Bug#501411: Took a quick look at this

[Damyan Ivanov]

-=| Damyan Ivanov, Fri, Oct 10, 2008 at 11:58:19AM +0300 |=-
> -=| Michael Casadevall, Fri, Oct 10, 2008 at 01:21:42AM -0400 |=-
> > I rebuilt it a few times, and it always passed. If it possible 
> > that this is just a random fluke, and this bug can be closed, or 
> > at least downgraded?
> 
> Could be. But if it is 'random', then it should fail from time to time 
> (i.e. when building it 100 times in a row). It isn't failing (here) so 
> the reason should be somewhere else.

I noticed that the build took an unusually long time. From the build  
log[1]:

    [1] http://newpeople.debian.org/~lucas/logs/2008/10/06/libdanga-socket-perl_1.59-1_llenny32.buildlog

Finished at 20081007-0349
Build needed 00:04:04, 292k disk space
DC-Build-Status: Failed 268.977224s
DC-Time-Estimation: 268.977224 versus expected 94 (r/m: 1.86145982978723 ; m: 94.0)

The failing tests check if a scheduled event comes after the specified 
time with a tollerance of one second. If the system is overloaded, 
this could fail.

I was able to produce such a failure when running the tests under 
'nice' and in parallel with two other[2] "nice -n -20 burnP6" 
processes.

  [2] I needed two as otherwise the multithreaded CPU handled the load 
  seamlesly.


Ways out that I see:

 1. Consider this to be not release-critical
 2. Modify t/05-postloop time tests not to fail, but only warn
 3. Modify t/05-postloop to skip the actial tests under high load.

-- 
dam            JabberID: dam at jabber.minus273.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
Url : http://lists.alioth.debian.org/pipermail/pkg-perl-maintainers/attachments/20081010/ffb1a483/attachment.pgp