From 2d8ca5da0888ace50c79900af6ef82761bea90b2 Mon Sep 17 00:00:00 2001 From: Dan Sugalski Date: Wed, 11 Jun 1997 12:00:00 +1200 Subject: [PATCH] [PATCH] Harness.pm bug w/perl5.004 & VMS Harness.pm's got a bug under VMS. $estatus is my'd and assigned to at the same time--unfortunately the assignment's a trinary conditional that uses $estatus, presumably before the my context is fully established. It generates an estatus doesn't exist error message for the first test only, and, after all the t/*.t tests are through, kills perl with an ACCVIO. The following patch just splits the my and assignment into two separate statements, which makes perl much happier. p5p-msgid: 3.0.1.32.19970530102300.008a2730@stargate.lbcc.cc.or.us --- lib/Test/Harness.pm | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/lib/Test/Harness.pm b/lib/Test/Harness.pm index 6979a11..f863717 100644 --- a/lib/Test/Harness.pm +++ b/lib/Test/Harness.pm @@ -104,7 +104,8 @@ sub runtests { } $fh->close; # must close to reap child resource values my $wstatus = $?; - my $estatus = ($^O eq 'VMS' + my $estatus; + $estatus = ($^O eq 'VMS' ? eval 'use vmsish "status"; $estatus = $?' : $wstatus >> 8); if ($wstatus) { -- 2.7.4