Hoff,

Thanks for your comments; they give me some hope my conclusions were 
correct.  Some very minor improvements in the docs would have saved me a lot 
of guessing time -- see below.

At 09:04 AM 3/21/00 -0500, [EMAIL PROTECTED] wrote:
[snip]

>   The creation of logical names containing "dollar" signs is reserved
>   to registered facilities.  (BTW, I am registering the "PERL" prefix,
>   have reworked configure.com and am creating a PCSI Perl kit, but that 
>   is another discussion entirely.)

%PERL-I-WEAREHAPPY, the VMSPERL list officially rejoices.

>   I was asked where Compaq OpenVMS documents this -- while the dollar 
>   prefix requirements are documented, it is impossible to document all 
>   of the different potential failure modes.  We do document what is 
>   supported, and I would be very surprised to find the redefinition of 
>   the DISK$mumble logical name documented.

Except that's not the problem exactly.  The problem is when you say

$ mount/sys device foo bar

and then redefine the logical name BAR, you get trouble.  BAR is not 
"reserved" in the sense that there are no dollar signs, it's just reserved 
in that it's permanently attached to the volume in a way that is not clearly 
documented anywhere that I could find.  Yes, you would get the same trouble 
by redefining DISK$FOO, but if you thought you were being good by redefining 
BAR and leaving DISK$FOO alone, you aren't being good enough. 

There are three very tiny doc changes that would be a huge help here:

1). The PCSI docs tell you that if you change a volume label you need to 
update your installations on that volume.  They say nothing about logical 
volume names.

2). The lib$fid_to_name docs say nothing about where the device part of the 
name comes from on output (they do on input).  How about "The device part of 
the returned file specification is the DVI$_LOGVOLNAM field created by the 
MOUNT command and returned by the $GETDVI system service."  If the docs are 
intentionally noncomittal, then they should say so: "You may get the lock 
name, you may get the physical device name, or you may get the logical 
volume name; we aren't going to tell you and you shouldn't depend on it."

3).  The docs on the MOUNT command make no claims about what the logical 
name is used for or even state that it is permanently attached to the 
volume.  How about simply, "The logical name passed as the third parameter 
to the MOUNT command (or DISK$volume-label if none specified) is permanently 
associated with the volume and must not be redefined."

>   As Dan succinctly commented: "don't do that."

Be sure to assign credit (and royalties) properly if you use this in the docs :-).


_______________________________________________
Craig A. Berry                                   
mailto:[EMAIL PROTECTED]

Reply via email to