In conjunction with trying to fix perldoc's problems with bleadperl (which has to do 
with the nonexistence of SYS$COMMAND in a spawned subprocess), I've been testing 
recent development Perls with Chuck Lane's pipe stress test located here:

<http://www.crinoid.com/perl560/test_pipe.pl>

It turns out the test hangs (precisely the sort of behavior the piping code was 
designed to avoid), and it hangs in 5.7.3 (on test 3) as well as in bleadperl before 
and after my perldoc fix (on test 7).  The hung processes are in LEF state, not RWMBX 
as I would have expected.  My current testing is all with Compaq C S6.5-002 on OpenVMS 
Alpha V7.3.

5.7.2 does not hang, so that leaves 10 months of patches to try to figure out what 
snuck in that is causing the problem.  The only pipe-related thing I can think of that 
has changed is that we now do system() and backticks via the piping code, but the test 
script already uses pipes so I don't see how that could be the issue.  Chuck, if 
you're out there, this could really use your understanding of how the pipe stuff 
works.  Otherwise this will very likely not get fixed before 5.8.0.  I think I'll go 
ahead and submit my patch that gets perldoc working since it makes things no worse 
than they already were.


-- 
____________________________________________
Craig A. Berry                   
mailto:[EMAIL PROTECTED]

"Literary critics usually know what they're
talking about. Even if they're wrong."
        -- Perl creator Larry Wall

Reply via email to