From @bbkr

[12​:21] <bbkr> rakudo​: use Test; eval_lives_ok '1<&1'
[12​:21] <p6eval> rakudo 27d54d​: OUTPUT«Use of uninitialized value in numeric
context�ok 1 - �»

The "Use of uninitialized value in numeric context" warning should not be
emited o STDERR, compare it to the case​:

$ perl6 -e 'use Test; eval_lives_ok "die 1"'
not ok 1 -

(no STDERR output is "leaking through" eval_lives_ok, not eval_lives_ok itself
emits one).

From @bbkr

Now it is​:

rakudo c2c02a​: OUTPUT«use of uninitialized variable $v of type Any in numeric context in block
<anon> at eval_0​:1â�¤â�¤ok 1 - â�¤Â»

From @coke

On Thu Sep 02 03​:36​:44 2010, pawel.pabian@​implix.com wrote​:

[12​:21] <bbkr> rakudo​: use Test; eval_lives_ok '1<&1'
[12​:21] <p6eval> rakudo 27d54d​: OUTPUT«Use of uninitialized value in
numeric
context�ok 1 - �»

The "Use of uninitialized value in numeric context" warning should not
be
emited o STDERR, compare it to the case​:

$ perl6 -e 'use Test; eval_lives_ok "die 1"'
not ok 1 -

(no STDERR output is "leaking through" eval_lives_ok, not
eval_lives_ok itself
emits one).

No longer leaking diagnostic fluid​:

01​:24 < [Coke]> rakudo​: use Test; eval_lives_ok '1<&1'
01​:24 <+camelia> rakudo 0d1d11​: OUTPUT«not ok 1 - â�¤# Error​: Whitespace required
  before < operatorâ�¤Â»

Closable with tests.
--
Will "Coke" Coleda

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

From @usev6

I'm afraid the original complaint ist still valid​:

$ perl6 -e 'use Test; eval_lives_ok "1 <&1"'
use of uninitialized value of type Nil in numeric context in block <unit> at EVAL_0​:1

ok 1 -

For some reason now a whitespace is needed before '<' (STD agrees here) and the parsing error prevents the leaking warning​:

$ perl6 -e 'use Test; eval_lives_ok "1<&1"'
not ok 1 -

# Failed test at -e line 1
# Error​: Whitespace required before < operator

From @labster

I'm not sure that this is a real bug. Compare​:

% perl -MTest​::Exception -w -e 'lives_ok sub {print undef}, "foo";'
Use of uninitialized value in print at -e line 1.
ok 1 - foo
...
% perl6 -MTest -e 'lives_ok sub { Str.Str }, "foo"'
use of uninitialized value of type Str in string context in sub at -e​:1

ok 1 - foo

Would these warnings interfere with TAP processing?

On Sat Oct 25 10​:04​:11 2014, bartolin@​gmx.de wrote​:

I'm afraid the original complaint ist still valid​:

$ perl6 -e 'use Test; eval_lives_ok "1 <&1"'
use of uninitialized value of type Nil in numeric context in block
<unit> at EVAL_0​:1

ok 1 -

For some reason now a whitespace is needed before '<' (STD agrees
here) and the parsing error prevents the leaking warning​:

$ perl6 -e 'use Test; eval_lives_ok "1<&1"'
not ok 1 -

# Failed test at -e line 1
# Error​: Whitespace required before < operator

From @jdv

Added tests to S24-testing/3-output.t.

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