Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84; site ucbvax.ARPA Path: utzoo!watmath!clyde!burl!ulysses!ucbvax!info-vax From: info-vax@ucbvax.ARPA Newsgroups: fa.info-vax Subject: Follow up - Group logical names Message-ID: <5908@ucbvax.ARPA> Date: Sat, 30-Mar-85 23:13:36 EST Article-I.D.: ucbvax.5908 Posted: Sat Mar 30 23:13:36 1985 Date-Received: Tue, 2-Apr-85 00:41:37 EST Sender: daemon@ucbvax.ARPA Organization: University of California at Berkeley Lines: 43 From: Richard Garland A few weeks ago I "complained" about the change in V4.0 in creating group logical names. The jist of it is that the V3.x scheme of changing UIC, and then doing DEFINE/GROUP no longer works. A number of people responded and for the good of all I will summarize. How to do it: V4.1 release notes explain 2 ways: use the /USERNAME qualifier on the SUBMIT command, or start a detached process running LOGINOUT with the UIC of the target group. See release notes for detailed example. Why: In V4.0, group logical names are in a "Group table" which in turn is listed in the system directory of logical name tables (and is in system space of course). The name of the table is LNM$GROUP_xxxxxx. The problem is not putting names into the table, but creating it in the first place. It must be created in KERNEL mode and that's the rub. The DCL CREATE/TABLE has no qualifier for KERNEL. The way a group table is created by the system is by LOGINOUT when it starts a process and notices no table for the group starting the process exists yet. Hence the 2 solutions (I might call them "workarounds") above both involve starting processes (running LOGINOUT) for the target group. Untill the table exists, all attempts to do the job with DEFINE/TABLE=... or similar commands will either put things in the wrong table, or create a table at a less privileged access mode than KERNEL and will fail. My apologies: Many people pointed out that this topic was mentioned at DECUS or brought up in V4.0 seminars (or discussed at length during field test) and I was not justified in calling it a "gotcha". Unfort- unately since I *only* read the published V4.0 documentation, I was not aware of it so for me it was a "gotcha". Suggestion: DEC should add the "/KERNEL" qualifier to the CREATE/TABLE command. Rg -------