I'll concur with what Craig has just said: "undef" should be fine for all
VMS for the time being.
I took a look at the latest C RTL docs at www.openvms.compaq.com
and found no nanosleep(). I will point out that the original patch was
correct had there been a nanosleep in the relevant RTL.
Note that it _might_ be possible with calls to sys$hiber() and sys$wake()
to emulate a nanosleep(). Such code would need to be introduced in
a place like vms/vms.c (my_nanosleep()?) most likely (if it would even
work with the needed resolution).
Peter Prymmer
"Craig A. Berry"
<craigberry@mac. To: "H.Merijn Brand"
<[EMAIL PROTECTED]>
com> cc: [EMAIL PROTECTED]
09/30/2002 12:31 Subject: Re: nanosleep
PM
At 10:18 AM 9/30/2002 +0200, H.Merijn Brand wrote:
>Would this be enough?
>
>--- configure.com 2002-09-27 13:11:04.000000000 +0200
>+++ configure.com.new 2002-09-30 09:55:13.000000000 +0200
Actually I don't see nanosleep present on any version of VMS, so I think
you'd simply want the following:
--- configure.com 2002-09-27 13:11:04.000000000 +0200
+++ configure.com.new 2002-09-30 09:55:13.000000000 +0200
@@ -5489,6 +5491,7 @@ $ WC "d_unlink_all_versions='undef'"
$ WC "d_unordered='undef'"
$ WC "d_usleep='" + d_usleep + "'"
$ WC "d_usleepproto='" + d_usleep + "'"
+$ WC "d_nanosleep='undef'"
$ WC "d_ustat='undef'"
$ WC "d_vendorarch='undef'"
$ WC "d_vendorlib='undef'"
