Date: Sun, 18 Oct 1998 10:20:14 +0800 From: Peter Wemm <peter@netplex.com.au> To: "Jordan K. Hubbard" <jkh@time.cdrom.com> Cc: Marc Slemko <marcs@znep.com>, current@FreeBSD.ORG Subject: Re: sleep() and Apache in release notes Message-ID: <199810180220.KAA12204@spinner.netplex.com.au> In-Reply-To: Your message of "Sat, 17 Oct 1998 18:58:01 MST." <24750.908675881@time.cdrom.com>
next in thread | previous in thread | raw e-mail | index | archive | help
"Jordan K. Hubbard" wrote:
> > (don't know who wrote what part of the release notes, but...)
>
> I'm not really sure who wrote that (possibly Garrett, but I wouldn't
> swear on it) but I'm sorry for the unfortunate choice of words. I
> should have proof-read the release notes more carefully - mea culpa!
>
> I don't think that any intentional slight was meant by it, just an
> off-the-cuff comment in engineer-speak that problably shouldn't have
> gone into the public docs. Once we figure out who added it (I'm on a
> slow link at the moment or I'd inspect the CVS logs), we can determine
> just what was meant by it.
Earlier on, it was a real problem because httpd was depending on alarm() to
wake up a sleep and depending on the SIGALRM handler not being called. This
was quite some time ago, my memory isn't all that good from back then. I
have a feeling that it was from the 1.1 or 1.2 beta releases. I am sure it
is no longer the case because the syscall that was originally added to
implement these semantics has since been removed. signanosleep(2) is gone
and sleep/usleep use normal nanosleep(2), so this note is out of date even.
> - Jordan
>
> >
> > It is somewhat disappointing to see:
> >
> > o sleep(3) and usleep(3) are now implemented in terms of signanosleep(2)
> > and now have correct SIGALRM interaction semantics and sleep(3) correctly
> > returns the time remaining. Some programs (notably apache httpd) bogusly
> > depend on a sleep() "absorbing" a SIGALRM from a timer that expires durin
g
> > the life of the sleep.
> >
> > in RELNOTES.TXT considering I have no idea where the bug in Apache that is
> > claiming would be and there have been no bug report or attempt to get this
> > fixed in Apache other than sticking lines in the release notes saying
> > "Apache is broken".
> >
> > If there is a problem in Apache, please point it out so it can be fixed.
> >
> > --
> > Marc Slemko | Apache Group member
> > marcs@znep.com | marc@apache.org
Cheers,
-Peter
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199810180220.KAA12204>