From @nwc10

On Tue, Dec 25, 2012 at 06​:22​:00PM -0500, George Greer wrote​:

Smoke logs available at http​://m-l.org/~perl/smoke/perl/linux/blead_clang_quick/log8d40577bdbdfa85ed3293f84bf26a313b1b92f55.log.gz

Automated smoke report for 5.17.8 patch 8d40577 v5.17.7.0-107-g8d40577
zwei​: Intel(R) Core(TM) i7 CPU 920 @​ 2.67GHz (GenuineIntel 2668MHz) (x86_64/8 cpu)

Testsuite was run only with 'harness'

Failures​: (common-args) -Accflags=-DPERL_POISON -Dcc=clang
[default]
../dist/Tie-File/t/29_downcopy.t............................FAILED
443-444
Bad plan. You planned 718 tests but ran 716.

I infer that the "bad plan" is because the skip count is wrong if the
timeout hits. The main test code looks like this​:

sub try {
  my ($pos, $len, $newlen) = @​_;

...

  if ($err) {
  if ($err =~ /^Alarm clock/) {
  print "# Timeout\n";
  print "not ok $N\n"; $N++;
  print "not ok $N\n"; $N++;
  return;
  } else {
  $@​ = $err;
  die;
  }
  }

...

  if (defined $len) {
  try($pos, undef, $newlen);
  }
}

Hence if try() is called with a $len defined and timeout hits, it prints out
two "not ok"s, then returns immediately. Whereas in the normal case (no
timeout) it runs 4 tests - the 2 at this level, and 2 more when it recurses.

Nicholas Clark

From @jkeenan

On Sun Dec 30 03​:04​:36 2012, nicholas wrote​:

On Tue, Dec 25, 2012 at 06​:22​:00PM -0500, George Greer wrote​:

Smoke logs available at http​://m-
l.org/~perl/smoke/perl/linux/blead_clang_quick/log8d40577bdbdfa85ed3293f84bf26a313b1b92f55.log.gz

Automated smoke report for 5.17.8 patch
8d40577 v5.17.7.0-107-g8d40577
zwei​: Intel(R) Core(TM) i7 CPU 920 @​ 2.67GHz (GenuineIntel 2668MHz)
(x86_64/8 cpu)

Testsuite was run only with 'harness'

Failures​: (common-args) -Accflags=-DPERL_POISON -Dcc=clang
[default]
../dist/Tie-File/t/29_downcopy.t............................FAILED
443-444
Bad plan. You planned 718 tests but ran 716.

I infer that the "bad plan" is because the skip count is wrong if the
timeout hits. The main test code looks like this​:

sub try {
my ($pos, $len, $newlen) = @​_;

...

if ($err) {
if ($err =~ /^Alarm clock/) {
print "# Timeout\n";
print "not ok $N\n"; $N++;
print "not ok $N\n"; $N++;
return;
} else {
$@​ = $err;
die;
}
}

...

if (defined $len) {
try($pos, undef, $newlen);
}
}

Hence if try() is called with a $len defined and timeout hits, it
prints out
two "not ok"s, then returns immediately. Whereas in the normal case
(no
timeout) it runs 4 tests - the 2 at this level, and 2 more when it
recurses.

Nicholas Clark

Would this be fixed simply by printing 'not ok' two more times?

The RT System itself - Status changed from 'new' to 'open'

From @dcollinsn

On Sun Dec 15 13​:56​:26 2013, jkeenan wrote​:

Would this be fixed simply by printing 'not ok' two more times?

I think this is the necessary patch, to print 2 fails if the timeout happens in the recursive call, or 4 fails if it happens in the parent.

--
Respectfully,
Dan Collins

From @dcollinsn

From 7b2ef8c50216b493a4de1405e6a90a516a0f333e Mon Sep 17 00:00:00 2001
From: Dan Collins <dcollinsn@gmail.com>
Date: Thu, 29 Sep 2016 11:23:00 -0400
Subject: [PATCH] RT #115980: Fail the right number of tests on timeout

---
 dist/Tie-File/t/29_downcopy.t | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/dist/Tie-File/t/29_downcopy.t b/dist/Tie-File/t/29_downcopy.t
index d75806d..72943ad 100644
--- a/dist/Tie-File/t/29_downcopy.t
+++ b/dist/Tie-File/t/29_downcopy.t
@@ -279,6 +279,11 @@ sub try {
       print "# Timeout\n";
       print "not ok $N\n"; $N++;
       print "not ok $N\n"; $N++;
+      if ($defined $len) {
+        # Fail the tests in the recursive call as well
+        print "not ok $N\n"; $N++;
+        print "not ok $N\n"; $N++;
+      }
       return;
     } else {
       $@ = $err;
-- 
2.9.3

From [Unknown Contact. See original ticket]

On Sun Dec 15 13​:56​:26 2013, jkeenan wrote​:

Would this be fixed simply by printing 'not ok' two more times?

I think this is the necessary patch, to print 2 fails if the timeout happens in the recursive call, or 4 fails if it happens in the parent.

--
Respectfully,
Dan Collins

From @jkeenan

On Thu Sep 29 08​:27​:52 2016, dcollinsn@​gmail.com wrote​:

On Sun Dec 15 13​:56​:26 2013, jkeenan wrote​:

Would this be fixed simply by printing 'not ok' two more times?

I think this is the necessary patch, to print 2 fails if the timeout
happens in the recursive call, or 4 fails if it happens in the parent.

Had to correct one syntax error in patch ($defined) and reference to incorrect RT number in commit message.

Pushed to blead in commit 5654fe3. Will monitor for no more than 7 days, then close ticket.

Thank you very much.

--
James E Keenan (jkeenan@​cpan.org)

From @jkeenan

On Thu Sep 29 11​:31​:54 2016, jkeenan wrote​:

On Thu Sep 29 08​:27​:52 2016, dcollinsn@​gmail.com wrote​:

On Sun Dec 15 13​:56​:26 2013, jkeenan wrote​:

Would this be fixed simply by printing 'not ok' two more times?

I think this is the necessary patch, to print 2 fails if the timeout
happens in the recursive call, or 4 fails if it happens in the
parent.

Had to correct one syntax error in patch ($defined) and reference to
incorrect RT number in commit message.

Pushed to blead in commit 5654fe3.
Will monitor for no more than 7 days, then close ticket.

Thank you very much.

No complaints registered; marking ticket Resolved.

--
James E Keenan (jkeenan@​cpan.org)

@jkeenan - Status changed from 'open' to 'resolved'