Re: "make check" failure with sort-compress on OSF and Linux

[Dan Hipschman]

On Fri, Feb 23, 2007 at 05:36:57PM +0100, Jim Meyering wrote:
> I was just poking around with gdb, started like this:
> 
>   seq -w 10000 > exp
>   gdb --args ./sort -S 1k exp
> 
> my jaw hit the floor when I saw the 500+ gzip zombies
> left hanging while I'm at a breakpoint in merge.

I'll try doing that when I get a chance to see why it's such a high
number.  There's technically no limit to the number of processes it
could fork, but it should keep the number of zombies down to about two.
I suppose it's possible that it just hasn't gotten a chance to clean
them up since you hit a breakpoint, but these tests show that you were
right all along and we should be avoiding the *need* for fork.