Bleadperl v5.23.5-240-g27e95af breaks PEVANS/Test-MemoryGrowth-0.02.tar.gz #15104
Comments
From @andkKudos once again goes to Slaven for bringing it to my attention. bisect 27e95af is the first bad commit regcomp.c: Try simplifications in some qr/[...]/d sample fail report http://www.cpantesters.org/cpan/report/8d109992-ad47-11e5-9f21-26bde0bfc7aa matrix http://matrix.cpantesters.org/?dist=Test-MemoryGrowth perl -V Summary of my perl5 (revision 5 version 23 subversion 6) configuration: Characteristics of this binary (from libperl): -- |
From @jkeenanOn Mon Dec 28 02:53:06 2015, andreas.koenig.7os6VVqR@franz.ak.mind.de wrote:
Problem confirmed on Linux x86_64: ########## Test Summary Report t/01leak.t (Wstat: 0 Tests: 2 Failed: 1) This is perl 5, version 23, subversion 7 (v5.23.7 (v5.23.6-31-g7c69ad4)) built for x86_64-linux It should be noted that Test-MemoryGrowth is, by the author's own statement (http://search.cpan.org/~pevans/Test-MemoryGrowth-0.02/lib/Test/MemoryGrowth.pm), currently a Linux-specific library. It does not pass tests on other platforms. Thank you very much. -- |
|
The RT System itself - Status changed from 'new' to 'open' |
From @tonycozOn Mon Dec 28 02:53:06 2015, andreas.koenig.7os6VVqR@franz.ak.mind.de wrote:
I tried looking into this but couldn't see the cause. It does behave differently without make test, eg. with make test: $ make test TEST_FILES=t/01leak.t TEST_VERBOSE=1 # Failed test 'no_growth push fails' but when run directly: $ PERL_DL_NONLAZY=1 ~/perl/blead/bin/perl5.23.7 -Mblib -w t/01leak.t The array *is* being extended in the failing case, and get_memusage() does appear to be working correctly - bracketing the call to no_growth() with: system "grep VmSize /proc/$$/status"; revealed no change in VmSize (which is what get_memusage() checks) under make test. I suspect qr// processing is now allocating a fairly large chunk of memory and releasing it, if I initialize @arr like: my $x = "a"; to consume some memory at runtime the tests pass under make test. Tony |
From @khwilliamsonOn 01/03/2016 07:26 PM, Tony Cook via RT wrote:
I think this is a bug in the test. The module author has not gotten
|
From @rjbsPaul believes that the error here is very likely with T::MG and not core. He will have a look into fixing the test, and I'm calling this not a blocker unless we get further evidence that it's something other than a busted test in Paul's library. -- |
|
@rjbs - Status changed from 'open' to 'resolved' |
Migrated from rt.perl.org#127054 (status was 'resolved')
Searchable as RT127054$
The text was updated successfully, but these errors were encountered: