BBC: Recent commit breaks test in B::DeparseTree

[jkeenan]

A recent commit to blead has broken a test file in CPAN distribution B-DeparseTree.

Sample test output:

[freebsd13: B-DeparseTree-v3.4.0-1] $ bleadprove -vb t/04-indirops.t
t/04-indirops.t .. 
ok 1 - print: lex=0 print() => print($_)
ok 2 - print: lex=1 print() => print($_)
ok 3 - print: lex=0 CORE::print() => print($_)
ok 4 - print: lex=1 CORE::print() => print($_)
...
ok 23 - printf: lex=0 CORE::printf($a, $b) => printf($a, $b)
ok 24 - printf: lex=1 CORE::printf($a, $b) => printf($a, $b)
Not enough arguments for sort at (eval 102) line 5, at EOF
 in sort at /usr/home/jkeenan/.cpan/build/B-DeparseTree-v3.4.0-1/t/helper.pm line 109, <$data_fh> line 33.
# Tests were run but no plan was declared and done_testing() was not seen.
# Looks like your test exited with 25 just after 24.
Dubious, test returned 25 (wstat 6400, 0x1900)
All 24 subtests passed 

Test Summary Report
-------------------
t/04-indirops.t (Wstat: 6400 Tests: 24 Failed: 0)
  Non-zero exit status: 25
  Parse errors: No plan found in TAP output
Files=1, Tests=24,  0 wallclock secs ( 0.07 usr  0.00 sys +  0.30 cusr  0.05 csys =  0.41 CPU)
Result: FAIL

The test failure has already been reported upstream by @eserte at RT 141404.

The most likely commit in Perl 5 blead to have triggered this problem is:

commit 78cc98885ff2800cd3a0eda0b60a0cbb113498a9
Author:     David Mitchell <davem@iabyn.nospamdeletethisbit.com>
AuthorDate: Wed Jan 19 20:05:28 2022 +0000
Commit:     David Mitchell <davem@iabyn.nospamdeletethisbit.com>
CommitDate: Wed Jan 19 20:38:02 2022 +0000

    make a sort with zero args a compile-time err
    
        @a = sort @empty; # unaffected
        @a = sort;        # now a compile-time error
    
    the main motivation for this is to potentially be able to treat a
    colon following the 'sort' keyword as part of the sort expression,
    to then be able to add sort attributes at some point, e.g.
        @a = sort :num, 1,2,3;
    rather than the colon being seen as part of a ?: conditional.
    
    See http://nntp.perl.org/group/perl.perl5.porters/262425.

@iabyn, can you investigate?

Thank you very much.
Jim Keenan

[jkeenan]

I have confirmed the first bad commit:

78cc98885ff2800cd3a0eda0b60a0cbb113498a9 is the first bad commit
commit 78cc98885ff2800cd3a0eda0b60a0cbb113498a9
Author: David Mitchell <davem@iabyn.com>
Date:   Wed Jan 19 20:05:28 2022 +0000

    make a sort with zero args a compile-time err

[hvds]

Looks like we're missing a perldelta entry for that commit, I'll try to put one together tomorrow.

B::DeparseTree will need to adapt for the changed definition, I imagine that won't be difficult - maybe as simple as removing the failing test, maybe making it conditional on perl version.

[hvds]

Looks like we're missing a perldelta entry for that commit, I'll try to put one together tomorrow.

Now put together as #19537.

[hvds]

Looks like we're missing a perldelta entry for that commit, I'll try to put one together tomorrow.

Now put together as #19537.

Oh, it was already in perl5358delta.pod, never mind me.

[jkeenan]

The upstream ticket was created by @eserte and can be found at RTC 141404.

@iabyn, can you take a look at this ticket?

Thank you very much.
Jim Keenan

[rjbs]

The tests for B-DeparseTree appear to generate and eval the code string sort (). So, there you go.

The tests should be updated. This library does not appear to be a dependency for anything. I expect a patch would not be particularly difficult, but the tests are definitely a bit complex, so I did not make even a modest attempt.

I will not yet label this non-5.36-blocker.

[rjbs]

I am now marking this a non-blocker. An issue has been filed with the maintainer (thanks, @jkeenan) and it includes the most basic fix possible. The maintainer is likely to know how to fix this more correctly.