delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp/1997/11/03/13:03:11

From: "Andrew Crabtree" <andrewc AT rosemail DOT rose DOT hp DOT com>
To: "Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il>
Cc: <djgpp AT delorie DOT com>
Subject: Re: Incompatibilities with NT
Date: Mon, 3 Nov 1997 10:02:53 -0800
Message-Id: <01bce882$b4ee2630$45111d0f@ros51675cra.rose.hp.com>
Mime-Version: 1.0

>Please run `df' from Fileutils and tell what does it print.

M:\>df --version
df (GNU fileutils) 3.16
M:\>df
Filesystem         1024-blocks  Used Available Capacity Mounted on
C                     999968  506144   493824     51%   c:/
\\HPRNLRD.ROSE.HP     999968  858688   141280     86%   e:/
\\BLUEBEAM\P1675CRA   999936       0   999936      0%   i:/
CCase                 999984  487984   512000     49%   m:/
CCase                 999984  487984   512000     49%   s:/
\\TYPHOON.ROSE.HP     999968  791648   208320     79%   t:/
CCase                 999984  487984   512000     49%   u:/
Drive V:              999968  676640   323328     68%   v:/
\\RYDER\WARP          999968  801728   198240     80%   w:/
Drive X:              999968  821728   178240     82%   x:/
Drive Y:              999968  673920   326048     67%   y:/
Drive Z:              999968  363744   636224     36%   z:/
M:\>

Some of the drives (x,y,z,v) are mapped via Intergraph Disk Access NFS.
They
are for clearcase too (each vob and view needs a separate mount point).  The
M drive is
where the multi-version file system resides, and is the only one I am
interested in working.

>I have no idea
>how this setting is done on NT, or even *if* it exists.  A call to
>`setdisk' will return the last possible drive.
>Don't bother.  It seems that NT doesn't know about TRUENAME.
It appears NT doesn't have a setdisk either.  Perhaps I can make a small
prog
that uses the sources from dj's libc for setdisk and truename?

>But you should be able to say "net use" and see where M: is mapped.
>I'd expect `df' to print the same mapping for M:, unless something is
>VERY wrong.
Doesn't look promising.


Status       Local     Remote                    Network
----------------------------------------------------------------------------
---
OK           E:        \\hprnlrd.rose.hp.com\public
                                                 Microsoft Windows Network
OK           I:        \\bluebeam\p1675cra       Microsoft Windows Network
             M:        \\view\                   Atria Multi-Version FS
             S:        \\view\andrewc_top        Atria Multi-Version FS
OK           T:        \\typhoon.rose.hp.com\andy_top
                                                 Microsoft Windows Network
             U:        \\view\andrewc_top        Atria Multi-Version FS
OK           W:        \\ryder\warp              Microsoft Windows Network
             V:        \\reebok\mnt\extfw2\sw-objects
                                                 NFS Network
             X:        \\reebok\mnt\extfw2\warp-extra14
                                                 NFS Network
             Y:        \\reebok\mnt\extfw2\warp-extra15
                                                 NFS Network
             Z:        \\aspen\warp-extra16      NFS Network
The command completed successfully.


>One possibility that I can think of is that Clearcase uses the NT
>SUBST command to create that drive.  It might be that SUBST on NT
>works differently from the DOS SUBST, and therefore that drive is
>invisible to DJGPP.  There are 3 functions of Interrupt 21h, AX=71AAh,
>described in the Interrupt List, that seem to support a new mechanism
>of SUBST under Windows 9X; maybe NT works this way also?  One of these
>functions allows to query what's the pathname associated with a given
>SUBST'ed drive.  I attach the info from the Interrupt List below; it
>would be interesting to see whether this function tells anything
>useful about your M: drive.
I'll play around with nasm and see if I can remember how to program real
mode assembly.


>Relax, I know what Clearcase is, we use it extensively on my job
>place.  I just don't see why do they need a drive.  Can't they just
>install it in a directory with a reserved name?

I'm really not happy with Clearcase environment on NT as compared to unix.
It is, however, roughly 10x faster on my PPro than on my old HP 715.
All access to files in done though view extended names basically.
On unix everyone would access our global makefile via /sw/code/Makefile
(/sw/code
being the vob mount point).

On NT it is m: (mvfs drive, can be changed though), view name
(andrewc_view), vob name
(code, only one level is supported, so /sw/code is illegal), then makefile
m:\andrew_view\code\Makefile.  Obviously, this changes based on the view,
which
makes writing makefiles a disaster.



>If you can ask them exactly how they create that drive, it might make
>the whole thing of figuring out this puzzle a lot more easier.
I will send a message to their technical support and see what it turns up.
They
haven't been overly helpful in the past.

Andrew

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019