t/op/lexsub.t: A harness curiosum #23648
Description
Observed today on blead at f69a90f. Begin by configuring and running make minitest_prep.
$ ./miniperl -Ilib t/op/lexsub.t 2>&1 | head -n 22
1..158
ok 1 - calling our sub from same package
...
ok 16 - our subs from other packages override all keywords
ok 17 - our sub pointing to "use constant" constant
ok 18 - our sub symbol translation does not affect meth names
#4242
ok 19 - our sub symbol translation & meth names after print
ok 20 - name_of_our_sub <newline> => is parsed properly
Now, complete the build (make test_prep) and run with ./perl:
$ ./perl -Ilib t/op/lexsub.t 2>&1 | head -n 22
1..158
ok 1 - calling our sub from same package
...
ok 16 - our subs from other packages override all keywords
ok 17 - our sub pointing to "use constant" constant
ok 18 - our sub symbol translation does not affect meth names
#4242
ok 19 - our sub symbol translation & meth names after print
ok 20 - name_of_our_sub <newline> => is parsed properly
But now, run with the harness (output trimmed):
$ cd t;./perl harness -v op/lexsub.t 2>&1 ; cd -
ok 1 - calling our sub from same package
ok 2 - calling our sub from same package (amper)
ok 3 - calling our sub from another package
ok 4 - calling our sub from another package (amper)
ok 5 - our sub falling out of scope
ok 6 - our sub falling out of scope (called via amper)
ok 7 - our sub invisible inside itself
ok 8 - our sub invisible inside itself (called via amper)
ok 9 - our sub visible inside itself after decl
ok 10 - our sub visible inside itself after decl (amper)
ok 11 - our sub foo; makes lex alias for existing sub
ok 12 - our sub foo; makes lex alias for existing sub (amper)
ok 13 - our sub foo; applies to subsequent sub foo {}
ok 14 - our sub with proto
ok 15 - lexical subs (even our) override all keywords
ok 16 - our subs from other packages override all keywords
ok 17 - our sub pointing to "use constant" constant
ok 18 - our sub symbol translation does not affect meth names
ok 19 - our sub symbol translation & meth names after print
ok 20 - name_of_our_sub <newline> => is parsed properly
ok 21 - sort our_sub
ok 22 - pad cleanup for a closure referring to an our sub
...
Where did the line with #4242 go?
But yet again, when run through the harness, then piped to head:
$ cd t;./perl harness -v op/lexsub.t 2>&1 | head -n 22; cd -
op/lexsub.t ..
1..158
ok 1 - calling our sub from same package
ok 2 - calling our sub from same package (amper)
ok 3 - calling our sub from another package
ok 4 - calling our sub from another package (amper)
ok 5 - our sub falling out of scope
ok 6 - our sub falling out of scope (called via amper)
ok 7 - our sub invisible inside itself
ok 8 - our sub invisible inside itself (called via amper)
ok 9 - our sub visible inside itself after decl
ok 10 - our sub visible inside itself after decl (amper)
ok 11 - our sub foo; makes lex alias for existing sub
ok 12 - our sub foo; makes lex alias for existing sub (amper)
ok 13 - our sub foo; applies to subsequent sub foo {}
ok 14 - our sub with proto
ok 15 - lexical subs (even our) override all keywords
ok 16 - our subs from other packages override all keywords
ok 17 - our sub pointing to "use constant" constant
ok 18 - our sub symbol translation does not affect meth names
#4242
ok 19 - our sub symbol translation & meth names after print
Now #4242 is back. Why?
Relevant part of `t/op/lexsub.t:
86 # our sub and method confusion
87 sub F::h { 4242 }
88 {
89 my $called;
90 our sub h { ++$called; 4343 };
91 is((h F),4242, 'our sub symbol translation does not affect meth names');
92 undef $called;
93 print "#";
94 print h F; # follows a different path through yylex to intuit_method
95 print "\n";
96 is $called, undef, 'our sub symbol translation & meth names after print'
97 }
What am I not grokking?