bug in jthread scheduling mechanism?

[Mihai Surdeanu]

> I'll think about it later.  One possibility may be to keep track of the 
> bytes written in the pipe and only read from the pipe if that number's

This was the first patch I used. Afterwards I realized that making sigPipe
nonblocking would also work since reading from the pipe happens only if
there was a write before, and never before. Thus we can not miss
anything, and the overhead seems smaller.

> greater than zero.  But making the pipe non-blocking as you suggest may 
> work too.

Regards,
Mihai