jgs@sfmag.UUCP (J.G.Smits) (10/01/86)
When the SVID was published, one of the goals behind it was to provide "an operating system environment that allows users to create applications software that is independent of any particular computer hardware." (From the Forward to Issue 2 of the SVID) To provide this specification, we did not just take the existing System V documentation and reformat it into a book. Rather, we selected the pieces of System V which we felt were necessary to provide a reasonable applications and user environment. When the questions of SIGCLD was raised, the feeling was that it wasn't necessary for a portable application to have the SIGCLD capability, so that particular feature was not made part of the SVID. What makes this issue more confusing is the cryptic warning discouraging the use of SIGCLD which appears in System V documentation. When SIGCLD was originally added to the system, its behavior was different from all other signals in that it wasn't reset when caught. At the time, there was hope that we could migrate the signal into a behavior that was more consistent with the other signals. However, the issues of providing an upwardly compatible interface prevent us from changing the SIGCLD behavior. In the System V, Release 3 documentation, the warning regard- ing SIGCLD usage has been removed. When you're putting together a standard, it can be difficult to make the call as to which features should or shouldn't go into it. If the sense of the net is that SIGCLD should be made part of the SVID, I will be more than happy to provide that feedback to the people who produce the SVID and see what I can do to get SIGCLD into the SVID. Jeff Smits AT&T Information Systems 190 River Rd. Summit, NJ 07901 (201)-522-6263 ...inhp4!attunix!jgs