by Well, this one sure is weird. Given a Solaris 2.4 Korn or Bourne
shell with no soft or hard limit on core dumps, first lower the soft
limit to 0. Then try and remove the limit by setting it to 2147483647
(0x7fffffff, the value that according to <sys/resource.h> is
RLIM_INFINITY). This should remove the limit, right?
Wrong. The shell curtly says "ulimit: bad ulimit", and refuses to
accept that value. In fact, ksh refuses to take anything higher than
20971519 (0x13fffff), and sh refuses anything above 4194303 (0x3fffff)!
Of course, one can always say "ulimit -S -c unlimited", but why should
perfectly acceptable (to setrlimit(2)) numbers be arbitrarily rejected
like this?
--
David Barts N5JRN | UW Civil Engineering, Box 352700 | Seattle, WA 98195-2700
1900 GMT T: 52 F wind: S 11 gust 21 mph P: 1009 mbar