home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.minix
- Path: sparky!uunet!ferkel.ucsb.edu!taco!rock!stanford.edu!agate!spool.mu.edu!think.com!rpi!batcomputer!munnari.oz.au!metro!ipso!runxtsa!bde
- From: bde@runx.oz.au (Bruce Evans)
- Subject: Re: alarm clock (Re:elvis)
- Message-ID: <1992Dec20.025326.19033@runx.oz.au>
- Organization: RUNX Un*x Timeshare. Sydney, Australia.
- References: <BzCpns.Mw@ucunix.san.uc.edu>
- Date: Sun, 20 Dec 92 02:53:26 GMT
- Lines: 17
-
- In article <BzCpns.Mw@ucunix.san.uc.edu> schoudhu@ucunix.san.uc.edu (Spandan Choudury) writes:
- > just wondering if a standard solution to the above bug has been
- > developed yet..
-
- It was fixed in elvis several versions ago. elvis-1.6 is the current version.
-
- I think the problem was a race condition in SYSV-style signal handling. The
- problem may have been amplified by slow Minix systems and/or Minix scheduling.
- I never saw it on a fast Minix system with different scheduling. Minix-1.6.23
- has better (POSIX) signal handling so the old version of elvis would probably
- work fine, at least if you compile the Minix-1.6.23 library signal.c with
- -DBASH to get reliable (BSD) signals. The standard signal() provides
- unreliable (SYSV) signals for backwards compatibility, but there aren't many
- applications that care. minicom is one. I think it forgets to unset a
- SIGALRM handler.
- --
- Bruce Evans (bde@runx.oz.au)
-