Hi,

any good and easy to use Version Control System in D3 ?? or how can it
be done via non-MV native product? (such as CVS)

thx

skypirate wrote:

Sure. It's pretty easy. You can use any source control system if you
keep your D3 source items as text files in a regular NT folder.

Just create an MD q-pointer entry to point to the NT folder on disk.

BP
001 Q
002
003 DOS:C:/MyData/BasicProgs

Note the forward slashes. You can also use NT instead of DOS as the
designation. I forget the subtleties between the two, but it seems to
make no difference for text-based source code files.

You can execute any D3 command against this file folder, just as if it
was a regular D3 file (ed, list, sort, etc).

Nice thing about this method is that you can then use a text editor to
modify your programs as they appear as single text items in the NT
folder.

--
Kevin Powick

Check out books or websites on CVS. It has a command-line interface
which you can access from D3 with !execute commands. To check-in,
write data to the host OS then !cvs as required. To check-out, !cvs
and read the item back. Lots more to it though... This can all be
managed with triggers or by putting wrappers around editors like WED,
ED, UP, etc, so that the operations are done more transparently behind
the scenes.

Call Susan Joslyn for PRC, a commercial source control environment.

I don't think there are any freebies for this, but I'm sure the above
mentioned interface to CVS can be written for anyone who wants to
commission it.

T

"skypirate" <singdad (AT) singdad (DOT) no-ip.com> wrote:

Hello.

CVS is fairly straightforward to implement on D3 and UniVision (two
products that I work with).

As Kevin suggests you need to remove the DATA section of your source
file and replace it with a super-q pointer to a directory within the
underlying operating system. This is your starting point and provides
the additional benefit of enabling you to use a better editor on your
source. On Linux you can access the source on your server directory
using NFS (Linux/Unix access) or SAMBA (Windows access). I use this
method at work.

The difficult part is getting aquanted with CVS. You might want to take
a look at Subversion. This is a second generation CVS tool with better
features and easier interface than 'CVS'.

http://subversion.tigris.org/

You might also like to consider Gruntspud. This is a GUI CVS client.
Written in Java it runs on both Linux and Windows.

http://gruntspud.sourceforge.net/

Hope this helps.


John Bend



skypirate wrote:

I have been using the single-user version of RCS from Component Software
with D3/Windows and have had good success.

http://www.componentsoftware.com/Pro...cMan/index.htm

I have a q-pointer to a Windows directory that is a mirror of the BP (as
well as corresponding directories for DICTS, PL, etc). Instead of tying RCS
directly to the BP files, I copy them over to the Windows directory, then I
check them files into RCS. One additional step, but not a big one. (I did
this mostly because I don't know how to make a super-q for only the data
section of a file...)

I really like the RCS, as it eliminates all the ".OLD, .OLD1, OLD2, .OLDER,
..OLDEST etc versions of programs that at times tend to end up in a BP. One
can compare any two revisions and all the changes are highlighted. And all
the revisions all have timestamps.

The only caveat is that file names with "*" in them cause the Windows side
of things to get a headache.


"skypirate" <singdad (AT) singdad (DOT) no-ip.com> wrote