You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@apr.apache.org by "William A. Rowe, Jr." <wr...@rowe-clan.net> on 2007/10/14 03:45:30 UTC

testcond failure on linux

pipe_consumer() test...

     /* naive fairness test */
     ABTS_INT_EQUAL(tc, 1, !!consumed);
}

at line 383 is failing (expected 1 but !!consumed evaluates to 0).

Can someone familiar with the condvar logic comment, is the test simply
too naive to trust?  Or is there an actual issue?  This happens on;

Linux s170.rcs 2.6.22.9-91.fc7 #1 SMP Thu Sep 27 23:10:59 EDT 2007 i686 i686 i386 GNU/Linux

Re: testcond failure on linux

Posted by Davi Arnaut <da...@apache.org>.
William A. Rowe, Jr. wrote:
> pipe_consumer() test...
> 
>      /* naive fairness test */
>      ABTS_INT_EQUAL(tc, 1, !!consumed);
> }
> 
> at line 383 is failing (expected 1 but !!consumed evaluates to 0).
> 
> Can someone familiar with the condvar logic comment, is the test simply
> too naive to trust?  Or is there an actual issue?  This happens on;
> 

It's a naive test that's supposed to show when one thread got starved
and never got a chance to read from the pipe. Safe to ignore, and I
forgot to remove it.

--
Davi Arnaut