Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!mnetor!uunet!husc6!mit-eddie!ll-xn!ames!hc!beta!mbr From: mbr@beta.UUCP (Mike Rose) Newsgroups: comp.os.vms Subject: Re: lnm tables Message-ID: <13019@beta.UUCP> Date: Wed, 2-Dec-87 21:49:23 EST Article-I.D.: beta.13019 Posted: Wed Dec 2 21:49:23 1987 Date-Received: Mon, 7-Dec-87 06:21:23 EST References: <37rrk@byuvax.bitnet> Reply-To: mbr@beta.UUCP (Mike Rose) Organization: Los Alamos Natl. Labs, Los Alamos, NM Lines: 25 In article <37rrk@byuvax.bitnet> rrk@byuvax.bitnet writes: >I believe the name of the JOB logical name table is constructed from the >JOB header address, which (I believe) IS available through GETJPI. Just >get it and do a FAO on a control string like "LNM$JOB_!XL". Currently the name of the job logical name table is built using the address of the Job Information Block (JIB), as described above. Unfortunately this is not available from $getjpi. If you really want it, the JIB address is stored in the PCB at offset PCB$L_JIB. I don't think DEC has stated officially that the jib address is, and will continue to be, used to build the name of the job logical name table. I doubt that they have. It seems a reasonable choice to me, except for it not being available from $getjpi. It might be like pid's: they provide a number (or in this case a name) and you make assumptions about how they generate that name at your own risk. This should certainly be a consideration for programs with a long expected lifetime. I missed the original posting on this. Why is the name of the job table needed? Mike Rose