Hi Tim,
On 10 Sep 2007, at 10:15, Tim Bunce wrote:
Hello Andy.
On Fri, Sep 07, 2007 at 05:38:28PM +0100, [EMAIL PROTECTED] wrote:
This is a computer-generated test report for DBI-1.59, created
automatically by CPAN::Reporter, version 0.99_08,
Test::Harness 2.99_03
Test::More 0.70
t/zvp_42prof_data...........ok
t/zvp_43prof_env............ No subtests run
t/zvp_80proxy............... No subtests run
t/zvxgp_43prof_env.......... No subtests run
t/zvxgp_80proxy............. No subtests run
Test Summary Report
-------------------
t/zvp_43prof_env.t (Wstat: 0 Tests: 0 Failed: 0)
Parse errors: No plan found in TAP output
t/zvp_80proxy.t (Wstat: 0 Tests: 0 Failed: 0)
Parse errors: No plan found in TAP output
t/zvxgp_43prof_env.t (Wstat: 0 Tests: 0 Failed: 0)
Parse errors: No plan found in TAP output
t/zvxgp_80proxy.t (Wstat: 0 Tests: 0 Failed: 0)
Parse errors: No plan found in TAP output
Files=121, Tests=6002, 76 wallclock secs (57.88 cusr + 4.86 csys
= 62.74 CPU)
Result: FAIL
Those tests are being falsely labeled as failures by Test::Harness
2.99_03
where previous versions have accepted the output from:
print "1..0 # Skipped: profiling not supported for
DBI::PurePerl\n";
exit 0;
as meaning to skip all tests. It may not be perfect modern TAP, but
it has
worked for many years.
Test::Harness needs fixing or you'll have a lot of upset authors and
confused users on your hands.
Yes, I'm really sorry about that. We've just discussed this on tapx-
dev and the consensus is that we need to get over ourselves about
these minor syntactical infractions and stop bothering people with
spurious failure reports.
We'll currently enforcing #\s+skip\b but we'll remove the \b for any
TAP that doesn't have an explicit version of 13 or higher.
Sorry to have bothered you.
--
Andy Armstrong, hexten.net
