stevens@hsi.UUCP (01/21/85)
We've found two rather major flaws with Venix/86, Version 2.0 on the
PC XT and was wondering if anyone else has encountered them. One we
reported to Unisource in December and have heard nothing back. The other
we have been waiting two days for a call back from Unisource.
(1) If you open an existing file for reading or for read/write, then
position the read/write pointer beyond the EOF with an lseek(),
then do a read() for N bytes, the read() call returns N bytes of
zero !! A quick check of our 4.2 BSD system and a Xenix 3.0 system
on the PC XT showed that this behavior just isn't right.
(2) The popen() routine sometimes hangs the entire system. The following
sequence hangs the system:
popen();
fopen();
after the call to fopen(). However, if you "prime" things with an
extraneous fopen()/fclose() before the popen(), then everything works:
fopen()
fclose();
popen();
fopen();
Perhaps something in their stdio isn't initialized right or popen()
just isn't right. Also, if you move the fopen() before the popen():
fopen();
popen();
then it works OK.
Richard Stevens
Health Systems International, New Haven, CT
ihnp4 ! hsi ! stevens