351#txn-1245355

So we were at the point of having to figure out:

* Do we have to change something in Perl 5 -- and does that warrant 
issuance of a Perl 5.18.3?

* Do we have to change something in Devel::Cover?

* Do we have to change something in both?

Tonight, Paul Johnson++ posted v1.07 of Devel::Cover on CPAN.  It works 
on 5.18.1, so I have closed RT #119351.  But we still have to address 
the first of the 3 bullet points listed above.  I am creating this 
ticket to track discussion of the following questions:

* Was e6c4c33 a bug fix that went awry?  Or was it an attempted 
refactoring for improvement's sake?  If the latter, why was it included 
in a maintenance release?

* Does e6c4c33 pose other problems that warrant issuance of a Perl 
5.18.3?  Would rolling back e6c4c33 be the fix, or is something else 
warranted?

Note:  I'm not trying to start a blame game here.  I assume there was a 
good rationale for e6c4c33 -- and I myself probably wouldn't have 
understood the rationale whether good or bad :-) .  But if this is 
something that should be fixed with a new maintenance release, we should 
start that discussion right away.

Thank you very much.
Jim Keenan



Thread Previous


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About