Bug #129197 for Audio-FindChunks: t/Audio-FindChunks.t failures

Mon Apr 15 16:27:30 2019 SREZIC [...] cpan.org - Ticket created

Subject:

t/Audio-FindChunks.t failures

On some of my smoker systems the test suite fails like this: ... "/usr/local/bin/perl" -MExtUtils::Command::MM -e 'cp_nonempty' -- FindChunks.bs blib/arch/auto/Audio/FindChunks/FindChunks.bs 644 PERL_DL_NONLAZY=1 "/usr/local/bin/perl" "-MExtUtils::Command::MM" "-MTest::Harness" "-e" "undef *Test::Harness::Switches; test_harness(0, 'blib/lib', 'blib/arch')" t/*.t # Test 8 got: "" (t/Audio-FindChunks.t at line 87) # Expected: "1" (threshold 1360.59680547217 < 1100) # t/Audio-FindChunks.t line 87 is: ok($t < 1100,1, "threshold $t < 1100"); # in v2.01/Intel: 1078.17744444174 (in 20.7122111714316 .. 7070.48043297351) # threshold: 1360.59680547217 (in 351.413154907691 .. 7079.3041586709) 4.4 =25.8 # n=1 duration 21.4; gap 4.4 (4.4 .. 25.8; 21.4) 27.7 =67 # n=2 duration 39.3; gap 1.9 (27.7 .. 1m07.0; 39.3) Frequency: 44100. Stride: 4; 2 channels. Chunk=0.1sec=17640bytes. ch0: -23296.0 .. 9999.0 (-3dB;-10dB). ch1: -23131.0 .. 9999.0 (-3dB;-10dB). 0: 0.0: 351.4= -36dB: #################### 1: 0.1: 351.4= -36dB: #################### 2: 0.2: 351.4= -36dB: #################### 3: 0.3: 351.4= -36dB: #################### 44: 4.4: 351.4= -36dB: #################### 45: 4.5: 351.4= -36dB: #################### 46: 4.6: 351.4= -36dB: #################### 47: 4.7: 351.4= -36dB: #################### 48: 4.8: 351.4= -36dB: #################### 49: 4.9: 351.4= -36dB: #################### 50: 5.0: 7079.3= -10dB: ############################> 51: 5.1: 7079.3= -10dB: ############################> 52: 5.2: 7079.3= -10dB: ############################> 53: 5.3: 7079.3= -10dB: ############################> # Test 22 got: "" (t/Audio-FindChunks.t at line 87 fail #2) # Expected: "1" (threshold 1360.59680547217 < 1100) t/Audio-FindChunks.t .. Failed 2/37 subtests ... This happens with all available perls (5.28.0 .. 5.29.9) on my freebsd13 smoker, and with one perl (perl 5.28.2-RC1) on my fedora28 smoker.

Thu Apr 18 02:20:58 2019 nospam-abuse [...] ilyaz.org - Correspondence added

CC:	;
Subject:	Re: [rt.cpan.org #129197] t/Audio-FindChunks.t failures
Date:	Wed, 17 Apr 2019 23:20:43 -0700
To:	Slaven_Rezic via RT <bug-Audio-FindChunks [...] rt.cpan.org>
From:	Ilya Zakharevich <nospam-abuse [...] ilyaz.org>

On Mon, Apr 15, 2019 at 04:27:31PM -0400, Slaven_Rezic via RT wrote: Show quoted text

> Mon Apr 15 16:27:30 2019: Request 129197 was acted upon. > Transaction: Ticket created by SREZIC > Queue: Audio-FindChunks > Subject: t/Audio-FindChunks.t failures > Broken in: 2.03 > Severity: (no value) > Owner: Nobody > Requestors: SREZIC@cpan.org > Status: new > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=129197 > > > > On some of my smoker systems the test suite fails like this:

Thanks, Slaven, this remains unsolved from 2008. I cannot reproduce it, and, judging by CPAN testers data, the conditions for failing are quite delicate. If you could help in testing this, this would be a lot of help! (This distribution has my engine for remote debugging of crashes — but I have no clue what to do remotely with “just failures”…) Just in case, I attach a copy of the initially generated data (tmp.rms) from my machine. Could you test it against tmp.rms in the build directory of a failing Audio::FindChunk? If we are lucky, the difference would be already visible there in this file… (And if it is visible, it may happen that the difference is in the generated .wav file — which is normally removed at the end of the test. If there are two configurations which fail and succeed, one could make clean export AUDIO_FH_TEST_UNLINK=0 then redo tests, and compare the files tmp.wav.) Thanks again, Ilya

Download tmp.rms-ok.bz2
application/octet-stream 465b

Message body not shown because it is not plain text.

