Craig, perhaps "misconfigured" is too strong a word, but it is an issue
which should be addressed. I believe this happened due to a Backup screwup
(Andy? screwup?), where the vms$common.dir;1 directory name got swapped out,
and syscommon.dir;1 got put into the IDENT area of the file header in its
place. This is wrong; the correct file name is vms$common.dir;1. The DSNLINK
site has instrucitons on how to correct this. So, I stand by my position: if
you look at your header with dump, as I suggested, and you find the IDENT
area showing syscom.dir;1, you've got the virus, and it should be fixed (not
a big deal, but fixing it is very scary, as I recall; I've only done it a
couple of times).[1]
As to my examples, the dka1's were typed in by hand; the dka100 stuff was
the real disk I was screwing around with on my test AlphaServer 800 (I
started writing, then I thought, gee I should see if my opinions really
work, experimental physicist that I once was). I was very surprised when the
f$getdvi returned the null string, but, that's what happened (again, it's
vms 7.2-1).
I'll pick up your program and try it out, but not tonight; I've done enough
yakking for tonight and I've got my home network to attend to yet.
Best, Carl
[1] if you want more details on this, I'll be glad to take it offline. If
you've ever built a common rooted directory system from scratch (I've done
that) you can appreciate the fact that syscommon.dir and the top level
sysexe.dir are alias created with the SET FILE/ENTER command. The FID is
invaluable in sorting this all out. Yhe first 9 FID's are magic, and are
assigned during INITIALIZE. If you are interested in this kind of nonsense
(I'm sort of an RMS and file system wizard, or was once upon a time), you
might want to read Kirby McCoy's "VMS File System Internals", isbn
1-55558-056-4, Digital Press, 1990, if it is still available. However, the
specifics on the common and rooted directory trees on the system disk can be
determined experimentally using dir/file and show log/full sys$sysroot :-)
-----Original Message-----
From: Craig A. Berry [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, March 21, 2000 9:57 PM
To: Carl Friedberg
Cc: 'Dan Sugalski'; Craig A. Berry; [EMAIL PROTECTED]
Subject: RE: Perl_cando(), rooted logicals, and volume labels
I don't think the system disk is misconfigured, is it? It's just that
SYSCOMMON.DIR pretends it's a top-level directory in order to make all the
cluster stuff work.
...
I'd agree except that in Perl_cando() we already know the physical device
name and the fid; if we can't get at the file armed with those two things,
then we really are in trouble.
>< my comment: take the device name from the input side; won't that avoid
the
><problem?
I'd be really curious to see what lib$fid_to_name returns; do you have my
fidvoltest program?
