Date: Tue, 19 May 1998 10:16:50 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: "Mark H. Wood" <mwood AT indyvax DOT iupui DOT edu>
cc: djgpp AT delorie DOT com
Subject: Re: cvs init gives ENODEV on Netware volume
In-Reply-To: <LJ3JkfdQj$WY@indyvax.iupui.edu>
Message-ID: <Pine.SUN.3.91.980519101634.21112C-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On 18 May 1998, Mark H. Wood wrote:

> 	\\DONOHUE\USERS\ADMIN\CVS
>
> This was done while sitting in K:\ADMIN\src\Launch32, mapped as follows:
> 
> K:\ADMIN\src\Launch32>map k:
> 
> Drive K: = DONOHUE_USERS: \ADMIN\SRC\LAUNCH32

Sorry, I don't understand this.  If k: is mapped to
\ADMIN\SRC\LAUNCH32, then K:\ADMIN\src\Launch32 should be
\ADMIN\SRC\LAUNCH32\ADMIN\src\Launch32, no?  But the TRUENAME results
says something else.  What am I missing here?

> Looking afterward, I see that a loginfo file was created and contains data.

Does that mean that ci actually succeeded, but somehow printed a bogus
message and exited with an error status?

> "+LFN=y" is already present in the global section of that file.  The Netware
> volume has LONG namespace installed, and I can use long filenames without any
> trouble when using DOS commands.

How about DJGPP programs--do they support long file names on drive K:?
Can you use e.g. `ls' from Fileutils, or compile a simple DJGPP
program and tell if they see long names?

> ci: loginfo,v: No such device (ENODEV)
> cvs.exe [init aborted]: failed to check in K:/Admin/cvs/CVSROOT/loginfo

If none of the above helps in any way, I will need to know where does
ci fail.  I'm afraid the only way to find that out would be to
download the RCS sources, compile them with -g, then run them under a
debugger to see where does that ENODEV error come from.  If you find
that out, please post the info here, and I will try to understand what
goes wrong, and how to solve it.