develooper Front page | perl.perl5.porters | Postings from August 2018

[perl #133429] dist/Time-HiRes/t/itimer.t: intermittent failures onLinux

Thread Next
From:
James E Keenan
Date:
August 6, 2018 21:18
Subject:
[perl #133429] dist/Time-HiRes/t/itimer.t: intermittent failures onLinux
Message ID:
rt-4.0.24-16006-1533590329-271.133429-75-0@perl.org
# New Ticket Created by  James E Keenan 
# Please include the string:  [perl #133429]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org/Ticket/Display.html?id=133429 >


For a couple of weeks I have been getting strange, intermittent test 
failures in dist/Time-HiRes/t/itimer.t when tested against Perl 5 blead 
on Linux.  The failures first appeared in the course of smoke-testing 
and have appeared in runs with various combinations of 
threaded/unthreaded and debugging/regular.  The smoke reports in 
question were:

http://perl5.test-smoke.org/report/68572

http://perl5.test-smoke.org/report/68549

http://perl5.test-smoke.org/report/68640

However, due to what I believe is a bug in the code for 
perl5.test-smoke.org, nothing about dist/Time-HiRes/t/itimer.t appears 
on those web pages.  The reason for the 'F' grades only appears when I 
examine the ./logs/smokecurrent/* files for each of the runs in 
question.  For example, for the run reported in report 68640, I see this 
in logs/smokecurrent/rptd342516410da8af9d1824e50179d7d6ff3c7ba95.rpt:

#####
Testsuite was run only with 'harness'

Tests skipped on user request:
     # One test name on a line
Failures: (common-args) none
[stdio]
[stdio/perlio] -DDEBUGGING -Duseithreads
../dist/Time-HiRes/t/itimer.t...............................??????
#####

And if I examine the overall log file for this run, I see 3 instances of 
this:

#####
error while running harness target 'test_harness': 2 at 
/home/jkeenan/p5smoke/lib/Test/Smoke/App/RunSm
oke.pm line 118.
[2018-08-06 15:35:36-0400] [unknown failure] 
../dist/Time-HiRes/t/itimer.t...............................
[2018-08-06 15:35:36-0400]
[2018-08-06 15:35:36-0400] 
../dist/Time-HiRes/t/itimer.t...............................??????
#####

Outside of a smoke-testing context, I built blead with threads and 
debugging on the same machine on which the smoke reports were generated 
(see attachment for configuration).  I got these results:

#####
$ cd t;./perl harness -v ../dist/Time-HiRes/t/itimer.t; cd -

ok 1 - ITIMER_VIRTUAL defined with sufficient granularity
ok 2 - time left should be zero
All 2 subtests passed

Test Summary Report
-------------------
../dist/Time-HiRes/t/itimer.t (Wstat: 26 Tests: 2 Failed: 0)
   Non-zero wait status: 26
Files=1, Tests=2,  0 wallclock secs ( 0.00 usr  0.00 sys +  0.58 cusr 
0.01 csys =  0.59 CPU)
Result: FAIL
/home/jkeenan/gitwork/perl
#####

So it appears that the unit tests in the file are PASSing but for some 
reason the file as a whole is being graded FAIL.

I haven't been conducting smoke testing on this machine for very long, 
so I don't have a sense of when this problem first appeared.  The 
problem tends to occur more with -Duseithreads and -DDEBUGGING, but as 
report 68640 above indicates, it can occur even with out either of those 
switches.  This platform may be more resource-constrained than other 
platforms on which I test blead.

Any ideas?

Thank you very much.
Jim Keenan

Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About