Re: SYSUAF logical name

On 11/10/2010 15:12, Richard Maher wrote:
[...snip...]
On a related note, note that a supervisor-mode logical name will
overwrite an executive-mode logical name. From an account with full
privs:

$ sh log/ful/all xxx
%SHOW-S-NOTRAN, no translation for logical name XXX
$ define/system/exec xxx yyy
$ sh log/ful/all xxx
"XXX" [exec] = "YYY" (LNM$SYSTEM_TABLE)
$ define/system xxx aaa
$ sh log/ful/all xxx
"XXX" [super] = "AAA" (LNM$SYSTEM_TABLE)

Bizarre! And not even an ss$_supersede. I look forward to reading that
description of this "supported" behaviour.

Then if you do a DEFINE/SYS/EXEC XXX ZZZ you get ss$_supersede but the
logical stays at supervisor mode. Intuitive?

I suspect this is one of those murky areas that nobody dares to
change for fear of breaking something that relies on the observed
behaviour.

Do, however, note that the /Name_Attributes=No_Alias is designed
to prevent re-defining a logical name in a lesser mode.

$ define/sys/exec/name=no_alias xxxx yyyy
$ define/sys xxxx yyyy
%SYSTEM-F-DUPLNAM, duplicate name
$
.

Relevant Pages

  • Re: SYSUAF logical name
    ... SYSNAM privilege, the DEFINE command ignores the qualifier and <--| ... $ sh log/ful/all xxx ... $ define/system/exec xxx yyy ... with cluster-wide logicals it is easier to check first while ...
    (comp.os.vms)
  • Re: SYSUAF logical name
    ... $ sh log/ful/all xxx ... $ define/system/exec xxx yyy ... without sufficient privs redefines it, ... executive-mode definition is gone. ...
    (comp.os.vms)