At 6:34 AM -0500 1/19/06, John E. Malmberg wrote:
>lib/ExtUtils/t/Manifest........................FAILED at test 26
I'm not quite sure what to do about this one. It's testing
manicopy() and then checking to see that the permissions on the
copied files are the same as the permissions on the original files.
The comment in the test says "executable if original was", but it's
also testing for "non-executable if original was non-executable" and
that's the test we are failing. On VMS, I don't think the
permissions on the source of a copy have anything to do with the
resulting permissions on the target. If they do, they're way down
the list after process defaults, the permissions on the directory
into which we're copying, permissions on lower versions of the same
file, access control lists, etc.
If I run the test with elevated privileges, it passes because
everything is considered readable, writable, and executable. If I
change my default permissions to exclude owner execute permission
from the default protection mask, then the test will pass because
both source and target are non-executable. But the default behavior
is that the source file will not be created with execute permission,
but the result of the copy will be. Something like the following may
be in order, but I'm going to cogitate on this a bit more.
--- Manifest.t;-0 Thu Nov 10 11:29:20 2005
+++ Manifest.t Thu Jan 19 15:08:20 2006
@@ -150,8 +150,10 @@
foreach my $orig (@copies) {
my $copy = "copy/$orig";
ok( -r $copy, "$copy: must be readable" );
- is( -w $copy, -w $orig, " writable if original was" );
- is( -x $copy, -x $orig, " executable if original was" );
+ # on VMS, the permission of the source file has little or nothing to do
+ # with the resulting permission on the target of a copy operation
+ is( -w $copy, -w $orig or $Is_VMS, " writable if original was" );
+ is( -x $copy, -x $orig or $Is_VMS, " executable if original was"
);
}
rmtree('copy');
[end of diff]
--
________________________________________
Craig A. Berry
mailto:[EMAIL PROTECTED]
"... getting out of a sonnet is much more
difficult than getting in."
Brad Leithauser
