You are right in general that if one of the tests fail then the test script
should indicate this by returning some non-zero exit value.
By the way can you point to me a perl module which implements or
a piece of documentation which describes this policy? I known that it may be
just common sense and good practice but it's always easier when you have some
clear guidlines.
In this particular case, I don't think that failure of the authentication
test (number 10) means defective software. So, I will change the script to
report the number of failed test except number 10. Is that ok?
Regards,
Nikolay
--
Original message:
This report is not because the one test failed on Solaris,
but because a test failure does not lead to failing of
the test phase (when using CPAN.pm). This could lead to
installation of defective software.
Running make test
PERL_DL_NONLAZY=1 /usr/local/bin/perl -Iblib/arch -Iblib/lib
-I/usr/local/perl/lib/5.6.1/sun4-solaris
-I/usr/local/perl/lib/5.6.1 test.pl
1..12
ok 1
---- The remaining tests will be run for service 'login', user 'kriegjcb'
and
---- device '/dev/pts/5'.
ok 2
ok 3
ok 4
ok 5
ok 6
ok 7
ok 8
ok 9
---- Now you will be prompted to enter your unix password. On some systems
---- this test will succeed only if you are running as 'root'.
Password:
not ok 10 (9 - Authentication failed)
ok 11
ok 12
/usr/local/gnu/bin/make test -- OK
Running make install
