Date: Sun, 22 Mar 1998 15:58:20 +0300 (IDT) From: Eli Zaretskii To: "Matthew H. Gerlach" cc: djgpp AT delorie DOT com Subject: Re: RCS woes In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Precedence: bulk On Fri, 20 Mar 1998, Matthew H. Gerlach wrote: > In the case of DJGPP, removing the target file name is attempted > first. Where do you see this? Is this how RCS works when compiled with DJGPP, or do you see this in a library function? From what I see on the library sources, the function `rename' first tries to rename (i.e. move) the file, and only if that fails, it will try to remove the target and attempt renaming again. Will renaming the file without removing the target first work in your case? > In remove.c, the file to be removed is made writable before removing it. > This fails as does the subsequent attempt to remove it. Can you explain why does this fail? Is this because the target file already exist, but is unwritable because another Unix user owns it? If so, you have several ways to deal with this: 1) Arrange, from the Unix side, for the RCS file to be writable, by either running some Unix command/script on it, or using one of the RCS options when checking the file in. Another possibility would be to change the group ownership/membership so that the file is writable by the other user(s) from the same group. 2) Set up your NFS software to make that file writable, or use one of the programs supplied with the NFS client to make the file writable from the DJGPP side. > I have had great luck with the DJGPP tools in general. As a longtime > PC hater, the DJGPP tools have made DOS and Win95 usable. For this > reason, I have a sad heart thinking I must abandon my efforts to get > RCS working, unless someone has a fresh idea. The main question is: how can that file be made writable from the PC side. If the answer is ``there's no way'', you cannot do anything. But if there is some way, you can solve the problem, provided that you describe the way here.