jkeenan1 has asked for the wisdom of the Perl Monks concerning the following question:

I ask the assistance of the monks concerning error reports I am getting from some -- but not all -- automated testers at

I have been uploading new versions of ExtUtils-ModuleMaker over the past week to get test reports on additional functionality. After extensive tweaking, the module passes a quite rigorous test suite on the two OSes to which I have daily access: Darwin and Win32. It has also passed on other Perl hackers' FreeBSD, Linux and Solaris boxes. But two automated testing reports have left me baffled. Here is the first part of one such report

t/01_ini..........................ok t/02_bad_constructor..............IO object version 1.21 does not matc +h bootstrap parameter 1.22 at /usr/local/perl-5.8.5/lib/5.8.5/sun4-so +laris-thread-multi/ line 253. Compilation failed in require at /net/sunu991/disc1/home/.cpanplus/5.8 +.5/build/IO-1.22/blib/lib/IO/ line 262. BEGIN failed--compilation aborted at /net/sunu991/disc1/home/.cpanplus +/5.8.5/build/IO-1.22/blib/lib/IO/ line 262. Compilation failed in require at (eval 6) line 3. ...propagated at /usr/local/perl-5.8.5/lib/5.8.5/ line 85. BEGIN failed--compilation aborted at /usr/local/perl-5.8.5/lib/5.8.5/F +ile/ line 145. Compilation failed in require at t/testlib/ line 29. BEGIN failed--compilation aborted at t/testlib/ line 29. Compilation failed in require at t/02_bad_constructor.t line 10. BEGIN failed--compilation aborted at t/02_bad_constructor.t line 12. # Looks like your test died before it could output anything. dubious Test returned status 255 (wstat 65280, 0xff00) DIED. FAILED tests 1-112

I don't call IO::Handle anywhere in my code or test suite. I know that 'bootstrap' and 'use Dynaloader' are seen in modules that employ XS, but I don't use any XS; it's all pure Perl.

The two automated testers who reported this error tested an earlier version (0.38) which did have problems which I have since corrected. But the problems they're reporting with v0.39 bear no relationship to those they reported in 0.38.

I believe that these two testers run Perl with threads, which I do not. But the test suite has passed all tests on other automated sites which employ threads.

I've posted messages to the cpan-testers list, but haven't heard from the humans behind these test reports. (And suspect I won't, since it seems that list consists overwhelmingly of automated reports.) So can anyone suggest (a) what these errors mean and (b) if there's anything I can do to avoid them, given that I can't reproduce them on my OSes?

Thank you very much.

Jim Keenan