Front page | perl.perl5.porters |
Postings from January 2013
On Wed, Jan 30, 2013 at 05:04:37AM -0800, James E Keenan via RT wrote: > On Tue Jan 29 22:35:55 2013, smueller@cpan.org wrote: > > On 01/30/2013 05:47 AM, Peter Rabbitson (via RT) wrote: > > > The test block > https://github.com/tsee/Data-Dumper/commit/db957a7e#L4R1476 > > > fails randomly (both pp and xs part) on any 5.8 perl (regardelss of what > > > version.pm is installed if any). While fixing this would be nice, it is > > > likely a bug in perl itself (otherwise the pp version would not be > failing > > > randomly). If this is the case I propose skipping the tests on $] < > 5.009, > > > as v-string support was not available on D::Ds that shipped with > 5.8, and > > > hence things like CPAN are unlikely to rely on it. > > > > I'm on board with this. Any objections? > > > > Jim, what's the status of your patches? I would hate to do this change > > just to cause a conflict for you! > > > > --Steffen > > > > > My first reaction would be to request output from the (admittedly > random) test failures so that we can directly evaluate them. Absolutely. A failure looks something like this (note - expected/got are reversed in the test itself): ... ok 391 - vstrings ok 392 ok 393 not ok 394 - xs vstrings --Expected-- #$a = \v65.66.67; #$b = \v65.66.067; #$c = \v65.66.6_7; #$d = \'ABC'; --Got-- #$a = \v65.66.67; #$b = \'ABC'; #$c = \v65.66.6_7; #$d = \'ABC'; ok 395 not ok 396 --Expected-- #$a = \v65.66.67; #$b = \v65.66.067; #$c = \v65.66.6_7; #$d = \'ABC'; --Got-- #$a = \v65.66.67; #$b = \'ABC'; #$c = \v65.66.6_7; #$d = \'ABC'; ok 397 - blessed overloaded globs ok 398 ... I am also attaching an xzip of this run: for i in $(seq 100); do prove -bv t/dumper.t >> dumper.t.log ; done Perl is 5.8.9 with -Dusethreads Cheers!Thread Previous | Thread Next