Tests blocking on "fifoor"

[Marcin Cieslak]

I had a quick look at this. Some tests (dup05, fcntl07B) wait
on close() of the O_RDWR fifo. It is usually enough to supply some
data to the fifo (located in /tmp/ somewhere) and the test exits cleanly (and 
usually passes).

I don't understand enough of FIFO blocking/non-blocking behaviour, but generally 
it's normal that process blocks when no other process is ready to read or write
from the other side of the pipe.

I will have look at this as soon as I get some fifo clue.

-- 
               << Marcin Cieslak // saper at system.pl >>