Here's a tiny little patch (to be applied on after the VMSPIPE patch)
for getting rid of the harmless %SYSTEM-F-ABORT messages in the
lib/vmsish.t tests.
I also tracked down the cause of the error messages in op/lex_assign.t
(the messages about ] not found before EOF) and they're a bit odder.
It seems that the test in question is doing a:
$a = `$^X -e"print qq[1\n]" `;
(i.e, running Perl in a subprocess with output collected in $a)
But for us that amounts to a command line something like:
$ perl -e"print qq[1
]"
Which obviously gives both Perl and DCL some problems.
One alternative is to simply skip the test on VMS.
Another possible solution is to double up the backslashes:
$a = `$^X -e"print qq[1\\n]" `;
^^
and this does give a "test success". But it's not entirely clear to
me what this particular test is trying to accomplish, and whether it
behaves differently on non-VMS systems.
So can anyone shed some light on this? Does op/lex_assign.t succeed
on unix systems, and do they return an error message or a "1\n" string
in $a?
Patch for the other test follows...
--- t/lib/vmsish.t-orig Wed Apr 19 10:39:55 2000
+++ t/lib/vmsish.t Wed Apr 19 10:39:37 2000
@@ -136,6 +136,7 @@
local *P;
open(P,'>vmsish_test.com') || die('not ok ?? : unable to open "vmsish_test.com"
for writing');
print P "\$ set message/facil/sever/ident/text\n";
+ print P "\$ define/nolog sys\$error nl:\n";
print P "\$ $Invoke_Perl @_\n";
close P;
my $x = `\@vmsish_test.com`;
--
Drexel University \V --Chuck Lane
----------------->--------*------------<[EMAIL PROTECTED]
(215) 895-1545 / \ Particle Physics [EMAIL PROTECTED]
FAX: (215) 895-5934 /~~~~~~~~~~~ [EMAIL PROTECTED]
