At 8:50 AM +0100 2/20/03, Vorl�nder, Martin wrote: >when rebuilding Perl 5.8.0 to include the tainting bug patch, I got >several test failures.
Do you have a subsystem identifier granted to the account where you are building and/or testing Perl? I found it impossible to get very far in the build with tainting enabled for everything. I suspect the same would be true of the test suite. However, the symptom should not be accvio. Incidentally, OVMS Engineering has apparently long since found and fixed the tainting bug in their supported versions of 5.5.3 and 5.6.1. See <http://www.openvms.compaq.com/openvms/products/ips/apache/csws_perl_relnotes.html#notes> It would be nice to find the ear of the right person and convince them that open source is a two-way street. They've got at least a couple of other bug fixes that are unlikely to make it into 5.8.1 unless they show up with their changes real soon now. Sorry, back to the problem at hand. >Upon looking closer into these, I saw that >most of them were caused by an ACCVIO in the backquote command called >by the runperl() function from t/test.pl. I'm running OpenVMS Alpha >V7.2-2. > >$^X is resolved to something like disk:[path.][000000]PERL.EXE >(I have installed it into the DCLTABLES) PERL_ROOT (and thus the above >disk:[path.]) and PERLSHR point to the correct images. Could the >unresolved ".][" have something to do with it? $^X should point to the Perl you are actually running, which in the test suite would ordinarily be the copy in the [.t] directory under the main source directory. To simulate what the test suite sees, you should be able to set default to the [.t] directory and do: $ define perlshr sys$disk:[-]perlshr.exe %DCL-I-SUPERSEDE, previous value of PERLSHR has been superseded $ mcr []perl. -e "print $^X;" disk$user:[homedir.perl.t]perl.;1 If instead this shows you what's in DCLTABLES, we may need to muck about with how $^X gets its info. The definition of PERL_ROOT should not be relevant for running the test suite. -- ________________________________________ Craig A. Berry mailto:[EMAIL PROTECTED] "... getting out of a sonnet is much more difficult than getting in." Brad Leithauser