Thu Apr 18 02:20:59 2019 The RT System itself - Status changed from 'new' to 'open'

Thu Apr 18 05:10:20 2019 nospam-abuse [...] ilyaz.org - Correspondence added

Subject:	Re: [rt.cpan.org #129197] t/Audio-FindChunks.t failures
Date:	Thu, 18 Apr 2019 02:10:11 -0700
To:	Slaven_Rezic via RT <bug-Audio-FindChunks [...] rt.cpan.org>
From:	Ilya Zakharevich <nospam-abuse [...] ilyaz.org>

On Wed, Apr 17, 2019 at 11:20:43PM -0700, Ilya Zakharevich wrote: Show quoted text

> > On some of my smoker systems the test suite fails like this:

> > Thanks, Slaven, > > this remains unsolved from 2008. I cannot reproduce it, and, judging > by CPAN testers data, the conditions for failing are quite delicate.

BTW, just to have it collected somewhere: before your reports, it was like the summary below (unfortunately, all this is a manual search; “Reports analysis” cannot find anything reasonable). Ilya XXX Keeping reports which look like exceptions separate: ###### has ptrsize=4 69120440 FAIL 8 arm-freebsd-thread-multi 8 -O 11.0-current 4 define 5.22.2 ###### have no -O3, or useithreads=undef 102750525 FAIL 8 x86_64-linux-thread-multi 8 -O2 -g 4.15.0-45-generic 8 define 5.29.8 69332235 FAIL 8 amd64-netbsd 8 -O 6.1.4 8 undef 5.20.1 69314827 FAIL 8 amd64-netbsd 8 -O 6.1.4 8 undef 5.20.0 ###### have -O3 and useithreads=undef 78830064 FAIL 8 cygwin 8 -O3 2.6.1(0.30553) 8 undef 5.18.3 78804849 FAIL 8 cygwin-multi 8 -O3 2.6.1(0.30553) 8 undef 5.20.1 The rest (11 reports) have -O3, ptrsize=alignbytes=nvsize=8, useithreads=define (48 reports with these conditions). 78979796 FAIL 8 cygwin-thread-multi 8 -O3 2.6.1(0.30553) 8 define 5.18.3 78979430 FAIL 8 cygwin-thread 8 -O3 2.6.1(0.30553) 8 define 5.18.1 78979233 FAIL 8 cygwin-thread-multi 8 -O3 2.6.1(0.30553) 8 define 5.18.0 78956430 FAIL 8 cygwin-thread-multi 8 -O3 2.6.1(0.30553) 8 define 5.22.1 78949215 FAIL 8 cygwin-thread-multi 8 -O3 2.6.1(0.30553) 8 define 5.16.2 78924652 FAIL 8 cygwin-thread-multi 8 -O3 2.6.1(0.30553) 8 define 5.22.3 78924523 FAIL 8 cygwin-thread 8 -O3 2.6.1(0.30553) 8 define 5.25.6 78894412 FAIL 8 cygwin-thread-multi 8 -O3 2.6.1(0.30553) 8 define 5.16.1 78887212 FAIL 8 cygwin-thread-multi 8 -O3 2.6.1(0.30553) 8 define 5.22.0 78884727 FAIL 8 cygwin-thread-multi 8 -O3 2.6.1(0.30553) 8 define 5.25.2 69175580 FAIL 8 darwin-thread-multi-2level 8 -O3 14.5.0 8 define 5.20.3

Fri Apr 19 07:22:36 2019 SREZIC [...] cpan.org - Correspondence added

RT-Send-CC:

nospam-abuse [...] ilyaz.org

On 2019-04-18 02:20:58, nospam-abuse@ilyaz.org wrote: Show quoted text

> On Mon, Apr 15, 2019 at 04:27:31PM -0400, Slaven_Rezic via RT wrote:

> > Mon Apr 15 16:27:30 2019: Request 129197 was acted upon. > > Transaction: Ticket created by SREZIC > > Queue: Audio-FindChunks > > Subject: t/Audio-FindChunks.t failures > > Broken in: 2.03 > > Severity: (no value) > > Owner: Nobody > > Requestors: SREZIC@cpan.org > > Status: new > > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=129197 > > > > > > > On some of my smoker systems the test suite fails like this:

> > Thanks, Slaven, > > this remains unsolved from 2008. I cannot reproduce it, and, judging > by CPAN testers data, the conditions for failing are quite delicate. > If you could help in testing this, this would be a lot of help! > > (This distribution has my engine for remote debugging of > crashes — but I have no clue what to do remotely with “just > failures”…) > > Just in case, I attach a copy of the initially generated data > (tmp.rms) from my machine. Could you test it against tmp.rms in the > build directory of a failing Audio::FindChunk? If we are lucky, the > difference would be already visible there in this file… > > (And if it is visible, it may happen that the difference is in the > generated .wav file — which is normally removed at the end of the > test. If there are two configurations which fail and succeed, one > could > make clean > export AUDIO_FH_TEST_UNLINK=0 > then redo tests, and compare the files tmp.wav.) >

If you like, then I can provide ssh access to the problematic VM. Just send me your ssh public key, and I'll send you the access information. Regards, Slaven