Bug #81449 for Tie-Hash-Indexed: Test failures due to hash randomisation in perl 5.17.6 and semantic change in scalar %hash in 5.25.3

Ditto +1 from me. The upcoming MCE 1.7 release will come with MCE::Shared allowing ordered hashes with full support for deep data sharing. Your module is one of three allowed. use Tie::Hash::Indexed; use MCE::Flow max_workers => 4; use MCE::Shared ordered => 1; my $db = mce_share_m {}; # Tie::Hash::Indexed behind the scene with MCE::Mutex locking $db->{key1} = 'foo'; $db->{key2} = 'bar'; $db->{key3} = 'baz'; $db->{counter} = 0; mce_flow sub { $db->lock(); $db->{counter} += 1; $db->unlock(); }; print $db->{counter}, "\n"; # 4 Unfortunately, am placed in an uncomfortable place and having to make a decision on wether to allow ordered keys from your incredible module. I am not totally sure if your module is working with Perl v5.18 and above. Look at T::H::I go... :-) 288,000 FETCH's/STORE's Tie::Hash::Indexed 0.307 secs Hash::Ordered 0.443 secs Tie::IxHash 0.519 secs +1 for Marcus and T::H::I Regards, - Mario

Hi Marcus, The 101_basic.t script is the reason for failing. The following diff allows Tie::Hash::Indexed to build on newer Perl versions. Thus, zero changes to the underlying XS code. diff -uar Tie-Hash-Indexed-0.05/t/101_basic.t Tie-Hash-Indexed-0.05-fix/t/101_basic.t --- Tie-Hash-Indexed-0.05/t/101_basic.t 2007-08-24 09:18:54.000000000 -0400 +++ Tie-Hash-Indexed-0.05-fix/t/101_basic.t 2016-05-11 09:18:30.000000000 -0400 @@ -39,20 +39,17 @@ ok(exists $h{foo}); ok(exists $h{bar}); ok(!exists $h{xxx}); -$s = &scalar_h; -skip($scalar, $s =~ /^(\d+)\/\d+$/ && $1 == scalar keys %h); +ok(scalar keys %h, 4); $h{xxx} = 5; ok(join(',', keys %h), 'foo,bar,zoo,baz,xxx'); ok(exists $h{xxx}); -$s = &scalar_h; -skip($scalar, $s =~ /^(\d+)\/\d+$/ && $1 == scalar keys %h); +ok(scalar keys %h, 5); $h{foo} = 6; ok(join(',', keys %h), 'foo,bar,zoo,baz,xxx'); ok(exists $h{foo}); -$s = &scalar_h; -skip($scalar, $s =~ /^(\d+)\/\d+$/ && $1 == scalar keys %h); +ok(scalar keys %h, 5); while (my($k,$v) = each %h) { $key .= $k; Regards, Mario

The following provides a parallel demonstration for sharing an ordered hash between workers. use strict; use warnings; use List::Util 'shuffle'; use Time::HiRes 'time'; use MCE::Hobo; use MCE::Shared; use Tie::Hash::Indexed; use MCE::Shared::Ordhash; use Hash::Ordered; srand 1618; my $c = shift || 1; # (1) Tie::Hash::Indexed # (2) MCE::Shared::Ordhash # (3) Hash::Ordered my $max = 1_000_000; my @rand = shuffle(1..$max); my $seq = MCE::Shared->sequence( { chunk_size => 500, bounds_only => 1 }, 1, $max ); my $oh1 = MCE::Shared->share( Tie::Hash::Indexed->TIEHASH ); my $oh2 = MCE::Shared->share( MCE::Shared::Ordhash->new ); my $oh3 = MCE::Shared->share( Hash::Ordered->new ); my $oh; if ($c eq '1') { print $oh1->blessed, "\n"; $oh = $oh1 } elsif ($c eq '2') { print $oh2->blessed, "\n"; $oh = $oh2 } elsif ($c eq '3') { print $oh3->blessed, "\n"; $oh = $oh3 } else { print "usage: $0 [ 1 | 2 | 3 ]\n" } my $s = time; sub _store { while ( my ( $beg, $end ) = $seq->next ) { $oh->STORE($_,$_) for $beg..$end; } }; MCE::Hobo->create('_store') for 1..8; MCE::Hobo->waitall; printf "store : %0.3f secs.\n", time - $s; $seq->rewind; my $d = time; sub _delete { while ( my ( $beg, $end ) = $seq->next ) { $oh->DELETE($_) for $beg..$end; } }; MCE::Hobo->create('_delete') for 1..8; MCE::Hobo->waitall; printf "delete : %0.3f secs.\n", time - $d; printf "elapse : %0.3f secs.\n", time - $s; print "\n"; __END__ $ perl bench_shared.pl 1 Tie::Hash::Indexed store : 2.803 secs. delete : 2.553 secs. elapse : 5.356 secs. $ perl bench_shared.pl 2 MCE::Shared::Ordhash store : 3.093 secs. delete : 3.810 secs. elapse : 6.903 secs. $ perl bench_shared.pl 3 Hash::Ordered store : 3.434 secs. delete : 4.325 secs. elapse : 7.760 secs. It is really nice and cool for T::H::I to run on Perl 5.18 and later :) Thank you, Mario

On Tue Apr 04 05:01:59 2017, NKH wrote: Show quoted text
And still failing with perl-5.27.9: ##### [Tie-Hash-Indexed-0.05] 522 $ bleadprove -vb t/101_basic.t t/101_basic.t .. 1..32 # Running under perl version 5.027009 for linux # Current time local: Wed Feb 21 21:33:52 2018 # Current time GMT: Thu Feb 22 02:33:52 2018 # Using Test.pm version 1.31 ok 1 ok 2 ok 3 ok 4 ok 5 ok 6 ok 7 not ok 8 ok 9 ok 10 not ok 11 ok 12 ok 13 not ok 14 ok 15 ok 16 ok 17 ok 18 ok 19 ok 20 ok 21 ok 22 ok 23 ok 24 ok 25 ok 26 ok 27 ok 28 ok 29 ok 30 ok 31 ok 32 # Failed test 8 in t/101_basic.t at line 43 # t/101_basic.t line 43 is: skip($scalar, $s =~ /^(\d+)\/\d+$/ && $1 == scalar keys %h); # Failed test 11 in t/101_basic.t at line 49 # t/101_basic.t line 49 is: skip($scalar, $s =~ /^(\d+)\/\d+$/ && $1 == scalar keys %h); # Failed test 14 in t/101_basic.t at line 55 # t/101_basic.t line 55 is: skip($scalar, $s =~ /^(\d+)\/\d+$/ && $1 == scalar keys %h); Failed 3/32 subtests Test Summary Report ------------------- t/101_basic.t (Wstat: 0 Tests: 32 Failed: 3) Failed tests: 8, 11, 14 Files=1, Tests=32, 0 wallclock secs ( 0.03 usr + 0.00 sys = 0.03 CPU) Result: FAIL #####

Bleadperl v5.25.2-28-g8bf4c40 introduced that scalar %hash == keys %hash and as such broke some tests in MHX/Tie-Hash-Indexed-0.05_03.tar.gz : https://perl5.git.perl.org/perl.git/commit/8bf4c40 The bleadperl BBC ticket for this change was https://rt.perl.org/Public/Bug/Display.html?id=128482, closed because nobody had reported Tie-Hash-Indexed-0.05_03 (possibly because it is a dev release).