On Thu, 06 Mar 2008 23:07:28 -0500, Jarkko Hietaniemi wrote:
Is it just a coincidence and just my problem that within the last
24 hours many of my CPAN modules have got failure reports from CPAN
testers who are running* 64-bit* Perls?
Time::HiRes, Graph, Statistics-Shannon: pretty independent modules,
only first of them XS, the middle one complex pure Perl, the last one
pure Perl and pretty trivial, and which one I haven't touched for ages?
The failures look nonsensical, but I'll try compiling a 64-bitint Perl
somewhere and see. The perls of the testers are 5.8.8 and 5.10.0.
I use Time::HiRes a lot and Graph occasionally both also on 64bitall
Just checked with 5.8.8-64all and 5.11.0-64all and no errors whatsoever
pc09 Linux 2.6.22.17-0.1-default/#1 x86_64 Xeon(R) CPU E5320 @ 1.86GHz/1596(4) x86_64 3956 Mb
Normally all my perls are non-threaded, but I have blead installed as
threaded for module tests
This is perl, v5.8.8 built for x86_64-linux
Locally applied patches:
defined-or
MAINT30961
This is perl, v5.11.0 DEVEL33450 built for x86_64-linux-thread-multi
DEVEL (33450)
Compile-time options: DEBUGGING MULTIPLICITY PERL_DONT_CREATE_GVSV
PERL_IMPLICIT_CONTEXT PERL_MALLOC_WRAP
PERL_TRACK_MEMPOOL USE_64_BIT_ALL USE_64_BIT_INT
USE_ITHREADS USE_LARGE_FILES USE_PERLIO
USE_REENTRANT_API