On Win32, there's a couple of minor things with the
ModPerl-Registry tests.
- in ModPerl-Registry/t/TEST.PL, Win32 likes to have
the volume letter in a different place, as in the
following diff:
Index: ModPerl-Registry/t/TEST.PL
===================================================================
RCS file: /home/cvs/modperl-2.0/ModPerl-Registry/t/TEST.PL,v
retrieving revision 1.5
diff -u -r1.5 TEST.PL
--- ModPerl-Registry/t/TEST.PL 23 Jan 2003 04:29:59 -0000 1.5
+++ ModPerl-Registry/t/TEST.PL 8 Mar 2003 07:41:35 -0000
@@ -8,6 +8,7 @@
use lib map {("../$_", "../../$_")} qw(lib Apache-Test/lib);
use Apache::TestRunPerl ();
+use Apache::TestConfig ();
# sub-class Apache::TestRunPerl
package MyTest;
@@ -25,7 +26,9 @@
my ($volume, $dir, $file) = splitpath Cwd::cwd(), 1;
my @dir = grep {length} splitdir $dir;
$dir = catdir @dir[0..($#dir-2)]; # remove two last dir segments
- my $base = rootdir() . catpath $volume, $dir, $file;
+ my $base = Apache::TestConfig::WIN32 ?
+ catpath $volume, rootdir(), $dir, $file :
+ rootdir() . catpath $volume, $dir, $file;
$self->{conf_opts}->{maxclients} = 2;
===================================================================
- in the ModPerl-Registry/t/basic.t test, some tests are
made which rely on the setting of the executable bit, which
isn't available on Win32. The following diff:
Index: ModPerl-Registry/t/basic.t
===================================================================
RCS file: /home/cvs/modperl-2.0/ModPerl-Registry/t/basic.t,v
retrieving revision 1.10
diff -u -r1.10 basic.t
--- ModPerl-Registry/t/basic.t 8 Mar 2003 03:20:36 -0000 1.10
+++ ModPerl-Registry/t/basic.t 8 Mar 2003 07:42:08 -0000
@@ -28,6 +28,10 @@
# test non-executable bit
for my $alias (@aliases) {
+ if (Apache::TestConfig::WIN32) {
+ skip "non-executable bit test for Win32", 0;
+ next;
+ }
my $url = "/$alias/not_executable.pl";
ok t_cmp(
============================================================
skips these tests for Win32.
There's also something with the closure.t test - in running
'nmake test', test 4 of closure.t fails, but running closure.t
individually allows all tests to pass. A somewhat similar issue
arose with the same interpreter mod_perl test (which now
passes consistently).
--
best regards,
randy
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
