Carl Friedberg wrote:
> Hi Peter
>
> I have a feeling that USER is one of those logical names that get's us in
> trouble.
Well there is nothing magic about the choice of the name other than it
coincides with a volume label.
> We've had a system level concealed logical name of USER since day one, as
> the system derives from (as in it's the same system management environment)
> MicroVMS 4.x. USER is defined that way on MicroVMS, or at least that was one
> of the default definitions that came with the box. I think I've whined about
> this once in the past; perl on VMS should be aware of this as a possibly
> "standard" vms logical name.
>
> Oh, I think I hit a problem in one of Dan's extensions, but that's not
> relevant here.
>
> So, you are saying that because you do a
>
> $ mount /system $1$DKB100: USER
>
> you get the systme logical disk$user (translates to _$1$DKB100:), and that
> you have the concealed logical
>
> $ DEFINE /SYSTEM /TRANSLATION_ATTRIBUTES=(CONCEAL,TERMINAL) USER
> DKB100:[u.]
>
> somehow lib$fid_to_name is confusing disk$user with the concealed logical
> user? That's a VMS bug...
No not necessarily a VMS bug. Rather it is a bug in vms perl's use of the
device name returned by the name returned by lib$fid_to_name, instead of
the device name returned by stat. We ought to work around it.
Peter Prymmer
