The secure popen() code worked after all

[Eric S. Raymond]

In <2285c429:2919 at snark.UUCP> I wrote:
>OK, I give up. Can someone *else* spot the bug in this code?

This will teach me not to post when I'm fried and frustrated. It turned out
that the code itself actually works fine -- I really did do the Right Things,
there was no such `bug in this code'.

The bug was in my *test* code...<blush>.

John Levine pointed out that the SIGPIPE abort had to mean the child process
was dying prematurely. This put me on the right track.

Now I'm not sure whether to feel more pleased about getting the tricky parts
right or upset about getting the `simple' part wrong.

I think there's a moral in here somewhere...
-- 
      Eric S. Raymond                     (the mad mastermind of TMN-Netnews)
      UUCP: {{uunet,rutgers,ihnp4}!cbmvax,rutgers!vu-vlsi,att}!snark!eric
      Post: 22 South Warren Avenue, Malvern, PA 19355   Phone: (215)-296-5718